Non Invasive Cardiac system (NICaS)

Whole Body Electrical Bio Impedance

Bio Impedance Technology

Study Review

January 2014
 Table of Content

No. Article Subject Page

1 Cardiac Output Monitoring - Reviewing the Evidence on Four NICaS 4
Systems, Health DEVICES Vol. 42 No. 6 June 2013 Validation
2 Yu Taniguchi et al. Noninvasive and Simple Assessmentof NICaS 15
Caridac Output and Pulmonary Vascular Resistance With Validation
Whole-Body Impedance Cardiography Is Useful for Monitoring
Patients With Pulmonary Hypertension, Circulation Journal, June
12,2012
3 SG Hall, J Garcia, DF Larson and R Smith. Cardiac power index: Cardiac Power 22
staging heart failure for mechanical circulatory support, Index Validation
Perfusion 2012 27:456-459
4 JGF Cleland et al. Non invasive measurements of Cardiac Output NICaS 29
9CO) and Cardiac Power Index (CPI) by whole body bio Validation
impedance in patients with heart failure. A report from SICA-HF
study 9FP7/2007-2013/241558). European Journal of Heart
Failure Supplements (2012) 11
5 Yoseph Rozenman et al, Detection of left ventricular systolic GGI Validation 30
dysfunction using a newly developed, laptop based, impedance
cardiography index, International Journal of Cardiology Volume
149, Issue 2: 248-250, 2 June 2011
6 Robert G. Turcott et al,Measurement Presision in the ICD for CRT 33
Optimization of Cardiac Resynchronization Therapy, Circulation Optimization
Heart Failure 2010;3:395-404, February 22, 2010 Validation
7 Yusuke Tanino et al. Whole Body Bioimpedance Monitoring for NICaS 44
Outpatient Chronic Heart Failure Follow up, circulation Journal Validation
73:1074-1079, June 2009
8 Konstantin M. Heinroth et al, Impedance Cardiography: a useful ICD for CRT 50
and reliable tool in optimization of cardiac resynchronization Optimization
devices. Europace (2007) 9, 744-750, 11 May 2007 Validation
9 Dorinna D. Mendoza et al. Cardiac Power Output predicts Cardiac Power 57
mortality across a broad spectrum of patients with acute Index Validation
cardiac disease. American heart Journal 153:366-370, March
2007
10 Oscar Luis Paredes et al. Impedance Cardiography for Cardiac NICaS 62
Output Estimation - Riliable of Wrist-to-Ankle Electrical Validation
Configuration -. Circulation Journal 2007; 70: 1164-1168,
September 2006
11 G Cotter, A Schachner, L Sasson, H Dekel and Y Moshkovitz. NICaS 67
Impedance cardiography revisited. Physiological Measurement Validation
27 (2006) 817-827, July 2006
12 Ronald D. Smith et al. Value of Noninvasive Hemodynamics to ICD for 79
Achieve Blood Pressure Control in Hypertension Subjects. Hypertension
Hypertension 2006;47:771-777, Mar 6, 2006 Validation

2
 Table of Content, Contd

No. Article Classification Page

13 Marina Leitman et al. Non-invasive measurement of cardiac NICaS 86
output by Whole-body bio-impedance during dobutamin stress Validation
echocardiography: Clinical implementations in patients with left
ventricular dysfunction and ischemia. The European Journal of
Heart Failure 8 (2006) 136-140
14 W. Frank Peacock et al. Impact of Impedance Cardiography on ICD Validation 91
Diagnosis and Therapy of Emergent Dyspnea: The ED-IMPACT for ER
Trial. Academic Emergency Medicine Vol. 13, No. 4:365-371, April
2006
15 Hector O. Ventura, Sandra J. Taler and John E. Strobeck. ICD Validation 98
Hypertension as a Hemodynamic Disease: The Role of for
Impedance Cardiography in Diagnostic, Prognostic and Hypertension
Therapeutic Decision Making. American Journal of Hypertension
Vol. 18, No. 2, Part 2:26S-43S, February 2005
16 Guillermo Torre-Amiot et al. Whole-Body Electrical Bio- NICaS 116
Impedance is accurate in Noninvasive Determination of Cardiac Validation
Output: A Thermodilution controlled, Prospective, Double
Blinded Evaluation. European Journal of Heart Failure, June 2004
17 Gad Cotter, Yaron Moshkovitz, Edo Kaluski, Amram J. Cohen, NICaS 117
Hilton Miller, Daniel Goor and Zvi Vered. Accurate, Noninvasive Validation
Continuous Monitoring of Cardiac Output by Whole-Body
Electrical bioimpedance. CHEST 2004;125;1431-1440
18 Gad Cotter et al. The role of cardiac power and systemic vascular Cardiac Power 129
resistancein the pathophysiology and diagnosis of patients with & Peripheral
acute congestive heart failure. The European journal of Heart Resistance
Failure 5 (2003) 443-451 Validation
19 John E. Strobeck et al. Beyond the Four Quadrants: The Critical ICD Validation 138
and Emergong Role of Impedance Cardiography in Heart Failure. for CHF
Editorial, April 2004 Supplement 2
20 Amram J. Cohen. Non-invasive measurement of cardiac output NICaS 144
during coronary artery bypass grafting, European Journal of Validation
Cardio-thoracic Surgery 14 (1998):64-69
21 C. Gresham Bayne. Evaluation and management of the elderly ICD Validation 150
homebound patient with congestive heart failure for CHF

3
4 of 158.
) t'k*ffiam

GC WITH THE FLOW

THERMODILUTION USING A PULMONARY ARTERY CATHETER IS CONSIDERED THE GOLD STANDARD
IN MEASURING CARDIAC OUTPUT. HOWEVER. DRAWBACKS ASSOCIATED WITH THE TECHNIQUE HAVE
PROMPTED CLINICIANS TO SEEK LESS INVASIVE OPTIONS. BUT THESE ALTERNATIVES HAVE THEIR OWN
LIMITATIONS. WE EXAMINE THE EVIDENCE ON FOUR DEVICES THAT MEASURE CARDIAC OUTPUT USING
AND NON INVASIVE TECH N IQU ES.
MI N IMALLY INVASIVE

\Ieasuring a patient's cardiac output-the amount
of blood pumped bv the heart per minute-can aid
in diagnosing diseases of the cardiovascular system
and in managing high-risk patients (e.g., those suf-
fering from burns, sepsis, or shock) and patients
undergoing maior surgicai procedures. The cardiac
output value can be derived lrom heart rate and
stroke volume (the amount of blood pumped br-
the heart per beat) using the relationship
cordioc oulpul = stroke volume x heort roie
The uaditional approach to monitoring car-
diac output is pulmonarv arterv catheterization
using thermodilution as the measurement method.
.Florvever, the draw-backs associated with this tech-
nique-including the risk of pulmonarr arterv
ruprure aod air emboli5m-h21.g led clinicians ro
seek less invasive alternatives.
In our December 2009 issue, rve looked at some
rninimallf invasive and noninvasive alternatives to
pulmonarl' arterv catheterization for cardiac output
monitoring and revierved the evidence regarding
dreir effectiveoe ss. This article setres as ao update to
that piece. We profiie four additional cardiac output
monitoring svsterns and rer.'ierv the evidence
regarding their accuracv compared to other cardiac
output monitodng techniques, particuiadr ther-
modilution. \Y'e also examine evidence regarding
improved clinical outcomes with use of these
derrices.
\\:e look at the following s,rstems:
) ConMed Ecollt. The literarure s'e revierved did
not show the ECOII to be ao acceptable alter-
native to the reference techniques; horvercr,
nvo studies found that the monitor rvas useful
in predicting fluid responsir.eness in abdominal
and cardiac surger\r patients. \I'hen the monitor
is used in the ()R electrosurgical equip-
"vith
ment, interruptions in monitoring could occur.
\\:e did flot 6od any sfudies addressing clinicai
outcomes.
F Edwords Llfesciences FloTroc sensor. Previousl,v
reviewed literarure showed the FloTrac sensor to
be comparable to thermodilution in stable car-
diac surgerv patients. Srudies on the latest gen-
eration of the FIoTrac sensor, horvever, [ound
that the device may be less reliable than thermo-
dilution in hemodynamicallv unstable patients
such as those undergoing abdominal surgen'

.l78 xmrraDEvlcEs JuNE2ot3 ) w.*ri.ors 02013 ECRI lnsiitute. Member hospitols moy reproduce this poge for internol distribution only.

5 of 158.
 Reuiewing the
Euideruce onFour
Cardiac Oatput
Monitoring Slstems
or iiver transplantation, those on vasopres-
sor therapt', or those with sepsis. Tivo srudies
reported improved clinical outcomes (reduced
complications and hospital length of star) rvhen
the FloTrac sensor was used to guide fluid man
agement in abdominal surgerv patients.
) Edwords Llfesciences VolumeView sensor. The
YolumeYierv sensor is relativelv nerv; therefore,
the evidence on it is limited. \Iore indepen-
dent srudies on raried patient populations are
required to determine its ef6cacr'. \\"e did not
6nd anr studies ad&essing clinical outcomes.
) Nl Medicql NlCoS. SLx srudies examined in our
Literarure review found that the NICaS is an
acceptable aiternative to the reference tech-
niques for cardiac output measurements in car-
diac patients and heart lailure patieots. Tvuo of
those studies found it to be more accurate than
thoracic impedance cardiographr. \V'e did not
6nd anr- srudies addressing cl-inical outcomes.
()fthese s)'stems, only \I \Iedicalt \ICaS is
noninvasive; the remaining three svstems are coo-
sidered minimallr- invasive.
O2O'13 ECRI lnstitute. Member hospitols moy reproduce this poge for internol distribution only.
6 of 158.
r:'i lli ii.ji{} i.iri\:iY n:,i.'r ii:r"l' ::'i'r,iiiiVlf;}1}i i,tii'ii.}:Li
Tiaditionally, cardiac output is measured using a
technique knorvn as pulmonar,v arterl.- thermodilu-
rion, *,hich incorporates the use of a pulmonarv
arteq'catheter, or PlC. PuimonarY atterv thermo-
dilution is considered the gold standard in cardiac
output monitoriag. There are two basic versions oF
this technique: bolus and continuous. In bolus, or
intermittent, thermodilution, a bolus of an iniectate
at a knorvn temperature is infused into the blood-
stream, and the change in temperarure caused bv
the iniectate is measured at some dorvnstream loca-
tion using temperature sensors incorporated into a
PlC. ()ne dras'back to bolus thermodilution is that
it does not provide a cootiouous measurement of
UMDNS ierm- Monilors, Bedside, Cordioc Output
I2O-1741
w.cri.org > HEALTHDEVICES JUNE2ots 179
 guidance
/
cardrac output and therefore cannot detect
rapid and sudden changes in the hemodr--
namic starus oi high-risk pauents.
Continuous rhermodilution inroh'es
measuring the change in the temperarure
oi blood at a dorvnstream location after
introducing a heating Elament upsream.
Ti-picallr', the heating filament is placed in
the right tentricle and pou'ered on intermit-
tenth', and blood tcmperarure is measured in
the pulmonarr arten-. The displaled cardiac
output is an average of measurements taken
ovet three to slr minutes.
The use of a P,\C adds a degree oi risk
to rhe technique: P-\Cs are both inrzsive
and associated with a number ol compli-
cations, most of s,hich are related to the
insertion of the catheter; these .include
pulmonarr- arterl rupfure, air embolism,
and arrhr-thmias. --\dditionalh', evidence
regarding the efficacv o[ P-\Cs in improv-
ing clinical outcomes is r'.'eak; therefore,
this technique is being used on a small
and declining percentage of critical care
padenrs (Schrvann et al. 201 1, \\'heeler
et aL.2006, Wiener et al. 2001).
The &arvbacks associated s'ith pulmonan-
arterr thermodilution have spurred the
development o[ less invasive alternatives
tbr monitoring cardiac outpttt. These
alternatir-e technologies, although less
invasi"'e than pulmonarv arterv thermodi-
lution, have their orvn dtaw-backs, and the
evidence regarding their usefulness can
be limited. 111 currentlv matketed cardiac
output monitoring svstems fall into one or
more oF rhe follorving categories.
Aderiql woveform onolysis. This minimallv
invasive technique uses an arterial pres-
sure lioe and is based on the principle that
changes in arterial blood volume during a
cardiac clcle are accompanied br changes
in pulse pressure. Stroke volume can be
determined bv conrcrting the change in
pulse pressure obtained from the arterial
pressure rvaveform to a corresPoflding
change io blood volume. Pulsc pressure
changes that cortespond to a change in
18O xrllrx DEvtcEg JUNE 2ot3 ) w.xrl.ors
7 of 158.
LIMITATIONS OF THE LITERATURE
There ore limitotions to the published sludies we revievred, ond reoders should be core{ul
vrhen drowing conclusions from our findings. One limitotion is intrinsic to fhe meosuremenl of
cordioc oulput: The occurocy o{ pulmonory oriery thermodilulion, the gold stondord method,
is highly dependent on user technique. As o result, thermodilulion's resulis ore not olrvoys
relioble. This in lurn offects the ossessmenl of ollernoiive lechniques such os those used by lhe
products we review in this orticle: Becouse of ihe possible inoccurocies in ihe PAC meosure-
menis thoiihe olternoii're techniques ore compored ogoinst, on olternotive is generolly con-
sidered to be occurote even i{ there exists up fo o 3096 discreponcy (percentoge error) rvhen
compored io the thermodllution re{erence meihod. This helps exploin the sometimes limited
ond inconclusive noture of evidence regording the occurocy o{ lhe producis in this orticle.
Thus, our review of the evidence olso took inio considerolion studies comporing the producis
io other cordioc output moniioring techniques {e.g., vrhether ihe resulls from the two products
correlote, or show ihe some irend).
Another limitotion o{ the siudies is thot, while they look oi the occurocy ond correlolion
o{ the cordioc output volue, ihey don't iypicolly provide ony insighi into lhe use o{ ihese
oliernoiive iechniques in clinicol proctice {e.g., how often they ore used, ihe iype of theropies
they guide, their clinicol outcomes). Thus, we recommenC thoi hospitois opply coution rvhen
considering these devices for use.
blood volume are dependent on atterial [described under Electrical, belorv]) and
characteristics such as aorric compliance the Edrvards Lifesciences FloTrac and
and peripheral resistance. \'olumeYiet'.
Esopho geo I Dopp le n f his rninirnalir inr-asive
Sr-stems using arterial rvar.elorm analr-
approach uses Doppler ultrasound tech-
sis often need to compensate for chaoges
niques to measure blood florv rn the aorta,
in individual arreialf aorac characteristics
that might occr-rr during the measurement and can be used in the absence of anr'
period, due either to phrsrologic changes peripl-reral or central line. In this technique,
or to changes associated rvith trcatment. an ultrasonic rvaye is directed torvard the
\Iost devices compensate for these aorta. \Ieasurements are perlormed using
changes using an independent measure ultrasouod probes that are inserted into
of stroke volume (e.g., thermodilution). dre esophagus and aimed at the aorta. The
Horvever, some devices don't use anv fte<luencr- of the rvave reflected back b,v the
external calibration and instead use pro- moving blood cells (in t1-re aorta) is difierent
prietan- algorithms that ftequenth- update ilom the origrnallr uansmined ultrasound
speci6c calculation parameters based on &equencl-; this shift in irequenci- can be
the patient's demographic and phvsiologic used to obtain the velocin'of the blood
in[ormation. in the aorta. That velocin can be used to
calculate the distance traveled br- the blood
Devices that use arterial rvar.eform
in the aorta during sr stole, knos'n as the
anah sis might not be accurate in parienrs
rvith certain arrhr-thrnias or arterial tree stroke distance. The stroke distance is then
used to estimate stroke volume using the
pathologies (conditions that afFect the
cross-secrional area of the aorta, lvhich is
shape of the arterial pressure s'aveform)
estimated usiog the panent! age, height, and
or in the presence of circulaton'assist
rveight. This technique is rypicall,v used on
devrces such as intra-aortic balloon pumps.
inrubated patients.
Three of the s1'stems rve rer.ierved use
this technique: the Conlled EC()\I (rvluch
Electricol.In this noninrzsive techoique,
cardiac output is deriwed br applving an
uses a combination of afterial wavefofm
electtical signal across certain areas of the
anallsis and impedance catdiographl
bodr- (e.g., chest, ankle) using electrodes
O2Oi3 ECRI lnsiitute. Member hospitols moy reproduce ihis poge for internql distribulion only.
 placed on the patient. Changes in the
applied signal are then measured during
each cardiac cr-cle. ()bsen'ed changes in
impedance and phase shift can be related
to changes in blood volume. Der,'ices using
these techniques are n'picallr susceprible
to motion artilacts that might prove a hin-
drance to continuous monitoring.
()ne o[ the more cornmon electrical
techniques is knorvn as impedance car-
diograph,r' (ICG), u.'hich is based on tl're
theon that changes in blood tolume in
the aorta during each cardiac cvcle tesult
in irnpedance changes that can be used
to estimate stroke volume. The imped-
ance changes can be derived br measuring
voltage changes to ao applied electrical
signal. Traditionalh-, the ICG technique
uses electrodes placed on the Patientt neck
and chesr (knorvn as thoracic ICG). ()ne
drarvback o[ these locat-ions is the suscep- excl.range) or intracardiac shunt Qlood
ubiliq' to noise: In addition to blood flow; Ilorv across the chambers ot the heart),
the electrodes ma,r pick up other signals pulmonan'blood flou'can be equated to
such as airflos, through the lungs. Two svstemic blood llorv or cardiac ourPut'
svstems rc revierv use ICG: the Conlled Devices using this technique mal be inac-
EC()\I (rvhich uses a combi'ration of ICG curate in patients with abnotmallr high
and arterial r'"aveform anaitsis) and the \I intrapulmonarv or iotracardiac shttnts
fledical \ICaS. (rvhich can occur in thc presence oi some
respiratort or cardrac diseases).
Rebreothing. The noninvasive rebreathing
technique is based on the principie that Tronsthoro(ic echocordiogrophy. In trans-
intake or elirnination o[ a gas (as iodicated thoracic echocardiography, or TTE, an
bv the ditference in the concentration of ulrasound probe is placed on the chest
the gas entering and exiring the lungs) is to measure the Yelocifr o[ blood across
proportional to the amount of blood florv the aort-ic or pulmonan= vaive. This, com-
to the lungs that is invol't'ed in the gas bined s,ith the cross-sectional area oF the
exchange. The amount of gas consumed/ valves, allot's calculation of ieft or right
produced is used as an indication of pul- stroke volume and cardiac output. This
monarv blood flos-, and in the absence of noninvasive technique mar be inaccurate
significant intrapulmonan shunt (blood in patients u'ith r,alr.e and florv-pattern
florv to thc lungs not invoh-ed in gas abnormalities.

PRODUCT REVIEWS
Pulmonarv arter't- thermodilution is consid-
ered the goid standard for cardiac ourput
measurement. Therefore, the tables that
accompanv our product reviervs comParc
the evidcnce on each product with the
evidence on thermodilution. But because
comparisons to other methods can also
be use6;1 (see the box on page 180), rhe
discussions in each revierv also take into
consideration srudies comparing the prod-
ucts to otier cardiac output techniques.
supplier. Coolled Corp. [101345], Utica,
\Y (US-\); +1 (800) 4-18-6s06, +1 (31s)
technique and uses this in conjunction
,x.ith the arterial s,zr.elorm analrsis tech-
nique to determine catdiac output.
The monitor consists of an endotracheal
rube with a printed elecrrode arrat that
applies the electrical signal required for ICG.
The elecrode arrav consists of a single
elecuode at the proximal end of the rube
and slr electrodes at the distai eod. \\'hen
the patient is innrbated, the distal eiectrodes
are located adjacent to the ascending aorta,
thus iacilitating aortic blood ilorv measure-
ments. The electrode arrav also detects
the R-rr,ar-e and uses the R-R interval ro
obtain l-reart rate measurements. ,lterial
pressure data is obtained from an esisting
elecuodes are in such ciose proximifi'to the
aorta, the EC()]I is less susceptible to noise
compared to traditional ICG methods.
clinicol uses, -\ccording to the vendor, the
trC()\I is used primarilv in the OR lor
short-term conLinuous morutoring and
is indicated for use in patients rvho are
expected to be inrubated for 24 hours ot
less. .\lthough ICG technologr in general is
susceptible to motion artilacts, this mav not
be an issue with this derice since inrubated
patients in the ()R are also npicallv sedated.
Disposobles, The monitor requites a propri-
etarr- disposable endotracheal rube.

7 97 -831 5 ; g'rvs,. conmed.com
Procedure. f'he EC()]I (rvhich stands for
Endotracheal Cardiac ()utput \Ionitor) is
a minimallf inrasive srstem that measures
cardiac output and other related hemodi--
namic parameters using a combination o[
trvo different techniques: ICG and arteriai
s.atetbrm analvsis. The monitor obtains
an impedance rvaveform rvith the ICG
t It.i
arterial pressure Line and, in addition to car-
diac output, facilitates estimation oi other
hemodvnamic pararneters such as s1'stemic
vascular resistance and stroke volume rzria-
tion (neit1-rer of u'hich can be obtained from
ICG).;\t the beginning oF the procedure,
the clinician must enter the patientt height, :3
rveight, and age, rvhich are used along with
the measured ICG signal to estimate suoke ConMed ECOM. (lmoge courtesy of
volume. -\ccording to the vendor, since the ConMed.)

O2OI3 ECRI lnstitute. Member hospitols moy reproduce lhis poge for inlernol distribution only. w.*ri.org > HEALTHDEVICES JUNEZoTS l8l

8 of 158.
) s"i'ton"

LITERATURE REVIEW: CONMED ECOM COMPARED TO THERMODILUTION

Study Polienlpopulolion Referencemethod Outcomes
Boll et ol. 2010 40 cordioc surgery Bolus pulmonory ortery With o percentoge error of 5O% ond o correlotion coe{ficient o{
potienls lhermodilution r : O.49, the h*o methods were not considered comporoble.
Gennort et ol. I 4 cordioc surgery Thermodilution. With percenioge errors of 419'5 ond 5970 (bosed on orteriol Iine
201 2 (conference potients lccotion), the iwo methods were not considered comporoble.
obskoct)
Mous et ol. 201 I 29 coronory orlery Bolus pulmonory ortery ther- The percentoge error wos 5070 compored to thermodilution ond
byposs groft poiients modilution ond tronsesopho- 48% compored to echocordicgrophy. However, lhe oulhors report
geol echocordiogrophy occeptoble trending of cordioc outpul.
Moller-Sorensen 25 coronory ortery Bolus pulmonory ortery VVith percentoge errors ronging hom 37o/o to 45% bosed on pctient
et ol. 20'12 byposs groft potients thermodilution posilion, the ivro methods v/ere not considered comporoble. The
study olso reported poollrending of cordioc oulpul by the ECOM
monilor.
Von der Kleii et ol 20 coronory ortery Pulmonory orlery The obskoct reports poor correlotion (r
= 0.23) beiween the ECOM
20 1 2 {conference byposs grofi potienis thermodilution" monilor ond lhe reference method in postoperolive potients.
obstroct)
- The obstroct does not speci{y whot lype of thermociilulion wos used.
'^ The obslrqcl does not specifo whelher lhe melhod of ihermodilulion used vros coniinuous or rnlermiilenl

Evidence. \\e revierved eight clinical srud-
ies that compared the EC()\I monitor to
other methods of cardiac output measure-
ment. The srudies involved cardiac surgerr
and abdominal surgerv patients, and the
number of patients in the srudies ranged
irom 12 to'10. In five of the eight snrdies,
the ECOII monitor was compared to pul-
monal\. arter\' thermodilution; these studres
are outlined in the table on this page. In
the remaining three srudies, the reierence
methods rvere transpulmonarr- thermodilu-
tion (rvhich uses a central venous catheter
rather than a P-\C) (F'ellahi er aL.201.2),
esophageal Doppler fl orgensen er al. 2Ol2),
and the FloTrac sensor @airamian et al.
201 1; see our rerierv of this slstem belorv-).
In addition, three ol the eight srudies
(Bairamian et al. 2011, Gennart et aL.2012,
\hn der Kieij et al. 201,2) rvere reported as
conference abstracts onlr; therefore, ther
ma\- not rePresent the 6nal results and con-
clusions oi the srudies.
In all eight srudies, the authors con-
cluded that the device \\,-as not comparable
to the reference method. The evidence
regarding the monitort abilirl'to trend
cardiac output is inconsistent, rvith one
srudv (\Iaus et al. 2011) reportiog good
trending (based on the correlatjon benveen
the EC()f t monitor and the reference
techniques) and rs'o studics (Jorgensen
er a|.2012, \Ioller-Soreosen et al. 2012)
reporting poor trending. livo srudies
(Bairamian et al. 20i 1, Fellahi et al. 2012)
conclude that the techoique mav be useful
for predicting fluid responsiveness (using
stroke volume rariation). Trvo other srud-
ies @all et al. 2010, ]Ioller Sorensen et al.
20i2) tbund that the EC()\I \i,as suscep-
tible to interference from electrocauterr-
equipment used during surgeries, leading
to intermittent loss of signal during elec-
trosurger\' and somedmes several seconds
after electrosurger\- \'v?s discontinued.
\\b d-rd not 6nd anr studies ad&essing
c[oical outcomes.
Supplier. Edrrards Li[esciences Corp.
[374501], Irvine, C-\ (US.\); +1 (800) 124-
, .,(e-+e) 250-2500; rwwv.edrvards.
::f
'.\lthough thc authrlrs usc rhe rcrm "clcctrocrutcn,"
*cloelierc tlrc intcnt wrs "clcctrcsurgcn."
Procedure. The FloTrac sensor uses
arterial rvaveflorm analrsis to estimate
continuous cardiac output using the Flo-
Trac algorithm. -\ccordiog to the vendor,
the algorithm constantlv monitors rhe arte-
rial pressure vaveform and compensates
for changes tn arterial f aottic characteris-
tics. The FloTrac sensor can be used with
either thc E\-1000 or thc \-igilco rnonitor
offered bv Eds'ards Lit'esciences.
The FloTrac sensor can be used onlv
on patients rv-ho har.e an existing arterial
line that can be connected to the sensor.
The sensor has nvo output lines: ()ne is
connected to the arterial pressure cable
of a bedside monitor, and the other is
connected to the E\-10o0/\'igileo moni-
tor. The sensor, once connected, sends
arterial pressure rv'aveform information
to both monitors. The cardiac output
monitor (either the E\'1000 or the \-igileo)
uses this informat.ion in coniunction with
patient demographic information entered
br the user (e.g., age, gender, height,
rveighr) to esrimate cardiac output.
\\'e covered the FloTrac sensor in our
December 2009 Guidance -{.rticle. Since
then, the vendor has updated the device's
algorithm. The hterarure review-ed in the

182 x;arrg DEvtcEs JUNE2or3 ) w.xri.org O2013 ECRI lnstilute. Member hospiiols moy reproduce this poge for inlernol distribution only.

9 of 158.
 LITERATURE REVIEW: EDWARDS LIFESCIENCES FLOTRAC COMPARED TO THERMODILUTION
Sludy Potient populotion method
Reference Oulcomes
Bioncofiore et ol. 2Ol I 2l liver tronsplontolion ortery
Bclus pulmonory With o percentoge error o{ 37o/" ond o correlotion coelficient
potients thermodilution of r = 0.65, lhe ogreement betueen the irvo methods vros noi
considered clinicoliy occeptoble.
De Bocker et ol. 20,l I 58 septic potients pulmo-
Bclus ond continuous \Vith o percentoge error ol 30.4o/o, the FloTroc sensorwos
thermodiluiion
nory ortery considered io be ot leosi os occurote os the reference methods
Morque ei ol. 201 3 18 seotic shock potienls Continuous pulmonory ortery With on overoll percenloge error o{ 640/o, ihe iwo methods
ihermodilution v/ere not considered comooroble.
Slogt el ol. 20 I 3
'l
9 septic shock potienis Bclus pulmonory ortery With on overoll percentoge error ol 53o/o, the two methods
thermodilution v/ere not considered comporoble. The oulhors conclude thot
the trending obility of the FloTroc sensor is foir.
Su et ol. 2012 28 liver tronsplontotion Conlinuous pulmonory ortery Wilh o percenloge error ol 75o/o,the irvo melhods were nol
potients ther'modllution considered comporoble. The outnors olso report poor trending
of cordioc outpuf by ihe FloTroc sensor.
Tsoi et ol. 20 I 2 20 liver tronsplontotion Bolus pulmonory ortery Wilh o percentoge error of 54.93o/o, ihe outhors report poor
potients thermodilution correiotion be,'ween the furo methods.
Vosdev et o1.2012 38 coronory ortery byposs Eclus pulmonory oriery With percentoge errors of 2Ao/o ond 2296 depending on the
groft poiients thermodiluiion monitoring site (rodiol vs. femorol), lhe turo melhods were
considered comporoble.

Evidence section belorv deals onlv rvith the to be comparable to rhermodilution in

perlormance oi the updated algorithm.
Glinicol uses.The Floftac sensor can be
used for [ong-term continuous monitoring
ot- patients rvho are alreadv connected to
an invasive blood pressure monitor rvith
preerisdng arterial iine access.
Disposobles. The srstem requires the pro-
prietarr', disposable FloTrac sensor. It does
not require anv proprietary catheters.
Evidenre. \\'e revierved 16 srudies published
since 2010 that compared the performance
ot- the nerver generation oi the FloTrac
sensor with other cardiac output monitor-
ing techoiques. Seven of the 16 srudies
(see the table abor.e) compared the FloTrac
to pulmonarr- arter\- thermodilution, four
studies compared it to uanspulmonarr'
thermodilution, and the remaining 6r.e
compared it ro Doppler echocardiographr.
:j
The srudies invoived cardiac sutgerv, lir.er
transplantation, sepsis, and abdorninal sur-
gery patients, and the number o[ patients
in the srudies ranged from i8 to 60.
Fdwords Lifesciences FloTroc sensor.
In our previous revierq the earlier gen-
(lmoge courtesy of Edvrords Lifesciences.)
eration of the FloTrac sensor',vas found
stable cardiac surgerr patients, and the
one recent stud\- we rerierr-ed that looked
at cardiac surgerl patients (\hsdev et al.
2012) reports that this is the case with the
nerv version oF the FloTrac.
\Iost srudies on the nerver generauon
oi the sensor examlned its performance
in hemodlnamicallr unstable patient
populations (e.g., septic patients, liver trans-
plantadon patients). In 10 o[ the srudies
examined in our current tevies', the authors
conclude that the latest generation of the
FloTrac sensor is not compatable to the
relerence techmques; patient populations
in these srudies include patients undergo-
ing abdominai surgerl-, patients undergoing
liver transplantauon, and patients under-
going vasopressor therapr- (rvhich could
affect arterial compliance). -\mong srudies
comparing FloTrac to thermodilution for
patients with sepsis, one srudr (De Backer
et al. 2011) reported good accurao',
rvhereas t$-o others (\Iarque et al. 2013,
Slagt et al. 2013) reported a high petcentage
error. Tr'"'o srudies (Benes et al. 2010, \Iarer
et al. 2010) looked at patient outcomes

020]3 ECRI lnslitute. Member hospitols moy reproduce this poge for inlernol dislribulion only. w.eri.ors > HEALTHDEVICES JuNE2Ol3 183

10 of 158.
 gaidance
rvhen the FloTrac sensor \\?s used for
Buidmanagementduringmajorabdominal
surger'compared to,standard fluid man-
agemenr ePPfoacnes (usmg mean arrefral
pi"..ur. and ..rrtr"l',,.rro,i. pressure ro
guide therapr). Both srudies shotved that
oi'rt^r incidence oI compLi- 39]^'J::1"-l liver tronsplonlolion
Ihe lcogth
^.,d ence obstroct)
catrons decreased rvhen the FloTrac senior
rvas used. $ote that the f lar-er studr rvas
funded br- Edtvards Littsciences.)
Supplier Edn'ards Lifesciences Corp.
[374501], kvrne, C-\ pS-{); -l
(800) -+24-
3218, + I (949) 250-2500; t'rvrv.edr','ards.com
Procedure. The \blumeVierv sensor is
part of the \blume\ierv set and is used
in conlunction with the E\-1000 monitor
for continuous cardiac ourput morutor-
ing. The sensor uses the arterial r'"'ar-eform
analrsis technique to estimate cont-inuous
cardiac output, and the arterill s,-aveform is
obtained using the \blumeYierv femoral
catheter. Horvever, uniike the FloTrac
sensor, s'hich does not use any exter-
nal calibration, the YolumeYierv sensor
requires calibration using an external
cardiac outpur measurement. Tvpicallv,
ot
t*,
Edwords Lifesciences VolumeView sensor.
(lmoge courtesy of Edwords Lilesciences.)
184 xglrraDEvlcts JUNE2or3 > ww.*ri.ors
11 of 158.
LIERATURE REVTEW: EDryARD5 LlFEsclENcEs voLUMEVtEW
coMpARED To THERMoDILUTIoN
Study Polient populotion Reference method Oulcomes
!^o1lo 9t
ol, 5 potienls undergoing Bolus ond continu- lVith p-mentoge errors
:J-',!'lT:::J "'t"ry
thermodrlutron
calibration is recommended tbr arte-
rial rvaveform analrsis techniques For
unstable or criticallv ill patients at defined
intervals of time (can range from e'r'err
30 minutes to everl eight hours based oo
patient condition) or before undertak-
ing anr therapeutic inten'endons. The
rcndor offers the abilin-to perform trans-
pulmonarv thermodilution' using the
\blumeYierv set, and the resultiog cardiac
output measurements can be used to peri-
odicallr calibrate the YolumeView sensor
measurements.
The transpulmonan' thermodiiution
technique can also be used to obtain r-o1-
ume measurements such as giobal
end-diastolic volume, global ejection lrac-
tion, and extra\ascular lung s,ater.
cllnicol uses. The YolumeVies, set can be
used for long-term contjnuous monitoring
of padents rvho have an existiag central
renous linc.
Disposobles. The YolumeYierv set, rvhich
inciudes the \blumeYiew sensor and the
\blume\-ierv [emorai catheter.
Evidence. The evidence on the \blume-
Yierv sensor is limited, since the derice
rvas introduced in 201 1 and is therefore
relativelr nes'. We onh- found nvo srudies
that compared the YolumeYiew sensor
rvith other cardiac outpur monitoring tech-
niques in human patienrs. Costa er al.
(2012) compared rhe cardiac outpur
- .\ salinc solutioo of kn()\\-o tcmpcriturc is injcctcd
into ao cxistinla ccntral rcnous cathctcr, end rhc chengc
in rcmpcmturc of thc injcctatc is mcasured in thc [cmo,
nl artcrr u:ing thc \irlumc\iicrv crtheter,
@2013 ECRI lnstiiute. Member hospitols moy reproduce ihis poge for inlernol disiribution only.
o[.49o/o
:ii
.olus!fond
i",*lT -':ilil"-dI
conl'nuous fhermo-
dilution, the VolumeView wos
nol considered comporoble lo
the re{erence methods.
measurements from the YolumeYierv sen-
sof to those obtained from pulmonarr
arten- thermodilution rvith P^\Cs in five
patients undergoing liver transplanra-
tion (see the table above). lhe srudr-
[ound that the \blumeYierv results rvere
not comparable to the ralues obtained
from thermodilut-ion. This srudr. s,zs onlr
repotted as a conference abstract; thus, the
results reported mav not represent the 6nal
results and conclusions of the srudr. The
other studl', Kiefer ct al. (2012), compared
the \blumeYierv sensor rvith aflother arte-
rial s,'ar-e[orm analrsis technique that a]so
uses transpulmonarr. thermodi.lution for
calibration. This srudr rrzs performed on 72
critjcallr ill patients and found that the Vol-
umeYierv results *-ere comparable to those
ftom the reference method; note that the
studv $?s funded br Edrvards Lifesciences.
\\'e did not 6nd anv studies addressing
cl-inical outcomes rvith this device.
supplier. \I \Iedical US-\ [457874], -\kron,
{)H (US-\); +1 (800) 979-2904; rvrv*r
nimedical.co.il
The \ICaS (rvhich stands for
Procedure.
\on Invasive Cardiac Svstem) consists
of a laptop aod one pair oF s'rist-ankle
ICG electrodes. The st'stem is based on
the ICG technique, rvhich is based on the
theorl that changes in the voltage oF an
electrical signal applied across an arcz of
the bodr are due to changes in impedance
(rvhich in rurn are associated *'ith changes
in blood volume during each cardiac clcle).
 impedance through nvo ICG electrodes- (e.g., special ke1'board covers) ma,v be
one on the patieot's $,rist and the other on needed, particularlv rvhen the device is
the contralateral ankle. used in the ()R.

-{t the beginning of the procedure, Disposobles. One pair of rvrist-ankle

the clinician must enter the patient's
gender, age, t'eight, and height. This data
is used in conjunct-ion with the impedance
data obtained from the rvrist and ankle
electrodes to estimate stroke r.olume. The
electrodes also provide heart rate and a
single-channel electrocardiogram (ECG).
The ECG s.aveform, the impedance rvave-
form, and the hemodi'namic parameters
are all displared on the laptop. Standard
ECG electrodes are provided as an option
with the monitor and can be used in
addition to the ICG eiectrodes to obtain
three-channel ECG.
Cllnicql uses. The ICG technologv is inher-
entll suscept-ible to motion artifacts, and
n picalh- patieots need to be still s'h.ile
measuremeflts are being takeo.'fhus, this
monitoq like others that use the ICG
Nl Medicol NlCoS. (lmqgg courtesy oI model, is more suitable for spot-check
Nl Medicol.) measurement of cardiac output and
other hemodvnamic parameters than for
continuous monitoring. -\nother issue
to consider is that since this monitor is
-\Iost cardiac output monitors that use this
model measure changes in thoracic imped- laptop-based, infection conftol risks must
ance wlth electrodes placed on the chest; be taken into account, as the presence of
horr,'e'r'eE the \ICaS monitor measures a kerboard complicates cleaning of the
device. -\s a result, additional precautioas
elecuodes.
Evidence. \1'e reviewrd slx srudies comPar-
ing cardiac ourput measurements from
the \ICaS monitor with those ftom other
reference methods. Five of these stud-
ies used thermodilution as the refetence
method (see the table on this page), and
the remaining srudr' (Leitman et al. 2006)
used Doppler echocardiographr'. ()ne of
the srudies (Goor 2006) rvas reported as
a conlerence abstract onlr; therefore, the
reported findings mav represent onlr- pre-
liminarr results and not the 6na1 results
and conclusions o[ the srudr. The srudies
involved cardiac patients, and the number
of patients in the srudies ranged from
35 to 122.
-\ll slr srudies found good correlation
betrveen the NICaS svstem and the refer-
ence method. Tivo of the srudies (Cotter
et al. 2006, Goor 2006) also compared
thoracic ICG, in addition to the \ICaS
monitor, to thermodilution for the same
patients, and Found the \ICaS svstem to
better correlate with thermodilution.
\\:e did not Efld anv studres addressing
clinical outcomes with this device.

LITERATURE REVIEW: Nl MEDICAL NICAS COMPARED TO THERMODILUTION

Study Potient populotion Reference melhod Oulcomes
Coiter, Moshkovilz, 122 cordioc potients Bolus pulmonory ortery lVilh o correlotion coefficient of r : 0.886, the two methods
et ol. 2004 thermodilution were considered comporoble.

Cotter, Torre-Amiot,
e1 ol. 2004
Cotter et o]t.2006
Goor 2006 (con{erence
obsiroct)
Poredes et ol. 2006
' The obslroct does nol specify whot
93 polienls wilh heorl
foilure
43 cordioc potienis
46 cordioc polienls
35 cqrdioc potients
\pe of thermodilulion
Bolus pulmonory ortery With o correlotion coefficient o{ r = 0.8.l, the two methods were
thermodilution considered comporoble
Bolus pulmonory ortery o correlotion coeflicieni oi r
Vy'ith = 0.9. ihe two meihods were
ihermodilulion considered comporoble.
Thermodiluiion* Wilh o percenioge error ol 2OYo, lhe two methods were consid-
ered comporoble.
Bolus pulmonory ortery Wiih o percenioge error ol 19.95o/o ond o correloiion coefficieni
ihermodiluiion of r : 0.9], the iwo methods were considered comporoble.
wos used.

O2O l 3 ECRI lnstitute. Member hospilols moy reproduce lhis poge for internol disiribution only. w.*ri,org > HEALTHDEVICES JUNEzoTA I85

12 of 158.
 &!,,!:!{,c{"u,o,
RE FERENC ES
ISairemianJIr, RrxrkJI:, :\llard \IU, et al. (irmpan-
son of stnrkc trtlumc rariation obtained from
an arrcrial prcssurc to cndorraclteal sensor svs-
tcm. 2(111 .\nnual \[ccting of thc Internatir>nal
-\ncsthcsia Rcsearch Socicq'; 201 1 \Ia1' 21 -2-l;
\hncourcr (Oanada).
\, of
tlall'I'lt, (iulp l)(i, Patel cr al. (irmparison
tlrc cndotruchcal crrditc tlurput m<rnilor ltr
thcrmodilution in canlirc surgcn plicnts
.l
(ttli,rlrird: I,b;; zlnt;th 2(llo ()cl.21(5):1(t2-6.
Bcncs J, Oh1'tn I, .Utmann I et al. Inrnropcradr c
fltrid optimization using strokc r-olumc varia-
tion in high risk surgical paticnts: results oI
prospcctirc randomizcd srudt (n7 (-az
2{)1{};l-1(3):Rl lll.
lliancofiorc G, Critchlcl'L.\, Lcc .\, ct al. Iilalua-
tion of a ncs' sr>fru-arc tersitln of rhc lrlo l rac/
Yigileo (r-crsion 3.1)2) and a comparison rvith
prcr-ious data in cirdrotic paticnts undcrgo-
ing fir-er trrnsplanr rurgc^-- /1n.ith -'lrulg2o11
Sep;1 13(3):515-22.
Broch () ltcnnerJ, Clruencrvald ]I, ct al- -\ compad-
son of third-gcncrarion scmi-inrasirc ilrrerial
rvavcftrrm analvsis rrith rhermodilution in
patients undergoing coronatr surgcn'.
.\' Licttti/itlUbiltlurnul 21l2;2tt 12:15 lt)81. ll lpub
2{)l 2 Jul 31 .l
Costa (1, Cccconct'l', Baron Q et al. Clrdiac out-
put moniroring in cirilrotic paticnts: IiYl()(Xl
r-crsus pulmonry arrerv catltctcr--prcliminarv
data fc<rn[crcncc abstractl. lt (rili;e! (qrt
(onfirc tt. 32nd Intcrnational -svmPosium
on Inte nsivc Care and I'imcrgcncr \lcdrcine
Ilrussels llclgium. 2012; (irnlcrcncc Srarr(r-ar.
pagings):20 1 2{)320-21}1 2(}S79. .\lso atailable:
htg://ccfrrrum.com/contcnr/pdi/cc I 0ti26.pdi
(irrtcr G, Ikrshkoritz \ Kaluski Ii, ct al. -\ccunte,
nonim-asir-e continuous monitoring of cadiac
outpur bv s'holc-bodv clectrical bioimpcdancc.
Ur.,t 2001 .\pc I 25(-1): 1.13 I -'lt).
Cottcr (1, Schachncr -\, Sasson L, ct al- Impcd
ancc carditrgraphv retisitcd. I'l1itl -\l*ti 2(l(l(t
Scp;27(9):U1 7-27.
(lottcr (1,'lirrrc--\mot (], \tred Z, ct al. Wht>lc-
bodv clcctrical bio-impedrncc is rccururc
in non invasirc de tcrmination oI cardiac
output: a tlrermodilution contnrllcd, PmsPec-
tivc, d<rublc blind cr aluation. I lr.l I lurt I iil
.\' upll 21ll.l;3(Supp);1 9. .\lso alailablc: httP;//
cur j htiupp.oxtbrd journais.org/conrcnt/3/
supplemcnt/ I9. l.extract.
-lan
Dc Backer I), \las
G, ,\, ct al. .\rtcrial
pressurc-bascd canJiac output monitoring: a
186 xrarrxDEvlcEs JUNE2ol3 > ww.Bri.org
13 of 158.
mulriccntcr r alidation oI thc third-gcncration (irxrr l) l'cripheral rersus thoracic impcdance c,rrdi-
strfnvare in scptic Paricnrs. InlLn;irt (.are -\[td (Uriphr lconfcrcncc abstncrl. | ( ,trdiu, l iil 20{)6
2t)1 1 Itb;37(2):233--11). .\ug; 12(6 Suppl):S5J-5.
li(iltl Insriturc: J<trgcnscn I IK. llisg.rarJ.l, ()ilsm l. (-,'mp.rrison oi
bioimpcdance and ocsophagcal I)t>pplcr cerdiac
I:ndotilcheal crrJirc ourpur mrtnirrr st stem
output monitoring during abdomin:rl atlrtic sur-
(Oon\led (lorp.) tor dctermining carditc output
ger.t-. Cit (,an 20t2;16(SuPPl l).
and strokc r-olume lcusrom product brict-1.
(lustom I Iotlinc Se n-icc. 3)l2 l)ec -|. Khs-aonimit B, Bhuralanontachai l{. Prcdicrion oi
lilo'I'mc scnsor Qidrvards Litlscicnccs, l.l.(-) ftrr Buid rc-iponsircncss in scptic shock patienrs:
noninr-asilc cardiac output mrlnitoring lcustom comprring -strokc rrrlumc r trriarirtn br likilrac/
product briefl. Cust()m I I()tlinc Scn ice. 201 2 Vigilur ancl rur(,mJtLd ptrlsc prcssurc r.triltion.
Dcc 6. I Lrr.l,,bttt.;l I tt ;io I ?{}1 2 treb;29(2) :6{_9.
\oninr-asilc cardiac slstcm (\I \Icdical. Lrd) Kre fcr\, I krtcr CK, \Iirn (i, et al. (llinicLrl v'rlida-
[<rr monitoring cardiac ()utPut lcustom pnrduct rion of a nc$ the rm()dilution svstcm [<rr thc
bncf]. Custom I Iotlinc Scn-ice. 2012 l)ec 5. asscssment of crrdilc ourput rnd 11)lumctdc
Stra(l:r from thc hcart: undcrstanding cardilc paramc tcrs. ( i t ( -a n 2tl 12 \ Iav 3{); I 6(3) :ll'9tt.
ourput monitr>ring tcchniques lguidancc erticlel. Kor,rkc \', \imada'I', \agata I I, ct al. 'liansicnt
I kt lr h l) t ir i 2009 Dec;3ll(l 2):.3tt6--lt) l. hcmodvnamic change and xccurJc\'oI arrcrial
YolumcYics set ([idrvards Litcscicnccs, I-1.(-) blrxrd prcssurc-bascd cardi,rc ()utPut.. lrc'l,
[r>r minimellt inr rsilc crrdilc output moninrring ,lnd32ttl t .\ug;t 13(2):272-a.
[custom producr brict-1. Cusrom I lorlinc -\cil'ice . Kusaka \', \irshit;rni K, Iric '1, ct irl. Clinical compari-
201-l I'eb 15.
s()n oi rn cchocardiograph-dcni-ed r-crsus pulsc
l;cllahi-ll-, |ischer \l() Rebct () ct al. -\ comprrison counte r-dcrir'.d c.trdi.rc outpur mcxsurLmrnt in
oI cndotracheal bioimpcdance cardi<>gr'rphr abdominrl ?()rric xntutrsm surgcr.t. | (trdiotlttrt'
and transpulmonarr thcrmodilution in ctrdi:c l. ; t .. I n, t h 2tt I 2 .\pr,26(2) :223' (,.
surgcrv padcnrs.. | (.*di o lfuiit ; I L, : . lrc l l, 2111 2 "t
I-cirman \I, Suclre r t:, Kaluski l'1, ct al. \on-invasir c
.\pr'26(?):217-22.
mcasuremcnt of crrdirc ourput bv rvholc-bodr
(]cnnart I;, lJeckcrs \trborgh
O, et al. Impacr
S, bi<>impedlncc during doburaminc strcss cclttt-
of;rrtcrial carhcrcr krcation tln thc
accuritcr' crrdiogr.rpltr: clJnrc,rl implic.trioni ln Prticnt j
of cardiac ourput proi-ided bl an cnd()trachcal s ith Ic[t rrntricul;rr dtsfunction and ischaemia.
bioimpcriancc dclice [c<>nfcrcncc abstr,rcrl. Cz7 I:Lr I I hnrt litil2t){)6 \lar;tt(2):13(r'l{}.
(rr, 2{)12; t(r(Suppl l).
DO YOU HAVE A QUESTION ABOUT CARDIAC OUTPUT
MONITORING?
For exomple, ore you wondering . . .
) Whol fociors do I need to consider when purchosing o minimolly invosive or noninvos ve
cordioc output monilor?
) How widespreod is the use of cordioc outpui monitoring in clinicol proclice?
) Whot is lhe most estoblished olternotive technique in monitoring cordioc ouipui?
You con get onswers to these ond couniless other quesiions through HD Advisor. Just
click the link in lhe top righi corner o{ your Heolth Devices home poge to occess your HD
Advisor doshboord ond storl your inquiry.
@ADVISOR
Ask the [xperts
O2O'13 ECRI lnsiitute. Member hospitols moy reproduce this poge for internol distribution only'
 \Ianluc S, (iros .\, (.himot L, et al- (:ardiac outpur n'irh rhcrmodilurion. ;li/a ;1rut.;tLt.'iol .\',und 2012 \I.\, lScute J, ct al. (i:rrdiac otrtpur
Slrrgt (i, de Lceurv
monitoring in scptic slrock: criluad()n oI thc . \pr;5(r(-1) ;-133-{t). mcasured br unca[bratcd arteriai prcssurc s.avc-
third gener:rrion l;krtrac-\'igilcoO. .l Qin,\lonil lorm analvsis br rcccntlr rclcased st>fnrare r-cr-
\krnnct X, .\ngue I \,.lozn iak \1, ct al. 'l hird-
(,q nOil r 2( l l 3 .l 0n;27 (3):27 -(). sion 3.()2 r'crsus thcrmodilution in scptic shock.
gLncration I'kil'rac/\'igrlco docs not reliablr
3

\Iaus'I'^\I, llcbe r 13, llanks I).\, ct al- Oardiac outpur rrack changcs in cardiac outpur induccd bl nor-
dcrcrmination lnrm cndorrirchcirllr mcasured epincplrrine in cridcalll ill p^ticnts. Br.[ ,Arp]rilh
impedancc cardiographv: cljnical evaluarion of 2r) I 2 .\or;1 0tt(.l) :61 5-2?.
cndotrechcrl cardiac output monitor. / (,tnlilla-
\lurdri Sll, I Iess JR, I lcss .\, ct el. Irocuscd rapid
rn: V'r ii .' Ift t h 2ll I I ()ct;25(5):771)-5.
echocardiognphic cr-aluation r-crsus vascular
\Iarer,J, lloldt.J, \Icngisru .\\I, ct al. (loal-directcd cf,thcr-brsLd [*l ;rsscssmcnt oI c:rrdirc ourput
inrrxrpcntire thcrapr bascd on eutocalibnrcd xnd tuocri()n in criticallr' ill tnuma prticnts../
artcrirl prcssurc s alclorm analvsis rcduccs ho-.-
'
l)u r mt . l,u t c (.o rc .\' m3 2t) 12 \ Iav;72(5) : I I 58-6-1.
pital stav in high-nsk surgical paticnts: r rand<>m-
\lutoh 'I, Ishikarra'l', Kobalashi S, ct al. l)ertirr-
ized, controlled tnal. (. i r ( ut 2o ll ); 1 -t(1 ) :l{ 1 tl.
mancc o I tlrird-gencrati<>n Irkiliac/\'igilc<r
I Uin .\[uir (tn\ar 2013 .\pc27(2):l7l-7.
Su l](-,'t-.ai Yl', Chcn C\', ct al, (lardiac ()utput de-
riled fiom artcrial pressurc q-ar-cfrrrm analvsis in
padcnrs undcrgoing lir-cr transplantation: i'alidiu
oi r third-gcncrarion dcticc. 'l-ratpltnl Prut2ll12
\Iar;$t(2):'12-l lt.
'lsai Yl; Su lXl, Lin (l(1, ct al. O;rrdirc ourput dcrii-td
trom arrcrial prcssure t avciorm enallsis: r alidl-
trrn oi tlrc rhir,.l-gcncr:tion sol-ro'rre in peticnrs
undcrgoing r>rthotopic lir-er rransplantation.
l raulllail Pro: 2ll12 \Ier;-l-l(2):-1.33-7.
'

\Icl.ean .\S, t luang S-f, Kot \I, ct al. Oomperison of srstcm during hrpcnlrnamic thcrrrpy lor deleled
cardiac output measu(cmrnts in criticallr ill pa- ccrebral isciremia aticr subarachnoid hemor- Van dcr Klcii S(-, Koolen B, (icrritse B, cr rr1. (-lini-
ticnts: ljlo'I r,rc/\'igileo rs transrlroracic [)oppler rhage . .1ir4:i -\irrr, / Iilt 2lll?;3'99. cal cr:rluatir>n ot- a nes' cndotrachcal impcd-
cclrocardi<rgrapll'. . lart,/h Iukr,il t Ltre 2O11 ince cardiogralhl mcthod. tlu,,/fu.,it 2l)12
\ational I Icarr, I-ung, and l]lood Instinrte .\cutc
-[ul;39(.+):.i9{).tl. lle spiraron Distrcss Slntlromc (.\Rl)S) Clinical .lul{t7 (7):729-33.
\Icng I., 'l ran \P, .\lcxandcr IlS, cr al. 'l'hc impacr 'lrials \cnrrrrk; \\'hcclcr .\l', llcrnard ()R, cr al. \ esdei- S, (-hauhan S, (ihoudhun \1, ct al. \rtcrixl
r>i phcnrlephrinc, cphcdrinc, and incrcascd Pulmonan'artcfl' rrrsus ccntral \ cnous cathcter pressure s ar-cfrrrm dcrir-cd cardiac output I;kr-
prcload on third-generation Yigileo-likil rac and to lluidu rreatmrnt of lcutc lung injun: \ I:41 | 'lrrc/\'igilco srstcm (third gcncration sofnvarc):
csophagcal l)opplcr cardiac output mcasurc- .\lid 2{t{t6 \la1' 25;35"1(21):221.3-2-l comparison of nvo monitoring sitcs r-ith thc
mcnts. -. h c't h .. 1na13 2{) I I ( )cs1 1.3(-l):75 1 -7. tlrcrmr>dilution cardiac ourput../ (l)r -\Iqnit (.rm-
Parcdcs ()1., Shire-[, Shinkc'I, ct al. Impedancc
\Ictzcldcr Coburn \I, Iirics \1, er al. Performancc
S, cardiograplrv [rrr cardiac output cstimarion: rcli- lrl 2012.\pr;26(2): 1 1 5-2t).
of cardiac output measurcmcnr de rir-cd from abilin oI rvrist-to-anklc elccrrodc conligurarion. \\'icncr R-S, \\llch I I(]. 'lrcnds in rhc usc of pulmo-
artcrial prcssurc s'ar-efrrrm anah sis in prticnts (.ir:.1 ?0o6 Sup;7tt(9):l l6+fl. nan irtcn'catheter in the United Srates, 1993-
requirin.q high-dosc \f,s()prcss(rr thcnpl. IJr./ ?ottl. l,'1 2r)l )7 -f ul 25;29i1({) :t23t. lD
Sclrl-ann \\I, I Iitlel Z, I krutl .\, ct al. I-ack of 1.. 1,\

n *,th 2{l I I Jun-10(t(():77 G{1-1.

eftectitcncss ol thc pulmonan' artcn' cath-
;1

\ftrllcr-Sorcnsen I [, I hnsen KL, Ostcrgaard \1, ct crcr in cardiac :urgerr. .'luitb t7nQg2\l1l
al. Lack oi agrermcnt and trcnding abilin oi the \or';l 1.j(5):99{-11){)2.
endorrachcal cardiac outpur monitor comparcd

OUR LATEST SURVEYS

Toke our physiologit moniloring syslem survey. We wont your thoughls on your physiologic
moniioring syslem's eose of use, functionolity, ond reliobility. Toke o {ew minules to fill out our
user experience survey, ovoiloble through your member home poge or ot hfips://survey.ecri.oq/
Survey.ospx?i=2178o362f1.244O2c9d0159tc11629Occ. When done, you'll be outomoticolly
enfered to win on iPod Mini.
See the resuhs of our CT suiyey. The results of our CT system user experience survey ore
now ovoiloble. Find out how your peers ronked lhe iop vendors ond models on lhe morket in
terms o{ lheir functions, {eolures, ond seruice. View the resulb ot https://members2.ecri.org/
Componenls/UserExperience/Poges/ct_sconners/userexp_moin.ospx.
ir

O2013 ECRI lnst:tule. Member hospitols moy reproduce ihis poge for internol distribulion only w.xri.ors > HEATTHDEVICE5 JUNE2ol3 187

14 of 158.
 Advance Publication by-J-STAGE
Circulation Journal
Official Journal of the Japanese Circulation Society
http://www. j-circ.or.jp

Noninvasive and Simple Assessment of Cardiac Output

and Pulmonary Vascular Resistance With Whole-Body
Impedance Cardiography Is Useful for Monitoring
Patients With Pulmonary Hypertension
Yu Taniguchi, MD; Noriaki Emoto, MD, PhD; Kazuya Miyagawa, MD, PhD;
Kazuhiko Nakayama, MD, PhD; Hiroto Kinutani, MD; Hidekazu Tanaka, MD, PhD;
Toshiro Shinke, MD, PhD; Ken-ichi Hirata, MD, PhD

Background: Right heart catheterization (RHC) is the gold standard for the diagnosis of pulmonary hypertension
(PH) and a useful tool for monitoring PH. However, there are some disadvantages in the regular use of RHC because
it is invasive. Noninvasive methods for monitoring hemodynamics are needed to manage patients with PH. In this
study, we aimed to evaluate the reliability of noninvasive hemodynamic assessment with whole-body impedance
cardiography (Non-Invasive Cardiac System [NICaS]) for PH.

MethodsandResults: We investigated 65 consecutive patients undergoing RHC. Two-thirds of them had pulmo-
nary arterial hypertension and one-third had chronic thromboembolic PH; 25% of the patients were receiving medi-
cal therapy. Cardiac output (CO) was estimated by NICaS (NI-CO), thermodilution (TD-CO), and the Fick method
(Fick-CO). There was a strong correlation between NI-CO and TD-CO (r=0.715, P<0.0001) and Fick-CO (r=0.653,
P<0.0001). Noninvasive pulmonary vascular resistance (PVR) was estimated using a conventional invasive equation
with NI-CO, mean pulmonary arterial pressure was calculated by echocardiographic measurement, and pulmonary
capillary wedge pressure was estimated at 10mmHg in all cases. NICaS-derived PVR was very strongly correlated
with invasive PVR (TD-PVR: r=0.704, P<0.0001; Fick-PVR: r=0.702, P<0.0001).

Conclusions: Noninvasive measurement of CO and PVR using NICaS and echocardiography is a useful tool for
the assessment of PH.

Key Words: Cardiac output; Noninvasive assessment; Pulmonary hypertension; Pulmonary vascular resistance;
Whole-body impedance cardiography

P
ulmonary arterial hypertension (PAH) is a progressive
disease characterized by elevated pulmonary vascular
resistance (PVR) because of pulmonary vascular re-
modeling. This leads to a decrease in cardiac output (CO) and
ultimately death. Recently, targeted medical therapy for PAH
patients with endothelin-receptor antagonists, phosphodies-
terase-5 inhibitors, and prostacyclin analogs has been estab-
lished,1 and the prognosis of PAH has improved.2 However,
there is no universally accepted consensus on the treatment
goals or follow-up strategy for PAH patients. Right heart cath-
eterization (RHC) is not only the gold standard for the diagno-
sis of PAH, but is also a useful tool for monitoring PAH, and
is recommended 36 months after new treatments and in the
case of clinical worsening.1 Hemodynamic monitoring with
RHC is predictive of survival and effective in a goal-oriented
treatment strategy,3,4 and has been recommended by a recent
guideline;1 however, there are some disadvantages in the reg-
ular use of RHC as a follow-up procedure, especially with
regard to invasiveness. Noninvasive and less complicated
methods for monitoring hemodynamics are needed to manage
patients with pulmonary hypertension (PH). Less invasive
hemodynamic monitoring has recently been suggested as fea-
sible in some situations.5 The Non-Invasive Cardiac System
(NICaS; NI Medical, Hod-Hasharon, Israel) is a device for
calculating CO noninvasively with whole-body impedance
cardiography (ICGWB). The NICaS-derived CO (NI-CO) has
been shown to be as reliable as the RHC-derived CO and is
applicable for the noninvasive assessment of cardiac function

Received February 3, 2013; revised manuscript received May 1, 2013; accepted May 2, 2013; released online June 12, 2012 Time for
primary review: 42 days
Division of Cardiovascular Medicine, Department of Internal Medicine, Kobe University Graduate School of Medicine, Kobe (Y.T., N.E.,
K.M., K.N., H.K., H.T., T.S., K.H.); Clinical Pharmacy, Kobe Pharmaceutical University, Kobe (N.E., K.N.), Japan
Mailing address: Noriaki Emoto, MD, PhD, Division of Cardiovascular Medicine, Department of Internal Medicine, Kobe University
Graduate School of Medicine, 7-5-1 Kusunoki, Chuo-ku, Kobe 650-0017, Japan. E-mail: emoto@med.kobe-u.ac.jp
ISSN-1346-9843doi:10.1253/circj.CJ-13-0172
All rights are reserved to the Japanese Circulation Society. For permissions, please e-mail: cj@j-circ.or.jp

15 of 158.
 Advance Publication by-J-STAGE
TANIGUCHI Y et al.

Table1. Clinical Characteristics of All Patients at Initial Table2. Hemodynamic Measurements

Hospitalization
Parameter Value
Age (years) 6214 TD-CO (L/min) 4.921.56
Female (%) 39 (65) Fick-CO (L/min) 3.871.24
Diagnosis (%) Echo-CO (L/min) 4.341.11
PAH 38 (63) NI-CO (L/min) 4.401.32
IPAH 12 (20) TD-PVR (dyns1cm5) 446249
CTD-PAH 24 (40) Fick-PVR (dyns1cm5) 583362
Po-PAH 2 (3) Echo-PVR (dyns1cm5) 660363
PH associated with respiratory disorders 3 (5) NI-PVR (dyns1cm5) 544316
CTPH 20 (33)
Echo-CO, echocardiography-derived cardiac output; Echo-PVR,
WHO-fc (%) echocardiography-derived PVR; Fick-CO, cardiac output derived
1 1 (1.7) by the modified Fick method; Fick-PVR, PVR derived by modified
Fick method; NI-CO, NICaS-derived cardiac output; NI-PVR,
2 22 (37.3)
NICaS with echocardiography-derived PVR; PVR, pulmonary
3 32 (54.2) vascular resistance; TD-CO, thermodilution-derived cardiac output;
4 4 (6.8) TD-PVR, thermodilution-derived PVR.
Treatment (%) 25 (24)
Bosentan 13 (22)
Sildenafil 14 (23) sis, and dialysis, all of which interfere with the accurate mea-
Beraprost 10 (17) surement of impedance-derived CO with NICaS, as previ-
Hemodynamic variables ously described.6 Patients with elevated pulmonary capillary
sPAP (mmHg) 53.921.3 wedge pressure (PCWP >15mmHg) on RHC were also ex-
mPAP (mmHg) 31.712.0 cluded; 5 patients were excluded because of the presence of an
RAP (mmHg) 3.74.2 intra-cardiac shunt; 29 patients were reevaluated 36 months
PCWP (mmHg) 7.04.3
after new treatments or at clinical worsening.
CO (TD) (L/min) 4.901.62
Hemodynamics
CO (Fick) (L/min) 3.922.08
Hemodynamic data were derived from standard RHC in all
PVR (TD) (dyns1cm5) 433244 patients using a 6Fr Swan-Ganz catheter (Baxter Healthcare,
PVR (Fick) (dyns1cm5) 581344 Irvine, CA, USA). The catheter was introduced into the pul-
HR (beats/min) 7311 monary artery under fluoroscopic guidance. Mean pulmonary
CO, cardiac output; CTD-PAH, collagen tissue disease associ- arterial pressure (mPAP), systolic and end-diastolic pulmonary
ated PAH; CTEPH, chronic thromboembolic pulmonary hyperten- arterial pressure (sPAP and dPAP), mean right atrial pressure,
sion; HR, heart rate; IPAH, idiopathic PAH; mPAP, mean PAP; and PCWP were measured. CO was measured using the fol-
PAH, pulmonary arterial hypertension; Po-PAH, portal PAH;
PCWP, pulmonary capillary wedge pressure; PVR, pulmonary
lowing techniques.
vascular resistance; RAP, right atrial pressure; sPAP, systolic Thermodilution-Derived CO (TD-CO) A 5-ml bolus of iced
pulmonary arterial pressure; TD, thermodilution; WHO-fc, World 5% glucose solution was injected 5 times at the same rate. The
Health Organization functional class. results of 3 injections within 15% of their extreme disparity
were averaged to derive the TD-CO value.
Modified Fick method (Fick-CO)Blood samples were ob-
in patients with left-sided chronic heart failure,6,7 but its feasi- tained from systemic and pulmonary arteries. All samples were
bility in patients with PH has not been evaluated. The purpose measured for oxygen saturation with the same device (Radi-
of this study was to evaluate the reliability of noninvasive ometer ABL 715, Copenhagen, Denmark).
measurement of CO and PVR with ICGWB in patients with PH. NI-COThese measurements were performed simultane-
ously with the measurement of TD-CO and Fick-CO during
Editorial p???? RHC. The measurement of NI-CO followed the method as
previously reported:6 an alternating electrical current of 1.4mA
with a 30-kHz frequency is passed through the patient via 2
Methods pairs of tetrapolar electrodes 1 pair placed on the wrist above
This study was approved by Kobe University Hospital Institu- the radial pulse, and the other pair placed on the contralateral
tional Review Board and the patients provided written informed ankle above the posterior tibialis arterial pulse. If the arterial
consent to participate. pulses in the legs are either absent or of poor quality, the sec-
ond pair of electrodes is placed on the contralateral wrist.
Patients The NICaS apparatus calculates the stroke volume (SV) by
We enrolled 65 consecutive patients with known or suspected Frinermans formula:8
pulmonary hypertension hospitalized in Kobe University Hos-
SV=dR/RL2/Ri(+)/KWHF
pital from April 2010 to August 2011. All patients who were
scheduled for RHC without fulfilling one of the exclusion where dR is the impedance change; R is the basal resistance;
criteria were eligible for the study. The exclusion criteria in- is the blood electrical resistivity; L is the patients height; Ri
cluded restlessness and/or unstable patient condition, severe is the corrected basal resistance according to sex and age; KW
aortic valve regurgitation and/or aortic stenosis, aortic aneu- is a correction factor for weight according to ideal values; HF
rysm, heart rate >130beats/min, intra- and extracardiac shunts, is the hydration factor, which takes into account the body
severe peripheral vascular disease, severe pitting edema, sep- water composition. + is equal to the ECG R-R wave interval

16 of 158.
 Advance Publication by-J-STAGE
Noninvasive Assessment of PH

Figure1. Linear correlation analysis between TD-CO and NI-CO (A), Fick-CO and NI-CO (B), TD-CO and Fick-CO (C), TD-CO
and Echo-CO (D), and Fick-CO and Echo-CO (E). TD-CO, thermodilution-derived cardiac output; NI-CO, NICaS-derived cardiac
output; Fick-CO, cardiac output derived by the modified Fick method; Echo-CO, echocardiography-derived cardiac output.

and is the diastolic time interval. To calculate the CO, SV is
multiplied by the heart rate. Because the NI-CO values are
calculated every 20s, the average of 3 measurements obtained
consecutively during 60s of monitoring is considered to be the
NI-CO value for each individual case.
Echocardiography
Echocardiography was performed using a Vivid 5 system and
a 3.5-MHz transducer (GE Vingmed Ultrasound AS, Horten,
Norway). Two-dimensional Doppler examinations were per-
formed in the usual manner. CO was measured by tracing the
left ventricular ejection flow (Echo-CO). Echo-sPAP was es-
timated from the peak velocity of the tricuspid regurgitation
jet plus estimated right atrial pressure (Echo-RAP).9
Measurement of PVR
PVR (dyns1cm5) was calculated using RHC from the equa-
tion:
PVR=80(mPAPPCWP)/COTD, Fick (TD-PVR, Fick-PVR).
PVR was also estimated noninvasively using a combination
of NICaS and echocardiography, and by echocardiography
alone.10 Echo-mPAP was calculated as Echo-sPAP0.61+
2mmHg, as previously described,11 and noninvasive PVR was
calculated as 80(Echo-mPAPPCWP)/COEcho,NI (Echo-PVR,
NI-PVR). PCWP for the calculation of noninvasive PVR was
estimated at 10mmHg in all cases.10
Statistical Analysis
Quantitative data are presented as meanSD. The correlations

17 of 158.
 Advance Publication by-J-STAGE
TANIGUCHI Y et al.

Figure2. Bland-Altman plots with mean difference (solid line) 2SD (dotted line) comparing TD-CO and NI-CO (A), Fick-CO and
NI-CO (B), TD-CO and Echo-CO (C), and Fick-CO and Echo-CO (D). TD-CO, thermodilution-derived cardiac output; NI-CO,
NICaS-derived cardiac output; Fick-CO, cardiac output derived by the modified Fick method; Echo-CO, echocardiography-derived
cardiac output; SD, standard deviation.

among TD-CO, Fick-CO, Echo-CO, and NI-CO and between
Echo-mPAP and mPAP measured by RHC (RHC-mPAP)
were determined by calculating the Spearmans rank corre-
lation coefficient. P<0.05 was considered to be significant.
Agreement between methods was analyzed by the Bland-Alt-
man method.12 The limits of the agreement were expressed as
the meanSD. The 95% confidence intervals (CIs) of the bias
were also calculated. Receiver-operating characteristic (ROC)
curves were generated for the detection of elevated PVR de-
fined as > 240dyns1cm5 (3 Wood units [WU]). The area not strong (r=0.512 or 0.461, P<0.0001, respectively). The
under the curve (AUC), cut-off value, sensitivity, and specific-
ity were estimated by the ROC curves. All statistical analyses
were performed using GraphPad Prism version 5 (GraphPad
Software, La Jolla, CA, USA).
Results
The baseline characteristics of all patients at initial hospital-
ization are summarized in Table1. Approximately two-thirds
of the patients had PAH (World Health Organization [WHO]
classification of PH group 1) and the other one-third of the
patients had chronic thromboembolic PH (CTEPH: WHO
group 4); 5% of the patients were classified as WHO group 3.
At enrollment, 24% of the patients were receiving medical
therapy.
Relationships Among Parameters
The mean CO values from all measurements in these sub-
jects for TD-CO, Fick-CO, Echo-CO, and NI-CO were 4.92
1.56L/min, 3.871.24L/min, 4.341.11L/min, and 4.40
1.32L/min, respectively (Table2). A significant and very
strong correlation was observed between TD-CO and NI-CO
(r=0.715, P<0.0001) and between TD-CO and Fick-CO (r=
0.795, P<0.0001) by 2-tailed Spearmans rank correlation test
(Figure1). There was a strong correlation between Fick-CO
and NI-CO (r=0.653, P<0.0001). However, the correlation
between Echo-CO and TD-CO or Fick-CO was significant but
differences between 2 measurements were plotted according
to the Bland-Altman method (Figure2). The mean bias and
limits of agreement between TD-CO and NI-CO, Fick-CO and
NI-CO, and TD-CO and Fick-CO were 0.501.08 (1.61 to
2.61) L/min, and 0.541.04 (2.57 to 1.49) L/min, and 1.02
0.86 (0.68 to 2.71) L/min, respectively. The limits of agree-
ment between TD-CO and Echo-CO, and Fick-CO and Echo-
CO were 0.641.33 (1.97 to 3.26) L/min and 0.421.18
(2.73 to 1.88) L/min, respectively. There was no clear differ-
ence in the measurements of CO among the patients with id-
iopathic PAH, collagen tissue disease associated PAH or
CTEPH (FigureS1).
Comparison of Invasive and Noninvasive Measurement of
mPAP and PVR
The mean values of all measurements of invasive mPAP and
Echo-mPAP were 32.91.28mmHg and 43.01.59mmHg, re-
spectively. There was a very strong correlation between inva-

18 of 158.
 Advance Publication by-J-STAGE
Noninvasive Assessment of PH

sive mPAP and Echo-mPAP (r=0.703, P<0.0001; Figure3A).

The limits of agreement between invasive mPAP and Echo-
mPAP were 9.6310.2 (29.6 to 10.4) mmHg (Figure3B).
The mean values of all measurements of TD-PVR, Fick-
PVR, Echo-PVR, and NI-PVR were 446249dyns1cm5,
583362dyns1cm5, 660363dyns1cm5, and 644316
dyns1cm5, respectively (Table2). There were significant
and very strong correlations between TD-PVR and NI-PVR
(r=0.704, P<0.0001), between Fick-PVR and NI-PVR (r=0.702,
P<0.0001), and between TD-PVR and Fick-PVR (r=0.942,
P<0.0001) (Figures4A,D). However, the correlation between
PVR measured by invasive methods and Echo-PVR was
not as strong (r=0.602 or 0.603, P<0.0001, respectively;
Figures4G,J) as that between invasive methods and NI-PVR.
Figure4B and dyns1cm5 shows the Bland-Altman plots of
the differences between TD-PVR, Fick-PVR, and NI-PVR.
The limits of agreement between TD-PVR and NI-PVR,
Fick-PVR and NI-PVR, and TD-PVR and Fick-PVR were
195265 (715 to 326) dyns1cm5, 35325 (673 to 603)
dyns1cm5, and 135164 (457 to 187) dyns1cm5,
respectively. The limits of agreement between TD-PVR and
Echo-PVR, and Fick-PVR and Echo-PVR were 191266
(713 to 330) dyns1cm5, and 33341 (703 to 635)
dyns1cm5, respectively (Figures4H,K). The AUC for NI-
PVR to detect increased PVR >240dyns1cm5 (3 WU)
against TD and Fick-PVR were 0.84 (95% CI, 0.720.96) and
0.92 (95% CI, 0.840.99), respectively (Figures3C,F), and
optimal cut-off values were 411dyns1cm5 (sensitivity:
81.3%, specificity: 75%) and 400dyns1cm5 (sensitivity:
80.3%, specificity: 100%), respectively. The AUC for Echo-
PVR against TD and Fick-PVR were lower: 0.75 (95% CI,
0.570.92) and 0.83 (95% CI, 0.660.99) (Figures4I,L) com- Figure3.Linear correlation analysis between RHC-mPAP
and Echo-mPAP (A). Bland-Altman plots with mean difference
pared with that for NI-PVR against TD and Fick-PVR. (solid line) 2SD (dotted line) comparing RHC-mPAP and
Echo-mPAP (B). RHC, right heart catheterization; mPAP,
mean pulmonary arterial pressure; Echo-mPAP, mPAP calcu-
Discussion lated by echocardiography; SD, standard deviation.
We report on the reliability of a noninvasive and simple meth-
od of assessing CO and PVR using ICGWB in patients with PH.
Previous reports have indicated the feasibility of hemody-
namic assessment using various methods in comparison with and echocardiography. Kouzu et al showed that tricuspid re-
RHC in a range of clinical settings;13 however, a reliable meth- gurgitant pressure gradient (TRPG)/right ventricular time-
od for the assessment of PH has not yet been established. In velocity integral (TVI) is reliable for the estimation of PVR.21
particular, there are few studies that have addressed the nonin- Although TRPG/TVI has been confirmed as a reliable method
vasive assessment of hemodynamics in PH. Hemodynamic for estimating PVR,16 accurate measurement of TVI needs a
assessment using cardiac magnetic resonance or echocardiog- skilled operator. Lindqvist et al reported the accuracy of a
raphy have been shown to be reliable,10,1417 but these methods simple Doppler-derived measurement of PVR with the con-
require expensive equipment and trained operators. Thoracic ventional invasive equation in patients with PH;10 however,
impedance cardiography has been used for the measurement of their study excluded patients with severe tricuspid regurgita-
CO noninvasively in PAH,18 and its reliability was shown to be tion, which causes inaccuracy in CO measurement using echo-
compromised in cardiac patients in a meta-analysis.19 cardiography. In the present study, we estimated PVR using
We demonstrated strong correlations among the NICaS, TD, the conventional invasive equation with the combination of
and the Fick methods for the measurement of CO. Although NI-CO and Echo-mPAP. There was very strong correlation
the limits of agreement between NI-CO and TD-CO or Fick- not only between invasive mPAP and Echo-mPAP, but also
CO estimated by the Bland-Altman approach were not small, between invasive PVR and NI-PVR. Furthermore, a stronger
they were acceptable when compared with previous reports.13 correlation of PVR was found for the use of NI-CO compared
Therefore, we believe that NICaS can be a reliable tool for the with Echo-CO. The limits of agreements estimated by the
noninvasive assessment of CO in PH. However, compared Bland-Altman analysis were large, but comparable with previ-
with NI-CO, the correlation between Echo-CO and TD-CO or ous reports that showed the feasibility of PVR derived by
Fick-CO was weaker and the limits of agreements were larger. echocardiography against invasive PVR.21,22 Furthermore, the
The relative inaccuracy of CO measured by echocardiography high AUC, sensitivity, and specificity for NI-PVR to detect
was consistent with a previous report,20 and may be a conse- increased PVR >240dyns1cm5 (3 WU) also indicates the
quence of using the Doppler method, severe tricuspid regurgi- reliability of noninvasive PVR assessment using NICaS.
tation, and operator-dependency. In our study, the value for TD-CO was significantly higher
We also demonstrated the feasibility of noninvasive and than the CO values with other methods, including Fick-CO,
simple measurement of PVR using a combination of NICaS and therefore, the value of TD-PVR was underestimated. This

19 of 158.
 Advance Publication by-J-STAGE
TANIGUCHI Y et al.

Figure4. Diagnostic reliability of NI-PVR compared with invasive PVR. Linear correlation analysis between TD-PVR and NI-PVR
(A), Fick-PVR and NI-PVR (D), TD-PVR and Echo-PVR (G), and Fick-PVR and Echo-PVR (J). Bland-Altman plot with mean differ-
ence (solid line) 2SD (dotted line) comparing TD-PVR and NI-PVR (B), Fick-PVR and NI-PVR (E), TD-PVR and NI-PVR (H), and
Fick-PVR and NI-PVR (K). Area under the receiver-operating characteristics curve with 95% confidence interval for NI-PVR and
Echo-PVR to detect increased PVR (>240dyns1cm5 [3 WU]) against TD-PVR (C,I) and Fick-PVR (F,L). PVR, pulmonary vascu-
lar resistance; NI-PVR, NICaS with echocardiography-derived PVR; TD-PVR, thermodiluyion-derived PVR; Fick-PVR, PVR derived
by the modified Fick method; Echo-PVR, echocardiography-derived PVR.

could be caused by overestimation of the value of TD-CO in
the presence of low CO, consistent with previous reports.23
Study Limitations
The main limitation of this study was the need to measure the
Doppler parameter for estimating NI-PVR. Proper alignment
of the ultrasound beam is crucial for the Doppler parameter to
be determined appropriately. This may have resulted in bias in
the measurement of NI-PVR. In our study, the Doppler param-
eter needed in order to estimate NI-PVR was only TRPG, and
there was no patient in whom we were unable to obtain that
value. Second, we used the conventional invasive equation for
estimating NI-PVR. We had to estimate PCWP at 10mmHg
in all cases as previously reported,10 which may also have re-

20 of 158.
 Advance Publication by-J-STAGE
Noninvasive Assessment of PH
sulted in the measurement of NI-PVR; however, in general, a
wide variation in PCWP is not usually observed among pa-
tients with PH. Third, because CO measurement using NICaS
in patients with cardiac shunts is known to be unreliable,24 we
excluded cases of PAH associated with cardiac shunts. Fourth,
noninvasive estimation of CO and PVR with NICaS was fea-
sible; however, there were some patients who had large diver-
gence between NI-CO or NI-PVR and invasive CO or PVR.
Further studies are needed to clarify the factors that lead to
inaccurate measurements of CO and PVR. Fifth, in our study,
the number of patients with WHO functional class 4 was
small. Most patients were WHO functional class 2 or 3. The
reliability of NICaS in patients with severe PH is to be exam-
ined in future studies. Finally, the study sample size was rela-
tively small and originated from a single center. We believe
that a larger, multicenter study is needed to appropriately
confirm the reliability of the method.
Although recent advances in treatment options and manage-
ment have improved the outcomes for patients with PH, treat-
ment goals and follow-up strategy are still not well defined.
Hemodynamic monitoring with RHC is recommended in a
goal-oriented treatment strategy for PH1 and is the gold stan-
dard for the assessment of PAH; however, the invasiveness of
RHC is a critical factoring its regular use as a follow-up pro-
cedure. A noninvasive, accurate, and simple method is re-
quired for the management of patients with PH. We have
demonstrated noninvasive measurement of CO and PVR using
only simple parameters. Echo-sPAP is needed to estimate
PVR, but Echo-sPAP has been established as a simple, reli-
able screening parameter for PH.25 This noninvasive, reliable,
and simple assessment can be a useful tool for monitoring and
managing patients with PH.
Conclusion
Noninvasive measurement of CO and PVR using NICaS is as
reliable as invasive RHC. This simple assessment could help
physicians to manage their patients with PH.
Disclosures
None.
References
1. Galie N, Hoeper MM, Humbert M, Torbicki A, Vachiery JL, Barbera
JA, et al. Guidelines for the diagnosis and treatment of pulmonary
hypertension: The Task Force for the Diagnosis and Treatment of
Pulmonary Hypertension of the European Society of Cardiology (ESC)
and the European Respiratory Society (ERS), endorsed by the Inter-
national Society of Heart and Lung Transplantation (ISHLT). Eur
Heart J 2009; 30: 24932537.
2. Benza RL, Miller DP, Barst RJ, Badesch DB, Frost AE, McGoon
MD. An evaluation of long-term survival from time of diagnosis in
pulmonary arterial hypertension from the REVEAL Registry. Chest
2012; 142: 448456.
3. Fukumoto Y, Shimokawa H. Recent progress in the management of
pulmonary hypertension. Circ J 2011; 75: 18011810.
4. Hoeper MM, Markevych I, Spiekerkoetter E, Welte T, Niedermeyer
J. Goal-oriented treatment and combination therapy for pulmonary
arterial hypertension. Eur Respir J 2005; 26: 858863.
5. Alhashemi JA, Cecconi M, Hofer CK. Cardiac output monitoring: An
integrative perspective. Crit Care 2011; 15: 214.
6. Paredes OL, Shite J, Shinke T, Watanabe S, Otake H, Matsumoto D,
et al. Impedance cardiography for cardiac output estimation: Reliabil-
ity of wrist-to-ankle electrode configuration. Circ J 2006; 70: 1164
1168.
7. Tanino Y, Shite J, Paredes OL, Shinke T, Ogasawara D, Sawada T, et
al. Whole body bioimpedance monitoring for outpatient chronic heart
failure follow up. Circ J 2009; 73: 10741079.
21 of 158.
8. Cohen AJ, Arnaudov D, Zabeeda D, Schultheis L, Lashinger J,
Schachner A. Non-invasive measurement of cardiac output during
coronary artery bypass grafting. Eur J Cardiothorac Surg 1998; 14:
6469.
9. Yeo TC, Dujardin KS, Tei C, Mahoney DW, McGoon MD, Seward
JB. Value of a Doppler-derived index combining systolic and dia-
stolic time intervals in predicting outcome in primary pulmonary
hypertension. Am J Cardiol 1998; 81: 11571161.
10. Lindqvist P, Soderberg S, Gonzalez MC, Tossavainen E, Henein MY.
Echocardiography based estimation of pulmonary vascular resistance
in patients with pulmonary hypertension: A simultaneous Doppler
echocardiography and cardiac catheterization study. Eur J Echocar-
diogr 2011; 12: 961966.
11. Chemla D, Castelain V, Humbert M, Hebert JL, Simonneau G,
Lecarpentier Y, et al. New formula for predicting mean pulmonary
artery pressure using systolic pulmonary artery pressure. Chest 2004;
126: 13131317.
12. Bland JM, Altman DG. Statistical methods for assessing agreement
between two methods of clinical measurement. Lancet 1986; 1: 307
310.
13. Mantha S, Roizen MF, Fleisher LA, Thisted R, Foss J. Comparing
methods of clinical measurement: Reporting standards for bland and
altman analysis. Anesth Analg 2000; 90: 593602.
14. Inaba T, Yao A, Nakao T, Hatano M, Maki H, Imamura T, et al.
Volumetric and functional assessment of ventricles in pulmonary
hypertension on 3-dimensional echocardiography. Circ J 2012; 77:
198206.
15. Garcia-Alvarez A, Fernandez-Friera L, Mirelis JG, Sawit S, Nair A,
Kallman J, et al. Non-invasive estimation of pulmonary vascular resis-
tance with cardiac magnetic resonance. Eur Heart J 2011; 32: 2438
2445.
16. Abbas AE, Fortuin FD, Schiller NB, Appleton CP, Moreno CA, Lester
SJ. A simple method for noninvasive estimation of pulmonary vas-
cular resistance. J Am Coll Cardiol 2003; 41: 10211027.
17. Kang KW, Chang HJ, Kim YJ, Choi BW, Lee HS, Yang WI, et al.
Cardiac magnetic resonance imaging-derived pulmonary artery dis-
tensibility index correlates with pulmonary artery stiffness and pre-
dicts functional capacity in patients with pulmonary arterial hyper-
tension. Circ J 2011; 75: 22442251.
18. Yung GL, Fedullo PF, Kinninger K, Johnson W, Channick RN. Com-
parison of impedance cardiography to direct Fick and thermodilution
cardiac output determination in pulmonary arterial hypertension. Con-
gest Heart Fail 2004; 10: 710.
19. Raaijmakers E, Faes TJ, Scholten RJ, Goovaerts HG, Heethaar RM.
A meta-analysis of published studies concerning the validity of tho-
racic impedance cardiography. Ann NY Acad Sci 1999; 873: 121127.
20. Fisher MR, Forfia PR, Chamera E, Housten-Harris T, Champion HC,
Girgis RE, et al. Accuracy of Doppler echocardiography in the he-
modynamic assessment of pulmonary hypertension. Am J Respir Crit
Care Med 2009; 179: 615621.
21. Kouzu H, Nakatani S, Kyotani S, Kanzaki H, Nakanishi N, Kitakaze
M. Noninvasive estimation of pulmonary vascular resistance by Dop-
pler echocardiography in patients with pulmonary arterial hyperten-
sion. Am J Cardiol 2009; 103: 872876.
22. Selimovic N, Rundqvist B, Bergh CH, Andersson B, Petersson S,
Johansson L, et al. Assessment of pulmonary vascular resistance by
Doppler echocardiography in patients with pulmonary arterial hyper-
tension. J Heart Lung Transplant 2007; 26: 927934.
23. Tournadre JP, Chassard D, Muchada R. Overestimation of low car-
diac output measured by thermodilution. Br J Anaesth 1997; 79: 514
516.
24. Kauppinen PK, Koobi T, Hyttinen J, Malmivuo J. Segmental com-
position of whole-body impedance cardiogram estimated by com-
puter simulations and clinical experiments. Clin Physiol 2000; 20:
106113.
25. Taleb M, Khuder S, Tinkel J, Khouri SJ. The diagnostic accuracy of
Doppler echocardiography in assessment of pulmonary artery systolic
pressure: A meta-analysis. Echocardiography 2013; 30: 258265.
Supplementary Files
Supplementary File 1
FigureS1.Linear correlation between NI-CO and TD-CO (A) or
Fick-CO (B) in IPAH ( , red line), CTD-PAH ( , blue line) and
CTEPH ( , green line).
Please find supplementary file(s);
http://dx.doi.org/10.1253/circj.CJ-13-0172
 Perfusion
http://prf.sagepub.com/

Cardiac power index: staging heart failure for mechanical circulatory support
SG Hall, J Garcia, DF Larson and R Smith
Perfusion 2012 27: 456 originally published online 13 June 2012
DOI: 10.1177/0267659112450933

The online version of this article can be found at:

http://prf.sagepub.com/content/27/6/456

Published by:

http://www.sagepublications.com

Additional services and information for Perfusion can be found at:

Email Alerts: http://prf.sagepub.com/cgi/alerts

Subscriptions: http://prf.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

>> Version of Record - Nov 1, 2012

OnlineFirst Version of Record - Jun 13, 2012

What is This?

22 of 158. Downloaded from prf.sagepub.com at Tel Aviv University on September 8, 2013

7659112450933Hall SG et al.Perfusion

Perfusion

Cardiac power index: staging heart failure for 27(6) 456461

The Author(s) 2012

mechanical circulatory support

Reprints and permission: sagepub.
co.uk/journalsPermissions.nav
DOI: 10.1177/0267659112450933
prf.sagepub.com

SG Hall, J Garcia, DF Larson and R Smith

Abstract
Cardiac power output has been shown to quantify cardiac reserve. Cardiac reserve is defined as the difference between
basal and maximal cardiac performance. We compared cardiac power index to other commonly used hemodynamic
parameters to validate its usefulness to stage heart failure patients and determine the optimal time for implantation of
mechanical circulatory support. A retrospective study of twenty-eight heart failure patients implanted with mechanical
circulatory support was analyzed at three levels of drug therapy. Subjects were further separated into two categories:
survived versus deceased. Cardiac power index was the only statistically significant hemodynamic parameter that identified
cardiac reserve (p<0.05) in this patient population. These results showed that a cardiac power index at or below 0.34
Watts/m2 resulted in increased mortality rate, ninety days post-implantation.
Conclusion: Cardiac reserve was a determinant of post-device survival; therefore, these data suggest that device implanta-
tion should occur prior to the 0.34 Watts/m2 threshold.

Keywords
cardiac power output; cardiac reserve; mechanical circulatory support; hemodynamics; cardiac power index

Introduction
Cardiac power output is the hydraulic energy required by the
heart to provide enough blood flow to the systemic circula-
tion. Cardiac power output measured at exercise reflects
the peak cardiac performance attainable by the heart.
Comparing maximum cardiac power output with resting
cardiac power output represents the hearts cardiac reserve.
In heart failure patients, cardiac reserve is severely limited.1
Patients performing exercise tests or positive inotrope infu-
sion tests tend to show little difference in cardiac perfor-
mance due to poor cardiac reserve. It has been demonstrated
in several studies that heart failure patients with higher car-
diac reserves have better survival rates. As heart failure pro-
gresses in these patients, cardiac reserve continues to decline.2
Once cardiac performance can no longer be improved
through inotropic drug therapy, heart failure patients may
require mechanical circulatory support. Because end-organ
damage is a predictor of poor prognosis in patients who
receive mechanical circulatory support, it is important to
implant devices before heart failure has become so severe
that irreparable systemic damage has occurred. Symptomatic
assessment, drug tolerance, and myocardial oxygen con-
sumption are commonly used to determine the severity
of heart failure, but are unable to provide all the informa-
tion necessary to reflect true cardiac reserve. Myocardial
oxygen consumption provides an indirect measurement of
cardiac output; however, given that cardiac power output
incorporates cardiac output as well as systemic pressures,
this distinguishes it as a rational marker of cardiac perfor-
mance when compared to myocardial oxygen consump-
tion. Cardiac power output has been used to assess cardiac
performance and prognosis in cardiogenic shock and acute
and chronic heart failure patients.3,4 The purpose of this
study was to retrospectively analyze the application of car-
diac power output to stage the severity of heart failure in
patients who have received mechanical circulatory support.
Methods
Twenty-eight patients, twenty-four male and four female,
with a history of heart failure were referred to the
Circulatory Sciences Graduate Perfusion Program, The University of
Arizona, Tucson, Arizona, USA
Corresponding author:
Doug F. Larson
4402 Arizona Health Sciences Center
1501 North Campbell Avenue
Tucson, AZ 85724, USA
Email: dflarson@u.arizona.edu
Presented at the 33rd Annual Seminar of The American Academy of
Cardiovascular Perfusion, New Orleans, Louisiana, 26-29 January 2012

23 of 158. Downloaded from prf.sagepub.com at Tel Aviv University on September 8, 2013

 Hall SG et al. 457

University of Arizona Medical Center for mechanical cir- hemodynamic standards were tested to determine which
culatory support. Patients were admitted due to increased hemodynamic parameter gave a better representation of
heart failure symptoms or were medically transferred the remaining contractile reserve. All data are presented as
from outside institutions. The studied population con- mean SEM.
sisted of New York Heart Association (NYHA) classes III
and IV and American College of Cardiology (ACC-
AHA) stages C and D. Patient data was taken from July Results
2008 to April 2011. Approval of the data review was Of the twenty-eight patients studied, twenty-one sur-
granted by the University of Arizona Medical Center, vived and seven died within ninety days of receiving an
Tucson and the Institutional Committee on Human implanted device. Six of the seven patients died from
Research. end-organ failure and one died from systemic inflamma-
Patient data was collected at three levels of drug ther- tory response syndrome within six months of receiving
apy from right heart catheterizations, Swan-Ganz cathe- mechanical circulatory support. The twenty-one patients
ters, and/or echocardiography exams. Hemodynamic who survived ninety days from receiving mechanical cir-
data was initially collected when patients were first admit- culatory support were classified as destination therapy
ted to the intensive care unit and before they received ino- or bridge to transplant patients.
tropic drug therapy. Patients were then assessed when Baseline hemodynamic values, seen in Table 1, indi-
they were first placed on inotropic support. The initial cated abnormal values for all subjects, with the exception
inotropic dose was documented within four hours of of heart rate and mean arterial pressure. Baseline values
administration. The most common inotropic drugs pre- were recorded as patients were on standard heart failure
scribed included one or more of the following: dobu- medications (-blocker, diuretics, angiotensin-converting-
tamine, milrinone, and dopamine. Throughout the enzyme (ACE)-inhibitors, and angiotensin receptor
patients admission, medical management was adjusted blockers (ARBs)). The other hemodynamic parameters
based on declining heart function and/or drug tolerance. (ejection fraction, left ventricular stroke work index, car-
The third level of drug therapy assessed was considered diac output, cardiac index, and cardiac power) were all
the peak inotropic dose prescribed where current ino- below healthy adult ranges. Table 2 displays documented
tropic doses were increased or inotropic type was changed mean inotropic therapy while patients were in the inten-
before the patients became non-responsive to medication sive care unit.
and required mechanical circulatory support. Patients Heart rate and mean arterial pressure (Figure 1A &
were further subdivided into two group: survived versus 1B) were not found to be statistically significant between
deceased. Patients in the survived group survived at least the survived and deceased groups. Heart rate increased
ninety days after receiving mechanical circulatory sup-
port. The deceased group consisted of patients who died
within ninety days of receiving mechanical circulatory Table 1. Baseline patient characteristics
support. Survival was assessed at ninety days to account
for delayed organ failure due to prolonged hypoperfu- Characteristic (n=28) Datum
sion. Age (year) 53 2.5
Six hemodynamic parameters were compared to BSA (m2) 2.0 0.03
assess the remaining cardiac reserve in the twenty-eight HF Etiology
heart failure patients studied: heart rate(HR) beats per Ischemic 13 (46%)
minute (bpm); mean arterial pressure mmHg, [(2 x Non-Ischemic 15 (54%)
diastolic) + systolic]/3; stroke volume ml, (cardiac ICD 19 (68%)
output(CO)/HR); ejection Fraction - %, (stroke volume HR (BPM) 81 7.5
(SV)/end-diastolic volume (EDV)) x 100; cardiac index MAP (mmHg) 82 4.6
(CI) - L/min/m2, [(HR x SV)/body surface area (BSA)]; SVR (dyns/cm5) 1724 197
and cardiac power index (CPI) - W/m2, [(MAP x CO) x CVP (mmHg) 21 2.1
0.0022]/BSA. PCWP (mmHg) 29 1.5
LVSWI (g/m2/beat) 15 1.92
EF (%) 22 3.5
Statistics CO (L/min) 3.14 0.3

Statistical analysis was done using Data Analysis and
Statistical Software Stata 11 (StataCorp LP, College Station,
TX, USA). An independent two-sample t-test with unequal
variance was used to identify if contractile reserve still
existed in the failing heart. Mechanical circulatory support
Abbreviations. BSA: body surface area; ICD: implantable cardioverter de-
fibrillator; HF: heart failure; HR: heart rate; MAP: mean arterial pressure;
SVR: systemic vascular resistance; CVP: central venous pressure; PCWP:
pulmonary capillary wedge pressure; LVSWI: left ventricular stroke work
index; EF: ejection fraction; CO: cardiac output.Values are expressed as
mean SEM or number (% within studied population).

24 of 158. Downloaded from prf.sagepub.com at Tel Aviv University on September 8, 2013

 458 Perfusion 27(6)
Table 2. Administered inotropes
Positive Inotrope Initial Dose (mcg/kg/min)
Dobutamine 4.7
Milrinone 0.47
Dopamine 4.5
Inotropic doses are expressed as mean values for the two levels of inotropic drug therapies studied.
in both the survived and deceased groups when com-
pared across all three levels of drug therapy. Mean arterial
pressure of the deceased group declined despite ino-
tropic drug intervention. Those in the survived group
maintained a consistent level of mean arterial pressure
across all three levels of drug therapy.
Stroke volume was well below the healthy adult ranges
(Figure 1D), but was not found to be statistically signifi-
cant between the survived and deceased groups.
Stroke volume did not increase in the deceased group
after the initial inotropic drug therapy; however, when
peak inotropic drug therapy was administered, stroke
volume decreased. On the other hand, the survived
group displayed an increase in stroke volume at each
level of drug therapy.
Left ventricular ejection fraction (Figure 1C) derived
from echocardiography was not significantly different
between the survived or deceased groups. When
placed on the initial inotropic treatment, both the sur-
vived and deceased groups had an increased ejection
fraction; however, the deceased was less (31.50 7.5%
to 22.76 13.17 %). There was a large variance in the
deceased group due to a select few having diastolic
heart failure with preserved ejection fraction. Those with
diastolic heart failure had poor cardiac performance, but
because their chamber size was unchanged or even
decreased despite ventricular wall thickening, the overall
percentage of blood ejected from the heart with each
contraction was still elevated due to a decrease in overall
ventricular chamber size.
Cardiac output was used to calculate cardiac index in
order to account for the distribution of blood across the
body surface area. In Figure 1E, the survived and
deceased patients presented with a cardiac index of
1.65 0.83 and 1.52 0.13 L/min/m2, respectively. With
the initial inotropic treatment, the survived groups
cardiac index increased to 2.14 0.54 L/min/m2, but
remained constant with the peak inotropic treatment at
2.17 0.47 L/min/m2. The deceased group displayed a
moderate increase in cardiac index during the initial ino-
tropic treatment (1.86 0.62 L/min/m2), however, this
decreased on peak inotropic support (1.69 0.26 L/min/
m2). Cardiac index was not a functional hemodynamic
marker to identify if contractile reserve remained or
whether it was able to detect survival outcomes between
the survived and deceased groups.
25 of 158. Downloaded from prf.sagepub.com at Tel Aviv University on September 8, 2013
Peak Dose (mcg/kg/min) P-value
6.0 >0.05
3.23 >0.05
3.78 >0.05
Cardiac power output was used to calculate cardiac
power index to adjust for blood volume distribution
across the body surface area. The baseline administered
inotropic dose increased cardiac power index, but
revealed no statistical difference when compared to the
deceased group (Figure 1F). When peak inotropic doses
were administered, cardiac power index continued to
decline in the deceased group, but was able to distin-
guish the contractile reserve between the two groups.
The mean cardiac power index of patients who survived
was 0.34 0.07 W/m2 at peak inotropic drug therapy. A
cardiac power index below 0.34 W/m2 identified patients
who did not survive ninety days past the implantation
of mechanical circulatory support because all patients
in the deceased group had a cardiac power index below
0.34 W/m2. Comparing cardiac power index between
the survived and deceased groups showed significantly
increased mortality rates once values fell below this
point (p<0.05).
Discussion
Current hemodynamic guidelines use left ventricular
ejection fraction and myocardial oxygen consumption as
criteria to assess cardiac and systemic degeneration for
possible mechanical circulatory support. As shown by
Marmor et al., the present study demonstrates that ejec-
tion fraction did not display any significant differences
between the survived and deceased groups during
maximal cardiac performance resulting from peak ino-
tropic drug therapy.9 In addition, results from prior
myocardial oxygen consumption studies on patients in
end-stage heart failure have shown to be inconclusive
when peak cardiac function was measured.3,10,11 Patients
with NYHA classes III and IV symptoms are severely
compromised when tested physically for myocardial oxy-
gen consumption due to progressive muscle decon-
ditioning, poor vascular circulation, compromised lung
capacity, and anemia. Although ejection fraction and
myocardial oxygen consumption tests have their respec-
tive roles in measuring specific parameters of heart func-
tion, they may not be completely reliable assessments of
cardiac function in these severely compromised patients.
Right heart catheterization provides very useful infor-
mation in gathering pressure and volume changes in spe-
cific areas of the heart. This method requires that patients
 Hall SG et al. 459

Figure 1. Hemodynamic parameters of the survived and deceased.

*Normal hemodynamic values are from reference numbers 5-8. Values are expressed as mean standard error of the mean. *p=0.009.

be placed in a supine position without simulating peak
cardiac contractility. This is a significant limitation, as
assessing heart function in this manner does not allow for
measuring the difference between basal and maximal car-
diac performance in order to determine true cardiac
reserve. Resting values alone do not provide a sufficient
amount of information on cardiac performance in healthy
versus diseased hearts; however, positive inotropic stimu-
lation that achieves peak contractility is when diseased
hearts can become distinguishable from healthy hearts.

26 of 158. Downloaded from prf.sagepub.com at Tel Aviv University on September 8, 2013

 460 Perfusion 27(6)
With the limited physical and physiological abilities of
this population, an effective method to assess true cardiac
reserve is through the use of stress tests using inotropic
drug infusion. This method has proven to be equivalent
to results produced by maximal exercise tests.9,12 In
severely compromised patients, this may be the only
practical method to use in order to determine existing
contractile energy. The current study used this concept by
gathering data while patients were placed on initial ino-
tropic support in order to stimulate the remaining con-
tractile reserve, if any existed. As heart failure progressed
in these patients and inotropic support was increased or
altered, therapeutic saturation was met, eventually, due to
the heart already functioning at its maximum contractile
capacity. This indicated a lack of existing reserve in these
patients, due to their basal cardiac performance equaling
their maximum.
Cardiac power index was the only hemodynamic
parameter that distinguished between the cardiac reserve
of patients who survived ninety days post-implant versus
those who died. Although invasive inotropic stimulation
may not be the first choice to assess cardiac contractile
reserve, measuring peak cardiac power index invasively
may provide information on whether pharmacological
support is still efficacious and if any contractile reserve
still remains. Non-invasive methods to evaluate cardiac
power output have been used for over five years and have
been shown to effectively determine the severity of heart
failure.3,8,10,13-16 With the availability of non-invasive
equipment utilizing an inert gas re-breathing technique,
patients in earlier stages of heart failure can be readily
assessed for existing cardiac reserve.17 Cardiac power
index can be a useful prognostic tool in order to deter-
mine if cardiac reserve exists in these patients and when
to intervene with pharmacological support or mechanical
circulatory support before the systemic effects of heart
failure become irreparable.
Limitations
Because this was a single-center study, our data were
non-randomized and our cohort of patients was rela-
tively small. Due to the novelty of the population, hav-
ing a large number of subjects is a difficult limitation to
overcome when performing studies on patients receiv-
ing mechanical circulatory support. The variation in
medical record charting between patients also limited
the number of hemodynamic values that could be eval-
uated. Subject data were collected either by right heart
catheterizations, Swan-Ganz catheters, and/or echocar-
diography exams, but these methods were inconsistent
across the sample population. A future prospective
study would be beneficial to reduce the lack of consis-
tency between available subject data and their collection
methods.
27 of 158. Downloaded from prf.sagepub.com at Tel Aviv University on September 8, 2013
Conclusions
In conclusion, our data suggest that attaining cardiac
index and mean arterial pressure, either invasively or
non-invasively, and applying these values to the cardiac
power index formula, can determine the severity of heart
failure based on the hearts maximal contractile reserve.
Peak cardiac power index values may also be a useful tool
to determine the appropriate therapeutic approach nec-
essary for these patients. Cardiac power index has proven
to recognize remaining contractile reserve in heart fail-
ure patients in comparison to other commonly used
hemodynamic parameters. In heart failure patients cate-
gorized as NYHA classes III and IV or ACC-AHA stages
C and D, a peak cardiac power index below 0.34 W/m2,
with or without the use of drug therapy should immedi-
ately be considered for implantation of mechanical cir-
culatory support. Patients at or below a cardiac power
index of 0.34 W/m2 have an increased incidence of mor-
tality.
Acknowledgement
This study was supported the Steinbronn Heart Failure Research
Award to DFL.
Conflict of Interest Statement
None declared.
References:
1. Williams SG, Barker D, Goldspink DF, Tan LB. A reap-
praisal of concepts in heart failure: central role of cardiac
power reserve. Arch Med Sci 2005; 1: 6574.
2. Cotter G, Williams SG, Z Vered Z, Tan LB. Role of car-
diac power in heart failure. Curr Opin Cardiol 2003; 18:
215222.
3. Lang CC, Karlin P, Haythe J, Lim TK, Mancini DM. Peak
cardiac power output, measured noninvasively is a pow-
erful predictor of outcome in chronic heart failure. Circ
Heart Fail 2009; 2: 3338.
4. Cotter G, Moshkovitz Y, Kaluski E, et al. The role of car-
diac power and systemic vascular resistance in the patho-
physiology and diagnosis of patients with acute congestive
heart failure. Eur J Heart Fail 2003; 5: 443451.
5. Pfisterer ME, Battler A, Zaret BL. Range of normal values
for left and right ventricular ejection fraction at rest and
during exercise assessed by radionuclide angiocardiogra-
phy. Eur Heart J 1985; 6: 647655.
6. Mayer SA, Lin J, Homma S, et al. Myocardial injury and
left ventricular performance after subarachnoid hemor-
rhage. Stroke 1999; 30: 780786.
7. Parry, Andrew. Inotropic drugs and their uses in critical
care. Nurs Crit Care 2012; 17: 1927.
8. Bromley PD, Hodges LD, Brodie DA. Physiological range
of peak cardiac power output in healthy adults. Clin
Physiol Funct Imaging 2006; 26: 240246.
9. Marmor A, Jain D, Cohen LS, Nevo E, Wackers FJ,
Zaret BL. Left ventricular peak power during exercise:
 Hall SG et al. 461
a noninvasive approach for assessment of contractile reserve.
J Nucl Med 1993; 34: 18771885.
10. Leitman M, Sucher E, Kaluski E, et al. Non-invasive meas-
urement of cardiac output by whole-body bio-impedance
during dobutamine stress echocardiography: clinical
implications in patients with left ventricular dysfunction
and ischaemia. Eur J Heart Fail 2006; 8: 136140.
11. Wilson JR, Rayos G, Yeoh TK, Gothard P, Bak K.
Dissociation between exertional symptoms and circula-
tory function in patients with heart failure. Circulation
1995; 92: 4753.
12. Marmor A, Schneeweiss A. Prognostic value of nonin-
vasively obtained left ventricular contractile reserve in
patients with severe heart failure. J Am Coll Cardiol 1997;
29: 422428.
13. Jakovljevic DG, George RS, Donovan G, et al. Comparison
of cardiac power output and exercise performance in
patients with left ventricular assist devices, explanted
28 of 158. Downloaded from prf.sagepub.com at Tel Aviv University on September 8, 2013
(recovered) patients, and those with moderate to severe
heart failure. Am J Cardiol 2010; 105: 17801785.
14. Jakovljevic DG, Donovan G, Nunan D, et al. The effect of
aerobic versus resistance exercise training on peak car-
diac power output and physical functional capacity in
patients with chronic heart failure. Int J Cardiol 2010; 145:
526528.
15. Shelton RJ, Ingle L, Rigby AS, Witte KK, Cleland JG, Clark
AL. Cardiac output does not limit submaximal exercise
capacity in patients with chronic heart failure. Eur J Heart
Fail 2010; 12: 983989.
16. Nicholls D, O'Dochartaigh C, Riley M. Circulatory power-
-a new perspective on an old friend. Eur Heart J 2002; 23:
12421245.
17. Agostoni P, Cattadori G, Apostolo A, et al. Noninvasive
measurement of cardiac output during exercise by inert
gas rebreathing technique: a new tool for heart failure
evaluation. J Am Coll Cardiol 2005; 46: 17791781.
 Non invasive measurements of Cardiac Output (CO) and Cardiac Power
Index (CPI) by whole body bio impedance in patients with heart failure. A
report from SICA- HF study (FP7/2007-2013/241558)
1 1 1 1 1 1 1 1
P Pellicori , E Wright , P Costanzo , S Smith , S Rimmer , J Hobkirk , A Torabi , T Mabote , J
1 1 1
Warden , JGF Cleland , University of Hull, Department of Academic Cardiology - Hull - United
Kingdom,

Objectives: Haemodynamic dysfunction is often used as part of the definition for heart
failure (HF), predicts an adverse outcome and could be an important target for therapy
but is rarely measured in routine practice, perhaps because simple, effective, inexpensive
technology is lacking. We assessed the ability of whole body electrical bioimpedance to
measure haemodynamics non-invasively.
Methods: Patients and controls enrolled in the Studies Investigating Co-morbidities
Aggravating Heart Failure (SICA) were included in this analysis if in sinus rhythm. Stroke
volume (SV), cardiac output (CO), cardiac power index (CPI) and total body water (TBW)
were measured non-invasively using whole body bio-impedance (NICaS) after two
minutes rest in the supine position. Results were compared amongst HF patients
according to tercile of amino-terminal pro-brain natriuretic peptide (NT-proBNP) and left
ventricular ejection fraction (LVEF).
Results: The median age of the 51 patients with HF was 71 years (IQR:63--77), 15 were
women (29%), median LVEF was 36% (IQR:29-44) and median body mass index (BMI)
2
was 30kg/m (IQR:26-33). The median age of 15 control subjects was 66 years (IQR:56-
2
75), 4 were women (27%) and median BMI was 29kg/m (IQR:26-37). As expected,
compared with controls, patients with HF had higher plasma NTproBNP, worse renal
2 2
function, lower CPI (0.640.25w/m vs. 0.470.18w/m ;p< 0.01) but TBW was similar
(479% vs. 468%;p=0.74).
Patients were divided into terciles of NTproBNP (lower and upper limits of the mid-tercile
were 373 and 1063ng/L). Patients in the highest tercile of NTproBNP had lower BMI
2 2 2
(325kg/m vs. 295kg/m vs.284kg/m respectively ; p=0.02), LVEF (425% vs. 379%
2 2 2
vs. 298%; p<0.01), CPI (0.520.2w/m vs. 0.500.2w/m vs. 0.380.1w/m ; p=0.04) and
poorer renal function (Creatinine: 9128mol/l vs. 11335mol/l vs. 13047mol/l;
p=0.02) and increased TBW (416% vs. 477% vs. 489%;p=0.02). No differences in
CO or SV were found. In the lowest tercile of LVEF, SV (8324ml vs. 6723ml
vs.589ml;p<0.01), CO (5.71.9l/min vs. 4.31.6l/min vs. 3.90.8l/min; p <0.01) and CPI
2 2 2
(0.580.2W/m vs. 0.45 0.2W/m vs. 0.380.1W/m ; p<0.01) were all significantly lower,
but TBW was similar across terciles (436% vs. 467% vs. 489%; p= 0.28).
Conclusion: In patients with HF and sinus rhythm, whole body bio-impedance might be
a useful method of monitoring the haemodynamic severity of heart failure that is quick,
simple and inexpensive. Whether it is as or more useful than NTproBNP as a marker of
outcome awaits the results of large long-term studies. (European Journal of Heart Failure
Supplements ( 2012 ) 11 ( S1 ), S91)

29 of 158.
 IJCA-13355; No of Pages 3
International Journal of Cardiology xxx (2011) xxxxxx

Contents lists available at ScienceDirect

International Journal of Cardiology

j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / i j c a r d

Letter to the Editor

Detection of left ventricular systolic dysfunction using a newly developed, laptop

based, impedance cardiographic index
Yoseph Rozenman a,,1, Renee Rotzak a, Robert P. Patterson b,1
a
Heart Institute, E. Wolfson Medical Center, Holon, Israel, Afliated to the Sackler Faculty of Medicine, Tel-Aviv University, Israel
b
Medical School, University of Minnesota, USA

a r t i c l e i n f o The regional ICG signal is obtained from 2 pairs of tetra-polar

Article history:
Received 8 February 2011
Accepted 10 February 2011
Available online xxxx
Keywords:
Impedance Cardiography
Congestive heart failure
Systolic dysfunction
Screening
Laptop
Congestive heart failure (CHF) is a major cause of morbidity and
mortality worldwide [1]. The development of left ventricular systolic
dysfunction (LVSD) is a marker of poor prognosis. Mild reduction in EF
progresses with time and when EF gets to b40%, most patients
develop CHF and their prognosis is dismal [2,3]. Simple, cheap
techniques that will reliably detect mild reduction of the EF, typical for
ALVSD, will thus be of enormous value to reduce the occurrence of
CHF and cardiac mortality [37].
Impedance Cardiography (ICG) is a noninvasive method of
determining hemodynamic status and reliable estimates of myocar-
dial contractility (and EF) can be obtained using indices based on
systolic time intervals [810]. We have recently developed an index
Granov Goor Index (GGI)that combines time interval and imped-
ance parameters in order to identify subjects with LVSD. The GGI is
obtained from a regional ICG signalmeasured using wrist and ankle
electrodes [1013].
In this manuscript we determined the accuracy of GGI in
identifying subjects with LVSD (in a training cohort of 100
individuals) and validated the ndings in an additional cohort of
201 subjects.
Corresponding author at: Cardiology Department, E. Wolfson Medical Center, POB
5, Holon, 58100, Israel. Tel.: +972 3 5028404; fax: +972 3 5028130.
E-mail address: yosephr@post.tau.ac.il (Y. Rozenman).
1
Consultants of NI Medical Ltd.
electrodes: one on the wrist, above the radial artery, and the other on
the contra-lateral ankle above the posterior tibial artery. The detailed
description of the system and measured parameters can be reviewed
in previous publications [1013]. The GGI is designed to assess the
systolic contractile function of the LV and is obtained from the
following formula:
GGI R=R   HRcorrected
where R is the basal impedance and R is the change in impedance,
is the time to peak R and HR is the heart rate. The GGI takes into
account both R as an estimate of stroke volume and as a time
interval parameter. HR and R are used as normalization factors so that
the result is a dimensionless index that reects the systolic function of
the LV.
The entire system (electrode, signal analysis and calculations) can
replace the CD ROM in a regular computer, transforming any laptop
into a simple diagnostic deviceideal for screening.
Echocardiography was performed using GE Vivid 7 (GE Vingmed
Ultrasound A/S, Norway). EF was calculated using the biplane
Simpson's technique [14]. Echocardiography and ICG were performed
on the same day.
Patients who were referred for screening echocardiography at the
outpatient clinic of the E. Wolfson Medical Center were considered for
this trial.
The study protocol conforms to the ethical guidelines of the 1975
Declaration of Helsinki and was approved by the institution's human
research committee.
Optimal cutoff value for the GGI for detection of LVSD was
determined from the Receiver Operating Characteristic (ROC) curve of
the training set. Sensitivity specicity and predictive accuracy of GGI
to detect LVSD were determined using standard technique. Statistical
analysis was performed using MedCalc software (Broekstraat,
Belgium version 9.5.2.0).
Tables 1 and 2 describe the demographic echocardiographic and
ICG results. The superiority of GGI in detecting LVSD is conrmed
using ROC curve analysis (Fig. 1). The area under curve (AUC) of ROC
curves of R/R and as predictors of LVSD were 0.84 and 0.75
respectively, signicantly lower (p b 0.001) than the value for GGI
(0.97 CI 0.92 to 0.99). For a GGI cutoff value of 10 the sensitivity,
specicity positive and negative predictive values are 86%, 100%, 100%
and 96% respectively.

0167-5273/$ see front matter 2011 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.ijcard.2011.02.029

Please cite this article as: Rozenman Y, et al, Detection of left ventricular systolic dysfunction using a newly developed, laptop based,
30 impedance
of 158. cardiographic index, Int J Cardiol (2011), doi:10.1016/j.ijcard.2011.02.029
2 Y. Rozenman et al. / International Journal of Cardiology xxx (2011) xxxxxx

Patients in the validation set were slightly older with a tendency Table 2
for lower prevalence of smoking. Echocardiographic and ICG values Echocardiographic ndings of subjects in the training and validation cohorts.

were similar in these two sets of patients (Tables 1 and 2). A GGI cutoff All Training cohort Validation cohort p
of 10 was 89% sensitive and 96% specic for detection of LVSD,
Echocardiography
conrming the excellent results of the training set. Positive and LVEF (%) 56 9 55 9.6 57 9.2 NS
negative predictive values were 78% and 98% respectively. LVEDD (cm) 4.8 0.6 4.9 0.6 4.8 0.6 NS
False positives and false negatives of GGI as a predictor of EF are LVESD (cm) 3.0 0.8 2.9 0.7 3.0 0.8 NS
IVS (cm) 1.0 0.1 1.0 0.16 1.0 0.16 NS
extremely rare (Fig. 2) reecting the accuracy of GGI to detect LVSD.
LVPW (cm) 1.0 0.1 1.0 0.1 1.0 0.8 NS
Careful observation of the data points suggests that rather than a LA diameter (cm) 3.8 0.6 3.8 0.55 3.8 0.6 NS
simple correlation, the relation between EF and GGI is more like a step MR (%) 27 27 26 NS
function with a GGI b 10 corresponding to a wide range of EF b 55% LVH (%) 22 26 19 NS
while a GGI 10 corresponds to varying levels of EF 55%. LA area (cm2) 18.9 5 18.5 4.9 19.0 5.0 NS
Diastolic dysfunction (%) 19 14 20 NS
In this study we were able to demonstrate that a simple regional
ICG
impedance based technology can reliably detect LVSD. Regional (as Cardiac index (l/m2) 3.5 0.8 3.5 0.9 3.6 0.75 NS
compared to thoracic) ICG enables more accurate determination of Stroke index (cc/m2) 50 9.3 50 9.5 49 9.2 NS
stroke volume and cardiac output [1012]. The addition of , a time TPR(dynes-sec-cm5) 1299341 1280 350 1308 337 NS
GGI 11.6 2.3 11.6 2.6 11.6 2.1 NS
interval parameter, with the resulting GGI adds signicantly to
GGI b10 (%) 17 18 16 NS
accuracy of this method to identify subjects with LVSD (increasing
the AUC from 0.84 to 0.97). Most importantly, also those at the LVEDD and LVESD are left ventricular end diastolic and end systolic diameters. IVS and
LVPW are LV septal and posterior wall thicknesses. MRmitral regurgitation (minimal
asymptomatic stage of mild LVSD are accurately detected. or mild), LVHleft ventricular hypertrophy, LAleft atrium and TPRtotal peripheral
Some of the energy of the LV contraction produces forward blood resistance.
ow during systole, while a signicant amount is briey stored as
potential energy in the distended arteriesmaintaining the forward
ow during diastole [15]. The information provided by relates to
this stored energy while the R is related to the stroke volume. The
incorporation of both these parameters into the GGI makes the GGI a Even though we excluded patients with symptoms of suggestive of
more reliable index of the energy generated by the contraction of the heart disease the decision to refer patients for echocardiography
LV. Time interval parameters were used in the past from the ICG curve biases the population so that data was not collected from the true
(or its derivative) to try to estimate the EF. Despite signicant target populationasymptomatic individuals with no prior history of
correlation [9] the coefcient values (typically approximately 0.5) are heart disease. However the likelihood that this will change the
not high enough for clinical purpose. Thus accurate estimation of the observed relation between GGI and EF is small.
value of EF is limited but, as we have showed, the GGI very accurately The GGI can accurately detect LVSD in subjects at risk (referred for
determines whether the LVEF is normal or abnormal (step function echocardiography). The GGI can be easily obtained from regional ICG
rather than simple correlation). signal using any laptop by replacing the CD ROM with the measuring
The natural history of patients with ALVSD is dismal [4]. Compared hardware and software. This laptop based device is cheap and easy to
with those with normal LV, patients with ALVSD had 45-fold operate making it an attractive tool to screen for LVSD. Additional
increase risk of CHF and death. Early detection of ALVSD requires a trials are required to prove the efcacy of this device in the target
search for CAD as a possible etiology. When CAD is causing reversible population of asymptomatic patients at risk.
dysfunction prompt revascularization is essential to improve out-
come; preventing CHF and early arrhythmic death [1619]. In those
with irreversible ALVSD; further deterioration can be attenuated by
the use of appropriate drug therapy [3].
Since early intervention in subjects with ALVSD is benecial,
screening programs need to be implemented to identify these
individuals. Echocardiographythe gold standard in the assessment
of LVSDis impractical for general screening since it is very expensive.
Thus an effective screening program should use an inexpensive
diagnostic test (such as GGI) to identify those who are likely to have
ALVSD.

Table 1
Clinical characteristics of subjects in the training and validation cohorts.

All Training Validation p

cohort cohort

Number 301 100 201

Male (%) 56 65 52 NS
Age (mean SD) 63 9 60 10 64 11 P = 0.0027
BMI (kg/m2), (mean SD) 28.2 3.5 28 4.5 28 4.4 NS
Hypertension (%) 44 41 44 NS
Diabetes (%) 20 27 17 NS
Dyslipidemia (%) 51 57 48 NS
Smoke (%) 20 28 16 P b 0.055
Family history of 52 58 49 NS
CAD (%)
Asymptomatic CAD (%) 29 28 29 NS
Fig. 1. Receiver operating characteristic (ROC) curve of R/R, HR and GGI as
BMIbody mass index. diagnostic tests for LVSD (EF b 55%).

Please cite this article as: Rozenman Y, et al, Detection of left ventricular systolic dysfunction using a newly developed, laptop based,
31 of 158.cardiographic index, Int J Cardiol (2011), doi:10.1016/j.ijcard.2011.02.029
impedance
 Y. Rozenman et al. / International Journal of Cardiology xxx (2011) xxxxxx
Fig. 2. Scatter plot of EF vs. GGI demonstrating the diagnostic power GGI (with cutoff value of 10) to detect LVSD (EF b 55%) (data from the entire cohort, n = 301).
Acknowledgement
This study was supported by NI Medical Ltd. Israel.
The authors of this manuscript have certied that they comply
with the Principles of Ethical Publishing in the International Journal of
Cardiology [20].
References
[1] Rosamond W, Flegal K, Friday G, et al. Heart disease and stroke statistics: 2007
update: a report from the American Heart Association Statistics Committee and
Stroke Statistics Subcommittee. Circulation 2007;115:e69e171.
[2] Goldberg LR, Jessup M. Stage B heart failure: management of asymptomatic left
ventricular systolic dysfunction. Circulation 2006;113:285160.
[3] Schocken DD, Benjamin EJ, Fonarow GC, et al. Prevention of heart failure: a
scientic statement from the American Heart Association Councils on Epidemi-
ology and Prevention, Clinical Cardiology, Cardiovascular Nursing, and High Blood
Pressure Research; Quality of Care and Outcomes Research Interdisciplinary
Working Group; and Functional Genomics and Translational Biology Interdisci-
plinary Working Group. Circulation 2008;117:254465.
[4] Wang TJ, Evans JC, Benjamin EJ, Levy D, LeRoy EC, Vasan RS. Natural history of
asymptomatic left ventricular systolic dysfunction in the community. Circulation
2003;108:97782.
[5] Redeld MM, Rodeheffer RJ, Jacobsen SJ, Mahoney DW, Bailey KR, Burnett Jr JC.
Plasma brain natriuretic peptide to detect preclinical ventricular systolic or
diastolic dysfunction: a community-based study. Circulation 2004;109:317681.
[6] McMurray JV, McDonagh TA, Davie AP, et al. Should we screen for asymptomatic
left ventricular dysfunction to prevent heart failure? Eur Heart J 1998;19:8426.
[7] Galasko GI, Barnes SC, Collinson P, Lahiri A, Senior R. What is the most cost-
effective strategy to screen for left ventricular systolic dysfunction: natriuretic
peptides, the electrocardiogram, hand-held echocardiography, traditional echo-
cardiography, or their combination? Euro Heart J 2006;27:193200.
[8] Cotter G, Moshkovitz Y, Kaluski E, et al. Accurate, noninvasive continuous
monitoring of cardiac output by whole-body electrical bioimpedance. Chest
2004;125:143140.
Please cite this article as: Rozenman Y, et al, Detection of left ventricular systolic dysfunction using a newly developed, laptop based,
32 impedance
of 158. cardiographic index, Int J Cardiol (2011), doi:10.1016/j.ijcard.2011.02.029
3
[9] Feng S, Okuda N, Fujinami T, Takada K, Nakano S, Ohte N. Detection of impaired left
ventricular function in coronary artery disease with acceleration index in the rst
derivative of the transthoracic impedance change. Clin Cardiol 1988;11:8437.
[10] Cotter G, Schachner A, Sasson L, Dekel H, Moshkovitz Y. Impedance cardiography
revisited. Physiol Measures 2006;27:81727.
[11] Paredes OL, Shite J, Shinke T, et al. Impedance cardiography for cardiac output
estimation. Reliability of wrist-to-ankle electrode conguration. Circ J 2006;70:
11648.
[12] Goor DA, Frinerman E, Granov I, Patterson R. Peripheral versus thoracic impedance
cardiography. J Cardiac Failure, Heart Failure Society of America (HFSA), 10th
Annual Scientic Meeting, September, 1013; 2006. p. S545.
[13] Rozenman Y, Goor DA, Rotzak R. Progress report on non-invasive diagnosis of
asymptomatic left ventricular systolic dysfunction. Abstract J Card Fail 2009;15:
68.
[14] Gottdiener JS, Bednarz J, Devereux RD, et al. American Society of Echocardiogra-
phy: recommendations for use of echocardiography in clinical trials, a report from
the American Society of Echocardiography's Nomenclature and Standards
Committee and the task force on echocardiography in clinical trials. J Am Soc
Echocardiogr 2004;17:1086119.
[15] Hardy CJ. Assessment of arterial elasticity by cardiovascular MRI. In: Kwong RY,
Libby P, editors. Cardiovascular magnetic resonance imaging (contemporary
cardiology). Totowa, New Jersey: Humana Press; 2008. p. 695710.
[16] Gheorghiade M, Sopko G, De Luca, et al. Navigating the crossroads of coronary
artery disease and heart failure. Circulation 2006;114:120213.
[17] Pagano D, Lewis ME, Townend JN, Davies P, Camici PG, Bonser RS. Coronary
revascularization for postischaemic heart failure: how myocardial viability affects
survival. Heart 1999;82:6848.
[18] Canty Jr JM, Suzuki G, Banas MD, Verheyen F, Borgers M, Fallavollita JA.
Hibernating myocardium. Chronically adapted to ischemia but vulnerable to
sudden death. Circ Res 2004;94:11429.
[19] Allman KC, Shaw LJ, Hachamovitch R, Udelson JE. Myocardial viability testing and
impact of revascularization on prognosis in patients with coronary artery disease
and left ventricular dysfunction: a meta-analysis. J Am Coll Cardiol 2002;39:
11518.
[20] Shewan LG, Coats AJ. Ethics in the authorship and publishing of scientic articles.
Int J Cardiol 2010;144:12.
 Measurement Precision in the Optimization of Cardiac Resynchronization Therapy
Robert G. Turcott, Ronald M. Witteles, Paul J. Wang, Randall H. Vagelos, Michael B. Fowler
and Euan A. Ashley

Circ Heart Fail. 2010;3:395-404; originally published online February 22, 2010;
doi: 10.1161/CIRCHEARTFAILURE.109.900076
Circulation: Heart Failure is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX
75231
Copyright 2010 American Heart Association, Inc. All rights reserved.
Print ISSN: 1941-3289. Online ISSN: 1941-3297

The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circheartfailure.ahajournals.org/content/3/3/395

Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Circulation: Heart Failure can be obtained via RightsLink, a service of the Copyright Clearance Center, not
the Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further information
about this process is available in the Permissions and Rights Question and Answer document.

Reprints: Information about reprints can be found online at:

http://www.lww.com/reprints

Subscriptions: Information about subscribing to Circulation: Heart Failure is online at:

http://circheartfailure.ahajournals.org//subscriptions/

Downloaded from http://circheartfailure.ahajournals.org/ by guest on December 9, 2013

33 of 158.
 Measurement Precision in the Optimization of Cardiac
Resynchronization Therapy
Robert G. Turcott, MD, PhD; Ronald M. Witteles, MD, FACC;
Paul J. Wang, MD; Randall H. Vagelos, MD, FACC;
Michael B. Fowler, MB, FRCP, FACC; Euan A. Ashley, MRCP, DPhil
BackgroundCardiac resynchronization therapy improves morbidity and mortality in appropriately selected patients.
Whether atrioventricular (AV) and interventricular (VV) pacing interval optimization confers further clinical
improvement remains unclear. A variety of techniques are used to estimate optimum AV/VV intervals; however, the
precision of their estimates and the ramifications of an imprecise estimate have not been characterized previously.
Methods and ResultsAn objective methodology for quantifying the precision of estimated optimum AV/VV intervals was
developed, allowing physiologic effects to be distinguished from measurement variability. Optimization using multiple
conventional techniques was conducted in individual sessions with 20 patients. Measures of stroke volume and dyssynchrony
were obtained using impedance cardiography and echocardiographic methods, specifically, aortic velocity-time integral,
mitral velocity-time integral, A-wave truncation, and septal-posterior wall motion delay. Echocardiographic methods yielded
statistically insignificant data in the majority of patients (62% 82%). In contrast, impedance cardiography yielded statistically
significant results in 84% and 75% of patients for AV and VV interval optimization, respectively. Individual cases
demonstrated that accepting a plausible but statistically insignificant estimated optimum AV or VV interval can result in worse
cardiac function than default values.
ConclusionsConsideration of statistical significance is critical for validating clinical optimization data in individual
patients and for comparing competing optimization techniques. Accepting an estimated optimum without knowledge of
its precision can result in worse cardiac function than default settings and a misinterpretation of observed changes over
time. In this study, only impedance cardiography yielded statistically significant AV and VV interval optimization data
in the majority of patients. (Circ Heart Fail. 2010;3:395-404.)
Key Words: cardiac resynchronization therapy pacing optimization AV delay interventricular interval
echocardiography impedance cardiography

C ardiac resynchronization therapy (CRT) decreases morbid-

ity and mortality in populations with reduced cardiac
function and conduction abnormalities13; however, approxi-
mately one third of these patients do not experience benefit.4
Atrioventricular (AV) and interventricular (VV) pacing interval
optimization improve hemodynamics acutely57 and may en-
hance the response to CRT relative to default interval settings
among both the traditional responder and nonresponder groups.
To date, however, results of prospective randomized trials have
been mixed.8 12
Clinical Perspective on p 404
The clarification of fundamental issues in pacing interval
optimization is still at an early stage. For example, the extent
to which optimum AV/VV intervals change with body
position and exertion remains unclear,1316 and data assessing
the stability of optima over time have been inconsistent.17,18
Even the definitions of pacing intervals vary among manu-
facturers, with identical programmed intervals corresponding
to very different ventricular pace timing.19 Perhaps the most
important fundamental weakness is the lack of statistical tools
to quantitatively evaluate the significance of the measured
data and to characterize the precision of the estimated
optimum AV/VV interval.11,20 22
In this study, we examined multiple commonly used
AV/VV interval optimization techniques. Central to our
approach is the recognition that the underlying dependence of
cardiac function on AV/VV interval is obscured to some
degree by measurement noise and that the optimum interval
identified by a given technique is in fact an estimation of the
true physiologic optimum. The degree to which measured
optimization data demonstrate a significant dependence on
pacing interval and the precision of estimated optima were
rigorously evaluated using new statistical tools based on

Received August 6, 2009; accepted January 20, 2010.

From the Division of Cardiovascular Medicine and Center for Biomedical Informatics Research, Stanford University School of Medicine, Stanford,
Calif.
Correspondence to Euan A. Ashley, MRCP, DPhil, Cardiomyopathy Center, Stanford University School of Medicine, Division of Cardiovascular
Medicine, Falk CVRC, 300 Pasteur Dr, MC5406, Stanford, CA 94305-5406. E-mail euan@stanford.edu
2010 American Heart Association, Inc.
Circ Heart Fail is available at http://circheartfailure.ahajournals.org DOI: 10.1161/CIRCHEARTFAILURE.109.900076

395
Downloaded from http://circheartfailure.ahajournals.org/ by guest on December 9, 2013
34 of 158.
 396 Circ Heart Fail May 2010

Table 1. Patient Demographics with aortic flow in the apical 5-chamber view (Figure 1B).20,21
For this and the other echocardiographic techniques, a 10- to
Patient No. Age, y Sex EF Etiology 20-second equilibration period followed each programming
1 80 M 31 Ischemic change. Data then were recorded over 1 to 2 respiratory cycles,
with premature and postpremature beats excluded.
2 73 M 30 Ischemic 3. Mitral inflow velocity-time integral (M-VTI), which is directly
3 70 M 28 Ischemic proportional to inflow volume, was obtained using pulsed-
4 66 M 29 Idiopathic wave Doppler with the sample volume placed just apical to the
mitral leaflets in the apical 4-chamber view (Figure 1C).20,21
5 62 M 42 Toxic 4. In contrast to the techniques described above, which attempt to
6 57 F 65 HCM* characterize cardiac function by estimates of forward flow,
7 55 F 70 HCM* septal-posterior wall-motion delay (SPWMD) provides an
assessment of left ventricular (LV) mechanical synchrony.20
8 80 M 26 Ischemic, AS, MR As shown in Figure 1D, the transducer signal was directed
9 45 F 20 Toxic across the septum and the posterior LV wall in the parasternal
long-axis view. Color M-mode Doppler was used to highlight
10 71 M 25 MR
the relative motion of the 2 walls, with minimum lag taken to
11 54 M 24 AS represent optimal ventricular synchrony.
12 57 F 23 MR 5. A-wave truncation identifies the optimum AV delay as the
shortest pacing interval that avoids truncation of the A-wave21
13 54 M 27 Idiopathic
and is based on the same Doppler waveform as M-VTI. Figure
14 60 M 27 Ischemic 1E shows an A-wave truncated at 150 milliseconds and
15 59 F 59 Idiopathic untruncated at 180 milliseconds.
16 75 M 33 Idiopathic Optimization techniques were compared in terms of their ability to
17 77 F 18 Idiopathic detect underlying physiologic changes with pacing interval. Specif-
18 60 M 20 Ischemic ically, the statistical significance of the data was quantitatively
estimated, as described below. Because A-wave truncation yields a
19 49 M 48 Ischemic, AV block binary assessment at each test interval (A wave is or is not
20 66 F 52 HCM truncated), it is not amenable to the analytic paradigm used for the
other techniques. Therefore, for A-wave truncation, we report the
AS indicates aortic stenosis; EF, ejection fraction; HCM, hypertrophic
number of times each of the 3 readers, blinded to other results, was
cardiomyopathy; MR, mitral regurgitation. able to estimate an optimum pacing interval.
*Dual-chamber device. Summary statistics were obtained based on all patients referred for
No LV capture. clinical optimization and, in addition, with hypertrophic cardiomy-
Frequent bigeminy. opathy patients excluded. Because ICG allows a greater number of
No atrial lead. data points to be obtained compared with echocardiographic meth-
ods, the AV interval optimization analysis was repeated for ICG data
bootstrapping, a computational approach that makes knowl- using the same number of measurements that were obtained in the
corresponding A-VTI data set at each test interval.
edge of the underlying statistical properties of the data
unnecessary.23 Statistical Methods
A third-degree polynomial was fit to the data, and the location of
Methods the maximum was taken as the estimated optimum interval. For
SPWMD, the location of the minimum absolute value of the
Patient Selection polynomial was taken as the optimum. Use of a continuous function,
Data were obtained from consecutive patients referred for clinical such as a polynomial, allows interpolation between test intervals and
pacing interval optimization with institutional review board ap- averaging both at and across test intervals provided that the number
proval. Patients with dual-chamber or biventricular pacemakers or of free parameters of the function is smaller than the number of
implantable cardioverter defibrillators were included without regard unique test intervals.
to underlying etiology (Table 1). The test for statistical significance was based on the formulation of
the alternative hypothesis that the measured data do not depend on
Optimization Techniques pacing interval. The probability of obtaining the observed data under
AV/VV interval optimization was conducted using multiple tech- this null hypothesis was estimated using bootstrapping.22,23 A test
niques, as follows: statistic s was defined to be equal to the area bounded above by the
best-fitting polynomial and below by the minimum value of the
1. Noninvasive, beat-to-beat estimates of stroke volume (SV) polynomial, as illustrated in Figure 2. The test statistic is not unique;
were acquired continuously using impedance cardiography other measures also would be acceptable if they possess the property
(ICG; BioZ, CardioDynamics, San Diego, Calif) (Figure that their value varies depending on how physiological the
1A).24 26 With ICG, changes in thoracic impedance are mea- optimization data are. Specifically, if an underlying physiologic
sured using surface electrodes and then processed by a propri- optimum exists, and if the range of test intervals was appropriately
etary algorithm to estimate SV and other hemodynamic pa- selected to span the optimum, then the test statistic should be larger
rameters. Each test interval was delivered for 60 seconds, with than what would be obtained under the null hypothesis in which there
all beats recorded during the last 30 seconds included in the is no dependence on pacing interval. The greater the difference in
analysis. cardiac function at the optimum pacing interval and the extreme of
2. The remaining techniques were derived from echocardiogra- the test range, the greater the value of s. For each optimization data
phy using a Philips iE33 System (Philips International B.V., set, 1000 surrogate bootstrapped data sets were generated by ran-
Amsterdam, The Netherlands). The aortic velocity-time inte- domly selecting data points from the original data set with replace-
gral (A-VTI), which is directly proportional to SV, was ment and without regard to test interval. This process yields
obtained by numerically integrating the ejection velocity en- surrogate data sets with the same number of data points at each test
velop obtained by continuous-wave Doppler directed in line interval as the original but replaces the original mapping between

Downloaded from http://circheartfailure.ahajournals.org/ by guest on December 9, 2013

35 of 158.
 Turcott et al Optimization of Cardiac Resynchronization Therapy 397

Figure 1. Raw optimization data. A, Impedance cardiography is measured noninvasively using surface electrodes and provides
estimates of multiple hemodynamic parameters, including stroke volume (SV). B, Aortic velocity-time integral is measured with
continuous-wave Doppler at the left ventricular outflow tract and is directly proportional to ejected SV. C, Mitral velocity-time inte-
gral (M-VTI) is measured with pulsed-wave Doppler echocardiography and is directly proportional to inflow volume. D, Septal-
posterior wall-motion delay is obtained using tissue Doppler imaging. The VV interval that minimizes the delay is taken as repre-
senting optimum synchrony. E, A-wave truncation uses the same images as M-VTI to identify the pacing interval at which the
A-wave becomes truncated by ventricular systole. The image shows the A-wave truncated when a 150-ms AV delay is used and
untruncated with a 180-ms AV delay.

Downloaded from http://circheartfailure.ahajournals.org/ by guest on December 9, 2013

36 of 158.
 398 Circ Heart Fail May 2010

Figure 2. Estimation of statistical significance using bootstrapping. Measures of cardiac function using impedance cardiography (ICG)
and aortic velocity-time integral (A-VTI) recorded from the same patient at the same optimization session. A test statistic s is defined as
the area that is bounded above by the best-fitting polynomial and below by the minimum of the polynomial. The greater the difference
in cardiac function at best and worst test intervals, the larger the value of s. Bootstrapping is used to obtain surrogate data sets (bot-
tom) from the original data that preserve the amplitude statistics of the original data but lack the underlying dependence on cardiac
function. The fraction p of bootstrapped tests statistics s that exceed s reflects the likelihood that the observed data occur in the
absence of a dependence of cardiac function on test interval. Here, the ICG data are highly significant (P0.001) and demonstrate a
physiologically plausible dependence on pacing interval. The data are well described by the continuous function despite the delivery of
test intervals in a random order and the relatively close spacing between test intervals. Furthermore, the 120-ms test interval delivered
4 times over the course of data acquisition shows good repeatability compared with the range of SVs over all test intervals. In contrast,
the A-VTI data are consistent with the null hypothesis: The large relative noise masks the underlying physiological dependence on pac-
ing interval, and the location of the optimum identified by A-VTI is an artifact of the statistical variability. In this patient, accepting the
optimum identified by A-VTI would result in worse cardiac function than the default interval (120 milliseconds). Solid curve indicates
best-fitting third-degree polynomial; X, location of estimated optimum pacing interval; horizontal line, 95% CI of estimated optimum. SV
indicates stroke volume.

measured data and test interval with a random pairing. Each
surrogate bootstrapped data set thus represents a single realization of
data that would be observed under the null hypothesis while
preserving the amplitude statistics of the original data. The test
statistic s was calculated for each surrogate data set, and the fraction
that was greater than or equal to that of the original data estimates the
probability that a test statistic at least as large as that associated with
the original data would be observed under the null hypothesis. For
P0.05, the null hypothesis was rejected, and the data were
interpreted as demonstrating a statistically significant dependence on
pacing interval.
The 95% CI of the estimated optimum pacing interval was
obtained by generating an additional 1000 bootstrapped data sets in
which the original data were randomly selected with replacement
while preserving the mapping to test interval. In this case, with the
mapping between test interval and measured data retained, the
collection of best-fitting polynomials of the bootstrapped data sets
reflects the variability in the best-fitting polynomial of the original
data that is attributable to the statistical variability in the measure-
ments. The locations of the maxima of the surrogate data sets were
ordered from smallest to largest, and the cutoff point of the smallest
and largest 2.5% of the surrogate optima were taken as the 95% CI
of optimum estimated from the original data.
Assessment of A-VTI Variability
The variability of the measured A-VTI data was quantified by
calculating the average and SD over all measurements made at an
AV delay of 120 ms, which by protocol were acquired over at least
1 respiratory cycle at 4 different times during the recording session.
To allow comparison with previously published values, a transfor-
mation was derived that converts the coefficient of variation (defined
as SD divided by mean) to the average difference of successive
measurements. Specifically, y2CV 2/, where y is the average
difference in successive measurements, and CV is the coefficient of
variation. This result is based on the transformation yixi1xi2/i,
where the xi represents successive A-VTI measurements at a given
pacing interval in a given patient; yi, the normalized difference in
A-VTI measures in a given patient, and the subscript i, patient-
specific values. The derivation assumes that xi1 and xi2 are indepen-
dently drawn from a Gaussian distribution whose mean i and SD
i can vary among patients but whose coefficient of variation CVi
remains fixed for all patients, that is, CVii/iconstant. Data
acquisition and device programming are summarized in Table 2.

Downloaded from http://circheartfailure.ahajournals.org/ by guest on December 9, 2013

37 of 158.
 Turcott et al
Table 2. Summary of Data Acquisition, Analysis, and
Pacemaker Programming
Data acquisition
AV and VV interval optimization were conducted independently, with
simultaneous biventricular pacing during AV optimization and an AV
delay of 120 ms during VV optimization
Data were recorded from each patient at a single session in the
following order: ICG (AV), ICG (VV), A-VTI (AV), M-VTI/A-Wave truncation
(AV), A-VTI (VV), SPWMD (VV)
Consecutive beat-to-beat measures of cardiac function were recorded
over at least 1 respiratory cycle
Test intervals were delivered in random order with 20 30 ms spacing
The 120 ms AV delay and 0 ms VV interval were repeated 4 times over
the duration of the recording to assess stability
Data Analysis
A 3rd degree polynomial was fit to the data, with the location of the
maximum serving as the estimated optimum (For SPWMD, the optimum
was the location with the minimum absolute delay)
A P value reflecting the statistical significance of measured data was
obtained
The 95% CI of the estimated optimum was obtained
For A-wave truncation, the shortest AV delay that avoided A-wave
truncation was selected as optimum, and the fraction of patients for
whom an optimum could be determined was noted
Pacemaker Programming
The statistical significance, 95% CI, and physiologic plausibility of the
measured data were considered
The pacemaker was left at default settings (AV delay 120 ms, VV
interval 0 ms) for statistically insignificant data sets, wide 95% CIs
that included the default setting, or physiologically implausible data.
Otherwise, the estimated optimum pacing interval was programmed.
Results
Patient demographics are presented in Table 1. Of the 20
sequential, unique patients referred for clinical pacing opti-
mization, 18 had biventricular pacemakers or implantable
cardioverter defibrillators and 2 had dual-chamber devices.
The LV lead in 1 patient failed to capture, and another patient
with chronic atrial fibrillation had a biventricular device
Figure 3. AV delay optimization. Data were obtained from an individual patient at a single optimization session. Impedance cardiogra-
phy (ICG) and aortic velocity-time integral (A-VTI) yielded statistically significant results, although the 95% CI of the estimated optimum
was much narrower for ICG than for A-VTI. Mitral velocity-time integral (M-VTI) data were not statistically significant. Solid curve indi-
cates best-fitting third-degree polynomial; X, location of estimated optimum pacing interval; horizontal line, 95% CI of estimated
optimum.
Downloaded from http://circheartfailure.ahajournals.org/ by guest on December 9, 2013
38 of 158.
Optimization of Cardiac Resynchronization Therapy 399
without an atrial lead. Patients who underwent AV delay
optimization were in sinus rhythm with the AV delay initiated
by an atrial sensed-event in all cases.
AV delay optimization data from a single patient are
presented in Figure 3. As with the ICG data shown here (left
panel), when statistically significant data were obtained they
typically exhibited an inverted U appearance, indicating that
the estimated optimum interval was in the interior of the
range of test intervals. Statistically insignificant data typically
had a best-fitting polynomial that was flat relative to the
intrinsic variability of the data.
The average and SD of the measured A-VTI data were
calculated over all beats (10 to 31, median 20) recorded from
each patient during AV optimization at the 120-ms test
interval. The ranges of the per-patient average and SD of the
A-VTI data were 6.7 to 56 cm and 0.73 to 5.1 cm, respec-
tively. The median sample coefficient of variation was 0.075,
which corresponds to an average normalized difference in
successive measurements of y0.12, consistent with previ-
ous reports of y0.10.1.27
The VV optimization data shown in Figure 4 are from the
same patient whose AV optimization results are presented in
Figure 3. As with this individual, for most patients, VV
optimization yielded insignificant results more frequently
than did AV optimization.
The number of statistically significant data sets is summa-
rized in Table 3. For both AV and VV optimization, A-VTI,
M-VTI, and SPWMD yielded data that were indistinguish-
able from the null hypothesis (ie, failed to demonstrate a
statistically significant dependence on pacing interval) the
majority of the time. As shown in Table 3, this finding
remained true whether patients with hypertrophic cardiomy-
opathy were included or excluded from the analysis. ICG was
significantly different from the null hypothesis 84% of the
time for AV delay optimization and 75% of the time for VV
interval optimization. Excluding patients with hypertrophic
cardiomyopathy from the analysis, the null hypothesis could
be rejected 81% and 75% of the time, respectively.
 400 Circ Heart Fail May 2010

Figure 4. VV interval optimization. Data were obtained from the same patient and optimization session as those presented in Figure 3.
Negative VV intervals indicate that the left ventricle is paced first, with extreme values corresponding to unichamber pacing. Excluding
the unichamber LV pacing from the septal-posterior wall-motion delay (SPWMD; VV140 ms) would yield statistically significant
results, although the physiological plausibility of extremely early right ventricular pacing it predicts is questionable, especially in light of
the insignificant results from impedance cardiography (ICG) and aortic velocity-time integral (A-VTI). Solid curve indicates best-fitting
third-degree polynomial; X, location of estimated optimum pacing interval; horizontal line, 95% CI of estimated optimum.

Among the statistically significant data sets, the optimum
AV delay estimated by ICG differed from the default value
(120 ms) by an average magnitude of 57 ms, and 63% of the
estimates differed from default by at least 50 ms. Among the
statistically significant, ICG-derived VV interval data sets,
estimated optima differed from the default value (0 ms) by an
average of 52 ms, and 75% of the estimates differed by at
least 30 ms. Repeating the ICG AV interval optimization
analysis using the same number of measurements at each test
interval as the A-VTI recordings continued to yield statisti-
cally significant results in the majority of patients, with 81%
of the reduced ICG data sets having P values 0.05.
An optimum AV delay could be estimated by 3 indepen-
dent readers using A-wave truncation 69%, 75%, and 94% of
the time. The estimates of each reader are compared with ICG
results in Table 4. The relationship between optima predicted
by ICG and A-wave truncation was variable, with the Pearson
correlation coefficient ranging from 0.02 to 0.67 for the 3
readers.
Although the analysis presented above is based on unique
patient visits, 1 patient underwent both ICG and echocardio-
graphic optimization on 2 separate occasions separated by 3.5
months. As shown in Figure 5, in both cases ICG yielded
similar-appearing, statistically significant results. The esti-
mates of the optimum pacing intervals were precise, with
narrow 95% CIs, and had good concordance, with 91 ms at
the first optimization and 67 ms 3.5 months later. In contrast,
in neither optimization session did A-VTI yield statistically
significant results. The wide 95% CI of the initial A-VTI
optimum predicted that subsequent optimization attempts
would be unlikely to identify a similar optimum interval.
Discussion
In this study, multiple clinically accepted AV and VV interval
optimization techniques were used in single sessions in a
heterogeneous group of 20 patients. For 62% to 86% of
patients, A-VTI, M-VTI, and SPWMD yielded data that were
statistically indistinguishable from the null hypothesis; that is,
Downloaded from http://circheartfailure.ahajournals.org/ by guest on December 9, 2013
39 of 158.
the majority of the time these measures yielded data that did
not show a significant dependence on AV or VV interval.
With A-wave truncation, it was possible for 3 independent
readers to estimate an optimum AV delay 69%, 75%, and
94% of the time. ICG performed better than the echocardi-
ography methods, yielding statistically significant data 84%
of the time for AV delay optimization and 75% of the time for
VV interval optimization. Notably, the echocardiography
techniques tested here represent the most commonly used
approaches to AV/VV interval optimization.28
In previous studies, SPWMD failed to predict a response to
CRT,29,30 and interval optimization using A-VTI neither
improved clinical outcomes12 nor yielded acute data that were
distinguishable from a negative control.31 The poor precision
of these techniques demonstrated in this study may account
for the lack of clinical utility seen in the earlier work.
AV/VV interval optimization traditionally has been con-
ducted without consideration of the intrinsic variability of the
measured data. Test intervals often are delivered in a nonran-
dom order, sweeping systematically from one end of the test
range to the other, and the AV or VV interval associated with
the best average measure of cardiac function typically is
taken at face value to represent the underlying physiologic
optimum.11,20 Although efficient, the traditional approach has
important drawbacks. A nonrandom order of test intervals
allows systematic error to be introduced into the measured
data in a way that cannot be subsequently corrected by
averaging (eg, with subtle drift of the Doppler angle over the
course of data collection). Furthermore, without considering
the intrinsic variability of the measured data, it is not possible
to characterize the precision of the estimated optimum
AV/VV interval.
In the absence of knowledge about its precision, an
estimated optimum AV/VV interval may in fact be spurious
and associated with worse cardiac function than the
population-derived, default setting (eg, a sensed AV delay of
120 milliseconds) (Figures 2, 3, and 5). In addition, although
multiple studies have suggested that optimum pacing inter-
 Turcott et al Optimization of Cardiac Resynchronization Therapy 401

Table 3. P Value Associated With Measured Optimization Data Table 4. Estimated Optima: ICG and A-Wave Truncation
AV Optimization VV Optimization ICG Estimated A-Wave
Optima, ms Truncation, ms
Patients ICG A-VTI M-VTI ICG A-VTI SPWMD Patient
No. Optimum 95% CI Reader 1 Reader 2 Reader 3
Individual patients
1 0.001 0.03 0.03 0.001 0.018 0.002 1 168 156 177 180 150 210

2 0.18 0.26 0.40 0.001 0.51 0.25 2 240 240 240 150 150 180

3 0.001 0.013 0.21 0.12 0.31 0.10 3 206 193 213 150 120 *

4 0.001 0.001 0.007 0.72 0.12 0.003 4 197 187 206 240 180 210

5 0.09 0.37 0.50 0.001 0.14 0.42 5 93 60 117 180 120 210

6 0.001 0.68 6 91 83 97 120 150 120

7 0.001 0.06 7 128 121 138

8 0.001 0.24 0.05 0.001 0.19 0.001 8 200 177 270 180 * 240

9 0.008 0.98 0.91 0.092 0.41 0.24 9 178 170 197 120 150 120

10 0.004 0.59 0.61 0.001 0.007 0.57 10 60 60 60 240 * *

11 0.001 0.17 0.44 0.001 0.73 0.93 11 60 60 60 150 120 120

12 0.016 0.04 0.17 12 60 60 60 180 * *

13 0.023 0.50 0.44 0.054 0.41 0.054 13 173 125 190 150 180 120

14 0.001 0.77 0.057 0.002 0.58 0.11 14 164 157 172 180 150 80

15 0.001 0.003 0.14 0.006 0.073 0.24 15 123 118 131 240 * *

16 0.051 0.05 0.16 0.002 0.061 0.081 16 218 30 248 * * 180

17 0.001 0.32 0.075 17

18 0.001 0.001 18 77 66 86

19 0.001 19 30 30 30 120 60 90

20 0.001 0.001 20 240 180 240

All patients Percent interpretable was as follows: Reader 1, 94%; Reader 2, 69%;
Reader 3, 75%. Correlation coefficients were as follows: Reader 1, 0.02;
Significant, n 16 6 3 12 2 3
Reader 2, 0.67; Reader 3, 0.48.
Attempted, n 19 16 14 16 14 14 *Reader indicated that A-wave truncation could not be reliably identified. ICG
Significant, % 84 38 21 75 14 21 indicates impedance cardiography; CI, confidence interval.
HCM excluded
Significant, n 13 6 3 12 2 3 the trial will not address the question of whether the experi-
mental technique is superior to default settings. In contrast,
Attempted, n 16 14 14 16 14 14
another clinical trial33,34 uses a 3-arm design in which patients
Significant, % 81 42 21 75 14 21
are randomized to the experimental technique, a control arm
HCM indicates hypertrophic cardiopmyopathy; AV, atrioventricular; VV, in which population-derived default settings are used, and a
interventricular; ICG, impedance cardiography; A-VTI, aortic velocity-time conventional optimization arm in which a specific optimiza-
integral; M-VTI, mitral velocity-time integral; SPWMD, septal-posterior wall-
motion delay.
tion technique is uniformly used. This study design allows
direct comparisons between the experimental technique and
both default settings and the specific conventional optimiza-
vals change over time,12,17,18 the lack of characterization of tion method, and for comparison of the conventional tech-
the precision of the estimated optima makes it impossible to nique to default settings.
determine whether the observed change reflects changes in Statistical significance testing may provide a useful way to
underlying physiology or is simply due to statistical variabil- evaluate the quality of competing optimization techniques. It
ity in measured data31 (Figure 5). avoids prespecifying a gold standard, and although demon-
The fact that any optimization technique carries some stration of improved clinical outcomes in well-designed trials
degree of imprecision requires caution when designing ran- ultimately is required, the approach presented here offers a
domized prospective trials. Two electrogram-based optimiza- way to narrow the very wide field of plausible optimization
tion techniques sponsored by competing manufacturers are techniques without the resource requirements of a prospective
currently undergoing clinical trials. In 1 design,32 patients are clinical trial.
randomized to the experimental optimization technique or a Although a rigorous, quantitative analysis is desirable,
control arm that may or may not include AV and VV interval statistical significance can be evaluated informally by plot-
optimization. If optimization is performed, it is conducted at ting the measured data against pacing interval. With test
the physicians discretion without a requirement to examine intervals delivered in a random order, if the data exhibit the
the precision of the measured data, raising the possibility of a expected inverted U shape and are tightly clustered about the
programmed AV or VV interval that yields worse cardiac overall curve, then one can be confident that the estimation of
function than population-derived default settings. By design, optimum AV/VV interval is precise; repeating the process
Downloaded from http://circheartfailure.ahajournals.org/ by guest on December 9, 2013
40 of 158.
 402 Circ Heart Fail May 2010
likely would yield a similar result. On the other hand, if the
plot is relatively flat compared with the intrinsic variability of
the measured data, then the estimate of the underlying
optimum is imprecise and heavily influenced by measure-
ment variability. In this case, repeating the optimization
process likely would yield a very different estimated opti-
mum AV/VV interval. The statistical tools used here along
with plotting capability have been made available on the
Internet.35
In the majority of patients examined in the present study,
ICG generated precise estimates and A-wave truncation
yielded data from which an optimum could be inferred,
although often with significant interreader variability and
marginal correlation with the ICG-predicted optima. Notably,
neither ICG nor A-wave truncation has been clinically vali-
dated in prospective interval optimization studies. In addition,
unanswered fundamental questions include whether the phys-
iologic optimum interval evolves over time and whether it
changes between rest and exertion or between supine and
upright posture. A study that compares estimated optimum
intervals obtained at both supine rest and upright exertion in
the same patient, perhaps using motion-tolerant ICG, would
add important insight into these basic questions. In the
absence of such data and given the theoretical potential
benefit of pacing optimization, our approach is to accept an
estimated optimum interval if quantitative and qualitative
analyses suggest that the estimate has good precision. Partic-
ular attention should be paid to the effect of outliers and the
overall shape of the curve compared with the measurement
variability.
Downloaded from http://circheartfailure.ahajournals.org/ by guest on December 9, 2013
41 of 158.
Figure 5. Stability of estimated optima
over time. Impedance cardiography and
aortic velocity-time integral were
obtained from the same patient at base-
line and 3.5 months later. In this patient,
the ICG data were highly significant, the
appearance of the data on repeat opti-
mization recapitulated the baseline data,
the estimated optima were concordant
between the 2 sessions, and the 95%
CIs of the estimates were narrow. In
contrast, the A-VTI data did not show a
statistically significant dependence on
pacing interval, and the estimated
optima were highly discordant between
the 2 optimization sessions. Accepting
the A-VTI-estimated optima without a
consideration of their statistical signifi-
cance or precision would lead to the
erroneous conclusion that the optimum
AV delay had changed in the intervening
3.5 months.
That ICG continued to yield statistically significant data in
the majority of patients, even when using an identical number
of data points as the A-VTI analysis, suggests that it has a
superior intrinsic signal-to-noise ratio and that the acquisition
time can be substantially shortened from the 60 s per test
interval that was used in this study. The superior noise
properties of ICG may be partly due to the automatic and
objective nature of data acquisition and analysis in contrast to
A-VTI, which requires the sonographer to physically hold the
probe in a fixed position and the reader to manually demar-
cate the envelope of the velocity waveform.
A wide variety of optimization techniques have been
advocated, including multiple approaches to the assessment
of systolic function, diastolic function, and electrical and
mechanical synchrony.9,19 21,24 26,30,36 40 Although each has a
rationale that is mechanistically plausible, consideration of
the neurohormonal derangements of heart failure and the
therapeutic interventions that have been successful lead us to
view SV and its surrogates as parameters that when optimized
are most likely to translate into clinical benefit. Specifically,
it is now well established that ameliorating the effects of
sympathetic tone in these patients leads to improved clinical
outcomes.41 For a given cardiac output, maximizing SV
would minimize sympathetic tone. Indeed, the effects on the
neurohormonal system of increased mechanical efficiency
may contribute to the salutary effects of CRT, which remains
the only contractility-enhancing intervention demonstrated to
prolong life.42 Theoretical arguments may not account for
important effects, however. For example, an increase in SV at
the expense of greater oxygen consumption may not benefit
 Turcott et al
the patient with ischemic heart failure. Ultimately, any
proposed optimization technique must be validated by ran-
domized prospective trials with hard clinical end points, a
goal which to date has not been frequently achieved.8 12
Limitations
The study was based on a small and heterogeneous popula-
tion comprising successive patients referred for clinical pac-
ing interval optimization. Multiple conventional optimization
techniques were examined for their ability to yield statisti-
cally significant results. This property is necessary but not
sufficient for improved clinical outcomes, which were not
examined in this study.
Conclusions
Optimization of AV and interventricular intervals in CRT
requires assessment of the variability of the measured data.
Accepting an estimated optimum without considering its
precision may result in worse cardiac function than default
settings in the individual patient and confound results in
clinical trials. In the small, heterogeneous pacemaker popu-
lation examined here, echocardiographic techniques yielded
statistically insignificant data in the majority of patients. In
contrast, ICG yielded precise estimates of the optimum AV
and VV interval in most patients. Further research is neces-
sary to confirm these results, to validate the accuracy of the
impedance-predicted optima, and to demonstrate clinical
improvement with pacing interval optimization compared to
population-derived default settings.
Sources of Funding
Dr Turcott was supported by a Heart Failure Society of America
Research Fellowship and by the National Library of Medicine
(Biomedical Informatics Training Grant LM 07033). Dr Ashley was
supported by the National Institutes of Health (grant K08
HL083914).
Disclosures
Dr Witteles has received honoraria from Medtronic. Dr Wang has
received honoraria and research support from and has served as a
consultant or advisor to Boston Scientific, Medtronic, and St Jude
Medical. Dr Fowler has received honoraria from Medtronic and
Boston Scientific.
References
1. Bristow MR, Saxon LA, Boehmer J, Krueger S, Kass DA, De Marco T,
Carson P, DiCarlo L, DeMets D, White BG, DeVries DW, Feldman AM.
Cardiac-resynchronization therapy with or without an implantable defi-
brillator in advanced chronic heart failure. N Engl J Med. 2004;350:
2140 2150.
2. Cleland JG, Daubert JC, Erdmann E, Freemantle N, Gras D, Kappen-
berger L, Tavazzi L. The effect of cardiac resynchronization on morbidity
and mortality in heart failure. N Engl J Med. 2005;352:1539 1549.
3. McAlister FA, Ezekowitz J, Hooton N, Vandermeer B, Spooner C,
Dryden DM, Page RL, Hlatky MA, Rowe BH. Cardiac resynchronization
therapy for patients with left ventricular systolic dysfunction: a systematic
review. J Am Med Assoc. 2007;297:25022514.
4. Yu CM, Wing-Hong Fung J, Zhang Q, Sanderson JE. Understanding
nonresponders of cardiac resynchronization therapy current and future
perspectives. J Cardiovasc Electrophysiol. 2005;16:11171124.
5. Breithardt OA, Stellbrink C, Franke A, Balta O, Diem BH, Bakker P,
Sack S, Auricchio A, Pochet T, Salo R. Acute effects of cardiac resyn-
chronization therapy on left ventricular Doppler indices in patients with
congestive heart failure. Am Heart J. 2002;143:34 44.
Downloaded from http://circheartfailure.ahajournals.org/ by guest on December 9, 2013
42 of 158.
Optimization of Cardiac Resynchronization Therapy 403
6. Auricchio A, Stellbrink C, Block M, Sack S, Vogt J, Bakker P, Klein H,
Kramer A, Ding J, Salo R, Tockman B, Pochet T, Spinelli J. Effect of
pacing chamber and atrioventricular delay on acute systolic function of
paced patients with congestive heart failure. Circulation. 1999;99:
29933001.
7. Kass DA, Chen CH, Curry C, Talbot M, Berger R, Fetics B, Nevo E.
Improved left ventricular mechanics from acute VDD pacing in patients with
dilated cardiomyopathy and ventricular conduction delay. Circulation. 1999;
99:15671573.
8. Boriani G, Muller CP, Seidl KH, Grove R, Vogt J, Danschel W,
Schuchert A, Djiane P, Biffi M, Becker T, Bailleul C, Trappe HJ.
Randomized comparison of simultaneous biventricular stimulation
versus optimized interventricular delay in cardiac resynchronization
therapy: The Resynchronization for the HemodYnamic Treatment for
Heart Failure Management II Implantable Cardioverter Defibrillator
(RHYTHM II ICD) study. Am Heart J. 2006;151:1050 1058.
9. Morales MA, Startari U, Panchetti L, Rossi A, Piacenti M. Atrioven-
tricular delay optimization by Doppler-derived left ventricular dP/dt
improves 6-month outcome of resynchronized patients. Pacing Clin
Electrophysiol. 2006;29:564 568.
10. Rao RK, Kumar UN, Schafer J, Viloria E, De Lurgio D, Foster E.
Reduced ventricular volumes and improved systolic function with cardiac
resynchronization therapy: a randomized trial comparing simultaneous
biventricular pacing, sequential biventricular pacing, and left ventricular
pacing. Circulation. 2007;115:2136 2144.
11. Sawhney NS, Waggoner AD, Garhwal S, Chawla MK, Osborn J, Faddis
MN. Randomized prospective trial of atrioventricular delay programming
for cardiac resynchronization therapy. Heart Rhythm. 2004;1:562567.
12. Boriani G, Biffi M, Muller CP, Seidl KH, Grove R, Vogt J, Danschel W,
Schuchert A, Deharo JC, Becker T, Boulogne E, Trappe HJ. A pro-
spective randomized evaluation of VV delay optimization in CRT-D
recipients: echocardiographic observations from the RHYTHM II ICD
study. Pacing Clin Electrophysiol. 2009;32(suppl 1):S120 S125.
13. Ritter P, Daubert C, Mabo P, Descaves C, Gouffault J. Haemodynamic
benefit of a rate-adapted A-V delay in dual chamber pacing. Eur Heart J.
1989;10:637 646.
14. Melzer C, Korber T, Theres H, Nienaber CA, Baumann G, Ismer B. How
can the rate-adaptive atrioventricular delay be programmed in atrioven-
tricular block pacing? Europace. 2007;9:319 324.
15. Scharf C, Li P, Muntwyler J, Chugh A, Oral H, Pelosi F, Morady F,
Armstrong WF. Rate-dependent AV delay optimization in cardiac resyn-
chronization therapy. Pacing Clin Electrophysiol. 2005;28:279 284.
16. Mehta D, Gilmour S, Ward DE, Camm AJ. Optimal atrioventricular delay
at rest and during exercise in patients with dual chamber pacemakers: a
non-invasive assessment by continuous wave Doppler. Br Heart J. 1989;
61:161166.
17. Zhang Q, Fung JW, Chan YS, Chan HC, Lin H, Chan S, Yu CM. The role
of repeating optimization of atrioventricular interval during interim and
long-term follow-up after cardiac resynchronization therapy. Int
J Cardiol. 2008;124:211217.
18. Valzania C, Biffi M, Martignani C, Diemberger I, Bertini M, Ziacchi M,
Bacchi L, Rocchi G, Rapezzi C, Branzi A, Boriani G. Cardiac resynchro-
nization therapy: variations in echo-guided optimized atrioventricular and
interventricular delays during follow-up. Echocardiography. 2007;24:
933939.
19. Burri H, Sunthorn H, Shah D, Lerch R. Optimization of device pro-
gramming for cardiac resynchronization therapy. Pacing Clin Electro-
physiol. 2006;29:1416 1425.
20. Agler DA, Adams DB, Waggoner AD. Cardiac resynchronization therapy
and the emerging role of echocardiography (part 2): the comprehensive
examination. J Am Soc Echocardiogr. 2007;20:76 90.
21. Jansen AH, Bracke FA, van Dantzig JM, Meijer A, van der Voort PH,
Aarnoudse W, van Gelder BM, Peels KH. Correlation of echo-Doppler
optimization of atrioventricular delay in cardiac resynchronization
therapy with invasive hemodynamics in patients with heart failure sec-
ondary to ischemic or idiopathic dilated cardiomyopathy. Am J Cardiol.
2006;97:552557.
22. Turcott RG, Ashley EA. Statistical methods for AV/VV optimization
[abstract]. J Cardiac Failure. 2008;14:S65.
23. Efron B, Tibshirani R. Statistical data analysis in the computer age.
Science. 1991;253:390 395.
24. Heinroth KM, Elster M, Nuding S, Schlegel F, Christoph A, Carter J,
Buerke M, Werdan K. Impedance cardiography: a useful and reliable tool
in optimization of cardiac resynchronization devices. Europace. 2007;9:
744 750.
 404 Circ Heart Fail May 2010

25. Braun MU, Schnabel A, Rauwolf T, Schulze M, Strasser RH. Impedance
cardiography as a noninvasive technique for atrioventricular interval optimi-
zation in cardiac resynchronization therapy. J Interv Card Electrophysiol.
2005;13:223229.
26. Tse HF, Yu C, Park E, Lau CP. Impedance cardiography for atrioven-
tricular interval optimization during permanent left ventricular pacing.
Pacing Clin Electrophysiol. 2003;26:189 191.
27. Zuber M, Toggweiler S, Quinn-Tate L, Brown L, Amkieh A, Erne P. A
comparison of acoustic cardiography and echocardiography for opti-
mizing pacemaker settings in cardiac resynchronization therapy. Pacing
Clin Electrophysiol. 2008;31:802 811.
28. Gras D, Gupta MS, Boulogne E, Guzzo L, Abraham WT. Optimization of
AV and VV delays in the real-world CRT patient population: an inter-
national survey on current clinical practice. Pacing Clin Electrophysiol.
2009;S236 S239.
29. Beshai JF, Grimm RA, Nagueh SF, Baker JH II, Beau SL, Greenberg SM,
Pires LA, Tchou PJ. Cardiac-resynchronization therapy in heart failure
with narrow QRS complexes. N Engl J Med. 2007;357:24612471.
30. Chung ES, Leon AR, Tavazzi L, Sun JP, Nihoyannopoulos P, Merlino J,
Abraham WT, Ghio S, Leclercq C, Bax JJ, Yu CM, Gorcsan J III, St John
Sutton M, De Sutter J, Murillo J. Results of the Predictors of Response to
CRT (PROSPECT) trial. Circulation. 2008;117:2608 2616.
31. Turcott RG, Das AK, Ashley EA. Re: A prospective randomized
evaluation of VV delay optimization in CRT-D recipients: echocar-
diographic observations from the RHYTHM II ICD study. Pacing Clin
Electrophysiol. 2009;32:1360 1361; author reply 13611362.
32. FREEDOM: A Frequent Optimization Study Using the QuickOpt
Method. http://www.clinicaltrials.gov. Identifier: NCT00418314.
Accessed January 2, 2010.
33. Stein KM, Ellenbogen KA, Gold MR, Lemke B, Lozano IF, Mittal S,
Spinale FG, van Eyk JE, Waggoner AD, Meyer TE. SmartDelay
Determined AV Optimization: a comparison of AV delay methods used
in cardiac resynchronization therapy (SMART-AV): rationale and design.
Pacing Clin Electrophysiol. 2010;33:54 63.
34. Comparison of AV Optimization Methods Used in Cardiac Resynchro-
nization Therapy (CRT) (SMART-AV). http://www.clinicaltrials.gov.
Identifier: NCT00677014. Accessed January 2, 2010.
35. Stanford University. Optimize CRT (Development Version) Available at
http://optimizecrt.stanford.edu. Accessed January 2, 2010.
36. Bertini M, Ziacchi M, Biffi M, Martignani C, Saporito D, Valzania C,
Diemberger I, Cervi E, Frisoni J, Sangiorgi D, Branzi A, Boriani G.
Interventricular delay interval optimization in cardiac resynchronization
therapy guided by echocardiography versus guided by electrocardio-
graphic QRS interval width. Am J Cardiol. 2008;102:13731377.
37. van Gelder BM, Meijer A, Bracke FA. The optimized V-V interval
determined by interventricular conduction times versus invasive mea-
surement by LV dP/dt(max). J Cardiovasc Electrophysiol. 2008;19:
939 944.
38. Burri H, Sunthorn H, Somsen A, Zaza S, Fleury E, Shah D, Righetti A.
Optimizing sequential biventricular pacing using radionuclide ventricu-
lography. Heart Rhythm. 2005;2:960 965.
39. Vidal B, Tamborero D, Mont L, Sitges M, Delgado V, Berruezo A,
Diaz-Infante E, Tolosana JM, Pare C, Brugada J. Electrocardiographic
optimization of interventricular delay in cardiac resynchronization
therapy: a simple method to optimize the device. J Cardiovasc Electro-
physiol. 2007;18:12521257.
40. Turcott RG, Pavek TJ. Hemodynamic sensing using subcutaneous pho-
toplethysmography. Am J Physiol Heart Circ Physiol. 2008;295:
H2560 H2572.
41. Brophy JM, Joseph L, Rouleau JL. Beta-blockers in congestive heart
failure. A Bayesian meta-analysis. Ann Intern Med. 2001;134:550 560.
42. Kass DA. Rescuing a failing heart: putting on the squeeze. Nat Med.
2009;15:24 25.

CLINICAL PERSPECTIVE
Cardiac resynchronization therapy improves morbidity and mortality in appropriately selected patients. Whether further
clinical benefit is possible with atrioventricular and interventricular pacing interval optimization remains unclear. Tools to
assess the statistical significance of the measured optimization data have not been available previously. In the study
reported here, an objective methodology for quantifying the statistical precision of estimated optimum pacing intervals was
developed and applied to a number of commonly used optimization techniques. Many of the techniques did not yield
statistically significant data in a majority of patients referred for atrioventricular and interventricular interval optimization,
a finding that raises questions about the ability of pacing interval optimization to enhance clinical outcomes. The data
demonstrate that accepting an estimated optimum interval without consideration of its statistical significance can result in
worse cardiac function than default settings and can lead to the erroneous conclusion that the physiological optimum has
changed over time. These results highlight the importance of evaluating the precision of measured data when conducting
pacing interval optimization for the individual patient and when interpreting the results of clinical trials.

Downloaded from http://circheartfailure.ahajournals.org/ by guest on December 9, 2013

43 of 158.
 ORIGINAL ARTICLE Heart Failure Circ J 2009; 73: 10741079

Whole Body Bioimpedance Monitoring for Outpatient

Chronic Heart Failure Follow up

Yusuke Tanino, MD; Junya Shite, MD; Oscar L Paredes, MD; Toshiro Shinke, MD;
Daisuke Ogasawara, MD; Takahiro Sawada, MD; Hiroyuki Kawamori, MD;
Naoki Miyoshi, MD; Hiroki Kato, MD; Naoki Yoshino, MD; Ken-ichi Hirata, MD

Background: Although cardiac output index (CI), stroke volume index (SVI), and total systemic vascular resis-
tance (TSVR) are important hemodynamic parameters for the prognosis of chronic heart failure (CHF), they are
difficult to measure in an outpatient setting. Whole body bioimpedance monitoring using a Non-Invasive Cardiac
System (NICaS) allows for easy, non-invasive estimation of these parameters. Here, whether NICaS-derived
hemodynamic parameters are clinically significant was investigated by relating them to other conventional car-
diovascular functional indices, and by evaluating their predictive accuracy for CHF readmission.
Methods and Results: Study subjects of 68 patients with CHF were enrolled in the study immediately upon
discharge from the hospital. NICaS-derived CI, -SVI, and -TSVR values obtained at an outpatient clinic were
significantly related with left ventricular ejection fraction (LVEF) measured by echocardiography, serum B-type
natriuretic peptide (BNP), and exercise tolerance. During the 10098 days follow-up, 15 patients were readmitted
to our hospital for CHF recurrence. Multivariate analysis indicated that LVEF, NICaS-derived CI, NICaS-derived
SVI, and plasma BNP were significant indicators (receiver operating characteristic curve cut-off point, LVEF: 37%,
NICaS-derived CI: 2.49Lmin1m2, NICaS-derived SVI: 27.2ml/m2, plasma BNP: 344pg/ml) for readmission.
Conclusions: Hemodynamic parameters derived by NICaS are applicable for the non-invasive assessment of
cardiac function in outpatient CHF follow up. (Circ J 2009; 73: 10741079)
Key Words: Bioimpedance; Cardiac output; Congestive heart failure; Non-invasive cardiac monitoring system;
Prediction of readmission

T
he growing geriatric population and increased number
of survivors of acute heart failure have dramatically
increased the number of outpatients with chronic
heart failure (CHF).12 Hemodynamic parameters, such as
stroke volume (SV), cardiac output (CO), total systemic
vascular resistance (TSVR), and plasma level of B-type
natriuretic peptide (BNP) are thought to be important for
predicting the long-term prognosis and guiding the optimal
treatment of CHF.35 Until recently, hemodynamic parame-
ters could only be obtained using the invasive thermodilu-
tion method with a SwanGanz catheter placed in the pul-
monary artery. A non-invasive and low-cost method for
measuring hemodynamic parameters would be useful for
the clinical assessment of CHF in an outpatient setting.
Several non-invasive technologies for measuring hemo-
dynamics are now available. Non-Invasive Cardiac System
(NICaS) (NI Medical; Hod-Hasharon, Israel) is a new
device for calculating SV, CO, and TSVR utilizing whole
body bioimpedance cardiography with electrodes placed on
one wrist and on the contralateral ankle. We previously
(Received September 8, 2008; revised manuscript received January 4,
2009; accepted January 20, 2009; released online April 16, 2009)
Division of Cardiovascular Medicine, Department of Internal Medi-
cine, Kobe University Graduate School of Medicine, Kobe, Japan
Conflict of Interest: none declared
Mailing address: Junya Shite, MD, Division of Cardiovascular and
Respiratory Medicine, Department of Internal Medicine, Kobe Uni-
versity Graduate School of Medicine, 7-5-1 Kusunoki-cho, Chuo-ku,
Kobe 650-0017, Japan. E-mail: shite@med.kobe-u.ac.jp
All rights are reserved to the Japanese Circulation Society. For permis-
sions, please e-mail: cj@j-circ.or.jp
established that NICaS-derived CO (NI-CO) is a reliable
parameter when compared with thermodilution CO or mod-
ified Fick CO6
The purpose of the present study was to evaluate the
feasibility of using NICaS in outpatients with CHF and to
assess whether NICaS-derived hemodynamic parameters in
CHF are clinically significant by relating them to other con-
ventional cardiovascular functional indices, and to assess
the predictive accuracy of NICaS-derived parameters for
CHF-related readmission.
Methods
Patient Selection and Exclusion Criteria
In the present study, patients who met the following cri-
teria were enrolled: (1) hospitalized because of acute CHF;
(2) had a stable condition after optimal medical therapy
including diuretics, vasodilators, ACE-inhibitors; (3) had no
renal dysfunction (defined as a creatinine value >2.0mg/dl);
and (4) had no significant severe valve diseases. The cardi-
nal manifestations of heart failure were dyspnea and fatigue
caused by cardiac dysfunction.
The study population comprised 68 CHF patients (56
men, 12 women, mean age 64.910.8 years); 20 patients
were classified as New York Heart Association (NYHA)
class I and 48 were NYHA class II patients (Table1). The
Institutional Ethics Committee of the Kobe University
Hospital approved the study protocol and all patients pro-
vided informed consent to participate in the study.
In most patients, CHF coexisted with multiple under-
lying heart diseases, as shown in Table1. Three patients

Circulation Journal Vol.73, June 2009

44 of 158.
 Bioimpedance Monitoring for Outpatient Heart Failure
had atrial fibrillation when CO was measured. In addition,
20 cases (29.4%) had a moderate degree of mitral valve
regurgitation, 7 cases (10.3%) had a moderate degree of
aortic valve regurgitation, and 1 case (1.5%) had moderate
aortic valve stenosis. The cardiovascular medicines taken
by the patients comprised: diuretics (n=44, 68.7%), digitalis
(n=7, 10.9%), angiotensin-converting enzyme inhibitors (n=
28, 43.7%), angiotensin receptor blockers (n=42, 65.6%),
beta blockers (n=53, 82.8%), nitrates (n=8, 12.5%), calcium
channel blockers (n=13, 20.3%), and oral inotropic agents
(n=10, 15.6%).
The exclusion criteria for the NICaS measurements
included restlessness and/or unstable patient condition,
severe aortic valve regurgitation and/or aortic stenosis,
aortic aneurysm, heart rate above 130beats/min, intra- and
extra-cardiac shunts, severe peripheral vascular disease,
severe pitting edema, sepsis, and dialysis, all of which inter-
fere with the accurate measurement of impedance-derived
SV,7 as previously described.
All patients were followed up at the outpatient clinic of
the Cardiovascular Division of Kobe University hospital and
had no walking disability. We measured the cardiac param-
eters once the patients attented the outpatient department,
which was the first routine visit to our hospital 1 month after
discharge from the hospital for the first heart failure admis-
sion. At the same visit, serum BNP levels were measured
and a questionnaire about exercise tolerance was adminis-
tered.
Cardiac Indicators Derived by NICaS
To measure NI-CO, an alternating electrical current of
1.4mA with a 30kHz frequency was passed through the
patient via 2 pairs of tetrapolar electrodes, 1 pair placed on
the wrist above the radial artery, and the other pair placed
on the contralateral side above the posterior tibialis artery.
If the arterial pulses were either absent or of poor quality, a
second pair of electrodes was placed on the contralateral
site. The NICaS apparatus calculated the SV using Tsoglin
and Frinermans formula:8
SV=dR/RL2/Ri(+)/KWHF
where dR is the change in impedance, R is the basal resis-
tance, is the blood electrical resistivity, L is the patient
height, Ri is the corrected basal resistance according to
gender and age,812 KW is a correction of weight according
to ideal values,913 HF is a hydration factor that takes into
account body water composition,11+ is equal to the ECG
RR wave interval, andis the diastolic time interval.
Because the NI-CO results are calculated every 20s, the
average of 3 measurements obtained consecutively during
60s of monitoring was considered to be the NI-CO value
for each individual case. By using the NI-CO value, the fol-
lowing parameters were calculated: NI-CO index (NI-CI=
NI-CO/body surface area), SV index (NI-SVI=NI-CI/heart
rate), and TSVR (NI-TSVR=Mean arterial pressure/CO*80
[dynes1cm5]).
Plasma BNP Measurements
Blood samples were obtained from the patient in a resting
condition after NICaS measurement. The samples were
drawn into plastic syringes, transferred to chilled siliconized
disposable tubes containing aprotinin (1,000 kallikrein inac-
tivator in units/ml; Ohkura Pharmaceutical, Kyoto, Japan)
and ethylenediaminetetraacetic acid (1mg/ml), immediately
placed on ice, and then centrifuged at 48C. An aliquot of
Circulation Journal Vol.73, June 2009
45 of 158.
1075
Table1. Patient Characteristics
No. of patients 68
Men 56 (82.4%)
Women 12 (17.6%)
Age (years) 64.910.8
NYHA I 20 (29.4%)
NYHA II 48 (70.6%)
Hypertension 8 (11.8%)
Diabetes mellitus 17 (25.0%)
Ischemic heart disease 18 (26.5%)
Dilated cardiomyopathy 28 (41.2%)
NYHA, New York Heart Association classification
the plasma was immediately frozen at 80C and thawed
only once at the time of the assay, which was performed
within 1 week. The plasma BNP concentration was measured
using a specific immunoradiometric assay kit (Shionogi
Co, Osaka, Japan), as previously reported.14
Echocardiography Measurement of the Ejection Fraction
(EF) and the Ratio of Early-to-Atrial Filling
of Transmitral Flow (E/A)
B-mode recordings were obtained with a commercially
available instrument (SONOS 5500; Agilent Technologies
Inc, Palo Alto, CA, USA) operating at 2.5MHz. Two-dimen-
sional imaging examinations were performed in the standard
manner from apical 4 and 2 chamber views, and the left ven-
tricular (LV) EFs were measured from B-mode images
according to the modified Simpsons method. Also, pulsed-
Doppler findings of transmitral flow were obtained to calcu-
late the E/A.
Exercise Tolerance Estimation
Each patient underwent a questionnaire-based interview
for the estimation of the exercise tolerance threshold (ET)
according to a specific activity scale15
Statistical Analysis
MedCalc version 9.6 (MedCalc Software, Mariakerke,
Belgium) was used for data analysis. The quantitative data
are expressed as meanSD. A Students t-test was used for
descriptive statistics. The MannWhitney U-test was used
for non-parametric data sets. A two-tailed Pearsons corre-
lation was used to compare the relationship between LVEF,
log BNP, and NI-CI, NI-SVI, and NI-TSVR. Spearmans
coefficient of rank correlation (rho) was used to estimate the
probability of correlations with ET. We used the multiple
regression analysis in order to detect the predictors for CHF
readmission, and also used receiver operating characteristic
curve (ROC) analysis for BNP, NI-CI, NI-SVI, NI-TSVR,
LVEF and ET to determine the thresholds between the read-
mission group and the non-readmission group. A two-tailed
P-value of less than 0.05 was considered to be significant.
Results
Relationship Between NICaS-Derived Hemodynamic
Parameters and LVEF, Plasma logBNP and ET
There were significant but weak correlations between
LVEF and NI-CI (LVEF=22.70+5.21*NI-CI, r=0.3148, P=
0.0089), NI-SVI (LVEF=22.80+0.38*NI-SVI, r=0.3257,
P=0.0067), NI-TSVR (LVEF=46.020.003*NI-TSVR, r=
0.2636, P=0.0298; Figure1). There were significant mod-
erate correlations between log BNP and NI-CI (log BNP=
 1076 TANINO Y et al.

70
70
A B
60
60

50
50

EF (%)

EF (%)
40 40

30 30

20 20

10 10
1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 10 20 30 40 50 60
NI-CI (Lmin1 m2) NI-SVI (ml/m2)
y = 22.70 + 5.21 x r=0.3148 P=0.0089 y = 22.80 + 0.38 x r=0.3257 P=0.0067

70
C
60

50
EF (%)

40

30

20

10
1000 2000 3000 4000 5000 6000
NI-TSVR (dynescm5 min1)
y = 46.02 0.003 x r=0.2636 P=0.0298

Figure1. Scatter plots showing the correlation between left ventricular ejection fraction (EF) and Non-Invasive Cardiac
System (NICaS) derived cardiac output index (NI-CI) [A]; EF and NICaS derived stroke volume index (NI-SVI) [B], EF and
NICaS derived total systemic vascular resistance (NI-TSVR) [C]. The 2-tailed Pearsons correlation test (r) was used for the
analyses.

10000
A B
10000

1000 1000
BNP (pg/ml)

BNP (pg/ml)

100 100

10 10

1 1
1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 10 20 30 40 50 60
NI-CI (Lmin m ) 1 2
NI-SVI (ml/m2)
Log(y) = 3.20 0.38 x r=0.4585 P<0.001 Log(y) = 3.20 0.028 x r=0.4737 P<0.0001

10000
C
1000
BNP (pg/ml)

100

10

1
1000 2000 3000 4000 5000 6000

NI-TSVR (dynescm5 min1)

Log(y) = 1.33 + 0.0003 x r=0.4803 P<0.0001

Figure2. Scatter plots showing the correlation between the logarithmic serum B-type natriuretic peptide (BNP) levels
with NI-CI [A]; BNP and NI-SVI [B], BNP and NI-TSVR [C]. The 2-tailed Pearsons correlation test (r) was used for the
analyses. NI, non-invasive; CI, cardiac output index; SVI, stroke volume index; TSVR, total systemic vascular resistance.

Circulation Journal Vol.73, June 2009

46 of 158.
 Bioimpedance Monitoring for Outpatient Heart Failure 1077

10

A B
10

9
9

8
8

ET (METs)

ET (METs)
7 7

6 6

5 5

4 4

3 3

2 2
1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 10 20 30 40 50 60
NI-CI (Lmin1 m2) NI-SVI (ml/m2)
Spearmans coefficient of rank correlation (rho) = 0.412 P=0.0025 Spearmans coefficient of rank correlation (rho) = 0.575 P<0.0001

10
C
9

8
ET (METs)

2
1000 2000 3000 4000 5000 6000

NI-TSVR (dynescm5 min1)

Spearmans coefficient of rank correlation (rho) = 0.357 P=0.0087

Figure3. Scatter plots showing the correlation between exercise tolerance threshold (ET) and NI-CI [A]; ET and NI-
SVI [B], and ET and NI-TSVR [C]. Spearmans coefficient of rank correlation test (rho) was used for analysis. NI, non-
invasive; CI, cardiac output index; SVI, stroke volume index; TSVR, total systemic vascular resistance.

3.200.38*NI-CI, r=0.4585, P=0.0001), NI-SVI (log BNP= 4.0

3.200.028*NI-SVI, r=0.4737, P<0.0001), and NI-TSVR

3.5
(log BNP=1.33+0.0003*NI-TSVR, r=0.4803, P<0.0001;
Figure2). There was a significant correlation between ET 3.0
and NI-CI (rho=0.412, P=0.0025), NI-SVI (rho=0.575, P<
0.0001), and NI-TSVR (rho=0.357, P=0.0087; Figure3). 2.5
There was no significant relationship between ET and E/A
(Figure4). 2.0
E/A

1.5
Predictive Factors for CHF Readmission
During the 10098 days (95%CI for the mean: 46154) 1.0
follow-up, 15 patients were readmitted to our hospital
because of CHF recurrence (readmission group) whereas 0.5

53 patients were not readmitted because of CHF recurrence

0.0
(no-readmission group). There was a significant difference in 2 4 6 8 10
NI-CI between the readmission and non-readmission groups ET (METs) P=NS
(1.860.37Lmin1m2 vs 2.920.63Lmin1m2, P<
0.0001). Similarly, there were significant differences between Figure4. Scatter plots showing the correlation between exercise
tolerance threshold (ET) with E/A of transmitral flow derived by
groups in NI-SVI (24.27.9mlmin1m2 vs 40.4mlmin1 Doppler echocardiogram. Spearmans coefficient of rank correlation
m2, P<0.0001), NI-TSVR (3,7861,162dynescm5m2 vs test (rho) was used for analysis.
2,420583dynescm5m2, P=0.0007), LVEF (27.48.4%
vs 39.311.7%, P=0.0005), ET (2.893.11 metabolic equiv-
alents vs 5.007.00 metabolic equivalents in 95%CI for plasma BNP were significant predictors of readmission
median, P<0.0001), and BNP (696.5pg/ml [95%CI; 487.5 (Table3).
995.1] vs 98.1pg/ml [95%CI; 69.8138.1], P<0.0001).
Receiver operator characteristic curves for sensitivity, spec-
ificity, positive predictive value, and negative predictive Discussion
value for readmission for each parameter are shown in We have previously reported the reliability of NICaS-
Table2. Multivariate analysis for readmission showed that derived hemodynamic parameters when compared with
LVEF measured by echocardiography, NI-CI, NI-SVI, and those obtained by the SwanGanz catheter. Overall, 2-tailed

Circulation Journal Vol.73, June 2009

47 of 158.
 1078 TANINO Y et al.

Table2. Receiver Operating Curve Analysis for Readmission

Sensitivity Specificit PPV NPV
AUC Cut-off
(%) (%) (%) (%)
BNP 93.3 88.7 70.0 97.9 0.95 344pg/ml
NI-CI 100 79.3 57.7 100 0.93 2.49Lmin1m2
NI-SVI 80 98.1 92.3 94.4 0.91 27.2ml/m2
NI-TSVR 100 68.6 46.7 100 0.89 2,597dynecm5min1
LVEF 86.6 64.2 40.6 94.4 0.80 37%
ET 100 71.4 52.0 100 0.88 4 METs

PPV, positive predictive value; NPV, negative predictive value; AUC, area under curve; BNP, B-type natriuretic peptide; NI-CI,
NICaS derived cardiac index; NI-SVI, NICaS derived stroke volume index; NI-TSVR, NICaS derived total systemic vascular resis-
tance; LVEF, left ventricular ejection fraction; ET, exercise tolerance threshold.

Table3. Multivariate Analysis for Factors of Readmission important that NICaS is compatible with the conventional
OR (95% confidence interval P value hemodynamic indicators derived by the SwanGanz cathe-
ter. Other hemodynamic indices including CI, SV are used
Age 1.0429 (0.95301.1414) 0.3611 when patients are in the intensive care unit. Once the patients
Gender 0.3235 (0.05931.7647) 0.1923 are moved to the general ward, however, these indices are
LVEF 0.8918 (0.82050.9693) 0.007069
NI-CI 0.0318 (0.00230.4421) 0.01023 not readily available, although there are some limited
NI-SVI 0.8121 (0.69140.9540) 0.01129 modalities similar to echocardiography. Even when the
BNP 1.0068 (1.00191.0117) 0.006053 patients are in the general ward or an outpatient department,
Heart rate 1.0456 (0.99051.1038) 0.1064 the hemodynamic indices derived by NICaS can be easily
SBP 1.0118 (0.97111.0542) 0.5750 taken and linked directly with the hemodynamic data derived
OR, odds ratio; SBP, systolic blood pressure. Other abbreviations as in by conventional methods. Therefore, another advantage of
Table2. NICaS is that patient hemodynamic data obtained inside
and outside the intensive care unit can be reported using the
same scale, allowing for easy comparison.
Pearsons correlation and BlandAltman limits of agree- According to the Ministry of Health, Labour, and Welfare
ment between NI-CO and thermodilution CO were r=0.91 in Japan, the number of outpatients with CHF has been
and 1.06 and 0.68L/min in 2SD and between NI-CO and increasing over the past several years.12 Recent studies have
Fick CO, r=0.80 and 1.52 and 0.88L/min in 2SD, respec- been initiated to better determine the actual prevalence of
tively.6 In the present study, NICaS-derived hemodynamic CHF in Japanese communities.18,19 Therefore, a less inva-
parameters were significantly correlated with LVEF, plasma sive method that is easier to use, sufficiently accurate, and
log BNP, and ET. Furthermore, NICaS-derived CI and SVI inexpensive is needed to categorize CHF patients accord-
were significant parameters for predicting the future recur- ing to their disease severity or probability of readmission.
rence of heart failure. An NICaS-derived CI of less than The exercise tolerance test is a major index for estimating a
2.49 Lmin1m2 had 100.0% sensitivity and 79.3% speci- patients practical cardiovascular performance, but there
ficity, and an NICaS-derived SVI of less than 27.2ml are a number of patients who cannot perform an exercise
min1m2 had 80.0% sensitivity and 98.1% specificity. In test because of their age, disability, or poor muscle con-
the readmission group, no cases of a preserved EF greater ditioning. In such cases, an alternative evaluation standard
than 55% were observed. ET was significantly worse in the of cardiovascular performance, such as NICaS, should be
readmission group than in the non-readmission group. The utilized.
overall average EF was 36.68%, and the average EF was
27.40% in the readmission group and 39.30% in the non- Study Limitations
readmission group. In this study, the NICaS-derived indices It is widely believed that exercise capacity is related to
seem to be good predictors of readmission with CHF second LV diastolic function20 but not to the LVEF, which in the
to BNP. present study was closely correlated to CI and SVI. We did
Serum BNP levels can be a good surrogate indicator for not, however, find a significant relationship between E/A
the long-term prognosis of CHF patients.16 Logeart et al. and ET. Also, in our study subjects, 3 patients were thought
reported that a serum BNP of more than 350ng/L at pre- to have a purely diastolic dysfunction because they had a
discharge predicted a worse prognosis in outpatients with preserved EF of more than 55% and an E/A with the trans-
CHF.17 According to our data, BNP had 93.3% sensitivity mitral Doppler echocardiogram of less than 1 during their
and 88.7% specificity for readmission when its cut-off value first visit to the outpatient department, and were not in the
was 344pg/ml, indicating that BNP had good predictive readmission group. Thus, in our study subjects, diastolic
power. Thus, in the present study, NICaS did not have a dysfunction did not have a significant effect, probably
decisive advantage over serum plasma BNP levels for pre- because of the small number of subjects used.
dicting readmission. In clinical use, however, NICaS has Although NICaS might be useful for estimating cardiac
sufficient predictive power. BNP measurement requires a performance in patients who are not exercise tolerant or who
peripheral venous blood sample, but some patients might have difficulty walking, it might not be reliable when applied
consider repeated blood sample collection an invasive pro- to patients who match the exclusion criteria described above,
cedure. NICaS has the advantage of not being invasive including those with arteriosclerosis obliterans and severe
because it only requires some electrodes to be placed on the pitting edema of the limbs.
patients skin, similar to an electrocardiogram. It is also

Circulation Journal Vol.73, June 2009

48 of 158.
 Bioimpedance Monitoring for Outpatient Heart Failure
Conclusions
NICaS-derived hemodynamic parameters obtained by
monitoring whole body bioimpedance are applicable for the
non-invasive assessment of cardiac function for the follow
up of outpatients with CHF.
References
1. Okamoto H, Tsutsui H. The epidemiology of heart failure in Japan.
Nihon Rinsho Extra Issue 2007; 913: 4954.
2. Okamoto H, Kitabatake A. Epidemiology of heart failure in Japan.
Nihon Rinsho 2003; 61: 709714.
3. Solomon SD, Anavekar N, Skali H, McMurray JJ, Swedberg K,
Yusuf S, et al. Influence of ejection fraction on cardiovascular out-
comes in a broad spectrum of heart failure patients. Circulation 2005;
112: 37383744.
4. Koglin J, Pehlivanli S, Schwaiblmair M, Vogeser M, Cremer P,
vonScheidt W. Role of brain natriuretic peptide in risk stratification
of patients with congestive heart failure. J Am Coll Cardiol 2001; 38:
19341941.
5. Berger R, Huelsman M, Strecker K, Bojic A, Moser P, Stanek B, et
al. B-type natriuretic peptide predicts sudden death in patients with
chronic heart failure. Circulation 2002; 105: 23922397. 17. Logeart D, Thabut G, Jourdain P, Chavelas C, Beyne P, Beauvais F.
6. Paredes OL, Shite J, Shinke T, Watanabe S, Otake H, Matsumoto D,
et al. Impedance cardiography for cardiac output estimation: Reliabil-
ity of wrist-to-ankle electrode configuration Circ J 2006; 70: 1164
1168.
7. Cotter G, Moshkovitz Y, Kaluski E, Kohen A, Miller H, Goor D, et
al. Accurate, noninvasive continuous monitoring of cardiac output by
whole-body electrical bioimpedance. Chest 2004; 125: 14311440.
8. Cohen AJ, Arnaudov D, Zabeeda D, Shultheis L, Lashinger J,
Schachner A. Non-invasive measurement of cardiac output during
coronary artery bypass grafting. Eur J Surg 1998; 14: 6469.
9. Organ LW, Bradham GB, Gore DT, Lozier SL. Segmental bioelectrical
impedance analysis: Theory and application of a new technique. J Appl
Physiol 1994; 77: 98112.
10. Lukaski H, Bolonchuk WW, Hall CB, Siders WA. Validation of tet-
rapolar bioelectrical impedance method to assess human body com-
position. J Appl Physiol 1986; 60: 13271332.
Circulation Journal Vol.73, June 2009
49 of 158.
1079
11. Hoffer EC, Meador CK, Simpson DC. A relationship between whole
body impedance and total body water volume. Ann NY Acad Sci 1970;
170: 452461.
12. Ward LC, Heitmann BL, Craid P, Stroud D, Azinge EC, Jebb S, et al.
Association between ethnicity, body mass index, and bioelectrical
impedance: Implications for the population specificity of prediction
equations. Ann NY Acad Sci 2000; 904: 199204.
13. Hamwi GT, Danowski TS. Changing dietary concepts in diabetes
mellitus: Diagnosis and treatment. New York: American Diabetes
Association; 1964; 7378.
14. Yasue H, Yoshimura M, Sumida H, Kikuta K, Kugiyama K, Jougasaki
M, et al. Localization and mechanism of secretion of B-type natri-
uretic peptide in comparison with those of A-type natriuretic peptide
in normal subjects and patients with heart failure. Circulation 1994;
90: 195203.
15. Sasayama S, Asanoi H, Ishizaka S, Miyagi K. Evaluation of functional
capacity of patients with congestive heart failure. In: Yasuda H,
Kawaguchi H, editors. New aspects in the treatment of failing heart.
Tokyo: Springer-Verlag; 1992; 113117.
16. Tsutamoto T, Wada A, Maeda K, Hisanaga T, Mabuchi N, Hayashi
M, et al. Plasma brain natriuretic peptide level as a biochemical
marker of morbidity and mortality in patients with asymptomatic or
minimally symptomatic left ventricular dysfunction: Comparison with
plasma angiotensin II and endothelin-1. Eur Heart J 1999; 20: 1799
1807.
B-Type natriuretic peptide assay for identifying patients at high risk
of re-admission after decompensated heart failure. J Am Coll Cardiol
2004; 43: 635641.
18. Tsutsui H, Tsuchihashi-Makaya M, Kinugawa S, Goto D, Takeshita
A; The JCARE-CARD Investigators. Clinical characteristics and
outcome of hospitalized patients with heart failure in Japan. Circ J
2006; 70: 16171623.
19. Tsutsui H, Tsuchihashi-Makaya M, Kinugawa S, Goto D, Takeshita
A; JCARE-GENERAL Investigators. Characteristics and outcomes
of patients with heart failure in general practices and hospitals. Circ J
2007; 71: 449454.
20. Parthenakis FI, Kanoupakis EM, Kochiadakis GE, Skalidis EI, Mezilis
NE, Simantirakis EN. Left ventricular diastolic filling pattern pre-
dicts cardiopulmonary determinants of functional capacity in patients
with congestive heart failure. Am Heart J 2000; 140: 338344.
 Europace (2007) 9, 744750
doi:10.1093/europace/eum086

Impedance cardiography: a useful and reliable tool in

optimization of cardiac resynchronization devices
Konstantin M. Heinroth*, Marcel Elster, Sebastian Nuding, Frithjof Schlegel, Arnd Christoph,
Justin Carter, Michael Buerke, and Karl Werdan
Department of Medicine III/Cardiology, Martin-Luther-University Halle-Wittenberg, Ernst-Grube-Strae 40,
06097 Halle (Saale), Germany
Received 18 October 2006; accepted after revision 7 April 2007; online publish-ahead-of-print 11 May 2007

KEYWORDS Aims Optimizing cardiac resynchronization therapy (CRT) devices has become more complex since
Cardiac resynchronization modication of both atrioventricular (AV) and interventricular (VV) stimulation intervals has become poss-

Downloaded from http://europace.oxfordjournals.org/ by guest on May 25, 2012

therapy;
Non-invasive optimization;
Impedance cardiography
ible. The current paper presents data from the routine use of impedance cardiography (IC)-based cardiac
output (CO) measurements to guide the optimization of AV- and VV-interval timing of CRT devices.
Methods and results Forty-six patients with heart failure (left ventricular ejection fraction ,35%, New York
Heart Asociation (NYHA) IIIIV) and left bundle branch block (.130 ms) in sinus rhythm were evaluated 35
days after implantation of a CRT device by means of IC. CO was measured without pacing and with biven-
tricular pacing using a standard protocol of VV- and AV-interval modication from 260 to 60 ms and 80 to
140 ms, respectively, in 20 ms steps. Mean CO without pacing was 3.66 + 0.85 L/min and signicantly
increased to 4.40 + 1.1 L/min (P , 0.05) with simultaneous biventricular pacing and an AV interval of
120 ms. Optimizing both VV and AV intervals further increased CO to 4.86 + 1.1 L/min (P , 0.05).
Maximum CO was measured in most patients with left ventricular pre-excitation. The proportion of non-
responders to CRT was reduced by 56% following AV- and VV-interval modication using IC guidance.
Conclusion Modication of both AV and VV intervals in patients with a CRT device signicantly improves CO
compared with standard simultaneous biventricular pacing and no pacing. IC is a useful non-invasive tech-
nique for guiding this modication. Marked variability of optimal AV and VV intervals between patients
requires optimization of these intervals for each patient individually.

Introduction
Cardiac resynchronization therapy (CRT) has developed from
an experimental method1 to an established adjunctive
treatment for patients with advanced heart failure. CRT
aims to improve cardiac output (CO) by lessening the inter-
and intra-ventricular conduction delay caused in part by
left bundle branch block (LBBB).2,3 CRTreduces clinical symp-
toms of heart failure and hospitalization,46 improves haemo-
dynamic parameters (including ventricular performance),7
and has been shown to reduce mortality.810 Acceptance of
this therapy is reected by its incorporation into current
guidelines for the management of heart failure.11,12
Successful CRT in any given patient depends upon many
variables, such as the appropriate positioning of the LV
lead,1315 and also on achieving optimal biventricular stimu-
lation timing.16,17 This latter variable has attracted particu-
lar interest recently in an attempt to reduce the substantial
proportion (2030%) of patients who derive no apparent
benet from CRT, despite meeting appropriate implantation
criteria and having had technically straightforward device
implantation (non-responders).18
* Corresponding author. Tel: 49 345 557 2601; fax: 49 345 557 2072.
E-mail address: konstantin.heinroth@medizin.uni-halle.de
CRT devices have become signicantly more complex
recently. Many now have the facility to programme different
atrioventricular (AV) and interventricular (VV) stimulation
timing intervals.19,20 Data exist to show that tailoring
these intervals to suit the patient in hand further improves
the haemodynamic benets brought by CRT. However, ques-
tions remain as to which method of haemodynamic assess-
ment is best for guiding the adjustment of the device
settings to suit any given post-implant patient.
In the measurement of cardiac function, invasive methods
(e.g. dp/dt estimation of contractility or the thermodilution
method for CO) are the gold standards but are not suitable
for routine use during routine follow-up of patients with
implanted CRT devices. Non-invasive methods used in opti-
mizing device settings include echocardiographic tech-
niques,2124 radionucleotide ventriculography,25 nger
photoplethysmography,26 and more recently, impedance
cardiography (IC).6,27
IC is an established technique for haemodynamic assess-
ment and is capable of calculating CO on a beat-to-beat
basis.28 It relies upon changes in impedance (resistance) to
current ow through the chest between strategically
placed electrodes. Given that most current takes the path

& The European Society of Cardiology 2007. All rights reserved. For Permissions, please e-mail: journals.permissions@oxfordjournals.org

50 of 158.
 Impedance cardiography 745

of least resistance (principally the blood-lled aorta) and
that the impedance changes with blood ow, these variables
can be used to calculate CO. IC has been used in patients
receiving dual chamber pacemakers2931 and in the assess-
ment of patients with heart failure.6,3234 IC as a technique
may supplement or even replace invasive measurements and
Doppler echocardiography6,27,35 in the optimization of
implanted CRT devices.
This paper presents data from the routine use of IC in
optimizing CRT by adjusting AV and VV intervals to obtain
maximum CO in post-implant patients. To our knowledge,
this is the rst report on the combined manipulation of
AV- and VV-interval timing in biventricular pacing devices
using IC.
Methods
Patient characteristics
Forty-six consecutive patients (37 males, 9 females, mean age
biventricular tracking of the intrinsic sinus rhythm (VAT-Mode).36
The AV interval was set at 120 ms as a standard value6 without LV
or RV pre-excitation (VV interval 0). This standard pacing
set-up was kept from implantation until optimization.
Impedance cardiography
Optimization of biventricular pacing was performed using a com-
mercially available system for IC (Task Force Monitor Systems,
CNSystems, Graz, Austria) as described by Braun et al. 6 Two electro-
des were placed bilaterally to the inferior chest wall in combination
with one electrode at the neck. Low-amplitude high-frequency
current was delivered via these surface electrodes, and transthor-
acic impedance (resistance) to this current ow was measured.
Changes in transthoracic impedance (mainly inuenced by changes
of systolic aortic blood ow) were measured by means of four
additional surface electrodes: one pair placed bilaterally to the
sternum and the second pair bilaterally to the abdomen. Cardiac
output was calculated on a beat-to-beat basis from the trans-
thoracic impedance signal.37 Figure 1 shows an example of the IC

Downloaded from http://europace.oxfordjournals.org/ by guest on May 25, 2012

63 + 9 years) with heart failure [left ventricular ejection fraction measurement acquired by the Task Force Monitor System.
(LVEF) ,35%; NYHA IIIIV], LBBB (.130 ms), and sinus rhythm
were evaluated 35 days after implantation of a CRT device.
Baseline characteristics of study patients are detailed in Table 1.
All patients had been receiving optimal (guideline compliant11,12)
medical treatment for heart failure for at least 1 month before
device implantation. Evaluation of patients before CRT included
12-lead surface ECG. Echocardiography was undertaken pre-
implant, and post-CRT device optimization according to a standard
template was performed with a System FiVe (GE VingMed) machine
coupled to a 2.5 MHz transducer. In all patients, measurement of
left ventricular (LV) dimensions, LVEF and evaluation for valve
disease, particularly mitral regurgitation was undertaken.
Transmitral ow was assessed using pulsed wave Doppler imaging
in the apical 4-chamber view. Coronary angiography was performed
in patients who had not previously had this done or where patients
had symptoms of active coronary disease requiring further imaging.
The echocardiography and coronary angiography form part of our
standard patient work-up for device implantation, but these
results were not part of the prospectively chosen datasets gathered
for the purpose of the study.

Device implantation
All patients received a CRT device in combination with a
cardioverterdebrillator (Contak Renewal, Guidant, St. Paul, MN,
USA) except one patient who received only a CRT device (Contak
Renewal). Forty-ve of the 46 CRT devices were implanted in our
catheterization laboratory by a cardiologist and a cardiac surgeon.
The other patient received a device with epicardial leads implanted
during surgical mitral valve reconstruction. Device implantation was
similar in all cases. Initially the right ventricular (RV) lead was
placed in the RV apex, then the coronary sinus lead was positioned
using the over-the-wire technique, and nally the right atrial lead
was implanted. The device was programmed in DDD-Mode with a
lower rate limit at 4050 bpm, producing atrial synchronous

Table 1 Patient characteristics

Patients 46
Age (years) 63 + 9 Figure 1 Example of the impedance cardiography measurement
Gender (female/male) 9/37 acquired by the Task Force Monitor System. There are two paced
% CAD/% DCM 20/80 beats and one beat without pacing. From top to bottom: lead II
QRS 187 + 26 ms from the surface ECG, blood pressure, and the rst derivative dZ/
PQ 197 + 40 ms dt of the impedance. DBP, diastolic blood pressure; SBP, systolic
LVEF 27 + 8% blood pressure; B, aortic valve opening; dZ/dt max, maximum of
the rst derivative of the impedance signal; X, aortic valve closure.

51 of 158.
 746 K.M. Heinroth et al.

Pacing study protocol
To optimize the CRT set-up, all patients were examined in
the supine position in a silent environment to reduce the
impact of sympathetic activation by external stimuli.
A standard protocol involving a period of stabilization and
equilibration followed by VV-interval optimization and then
AV-interval optimization was employed in all patients.
Pacemakers were programmed in a DDD-Mode with a lower
rate limit of 40 bpm to avoid effects of atrial pacing on
the AV interval.36 During data acquisition, telemetry
between the implanted device and the programmer was dis-
connected to prevent interference with the measurement of
impedance. AV-interval values relate to the atrio-RV stimu-
lation interval, and VV intervals relate to interventricular
interval with negative gures implying LV pre-excitation.
The rst stage of the pacing protocol was a period of
stabilization and equilibration. The baseline CO without
pacing was recorded, and this was alternated three times
with standard simultaneous biventricular pacing with an
There were no signicant differences between the heart
rates without pacing (75.1 + 10.7 bpm), with initial simul-
taneous biventricular pacing (74.7 + 10.9 bpm), or opti-
mized biventricular pacing (74.8 + 10.2 bpm, P . 0.05).
Thus, results for CO obtained at each CRT set-up were unaf-
fected by the heart rate.
Cardiac output without pacing vs. simultaneous
biventricular pacing vs. optimized biventricular
pacing
The alternation between three cycles of no pacing with sim-
ultaneous biventricular pacing during stabilization and equi-
libration revealed (across all patients) a mean increase in CO
of 21.6 + 22.1% (P , 0.05) with pacing. The mean range
(i.e. variation) between the three recordings (in each
patient) without pacing was 10.1% and between the three
recordings with pacing was 11.8%.
Figure 2 shows data obtained for a typical patient at

Downloaded from http://europace.oxfordjournals.org/ by guest on May 25, 2012

different VV intervals with a xed AV interval of 120 ms. In
AV interval of 120 ms for 50 s at each setting. Once consis-
this patient, the maximum CO was measured during LV pre-
tent values for CO in both modes were conrmed, we pro-
excitation of 20 and 40 ms relative to RV stimulation. Still,
ceeded to the optimization stage of the pacing protocol.
earlier LV pre-excitation, simultaneous biventricular
With the AV interval xed at 120 ms, VV intervals were
pacing, and RV pre-excitation resulted in a lower CO.
adjusted through a set of seven steps ranging from LV pre-
Data collected from all patients at each stimulation set-up
excitation of 260 ms to an LV delay of 60 ms in steps of
with VV optimization at a xed AV interval of 120 ms and
20 ms (as reported previously38). The CO was recorded for
with AV optimization at a VV interval set to zero are
50 s at each setting. Then, the VV interval was set at zero
shown in Figure 3.
while the AV interval was adjusted from 80 to 140 ms step-
In detail, the maximum CO was achieved with the follow-
wise in 20 ms increments. Again, CO at each setting was
ing VV intervals (when combined with an AV interval of
recorded. We found that 10 s were sufcient for stabilization
120 ms): 260 ms in 15 pts, 240 ms in 6 pts, 220 ms in
in values after any change in the stimulation set-up. Finally,
9 pts, +0 ms in 5 pts, 20 ms in 2 pts, 40 ms in 2 pts,
the device was set at whichever combination of AV and VV
and 60 ms in 3 pts, so the majority of patients (30 of
intervals produced highest CO in that patient.
46 patients) achieved peak CO with LV pre-excitation. The
We dened an increase in CO of greater than 10% above
remaining four patients achieved maximum CO with simul-
baseline (i.e. without pacing) as a positive response to
taneous biventricular pacing, two with an AV interval of
biventricular pacing, in keeping with data previously pub-
80 ms, and two with an AV interval of 140 ms.
lished.39 Where maximum CO was identied by IC to occur
The combination of an AV interval of 120 ms and LV
at very short AV intervals (i.e. A to RV or A to LV of
pre-excitation of 40 ms yielded the highest mean CO of
,80 ms), an echocardiogram was performed to exclude a
truncated A wave caused by atrial contraction against
closed AV valves.

Statistics
All data are presented as means + standard error. Statistical
analysis was performed using multiple analysis of variance,
KruskallWallis test, and Fishers exact test to compare
more than two sets of data. For a comparison of two sets
of data, a students t-test was performed. A P-value of
, 0.05 was considered to be signicant. Data was processed
using commercially available software (Statgraphics Plus for
Windows).

Results
Cardiac resynchronization therapy
Forty-six CRT devices were successfully implanted, 45 by the
transvenous approach and 1 by the transthoracic approach
during surgical mitral valve repair. The LV lead was placed
in a posterolateral (n 35) or an anterolateral (n 10)
side branch. There were no complications.
Figure 2 Cardiac output measured by impedance cardiography at
different ventriculo-ventricular intervals at an atrioventricular
interval of 120 ms in a representative patient. Optimum cardiac
output was yielded at a left ventricular pre-excitation of 240 and
220 ms.

52 of 158.
 Impedance cardiography 747

Figure 3 Cardiac output measured by impedance cardiography at different ventriculo-ventricular and atrioventricular intervals for all
patients. Data are presented as mean + standard deviation. Maximum cardiac output was yielded at a left ventricular pre-excitation of
240 ms and an atrioventricular interval of 120 ms.

Downloaded from http://europace.oxfordjournals.org/ by guest on May 25, 2012

Figure 4 Cardiac output measured by impedance cardiography without pacing, with simultaneous biventricular pacing (and an atrioventri-
cular interval of 120 ms), and with optimized biventricular pacing. Asterisk represents P ,0.05, simultaneous biventricular pacing vs. no
pace; ash represents P ,0.05, optimized biventricular pacing vs. simultaneous biventricular pacing and vs. no pace.

4.57 + 1.1 L/min. Similar values were obtained by LV pre-
excitation of 20 and 60 ms (4.51 + 1.3 and 4.50 + 1.2 L/
min, respectively). The optimal pacing set-up varied
widely from patient to patient. As a result of this variability,
there was no signicant difference in CO when the results
from different AV and VV intervals were compared. No
single combination of AV and VV intervals could be recom-
mended for application to the whole study population,
because no single combination of intervals showed statisti-
cally signicant superiority over other combinations.
For the population as a whole, when the CO obtained with
optimized biventricular pacing (i.e. the CO measured at
whichever AV and VV intervals produced the highest value
in that patient) is compared with the CO obtained with
standard simultaneous biventricular pacing and compared
with the CO obtained with no pacing, a statistically signi-
cant difference is seen (Figure 4). CO without pacing
was 3.66 + 0.85 L/min. CO increased to 4.40 + 1.1 L/min
(P , 0.05) with simultaneous biventricular pacing using a
standard AV interval of 120 ms.
The mean increase in CO changing from no pacing to sim-
ultaneous biventricular pacing was 21.6 + 22.1%, P , 0.05).
Optimizing VV and AV intervals further increased the mean
CO to 4.86 + 1.1 L/min. This corresponds to an increase in
mean CO of 32.8% without pacing and an increase of 11.2%
compared with simultaneous biventricular pacing, all three
set-ups resulting in signicantly different CO (P , 0.05,
Figure 4).
Taking an increase in CO 10% as a denition of a positive
haemodynamic response to CRT,39 16 of the 46 patients
(35%) were non-responders with standard simultaneous
biventricular CRT. In ve of these patients, CO with standard
biventricular pacing was lower than without any pacing. The
mean CO in this group of 16 non-responders (with standard
simultaneous biventricular CRT) was 100.8 + 4.9% (where
CO without pacing was 100%). After optimization, 9 of
these 16 patients (who were non-responders to standard
simultaneous biventricular CRT) experienced an increase
of 10% in CO to become responders. In this group of
nine patients, the mean CO increased signicantly from
101.3 + 3.6% to 121 + 5.1% (P  0.05). In the seven patients
who failed to respond despite optimization, the mean CO
was 106.6 + 2.6% with optimized pacing. Nevertheless,
after modication of AV and VV intervals to produce the

53 of 158.
 748
Figure 5 Comparison of the percentage of non-responders to
cardiac resynchronization therapy (denition see text) with simul-
taneous biventricular pacing (with an atrioventricular interval of
120 ms) vs. atrioventricular- and ventriculo-ventricular-interval
optimized biventricular pacing according to the cardiac output
measured by impedance cardiography for the whole study
population.
maximum CO, the number of haemodynamic non-
responders was reduced to 7 of the 46 patients, i.e. the
number of non-responders was reduced by 56% (Figure 5).
Discussion
In the present study, we described the use of IC to guide
optimization of the AV and VV timing following CRT. To our
knowledge, this is the rst report of both AV and VV manipu-
lation using IC as a guide in CRT optimization. We found IC to
be simple to apply and capable of yielding rapid results in a
reliable and reproducible fashion. We found that the
optimal CRT-timing settings varied substantially between
patients underlining the need to optimize each patients
device individually in order to gain most benet from
device. Using this technique, we were able to signicantly
improve CO by manipulating both AV and VV intervals. By
optimizing both settings, we were able to further improve
the CRT-related increase in CO in responders and, further-
more, could produce a signicant increase in CO in patients
who, with standard or simultaneous biventricular pacing,
had demonstrated no increase in CO (non-responders). In
part, this reects the importance of electrical reposition-
ing of the LV lead by VV-interval manipulation. Thus,
using this technique, we were able to reduce the rate of
non-responders to CRT by 56% (from 35 to 15%) by IC haemo-
dynamic criteria.
Assessment of optimal biventricular stimulation:
current methods
Although invasive measurements of CO and other par-
ameters remain the gold standard for the evaluation of
heamodynamics, they are not suitable for routine follow-up
and optimization of CRT device settings because such inva-
sive procedures are unpleasant for the patient and have
potentially serious complications.8,40
54 of 158.
K.M. Heinroth et al.
Most units favour non-invasive methods for the optimiza-
tion of CRT devices. The use of echocardiographic tech-
niques for this purpose has received more attention in the
literature, although there is no consensus as to which of
the many echo-based parameters is the best surrogate for
CO in the context of CRT optimization.
Mitral41 and aortic42 valve Doppler velocity time integrals
(VTI) as well as several other24 echocardiographic para-
meters have been assessed as an alternative to invasive
measures and as a surrogate for CO in the optimization of
pacing devices. Recently, there has been promising data
using advanced tissue Doppler imaging techniques and real-
time 3-D echo (particularly in selecting the optimal lead
positioning for biventricular pacing), but despite these
advances, the echocardiographic lead optimization of CRT
remains complex and time consuming. It requires an experi-
enced operator and has signicant problems relating to
reproducibility and objectivity.
Downloaded from http://europace.oxfordjournals.org/ by guest on May 25, 2012
Impedance cardiography: a reliable
alternative technique?
Several studies have demonstrated that IC is capable of pro-
viding a reliable and accurate measurement of CO when
compared with invasive methods.28,32,43 The utility of IC
has been shown in patients with decompensated cardiac
failure33 and in optimizing dual chamber pacemakers.44
A direct comparison of IC and echo techniques in optimizing
AV intervals for pacemakers has been made in several
studies30,31 with a recurring theme being that optimal AV
intervals calculated by IC tend to be shorter than those
obtained by echocardiography.45 Recently, Braun et al. 6 pro-
vided data using IC to guide the manipulation (primarily) of
AV intervals and showed that IC-based optimization is com-
parable with transaortic VTI, but IC was felt to be more sen-
sitive to small changes in CO and easier to apply.
Limitations of impedance cardiography
The limitations of IC was usefully reviewed by Kinderman.46
In a study of 14 patients with dual chamber pacemakers,
data from IC overestimated CO if very short AV intervals
were programmed. This phenomenon seems to be attribu-
table to a decrease in the thoracic impedance caused by a
retrograde ow into the great thoracic veins induced by
atrial contraction against closed AV valves. In these cases,
optimization of AV interval solely according to the impe-
dance signal would result in a truncated A wave of the trans-
mitral ow, causing potentially deleterious effects. In our
study, if IC indicated that optimum CO was seen at more
than one AV or VV interval, the longer AV interval and the
shorter VV interval were programmed. If IC indicated
optimum CO at short AV intervals (below 80 ms), transmitral
ow was checked by Doppler echocardiography to exclude a
truncated A wave. Of the 46 patients in the present study,
this echocardiographic check was required in two,
neither of whom required a lengthening of AV interval.
Importance of optimizing both
VV and AV intervals
The early generations of CRT devices enabled programming
of the AV interval only.
 Impedance cardiography
Prospective randomized trial data have shown that
AV-interval optimization not only improves the haemo-
dynamic response to CRT, but also improves NYHA functional
class and quality of life scores.47
The current generation of devices allows the manipulation
of both AV and VV intervals.
Since activation of the interventricular septum is mainly
inuenced by the RV lead, modication of the VV interval
affects not only inter-ventricular dyssynchrony, but also
intra-ventricular dyssynchrony of the LV. The capacity to
manipulate VV intervals may thus further improve the
haemodynamic response to CRT.
There is increasing evidence that sequential biventricular
pacing is superior to simultaneous biventricular pacing in
many patients with a CRT device.
Perego et al. examined the impact of sequential biventri-
cular pacing while invasively monitoring dp/dt in both ven-
tricles simultaneously. Simultaneous ventricular pacing
produced an increase in dp/dt of 29% over baseline, but
sequential pacing produced a signicantly greater increase
of 35% (P , 0.01). The mean optimal VV interval was LV pre-
excitation of 25 ms, and there was no detriment to RV func-
tion.38 Similar ndings were reported by van Gelder et al. 20
This pattern of haemodynamic improvement with both AV
and VV optimization has been veried using Doppler echo-
cardiography,48 3-D echocardiography49 and radionuclide
ventriculography.25 The largest benet is obtained in most
cases by LV pre-excitation. Our study is the rst to validate
the previous data, showing the additional benet of optimiz-
ing both AV and VV intervals using IC as the method of
assessment.
Where no facility for optimizing CRT devices in individual
patients exists, we would recommend an AV interval of
120 ms and LV pre-excitation of 2040 ms as empirical
intervals in CRT programming. However, due to signicant
heterogeneity between patients, we were unable to identify
any single set of AV and VV intervals as being superior and
thus suitable for application to the whole CRT population
on an empirical basis. This variation between patients is to
be expected, since there is a substantial heterogeneity
between patients of conduction pattern, ventricular size,
scar tissue formation, ventricular function (especially with
respect to regional wall motion abnormalities) CRT lead pla-
cement positions, and other variables. The facility to mani-
pulate VV intervals in effect allows the operator to
electrically reposition the LV lead to help overcome some
of these variations to maximize the haemodynamic response
to the device.
This work has been an early, descriptive study of the feasi-
bility of using IC to manipulate both AV and VV intervals in
optimizing CRT devices. Further work is required before
this technique can enter mainstream use. For instance, in
any given patient, since AV and VV intervals are likely to
be interdependent, the ideal IC protocol would test every
AV-interval setting against every VV-interval setting and
vice versa. With current equipment, this would be impracti-
cal. Another simpler method of combining these interdepen-
dent values of AV and VV intervals would be to identify the
optimal VV interval (i.e. A to LV interval) and keeping this
xed (rather than re-setting this to zero as our protocol
did), and then adjusting the AV (i.e. A to V interval) to iden-
tify the optimal CO. We suspect that this method of com-
bined optimization of AV and VV intervals by IC may yield
55 of 158.
749
still higher CO values, and hence hopefully reduce the
number of non-responders. Further studies are required.
Conclusions
The present paper is the rst to report the utility of IC in the
optimization of both AV and VV intervals in biventricular
pacemaker set-up in routine clinical practice. Using IC, we
demonstrated that manipulating both AV and VV intervals
is feasible and may result in a signicant improvement in
the haemodynamic response to CRT compared with stan-
dard simultaneous biventricular pacing and no pacing. The
results show the importance of optimizing CRT set-up in
each patient individually, since the variability between
patients means that no single set of intervals can be identi-
ed as being suitable for all or even most patients.
In our opinion, IC is a promizing method of CRT optimiz-
ation and may compare well with other techniques in
routine use. It carries less risk of complications than invasive
Downloaded from http://europace.oxfordjournals.org/ by guest on May 25, 2012
techniques and is more comfortable for the patient. The
validity and reliability of IC has now been reported by
several authors.28,3133,43 Notwithstanding its limitations,
which include a tendency to relative CO overestimation
and the potential for selecting inappropriately short AV
intervals, IC has an emerging role in the optimization of
CRT parameters in routine clinical practice. Further work
is required to make best use of IC derived data in combining
AV and VV intervals to ideally suit any given patient.
Conict of interest: none declared.
References
1. Leclercq C, Cazeau S, Le BH, Ritter P, Mabo P, Gras D et al. Acute hemo-
dynamic effects of biventricular DDD pacing in patients with end-stage
heart failure. J Am Coll Cardiol 1998;32:182531.
2. Cohen TJ, Klein J. Cardiac resynchronization therapy for treatment of
chronic heart failure. J Invasive Cardiol 2002;14:4853.
3. Cazeau S, Alonso C, Jauvert G, Lazarus A, Ritter P. Cardiac resynchroniza-
tion therapy. Europace 2004;5(Suppl. 1):S428.
4. Abraham WT, Fisher WG, Smith AL, Delurgio DB, Leon AR, Loh E et al.
Cardiac resynchronization in chronic heart failure. N Engl J Med
2002;346:184553.
5. Cazeau S, Leclercq C, Lavergne T, Walker S, Varma C, Linde C et al.
Effects of multisite biventricular pacing in patients with heart failure
and intraventricular conduction delay. N Engl J Med 2001;344:87380.
6. Braun MU, Schnabel A, Rauwolf T, Schulze M, Strasser RH. Impedance
cardiography as a noninvasive technique for atrioventricular interval
optimization in cardiac resynchronization therapy. J Interv Card
Electrophysiol 2005;13: 223229.
7. Bakker PF, Meijburg HW, de Vries JW, Mower MM, Thomas AC, Hull ML
et al. Biventricular pacing in end-stage heart failure improves functional
capacity and left ventricular function. J Interv Card Electrophysiol 2000;
4:395404.
8. Bristow MR, Saxon LA, Boehmer J, Krueger S, Kass DA, De MT et al.
Cardiac-resynchronization therapy with or without an implantable de-
brillator in advanced chronic heart failure. N Engl J Med 2004;
350:214050.
9. Salukhe TV, Dimopoulos K, Francis D. Cardiac resynchronisation may
reduce all-cause mortality: meta-analysis of preliminary COMPANION
data with CONTAK-CD, InSync ICD, MIRACLE and MUSTIC. Int J Cardiol
2004;93:1013.
10. Cleland JG, Daubert JC, Erdmann E, Freemantle N, Gras D, Kappenberger L
et al. The effect of cardiac resynchronization on morbidity and mortality in
heart failure. N Engl J Med 2005;352:153949.
11. Hoppe UC, Bohm M, Dietz R, Hanrath P, Kroemer HK, Osterspey A et al.
Guidelines for therapy of chronic heart failure. Z Kardiol 2005;
94:488509.
 750
12. Hunt SA. ACC/AHA 2005 guideline update for the diagnosis and manage-
ment of chronic heart failure in the adult: a report of the American
College of Cardiology/American Heart Association Task Force on
Practice Guidelines (Writing Committee to Update the 2001 Guidelines
for the Evaluation and Management of Heart Failure). J Am Coll
Cardiol 2005;46:e182.
13. Butter C, Auricchio A, Stellbrink C, Fleck E, Ding J, Yu Y et al. Effect of
resynchronization therapy stimulation site on the systolic function of
heart failure patients. Circulation 2001;104:30269.
14. Gold MR, Auricchio A, Hummel JD, Giudici MC, Ding J, Tockman B et al.
Comparison of stimulation sites within left ventricular veins on the
acute hemodynamic effects of cardiac resynchronization therapy. Heart
Rhythm 2005;2:37681.
15. Albertsen AE, Nielsen JC, Pedersen AK, Hansen PS, Jensen HK, Mortensen
PT. Left ventricular lead performance in cardiac resynchronization
therapy: impact of lead localization and complications. Pacing Clin
Electrophysiol 2005;28:4838.
16. Auricchio A, Ding J, Spinelli JC, Kramer AP, Salo RW, Hoersch W et al.
Cardiac resynchronization therapy restores optimal atrioventricular
mechanical timing in heart failure patients with ventricular conduction
delay. J Am Coll Cardiol 2002;39:11639.
17. Doshi RN. Optimizing resynchronization therapy: can we increase the
number of true responders? J Cardiovasc Electrophysiol 2005;16(Suppl.
1):S4851.
18. Birnie DH, Tang AS. The problem of non-response to cardiac resynchroni-
zation therapy. Curr Opin Cardiol 2006;21:206.
19. ODonnell D, Nadurata V, Hamer A, Kertes P, Mohammed W. Long-term
variations in optimal programming of cardiac resynchronization therapy
devices. Pacing Clin Electrophysiol 2005;28(Suppl. 1):S246.
20. van Gelder BM, Bracke FA, Meijer A, Lakerveld LJ, Pijls NH. Effect of opti-
mizing the VV interval on left ventricular contractility in cardiac resyn-
chronization therapy. Am J Cardiol 2004;93:15003.
21. Breithardt OA, Sinha AM, Franke A, Hanrath P, Stellbrink C.
Echocardiography in cardiac resynchronization therapy: identication
of suitable patients, follow-up and therapy optimization. Herz 2003;
28:61527.
22. Bax JJ, Ansalone G, Breithardt OA, Derumeaux G, Leclercq C, Schalij MJ
et al. Echocardiographic evaluation of cardiac resynchronization
therapy: ready for routine clinical use? A critical appraisal. J Am Coll
Cardiol 2004;44:19.
23. Lane RE, Chow AW, Chin D, Mayet J. Selection and optimisation of biven-
tricular pacing: the role of echocardiography. Heart 2004;90(Suppl.
6):vi106.
24. Melzer C, Knebel F, Ismer B, Bondke H, Nienhaber CA, Baumann G et al.
Inuence of the atrio-ventricular delay optimization on the intra left ven-
tricular delay in Cardiac Resynchronization Therapy. Cardiovasc
Ultrasound 2006;4:5.
25. Burri H, Sunthorn H, Somsen A, Zaza S, Fleury E, Shah D et al. Optimizing
sequential biventricular pacing using radionuclide ventriculography.
Heart Rhythm 2005;2:9605.
26. Butter C, Stellbrink C, Belalcazar A, Villalta D, Schlegl M, Sinha A et al.
Cardiac resynchronization therapy optimization by nger plethysmogra-
phy. Heart Rhythm 2004;1:56875.
27. Tse HF, Yu C, Park E, Lau CP. Impedance cardiography for atrioventricular
interval optimization during permanent left ventricular pacing. Pacing
Clin Electrophysiol 2003;26:18991.
28. Fortin J, Habenbacher W, Heller A, Hacker A, Grullenberger R, Innerhofer J
et al. Non-invasive beat-to-beat cardiac output monitoring by an improved
method of transthoracic bioimpedance measurement. Comput Biol Med
2006;36:1185203.
29. Ovsyshcher I, Gross JN, Blumberg S, Furman S. Precision of impedance
cardiography measurements of cardiac output in pacemaker patients.
Pacing Clin Electrophysiol 1992;15:19236.
30. Kindermann M, Frohlig G, Doerr T, Schieffer H. Optimizing the AV delay in
DDD pacemaker patients with high degree AV block: mitral valve Doppler
versus impedance cardiography. Pacing Clin Electrophysiol
1997;20:245362.
31. Kolb HJ, Bohm U, Mende M, Neugebauer A, Pfeiffer D, Rother T. -
Assessment of the optimal atrioventricular delay in patients with
56 of 158.
K.M. Heinroth et al.
dual-chamber pacemakers using impedance cardiography and Doppler
echocardiography. J Clin Basic Cardiol 1999;2:23740.
32. Belardinelli R, Ciampani N, Costantini C, Blandini A, Purcaro A. -
Comparison of impedance cardiography with thermodilution and direct
Fick methods for noninvasive measurement of stroke volume and
cardiac output during incremental exercise in patients with ischemic car-
diomyopathy. Am J Cardiol 1996;77:1293301.
33. Albert NM, Hail MD, Li J, Young JB. Equivalence of the bioimpedance and
thermodilution methods in measuring cardiac output in hospitalized
patients with advanced, decompensated chronic heart failure. Am J
Crit Care 2004;13:46979.
34. Adachi H, Hiratsuji T, Sakurai S, Tada H, Toyama T, Naito S et al.
Impedance cardiography and quantitative tissue Doppler echocardiogra-
phy for evaluating the effect of cardiac resynchronization therapy: a
case report. J Cardiol 2003;42:3742.
35. Santos JF, Caetano F, Parreira L, Madeira J, Cardoso P, Fonseca N et al.
Tissue Doppler echocardiography for evaluation of patients with ventricu-
lar resynchronization therapy. Rev Port Cardiol 2003;22:136371.
36. Bernheim A, Ammann P, Sticherling C, Burger P, Schaer B, Brunner-La
Rocca HP et al. Right atrial pacing impairs cardiac function during resyn-
chronization therapy: acute effects of DDD pacing compared to VDD
pacing. J Am Coll Cardiol 2005;45:14827.
37. Ovsyshcher I, Furman S. Impedance cardiography for cardiac output esti-
mation in pacemaker patients: review of the literature. Pacing Clin
Downloaded from http://europace.oxfordjournals.org/ by guest on May 25, 2012
Electrophysiol 1993;16:141222.
38. Perego GB, Chianca R, Facchini M, Frattola A, Balla E, Zucchi S et al.
Simultaneous vs. sequential biventricular pacing in dilated cardiomyopa-
thy: an acute hemodynamic study. Eur J Heart Fail 2003;5:30513.
39. Bleeker GB, Bax JJ, Fung JW, van der Wall EE, Zhang Q, Schalij MJ et al.
Clinical versus echocardiographic parameters to assess response to
cardiac resynchronization therapy. Am J Cardiol 2006;97:2603.
40. Auricchio A, Stellbrink C, Sack S, Block M, Vogt J, Bakker P et al. The
Pacing Therapies for Congestive Heart Failure (PATH-CHF) study: ration-
ale, design, and endpoints of a prospective randomized multicenter
study. Am J Cardiol 1999;83:130D5D.
41. Jansen AH, Bracke FA, van Dantzig JM, Meijer A, van der Voort PH,
Aarnoudse W et al. Correlation of echo-Doppler optimization of atrioven-
tricular delay in cardiac resynchronization therapy with invasive hemody-
namics in patients with heart failure secondary to ischemic or idiopathic
dilated cardiomyopathy. Am J Cardiol 2006;97:5527.
42. Kerlan JE, Sawhney NS, Waggoner AD, Chawla MK, Garhwal S, Osborn JL
et al. Prospective comparison of echocardiographic atrioventricular
delay optimization methods for cardiac resynchronization therapy.
Heart Rhythm 2006;3:14854.
43. Drazner MH, Thompson B, Rosenberg PB, Kaiser PA, Boehrer JD, Baldwin
BJ et al. Comparison of impedance cardiography with invasive hemody-
namic measurements in patients with heart failure secondary to ischemic
or nonischemic cardiomyopathy. Am J Cardiol 2002;89:9935.
44. Ovsyshcher I, Zimlichman R, Katz A, Bondy C, Furman S. Measurements of
cardiac output by impedance cardiography in pacemaker patients at rest:
effects of various atrioventricular delays. J Am Coll Cardiol
1993;21:7617.
45. Santos JF, Parreira L, Madeira J, Fonseca N, Soares LN, Ines L. Non inva-
sive hemodynamic monitorization for AV interval optimization in patients
with ventricular resynchronization therapy. Rev Port Cardiol
2003;22:10918.
46. Kindermann M. Impedanzkardiograpie. Herzschrittmacherther
Elektrophysiol 2004;15:11029.
47. Sawhney NS, Waggoner AD, Garhwal S, Chawla MK, Osborn J, Faddis MN.
Randomized prospective trial of atrioventricular delay programming for
cardiac resynchronization therapy. Heart Rhythm 2004;1:5627.
48. Riedlbauchova L, Kautzner J, Fridl P. Inuence of different atrioventricu-
lar and interventricular delays on cardiac output during cardiac resyn-
chronization therapy. Pacing Clin Electrophysiol 2005;28(Suppl. 1):
S1923.
49. Sogaard P, Egeblad H, Pedersen AK, Kim WY, Kristensen BO, Hansen PS
et al. Sequential versus simultaneous biventricular resynchronization
for severe heart failure: evaluation by tissue Doppler imaging.
Circulation 2002;106:207884.
 Cardiac power output predicts mortality across a
broad spectrum of patients with acute cardiac disease
Dorinna D. Mendoza, MD, Howard A. Cooper, MD, and Julio A. Panza, MD Washington, DC

Background Cardiac power output (CPO) is a novel hemodynamic measurement that represents cardiac
pumping ability. The prognostic value of CPO in a broad spectrum of patients with acute cardiac disease undergoing
pulmonary artery catheterization (PAC) has not been examined.

Methods Consecutive patients with a primary cardiac diagnosis who were undergoing PAC in a single coronary
care unit were included. The relationship between initial CPO [(mean arterial pressure  cardiac output [CO])/451] and
inhospital mortality was evaluated. CPO was analyzed both as a dichotomous variable (using a cutoff value previously
established among patients with cardiogenic shock) and as a continuous variable.

Results Data were available for 349 patients. The mean CPO was 0.88 F 0.37 W. The inhospital mortality rate
was significantly higher among patients with a CPO V0.53 W (n = 53) compared with those with a CPO N0.53 W (n = 296)
(49% vs 20%, P b .001). In separate multivariate analyses, both CPO and CO were associated with inhospital mortality.
However, when both terms were included simultaneously, CPO remained strongly associated with mortality (odds ratio
0.63, 95% CI 0.43-0.91, P = .01), whereas CO did not (odds ratio 1.05, 95% CI 0.75-1.48, P = .78).

Conclusions Cardiac power output is a strong, independent predictor of inhospital mortality in a broad spectrum
of patients with primary cardiac disease undergoing PAC. (Am Heart J 2007;153:366270.)

Pulmonary artery catheterization (PAC) remains in
common usage in the coronary care unit (CCU) setting
despite questions regarding the use of this procedure in
critically ill patients.1-3 The decision to use PAC is based
on the concept that invasive hemodynamic measure-
ments provide a more precise understanding of the
status of the cardiovascular system, which in turn allows
close monitoring of the response to therapeutic inter-
ventions. However, traditional hemodynamic measure-
ments obtained from PAC, such as cardiac output (CO),
do not fully describe the fundamental state of the
cardiovascular system in the setting of severe illness.4 In
particular, because of its critical dependence on preload
and afterload, measured CO provides a limited estima-
tion of ventricular function.
Cardiac power output (CPO) is a novel hemodynamic
measure that is the product of simultaneously measured
CO and mean arterial pressure (MAP).4,5 By incorporating
both the pressure and flow domains of the cardiovascular
From the Division of Cardiology, Washington Hospital Center, Washington, DC.
This work was supported by the Cardiovascular Research Institute, Washington Hospital
Center, Washington, DC.
Submitted July 13, 2006; accepted November 21, 2006.
Reprint requests: Howard A. Cooper, MD, Division of Cardiology, Washington Hospital
Center, 110 Irving Street NW, Suite NA1103, Washington, DC 20010.
E-mail: howard.a.cooper@medstar.net
0002-8703/$ - see front matter
n 2007, Mosby, Inc. All rights reserved.
doi:10.1016/j.ahj.2006.11.014
system, CPO is an integrative measure of cardiac
hydraulic pumping ability. Cardiac power output has
been shown to be a powerful predictor of mortality in
patients with chronic heart failure (HF) and in those with
cardiogenic shock.6-8 Recently, a report from the SHOCK
Trial Registry found that CPO was the strongest inde-
pendent hemodynamic correlate of inhospital mortality
in patients with cardiogenic shock.4 We hypothesized
that CPO would provide important prognostic informa-
tion across a broad spectrum of patients with acute
cardiac disease. Accordingly, we analyzed the relation-
ship between CPO and short-term mortality in a consec-
utive series of patients undergoing PAC in a single CCU.
Methods
Study design and data collection
The institutional review board for human research at the
Washington Hospital Center approved this research protocol.
Data were collected prospectively for consecutive patients
admitted to the CCU between October 2002 and February
2006. Based on all available information, the attending
physician determined the primary clinical diagnosis requiring
admission to the CCU. Patients with a noncardiac primary
diagnosis were excluded from this analysis. Persistent cardio-
genic shock was determined to be present by using conven-
tional clinical criteria of hypotension and signs of peripheral
hypoperfusion in the presence of pulmonary congestion that
did not rapidly resolve. Left ventricular ejection fraction (EF)
was visually estimated from 2-dimensional echocardiography

57 of 158.
 American Heart Journal
Mendoza et al 367
Volume 153, Number 3

obtained during the index admission, or from left ventricular

angiography (when available) if echocardiography was Table I. Primary diagnosis
not performed. Primary diagnosis (N = 349) No. of patients (%)
For each patient, the first set of hemodynamic measure-
ments obtained in the CCU was used for analysis. Measure- ST-segment elevation myocardial infarction 122 (35)
ments included those made while receiving support with NonST-segment elevation 79 (23)
inotropes and/or an intra-aortic balloon pump. Cardiac output myocardial infarction
was measured by the thermodilution method, taking the Ischemic cardiomyopathy 38 (11)
average of 3 to 5 readings. Mean arterial pressure was Nonischemic cardiomyopathy 33 (9)
determined by invasive arterial monitoring if available; other- Coronary atherosclerosis 17 (5)
Unstable angina 16 (5)
wise, automated sphygmomanometry was used. Cardiac
Cardiac arrest 15 (4)
power output was calculated as MAP  CO H 451.4 Acute myocarditis 4 (1)
Treatment decisions were made without specific regard to Hypertrophic cardiomyopathy 4 (1)
CPO measurements. Ventricular fibrillation 4 (1)
Tako-Tsubo cardiomyopathy 3 (0.9)
Statistical considerations Aortic stenosis 2 (0.6)
Previously, the SHOCK investigators determined that the Atrial fibrillation 2 (0.6)
Mitral regurgitation 2 (0.6)
cutoff value of CPO that most accurately predicted inhospital
Mitral stenosis 2 (0.6)
mortality in their patient cohort was 0.53 W.4 Accordingly, we Pericardial tamponade 2 (0.6)
analyzed CPO both as a dichotomous variable (N0.53 W vs Ventricular tachycardia 2 (0.6)
V0.53 W) and as a continuous variable (per 0.2-W increments). Complete heart block 1 (0.3)
Baseline variables were compared using standard statistical Supraventricular tachycardia 1 (0.3)
tests. The relationship between baseline variables and inhos-
pital mortality was assessed with univariate logistic regression
analysis. Multivariate logistic regression analysis was performed
to examine the independent relationship between CPO (as a
continuous variable) and inhospital mortality. A comprehen- be women (64% vs 34%, P b .001). As expected, those
sive logistic regression model was constructed including the with CPO V0.53 W had a lower MAP, CO, and cardiac
following covariates: age, sex, race (white or nonwhite), index, and a higher systemic vascular resistance (all
treated diabetes mellitus, current or recent smoking, chronic P b .001). Of note, the EF was similar among patients
renal failure requiring dialysis, prior coronary artery bypass with CPO N0.53 W and patients with CPO V0.53 W (31%
grafting, intra-aortic balloon counterpulsation, mechanical
vs 28%, P = not significant).
ventilation, inotrope use, persistent cardiogenic shock, EF,
heart rate, central venous pressure, pulmonary capillary wedge
pressure, systemic vascular resistance, and CPO. A reduced Outcomes
model was then created that included only those covariates Patients with a CPO V0.53 W had a significantly higher
significantly associated with inhospital mortality. Cardiac inhospital mortality rate than those with a CPO N0.53 W
output was subsequently forced into this reduced model. A (49% vs 20%, P b .001), corresponding to a positive
2-sided P value less than .05 was considered to represent predictive value of 49% and a negative predictive value
statistical significance. of 80%. On univariate analysis, predictors of mortality
were age, diabetes mellitus, need for mechanical
ventilation, persistent cardiogenic shock, MAP, CO,
Results cardiac index, heart rate, central venous pressure, and
Baseline characteristics CPO (Table III). In the final reduced multivariate model,
Complete data were available for 349 patients. The CPO remained independently associated with a reduced
primary admission diagnoses are listed in Table I. The risk of inhospital mortality (odds ratio [OR] 0.65, 95% CI
most common diagnoses were acute myocardial infarc- 0.54-0.78, P b .001) (Table IV). Other independent
tion, cardiomyopathy, coronary atherosclerosis, and predictors were heart rate ( P = .01), diabetes mellitus
unstable angina. Baseline patient characteristics and ( P = .02), and the need for mechanical ventilation
hemodynamics are presented in Table II. The mean age ( P b .001). When CO was forced into this model, the
was 64 years, 59% were men, and 39% were of nonwhite relationship between CPO and inhospital mortality was
race. Patients had severe illness, as evidenced by the fact not substantially altered (OR 0.63, 95% CI 0.43-0.91,
that 50% required an intra-aortic balloon pump, 51% P = .01), whereas there was no significant relationship
required mechanical ventilation, and 55% required between CO and mortality (OR 1.05, 95% CI 0.75-1.48,
intravenous inotropic support. P = .78).
The CPO ranged from 0.17 to 2.48 W, with a mean of
0.88 F 0.37 W. Compared with patients with a CPO Cardiogenic shock
N0.53 W (n = 296), patients with a CPO V0.53 W (n = 53) Persistent cardiogenic shock was diagnosed in
were older (69 vs 63 years, P = .004) and more likely to 73 patients (21%). Mortality in patients with persistent

58 of 158.
 American Heart Journal
368 Mendoza et al
March 2007

Table II. Baseline characteristics and initial hemodynamic measurements

Characteristic All patients (N = 349) CPO >0.53 (n = 296) CPO <
_ 0.53 (n = 53) P

Age (y) 64 F 14 63 F 14 69 F 11 .004

Male, n (%) 206 (59) 188 (64) 18 (34) b.001
Nonwhite race, n (%) 135 (39) 113 (38) 22 (42) .65
History of
Coronary artery bypass grafting, n (%) 61 (18) 51 (17) 10 (19) .77
Percutaneous coronary intervention, n (%) 46 (13) 37 (13) 9 (17) .37
Diabetes mellitus, n (%) 79 (23) 67 (23) 12 (23) 1.0
Smoking, n (%) 27 (8) 25 (8) 2 (4) .40
Chronic renal failure, n (%) 14 (4) 13 (4) 1 (2) .70
Intra-aortic balloon pump, n (%) 175 (50) 155 (52) 20 (38) .05
Mechanical ventilation, n (%) 177 (51) 147 (50) 30 (57) .35
Acute myocardial infarction, n (%) 201 (58) 175 (59) 26 (49) .17
Persistent cardiogenic shock, n (%) 73 (21) 60 (20) 13 (25) .48
Mean arterial pressure (mm Hg) 80 F 13 82 F 13 70 F 11 b.001
Cardiac output (L/min) 4.9 F 1.8 5.2 F 1.7 2.8 F 0.7 b.001
Cardiac index (L min 1 m 2) 2.5 F 0.8 2.6 F 0.8 1.6 F 0.4 b.001
Systemic vascular resistance (dyne s/cm5) 1297 F 597 1207 F 535 1824 F 668 b.001
Heart rate (beats/min) 88 F 19 90 F 18 86 F 19 .23
Pulmonary capillary wedge pressure (mm Hg) 20 F 7 19 F 7 21 F7 .15
Central venous pressure (mm Hg) 13 F 7 12 F 5 15 F 12 .12
Any inotrope, n (%) 193 (55) 161 (54) 32 (60) .42
Ejection fraction (%) 30 F 13 31 F 13 28 F 12 .14

Table III. Univariate logistic regression results for inhospital Table IV. Multivariate logistic regression analysis for
mortality inhospital mortality
Variable OR (95% CI) P Variable OR (95% CI) P

Age 1.02 (1.00-1.04) .03 CPO 0.65 (0.54-0.78) b.001

Male sex 0.75 (0.46-1.22) .24 Heart rate 1.02 (1.01-1.04) .01
Nonwhite race 0.74 (0.44-1.24) .25 Diabetes mellitus 2.10 (1.13-3.89) .02
Prior coronary artery bypass grafting 1.15 (0.61-2.16) .66 Mechanical ventilation 3.76 (2.08-6.82) b.001
Prior percutaneous coronary intervention 1.13 (0.56-2.30) .73
Diabetes mellitus 1.94 (1.12-3.36) .02
Smoker 0.37 (0.11-1.27) .11
Chronic renal failure 0.89 (0.23-3.14) .81
Acute myocardial infarction 1.04 (0.63-1.71) .88 Discussion
Intra-aortic balloon pump 1.12 (0.69-1.84) .64 To our knowledge, the current study is the largest
Mechanical ventilation 4.30 (2.46-7.52) b.001 reported analysis of CPO in a broad spectrum of
Persistent cardiogenic shock 2.26 (1.29-3.94) .004 patients with critical cardiac illness undergoing PAC.
Any inotrope 1.18 (0.72-1.93) .52
MAP 0.96 (0.93-0.98) b.001
The major findings were the following: (1) inhospital
CO 0.74 (0.63-0.88) b.001 mortality was substantially higher in patients with a low
Cardiac index 0.47 (0.32-0.68) b.001 initial CPO, (2) CPO was independently associated with
Systemic vascular resistance 1.00 (1.00-1.00) .44 inhospital mortality after controlling for important
Heart rate 1.01 (1.00-1.03) .03
covariates, and (3) the association between CPO and
Pulmonary capillary wedge pressure 1.01 (0.98-1.05) .45
Central venous pressure 1.05 (1.01-1.10) .01 inhospital mortality was stronger than that of the more
EF 1.00 (0.98-1.02) .98 traditional CO.
CPO 0.66 (0.55-0.79) b.001 Our findings suggest that measurement of CPO is
useful for risk stratification in patients with primary
cardiac disease undergoing PAC in the CCU setting.
CPO represents the rate of energy input that the
cardiogenic shock was higher than mortality among systemic vasculature receives from the heart, incorpo-
patients without this condition (37% vs 21%, P b .001). rating both pressure and flow domains of the cardio-
After excluding patients with persistent cardiogenic vascular system.4 It is therefore logical that such a
shock, CPO remained independently associated with measure of cardiac pumping ability would predict
inhospital mortality (OR 0.64, 95% CI 0.50 to 0.81, outcomes for patients with cardiogenic shock and
P b .001). severe HF, as has been demonstrated previously.4,6-8

59 of 158.
 American Heart Journal
Volume 153, Number 3
Our analysis extends these findings to a broader
spectrum of patients with critical cardiac illness,
including those with acute coronary syndromes, car-
diomyopathies of various etiologies, valvular heart
disease, cardiac arrhythmias, and complicated post
percutaneous coronary intervention courses.
As expected, patients in our analysis had CPO values
considerably higher (mean 0.88 W) than those in a
more select group with acute myocardial infarction and
cardiogenic shock in the SHOCK trial registry (mean
0.62 W). Despite the substantial differences in patient
populations and absolute CPO values, however, the
relationship between CPO and inhospital mortality was
remarkably similar in the 2 analyses: an OR of 0.65 per
0.2-W increase in CPO in the current study compared
with an OR of 0.60 in the SHOCK trial registry.4
Furthermore, among the subset of our patients diag-
nosed with persistent cardiogenic shock (n = 73), the
positive and negative predictive values using a cutoff of
0.53 W were nearly identical in the 2 studies (62% vs
58% and 68% vs 71%, respectively). This concordance
of results suggests both that the relationship between
CPO and short-term mortality is consistent across a
broad spectrum of cardiac diagnoses and a wide range
of CPO values, and is reproducible across studies.
Cardiac power output is calculated from the more
familiar hemodynamic variables, CO and MAP. There-
fore, from a statistical standpoint, it cannot provide
additional prognostic information beyond that contained
within the 2 individual components. However, CPO
combines information from these 2 separate hemody-
namic axes (flow and pressure) into a single, easily
understandable variable that can potentially be targeted
for therapy. For this reason, the concept of CPO appears
to provide clinically useful information above and
beyond that provided by standard hemodynamic varia-
bles alone.
Recent studies have questioned the use of PAC in
guiding treatment in critically ill patients, including
patients with severe CHF.3,9-13 Possible explanations for
these negative results include complications related to
the PAC procedure itself (arrhythmias, infection);
misinterpretation of the obtained data; incorrect action
in response to even appropriately interpreted data; or
unanticipated deleterious effects of treatments imple-
mented in response to these data. In patients with
cardiac disease, treatments are typically targeted to the
CO (or cardiac index) and pulmonary capillary wedge
pressure. However, CO is a measure of cardiovascular
flow, not cardiac contractility, whereas pulmonary
capillary wedge pressure is a measure of intracardiac
pressure not directly representing cardiac perfor-
mance.1 Theoretically, therefore, CPOa measure of
cardiac pumping abilitymight be a more appropriate
target for medical therapy. In addition, CPO can be
simply and accurately determined noninvasively in most
60 of 158.
Mendoza et al 369
patients by using whole-body electrical bioimpedance
to determine CO, thereby eliminating any direct
complications of PAC.14,15 Clinical studies in which
noninvasively determined CPO is targeted for treatment
appear to be indicated.
Our study has several potential limitations. First, this
was a single-center retrospective analysis, subject to the
limitations inherent in such a study design. However,
measurements of CPO were not used by the CCU
physicians to make therapeutic decisions. Hence, we
are confident that the observed prognostic value of
CPO was not influenced by bias introduced from
knowledge of this information. Second, although we
attempted to control for important differences in
baseline characteristics between patients, residual con-
founding might remain. Third, we were unable to
distinguish between hemodynamic measurements made
with and without support from inotropes and/or an
intra-aortic balloon pump. Because both types of
interventions are likely to increase CPO, this might
affect the relationship between CPO and mortality.
However, this relationship was maintained after in-
cluding the use of these interventions in the multivar-
iate model. Indeed, the relationship between CPO and
mortality might be stronger in the absence of such
support measures. Finally, we must emphasize that our
findings demonstrate an association between CPO and
outcome without providing a direct assessment of
whether CPO should be used as a clinical end point for
therapeutic interventions.
Conclusions
Cardiac power output is a novel and useful
prognostic tool in patients with acute cardiac disease
undergoing PAC in the CCU setting. Further studies
in which CPO is targeted for treatment appear to
be warranted.
References
1. Connors Jr AF, Speroff T, Dawson NV, et al. The effectiveness of
right heart catheterization in the initial care of critically ill patients.
SUPPORT Investigators. JAMA 1996;276:889 - 97.
2. Dalen JE, Bone RC. Is it time to pull the pulmonary artery catheter?
JAMA 1996;276:916 - 8.
3. Binanay C, Califf RM, Hasselblad V, et al. Evaluation study of
congestive heart failure and pulmonary artery catheterization
effectiveness: the ESCAPE trial. JAMA 132005;294:1625 - 33.
4. Fincke R, Hochman JS, Lowe AM, et al. Cardiac power is the
strongest hemodynamic correlate of mortality in cardiogenic shock:
a report from the SHOCK trial registry. J Am Coll Cardiol
2004;44:340 - 8.
5. Cotter G, Williams SG, Vered Z, et al. Role of cardiac power in
heart failure. Curr Opin Cardiol 2003;18:215 - 22.
6. Williams SG, Cooke GA, Wright DJ, et al. Peak exercise cardiac
power output; a direct indicator of cardiac function strongly

370 Mendoza et al
predictive of prognosis in chronic heart failure. Eur Heart J
2001;22:1496 - 503.
7. Tan LB, Littler WA. Measurement of cardiac reserve in cardiogenic
shock: implications for prognosis and management. Br Heart J
1990;64:121 - 8.
8. Tan LB. Cardiac pumping capability and prognosis in heart failure.
Lancet 1986;2:1360 - 3.
9. Sandham JD, Hull RD, Brant RF, et al. A randomized, controlled trial
of the use of pulmonary-artery catheters in high-risk surgical
patients. N Engl J Med 2003;348:5 - 14.
10. Richard C, Warszawski J, Anguel N, et al. Early use of the
pulmonary artery catheter and outcomes in patients with shock and
acute respiratory distress syndrome: a randomized controlled trial.
JAMA 2003;290:2713 - 20.
11. Harvey S, Harrison DA, Singer M, et al. Assessment of the clinical
effectiveness of pulmonary artery catheters in management of
61 of 158.
American Heart Journal
March 2007
patients in intensive care (PAC-Man): a randomised controlled trial.
Lancet 2005;366:472 - 7.
12. Rhodes A, Cusack RJ, Newman PJ, et al. Bennett ED. A randomised,
controlled trial of the pulmonary artery catheter in critically ill
patients. Intensive Care Med 2002;28:256 - 64.
13. Wheeler AP, Bernard GR, Thompson BT, et al. Pulmonary-artery
versus central venous catheter to guide treatment of acute lung
injury. N Engl J Med 2006;354:2213 - 24.
14. Cotter G, Moshkovitz Y, Kaluski E, et al. Accurate, noninvasive
continuous monitoring of cardiac output by whole-body electrical
bioimpedance. Chest 2004;125:1431 - 40.
15. Leitman M, Sucher E, Kaluski E, et al. Non-invasive measurement of
cardiac output by whole-body bio-impedance during dobutamine
stress echocardiography: clinical implications in patients with left
ventricular dysfunction and ischaemia. Eur J Heart Fail 2006;8:
136 - 40.
 Circ J 2006; 70: 1164 1168

Impedance Cardiography for Cardiac Output Estimation

Reliability of Wrist-to-Ankle Electrode Configuratio

Oscar Luis Paredes, MD; Junya Shite, MD; Toshiro Shinke, MD;
Satoshi Watanabe, MD; Hiromasa Otake, MD; Daisuke Matsumoto, MD;
Yusuke Imuro, MD; Daisuke Ogasawara, MD;
Takahiro Sawada, MD; Mitsuhiro Yokoyama, MD

Background Non-invasive measurement of cardiac output (CO) may become an important modality for the
treatment of heart failure. Among the several methods proposed, impedance cardiography (ICG) has gained
particular attention. There are 2 basic technologies of ICG: thoracic and whole-body ICG whereby the electrodes
are applied either to the chest or to the limbs. The present study is aimed to test the effectiveness of the Non-
Invasive Cardiac System (NICaS), a new ICG device working with a wrist-to-ankle configuration
Methods and Results To evaluate the reliability of NICaS derived CO (NI-CO), 50 CO measurements were
taken simultaneously with thermodilution (TD-CO) and modified Fick (Fick-CO) in 35 cardiac patients, with the
TD-CO serving as the gold-standard for the evaluation. Overall, 2-tailed Pearsons correlation and Bland-Altman
limits of agreement between NI-CO and TD-CO were r=0.91 and 1.06 and 0.68 L/min and between Fick-CO
and TD-CO, r=0.80 and 1.52 and 0.88 L/min, respectively. Good correlation was observed in patients with load-
ing conditions altered by nitroglycerin and also in patients with moderate valvular diseases.
Conclusion Agreement between NI-CO and TD-CO is within the boundaries of the FDA guidelines of bio-
equivalence. NI-CO is applicable for non-invasive assessment of cardiac function. (Circ J 2006; 70: 1164
1168)
Key Words: Cardiac output; Impedance cardiography; Thermodilution method

S
everal studies suggest the importance of cardiac power ICG (ICGT), which was introduced in 1964, and the elec-
output calculation, which is derived from cardiac trodes are placed on the root of the neck and on the lower
output (CO) and mean blood pressure, to predict the chest. When the CO is measured in subjects with healthy
prognosis in heart failure patients not only in hospital but hearts, the results from both these technologies are usually
also in the outpatient setting.13 CO measured by the ther- reliable, but the reliability of CO measurements taken by
modilution method with a Swan-Ganz catheter placed in ICGT is compromised in patients with cardiac diseases.1216
the pulmonary artery has become one of the most widely According to the Food and Drug Administration (FDA)
accepted and used methods of monitor cardiac function, standard of bio-equivalence,17 the disparity between 2 tech-
despite its certain limitations.1,3,4 A noninvasive and low
cost method for measuring CO would be relevant for the
widespread clinical use of cardiac power output.
Some noninvasive techniques of measuring CO have been
proposed over the past years. The indirect Fick method of
re-breathing carbon dioxide5,6 and Doppler flow measure-
ment of the left ventricular outflow tract have been shown
to be accurate;7 however, their applications require expen-
sive equipments and trained operators. Other promising
results have been observed with devices based on electrical
bioimpedance technology,8 and 2 basic technologies of
impedance cardiography (ICG) are currently in use. The
first is called whole-body ICG9,10 (ICGWB), which was
introduced in 1948,11 in which the electrodes are placed on
the distal portion of the limbs. The second one is thoracic

(Received March 30, 2006; revised manuscript received June 1,

2006; accepted June 21, 2006)
Division of Cardiovascular and Respiratory Medicine, Department of
Internal Medicine, Kobe University Graduate School of Medicine,
Kobe, Japan
Mailing address: Junya Shite, MD, Division of Cardiovascular and
Respiratory Medicine, Department of Internal Medicine, Kobe Uni-
versity Graduate School of Medicine, 7-5-1 Kusunoki-cho, Chuo-ku,
Kobe 650-0017, Japan. E-mail: shite@med.kobe-u.ac.jp
Fig 1. Schema of Non-Invasive Cardiac System, showing the imped-
ance box, which is connected to the patient via proprietary electrodes
and the computer that interprets the collected data. Three electrodes
are applied to patients chest for the ECG monitor. One pair of imped-
ance electrodes is applied to the left wrist and another pair to the right
ankle (tetra polar mode).

Circulation Journal Vol.70, September 2006

62 of 158.
 Wrist-to-Ankle Impedance for CO Measurement
Table 1 Statistical Analysis of CO Measurements Performed by NI-CO and Indirect Fick-CO Compared With TD-CO
All measurements in 35 patients
(n=50)
NI-CO Fick-CO
TD-
vs vs
CO
TD-CO TD-CO
Correlation 0.911a 0.801b
Bias (L/min) 0.18 0.32
SD (L/min) 0.43 0.60
Lower level of agreement (L/min) 1.06 1.52
Upper level of agreement (L/min) 0.68 0.88
Average CO SD (L/min) 4.181.01 4.361.03 4.050.89
Values are mean SD. Correlation was calculated using Pearsons 2-tailed test.
1ap<0.0001, 1bp<0.0001, 2ap<0.0001, 2bp=0.0001.
CO, cardiac output; NI-CO, Non-Invasive Cardiac System derived CO; Fick-CO, modified Fick derived CO; TD-CO, thermodilution-derived CO; NTG, nitro-
glycerine.
nologies should not exceed the range of 20%. The purpose
of this study was to evaluate the reliability and feasibility of
the new Non-Invasive Cardiac System (NICaS: NI Medical,
Hod-Hasharon, Israel), which calculates the CO by measur-
ing ICG in a tetra polar mode, derived from electrodes
placed on one wrist and the contra-lateral ankle (Fig 1).
Methods
The Trial
This study prospectively enrolled 35 patients. The ther-
modilution-derived CO (TD-CO) was measured using a
Swan-Ganz catheter (Baxter Healthcare, Irvine, CA, USA),
followed immediately by the modified Fick (Fick-CO) and
with the NICaS (NI-CO). In 15 subjects, a second round of
CO measurements was carried out using the 3 technologies
after 24 mg nitroglycerin injection into the arteries. Thus,
a total of 50 comparative CO measurements were per-
formed. The Institutional Ethics Committee of the Kobe
University Hospital approved the study protocol and all
patients gave their consent.
Patient Selection and Exclusion Criteria
All cardiac disease patients who were scheduled for
routine right heart catheterization or emergency procedures
that required the deployment of a Swan Ganz catheter for
continuous cardiac function monitoring were eligible for
the study, unless they fulfilled one of the exclusion criteria
The exclusion criteria for the NI-CO measurements in-
cluded: restlessness and/or chaotic patient condition, severe
aortic valve regurgitation and/or aortic stenosis, aortic
aneurysm, heart rate above 130 beats/min, intra- and extra-
cardiac shunts, severe peripheral vascular disease, severe
pitting edema, sepsis and dialysis, all of which interfere
with the proper measurements of impedance derived stroke
volume.4
The study population comprised 35 patients (17 men, 18
women, mean age 65.513.7 years). In most patients there
was coexistence of multiple underlying heart diseases,
including hypertension (n= 15), diabetes mellitus (n= 13),
coronary artery disease (n= 21) and idiopathic dilated car-
diomyopathy (n=3). Twelve patients presented with conges-
tive heart failure (CHF) and 4 patients had atrial fibrillatio
when CO was measured. In addition, our study subjects
included 7 cases of moderate degree of valve regurgitation
(aortic regurgitation 2 cases, mitral regurgitation 4 cases,
mild tricuspid regurgitation 1 case) and 7 cases of moderate
Circulation Journal Vol.70, September 2006
63 of 158.
1165
Measurements after NTG injection
(n=15)
NI-CO Fick-CO
NI- Fick- TD- NI- Fick-
vs vs
CO CO CO CO CO
TD-CO TD-CO
0.962a 0.822b
0.15 0.34
0.50 0.69
1.16 1.73
0.85 1.04
3.590.76 3.850.85 3.670.79
aortic valve stenosis.
Measuring CO
All operators were unaware of the CO results obtained
by the various measuring techniques.
TD-CO Right heart catheterization using a 6 or 8Fr
Swan-Ganz catheter was performed according to the stan-
dard institutional protocol. The catheter was advanced to
the pulmonary artery under fluoroscopic guidance and veri-
fied with the pressure waveforms registered on the poly-
graph. TD-CO was measured 5 times by injecting 5 ml
bolus of iced 9% saline solution at the same rate. There-
after, the 3 results of the saline injections that were within
15% of their extreme disparity were averaged for the TD-
CO result.
Fick-CO For arterial oxygen saturation, blood samples
were obtained from the arterial access sheath, and for
venous oxygen saturation, blood samples were withdrawn
using the distal edge lumen of the Swan-Ganz catheter
placed in the pulmonary artery. All samples were immedi-
ately measured for oxygen saturation using the same device
(Radiometer ABL 715, Copenhagen, Denmark).
NI-CO To measure the CO with the NICaS apparatus,
an alternating electrical current of 1.4 mA with a 30 kHz
frequency is passed through the patient via 2 pairs of
tetrapolar electrodes, one pair placed on the wrist above the
radial pulse, and the other pair on the contralateral ankle
above the posterior tibialis arterial pulse. If the arterial
pulses in the legs are either absent or of poor quality, the
second pair of electrodes is placed on the contralateral
wrist. The NICaS apparatus calculates the stroke volume
by Frinermans formula:18
Stroke volume = dR/R L2 /Ri (+)/ KW HF
where dR is the impedance change, R is the basal resis-
tance,is the blood electrical resistivity, L is the patients
height, Ri is the corrected basal resistance according to
gender and age,1823 KW is a correction of weight according
to ideal values,1924 HF is the hydration factor, which takes
into account the body water composition,21 and + is
equal to the ECG RR wave interval, andis the diastolic
time interval.
Because the NI-CO results are calculated every 20 s, the
average of 3 measurements obtained consecutively during
60 s of monitoring was considered to be the NI-CO value
for each individual case.
 1166 PAREDES O L et al.

Fig 2. Scatter plots of cardiac output (CO) showing the correlation between Non-Invasive Cardiac System (NICaS)
derived CO (NI-CO) with thermodilution-derived CO (TD-CO) [A]; modified Fick derived CO (Fick-CO) with TD-CO
[B] and NI-CO with Fick-CO [C]. The 2-tailed Pearsons correlation test (r) was used for the analyses.

Fig 3. [A] Bland-Altman scatter plot of difference against average of cardiac output (CO) results measured by Non-
Invasive Cardiac System (NICaS) and thermodilution method. [B] Bland-Altman scatter plot of difference against
average of CO results measured by the modified Fick and thermodilution methods. NI-CO, NICaS derived CO; TD-CO,
thermodilution-derived CO; Fick-CO, modified Fick derived CO

Statistical Analysis ered to be significant

The quantitative data are expressed as mean SD. For de-
scriptive statistic, Students t-test was used. To compare the
results of NI-CO, Fick-CO and TD-CO, 2-tailed Pearsons
correlation and the Bland-Altman25 limits of agreement
were used. The gold-standard for determining accuracy of
the results was the TD-CO. Values of p<0.05 were consid-
Results
The average values of CO in the study subjects for TD-
CO, NI-CO and Fick-CO were 4.181.01 L/min, 4.36
1.03 L/min, and 4.050.89 L/min, respectively. There were

Circulation Journal Vol.70, September 2006

64 of 158.
 Wrist-to-Ankle Impedance for CO Measurement
no significant differences between the 3 groups (Table 1).
The overall results of the Pearson correlation analysis were
as follows: NI-CO vs TD-CO: r=0.91, p<0.0001; Fick-CO
vs TD-CO: r= 0.80, p<0.0001 and NI-CO vs Fick-CO: r=
0.85, p<0.0001 (Fig 2). The Bland-Altman 2-standard devia-
tion limit of agreement between the NI-CO and TD-CO
was 0.87 (1.06 and 0.68) L/min, and the agreement
between the Fick-CO and TD-CO was 1.20 (1.52 and
0.88) L/min (Fig 3). The calculated percentage of disparity
between the NI-CO and TD-CO would thus be 19.95%
(0.87 L/min 4.36 L/min), which was less than that between
Fick-CO and TD-CO (29.63% [1.20 L/min 4.05 L/min]).
Nevertheless, there are 3 cases in this series in which the
disparity between the NI-CO and TD-CO was greater than
20%, indicating that the disparity here is not equal, but
close to FDA bio-equivalence (Note that the mathematical
model of the FDA for determining bio-equivalence was not
used here. Yet the simple model of limits of agreement,
which was used, offers an acceptable appraisal of the good
interrelationships between the 2 statistical approaches).
When we analyzed the subgroup of measurements be-
fore and after nitroglycerine injection to alter vascular re-
sistance, identical changes in CO were observed with the
NI-CO and TD-CO (TD-CO: 4.421.00 L/min 3.59
0.76 L/min, NI-CO: 4.071.07 L/min 3.850.85 L/min
(Fig 4)). The relation between NI-CO and TD-CO after
nitroglycerine injection was r= 0.96, p<0.0001, n= 15. All
other statistical details are summarized in Table 1.
In the other subgroups of moderate degree of valvular
regurgitation (n= 7) and aortic valve stenosis (n= 7), the
correlation of NI-CO with TD-CO was r=0.92, p<0.0001,
with a lower limit of agreement of 0.97 L/min and upper
limit of agreement of 0.73 L/min.
Discussion
According to Bland and Altman26 and Raaijmakers et
al,15 when averages of repeated measurement results are
used to compare the performance of a new medical device
with a gold-standard, there is an overestimation of the cor-
relation coefficient. In the present investigation the prefera-
ble single measurement design was used; namely, each test
consisted of a TD-CO measurement, followed immediately
by a NI-CO and a Fick-CO measurement. In 20 patients,
only 1 study was performed, whereas in the remaining 15
patients 2 studies were done: before and after nitroglycer-
ine injection. However, each of the second tests was con-
ducted in the manner of an independent comparative
measurement.
According to the definition of the FDA, if there is bio-
equivalence between the gold-standard and a new technol-
Table 2 Summary of Previous Reports of the Accuracy of CO Measurements by ICG (ICGT-CO, ICGWB-CO, and
NI-CO vs TD-CO)
Authors Method Condition
Drazner M, et al27 ICGT CHF
Van De Water JM, et al28 ICGT CABG
Leslie S J, et al29 ICGT CHF
Koobi T, et al10 ICGWB CABG
Cotter G, et al4 ICGNICaS CABG, CHF
Bipolar
Values are mean. ICG, impedance cardiography; ICGT, thoracic ICG; ICGWB, whole-body ICG; TD, thermodilution; CHF, congestive
heart failure; CABG, coronary artery bypass graft; ICGNICaS, Non-Invasive Cardiac System ICG. Other abbreviations as in Table 1.
Circulation Journal Vol.70, September 2006
65 of 158.
1167
Fig 4. Changes in cardiac output in 15 patients with repeated mea-
surements, pre and post nitroglycerine (NTG) administration. The
graph compares the results obtained by Non-Invasive Cardiac System
derived cardiac output (NI-CO) and thermodilution-derived cardiac
output (TD-CO). Values are mean SD.
ogy, all the comparative results should be within a range of
20% disparity. Previous studies reported limits of agree-
ment between ICGT-CO vs TD-CO of 2.2 to 2.2 L/min at
best,2729 ICGWB-CO (JR Medical-Tallinn, Estonia) vs TD-
CO of 1.37 to 1.87 L/min,10 and bipolar NI-CO vs TD-CO
of 1.25 to 1.30 L/min (Table 2).4 Based on these reports,
it can be calculated that the disparity between ICGT-CO,
ICGWB-CO, bipolar NI-CO vs TD-CO is 40%, 32.4% and
26%, respectively. Our result for tetra polar NI-CO vs TD-
CO (1.06 to 0.68 L/min, disparity 20%) is better than
those results. The reason for the better result with NICaS
may be related to the most upgraded calculation formula.
An important fringe benefit of this trial is the data pro-
duced by the Fick-CO technology. This method, which
enjoys increasing popularity among practical cardiologists,
is still considered controversial.30 According to the present
results, the limits of agreement between Fick-CO and TD-
CO are 1.52 and 0.88 L/min (average 1.20 L/min). In the
presence of a mean Fick-CO of 4.05 L/min, the disparity
between the 2 technologies is 30%, better than that of ICGT
but inferior to that of NI-CO.
Among the established exclusion criteria related to the
use of the NICaS are: severe aortic stenosis, in which the
NI-CO is usually underestimated, and significant aortic re-
gurgitation, in which the NI-CO tends to be overestimated.
In the present trial, however, 7 cases of moderate aortic
stenosis and 2 with mild moderate aortic regurgitation
were involved, indicating that NICaS is applicable in cases
of mild to moderate aortic valve disease.
From the 15 measurements that were obtained after nitro-
glycerine injection, we observed a decrease in the mean CO
Limits of agreement between
Year published
ICG and TD (L/min)
2002 2.2 to 2.2
2003 2.36 to 1.99
2004 2.2 to 2.2
1997 1.37 to 1.87
2004 1.25 to 1.30
 1168
levels. However, the accuracy of the results remained un-
altered in NI-CO, which suggests that NICaS is also applic-
able even when the arterial resistance has changed.
Clinical Implications
Recent studies have shown that the calculation of cardiac
power output (CO mean arterial pressure) and systemic
vascular resistance are important for the management of
various cardiac diseases.13 Cardiac power output was found
to be the strongest independent predictor of in-hospital
mortality in patients admitted with cardiogenic shock2 and
is an important tool for assessing the clinical response to
drug therapy. In addition, there is enough evidence that
ambulatory monitoring of cardiac power would benefit pa-
tients with CHF, resulting in better titration of medication
and possibly less readmission to hospital. By introducing
NICaS apparatus, wide-spread clinical use of cardiac
power calculation would become feasible.
Conclusion and Limitations of NICaS
The present study indicates that NICaS performs at least
as accurately as the thermodilution method. However, the
reliability of the NICaS method depends on an alignment
with exclusion criteria. This allows for the use of NICaS in
approximately 8085% of patients needing the examina-
tion.
References
1. Nohria A, Mielniczuk L, Warner L. Evaluation and monitoring of
patients with acute heart failure syndromes. Am J Cardiol 2005;
96(Suppl): 32G 40G.
2. Fincke R, Hochman J, Lowe A, Menon V, Slater JN, Webb JG, et al
(for the SHOCK investigators). Cardiac power is the strongest hemo-
dynamic correlate of mortality in cardiogenic shock: A report from
the SHOCK trial registry. J Am Coll Cardiol 2004; 40: 340 348.
3. Cotter G, Moshkovitz Y, Kaluski E, Milo O, Nobikov Y, Schneeweiss
A, et al. The role of cardiac power and systemic vascular resistance
in the pathophysiology and diagnosis of patients with acute conges-
tive heart failure. Eur J Heart Fail 2003; 5: 443 451.
4. Cotter G, Moshkovitz Y, Kaluski E, Kohen A, Miller H, Goor D, et
al. Accurate, noninvasive continuous monitoring of cardiac output by
whole-body electrical bioimpedance. Chest 2004; 125: 1431 1440.
5. Haryadi DG, Orr JA, Kuck K, Mc James S, Westenskow DR. Partial
CO2 rebreathing indirect Fick technique for non-invasive measure-
ment of cardiac output. J Clin Monit Comput 2000; 16: 361 374.
6. Jover JL, SoroM, Belda FJ, Aguilar G, Caro P, Ferrandis R. Measure-
ment of cardiac output after cardiac surgery: Validation of a partial
carbon dioxide rebreathing (NICO) system in comparison with con-
tinuous thermodilution with a pulmonary artery catheter. Rev Esp
Anestesiol Reanim 2005; 52: 256 262 (in Spanish).
7. Turner MA. Doppler-based hemodynamic monitoring: A minimally
invasive alternative. AACN Clin Issues 2003; 14: 220 231.
8. Moshkovitz Y, Kaluski E, Milo O, Vered Z, Cotter G. Recent devel-
opments in cardiac output determination by bioimpedance: Compari-
son with invasive cardiac output and potential cardiovascular applica-
tions. Curr Opin Cardiol 2004; 19: 229 237.
9. Tischenko MI. Estimation of stroke volume by integral rheogram of
the human body. Sechenov Phisiological J 1973; 59: 1216 1224 (in
Russian).
66 of 158.
PAREDES O L et al.
10. Koobi T, Kaukinen S, Turjanmaa VM. Cardiac output can be reliably
measured noninvasively after coronary artery bypass grafting opera-
tion. Crit Care Med 1999; 27: 2206 2211.
11. Kedrov AA. An attempt to quantify assessment of the central and
peripheral circulation by electrometrical method. Klin Med 1948; 26:
32 51 (in Russian).
12. Patterson RP, Kubicek WG, Kinnen E, Witsoe DA, Noren G. Devel-
opment of an electrical impedance plethysmography system to moni-
tor cardiac output. Proceedings of the 1st Annual Rocky Mountains
Bioengineering Symposium 1964; 56 71.
13. Kubicek WG, Karnegis JN, Patterson RP, Witsoe DA, Mattson RH.
Development and evaluation of an impedance cardiac output system.
Aerosp Med 1966; 37: 1208 1212.
14. Kubicek WG, Kottke FJ, Ramos MU, Patterson RP, Witsol DA,
Labree JW, et al. The Minnesota impedance cardiograph: Theory and
applications. Biomed Eng 1974; 9: 410 416.
15. Raaijmakers E, Faes ThJC, Scholten RJPM, Goovaerts HG, Heethaar
R. A meta-analysis of published studies concerning the validity of
thoracic impedance cardiography. Ann NY Acad Sci 1999; 873: 121
134.
16. Handelsman H. Public health service reassessment: Measuring car-
diac output by electrical bioimpedance. Health Technology Assess-
ment Reports, US Dept Health and Human Services, Public Agency
for Health Care Policy and Research 1991; 6: 1 13.
17. Guidance for Industry. Statistical approach to establishing bioequiv-
alence: US department of health and human services. Food and Drug
Administration, Center For Drug Evaluation and Research (CDER)
January 2001, BP.
18. Cohen AJ, Arnaudov D, Zabeeda D, Shultheis L, Lashinger J,
Schachner A. Non-invasive measurement of cardiac output during
coronary artery bypass grafting. Eur J Surg 1998; 14: 64 69.
19. Organ LW, Bradham GB, Gore DT, Lozier SL. Segmental bioelectri-
cal impedance analysis: Theory and application of a new technique. J
Appl Physiol 1994; 77: 98 112.
20. Lukaski H, Bolonchuk WW, Hall CB, Siders WA. Validation of tet-
rapolar bioelectrical impedance method to assess human body
composition. J Appl Physiol 1986; 60: 1327 1332.
21. Hoffer EC, Meador CK, Simpson DC. A relationship between whole
body impedance and total body water volume. Ann NY Acad Sci
1970; 170: 452 461.
22. Lamberts R, Visser KR, Zijlstra WG. Impedance cardiography.
Assen, The Netherlands; Van Gorkum 1984; 21 94.
23. Ward LC, Heitmann BL, Craid P, Stroud D, Azinge EC, Jebb S, et al.
Association between ethnicity, body mass index, and bioelectrical
impedance: Implications for the population specificity of prediction
equations. Ann NY Acad Sci 2000; 904: 199 204.
24. Hamwi GT, Danowski TS. Changing dietary concepts in diabetes
mellitus: Diagnosis and treatment. New York: American Diabetes
Association 1964; 73 78.
25. Bland JM, Altman DG. Statistical methods for assessing agreement
between two methods of clinical measurement. Lancet 1986; I: 307
310.
26. Bland JM, Altman DG. Correlation, regression and repeated data.
BMJ 1994; 308: 896.
27. Drazner M, Thompson B, Rosenberg P, Kaiser PA, James MSN,
Boehrer D, et al. Comparison of impedance cardiography with inva-
sive hemodynamic measurements in patients with heart failure sec-
ondary to ischemic or nonischemic cardiomyopathy. Am J Cardiol
2002; 89: 993 995.
28. Van De Water JM, Miller TW, Vogel RL, Mount BE, Dalton ML.
Impedance cardiography: The next vital sign technology? Chest
2003; 123: 2028 2033.
29. Leslie S J, McKee S, Newby DE, Webb DJ, Denvir MA. Non-inva-
sive measurement of cardiac output in patients with chronic heart
failure. Blood Press Monit 2004; 9: 277 280.
30. Dhingra VK, Fenwick JC, Walley KR, Chittock D, Ronco JJ. Lack
of agreement between thermodilution and Fick cardiac output in criti-
cally ill patients. Chest 2002; 122: 990 997.
Circulation Journal Vol.70, September 2006
 INSTITUTE OF PHYSICS PUBLISHING PHYSIOLOGICAL MEASUREMENT

Physiol. Meas. 27 (2006) 817827 doi:10.1088/0967-3334/27/9/005

Impedance cardiography revisited

G Cotter1, A Schachner2, L Sasson2, H Dekel2 and Y Moshkovitz3
1 Divisions of Clinical Pharmacology and Cardiology, Duke University Medical Center, Durham,

NC, USA
2 Angela & Sami Sharnoon Cardiothoracic Surgery Department, Wolfson Medical Center, Israel
3 Department of Cardiac Surgery, Assuta Hospital, Petah Tikva, Israel

E-mail: gad.cotter@duke.edu

Received 7 February 2006, accepted for publication 9 June 2006

Published 5 July 2006
Online at stacks.iop.org/PM/27/817

Abstract
Previously reported comparisons between cardiac output (CO) results in
patients with cardiac conditions measured by thoracic impedance cardiography
(TIC) versus thermodilution (TD) reveal upper and lower limits of agreement
with two standard deviations (2SD) of approximately 2.2 l min1, a 44%
disparity between the two technologies. We show here that if the electrodes
are placed on one wrist and on a contralateral ankle instead of on the chest, a
configuratio designated as regional impedance cardiography (RIC), the 2SD
limit of agreement between RIC and TD is 1.0 l min1, approximately
20% disparity between the two methods. To compare the performances of
the TIC and RIC algorithms, the raw data of peripheral impedance changes
yielded by RIC in 43 cardiac patients were used here for software processing
and calculating the CO with the TIC algorithm. The 2SD between the TIC
and TD was 1.7 l min1, and after annexing the correcting factors of the
RIC formula to the TIC formula, the disparity between TIC and TD further
declined to 1.25 l min1. Conclusions: (1) in cardiac conditions, the RIC
technology is twice as accurate as TIC; (2) the advantage of RIC is the use
of peripheral rather than thoracic impedance signals, supported by correcting
factors.

Keywords: cardiac output measurements, thoracic bioimpedance, whole-body

bioimpedance, impedance cardiography

Introduction

Three basic technologies are currently in use for impedance cardiography (ICG): (1) the
thoracic ICG (TIC), where the electrodes are placed on the root of the neck and the lower
part of the chest, being the dominant method in the market (Patterson et al 1964, Kubicek
0967-3334/06/090817+11$30.00 2006 IOP Publishing Ltd Printed in the UK 817

67 of 158.
 818 G Cotter et al

et al 1966, 1974); (2) the whole-body ICG (ICGWB), where four pairs of electrodes are used,
one pair on each limb (Tischenko 1973, Koobi et al 1999); (3) the regional ICG (RIC), a
technology which is used by the NICaS (noninvasive cardiac system). In this technology,
which is the subject of this report, only two pairs of electrodes are used, performing best
when placed on one wrist and on the contralateral ankle (Cohen et al 1998, Cotter et al 2004,
Torre-Amione et al 2004).
Two comprehensive reviews of the literature on clinical experience in measuring the
cardiac output (CO) by TIC determined that in patients with cardiac conditions the TIC-CO
results are unreliable (Raaijmakers et al 1999, Handelsman 1991). According to Patterson
(1985) and Wang et al (2001), a number of sources in the chest, such as the lungs, vena cava, and
systemic and pulmonary arterial vasculatures, generate systolic impedance reductions, while
the heart generates signals of increased impedance. In addition to these multifarious sources of

Z,4 variations in the electrical conductivities between the sources of impedance changes and
the TIC electrodes (Kim et al 1988, Kauppinen et al 1998), and between the various impedance
sources (Wtorek 2000) have been described. These model experimentations indicated that
the thoracic
Z is not a reliable signal for calculation of the SV due to the multiple sources
of dZ/dt (Kim et al 1988, Wang and Patterson 1995, Kauppinen et al 1998, Wtorek 2000),
providing the explanations for the above-mentioned unsatisfactory clinical results obtained by
TIC (Raaijmakers et al 1999, Handelsman 1991).
In this report, an attempt is made to defin the differences between the peripheral and
thoracic impedance signals, and based on this, to explain the differences in the performance
of RIC and TIC.
As we are capable of saving raw data from the wristankle (peripheral) impedance
signals, we were able to use the peripheral impedance waveforms and base impedance values
to calculate stroke volumes, using various algorithms that have been associated with TIC
calculations. This enabled us to prove that (1) the performance of RIC is twice as accurate
as reported TIC results; (2) the reasons for this are as follows: (a) the impedance changes
which are yielded by the limb electrodes are more suitable than the impedance changes of
the thoracic electrodes for calculating the stroke volume and (b) the use of properly designed
coefficient improved the accuracy of the CO results by at least an additional 25%.

Methods

The data for this project were gathered from two patient series. In both, comparisons were made
between cardiac output results measured by the NICaS versus thermodilution. One series,
which was studied in hospital A, consisted of 30 patients who were studied immediately upon
arrival at the ICU following an open heart operation. In 11 (36%), despite the intravenous
administration of adrenalin, cardiac index (CI) was lower than 2.5 l min1 m2. The second
series included 13 cases of acute heart failure that were studied in hospital B. CI was lower
than 2.5 l min1 m2 in seven (54%), and in the combined group of 43 cases of the two
hospitals, it was lower than 2.5 l min1 m2 in 18 (43%).
The purpose of this study was to use peripheral impedance waveforms to calculate stroke
volume by means of four different ICG algorithms and to compare each of these SV values
with the thermodilution SV result.
Of the 55 and 31 studies conducted in hospitals A and B, respectively, raw data were
successfully retrieved from only the last 30 consecutive patients of hospital A and the last 13
4 In the ICGWB and RIC, where the impedance changes are depicted in the periphery, the impedance value is
automatically converted into the real parts (R0 and
R) of the measured impedance signals (Lamberts et al 1984).

68 of 158.
 Impedance cardiography revisited 819

ECG

Figure 1. Unedited print of the ECG and
R waveforms to demonstrate the definition of
and .

consecutive cases of hospital B. It quickly became apparent that these 43 source data were
retrieved from only one of the two NICaS devices that were being used in the two hospitals. The
reason for this was that we did not realize that the software designed for source data retrieval
was not implemented in one of the two NICaS devices that were used interchangeably during
these trials.

Measuring the CO
Thermodilution. The standard techniques for measuring the thermodilution CO (TDCO)
used in these two hospitals have been reported elsewhere (Cotter et al 2004).

RIC (NICaS). To measure the CO with the NICaS (NICO), which is a tetrapolar device, two
electrodes are placed on a wrist above the radial pulse and two on the contralateral ankle above
the posterior tibialis arterial pulse. If the arterial pulses in the legs are either absent or of poor
quality, the second pair of electrodes is placed on the contralateral wrist. The NICaS device
calculates the SV by means of the following Frinerman formula5 (Cohen et al 1998):

RL2 ( + )Kw HF
SV = , (2)
RRi
where SV is the cardiac stroke volume (ml),
R is the resistance change during the cardiac
cycle (), R is the basal resistance (), Ri is a corrected basal R (), is the blood electrical
resistivity ( cm), L is the patients height (cm), Kw is a correcting factor for the body weight,
HF is the hydration factor related to the body water composition and ( + )/ is the ratio
of the systolic time plus the diastolic time divided by the diastolic time of the
R waveform
(figur 1).
5 The present formula is principally the same as the formula in Cohen et al (1998), but is written differently. The
original formula was

Hctcorr 2
R
IB Hcorr ( + )
SV = Kel Kweight . (1)
Ksex R
The values of the hematocrit (Hct) and Na( (el) are now represented by , which is the electrical resistivity of the blood.
Ksex age, which affects the basal resistance (R), is now represented by R i, and IB (index body) is now represented by
HF (hydration factor). The correction of H 2 is no longer included in the formula, but in patients whose arms are
disproportionately long, the electrodes should be placed 5 cm proximally to their regular position.

69 of 158.
 820 G Cotter et al

Definition and principles of the correcting coefficients of the Frinerman formula

The basic impedance part of Frinermans formula consists of the following Bonjer equation
(1950):
dRL2
SV = , (3)
R2
to which Frinermans correcting factors were added. These coefficient were designed to
neutralize the individual effects of gender, age, body water composition and anthropomorphic
variabilities of each patient.
According to RIC studies, the basal R0 is higher in females than in males, a fact which
is in accordance with the literature (Organ et al 1994, Lukaski et al 1986, Hoffer et al 1970,
Lamberts et al 1984, Ward et al 2000), and it tends to rise with age (Organ et al 1994).
Based on reported values of basal resistances related to sex and age (Organ et al 1994,
Lukaski et al 1986, Ward et al 2000), a coefficien is calculated to determine an ideal value
of R0 for the studied patient. Ideal here means an ideal weight compared to height in healthy
condition. By dividing the measured R0 of the patient by the calculated ideal R0, a correcting
coefficien is obtained. By multiplying the measured R by the correcting coefficient the Ri
variable is determined and placed in the denominator of the formula.
Similarly, a correcting coefficien (Kw) for the body weight is calculated by dividing the
measured weight by the calculated weight according to Hamwis formula (1964) of ideal
weight. Kw is required for compensation of the reduced electrical conductance in the fat, and
this coefficien is placed in the numerator of the formula, affecting the value of
R. In a similar
fashion, the hydration factor, HR, is calculated by the patients body water composition, and
this coefficien is placed in the numerator of the formula, either to reduce or to increase
R in
over- or under-hydrated patients, respectively.
The variable ( + )/ is based on the Windkessel principle (Frank 1926, Faes et al
1999), where a distinction is made between the aortic infl w , which is incurred by the
systolic left ventricular ejection, and the aortic outfl w ( + ), which extends throughout the
cardiac cycle (figur 1).

Comparative calculation of CO by software simulation

The impedance raw data of the compiled 43 investigative patients were retrieved and their
CO were calculated by the following formulae: (1) Frinermans formula, as measured by
the NICaS; (2) Bonjers equation of 1950 (Bonjer 1950), which was expressed differently by
Pattersons f rst version of the TIC formula (Patterson et al 1964):

2
L

V =
Z; (4)
Z
(3) Pattersons formula of the f rst derivative (Patterson 1965):

2
dZ L
SV = T , (5)
dt Z
where SV is the cardiac stroke volume (cm), dZ/dt is the peak of the f rst derivative of the
impedance change during systole ( s1), T is the cardiac ejection time (s), L is the length
between voltage pick-up electrodes (cm) and Z is the base impedance value (), which is still
used by the TIC technology (Kubicek et al 1974); (4) a combined formula which includes
Pattersons equation together with Frinermans correcting factors (table 1, figure 2 and 3):

70 of 158.
 Impedance cardiography revisited 821

11.0

9.0

ICG-CO (lit/min)

7.0

5.0

3.0

1.0
1.0 3.0 5.0 7.0 9.0 11.0

TDCO (lit/min)

Frinerman NICaS Bonjer

Kubicek-Patterson(first derivative) Kubicek & Frinerman Correcting Factors

Figure 2. Scatter plot comparing CO results calculated by four impedance algorithms versus
thermodilution.

Table 1. Correlations and limits of agreement between four algorithms versus TD. Comparison of
four different algorithms in 43 patients for measuring cardiac output (CO) of the same raw data:
(a) Frinerman (NICaS); (b) Bonjer; (c) PattersonKubicek; (d) PattersonKubicekFrinerman.
Each of these was compared with thermodilution.

Frinerman Patterson PattersonKubicek

TD NICaS Bonjer Kubicek Frinerman

Average CO (l min1) 5.12 5.05 5.33 5.00 5.09

SD 1.952 2.043 2.218 1.781 1.986
Correlation with TD 0.969 0.841 0.897 0.950
p value versus NICaSa <0.001 <0.0005 0.33
Bias 0.0698 0.2107 0.1116 0.0302
SD of Bias 0.509 1.203 0.864 0.624
LL agreement 1.088 2.196 1.840 1.277
UL agreement 0.949 2.617 1.617 1.217
SD = standard deviation; LL agreement = lower limit of agreement; UL agreement = upper limit
of agreement.
a Bonferroni adjusted p values for testing equality of correlation NICaS TD as compared with Bonjer

TD (<0.0001), PattersonKubicek TD (<0.0005), PattersonKubicekFrinerman TD (0.33).

dR T L2 ( + )
SV = Kw HF, (6)
dt R Ri
where Frinermans
R was replaced by dZ/dt T and ( + )/ was deleted.

Patient selection
Restlessness, aortic insufficien y, abdominal aneurysm, heart rate above 130 beats min1,
arrhythmia with significan irregular heart rate, severe peripheral vascular disease, two or more

71 of 158.
 822 G Cotter et al

(A) Frinerman (B) Bonjer

3.0 3.0
+2SD
2.0
Difference(NICaS-TD) CO

2.0

Difference(Bonjer-TD) CO
1.0 +2SD
1.0

lit/min
0.0 0.0
lit/min

-1.0 -2SD -1.0

-2.0 -2.0 - 2SD

-3.0 -3.0
0 2 4 6 8 10 12 0 2 4 6 8 10 12
Average(NICaS+TD)/2 CO lit/min Average(Bonjer+TD)/2 CO lit/min

(C) Patterson-Kubicek (D) Patterson-Kubicek-Frinerman

3.0 3.0

Difference(Patterson-Kubicek-
Difference(Patterson-Kubicek-

2.0 2.0

Frinerman-TD) CO lit/min
+2SD
+2SD
1.0 1.0
TD) CO lit/min

0.0 0.0

-1.0 -1.0
-2SD
-2SD
-2.0 -2.0

-3.0 -3.0
0 2 4 6 8 10 12 0 2 4 6 8 10 12
Average(Patterson-Kubicek+TD)/2 CO lit/min Average(Patterson-Kubicek-Frinerman+TD)/2 CO lit/min

Figure 3. Scatter plots of BlandAltman difference-against-average of the CO results calculated

by the four impedance algorithms versus thermodilution results (based on table 1).

pitting edema and dialysis all interfere with the proper measurements of the SV; therefore, this
patient population, which amounts to 15% of the candidates, is a priori excluded from the ICG
studies. Also excluded are RIC measurements of the CO in thoracoabdominal operations, such
as esophagectomies and pancreatectomies, where plasma loss and heavy intravenous volume
loads due to significan bleeding undermine the stability of the basal R and the reliability of
the CO results (Critchley et al 1996).

Statistical analysis

Thermodilution (TD) was used as the gold standard, and agreement between the four ICG
formulae and TD was analyzed according to the BlandAltman approach (Bland and Altman
1986).

Results6

The results of the four algorithms revealed a rather similar average of the NICO in the range
of 5 l min1 (table 1). The Pearson correlation coefficient were r = 0.969, 0.841, 0.897 and
0.950 with the formulae of Frinerman, Bonjer (1950), PattersonKubicek (Patterson 1965,
Kubicek et al 1974) and the combined PattersonKubicekFrinerman formula, respectively
(table 1, figure 2 and 3). The BlandAltman 2SD lower and upper limits of agreement with
TD were 1.088 and 0.949, 2.196 and 2.617, 1.840 and 1.617, and 1.277 and 1.217,
according to the same order of formulae as above (table 1, figur 3).

6 To calculate the SV with the Bonjer, PattersonKubicek and PattersonKubicekFrinerman formulae, the raw data
of the 43 patients were e-mailed to Professor Robert Patterson, the inventor of the TIC technology (Patterson et al
1964, Patterson 1965), who made these calculations at the University of Minnesota.

72 of 158.
 Impedance cardiography revisited 823

Discussion

To compare the performance of the currently used TIC technology with RIC, we used three
recent reports representing TIC results in patients with chronic stable congestive heart failure,
which revealed similar values (Drazner et al 2002, Van De Water et al 2003, Leslie et al 2004)
and could serve as a paradigm of the performance of TIC in the presence of cardiac conditions.
A relatively recent publication by Sageman et al (2002) could not be included here because
(a) only CI values were provided, (b) these values are characteristic of patients with normal
cardiac functions and (c) the patients CO results were averages of up to 20 measurements, an
approach which is not used clinically.
The average cardiac output in each of the three TIC series was in the range of 5 l min1.
Comparison in each series of the results with thermodilution (TD) revealed a BlandAltman
limit of agreement with a 2SD of approximately 2.2 l min1a 44% deviation compared to
TD (Drazner et al 2002, Van De Water et al 2003, Leslie et al 2004). Despite the fact that the
present series did not include cases of chronic congestive heart failure, it is comparable with
the TIC studies because (a) the patients were actively managed for cardiac conditions, (b) a
CI < 2.5 l min1 m2 was observed in 43% of the cases and (c) the average CO was in the
range of 5 l min1, as in the TIC paradigm.
When the PattersonKubicek algorithm was fed by impedance raw data yielded by the
peripheral wristankle rather than by thoracic electrodes, the disparity compared to TD was
1.7 l min1, which is a 34% difference (figure 2 and 3). When the PattersonKubicek
algorithm was reinforced by the correcting factors of the NICaS formula and the peripheral
raw data were used, there was a further decline in the disparity between ICG and TD to the
range of 1.25 l min1, a 25% difference. This is almost as good as the 20% disparity between
the NICaS and the TD.
Judging by the 20% deviation between the NICaS and TD CO results, and by the 44%
deviation between the reported TIC and TD CO results, it is evident that the accuracy of the
NICaS is higher by a factor of more than 2, than the reported values of TIC (Raaijmakers
et al 1999, Handelsman 1991, Drazner et al 2002, Van De Water et al 2003, Leslie et al 2004).
Moreover, according to the FDA standard of bioequivalence (Guidance for Industry 2001),
the comparative results of new and gold-standard technologies should be contained within a
range of 20% disparity. This 20% range is determined by the 10% repeatability rate of each
of the two methods. Judging by figur 3(A), the agreement between the NICaS and the TD is
in very close proximity to the FDA standard bioequivalence.

About the validity of correcting coefficients

To appreciate the competence of the correcting factors, it is suffic to compare the CO results
of Frinermans and Bonjers formulae (figure 3(A) and (B)). As stated earlier, the only
difference between these two equations is the addition of Frinermans correcting factors to
Bonjers formula. Ri, for example, which is the corrected R0, may reach twice the value of
the measured R0. Kw may increase the measured
R by up to 45%. The HF may either
reduce or increase the value of
R by only slight to moderate degrees. Thus, it can be
discerned that when using Bonjers equation alone (figur 3(B)) the 2SD limits of agreement
are approximately 2.4 l min1, and when Frinermans variables are added (figur 3(A)) the
2SD limits of agreement are 1.0 l min1. There is a 20% disparity between the NICaS and
the TD, and a 45% improvement of Bonjers reinforced performance.
Similarly, annexing the Frinerman correcting coefficient to the TIC dZ/dt T
formula dramatically improved the calculated CO results, from 2SD limits of agreement of

73 of 158.
 824 G Cotter et al

approximately 1.7 l min1 (table 1, figur 3(C)) to 2SD limits of agreement of approximately
1.25 l min1 (table 1, figur 3(D)).
It has been repeatedly shown by Hoffer et al (1970), Lukaski et al (1986), Organ et al
(1994), Ward et al (2000) and others that the body is not a homogeneous conductor, and its
basal resistance, and consequently the spatial distribution of conductivities, is a function of
the body composition. It is also recognized that the two main factors which determine the
impedance variabilities are the percentages of body water and fat. As shown here, by using the
RIC approach, the variabilities of age, sex, height and weight can be quantitated and translated
to the effective correcting coefficients

About the validity of the wristankle electrodes

The most significan advantage of RIC in comparison to TIC is the use of the peripheral
impedance signal for calculation of the SV. According to Kauppinen et al (2000), 75% of the
peripheral (RIC) impedance waveform is borne by the systolic blood volume pulsations of
the arterial vasculature of the upper and lower limbs, and the remaining 25% arrives from the
trunk (thorax). Still, the sole source of the peripheral signal is generated by the blood volume
pulse of the arterial vasculature.
It must be borne in mind that the aorta and its peripheral ramification comprise a single
anatomophysiological structure, and the pulse waveform, which evolves throughout its length,
occurs almost at the same time. This is possible only because the velocity of the pulse wave is
so rapid that the arterial expansion is completed before the termination of the left ventricular
contraction (Guyton and Hall 2000). The mean value of the pulse wave velocity (PWV) from
the thorax to the calves is approximately 710 m s1 (Guyton and Hall 2000).7 Similarly, the
impedance volume pulse travels at approximately the same rate as reported by Risacher et al
(1995).
Thus, in accordance with the existing knowledge of the cardiac role in the formation of
the pulse (McVeigh et al 2002), the RIC peripheral volumetric signal is borne throughout the
length of the arterial tree beginning with the left ventricular stroke volume ejection.
In contrast, the thoracic (TIC) waveform is generated by multiple sources, including
the aorta, lungs, atria, vena cava and artifacts due to heart movements (Wang et al 2001,
Kauppinen et al 1998, Wtorek 2000). In normal people, some studies have shown that TIC
gives reasonably reliable CO results (Raaijmakers et al 1999), but in the presence of cardiac
conditions, distortions of the TIC waveforms were already observed in the earliest days of the
emergence of this technology (Kubicek et al 1974). This is contrary to RIC, where we see
very few changes, if any, in the waveform shape incurred by the different cardiac conditions.
The immunity of the peripheral impedance waveform to the distorting influenc of cardiac
conditions is reflecte by the results of the present study. The average disparity between TIC-
CO and TD-CO measured in cardiac patients is 44%. In the present trial, the average disparity
between TD-CO and ICG-CO, which was calculated by the combination of the TIC algorithm
and a peripheral impedance waveform, was 34%. This 23% higher accuracy, which is obtained
by the standard TIC algorithm, can be attributed only to one factor, the reliance for calculation
of the SV on the peripheral, rather than the thoracic, impedance signal.
The fact that the RIC region consists of only part of the whole body but can be used
to calculate the CO of the whole body is explained by the electrophysiological relationships
which exist between RIC and ICGWB. In the ICGWB technique, electrodes are placed on all
four limbs, and yet, despite the fact that the head which consumes 13% of the CO is not
7 Significan differences exist in these values, depending on the elasticity of the arterial vasculature.

74 of 158.
 Impedance cardiography revisited 825

included in the electrical field the reliability of the CO results of this technology has been
validated (Tischenko 1973, Koobi et al 1999).
The common electrophysiological denominator of RIC and ICGWB is in that the same
basal R of a body that is measured by RIC is twice the value measured by ICGWB. As
measured in this series, and by and large as measured by others (Kauppinen et al 2000,
Lukaski et al 1986, Organ et al 1994), the basal R of the region between the wrist and the
ankle is approximately 450 .8 This, when the value of the basal R of ICGWB is in the range
of 200250  (Tischenko 1973, Lamberts et al 1984, Kauppinen et al 2000). The 2:1 ratio of
the basal R in RIC versus ICGWB is attributed to the fact that in ICGWB the two upper and the
two lower limbs are measured in parallel. Consequently, the cross-sectional area of the limbs
is twice as large in ICGWB, resulting in the reduction of the electrical resistance to one-half.
Thus, it is the same physiological mechanism which facilitates the appropriate measurement
of the total CO by ICGWB, which also holds for the exceptionally accurate RIC results.

Conclusions

The present CO results measured by the NICaS device indicate that, with regard to accuracy
of measuring the CO, the RIC technology outperforms any other ICG technology, being twice
as accurate as TIC.
The peripheral systolic impedance changes are more reliable than the TIC impedance
changes for calculating the cardiac stroke volume.
The main disadvantage of the RIC technology is that in approximately 15% of the patients
who need the test, there are exclusion criteria which preclude the use of the NICaS, a problem
which is also shared by other impedance technologies.

References

Bland J M and Altman D G 1986 Statistical methods for assessing agreement between two methods of clinical
measurement Lancet 1 30710
Bonjer F H 1950 Circulatieonderzoek door Impedantiemeting Groningen, Drukkerij I, Oppenheim NV
Bonjer F H, Van Den Berg J W and Direden M N J 1952 The origin of the variations of body impedance occurring
during the cardiac cycle Circulation 6 41520
Cohen A J, Arnaudov D, Zabeeda D, Shultheis L, Lashinger J and Schachner A 1998 Non-invasive measurement of
cardiac output during coronary artery bypass grafting Eur. J. Cardiothorac. Surg. 14 649
Cotter G, Moshkovitz Y, Kaluski E, Cohen A J, Miller H, Goor D A and Vered Z 2004 Accurate, noninvasive
continuous monitoring of cardiac output by whole-body electrical bioimpedance Chest 125 143140
Critchley L A H, Leung D H Y and Short T G 1996 Abdominal surgery alters the calibration of bioimpedance cardiac
output Int. J. Clin. Monit. Comput. 13 18
Drazner M, Thompson B, Rosenberg P, Kaiser P A, James M S N, Boehrer D, Baldwin B J, Dries D L and
Yancy C W 2002 Comparison of impedance cardiography with invasive hemodynamic measurements in patients
with heart failure secondary to ischemic or nonischemic cardiomyopathy Am. J. Cardiol. 89 9935
Faes Th J C, Raaijmakers J H, Meijer H G, Goovaerts H G and Heethaar R M 1999 Towards a theoretical understanding
of stroke volume estimation with impedance cardiography Ann. New York Acad. Sci. 873 12834
Frank O 1926 Die Theorie der Pulswellen Z. Biol. 85 95180
Guidance for Industry Statistical Approach to Establishing Bioequivalence 2001 US Department of Health and Human
Services, Food and Drug Administration, Center for Drug Evaluation and Research (CDER), BP
Guyton A C and Hall J E 2000 Textbook of Medical Physiology 10th edn (Philadelphia, PA: WB Saunders)

8 It should be borne in mind that in females the electrical resistance between the wrist and the ankle is higher than
in males by 80100  (Organ et al 1994, Lukaski et al 1986, Hoffer et al 1970). Furthermore, there are differences
in the basal R values of different ethnic populations (Ward et al 2000), where different antropometric characteristics
exist. In the present series, the average basal R of the postoperative cases was lower by 100  compared to the
non-operative cardiac patients, and this is attributed to the over-hydration induced during the surgery.

75 of 158.
 826 G Cotter et al

Hamwi G T 1964 Therapy Changing Dietary Concepts in Diabetes Mellitus: Diagnosis and Treatment
ed T S Danowski (New York: American Diabetes Association) pp 738
Handelsman H 1991 Public health service reassessment: measuring cardiac output by electrical bioimpedance Health
Technology Assessment Report 6 (US Dept Health and Human Services, Public Agency for Health Care Policy
and Research) pp 113
Hoffer E C, Meador C K and Simpson D C 1970 A relationship between whole body impedance and total body water
volume Ann. New York Acad. Sci. 170 45261
Kauppinen P K, Hyttinen J A and Malmivuo J A 1998 Sensitivity distributions of impedance cardiography using band
and spot electrodes analyzed by a three-dimensional computer model Am. Biomed. Eng. 26 694702
Kauppinen P K, Koobi T, Hyttinen J and Malmivuo J 2000 Segmental composition of whole-body impedance
cardiogram estimated by computer simulations and clinical experiment Clin. Physiol. 20 10613
Kim D W, Baker L E, Pearce J A and Kim W K 1988 Origins of the impedance change in impedance cardiography
by a three-dimensional f nite element model IEEE Trans. Biomed. Eng. 35 12
Koobi T, Kaukinen S and Kauppinen P 2001 Comparison of methods of cardiac output measurements Crit. Care
Med. 29 1092 (Letter to the Editor)
Koobi T, Kaukinen S and Turjanmaa V M 1999 Cardiac output can be reliably measured noninvasively after coronary
artery bypass grafting operation Crit. Care Med. 27 220611
Kubicek W G, Karnegis J N, Patterson R P, Witsoe D A and Mattson R H 1966 Development and evaluation of an
impedance cardiac output system Aerospace Med. 37 12082
Kubicek W G, Kottke F J, Ramos M U, Patterson R P, Witsol D A, Labree J W, Remole W, Layman T E, Schoening H
and Garamela J T 1974 The Minnesota impedance cardiograph: theory and applications Biomed. Eng. 9 4106
Lamberts R, Visser K R and Zijlstra W G 1984 Impedance Cardiography (Assen, The Netherlands: Van Gorkum)
pp 2194
Leslie S J, McKee S, Newby D E, Webb D J and Denvir M A 2004 Non-invasive measurement of cardiac output in
patients with chronic heart failure Blood Pressure Monit. 9 27780
Lukaski H, Bolonchuk W W, Hall C B and Siders W A 1986 Validation of tetrapolar bioelectrical impedance method
to assess human body composition J. Appl. Physiol. 60 132732
McVeigh G E, Hamilton P K and Morgan D R 2002 Evaluation of mechanical arterial properties: clinical, experimental
and therapeutic aspects Clin. Sci. 102 5167
Nyboer J 1950 Electrical impedance plethysmography: a physical and physiologic approach to peripheral vascular
study Circulation 2 81121
Organ L W, Bradham G B, Gore D T and Lozier S L 1994 Segmental bioelectrical impedance analysis: theory and
application of a new technique J. Appl. Physiol. 77 98112
Patterson R P 1965 Cardiac output determinations using impedance plethysmography MSc Thesis University of
Minnesota
Patterson R P 1985 Sources of thoracic cardiogenic electrical impedance signal as determined by a model Med. Biol.
Eng. Comput. 23 4117
Patterson R 1989 Body flui determinations using multiple impedance measurements IEEE Eng. Med. Biol. 8 169
Patterson R P, Kubicek W G, Kinnen E, Witsoe D A and Noren G 1964 Development of an electrical impedance
plethysmography system to monitor cardiac output Proc. 1st Annual Rocky Mt Bioeng. Symp. pp 5671
Raaijmakers E, Faes Th J C, Scholten R J P M, Goovaerts H G and Heethaar R 1999 A meta-analysis of published
studies concerning the validity of thoracic impedance cardiography Ann. New York Acad. Sci. 873 12134
Risacher F, Jossinet J and Schmitt M 1995 Comparison of global and local pulse wave velocity in man measured by
multichannel impedance plethysmography Proc. 9th Int. Conf. on Electrical Bioimpedance
Sageman W S, Riffenburgh R H and Spiess B D 2002 Equivalence of bioimpedance and thermodilution in measuring
cardiac index after cardiac surgery J. Cardiothorac. Vasc. Anesth. 16 814
Thomasset A 1962 Bioelectrical properties of tissue impedance measurements Lyon Med. 207 10718
Tischenko M I 1973 Estimation of stroke volume by integral rheogram of the human body Sechenov Physiol. J. 59
121624 (in Russian)
Torre-Amione G et al 2004 Non-invasive measurement of cardiac output by whole-body electrical bioimpedance in
patients treated for acute heart failure: a prospective, double-blind comparison with thermodilution J. Am. Coll.
Cardiol. 5 209A (abstract)
Van De Water J M, Miller T W, Vogel R L, Mount B E and Dalton M L 2003 Impedance cardiographythe next
vital sign technology? Chest 123 202833
Wang Y, Haynor D R and Kim Y 2001 A f nite-element study of the effects of electrode position on the measured
impedance change in impedance cardiography IEEE Trans. Biomed. Eng. 48 1390401
Wang L and Patterson L 1995 Multiple sources of the impedance cardiogram based on 3D f nite difference human
thorax models IEEE Trans. Biomed. Eng. 42 1418

76 of 158.
 Impedance cardiography revisited 827

Ward L C et al 2000 Association between ethnicity, body mass index, and bioelectrical impedance. Implications for
the population specificit of prediction equations Ann. New York Acad. Sci. 904 199204
Wotton M J, Thomas B J, Cornish B H and Ward L C 2000 Comparison of whole body and segmental bioimpedance
methodologies for estimating total body water Ann. New York Acad. Sci. 904 1816
Wtorek J 2000 Relations between components of impedance cardiogram analyzed by means of f nite element model
and sensitivity theorem Ann. Biomed. Eng. 28 135261

77 of 158.
 Value of Noninvasive Hemodynamics to Achieve Blood Pressure Control in
Hypertensive Subjects
Ronald D. Smith, Pavel Levy, Carlos M. Ferrario and for the Consideration of
Noninvasive Hemodynamic Monitoring to Target Reduction of Blood Pressure Levels
Study Group
Hypertension 2006;47;771-777; originally published online Mar 6, 2006;
DOI: 10.1161/01.HYP.0000209642.11448.e0
Hypertension is published by the American Heart Association. 7272 Greenville Avenue, Dallas, TX
72514
Copyright 2006 American Heart Association. All rights reserved. Print ISSN: 0194-911X. Online
ISSN: 1524-4563

The online version of this article, along with updated information and services, is
located on the World Wide Web at:
http://hyper.ahajournals.org/cgi/content/full/47/4/771

Subscriptions: Information about subscribing to Hypertension is online at

http://hyper.ahajournals.org/subscriptions/

Permissions: Permissions & Rights Desk, Lippincott Williams & Wilkins, a division of Wolters
Kluwer Health, 351 West Camden Street, Baltimore, MD 21202-2436. Phone: 410-528-4050. Fax:
410-528-8550. E-mail:
journalpermissions@lww.com

Reprints: Information about reprints can be found online at

http://www.lww.com/reprints

Downloaded from hyper.ahajournals.org by on June 11, 2011

78 of 158.
 Value of Noninvasive Hemodynamics to Achieve Blood
Pressure Control in Hypertensive Subjects
Ronald D. Smith, Pavel Levy, Carlos M. Ferrario; for the Consideration of Noninvasive
Hemodynamic Monitoring to Target Reduction of Blood Pressure Levels Study Group

AbstractAbnormal hemodynamics play a central role in the development and perpetuation of high blood pressure. We
hypothesized that hypertension therapy guided by noninvasive hemodynamics with impedance cardiography could aid
primary care physicians in reducing blood pressure more effectively. Uncontrolled hypertensive patients on 1 to 3
medications were randomized by 3:2 ratio to either a standard arm or hemodynamic arm that used impedance
cardiography (BioZ, CardioDynamics). Each patient completed 5 study visits with a 2-week washout period followed
by 3 months of treatment. A total of 164 patients from 11 centers completed the study, 95 in the standard arm and 69
in the hemodynamic arm. At baseline and after washout, there were no differences between arms in number of
medications or demographic, blood pressure, or hemodynamic characteristics. Systolic blood pressure reductions in the
hemodynamic arm were greater from baseline (19 mm Hg versus 11 mm Hg; P0.01) and after washout (25 mm Hg
versus 19 mm Hg; P0.05). Diastolic blood pressure reductions were also greater in the hemodynamic arm from
baseline (12 mm Hg versus 5 mm Hg; P0.001) and after washout (17 mm Hg versus 10 mm Hg; P0.001). The
hemodynamic arm achieved goal blood pressure (140/90 mm Hg) more frequently (77% versus 57%; P0.01) and
a more aggressive blood pressure level (130/85 mm Hg) more frequently (55% versus 27%; P0.0001). These study
results indicate that antihypertensive therapy guided by impedance cardiography in uncontrolled hypertensive patients
on 1 medications is more effective than standard care. (Hypertension. 2006;47:771-777.)
Key Words: hemodynamics cardiac output vascular resistance hypertension, arterial
hypertension, essential blood pressure

A pproximately 65 million people in the United States1

and 1 billion people worldwide2 have hypertension; it
is the most common reason adults visit US physicians.3
Hypertension is a major public health concern, because it
significantly increases risk of coronary artery disease,
heart failure, renal disease, and stroke.4 In spite of major
public health and medical education efforts and availability
of effective antihypertensive agents, blood pressure (BP) con-
trol rates in the United States remain low, with only 31% of
hypertensives and 54% of those actively treated and taking
medications achieving BP 140/90 mm Hg.5 Why is BP control
such an elusive goal? The reasons are numerous and complex.
However, inadequate pharmacological treatment remains the
most common cause of uncontrolled BP in actively treated
patients.6
Hypertension is a hemodynamic-related disorder. BP rises
as the result of increased systemic vascular resistance (SVR),
cardiac output (CO), fluid volume, or a combination of these
factors.7,8 Consequently, antihypertensive agents lower BP by
reducing SVR, CO, fluid volume, or combinations thereof.9
Previous authors hypothesized that hemodynamic informa-
tion could help tailor therapy and subsequently improve BP
control.10 Invasive procedures for hemodynamic profiling are
not warranted in outpatient clinics, and noninvasive proce-
dures, such as echocardiography, are costly and operator
dependent.11
Impedance cardiography (ICG) has emerged as a reliable
noninvasive method to measure hemodynamics in physician
offices. In a randomized, controlled trial, ICG-guided treat-
ment improved BP control rates in resistant hypertension
treated by hypertension specialists.12 We hypothesized that
ICG-guided treatment could aid physicians in reducing BP
more effectively than standard care in a population of
uncontrolled hypertensive patients receiving 1 to 3 medica-
tions in a primary care setting.
Methods
Eligibility
Male and female patients (age range, 18 to 75 years) were eligible if
they had a diagnosis of essential hypertension and were currently on

Received August 10, 2005; first decision August 29, 2005; revision accepted December 12, 2005.
From the Hypertension and Vascular Disease Center, Wake Forest University School of Medicine, Winston-Salem, NC.
R.D.S. and C.M.F. received honorarium from CardioDynamics for consultation and speaker fees. Site investigators received study support from
CardioDynamics.
Correspondence to Carlos Ferrario, Hypertension and Vascular Disease Center, Wake Forest University School of Medicine, Medical Center Blvd,
Winston-Salem, NC 27157-1032. E-mail cferrari@wfubmc.edu
2006 American Heart Association, Inc.
Hypertension is available at http://www.hypertensionaha.org DOI: 10.1161/01.HYP.0000209642.11448.e0

771
Downloaded from hyper.ahajournals.org by on June 11, 2011
79 of 158.
 772 Hypertension April 2006
Figure 1. Study design comparing efficacy of standard arm vs hemodynamic arm treatment of high BP.
1 to 3 antihypertensive medications with systolic BP of 140 to
179 mm Hg and/or diastolic BP of 90 to 109 mm Hg. Patients were
excluded if they were on 3 antihypertensive agents; had abnormal
laboratory findings; or had history of heart failure, left ventricular
ejection fraction 40%, atrial fibrillation, severe valvular disease,
cerebrovascular event within 3 months, severe renal disease, ne-
phrotic syndrome, or cirrhosis. Patients in whom ICG might be
subject to technical limitations were excluded (height 47 or 75
inches or weight 66 or 341 pounds, presence of activated
minute-ventilation pacemaker, known hypersensitivity to sensor gel
or adhesive, or skin lesion interfering with sensor placement).
Patients who were enrolled and subsequently found to have not met
the inclusion/exclusion criteria were terminated and excluded from
analysis. Therefore, this was not an intention-to-treat analysis. The
study was approved by an independent institutional review board,
which adheres to the Declaration of Helsinki and US Code of Federal
Regulations. These hypertensive outpatients provided written in-
formed consent and had study procedures consistent with the
protocol (no. 20021400).
Hemodynamic Evaluation Assignment
Eligible patients (N164) underwent a 2-week washout period at
which time all of the antihypertensive medications were discontinued
according to the manufacturers recommendations. After screening
and medication washout, each patient had 3 monthly office visits
Figure 2. Suggested treatment strategy for hemodynamic arm; BB indicates blocker; CAA, central acting agent; and VD, vasodilator.
Downloaded from hyper.ahajournals.org by on June 11, 2011
80 of 158.
(Figure 1). After the 2-week washout period, patients meeting
inclusion/exclusion criteria were randomized in a 3:2 ratio to the
standard arm (n95) or ICG-aided hemodynamic arm (n69) using
a central telephone service and stratified by site. All of the physicians
were educated on the hemodynamic treatment strategy illustrated in
Figure 2.
Procedures
BP determinations were made in the seated position using the
oscillometric technique. ICG data were collected by trained techni-
cians at each visit in all of the patients, but ICG findings were not
revealed in the standard arm to treating physicians or patients. ICG
was performed with patients in the supine position, resting for 5
minutes before measurement (BioZ ICG Monitor, CardioDynamics).
ICG involves the measurement of thoracic impedance through
placement of 4 dual sensors, 2 on the neck and 2 on the chest.
Electrical impedance changes are digitally processed to calculate
CO, SVR, and thoracic fluid content (TFC).13 CO and SVR are
normalized for body size by indexing to each patients body sur-
face area to obtain cardiac index (CI) and SVR index (SVRI). TFC
is the inverse of baseline chest impedance, and any changes in TFC
are directly proportional to total fluid (intravascular and extravascu-
lar) changes.14 TFC has different normal ranges for each gender that
are displayed and printed for reference. The reproducibility of this
ICG device in stable outpatients has been established,15 and accuracy
 Smith et al Noninvasive Hemodynamics and BP Control 773

has been validated versus invasive methods in patients with various enrollment criteria (BP 140/90 mm Hg at screening) and
cardiovascular disorders.16,17 were removed as protocol violations. No reported adverse
events (minor or serious) were attributable to ICG.
Outcome Measures
There were no differences in the number of antihyperten-
Physician investigators were instructed that the treatment goal was to
reduce systolic and diastolic BP as low as they believed would be sive medications, patient demographic, clinical, BP, or ICG
beneficial to the patient and to achieve sustained BP 140/ variables at baseline or after washout (Table 1). At baseline,
90 mm Hg. The primary study end points were reductions in systolic there were no differences in the percentage of patients in the
and diastolic BP from baseline and post-washout visit. Addi- standard versus hemodynamic arm on 1 (42% versus 45%;
tional study end points were achievement of: (1) goal BP 140/
90 mm Hg, (2) more aggressive BP of 130/85 mm Hg, and (3) BP P0.05), 2 (48% versus 44%; P0.05), or 3 (6% versus
of 140/90 mm Hg with normal values of CI and SVRI. Normal 10%; P0.05) medications. Baseline medication usage in the
range for CI was defined as 2.5 to 4.2 L/min per m2 and for SVRI as
1680 to 2580 dynesm2/cm5. Isolated systolic hypertension was
defined as systolic BP 140 mm Hg and diastolic BP 90 mm Hg. TABLE 1. Patient Characteristics
Standard Hemodynamic
Interventions Care Care P
After randomization, therapy was initiated in all of the patients at the Variable (n95) (n69) Value
post-washout visit, 2 weeks after screening. Physician investigators
prescribed medications consistent with published guidelines, their Age, y 54.59.4 55.29.2 ns
usual practice patterns, and patient clinical characteristics. In the Body mass index, kg/m2 30.26.3 30.85.1 ns
hemodynamic arm, the treating physician was also encouraged to use Men 51 (53.4) 38 (55.1) ns
a hemodynamic treatment strategy to guide therapeutic decisions
about pharmacological agents and dosing (Figure 2). Ethnicity
Physicians could share ICG information with patients in the White, non-Hispanic 75 (79.0) 53 (76.8) ns
hemodynamic arm, and patients in both arms received education on
White, Hispanic 7 (7.4) 5 (7.3) ns
the importance of medication compliance, which was reinforced with
a nurse telephone call midway between each study visit. Compliance Black 8 (8.4) 6 (8.7) ns
was assessed at each visit by asking patients to estimate the Asian 3 (3.2) 3 (4.4) ns
percentage of prescribed pills they had taken over the previous
History
month. Patients were considered compliant with the prescribed
protocol if pill count was 85% over the prior month. Type II diabetes mellitus 4 (4.2) 3 (4.4) ns
Ischemic heart disease 2 (2.1) 5 (7.3) ns
Statistical Analysis Hyperlipidemia 14 (14.7) 12 (17.4) ns
Data from case report forms were collected by study coordinators
and entered into a locked database. Statistical analysis was per- Baseline BP and
formed with SAS statistical analysis software, version 8.2. Contin- hemodynamics
uous variables are expressed as meanSD and categorical variables Systolic BP, mm Hg 1479 14812 ns
as n (%). Differences in continuous variables between treatment
Diastolic BP, mm Hg 8710 898 ns
groups were examined by the Student t test and by ANOVA and in
discrete variables using Fishers exact tests. Subgroup analysis was Heart rate, bpm 7512 7413 ns
performed in subjects with isolated systolic hypertension, age 55 Cardiac index, L/min/m2 2.80.5 2.90.6 ns
years, and those receiving a thiazide diuretic. Additional evaluation
of age-specific results was performed by a 2-way ANOVA for Systemic vascular 2933576 2956605 ns
achievement of BP end points, in which treatment arm and dichot- resistance index,
omized age (55 years) were included in the model. In combination dynesm2/cm5
agents, each class and dosage was counted separately for analysis. Thoracic fluid content, 28.64.9 28.04.8 ns
Equivalency of defined daily doses for each class of medication was /kOhm
calculated using World Health Organization criteria.18 Medication
Isolated systolic hypertension 46 (48.4) 31 (44.9) ns
changes were evaluated in visits where such changes affected BP end
points (visits 2 versus 1, 3 versus 2, and 4 versus 3). Medication class at baseline
and dose were compared with the prior study visit, with any change Post-washout BP and
in class or dose counted separately. Sample size was powered using hemodynamics
5 mm Hg as the detectable difference between treatment groups with Systolic BP, mm Hg 15613 15513 ns
a type I error of 5% and type II error of 20%. The expected
heterogeneity in treatment approach for patients in the standard arm Diastolic BP, mm Hg 929 949 ns
was offset by the greater number of patients randomized to the Heart rate, bpm 7912 7814 ns
standard arm. Although this approach increased the probability that
Cardiac index, L/min/m2 2.90.5 2.90.5 ns
the standard arm results would reflect actual practice patterns, it
required a larger sample size to power the study. Systemic vascular 3083630 3122672 ns
resistance index,
Results dynesm2/cm5
Eleven primary care centers participated in the study between Thoracic fluid content, 29.15.0 28.44.3 ns
November 2002 and November 2004. Of 262 patients /kOhm
screened, 184 were randomized. A total of 164 patients (95 in Medications
the standard arm and 69 in the hemodynamic arm) completed Total antihypertensive 1.70.8 1.70.7 ns
the study and were analyzed. There were 20 early termina- medications
tions, including 2 who withdrew and 18 who were random- Categorical variables expressed as n (%), continuous variables as
ized but were subsequently found not to have met BP meanSD; ns indicates not significant.

Downloaded from hyper.ahajournals.org by on June 11, 2011

81 of 158.
 774 Hypertension April 2006

standard versus hemodynamic arm was as follows: blockers

(2.1% versus 1.4%; P0.05), angiotensin converting enzyme
inhibitors (ACEI; 53.7% versus 47.8%; P0.05), angiotensin
II receptor blockers (ARB; 14.7% versus 29.0%; P0.05),
blockers (23.2% versus 13.0%; P0.05), calcium channel
blockers (CCB; 33.7% versus 39.1%; P0.05), central acting
agents (0% versus 1.4%; P0.05), diuretics (31.6% versus
26.1%; P0.05), and other vasodilators (0% versus 0%;
P0.05).
BP and ICG values at the final visit and their differences from
baseline and post-washout visits are shown in Table 2. Systolic
BP reductions were greater in the hemodynamic arm from
baseline (1917 versus 1118 mm Hg; P0.01) and post-
washout (2518 versus 1917 mm Hg; P0.05). Diastolic
BP reductions were also greater in the hemodynamic arm
from baseline (1211 versus 512 mm Hg; P0.001) and
post-washout (1712 versus 1011 mm Hg; P0.001).
Final BP was lower in the hemodynamic arm (129/7614/11
versus 136/8215/10 mm Hg; P0.01). Figure 3 demon-
strates that goal BP (140/90 mm Hg) was achieved more
Figure 3. Target BP achievement; with 95% CI; *P0.01 vs
frequently in the hemodynamic arm (77% versus 57%; standard care, P0.0001 vs standard care.
P0.01), and the more aggressive BP (130/85 mm Hg) was
also achieved more often (55% versus 27%; P0.0001). baseline (2216 versus 1117 mm Hg; P0.01) and post-
Patients with isolated systolic hypertension in the hemo- washout (2816 versus 1816 mm Hg; P0.05) than those
dynamic arm (n31) had greater systolic BP reductions from in the standard arm (n46). Patients 55 years in the
hemodynamic arm (n33) had greater systolic BP reductions
TABLE 2. Final BP and Hemodynamic Values compared with the standard arm (n51) from baseline (2117
Standard Hemodynamic versus 1120 mm Hg; P0.05) and trended greater from
Care Care P post-washout (2620 versus 2119 mm Hg; P0.05). Diastol-
Variable (n95) (n69) Value ic BP reductions were also greater in those 55 years
Systolic BP, mm Hg in the hemodynamic arm from baseline (1311 versus
Final 13615 12914 0.01 412 mm Hg; P0.001) and post-washout (1611 versus
baseline to final 1118 1917 0.01 1012 mm Hg; P0.05). In patients 55 years, goal BP
(140/90 mm Hg) was achieved more frequently in the hemo-
post-washout to final 1917 2518 0.05
dynamic arm (76% versus 53%; P0.05), and the more aggres-
Diastolic BP, mm Hg
sive BP (130/85 mm Hg) was also achieved more often (58%
Final 8210 7611 0.01 versus 27%; P0.01). ANOVA also indicated that age 55
baseline to final 512 1211 0.001 years had no effect on study end points (P0.05).
post-washout to final 1011 1712 0.001 SVRI was reduced to a greater extent in the hemodynamic
Heart rate, bpm arm than in the standard arm from baseline and post-washout.
Final 7713 7611 ns There were no significant differences between arms at the
baseline to final 112 213 ns final visit for heart rate, CI, or TFC. However, the standard
arm had a small but significant reduction in TFC from
post-washout to final 213 213 ns
post-washout to final. The percentage of patients achieving
Cardiac index, L/min/m2
normal hemodynamic values defined as simultaneously nor-
Final 2.90.5 2.90.5 ns mal values of BP, CI, and SVRI was 52% in the hemody-
baseline to final 0.10.5 0.00.5 ns namic arm and 29% in the standard arm (P0.01). Patients in
post-washout to final 0.00.5 0.00.5 ns either arm who achieved BP 130/85 mm Hg had lower
Systemic vascular resistance SVRI (2646592 versus 2855606 dynesm2/cm5; P0.05)
index, dynesm2/cm5 and lower CI (2.70.5 versus 2.90.5 L/min/m2; P0.05) than
Final 2714619 2523581 0.05 those who did not achieve BP 130/85 mm Hg. Patients in
baseline to final 219667 433660 0.05 the hemodynamic arm who achieved BP 130/85 mm Hg
post-washout to final 369642 599738 0.05 trended toward lower SVRI (2446580 versus 2573612
dynesm2/cm5; P0.05) and higher CI (2.80.5 versus
Thoracic fluid content, /kOhm
2.60.5 L/min/m2; P0.05) than those in the standard care
Final 27.84.1 28.24.9 ns
arm who achieved BP 130/85 mm Hg.
baseline to final 0.83.6 0.13.0 ns In the visit after medication washout, patients in the
post-washout to final 1.23.3 0.22.7 0.05 hemodynamic arm were more likely to be prescribed an
Variables expressed as meanSD; ns indicates not significant. ACEI, ARB, or CCB (92.8% versus 80.0%; P0.05). Over
Downloaded from hyper.ahajournals.org by on June 11, 2011
82 of 158.
 Smith et al Noninvasive Hemodynamics and BP Control 775

the course of the study, patients in the hemodynamic arm hemodynamic arms (44.2% versus 40.2%; P0.05), and use
were more likely to be prescribed an ACEI, ARB, or CCB was similar at the final visit (33.7% versus 34.8%; P0.05).
when their SVRI was high, per the hemodynamic treatment Medication doses were not different between arms except that
strategy (78.3% versus 67.1%; P0.05). However, there patients in the standard arm were on higher doses of thiazide
were no differences in the other 2 treatments encouraged by diuretics (18.98.3 versus 13.02.6 mg/day; P0.01).
the hemodynamic treatment strategy, blocker use based on There were no differences in the hemodynamic arm in the
high CI, or in diuretic use when TFC did not decrease in dosing of ACEIs (19.1 versus 19.1 mg/day; P0.05), ARBs
response to diuretic initiation or increase. Patients in the (93.9 versus 87.0 mg/day; P0.05), blockers (65.6 versus
hemodynamic arm were more likely to avoid blocker use or 80.9 mg/day; P0.05), or CCBs (7.9 versus 7.9 mg/day;
to have their blocker reduced in the presence of low or P0.05). The greater mean dose of thiazide diuretics was
normal CI (85.4% versus 77.0%; P0.05) as the hemody- because of a higher percentage of patients taking 25 mg/day
namic strategy suggested. Direct vasodilators were not used, versus 12.5 mg/day in the standard arm (40.1% versus 8.3%;
and, therefore, changes in vasodilator use in the presence of P0.05). When the study end points were analyzed only for
normal SVRI were not evaluated. Table 3 lists all of the patients on a thiazide diuretic in the final visit, patients in
medications at the final visit. Patients in the standard arm hemodynamic arm had greater decreases in systolic BP from
were on 2.00.8 medications compared with 2.10.9 for the baseline (2619 versus 817 mm Hg; P0.001) and post-
hemodynamic arm (P0.05). In the hemodynamic arm, ARB washout (3617 versus 2120 mm Hg; P0.01) and greater
use was higher (46.4% versus 30.5%; P0.05), and ACEI decreases in diastolic BP from baseline (1611 versus
use was similar (49.3% versus 53.7%; P0.05). However, 314 mm Hg; P0.001) and post-washout (2012 versus
the percentage of patients in the hemodynamic arm who were 1113 mm Hg; P0.01). There were no differences in
prescribed either an ACEI or ARB was not significantly patient-reported compliance between the standard and hemo-
different (87.0% versus 76.8%; P0.05). There were no dynamic arm in visit 3 (96.8% versus 97.1%; P0.05), 4
differences in the percentage of patients in the hemodynamic (96.8% versus 98.6%; P0.05), or 5 (100% versus 100%;
care on 1 (25% versus 26%), 2 (48% versus 53%), 3 (19% P0.05) or when these visits were combined (97.9% versus
versus 15%), 4 (9% versus 5%), or 5 (0% versus 1%) 98.6%; P0.05).
medications at the final visit (P0.05 for all). There were a
greater number of medication dose increases in the standard Discussion
versus hemodynamic arm (3.61.3 versus 3.01.2; P0.001), Our results demonstrate that ICG-guided antihypertensive
as well as a greater number of dose decreases (2.71.3 versus treatment was more effective in reducing BP than standard
1.71.0; P0.001). Medication class changes in the standard therapy and empiric selection of antihypertensive medica-
and hemodynamic arm were similar in both class initiation tions. Patients in the 2 arms of our study were not signifi-
(1.00.9 versus 1.10.9; P0.05) and removal (0.80.8 ver- cantly different at baseline, and each patient underwent a
sus 0.70.8; P0.05). medication washout period to additionally equalize the 2
Thiazide diuretic use at baseline was similar in the standard groups. The 57% BP control rate in the standard arm was
versus hemodynamic arm (28.4% versus 24.6%; P0.05). A substantial and compared favorably to BP control rates of
similar proportion of patients were prescribed thiazide diuret- long durations in large antihypertensive trials.19,20 However,
ics at some point during the trial in both the standard and the 77% BP control rate in the hemodynamic arm was even
more impressive with an 8/7 mm Hg greater BP reduction
TABLE 3. Final Antihypertensive Medications from baseline and a 6/7 mm Hg greater BP reduction from
post-washout. As a result, patients in the hemodynamic arm
Standard Hemodynamic achieved goal BP of 140/90 mm Hg 35% more often (77%
Antihypertensive Care Care P
versus 57%) and the more aggressive level of BP control
Medication (n95) (n69) Value
(130/85 mm Hg) 104% more often (55 versus 27%) than
No. at final visit 2.00.8 2.10.9 ns those in the standard arm. The hemodynamic arm maintained
Blocker 1 (1.0) 1 (1.4) ns superiority in 3 key subgroups: patients who were older, on
ACEI 51 (53.7) 34 (49.3) ns thiazide diuretics, or had isolated systolic hypertension.
ARB 29 (30.5) 32 (46.4) 0.05 Why did the hemodynamic arm achieve greater reductions
Blocker 18 (19.0) 6 (8.7) ns in BP and higher BP control rates than the standard arm? The
Calcium channel blocker, 36 (37.9) 28 (40.6) ns fundamental difference between the two arms was that patient
dihydropyridine treatment in the hemodynamic arm was individualized and
Calcium channel blocker, 6 (6.3) 7 (10.1) ns targeted at the hemodynamic abnormality associated with the
nondihydropyridine elevated BP. This approach led to greater reductions in SVRI
Central acting agent 0 (0.0) 1 (1.4) ns in the hemodynamic arm, which allowed greater decreases in
both systolic and diastolic BP. The mechanistic and hemo-
Diuretic, thiazide 32 (33.7) 24 (34.8) ns
dynamically based improvement in BP was also demon-
Diuretic, loop 1 (1.1) 0 (0.0) ns
strated in patients achieving BP 130/85 mm Hg through
Diuretic, potassium sparing 6 (6.3) 3 (4.3) ns significantly lower SVRI and higher CI in both arms. In
Vasodilator 0 (0.0) 0 (0.0) ns theory, the larger drop in SVRI and BP levels in the
Categorical variables expressed as n (%); ns indicates not significant. hemodynamic arm could have occurred through use of more
Downloaded from hyper.ahajournals.org by on June 11, 2011
83 of 158.
 776 Hypertension April 2006
medications, more effective medications, greater dosing in-
tensity, more effective combination therapy, or better patient
compliance. Our study allowed full discretion by the physi-
cian in choosing the agents, and a multitude of classes and
doses within classes were used. The study was not powered to
find small disparities in medication use, and most medication
differences did not reach statistical significance.
On the other hand, some differences are worth noting.
Patients in the standard arm were more likely to experience
both increases and decreases in their medication doses,
whereas medication class changes were not different between
arms. This result might have been expected, because treat-
ment in the standard arm followed guidelines and usual
practice patterns in which a stepped approach to therapy
contributes to a trial-and-error method of determining
whether agents and doses are working. In the hemodynamic
arm, the initial selection of antihypertensive medications
appears to have been influenced by the hemodynamic data,
because these patients were more likely to be prescribed a
vasodilating agent to reduce SVRI. Additionally, the hemo-
dynamic treatment strategy influenced medication use when
SVRI was considered high, because patients in the hemody-
namic arm were more likely to have received an ACEI, ARB,
or CCB, as was suggested. The hemodynamic treatment
strategy did not influence the prescription of blockers in the
presence of high CI or in diuretic use in response to TFC
changes. However, -blocker use was lower or reduced in the
presence of low or normal CI in the hemodynamic arm.
Although the final number of antihypertensive medications
given to patients in both arms of the study was similar,
patients in the hemodynamic arm were more likely to be
prescribed an ARB. However, when ACEI and ARB use was
combined into a single category (reninangiotensinaldoste-
rone system inhibitors), the hemodynamic arm only trended
toward greater use at the final visit (87.0% versus 76.8%).
Thiazide diuretic use increased during the study but was
lower than in some pharmacological trials and what hyper-
tension guidelines currently suggest. However, the percent-
age of patients in both arms who were prescribed a thiazide
diuretic at the final visit was very similar to the 35.6% usage
that was reported in recent analysis of over 25 000 hyperten-
sive patients.21 The lower use of diuretics and blockers also
follows the previously recognized physician preference for
other antihypertensive agents.22 Some might hypothesize that
greater BP reductions could have been achieved in the
standard arm if diuretics were used more frequently. How-
ever, when patients taking a thiazide diuretic were examined
as a subgroup, the hemodynamic arm maintained its superi-
ority. Additionally, although the higher doses of thiazide
diuretics in the standard arm may have contributed to a
greater drop in TFC from the post-washout visit, they did not
lead to better BP control.
Our study was not intended to evaluate whether a particular
antihypertensive agent was more effective at reducing BP
than another. Rather, it was designed to determine whether
providing hemodynamic data to the physician and patient
could more effectively reduce BP. Whether hemodynamic
data led to a more tailored approach to selection and moni-
toring of antihypertensive agents or by other factors, it
Downloaded from hyper.ahajournals.org by on June 11, 2011
84 of 158.
resulted in greater reduction in BP and SVRI and better BP
control. Physicians cannot adequately estimate hemodynam-
ics from routine clinical examination or BP measurements,23
because at similar levels of BP, SVR and CO can vary widely.
Therefore, the addition of accurate, noninvasive, and readily
obtainable hemodynamic measurements is clinically relevant.
Importantly, the current study also showed that patients in
the hemodynamic arm were almost twice as likely to achieve
BP control with normalization of both CI and SVRI. Im-
provements in vascular resistance may result in greater
benefits in reducing cardiovascular risk than improvement in
BP alone,24 and differences in SVRI at the same BP may
explain poorer prognosis for men versus women25 and black
versus nonblack patients.26 Hemodynamics are also known to
change with age. In older subjects, decreased arterial com-
pliance and CI lead to increased SVRI, arterial BP, and pulse
pressure.27 In spite of the expected differences in the hemo-
dynamics of older patients, this study demonstrated that
hemodynamically driven, individualized therapy was simi-
larly effective regardless of age or existence of isolated
systolic hypertension.
The use of ICG to achieve greater BP control offers the
potential for better short-term use of healthcare resources. In
addition, the long-term benefits of even small levels of BP
reduction are well known. A sustained BP reduction of
4/3 mm Hg is expected to reduce stroke risk 23%, coronary
heart disease events 15%, heart failure 16%, and overall
mortality 14%.28 Accordingly, a recent meta-analysis of
major hypertension trials reveals that an antihypertensive
agent is judged favorably when it produces mean BP im-
provements versus placebo of only 3 or 4 mm Hg or versus
another antihypertensive agent of only 1 or 2 mm Hg.29
Previously, ICG has been used to profile hemodynamic
variability across BP values30 and to identify left ventricular
dysfunction.31 Changes in ICG parameters have demonstrated
the hemodynamic effect of antihypertensive agents32,33 and
dietary sodium.34 ICG-guided therapy has shown benefit in a
case series,35 observational study,36 and a randomized trial in
resistant hypertensive patients.12 In the randomized trial,
ICG-guided therapy resulted in better final BP and greater BP
control. Similar to our study, that study showed no differ-
ences in the number of medications between arms. In contrast
to our study with lower diuretic doses and fewer medication
changes in the hemodynamic arm, resistant hypertension
patients receiving ICG-guided therapy had higher diuretic
doses and more medication changes. The differences between
the studies might be expected because of the difference in
patients (severe hypertension on more medications versus
milder hypertension on fewer medications) and setting (spe-
cialist versus generalist). However, in both studies, ICG-
guided therapy led to more effective treatment as evidenced
by better BP outcomes.
The conclusions of this study may be limited to its duration
of 3 months. However, in pharmacological trials, short-term
reductions in BP are typically sustained over longer periods.37
Another limitation may be in our use of patient-reported
medication compliance. Without using automatic counting
procedures, our goal was to educate both arms equally and to
reinforce patient compliance with follow-up phone calls.
 Smith et al
Lastly, treatment differences in the hemodynamic arm do not
imply superiority of one medication over another, because the
study was not designed to evaluate this question.
Perspectives
The results of this study indicate that ICG-guided antihyper-
tensive therapy in uncontrolled hypertensive patients on 1 to
3 antihypertensive medications is more effective than stan-
dard care. This was evident by greater reductions in systolic
and diastolic BP and by achieving a better level of BP control.
Our study showed that, in clinical practice, inclusion of ICG
hemodynamic assessment may improve BP control rates in
patients who are not controlled on initial therapy.
Acknowledgments
Consideration of Noninvasive Hemodynamic Monitoring to Target
Reduction of Blood Pressure Levels (CONTROL) Site Investigators:
Fred E. Abbo, Douglas C. Beatty, Donald M. Brandon, Milan L.
Brandon, Anthony V. Dallas, Jr, Lawrence Dinenberg, Mazhar El
Amir, Nigar Enayat, Neil W. Hirschenbein, Dorothy Lebeau, Ber-
nard A. Michlin, John Millspaugh, Nell Nestor, Melissa Noble,
Robin F. Spiering, Joseph Taylor, and Allen R. Walker. Statistical
support was provided by Gerard Smits and the study was sponsored
by CardioDynamics (San Diego, CA).
References
1. Fields LE, Burt VL, Cutler JA, Hughes J, Roccella EJ, Sorlie P. The
burden of adult hypertension in the United States 1999 to 2000: a rising
tide. Hypertension. 2004;44:398 404.
2. Kearney PM, Whelton M, Reynolds K, Muntner P, Whelton PK, He J.
Global burden of hypertension: analysis of worldwide data. Lancet. 2005;
365:217223.
3. Chery DK, Woodwell DA. National Ambulatory Care Medical Survey:
2000 summary. Advance Data. 2002;328:132.
4. American Heart Association. Heart Disease and Stroke Statistics 2004
Update. Dallas, Texas: American Heart Association; 2003.
5. Wang Y, Wang QJ. The prevalence of prehypertension and hypertension
among US adults according to the new joint national committee guidelines:
new challenges of the old problem. Arch Intern Med. 2004;164:21262134.
6. Yakovlevitch M, Black HR. Resistant hypertension in a tertiary care
clinic. Arch Intern Med. 1991;151:1786 1792.
7. Ferrario CM, Page IH. Current views concerning cardiac output in the
genesis of experimental hypertension. Circ Res. 1978;43:821 831.
8. Davidson RC, Ahmad S. Hemodynamic profiles in essential and sec-
ondary hypertension. In: Izzo JL, Black HR, eds. Hypertension Primer.
3rd ed. Dallas, TX: Council on High Blood Pressure Research, American
Heart Association; 2003:349 351.
9. Sullivan JM, Schoeneberger TE, Ratts ET, Palmer ET, Samaha JK,
Mance CJ, Muirhead EE. Short-term therapy of severe hypertension.
Hemodynamic correlates of the antihypertensive response in man. Arch
Intern Med. 1979;139:12331239.
10. Egan B, Schmouder R. The importance of hemodynamic considerations
in essential hypertension. Am Heart J. 1988;116:594 599.
11. Northridge DB, Findlay IN, Wilson J, Henderson E, Dargie HJ. Non-
invasive determination of cardiac output by Doppler echocardiography
and electrical bioimpedance. Br Heart J. 1990;63:9397.
12. Taler SJ, Textor SC, Augustine JE. Resistant hypertension: Comparing he-
modynamic management to specialist care. Hypertension. 2002;39:982988.
13. Ventura HO, Taler SJ, Strobeck JE. Hypertension as a hemodynamic
disease: the role of impedance cardiography in diagnostic, prognostic, and
therapeutic decision making. Am J Hypertens. 2005;18:26S 43S.
14. Van De Water JM, Mount BE, Chandra KM, Mitchell BP, Woodruff TA,
Dalton ML. TFC (thoracic fluid content): a new parameter for assessment
of changes in chest fluid volume. Am Surg. 2005;71:81 86.
15. Treister N, Wagner K, Jansen PR. Reproducibility of impedance cardi-
ography parameters in outpatients with clinically stable coronary artery
disease. Am J Hypertens. 2005;18:44S50S.
16. Van De Water JM, Miller TW. Impedance cardiography: the next vital
sign technology? Chest. 2003;123:2028 2033.
17. Yung G, Fedullo P, Kinninger K, Johnson W, Channick R. Comparison
of impedance cardiography to direct fick and thermodilution cardiac
Downloaded from hyper.ahajournals.org by on June 11, 2011
85 of 158.
Noninvasive Hemodynamics and BP Control 777
output determination in pulmonary arterial hypertension. Congest Heart
Fail. 2004;10(2 suppl 2):710.
18. Defined Daily Dose (DDD) developed by the World Health Organization
for use in drug utilization studies. WHO Collaborating Centre for Drug
Statistics Methodology. Anatomical Therapeutic Chemical (ATC) Clas-
sification Index With Defined Daily Doses (DDDs). Oslo, Norway: WHO
Collaborating Centre; 2000.
19. Dahlof B, Devereux RB, Kjeldsen SE, Julius S, Beevers G, de Faire U,
Fyhrquist F, Ibsen H, Kristiansson K, Lederballe-Pedersen O, Lindholm
LH, Nieminen MS, Omvik P, Oparil S, Wedel H; LIFE Study Group.
Cardiovascular morbidity and mortality in the Losartan Intervention for
Endpoint reduction in hypertension study (LIFE): a randomised trial
against atenolol. Lancet. 2002;359:9951003.
20. ALLHAT Officers and Coordinators for the ALLHAT Collaborative
Research Group. The Antihypertensive and Lipid-Lowering Treatment to
Prevent Heart Attack Trial. Major outcomes in moderately hypercholes-
terolemic, hypertensive patients randomized to pravastatin vs usual care:
The Antihypertensive and Lipid-Lowering Treatment to Prevent Heart
Attack Trial (ALLHAT-LLT). JAMA. 2002;288:2998 3007.
21. Singh H, Johnson ML. Prescribing patterns of diuretics in multi-drug
antihypertensive regimens. J Clin Hypertens. 2005;7:81 87.
22. Ubel PA, Jepson C, Asch DA. Misperceptions about -blockers and
diuretics: a national survey of primary care physicians. J Gen Intern Med.
2003;18:1060 1061.
23. Connors AF, Dawson NV, Shaw PK, Montenegro HD, Nara AR, Martin
L. Hemodynamic status in critically ill patients with and without acute
heart disease. Chest. 1990;98:1200 1206.
24. Fagard RH, Pardaens K, Staessen JA, Thijs L. Prognostic value of
invasive hemodynamic measurements at rest and during exercise in
hypertensive men. Hypertension. 1996;28:3136.
25. Messerli FH, Garavaglia GE, Schmieder RE, Sundgaard-Riise K, Nunez
BD, Amodeo C. Disparate cardiovascular findings in men and women
with essential hypertension. Ann Intern Med. 1987;107:158 161.
26. Hinderliter A, Blumenthal J, Waugh R, Chilukuri M, Sherwood A. Ethnic
differences in left ventricular structure: Relations to hemodynamics and
diurnal blood pressure variation. Am J Hypertens. 2004;17:43 49.
27. Alfie J, Galarza C, Waisman G. Noninvasive hemodynamic assessment of
the effect of mean arterial pressure on the amplitude of pulse pressure.
Am J Hypertens. 2005;18:60S 64S.
28. Blood Pressure Lowering Treatment Trialists Collaboration. Effects of
different blood-pressure-lowering regimens on major cardiovascular
events: results of prospectively-designed overviews of randomized trials.
Lancet. 2003;362:15271545.
29. Williams B. Recent hypertension trials: implications and controversies.
J Am Coll Cardiol. 2005;45:813 827.
30. Abdelhammed AI, Smith RD, Levy P, Smits GJ, Ferrario CM. Nonin-
vasive hemodynamic profiles in hypertensive subjects. Am J Hypertens.
2005;18:51S59S.
31. Bhalla V, Isakson S, Bhalla MA, Lin JP, Clopton P, Gardetto N, Maisel
AS. Diagnostic ability of B-type natriuretic peptide and impedance car-
diography: testing to identify left ventricular dysfunction in hypertensive
patients. Am J Hypertens. 2005;18:73S 81S.
32. Mitchell A, Buhrmann S, Saez AO, Rushentsova U, Schafers RF, Philipp
T, Nurnberger J. Clonidine lowers blood pressure by reducing vascular
resistance and cardiac output in young, healthy males. Cardiovasc Drugs
Ther. 2005;19:49 55.
33. Breithaupt-Grogler K, Gerhardt G, Lehmann G, Notter T, Belz GG.
Blood pressure and aortic elastic propertiesverapamil SR/trandolapril
compared to a metoprolol/hydrochlorothiazide combination therapy. Int
J Clin Pharmacol Ther. 1998;36:425 431.
34. Ashida T, Nishioeda Y, Kimura G, Kojima S, Kawamura M, Imanishi M,
Abe H, Kawano Y, Yoshimi H, Yoshida K, Kuramochi M, Omae T.
Effects of salt, prostaglandin, and captopril on vascular responsiveness in
essential hypertension. Amer J Hypertens. 1989;2:640 642.
35. Sharman DL, Gomes CP, Rutherford JP. Improvement in blood pressure
control with impedance cardiograph-guided pharmacologic decision
making. Congest Heart Fail. 2004;10:54 58.
36. Sramek BB, Tichy JA, Hojerova M, Cervenka V. Normohemodynamic
goal-oriented antihypertensive therapy improves the outcome (abstract).
Amer J Hypertens. 1996;9:141A.
37. Julius S, Kjeldsen SE, Weber M, Brunner HR, Ekman S, Hansson L, Hua
T, Laragh J, McInnes GT, Mitchell L, Plat F, Schork A, Smith B,
Zanchetti A. VALUE trial group. Outcomes in hypertensive patients at
high cardiovascular risk treated with regimens based on valsartan or
amlodipine: the VALUE randomised trial. Lancet. 2004;363:20222031.
 The European Journal of Heart Failure 8 (2006) 136 140
www.elsevier.com/locate/heafai

Non-invasive measurement of cardiac output by whole-body

bio-impedance during dobutamine stress echocardiography: Clinical
implications in patients with left ventricular dysfunction and ischaemia
Marina Leitman a, Edgar Sucher a, Edo Kaluski a, Ruth Wolf a, Eli Peleg a, Yaron Moshkovitz b,
Olga Milo-Cotter a, Zvi Vered a, Gad Cotter c,*
a
Cardiology Department, Assaf-Harofeh Medical Center, Israel
b
Cardiac Surgical Department, Ramat-Marpe-Assuta Medical Center, Petah Tikva, Sackler School of Medicine, Tel-Aviv University, Israel
c
Duke Clinical Research Institute, PO Box 17969, Durham NC, 27715, USA

Received 20 December 2004; accepted 30 June 2005

Available online 29 September 2005

Abstract

Objectives: To compare non-invasive determination of cardiac index (CI) by whole body electrical bioimpedance using the NICaS apparatus
and Doppler echocardiography, and the role of cardiac power index (Cpi) and total peripheral resistance index (TPRi) calculation during
dobutamine stress echocardiography (DSE).
Subjects and methods: We enrolled 60 consecutive patients undergoing DSE. Patients were prospectively divided into 3 groups: Group 1
(n = 20): normal DSE (control). Group 2 (n = 20): EF < 40% without significant ischaemia. Group 3 (n = 20): patients with significant
ischaemia on DSE. Measurements of CI were performed at the end of each stage of DSE by both echocardiographic left ventricular
outflow track flow and the NICaS apparatus, using whole-body bio-impedance. MAP was measured simultaneously and TPRi and Cpi
were calculated.
Results: The correlation between non-invasive CI as determined by NICaS and echocardiography was 0.81, although Echocardiographic
readings of CI were higher during administration of higher doses of dobutamine. Lower EF correlated with lower Cpi, especially stress
induced Cpi. Hence, patients with reduced EF (group 2) had a blunted increase in Cpi during stress. Patients with ischaemia (group 3) had a
blunted increase in Cpi as well as a decrease in Cpi and increase in TPRi during the last stages of DSE.
Conclusion: Measurement of CI by NICaS correlated well with Doppler derived CI. The calculation of Cpi and TPRi changes during
dobutamine stress may provide important clinical information.
D 2005 European Society of Cardiology. Published by Elsevier B.V. All rights reserved.

Keywords: Cardiac index; Cardiac power; Bioimpedance

1. Background and aims
Cardiac power index (Cpi) is the product of simulta-
neously measured cardiac index (CI) and mean arterial
blood pressure (MAP). Cpi increases cardiac power
reserve during stress [1,2] or dobutamine administration
[3] was shown in previous studies to be an important
* Corresponding author. Duke Clinical Research Institute, 2400 Pratt
Street, Durham, 27715 NC, USA.
E-mail address: cotte013@mc.duke.edu (G. Cotter).
1388-9842/$ - see front matter D 2005 European Society of Cardiology. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.ejheart.2005.06.006
measure of systolic cardiac contractile reserve, better than
VO 2 and echocardiographic ejection fraction (EF).
Recently, Samejima et al. [4] demonstrated that CO
increase during stress using non-invasive CO determina-
tion with bioimpedance was correlated with stress induced
dyspnea.
The use of hemodynamic measures such as increase in
vascular resistance for the detection of ischaemia was
suggested almost a decade ago [5]. However, this research
avenue has not been pursued due to the lack of simple non-
invasive devices for CO measurement. Recently Weiss et al.

86 of 158.
 M. Leitman et al. / The European Journal of Heart Failure 8 (2006) 136 140 137

Table 1 2. Patients and methods

Baseline characteristics of patients in the three groups
Normal Ischaemia EF < 40% P value We enrolled 60 consecutive patients undergoing stand-
DSE by DSE
ard DSE using incremental dobutamine infusion from 10
N 20 20 20 to 40 Ag/min and atropine up to 1 mg as required to reach
Age (years) 59 T 10 59 T 17 54 T 20 0.6
the pre-determined target heart rate. Patients were recruited
Prior MI 25% 55% 90% 0.003
Hypertension 50% 75% 80% 0.12 in our outpatient clinic during a once weekly session. All
Diabetes mellitus 25% 45% 50% 0.46 consecutive patients attending the clinic during that day for
Hyperlipidaemia 35% 55% 40% 0.6 the purpose of DSE were considered for the study. Patients
Prior smoking 30% 35% 30% 0.7 were divided into 3 groups: Group 1: Control. Normal
Family history of ischaemic 30% 10% 10% 0.3
DSE, including baseline EF >40% and no significant
heart disease
ischaemia, Group 2: LV systolic dysfunction as determined
Medical therapy by baseline echocardiographic EF < 40% without signifi-
Aspirin 75% 95% 85% 0.17 cant ischaemia and Group 3: Significant ischaemia as
Beta-blockers 45% 75% 80% 0.03 determined by improvement of contractility during low-
ACEi/AT II antagonists 30% 55% 90% 0.04
dose dobutamine infusion followed by decreased contrac-
Diuretics 10% 45% 90% 0.03
Statins 45% 80% 75% 0.1 tility during high dose dobutamine infusion in at least one
non-infarcted myocardial segment.
Echocardiographic findings Exclusion criteria were inability to achieve good
Left ventricular EF (%) 57 T 6 50 T 11 28 T 7 <0.001 echocardiographic visualization, significant hypotensive or
LVH 40% 50% 45% 0.78
hypertensive reactions or tachy or bradyarrhythmias during
EF: Ejection fraction, MI: Myocardial infarction, ACEi: Angiotensin dobutamine infusion and inability to reach the pre-deter-
converting enzyme inhibitors, ATII antagonists angiotensin II antago-
nists, LVH: Left ventricular hypertrophy.
mined heart rate.

2.1. Study protocol

[6] demonstrated that in patients with significant ischaemia
during stress, CO increase by bio-impedance is lower than
in patients without ischaemia.
The NICaS apparatus uses whole body bio-impedance and
the Tsoglin Frinerman formula for non-invasive determi-
nation of CI [7]. In short, a small electrical current is
transferred from the left wrist to the right foot, and the
impedance to its transit is detected (termed whole-body bio-
impedance). The instantaneous change in bio-impedance has
previously been shown to be related to the pulsatile changes
in the volume of the great arteries. The Tsoglin Frinerman
formula uses this change in bio-impedance (DR) as well as
population based constants correcting for age, sex, weight
and body composition (electrolytes, haematocrit and changes
in baseline bio-impedance) to calculate the stroke volume
(SV). Thereafter by electrocardiographically measuring the
pulse rate it calculates cardiac output and CI. In a few recently
published studies [7 9], NICaS measurements of CI in
patients with various cardiac conditions showed good
DSE was performed according to a standard protocol by
2 DSE teams. NICaS CI was measured by one NICaS
operator. CI was determined by both echocardiographic left
ventricular outflow track (LVOT) diameter and flow
velocity as well as the NICaS apparatus. Operators
measuring CI by one method were blinded to the result
of the other method throughout the examination. There-
after, based on NICaS determined CI, we calculated Cpi
and total peripheral vascular resistance (TPRi) for each of
the above mentioned time-points. NICaS and Doppler
determined CI were not calculated during dobutamine
administration in 3 patients (5%) in whom it was judged by
10
9
R=0.81
8
7
6
NICas CI

reproducibility and correlated well with thermodilution

(R = 0.8 0.9), with no bias and precision of approximately 5
0.6 L/min/M2. 4
The aim of the present study was two fold: first, to 3
compare CI measurements by NICaS and Doppler echocar- 2
diography over a wide range of values during dobutamine 1
stimulation and, secondly, to determine whether the non- 0
invasive continuous measurement of CI and MAP and 0 2 4 6 8 10
calculation of Cpi and TPRi changes during dobutamine Echo CI
stress could be used for diagnosis of significant left Fig. 1. Correlation between Doppler echocardiography CI and NICaS CI
ventricular (LV) dysfunction or myocardial ischaemia as measurements. Horizontal axis CI measured by echocardiography, L/
determined by dobutamine stress echocardiography (DSE). min/M2, Vertical axis: CI measured by NICaS, L/min/M2.

87 of 158.
 138 M. Leitman et al. / The European Journal of Heart Failure 8 (2006) 136 140

Difference in CI (Echo Doppler, NICaS)

3

-1

-2

-3

-4

-5
0 1 2 3 4 5 6 7 8 9
Mean CI (Echo Doppler; NICaS)
Fig. 2. Difference of CI measurements by Doppler echocardiography and NICaS vs. their mean (Bland and Altman analysis).

the operator that LVOT obstruction occurred due to Analysis of variance with repeated measurements (time *
dobutamine administration. group) was used to compare changes in Cpi and TPRi
over time in the different groups. All P values < 0.05
2.2. Study end-points were considered significant.

(1) The correlation between NICaS and Doppler echo-

cardiography derived CI and (2) the absolute and relative 3. Results
changes in NICaS determined Cpi and TPRi during dobut-
amine stress in the 3 groups. Sixty consecutive patients where enrolled in the 3
prospectively defined groups. The baseline characteristics
2.3. Statistical methods of the three groups are presented in Table 1. As expected,
patients differed with respect to baseline echocardio-
All data is reported based on pre-determined group graphic EF, age and severity of background diseases.
allocations. To compare the baseline characteristics of the The correlation between CI as determined by echocardio-
groups we used the Students t-test to compare continuous graphic LVOT area and mean flow velocity and the NICaS
variables and the chi-square test to compare categorical apparatus was R = 0.81 (Fig. 1). The Bland Altman
variables. Since the cardiac index and cardiac power distribution of CI measurements is depicted in Fig. 2. The
measurements were not normally distributed we used the correlation was better for CI measurements at baseline and
Spearman-rank test for comparison of NICaS and echo- during the infusion of dobutamine at doses of up to 20 Ag/
cardiographically determined CI and for comparison of min than for CI determinations during administration of
resting EF and Cpi during the different stages of DSE. dobutamine at a rate of 30 and 40 Ag/min (Table 2, Fig. 2),
Since both NICaS and Doppler echocardiography are not due to increasing bias (i.e., CI measurements by Doppler
regarded as gold standard for CI determination, when echocardiography tending to be higher) and lower precision.
comparing the two methods we used the Bland and Therefore, the CI increase in the different stages of DSE was
Altman [10] recommendations and for each dobutamine similar by both techniques (Fig. 3). Again, a slight tendency
stage, as well as for the whole cohort, we determined bias was observed for higher increase in the Doppler echocar-
(mean difference between the 2 methods) and the limits of diography determined CI during the last phase of dobut-
agreement (precision) calculated as 2 SD of the bias. amine infusion.

Table 2
CI measurements obtained by Doppler echocardiography and NICaS during the different DSE stages
Dobutamine infusion Mean NICaS-CI Mean echo-Doppler CI Spearman Rank Bias Precision
rate Correlation
Baseline 2.9 T 0.6 2.8 T 0.6 0.77 0.06 0.39
10 Ag/min 3.4 T 0.9 3.4 T 1.1 0.81 0.04 0.71
20 Ag/min 4.1 T1.1 4.2 T 1.5 0.87 0.16 0.8
30 Ag/min 4.7 T 1.4 4.9 T 1.5 0.81 0.24 1.16
40 Ag/min 4.7 T 1.2 5.2 T 1.4 0.62 0.49 1.18
All measurements 3.9 T 1.3 4.1 T1.5 0.81 0.17 0.9

88 of 158.
 M. Leitman et al. / The European Journal of Heart Failure 8 (2006) 136 140 139

5.5 1.4
5 1.2
CI (Liter/min/M2)

Cpi (watt/M2)
4.5 1
4 0.8
3.5 0.6
3 0.4
2.5 0.2
0 10 20 30 40
Dobutamine dose (mcg/min) 0
0 10 20 30 40
NICaS CI Echo-Doppler CI
Dobutamine dose(mcg/min.)
Fig. 3. Trend changes in CI during dobutamine infusion: Doppler Normal DSE Ischaemia EF<40%
echocardiography vs. NICaS CI.
Fig. 4. Changes in Cpi during dobutamine infusion in the 3 groups.

We have observed a significant correlation between EF and

Cpi based both on echocardiographically determined CI and p = 0.0002) and increase in TPRi. (+ 279 T 636 vs. 59 T
on NICAS-CI (Table 3). Interestingly, and consistent with 169 dynes s M2, P = 0.022).
previous studies [11], this correlation was better for stress No significant adverse events were recorded during the
Cpi than for rest Cpi. dobutamine stress test.
During dobutamine infusion, NICaS determined Cpi
increase was significantly different in the three groups
(Fig. 4). Cpi increase was smallest in the group of patients 4. Discussion
with LV dysfunction followed by the group of patients
with significant ischaemia and highest in patients with The results of the present study demonstrate that CI
normal DSE ( p = 0.002 comparing LV dysfunction with determination by whole-body bio-impedance using the
normal DSE, p = 0.03 comparing patients with ischaemia NICaS device is correlated well with CI determination by
with patients with normal DSE and p = 0.02 comparing Doppler echocardiography. However, during infusion of
patients with ischaemia to patients with LV dysfunction). higher doses of dobutamine ( 30 Ag/min), the correlation
Baseline TPRi was significantly higher at baseline in the became less accurate, mainly due to significant bias, i.e.
LV systolic dysfunction group as compared to patients with Doppler echocardiography CI measurements tended to be
normal DSE (3120 T 1020 vs. 2450 T 940 dynes s M2, significantly higher than NICaS readings.
p = 0.04) however, during dobutamine stress it decreased In the overall cohort we found a correlation between
steeply in all 3 groups, to a similar degree. severity of LV dysfunction by resting EF and Cpi at rest and
A significant difference was observed in Cpi and TPRi especially Cpi during the dobutamine stress; i.e., the lower
changes during the last phase of dobutamine infusion in the EF the lower the rest and peak DSE Cpi. Hence, in the
patients in the significant ischaemia group. As compared to group of patients with reduced baseline EF (group 2), Cpi
patients in the control as well as the systolic LV dysfunction increase during exercise was significantly blunted. This
group, patients who were found by DSE to have significant finding is substantiated by the results of previous studies
ischaemia had during the last phase of DES a significant showing that lower Cpi increase during stress (lower cardiac
decrease in Cpi ( 0.16 T 0.15 vs. + 0.1 T 0.15 W/M2, power reserve) is correlated with poor outcome although
this correlation was superior to the correlation of EF and
outcome. In the present study, the small number of patients
Table 3 enrolled did not allow for outcome analysis, however, it is
Correlations between resting EF and Cpi by NICaS and echocardiography possible that accurate non-invasive CI determination and
during the different DSE stages calculation of Cpi reserve by whole body bio-impedance
Dobutamine Spearman Rank Spearman Rank Correlation during stress may become a useful predictor of outcome in
infusion rate Correlation between between resting EF and patients with reduced left ventricular function.
resting EF and NICaS-Cpi at different The results of the present cohort, in concordance with
echo-Doppler Cpi at DSE stages
different DSE stages
previous studies [5,6] show that during significant ischae-
mia, cardiac power tends to decrease and vascular resistance
Baseline 0.44 0.44
10 Ag/min 0.56 0.39
increases. Again, such non invasive calculation may enable
20 Ag/min 0.61 0.54 an additional important indication for significant ischaemia
30 Ag/min 0.58 0.53 in addition to conventional signs on DSE.
40 Ag/min 0.59 0.45 Importantly, all haemodynamic data was obtained in the
All measurements 0.57 0.52 present study using a simple non-invasive device. Although

89 of 158.
 140 M. Leitman et al. / The European Journal of Heart Failure 8 (2006) 136 140
the size of the present cohort does not allow for far-reaching
conclusions, if these results are reproduced by additional,
larger studies, cardiac power and vascular resistance
changes could be used for non-invasive detection of left
ventricular dysfunction and ischaemia during simple stress
tests such as electrocardiographic exercise stress test or
mental stress test. Moreover, serial changes in these
measures may be useful for improving detection of
ischaemia in patients at home using telemedicine.
4.1. Study limitations
The study included a small, select group of patients
referred for DSE due to various symptoms. Hence, the
results require further confirmation in a larger study
including a non-selected group of outpatients.
5. Conclusion
The results of the present study suggest that a simple
stress test using dobutamine infusion and non-invasive
determination of CI by the NICaSi 2001 apparatus and
calculation of Cpi and TPRi can be used for easy out-patient
screening of patients for systolic LV dysfunction and
myocardial ischaemia.
References
[1] Williams SG, Cooke GA, Wright DJ, Parsons WJ, Riley RL, Marshall
P, et al. Peak exercise cardiac power output. A direct indicator of
90 of 158.
cardiac function strongly predictive of prognosis in chronic heart
failure. Eur Heart J 2001;22:1496 503.
[2] Cohen-Solal A, Tabet JY, Logeart D, Bourgoin P, Tokmakova M,
Dahan M. A non-invasively determined surrogate of cardiac power
(Fcirculatory power_) at peak exercise is a powerful prognostic factor
in chronic heart failure. Eur Heart J 2002;23:806 14.
[3] Marmor A, Schneeweiss A. Prognostic value of noninvasively
obtained left ventricular contractile reserve in patients with severe
heart failure. J Am Coll Cardiol 1997;29:422 8.
[4] Samejima H, Omiya K, Uno M, Inoue K, Tamura M, Itoh K, et al.
Relationship between impaired chronotropic response, cardiac output
during exercise, and exercise tolerance in patients with chronic heart
failure. Jpn Heart J 2003;44:515 25.
[5] Mohr R, Rath S, Meir O, Smolinsky A, Har-Zahav Y, Neufeld HN,
et al. Changes in systemic vascular resistance detected by the arterial
resistometer: preliminary report of a new method tested during
percutaneous transluminal coronary angioplasty. Circulation 1986;
74:780 5.
[6] Weiss SJ, Ernst AA, Godorov G, Diercks DB, Jergenson J, Kirk JD.
Bioimpedance-derived differences in cardiac physiology during
exercise stress testing in low-risk chest pain patients. South Med J
2003;96:1121 7.
[7] Cohen AJ, Arnaudov D, Zabeeda D, Schultheis L, Lashinger J,
Schachner A. Non-invasive measurement of cardiac output during
coronary artery bypass grafting. Eur J Cardiothorac Surg 1998;14:
64 9.
[8] Cotter G, Moshkovitz Y, Kaluski E, Cohen AJ, Miller M, Goor DA,
et al. Accurate, non-invasive continuous monitoring of cardiac output
by whole body electrical bio-impedance. Chest 2004;125:1431 40.
[9] Torre-Amione G, Cotter G, Kaluski E, Salah A, Milo O, Richter C,
et al. Non-invasive measurement of cardiac output by whole-body
electrical bioimpedance in patients treated for acute heart failure: a
prospective, double blind comparison with thermodilution. J Am
Coll Cardiol 2004;43:109 [abstract].
[10] Bland JM, Altman DG. Statistical methods for assessing agreement
between two methods of clinical measurement. Lancet 1986;1:307 10.
[11] Cotter G, Williams SG, Vered Z, Tan LB. The role of cardiac power in
heart failure. Curr Opin Cardiol 2003;18:215 22.
 CLINICAL INVESTIGATIONS

Impact of Impedance Cardiography on

Diagnosis and Therapy of Emergent
Dyspnea: The ED-IMPACT Trial
W. Frank Peacock, MD, Richard L. Summers, MD, Jody Vogel, MD, Charles E. Emerman, MD

Abstract
Background: Dyspnea is one of the most common emergency department (ED) symptoms, but early diag-
nosis and treatment are challenging because of multiple potential causes. Impedance cardiography (ICG) is
a noninvasive method to measure hemodynamics that may assist in early ED decision making.
Objectives: To determine the rate of change in working diagnosis and initial treatment plan by adding ICG
data during the course of ED clinical evaluation of elder patients presenting with dyspnea.
Methods: The authors studied a convenience sample of dyspneic patients 65 years and older who were
presenting to the EDs of two urban academic centers. The attending emergency physician was initially
blinded to the ICG data, which was collected by research staff not involved in patient care. At initial ED
presentation, after history and physical but before central lab or radiograph data were returned, the at-
tending ED physician completed a case report form documenting diagnosis and treatment plan. The phy-
sician then was shown the ICG data and the same information was again recorded. Pre- and post-ICG
differences were analyzed.
Results: Eighty-nine patients were enrolled, with a mean age of 74.8  7.0 years; 52 (58%) were African
American, 42 (47%) were male. Congestive heart failure and chronic obstructive pulmonary disease
were the most common final diagnoses, occurring in 43 (48%), and 20 (22%), respectively. ICG data
changed the working diagnosis in 12 (13%; 95% CI = 7% to 22%) and medications administered in 35
(39%; 95% CI = 29% to 50%).
Conclusions: Impedance cardiography data result in significant changes in ED physician diagnosis and
therapeutic plan during the evaluation of dyspneic patients 65 years and older.
ACADEMIC EMERGENCY MEDICINE 2006; 13:365371 2006 by the Society for Academic Emergency
Medicine
Keywords: dyspnea, hemodynamics, impedance cardiography, bioimpedance, cardiac output, systemic
vascular resistance, noninvasive

D
yspnea is one of the most common emergency
department (ED) symptoms in older patients.1
Various conditions, including heart failure,
chronic obstructive pulmonary disease, pneumonia, pul-
monary embolus, and acute coronary syndromes, may
occur alone or in combination in a given patient, adding
uncertainty to diagnosis and treatment. In patients with
both cardiac and pulmonary disease, the initial assess-
ment and therapy in the ED are challenging.
Because cardiovascular disease, specifically decom-
pensated heart failure (HF), is a relatively common cause
of dyspnea in elders, an assessment of hemodynamics,

From the Department of Emergency Medicine, Cleveland Clinic
(WFP), Cleveland, OH; Department of Emergency Medicine,
University of Mississippi (RLS), Jackson, MS; Wayne State
University (JV), Detroit, MI; and Department of Emergency
Medicine, Case Western Reserve University (CEE), Cleveland, OH.
Received September 6, 2005; revision received November 7,
2005; accepted November 15, 2005.
Supported by GE Medical Systems (Milwaukee, WI), which
provided devices and disposables for this study, and by Cardio-
Dynamics (San Diego, CA), which provided a study grant for
support of research assistants. In addition, CardioDynamics
participated in the creation of the educational process for the par-
ticipating physicians before study commencement. The sponsor
had no role in data collection or statistical analyses, and the man-
uscript is the sole responsibility of the authors. W.F.P. and R.L.S.
received honoraria in 2003 for speaking for CardioDynamics.
Presented as a moderated poster at the American College of
Emergency Physicians Research Forum, October 2003.
Address for correspondence and reprints: W. Frank Peacock,
MD, The Cleveland Clinic Foundation, Department of Emer-
gency Medicine, Desk E-19, 9500 Euclid Avenue, Cleveland,
OH 44195. Fax: 216-445-4552; e-mail: peacocw@ccf.org.

2006 by the Society for Academic Emergency Medicine ISSN 1069-6563

doi: 10.1197/j.aem.2005.11.078 PII ISSN 1069-6563583 365

91 of 158.
366 Peacock et al.
including cardiac output, systemic vascular resistance,
and fluid status, may provide important information
and aid decision making beyond what is possible from
history and physical examination alone. Unfortunately,
hemodynamic parameters cannot be accurately deter-
mined by patient history or physical examination.25 Until
recently, hemodynamic data could only be obtained by
pulmonary artery catheterization. Because this invasive
procedure is not practical in the ED, physicians typically
are left to make diagnosis and treatment decisions with-
out reliable information about a patients hemodynamic
status.
Noninvasive hemodynamic monitoring by impedance
cardiography (ICG) has been used in more than four
million patients. Cardiac output (CO) by ICG has been
shown to correlate well with CO obtained by invasive
methods in hospitalized patient populations with correla-
tion coefficients for CO by ICG and thermodilution rang-
ing from 0.76 to 0.89.610 ICG also has been used as an
alternative to invasive monitoring in the critical care
setting.11 In the ED setting, ICG has been studied for the
differential diagnosis of dyspnea1214 and the identification
of pulmonary edema15,16 and provides prognostic infor-
mation about hospitalization costs and length of stay.17
ICG results are available within a few minutes, allowing
more rapid patient evaluation than that afforded by
radiographic or laboratory studies.
Given the high rate of morbidity, mortality, and hospi-
tal readmissions for patients with dyspnea and acute de-
compensated HF, there is an urgent need to examine
technologies that could lead to improvements in care in
the ED. The present study examines an aspect of thera-
peutic efficacy18 as it relates to ICG and the acutely dysp-
neic emergency patient, and not simply the performance
of ICG as a testing modality. Put into context, previous
studies of commonly used ED tools, such as pulse oxim-
etry19,20 and B-type natriuretic peptide (BNP) testing,21
suggest that a 5% to 11% rate of change in diagnosis,
or 10% rate of change in therapy, is clinically relevant.
The effect of ICG-derived hemodynamics on diagnosis
and treatment of dyspnea in the ED is not yet known.
The purpose of this study was to determine the rate of
change in diagnosis and therapy resulting from the avail-
ability of ICG data during the initial evaluation of older
ED patients presenting with dyspnea.
METHODS
Study Design
This was a prospective study of dyspneic patients that
was designed to determine the frequency of change in
the ED physicians initial diagnosis and therapeutic plan
after physician access to noninvasive ICG hemodynamic
data. The study was approved by each hospitals institu-
tional review board. All patients gave informed consent
before enrollment in the study.
Study Setting and Population
The setting was two large urban academic EDs with
experience in using noninvasive hemodynamic monitor-
ing. A convenience sample was obtained from patients
age 65 years or older who were presenting with a chief
complaint of dyspnea or symptoms of HF, as determined
92 of 158.
 IMPEDANCE CARDIOGRAPHY IN EMERGENT DYSPNEA
by the ED physician. Because ICG is not currently recom-
mended (per U.S. Food and Drug Administration guide-
lines) for the diagnosis of acute coronary syndromes,
including acute myocardial infarction (MI), patients were
excluded if electrocardiogram (ECG) or serum markers
were positive for acute MI. Additional exclusions in-
cluded the following: if ICG monitoring was not possible
because of inability to place electrodes, if the patients
weight was greater than 341 pounds, or if the patient
had an activated minute ventilation pacemaker (which
uses an impedance signal). Also, although severe aortic
regurgitation that could give a falsely elevated ICG CO
is rare and generally evident on ED evaluation, we ex-
cluded those with aortic regurgitation by past history,
and those with the typical diastolic murmur. Last, because
the treatment and disposition actions for patients needing
immediate intubation and mechanical ventilation are gen-
erally well defined from the emergency physician point
of view, and because it was our intent to study the
diagnostically most challenging patients, those requiring
urgent intubation and mechanical ventilation upon
presentation to the ED were excluded.
Study Protocol
Project coordinators screened candidates and an inde-
pendent research nurse, not involved in the diagnosis
or treatment of the patient, obtained hemodynamic
data. Hemodynamic data were collected by using the
BioZ ICG monitor (CardioDynamics, San Diego, CA), as
has been described elsewhere.22 ICG data are obtained
by the following technique: four dual sensors (each sen-
sor consisting of two electrodes) are placed on the
patient, as shown in Figure 1, on opposite sides of the
neck at a level between the ears and shoulders and on ei-
ther side of the chest in the mid-axillary line at the level of
the xiphoid process. The outer electrodes in each sensor
transmit a low-amplitude, high-frequency current (2.5
mA, 70 kHZ), and the inner electrodes detect thoracic
voltage changes. Changes in voltage are used to calculate
changes in impedance (Z). Baseline, static impedance is
indicative of chest fluid volume, and dynamic impedance
Figure 1. Front view of impedance cardiography method.
 ACAD EMERG MED  April 2006, Vol. 13, No. 4  www.aemj.org
is affected by aortic blood volume and velocity. Beat-
to-beat changes in thoracic impedance are processed to
calculate blood flow per heartbeat (stroke volume) and
per minute (cardiac output). By using standard equations,
other hemodynamic parameters, such as systemic vascu-
lar resistance, are calculated. The reciprocal of baseline
thoracic impedance can provide an index of intrathoracic
fluid and is termed thoracic fluid content (TFC). TFC has
been used to identify intravascular and extravascular
fluid changes23,24 and to titrate diuretic therapy.25
Before study initiation, participating physicians re-
ceived instruction regarding the interpretation of the
hemodynamic values obtained by the ICG device.
Attending physicians, all of whom were board-certified
or board-eligible in emergency medicine, were given a
description of ICG technology and hemodynamic param-
eters provided on the ICG report, including definitions
and normal values for cardiac index (CI), resistance, tho-
racic fluid content, and measures reflecting left ventricu-
lar performance. This was performed at departmental
grand rounds and at the monthly attending-physician
staff meeting. Additional information was disseminated
in hardcopy by mailing and was duplicated by e-mail.
The pathophysiology of HF and hemodynamic findings
most suggestive of dyspnea caused by decompensated
HF (reduced CI, elevated systemic vascular resistance,
and increased TFC) were described. The expected effects
of various medications on hemodynamic parameters
were discussed, including use of diuretics, vasodilators,
and drugs affecting contractility. Additionally, physicians
were provided personal reference cards for use at their
discretion that detailed normative values for all ICG
data. Copies of the data card were also kept fixed to the
ICG device. These data were also shown at the time of
ICG unblinding. For any given parameter, ICG data are
presented as a bar indicating the normal human range.
The average result and the currently measured data point
then are indicated on this bar, such that variations from
normal are readily apparent.
All staff involved with patient care were blinded to the
ICG data until after the initial history and physical exam-
ination by the attending physician. After the initial his-
tory and physical exam, but before initiation of therapy
(other than supplemental oxygen), and before obtaining
any central laboratory or radiographic data, the attend-
ing physician completed a case report form indicating
his or her working diagnoses and short-term therapeutic
plans. The physician was then immediately shown the
ICG hemodynamic data and was asked to complete the
case report form again, this time with consideration of
the ICG data. All patient care then proceeded according
to usual ED routine. Blood tests, including electrolytes,
blood urea nitrogen (BUN), serum creatinine (Cr), and
BNP levels, were obtained in the majority of cases.
Although these data were not mandated as part of the
protocol, they were used in most cases to determine final
ED diagnosis.
Measures
The two primary endpoints were 1) the rates of change
in working diagnosis and 2) medical therapy after the
addition of ICG data to the physicians initial clinical
assessment and therapeutic plan. In the cases in which
93 of 158.
367
the diagnosis changed on the basis of ICG data, a com-
parison to the final diagnosis was made to determine
whether the pre- or post-ICG diagnosis was more consis-
tent with the final ED diagnosis. The final ED primary di-
agnosis was defined as the principal diagnosis at the end
of the ED visit after all diagnostic testing was completed
and reviewed by the ED physician responsible for dispo-
sition. A change in therapy was defined as the addition or
subtraction of a drug or procedure. Changing the dose of
a previously ordered drug was not considered a thera-
peutic change. Adverse events were defined as cardiac
arrest, intubation for respiratory failure, urgent cardio-
version, or blood transfusion.
Data Analysis
The size of the study was prospectively determined on the
basis of the number needed to detect a 5% rate of change
in diagnosis or therapy. Given an alpha of 0.05 and a beta
of 0.20, a sample size of 100 was needed to detect a statis-
tically significant change. Data were analyzed by an inde-
pendent statistician using SAS Software (Cary, NC).
Demographic data are reported descriptively. Continu-
ous variables are reported as mean  standard deviation
(SD). Rates of change were calculated by dividing the
number of patients in whom diagnosis or therapeutic
plan changed by the total number of patients and were
reported as percentages. An analysis of variance was
performed to assess for differences among vital signs
and ICG parameters in the final diagnosis categories.
RESULTS
Eighty-nine patients, cared for by 31 ED staff physicians,
were enrolled from December 2001 through July 2003
and are included in the analysis. No adverse event,
defined as cardiac arrest, intubation, cardioversion, or
blood transfusion, occurred during the course of the
ED observation during this study.
The patient characteristics and vital signs are summa-
rized in Table 1. The mean (SD) age of the subjects
was 74.8 (7.0) years. Fifty-eight percent of the patients
were African American, and 61% had a history of HF,
Table 1
Patient Characteristics and Vital Signs
Patient Characteristic Value
Male, n (%) 42 (47)
Ethnic background, n (%)
African American 52 (58)
White 36 (41)
Other 1 (1)
Medical history, n (%)
History of heart failure 43 (48)
History of chronic obstructive 34 (38)
pulmonary disease or asthma
History of heart failure and COPD 11 (13)
Vital signs (mean  SD)
Temperature (C) 36.5  0.7
Respiratory rate (min1) 22.2  5.1
Heart rate (min1) 84.6  18.8
Systolic blood pressure (mm Hg) 145.6  29.3
N = 89.
368 Peacock et al.  IMPEDANCE CARDIOGRAPHY IN EMERGENT DYSPNEA

Table 2 Diagnosis Changes

Hemodynamic Values at Presentation
Hemodynamic Parameter Mean  SD Normal Range
Thoracic fluid content 34.2  10.4 3050 (male)
(kOhm1) 2137 (female)
Systemic vascular resis- 1,685  579
tance (dyne sec cm5)
Cardiac index (L/min/m2) 2.6  0.6
Stroke index (mL/m2) 31.8  10.1
N = 89.
including 13% with a history of both HF and chronic lung
disease. The prevalence of chronic lung disease, includ-
ing asthma, was 38%. The average respiratory rate was
22.2 (5.1) min1 with systolic BP and heart rate of
145.5 (29.3) mm Hg and 84.6 (18.8) min1, respectively.
The hemodynamic values for the population as a whole
are listed in Table 2.
Patients could be categorized by final primary diagno-
sis at the time of ED discharge or hospital admission into
three major groups: 1) HF (n = 43); 2) chronic obstructive
pulmonary disease (COPD; 20); and 3) other (26). The
other group included other cardiovascular and lung
conditions not included in the HF or COPD groups: atrial
fibrillation (n = 4), bronchitis (4), hypertension (2), pneu-
monia (2), pulmonary hypertension (2), anemia (1), influ-
enza (1), lung cancer (1), palpitations (1), upper
respiratory infection (1), atypical chest pain (1), hypoxia
(1), intra-abdominal abscess (1), non-cardiac shortness
of breath (1), pulmonary fibrosis (1), vertigo (1), and de-
hydration (1).
Chest radiographs and ECG results were recorded by
the ED physician in 85 patients (96%). The various ECG
and radiographic findings are summarized in Table 3.
ECG findings were described as normal or nonspecific
in the vast majority (82%), and the chest radiograph
was normal or nondiagnostic in nearly half, with only
16% showing either HF or upper zone redistribution
consistent with pulmonary venous congestion.
Table 3
Electrocardiographic and Chest Radiographic Findings
Parameter
Electrocardiographic findings
Normal, nonspecific, or nondiagnostic ECG
Left ventricular hypertrophy
Atrial fibrillation
Prior myocardial infarction
Left bundle branch block
Relevant chest radiography findings
Normal or no acute disease
Increased cardiothoracic ratio (>0.5)
or cardiomegaly
Pleural effusion
Pulmonary edema or heart failure
Upper zone redistribution
Pulmonary infiltrate
Atelectasis
Hyperinflation
For each finding group, n = 85. Because of multiple findings, total for ECG
diagnoses is greater than 85.
A summary of the rates of diagnosis and therapy change
that resulted from ICG data are presented in Figure 2.
ICG data changed the working diagnosis in 12 (13%;
95% CI = 7% to 22%). When diagnoses were categorized
as either cardiac or noncardiac, the post-ICG diagnosis
7421,378
was the same as the final diagnosis in 8 of 12 patients
2.54.2 in whom ICG resulted in a change (67%, 95% CI = 35%
3565 to 90%). Of the four patients in whom a change in diag-
nosis after ICG did not match the final ED diagnosis,
one who was ultimately diagnosed with a cardiac cause
of dyspnea had normal hemodynamic parameters, sug-
gesting a pulmonary cause. In another patient, altered
hemodynamic parameters suggested cardiac dyspnea
that was later attributed to an exacerbation of COPD.
One patient with lung cancer had hemodynamic findings
consistent with diastolic HF. Finally, one patient who
initially was thought to have pulmonary dyspnea had
altered hemodynamic findings that were believed to be
nondiagnostic by the evaluating physician; that patient
was ultimately treated for fluid overload and discharged
home.
Results Grouped by Final Diagnosis
A summary of the patient vital signs and hemodynamic
characteristics grouped by final ED diagnosis is listed
in Table 4. No diagnosis group had vital sign data that
were significantly different from those of any other
group (p = 0.1332). Of the hemodynamic parameters,
cardiac index, systemic vascular resistance index, and
thoracic fluid content had one diagnosis group that
differed significantly from the other two (p < 0.02). HF
patients had greater amounts of lung water, as reflected
by a mean TFC (38.5  12.3 kOhm1) that was signifi-
cantly higher than that of the other two diagnosis groups
(30.0  6.17 and 30.4  5.6 for the COPD and other
groups, respectively). Patients with COPD had higher
CI (3.08  0.57 vs. 2.39  0.56 and 2.48  0.65) and lower
SVR (1,361  407 vs. 1,772  565 and 1,789  638) than did
patients in the HF or other groups, respectively.
Laboratory measurements, including electrolytes,
BUN, Cr, WBC, Hgb, and BNP were analyzed by
final diagnosis. Of the laboratory measurements, only
BNP, measured in 72 patients, exhibited a statistically
n (%)
70 (82)
15 (18)
11 (13)
8 (9)
3 (4)
42 (49)
13 (15)
8 (9)
8 (9)
6 (7)
5 (6)
2 (2)
1 (1)
Figure 2. Category rate of change from pre-ICG to post-ICG
(mean values with 95% condence intervals).

94 of 158.
 ACAD EMERG MED  April 2006, Vol. 13, No. 4  www.aemj.org 369

Table 4 final diagnosis of COPD, and 17 medication changes in

Initial Vital Signs and Selected Hemodynamic Parameters by the 26 patients in the other group. Twenty-three of 54
Final Diagnosis (43%) medication changes that resulted from the avail-
ability of hemodynamic information were changes in
Final Diagnosis
the use of diuretics or bronchodilators.
Chronic
Obstructive
Heart Pulmonary DISCUSSION
Failure Disease Other
Parameter (n = 43) (n = 20) (n = 26)
In cases of elder dyspneic patients who may require
urgent treatment, the ED physician must assess status,
Temperature (C) 36.4  0.7 36.6  0.8 36.7  0.5
Systolic blood 149.5  30.5 135.6  19.1 146.8  32.2
formulate a working diagnosis, and institute therapy, in
pressure (mm Hg) many cases before all information is available. Hemo-
Diastolic blood 80.2  18.0 75.7  13.0 78.5  21.2 dynamic information, which reflects the contribution of
pressure (mm Hg) the cardiovascular system to the current presentation,
Heart rate (min1) 84.1  18.0 88.2  19.4 82.9  20.1 may have an important impact on the process of care.
Respiration rate (min1) 22.1  5.7 23.8  3.8 21.0  4.6 Our results demonstrate that knowledge of ICG data
Thoracic fluid content 38.5  12.3* 30.0  6.17 30.4  5.6 leads to a change of working primary diagnosis in 13%
(kOhm1) of elder patients presenting with dyspnea to the ED.
Systemic vascular 1,772  565 1,361  407* 1,789  638 When changes in diagnosis were made, they were con-
resistance (dyne
sistent with the final diagnosis at time of ED disposition
sec cm5)
Cardiac index 2.39  0.56 3.08  0.57* 2.48  0.65
in two-thirds of cases. In addition to changes in diagno-
(L/min/m2) sis, ED physicians made medication changes on the basis
Mean arterial 100.0  20.0 98.5  14.2 100.7  21.8 of ICG-derived hemodynamic information in 39% of
pressure (mm Hg) cases. Finally, unlike vital signs, which were similar across
Stroke index (mL/m2) 29.7  9.9 36.3  9.7 31.6  10.0 the various diagnostic groups, hemodynamic data varied
Data are mean  SD. based on causes of dyspnea. These findings are consistent
* Different from other two categories, p < 0.02. with the hypothesis that hemodynamic information is
relevant and actionable in the ongoing evaluation and
treatment of such patients.
significant difference (p < 0.0001) among the three diag- Patients presenting with dyspnea are commonly at risk
nosis groups. The HF group had a significantly higher for exacerbation of either cardiac or pulmonary disease.
mean BNP level (940 pg/mL) than did the other diagnosis Those with acute HF typically have reduced cardiac
groups (137 pg/mL, and 357 pg/mL for COPD and other output and elevated vascular resistance. Those with a
groups, respectively). pulmonary or other noncardiac cause of their dyspnea
typically have normal cardiac output and hemodynamic
Treatment Changes parameters. Because the accuracy and reproducibility
Changes in planned medication orders, occurring after of ICG have been validated in a variety of patient popula-
ICG information was revealed (and without other input tions and settings, it is not surprising that physicians
to the treating physician), are shown in Table 5. Thirty- used this information to help guide diagnosis and treat-
five patients (39%, 95% CI = 29% to 50%) had a total of ment in dyspneic patients. Our finding of different values
54 changes in the medication plan after initial assessment of hemodynamic parameters among the diagnostic groups
and review of ICG data. Use of diuretics most often was is consistent with this paradigm. The relatively high rate
altered based on ICG findings, suggesting fluid overload of change of diagnosis, when ICG-derived information
or cardiac cause of dyspnea. When looking at medication was revealed to treating physicians, suggests acceptance
changes by category of final diagnosis, there were 17 of the technical and diagnostic accuracy efficacies of the
medication changes in the 43 patients with a final diagno- test. Not only does the information result in altered diag-
sis of HF, 20 medication changes in the 20 patients with a nosis, but the noninvasive hemodynamic data provided
by ICG was applied by the physicians to therapeutic de-
cision making, an indication of therapeutic efficacy, as
Table 5 defined by Pearl.18 Thus, our results support the potential
Therapeutic Changes Post- vs. Pre-ICG Listed by Medication value of such information and support a practical role for
Class this technology in the ED assessment of such patients.
Medication Class n (%) [95% CI] Recently, BNP testing has been shown to be a useful
bedside tool to aid in diagnosis of patients presenting
Diuretics 12 (13) [7%22%]
Nitroglycerin 4 (4) [1%11%] with shortness of breath.26 However, despite the avail-
Bronchodilators 11 (12) [8%24%] ability of point-of-care laboratory testing, real-time diag-
Steroids 6 (7) [3%14%] nosis and treatment can be delayed. In fact, one large
Antibiotics 6 (7) [3%14%] trial of cardiac markers found that even with point-of-
Anticoagulants 5 (6) [2%13%] care testing, the door-to-brain time (the time from ED
Other 10 (11) [6%20%] arrival until cardiac marker results are available for the
There were 54 total medication changes in 35 patients. Because of mul- physician to act upon) exceeded one hour.27 ICG data
tiple therapeutic changes, total changes are greater than the number of are available within several minutes. And, unlike the he-
patients.
modynamic information obtained by a pulmonary artery
95 of 158.
370 Peacock et al.
catheter, performance of ICG is noninvasive and can be
readily accomplished in the ED without specialized train-
ing and at minimal risk to the patient.
The magnitude of the changes in diagnosis and treat-
ment resulting from ICG-derived hemodynamic data
can be compared with that from other technologies
that are currently the standard of care in most EDs. His-
torically, changes in therapy on the order of 5% to 11%
appear to define utility of testing in the ED. In one study,
Summers et al.19 reported that the ED physician assess-
ment of patient severity of illness was changed by pulse
oximetry in 3% of cases. Kosowky et al.,21 evaluating
BNP testing in patients older than 40 years of age, found
that BNP data changed the diagnosis in 10%, and treat-
ment in 11%, of cases. These trials suggest that the rate
of change that resulted from ICG use in the present study
would be clinically significant in the ED environment.
Moreover, ICG can be performed concurrently with
existing diagnostic and therapeutic strategies, such that
the information is incremental in the decision-making
process.
The changes in ED decision making from esophageal
Doppler results, a more invasive and less common form
of cardiac output measurement, have been studied else-
where.28 Those investigators found a change in man-
agement decisions in 31% of cases. Our results show a
greater change in therapy alone, perhaps because of
the incremental information provided by systemic vascu-
lar resistance and thoracic fluid content parameters.
Although most ED physicians would not subject a patient
to esophageal monitoring to obtain hemodynamic mea-
surements, it is likely that many would consider the
collection of ICG data, which requires little more time or
inconvenience than obtaining an electrocardiogram.
We did not measure the time required to obtain ICG
data; however, in routine use, these data can be obtained
in about 3 to 5 minutes and require 30 to 60 seconds
to interpret. Because ICG provides early and accurate
data, there is a potential for significant clinical impact
from its use. We did not specifically study the financial
effects of ICG in this study or how it might have affected
length of ED stay or hospital admission rate. However,
at a procedural cost for each test of less than $20 and
with the cost of a day in intensive care at more than
$1,000, the provision of ICG would be cost-effective
even if, for example, it reduced hospital length of stay
by only one day for every 50 patients monitored.
LIMITATIONS
We acknowledge several limitations. This study evalu-
ated the effect of ICG on working diagnosis and initial
treatment plan before the results of chest radiograph,
ECG, or BNP level. Thus, it is impossible to gauge the rel-
ative importance of the information obtained from ICG
to that obtained by these other tests or to judge the addi-
tional contribution of ICG for cases in which the results
of other tests were available before performing ICG. Al-
though blood work and various ancillary testing such as
chest radiography are part of the complete ED evaluation
of such patients, the results are generally not available
within the first few minutes of patient assessment. By de-
sign, this study evaluated ICGs effect on working diag-
96 of 158.
 IMPEDANCE CARDIOGRAPHY IN EMERGENT DYSPNEA
nosis and therapy in a manner that would be consistent
with clinical practice in the ED, where patients present-
ing with dyspnea might be evaluated with ICG either
before or within minutes of the ED physicians initial
assessment. Furthermore, as seen in our study, the find-
ings of ECGs and chest radiographs are often normal or
nonspecific and may not provide significant diagnostic
certainty. Because ICG is not part of the diagnostic crite-
ria for acute coronary syndromes, including acute MI, we
did exclude patients with evidence of myocardial necro-
sis from analysis. Therefore, the role of ICG in providing
possible clues in the evaluation of patients with dyspnea
as a manifestation of MI cannot be assessed by the pres-
ent study.
Our study was also limited by the use of the final ED
diagnosis as the criterion standard for diagnostic cate-
gorization. Although it is possible that this diagnosis
was incorrect or incomplete in some patients, this repre-
sents the real-life diagnosis based on current evaluation
strategies during the patients ED visit. It is also possible
that a physician had the right diagnosis and treatment
plan before reviewing ICG results and that ICG data
resulted in inappropriate therapies. A larger prospective
outcome-based study will be required to determine the
potential for this to occur.
In our study, ICG data were available and likely con-
tributed to the final ED diagnosis, thereby introducing
possible bias. However, the goal of this study was not
to assess technical accuracy of the technology, which
has been evaluated in previous studies. In contrast, this
study was designed to assess whether physicians would
incorporate early hemodynamic information into the
process of formulating an initial working diagnosis and
treatment plan. In addition, the study design does not al-
low us to draw conclusions about the sensitivity or spec-
ificity of ICG criteria, or to compare diagnostic accuracy
to other measures, such as BNP or chest radiography.
The accuracy of the post-ICG diagnosis based on these
hemodynamic criteria could only be verified by a more
standardized diagnostic approach including cardiac
imaging studies, blinded reviews of subsequent hospital
records with adjudication of discordant diagnoses, and
long-term follow-up, which were not within the scope
of the current study.
CONCLUSIONS
Knowledge of ICG data early in the ED evaluation of pa-
tients older than 65 years of age presenting with dyspnea
results in significant changes in diagnosis and treatment
plan. Whether changes in diagnosis, diagnostic certainty,
or therapy from ICG improve outcomes or are cost-effec-
tive will require a prospective, randomized clinical trial
with longer periods of clinical follow-up.
The authors thank Gerard Smits, PhD, for his statistical assis-
tance.
References
1. McCraig LF, Burt CW. National hospital ambulatory
medical care survey: 2001 emergency department
summary. Report from Centers for Disease Control
 ACAD EMERG MED  April 2006, Vol. 13, No. 4  www.aemj.org
and Prevention, National Center for Health Statistics.
Advance Data Vital Health Stat. 2003; 335:18.
2. Eisenberg PR, Jaffe AS, Schuster DP. Clinical evalua-
tion compared to pulmonary artery catheterization in
the hemodynamic assessment of critically ill patients.
Crit Care Med. 1984; 12:54953.
3. Speroff T, Connors AF Jr, Dawson NV. Lens model
analysis of hemodynamic status in the critically ill.
Med Decis Making. 1989; 9:24352.
4. Neath SX, Lazio L, Guss DA. Utility of impedance car-
diography to improve physician estimation of hemo-
dynamic parameters in the emergency department.
Congest Heart Fail. 2005; 11:1720.
5. Van De Water JM, Dalton ML, Parish DC, Vogel RL,
Beatty JC, Adeniyi SO. Cardiopulmonary assess-
ment: is improvement needed? World J Surg. 2005;
29(Suppl 1):S958.
6. Sageman WS, Riffenburgh RH, Spiess BD. Equiva-
lence of bioimpedance and thermodilution in mea-
suring cardiac index after cardiac surgery. J
Cardiothorac Vasc Anesth. 2002; 16:814.
7. Drazner M, Thompson B, Rosenberg P, Yancy C.
Comparison of impedance cardiography with inva-
sive hemodynamic measurements in patients with
heart failure secondary to ischemic or nonischemic
cardiomyopathy. Am J Cardiol. 2002; 89:9935.
8. Van De Water JM, Miller TW. Impedance cardiogra-
phy: the next vital sign technology? Chest. 2003; 123:
202833.
9. Yung GL, Fedullo PF, Kinninger K, Johnson FW,
Channick RN. Comparison of impedance cardiogra-
phy to direct Fick and thermodilution cardiac output
determination in pulmonary arterial hypertension.
Congest Heart Fail. 2004; 10(2 Suppl 2):710.
10. Albert NM, Hail MD, Li J, Young JB. Equivalence of
bioimpedance and thermodilution in measuring car-
diac output in hospitalized patients with advanced,
decompensated chronic heart failure. Am J Crit
Care. 2004; 3:46979.
11. Silver M, Cianci P, Brennan S, Longeran-Thomas H,
Ahmad F. Evaluation of impedance cardiography as
an alternative to pulmonary artery catheterization
in critically ill patients. Congest Heart Fail. 2004; 10(2
Suppl 2):1721.
12. Springfield C, Sebat F, Johnson D, Lengle S, Sebat C.
Utility of impedance cardiography to determine
cardiac vs. noncardiac cause of dyspnea in the
emergency department. Congest Heart Fail. 2004;
10(2 Suppl 2):146.
13. Han J, Lindsell C, Tsurov B, Storrow A. The clinical
utility of impedance cardiography in diagnosing con-
gestive heart failure in dyspneic emergency depart-
ment patients [abstract]. Acad Emerg Med. 2002; 9:
43940.
14. Barcarse E, Kazanegra R, Chen A, Chiu A, Clopton
P, Maisel A. Combination of B-type natriuretic pep-
97 of 158.
371
tide levels and non-invasive hemodynamic parame-
ters in diagnosing congestive heart failure in the
emergency department. Congest Heart Fail. 2004;
10:1716.
15. Peacock WF, Albert NM, Kies P, White RD, Emerman
CL. Bioimpedance monitoring: better than chest
x-ray for predicting abnormal pulmonary fluid? Con-
gest Heart Fail. 2000; 6(2):325.
16. Newman RB, Pierre H, Scardo J. Thoracic fluid con-
ductivity in peripartum women with pulmonary
edema. Obstet Gynecol. 1999; 94:4851.
17. Milzman D, Morrisey J, Pugh C, Napoli A, Gerace T,
Fernandez E. Occult perfusion deficits in heart failure
patients: identification through noninvasive central
hemodynamic monitoring [abstract]. Crit Care Med.
1999; 27:A88.
18. Pearl WS. A hierarchical outcomes approach to test
assessment. Ann Emerg Med. 1999; 33:7784.
19. Summers R, Anders R, Woodward L, Jenkins A, Galli
R. Effect of routine pulse oximetry measurements on
ED triage classification. J Emerg Med. 1999; 16:57.
20. Mower WR, Sachs C, Nicklin EL, Safa P, Baraff LJ.
Effect of routine emergency department triage pulse
oximetry screening on medical management. Chest.
1995; 108:1297302.
21. Kosowsky JM, Weiner C, Morrissey JH. Impact of
B-type natriuretic peptide testing on medical deci-
sion-making for older patients with dyspnea. Ann
Emerg Med. 2003; 42:S11.
22. Summers R, Schoemaker W, Peacock WF, Ander D,
Coleman T. Bench to bedside series: impedance car-
diography (ICG). Acad Emerg Med. 2003; 10:66980.
23. Luepker R, Michael JR, Warbasse JR. Transthoracic
electrical impedance: quantitative evaluation of a
noninvasive measure of thoracic fluid volume. Am
Heart J. 1973; 85:8393.
24. Van de Water JM, Mount BE, Chandra KM, Mitchell
BP, Woodruff TA, Dalton ML. TFC (thoracic fluid
content): a new parameter for assessment of changes
in chest fluid volume. Am Surg. 2005; 71:816.
25. Taler SJ, Textor SC, Augustine JE. Resistant hyper-
tension: comparing hemodynamic management to
specialist care. Hypertension. 2002; 39:9828.
26. Maisel AS, Krishnaswamy P, Nowak RM, et al. Rapid
measurement of B-type natriuretic peptide in the
emergency diagnosis of heart failure. N Engl J Med.
2002; 347:1617.
27. Peacock WF, Roe MT, Chen AY, et al. Vein-to-brain
time: an emergency department quality of care
marker for non-ST-segment elevation acute coro-
nary syndrome [abstract]. Acad Emerg Med. 2004;
11:569.
28. Urrunaga J, Rivers E, Mullen M, et al. Hemodynamic
assessment of the critically ill: the clinician versus
esophageal Doppler monitoring (EDM) [abstract].
Acad Emerg Med. 2000; 7:587b.
 AJH 2005; 18:26S 43S
Featured Review Article
Hypertension as a Hemodynamic Disease: The
Role of Impedance Cardiography in Diagnostic,
Prognostic, and Therapeutic Decision Making
Hector O. Ventura, Sandra J. Taler, and John E. Strobeck
Hypertension is the most common cardiovascular disease,
affecting approximately 60 million Americans. Despite the
importance of this condition, only the minority of patients
are appropriately identifie and treated to reach recom-
mended blood pressure (BP) goals. Although historically
define as an elevation of BP alone, hypertension is char-
acterized by abnormalities of cardiac output, systemic
vascular resistance, and arterial compliance. These hemo-
dynamic aspects of hypertension have implications for
diagnosis, risk stratification and treatment. Impedance
cardiography (ICG) has emerged as a unique and highly
accurate noninvasive tool that is used to assess hemody-
hen functioning properly, the cardiovascular
W system provides normal blood flo to the var-
ious tissues of the body under normal arterial
blood pressure (BP). Historically, BP is the most com-
monly measured parameter of cardiovascular function. Hy-
pertensiontypically define by BP levels of 140/90 mm Hg
and higherleads to increased rates of coronary artery
disease, heart failure, renal disease, and stroke. Therefore,
BP control is of paramount importance for both individual
and public health considerations.
Blood pressure by itself is an incomplete indicator of
the status of the cardiovascular system. Mean arterial
pressure (MAP) is the product of two hemodynamic com-
ponents: cardiac output (CO), the flo of blood pumped
by the heart each minute; and systemic vascular resistance
(SVR), the force the left ventricle must overcome to expel
blood into the systemic vasculature, also called total pe-
ripheral resistance. Hypertension results from elevations
of CO, SVR, or both. Because hemodynamics literally
refers to blood flowrelate parameters of the arterial
system, CO and SVR are fundamental to obtaining greater
insight into the pathophysiology of hypertension, and they
can help to guide diagnostic, prognostic, and therapeutic
management decisions. Thus, the hemodynamic model of
Received October 29, 2004. Accepted November 4, 2004.
From the Ochsner Clinic Foundation (HOV), New Orleans,Loui-
siana; Mayo Clinic College of Medicine (SJT), Rochester, Minnesota;
and Cabrini Medical Center (JES), New York, New York.
0895-7061/05/$30.00
98 doi:10.1016/j.amjhyper.2004.11.002
of 158.
namic parameters. Measurement of the various hemody-
namic components using ICG in those with hypertension
allows more complete characterization of the condition, a
greater ability to identify those at highest risk, and allows
more effectively targeted drug management. This article
reviews the importance of hemodynamic factors in hyper-
tension and the evolving role of ICG technology in the
assessment and management of this important cardiovas-
cular condition. Am J Hypertens 2005;18:26S 43S
2005 American Journal of Hypertension, Ltd.
Key Words: Hypertension, hemodynamics, impedance
cardiography.
hypertension has intrigued scientists and clinicians since
the early part of the last century1 and has been reviewed
extensively by leaders in the field 27
The hemodynamic components of BP, CO, and SVR,
and other related parameters such as arterial compliance
provide insight into mechanisms of hypertension8 and
have implications for management of patients with this
condition. Historically, most hemodynamic information
used in research has been obtained using invasive tech-
niques, including arterial cannulation and placement of a
pulmonary artery catheter for the measurement of cardiac
output and determination of SVR. However, invasive pro-
cedures are not feasible in the routine care of patients with
hypertension. Echocardiography provided early noninva-
sive measurement of cardiac output, but it is costly and
highly operator dependent, and it is therefore impractical
for frequent serial measurements in the clinical setting.
Recent advancements in noninvasive hemodynamic mon-
itoring with impedance cardiography (ICG) have been
achieved, elevating its role as a unique and valuable non-
invasive tool for the assessment of hemodynamic status in
patients with hypertension.
This review describes the historical use of hemodynam-
ics in hypertension and reveals the growing body of evi-
Address correspondence and reprint requests to Dr. Hector O.
Ventura, Ochsner Clinic Foundation, 1514 Jefferson Highway, Room
3E419, New Orleans, LA 70121; e-mail: Hventura@ochsner.org
2005 by the American Journal of Hypertension, Ltd.
Published by Elsevier Inc.
 AJHFebruary 2005VOL. 18, NO. 2, Part 2
Table 1. JNC 7 Classification of blood pressure for
adults 18 years of age
Systolic Diastolic
BP BP
BP classification (mm Hg) (mm Hg)
Normal 120 and 80
Prehypertension 120139 or 8089
Stage 1 hypertension 140159 or 9099
Stage 2 hypertension 160 or 100
JNC Joint National Committee on the Prevention, Detection, Eval-
uation, and Treatment of High Blood Pressure.
dence using noninvasive monitoring of hemodynamics
with ICG. The specifi role of hemodynamics in diagnos-
tic, prognostic, and therapeutic decision making in the
patient with hypertension is reviewed in detail.
Hypertension: Definition and
Clinical Presentation
Hypertension is most commonly define as a systolic BP
(SBP) of 140 mm Hg or a diastolic BP (DBP) of 90
mm Hg. In patients at high risk for complications from
elevated BP levels, such as those with diabetes or chronic
renal disease, lower levels of BP (eg, 130/80 mm Hg)
are recommended. Between BP levels of 115/75 mm Hg
and 185/115 mm Hg, each 20 mm Hg increase in SBP or
10 mm Hg increase in DBP doubles the risk of a cardio-
vascular event.9 In recognition of this increase in risk from
levels as low as 115/75 mm Hg, the Seventh Report of the
Joint National Committee on Prevention, Detection, Eval-
uation, and Treatment of High Blood Pressure (JNC 7)9
recently identifie prehypertension (define as a BP of
120 to 139 / 80 to 89 mm Hg) as a significan potential
health problem associated with increased risk for pro-
gression to hypertension. The JNC 7 recommended life-
style modificatio for the management of prehypertension.
Table 1 lists the stages of hypertension as define in the
JNC 7 classification
Secondary hypertension results from an identifiabl
cause such as renal, adrenal, or vascular pathology. In
contrast, 90% of patients have no specifi identifiabl
cause of their BP elevation10 and are thus diagnosed with
essential hypertension. Most patients with hypertension
are asymptomatic and do not show evidence of acute
pathologic changes; their clinical presentation has been
termed benign, although their long-term risk of cardio-
vascular complications is significantl greater than in nor-
motensive persons without so-called benign hypertension.
Severe elevation of BP, associated with papilledema on
fundoscopic examination, is termed malignant hyperten-
sion. Similar levels of BP elevation with congestive heart
failure, anginal symptoms, or other evidence for acceler-
ated end-organ injury (but without papilledema) are
termed hypertensive urgencies or hypertensive emer-
99 of 158.
HYPERTENSION AS A HEMODYNAMIC DISEASE 27S
gencies. The etiologies of hypertension (essential versus
secondary) and the various clinical presentations (benign,
malignant, unspecified and with or without associated
co-morbidities) are reflecte in the World Health Organi-
zations International Classificatio of Diseases, Ninth
Revision (ICD-9), coding for hypertension (Table 2).
Hypertension: Magnitude of the
Problem
Hypertension affects up to 60 million Americans and as
many as 1 billion persons worldwide; and it is the most
common reason that patients in the United States visit their
physicians.9,11 The incidence of hypertension increases
significantl with advancing age (Fig. 1), such that a
normotensive adult in the United States 55 years of age
still has a 90% lifetime risk of developing hypertension.9
In fact, the most common group with hypertension is
comprised of elderly patients with systolic hypertension.12
Although controlling BP levels reduces the incidence of
stroke and other cardiovascular complications, BP control
in the US is well below stated goals. For an individual
patient, this may be due to the lack of recognition of the
condition, failure to institute effective treatment, or the result
of a suboptimal long-term medical regimen (Table 3). There
remains a substantial need for improvement in the effective-
ness of hypertension treatment. As reported in JNC 7, only
34% of adults aged 18 to 74 years with hypertension have
achieved BP control, despite a published goal of 50%. In
the elderly population, BP control is even less successful:
fewer than 20% of treated patients 70 years or more of age
attain BP levels of 140/90 mm Hg.9
Hypertension treatment commonly requires multiple
medications. Refractory hypertension has been define
as hypertension that is not controlled on two or more
antihypertensive medications.13 Resistant hypertension
has been define by some as BP readings of 140/90 mm
Hg despite an optimal two-drug regimen that has had
adequate time to work (at least 1 month since last drug or
dosage adjustment).14 As define by Gifford, resistant
hypertension is the failure to reach goal BP in patients who
are adhering to full doses of an appropriate three-drug
regimen that includes a diuretic.15 The JNC 7 recommends
a goal BP of 140/90 mm Hg for the general population,
with the tighter goal of 130/80 mm Hg for persons with
chronic renal disease or diabetes mellitus.
Hypertension substantially increases the incidence of
cardiovascular events, especially the risk of stroke. Wilk-
ing et al,16 in data from the Framingham study, reported on
the prognostic significanc of systolic hypertension. They
found that for men and women, the relative risk of car-
diovascular disease event adjusted for age was approxi-
mately 2.5 times greater for persons with isolated systolic
hypertension compared with those with BP levels
140/95 mm Hg. Lower BP levels are thus associated
with improved prognosis and decreased incidence of mor-
bidity and mortality. From pooled data of more than 60
 28S HYPERTENSION AS A HEMODYNAMIC DISEASE AJHFebruary 2005VOL. 18, NO. 2, Part 2

Table 2. International Classification of Diseases codes for hypertension

Essential hypertension
401.0 Essential hypertension; Malignant
401.1 Benign essential hypertension
401.9 Unspecified essential hypertension
Hypertensive heart disease
402.00 Hypertensive heart disease; malignant; without congestive heart failure
402.01 Hypertensive heart disease; malignant; with congestive heart failure
402.10 Hypertensive heart disease; benign; without congestive heart failure
402.11 Hypertensive heart disease; benign; with congestive heart failure
402.90 Hypertensive heart disease; unspecified; without congestive heart failure
402.91 Hypertensive heart disease; unspecified; with congestive heart failure
Hypertensive renal disease
403.00 Hypertensive renal disease; malignant; without mention of renal failure
403.01 Hypertensive renal disease; malignant; with renal failure
403.10 Hypertensive renal disease; benign; without mention of renal failure
403.11 Hypertensive renal disease; benign; with renal failure
403.90 Hypertensive renal disease; unspecified; without mention of renal failure
403.91 Hypertensive renal disease; unspecified; with renal failure
Hypertensive heart and
renal disease
404.00 Hypertensive heart and renal disease; malignant; w/o mention of congestive
heart failure or renal failure
404.01 Hypertensive heart and renal disease; malignant; with congestive heart failure
404.02 Hypertensive heart and renal disease; malignant; with renal failure
404.03 Hypertensive heart and renal disease; malignant; with congestive heart failure
and renal failure
404.10 Hypertensive heart and renal disease; benign; w/o mention of congestive
heart failure and renal failure
404.11 Hypertensive heart and renal disease; benign; with congestive heart failure
404.12 Hypertensive heart and renal disease; benign; with renal failure
404.13 Hypertensive heart and renal disease; benign; with congestive heart failure
and renal failure
404.90 Hypertensive heart and renal disease; unspecified; w/o mention of congestive
heart failure or renal failure
404.91 Hypertensive heart and renal disease; unspecified; with congestive heart
failure
404.92 Hypertensive heart and renal disease; unspecified; with renal failure
404.93 Hypertensive heart and renal disease; unspecified; with congestive heart
failure and renal failure
Secondary hypertension
405.01 Secondary hypertension; malignant; renovascular
405.09 Secondary hypertension; malignant; other
405.11 Secondary hypertension; benign; renovascular
405.09 Secondary hypertension; benign; other
405.91 Secondary hypertension; unspecified; renovascular
405.99 Secondary hypertension; unspecified; other

prospective studies and 1 million patients, Lewington et al17
report that 10 mm Hg reductions in systolic BP would be
expected to reduce stroke mortality by as much as 40%.
FIG. 1 Prevalence of hypertension by age.
Importantly, the authors note that even a 2mm Hg reduc-
tion in systolic BP is associated with a 10% lower death
rate from stroke. These reductions in risk apply all the way
to BP levels of 115/75 mm Hg. Thus, the failure to lower
BP even modestly in patients with hypertension is respon-
sible for a significan number of preventable cardiovascu-
lar events each year.
The financia implications of hypertension and hyper-
tension management are substantial.18 The direct costs of
treating hypertension exceeded $37 billion in the year
2003, and additional costs due to loss of productivity were
more than $13 billion (Table 4). Of the ten medical con-
ditions evaluated for their effects on absenteeism from
work and loss of productivity, hypertension was the most

100 of 158.
 AJHFebruary 2005VOL. 18, NO. 2, Part 2 HYPERTENSION AS A HEMODYNAMIC DISEASE 29S

Table 3. Trends in awareness, treatment, and control of high blood pressure in adults with hypertension 18
to 74 years of age

II III-1 III-2
NHANES (19761980) (19881991) (19911994) 19992000
Awareness 51 73 68 70
Treatment 31 55 54 59
Control 10 29 27 34

NHANES National Health and Nutrition Examination Survey.

All numbers expressed as percentages.

expensive costing businesses an average of $392 per
eligible employee per year.19 Improving the efficienc and
effectiveness of drug management in the hypertensive
population would likely reduce these costs in addition to
decreasing morbidity and mortality associated with the
condition.
Hemodynamic Measurements
Using ICG
The historical use of BP without CO or SVR is, in part,
because it has been impractical to estimate or measure
these parameters in most clinical settings. The assessment
of the hemodynamic components of hypertension from
clinical evaluation alone is unreliable. Even in patients
with acute conditions such as those requiring the emer-
gency department or patients with decompensated conges-
tive heart failure (in whom hemodynamic derangements
are greater than those in patients with essential hyperten-
sion), clinicians are generally unable to estimate CO or
SVR with accuracy.20,21
Echocardiography has been used to measure cardiac
output and in some studies has demonstrated acceptable
correlation with invasive techniques.22 However, in a
comparison with ICG, echocardiography is considerably
more time consuming and technically demanding.23 In the
offic setting, CO is not generally reported by most phy-
sicians interpreting echocardiograms in clinical practice.
Thus, until recently, CO and SVR were commonly mea-
sured only in the intensive care setting or catheterization
laboratory setting using invasive means such as a pulmo-
nary artery catheter.
In recent years, ICG has emerged as an accurate, safe,
and inexpensive tool with which to measure hemodynamic
parameters by noninvasive means. The procedure is most
commonly performed in the physician offic setting by
medical assistants or nurses, requiring about 5 min to
complete the test. Using four sets of paired sensors on the
neck and chest, ICG measures the instantaneous change of
an electrical signal across the thoracic cavity (Fig. 2). As
the changes of thoracic impedance during the cardiac
cycle are most dependent on the changes in the size and
the blood volume of the thoracic aorta, ICG is able to
calculate the amount of blood ejected from the left ven-
tricle (that is, the stroke volume [SV]). The product of
heart rate (HR) and SV yields CO. In addition, ICG-
derived parameters related to the changes of thoracic im-
pedance are indicative of aortic blood velocity and
acceleration, and they correlate with measures of inotropic
state and cardiac performance. As flui is the best con-
ductor of the electrical signal through the chest (when

Table 4. Direct and indirect costs attributable to

hypertension

Cost (in $
Type of cost billion)
Direct costs
Inpatient 8.7
Professional services 9.2
Drugs and medical durables 17.8
Home health care 1.5
Total direct costs 37.2
Indirect costs of lost productivity
Related to morbidity 7.0
Related to mortality 6.1
Total indirect costs 13.1
Total costs 50.3
FIG. 2 Measurement of impedance signal using four sets of paired
Data are from Heart Disease and Stroke Statistics2003 Update. sensors. Sensors transmit and record electrical signal from which
American Heart Association; 2002. multiple hemodynamic parameters are derived.

101 of 158.
 30S HYPERTENSION AS A HEMODYNAMIC DISEASE AJHFebruary 2005VOL. 18, NO. 2, Part 2

tients disease progression, response to therapy, and

FIG. 3 Representative simultaneous tracings of electrocardiogram
(ECG), thoracic impedance (Z), and first-time derivative of imped-
ance (dZ/dt), and dZ/dt waveforms. Fiducial points are identified on
the dZ/dt tracing from which various hemodynamic and parameters
and timing intervals are derived.
compared with bone, air and fat, in particular), the total
thoracic impedance is inversely related to an index of flui
termed the thoracic flui content (TFC). Finally, using a
simultaneous electrocardiographic recording, ICG mea-
sures the pre-ejection period and LV ejection timetim-
ing intervals that relate to cardiac performance.
Representative simultaneous tracings of an electrocar-
diogram, change in thoracic impedance (Z), and firs
derivative of impedance (dZ/dt) are shown in Fig. 3. From
the measured variables and from HR and mean BP deter-
mined by oscillometry, SVR and other parameters are
calculated and displayed. An ICG test report is shown in
Fig. 4. A more detailed description of selected parameters
is provided in Table 5.
need for further intervention. Thermodilution, using a
pulmonary artery catheter, has traditionally been the
standard to which ICG has been compared. Van De
Water et al24 assessed the relative reproducibility of
ICG and thermodilution cardiac outputs in hospitalized
patients in whom a pulmonary artery catheter was
placed for hemodynamic monitoring after bypass sur-
gery. Serial ICG measurements in a given patient
showed better reproducibility than serial CO measure-
ments using thermodilution technique (Table 7). The
investigators concluded that current ICG technology has
advanced such that ICG provides a level of agreement
that is equivalent to thermodilution. Their finding
support the clinical utility of ICG for serial measure-
ments in patients with cardiovascular disease.
In a stable group of patients in the outpatient setting,
Verhoeve et al30 demonstrated a high reproducibility of
measurements performed on the same day and appropriate
sensitivity for the physiologic variations expected from
day to day. The variation in the average of readings for
CO, SVR, and thoracic flui content (TFC) ranged be-
tween 3% and 7% for serial measurements 1 week apart.
Figure 5 illustrates the high degree of correlation between
stroke index measured on day 1 and then 1 week later in
96 patients who were clinically stable.
The ICG technique is widely applicable, and reliable
information can be obtained in minutes at virtually no

Validation of Current ICG

Technology
Placement of a pulmonary artery catheter is a costly pro-
cedure requiring special training and expertise; and it is
associated with risks of bleeding, infection, and damage to
vascular and other structures. Because of the risks inherent
in invasive methods for measuring hemodynamics, studies
comparing ICG to invasive techniques of hemodynamic
measurement are only available from populations with
significan underlying cardiovascular conditions or situa-
tions that justify the risks associated with pulmonary ar-
tery catheter placement. In such clinical settings and
patient populations, multiple studies have shown that cur-
rent ICG technology, using advanced data processing and
modeling techniques, yields data that are significantl
more accurate than those obtained with prior generations
of ICG devices.24 Five additional validation studies of
ICG presented since 1998, 2529 using refine ICG tech-
nology (BioZ ICG Monitor, CardioDynamics, San Diego,
CA), demonstrate the high correlation and accuracy avail-
able with ICG when compared with invasive techniques
(Table 6).
The ability to measure changes in hemodynamic
parameters reliably in a given patient is critically im-
portant from a clinical perspective, as the changes in
serial measurements are the basis for evaluating pa- FIG. 4 Hemodynamic status report.

102 of 158.
 AJHFebruary 2005VOL. 18, NO. 2, Part 2 HYPERTENSION AS A HEMODYNAMIC DISEASE 31S

Table 5. Impedance cardiography variables

Impedance Cardiography
Variable Units Measurement/Calculation
Blood flow
Stroke volume mL VI LVET VEPT (Z MARC algorithm)
Stroke index mL/m2 SV/BSA
Cardiac output L/min SV HR
Cardiac index L/min/m2 CO/BSA
Resistance
Systemic vascular resistance dyne sec cm5 [(MAP - CVP)/CO] 80
Systemic vascular resistance index dyne sec cm5 m2 [(MAP - CVP)/CI] 80
Contractility
1000 first-time derivative of Zmax/
Velocity index /1000/sec baseline impedance
100 second-time derivative of
Acceleration index /100/sec2 Zmax/baseline impedance
Pre ejection period msec ECG Q wave to aortic valve opening
Left ventricular ejection time msec Aortic valve opening to closing
Systolic time ratio - PEP/LVET
Cardiac work
Left cardiac work index kg m/m2 (MAP - PCWP) CI 0.0144
Fluid status
Thoracic fluid content /kOhm 1000 1/baseline impedance

BSA body surface area; cm centimeter; CVP central venous pressure (estimated value of 6 mmHg); ECG electrocardiography; HR
heart rate; ICG impedance cardiography; MAP mean arterial pressure; PCWP pulmonary capilary wedge pressure (estimated value
of 10 mmHg); R to R interval 60/ heart rate: VEPT volume of electrically participating tissue; Z MARC impedance modulating aortic
compliance.

risk to the patient. However, ICG has some limitations
related to the technology and patient factors. Although
ICG equations have demonstrated accuracy over a wide
range of conditions and patient populations, ICG has
not been evaluated extensively in patients 66 pounds
or 342 pounds. Severe aortic insufficienc may affect
ICG reliability, but it has not been fully studied and
validated in such patients. In addition, a few models of
permanent pacemakers use impedance technology to
measure minute ventilation. If the minute ventilation
function is activated, the paced rate may increase be-
cause of ICG signals31; therefore, patients with such
pacemakers must have the minute ventilation sensor
function inactivated before ICG testing. In patients with
atrial fibrillatio or frequent premature ventricular con-
tractions, marked irregularity in heart rhythm can affect
data collection and analysis of wave forms.
Hemodynamic Parameters in
Hypertension
Hypertension is the result of complex cardiac, renal, neu-
rohormonal, and vascular mechanisms that are modulated
by both genetic and environmental factors.10,32 The inter-
actions of these many factors result in endothelial dysfunc-
tion and hemodynamic derangements of arterial
compliance, CO, and SVR. As noted earlier, MAP is the
product of CO and SVR, and elevations of BP can result

Table 6. Validation studies of impedance cardiography (ICG)

Population AuthorRef Parameter Comparison r Value Bias Precision

25
HF in ICU Albert et al CO ICG-TD 0.89 0.08 1.38
HF in catheterization
laboratory Drazner et al26 CO ICG-Fick 0.73 0.74 1.1
TD-Fick 0.81 0.75 0.95
ICG-TD 0.76 0.03 1.1
Mechanical ventilation Ziegler et al27 CO ICG-TD 0.89 0.45 1.2
Post-CABG Sageman et al28 CI ICG-TD 0.92 0.07 0.40
Post-CABG Van De Water et al24 CO ICG-TD 0.81 0.17 1.09
Pulmonary hypertension Yung et al29 CO ICG-Fick 0.84 0.24 0.87
TD-Fick 0.89 0.19 0.76
ICG-TD 0.80 0.43 1.01

CABG coronary artery bypass surgery; CO cardiac output; HF heart failure; ICU intensive care unit. TD thermodilution;

103 of 158.
 32S HYPERTENSION AS A HEMODYNAMIC DISEASE
Table 7. Reproducibility of serial measurements:
impedance cardiography (ICG) versus thermodilu-
tion (TD)
Correlation
Comparison (r value) SD (L/min)
TD 2 v TD 1 0.83 1.02
TD 3 v TD 2 0.84 1.01
TD 3 v TD 1 0.83 1.07
ICG 2 v ICG 1 0.97 0.44
ICG 3 v ICG 2 0.98 0.39
ICG 3 v ICG 1 0.97 0.43
Adapted from Van De Water et al.24
from elevation of either or both of these hemodynamic
parameters. Pulse pressure (PP), that is, the difference
between systolic BP and diastolic BP, is determined by SV
and total arterial compliance (TAC). Arterial compliance
is a complex parameter that is most closely approximated
using a complicated and sophisticated model (the three-
element Windkessel model) that incorporates the ratio of
the decay time constant to peripheral resistance.33,34 True
arterial compliance is thus tedious and time-consuming to
measure and is not clinically useful. However, studies
have shown that arterial compliance can be reliably esti-
mated as the ratio of SV to PP.34,35 Relationships among
the hemodynamic parameters including PP, SV, MAP,
CO, and SVR are shown in Fig. 6. The hemodynamics of
hypertension have been studied for decades, and previ-
ously various aspects have been extensively re-
viewed.3,36 39
Hemodynamics of Hypertension:
Diagnostic Considerations
Numerous studies using either invasive or noninvasive
techniques have demonstrated that there are distinct he-
modynamic subsets among various groups of patients with
hypertension. Hemodynamic measurements allow the dif-
ferentiation of patients with primarily elevated CO from
those in which elevated SVR (signifying a vasoconstricted
state) is the primary mechanism of their hypertension.
FIG. 5 Reproducibility of stroke index (SI) measurements made 1
week apart. Adapted from Verhoeve et al.30
104 of 158.
AJHFebruary 2005VOL. 18, NO. 2, Part 2
FIG. 6 Components of mean arterial pressure (MAP) and pulse
pressure (PP). CO cardiac output; SV stroke volume; SVR
systemic vascular resistance.
Moreover, hemodynamic measurements can elucidate the
relative contributions of SV and arterial compliance to
elevations in PP.
Invasive Hemodynamic and
Echocardiographic Studies
In the Tecumseh, Michigan study, 40 patients were studied
using echocardiographic techniques and investigators
found that 37% of patients with hypertension were hy-
perkinetic, as define by increased cardiac index, HR,
forearm blood flow and plasma norepinephrine levels.
The distribution of cardiac index in this population study
is shown in Fig. 7. The wide distribution of cardiac index
values in these patients provides corroboration that hyper-
tension represents a heterogeneous mix of various hemo-
dynamic subsets.
In general, aging is associated with decreases in CO
and increases in SVR, as shown in Fig. 8. In young adults,
hypertension may be more commonly associated with
increased CO, whereas in older adults it is more com-
monly associated with elevated SVR. Lund-Johansen41
found a change in hemodynamic pattern in patients with
borderline hypertension at 10 and 17 years of follow-up.
There was a significan and progressive decrease in CO
over time, associated with an increase in SVR.
Age-related changes in hemodynamic status, as evi-
denced by changes in arterial compliance, occur in patients
with hypertension even in the absence of changes in CO or
SVR. Slotwiner et al42 used echocardiographic estimates
of cardiac output to study hemodynamic parameters in 272
patients who were 25 to 80 years of age and had mild
hypertension. These investigators found that in their study
group, CO and SVR levels did not vary significantl with
age. However, vascular stiffness, as reflecte by the ratio
of PP to SV (the reciprocal of TAC) increased with age,
which is possibly the mechanism for increased rates of
cardiovascular events in elderly individuals. Others have
noted that arterial stiffness exerts deleterious effects due to
increases in central aortic pressureanother hemody-
namic mechanism that is key in the pathophysiology of
hypertensive cardiovascular disease.43
 AJHFebruary 2005VOL. 18, NO. 2, Part 2
FIG. 7 Distribution of cardiac index values in Tecumseh, Michigan
study. Cardiac index values for the hypertensive population show a
bimodal distribution. From Julius et al.40
Other factors besides age appear to predict general
trends in the hemodynamic parameters in hypertensive
populations. Hemodynamic parameters differ between hy-
pertensive men and women. Messerli et al44 measured PP,
CO, and SVR using invasive techniques in 200 subjects.
Despite equal levels of arterial BP, women had signifi
cantly higher CO, PP, and lower SVR compared with men.
Isometric exercise was associated with an increase in
arterial pressure that was nearly 50% greater in men than
in women. The hemodynamic differences between men
and women were confine to premenopausal women, sug-
gesting that estrogens play a significan role in the cardio-
vascular and hemodynamic responses in patients with
hypertension. The mechanisms of hypertension seen with
acute stressors, such as public speaking or mental arith-
metic, also vary based on gender. Studies have shown that
men and postmenopausal women have a more significan
increase in SVR in response to acute stressors, whereas
premenopausal women exhibit a hypertensive response
that is due primarily to increases in CO. Some studies have
suggested that hypertension early in the course of diabetes
and with obesity are associated with increased CO and
relatively normal SVR.45 Others have shown that the
earliest hemodynamic abnormalities may be changes in
arterial compliance.46 To a significan degree, hyperten-
sion in patients on dialysis results from volume expansion
and can be associated with signs of sympathetic stimula-
tion such as increased HR and SV.
Impedance Cardiographic Studies
Impedance cardiography has been used to evaluate the
hemodynamic parameters in normotensive individuals at
different ages and in various hypertensive populations. In
a study of 640 normal subjects evaluated as renal trans-
plant donors, Taler et al47 found that increasing age was
associated with increasing BP, increasing SVR, and de-
creasing CO due to decreased SV. Hemodynamic changes
with age were similar in men and women, although BP and
105 of 158.
HYPERTENSION AS A HEMODYNAMIC DISEASE 33S
CO were lower and HR and SVR were higher in women.
Age-related changes included an increase in total thoracic
impedance, equivalent to a decrease in its reciprocal TFC
and consistent with decreasing cardiopulmonary volume
or muscle mass or both.
In a study comparing hemodynamic variables between
pre-menopausal and post-menopausal women, Hinderliter
et al48 showed that post-menopausal women had lower CO
and higher SVR for any given BP level compared with
pre-menopausal women. Importantly, these significan
changes in CO and SVR occurred without significan
changes in BP levels, suggesting that the hemodynamic
parameters underlying BP provide more information than
does MAP alone. This is also seen in data from a study by
Galarza et al,49 in which, despite relatively stable DBP
levels in patients from the third to seventh decades of life,
the investigators found significan increases in SVRI of
nearly 50% and decreases in cardiac index of 27%. Alfi
et al50 used impedance techniques to show that elevations
in the difference between SBP and DBP (pulse pressure
[PP]) occurred due to different hemodynamic mechanisms
in men 30 years of age compared with those middle aged
and older. In younger men, increased PP was associated
with increases in stroke index, reflectin preserved hemo-
dynamic load with normal arterial compliance. In contrast,
after age 50 years, men showed increases in PP associated
with decreases in stroke index, reflectin age-related de-
creases in arterial compliance. Thus, BP readings alone
did not reflec the underlying hemodynamic differences in
groups with presumably different cardiovascular risk de-
spite similar levels of MAP and PP. Gender differences are
seen in impedance studies of the hypertensive response to
caffeine: men who show hypertensive responses to caf-
feine increase their SVR, whereas women primarily in-
crease SV and CO.51 Yu et al52 studied hemodynamic
parameters in patients with different mood states. Findings
of correlation of CO and SVR but not SBP, DBP, or
FIG. 8 Age-related changes in cardiac output and peripheral re-
sistance. With increasing age, peripheral resistance rises in the
hypertensive population and at higher levels it is associated with
decreasing cardiac output and ultimately congestive heart fail-
ure.
 34S HYPERTENSION AS A HEMODYNAMIC DISEASE
Table 8. Improved predictive power of pulse pres-
sure (PP) to stroke volume index ratio (SVIR) com-
pared with pulse pressure alone*
Hazard rate, Hazard rate,
CV event mortality
PP 1.46 (0.912.35) 1.65 (0.833.30)
PP to SVIR 1.79 (1.152.80) 2.05 (1.153.65)
CV cardiovascular.
Values and 95% confidence intervals for 1-SD increase in the
variable of interest.
Adapted from Fagard et al.56
* Adjusted for age, gender, mean arterial pressure, and heart
rate; P NS; P .01.
MAPwith affective state is further evidence of the het-
erogeneity of hemodynamic subsets within various groups
clinically identified
Hinderliter53 reported that African American men and
women had increased SVR, decreased CO, and associated
LV remodeling compared with Caucasian men and women
despite similar BP readings. In normotensive African
Americans, Calhoun et al54 postulated that vasoconstrictor
responses seen with mental stress and cold presser testing
may contribute to elevated SVR and the development of
hypertension.
These and other studies using ICG technology show
that within any given population SVR, CO, and TAC show
significan variation. The heterogeneity of hemodynamic
finding within various cohorts is evidence that the spe-
cifi hemodynamic status of an individual patient cannot
reliably be predicted on the basis of age, gender, or ethnic
background. Moreover, that hemodynamic values cannot
be identifie by BP levels or clinical assessment alone
supports the need for measurement of hemodynamic pa-
rameters in individual patients with hypertension.
Hemodynamics of Hypertension:
Prognostic Considerations
The underlying hemodynamic abnormalities in hyperten-
sion result in structural and functional changes in the
cardiovascular system that adversely affect prognosis, ie,
that increase risk of morbidity or mortality. Increases in
hemodynamic measures such as SVR and reductions in
arterial compliance provide prognostic information in ad-
dition to that obtained by BP measurements alone.
Invasive Hemodynamic and
Echocardiographic Studies
Elevated arterial BP is the result of increased arterial
stiffness and increased SVR. This results in increased LV
wall stress, the best measure of LV afterload. Using cath-
eterization techniques, Fagard et al55 demonstrated that
SBP and SVR at rest and during exercise correlated with
the risk of cardiovascular events and total mortality at an
106 of 158.
AJHFebruary 2005VOL. 18, NO. 2, Part 2
average of 16.2 years of follow-up. In this study, exercise
SVR but not exercise BPadded prognostic value to
parameters measured at rest, suggesting that hemodynamic
variables other than BP might have greater prognostic
value.
A subsequent article56 reported the relationship between
another hemodynamic parameter, namely, the ratio of PP to
stroke index (PP-to-SVi ratio), and outcomes in patients
followed for an average of 16.5 years. In this study, the
PP-to-SVi ratio, that is, the reciprocal of TAC index, was
independently associated with cardiovascular events or death.
Each increase in PP-to-SVi ratio of 0.75 mm Hg/(mL/m2)
was associated with a 79% increase in the risk of a cardio-
vascular event (P .01) and greater than double the risk of
all-cause mortality (P .01). As shown in Table 8, the
increased hazard rates with PP-to-SVi ratio compared with
PP alone demonstrates the additional predictive value when
the hemodynamic parameter of flo (SVi) is added to the
pressure measurement alone.
De Simone et al57 studied the effects of TAC on car-
diovascular events over a 10-year period. They found that
risks of fatal and total cardiovascular events were inde-
pendently correlated with age, LV mass, and lower levels
of arterial compliance, define as decreasing values of the
measured ratio of echocardiographic SV to PP (SV/PP) to
that predicted from previously developed equations (%
SV/PP). Moreover, consistent with the results of Fagard et
al,56 the investigators found that systolic BP, mean BP, or
PP alone (without including the flow-relate hemody-
namic parameter of SV) were not independent predictors
of prognosis. After adjustment for age and LVH there
remained an independent effect of % SV/PP on cardiovas-
cular endpoints at 10-year follow-up (Fig. 9). These in-
vestigators found that hemodynamic parameters such as
arterial compliance and percent predicted arterial compli-
ance correlated better with changes in cardiac structure (ie,
hypertrophy and remodeling) than did BP levels alone.
FIG. 9 Event-free survival based on predicted ratio of stroke vol-
ume (SV) to pulse pressure (PP). Event-free survival is reduced in
hypertensive patients with reduced arterial compliance, as defined
as a low predicted SV/PP ratio. From De Simone et al.57
 AJHFebruary 2005VOL. 18, NO. 2, Part 2
Gender has long been known to affect prognosis in
patients with hypertension. In 1913, Janeway58 published
the observation that women with hypertension tended to
have a better prognosis than men. More recent studies
have suggested that this difference may be related to
different hemodynamic substrates in men compared to
women with elevated BP levels. Messerli et al44 suggested
that the disparate prognosis between men and women
might be explained on the basis of differing hemodynamic
mechanisms: For any level of arterial pressure, total
peripheral resistance (and therefore the risk of hyperten-
sive cardiovascular disease) was lower in women than in
men.
The mechanism of the adverse prognosis from hyper-
tension is in part related to structural changes in the heart
that result from elevated wall stress. Prolonged increases
in wall stress lead to left ventricular (LV) structural
changes with relative increases in wall thickness, overall
LV mass or both.59 In concentric remodeling, there is a
relative increase in LV thickness without increase in over-
all LV mass. This structural change appears to be related
to increased pressure load but with relative decrease in
volume as evidenced by low normal CO (termed volume
underload). Concentric left ventricular hypertrophy
(LVH) is characterized by an increase in wall thickness
with increase in LV mass or mass index and also results
from pressure overload caused by long-term hypertension.
Eccentric hypertrophy is define as increased LV mass
index with preserved relative wall thickness and is asso-
ciated with both pressure and volume overload. This pat-
tern, a common result of the afterload and volume
excesses in long-standing severe aortic insufficiency re-
sults in spherical remodeling of the LV. Interestingly, in
this study of hypertensive individuals,59 both eccentric
hypertrophy and concentric remodeling were more com-
mon than the classic pattern of hypertensive heart dis-
ease, namely, concentric LVH.
Nonetheless, LVH is a powerful predictor of cardiovas-
cular risk and is independently associated with mortality in
patients with coronary artery disease.60 63 For example,
Vakili et al63 reported on the pooled results of 20 pub-
lished studies of LVH as define by electrocardiographic
or echocardiographic criteria. They demonstrated a
weighted mean relative risk of cardiovascular morbidity
from LVH of 2.3, independent of all covariates analyzed.
As reported by Ichkhan et al,64 LVH is associated with
abnormalities of ventricular repolarization and at least a
twofold increase in the risk of serious ventricular arrhyth-
mias.
Hypertension results in abnormalities of endothelial
function, affecting hemodynamic factors such as arterial
compliance. Gomez-Cerezo et al65 demonstrated impaired
brachial artery flowmediate dilation, a common test of
endothelial function, in patients with sustained or labile
hypertension. They found that flow-mediate dilation was
abnormal to a similar degree in patients with sustained
essential hypertension or white-coat hypertension but
107 of 158.
HYPERTENSION AS A HEMODYNAMIC DISEASE 35S
was normal in individuals with normal BP levels. Others
have evaluated measures of arterial compliance (or, alter-
natively, arterial stiffness) using the measure of pulse
wave velocity.66,67
Additional structural changes occuring at the level of
the heart and blood vessel have prognostic significanc in
persons with hypertension, including vascular remodeling
with changes in lumen to wall thickness.68,69 Apoptosis, or
programmed cell death, contributes to the vascular
changes (ie, remodeling) in hypertension. Inflammatio
and fibrosi similarly contribute with the accumulation of
various components in the extracellular matrix such as
collagen and fibronectin Intengan and Schiffrin69 re-
viewed the factors that result in arterial remodeling and
altered hemodynamic parameters in patients with hyper-
tension.
Importantly, studies have shown that treatment with
antihypertensive agents may result in regression of the
structural abnormalities caused by long-standing hyperten-
sion and may result in improved prognosis.70 72 Mathew
et al,70 reporting on data from the Heart Outcomes Pre-
vention Evaluation (HOPE) study, demonstrated that treat-
ment with the angiotensin-converting enzyme (ACE)
inhibitor ramipril was associated with regression of LVH
by electrocardiographic criteria compared with placebo
control. That BP showed minimal difference between the
treatment and control groups is consistent with other stud-
ies demonstrating improvement in overall hemodynamics
(as shown by significan decreases in SVR and parallel
increases in CO) that are not evident from BP levels alone.
Ofil et al,73 in an echocardiographic substudy of the
Systolic Hypertension in the Elderly Program (SHEP),
demonstrated partial regression of LVH in patients treated
with a diuretic-based regimen for a minimum of 3 years. In
a meta-analysis of 1000 patients with serial echocardi-
ography during treatment of essential hypertension, Ver-
decchia et al74 demonstrated that patients whose LVH
regressed during treatment had significantl fewer cardio-
vascular events compared with those in whom LV mass
increased, consistent with the hypothesis that improve-
ments in hemodynamics correlate with improved progno-
sis in patients with appropriately treated hypertension.
Impedance Cardiographic Studies
The ICG technique has been used to explore age-related
changes in hemodynamic variables and their correlation
with cardiovascular risk and the adverse prognosis. These
studies support previous finding that future risk in pa-
tients with hypertension may not be reflecte in BP levels
alone. Alfi et al50 demonstrated that despite similar ele-
vations in PP, younger men had preserved stroke index
(and arterial compliance) compared with older men. They
concluded that preserved arterial compliance and cardiac
pump function may explain the lack of prognostic signif-
icance of elevated PP in younger men. These finding lend
further support to the value of the incremental information
 36S HYPERTENSION AS A HEMODYNAMIC DISEASE
FIG. 10 Hemodynamic components of hypertension. BP blood
pressure; CO cardiac output; HR heart rate; SV stroke vol-
ume; SVR systemic vascular resistance.
provided by the hemodynamic components of BP and PP
levels.
Hemodynamic differences have been demonstrated in
patients who have experienced complications of hyperten-
sion compared with those with hypertension alone. In a
study of hemodynamic status in hypertensive patients with
and without a history of stroke, Galarza et al75 found lower
cardiac index and higher SVR index in those with history
of stroke. These differences occurred in the absence of
differences in BP or antihypertensive treatment, providing
another example of the unreliability of BP to reflec the
severity of underlying hemodynamic abnormalities.
Hemodynamics of Hypertension:
Therapeutic Considerations
Hypertension management includes hygienic measures
such as sodium restriction and weight loss; and, in most
cases, it requires the use of one or more antihypertensive
agents. Antihypertensive medications exert their BP-low-
ering effects by reductions in SVR or CO. Hemodynamic
effects can be used to classify antihypertensive agents,
predict the response to antihypertensive therapy, and guide
both the initiation and titration of these agents.76 78
Just as interpretation and treatment of serum cholesterol
level improves when its components (HDL-cholesterol and
LDL-cholesterol) are measured, hypertension may be better
diagnosed and treated by examining its hemodynamic com-
ponents (CO and SVR). As MAP is the product of CO and
SVR, elevated mean BP results from elevated CO, SVR, or
both. As shown in Fig. 10, CO is the product of HR and SV.
Stroke volume is determined in part by LV fillin (preload)
and contractile (inotropic) state. Hypertension can thus result
from increases in SVR (vasoconstriction), HR (hyperchro-
notropy), preload (hypervolemia), or contractility (hyperinot-
ropy).
Invasive and Echocardiographic Studies
In a small group of men with severe hypertension, Sullivan
et al79 studied the relationship between baseline hemody-
namic status and the response to various antihypertensive
108 of 158.
AJHFebruary 2005VOL. 18, NO. 2, Part 2
agents that were randomly selected. Patients with elevated
SVR responded with decreases in SVR, and those with
elevated CO had BP control associated with normalization
of CO.
Treatment targeted at the specifi hemodynamic cause
of hypertension has predictable and appropriate results.
Easterling et al80 studied noninvasive hemodynamic pa-
rameters using Doppler echocardiography in 19 pregnant
hypertensive women. Ten patients had elevated CO,
whereas nine patients had elevated SVR, demonstrating
hemodynamic heterogeneity within this apparently homo-
geneous population. Patients with elevated CO were
treated with a -blocker (atenolol) and those with elevated
SVR were treated with hydralazine, a vasodilator targeted
at elevated SVR. Patients given hydralazine had dramatic
improvements in CO in association with decreases in
SVR; those given atenolol for elevated CO had improve-
ment in BP and normalization of CO. The investigators
suggest that the failure of previous studies to show con-
sistent results in the drug management of hypertension in
pregnancy may have resulted from treating a heteroge-
neous hemodynamic group with a single regimen. The
implication of their study is that hemodynamically guided
therapy would be expected to show more consistent results
in hemodynamically diverse populations.
Differential effects of antihypertensive medications on
hemodynamic variables may not be evident from changes
in BP alone. Resnick and Lester81 studied the effects of
various BP medications on arterial compliance in patients
referred to an outpatient practice specializing in hyperten-
sion. The changes in compliance of the large arteries
(capacitive compliance) and in smaller arteries (reflectiv
compliance) were evaluated during treatment with ACE
inhibitors, angiotensin-receptor blockers (ARBs), calcium
channel blockers (CCBs), and -blockers. The researchers
found that despite similar changes in SBP, DBP, and PP
during treatment, there were improvements in arterial
compliance with ACE inhibitors, ARB, and CCB but not
with -blockers. These researchers suggest that choosing
medications that have favorable effects on both BP and
arterial compliance might further enhance the potential
clinical benefi of drug therapy in hypertension. Simi-
larly, Zusman78 reported that despite similar degrees of BP
reduction, the hemodynamic effects of the CCB nifedipine
were favorable when compared with the -blocker ateno-
lol, resulting in decreased SVR, increased CO and im-
proved measures of LV contractility and diastolic
function. Others have shown significantl different hemo-
dynamic effects between various -blockers such as be-
tween metoprolol and carvedilol due to the -adrenergic
blocking properties of the latter.82
Studies suggest that most patients require multiple
medications to achieve BP control. In the Antihyperten-
sive and Lipid-Lowering Treatment to Prevent Heart At-
tack Trial (ALLHAT),83 a large randomized trial
comparing outcomes in patients treated with different
classes of antihypertensive agents, 90% of patients were
 AJHFebruary 2005VOL. 18, NO. 2, Part 2 HYPERTENSION AS A HEMODYNAMIC DISEASE 37S

on treatment at time of pre-randomization visit, although

-blockers
with ISA

2 or NC
2 or NC
NC or 1
only 27% had adequate BP control. After 5 years of treat-

NC
ment, 66% had achieved levels of BP 140/90 mm Hg. Of

1
2
the participants whose hyptertension was controlled, 63%

ACE angiotensin converting enzyme; CO cardiac output; LVH left ventricular hypertrophy; NC no change; SVR systemic vascular resistance; 1 Increase, 2 decrease.
were on two or more medications, indicating that BP
control required combination therapy in the majority of
cases. Yakovlevitch and Black84 reviewed 436 charts to

- and -
blockers

2 or NC
2 or NC
NC or 1
2 or NC
2 or NC
identify 91 cases of resistant hypertension referred to their

2
hypertension clinic of a tertiary care center and evaluated
for possible causes of resistance, including medication
noncompliance, secondary causes of hypertension, drug
interactions, and the appropriateness of the medical regi-

vasodilators
men. They found that the most common cause of inade-
quate BP control in the 91 patients identifie with resistant

Direct

1
1
1
1
2
1
hypertension was an inappropriate medical regimen.
The hemodynamic responses of various classes of an-
tihypertensive medications have been categorized in an
extensive review of hypertension by Houston,85 and se-
lected hemodynamic effects are summarized in Table 9.

blockers

NC or 1
NC or 1
NC or 1
NC
-

2
2
Rationale for an ICG-Guided
Approach to Antihypertensive

-agonists
Therapy

2 or NC
2 or NC
Central
Table 9. Selected hemodynamic effects of various antihypertensive agents by class

NC
NC

2
2
The foregoing discussion has presented data on the role of
hemodynamic information for diagnostic, prognostic, and
therapeutic decision making for patients with hyperten-
sion. The use of ICG-derived hemodynamic information to
inhibitors

improve BP control requires accurate assessment of base-

line hemodynamic state, creation of a therapeutic regimen
ACE

NC
2
2
1
2
1

based on hemodynamic status, and timely measurement of

changes in various hemodynamic parameters in response
to therapy.86 Studies have shown that it is very difficulti
not impossibleto make an accurate assessment of CO
blockers
Calcium
channel

1 or NC
1 or NC
NC or 2

and SVR at the bedside by physical examination

2
2
2

alone.20,21,87,88 Therefore, it is not likely to be possible to

use physical examination to reliably identify baseline he-
modynamic subsets or changes in hemodynamic status so
blockers

as to optimize therapy.
1 or NC

Clinicians have used ICG in various patient care set-

NC
2
1
2
2

tings to assess its applicability in the assessment and

treatment of hypertension.89 95 As noted earlier here and
in Fig. 7, there is significan hemodynamic heterogeneity
Diuretics

among individuals with hypertension, suggesting that BP

1 or NC
2 or NC

level alone is not adequate to categorize patients into

1
2
2
2

clinically meaningful subgroups. De Divitiis et al91 used

ICG to confir the presence of distinct hemodynamic
profile in patients with hypertension: 1) elevated CO in
Intravascular volume

Adapted from Houston.85

association with normal or nearly normal SVR, and 2)

predominantly elevated SVR. Margulis et al96 evaluated
other hemodynamic parameters in untreated patients with
Stroke volume

hypertension. They found impairment of cardiac perfor-

mance with decreased indices of contractility and evidence
Heart rate

for increased thoracic flui content, suggesting increased

Effect

water content of the lungs or thoracic wall tissues.

SVR

LVH
CO

Thoracic flui content, the reciprocal of total thoracic

109 of 158.
 38S HYPERTENSION AS A HEMODYNAMIC DISEASE
impedance, is strongly correlated with amount of flui in
the chest cavity, whether intravascular or extravascular. In
patients undergoing thoracentesis, Petersen et al97 demon-
strated a strong correlation between the volume of pleural
flui removed and the change in total thoracic impedance
(correlation coefficient 0.97). In studies using lower-body
negative pressure to create pooling of venous blood in the
lower extremities, Ebert et al98 found a nearly perfect
linear correlation with changes in central venous pressure
and changes in thoracic impedance. Thus, TFC has been
used to monitor changes in flui volume and guide diuretic
therapy in patients with hypertension.
Linb et al89 reported that BP reductions resulted from
improvements in baseline hemodynamic abnormalities;
patients with elevated CO responded to targeted therapy
with a -blocker (propranolol), whereas those with ele-
vated SVR responded to treatment with the vasodilating
CCB (nifedipine). Mattar et al99 showed that an intensive
regimen of diet and exercise resulted in improvements in
hemodynamic parameters with substantial increases in CO
and decreases in SVR despite only modest changes in
MAP. Moreover, the investigators speculated that failure
of some hypertensive patients to show hemodynamic im-
provement on serial ICG measures resulted from the in-
appropriate choice of medications that were not targeted
toward the underlying hemodynamic abnormalities.
Hemodynamic parameters derived from ICG have been
used to evaluate the differential effects of medications in
patients with essential hypertension. In a study of the
effects of a cardioselective -blocker compared with a
-blocker with intrinsic sympathomimetic activity, Toth et
al100 studied 57 patients randomized to treatment with
either atenolol or pindolol for 12 weeks. Pindolol therapy
was associated with a 12% decrease in SVR compared
with minimal change with atenolol. Atenolol-related im-
provement in BP resulted from decrease in HR and cardiac
index. Breithaupt-Grogler et al67 reported on the differen-
tial hemodynamic effects of combination therapy with
verapamil/trandolapril (Vera/Tran) compared with meto-
prolol/hydrochlorothiazide (Meto/HCTZ) in 26 patients
after 6 months of therapy. In addition to ICG-derived CO
and SVR, the authors measured carotidofemoral pulse
wave velocity as a measure of arterial stiffness. The com-
bination of CCB and ACE inhibitor (Vera/Tran) reduced
diastolic BP to a greater degree than Meto/HCTZ and
lowered SVR by about 15% compared with minimal
change with the -blocker/diuretic combination. Treat-
ment with Meto/HCTZ was associated with a significan
reduction in CO compared with baseline, which was not
seen with Vera/Tran. However, pulse wave velocity de-
creased with Vera/Tran but not with Meto/HCTZ, suggest-
ing an improvement in the elastic properties of the aorta
with the former drug regimen.
The ICG technique has been used to assess the hemo-
dynamic effects of sodium restriction in a small group of
subjects with mild hypertension.101 During sodium restric-
tion, ICG-derived measures of SV decreased in association
110 of 158.
AJHFebruary 2005VOL. 18, NO. 2, Part 2
with fall in diastolic BP. An increase in overall thoracic
impedance (the reciprocal of TFC) was consistent with a
decrease in extracellular flui volume. In addition, ICG
has been used to explore the mechanisms of responses to
ACE inhibitors and prostaglandin inhibitors in patients
who are either salt sensitive or salt insensitive.102 These
examples illustrate the use of ICG in assessing the mech-
anisms of BP elevation and the hemodynamic effects of
nonpharmacologic interventions in hypertension.
As noted above, antihypertensive medications ulti-
mately act on one or more of the hemodynamic compo-
nents that determine BP.79,80 Once the baseline
hemodynamic status is known, an appropriate medical
regimen can be designed based on the expected hemody-
namic effects of various medications. However, individual
patients vary in their responses to antihypertensive drugs
such that the actual hemodynamic effects and side effects
cannot reliably be predicted. Therefore, empiric selection
of drug combinations based on their general hemodynamic
actions as a class may not be successful in managing a
specifi patient, even if the baseline hemodynamic status is
known. The ICG technique is unique in that it can provide
not only an accurate hemodynamic profil noninvasively
but can guide therapy toward a drug regimen that is most
appropriate for the specifi patient based on serial mea-
surements. Periodic measurements of hemodynamic status
allow the physician to monitor therapy when results are
suboptimal or unexpected. It is for these reasons that ICG
has emerged as a valuable tool in the evaluation and
treatment of patients with hypertension.
Outcome Studies Using ICG-
Guided Therapy in Hypertension
The observation that hypertension is a hemodynamic dis-
ease implies that measurement of hemodynamic parame-
ters can be used to guide medication selection, to titrate
dose, and to evaluate efficac of the medical regimen.
Several studies have used ICG to evaluate hemodynamic
parameters and demonstrated that ICG-guided therapy im-
proves BP control. Taler et al103 randomized 104 patients
with hypertension uncontrolled on two or more drugs to a
3-month trial of ICG-guided therapy or standard therapy
directed by a hypertension specialist. In this study, BP
control (define as achieving BP 140/90 mm Hg) oc-
curred 70% more often in the ICG-guided group (Fig. 11).
Use of ICG resulted in greater reductions in SVR index
and more intensive use of diuretic therapy, guided by
levels of TFC. According to the study investigators, mea-
surement of hemodynamic and impedance parameters was
more effective than clinical judgment alone in guiding
selection of antihypertensive therapy patients resistant to
empiric therapy.
Sharman et al104 studied a cohort of patients in the
primary care offic setting with drug-resistant hyperten-
sion, define as systolic BP 140 mm Hg or diastolic BP
90 mm Hg during treatment with two antihypertensive
 AJHFebruary 2005VOL. 18, NO. 2, Part 2
FIG. 11 Percentage of patients achieving blood pressure (BP)
control using impedance cardiography (ICG) guided therapy
compared with nonICG-guided therapy. Adapted from Taler et
al.103
medications. Patients were treated based on a published
ICG-guided treatment algorithm (Fig. 12) for an average
of 7 months. In this study, ICG resulted in BP control in
57.1% of patients who were not controlled before the use
of ICG-guided therapy. The average number of medica-
tions increased from 2.0 at time of entry to 2.5 0.7 at the
end of the study period. The observation that hemody-
namic information derived from ICG resulted in BP con-
trol with two medications in some patients and three or
more in others is consistent with both higher intensity and
more appropriate medical regimens. The investigators
concluded that ICG is safe and cost-effective and could
assist community-based physicians in treating uncon-
trolled hypertension.
Sramek et al105 reported on a series of 322 patients
with hypertension uncontrolled despite previous ther-
apy with two or more antihypertensive agents for peri-
ods of 2 years or more. The researchers directed the
management of hypertension at both control of BP and
improvement in underlying hemodynamic parameters
including CO and SVR. At baseline, 16% of subjects
had significantl reduced CO (ie, were considered hy-
podynamic) and approximately 19% were hyperdy-
namic. In this large series of patients treated using the
results of ICG evaluation, so-called normodynamic
goal-oriented therapy controlled BP in 203 subjects
(63%) within several weeks. The investigators highlight
the observation that ICG was able to identify medica-
tions that were optimal and specifi for the individual
patients, resulting in an approach superior to the con-
ventional trial-and-error method.
Additional Roles of ICG in Patients
With Hypertension
The evidence cited above supports the use of ICG-derived
hemodynamic information in guiding the selection, initi-
ation, titration, and evaluation of antihypertensive medi-
cation. However, patients and their physicians fail to
111 of 158.
HYPERTENSION AS A HEMODYNAMIC DISEASE 39S
achieve adequate BP control for reasons other than the
responses to specifi medications. Common barriers to BP
control include lack of awareness of the condition, inabil-
ity to make necessary dietary and other lifestyle modifi
cations, noncompliance with medications, complicating
factors such as drug interactions, secondary causes of
hypertension, and comorbidities such as kidney disease.
Testing with ICG using currently available equipment
may favorably affect each of these issues. Oscillometric
measurements of BP, as with the most widely used ICG
equipment, are more reliable and less operator dependent
than standard BP techniques. The accurate and reproduc-
ible measures of CO and SVR identify patients with ab-
normal hemodynamic states and may increase clinical
suspicion and diagnostic sensitivity for those with border-
line or prehypertensive BP readings. The ICG reports are
useful teaching tools for patients and may provide moti-
vation for the dietary and other lifestyle changes that assist
in BP control.
Changes in hemodynamic parameters may identify in-
stances when patients stop their medications or when there
are complicating factors such as worsening renal function
or interactions with medications such as over-the-counter
nonsteroidal anti-inflammator drugs. As noted,103 an in-
crease in one class of hypertensive agents may result in
compensatory flui retention, leading to an increase in
TFC as measured by ICG and the need for higher doses of
diuretics. Similarly, flui retention resulting from the renal
effects of anti-inflammator medications may be recog-
nized by changes in TFC. Importantly, ICG-derived mea-
sures of cardiac performance, such as velocity index or
FIG. 12 Algorithm for hemodynamically guided therapy of hyper-
tension. Adapted from Sharman et al105 and Taler et al.104 In the
latter study, postural changes in total body impedance (TBI), the
reciprocal of thoracic fluid content (TFC), were used as the criteria
for decisions reguarding fluid status.
 40S HYPERTENSION AS A HEMODYNAMIC DISEASE
systolic time ratio, may be the initial signs of the devel-
opment of LV dysfunction.106,107
Implications of Hemodynamics
and Future Considerations
In addition to improving the diagnosis and therapy of
hypertension, hemodynamic measurements provide in-
sights into other aspects of cardiovascular function. For
example, studies have shown the importance of endothe-
lial function in the development and progression of car-
diovascular disease.108 Endothelial dysfunction, as
measured by reduced flow-mediate arterial dilation, is
associated with abnormal hemodynamic measures, includ-
ing elevated SVR.109 In the HOPE study,110 an ACE
inhibitora drug that both lowers SVR and improves
endothelial functionreduced mortality from cardiovas-
cular disease despite only minor effects on BP. Future
studies will likely examine the significanc of elevated
SVR and arterial compliance in individuals with hyperten-
sion or prehypertension and will correlate ICG-derived
hemodynamic parameters with other evolving markers of
increased cardiovascular risk such as C-reactive protein,
homocysteine level, and the metabolic syndrome.
The studies included in this supplement of the journal add
to the growing body of literature that supports the accuracy,
reliability, and clinical utility of ICG in diagnostic and prog-
nostic assessment and therapeutic management of patients
with hypertension. The use of ICG has added significantl to
our understanding of hypertension as a disease with both
hemodynamic causes and hemodynamic consequences. Just
as congestive heart failure reflect abnormal flo or inappro-
priate ventricular fillin pressures, hypertension occurs when
there is abnormal flo or inappropriate vascular resistance or
compliance. When hypertension impairs LV performance
(either systolic or diastolic), heart failure ensues. Although
these conditions often co-exist, in many cases hypertension is
a step in a hemodynamic continuum that leads to further
hemodynamic derangement and heart failure. It is believed
that future studies will confir recent finding that hemody-
namic measurements in individual patients will improve di-
agnosis, risk assessment, and treatment for these patients. It is
also possible that further exploration of the implications of
hypertension as a hemodynamic disease will lead to studies
demonstrating that earlier detection and treatment of the
hemodynamic components of hypertension may change the
natural history of this disease process.
References
1. Wiggers CJ: Hemodynamic changes in hypertension. Am Heart J
1938;16:515525.
2. Freis ED: Hemodynamics of hypertension. Physiol Rev 1960;40:
2754.
3. Freis ED: Studies in hemodynamics and hypertension. Hypertension
2001;38:15.
112 of 158.
AJHFebruary 2005VOL. 18, NO. 2, Part 2
4. Lund-Johansen P: Newer thinking on the hemodynamics of hyper-
tension. Curr Opin Cardiol 1994;9:505511.
5. Julius S: Autonomic nervous system dysregulation in human hyper-
tension. Am J Cardiol 1991;67:3B7B.
6. Susic D, Frohlich ED: Hypertension and the heart. Curr Hypertens
Rep 2000;6:565569.
7. Messerli FH, DeCarvalho JG, Christie B, Frohlich ED: Essential
hypertension in black and white subjects. Hemodynamic finding
and flui volume state. Am J Med 1979;67:2731.
8. Ventura H, Messerli FH, Oigman W, Suarez DH, Dreslinski GR,
Dunn FG, Reisin E, Frohlich ED: Impaired systemic arterial com-
pliance in borderline hypertension. Am Heart J 1984;108:132136.
9. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA,
Izzo JL, Jones DW, Materson BJ, Oparil S, Wright JT, Roccella EJ:
The Seventh Report of the Joint National Committee on Prevention,
Detection, Evaluation, and Treatment of High Blood Pressure. The
JNC 7 Report. J Am Med Assoc 2003;289:2560 2572.
10. Oparil S, Zaman MA, Calhoun DA: Pathogenesis of hypertension.
Ann Intern Med 2003;139:761776.
11. Schappert SM, Nelson C: National ambulatory medical care survey:
199596 summary. Vital and health statistics, series 13, no. 142.
Washington, DC: Government Printing Office November 1999.
DHHS publication no. (PHS) 2000-1713.
12. Hyman DJ, Pavlik VN: Characteristics of patients with uncontrolled
hypertension in the United States. N Engl J Med 2001;345:479
486.
13. Setaro JF, Black HR: Refractory hypertension. N Engl J Med
1992;327:543547.
14. Oparil S, Calhoun DA: Managing the patient with hard-to-control
hypertension. Am Fam Phys 1998;57:10071014.
15. Gifford RW Jr. Resistant hypertension: introduction and definitions
Hypertension 1988;11(Suppl II):II-65II-66.
16. Wilking SV, Belanger A, Kannel WB, DAgostino RB, Steel K:
Determinants of isolated systolic hypertension. J Am Med Assoc
1988;260:34513455.
17. Lewington S, Clarke R, Qizilbash N, Peto R, Collins R: Age-
specifi relevance of usual blood pressure to vascular mortality: a
meta-analysis of individual data for one million adults in 61 pro-
spective studies. Lancet 2002;360:19031913.
18. Elliott WJ: The economic impact of hypertension. J Clin Hypertens
2003;5(3 Suppl 2):313.
19. Goetzel RZ, Long SR, Ozminkowski RJ, Hawkins K, Wang S,
Lynch W: Health, absence, disability, and presenteeism cost esti-
mates of certain physical and mental health conditions affecting U.S.
employers. J Occup Environ Med 2004;46:398 412.
20. Steingrub JS, Celoria G, Vickers-Lahti M, Teres D, Bria W: Ther-
apeutic impact of pulmonary artery catheterization in a medical/
surgical ICU. Chest 1991;99:14511455.
21. Connors AF, Dawson NV, Shaw PK, Montenegro HD, Nara AR,
Martin L: Hemodynamic status in critically ill patients with and
without acute heart disease. Chest 1990;98:1200 1206.
22. Hoit BD, Rashwan M, Watt C, Sahn DJ, Bhargava V: Calculating
cardiac output from transmitral volume flo using Doppler and
M-mode echocardiography. Am J Cardiol 1988;62:131135.
23. Northridge DB, Findlay IN, Wilson J, Henderson E, Dargie HJ:
Non-invasive determination of cardiac output by Doppler echocar-
diography and electrical bioimpedance. Br Heart J 1990;63:9397.
24. Van De Water JM, Miller TW: Impedance cardiography: the next
vital sign technology? Chest 2003;123:2028 2033.
25. Albert N, Hail M, Li J, Young J: Equivalence of bioimpedance and
thermodilution in measuring cardiac output and index in patients
with advanced, decompensated chronic heart failure hospitalized in
critical care. Am J Crit Care 2004;13:469 479.
26. Drazner M, Thompson B, Rosenberg P, Kaiser PA, Boehrer JD,
Baldwin BJ, Dries DL, Yancy CW: Comparison of impedance
cardiography with invasive hemodynamic measurements in patients
with heart failure secondary to ischemic or nonischemic cardiomy-
opathy. Am J Cardiol 2002;89:993995.
 AJHFebruary 2005VOL. 18, NO. 2, Part 2
27. Ziegler D, Grotti L, Krucke G: Comparison of cardiac output
measurements by TEB vs. intermittent bolus thermodilution in
mechanical ventilated patients (Abstract). Chest 1999;116:281S.
28. Sageman WS, Riffenburgh RH, Spiess BD: Equivalence of bioim-
pedance and thermodilution in measuring cardiac index after cardiac
surgery. J Cardiothoracic Vasc Anesth 2002;16:8 14.
29. Yung, G, Fedullo P, Kinninger K, Johnson W, Channick R: Com-
parison of impedance cardiography to direct Fick and thermodilu-
tion cardiac output determination in pulmonary arterial
hypertension. Congest Heart Fail 2004;10(Suppl 2):710.
30. Verhoeve PE, Cadwell CA, Tsadok S: Reproducibility of noninva-
sive bioimpedance measurements of cardiac function (Abstract).
J Card Fail 1998;4(Suppl):S3.
31. Aldrete AJ, Brown C, Daily J, Buerke V: Pacemaker malfunction
due to microcurrent injection from a bioimpedance noninvasive
cardiac output monitor. J Clin Monit 1995;11:131133.
32. Frohlich ED: Local hemodynamic changes in hypertension. Insights
for therapeutic preservation of target organs. Hypertension 2001;38:
1388 1394.
33. Segers P, Brimioulle S, Stergiopulos N, Westerhof N, Naeije R,
Maggiorini M, Verdonck P: Pulmonary arterial compliance in dogs
and pigs: the three-element Windkessel model revisited. Am J
Physiol 199;277:H725H731.
34. Randall OS, Westerhof N, van den Bos GC, Alexander B: Reliabil-
ity of stroke volume to pulse pressure ratio for estimating and
detecting changes in arterial compliance. J Hypertens 1986;4(Suppl
5):S293S296.
35. Chemla D, Hebert J-L, Coirault C, Zamani K, Suard I, Colin P,
Lecarpentier Y: Total arterial compliance estimated by stroke vol-
ume-to-aortic pulse pressure ratio in humans. Am J Physiol 1998;
274: (Heart Circ Physiol 43) H500 H505.
36. Egan B, Schmouder R: The importance of hemodynamic consider-
ations in essential hypertension. Am Heart J 1988;116:594 599.
37. Izzo J: Arterial stiffness and the systolic hypertension syndrome.
Curr Opin Cardiol 2004;19:341352.
38. Guyton AC: The relationship of cardiac output and arterial pressure
control. Circulation 1981;64:1079 1088.
39. Fouad-Tarazi F, Izzo JL: Hemodynamic profile and responses, in
Izzo JL, Black HR (eds): Hypertension Primer, : The Essentials of
High Blood Pressure. Council on High Blood Pressure Research
(American Heart Association), Dallas, 1999, pp 115117. Second
Edition
40. Julius S, Krause L, Schork NJ, Mejia AD, Jones KA, van de Ven C,
Johnson EH, Sekkarie MA, Kjeldsen SE, Petrin J: Hyperkinetic
borderline hypertension in Tecumseh, Michigan. J Hypertens 1991;
9:77 84.
41. Lund-Johansen P: Hemodynamic patterns in the natural history of
borderline hypertension. J Cardiovasc Pharm 1986;8(Suppl 5):S8
S14.
42. Slotwiner DJ, Devereux RB, Schwartz JE, Pickering TG, de Simone
G, Roman MJ: Relation of age to left ventricular function and
systemic hemodynamics in uncomplicated mild hypertension. Hy-
pertension 2001;37:1404 1409.
43. Franklin SS: Blood pressure and cardiovascular disease: what re-
mains to be achieved? J Hypertens 2001;19(Suppl 3):S3S8.
44. Messerli FH, Garavaglia GE, Schmieder RE, Sundgaard-Riise K,
Nunez BD, Amodeo C: Disparate cardiovascular finding in men
and women with essential hypertension. Ann Intern Med 1987;107:
158 161.
45. Davidson RC, Ahmad S: Hemodynamic profile in essential and
secondary hypertension, in Izzo JL, Black HR (eds): Hypertension
Primer: The Essentials of High Blood Pressure. Council on High
Blood Pressure Research (American Heart Association), Dallas,
2003, pp 349 351. Third Edition
46. Romney JS, Lewanczuk RZ: Vascular compliance is reduced in
the early stages of type 1 diabetes. Diabetes Care 2001;24:2102
2106.
47. Taler S, Driscoll N, Tibor M, Sprau G, Augustine J, Larson T,
113 of 158.
HYPERTENSION AS A HEMODYNAMIC DISEASE 41S
Stegall M, Textor S: Changes in hemodynamic patterns with age
in normotensive subjects (Abstract). Am J Hypertens 2004;17:
373.
48. Hinderliter AL, Sherwood A, Blumenthal JA, Light KC, Girdler SS,
McFetridge J, Johnson K, Waugh R: Changes in hemodynamics and
left ventricular structure after menopause. Am J Cardiol 2002;89:
830 833.
49. Galarza CR, Alfi J, Waisman GD, Mayorga LM, Camera LA, del
Rio M, Vasvari F, Limansky R, Farias J, Tessler J, Camera MI:
Diastolic pressure underestimates age-related hemodynamic impair-
ment. Hypertension 1997;30:809 816.
50. Alfi J, Gabriel D, Waisman G, Camera MI: Contribution of stroke
volume to the change in pulse pressure pattern with age. Hyperten-
sion 1999;34:808 812.
51. Hartley TR, Lovallo WR,Whitsett TL: Cardiovascular effects
of caffeine in men and women. Am J Cardiol 2004;93:
10221026.
52. Yu BH, Nelesen R, Ziegler MG, Dimsdale JE: Mood states and
impedance cardiography-derived hemodynamics. Ann Behav Med
2001;23:2125.
53. Hinderliter A, Blumenthal J, Waugh R, Chilukuri M, Sherwood A:
Ethnic differences in left ventricular structure: relations to hemody-
namics and diurnal blood pressure variation. Am J Hypertens 2004;
17:43 49.
54. Calhoun DA, Mutinga ML, Collins AS, Wyss JM, Oparil S: Nor-
motensive blacks have heightened sympathetic response to cold
pressor test. Hypertension 1993;22:801 805.
55. Fagard RH, Pardaens K, Staessen JA, Thijs L: Prognostic value of
invasive hemodynamic measurements at rest and during exercise in
hypertensive men. Hypertension 1996;28:3136.
56. Fagard RH, Pardaens K, Staessen JA, Thijs L: The pulse pres-
sure-to-stroke index ratio predicts cardiovascular events and
death in uncomplicated hypertension. J Am Coll Cardiol 2001;
38:22731.
57. de Simone G, Roman MJ, Koren MJ, Mensah GA, Ganau A,
Devereux RB: Stroke volume/pulse pressure ratio and cardiovas-
cular risk in arterial hypertension. Hypertension 1999;33:800
805.
58. Janeway TC: A clinical study of hypertensive cardiovascular dis-
ease. Arch Intern Med 1913;12:755798.
59. Ganau A, Devereux RB, Roman MJ, de Simone G, Pickering TG,
Saba PS, Vargiu P, Simongini I, Laragh JH: Patterns of left ven-
tricular hypertrophy and geometric remodeling in essential hyper-
tension. J Am Coll Cardiol 1992;19:1550 1558.
60. Kahan T: The importance of left ventricular hypertrophy in human
hypertension. J Hypertens 1998;16(Suppl):S23S29.
61. Verdecchia P, Schillaci G, Borgioni C, Ciucci A, Gattobigio R,
Zampi I, Santucci A, Santucci C, Reboldi G, Porcellati C: Prognos-
tic value of left ventricular mass and geometry in systemic hyper-
tension with left ventricular hypertrophy. Am J Cardiol 1996;78:
197202.
62. Verdecchia P, Carini G, Circo A, Dovellini E, Giovannini E, Lom-
bardo M, Solinas P, Gorini M, Maggioni AP: Left ventricular mass
and cardiovascular morbidity in essential hypertension: the MAVI
study. J Am Coll Cardiol 2001;38:1829 1835.
63. Vakili BA, Okin PM, Devereux RB: Prognostic implications of left
ventricular hypertrophy. Am Heart J 2001;141:334 341.
64. Ichkhan K, Molnar J, Somberg J: Relation of left ventricular mass
and QT dispersion in patients with systematic hypertension. Am J
Cardiol 1997;79:508 511.
65. Gomez-Cerezo J, Rios Blanco JJ, Suarez Garcia I, Moreno Anaya P,
Garcia Raya P, Vazquez-Munoz E, Barbado Hernandez FJ: Nonin-
vasive study of endothelial function in white coat hypertension.
Hypertension 2002;40:304 309.
66. London GM, Cohn JN: Prognostic application of arterial stiffness:
task forces. Am J Hypertens 2002;15:754 758.
 42S HYPERTENSION AS A HEMODYNAMIC DISEASE
67. Breithaupt-Grogler K, Gerhardt G, Lehmann G, Notter T, Belz GG:
Blood pressure and aortic elastic propertiesverapamil SR/tran-
dolapril compared to a metoprolol/hydrochlorothiazide combination
therapy. Int J Clin Pharmacol Ther 1998;36:425 431.
68. Glasser SP: Hypertension syndrome and cardiovascular events.
Postgrad Med 2001;110:29 35.
69. Intengan HD, Schiffrin EL: Vascular remodeling in hypertension:
roles of apoptosis, inflammation and fibrosis Hypertension 2001;
38:581587.
70. Mathew J, Sleight P, Lonn E, Johnstone D, Pogue J, Yi Q, Bosch J,
Sussex B, Probstfiel J, Yusuf S: Reduction of cardiovascular risk
by regression of electrocardiographic markers of left ventricular
hypertrophy by the angiotensin-converting enzyme inhibitor
ramipril. Circulation 2001;104:16151621.
71. Corea L, Bentivoglio M, Verdecchia P, Providenza M, Motolese M:
Left ventricular hypertrophy regression in hypertensive patients
treated with metoprolol. Int J Pharmacol Ther Toxicol 1984;22:365
370.
72. Yoshitomi Y, Nishikimi T, Abe H, Nagata S, Kuramochi M, Mat-
suoka H, Omae T: Left ventricular systolic and diastolic function
and mass before and after antihypertensive treatment in patients
with essential hypertension. Hypertens Res 1997;20:2328.
73. Ofil EO, Cohen JD, St Vrain JA, Pearson A, Martin TJ, Uy ND,
Castello R, Labovitz AJ: Effect of treatment of isolated systolic
hypertension on left ventricular mass. J Am Med Assoc 1998;279:
778.
74. Verdecchia P, Angeli F, Borgioni C, Gattobigio R, de Simone G,
Devereux RB, Porcellati C: Changes in cardiovascular risk by
reduction of left ventricular mass in hypertension: a meta-analysis.
Am J Hypertens 2003;16:895 899.
75. Galarza C, Alfi J, Waisman G, Burgos M, Zaniello G, del Rio M,
Serra L, Toselli P, Vasvari F, Camera M: Severe systemic hemo-
dynamic impairment in patients with stroke (Abstract). Am J Hy-
pertens 1996;9(Suppl 1):172A.
76. Johns DW, Peach MJ: Factors that contribute to resistant forms of
hypertension. Pharmacological considerations. Hypertension 1988;
11(Suppl II):II-88 II-95.
77. Tsukiyama H: Choice of antihypertensive agents in hemodynamic
aspects to match pathophysiology and pharmacology in essential
hypertension. Jpn J Med 1989;28:261264.
78. Zusman RM: Left ventricular function in hypertension. Relevance
to the selection of antihypertensive therapy. Am J Hypertens 1989;
2:200S206S.
79. Sullivan JM, Schoeneberger TE, Ratts ET, Palmer ET, Samaha JK,
Mance CJ, Muirhead EE: Short-term therapy of severe hyperten-
sion. Hemodynamic correlates of the antihypertensive response in
man. Arch Intern Med 1979;139:12331239.
80. Easterling TR, Benedetti TJ, Schmucker BC, Carlson KL: Antihy-
pertensive therapy in pregnancy directed by noninvasive hemody-
namic monitoring. Am J Perinatol 1989;6:86 89.
81. Resnick LM, Lester MH: Differential effects of antihypertensive
drug therapy on arterial compliance. Am J Hypertens 2002;15:
1096 1100.
82. Weber K, Bohmeke T, van der Does R, Taylor SH: Hemodynamic
differences between metoprolol and carvedilol in hypertensive pa-
tients. Am J Hypertens 1998;11:614 617.
83. ALLHAT Coordinators and Officers Major outcomes in high-risk
hypertensive patients randomized to angiotensin-converting enzyme
inhibitor or calcium channel blocker vs. diuretic. The Antihyperten-
sive and Lipid-Lowering Treatment to Prevent Heart Attack Trial
(ALLHAT). J Am Med Assoc 2002;288:29812997.
84. Yakovlevitch M, Black HR: Resistant hypertension in a tertiary care
clinic. Arch Intern Med 1991;151:1786 1792.
85. Houston MC: Hypertension strategies for therapeutic interven-
tion and prevention of end-organ damage. Prim Care 1991;18:
713753.
86. Taler SJ, Augustine J, Textor SC: A hemodynamic approach to
resistant hypertension. Congest Heart Fail 2000;6:90 93.
114 of 158.
AJHFebruary 2005VOL. 18, NO. 2, Part 2
87. Bailey JM, Levy JH, Kopel MA: Relationship between clinical
evaluation of peripheral perfusion and global hemodynamics in
adults after cardiac surgery. Crit Care Med 1990;18:1353
1356.
88. Tibby SM, Hatherill M, Marsh MJ, Murdoch IA: Clinicians abili-
ties to estimate cardiac index in ventilated children and infants. Arch
Dis Child 1997;77:516 518.
89. Linb G, Eisenberg BM: Noninvasive techniques for evaluation of
heart function and hemodynamics in arterial hypertension. Acta
Cardiol 1990;45:133139.
90. Abdelhamed AI, Levy P, Smith KS, Wilkins J, Smith R, Ferrario
CM: Value of non-invasive hemodynamic measurements in hyper-
tensive patients. Am J Hypertens 2002;15(Suppl 1):A89.
91. de Divitiis O, Di Somma S, Liguori V, Petitto M, Magnotta C,
Ausiello M, Natale N, Brignoli M, Galderisi M: Effort blood pres-
sure control in the course of antihypertensive treatment. Am J Med
1989;87:(Suppl 3C)46S56S.
92. Eliot RS: The dynamics of hypertensionan overview: present
practices, new possibilities, and new approaches. Am Heart J 1988;
116:583589.
93. Taler SJ, Augustine J, Burnett JC, Textor SC: Hemodynamic and
volume changes during intensive treatment (Rx) for resistant hyper-
tension (ResHTN) (Abstract). Am J Hypertens 2000;13:61A 62A.
94. Lang M: Noninvasive hemodynamic-driven care management.
Achieving quality patient outcomes across a continuum of care.
Cardiovasc Dis Manage 2001;7173.
95. Vuurmans T, Boper P, Koomans H: Effects of endothelin-1 and
endothelin-1 receptor blockade on cardiac output, aortic pressure,
and pule wave velocity in humans. Hypertension 2003;41:1253
1258.
96. Margulis F, Dalton R, Killinger C, Pisarello J, Alvarez C: Decreased
myocardial contractility and increased thoracic flui content de-
tected by impedance cardiography in untreated hypertensive patients
(Abstract). Am J Hypertens 1997;10(Suppl 1):55A.
97. Petersen JR, Jensen BV, Drabaek H, Viskum K, Mehlsen J: Elec-
trical impedance measured changes in thoracic flui content during
thoracentesis. Clin Physiol 1994;14:459 466.
98. Ebert TJ, Smith JJ, Barney JA, Merrill DC, Smith GK: The use of
thoracic impedance for determining thoracic blood volume changes
in man. Aviat Space Environ Med 1986;57:49 53.
99. Mattar JA, Salas CE, Bernstein DP, Lehr D, Bauer R: Hemody-
namic changes after an intensive short-term exercise and nutrition
program in hypertensive and obese patients with and without cor-
onary disease. Arq Bras Cardiol 1990;54:307312.
100. Toth PD, Demeter RJ, Woods, JR, Nyhuis AW, Judy WV:
Comparison of the effects of pindolol and atenolol on hemody-
namic function in systemic hypertension. Am J Cardiol 1988;62:
413 418.
101. Koga Y, Gillum RF, Kubicek WG: An impedance cardiographic
study of the mechanism of blood pressure fall after moderate dietary
sodium restriction. Jpn Heart J 1985;26:197207.
102. Ashida T, Nishioeda Y, Kimura G, Kojima S, Kawamura M,
Imanishi M, Abe H, Kawano Y, Yoshimi H, Yoshida K, Kuramo-
chi M, Omae T: Effects of salt, prostaglandin, and captopril on
vascular responsiveness in essential hypertension. Am J Hyper-
tens 1989;2:640 642.
103. Taler SJ, Textor SC, Augustine JE: Resistant hypertension: com-
paring hemodynamic management to specialist care. Hypertension
2002;39:982988.
104. Sharman DL, Gomes CP, Rutherford JP: Improvement in blood
pressure control with impedance cardiograph-guided pharmacologic
decision making. Congest Heart Fail 2004;10:54 58.
105. Sramek BB, Tichy JA, Hojerova M, Cervenka V: Normohemody-
namic goal-oriented antihypertensive therapy improves the out-
come. Am J Hypertens 1996;9:141A.
106. Parrott C, Burnham K, Quale C, Lewis D: Comparison of changes
in ejection fraction to changes in impedance cardiography cardiac
 AJHFebruary 2005VOL. 18, NO. 2, Part 2
index and systolic time ratio. Congest Heart Fail 2004;10(2 Suppl
2):1113.
107. Thomson B, Drazner M, Dries D, Yancy C: Is impedance cardiog-
raphy-derived systolic time ratio a useful method to determine left
ventricular systolic dysfunction in heart failure (Abstract)? J Card
Failure 2004;10(Suppl 4):S38.
108. Ross R: Atherosclerosisan inflammator disease. N Engl J Med
1999;340:115126.
115 of 158.
HYPERTENSION AS A HEMODYNAMIC DISEASE 43S
109. Panza JA: High-normal blood pressuremore high than nor-
mal.N Engl J Med 2001;345:13371340.
110. Dagenais GR, Yusuf S, Bourassa MG, Yi Q, Bosch J, Lonn EM,
Kouz S, Grover J: Effects of ramipril on coronary events in high-risk
persons: results of the Heart Outcomes Prevention Evaluation
Study. Circulation 2001;104:522526.
111. Herrera-Acosta J, Perez-Grovas H, Fernandez M, Arriaga J: Enala-
pril in essential hypertension. Drugs 1985;30(Suppl 1):35 46.
 European Journal of Heart Failure, June 2004

Whole-Body Electrical Bio-Impendance is accurate in Non Invasive Determination of

Cardiac Output: A Thermodilution controlled, Prospective, Double Blind Evaluation.
1 2 2 2 3
Guillermo Torre-Amiot MD, Gad Cotter MD, Zvi Vered MD, Edo Kaluski MD, Karl Stang MD.
1 2
From Baylor College of Medicine, Houston, TX, USA ( ), Assaf-Harofeh Medical Center, Israel ( )
3
and Charite Campus, Berlin, Germany ( ).
TM
Background: The NICaS is a novel non-invasive apparatus based on whole body electrical
bio-impedance for simple non-invasive continuous CO determination.

Patients and Methods: Patients were recruited while randomized in a study evaluating the
efficacy of Tezosentan (a ET-A/B endothelin antagonist) in patients admitted due to acute heart
failure (CHF). Patients were randomized after having been hospitalized due to acute heart failure
2
with dyspnea at rest, CI < 2.5 L/min/m and PCWP 20 mmHg. Study Protocol: At baseline and
during treatment with study drug at the pre-specified time points of 0.5,1,2,3,4 and 6 hours from
TM
randomization CO was determined by both thermodilution and the NICaS 2001 apparatus. At
TM
each time point CO was determined by thermodilution and NICaS 2001 apparatus by a two
independed, blinded operators.

Results: Out of 130 patients enrolled, in 93 CO was measured simultaneously by both methods
at all the pre-determined time points. The overall Correlation between the two methods was
R=0.81 (Figure). Precision and bias were 0.010.6 L/min. There was a difference between the
2
two methods in cardiac output readings. When Mean CI (of both methods) was < 2 L/min/M CO
2
readings were statistically significantly lower by NICaS while when CI was >3 L/min/M , CO
readings were statistically significantly higher by NICaS. We have calculated the cardiac power
index (Cpi=CI* mean arterial pressure), and found that low Cpi (indicating reduced myocardial
contractile reserve) was related to higher recurrent CHF. However, Cpi based on NICaS CI
measurement (NICaS Cpi) was a better predictor of recurrent CHF then thermodilution Cpi (Th
Cpi), due to less accurate prediction in patients with high Cpi.

Conclusions: NICaS is a novel accurate non-invasive method for CO determination. The results
of the present study suggest that NICaS might be more accurate then thermodilution for CO
determination due to the tendency of thermodilution to under estimate CO when high and over
estimate it when low.

116 of 158.
 Accurate, Noninvasive Continuous
Monitoring of Cardiac Output by Whole-Body
Electrical Bioimpedance
Gad Cotter, Yaron Moshkovitz, Edo Kaluski, Amram J. Cohen, Hilton
Miller, Daniel Goor and Zvi Vered

Chest 2004;125;1431-1440
DOI 10.1378/chest.125.4.1431
The online version of this article, along with updated information
and services can be found online on the World Wide Web at:
http://chestjournals.org/cgi/content/abstract/125/4/1431

CHEST is the official journal of the American College of Chest

Physicians. It has been published monthly since 1935. Copyright 2007
by the American College of Chest Physicians, 3300 Dundee Road,
Northbrook IL 60062. All rights reserved. No part of this article or PDF
may be reproduced or distributed without the prior written permission
of the copyright holder
(http://www.chestjournal.org/misc/reprints.shtml). ISSN: 0012-3692.

Downloaded from chestjournals.org on April 4, 2007

117 of 158. Copyright 2004 by American College of Chest Physicians
 clinical investigations in critical care
Accurate, Noninvasive Continuous
Monitoring of Cardiac Output by Whole-
Body Electrical Bioimpedance*
Gad Cotter, MD; Yaron Moshkovitz, MD; Edo Kaluski, MD;
Amram J. Cohen, MD, FCCP; Hilton Miller, MD; Daniel Goor, MD; and
Zvi Vered, MD

Study objectives: Cardiac output (CO) is measured but sparingly due to limitations in its measure-
ment technique (ie, right-heart catheterization). Yet, in recent years it has been suggested that CO
may be of value in the diagnosis, risk stratification, and treatment titration of cardiac patients,
especially those with congestive heart failure (CHF). We examine the use of a new noninvasive,
continuous whole-body bioimpedance system (NICaS; NI Medical; Hod-Hasharon, Israel) for mea-
suring CO. The aim of the present study was to test the validity of this noninvasive cardiac output
system/monitor (NICO) in a cohort of cardiac patients.
Design: Prospective, double-blind comparison of the NICO and thermodilution CO determinations.
Patients: We enrolled 122 patients in three different groups: during cardiac catheterization (n 40);
before, during, and after coronary bypass surgery (n 51); and while being treated for acute
congestive heart failure (CHF) exacerbation (n 31).
Measurements and intervention: In all patients, CO measurements were obtained by two inde-
pendent blinded operators. CO was measured by both techniques three times, and an average was
determined for each time point. CO was measured at one time point in patients undergo-
ing coronary catheterization; before, during, and after bypass surgery in patients undergoing
coronary bypass surgery; and before and during vasodilator treatment in patients treated for acute
heart failure.
Results: Overall, 418 paired CO measurements were obtained. The overall correlation between the
NICO cardiac index (CI) and the thermodilution CI was r 0.886, with a small bias (0.0009 0.684
L) [mean 2 SD], and this finding was consistent within each group of patients. Thermodilution
readings were 15% higher than NICO when CI was < 1.5 L/min/m2, and 5% lower than NICO when
CI was > 3 L/min/m2. The NICO has also accurately detected CI changes during coronary bypass
operation and vasodilator administration for acute CHF.
Conclusion: The results of the present study indicate that whole-body bioimpedance CO measure-
ments obtained by the NICO are accurate in rapid, noninvasive measurement and the follow-up of
CO in a wide range of cardiac clinical situations. (CHEST 2004; 125:14311440)

Key words: cardiac function test; cardiac output; congestive heart failure

Abbreviations: CABG coronary artery bypass graft; CHF congestive heart failure; CI cardiac index; CO cardiac
output; Cpi cardiac power index; ISDN isosorbide-dinitrate; MAP mean arterial BP; NICO noninvasive cardiac
output system/monitor; SV stroke volume; SVRi systemic vascular resistance index; TEB thoracic electrical bioim-
pedance; WBEB whole-body electrical bioimpedance

M easurement of cardiac output (CO) and the cal-

culation of cardiac index (CI) has been used
selectively over the last 2 decades, mainly due to the
fact that CI measurement requires the invasive proce-
dure of right-heart catheterization and placement of a
Swan-Ganz catheter (Baxter Healthcare; Irvine, CA).
Hence, the experience with its application for monitor-
ing and risk stratification of cardiac patients is limited.
In recent years, however, evidence has accumu-
lated to the effect that CI measurement and the
calculation of systemic vascular resistance index
(SVRi) and cardiac power index (Cpi, the product of CI

www.chestjournal.org CHEST / 125 / 4 / APRIL, 2004 1431

Downloaded from chestjournals.org on April 4, 2007
118 of 158. Copyright 2004 by American College of Chest Physicians
 multiplied by mean arterial BP [MAP]) measured
simultaneously might be instrumental in the monitor-
ing and risk stratification of cardiac patients, especially
those with acute and chronic congestive heart failure
(CHF) and patients admitted with cardiogenic shock.1
In cardiogenic shock, a recent analysis of the SHOCK
(SHould We Emergently Revascularize Occluded Cor-
onaries in Cardiogenic ShocK) registry data has shown
Cpi to be the strongest independent predictor of
in-hospital mortality,2 while in acute heart failure Cpi
was found to be an important tool for diagnosis and risk
stratification.3,4 In patients with chronic CHF, a few
studies57 have shown that noninvasive Cpi reserve (the
increase in Cpi during exercise or dobutamine stress) is
the strongest predictor of outcome (a better oxygen
consumption and echocardiographic ejection fraction)
in such patients. Furthermore, changes in acute SVRi
may be useful for early detection of myocardial isch-
emia.8 Moreover, in two separate studies9 11 examining
the efficacy of vasodilators in patients with acute CHF,
medication was found to be effective mainly in patients
who were submitted to hemodynamic monitoring, im-
plying that perhaps in order to be efficacious, vaso-
dilator treatment should be monitored attentively to
prevent overtreatment and undertreatment. In differ-
ent studies,12,13 we have also demonstrated that careful
titration of vasodilator treatment administered for acute
heart failure and acute coronary syndromes is impor-
tant to optimize its efficacy. In the present study,
we evaluated the accuracy of a novel method of CI
measurement (whole-body electrical bioimpedance
[WBEB]) in different cardiac clinical settings (during
cardiac catheterization and coronary artery bypass graft
[CABG] surgery, and for monitoring patients with
acute CHF) and over a wide range of CI values and
severity of left ventricular dysfunction.
Materials and Methods
Patient Populations
Group 1: Group 1 consisted of 40 patients with coronary artery
disease referred during March to July 1993 for left and right
*From the Cardiology Department (Drs. Cotter, Kaluski, and
Vered), AssafHarofeh Medical Center, Zerifin; the Cardiac
Surgery Department (Drs. Moshkovitz and Goor), Sackler School
of Medicine, Tel Aviv University, Tel Aviv; the Department of
Cardiology (Dr. Miller), Sourasky Medical Center, Tel-Aviv; and
the Department of Cardiac Surgery (Dr. Miller), the Edith
Wolfson Medical Center, Holon, Israel.
Deceased.
Manuscript received March 19, 2003; revision accepted Septem-
ber 19, 2003.
Reproduction of this article is prohibited without written permis-
sion from the American College of Chest Physicians (e-mail:
permissions@chestnet.org).
Correspondence to: Gad Cotter, MD, Cardiology Department,
Assaf-Harofeh Medical Center, 70300, Zerifin, Israel; e-mail:
cotterg@hotmail.com
1432
Downloaded from chestjournals.org on April 4, 2007
119 of 158. Copyright 2004 by American College of Chest Physicians
cardiac catheterization based on conventional clinical indications.
During the right-heart study, a pulmonary artery catheter was
introduced under fluoroscopy; at a single time point, a paired
measurement of CI by a noninvasive cardiac output system/
monitor (NICO) [NICaS; NI Medical; Hod-Hasharon, Israel]
and by thermodilution was performed.
Group 2: Group 2 included 51 patients undergoing CABG
operations. The first 15 patients were studied at Wolfson Medical
Center (Israel) during October to November 1994. The next 16
patients were studied at Johns Hopkins Medical Center during
April to May 1995. The remaining 20 patients were studied again
at Wolfson Medical Center during August to October 1995. A
balloon-guided, Swan-Ganz catheter was introduced after the
induction of anesthesia, and five paired NICO CI and thermodi-
lution CI measurements were obtained at specific operative and
postoperative stages: immediately prior to the skin incision; after
sternotomy; after pericardiotomy; 10 min after weaning from the
pump; and immediately after arrival to the ICU. The results
obtained from the first 31 cases of this series have already been
published.14
Group 3: Group 3 consisted of 31 patients admitted during
September to December 2001 to the ICU of AssafHarofeh
Medical Center (Israel) because of an acute exacerbation of
CHF. Prior to admittance, they underwent a right-heart study in
the catheterization laboratory for the assessment of their cardiac
condition, and a Swan-Ganz catheter was inserted under fluoros-
copy. CO measurement was begun on arrival to the ICU, where
three baseline measurements were obtained, 15 min apart. In 17
patients who required vasodilator therapy, four measurements
were obtained during the initiation and up-titration of IV isosor-
bide-dinitrate (ISDN) treatment. In the 14 patients who did not
require ISDN treatment, an additional (fourth) baseline mea-
surement was obtained. All the studies were approved by the
review boards and the Helsinki committees of the various
hospitals. Consents for the studies were obtained from each
patient.
Measuring CI
Thermodilution: In all three study groups, a No. 7F Swan-Ganz
balloon flotation catheter was placed in the pulmonary artery. In
groups 1 and 2, the Swan-Ganz catheter was introduced in the
catheterization room, and in group 2 on the operating table,
following anesthesia. In the 16 patients of group 2 who were
studied at Johns Hopkins Medical Center, CO measures were
obtained by the 7010 Series Marquette (Marquette; Madison,
WI). In the remaining 35 patients of group 2 studied at Wolfson
Medical Center, and also in group 1 patients at the Tel Aviv
Medical Center (Israel), the Horizon 2000 (Mennen Medical;
Rohovot, Israel) was used. In group 3, the CO was measured by
Marquette 8000 Clinical Information Center 419897-015 (Mar-
quette).
A volume of 10 mL of 5% dextrose at room temperature was
injected in all patients via the proximal port. Temperature
changes were measured via the distal port located in the pulmo-
nary artery, ascertained by fluoroscopy, oxygen saturation, and
wedge pressure measurements. In all patients, three CI measure-
ments were obtained; when a 15% disparity occurred between
the two extreme measurements, two further injections, or more,
were administered until an average of three measurements within
the 15% range was obtained.
NICO WBEB Technology
When transmitting a small electrical current through the body,
an impedance to its transmission (restivity, R) is being measured.
Clinical Investigations in Critical Care
 This resistivity is called bioimpedance. According to Kirchovs
law, electric current passes through conduits of higher conduc-
tance (lowest resistivity). The resistivity of blood and plasma is
the lowest in the body (resistivity of blood is 150; plasma, 63;
cardiac muscle, 750; lungs, 1,275; and fat, 2,500 ohm/cm).15
Thus, when an alternating current of 20 to 100 kHz is applied to
the body, it is primarily distributed via the extracellular fluid and
the blood. The changes in the body resistivity (R) over time
(milliseconds) are therefore related to the dynamic changes of
the blood and plasma volume. This pertains particularly to the
passage of the stroke volume (SV) from the left ventricle into
the aorta and its branches. However, in the capillaries and in the
venous system the blood volume is relatively constant, because
the flow in these vessels is not pulsatile. Consequently, each
systolic increase in the aortic blood volume is associated with a
proportional increase in the measurable conductance of the
whole body (Fig 1).16 Thus, for measuring the aortic SV by means
of its impedance change, Frinerman and Tsoglin developed the
following algorithm14:
Hctcorr H2corr R
SV Kel Kweight IB placed at the root of the neck and another set around the lower
Ksex, age R
in which the R/R is corrected for hematocrit (Hctcorr), elec-
trolytes (Kel), body composition (K sex, age), weight (Kweight),
time characteristics ( systolic time, diastolic time), and
index balance (IB), which measures the body water composition.
To collect patient signals, the NICO uses proprietary elec-
trodes arranged in a wrist-to-ankle configuration (Fig 2); in
certain conditions when this particular form is not applicable (as
with severe peripheral edema or severe peripheral vascular
disease), a wrist-to-wrist configuration is used. The precise
positioning of the NICO electrodes is not critical; an untrained
operator can make the attachments. An alternating current of 30
kHz, 1.4 mA is delivered through the two electrodes, and the
bioimpedance and its fluctuations over time are measured. In
addition, a standard three-lead ECG connection is made for
measuring the pulse rate. The other variables required for SV and
CI calculation (age, gender, weight, height, hematocrit, electro-
lytes) are introduced into the machine only at the start of
monitoring. For measuring the CO, the SV is multiplied by the
heart rate.
Although the idea of the WBEB was conceived and tested in
the Soviet Union in 1941,17 it remained dormant until re-
cently. Meanwhile another bioimpedance approach for mea-
suring the CO was initially introduced by Kubicek et al18 in the
United States in 1966, and further refined in 1974.19 His
method is called thoracic electrical bioimpedance (TEB), and
the tools of this approach are commercially available. Noted
here are dissimilarities in the operation of the two technolo-
gies. In the NICO, one electrode is applied to the wrist and
the other to the ankle; in TEB, a number of electrodes are
part of the chest cage. In WBEB, the SV is measured by the
impedance variation (Z or R) induced by the systolic
volume ejected into the aorta. In the original TEB formula of
1966,18 the SV measurement was based on a similar principle.
However, since when the electrodes are placed on the chest,
the Z wave could be hardly detected, Kubicek et al19 adopted
another principle, whereby the SV measurement is based on
the depiction of the aortic systolic dp/dt (instantaneous pres-
sure change over time) for calculation of the systolic blood
flow into the great arteries.

Figure 1. Recordings of the interrelation between impedance variations and hemodynamic parame-
ters, according to Djordjevich and Sadove.16

www.chestjournal.org CHEST / 125 / 4 / APRIL, 2004 1433

Downloaded from chestjournals.org on April 4, 2007
120 of 158. Copyright 2004 by American College of Chest Physicians
 Figure 2. Wrist-to-ankle configuration of the electrodes in WBEB. I electric current; V electric
voltage.

Measuring CI by the NICO Results

For each CI determination, three NICO measurements were No significant differences between the means of
obtained. Since the CO results that are displayed on the scope are
updated every 20 s, for the determination of the CI an average
NICO CI and thermodilution CI in the three clinical
CO is derived from a 60-s monitoring. groups, as well as the whole cohort, were observed
(Table 1). A significant, high correlation was found
Statistical Methods between the NICO CI and the thermodilution CI
measurements: 0.886 in the whole cohort, and 0.881,
Agreement between the CI values of the NICO and thermodi- 0.902, and 0.851 in the catheterization, bypass sur-
lution was evaluated in three ways: the mean CI values of the
NICO and thermodilution were compared by a paired Student t
gery, and CHF groups, respectively. All correlation
test; correlation between these values was evaluated by calculat- coefficients were statistically significant (p 0.0001).
ing the Pearson correlation coefficient and by applying a linear The results of applying linear regression models to
regression model of the NICO on thermodilution; the differences the data (Table 2, Fig 3) demonstrate similar models
between the paired CI values of the NICO and thermodilution in the three clinical groups, as the intercepts and
were plotted against the average CI values of both methods,
instead of against thermodilution alone. This statistical method
slopes of the regression lines are not significantly
was recommended by Bland and Altman20 for evaluating a new different (intercepts, p 0.2398; slopes, p 0.2310).
device (NICaS) against an established method (thermodilution), Figure 4 shows differences between CI values plot-
which has its own inaccuracies. Bias was defined as the mean ted against the average value of the two methods
difference between the NICO CI and thermodilution CI values. with limits of agreement: two SDs from the mean
Limits of agreement (precision) were calculated as bias 2 SD
of the differences between the NICO CI and thermodilution CI
difference.
values. All three analyses were carried for the whole sample and Significant differences between the NICO CI and
for each specific clinical group: cardiac catheterization patients thermodilution CI were observed when comparing
(group 1), CABG patients (group 2), and CHF patients (group 3). the average CI of the four CI ranges (Table 3). When
We have examined the differences between the CI determi- CI is 2 L/min/m2, the thermodilution CI readouts
nation by the NICO and thermodilution at different CI readings
by classifying the readings into four subgroups according to the
mean CI levels: CI 1.5, CI 1.5 and 2, and CI 2 and 3,
and CI 3. CI readings by the NICO and thermodilution in each
group were compared using the paired Student t test and
Table 1Comparison Between the Mean CI Values of
presented as mean and SDs.
the NICO and Thermodilution in the Three Clinical
The sensitivity of the two methods (NICO and thermodilution)
Groups and in the Whole Cohort*
to a change in a specific medical condition was compared in the Thermodilution NICO CI,
CABG group at five operative and postoperative stages. A CI, L/min/m2 L/min/m2
variance analysis with repeated measures (type of method and
stage) was performed, followed by contrast analysis that com- Group No. Mean SD Mean SD p Value
pared successive stages. This analysis could be performed only for
Whole sample 418 2.39 0.70 2.38 0.73 NS
patients with complete data at all the five stages. A similar
Catheterization 40 2.81 0.72 2.81 0.68 NS
analysis was performed at seven time points in CHF patients who
CABG 208 2.33 0.72 2.31 0.77 NS
had been treated with an IV vasodilatory drug. All statistical
CHF 170 2.35 0.63 2.38 0.66 NS
analyses were performed using the SAS System for Windows
(version 8.01; SAS Institute; Cary, NC). *NS not significant.

1434 Clinical Investigations in Critical Care

Downloaded from chestjournals.org on April 4, 2007
121 of 158. Copyright 2004 by American College of Chest Physicians
 Table 2Linear Regression Analysis, Bias, and Precision in the Three Clinical Groups and in the Whole Cohort

Bias (Mean Between-

Method Difference), Precision (Mean SD),
Sample Intercept Slope R2 L/min L/min

Whole 0.18 0.92 0.79 0.0009 0.6849 0.6831

Catheterization 0.45 0.84 0.77 0.0040 0.7134 0.6393
CABG 0.08 0.96 0.81 0.0247 0.6789 0.7331
CHF 0.29 0.89 0.72 0.0271 0.684 0.692

are significantly higher than the NICO CI; when CI
is 3 L/min/ m2, the thermodilution CI is lower
than the NICO CI (Table 3). When CI was between
2 L/min/m2 and 3 L/min/m2, there was a slight
difference of only 3.28% between the two methods,
with a borderline significance (Table 3).
Twenty-five patients with CABG (subgroup A)
had complete information of the NICO CI and
thermodilution CI results at five operative and post-
operative stages (Table 4, Fig 5, top). A further 26
CABG patients (subgroup B) had incomplete data
( 5 paired measurements for each patient), yielding
additional 80 paired measurements (Table 4). There
were time-related changes in the CI, and the NICO
and the thermodilution followed these changes
by producing similar results at each time point
(p 0.0035 and p 0.0058, respectively; Fig 5,
top). A contrast analysis, performed to compare the
CO in successive stages of the measurements, found
that the difference between stage 3 and stage 4 was
statistically significant according to the two measure-
ments in subgroup A (NICO CI, p 0.0135; ther-
modilution CI, p 0.002; Fig 5, top). In 17 of the
CHF patients in group 3 who were treated with
an IV vasodilator agent, and in whom the CI was
measured simultaneously at seven time points during
the treatment, the total time trend was significant in
the NICO (p 0.0056) but not significant in ther-
modilution (Fig 5, bottom).
Discussion
In recent years it has been suggested that CO and
MAP measurement and the calculation of CI, Cpi,
and SVRi might be instrumental in the diagnosis,
treatment titration, and risk stratification of cardiac
patients, especially those with CHF.710 However,

Figure 3. Plot of CI values measured by the NICO and thermodilution.

www.chestjournal.org CHEST / 125 / 4 / APRIL, 2004 1435

Downloaded from chestjournals.org on April 4, 2007
122 of 158. Copyright 2004 by American College of Chest Physicians
 Figure 4. Differences between CI values (NICO vs thermodilution) plotted against the average values
of the two methods.

CI has been measured only during invasive right-
heart catheterization, which requires intensive care
admission and may be associated with complica-
tions.2123 Hence, CI was measured only rarely, and
in the sickest patients. Therefore, a simple, reliable,
noninvasive, and continuous method for CI mea-
surement has become necessary in order to enable
its application to cardiac patients with different
degrees of medical severity and in diverse settings.
Currently there are four accepted methods for
noninvasive CO measurement. The Doppler echo-
cardiogram obtained from the left ventricular out-
flow track and CO2 rebreathing techniques have
been shown to be accurate in measuring CI. But
these methods are limited by the requirement for
expensive equipment and specialized personnel for
their application and therefore are not simple to use,
and moreover do not enable continuous measure-
ments. Thoracic bioimpedance has been used in the
last decade for continuous CO measurement. Judg-
ing by the literature,24 26 as long as the heart func-
tion is intact, TEB can be useful for monitoring the
hemodynamic state in various clinical conditions
such as trauma, massive surgery, sepsis, etc.24 But
when it comes to monitoring and managing patho-
logic cardiac conditions TEB requires further im-
provement.24 26
Thus far, eight groups of patients who submitted
to CO measurements by WBEB have been reported
in six published articles. Kedrov,17 who was the first,
compared the average CI measured by the WBEB
in 57 subjects with normal hearts in published results
of the Fick method, revealing 3.3 28% vs 3.31
L/min/m2 (range, 2.4 to 4.2 L/min/m2), respectively.
Tischenko27 compared the CI results measured by
WBEB in three groups of subjects with normal
hearts vs three standard methods. There were 31
cases vs acetylen (r 0.84), 28 cases vs thermodilu-
tion (r 0.95), and 28 cases vs Fick (r 0.99).
Using a modified Tischenko algorithm vs thermodi-

Table 3Differences Between the NICO CI and Thermodilution CI Within the CI Ranges

Results, NICO CI, Thermodilution Relative

CI Ranges by NICO CI No. Mean CI, Mean Difference, % Significance

1.5 30 1.278 0.16 1.515 0.35 12.99 0.0002

1.5 CI 2 98 1.749 0.14 1.876 0.33 4.65 0.0001
2 CI 3 220 2.433 0.28 2.392 0.40 3.28 0.0484
CI 3 70 3.594 0.57 3.449 0.64 5.44 0.0045

1436 Clinical Investigations in Critical Care

Downloaded from chestjournals.org on April 4, 2007
123 of 158. Copyright 2004 by American College of Chest Physicians
 Table 4 Trend Follow-up of the NICO CI and Thermodilution CI During the Five Operative and
Postoperative Stages
NICO CI
Subgroup A Subgroups A B
(n 25) (n 51)
Method Mean SD Mean
1 2.28 0.87 2.25
2 2.09 0.58 2.09
3 2.17 0.62 2.26
4 2.72 0.79 2.54
5 2.56 0.74 2.46
lution, Koobi et al28 obtained simultaneous measure-
ments in 74 patients with coronary disease, reaching
a bias between the two methods of 0.25 0.8 L/min
(SD), where the limits of agreement (2 SD) were
1.37 L/min and 1.897 L/min, respectively. Using
the NICaS apparatus, Cohen et al14 compared its
performance against thermodilution by measuring
the CO in patients undergoing CABG operations,
with a correlation of r 0.91. Thus, this six-clinical
series, which included 274 subjects, revealed similar
correlation coefficients between the compared
methods, just as in the present series. Moreover, in
none of these publications did the authors express
reservations about the performance of the WBEB.
There were, nonetheless, two publications in
which no correlation was found between WBEB and
thermodilution; in both instances, the underlying
clinical conditions are listed in the exclusion criteria
of the NICO. Lamberts et al29 compared the original
Tischenko equation with dye dilution CO in 10
patients, 4 of whom had significant aortic regurgita-
tion and 1 had coarctation of the aorta. The NICO
apparatus cannot measure the CO in such condi-
tions.
Imhoff et al30 compared the NICO apparatus
against thermodilution in 22 postpancreatectomy or
esophagectomy patients. They were all managed
postoperatively by Swan-Ganz catheters for boosting
the oxygen delivery to 600 mL/min/m2 and the CI to
4.5 L/min/m2. Hence, in these patients the radical
surgical procedures were followed by massive inter-
compartmental volume shifts due to IV administra-
tion of up to 6 L per 24 h of crystalloids and plasma,
often accompanied by massive peripheral edema. In
such hemodynamic situations, the baseline imped-
ance should properly become distorted, preventing
an accurate measurement of the systolic impedance
variation.
In the present study, the agreement between
NICO CI and thermodilution CI as tested by com-
parisons of the means is highly significant. The
similarity between the means in the entire cohort as
www.chestjournal.org
Downloaded from chestjournals.org on April 4, 2007
124 of 158. Copyright 2004 by American College of Chest Physicians
Thermodilution CI
Subgroup B, Subgroups A B,
(n 26) (n 51)
SD Mean SD Mean SD
0.83 2.22 0.69 2.22 0.67
0.57 2.14 0.57 2.14 0.52
0.69 2.15 0.68 2.22 0.70
0.81 2.80 0.74 2.61 0.77
0.85 2.61 0.83 2.53 0.87
well as in each diagnostic group, together with the
relatively large sample size, further endorses the
significance of the results.
The mean difference between the NICO and
thermodilution in the entire sample range was
0.0009 L/min (Table 2, Fig 4), ranging from 0.0040
to 0.0271 L/min in the three diagnostic groups. This
disparity is smaller than the level of accuracy of
thermodilution, which is defined to a 15% range.31
Linear regression applied to the data revealed that
the line slope was close to 1.00 in the entire sample
range and in each specific diagnostic group. There
were no significant differences between the slopes
and the intercepts of the three diagnostic groups.
This indicates that the relation between the NICO
and thermodilution is similar in all diagnostic groups.
Following the recommendations of Bland and Alt-
man,20 the differences between the two measure-
ments were plotted against their means. This plot
demonstrates that the range of differences is similar
along the different values of the average.
Although the main purpose of this work was to
compare the performance of the NICO vs thermo-
dilution, following the suggestion in previous stud-
ies3135 that thermodilution tends to overestimate CI
when low and underestimate it when high, we have
compared the NICO CI and thermodilution CI in
the different CI ranges. The results of this analysis
have shown that when the CI was 2 L/min/m2, the
thermodilution results were higher than the NICO
results; when the CI was 3 L/min/m2, the ther-
modilution results were slightly lower than the
NICO results (Table 3). As a consequence, the
differences of the hemodynamic responses to vaso-
dilation therapy may be better depicted by the
NICO when compared with thermodilution (Fig 5,
bottom).
The link between a low CI and a higher thermo-
dilution readout, and between a higher CI with a
low thermodilution readout, is also expressed in the
stability of the range of differences along the differ-
ent values of the average (Fig 4). Furthermore, the
CHEST / 125 / 4 / APRIL, 2004 1437
 Figure 5. Top: Comparisons between the NICO CI and thermodilution CI at five operative and
postoperative stages. Bottom: NICO CI and thermodilution CI responses to vasodilating titration
therapy in patients with acute CHF.

almost identical results of thermodilution and the Limitations of the NICO

NICO observed in the CI 2 to 3 range (Table 3)
yield a mutual endorsement of the technical opera-
tions of the measuring devices by the two methods.
It should also be noted that the two methods switch
their interrelations at CI of approximately 2.5
L/min/m2 (Fig 5, bottom), which is close to the lower
border.
Diseases of the Aorta and Aortic Valve: The NICO
measures the SV of the aorta and its derivatives
(including the peripheral arteries). For determining
the CO, SV is multiplied by the heart rate. Any
aberration in the hemodynamics of the aorta and its
derivatives may interfere with the proper function of

1438 Clinical Investigations in Critical Care

Downloaded from chestjournals.org on April 4, 2007
125 of 158. Copyright 2004 by American College of Chest Physicians
 the NICO. In aortic insufficiency, the SV is patho-
logically increased. Coarctation and significant aneu-
rysms of the aorta and also severe peripheral arterial
disease may produce inaccurate readouts of the SV.
Significant (3/4) Peripheral Edema: In signif-
icant edema, the baseline impedance of the body is
occasionally very low, currently interfering with the
accuracy of the results.
Shunts: Our experience with congenital heart
disease is limited, but we assume that since there is
no complete separation between the two circulations
the impedance technique may not be appropriate.
Restlessness: The impedance test requires a mo-
tionless condition; in addition, restlessness is associ-
ated with fluctuating activation of the sympathetic
system, resulting in an unstable hemodynamic state.
Arrhythmias: In most cases, there is no interfer-
ence with the CO measurement, although when
associated with a heart rate 150 beats/min or when
there is a severe disarray of the complexes (ECG and
SV), the results may become inaccurate, as may be
for any technique measuring CO.
Resections: Major abdominal and thoracic surgical
resections, especially those that include major rapid
shifts in fluid distribution between the intravascular
and extravascular space.
Conclusions
In spite of these limitations, the NICO apparatus
offers a simple, noninvasive, reliable, and continuous
measurement of CI in cardiac patients with particu-
lar emphasis on CHF. This measurement combined
with MAP measurement and the calculation of Cpi
and SVRi is destined to become a safe, simple, rapid,
noninvasive method for evaluating and treating pri-
marily coronary patients sustaining chronic and acute
exacerbations of CHF.
References
1 Cotter G, Williams SG, Vered Z, et al. The role of cardiac
power in heart failure. Curr Opin Cardiol 2003; 18:215222
2 Cotter G, Fincke R, Lowe A, et al. Hemodynamic parameters
in cardiogenic shock due to myocardial infarction: a report
from the SHOCK trial registry [abstract]. Circulation 2003;
108:539
3 Cotter G, Moshkovitz Y, Milovanov O, et al. Acute heart
failure: a novel approach to its pathogenesis and treatment.
Eur J Heart Fail 2002; 4:227234
4 Cotter G, Moshkovitz Y, Kaluski E, et al. The role of cardiac
power and systemic vascular resistance in the pathophysiology
and diagnosis of patients with acute congestive heart failure.
Eur J Heart Fail 2003; 5:443 451
5 Marmor A, Schneeweiss A. Prognostic value of noninvasively
obtained left ventricular contractile reserve in patients with
severe heart failure. J Am Coll Cardiol 1997; 29:422 428
6 Williams SG, Cooke GA, Wright DJ, et al. Peak exercise
www.chestjournal.org
Downloaded from chestjournals.org on April 4, 2007
126 of 158. Copyright 2004 by American College of Chest Physicians
cardiac power output: a direct indicator of cardiac function
strongly predictive of prognosis in chronic heart failure. Eur
Heart J 2001; 22:1496 1503
7 Cohen-Solal A, Tabet JY, Logeart D, et al. A non-invasively
determined surrogate of cardiac power (circulatory power)
at peak exercise is a powerful prognostic factor in chronic
heart failure. Eur Heart J 2002; 23:806 814
8 Mohr R, Rath S, Meir O, et al. Changes in systemic vascular
resistance detected by the arterial resistometer: preliminary
report of a new method tested during percutaneous translu-
minal coronary angioplasty. Circulation 1986; 74:780 785
9 Torre-Amione G, Young JB, Colucci ES, et al. Hemodynamic
and clinical effects of tezosentan, an intravenous dual endo-
thelin receptor antagonist, in patients hospitalized for acute
decompensated heart failure J Am Coll Cardiol 2003; 42:140
147
10 Teerlink JR, Massie BM, Cleland JG, et al, for the Ritz 1
Investigators. A double-blind, parallel-group, multi-center,
placebo-controlled study to investigate the efficacy and safety
of tezosentan in reducing symptoms in patients with acute
decompensated heart failure (RITZ 1) [abstract]. Circulation
2001; 104:II-526
11 VMAC Investigators. Intravenous nesiritide vs nitroglycerin
for treatment of decompensated congestive heart failure: a
randomized trial. JAMA 2002; 287:15311540
12 Cotter G, Faibel H, Barash P, et al. High-dose nitrates in the
immediate management of unstable angina: optimal dosage,
route of administration, and therapeutic goals. Am J Emerg
Med 1998; 16:219 224
13 Kaluski E, Kobrin I, Zimlichman R, et al. RITZ-5: random-
ized intravenous tezosentan (an endothelin ET-A/B antago-
nist) for the treatment of pulmonary edema: a prospective
randomized, multicenter, double-blind placebo controlled
study. J Am Coll Cardiol 2003; 41:204 210
14 Cohen AJ, Arnaudov D, Zabeeda D, et al. Non-invasive
measurement of cardiac output during coronary artery bypass
grafting. Eur J Cardiothorac Surg 1998; 14:64 69
15 Baker LE. Principles of impedance technique. IEEE Eng
Med Biol 1989; 3:1115
16 Djordjevich L, Sadove MS. Basic principles of electrohaemo-
dynamics. J Biomed Eng 1981; 3:2533
17 Kedrov AA. An attempt of the quantify assessment of the
central and peripheral circulation by electrometrical method.
Klin Med 1948; 26:3251
18 Kubicek WG, Karnegis JN, Patterson RP, et al. Development
and evaluation of an impedance cardiac output system.
Aerosp Med 1966; 37:1208 1212
19 Kubicek WG, Kottke FJ, Ramos MU, et al. The Minnesota
impedance cardiograph: theory and applications. Biomed
Eng 1974; 9:410 416
20 Bland JM, Altman DG. Statistical methods for assessing
agreement between two methods of clinical measurement.
Lancet 1986; I:307310
21 Robin ED, McCauley RF. Monitor wizards can be dangerous.
Chest 1998; 114:11511153
22 Dalen JE, Bone RC. Is it time to pull the pulmonary artery
catheter? JAMA 1996; 276:916 918
23 Connors AF, Speroff T, Dawson NV, et al, for the SUPPORT
investigators. The effectiveness of right heart catheterization
in the initial care of critically ill patients. JAMA 1996;
276:889 897
24 Raaijmakers E, Faes ThJC, Scholten RJPM, et al. A meta-
analysis of published studies concerning the validity of tho-
racic impedance cardiography. Ann NY Acad Sci 1999;
873:121134
CHEST / 125 / 4 / APRIL, 2004 1439
 25 Patterson RP, Witsoe DA, From A. Impedance stroke volume
compared with dye and electromagnetic flowmeter values
during drug-induced inotropic and vascular changes in dogs.
Ann N Y Acad Sci 1999; 873:143148
26 Marik PE, Pendelton JE, Smith R. A comparison of hemo-
dynamic parameters derived from transthoracic electrical
bioimpedance with those parameters obtained by thermodi-
lution and ventricular angiography. Crit Care Med 1997;
25:15451550
27 Tischenko MI. Estimation of stroke volume by integral
rheogram of the human body [in Russian]. Sechenov Phisio-
logical J 1973; 59:1216 1224
28 Koobi T, Kaukinen S, Turjanmaa VM, et al. Whole-body
impedance cardiography in the measurement of cardiac
output. Crit Care Med 1997; 25:779 785
29 Lamberts R, Visser KR, Zijlstra WG. Impedance cardiogra-
phy. Assen, The Netherlands: Van Gorkum, 1984; 2194
30 Imhoff M, Lehner JH, Lohelin D. Noninvasive whole-body
electrical bioimpedance cardiac output and invasive ther-
1440
Downloaded from chestjournals.org on April 4, 2007
127 of 158. Copyright 2004 by American College of Chest Physicians
modilution cardiac output in high-risk surgical patients. Crit
Care Med 2000; 28:28122818
31 Davidson CJ, Fishman RF, Bonow RD. Cardiac catheteriza-
tion. In: Branwald E, ed. Heart disease: a textbook of
cardiovascular medicine. 5th ed. Philadelphia, PA: W.B.
Saunders Company, 1997; 192
32 van Grondelle A, Ditchey RV, Groves BM, et al. Thermodi-
lution method overestimates low cardiac output in humans.
Am J Physiol 1983; 245:690 692
33 Kohanna FH, Cunningham JN. Monitoring of cardiac output
by thermodilution after open-heart surgery. J Thorac Cardio-
vasc Surg 1977; 73:451 457
34 Espersen K, Jensen EW, Rosenborg D, et al. Comparison of
cardiac output measurement techniques: thermodilution,
Doppler, CO2-rebreathing and direct Fick method. Acta
Anaesthesiol Scand 1995; 39:245251
35 Nanas JN, Anastasiou-Nana MI, Sutton RB, et al. Compari-
son of Fick and dye cardiac outputs during rest and exercise
in 1,022 patients. Can J Cardiol 1986; 2:195199
Clinical Investigations in Critical Care
 Accurate, Noninvasive Continuous Monitoring of Cardiac Output by
Whole-Body Electrical Bioimpedance
Gad Cotter, Yaron Moshkovitz, Edo Kaluski, Amram J. Cohen, Hilton Miller,
Daniel Goor and Zvi Vered
Chest 2004;125;1431-1440
DOI 10.1378/chest.125.4.1431
This information is current as of April 4, 2007

Updated Information Updated information and services, including

& Services high-resolution figures, can be found at:
http://chestjournals.org/cgi/content/full/125/4/1431
References This article cites 31 articles, 10 of which you can access
for free at:
http://chestjournals.org/cgi/content/full/125/4/1431#BIBL
Citations This article has been cited by 1 HighWire-hosted
articles:
http://chestjournals.org/cgi/content/full/125/4/1431
Permissions & Licensing Information about reproducing this article in parts
(figures, tables) or in its entirety can be found online at:
http://chestjournals.org/misc/reprints.shtml
Reprints Information about ordering reprints can be found online:
http://chestjournals.org/misc/reprints.shtml
Email alerting service Receive free email alerts when new articles cite this
article sign up in the box at the top right corner of the
online article.
Images in PowerPoint format Figures that appear in CHEST articles can be
downloaded for teaching purposes in PowerPoint slide
format. See any online article figure for directions.

Downloaded from chestjournals.org on April 4, 2007

128 of 158. Copyright 2004 by American College of Chest Physicians
 The European Journal of Heart Failure 5 (2003) 443451

The role of cardiac power and systemic vascular resistance in the

pathophysiology and diagnosis of patients with acute congestive heart
failure
Gad Cottera,*, Yaron Moshkovitzb, Edo Kaluskia, Olga Miloa, Ylia Nobikovc, Adam Schneeweissa,
Ricardo Krakovera, Zvi Vereda
a
Cardiology Department, Assaf-Harofeh Medical Center, 70300 Zerifin, Israel
b
Cardiac Surgery Department, Ramat-Marpe Hospital, Petah-Tikva, Israel
c
Biostatistical Department, Sheba Medical Center, Sackler School of Medicine, Tel-Aviv University, Tel-Hashomer, Israel

Received 16 August 2002; received in revised form 6 January 2003; accepted 21 January 2003

Downloaded from eurjhf.oxfordjournals.org by guest on June 14, 2011

Abstract

Objective: Conventional hemodynamic indexes (cardiac index (CI), and pulmonary capillary wedge pressure) are of limited
value in the diagnosis and treatment of patients with acute congestive heart failure (CHF). Patients and methods: We measured
CI, wedge pressure, right atrial pressure (RAP) and mean arterial blood pressure (MAP) in 89 consecutive patients admitted due
to acute CHF (exacerbated systolic CHF, ns56; hypertensive crisis, ns5; pulmonary edema, ns11; and cardiogenic shock, ns
17) and in two control groups. The two control groups were 11 patients with septic shock and 20 healthy volunteers. Systemic
vascular resistance index (SVRi) was calculated as SVRis(MAPyRAP)yCI. Cardiac contractility was estimated by the cardiac
power index (Cpi), calculated as CI=MAP. Results and discussion: We found that CI-2.7 lyminym2 and wedge pressure )12
mmHg are found consistently in patients with acute CHF. However, these measures often overlapped in patients with different
acute CHF syndromes, while Cpi and SVRi permitted more accurate differentiation. Cpi was low in patients with exacerbated
systolic CHF and extremely low in patients with cardiogenic shock, while SVRi was increased in patients with exacerbated
systolic CHF and extremely high in patients with pulmonary edema. By using a two-dimensional presentation of Cpi vs. SVRi
we found that these clinical syndromes can be accurately characterized hemodynamically. The paired measurements of each
clinical group segregated into a specific region on the CpiySVRi diagnostic graph, that could be mathematically defined by a
statistically significant line (Lambdas0.95). Therefore, measurement of SVRi and Cpi and their two-dimensional graphic
representation enables accurate hemodynamic diagnosis and follow-up of individual patients with acute CHF.
2003 European Society of Cardiology. Published by Elsevier B.V. All rights reserved.

Keywords: Cardiac power; Vascular resistance; Acute congestive heart failure

1. Introduction these clinical syndromes of acute CHF may be difficult,

Acute congestive heart failure (CHF) is a common
disease, accounting for over 700 000 annual admissions
to hospitals in the USA alone. We have recently sug-
gested that this disease can be divided into four major
clinical syndromes w1x: (1) pulmonary edema, (2) car-
diogenic shock, (3) hypertensive (HTN) crisis and (4)
exacerbated systolic CHF. However, the diagnosis of
*Corresponding author. Tel.: q972-8-977-9778; fax: q972-8-977-
9779.
E-mail address: cotterg@hotmail.com (G. Cotter).
due to an overlap in symptoms and signs among the
different syndromes as well as lack of objective criteria
for their diagnosis. For example, both cardiogenic shock
and pulmonary edema patients present with severe cir-
culatory and respiratory distress and in both cases CI is
low and wedge pressure is high. However, these two
clinical syndromes have a completely different course
(mortality rates at 1-month in the SHOCK study w2x
were approximately 60%, compared with a mortality
rate of approximately 10% for pulmonary edema patients
during the first 30 days in the RITZ-5 study w3x). In
addition, the pathophysiology of the two clinical syn-

1388-9842/03/$ - see front matter 2003 European Society of Cardiology. Published by Elsevier B.V. All rights reserved.
PII: S 1 3 8 8 - 9 8 4 2 0 3 . 0 0 1 0 0 - 4
129 of 158.
 444 G. Cotter et al. / The European Journal of Heart Failure 5 (2003) 443451
dromes is completely different and their treatment is
almost opposite. Moreover, even the diagnosis of acute
CHF is sometimes difficult, due to an overlap in signs
and symptoms with those of acute exacerbations of
obstructive or restrictive lung diseases and the occasion-
al difficulty in differentiation between septic and cardi-
ogenic shock.
Measurement of invasive hemodynamic variables,
including CI and pulmonary capillary wedge pressure
has been used in patients with acute decompensated and
chronic compensated CHF, as well as cardiogenic shock,
for more than two decades. However, despite extensive
experience and numerous studies, no specific diagnostic
criteria or accurate cut-off points have been determined
w2,46x. In some studies, trends and changes in CI and
wedge pressure have been used, however, no reproduc-
ible criteria for diagnosis and follow-up could be
established.
Cardiac power, an index of cardiac contractility, is
calculated based on the classical physical rule of fluids,
i.e. powersflow=pressure, hence cardiac power index
(Cpi) is the product of simultaneously measured mean
arterial blood pressure (MAP) and cardiovascular flow
(CI): CpisMAP=CI=0.0022 w1,79x. The units are
Wym2. Cpi has been used extensively during recent
years to evaluate patients with chronic and acute CHF.
In three separate studies, Marmor and Schneeweiss w7x,
Tan et al. w8x and Cohen-Solal et al. w9x have demon-
strated that Cpi increases during exercise (cardiac power
reserve) and is the strongest predictor of outcome in
patients with chronic CHF, stronger than O2 consumption
and echocardiographic ejection fraction. We have pre-
viously demonstrated w1x that in patients with exacer-
bated systolic CHF, baseline Cpi at admission is the
strongest predictor of short- and long-term outcome. On
the other hand, the main event preceding recurrent
worsening heart failure was a steep increase in SVRi.
In a recent analysis, we have also found that Cpi at
baseline and during follow-up was the strongest predic-
tor of outcome in a large cohort of cardiogenic shock
patients (unpublished data).
The main hypothesis of the present study was that in
patients with acute CHF, as Cpi decreases, SVRi should
concomitantly increase. Therefore, for each Cpi decrease
the SVRi increase may be adequate, too high or too low
and, thus, CpiySVRi coupling may characterize the
clinical-hemodynamic state. Therefore, in the present
study, we examined in a two-dimensional representation,
the relationship between changes in Cpi (pump work)
and SVRi (resistance or work load) in the four clinical
syndromes of acute CHF, (i.e. exacerbated systolic CHF,
pulmonary edema, cardiogenic shock and HTN crisis),
as well as in two control groups: (i.e. septic shock and
normal subjects).
130 of 158.
2. Methods
2.1. Inclusion criteria
Hemodynamic data was obtained in all patients under-
going right heart catheterization who were diagnosed by
the usual clinical criteria as having acute CHF. We also
enrolled two control groups: these were 11 patients with
septic shock and 20 healthy volunteers.
2.2. Exclusion criteria
Significant valvular disease, significant brady- or
tachy-arrhythmias or renal failure (creatinine )2.5 mgy
dl).
2.3. Clinical diagnosis criteria
2.3.1. Exacerbated systolic CHF
Patients admitted due to signs and symptoms of
worsening CHF, who were in a stable clinical condition;
Downloaded from eurjhf.oxfordjournals.org by guest on June 14, 2011
not fulfilling the criteria for cardiogenic shock, pulmo-
nary edema and HTN crisis and who had EF-35% on
echocardiography. (The echocardiographic criteria were
used to ensure that the symptoms of dyspnea were
indeed due to acute CHF.)
2.3.2. Pulmonary edema
Patients admitted due to clinical symptoms and signs
of acute pulmonary congestion accompanied by findings
of lung edema on chest X-ray who had severe respiratory
distress accompanied by O2 saturation -90% in room
air by pulse oxymetery during the invasive
measurements.
2.3.3. Cardiogenic shock
Systolic blood pressure -100 mmHg for at least 1 h,
not responsive to percutaneous revascularization,
mechanical ventilation, intra-aortic balloon-pump
(IABP), IV fluid administration and dopamine of at
least 10 mgykgymin and accompanied by signs of end
organ hypoperfusion but not accompanied by fever
)388 or a systemic inflammatory syndrome.
2.3.4. HTN crisis
Patients with signs and symptoms of acute CHF
accompanied by high blood pressure (MAP)130
mmHg during invasive measurements); not fulfilling the
criteria for pulmonary edema.
2.3.5. Septic shock
Systolic blood pressure -100 mmHg accompanied
by fever )388, systemic inflammatory syndrome and
signs of end organ hypoperfusion for at least 3 h not
responsive to IV fluids and IV dopamine of at least 10
mgykgymin. No evidence of an acute cardiac event.
 G. Cotter et al. / The European Journal of Heart Failure 5 (2003) 443451
2.4. Assessment of hemodynamic variables
Prior to enrolment in this study all patients gave
written informed consent. The study protocol was
approved by the local ethics review board. In all patients
the hemodynamic variables were obtained during right
heart catheterization using a SwanGanz catheter placed
under fluoroscopic guidance. All measurements were
obtained while patients were at least 30 s without IABP
while on the same treatment used at the time the clinical
diagnosis was made. Measurement of hemodynamic
variables was performed at least 6 h after the last intake
of an oral drug and 2 h after intravenous drug therapy.
CI was measured by thermodilution, using the mean
of at least three consecutive measurements within a
range of -15%. In normal subjects, right heart cathe-
terization was not performed for ethical reasons. The
values used in this cohort were obtained by standard
noninvasive cuff blood pressure measurement and eval-
uation of CI by the FDA approved NICaS 2001, a
noninvasive continuous cardiac output monitor w10x.
Therefore, wedge pressure was not assessed in normal
subjects. Instead, we used standard values documented
in the literature w11x.
2.5. Calculation of hemodynamic variables
Cpi was determined as MAP=CI=0.0022 and SVRi
was determined as (MAPyright atrial pressure (RAP)y
CI. As RAP was not measured in normal subjects, it
was estimated to be 10% of MAP w11x.
2.6. Echocardiographic evaluation
All patients underwent routine echocardiographic
evaluation after initial stabilization. This included visual
estimation of cardiac function, evaluation of valvular
function and gross estimation of signs of diastolic
dysfunction.
2.7. Statistical methods
The five clinical groups were compared with regard
to all parameters using a one-way analysis of variance
(ANOVA). Therafter, the RyanEinotGabrielWelsch
Multiple Range Test was used for pair-wise comparisons
between the groups, while Dunnetts t-test was used to
compare all groups to the healthy controls.
A one-sample t-test was performed to compare mean
wedge pressure in each group to the wedge pressure of
normal people (-12 mmHg).
In order to determine the usefulness of the hemodyn-
amic parameters to discriminate between the clinical
syndromes, ROC curves, derived from a logistic regres-
sion model were applied to the data to determine the
131 of 158.
445
best cut-off point for the various parameters, in terms
of highest sensitivity and specificity.
2.8. CpiySVRi diagnostic graph
A classification rule was developed using second
order discriminant analysis. The normality of the distri-
bution of Cpi and SVRi was examined by the Wilk
Shapiro test. Due to the skewness of the data in some
groups, both variables (CPi and SVRi) were transformed
into log scale for better approximation to normality.
Since the number of patients with HT crisis was small
they were considered together with the exacerbated
systolic CHF group. The classification used two steps.
In the first step the rule separated three classes: septic
shock, cardiogenic shock and a combined group, which
included the normal controls, compensated CHF and
pulmonary edema patients (NCP). If, after the first
step the patient was defined as NCP, the second
classification was used to differentiate between the
normal, exacerbated systolic CHF and pulmonary edema
Downloaded from eurjhf.oxfordjournals.org by guest on June 14, 2011
subgroups.
All calculations were performed by SAS 6.12 (SAS
Institute Inc., Cary, NC) using procedures FREQ,
MEANS, GLM, DISCRIM, GPLOT.
Alfa level: 5%.
3. Results
Eighty-nine consecutive patients admitted due to acute
CHF and LV dysfunction (exacerbated systolic CHF,
ns56; pulmonary edema, ns11; cardiogenic shock,
ns17; and HTN crisis, ns5) as well as 11 patients
with septic shock and 20 healthy volunteers were
enrolled in the study. Baseline characteristics of the
different patient groups are presented in Table 1. The
mean CI, wedge pressure, MAP, SVRi and Cpi according
to clinical diagnosis are presented in Table 2.
3.1. Hemodynamic variables
3.1.1. Cardiac index (Fig. 1)
The average values of CI were significantly lower in
patients with acute CHF and higher in patients with
septic shock. ROC analysis found a cut-off point of
CI-2.7 lyminym2 useful for the determination that a
patient had acute CHF (sensitivitys1, specificitys
0.99). However, values between 1.2 and 2.7 lyminym2
could be found in all patients with exacerbated systolic
CHF and HTN crisis as well as 73% of patients with
pulmonary edema and 47% of patients with cardiogenic
shock. Moreover, the mean CI of patients with pulmo-
nary edema and cardiogenic shock was found to be
almost identical (1.4"0.4 vs. 1.35"0.7 lyminym2, Ps
ns).
 446 G. Cotter et al. / The European Journal of Heart Failure 5 (2003) 443451

Table 1
Baseline characteristics of the five patient groups

Normal Septic Exacerbated Pulmonary Cardiogenic

volunters shock systolic CHF edema shock
Sex (male:female) 12:8 7:4 51:10 5:6 11:6
Age (years) 60"8 55"11 69"10 73"12 67"11
Weight (kg) 79"14 77"10 72"8 70"9 80"14
Body surface area (m2) 1.92"0.22 1.91"0.23 1.88"0.21 1.81"0.24 1.92"0.24
IHD (%) 0 18 79 73 100
Previous MI 0 9 62 55 18
EF (%) 55"3 46"9 27"5 41"10 24"6
Diabetes mellitus (%) 20 73 66 66 53
Current smokers (%) 60 55 36 44 53
Hypertension (%) 50 45 56 88 71
Hyperlipidemia (%) 65 73 66 66 53
Baseline creatinine 135"75 124"55 144"81 110"47
Medications for CHF
Diuretic (%) 0 0 82 91 6
Digoxin (%) 0 0 33 45 0
ACE inhibitoryAII blocker (%) 0 9 95 91 29
Beta-blocker (%) 0 9 62 55 24
Nitrate (%) 0 0 41 55 0
IHD, ischemic heart disease; MI, myocardial infarction; EF, ejection fraction; CHF, congestive heart failure.

Downloaded from eurjhf.oxfordjournals.org by guest on June 14, 2011

3.1.2. Mean arterial blood pressure
By virtue of their clinical definition, the average
values of MAP were higher in patients with HTN crisis
and lower in those with septic and cardiogenic shock.
However, large areas of overlap were found between
pulmonary edema, HTN crisis and exacerbated systolic
CHF (MAP)100 mmHg) and between exacerbated
systolic CHF, cardiogenic shock and septic shock
(MAP-100 mmHg).
3.1.3. Pulmonary capillary wedge pressure (Fig. 2)
The mean wedge pressure was significantly higher in
patients with acute CHF and lower in patients with
septic shock. The analysis was based on normal values
reported in the literature (-12 mmHg w8x) (Ps0.001).
However, the overlap of wedge pressure in the different
acute CHF groups was extensive. Values between 12
and 38 mmHg were found in 82, 64, 76 and 18% of
patients with exacerbated systolic CHF, pulmonary ede-
ma, cardiogenic shock and septic shock, respectively.
3.1.4. Cardiac power index (Fig. 3)
Compared with the normal controls, the mean values
of Cpi were low in patients with exacerbated systolic
CHF and pulmonary edema and extremely low in

Table 2
Baseline distribution of the various hemodynamic parameters in the six diagnosis groups presented as means and S.D.

Exacerbated Pulmonary Cardiogenic HTN Septic Normal

systolic CHF edema shock crisis shock
N 56 11 17 5 11 20
SVRi 44.9"8.0 88.2"16.7 55.6"31.1 54.3"3.2 11.8"1.1 25.2"4.1
Cpi 0.47"0.13 0.4"0.13 0.22"0.08 0.75"0.04 0.8"0.13 0.62"0.08
Wedge 25.5"7.2 32.7"8.6 23.3"6.5 28.5"4.5 11.4"7.7
MAP 101"18 131.4"12.7 72.2"11.3 150"10.5 68.2"5.4 87.9"8.85
CI 2.06"0.33 1.37"0.32 1.42"0.64 2.24"0.37 5.2"0.5 3.2"0.36
The results of the ANOVA for comparisons between exacerbated systolic CHF (CHF), pulmonary edema
(edema) and cardiogenic shock (shock) patients
Parameter P-value for overall group P-value for paired comparisons (RyanEinotGabri-
groups comparison (ANOVA) elWelsch multiple range test)
CHF-edema CHF-shock Edema-shock
CI 0.0001 q q
Wedge 0.0037 q q
SVRi 0.0001 q q q
CPi 0.0001 q q
N, number of patients; SVRi, systemic vascular resistance index (wood m2 ); Cpi, cardiac power index (mmHg lyminym2 ); wedge, pulmonary
capillary wedge pressure (mmHg); MAP, mean arterial blood pressure (mmHg); CI: cardiac index (lyminym2 ). q, Significant group difference.

132 of 158.
 G. Cotter et al. / The European Journal of Heart Failure 5 (2003) 443451
Fig. 1. CI (lyminym2) Box-plots (median and 2575% percentile
range) in patients with the different syndromes of acute CHF and
patients with septic shock and normal controls.
patients with cardiogenic shock. However, some overlap
was encountered among the five groups.
3.1.5. Systemic vascular resistance index (Fig. 4)
Average values of SVRi were significantly higher in
all patients with exacerbated systolic CHF, HTN crisis
and extremely high in patients with pulmonary edema,
but were lower in patients with septic shock. SVRi was
found to be instrumental in the diagnosis of pulmonary
edema. All patients with this clinical syndrome had
SVRi)67 wood m2, while SVRi values in all other
patient groups, as well as in normal subjects, were
significantly lower than this value.
Fig. 2. Pulmonary capillary wedge pressure (mmHg) Box-plots
(median and 2575% percentile range) in patients with the different
syndromes of acute CHF and patients with septic shock and normal
controls.
133 of 158.
447
Fig. 3. Cpi (Wym2) Box-plots (median and 2575% percentile range)
in patients with the different syndromes of acute CHF and patients
with septic shock and normal controls.
3.2. CpiySVRi diagnostic graph (Fig. 5)
Downloaded from eurjhf.oxfordjournals.org by guest on June 14, 2011
Since the number of patients with HTN crisis was
small they were included in the exacerbated systolic
CHF group. Distributions of SVRi and Cpi were highly
skewed. The normality of the distribution of Cpi and
SVRi was assessed by the WilkShapiro test. The results
showed that Cpi and SVRi distribution was not normal
for the normal volunteers (Ps0.03 for Cpi), the exac-
erbated systolic CHF patients(Ps0.007 for Cpi and Ps
0.04 for SVRi) and the cardiogenic shock patients(Ps
0.016 for SVRi).
However, log(SVRi) and log(CPi) were normally
distributed (Psns for both Cpi and SVRi in all patient
groups). Therefore, for the analysis we used only log
Fig. 4. Systemic vascular resistance index (SVRi) (wood m2) Box-
plots (median and 2575% percentile range) in patients with the dif-
ferent syndromes of acute CHF and patients with septic shock and
normal controls.
 448 G. Cotter et al. / The European Journal of Heart Failure 5 (2003) 443451

numbers of observations in two of the groups prevented

us from using more flexible kernel functions.
Due to large variability of variances of the parameters
in the five groups, we could not suppose equal covari-
ance matrices in the groups. (The test of homogeneity
of within covariance matrices gives Ps0.)

3.2.1. Classification rule

The calculations leading to the classification rule and
CpiySVRi diagnostic graph are given in Appendix A.

3.2.2. Classification results

The results of the application of the classification rule
to the sample are presented in Table 4.

3.2.3. Performance of the classification rule

Fig. 5. Diagnostic graph for classification of the hemodynamic status
of patients with different syndromes of acute CHF.
values. The distribution of the two log parameters (Cpi
and SVRi) was different among the five groups, how-
The performance of a diagnostic procedure with only
two possible results and two classes of patients is usually
expressed using measures like positive (negative) pre-
dictive value w12x or diagnostic odds ratio w13x. For
more complex tests with many outcomes and many
classes of patients the overall performance may be

Downloaded from eurjhf.oxfordjournals.org by guest on June 14, 2011

ever, none enabled separation of the groups (Table 3).
These data suggested that separation may be possible
using two-dimensional discriminant analysis. We used
classical discriminant analysis for normal distributions
with unequal covariance matrices because the small
expressed through the difference between proportion of
erroneously classified patients with and without using
the test. This measure is usually called Lambda asym-
metric (RNC), where R (rows) is the true group and C
(column) is the group where the patient was classified.

Table 3
Number of observations classified into the correct clinical group using log(Cpi) or log(SVRi) only

Group Cardiogenic Exacerbated Normal Pulmonary Septic Total

shock systolic CHF edema shock
(a) Number of observations classified into appropriate groups: classification using log(CPi) only
Cardiogenic Shock 13 4 0 0 0 17
Exacerbated systolic CHF 1 44 14 0 2 61
Normal 0 9 8 0 3 20
Pulmonary edema 1 9 1 0 0 11
Septic shock 0 0 3 0 8 11

(b) Number of observations classified into appropriate groups using log(SVRi) only
Cardiogenic shock 2 12 1 2 0 17
Exacerbated systolic CHF 0 58 3 0 0 61
Normal 0 3 17 0 0 20
Pulmonary edema 2 0 0 9 0 11
Septic shock 0 0 0 0 11 11

Table 4
Number of observations classified into the correct clinical group using log(SVRi) and log(CPi)

Group Cardiogenic Exacerbated Normal Pulmonary Septic Total

shock systolic CHF edema shock
Cardiogenic shock 15 2 0 0 0 17
Exacerbated systolic CHF 0 60 1 0 0 61
Normal 0 0 20 0 0 20
Pulmonary edema 2 0 0 11 0 11
Septic shock 0 0 0 0 11 11

134 of 158.
 G. Cotter et al. / The European Journal of Heart Failure 5 (2003) 443451
For our data Lambda (RNC)s0.95 (S.D. (Lambda)s
0.03) which corresponds to three errors of classification
according to the classification rule, instead of 59 errors
of classification according to the prior probabilities.
4. Discussion
During the last decade the pathogenesis of CHF has
become clearer. The emphasis on cardiac performance
as the sole pathogenic mechanism of CHF has changed
to a more comprehensive understanding of the impor-
tance of the interaction between cardiac contractility,
neurohormonal and inflammatory control mechanisms
and vascular resistance. We have recently studied the
treatment of the acute CHF syndromes of pulmonary
edema and cardiogenic shock and have shown that
treatment modalities with significant vascular effect are
effective in improving the outcome of these patients
w1416x. These findings substantiated our theory that
the SVRi reaction to the decrease in Cpi determines the
hemodynamic condition and clinical syndrome of
patients with acute CHF.
4.1. Classical hemodynamic monitoring
CI is the most popular parameter used in invasive
hemodynamic monitoring of patients with acute CHF.
However, the results of the present study as well as
previous ones w2,46x show that CI measurements are
not sufficient for the diagnosis and treatment titration in
patients with acute CHF. This might be explained by
the fact that CI is actually a measure of cardiovascular
flow. Hence, CI (flow) is determined by both cardiac
contractility and vascular resistance and, therefore, may
change dramatically when Cpi decreases but also with
even mild changes in SVRi. Pulmonary capillary wedge
pressure is the second most popular hemodynamic var-
iable used in hemodynamic monitoring, since it repre-
sents the hydraulic pressure transmitted backwards to
the pulmonary circulation, and hence, is an important
determinant of pulmonary edema. However, wedge pres-
sure cannot be used for the exact diagnosis of the
different clinical syndromes of acute CHF, due to the
extent of overlapping values between patients with
exacerbated systolic CHF, HTN crisis and even cardio-
genic shock.
4.2. Cpi and SVRi and their role in patients with acute
CHF
Cpi as measured in the present study w1,8,9x is a
simplified version of a previously described method of
measuring cardiac contractility w7x. This value is derived
from the entire cardiac cycle (instead of instantaneous
measurements) and is the product of the mean pressure
and flow. Cpi has been shown to be the best predictor
135 of 158.
449
of outcome in chronic CHF patients w79x, exacerbated
systolic CHF w1x and cardiogenic shock (unpublished
data).
In the present study, we found that in patients with
exacerbated systolic CHF either Cpi was decreased or
SVRi was increased or both changes occurred. In a
previous study, we described the sequence of events
leading to acute heart failure w1x. In most patients an
acute CHF event starts with a progressive decrease in
cardiac contractility and power (Cpi). Thereafter, as Cpi
decreases, neurohormonal vascular control mechanisms
are activated and SVRi increases w1,17x. This increase
is a very important protective mechanism for two
reasons:
1. The increase in SVRi in the face of decreased
contractility maintains blood pressure and the perfu-
sion of vital organs.
2. This afterload increase (while within certain limits)
may improve contractility (possibly through the
Gregg phenomenon w18x), which may account for the
normal Cpi we observed in some patients with
Downloaded from eurjhf.oxfordjournals.org by guest on June 14, 2011
echocardiographically demonstrated systolic
dysfunction.
However, SVRi increase in response to Cpi decrease
is not uniform. It can be appropriate (thus, leading to a
compensated state), inappropriately low (thus, leading
to low blood pressure, forward hypoperfusion and car-
diogenic shock) or inappropriately high (thus, inducing
an extreme afterload mismatch leading to pulmonary
edema).
Indeed, in the present study, in patients who were
clinically diagnosed as cardiogenic shock, Cpi was found
to be extremely low, however, SVRi was only slightly
increased. This imbalance between very low Cpi and
inadequate increase in SVRi probably resulted in low
blood pressure and decreased perfusion pressure of vital
organs including the heart. This decrease in coronary
perfusion might lead to decreased contractility inducing
a vicious cycle of low contractility, low SVRi and
reduced perfusion. For this reason, in a previous study
w15x we treated patients with cardiogenic shock, by
short-term administration of a peripheral vasoconstrictor
(L-NMMA) with good clinical response.
On the other hand, in patients diagnosed as pulmonary
edema, despite what appears to be a similar clinical
presentation (pulmonary congestion, clammy extremi-
ties, low CI and high wedge pressure), the pathophy-
siological findings as well as the treatment, are the
complete opposite. In patients with pulmonary edema,
we measured Cpi values similar to those in exacerbated
systolic CHF, however, SVRi was markedly increased.
These findings are collaborated by the study by Gandhi
et al. w19x showing a dramatic increase in blood pressure
in patients with pulmonary edema. We hypothesize that
this increase in SVRi might be an inappropriate
 450 G. Cotter et al. / The European Journal of Heart Failure 5 (2003) 443451
response, related to neurohormonal, endothelial and
perhaps inflammatory activation w20,21x. This remark-
able increase in SVRi induces an afterload mismatch,
reducing CI and increasing intracardiac pressures,
LVEDP and wedge pressure, resulting in the severe
congestive symptoms of pulmonary edema. Therefore,
as previously suggested w14,16,22,23x, vasodilator treat-
ment is effective in the treatment of pulmonary edema.
4.3. Two-dimensional graphic representation of Cpiy
SVRi and its use in the treatment of cardiogenic shock
and pulmonary edema
In the present study, we found that when plotting on
a two-dimensional graph the results of Cpi and SVRi
for individual patients, each clinical group of patients
could be segregated into a specific area on the graph
which could be bound by a mathematically defined line
(Fig. 5). This graph enables exact clinical diagnosis of
most (95%) patients with exacerbated systolic CHF,
pulmonary edema, HTN crisis, cardiogenic shock and
septic shock. Of course, the boundaries on the graph are
somewhat arbitrary, since the definitions of the syn-
dromes are as used by the medical community, and
therefore, arbitrary. However, this two-dimensional rep-
resentation enables a better understanding of the patho-
physiology of the different syndromes of acute CHF.
We believe that this new and simple diagnostic tool
may become useful for the initial evaluation of acutely
decompensated patients, while the clinical diagnosis has
not yet been established and initiation of appropriate
disease-specific treatment is crucial. This might become
even more important with the advent of new devices
that accurately measure CI noninvasively. The combi-
nation of noninvasive MAP and CI measurement, Cpi
and SVRi calculation and the two-dimensional Cpiy
SVRi graph could enable improved diagnosis of patients
even in paramedic units and in emergency rooms.
Furthermore, this method may become an important
tool for monitoring the patients response to treatment.
4.4. Limitations
The results of the present study are based on a
relatively small number of patients, and therefore, need
confirmation by prospectively evaluating a larger group
of patients with acute CHF. Also, the measurements in
the present study were performed by thermodilution,
which has an inherent 1015% deviation in measuring
cardiac output. Finally, cardiac power calculations were
performed using whole-cycle measurements (cardiac
output and MAP). Although we recognize that the
cardiovascular system operates in a pulsatile manner,
cardiac power calculated according to the methodology
used in the present study has previously been shown to
be a useful measure of cardiac contractility and contrac-
136 of 158.
tility reserve in chronic and acute CHF as well as in
cardiogenic shock.
Appendix A: Classification rule
Given a patient with measured values of SVRi and
Cpi, the classification may be performed either (A)
through special calculations or (B) using the Graph for
classification of CHF patients (CpiySVRi graph).
(A) Classification using calculations
Step 1. Calculate three values v1, v2, v3 according to
the formulas below.
v1sLCPi2=21.54q2=LCPi=LSVRi=10.61
qLSVRi2=59.44yLCPi=305.24yLSVRi
=417.70q1408.89
v2sLCPi2=10.12q2=LCPi=LSVRi=5.67yLSVRi2
=4.99yLCPi=135.81yLSVRi=90.11q482.61
Downloaded from eurjhf.oxfordjournals.org by guest on June 14, 2011
v3sLCPi2=7.29qLCPi=LSVRi=2.57qLSVRi2
=4.09yLCPi=97.41yLSVRi=58.22q368.16
Classify the patient
into the group Septic shock, if v1 is the smallest value
into the group Cardiogenic Shock, if v2 is the smallest
value
if v3 is the smallest value go to step 2
Step 2. Calculate three values v4, v5, v6 according to
the formula below.
v4sLCPi2=6.45y2=LCPi=LSVRi=0.45q
LSVRi2=16.01yLCPi=65.16yLSVRi=116.53q
391.67
v5sLCPi2=17.75q2=LCPi=LSVRi=26.56q
LSVRi2=54.27yLCPi=420.26ySVRi=758.55q
2775.78
v6sLCPi2=32.95q2=LCPi=LSVRi=3.09q
LSVRi2=19.72yLCPi=390.74yLSVRi=161.49q
1355.57
Classify the patient
into the group Exacerbated Systolic CHF, if v4 is the
smallest value among v4, v5, v6 and LSVRi-log(67)
into the group Pulmonary Edema, if v5 is the smallest
value among v4, v5, v6 and LSVRi)log(67)
into the group Normal, if v6 is the smallest value
among v4, v5, v6
The value of SVRis67 was used to separate patients
with exacerbated systolic CHF from patients with pul-
monary edema since the group of pulmonary edema
was rather small and by classifying these patients
according to the usual rule we did not receive a
separating line for Cpi measures)250 Wym2.
 G. Cotter et al. / The European Journal of Heart Failure 5 (2003) 443451
(B) Classification using the diagnostic graph
Put the point (CPi, SVRi) on the diagnostic graph
Fig. 5 or point (LCPi, LSVRi) and classify the patient
according to the area of the graph, where the point is
located.
References
w1x Cotter G, Moshkovitz Y, Milovanov O, et al. Acute congestive
heart failure: a novel approach to its pathogenesis and treat-
ment. Eur J Heart Fail 2002;4:227 34 (review).
w2x Menon V, White H, LeJemtel T, Webb JG, Sleeper LA,
Hochman JS. The clinical profile of patients with suspected
cardiogenic shock due to predominant left ventricular failure:
a report from the SHOCK Trial Registry. Should we emergently
revascularize occluded coronaries in cardiogenic shock? J Am
Coll Cardiol 2000;36:1071 6.
w3x Kaluski E, Kobrin I, Zimlichman R, et al. RITZ-5: Randomized
Intravenous Tezosentan (an Endothelin ET-AyB Antagonist)
for the treatment of pulmonary edema. A prospective random-
ized, multicenter, double-blind placebo controlled study. J Am
Coll Cardiol 2003;41:20410.
w4x Morley D, Brozena SC. Assessing risk by hemodynamic profile
in patients awaiting cardiac transplantation. Am J Cardiol
1994;73:379 83.
w5x Anguita M, Arizon JM, Bueno G, et al. Clinical and hemodyn-
amic predictors of survival in patients aged -65 years with
severe congestive heart failure secondary to ischemic or non-
ischemic dilated cardiomyopathy. Am J Cardiol 1993;72:413
7.
w6x Levine TB, Levine AB, Goldberg D, Narins B, Goldstein S,
Lesch M. Reversal of end-stage heart failure is predicted by
long-term therapeutic response rather than initial hemodynamic
and neurohormonal profile. J Heart Lung Transplant
1996;15:297 303.
w7x Marmor A, Schneeweiss A. Prognostic value of noninvasively
obtained left ventricular contractile reserve in patients with
severe heart failure. J Am Coll Cardiol 1997;29:422 8.
w8x Williams SG, Cooke GA, Wright DJ, et al. Peak exercise
cardiac power output; a direct indicator of cardiac function
strongly predictive of prognosis in chronic heart failure. Eur
Heart J 2001;22:1496 503.
w9x Cohen-Solal A, Tabet JY, Logeart D, Bourgoin P, Tokmakova
M, Dahan M. A non-invasively determined surrogate of cardiac
power (circulatory power) at peak exercise is a powerful
prognostic factor in chronic heart failure. Eur Heart J
2002;23:806 14.
137 of 158.
451
w10x Cohen JA, Arnaudov D, Zabeeda D, Schlthes L, Lashinger J,
Schachner A. Non-invasive measurement of cardiac output
during coronary artery bypass grafting. Eur J Card Thoracic
Surg 1998;14:64 9.
w11x Lange RA, Hillis LD. Cardiac catheterization and hemodyn-
amic assessment. In: Topol EJ, Textbook of Cardiovascular
Medicine. page 1653.
w12x Sasieni PD. Statistical analysis of the performance of diagnostic
tests. Cytopathology 1999;10:73 8 (invited review).
w13x Lijmer JG, Willen Mol B, Heisterkamp S, et al. Empirical
evidence of design related bias in studies of diagnostic tests. J
Am Med Assoc 1999;282:1061 6.
w14x Cotter G, Metzkor E, Kaluski E, et al. Randomised trial of
high-dose isosorbide dinitrate plus low-dose furosemide versus
high-dose furosemide plus low-dose isosorbide dinitrate in
severe pulmonary oedema. Lancet 1998;351:389 93.
w15x Cotter G, Kaluski E, Blatt A, et al. L-NMMA (a nitric oxide
synthase inhibitor) is effective in the treatment of cardiogenic
shock. Circulation 2000;101:1358 61.
w16x Sharon A, Shpirer I, Kaluski E, et al. High-dose intravenous
isosorbide-dinitrate is safer and better than Bi-PAP ventilation
combined with conventional treatment for severe pulmonary
edema. J Am Coll Cardiol 2000;36:832 7.
w17x Mohr R, Rath S, Meir O, et al. Changes in systemic vascular
resistance detected by the arterial resistometer: preliminary
Downloaded from eurjhf.oxfordjournals.org by guest on June 14, 2011
report of a new method tested during percutaneous transluminal
coronary angioplasty. Circulation 1986;74:780 5.
w18x Karunanithi MK, Young JA, Kalnins W, Kesteven S, Dip G,
Feneley MP. Response of the intact canine left ventricle to
increased afterload and increased coronary perfusion pressure
in the presence of coronary flow autoregulation. Circulation
1999;100:1562 8.
w19x Gandhi SK, Powers JC, Nomier AM, et al. The pathogenesis
of acute pulmonary edema associated with hypertension. New
Engl J Med 2001;344:17 22.
w20x Bank AJ, Lee PC, Kubo SH. Endothelial dysfunction in patients
with heart failure: relationship to disease severity. J Card Fail
2000;6:29 36.
w21x Milo O, Cotter G, Kaluski E, Blatt A, Krakover R, Vered Z,
Hershkovitz A. Inflammatory and neurohormonal activation in
cardiogenic pulmonary edema: Implications on the pathogene-
sis and outcome of acute ischemic versus non-ischemic acute
heart failure. Am J Cardiol (In Print).
w22x Northridge D. Furosamide or nitrates for acute heart failure?
Lancet 1996;347:667 8.
w23x Cotter G, Kaluski E, Moshkovitz Y, Milovanov O, Krakover
R, Vered Z. Pulmonary edema: new insights on pathogenesis
and treatment. Curr Opin Card 2001;16:159 63.

Beyond the Four Quadrants: The Critical and Emerging Role
of Impedance Cardiography in Heart Failure
H eart failure (HF) is a disorder
characterized by hemodynamic
abnormalities including a reduction in
the hearts ability to deliver oxygen-
ated blood to the body. HF is also asso-
ciated with important neurohormonal
abnormalities, including activation
of the renin-angiotensin-aldosterone
and sympathetic nervous systems and
their resulting effects on the heart
and vascular endothelium. Our under-
standing of the neurohormonal role
in the progression of HF has greatly
improved in the past 10 years,1 and
many of the therapies that significant-
ly improve the symptoms and progno-
sis of patients with HF now target the
underlying neurohormonal abnormali-
ties. As shown in Figure 1, neurohor-
monal activation can lead to progres-
sion of hemodynamic abnormalities
resulting in reduced cardiac output
(CO); increased filling pressures; and
ultimately worsening symptoms of
fatigue, dyspnea, and decreased exer-
cise tolerance. Although the neuro-
hormonal mechanisms may cause pro-
gression of the disease process, nearly
all medications used in HF treatment
have demonstrable effects on hemo-
dynamics. Current acute HF treat-
ment is aimed directly at stabilizing
and improving a patients short-term
hemodynamic condition; chronic HF
treatments can alter short-term and
improve long-term hemodynamics.
Specific hemodynamic measure-
ments such as CO and systemic vascu-
John E. Strobeck, MD, PhD;1 Marc A. Silver, MD,2 Co-Editors in Chief
From the Heart-Lung Center, Hawthorne, NJ;1 and the Department of Medicine and the Heart Failure Institute, Advocate Christ
Medical Center, Oak Lawn, IL2
Address for correspondence: Marc A. Silver, MD, Advocate Christ Medical Center, 4440 West 95th Street, Suite 428 South,
Oak Lawn, IL 60453-2600
E-mail: marc.silver@advocatehealth.com
editorial: emerging role of ICG in HF
138 of 158.
CHF-MarchApril-Suppl-04.indd 1
EDITORIAL
Adrenergic System
60%
Ejection Fraction
Secondary
Damage
20%
Time
Hemodynamics
Reduced cardiac
output, increased
filling pressures
Figure 1. Neurohormonal activation and
resultant hemodynamic and symptom
changes. RAS=renin-angiotensin-aldo-
sterone system
lar resistance are generally obtained for
only the most critically ill HF patients,
in large part due to the risk, discom-
fort, and cost of invasive procedures
such as pulmonary artery catheter-
ization.2 Nonetheless, understanding
and measuring the factors that affect
CO are central to the assessment,
prognosis, and treatment of patients
with HF. The four determinants of
CO are the rate of the pump (heart
rate), the volume of blood available to
pump (preload), the pumping strength
(contractility), and the force the heart
must overcome to pump (afterload,
generally approximated by systemic
vascular resistance). Symptoms
physical findings like vital signsand
www.lejacq.com ID:3405
Congestion at Rest
X NO YES
X RAS A B
NO
Low Warm & Dry Warm & Wet
Perfusion
at Rest (Low Profile) (Complex)
YES L C
Cold & Dry Cold & Wet
Possible Evidence of Low Perfusion:
Narrow pulse pressure, cool extremities, sleepy/obtunded,
Symptoms
Fatigue, dyspnea,
hypotension with ACE inhibitor, low serum sodium, renal/hepatic
dysfunction
decreased exercise
Signs/Symptoms of Orthopnea/PND JV distension, Hepatomegaly,
tolerance
Congestion: Edema, Rales, Abdominal-jugular reflex
Figure 2. Clinical profiles in heart failure.
PND=paroxysmal nocturnal dyspnea;
JV=jugular vein; ACE=angiotensin-
converting enzyme. Adapted from J Am
Coll Cardiol. 2003;41(10):17971804.5
laboratory findings such as blood tests
and chest radiographs are imprecise
measures of hemodynamic function.
Unfortunately, they are the only data
many clinicians have at their disposal
when making important decisions in
the care of patients with HF.
The direct cost of treating HF is esti-
mated to be $56 billion per year in the
United States3 and the number of HF
patients in this country may reach 10
million by 2010.4 A significant portion
of the cost of HF care is the high cost of
hospitalizations for patients with acute
decompensation. Through careful sur-
veillance of patients with chronic HF
using improved methods for measuring
hemodynamic and neurohormonal
march . april 2004 . supplement 2 1
3/30/04 9:40:41 AM

Figure 3. Front view of impedance
cardiography method
status, primary care physicians and car-
diologists may be able to intervene in a
timely manner and prevent acute epi-
sodes leading to hospitalization, major
morbidity, or death.
Warner-Stevenson5 has developed
and popularized the concept of categoriz-
ing HF patients by hemodynamic sub-
set based on perfusion with CO (warm
vs. cold) and congestion with pulmo-
nary artery wedge pressure (wet vs. dry).
The four quadrants, representing the
four hemodynamic classes, are shown
in Figure 2. Studies have suggested that
these profiles provide a useful framework
to risk stratify patients with HF, pre-
dict outcomes, and identify therapeutic
Table I. Impedance Cardiography Parameters
PARAMETER
Flow
Stroke volume
Stroke index
Cardiac output
Cardiac index
Resistance
Systemic vascular resistance
Systemic vascular resistance index
Contractility
Preejection period
Left ventricular ejection time
Systolic time ratio
Velocity index
Acceleration index
Left cardiac work index
Fluid Status
Thoracic fluid content
VEPT=volume of electrically participating tissue; Z MARC=impedance modulating aortic compliance; CVP=central venous pressure (estimated value
of 6 mm Hg); MAP=mean arterial pressure; ECG=electrocardiogram; PCWP=pulmonary capillary wedge pressure (estimated value of 10 mm Hg)
2 editorial: emerging role of ICG in HF
139 of 158.
CHF-MarchApril-Suppl-04.indd 2
patients and evaluate the effectiveness of
subsequent pharmacologic and implant-
able interventions.
Impedance cardiography (ICG) is
a noninvasive method of determining
hemodynamic status. In the past, studies
questioned the reliability of ICG technol-
ogy,7,8 leading some to conclude that the
technology did not have value in clinical
decision making. However, refinements
in signal processing and CO algorithms
have greatly improved the reliability of
ICG technology. The latest generation
of ICG devices (BioZ ICG Monitor,
CardioDynamics, San Diego, CA; and
BioZ ICG Module, GE Medical Systems
Information Technologies, Milwaukee,
WI) are both highly reproducible and
accurate in a number of clinical settings,
Figure 4. Lateral view of impedance including HF.911 A recent search of the
cardiography method literature failed to show a single citation
since US Food and Drug Administration
options. However, this framework is based 510(k) clearances of these particular
on invasive pulmonary artery catheteriza- devices that suggests they are not valid
tion, with its requisite risk and cost, or on for clinical applications.
physical examination and patient history, ICG is a form of plethysmography
which have been shown to lack sensitiv- that utilizes changes in thoracic electri-
ity and specificity, even in the hands of cal impedance to estimate changes in
experienced clinicians.6 HF management blood volume in the aorta and changes
using hemodynamic subsets could be in fluid volume in the thorax. As shown
substantially improved by the existence of in Figures 3 and 4, the ICG procedure
more objective data with which to classify involves the placement of four dual
ABBREVIATION MEASUREMENT/CALCULATION UNITS
SV VI LVET VEPT (Z MARC algorithm) mL
SI SV/body surface area mL/m2
CO SV heart rate L/min
CI CO/body surface area L/min/m2
SVR ([MAP CVP]/CO) 80 dyne s cm-5
SVRI ([MAP CVP]/CI) 80 dyne s cm-5 m2
PEP ECG Q wave to aortic valve opening ms
LVET Aortic valve opening to closing ms
STR PEP/LVET No units
VI First time derivativemax/baseline impedance /1000/s
ACI Second time derivativemax/baseline impedance /100/s2
LCWI (MAP PCWP) CI 0.0144 kg m/m2
TFC 1/baseline impedance /kOhm
march . april 2004 . supplement 2
3/30/04 9:40:42 AM
 sensors on a patients neck and chest.
A low-amplitude, high-frequency alter-
nating current is delivered from the
four outer sensors and the four inner
sensors detect instantaneous changes
in voltage. As suggested by Ohms law,
when a constant current is applied to
the thorax, the changes in voltage are
directly proportional to the changes in
measured impedance. The overall tho-
racic impedance, called base impedance
(Z0) is the sum of the impedances of the
components of the thorax, including fat,
cardiac and skeletal muscle, lung and
vascular tissue, bone, and air. Changes
from Z0 occur due to changes in lung
volumes with respiration and changes
in the volume and velocity of blood in
the great vessels during systole and dias-
tole. The rapidly changing component
of chest impedance (Z) is filtered to
remove the respiratory variation, leaving
the impedance changes due to ventricu-
lar ejection. Figure 5 details the elements
contributing to Z0 and Z, and Figure 6
illustrates how the first derivative of the
impedance waveform (Z/t) is used
Table II. Summary of Impedance Cardiography Applications in Heart Failure
APPLICATION DESCRIPTION
Assessment and Establish baseline hemodynamics
diagnostic Trend changes to gauge level of hemodynamic decompensation
Determine whether symptoms are due to hemodynamic deterioration
Aid in differentiation of systolic vs. diastolic dysfunction
Prognostic Emergency department values predictive of length of stay and
hospital charges
Improvements associated with improving NYHA class, quality-of-life
measures
Abnormal values associated with mortality
Treatment Determine stability for initiation and up-titration of -blocker and
ACE-inhibitor therapy
Assist in selection of drug agents and dosing
Measure response to adjustments in therapy
Determine need and optimal selection/dosing of IV therapy
(dobutamine, milrinone, nesiritide)
Optimize LVAD settings and wean patients from LVAD support
Determine optimal pacemaker settings in patients with AV sequential
pacemakers
Detect hemodynamic changes due to compensation, medication, and
diet compliance
Provide an adjunct to post-transplant myocardial biopsies
NYHA=New York Heart Association; ACE=angiotensin-converting enzyme; LVAD=left
ventricular assist device
Adapted from Yancy C, Abraham W. Noninvasive hemodynamic monitoring in heart
failure: utilization of impedance cardiography. Congest Heart Fail. 2003:9(5):241250.
Blood volume
Venous return

with an electrocardiogram to determine Posture

the beginning of electrical systole, aortic Heart rate Preload

valve opening, maximal deflection of the Exercise

Contractility

Z/t waveform, and the closing of the Intrinsic

contractility
Stroke
volume
aortic valve. From these fiducial points, Myocardial
compliance

a variety of measured and calculated

Afterload
parameters (Table I) are continuously Arterial blood

displayed on the ICG device screen for pressure

Aortic
compliance
monitoring purposes, or in a printed Age Z
report for review (Figure 7). Pulmonary
artery
Pulmonary
artery
(Systole)

The hemodynamic parameters compliance compliance

derived from ICG can aid in the Thoracic fluid content

diagnostic and prognostic evaluation Body anatomy/composition

of patients with HF. Using ICG, a Cardiogenic pulmonary edema

Specific
resistance Z0
of thorax
clinician is able to evaluate direct or Noncardiogenic pulmonary
edema
indirect measures of each of the four Emphysema

major determinants of CO (preload,

afterload, contractility, and heart Hematocrit Specific
resistance

rate). Figure 8 is a conceptual dia- Orientation of erythrocytes of blood

gram of CO and its determinants,

ICG parameters associated with the
determinants, and the effects of
pharmacologic agent classes on each
determinant. Due to greater accep-
tance of ICG in clinical and research
settings, clinicians are now able to
use ICG-derived hemodynamic data
to help decide when to initiate and
Figure 5. Contributing elements to thoracic impedance. Z0=baseline impedance;
Z=change in impedance
Adapted from Osypka MJ, Bernstein DP. Electrophysiologic principles and theory
of stroke volume determination by thoracic electrical bioimpedance. AACN Clin
Issues. 1999;10(3):385399.
titrate these types of medications. A In this supplement to Congestive
summary of applications of ICG in HF Heart Failure, we seek to further define
is presented in Table II, demonstrating the role of ICG through a series of
its broad clinical applicability. original contributions. The study by

editorial: emerging role of ICG in HF march . april 2004 . supplement 2 3

140 of 158.

CHF-MarchApril-Suppl-04.indd 3 3/30/04 9:40:43 AM


ECG
Z onstrate the ability of ICG to simply
Z/t
Figure 6. Fiducial points derived from electrocardiogram (ECG) and impedance
waveforms. Z=change in impedance; Z/t=first derivative of the impedance
waveform; PEP=preejection period; LVET=left ventricular ejection time
Figure 7. Impedance cardiography hemodynamic status report (BioZ ICG Monitor,
CardioDynamics, San Diego, CA)
Yung et al. (p. 7) validates the accu-
racy of ICG in patients with pulmonary
hypertension by comparing ICG to both
direct Fick method and thermodilution
4 editorial: emerging role of ICG in HF
141 of 158.
CHF-MarchApril-Suppl-04.indd 4
et al. (p. 11) compare changes in ejec-
Q =Start of ventricular
depolarization tion fraction by echocardiography to
changes in ICG parameters in estab-
B =Opening of pulmonic
and aortic valves lished HF patients. Their findings dem-
C =Maximal deflection
of the Z/ t and cost-effectively identify changes in
X =Closure of aortic
ventricular function. While pulmonary
valve artery catheterization in patients with
Y =Closure of pulmonic
HF has been criticized and is largely
valve unproven by clinical trial, an estimated
O =Mitral opening/
2 million such catheters are sold world-
rapid ventricular
filling
wide each year.12 Springfield et al. (p.
14) illustrate the role of ICG in the
differential diagnosis of patients with
dyspnea. Although B-type natriuretic
peptide testing has gained wide atten-
tion recently as an aid to diagnose
HF in the emergency department,13
ICG may also have a diagnostic role
and provides additional value because
of its ability to identify appropriate
therapeutic options and monitor the
response to therapy in real time. Silver
et al. (page 17) report on the abil-
ity of ICG to replace pulmonary artery
catheterization, which has tremendous
cost implications for hospitals caring
for such patients. Vijayaraghavan et al.
(page 22) demonstrate the prognostic
role of ICG in patients with chronic
HF, and show strong association of ICG
changes to changes in functional status
and quality-of-life measures. Summers
et al. (page 28) provide a series of
case reports that illustrate ICGs prac-
tical role in the initiation and titra-
tion of neurohormonal agents and their
patient-specific hemodynamic effects.
This compilation of studies adds to
the growing body of data supporting
the role of ICG in the management of
patients with HF. Within a year, the
results of two multicenter trials studying
key roles for ICG should be available:
PRospective Evaluation and identifica-
tion of Decompensation by Impedance
Cardiography Test (PREDICT), con-
ducted in patients with chronic HF;
and the BioImpedance cardioGraphy
(BIG) substudy of the Evaluation
Study of Congestive Heart Failure
CO. In doing so, the authors demon- and Pulmonary Artery Catheterization
strate the potential hazard of using Effectiveness (ESCAPE).14 PREDICT
thermodilution as the only reference specifically addresses the ability of
standard for CO measurement. Parrott ICG-derived hemodynamic data to
march . april 2004 . supplement 2
3/30/04 9:40:45 AM
 identify patients at risk for death, hos-
pitalization, or emergency department Cardiac output
visit. The BIG substudy will evaluate (CO/CI)

the diagnostic and prognostic role of

ICG in both arms of a randomized,
controlled trial in pulmonary artery Heart rate
Stroke
catheterhemodynamic-guided man- volume
(HR)
(SV/SI)
agement of patients admitted with an
acute episode of HF.
There is now a compelling body of Contractility
Preload Afterload
(ACI, VI, PEP,
literature that demonstrates the validity (TFC) (SVR/SVRI)
LVET, STR)
of ICG using the most current technol-
ogy. More and more studies have shown
the value of ICG in clinical settings (-) Neg. Chronotropes
(+) Pos. Chronotropes
(-) Diuretics
(+) Volume Expanders
(-) Vasodilators
(+) Vasoconstrictors
(-) Negative Inotropes
(+) Positive Inotropes

in addition to HF, including dyspnea,15

hypertension,16 and atrioventricular
sequential pacemakers.17 The studies
presented in this issue of Congestive
Heart Failure further define the role of
this valuable, noninvasive technology in
clinical medicine. It is likely that these
and other studies of ICG in HF will be
used to refine our understanding and
ability to assess patients and predict
prognosis, expanding on the concept of
the four quadrants presented in Figure
Figure 8. Pharmacologic agent effect on cardiac output determinants and imped-
ance cardiography parameters. CO=cardiac output; CI=cardiac index; HR=heart
rate; SV=stroke volume; SI=stroke index; TFC=change in thoracic fluid
content; SVR=systemic vascular resistance; SVRI=systemic vascular resistance
index; VI=velocity index; PEP=preejection period; ACI=acceleration index;
LVET=left ventricular ejection time; STR=systolic time ratio
Congestion at Rest
NO YES
TFC TFC

2. The impact of adding ICG hemo- A B

dynamic data to the four quadrants NO CI
Warm & Dry Warm & Wet
is depicted in Figure 9. Knowledge of Low
Perfusion
stroke index, cardiac index, systemic at Rest (Low Profile) (Complex)
vascular resistance index, and chang- YES CI L C
es in fluid with thoracic fluid content Cold & Dry Cold & Wet
would likely provide more quantitative, SVRI SVRI
objective, and sensitive measurements
of hemodynamic factors, and has signifi- Role of impedance cardiography

cant implications for the management of
patients with HF.
Incorporating this model of assess-
ment into a proposed therapeutic algo-
rithm is shown in Figure 10. Ideally,
a baseline measurement of ICG in
addition to other standard clinical
variables would be collected and uti-
lized in combination to more precisely
assess a patients perfusion, conges-
tion, and vasoactive status. This assess-
ment would lead to a categorization
of the patients absolute or relative
change in hemodynamic profile, facili-
tating assessment of short-term risk for
adverse HF-related events. The change
in hemodynamic status and assessment
1. Establish baseline cardiac index (CI), systemic vascular resistance index (SVRI) and thoracic fluid content (TFC)
2. Monitor changes in CI to objectify perfusion assessment
3. Monitor changes in TFC to objectify congestion assessment
4. Monitor changes in SVRI to objectify vasoactive status
Figure 9. Model for clinical profiles in heart failure utilizing impedance cardiogra-
phy hemodynamic measurements
of higher risk may lead to increased Note: This supplement to Congestive
clinical surveillance or a decision to Heart Failure contains articles dealing
intervene to prevent a negative patient with ICG. Readers are reminded that
outcome. In addition, ICG parameters positive statements about the clinical
may aid in the assessment of a stable, utility of ICG, and the BioZ ICG Monitor
low-risk hemodynamic profile toward in particular, are solely the opinions of
the initiation and up-titration of neuro- the authors and do not represent an offi-
hormonal agents that are often under- cial endorsement by Congestive Heart
prescribed but are known to improve Failure, its Editors or Editorial Board, or
event-free survival. the Heart Failure Society of America.

REFERENCES
1 Packer M. How should physicians view heart evolution of three conceptual models of the 2 ACC/AHA guidelines for the evaluation and
failure? The philosophical and physiological disease. Am J Cardiol. 1993;71(9):3C11C. management of chronic heart failure in the

editorial: emerging role of ICG in HF march . april 2004 . supplement 2 5

142 of 158.

CHF-MarchApril-Suppl-04.indd 5 3/30/04 9:40:47 AM

 PREDICTED
HEMODYNAMIC THERAPY
ASSESSMENT SHORT TERM
PROFILE OPTIONS
RISK

COLD & WET

Increase or add diuretic
Lower SI / CI
Higher Risk Need to decrease afterload?
Higher TFC Increase or add vasodilating agent
Baseline Evaluation Higher SVRI
History & Physical
Lab tests
Diagnostic tests
ICG hemodynamics WARM & WET

Higher SI / CI Increase or add diuretic

+ Higher TFC Moderate Risk Need to decrease afterload?

Increase or add vasodilating agent
Higher SVRI?
Current Visit Evaluation
History & Physical
Lab tests
Diagnostic tests
COLD & DRY
ICG hemodynamics
Lower SI / CI
Increase or add vasodilating agent
Lower TFC
Moderate Risk

Higher SVRI

Is patient currently taking neurohormonal

WARM & DRY antagonists?

Higher SI / CI 1. ACEI or ARB

Lo w e r Risk 2. Beta blocker
Lower TFC 3. Aldosterone inhibitor

Higher SVRI? If no, initiate

If yes, uptitrate toward target dose

Figure 10. Therapeutic algorithm for incorporating impedance cardiography (ICG) parameters into clinical assessment of
heart failure. SI=stroke index; CI=cardiac index; TFC=thoracic fluid content; SVRI=systemic vascular resistance index;
ACEI=angiotensin-converting enzyme inhibitor; ARB=angiotensin-receptor blocker

adult: executive summary. A report of the
American College of Cardiology/American
Heart Association Task Force on Practice. J
Am Coll Cardiol. 2001;38(7):21012113.
3 OConnell JB. The economic burden of heart fail-
ure. Clin Cardiol. 2000;23(3 suppl):III6III10.
4 Heart Disease and Stroke Statistics2003
Update. Dallas, TX: American Heart
Association; 2002.
5 Nohria A, Tsang SW, Fang JC, et al. Clinical
assessment identifies hemodynamic profiles
that predict outcomes in patients admit-
ted with heart failure. J Am Coll Cardiol.
2003;41(10):17971804.
6 Shah MR, Hasselblad V, Stinnett SS, et al.
Hemodynamic profiles of advanced heart
failure: association with clinical character-
istics and long-term outcomes. J Card Fail.
2001;7(2):105113.
7 Sageman WS, Amundson DE. Thoracic elec-
trical bioimpedance measurement of cardiac
output in postaortocoronary bypass patients.
Crit Care Med. 1993;21(8):11391142.
8 Marik PE, Pendelton JE, Smith R. A compari-
son of hemodynamic parameters derived from
transthoracic electrical bioimpedance with
those parameters obtained by thermodilution
and ventricular angiography. Crit Care Med.
1997;25(9):15451550.
9 Greenberg BH, Hermann DD, Pranulis MF,
et al. Reproducibility of impedance cardiog-
raphy hemodynamic measures in clinically
stable heart failure patients. Congest Heart
Fail. 2000;6(2):7480.
10 Drazner M, Thompson B, Rosenberg P, et
al. Comparison of impedance cardiography
with invasive hemodynamic measurements
in patients with heart failure secondary to
ischemic or nonischemic cardiomyopathy. Am
J Cardiol. 2002;89(8):993995.
11 Kazuhiko N, Kawasaki M, Kunihiko T.
Usefulness of thoracic electrical bioimpedance
cardiography: noninvasive monitoring of car-
diac output. J Card Fail. 2003;9(suppl):S170.
12 Ginosar Y, Sprung CL. The Swan-Ganz cath-
eter: twenty-five years of monitoring. Crit
Care Clin. 1996;12:771776.
13 Mueller C, Scholer A, Laule-Kilian K, et al.
Use of B-type natriuretic peptide in the evalu-
ation and management of acute dyspnea. N
Engl J Med. 2004;350:647654.
14 Shah MR, OConnor CM, Sopko G, et
al. Evaluation Study of Congestive Heart
Failure and Pulmonary Artery Catheterization
Effectiveness (ESCAPE): design and rationale.
Am Heart J. 2001;141(4):528535.
15 Peacock F, Summers R, Emerman C. Emergent
dyspnea impedance cardiography-aided
assessment changes therapy: The ED IMPACT
trial. Ann Emerg Med. 2003;42(4):S82.
16 Taler SJ, Textor SC, Augustine JE. Resistant
hypertension: comparing hemodynamic man-
agement to specialist care. Hypertension.
2002;39:982988.
17 Santos JF, Parreira L, Madeira J, et al.
Noninvasive hemodynamic monitorization for
AV interval optimization in patients with ven-
tricular resynchronization therapy. Rev Port
Cardiol. 2003;22(9):10911098.

6 editorial: emerging role of ICG in HF march . april 2004 . supplement 2

143 of 158.

CHF-MarchApril-Suppl-04.indd 6 3/30/04 9:40:49 AM

 European Journal of Cardio-thoracic Surgery 14 (1998) 6469

Non-invasive measurement of cardiac output during

coronary artery bypass grafting

Amram J. Cohen a , c ,*, Dimitri Arnaudov b , c, Deeb Zabeeda b , c,

Lex Schultheis d, John Lashinger e, Arie Schachner a , c
a
Department of Cardiovascular Surgery, The Edith Wolfson Medical Center, Holon, 58100 Israel
b
Department of Anesthesiology, The Edith Wolfson Medical Center, Holon, 58100 Israel
c
Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel
d
Department of Anesthesiology and Critical Care Medicine, Johns Hopkins Medical Center, Baltimore, MD, USA
e
Department of Cardiovascular Surgery, Johns Hopkins Medical Center, Baltimore, MD, USA

Received 18 November 1997; revised version received 23 March 1998; accepted 15 April 1998

Abstract

Objective: A new device, using whole body bioresistance measurements and a new equation for calculating stroke volume has been
developed. Using this equation, an attempt was made to correlate whole body bioresistance cardiac output with thermodilution cardiac
output in patients undergoing coronary artery bypass grafting. Methods: Thirty-one adults undergoing elective coronary artery bypass
grafting were studied prospectively. Simultaneous paired cardiac output measurements by whole body bioresistance and thermodilution
were made at fiv time points during coronary artery bypass grafting: in anesthetized patients before incision (T1), after sternotomy (T2),
after opening the pericardium (T3), ten min post bypass (T4), and in the intensive care unit (T5). The patients had a mean of three
thermodilution cardiac outputs compared with a mean of three bioimpedance measurements at each time point. The bias and precision
between the methods were calculated. Results: There was good correlation between bioresistance cardiac output (nCO) and thermodilution
cardiac output (ThCO) measurements in both groups for all recorded times. The patients mean ThCO and nCO, as well as bias and
precision between methods were calculated. Mean ThCO ranged between 4.14 and 5.06 l/min; mean nCO ranged between 4.12 and 4.97 l/
min. Bias calculations ranged between 0.072 and 0.104 l/min. Precision (2 SD) calculations ranged between 0.873 and 1.228 l/min for
95% confidenc intervals. Pearsons correlation ranged from 0.919 to 0.938. Conclusions: Cardiac output measured with the new device
correlates well with the thermodilution measurements of cardiac output during and immediately following coronary artery bypass grafting.
The overall agreement between the two methods was good. The new device is an accurate non-invasive method of measuring cardiac output
during coronary artery bypass grafting. 1998 Elsevier Science B.V. All rights reserved

Keywords: Cardiac output; Cardiac surgery; Thermodilution; Electrical bioimpedance; Bioresistance; Hemodynamics

1. Introduction A number of attempts have been made to determine CO

Hemodynamic monitoring continues to be an integral part
of peri- and postoperative care for patients undergoing car-
diac surgery. Cardiac output (CO) is an important parameter
in these measurements. To date, the clinical tool used to
measure CO is the SwanGanz catheter using a thermodilu-
tion technique. However, the procedure is invasive, expen-
sive, and may lead to complications [1,2].
* Corresponding author. Tel.: +972 3 5028723; fax: +972 3 5028735.
in a non-invasive manner [36]. Using thoracic electrical
bioimpedance techniques (TEB), moderate success has been
achieved in some clinical settings [3,4,68]. The technique
has been unsuccessful when applied to patients undergoing
cardiac surgery [911]. A new non-invasive cardiac system
device has been developed to utilize whole body bioresis-
tance in a semi-empirical formula to determine the CO.
Accuracy of the CO measurement using this device was
established comparing thermodilution CO (ThCO) for
patients undergoing right and left heart catheterization
[12]. The purpose of this investigation was to compare the

1010-7940/98/$19.00 1998 Elsevier Science B.V. All rights reserved

PII S1010-7940 (98 )0 0135-3
144 of 158.
 A.J. Cohen et al. / European Journal of Cardio-thoracic Surgery 14 (1998) 6469
new technology for measuring cardiac output with thermo-
dilution measured CO in patients undergoing coronary
artery bypass grafting (CABG).
2. Patients and methods
2.1. Patients
Thirty-one patients undergoing elective, primary CABG
were prospectively studied; 15 at Wolfson Medical Center
and 16 at Johns Hopkins University School of Medicine.
Patients with aortic valve insufficien y, mitral and tricuspid
regurgitation or cardiac shunt were not included. The
CABG was similar in both hospitals. All patients underwent
a median sternotomy. Cardiopulmonary bypass was estab-
lished using an aortic and single venous cannula. Diastolic
arrest was achieved with cold sanguineous potassium cardi-
oplegia in both antegrade and retrograde fashion. The inter-
nal mammary artery was used to graft the left anterior
coronary artery (LAD) in all cases, and saphenous vein
grafts were used to bypass the other obstructed vessels.
Proximal anastomoses were performed using a partial
cross clamp. During the study, no patient had atrial flutte
nor atrial fibrillation
2.2. Protocol
The study protocol was approved by Edith Wolfson and
Johns Hopkins Institutional Review Boards. For each
patient demographic and clinical data was tabulated. In
each patient, CO measurements were taken at fiv different
time intervals during the procedure; after induction of
anesthesia but before incision (T1), after sternotomy (T2),
after creation of a pericardial pocket (T3), ten min after
weaning from bypass (T4), and immediately after arrival
in the intensive care unit (T5).
At each time interval, three adequate thermodilution mea-
surements were made. All measurements were made at end
expiration during the respiratory cycle. Thirty-one patients
underwent 155 thermodilution CO measurements. Nineteen
were excluded because of 10% variation between the mea-
surements. Simultaneously, three bioresistance measure-
ments were taken at each interval, and their average was
considered the bioresistance CO.
Each bioresistance measurement took 20 s. Since these
measurements were continuous, there was no time between
measurements such that to achieve an average bioresistance
CO measurement took 1 min.
2.3. Thermodilution method
At both hospitals seven French true size thermodilution
catheters (Baxter Healthcare, Irvine, CA) were inserted in
the operating room after induction of anesthesia. The
proper location of the thermodilution catheter was con-
145 of 158.
65
f rmed by hemodynamic measurements and by post-
operative chest roentgenograms. The pulmonary artery
catheter was connected to a thermodilution cardiac output
monitor (Horizon 2000, Mennen Medical, Israel was used
at Wolfson Medical Center, and 7010 Series Marquette,
Madison, WI at Johns Hopkins Institutions). Ten milliliters
of room temperature 5% dextrose solution was injected
manually at end-expiration by an experienced anes-
thesiologist who was unaware of the bioresistance CO
results.
2.4. Whole body resistance method
NICaSy 2001, a non-invasive cardiac system device
(NICaSy 2001, Teledyne-NIM, LLC, North Andover,
MA, 510K, certificat number K942227), was utilized for
measuring bioresistance CO. Two proprietary designed,
NICaSy disposable electrodes (510K, certificat number
K972002) were applied to each wrist, and were then con-
nected to the non-invasive cardiac system device. The sys-
tem passed an AC current 30 kHz and 1.4 mA through the
whole body. It then measured the resistive portion of the
bioimpedance. Since the electrodes were placed on the dis-
tal aspect of the extremities, the current passed through all
major arteries and veins of the systemic circulation. As
such, the system measured the bioresistance of the systemic
circulation and allowed calculation of the stroke volume.
In 1992, Tsoglin and Frinerman and developed a new
semi-empirical formula for calculating the stroke volume
(SV) [13].
Hctcorr H 2 DR a+b
SV = kel Kweight IB corr
Ksex, age R b
where Kel is a coefficien related to blood electrolytes,
Kweight is a ratio of a patients weight to ideal weight, IB
is the index balance equal to the ratio of the measured
extracellular flui (as measured from the baseline impe-
dance) to the expected extracellular flui volume, HCTcorr
is a hematocrit correction factor proportional to measured
hematocrit, Ksex,age is a coefficien depending upon a
patients sex and age, H is a patients height, DR is the
incremental change of the resistive portion of the bioimpe-
dance, R is the baseline whole body bioresistance and
a + b/b is the ratio of the sum of the systole time plus
diastole time divided by the diastole time derived from
the fin structure of the varying portion of the bioresistance.
Using whole body resistance measurements and the
above formula, SV was calculated and converted to CO.
2.4.1. Statistical analysis
The correlation between methods at each time point was
evaluated by Pearsons correlation coeff cient and simple
linear regression. The degree of agreement between meth-
ods at each time point was evaluated by calculation of the
bias (mean between-method difference) and precision
(mean bias 2 SD) [14].
 66 A.J. Cohen et al. / European Journal of Cardio-thoracic Surgery 14 (1998) 6469

Table 1 where SV is related to the resistivity of blood (r), dZ/dt is

Patient demographic and clinical data the firs peak of the derivative of the bioimpedance curve, L
is the distance between the electrodes, Z0 is the mean time
Wolfson Medical Johns Hopkins
Center Medical Center
averaged thoracic bioimpedance and T is the ventricular
ejection fraction.
Males (N) 11 13 The equation was modif ed by Bernstein [16] to:
Females (N) 4 3
Average age (years) 63.26 9.94 64.2 9.37 SV = VEPT ( dZ =dt)=Z0 )T
Minimum age (years) 46 51
Maximum age (years) 79 81 where VEPT is a coeff cient that represents the volume of
Average weight (kg) 73.13 14.74 87.14 13.29 electrically participating tissue. Using this equation, limited
Minimum weight (kg) 42 59 clinical success has been achieved in determining CO using
Maximum weight (kg) 102 108
Average height (cm) 166.3 10.4 176.92 8.96
bioimpedance techniques. Correlation to thermodilution
Minimum height (cm) 140 160 techniques have been achieved in healthy volunteers,
Maximum height (cm) 183 185 patients undergoing operations without cardiopulmonary
Temperature (C)a 26.8 0.77 25.56 0.403 bypass, patients undergoing procedures in the cardiac
Number of vessels bypassed 3.33 0.96 2.55 0.73 catheterization laboratory, and small numbers of intensive
Diabetes (%) 46.67 18.75
Hypertension (%) 53.33 56.25
care unit patients [3,4,7,1720].
This equation has been problematic due to diff culties in
a
Lowest temperature during the operation. accurately computing VEPT, and its application becomes
impractical in a patient undergoing rapid hemodynamic
3. Results changes. Furthermore, determination of VEPT is dependent
upon the accurate placement of electrodes, which is not
Thirty-one patients were evaluated in the study. Demo- always possible during open heart surgery. In addition,
graphic and clinical data are shown in Table 1. The agree- this equation still depends upon the firs derivative of the
ment between the average CO measured by bioresistance bioimpedance curve, which is a rapidly fluctuatin factor.
versus thermodilution is shown in Table 2. The plot of the As a result, attempts to correlate bioimpedance CO in the
regression analysis and difference versus mean for all mea- patient with thermodilution techniques during CABG have
surements at all times in the 31 patients is shown in Fig. been unsuccessful [9,10].
1a,b. There was good correlation between ThCO and nCO In addition to practical problems, there is a conceptual
ranges of cardiac output during all phases of CABG and the problem in applying TEB to measure stroke volume. TEB
immediate postoperative period. measurements of the bioimpedance includes both systemic
and pulmonary circulations. The two circulations cannot be
separated by measurements and variation in proportion of
4. Discussion these circulation will lead to inaccurate measurements of SV
[21,22]. This will not happen in whole body bioresistance
The initial attempt to obtain CO by measuring the stroke measurements where the systemic circulation dominates the
volume (SV) through thoracic electrical bioimpedance measurement and reflect LV SV.
(TEB) was performed by Kubicek [15] at the National Aero- The physiological and physical basis of bioimpedance
nautics Space Agency (NASA) where he introduced the has been studied by many authors [2325]. However, the
equation that became the basis for bioimpedance cardiogra- variation between different tissues and their complex struc-
phy: tures make it difficu t to model their electrical behavior. In
terms of bioresistance, the body may be divided into the
SV = r((dZ =dt)) ((L2 T))=(Z02 ) blood compartment and tissue compartment. In the

Table 2
Agreement of CO by NICaS and thermodilution measurements (n = 31)

Time Mean Mean R/R 2 Regression SEE y intercept Bias (mean between- Precision mean 2
ThCO nCO slopea method difference) SD (l/min)
(l/min)

After anesthesia 4.19 4.19 0.93/0.86 1.128 0.328 0.529 0.0086 1.1131.131
After sternotomy 4.14 4.12 0.92/0.85 0.898 0.185 0.404 0.019 0.8630.823
After pericardiotomy 4.14 4.24 0.93/0.86 0.915 0.326 0.456 0.104 1.0151.223
Immediately after bypass 5.06 4.97 0.92/0.85 0.967 0.367 0.324 0.0483 1.2341.138
ICU admission 4.71 4.62 0.94/0.88 0.891 0.306 0.424 0.072 1.1561.012

SEE, standard error of the estimate; ICU, intensive care unit.

a
y, cardiac output by NICaS 2001 bioimpedance; x, cardiac output by thermodilution.

146 of 158.
 A.J. Cohen et al. / European Journal of Cardio-thoracic Surgery 14 (1998) 6469 67

blood compartment, conductivity is high and therefore a
current introduced into the body will travel primarily in this
compartment. The resistance changes in this compartment
as the blood volume varies in the great arteries due to pul-
satile flow In the tissue compartment, there is less signif-
icant current flo and constant resistance. In each
individual, the resistance in the tissue compartment is
determined by the patients height, lean body weight (mus-
cle) to fat ratio, sex, age, body build, extracellular flui
volume, and electrolytes. In the blood compartment, the

Fig. 1. (a) Linear regression analysis comparing bioresistance measured cardiac output with thermodilution cardiac output for all averaged measurements in
the study. (b) Mean difference between bioresistance measured cardiac output and thermodilution cardiac output for all averaged measurements in the study.

147 of 158.
 68 A.J. Cohen et al. / European Journal of Cardio-thoracic Surgery 14 (1998) 6469
resistance is determined by hematocrit, electrolytes and the
pulsatile blood volume changes within the system. To mea-
sure SV, the pulsatile changes within the blood compartment
and baseline bioresistance must be measured accurately and
all other factors must be accounted for with appropriate
correction factors. The measurement of whole body biore-
sistance allows for the accurate measurement of pulsatile
changes and baseline resistance, and the proposed correc-
tion factors allow for the accurate calculation of the sys-
temic bioresistance with which accurate LV SV can be
calculated.
The method utilized in this study has been proven accu-
rate in patients undergoing catheterization [12]. It was also
shown to be accurate in a pilot study in patients under-
going CABG [26]. Compared with previous attempts to
correlate bioresistance with SV, the equation has major
advantages.
1. The equation does not depend on rapidly fluct ating
time derivatives of bioimpedance.
2. The equation uses empirically derived coeffici nts that
are obtained from laboratory values and simultaneously
measured bioimpedance values instead of the artifi ial
and difficu t to approximate VEPT.
3. The precise placement of the electrodes is not a critical
factor.
4. The NICaSy electrode arrangement is optimal to mea-
sure the left ventricular SV.
5. The respiration has almost no effect on the NICaSy CO
measurements in real time.
The new methodology showed good correlation between
thermodilution and bioresistance CO during all phases of
CABG and the immediate postoperative period in two inde-
pendent hospital populations. Correlation was good in all
ranges, including low cardiac output. This fact is important
since patients undergoing CABG frequently have low car-
diac output. This was even true in the immediate post car-
diopulmonary bypass period where changes in the patients
volume status, temperature, blood electrolytes and hemody-
namics are changing rapidly. Such a correlation would be
impossible using previous techniques.
The new device has certain limitations. First, it cannot
measure CO while using diathermy. Second, the device is
sensitive to movement so that the patient cannot be manipu-
lated while the CO is being measured. These two factors
require that the operation stop for the 2030 s required to
measure CO. Finally, the device measures SV and multiplies
it by heart rate to measure CO. Arrythmias in which the
heart rate varies signific ntly will result in a non-represen-
tative measurement of CO.
In summary, a new device using a semi-empirical equa-
tion relating SV to changes in the resistive portion of the
patients bioresistance has been developed. Information
about specif c patient data, blood components and body
habitus inserted into the analytical software has signific ntly
improved the ability to calculate SV from the bioresistance
148 of 158.
data. The device allows accurate, easy to obtain, non-inva-
sive CO during cardiac surgery. Within the above stated
limitations, we have confir ed the validity of the new biore-
sistance methodology in 31 patients who underwent CABG.
Bioresistance measured cardiac output correlated well with
ThCO during the CABG procedure and the immediate post-
operative period.
Acknowledgements
This paper was prepared in consultation with Diklah
Geva who prepared the statistical calculations, and with
the technical assistance of Sally Esakov.
References
[1] Mermel LA, Maki DG. Infectious complications of SwanGanz
pulmonary artery catheters: pathogenesis, epidemiology, prevention,
and management. Am J Respir Crit Care Med 1994;149:1020
1036.
[2] Choh JH, Khazei AH, Ihm HJ, Thatcher WC, Batty PR. Catheter
induced pulmonary arterial perforation during open heart surgery. J
Cardiovasc Surg (Torino) 1994;35:6164.
[3] Denniston JC, Maher JT, Reeves JT, Cruz JC, Cymerman A, Grover
RF. Measurement of cardiac output by electrical impedance at rest
and during exercise. J Appl Physiol 1976;42:9195.
[4] Smith SA, Russell AE, West MJ, Chalmers J. Automated non-inva-
sive measurement of cardiac output: comparison of electrical bioim-
pedance and carbon dioxide rebreathing techniques. Br Heart J
1988;59:292298.
[5] Reybrouck T, Fagard R. Assessment of cardiac output at rest and
during exercise by a carbon dioxide rebreathing method. Eur Heart J
1990;11 (suppl I ):2125.
[6] Belardinelli R, Ciampani N, Costantini C, Blandini A, Purcaro A.
Comparison of impedance cardiography with thermodilution and
direct Fick methods for non-invasive measurement of stroke volume
and cardiac output during incremental exercise in patients with
ischemic cardiomyopathy. Am J Cardiol 1996;77:12931301.
[7] White SW, Quail AW, DeLeeuw PW, Traugott FM, Brown WJ,
Porges WL, Cottee DB. Impedance cardiography for cardiac output
measurement: an evaluation of accuracy and limitation. Eur Heart J
1990;11 (Suppl ):7992.
[8] Teo KK, Hetherington MD, Haennel R, Greenwood PV, Rossall RE,
Kappagoda T. Cardiac output measured by impedance cardiography
during maximal exercise tests. Cardiovasc Res 1985;19:737743.
[9] Sageman WS, Amundson DE. Thoracic electrical bioimpedance
measurement of cardiac output in postaortocoronary bypass
patients. Crit Care Med 1993;21 (8 ):11391142.
[10] Clarke DE, Raffi TA. Thoracic electrical bioimpedance measure-
ment of cardiac output not ready for prime time. Crit Care Med
1993;21 (8 ):11111112.
[11] Thomas AN, Ryan J, Doran BR, Pollard BJ. Bioimpedance versus
thermodilution cardiac output measurement: the Bomed NCCOM3
after coronary bypass surgery. Intensive Care Med 1991;17 (7):383
386.
[12] Miller HI, Frinerman E, Tsoglin A, Frinerman A, Rosenschein U,
Keren G, Roth A, Laniado S. Non-invasive bioimpedance measure-
ment of cardiac output: a validation study (Abstr). Second Confer-
ence of the Israel Heart Society, 1993:58.
[13] Tsoglin, A., Frinerman, E, N.I. Medical, Ltd. Non-invasive method
and device for collecting measurements representing body activity
 A.J. Cohen et al. / European Journal of Cardio-thoracic Surgery 14 (1998) 6469
and determining cardiorespiratory parameters of the human body
based upon the measurements collected. US Patent No. 5,469,859,
Nov. 28 1995.
[14] Bland, J.M., Altman, D.G. Statistical methods for assessing agree-
ment between two methods of clinical measurement. Lancet
1986;Feb 81(8476):307310.
[15] Kubicek WG, Patterson RP, Witsoe DA. Impedance cardiography as
a non-invasive method of monitoring cardiac function and other
parameters of the cardiovascular system. Ann NY Acad Sci
1970;170:724729.
[16] Bernstein DP. A new stroke volume equation for thoracic electrical
bioimpedance: theory and rationale. Crit Care Med 1986;14:904
909.
[17] Donovan KD, Dobb GJ, Woods WP, Hockings BE. Comparison of
transthoracic electrical impedance and thermodilution methods for
measuring cardiac output. Crit Care Med 1986;14:10381044.
[18] Salandin V, Zussa C, Risica G, Michielon P, Paccagnella A, Cipolotti
G, Simini G. Comparison of cardiac output estimates by thoracic
electrical bioimpedance, thermodilution and Flick methods. Crit
Care Med 1988;16:11571158.
[19] Thomas SHL. Impedance cardiography using Sramek-Bernstein
method: accuracy and variability at rest and during exercise. Br J
Clin Pharmacacol 1992;34:467476.
149 of 158.
69
[20] Wong DH, Onishi R, Tremper KK, Reeves C, Zaccari J, Wong AB,
Miller JB, Cordero V, Davidson J. Thoracic bioimpedance and Dop-
pler cardiac output measurement: learning curve and interobserver
reproducibility. Crit Care Med 1989;17:11941198.
[21] Fuller HP. Evaluation of left ventricular function by impedance car-
diography: a review. Prog Cardiovasc Dis 1994;36:267273.
[22] Patterson RP. Sources of the thoracic cardiogenic electrical impe-
dance signal as determined by a model. Med Biol Eng Comput
1985;23:41417.
[23] Karnegis JN, Kubicek WG. Physiological correlates of the cardiac
thoracic impedance waveform. Am Heart J 1970;79:519523.
[24] Sakamoto K, Muto K, Kanai H, Iizuka M. Problems of impedance
cardiography. Med Biol Eng Comput 1979;17:697709.
[25] Paul E, Marik MD, Judy E. A comparison of hemodynamic para-
meters divided from transthoracic electrical bioimpedance with those
parameters obtained by thermodilution and ventricular angiography.
Crit Care Med 1997;25:15451550.
[26] Cohen A, Frinerman E, Katz M, Ezra S, Dotan A, Weissberg A,
Schachner A. Validity of bioimpedance hemodynamic measurements
during coronary artery bypass grafting (CABG) (Abstr). J Cardiovasc
Diag Proc 1996;13:57.
 EVALUATION AND MANAGEMENT OF THE ELDERLY HOMEBOUND
PATIENT WITH CONGESTIVE HEART FAILURE
C. Gresham Bayne MD

BACKGROUND: The frail elderly patient population is characterized not only by age
demographics, but by the inadequacy of the office-based medical model to meet their
needs when they can no longer reach the physicians office on a regular basis. In
addition, we are often called to see the new patient with presenting symptoms of heart
failure precipitated by one of many proximate causes.

Congestive Heart Failure (CHF) is the leading cause for hospitalization after the age of
65. There are one million admissions for CHF annually with 80% of the patients over 65
years of age. CHF is the most costly DRG in the United States with annual costs
exceeding $10 billion. The incidence and prevalence of CHF are rising in the population
with prevalence doubling each decade after age 45. The one year survival for Class III and
IV CHF elder patients is still only about 60%. The best objector predictor for survival
has been the cardiothoracic ratio on standard chest Xray.

There is a close and perhaps causative relationship between hypertension and CHF in the
elderly due to the complex neurohormonal reflexes involved (see below). As people age
the cardiovascular system normally increases its electrical impedance, has lowered beta-
adrenergic responsiveness, alters its myocardial energy metabolism, with the heart usually
showing impaired diastolic relaxation and compliance. The net effect is a marked
reduction in the cardiovascular reserve.

PRECIPITATING EVENTS:
These precipitating events must be evaluated first, before the patients failure can be
addressed. The most common of these is the failure to take their medications due to
financial, social, or cognitive problems. Despite the medication failure, the benefit might
be afforded us to switch them to the more acceptable geriatric approach to congestive
heart failure than has often been the case in the past. Briefly stated, the philosophy of
treatment has changed dramatically in past years from a diuretic-based ramp-up to the
primary use of ACE inhibitors and other afterload reducers. This modern approach will
be discussed more fully later.

Other precipitating factors common to our population include anemia, hyperthyroidism,

infection, new onset arrhythmias such as atrial fibrillation, change in diet, silent
myocardial infarction, worsening valvular disease, inappropriate use of the sodium-
retaining NSAIDS, and environmental changes such as increased ambient humidity or
temperature, and stress. Obviously, these historical and clinical factors must be evaluated
in perspective.

The absence of a readily available medical records and EKG/Xray documentation

mandates the use of appropriate technology when patients present with symptoms
suggestive of failure. Even in non-acute settings, the ability to have local physicians copy

150 of 158.
 and send forward medical records is so remote, it should not be anticipated that a chart is
forthcoming. In fact, many of the patients will be well-served by an entirely new
approach to their clinical symptoms, especially those on multiple medications.
Obviously, precipitating causes should be addressed before redirecting the primary
interest in care to the failure itself.

PHYSIOLOGY REVIEW:

The function of the heart is to pump arterialized blood forward under sufficient pressure
to meet the peripheral tissue metabolic needs. Given adequate oxygenation and a normal
hemoglobin, the cardiac output, defined as the pulse times the stroke volume, should
pump enough blood to prevent angina (ischemia to the heart), cold knee caps (ischemia to
the skin), oliguria or azotemia (ischemia to the kidney) and confusion (ischemia to the
brain). Thus, a quick iSTAT and pulse oximetry is required to rule out azotemia,
hypoxemia, and a metabolic acidosis demonstrated by the base deficit. Even a venous
sample with a normal base excess is adequate to prove the absence of a low perfusion
state.

NOTE that the above definition of inadequate cardiac output (CO) did NOT mention the
blood pressure at all! In fact, the only time hypotension is an emergency for bedbound
patients is when the diastolic pressure is below 60 mmHg and the patient experiences
angina or failure symptoms. Since coronary arterial blood flow is 85% during diastole,
the aortic root pressure is critical in determining myocardial perfusion. Thus, in the
emergency room, some patients in shock with angina find that dopamine is the best
analgesic!

To evaluate the failing heart, one needs to think in terms of myocardial efficiency, NOT
blood pressure. Efficiency of the heart is measure in terms of Oxygen Demand by the
heart as balanced against the Stroke Work Index. We will first address the Oxygen
Demand, which is determined by five factors, of which only three are clinically relevant:
the pulse, the mean arterial pressure and the wall tension index (as measured by the CXR
assessment of cardiomegaly). Myocardial oxygen demand is linearly increased as the
pulse increases over 100, dramatically increased by mean arterial pressure (usually
estimated by a value one-third the distance between the diastolic and systolic pressures),
and increases despite the same cardiac output with increases in the left ventricular end-
diastolic volume (cardiomegaly).

Thus, the patient with a pulse of 110, a blood pressure of 140/100 and cardiomegaly on
CXR is demanding much more oxygen for their heart than a patient of normal heart size
and vital signs. When the presenting symptoms include angina, the approach to failure is
first to treat the increased demand by using nitrates or nifedipine to lower the MAP, slow
the pulse, and vasodilate the coronary arteries. Once stabilized, the failure may be
addressed.

151 of 158.
 Stroke Work Index (SWI) is a complicated clinical measurement requiring the use of a
pulmonary arterial catheter, but can be thought of in clinical terms quite simply.
Remember, the heart does not have to create higher than normal pressures to increase its
cardiac output (CO) if the peripheral vascular resistance (PVR) is reduced. The SWI is
calculated basically by multiplying the cardiac output times the PVR. The human heart is
a very efficient flow generator, but a very inefficient pressure generator. Thus, oxygen
demand to increase cardiac output is efficiently handled, but requirements for
hypertension are disastrous. It is important to note that the failing heart may, due to its
inefficiencies and maladaptive neurohumoral reflexes, simply increase the blood pressure
without generating much increase in the CO. Thus, hypertension is the enemy of both
ischemic heart disease and congestive heart failure.

The physiological changes in the elderly which occur in the chronic state of heart failure
are numerous and complex. They include:
1. Increased renin levels, which increases the conversion of Angiotension I to
Angiotension II, the most potent vasoconstrictor known
2. Increased renin levels promoting the kidney to increase aldosterone leading to
increased retention of sodium and water
3. Increased adrenergic background tone by the adrenal cortex with elevated
circulating norepinephrine levels and inappropriate hypertension
4. Decreased levels of adenyl cyclase leading to decreased levels of cyclic AMP
which can lead to lower protein kinase, calcium entry levels and calcium re-
uptake by the myofibrils

The net result is a chronic fight or flight background level of physiologic stress which
ultimately leads to cardiac failure. I think of my elderly patients as on an endogenous
catecholamine drip producing more hypertension than forward flow. Since it is unlikely
that the patient will ever be normal, it is useful to think in terms of therapy to maximize
forward blood flow at the lowest physiologic cost.

TREATMENT PRINCIPLES:

To maximize the cardiac output, we typically intervene with the three major forces we
can easily effect with medication: Preload, Contractility, and Afterload.

Preload: The hypovolemic patient may be in shock or hypotensive simply because of

fluid depletion. This is often easily discerned by the history, physical and iSTAT testing,
but often problematic in the bedbound, demented patient. The hypervolemic patient is
often much harder to evaluate, since chronic rales, interstial lung disease, emphysematous
changes, obesity, patient intolerance to the exam, may all affect our confidence level.
Therefore, some simple in-home testing may be useful. The CXR showing cardiomegaly
is strong support for a chronic hypervolemic state and elevated left ventricular filling
pressures. The presence of LVH or left ventricular strain pattern in the EKG (R waves in
V5 and V6 add up to more that 25mv) is often associated with overload.

152 of 158.
 A final test, done only in the normotensive or hyptertensive patient, is the so-called
Chatterjee test when one monitors the pulse before and after a single dose of sublingual
nitroglycerine. Since NTG predictably lowers the pulmonary capillary wedge pressure
(PCWP) (left ventricular filling pressure), the hypervolemic patient will NOT have a
tachycardic change in pulse with NTG. The normal or hypovolemic patient will increase
their pulse by 10% or more. Thus, the absence of tachycardia in response to NTG is de
facto evidence of increase PCWP and indicates the need for (temporary) diuresis.

The use of nitrates alone to control preload and subsequent CHF symptoms is
complicated by the tolerance that all patients develop for the longer acting versions. The
use of nitrates in the elderly is most often useful on the acute housecall when the patient
has suddenly decompensated and has hypertension in addition to the symptoms of CHF.
One must be cautious in the home when using nitrates to be able to place the patient in
recumbency before administration, as the nitrate syncope syndrome is much more likely
in our patients.

The use of diuretics has been the mainstay of treatment for both hypertension and failure
for some forty years, despite the fact that diuretics have yet to be shown in controlled
prospective clinical trials to increase survival rates. Currently, some 35-40% of all
seniors over the age of 65 are taking a diuretic in some form. However, cumulative
studies are now overwhelming in their condemnation of the indiscriminate application of
either loop or thiazide diuretics due to the following known side effects:
1. 60% of all toxic drug reactions are due to diuretics in the elderly
2. diuretics are the most common drug reaction requiring hospitalization
3. both types cause increased serum cholesterol, although subsequent cardiac
disease is unproven
4. both cause increased uric acid levels with subsequent tophaceous gout
5. both cause glucose intolerance in Type 2 prone patients, and the instigation of
Type 2 diabetes is NOT reversible by stopping the diuretic (bumex less than
lasix)
6. both cause deafness, even in chronic oral dosing regimes (bumex less than
lasix)
7. both cause decreased calcium absorption and worsen osteoporosis (loop worse
than thiazides)
8. both cause behavioral risks with incontinence in the female, urinary retention
in the male
9. both can lead to hyperkalemia in the elderly as well as hypokalemia
10. both can cause severe hyponatremia in the elderly

Thus, the use of diuretics is now considered second or third line therapy, after more
contemporary treatment for both hypertension and/or failure has failed.

Contractility: The state of inotropy of the heart in failure patients is compromised by

many of the maladaptive neurohormonal changes discussed above as well as the loss of
cardiac tissue from current ischemic heart disease and past infarcts. Unfortunately, the

153 of 158.
 use of phosphodiesterase inhibitors in the elderly to increase inotropy has been clinically
unsuccessful and the use of adrenergic agents such as dopamine and dolbutamine requires
monitoring and infusion therapy along with the hassles of an intravenous. Early studies
on NY Stage IV CHF patients on home dobutrex drips have prolonged life but at a
significant cost on the quality of life.

What we have learned in the ICU from the two adrenergic agents in most common use
today is important. Whereas, dopamine causes a more potent increase in inotropy and
blood pressure, it has little or no effect on PCWP, increases ventricular irritability, and
requires much more care. Dolbutamine, on the other hand, not only has less toxicity at
the same cardiac output, but lowers the PCWP dramatically, thereby restoring the normal
cardiac anatomy and reducing myocardial oxygen demand.

To reduce the oxygen demand of the heart, the eighties and early nineties saw an increase
in the use of beta blockers even in failure patients. Unfortunately, the use of beta
blockers to control hypertension in the elderly involves much more toxicity, and (other
than post MI) has shown to increase mortality due to side effects, including an increase in
cholesterol for all age groups.

The use of digoxin remains controversial in patients without the need for rate control in
atrial fibrillation. Although digoxin is the oldest and most proven inotropic agent, and
the only one which does so while decreasing the wall tension index (cardiomegaly), its
toxicity and need for periodic serum measurements makes it of much more problematic in
the demented, homebound elderly CHF patient. A NEJM study in 1997 on Class II or III
CHF patients showed no change in mortality with digoxin therapy in addition to ACE
inhibition (see below), but a significant improvement in functional status and decrease in
hospital days. Most patients with an ejection fraction of over 45% will do well on
digoxin alone for control of symptoms of CHF. Obviously, the ability to measure directly
an improvement in cardiac output with digitalis would allow better patient selection in the
future.

Suffice to say, patients with cardiomegaly and/or a S3 gallop should have a trial on
digoxin alone before using multiple drug regimes. Finally, no one argues about the use of
digoxin to control rapid ventricular response in the patient with atrial fibrillation.

Afterload: If the SWI can be reduced by reducing the MAP, one might be able to both
increase the forward flow of blood (CO) as well as decrease the myocardial oxygen
demand. This would be the best of both worlds! Indeed, we now know that peripheral
vasodilatation has less morbidity and better survival rates with better functional status
than conventional beta blocker/diuretic regimes. The geriatric meetings are replete with
studies confirming the observation that the ACE inhibitors by themselves can often
control hypertension and should be used for patients with risk of CHF (such as LVH or
cardiomegaly on CXR), and especially diabetics even when they are asymptomatic.

154 of 158.
 By using ACE inhibition at lower doses beginning at night, one can often avoid
orthostatic hypotension and preserve the peripheral vasomotor reflexes so important to
prevent side effects. Many patients can be converted from 2-4 older medications
(digoxin, lasix, KCL, tenormin) to a single dose of benzapril 10 mg hs. Titrating the
benzapril up to 80 mg daily may control the blood pressure while relieving the symptoms
of failure as well. The therapeutic effects must be evaluated by the patients subjective
reports of exercise tolerance or orthopnea, as well as objective weights. If CHF is to be
controlled on ACE inhibitors alone, the patient will lose weight on it.

In addition, the Cooperative North Scandinavian Enalapril Survival Study showed that
ACE inhibition improves symptoms and exercise tolerance, while affording significant
survival benefits among patient with severe CHF. These agents not only decrease
afterload, but preload as well, and deter activation of the neuroendocrine system leading
to reduced risk of hypertensive episodes, stress reduction, less fluid retention, and perhaps
less arrhythmogenicity.

Before reverting back to the diuretics which some patients simply must have, adding a
calcium channel blocker in angina patients or alpha receptor blockade may be useful to
improve inotropy by increasing coronary blood flow and further decreasing the peripheral
vascular resistance. In addition, the newer calcium-channel blockers such as amlodipine
have been shown to increase end-diastolic ventricular relaxation, promoting ventricular
filling and increasing stroke volume at the same filling pressures. Once the afterload is
under control, the SWI is decreased enough to usually require no further therapy
requirements for heart failure. Of course, if you cannot know the cardiac output, you
cannot calculate afterload, so using BP as a surrogate if fraught with undertreatment
biases. Should further therapy be required, a diuretic may be cautiously introduced since
their effect is additive to both ACE inhibition and calcium channel blockade.

What has been lacking in both the critical care and outpatient arenas is an ability to know
the systemic vascular resistance. Thus, we are still basing our clinical decisions on non-
physiological parameters: i.e., mean arterial pressure. Although MAP may be a useful
guideline for stroke prevention, many patients with borderline /low cardiac outputs have
normotension while their SVR is markedly elevated. We cannot know this without
measurement of the cardiac output, which heretofore has been invasive and expensive
with the pulmonary arterial catheter. The advent of non-invasive, inexpensive cardiac
impedance studies in the outpatient and other settings has opened up a dramatic
opportunity to not only titrate therapies toward maximum lowering of the SVR to
improve cardiac output, but also to titrate therapies both acutely and chronically toward
the best cardiac efficiency factors, since impedance can track the pre-ejection period and
minimize isovolumetric contraction time.

The Use of Cardiac Impedance in the Home-Bound Patient

Background: For over twenty-five years it has been possible to provide a non-invasive
monitoring system capable of measuring the trends in cardiac output, stroke volume,

155 of 158.
 ejection fraction and total intra-thoracic fluid. Since the chest is essentially a box
containing a variable amount of non-conducting air in the lungs and airways, and a
variable amount of salt water in tissue and blood, measuring the resistance to the flow of
electricity through the chest should correlate to fluctuations in air and/or fluid.

In fact, this measurement, known as impedance, can easily be measured across the chest
by a series of paired electrodes placed on the neck and lateral chest wall. The static
values measured on an insensitive scale correlate to Total Fluid Content or TFC of the
chest and will change over time with things like pleural effusions, extravascular lung
water, and congestive heart failure. Using a very sensitive scale, one can see alterations
in impedance conforming to the pumping action of the heart during each systole. Since
the heart pumps blood out of the chest once the aortic valve opens, the velocity of this
pumping action can be calculated from the first derivative of impedance or dZ/dT. The
peak value of dZ/dT correlates closely with the stroke volume (SV) of each heartbeat.
Since the cardiac output is easily calculated by multiplying the SV time heart rate, one
can watch online as therapies affect the cardiac output.

In the 1960s impedance research was done primarily in Israel where engineers were best
in signal acquisition technology or Russia, where cognitive strengths were focused on
complex algorithms to reduce the signal noise from motion or respiratory artifact.
Recently, a combination of these talents has come together to produce an FDA-approved
non-invasive, whole-body impedance measurement which has the distinct advantage of
not having to undress female patients (electrodes are placed on two extremities). There
continues to be controversy over which is best, trans-thoracic or whole body
impedance. In addition, there remains argument over which proprietary algorithm should
be used. Proponents of the Sramek Equation believe selecting the single best systole for
detailed examination and reporting is superior to averaging data from many systolic
waveforms during a one minute sample (for example) analyzed by the Kubicek Equation.

The second derivative of impedance varies with the acceleration of the blood as it exits
the left ventricle and correlates closely with the inotropic state of the heart. Since
electrodes can easily sense the EKG tracing of a patient, one can pinpoint the R wave and
measure the time between the initiation of systole and the time when the aortic valve
opens, sensed as the onset of the acceleration of blood or the second derivative of
impedance. This period is called the PEP or Pre-Injection Period, a time when the heart
is maximally consuming oxygen during systole, but doing no useful work, since no
bloodflow has yet occurred in the aorta. The longer the PEP, the more inefficient the
heart is for any given cardiac output. Examples of clinical conditions which prolong the
opening of the aortic valve or PEP are hypertension, myocardial dyskinesis, ventricular
aneurysm, and aortic stenosis. Finding the maximum cardiac output for the minimum
PEP has now become routine in setting time intervals for two channel pacemakers.

Since the closing of the aortic valve is also marked by an abrupt deceleration of blood
flow in the aorta, one can use impedance to mark the closure of the aortic valve. Thus,
the LVET, or Left Ventricular Ejection Time, when the left ventricle is actually pumping

156 of 158.
 blood, can now be tracked with non-invasive monitoring. More importantly, it turns out
that clinical studies of patients with heart disease, including mitral valve disease and atrial
fibrillation, have shown an extremely close inverse correlation of the left ventricular
ejection fraction and the PEP/LVET ratio.

Clinical Significance in the Home Bound Patient:

Many patients over the age of 80 are not only homebound, but extremely demented and
frail. Not only do they have a high prevalence of congestive heart failure, the number one
cause of death in the elderly, but they cannot get to the cardiologist for appropriate
studies. Even if they could, they cannot tolerate a treadmill or other means of evaluating
their heart. For those that can make it to the cardiologists office, Medicare policy limits
payment for the gold standard of doppler ultrasound to annual usage in most cases. Thus,
the reality is that the homebound patient is usually not followed objectively in the
management of their cardiac function during therapeutic changes directed at blood
pressure or congestive heart failure.

It is clear now that afterload and CHF control are critical to maximizing both the lifespan
and the functional status of these patients. Unfortunately, we have not had access to
objective measurements for any of our therapies in the home. For instance, it is generally
agreed that ACE inhibition therapy should be maximized to a physiologic endpoint, but
daily weights, clinical exam, and pulse oximetry are late findings and too non-specific
with too many covariables to give us confidence.

The use of cardiac impedance as a non-invasive static and dynamic monitoring system to
document cardiac function in the home represents a break-through technology allowing
us to pursue better quality of care for these patients. Using the priniciple of transthoracic
bioimpedance, intensivists and surgeons have known for over a decade that cardiac output
and ejection fractions can be monitored with as much reliability as thermodilution Swan-
Ganz catheter measurements. Not only is the data correlated with direct Fick
measurements as well as thermodilution, the bioimpedance measurements show less
inter-operator variability and higher reproducibility for trend analysis with a given patient.

In summary, the use of FDA-approved portable bioimpedance devices (paid by Medicare

at about $40/test), such as the (www.cardiodynamics.com ), the HemoSapiens
(www.hemosapiens.com) or NIMedical (www.ni-medical.com) offers another step
forward in low-cost, comprehensive testing and monitoring of patients outside of the
hospital environment: whether in the home, the outpatient clinic, or the physicians
office. It is likely that the ease of use and higher physiological parameters measured will
make this monitoring a mainstay for both future cardiologists and housecall physicians,
although the high capital costs of machines and low per-test reimbursement have limited
growth in the fee-for-service sector.

157 of 158.
 Examples of theoretical usage of non-invasive measurement of trans-
thoracic impedance in the homebound patient:

1. Evaluation of the new patient in heart failure

2. Weaning CHF patients off their high-dose beta-blockers and calcium
channel/diuretic therapies
3. Titrating to the maximum dose of ACEI/ARB consistent with renal
impairment
4. Monitoring diuretic therapies in the bedbound patient or others that
cannot be weighed
5. Adjustment of pacemaker rate settings for maximum cardiac output in
marginal patients
6. Titrate hypertensive therapies to the point of maximum cardiac
efficiency
7. Evaluate the patient with frequent PVCs, AF or other arrhythmia
affecting output
8. Evaluate starting or stopping of digoxin therapy
9. Safety in rapid rehydration or transfusion therapy at home
10. Monitoring the terminal patient remotely to aid the family in planning,
etc.
11. Monitoring patient compliance with dietary restrictions or fluid
hydration overnight
12. Use in pharmcologic stress test for IHD evaluations in the bedbound
patient
13. Monitoring of pleural effusions over time or during thoracentesis
14. Screening for CHF in patients with poor histories and weakness
15. Evaluate Cyclosporine or EPO-induced hypertension
16. Short term stabilization of the patient in pulmonary edema

158 of 158.