Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Study Review
January 2014
Table of Content
2
Table of Content, Contd
3
4 of 158.
) t'k*ffiam
\Ieasuring a patient's cardiac output-the amount regarding their accuracv compared to other cardiac
of blood pumped bv the heart per minute-can aid output monitodng techniques, particuiadr ther-
in diagnosing diseases of the cardiovascular system modilution. \Y'e also examine evidence regarding
and in managing high-risk patients (e.g., those suf- improved clinical outcomes with use of these
fering from burns, sepsis, or shock) and patients derrices.
undergoing maior surgicai procedures. The cardiac \\:e look at the following s,rstems:
output value can be derived lrom heart rate and
stroke volume (the amount of blood pumped br-
) ConMed Ecollt. The literarure s'e revierved did
the heart per beat) using the relationship
not show the ECOII to be ao acceptable alter-
native to the reference techniques; horvercr,
cordioc oulpul = stroke volume x heort roie
nvo studies found that the monitor rvas useful
The uaditional approach to monitoring car- in predicting fluid responsir.eness in abdominal
diac output is pulmonarv arterv catheterization and cardiac surger\r patients. \I'hen the monitor
using thermodilution as the measurement method. is used in the ()R electrosurgical equip-
"vith
.Florvever, the draw-backs associated with this tech- ment, interruptions in monitoring could occur.
nique-including the risk of pulmonarr arterv \\:e did flot 6od any sfudies addressing clinicai
ruprure aod air emboli5m-h21.g led clinicians ro outcomes.
seek less invasive alternatives. F Edwords Llfesciences FloTroc sensor. Previousl,v
In our December 2009 issue, rve looked at some reviewed literarure showed the FloTrac sensor to
rninimallf invasive and noninvasive alternatives to be comparable to thermodilution in stable car-
pulmonarl' arterv catheterization for cardiac output diac surgerv patients. Srudies on the latest gen-
monitoring and revierved the evidence regarding eration of the FIoTrac sensor, horvever, [ound
dreir effectiveoe ss. This article setres as ao update to that the device may be less reliable than thermo-
that piece. We profiie four additional cardiac output dilution in hemodynamicallv unstable patients
monitoring svsterns and rer.'ierv the evidence 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
O2O'13 ECRI lnstitute. Member hospitols moy reproduce this poge for internol distribution only. w.cri.org > HEALTHDEVICES JUNE2ots 179
6 of 158.
guidance
/
18O xrllrx DEvtcEg JUNE 2ot3 ) w.xrl.ors O2Oi3 ECRI lnsiitute. Member hospitols moy reproduce ihis poge for internql distribulion only.
7 of 158.
placed on the patient. Changes in the drarvback o[ these locat-ions is the suscep- excl.range) or intracardiac shunt Qlood
applied signal are then measured during ubiliq' to noise: In addition to blood flow; Ilorv across the chambers ot the heart),
each cardiac cr-cle. ()bsen'ed changes in the electrodes ma,r pick up other signals pulmonan'blood flou'can be equated to
impedance and phase shift can be related such as airflos, through the lungs. Two svstemic blood llorv or cardiac ourPut'
to changes in blood volume. Der,'ices using svstems rc revierv use ICG: the Conlled Devices using this technique mal be inac-
these techniques are n'picallr susceprible EC()\I (rvhich uses a combi'ration of ICG curate in patients with abnotmallr high
to motion artilacts that might prove a hin- and arterial r'"aveform anaitsis) and the \I intrapulmonarv or iotracardiac shttnts
drance to continuous monitoring. fledical \ICaS. (rvhich can occur in thc presence oi some
()ne o[ the more cornmon electrical respiratort or cardrac diseases).
Rebreothing. The noninvasive rebreathing
techniques is knorvn as impedance car- technique is based on the principie that Tronsthoro(ic echocordiogrophy. In trans-
diograph,r' (ICG), u.'hich is based on tl're intake or elirnination o[ a gas (as iodicated thoracic echocardiography, or TTE, an
theon that changes in blood tolume in bv the ditference in the concentration of ulrasound probe is placed on the chest
the aorta during each cardiac cvcle tesult the gas entering and exiring the lungs) is to measure the Yelocifr o[ blood across
in irnpedance changes that can be used proportional to the amount of blood florv the aort-ic or pulmonan= vaive. This, com-
to estimate stroke volume. The imped- to the lungs that is invol't'ed in the gas bined s,ith the cross-sectional area oF the
ance changes can be derived br measuring exchange. The amount of gas consumed/ valves, allot's calculation of ieft or right
voltage changes to ao applied electrical produced is used as an indication of pul- stroke volume and cardiac output. This
signal. Traditionalh-, the ICG technique monarv blood flos-, and in the absence of noninvasive technique mar be inaccurate
uses electrodes placed on the Patientt neck significant intrapulmonan shunt (blood in patients u'ith r,alr.e and florv-pattern
and chesr (knorvn as thoracic ICG). ()ne florv to thc lungs not invoh-ed in gas abnormalities.
PRODUCT REVIEWS
Pulmonarv arter't- thermodilution is consid- technique and uses this in conjunction elecuodes are in such ciose proximifi'to the
ered the goid standard for cardiac ourput ,x.ith the arterial s,zr.elorm analrsis tech- aorta, the EC()]I is less susceptible to noise
measurement. Therefore, the tables that nique to determine catdiac output. compared to traditional ICG methods.
accompanv our product reviervs comParc The monitor consists of an endotracheal clinicol uses, -\ccording to the vendor, the
the evidcnce on each product with the rube with a printed elecrrode arrat that trC()\I is used primarilv in the OR lor
evidence on thermodilution. But because applies the electrical signal required for ICG. short-term conLinuous morutoring and
comparisons to other methods can also The elecrode arrav consists of a single is indicated for use in patients rvho are
be use6;1 (see the box on page 180), rhe elecuode at the proximal end of the rube expected to be inrubated for 24 hours ot
discussions in each revierv also take into and slr electrodes at the distai eod. \\'hen less. .\lthough ICG technologr in general is
consideration srudies comparing the prod- the patient is innrbated, the distal eiectrodes susceptible to motion artilacts, this mav not
ucts to otier cardiac output techniques. are located adjacent to the ascending aorta, be an issue with this derice since inrubated
thus iacilitating aortic blood ilorv measure- patients in the ()R are also npicallv sedated.
ments. The electrode arrav also detects Disposobles, The monitor requites a propri-
the R-rr,ar-e and uses the R-R interval ro etarr- disposable endotracheal rube.
obtain l-reart rate measurements. ,lterial
supplier. Coolled Corp. [101345], Utica, pressure data is obtained from an esisting
\Y (US-\); +1 (800) 4-18-6s06, +1 (31s)
t It.i
arterial pressure Line and, in addition to car-
7 97 -831 5 ; g'rvs,. conmed.com diac output, facilitates estimation oi other
Procedure. f'he EC()]I (rvhich stands for hemodvnamic pararneters such as s1'stemic
Endotracheal Cardiac ()utput \Ionitor) is vascular resistance and stroke volume rzria-
a minimallf inrasive srstem that measures tion (neit1-rer of u'hich can be obtained from
cardiac output and other related hemodi-- ICG).;\t the beginning oF the procedure,
namic parameters using a combination o[ the clinician must enter the patientt height, :3
trvo different techniques: ICG and arteriai rveight, and age, rvhich are used along with
s.atetbrm analvsis. The monitor obtains the measured ICG signal to estimate suoke ConMed ECOM. (lmoge courtesy of
an impedance rvaveform rvith the ICG 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"
Evidence. \\e revierved eight clinical srud- the EC()f t monitor and the reference Procedure. The FloTrac sensor uses
ies that compared the EC()\I monitor to techniques) and rs'o studics (Jorgensen arterial rvaveflorm analrsis to estimate
other methods of cardiac output measure- er a|.2012, \Ioller-Soreosen et al. 2012) continuous cardiac output using the Flo-
ment. The srudies involved cardiac surgerr reporting poor trending. livo srudies Trac algorithm. -\ccordiog to the vendor,
and abdominal surgerv patients, and the (Bairamian et al. 20i 1, Fellahi et al. 2012) the algorithm constantlv monitors rhe arte-
number of patients in the srudies ranged conclude that the techoique mav be useful rial pressure vaveform and compensates
irom 12 to'10. In five of the eight snrdies, for predicting fluid responsiveness (using for changes tn arterial f aottic characteris-
the ECOII monitor was compared to pul- stroke volume rariation). Trvo other srud- tics. The FloTrac sensor can be used with
monal\. arter\' thermodilution; these studres ies @all et al. 2010, ]Ioller Sorensen et al. either thc E\-1000 or thc \-igilco rnonitor
are outlined in the table on this page. In 20i2) tbund that the EC()\I \i,as suscep- offered bv Eds'ards Lit'esciences.
the remaining three srudies, the reierence tible to interference from electrocauterr- The FloTrac sensor can be used onlv
methods rvere transpulmonarr- thermodilu- equipment used during surgeries, leading on patients rv-ho har.e an existing arterial
tion (rvhich uses a central venous catheter to intermittent loss of signal during elec- line that can be connected to the sensor.
rather than a P-\C) (F'ellahi er aL.201.2), trosurger\' and somedmes several seconds The sensor has nvo output lines: ()ne is
esophageal Doppler fl orgensen er al. 2Ol2), after electrosurger\- \'v?s discontinued. connected to the arterial pressure cable
and the FloTrac sensor @airamian et al. \\b d-rd not 6nd anr studies ad&essing of a bedside monitor, and the other is
201 1; see our rerierv of this slstem belorv-). c[oical outcomes. connected to the E\-10o0/\'igileo moni-
In addition, three ol the eight srudies tor. The sensor, once connected, sends
(Bairamian et al. 2011, Gennart et aL.2012, arterial pressure rv'aveform information
\hn der Kieij et al. 201,2) rvere reported as to both monitors. The cardiac output
conference abstracts onlr; therefore, ther monitor (either the E\'1000 or the \-igileo)
ma\- not rePresent the 6nal results and con-
uses this informat.ion in coniunction with
clusions oi the srudies. patient demographic information entered
In all eight srudies, the authors con- Supplier. Edrrards Li[esciences Corp. br the user (e.g., age, gender, height,
cluded that the device \\,-as not comparable [374501], Irvine, C-\ (US.\); +1 (800) 124- rveighr) to esrimate cardiac output.
to the reference method. The evidence , .,(e-+e) 250-2500; rwwv.edrvards. \\'e covered the FloTrac sensor in our
regarding the monitort abilirl'to trend ::f December 2009 Guidance -{.rticle. Since
cardiac output is inconsistent, rvith one then, the vendor has updated the device's
srudv (\Iaus et al. 2011) reportiog good algorithm. The hterarure review-ed in the
'.\lthough thc authrlrs usc rhe rcrm "clcctrocrutcn,"
trending (based on the correlatjon benveen
*cloelierc tlrc intcnt wrs "clcctrcsurgcn."
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.
020]3 ECRI lnslitute. Member hospitols moy reproduce this poge for inlernol dislribulion only. w.eri.ors > HEALTHDEVICES JuNE2Ol3 183
10 of 158.
gaidance
ot
Procedure.
that compared the YolumeYiew sensor \on Invasive Cardiac Svstem) consists
rvith other cardiac outpur monitoring tech- of a laptop aod one pair oF s'rist-ankle
niques in human patienrs. Costa er al. ICG electrodes. The st'stem is based on
t*,
(2012) compared rhe cardiac outpur the ICG technique, rvhich is based on the
theorl that changes in the voltage oF an
electrical signal applied across an arcz of
- .\ salinc solutioo of kn()\\-o tcmpcriturc is injcctcd the bodr are due to changes in impedance
Edwords Lifesciences VolumeView sensor. into ao cxistinla ccntral rcnous cathctcr, end rhc chengc (rvhich in rurn are associated *'ith changes
(lmoge courtesy of Edwords Lilesciences.) in rcmpcmturc of thc injcctatc is mcasured in thc [cmo, in blood volume during each cardiac clcle).
nl artcrr u:ing thc \irlumc\iicrv crtheter,
184 xglrraDEvlcts JUNE2or3 > ww.*ri.ors @2013 ECRI lnstiiute. Member hospitols moy reproduce ihis poge for inlernol disiribution only.
11 of 158.
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.
Cotter, Torre-Amiot, 93 polienls wilh heorl Bolus pulmonory ortery With o correlotion coefficient o{ r = 0.8.l, the two methods were
e1 ol. 2004 foilure thermodilution considered comporoble
Cotter et o]t.2006 43 cordioc potienis Bolus pulmonory ortery o correlotion coeflicieni oi r
Vy'ith = 0.9. ihe two meihods were
ihermodilulion considered comporoble.
Goor 2006 (con{erence 46 cordioc polienls Thermodiluiion* Wilh o percenioge error ol 2OYo, lhe two methods were consid-
obsiroct) ered comporoble.
Poredes et ol. 2006 35 cqrdioc potients 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.
' The obslroct does nol specify whot \pe of thermodilulion 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
mulriccntcr r alidation oI thc third-gcncration (irxrr l) l'cripheral rersus thoracic impcdance c,rrdi-
ISairemianJIr, RrxrkJI:, :\llard \IU, et al. (irmpan-
strfnvare in scptic Paricnrs. InlLn;irt (.are -\[td (Uriphr lconfcrcncc abstncrl. | ( ,trdiu, l iil 20{)6
son of stnrkc trtlumc rariation obtained from
an arrcrial prcssurc to cndorraclteal sensor svs- 2t)1 1 Itb;37(2):233--11). .\ug; 12(6 Suppl):S5J-5.
tcm. 2(111 .\nnual \[ccting of thc Internatir>nal li(iltl Insriturc: J<trgcnscn I IK. llisg.rarJ.l, ()ilsm l. (-,'mp.rrison oi
-\ncsthcsia Rcsearch Socicq'; 201 1 \Ia1' 21 -2-l; bioimpcdance and ocsophagcal I)t>pplcr cerdiac
I:ndotilcheal crrJirc ourpur mrtnirrr st stem
\hncourcr (Oanada). output monitoring during abdomin:rl atlrtic sur-
(Oon\led (lorp.) tor dctermining carditc output
\, of ger.t-. Cit (,an 20t2;16(SuPPl l).
tlall'I'lt, (iulp l)(i, Patel cr al. (irmparison and strokc r-olume lcusrom product brict-1.
tlrc cndotruchcal crrditc tlurput m<rnilor ltr (lustom I Iotlinc Se n-icc. 3)l2 l)ec -|. Khs-aonimit B, Bhuralanontachai l{. Prcdicrion oi
thcrmodilution in canlirc surgcn plicnts lilo'I'mc scnsor Qidrvards Litlscicnccs, l.l.(-) ftrr Buid rc-iponsircncss in scptic shock patienrs:
.l
(ttli,rlrird: I,b;; zlnt;th 2(llo ()cl.21(5):1(t2-6. noninr-asilc cardiac output mrlnitoring lcustom comprring -strokc rrrlumc r trriarirtn br likilrac/
Bcncs J, Oh1'tn I, .Utmann I et al. Inrnropcradr c 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.
fltrid optimization using strokc r-olumc varia-
tion in high risk surgical paticnts: results oI \oninr-asilc cardiac slstcm (\I \Icdical. Lrd) Kre fcr\, I krtcr CK, \Iirn (i, et al. (llinicLrl v'rlida-
prospcctirc randomizcd srudt (n7 (-az [<rr monitoring cardiac ()utPut lcustom pnrduct rion of a nc$ the rm()dilution svstcm [<rr thc
2{)1{};l-1(3):Rl lll. bncf]. Custom I Iotlinc Scn-ice. 2012 l)ec 5. asscssment of crrdilc ourput rnd 11)lumctdc
lliancofiorc G, Critchlcl'L.\, Lcc .\, ct al. Iilalua- Stra(l:r from thc hcart: undcrstanding cardilc paramc tcrs. ( i t ( -a n 2tl 12 \ Iav 3{); I 6(3) :ll'9tt.
tion of a ncs' sr>fru-arc tersitln of rhc lrlo l rac/ ourput monitr>ring tcchniques lguidancc erticlel. Kor,rkc \', \imada'I', \agata I I, ct al. 'liansicnt
Yigileo (r-crsion 3.1)2) and a comparison rvith I kt lr h l) t ir i 2009 Dec;3ll(l 2):.3tt6--lt) l. hcmodvnamic change and xccurJc\'oI arrcrial
prcr-ious data in cirdrotic paticnts undcrgo- YolumcYics set ([idrvards Litcscicnccs, I-1.(-) blrxrd prcssurc-bascd cardi,rc ()utPut.. lrc'l,
ing fir-er trrnsplanr rurgc^-- /1n.ith -'lrulg2o11 [r>r minimellt inr rsilc crrdilc output moninrring ,lnd32ttl t .\ug;t 13(2):272-a.
Sep;1 13(3):515-22. [custom producr brict-1. Cusrom I lorlinc -\cil'ice . Kusaka \', \irshit;rni K, Iric '1, ct irl. Clinical compari-
Broch () ltcnnerJ, Clruencrvald ]I, ct al- -\ compad- 201-l I'eb 15.
s()n oi rn cchocardiograph-dcni-ed r-crsus pulsc
son of third-gcncrarion scmi-inrasirc ilrrerial l;cllahi-ll-, |ischer \l() Rebct () ct al. -\ comprrison counte r-dcrir'.d c.trdi.rc outpur mcxsurLmrnt in
rvavcftrrm analvsis rrith rhermodilution in oI cndotracheal bioimpcdance cardi<>gr'rphr abdominrl ?()rric xntutrsm surgcr.t. | (trdiotlttrt'
patients undergoing coronatr surgcn'. and transpulmonarr thcrmodilution in ctrdi:c l. ; t .. I n, t h 2tt I 2 .\pr,26(2) :223' (,.
.\' Licttti/itlUbiltlurnul 21l2;2tt 12:15 lt)81. ll lpub 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
2{)l 2 Jul 31 .l .\pr'26(?):217-22.
mcasuremcnt of crrdirc ourput bv rvholc-bodr
Costa (1, Cccconct'l', Baron Q et al. Clrdiac out- (]cnnart I;, lJeckcrs \trborgh
O, et al. Impacr
S, bi<>impedlncc during doburaminc strcss cclttt-
put moniroring in cirilrotic paticnts: IiYl()(Xl of;rrtcrial carhcrcr krcation tln thc
accuritcr' crrdiogr.rpltr: clJnrc,rl implic.trioni ln Prticnt j
r-crsus pulmonry arrerv catltctcr--prcliminarv of cardiac ourput proi-ided bl an cnd()trachcal s ith Ic[t rrntricul;rr dtsfunction and ischaemia.
data fc<rn[crcncc abstractl. lt (rili;e! (qrt bioimpcriancc dclice [c<>nfcrcncc abstr,rcrl. Cz7 I:Lr I I hnrt litil2t){)6 \lar;tt(2):13(r'l{}.
(onfirc tt. 32nd Intcrnational -svmPosium (rr, 2{)12; t(r(Suppl l).
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 DO YOU HAVE A QUESTION ABOUT CARDIAC OUTPUT
(irrtcr G, Ikrshkoritz \ Kaluski Ii, ct al. -\ccunte, MONITORING?
nonim-asir-e continuous monitoring of cadiac
outpur bv s'holc-bodv clectrical bioimpcdancc. For exomple, ore you wondering . . .
bodv clcctrical bio-impedrncc is rccururc You con get onswers to these ond couniless other quesiions through HD Advisor. Just
in non invasirc de tcrmination oI cardiac click the link in lhe top righi corner o{ your Heolth Devices home poge to occess your HD
output: a tlrermodilution contnrllcd, PmsPec- Advisor doshboord ond storl your inquiry.
tivc, d<rublc blind cr aluation. I lr.l I lurt I iil
@ADVISOR
.\' 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 Ask the [xperts
pressurc-bascd canJiac output monitoring: a
186 xrarrxDEvlcEs JUNE2ol3 > ww.Bri.org O2O'13 ECRI lnsiitute. Member hospitols moy reproduce this poge for internol distribution only'
13 of 158.
\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- I Uin .\[uir (tn\ar 2013 .\pc27(2):l7l-7.
dcrcrmination lnrm cndorrirchcirllr mcasured epincplrrine in cridcalll ill p^ticnts. Br.[ ,Arp]rilh Su l](-,'t-.ai Yl', Chcn C\', ct al, (lardiac ()utput de-
impedancc cardiographv: cljnical evaluarion of 2r) I 2 .\or;1 0tt(.l) :61 5-2?. riled fiom artcrial pressurc q-ar-cfrrrm analvsis in
cndotrechcrl cardiac output monitor. / (,tnlilla- padcnrs undcrgoing lir-cr transplantation: i'alidiu
\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. oi r third-gcncrarion dcticc. 'l-ratpltnl Prut2ll12
echocardiognphic cr-aluation r-crsus vascular
\Iar;$t(2):'12-l lt.
\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../ 'lsai Yl; Su lXl, Lin (l(1, ct al. O;rrdirc ourput dcrii-td
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. trom arrcrial prcssure t avciorm enallsis: r alidl-
pital stav in high-nsk surgical paticnts: r rand<>m- trrn oi tlrc rhir,.l-gcncr:tion sol-ro'rre in peticnrs
\lutoh 'I, Ishikarra'l', Kobalashi S, ct al. l)ertirr-
ized, controlled tnal. (. i r ( ut 2o ll ); 1 -t(1 ) :l{ 1 tl. undcrgoing r>rthotopic lir-er rransplantation.
mancc o I tlrird-gencrati<>n Irkiliac/\'igilc<r
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,\
\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
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
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 RHC is predictive of survival and effective in a goal-oriented
disease characterized by elevated pulmonary vascular treatment strategy,3,4 and has been recommended by a recent
resistance (PVR) because of pulmonary vascular re- guideline;1 however, there are some disadvantages in the reg-
modeling. This leads to a decrease in cardiac output (CO) and ular use of RHC as a follow-up procedure, especially with
ultimately death. Recently, targeted medical therapy for PAH regard to invasiveness. Noninvasive and less complicated
patients with endothelin-receptor antagonists, phosphodies- methods for monitoring hemodynamics are needed to manage
terase-5 inhibitors, and prostacyclin analogs has been estab- patients with pulmonary hypertension (PH). Less invasive
lished,1 and the prognosis of PAH has improved.2 However, hemodynamic monitoring has recently been suggested as fea-
there is no universally accepted consensus on the treatment sible in some situations.5 The Non-Invasive Cardiac System
goals or follow-up strategy for PAH patients. Right heart cath- (NICaS; NI Medical, Hod-Hasharon, Israel) is a device for
eterization (RHC) is not only the gold standard for the diagno- calculating CO noninvasively with whole-body impedance
sis of PAH, but is also a useful tool for monitoring PAH, and cardiography (ICGWB). The NICaS-derived CO (NI-CO) has
is recommended 36 months after new treatments and in the been shown to be as reliable as the RHC-derived CO and is
case of clinical worsening.1 Hemodynamic monitoring with 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.
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 Measurement of PVR
multiplied by the heart rate. Because the NI-CO values are PVR (dyns1cm5) was calculated using RHC from the equa-
calculated every 20s, the average of 3 measurements obtained tion:
consecutively during 60s of monitoring is considered to be the
PVR=80(mPAPPCWP)/COTD, Fick (TD-PVR, Fick-PVR).
NI-CO value for each individual case.
PVR was also estimated noninvasively using a combination
Echocardiography of NICaS and echocardiography, and by echocardiography
Echocardiography was performed using a Vivid 5 system and alone.10 Echo-mPAP was calculated as Echo-sPAP0.61+
a 3.5-MHz transducer (GE Vingmed Ultrasound AS, Horten, 2mmHg, as previously described,11 and noninvasive PVR was
Norway). Two-dimensional Doppler examinations were per- calculated as 80(Echo-mPAPPCWP)/COEcho,NI (Echo-PVR,
formed in the usual manner. CO was measured by tracing the NI-PVR). PCWP for the calculation of noninvasive PVR was
left ventricular ejection flow (Echo-CO). Echo-sPAP was es- estimated at 10mmHg in all cases.10
timated from the peak velocity of the tricuspid regurgitation
jet plus estimated right atrial pressure (Echo-RAP).9 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 jects for TD-CO, Fick-CO, Echo-CO, and NI-CO were 4.92
Echo-mPAP and mPAP measured by RHC (RHC-mPAP) 1.56L/min, 3.871.24L/min, 4.341.11L/min, and 4.40
were determined by calculating the Spearmans rank corre- 1.32L/min, respectively (Table2). A significant and very
lation coefficient. P<0.05 was considered to be significant. strong correlation was observed between TD-CO and NI-CO
Agreement between methods was analyzed by the Bland-Alt- (r=0.715, P<0.0001) and between TD-CO and Fick-CO (r=
man method.12 The limits of the agreement were expressed as 0.795, P<0.0001) by 2-tailed Spearmans rank correlation test
the meanSD. The 95% confidence intervals (CIs) of the bias (Figure1). There was a strong correlation between Fick-CO
were also calculated. Receiver-operating characteristic (ROC) and NI-CO (r=0.653, P<0.0001). However, the correlation
curves were generated for the detection of elevated PVR de- between Echo-CO and TD-CO or Fick-CO was significant but
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- differences between 2 measurements were plotted according
ity were estimated by the ROC curves. All statistical analyses to the Bland-Altman method (Figure2). The mean bias and
were performed using GraphPad Prism version 5 (GraphPad limits of agreement between TD-CO and NI-CO, Fick-CO and
Software, La Jolla, CA, USA). 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-
Results ment between TD-CO and Echo-CO, and Fick-CO and Echo-
The baseline characteristics of all patients at initial hospital- CO were 0.641.33 (1.97 to 3.26) L/min and 0.421.18
ization are summarized in Table1. Approximately two-thirds (2.73 to 1.88) L/min, respectively. There was no clear differ-
of the patients had PAH (World Health Organization [WHO] ence in the measurements of CO among the patients with id-
classification of PH group 1) and the other one-third of the iopathic PAH, collagen tissue disease associated PAH or
patients had chronic thromboembolic PH (CTEPH: WHO CTEPH (FigureS1).
group 4); 5% of the patients were classified as WHO group 3.
At enrollment, 24% of the patients were receiving medical Comparison of Invasive and Noninvasive Measurement of
therapy. mPAP and PVR
The mean values of all measurements of invasive mPAP and
Relationships Among Parameters Echo-mPAP were 32.91.28mmHg and 43.01.59mmHg, re-
The mean CO values from all measurements in these sub- spectively. There was a very strong correlation between inva-
18 of 158.
Advance Publication by-J-STAGE
Noninvasive Assessment of PH
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 be determined appropriately. This may have resulted in bias in
the presence of low CO, consistent with previous reports.23 the measurement of NI-PVR. In our study, the Doppler param-
eter needed in order to estimate NI-PVR was only TRPG, and
Study Limitations there was no patient in whom we were unable to obtain that
The main limitation of this study was the need to measure the value. Second, we used the conventional invasive equation for
Doppler parameter for estimating NI-PVR. Proper alignment estimating NI-PVR. We had to estimate PCWP at 10mmHg
of the ultrasound beam is crucial for the Doppler parameter to 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 8. Cohen AJ, Arnaudov D, Zabeeda D, Schultheis L, Lashinger J,
wide variation in PCWP is not usually observed among pa- Schachner A. Non-invasive measurement of cardiac output during
coronary artery bypass grafting. Eur J Cardiothorac Surg 1998; 14:
tients with PH. Third, because CO measurement using NICaS 6469.
in patients with cardiac shunts is known to be unreliable,24 we 9. Yeo TC, Dujardin KS, Tei C, Mahoney DW, McGoon MD, Seward
excluded cases of PAH associated with cardiac shunts. Fourth, JB. Value of a Doppler-derived index combining systolic and dia-
noninvasive estimation of CO and PVR with NICaS was fea- stolic time intervals in predicting outcome in primary pulmonary
sible; however, there were some patients who had large diver- hypertension. Am J Cardiol 1998; 81: 11571161.
10. Lindqvist P, Soderberg S, Gonzalez MC, Tossavainen E, Henein MY.
gence between NI-CO or NI-PVR and invasive CO or PVR. Echocardiography based estimation of pulmonary vascular resistance
Further studies are needed to clarify the factors that lead to in patients with pulmonary hypertension: A simultaneous Doppler
inaccurate measurements of CO and PVR. Fifth, in our study, echocardiography and cardiac catheterization study. Eur J Echocar-
the number of patients with WHO functional class 4 was diogr 2011; 12: 961966.
11. Chemla D, Castelain V, Humbert M, Hebert JL, Simonneau G,
small. Most patients were WHO functional class 2 or 3. The Lecarpentier Y, et al. New formula for predicting mean pulmonary
reliability of NICaS in patients with severe PH is to be exam- artery pressure using systolic pulmonary artery pressure. Chest 2004;
ined in future studies. Finally, the study sample size was rela- 126: 13131317.
tively small and originated from a single center. We believe 12. Bland JM, Altman DG. Statistical methods for assessing agreement
between two methods of clinical measurement. Lancet 1986; 1: 307
that a larger, multicenter study is needed to appropriately 310.
confirm the reliability of the method. 13. Mantha S, Roizen MF, Fleisher LA, Thisted R, Foss J. Comparing
Although recent advances in treatment options and manage- methods of clinical measurement: Reporting standards for bland and
ment have improved the outcomes for patients with PH, treat- altman analysis. Anesth Analg 2000; 90: 593602.
14. Inaba T, Yao A, Nakao T, Hatano M, Maki H, Imamura T, et al.
ment goals and follow-up strategy are still not well defined. Volumetric and functional assessment of ventricles in pulmonary
Hemodynamic monitoring with RHC is recommended in a hypertension on 3-dimensional echocardiography. Circ J 2012; 77:
goal-oriented treatment strategy for PH1 and is the gold stan- 198206.
dard for the assessment of PAH; however, the invasiveness of 15. Garcia-Alvarez A, Fernandez-Friera L, Mirelis JG, Sawit S, Nair A,
RHC is a critical factoring its regular use as a follow-up pro- Kallman J, et al. Non-invasive estimation of pulmonary vascular resis-
tance with cardiac magnetic resonance. Eur Heart J 2011; 32: 2438
cedure. A noninvasive, accurate, and simple method is re- 2445.
quired for the management of patients with PH. We have 16. Abbas AE, Fortuin FD, Schiller NB, Appleton CP, Moreno CA, Lester
demonstrated noninvasive measurement of CO and PVR using SJ. A simple method for noninvasive estimation of pulmonary vas-
only simple parameters. Echo-sPAP is needed to estimate cular resistance. J Am Coll Cardiol 2003; 41: 10211027.
17. Kang KW, Chang HJ, Kim YJ, Choi BW, Lee HS, Yang WI, et al.
PVR, but Echo-sPAP has been established as a simple, reli- Cardiac magnetic resonance imaging-derived pulmonary artery dis-
able screening parameter for PH.25 This noninvasive, reliable, tensibility index correlates with pulmonary artery stiffness and pre-
and simple assessment can be a useful tool for monitoring and dicts functional capacity in patients with pulmonary arterial hyper-
managing patients with PH. 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-
Conclusion gest Heart Fail 2004; 10: 710.
Noninvasive measurement of CO and PVR using NICaS is as 19. Raaijmakers E, Faes TJ, Scholten RJ, Goovaerts HG, Heethaar RM.
A meta-analysis of published studies concerning the validity of tho-
reliable as invasive RHC. This simple assessment could help racic impedance cardiography. Ann NY Acad Sci 1999; 873: 121127.
physicians to manage their patients with PH. 20. Fisher MR, Forfia PR, Chamera E, Housten-Harris T, Champion HC,
Girgis RE, et al. Accuracy of Doppler echocardiography in the he-
Disclosures modynamic assessment of pulmonary hypertension. Am J Respir Crit
Care Med 2009; 179: 615621.
None. 21. Kouzu H, Nakatani S, Kyotani S, Kanzaki H, Nakanishi N, Kitakaze
M. Noninvasive estimation of pulmonary vascular resistance by Dop-
References pler echocardiography in patients with pulmonary arterial hyperten-
sion. Am J Cardiol 2009; 103: 872876.
1. Galie N, Hoeper MM, Humbert M, Torbicki A, Vachiery JL, Barbera 22. Selimovic N, Rundqvist B, Bergh CH, Andersson B, Petersson S,
JA, et al. Guidelines for the diagnosis and treatment of pulmonary Johansson L, et al. Assessment of pulmonary vascular resistance by
hypertension: The Task Force for the Diagnosis and Treatment of Doppler echocardiography in patients with pulmonary arterial hyper-
Pulmonary Hypertension of the European Society of Cardiology (ESC) tension. J Heart Lung Transplant 2007; 26: 927934.
and the European Respiratory Society (ERS), endorsed by the Inter- 23. Tournadre JP, Chassard D, Muchada R. Overestimation of low car-
national Society of Heart and Lung Transplantation (ISHLT). Eur diac output measured by thermodilution. Br J Anaesth 1997; 79: 514
Heart J 2009; 30: 24932537. 516.
2. Benza RL, Miller DP, Barst RJ, Badesch DB, Frost AE, McGoon 24. Kauppinen PK, Koobi T, Hyttinen J, Malmivuo J. Segmental com-
MD. An evaluation of long-term survival from time of diagnosis in position of whole-body impedance cardiogram estimated by com-
pulmonary arterial hypertension from the REVEAL Registry. Chest puter simulations and clinical experiments. Clin Physiol 2000; 20:
2012; 142: 448456. 106113.
3. Fukumoto Y, Shimokawa H. Recent progress in the management of 25. Taleb M, Khuder S, Tinkel J, Khouri SJ. The diagnostic accuracy of
pulmonary hypertension. Circ J 2011; 75: 18011810. Doppler echocardiography in assessment of pulmonary artery systolic
4. Hoeper MM, Markevych I, Spiekerkoetter E, Welte T, Niedermeyer pressure: A meta-analysis. Echocardiography 2013; 30: 258265.
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. Supplementary Files
6. Paredes OL, Shite J, Shinke T, Watanabe S, Otake H, Matsumoto D, Supplementary File 1
et al. Impedance cardiography for cardiac output estimation: Reliabil-
FigureS1.Linear correlation between NI-CO and TD-CO (A) or
ity of wrist-to-ankle electrode configuration. Circ J 2006; 70: 1164
Fick-CO (B) in IPAH ( , red line), CTD-PAH ( , blue line) and
1168.
CTEPH ( , green line).
7. Tanino Y, Shite J, Paredes OL, Shinke T, Ogasawara D, Sawada T, et
al. Whole body bioimpedance monitoring for outpatient chronic heart Please find supplementary file(s);
failure follow up. Circ J 2009; 73: 10741079. http://dx.doi.org/10.1253/circj.CJ-13-0172
21 of 158.
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
Published by:
http://www.sagepublications.com
Subscriptions: http://prf.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.com/journalsPermissions.nav
What is This?
Perfusion
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 incorporates cardiac output as well as systemic pressures,
heart to provide enough blood flow to the systemic circula- this distinguishes it as a rational marker of cardiac perfor-
tion. Cardiac power output measured at exercise reflects mance when compared to myocardial oxygen consump-
the peak cardiac performance attainable by the heart. tion. Cardiac power output has been used to assess cardiac
Comparing maximum cardiac power output with resting performance and prognosis in cardiogenic shock and acute
cardiac power output represents the hearts cardiac reserve. and chronic heart failure patients.3,4 The purpose of this
In heart failure patients, cardiac reserve is severely limited.1 study was to retrospectively analyze the application of car-
Patients performing exercise tests or positive inotrope infu- diac power output to stage the severity of heart failure in
sion tests tend to show little difference in cardiac perfor- patients who have received mechanical circulatory support.
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- Methods
gresses in these patients, cardiac reserve continues to decline.2 Twenty-eight patients, twenty-four male and four female,
Once cardiac performance can no longer be improved with a history of heart failure were referred to the
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 Circulatory Sciences Graduate Perfusion Program, The University of
receive mechanical circulatory support, it is important to Arizona, Tucson, Arizona, USA
implant devices before heart failure has become so severe
Corresponding author:
that irreparable systemic damage has occurred. Symptomatic Doug F. Larson
assessment, drug tolerance, and myocardial oxygen con- 4402 Arizona Health Sciences Center
sumption are commonly used to determine the severity 1501 North Campbell Avenue
of heart failure, but are unable to provide all the informa- Tucson, AZ 85724, USA
tion necessary to reflect true cardiac reserve. Myocardial Email: dflarson@u.arizona.edu
oxygen consumption provides an indirect measurement of Presented at the 33rd Annual Seminar of The American Academy of
cardiac output; however, given that cardiac power output Cardiovascular Perfusion, New Orleans, Louisiana, 26-29 January 2012
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 Abbreviations. BSA: body surface area; ICD: implantable cardioverter de-
Statistical Software Stata 11 (StataCorp LP, College Station, fibrillator; HF: heart failure; HR: heart rate; MAP: mean arterial pressure;
SVR: systemic vascular resistance; CVP: central venous pressure; PCWP:
TX, USA). An independent two-sample t-test with unequal pulmonary capillary wedge pressure; LVSWI: left ventricular stroke work
variance was used to identify if contractile reserve still index; EF: ejection fraction; CO: cardiac output.Values are expressed as
existed in the failing heart. Mechanical circulatory support mean SEM or number (% within studied population).
in both the survived and deceased groups when com- Cardiac power output was used to calculate cardiac
pared across all three levels of drug therapy. Mean arterial power index to adjust for blood volume distribution
pressure of the deceased group declined despite ino- across the body surface area. The baseline administered
tropic drug intervention. Those in the survived group inotropic dose increased cardiac power index, but
maintained a consistent level of mean arterial pressure revealed no statistical difference when compared to the
across all three levels of drug therapy. deceased group (Figure 1F). When peak inotropic doses
Stroke volume was well below the healthy adult ranges were administered, cardiac power index continued to
(Figure 1D), but was not found to be statistically signifi- decline in the deceased group, but was able to distin-
cant between the survived and deceased groups. guish the contractile reserve between the two groups.
Stroke volume did not increase in the deceased group The mean cardiac power index of patients who survived
after the initial inotropic drug therapy; however, when was 0.34 0.07 W/m2 at peak inotropic drug therapy. A
peak inotropic drug therapy was administered, stroke cardiac power index below 0.34 W/m2 identified patients
volume decreased. On the other hand, the survived who did not survive ninety days past the implantation
group displayed an increase in stroke volume at each of mechanical circulatory support because all patients
level of drug therapy. in the deceased group had a cardiac power index below
Left ventricular ejection fraction (Figure 1C) derived 0.34 W/m2. Comparing cardiac power index between
from echocardiography was not significantly different the survived and deceased groups showed significantly
between the survived or deceased groups. When increased mortality rates once values fell below this
placed on the initial inotropic treatment, both the sur- point (p<0.05).
vived and deceased groups had an increased ejection
fraction; however, the deceased was less (31.50 7.5% Discussion
to 22.76 13.17 %). There was a large variance in the
deceased group due to a select few having diastolic Current hemodynamic guidelines use left ventricular
heart failure with preserved ejection fraction. Those with ejection fraction and myocardial oxygen consumption as
diastolic heart failure had poor cardiac performance, but criteria to assess cardiac and systemic degeneration for
because their chamber size was unchanged or even possible mechanical circulatory support. As shown by
decreased despite ventricular wall thickening, the overall Marmor et al., the present study demonstrates that ejec-
percentage of blood ejected from the heart with each tion fraction did not display any significant differences
contraction was still elevated due to a decrease in overall between the survived and deceased groups during
ventricular chamber size. maximal cardiac performance resulting from peak ino-
Cardiac output was used to calculate cardiac index in tropic drug therapy.9 In addition, results from prior
order to account for the distribution of blood across the myocardial oxygen consumption studies on patients in
body surface area. In Figure 1E, the survived and end-stage heart failure have shown to be inconclusive
deceased patients presented with a cardiac index of when peak cardiac function was measured.3,10,11 Patients
1.65 0.83 and 1.52 0.13 L/min/m2, respectively. With with NYHA classes III and IV symptoms are severely
the initial inotropic treatment, the survived groups compromised when tested physically for myocardial oxy-
cardiac index increased to 2.14 0.54 L/min/m2, but gen consumption due to progressive muscle decon-
remained constant with the peak inotropic treatment at ditioning, poor vascular circulation, compromised lung
2.17 0.47 L/min/m2. The deceased group displayed a capacity, and anemia. Although ejection fraction and
moderate increase in cardiac index during the initial ino- myocardial oxygen consumption tests have their respec-
tropic treatment (1.86 0.62 L/min/m2), however, this tive roles in measuring specific parameters of heart func-
decreased on peak inotropic support (1.69 0.26 L/min/ tion, they may not be completely reliable assessments of
m2). Cardiac index was not a functional hemodynamic cardiac function in these severely compromised patients.
marker to identify if contractile reserve remained or Right heart catheterization provides very useful infor-
whether it was able to detect survival outcomes between mation in gathering pressure and volume changes in spe-
the survived and deceased groups. cific areas of the heart. This method requires that patients
be placed in a supine position without simulating peak reserve. Resting values alone do not provide a sufficient
cardiac contractility. This is a significant limitation, as amount of information on cardiac performance in healthy
assessing heart function in this manner does not allow for versus diseased hearts; however, positive inotropic stimu-
measuring the difference between basal and maximal car- lation that achieves peak contractility is when diseased
diac performance in order to determine true cardiac hearts can become distinguishable from healthy hearts.
a noninvasive approach for assessment of contractile reserve. (recovered) patients, and those with moderate to severe
J Nucl Med 1993; 34: 18771885. heart failure. Am J Cardiol 2010; 105: 17801785.
10. Leitman M, Sucher E, Kaluski E, et al. Non-invasive meas- 14. Jakovljevic DG, Donovan G, Nunan D, et al. The effect of
urement of cardiac output by whole-body bio-impedance aerobic versus resistance exercise training on peak car-
during dobutamine stress echocardiography: clinical diac power output and physical functional capacity in
implications in patients with left ventricular dysfunction patients with chronic heart failure. Int J Cardiol 2010; 145:
and ischaemia. Eur J Heart Fail 2006; 8: 136140. 526528.
11. Wilson JR, Rayos G, Yeoh TK, Gothard P, Bak K. 15. Shelton RJ, Ingle L, Rigby AS, Witte KK, Cleland JG, Clark
Dissociation between exertional symptoms and circula- AL. Cardiac output does not limit submaximal exercise
tory function in patients with heart failure. Circulation capacity in patients with chronic heart failure. Eur J Heart
1995; 92: 4753. Fail 2010; 12: 983989.
12. Marmor A, Schneeweiss A. Prognostic value of nonin- 16. Nicholls D, O'Dochartaigh C, Riley M. Circulatory power-
vasively obtained left ventricular contractile reserve in -a new perspective on an old friend. Eur Heart J 2002; 23:
patients with severe heart failure. J Am Coll Cardiol 1997; 12421245.
29: 422428. 17. Agostoni P, Cattadori G, Apostolo A, et al. Noninvasive
13. Jakovljevic DG, George RS, Donovan G, et al. Comparison measurement of cardiac output during exercise by inert
of cardiac power output and exercise performance in gas rebreathing technique: a new tool for heart failure
patients with left ventricular assist devices, explanted evaluation. J Am Coll Cardiol 2005; 46: 17791781.
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
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.
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 3
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 [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.
This study was supported by NI Medical Ltd. Israel. [10] Cotter G, Schachner A, Sasson L, Dekel H, Moshkovitz Y. Impedance cardiography
The authors of this manuscript have certied that they comply revisited. Physiol Measures 2006;27:81727.
with the Principles of Ethical Publishing in the International Journal of [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:
Cardiology [20]. 11648.
[12] Goor DA, Frinerman E, Granov I, Patterson R. Peripheral versus thoracic impedance
References cardiography. J Cardiac Failure, Heart Failure Society of America (HFSA), 10th
Annual Scientic Meeting, September, 1013; 2006. p. S545.
[1] Rosamond W, Flegal K, Friday G, et al. Heart disease and stroke statistics: 2007 [13] Rozenman Y, Goor DA, Rotzak R. Progress report on non-invasive diagnosis of
update: a report from the American Heart Association Statistics Committee and asymptomatic left ventricular systolic dysfunction. Abstract J Card Fail 2009;15:
Stroke Statistics Subcommittee. Circulation 2007;115:e69e171. 68.
[2] Goldberg LR, Jessup M. Stage B heart failure: management of asymptomatic left [14] Gottdiener JS, Bednarz J, Devereux RD, et al. American Society of Echocardiogra-
ventricular systolic dysfunction. Circulation 2006;113:285160. phy: recommendations for use of echocardiography in clinical trials, a report from
[3] Schocken DD, Benjamin EJ, Fonarow GC, et al. Prevention of heart failure: a the American Society of Echocardiography's Nomenclature and Standards
scientic statement from the American Heart Association Councils on Epidemi- Committee and the task force on echocardiography in clinical trials. J Am Soc
ology and Prevention, Clinical Cardiology, Cardiovascular Nursing, and High Blood Echocardiogr 2004;17:1086119.
Pressure Research; Quality of Care and Outcomes Research Interdisciplinary [15] Hardy CJ. Assessment of arterial elasticity by cardiovascular MRI. In: Kwong RY,
Working Group; and Functional Genomics and Translational Biology Interdisci- Libby P, editors. Cardiovascular magnetic resonance imaging (contemporary
plinary Working Group. Circulation 2008;117:254465. cardiology). Totowa, New Jersey: Humana Press; 2008. p. 695710.
[4] Wang TJ, Evans JC, Benjamin EJ, Levy D, LeRoy EC, Vasan RS. Natural history of [16] Gheorghiade M, Sopko G, De Luca, et al. Navigating the crossroads of coronary
asymptomatic left ventricular systolic dysfunction in the community. Circulation artery disease and heart failure. Circulation 2006;114:120213.
2003;108:97782. [17] Pagano D, Lewis ME, Townend JN, Davies P, Camici PG, Bonser RS. Coronary
[5] Redeld MM, Rodeheffer RJ, Jacobsen SJ, Mahoney DW, Bailey KR, Burnett Jr JC. revascularization for postischaemic heart failure: how myocardial viability affects
Plasma brain natriuretic peptide to detect preclinical ventricular systolic or survival. Heart 1999;82:6848.
diastolic dysfunction: a community-based study. Circulation 2004;109:317681. [18] Canty Jr JM, Suzuki G, Banas MD, Verheyen F, Borgers M, Fallavollita JA.
[6] McMurray JV, McDonagh TA, Davie AP, et al. Should we screen for asymptomatic Hibernating myocardium. Chronically adapted to ischemia but vulnerable to
left ventricular dysfunction to prevent heart failure? Eur Heart J 1998;19:8426. sudden death. Circ Res 2004;94:11429.
[7] Galasko GI, Barnes SC, Collinson P, Lahiri A, Senior R. What is the most cost- [19] Allman KC, Shaw LJ, Hachamovitch R, Udelson JE. Myocardial viability testing and
effective strategy to screen for left ventricular systolic dysfunction: natriuretic impact of revascularization on prognosis in patients with coronary artery disease
peptides, the electrocardiogram, hand-held echocardiography, traditional echo- and left ventricular dysfunction: a meta-analysis. J Am Coll Cardiol 2002;39:
cardiography, or their combination? Euro Heart J 2006;27:193200. 11518.
[8] Cotter G, Moshkovitz Y, Kaluski E, et al. Accurate, noninvasive continuous [20] Shewan LG, Coats AJ. Ethics in the authorship and publishing of scientic articles.
monitoring of cardiac output by whole-body electrical bioimpedance. Chest Int J Cardiol 2010;144:12.
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
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.
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
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.
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 and largest 2.5% of the surrogate optima were taken as the 95% CI
surrogate bootstrapped data set thus represents a single realization of of optimum estimated from the original data.
data that would be observed under the null hypothesis while
preserving the amplitude statistics of the original data. The test Assessment of A-VTI Variability
statistic s was calculated for each surrogate data set, and the fraction The variability of the measured A-VTI data was quantified by
that was greater than or equal to that of the original data estimates the calculating the average and SD over all measurements made at an
probability that a test statistic at least as large as that associated with AV delay of 120 ms, which by protocol were acquired over at least
the original data would be observed under the null hypothesis. For 1 respiratory cycle at 4 different times during the recording session.
P0.05, the null hypothesis was rejected, and the data were To allow comparison with previously published values, a transfor-
interpreted as demonstrating a statistically significant dependence on mation was derived that converts the coefficient of variation (defined
pacing interval. as SD divided by mean) to the average difference of successive
The 95% CI of the estimated optimum pacing interval was measurements. Specifically, y2CV 2/, where y is the average
obtained by generating an additional 1000 bootstrapped data sets in difference in successive measurements, and CV is the coefficient of
variation. This result is based on the transformation yixi1xi2/i,
which the original data were randomly selected with replacement
where the xi represents successive A-VTI measurements at a given
while preserving the mapping to test interval. In this case, with the pacing interval in a given patient; yi, the normalized difference in
mapping between test interval and measured data retained, the A-VTI measures in a given patient, and the subscript i, patient-
collection of best-fitting polynomials of the bootstrapped data sets specific values. The derivation assumes that xi1 and xi2 are indepen-
reflects the variability in the best-fitting polynomial of the original dently drawn from a Gaussian distribution whose mean i and SD
data that is attributable to the statistical variability in the measure- i can vary among patients but whose coefficient of variation CVi
ments. The locations of the maxima of the surrogate data sets were remains fixed for all patients, that is, CVii/iconstant. Data
ordered from smallest to largest, and the cutoff point of the smallest acquisition and device programming are summarized in Table 2.
Table 2. Summary of Data Acquisition, Analysis, and without an atrial lead. Patients who underwent AV delay
Pacemaker Programming optimization were in sinus rhythm with the AV delay initiated
Data acquisition by an atrial sensed-event in all cases.
AV and VV interval optimization were conducted independently, with AV delay optimization data from a single patient are
simultaneous biventricular pacing during AV optimization and an AV presented in Figure 3. As with the ICG data shown here (left
delay of 120 ms during VV optimization panel), when statistically significant data were obtained they
Data were recorded from each patient at a single session in the typically exhibited an inverted U appearance, indicating that
following order: ICG (AV), ICG (VV), A-VTI (AV), M-VTI/A-Wave truncation the estimated optimum interval was in the interior of the
(AV), A-VTI (VV), SPWMD (VV) range of test intervals. Statistically insignificant data typically
Consecutive beat-to-beat measures of cardiac function were recorded had a best-fitting polynomial that was flat relative to the
over at least 1 respiratory cycle intrinsic variability of the data.
Test intervals were delivered in random order with 20 30 ms spacing The average and SD of the measured A-VTI data were
The 120 ms AV delay and 0 ms VV interval were repeated 4 times over calculated over all beats (10 to 31, median 20) recorded from
the duration of the recording to assess stability each patient during AV optimization at the 120-ms test
Data Analysis interval. The ranges of the per-patient average and SD of the
A 3rd degree polynomial was fit to the data, with the location of the A-VTI data were 6.7 to 56 cm and 0.73 to 5.1 cm, respec-
maximum serving as the estimated optimum (For SPWMD, the optimum tively. The median sample coefficient of variation was 0.075,
was the location with the minimum absolute delay)
which corresponds to an average normalized difference in
A P value reflecting the statistical significance of measured data was
successive measurements of y0.12, consistent with previ-
obtained
ous reports of y0.10.1.27
The 95% CI of the estimated optimum was obtained
The VV optimization data shown in Figure 4 are from the
For A-wave truncation, the shortest AV delay that avoided A-wave
same patient whose AV optimization results are presented in
truncation was selected as optimum, and the fraction of patients for
whom an optimum could be determined was noted
Figure 3. As with this individual, for most patients, VV
optimization yielded insignificant results more frequently
Pacemaker Programming
than did AV optimization.
The statistical significance, 95% CI, and physiologic plausibility of the
measured data were considered
The number of statistically significant data sets is summa-
rized in Table 3. For both AV and VV optimization, A-VTI,
The pacemaker was left at default settings (AV delay 120 ms, VV
interval 0 ms) for statistically insignificant data sets, wide 95% CIs M-VTI, and SPWMD yielded data that were indistinguish-
that included the default setting, or physiologically implausible data. able from the null hypothesis (ie, failed to demonstrate a
Otherwise, the estimated optimum pacing interval was programmed. statistically significant dependence on pacing interval) the
majority of the time. As shown in Table 3, this finding
Results remained true whether patients with hypertrophic cardiomy-
Patient demographics are presented in Table 1. Of the 20 opathy were included or excluded from the analysis. ICG was
sequential, unique patients referred for clinical pacing opti- significantly different from the null hypothesis 84% of the
mization, 18 had biventricular pacemakers or implantable time for AV delay optimization and 75% of the time for VV
cardioverter defibrillators and 2 had dual-chamber devices. interval optimization. Excluding patients with hypertrophic
The LV lead in 1 patient failed to capture, and another patient cardiomyopathy from the analysis, the null hypothesis could
with chronic atrial fibrillation had a biventricular device be rejected 81% and 75% of the time, respectively.
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.
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 the majority of the time these measures yielded data that did
AV delay estimated by ICG differed from the default value not show a significant dependence on AV or VV interval.
(120 ms) by an average magnitude of 57 ms, and 63% of the With A-wave truncation, it was possible for 3 independent
estimates differed from default by at least 50 ms. Among the readers to estimate an optimum AV delay 69%, 75%, and
statistically significant, ICG-derived VV interval data sets, 94% of the time. ICG performed better than the echocardi-
estimated optima differed from the default value (0 ms) by an ography methods, yielding statistically significant data 84%
average of 52 ms, and 75% of the estimates differed by at of the time for AV delay optimization and 75% of the time for
least 30 ms. Repeating the ICG AV interval optimization VV interval optimization. Notably, the echocardiography
analysis using the same number of measurements at each test techniques tested here represent the most commonly used
interval as the A-VTI recordings continued to yield statisti- approaches to AV/VV interval optimization.28
cally significant results in the majority of patients, with 81% In previous studies, SPWMD failed to predict a response to
of the reduced ICG data sets having P values 0.05. CRT,29,30 and interval optimization using A-VTI neither
An optimum AV delay could be estimated by 3 indepen- improved clinical outcomes12 nor yielded acute data that were
dent readers using A-wave truncation 69%, 75%, and 94% of distinguishable from a negative control.31 The poor precision
the time. The estimates of each reader are compared with ICG of these techniques demonstrated in this study may account
results in Table 4. The relationship between optima predicted for the lack of clinical utility seen in the earlier work.
by ICG and A-wave truncation was variable, with the Pearson AV/VV interval optimization traditionally has been con-
correlation coefficient ranging from 0.02 to 0.67 for the 3 ducted without consideration of the intrinsic variability of the
readers. measured data. Test intervals often are delivered in a nonran-
Although the analysis presented above is based on unique dom order, sweeping systematically from one end of the test
patient visits, 1 patient underwent both ICG and echocardio- range to the other, and the AV or VV interval associated with
graphic optimization on 2 separate occasions separated by 3.5 the best average measure of cardiac function typically is
months. As shown in Figure 5, in both cases ICG yielded taken at face value to represent the underlying physiologic
similar-appearing, statistically significant results. The esti- optimum.11,20 Although efficient, the traditional approach has
mates of the optimum pacing intervals were precise, with important drawbacks. A nonrandom order of test intervals
narrow 95% CIs, and had good concordance, with 91 ms at allows systematic error to be introduced into the measured
the first optimization and 67 ms 3.5 months later. In contrast, data in a way that cannot be subsequently corrected by
in neither optimization session did A-VTI yield statistically averaging (eg, with subtle drift of the Doppler angle over the
significant results. The wide 95% CI of the initial A-VTI course of data collection). Furthermore, without considering
optimum predicted that subsequent optimization attempts the intrinsic variability of the measured data, it is not possible
would be unlikely to identify a similar optimum interval. to characterize the precision of the estimated optimum
AV/VV interval.
Discussion In the absence of knowledge about its precision, an
In this study, multiple clinically accepted AV and VV interval estimated optimum AV/VV interval may in fact be spurious
optimization techniques were used in single sessions in a and associated with worse cardiac function than the
heterogeneous group of 20 patients. For 62% to 86% of population-derived, default setting (eg, a sensed AV delay of
patients, A-VTI, M-VTI, and SPWMD yielded data that were 120 milliseconds) (Figures 2, 3, and 5). In addition, although
statistically indistinguishable from the null hypothesis; that is, multiple studies have suggested that optimum pacing inter-
Downloaded from http://circheartfailure.ahajournals.org/ by guest on December 9, 2013
39 of 158.
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
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
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 * *
18 0.001 0.001 18 77 66 86
19 0.001 19 30 30 30 120 60 90
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 That ICG continued to yield statistically significant data in
plot is relatively flat compared with the intrinsic variability of the majority of patients, even when using an identical number
the measured data, then the estimate of the underlying of data points as the A-VTI analysis, suggests that it has a
optimum is imprecise and heavily influenced by measure- superior intrinsic signal-to-noise ratio and that the acquisition
ment variability. In this case, repeating the optimization time can be substantially shortened from the 60 s per test
process likely would yield a very different estimated opti- interval that was used in this study. The superior noise
mum AV/VV interval. The statistical tools used here along properties of ICG may be partly due to the automatic and
with plotting capability have been made available on the objective nature of data acquisition and analysis in contrast to
Internet.35 A-VTI, which requires the sonographer to physically hold the
In the majority of patients examined in the present study, probe in a fixed position and the reader to manually demar-
ICG generated precise estimates and A-wave truncation cate the envelope of the velocity waveform.
yielded data from which an optimum could be inferred, A wide variety of optimization techniques have been
although often with significant interreader variability and advocated, including multiple approaches to the assessment
marginal correlation with the ICG-predicted optima. Notably, of systolic function, diastolic function, and electrical and
neither ICG nor A-wave truncation has been clinically vali- mechanical synchrony.9,19 21,24 26,30,36 40 Although each has a
dated in prospective interval optimization studies. In addition, rationale that is mechanistically plausible, consideration of
unanswered fundamental questions include whether the phys- the neurohormonal derangements of heart failure and the
iologic optimum interval evolves over time and whether it therapeutic interventions that have been successful lead us to
changes between rest and exertion or between supine and view SV and its surrogates as parameters that when optimized
upright posture. A study that compares estimated optimum are most likely to translate into clinical benefit. Specifically,
intervals obtained at both supine rest and upright exertion in it is now well established that ameliorating the effects of
the same patient, perhaps using motion-tolerant ICG, would sympathetic tone in these patients leads to improved clinical
add important insight into these basic questions. In the outcomes.41 For a given cardiac output, maximizing SV
absence of such data and given the theoretical potential would minimize sympathetic tone. Indeed, the effects on the
benefit of pacing optimization, our approach is to accept an neurohormonal system of increased mechanical efficiency
estimated optimum interval if quantitative and qualitative may contribute to the salutary effects of CRT, which remains
analyses suggest that the estimate has good precision. Partic- the only contractility-enhancing intervention demonstrated to
ular attention should be paid to the effect of outliers and the prolong life.42 Theoretical arguments may not account for
overall shape of the curve compared with the measurement important effects, however. For example, an increase in SV at
variability. the expense of greater oxygen consumption may not benefit
Downloaded from http://circheartfailure.ahajournals.org/ by guest on December 9, 2013
41 of 158.
Turcott et al Optimization of Cardiac Resynchronization Therapy 403
the patient with ischemic heart failure. Ultimately, any 6. Auricchio A, Stellbrink C, Block M, Sack S, Vogt J, Bakker P, Klein H,
proposed optimization technique must be validated by ran- Kramer A, Ding J, Salo R, Tockman B, Pochet T, Spinelli J. Effect of
pacing chamber and atrioventricular delay on acute systolic function of
domized prospective trials with hard clinical end points, a paced patients with congestive heart failure. Circulation. 1999;99:
goal which to date has not been frequently achieved.8 12 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
Limitations dilated cardiomyopathy and ventricular conduction delay. Circulation. 1999;
The study was based on a small and heterogeneous popula- 99:15671573.
tion comprising successive patients referred for clinical pac- 8. Boriani G, Muller CP, Seidl KH, Grove R, Vogt J, Danschel W,
ing interval optimization. Multiple conventional optimization Schuchert A, Djiane P, Biffi M, Becker T, Bailleul C, Trappe HJ.
Randomized comparison of simultaneous biventricular stimulation
techniques were examined for their ability to yield statisti- versus optimized interventricular delay in cardiac resynchronization
cally significant results. This property is necessary but not therapy: The Resynchronization for the HemodYnamic Treatment for
sufficient for improved clinical outcomes, which were not Heart Failure Management II Implantable Cardioverter Defibrillator
(RHYTHM II ICD) study. Am Heart J. 2006;151:1050 1058.
examined in this study.
9. Morales MA, Startari U, Panchetti L, Rossi A, Piacenti M. Atrioven-
tricular delay optimization by Doppler-derived left ventricular dP/dt
Conclusions improves 6-month outcome of resynchronized patients. Pacing Clin
Optimization of AV and interventricular intervals in CRT Electrophysiol. 2006;29:564 568.
10. Rao RK, Kumar UN, Schafer J, Viloria E, De Lurgio D, Foster E.
requires assessment of the variability of the measured data. Reduced ventricular volumes and improved systolic function with cardiac
Accepting an estimated optimum without considering its resynchronization therapy: a randomized trial comparing simultaneous
precision may result in worse cardiac function than default biventricular pacing, sequential biventricular pacing, and left ventricular
pacing. Circulation. 2007;115:2136 2144.
settings in the individual patient and confound results in
11. Sawhney NS, Waggoner AD, Garhwal S, Chawla MK, Osborn J, Faddis
clinical trials. In the small, heterogeneous pacemaker popu- MN. Randomized prospective trial of atrioventricular delay programming
lation examined here, echocardiographic techniques yielded for cardiac resynchronization therapy. Heart Rhythm. 2004;1:562567.
statistically insignificant data in the majority of patients. In 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-
contrast, ICG yielded precise estimates of the optimum AV spective randomized evaluation of VV delay optimization in CRT-D
and VV interval in most patients. Further research is neces- recipients: echocardiographic observations from the RHYTHM II ICD
sary to confirm these results, to validate the accuracy of the study. Pacing Clin Electrophysiol. 2009;32(suppl 1):S120 S125.
13. Ritter P, Daubert C, Mabo P, Descaves C, Gouffault J. Haemodynamic
impedance-predicted optima, and to demonstrate clinical benefit of a rate-adapted A-V delay in dual chamber pacing. Eur Heart J.
improvement with pacing interval optimization compared to 1989;10:637 646.
population-derived default settings. 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.
Sources of Funding 15. Scharf C, Li P, Muntwyler J, Chugh A, Oral H, Pelosi F, Morady F,
Dr Turcott was supported by a Heart Failure Society of America Armstrong WF. Rate-dependent AV delay optimization in cardiac resyn-
Research Fellowship and by the National Library of Medicine chronization therapy. Pacing Clin Electrophysiol. 2005;28:279 284.
(Biomedical Informatics Training Grant LM 07033). Dr Ashley was 16. Mehta D, Gilmour S, Ward DE, Camm AJ. Optimal atrioventricular delay
supported by the National Institutes of Health (grant K08 at rest and during exercise in patients with dual chamber pacemakers: a
HL083914). 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
Disclosures of repeating optimization of atrioventricular interval during interim and
Dr Witteles has received honoraria from Medtronic. Dr Wang has long-term follow-up after cardiac resynchronization therapy. Int
received honoraria and research support from and has served as a J Cardiol. 2008;124:211217.
consultant or advisor to Boston Scientific, Medtronic, and St Jude 18. Valzania C, Biffi M, Martignani C, Diemberger I, Bertini M, Ziacchi M,
Medical. Dr Fowler has received honoraria from Medtronic and Bacchi L, Rocchi G, Rapezzi C, Branzi A, Boriani G. Cardiac resynchro-
Boston Scientific. nization therapy: variations in echo-guided optimized atrioventricular and
interventricular delays during follow-up. Echocardiography. 2007;24:
933939.
References 19. Burri H, Sunthorn H, Shah D, Lerch R. Optimization of device pro-
1. Bristow MR, Saxon LA, Boehmer J, Krueger S, Kass DA, De Marco T, gramming for cardiac resynchronization therapy. Pacing Clin Electro-
Carson P, DiCarlo L, DeMets D, White BG, DeVries DW, Feldman AM. physiol. 2006;29:1416 1425.
Cardiac-resynchronization therapy with or without an implantable defi- 20. Agler DA, Adams DB, Waggoner AD. Cardiac resynchronization therapy
brillator in advanced chronic heart failure. N Engl J Med. 2004;350: and the emerging role of echocardiography (part 2): the comprehensive
2140 2150. examination. J Am Soc Echocardiogr. 2007;20:76 90.
2. Cleland JG, Daubert JC, Erdmann E, Freemantle N, Gras D, Kappen- 21. Jansen AH, Bracke FA, van Dantzig JM, Meijer A, van der Voort PH,
berger L, Tavazzi L. The effect of cardiac resynchronization on morbidity Aarnoudse W, van Gelder BM, Peels KH. Correlation of echo-Doppler
and mortality in heart failure. N Engl J Med. 2005;352:1539 1549. optimization of atrioventricular delay in cardiac resynchronization
3. McAlister FA, Ezekowitz J, Hooton N, Vandermeer B, Spooner C, therapy with invasive hemodynamics in patients with heart failure sec-
Dryden DM, Page RL, Hlatky MA, Rowe BH. Cardiac resynchronization ondary to ischemic or idiopathic dilated cardiomyopathy. Am J Cardiol.
therapy for patients with left ventricular systolic dysfunction: a systematic 2006;97:552557.
review. J Am Med Assoc. 2007;297:25022514. 22. Turcott RG, Ashley EA. Statistical methods for AV/VV optimization
4. Yu CM, Wing-Hong Fung J, Zhang Q, Sanderson JE. Understanding [abstract]. J Cardiac Failure. 2008;14:S65.
nonresponders of cardiac resynchronization therapy current and future 23. Efron B, Tibshirani R. Statistical data analysis in the computer age.
perspectives. J Cardiovasc Electrophysiol. 2005;16:11171124. Science. 1991;253:390 395.
5. Breithardt OA, Stellbrink C, Franke A, Balta O, Diem BH, Bakker P, 24. Heinroth KM, Elster M, Nuding S, Schlegel F, Christoph A, Carter J,
Sack S, Auricchio A, Pochet T, Salo R. Acute effects of cardiac resyn- Buerke M, Werdan K. Impedance cardiography: a useful and reliable tool
chronization therapy on left ventricular Doppler indices in patients with in optimization of cardiac resynchronization devices. Europace. 2007;9:
congestive heart failure. Am Heart J. 2002;143:34 44. 744 750.
25. Braun MU, Schnabel A, Rauwolf T, Schulze M, Strasser RH. Impedance Determined AV Optimization: a comparison of AV delay methods used
cardiography as a noninvasive technique for atrioventricular interval optimi- in cardiac resynchronization therapy (SMART-AV): rationale and design.
zation in cardiac resynchronization therapy. J Interv Card Electrophysiol. Pacing Clin Electrophysiol. 2010;33:54 63.
2005;13:223229. 34. Comparison of AV Optimization Methods Used in Cardiac Resynchro-
26. Tse HF, Yu C, Park E, Lau CP. Impedance cardiography for atrioven- nization Therapy (CRT) (SMART-AV). http://www.clinicaltrials.gov.
tricular interval optimization during permanent left ventricular pacing. Identifier: NCT00677014. Accessed January 2, 2010.
Pacing Clin Electrophysiol. 2003;26:189 191. 35. Stanford University. Optimize CRT (Development Version) Available at
27. Zuber M, Toggweiler S, Quinn-Tate L, Brown L, Amkieh A, Erne P. A http://optimizecrt.stanford.edu. Accessed January 2, 2010.
comparison of acoustic cardiography and echocardiography for opti- 36. Bertini M, Ziacchi M, Biffi M, Martignani C, Saporito D, Valzania C,
mizing pacemaker settings in cardiac resynchronization therapy. Pacing Diemberger I, Cervi E, Frisoni J, Sangiorgi D, Branzi A, Boriani G.
Clin Electrophysiol. 2008;31:802 811. Interventricular delay interval optimization in cardiac resynchronization
28. Gras D, Gupta MS, Boulogne E, Guzzo L, Abraham WT. Optimization of therapy guided by echocardiography versus guided by electrocardio-
AV and VV delays in the real-world CRT patient population: an inter- graphic QRS interval width. Am J Cardiol. 2008;102:13731377.
37. van Gelder BM, Meijer A, Bracke FA. The optimized V-V interval
national survey on current clinical practice. Pacing Clin Electrophysiol.
determined by interventricular conduction times versus invasive mea-
2009;S236 S239.
surement by LV dP/dt(max). J Cardiovasc Electrophysiol. 2008;19:
29. Beshai JF, Grimm RA, Nagueh SF, Baker JH II, Beau SL, Greenberg SM,
939 944.
Pires LA, Tchou PJ. Cardiac-resynchronization therapy in heart failure
38. Burri H, Sunthorn H, Somsen A, Zaza S, Fleury E, Shah D, Righetti A.
with narrow QRS complexes. N Engl J Med. 2007;357:24612471.
Optimizing sequential biventricular pacing using radionuclide ventricu-
30. Chung ES, Leon AR, Tavazzi L, Sun JP, Nihoyannopoulos P, Merlino J, lography. Heart Rhythm. 2005;2:960 965.
Abraham WT, Ghio S, Leclercq C, Bax JJ, Yu CM, Gorcsan J III, St John 39. Vidal B, Tamborero D, Mont L, Sitges M, Delgado V, Berruezo A,
Sutton M, De Sutter J, Murillo J. Results of the Predictors of Response to Diaz-Infante E, Tolosana JM, Pare C, Brugada J. Electrocardiographic
CRT (PROSPECT) trial. Circulation. 2008;117:2608 2616. optimization of interventricular delay in cardiac resynchronization
31. Turcott RG, Das AK, Ashley EA. Re: A prospective randomized therapy: a simple method to optimize the device. J Cardiovasc Electro-
evaluation of VV delay optimization in CRT-D recipients: echocar- physiol. 2007;18:12521257.
diographic observations from the RHYTHM II ICD study. Pacing Clin 40. Turcott RG, Pavek TJ. Hemodynamic sensing using subcutaneous pho-
Electrophysiol. 2009;32:1360 1361; author reply 13611362. toplethysmography. Am J Physiol Heart Circ Physiol. 2008;295:
32. FREEDOM: A Frequent Optimization Study Using the QuickOpt H2560 H2572.
Method. http://www.clinicaltrials.gov. Identifier: NCT00418314. 41. Brophy JM, Joseph L, Rouleau JL. Beta-blockers in congestive heart
Accessed January 2, 2010. failure. A Bayesian meta-analysis. Ann Intern Med. 2001;134:550 560.
33. Stein KM, Ellenbogen KA, Gold MR, Lemke B, Lozano IF, Mittal S, 42. Kass DA. Rescuing a failing heart: putting on the squeeze. Nat Med.
Spinale FG, van Eyk JE, Waggoner AD, Meyer TE. SmartDelay 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.
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 established that NICaS-derived CO (NI-CO) is a reliable
of survivors of acute heart failure have dramatically parameter when compared with thermodilution CO or mod-
increased the number of outpatients with chronic ified Fick CO6
heart failure (CHF).12 Hemodynamic parameters, such as The purpose of the present study was to evaluate the
stroke volume (SV), cardiac output (CO), total systemic feasibility of using NICaS in outpatients with CHF and to
vascular resistance (TSVR), and plasma level of B-type assess whether NICaS-derived hemodynamic parameters in
natriuretic peptide (BNP) are thought to be important for CHF are clinically significant by relating them to other con-
predicting the long-term prognosis and guiding the optimal ventional cardiovascular functional indices, and to assess
treatment of CHF.35 Until recently, hemodynamic parame- the predictive accuracy of NICaS-derived parameters for
ters could only be obtained using the invasive thermodilu- CHF-related readmission.
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 Methods
the clinical assessment of CHF in an outpatient setting. Patient Selection and Exclusion Criteria
Several non-invasive technologies for measuring hemo- In the present study, patients who met the following cri-
dynamics are now available. Non-Invasive Cardiac System teria were enrolled: (1) hospitalized because of acute CHF;
(NICaS) (NI Medical; Hod-Hasharon, Israel) is a new (2) had a stable condition after optimal medical therapy
device for calculating SV, CO, and TSVR utilizing whole including diuretics, vasodilators, ACE-inhibitors; (3) had no
body bioimpedance cardiography with electrodes placed on renal dysfunction (defined as a creatinine value >2.0mg/dl);
one wrist and on the contralateral ankle. We previously and (4) had no significant severe valve diseases. The cardi-
nal manifestations of heart failure were dyspnea and fatigue
(Received September 8, 2008; revised manuscript received January 4,
caused by cardiac dysfunction.
2009; accepted January 20, 2009; released online April 16, 2009) The study population comprised 68 CHF patients (56
Division of Cardiovascular Medicine, Department of Internal Medi- men, 12 women, mean age 64.910.8 years); 20 patients
cine, Kobe University Graduate School of Medicine, Kobe, Japan were classified as New York Heart Association (NYHA)
Conflict of Interest: none declared class I and 48 were NYHA class II patients (Table1). The
Mailing address: Junya Shite, MD, Division of Cardiovascular and Institutional Ethics Committee of the Kobe University
Respiratory Medicine, Department of Internal Medicine, Kobe Uni-
versity Graduate School of Medicine, 7-5-1 Kusunoki-cho, Chuo-ku,
Hospital approved the study protocol and all patients pro-
Kobe 650-0017, Japan. E-mail: shite@med.kobe-u.ac.jp vided informed consent to participate in the study.
All rights are reserved to the Japanese Circulation Society. For permis- In most patients, CHF coexisted with multiple under-
sions, please e-mail: cj@j-circ.or.jp lying heart diseases, as shown in Table1. Three patients
had atrial fibrillation when CO was measured. In addition, Table1. Patient Characteristics
20 cases (29.4%) had a moderate degree of mitral valve
No. of patients 68
regurgitation, 7 cases (10.3%) had a moderate degree of Men 56 (82.4%)
aortic valve regurgitation, and 1 case (1.5%) had moderate Women 12 (17.6%)
aortic valve stenosis. The cardiovascular medicines taken Age (years) 64.910.8
by the patients comprised: diuretics (n=44, 68.7%), digitalis NYHA I 20 (29.4%)
(n=7, 10.9%), angiotensin-converting enzyme inhibitors (n= NYHA II 48 (70.6%)
28, 43.7%), angiotensin receptor blockers (n=42, 65.6%), Hypertension 8 (11.8%)
Diabetes mellitus 17 (25.0%)
beta blockers (n=53, 82.8%), nitrates (n=8, 12.5%), calcium Ischemic heart disease 18 (26.5%)
channel blockers (n=13, 20.3%), and oral inotropic agents Dilated cardiomyopathy 28 (41.2%)
(n=10, 15.6%).
The exclusion criteria for the NICaS measurements NYHA, New York Heart Association classification
included restlessness and/or unstable patient condition,
severe aortic valve regurgitation and/or aortic stenosis,
aortic aneurysm, heart rate above 130beats/min, intra- and the plasma was immediately frozen at 80C and thawed
extra-cardiac shunts, severe peripheral vascular disease, only once at the time of the assay, which was performed
severe pitting edema, sepsis, and dialysis, all of which inter- within 1 week. The plasma BNP concentration was measured
fere with the accurate measurement of impedance-derived using a specific immunoradiometric assay kit (Shionogi
SV,7 as previously described. Co, Osaka, Japan), as previously reported.14
All patients were followed up at the outpatient clinic of
the Cardiovascular Division of Kobe University hospital and Echocardiography Measurement of the Ejection Fraction
had no walking disability. We measured the cardiac param- (EF) and the Ratio of Early-to-Atrial Filling
eters once the patients attented the outpatient department, of Transmitral Flow (E/A)
which was the first routine visit to our hospital 1 month after B-mode recordings were obtained with a commercially
discharge from the hospital for the first heart failure admis- available instrument (SONOS 5500; Agilent Technologies
sion. At the same visit, serum BNP levels were measured Inc, Palo Alto, CA, USA) operating at 2.5MHz. Two-dimen-
and a questionnaire about exercise tolerance was adminis- sional imaging examinations were performed in the standard
tered. manner from apical 4 and 2 chamber views, and the left ven-
tricular (LV) EFs were measured from B-mode images
Cardiac Indicators Derived by NICaS according to the modified Simpsons method. Also, pulsed-
To measure NI-CO, an alternating electrical current of Doppler findings of transmitral flow were obtained to calcu-
1.4mA with a 30kHz frequency was passed through the late the E/A.
patient via 2 pairs of tetrapolar electrodes, 1 pair placed on
the wrist above the radial artery, and the other pair placed Exercise Tolerance Estimation
on the contralateral side above the posterior tibialis artery. Each patient underwent a questionnaire-based interview
If the arterial pulses were either absent or of poor quality, a for the estimation of the exercise tolerance threshold (ET)
second pair of electrodes was placed on the contralateral according to a specific activity scale15
site. The NICaS apparatus calculated the SV using Tsoglin
and Frinermans formula:8 Statistical Analysis
MedCalc version 9.6 (MedCalc Software, Mariakerke,
SV=dR/RL2/Ri(+)/KWHF
Belgium) was used for data analysis. The quantitative data
where dR is the change in impedance, R is the basal resis- are expressed as meanSD. A Students t-test was used for
tance, is the blood electrical resistivity, L is the patient descriptive statistics. The MannWhitney U-test was used
height, Ri is the corrected basal resistance according to for non-parametric data sets. A two-tailed Pearsons corre-
gender and age,812 KW is a correction of weight according lation was used to compare the relationship between LVEF,
to ideal values,913 HF is a hydration factor that takes into log BNP, and NI-CI, NI-SVI, and NI-TSVR. Spearmans
account body water composition,11+ is equal to the ECG coefficient of rank correlation (rho) was used to estimate the
RR wave interval, andis the diastolic time interval. probability of correlations with ET. We used the multiple
Because the NI-CO results are calculated every 20s, the regression analysis in order to detect the predictors for CHF
average of 3 measurements obtained consecutively during readmission, and also used receiver operating characteristic
60s of monitoring was considered to be the NI-CO value curve (ROC) analysis for BNP, NI-CI, NI-SVI, NI-TSVR,
for each individual case. By using the NI-CO value, the fol- LVEF and ET to determine the thresholds between the read-
lowing parameters were calculated: NI-CO index (NI-CI= mission group and the non-readmission group. A two-tailed
NI-CO/body surface area), SV index (NI-SVI=NI-CI/heart P-value of less than 0.05 was considered to be significant.
rate), and TSVR (NI-TSVR=Mean arterial pressure/CO*80
[dynes1cm5]).
Results
Plasma BNP Measurements Relationship Between NICaS-Derived Hemodynamic
Blood samples were obtained from the patient in a resting Parameters and LVEF, Plasma logBNP and ET
condition after NICaS measurement. The samples were There were significant but weak correlations between
drawn into plastic syringes, transferred to chilled siliconized LVEF and NI-CI (LVEF=22.70+5.21*NI-CI, r=0.3148, P=
disposable tubes containing aprotinin (1,000 kallikrein inac- 0.0089), NI-SVI (LVEF=22.80+0.38*NI-SVI, r=0.3257,
tivator in units/ml; Ohkura Pharmaceutical, Kyoto, Japan) P=0.0067), NI-TSVR (LVEF=46.020.003*NI-TSVR, r=
and ethylenediaminetetraacetic acid (1mg/ml), immediately 0.2636, P=0.0298; Figure1). There were significant mod-
placed on ice, and then centrifuged at 48C. An aliquot of erate correlations between log BNP and NI-CI (log BNP=
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
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.
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
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.
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
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
11. Hoffer EC, Meador CK, Simpson DC. A relationship between whole
Conclusions body impedance and total body water volume. Ann NY Acad Sci 1970;
NICaS-derived hemodynamic parameters obtained by 170: 452461.
12. Ward LC, Heitmann BL, Craid P, Stroud D, Azinge EC, Jebb S, et al.
monitoring whole body bioimpedance are applicable for the Association between ethnicity, body mass index, and bioelectrical
non-invasive assessment of cardiac function for the follow impedance: Implications for the population specificity of prediction
up of outpatients with CHF. 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
References Association; 1964; 7378.
14. Yasue H, Yoshimura M, Sumida H, Kikuta K, Kugiyama K, Jougasaki
1. Okamoto H, Tsutsui H. The epidemiology of heart failure in Japan. M, et al. Localization and mechanism of secretion of B-type natri-
Nihon Rinsho Extra Issue 2007; 913: 4954. uretic peptide in comparison with those of A-type natriuretic peptide
2. Okamoto H, Kitabatake A. Epidemiology of heart failure in Japan. in normal subjects and patients with heart failure. Circulation 1994;
Nihon Rinsho 2003; 61: 709714. 90: 195203.
3. Solomon SD, Anavekar N, Skali H, McMurray JJ, Swedberg K, 15. Sasayama S, Asanoi H, Ishizaka S, Miyagi K. Evaluation of functional
Yusuf S, et al. Influence of ejection fraction on cardiovascular out- capacity of patients with congestive heart failure. In: Yasuda H,
comes in a broad spectrum of heart failure patients. Circulation 2005; Kawaguchi H, editors. New aspects in the treatment of failing heart.
112: 37383744. Tokyo: Springer-Verlag; 1992; 113117.
4. Koglin J, Pehlivanli S, Schwaiblmair M, Vogeser M, Cremer P, 16. Tsutamoto T, Wada A, Maeda K, Hisanaga T, Mabuchi N, Hayashi
vonScheidt W. Role of brain natriuretic peptide in risk stratification M, et al. Plasma brain natriuretic peptide level as a biochemical
of patients with congestive heart failure. J Am Coll Cardiol 2001; 38: marker of morbidity and mortality in patients with asymptomatic or
19341941. minimally symptomatic left ventricular dysfunction: Comparison with
5. Berger R, Huelsman M, Strecker K, Bojic A, Moser P, Stanek B, et plasma angiotensin II and endothelin-1. Eur Heart J 1999; 20: 1799
al. B-type natriuretic peptide predicts sudden death in patients with 1807.
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, B-Type natriuretic peptide assay for identifying patients at high risk
et al. Impedance cardiography for cardiac output estimation: Reliabil- of re-admission after decompensated heart failure. J Am Coll Cardiol
ity of wrist-to-ankle electrode configuration Circ J 2006; 70: 1164 2004; 43: 635641.
1168. 18. Tsutsui H, Tsuchihashi-Makaya M, Kinugawa S, Goto D, Takeshita
7. Cotter G, Moshkovitz Y, Kaluski E, Kohen A, Miller H, Goor D, et A; The JCARE-CARD Investigators. Clinical characteristics and
al. Accurate, noninvasive continuous monitoring of cardiac output by outcome of hospitalized patients with heart failure in Japan. Circ J
whole-body electrical bioimpedance. Chest 2004; 125: 14311440. 2006; 70: 16171623.
8. Cohen AJ, Arnaudov D, Zabeeda D, Shultheis L, Lashinger J, 19. Tsutsui H, Tsuchihashi-Makaya M, Kinugawa S, Goto D, Takeshita
Schachner A. Non-invasive measurement of cardiac output during A; JCARE-GENERAL Investigators. Characteristics and outcomes
coronary artery bypass grafting. Eur J Surg 1998; 14: 6469. of patients with heart failure in general practices and hospitals. Circ J
9. Organ LW, Bradham GB, Gore DT, Lozier SL. Segmental bioelectrical 2007; 71: 449454.
impedance analysis: Theory and application of a new technique. J Appl 20. Parthenakis FI, Kanoupakis EM, Kochiadakis GE, Skalidis EI, Mezilis
Physiol 1994; 77: 98112. NE, Simantirakis EN. Left ventricular diastolic filling pattern pre-
10. Lukaski H, Bolonchuk WW, Hall CB, Siders WA. Validation of tet- dicts cardiopulmonary determinants of functional capacity in patients
rapolar bioelectrical impedance method to assess human body com- with congestive heart failure. Am Heart J 2000; 140: 338344.
position. J Appl Physiol 1986; 60: 13271332.
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-
Introduction
Cardiac resynchronization therapy (CRT) has developed from CRT devices have become signicantly more complex
an experimental method1 to an established adjunctive recently. Many now have the facility to programme different
treatment for patients with advanced heart failure. CRT atrioventricular (AV) and interventricular (VV) stimulation
aims to improve cardiac output (CO) by lessening the inter- timing intervals.19,20 Data exist to show that tailoring
and intra-ventricular conduction delay caused in part by these intervals to suit the patient in hand further improves
left bundle branch block (LBBB).2,3 CRTreduces clinical symp- the haemodynamic benets brought by CRT. However, ques-
toms of heart failure and hospitalization,46 improves haemo- tions remain as to which method of haemodynamic assess-
dynamic parameters (including ventricular performance),7 ment is best for guiding the adjustment of the device
and has been shown to reduce mortality.810 Acceptance of settings to suit any given post-implant patient.
this therapy is reected by its incorporation into current In the measurement of cardiac function, invasive methods
guidelines for the management of heart failure.11,12 (e.g. dp/dt estimation of contractility or the thermodilution
Successful CRT in any given patient depends upon many method for CO) are the gold standards but are not suitable
variables, such as the appropriate positioning of the LV for routine use during routine follow-up of patients with
lead,1315 and also on achieving optimal biventricular stimu- implanted CRT devices. Non-invasive methods used in opti-
lation timing.16,17 This latter variable has attracted particu- mizing device settings include echocardiographic tech-
lar interest recently in an attempt to reduce the substantial niques,2124 radionucleotide ventriculography,25 nger
proportion (2030%) of patients who derive no apparent photoplethysmography,26 and more recently, impedance
benet from CRT, despite meeting appropriate implantation cardiography (IC).6,27
criteria and having had technically straightforward device IC is an established technique for haemodynamic assess-
implantation (non-responders).18 ment and is capable of calculating CO on a beat-to-beat
basis.28 It relies upon changes in impedance (resistance) to
* Corresponding author. Tel: 49 345 557 2601; fax: 49 345 557 2072. current ow through the chest between strategically
E-mail address: konstantin.heinroth@medizin.uni-halle.de 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 biventricular tracking of the intrinsic sinus rhythm (VAT-Mode).36
that the impedance changes with blood ow, these variables The AV interval was set at 120 ms as a standard value6 without LV
can be used to calculate CO. IC has been used in patients or RV pre-excitation (VV interval 0). This standard pacing
receiving dual chamber pacemakers2931 and in the assess- set-up was kept from implantation until optimization.
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 Impedance cardiography
implanted CRT devices. Optimization of biventricular pacing was performed using a com-
This paper presents data from the routine use of IC in mercially available system for IC (Task Force Monitor Systems,
optimizing CRT by adjusting AV and VV intervals to obtain CNSystems, Graz, Austria) as described by Braun et al. 6 Two electro-
maximum CO in post-implant patients. To our knowledge, des were placed bilaterally to the inferior chest wall in combination
this is the rst report on the combined manipulation of with one electrode at the neck. Low-amplitude high-frequency
AV- and VV-interval timing in biventricular pacing devices current was delivered via these surface electrodes, and transthor-
using IC. 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
Methods additional surface electrodes: one pair placed bilaterally to the
sternum and the second pair bilaterally to the abdomen. Cardiac
Patient characteristics output was calculated on a beat-to-beat basis from the trans-
Forty-six consecutive patients (37 males, 9 females, mean age thoracic impedance signal.37 Figure 1 shows an example of the IC
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
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 There were no signicant differences between the heart
rates without pacing (75.1 + 10.7 bpm), with initial simul-
To optimize the CRT set-up, all patients were examined in taneous biventricular pacing (74.7 + 10.9 bpm), or opti-
the supine position in a silent environment to reduce the mized biventricular pacing (74.8 + 10.2 bpm, P . 0.05).
impact of sympathetic activation by external stimuli. Thus, results for CO obtained at each CRT set-up were unaf-
A standard protocol involving a period of stabilization and fected by the heart rate.
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 Cardiac output without pacing vs. simultaneous
rate limit of 40 bpm to avoid effects of atrial pacing on biventricular pacing vs. optimized biventricular
the AV interval.36 During data acquisition, telemetry pacing
between the implanted device and the programmer was dis-
The alternation between three cycles of no pacing with sim-
connected to prevent interference with the measurement of
ultaneous biventricular pacing during stabilization and equi-
impedance. AV-interval values relate to the atrio-RV stimu-
libration revealed (across all patients) a mean increase in CO
lation interval, and VV intervals relate to interventricular
of 21.6 + 22.1% (P , 0.05) with pacing. The mean range
interval with negative gures implying LV pre-excitation.
(i.e. variation) between the three recordings (in each
The rst stage of the pacing protocol was a period of
patient) without pacing was 10.1% and between the three
stabilization and equilibration. The baseline CO without
recordings with pacing was 11.8%.
pacing was recorded, and this was alternated three times
Figure 2 shows data obtained for a typical patient at
with standard simultaneous biventricular pacing with an
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 Figure 2 Cardiac output measured by impedance cardiography at
transvenous approach and 1 by the transthoracic approach different ventriculo-ventricular intervals at an atrioventricular
during surgical mitral valve repair. The LV lead was placed interval of 120 ms in a representative patient. Optimum cardiac
in a posterolateral (n 35) or an anterolateral (n 10) output was yielded at a left ventricular pre-excitation of 240 and
side branch. There were no complications. 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.
4.57 + 1.1 L/min. Similar values were obtained by LV pre- Optimizing VV and AV intervals further increased the mean
excitation of 20 and 60 ms (4.51 + 1.3 and 4.50 + 1.2 L/ CO to 4.86 + 1.1 L/min. This corresponds to an increase in
min, respectively). The optimal pacing set-up varied mean CO of 32.8% without pacing and an increase of 11.2%
widely from patient to patient. As a result of this variability, compared with simultaneous biventricular pacing, all three
there was no signicant difference in CO when the results set-ups resulting in signicantly different CO (P , 0.05,
from different AV and VV intervals were compared. No Figure 4).
single combination of AV and VV intervals could be recom- Taking an increase in CO 10% as a denition of a positive
mended for application to the whole study population, haemodynamic response to CRT,39 16 of the 46 patients
because no single combination of intervals showed statisti- (35%) were non-responders with standard simultaneous
cally signicant superiority over other combinations. biventricular CRT. In ve of these patients, CO with standard
For the population as a whole, when the CO obtained with biventricular pacing was lower than without any pacing. The
optimized biventricular pacing (i.e. the CO measured at mean CO in this group of 16 non-responders (with standard
whichever AV and VV intervals produced the highest value simultaneous biventricular CRT) was 100.8 + 4.9% (where
in that patient) is compared with the CO obtained with CO without pacing was 100%). After optimization, 9 of
standard simultaneous biventricular pacing and compared these 16 patients (who were non-responders to standard
with the CO obtained with no pacing, a statistically signi- simultaneous biventricular CRT) experienced an increase
cant difference is seen (Figure 4). CO without pacing of 10% in CO to become responders. In this group of
was 3.66 + 0.85 L/min. CO increased to 4.40 + 1.1 L/min nine patients, the mean CO increased signicantly from
(P , 0.05) with simultaneous biventricular pacing using a 101.3 + 3.6% to 121 + 5.1% (P 0.05). In the seven patients
standard AV interval of 120 ms. who failed to respond despite optimization, the mean CO
The mean increase in CO changing from no pacing to sim- was 106.6 + 2.6% with optimized pacing. Nevertheless,
ultaneous biventricular pacing was 21.6 + 22.1%, P , 0.05). after modication of AV and VV intervals to produce the
53 of 158.
748 K.M. Heinroth et al.
54 of 158.
Impedance cardiography 749
Prospective randomized trial data have shown that still higher CO values, and hence hopefully reduce the
AV-interval optimization not only improves the haemo- number of non-responders. Further studies are required.
dynamic response to CRT, but also improves NYHA functional
class and quality of life scores.47
The current generation of devices allows the manipulation Conclusions
of both AV and VV intervals.
Since activation of the interventricular septum is mainly The present paper is the rst to report the utility of IC in the
inuenced by the RV lead, modication of the VV interval optimization of both AV and VV intervals in biventricular
affects not only inter-ventricular dyssynchrony, but also pacemaker set-up in routine clinical practice. Using IC, we
intra-ventricular dyssynchrony of the LV. The capacity to demonstrated that manipulating both AV and VV intervals
manipulate VV intervals may thus further improve the is feasible and may result in a signicant improvement in
haemodynamic response to CRT. the haemodynamic response to CRT compared with stan-
There is increasing evidence that sequential biventricular dard simultaneous biventricular pacing and no pacing. The
pacing is superior to simultaneous biventricular pacing in results show the importance of optimizing CRT set-up in
many patients with a CRT device. each patient individually, since the variability between
Perego et al. examined the impact of sequential biventri- patients means that no single set of intervals can be identi-
cular pacing while invasively monitoring dp/dt in both ven- ed as being suitable for all or even most patients.
tricles simultaneously. Simultaneous ventricular pacing In our opinion, IC is a promizing method of CRT optimiz-
produced an increase in dp/dt of 29% over baseline, but ation and may compare well with other techniques in
sequential pacing produced a signicantly greater increase routine use. It carries less risk of complications than invasive
55 of 158.
750 K.M. Heinroth et al.
12. Hunt SA. ACC/AHA 2005 guideline update for the diagnosis and manage- dual-chamber pacemakers using impedance cardiography and Doppler
ment of chronic heart failure in the adult: a report of the American echocardiography. J Clin Basic Cardiol 1999;2:23740.
College of Cardiology/American Heart Association Task Force on 32. Belardinelli R, Ciampani N, Costantini C, Blandini A, Purcaro A. -
Practice Guidelines (Writing Committee to Update the 2001 Guidelines Comparison of impedance cardiography with thermodilution and direct
for the Evaluation and Management of Heart Failure). J Am Coll Fick methods for noninvasive measurement of stroke volume and
Cardiol 2005;46:e182. cardiac output during incremental exercise in patients with ischemic car-
13. Butter C, Auricchio A, Stellbrink C, Fleck E, Ding J, Yu Y et al. Effect of diomyopathy. Am J Cardiol 1996;77:1293301.
resynchronization therapy stimulation site on the systolic function of 33. Albert NM, Hail MD, Li J, Young JB. Equivalence of the bioimpedance and
heart failure patients. Circulation 2001;104:30269. thermodilution methods in measuring cardiac output in hospitalized
14. Gold MR, Auricchio A, Hummel JD, Giudici MC, Ding J, Tockman B et al. patients with advanced, decompensated chronic heart failure. Am J
Comparison of stimulation sites within left ventricular veins on the Crit Care 2004;13:46979.
acute hemodynamic effects of cardiac resynchronization therapy. Heart 34. Adachi H, Hiratsuji T, Sakurai S, Tada H, Toyama T, Naito S et al.
Rhythm 2005;2:37681. Impedance cardiography and quantitative tissue Doppler echocardiogra-
15. Albertsen AE, Nielsen JC, Pedersen AK, Hansen PS, Jensen HK, Mortensen phy for evaluating the effect of cardiac resynchronization therapy: a
PT. Left ventricular lead performance in cardiac resynchronization case report. J Cardiol 2003;42:3742.
therapy: impact of lead localization and complications. Pacing Clin 35. Santos JF, Caetano F, Parreira L, Madeira J, Cardoso P, Fonseca N et al.
Electrophysiol 2005;28:4838. Tissue Doppler echocardiography for evaluation of patients with ventricu-
16. Auricchio A, Ding J, Spinelli JC, Kramer AP, Salo RW, Hoersch W et al. lar resynchronization therapy. Rev Port Cardiol 2003;22:136371.
Cardiac resynchronization therapy restores optimal atrioventricular 36. Bernheim A, Ammann P, Sticherling C, Burger P, Schaer B, Brunner-La
mechanical timing in heart failure patients with ventricular conduction Rocca HP et al. Right atrial pacing impairs cardiac function during resyn-
delay. J Am Coll Cardiol 2002;39:11639. chronization therapy: acute effects of DDD pacing compared to VDD
17. Doshi RN. Optimizing resynchronization therapy: can we increase the pacing. J Am Coll Cardiol 2005;45:14827.
number of true responders? J Cardiovasc Electrophysiol 2005;16(Suppl. 37. Ovsyshcher I, Furman S. Impedance cardiography for cardiac output esti-
1):S4851. mation in pacemaker patients: review of the literature. Pacing Clin
56 of 158.
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 system, CPO is an integrative measure of cardiac
common usage in the coronary care unit (CCU) setting hydraulic pumping ability. Cardiac power output has
despite questions regarding the use of this procedure in been shown to be a powerful predictor of mortality in
critically ill patients.1-3 The decision to use PAC is based patients with chronic heart failure (HF) and in those with
on the concept that invasive hemodynamic measure- cardiogenic shock.6-8 Recently, a report from the SHOCK
ments provide a more precise understanding of the Trial Registry found that CPO was the strongest inde-
status of the cardiovascular system, which in turn allows pendent hemodynamic correlate of inhospital mortality
close monitoring of the response to therapeutic inter- in patients with cardiogenic shock.4 We hypothesized
ventions. However, traditional hemodynamic measure- that CPO would provide important prognostic informa-
ments obtained from PAC, such as cardiac output (CO), tion across a broad spectrum of patients with acute
do not fully describe the fundamental state of the cardiac disease. Accordingly, we analyzed the relation-
cardiovascular system in the setting of severe illness.4 In ship between CPO and short-term mortality in a consec-
particular, because of its critical dependence on preload utive series of patients undergoing PAC in a single CCU.
and afterload, measured CO provides a limited estima-
tion of ventricular function.
Cardiac power output (CPO) is a novel hemodynamic Methods
measure that is the product of simultaneously measured Study design and data collection
CO and mean arterial pressure (MAP).4,5 By incorporating The institutional review board for human research at the
both the pressure and flow domains of the cardiovascular Washington Hospital Center approved this research protocol.
Data were collected prospectively for consecutive patients
admitted to the CCU between October 2002 and February
From the Division of Cardiology, Washington Hospital Center, Washington, DC.
2006. Based on all available information, the attending
This work was supported by the Cardiovascular Research Institute, Washington Hospital physician determined the primary clinical diagnosis requiring
Center, Washington, DC. admission to the CCU. Patients with a noncardiac primary
Submitted July 13, 2006; accepted November 21, 2006. diagnosis were excluded from this analysis. Persistent cardio-
Reprint requests: Howard A. Cooper, MD, Division of Cardiology, Washington Hospital genic shock was determined to be present by using conven-
Center, 110 Irving Street NW, Suite NA1103, Washington, DC 20010.
tional clinical criteria of hypotension and signs of peripheral
E-mail: howard.a.cooper@medstar.net
0002-8703/$ - see front matter
hypoperfusion in the presence of pulmonary congestion that
n 2007, Mosby, Inc. All rights reserved. did not rapidly resolve. Left ventricular ejection fraction (EF)
doi:10.1016/j.ahj.2006.11.014 was visually estimated from 2-dimensional echocardiography
57 of 158.
American Heart Journal
Mendoza et al 367
Volume 153, Number 3
58 of 158.
American Heart Journal
368 Mendoza et al
March 2007
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
59 of 158.
American Heart Journal
Mendoza et al 369
Volume 153, Number 3
Our analysis extends these findings to a broader patients by using whole-body electrical bioimpedance
spectrum of patients with critical cardiac illness, to determine CO, thereby eliminating any direct
including those with acute coronary syndromes, car- complications of PAC.14,15 Clinical studies in which
diomyopathies of various etiologies, valvular heart noninvasively determined CPO is targeted for treatment
disease, cardiac arrhythmias, and complicated post appear to be indicated.
percutaneous coronary intervention courses. Our study has several potential limitations. First, this
As expected, patients in our analysis had CPO values was a single-center retrospective analysis, subject to the
considerably higher (mean 0.88 W) than those in a limitations inherent in such a study design. However,
more select group with acute myocardial infarction and measurements of CPO were not used by the CCU
cardiogenic shock in the SHOCK trial registry (mean physicians to make therapeutic decisions. Hence, we
0.62 W). Despite the substantial differences in patient are confident that the observed prognostic value of
populations and absolute CPO values, however, the CPO was not influenced by bias introduced from
relationship between CPO and inhospital mortality was knowledge of this information. Second, although we
remarkably similar in the 2 analyses: an OR of 0.65 per attempted to control for important differences in
0.2-W increase in CPO in the current study compared baseline characteristics between patients, residual con-
with an OR of 0.60 in the SHOCK trial registry.4 founding might remain. Third, we were unable to
Furthermore, among the subset of our patients diag- distinguish between hemodynamic measurements made
nosed with persistent cardiogenic shock (n = 73), the with and without support from inotropes and/or an
positive and negative predictive values using a cutoff of intra-aortic balloon pump. Because both types of
0.53 W were nearly identical in the 2 studies (62% vs interventions are likely to increase CPO, this might
58% and 68% vs 71%, respectively). This concordance affect the relationship between CPO and mortality.
of results suggests both that the relationship between However, this relationship was maintained after in-
CPO and short-term mortality is consistent across a cluding the use of these interventions in the multivar-
broad spectrum of cardiac diagnoses and a wide range iate model. Indeed, the relationship between CPO and
of CPO values, and is reproducible across studies. mortality might be stronger in the absence of such
Cardiac power output is calculated from the more support measures. Finally, we must emphasize that our
familiar hemodynamic variables, CO and MAP. There- findings demonstrate an association between CPO and
fore, from a statistical standpoint, it cannot provide outcome without providing a direct assessment of
additional prognostic information beyond that contained whether CPO should be used as a clinical end point for
within the 2 individual components. However, CPO therapeutic interventions.
combines information from these 2 separate hemody-
namic axes (flow and pressure) into a single, easily
understandable variable that can potentially be targeted Conclusions
for therapy. For this reason, the concept of CPO appears Cardiac power output is a novel and useful
to provide clinically useful information above and prognostic tool in patients with acute cardiac disease
beyond that provided by standard hemodynamic varia- undergoing PAC in the CCU setting. Further studies
bles alone. in which CPO is targeted for treatment appear to
Recent studies have questioned the use of PAC in be warranted.
guiding treatment in critically ill patients, including
patients with severe CHF.3,9-13 Possible explanations for
these negative results include complications related to References
the PAC procedure itself (arrhythmias, infection); 1. Connors Jr AF, Speroff T, Dawson NV, et al. The effectiveness of
misinterpretation of the obtained data; incorrect action right heart catheterization in the initial care of critically ill patients.
in response to even appropriately interpreted data; or SUPPORT Investigators. JAMA 1996;276:889 - 97.
unanticipated deleterious effects of treatments imple- 2. Dalen JE, Bone RC. Is it time to pull the pulmonary artery catheter?
mented in response to these data. In patients with JAMA 1996;276:916 - 8.
cardiac disease, treatments are typically targeted to the 3. Binanay C, Califf RM, Hasselblad V, et al. Evaluation study of
CO (or cardiac index) and pulmonary capillary wedge congestive heart failure and pulmonary artery catheterization
pressure. However, CO is a measure of cardiovascular effectiveness: the ESCAPE trial. JAMA 132005;294:1625 - 33.
4. Fincke R, Hochman JS, Lowe AM, et al. Cardiac power is the
flow, not cardiac contractility, whereas pulmonary
strongest hemodynamic correlate of mortality in cardiogenic shock:
capillary wedge pressure is a measure of intracardiac
a report from the SHOCK trial registry. J Am Coll Cardiol
pressure not directly representing cardiac perfor- 2004;44:340 - 8.
mance.1 Theoretically, therefore, CPOa measure of 5. Cotter G, Williams SG, Vered Z, et al. Role of cardiac power in
cardiac pumping abilitymight be a more appropriate heart failure. Curr Opin Cardiol 2003;18:215 - 22.
target for medical therapy. In addition, CPO can be 6. Williams SG, Cooke GA, Wright DJ, et al. Peak exercise cardiac
simply and accurately determined noninvasively in most power output; a direct indicator of cardiac function strongly
60 of 158.
American Heart Journal
370 Mendoza et al
March 2007
predictive of prognosis in chronic heart failure. Eur Heart J patients in intensive care (PAC-Man): a randomised controlled trial.
2001;22:1496 - 503. Lancet 2005;366:472 - 7.
7. Tan LB, Littler WA. Measurement of cardiac reserve in cardiogenic 12. Rhodes A, Cusack RJ, Newman PJ, et al. Bennett ED. A randomised,
shock: implications for prognosis and management. Br Heart J controlled trial of the pulmonary artery catheter in critically ill
1990;64:121 - 8. patients. Intensive Care Med 2002;28:256 - 64.
8. Tan LB. Cardiac pumping capability and prognosis in heart failure. 13. Wheeler AP, Bernard GR, Thompson BT, et al. Pulmonary-artery
Lancet 1986;2:1360 - 3. versus central venous catheter to guide treatment of acute lung
9. Sandham JD, Hull RD, Brant RF, et al. A randomized, controlled trial injury. N Engl J Med 2006;354:2213 - 24.
of the use of pulmonary-artery catheters in high-risk surgical 14. Cotter G, Moshkovitz Y, Kaluski E, et al. Accurate, noninvasive
patients. N Engl J Med 2003;348:5 - 14. continuous monitoring of cardiac output by whole-body electrical
10. Richard C, Warszawski J, Anguel N, et al. Early use of the bioimpedance. Chest 2004;125:1431 - 40.
pulmonary artery catheter and outcomes in patients with shock and 15. Leitman M, Sucher E, Kaluski E, et al. Non-invasive measurement of
acute respiratory distress syndrome: a randomized controlled trial. cardiac output by whole-body bio-impedance during dobutamine
JAMA 2003;290:2713 - 20. stress echocardiography: clinical implications in patients with left
11. Harvey S, Harrison DA, Singer M, et al. Assessment of the clinical ventricular dysfunction and ischaemia. Eur J Heart Fail 2006;8:
effectiveness of pulmonary artery catheters in management of 136 - 40.
61 of 158.
Circ J 2006; 70: 1164 1168
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
Table 1 Statistical Analysis of CO Measurements Performed by NI-CO and Indirect Fick-CO Compared With TD-CO
All measurements in 35 patients Measurements after NTG injection
(n=50) (n=15)
NI-CO Fick-CO NI-CO Fick-CO
TD- NI- Fick- TD- NI- Fick-
vs vs vs vs
CO CO CO CO CO CO
TD-CO TD-CO TD-CO TD-CO
Correlation 0.911a 0.801b 0.962a 0.822b
Bias (L/min) 0.18 0.32 0.15 0.34
SD (L/min) 0.43 0.60 0.50 0.69
Lower level of agreement (L/min) 1.06 1.52 1.16 1.73
Upper level of agreement (L/min) 0.68 0.88 0.85 1.04
Average CO SD (L/min) 4.181.01 4.361.03 4.050.89 3.590.76 3.850.85 3.670.79
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 aortic valve stenosis.
of this study was to evaluate the reliability and feasibility of
the new Non-Invasive Cardiac System (NICaS: NI Medical, Measuring CO
Hod-Hasharon, Israel), which calculates the CO by measur- All operators were unaware of the CO results obtained
ing ICG in a tetra polar mode, derived from electrodes by the various measuring techniques.
placed on one wrist and the contra-lateral ankle (Fig 1). 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
Methods the pulmonary artery under fluoroscopic guidance and veri-
The Trial fied with the pressure waveforms registered on the poly-
This study prospectively enrolled 35 patients. The ther- graph. TD-CO was measured 5 times by injecting 5 ml
modilution-derived CO (TD-CO) was measured using a bolus of iced 9% saline solution at the same rate. There-
Swan-Ganz catheter (Baxter Healthcare, Irvine, CA, USA), after, the 3 results of the saline injections that were within
followed immediately by the modified Fick (Fick-CO) and 15% of their extreme disparity were averaged for the TD-
with the NICaS (NI-CO). In 15 subjects, a second round of CO result.
CO measurements was carried out using the 3 technologies Fick-CO For arterial oxygen saturation, blood samples
after 24 mg nitroglycerin injection into the arteries. Thus, were obtained from the arterial access sheath, and for
a total of 50 comparative CO measurements were per- venous oxygen saturation, blood samples were withdrawn
formed. The Institutional Ethics Committee of the Kobe using the distal edge lumen of the Swan-Ganz catheter
University Hospital approved the study protocol and all placed in the pulmonary artery. All samples were immedi-
patients gave their consent. ately measured for oxygen saturation using the same device
(Radiometer ABL 715, Copenhagen, Denmark).
Patient Selection and Exclusion Criteria NI-CO To measure the CO with the NICaS apparatus,
All cardiac disease patients who were scheduled for an alternating electrical current of 1.4 mA with a 30 kHz
routine right heart catheterization or emergency procedures frequency is passed through the patient via 2 pairs of
that required the deployment of a Swan Ganz catheter for tetrapolar electrodes, one pair placed on the wrist above the
continuous cardiac function monitoring were eligible for radial pulse, and the other pair on the contralateral ankle
the study, unless they fulfilled one of the exclusion criteria above the posterior tibialis arterial pulse. If the arterial
The exclusion criteria for the NI-CO measurements in- pulses in the legs are either absent or of poor quality, the
cluded: restlessness and/or chaotic patient condition, severe second pair of electrodes is placed on the contralateral
aortic valve regurgitation and/or aortic stenosis, aortic wrist. The NICaS apparatus calculates the stroke volume
aneurysm, heart rate above 130 beats/min, intra- and extra- by Frinermans formula:18
cardiac shunts, severe peripheral vascular disease, severe
Stroke volume = dR/R L2 /Ri (+)/ KW HF
pitting edema, sepsis and dialysis, all of which interfere
with the proper measurements of impedance derived stroke where dR is the impedance change, R is the basal resis-
volume.4 tance,is the blood electrical resistivity, L is the patients
The study population comprised 35 patients (17 men, 18 height, Ri is the corrected basal resistance according to
women, mean age 65.513.7 years). In most patients there gender and age,1823 KW is a correction of weight according
was coexistence of multiple underlying heart diseases, to ideal values,1924 HF is the hydration factor, which takes
including hypertension (n= 15), diabetes mellitus (n= 13), into account the body water composition,21 and + is
coronary artery disease (n= 21) and idiopathic dilated car- equal to the ECG RR wave interval, andis the diastolic
diomyopathy (n=3). Twelve patients presented with conges- time interval.
tive heart failure (CHF) and 4 patients had atrial fibrillatio Because the NI-CO results are calculated every 20 s, the
when CO was measured. In addition, our study subjects average of 3 measurements obtained consecutively during
included 7 cases of moderate degree of valve regurgitation 60 s of monitoring was considered to be the NI-CO value
(aortic regurgitation 2 cases, mitral regurgitation 4 cases, for each individual case.
mild tricuspid regurgitation 1 case) and 7 cases of moderate
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
Table 2 Summary of Previous Reports of the Accuracy of CO Measurements by ICG (ICGT-CO, ICGWB-CO, and
NI-CO vs TD-CO)
levels. However, the accuracy of the results remained un- 10. Koobi T, Kaukinen S, Turjanmaa VM. Cardiac output can be reliably
altered in NI-CO, which suggests that NICaS is also applic- measured noninvasively after coronary artery bypass grafting opera-
tion. Crit Care Med 1999; 27: 2206 2211.
able even when the arterial resistance has changed. 11. Kedrov AA. An attempt to quantify assessment of the central and
peripheral circulation by electrometrical method. Klin Med 1948; 26:
Clinical Implications 32 51 (in Russian).
Recent studies have shown that the calculation of cardiac 12. Patterson RP, Kubicek WG, Kinnen E, Witsoe DA, Noren G. Devel-
opment of an electrical impedance plethysmography system to moni-
power output (CO mean arterial pressure) and systemic tor cardiac output. Proceedings of the 1st Annual Rocky Mountains
vascular resistance are important for the management of Bioengineering Symposium 1964; 56 71.
various cardiac diseases.13 Cardiac power output was found 13. Kubicek WG, Karnegis JN, Patterson RP, Witsoe DA, Mattson RH.
to be the strongest independent predictor of in-hospital Development and evaluation of an impedance cardiac output system.
Aerosp Med 1966; 37: 1208 1212.
mortality in patients admitted with cardiogenic shock2 and 14. Kubicek WG, Kottke FJ, Ramos MU, Patterson RP, Witsol DA,
is an important tool for assessing the clinical response to Labree JW, et al. The Minnesota impedance cardiograph: Theory and
drug therapy. In addition, there is enough evidence that applications. Biomed Eng 1974; 9: 410 416.
ambulatory monitoring of cardiac power would benefit pa- 15. Raaijmakers E, Faes ThJC, Scholten RJPM, Goovaerts HG, Heethaar
tients with CHF, resulting in better titration of medication R. A meta-analysis of published studies concerning the validity of
thoracic impedance cardiography. Ann NY Acad Sci 1999; 873: 121
and possibly less readmission to hospital. By introducing 134.
NICaS apparatus, wide-spread clinical use of cardiac 16. Handelsman H. Public health service reassessment: Measuring car-
power calculation would become feasible. 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.
Conclusion and Limitations of NICaS 17. Guidance for Industry. Statistical approach to establishing bioequiv-
alence: US department of health and human services. Food and Drug
The present study indicates that NICaS performs at least Administration, Center For Drug Evaluation and Research (CDER)
as accurately as the thermodilution method. However, the January 2001, BP.
18. Cohen AJ, Arnaudov D, Zabeeda D, Shultheis L, Lashinger J,
reliability of the NICaS method depends on an alignment Schachner A. Non-invasive measurement of cardiac output during
with exclusion criteria. This allows for the use of NICaS in coronary artery bypass grafting. Eur J Surg 1998; 14: 64 69.
approximately 8085% of patients needing the examina- 19. Organ LW, Bradham GB, Gore DT, Lozier SL. Segmental bioelectri-
tion. 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-
References rapolar bioelectrical impedance method to assess human body
composition. J Appl Physiol 1986; 60: 1327 1332.
1. Nohria A, Mielniczuk L, Warner L. Evaluation and monitoring of 21. Hoffer EC, Meador CK, Simpson DC. A relationship between whole
patients with acute heart failure syndromes. Am J Cardiol 2005; body impedance and total body water volume. Ann NY Acad Sci
96(Suppl): 32G 40G. 1970; 170: 452 461.
2. Fincke R, Hochman J, Lowe A, Menon V, Slater JN, Webb JG, et al 22. Lamberts R, Visser KR, Zijlstra WG. Impedance cardiography.
(for the SHOCK investigators). Cardiac power is the strongest hemo- Assen, The Netherlands; Van Gorkum 1984; 21 94.
dynamic correlate of mortality in cardiogenic shock: A report from 23. Ward LC, Heitmann BL, Craid P, Stroud D, Azinge EC, Jebb S, et al.
the SHOCK trial registry. J Am Coll Cardiol 2004; 40: 340 348. Association between ethnicity, body mass index, and bioelectrical
3. Cotter G, Moshkovitz Y, Kaluski E, Milo O, Nobikov Y, Schneeweiss impedance: Implications for the population specificity of prediction
A, et al. The role of cardiac power and systemic vascular resistance equations. Ann NY Acad Sci 2000; 904: 199 204.
in the pathophysiology and diagnosis of patients with acute conges- 24. Hamwi GT, Danowski TS. Changing dietary concepts in diabetes
tive heart failure. Eur J Heart Fail 2003; 5: 443 451. mellitus: Diagnosis and treatment. New York: American Diabetes
4. Cotter G, Moshkovitz Y, Kaluski E, Kohen A, Miller H, Goor D, et Association 1964; 73 78.
al. Accurate, noninvasive continuous monitoring of cardiac output by 25. Bland JM, Altman DG. Statistical methods for assessing agreement
whole-body electrical bioimpedance. Chest 2004; 125: 1431 1440. between two methods of clinical measurement. Lancet 1986; I: 307
5. Haryadi DG, Orr JA, Kuck K, Mc James S, Westenskow DR. Partial 310.
CO2 rebreathing indirect Fick technique for non-invasive measure- 26. Bland JM, Altman DG. Correlation, regression and repeated data.
ment of cardiac output. J Clin Monit Comput 2000; 16: 361 374. BMJ 1994; 308: 896.
6. Jover JL, SoroM, Belda FJ, Aguilar G, Caro P, Ferrandis R. Measure- 27. Drazner M, Thompson B, Rosenberg P, Kaiser PA, James MSN,
ment of cardiac output after cardiac surgery: Validation of a partial Boehrer D, et al. Comparison of impedance cardiography with inva-
carbon dioxide rebreathing (NICO) system in comparison with con- sive hemodynamic measurements in patients with heart failure sec-
tinuous thermodilution with a pulmonary artery catheter. Rev Esp ondary to ischemic or nonischemic cardiomyopathy. Am J Cardiol
Anestesiol Reanim 2005; 52: 256 262 (in Spanish). 2002; 89: 993 995.
7. Turner MA. Doppler-based hemodynamic monitoring: A minimally 28. Van De Water JM, Miller TW, Vogel RL, Mount BE, Dalton ML.
invasive alternative. AACN Clin Issues 2003; 14: 220 231. Impedance cardiography: The next vital sign technology? Chest
8. Moshkovitz Y, Kaluski E, Milo O, Vered Z, Cotter G. Recent devel- 2003; 123: 2028 2033.
opments in cardiac output determination by bioimpedance: Compari- 29. Leslie S J, McKee S, Newby DE, Webb DJ, Denvir MA. Non-inva-
son with invasive cardiac output and potential cardiovascular applica- sive measurement of cardiac output in patients with chronic heart
tions. Curr Opin Cardiol 2004; 19: 229 237. failure. Blood Press Monit 2004; 9: 277 280.
9. Tischenko MI. Estimation of stroke volume by integral rheogram of 30. Dhingra VK, Fenwick JC, Walley KR, Chittock D, Ronco JJ. Lack
the human body. Sechenov Phisiological J 1973; 59: 1216 1224 (in of agreement between thermodilution and Fick cardiac output in criti-
Russian). cally ill patients. Chest 2002; 122: 990 997.
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
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.
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
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)
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.
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
2.0
Difference(Bonjer-TD) CO
1.0 +2SD
1.0
lit/min
0.0 0.0
lit/min
-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
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
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.
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
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
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
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
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 (Figure 1). After the 2-week washout period, patients meeting
179 mm Hg and/or diastolic BP of 90 to 109 mm Hg. Patients were inclusion/exclusion criteria were randomized in a 3:2 ratio to the
excluded if they were on 3 antihypertensive agents; had abnormal standard arm (n95) or ICG-aided hemodynamic arm (n69) using
laboratory findings; or had history of heart failure, left ventricular a central telephone service and stratified by site. All of the physicians
ejection fraction 40%, atrial fibrillation, severe valvular disease, were educated on the hemodynamic treatment strategy illustrated in
cerebrovascular event within 3 months, severe renal disease, ne- Figure 2.
phrotic syndrome, or cirrhosis. Patients in whom ICG might be
subject to technical limitations were excluded (height 47 or 75 Procedures
inches or weight 66 or 341 pounds, presence of activated BP determinations were made in the seated position using the
minute-ventilation pacemaker, known hypersensitivity to sensor gel oscillometric technique. ICG data were collected by trained techni-
or adhesive, or skin lesion interfering with sensor placement). cians at each visit in all of the patients, but ICG findings were not
Patients who were enrolled and subsequently found to have not met revealed in the standard arm to treating physicians or patients. ICG
the inclusion/exclusion criteria were terminated and excluded from was performed with patients in the supine position, resting for 5
analysis. Therefore, this was not an intention-to-treat analysis. The minutes before measurement (BioZ ICG Monitor, CardioDynamics).
study was approved by an independent institutional review board, ICG involves the measurement of thoracic impedance through
which adheres to the Declaration of Helsinki and US Code of Federal placement of 4 dual sensors, 2 on the neck and 2 on the chest.
Regulations. These hypertensive outpatients provided written in- Electrical impedance changes are digitally processed to calculate
formed consent and had study procedures consistent with the CO, SVR, and thoracic fluid content (TFC).13 CO and SVR are
protocol (no. 20021400). normalized for body size by indexing to each patients body sur-
face area to obtain cardiac index (CI) and SVR index (SVRI). TFC
Hemodynamic Evaluation Assignment is the inverse of baseline chest impedance, and any changes in TFC
Eligible patients (N164) underwent a 2-week washout period at are directly proportional to total fluid (intravascular and extravascu-
which time all of the antihypertensive medications were discontinued lar) changes.14 TFC has different normal ranges for each gender that
according to the manufacturers recommendations. After screening are displayed and printed for reference. The reproducibility of this
and medication washout, each patient had 3 monthly office visits ICG device in stable outpatients has been established,15 and accuracy
Figure 2. Suggested treatment strategy for hemodynamic arm; BB indicates blocker; CAA, central acting agent; and VD, vasodilator.
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.
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- resulted in greater reduction in BP and SVRI and better BP
tensity, more effective combination therapy, or better patient control. Physicians cannot adequately estimate hemodynam-
compliance. Our study allowed full discretion by the physi- ics from routine clinical examination or BP measurements,23
cian in choosing the agents, and a multitude of classes and because at similar levels of BP, SVR and CO can vary widely.
doses within classes were used. The study was not powered to Therefore, the addition of accurate, noninvasive, and readily
find small disparities in medication use, and most medication obtainable hemodynamic measurements is clinically relevant.
differences did not reach statistical significance. Importantly, the current study also showed that patients in
On the other hand, some differences are worth noting. the hemodynamic arm were almost twice as likely to achieve
Patients in the standard arm were more likely to experience BP control with normalization of both CI and SVRI. Im-
both increases and decreases in their medication doses, provements in vascular resistance may result in greater
whereas medication class changes were not different between benefits in reducing cardiovascular risk than improvement in
arms. This result might have been expected, because treat- BP alone,24 and differences in SVRI at the same BP may
ment in the standard arm followed guidelines and usual explain poorer prognosis for men versus women25 and black
practice patterns in which a stepped approach to therapy versus nonblack patients.26 Hemodynamics are also known to
contributes to a trial-and-error method of determining change with age. In older subjects, decreased arterial com-
whether agents and doses are working. In the hemodynamic pliance and CI lead to increased SVRI, arterial BP, and pulse
arm, the initial selection of antihypertensive medications pressure.27 In spite of the expected differences in the hemo-
appears to have been influenced by the hemodynamic data, dynamics of older patients, this study demonstrated that
because these patients were more likely to be prescribed a hemodynamically driven, individualized therapy was simi-
vasodilating agent to reduce SVRI. Additionally, the hemo- larly effective regardless of age or existence of isolated
dynamic treatment strategy influenced medication use when systolic hypertension.
SVRI was considered high, because patients in the hemody- The use of ICG to achieve greater BP control offers the
namic arm were more likely to have received an ACEI, ARB, potential for better short-term use of healthcare resources. In
or CCB, as was suggested. The hemodynamic treatment addition, the long-term benefits of even small levels of BP
strategy did not influence the prescription of blockers in the reduction are well known. A sustained BP reduction of
presence of high CI or in diuretic use in response to TFC 4/3 mm Hg is expected to reduce stroke risk 23%, coronary
changes. However, -blocker use was lower or reduced in the heart disease events 15%, heart failure 16%, and overall
presence of low or normal CI in the hemodynamic arm. mortality 14%.28 Accordingly, a recent meta-analysis of
Although the final number of antihypertensive medications major hypertension trials reveals that an antihypertensive
given to patients in both arms of the study was similar, agent is judged favorably when it produces mean BP im-
patients in the hemodynamic arm were more likely to be provements versus placebo of only 3 or 4 mm Hg or versus
prescribed an ARB. However, when ACEI and ARB use was another antihypertensive agent of only 1 or 2 mm Hg.29
combined into a single category (reninangiotensinaldoste- Previously, ICG has been used to profile hemodynamic
rone system inhibitors), the hemodynamic arm only trended variability across BP values30 and to identify left ventricular
toward greater use at the final visit (87.0% versus 76.8%). dysfunction.31 Changes in ICG parameters have demonstrated
Thiazide diuretic use increased during the study but was the hemodynamic effect of antihypertensive agents32,33 and
lower than in some pharmacological trials and what hyper- dietary sodium.34 ICG-guided therapy has shown benefit in a
tension guidelines currently suggest. However, the percent- case series,35 observational study,36 and a randomized trial in
age of patients in both arms who were prescribed a thiazide resistant hypertensive patients.12 In the randomized trial,
diuretic at the final visit was very similar to the 35.6% usage ICG-guided therapy resulted in better final BP and greater BP
that was reported in recent analysis of over 25 000 hyperten- control. Similar to our study, that study showed no differ-
sive patients.21 The lower use of diuretics and blockers also ences in the number of medications between arms. In contrast
follows the previously recognized physician preference for to our study with lower diuretic doses and fewer medication
other antihypertensive agents.22 Some might hypothesize that changes in the hemodynamic arm, resistant hypertension
greater BP reductions could have been achieved in the patients receiving ICG-guided therapy had higher diuretic
standard arm if diuretics were used more frequently. How- doses and more medication changes. The differences between
ever, when patients taking a thiazide diuretic were examined the studies might be expected because of the difference in
as a subgroup, the hemodynamic arm maintained its superi- patients (severe hypertension on more medications versus
ority. Additionally, although the higher doses of thiazide milder hypertension on fewer medications) and setting (spe-
diuretics in the standard arm may have contributed to a cialist versus generalist). However, in both studies, ICG-
greater drop in TFC from the post-washout visit, they did not guided therapy led to more effective treatment as evidenced
lead to better BP control. by better BP outcomes.
Our study was not intended to evaluate whether a particular The conclusions of this study may be limited to its duration
antihypertensive agent was more effective at reducing BP of 3 months. However, in pharmacological trials, short-term
than another. Rather, it was designed to determine whether reductions in BP are typically sustained over longer periods.37
providing hemodynamic data to the physician and patient Another limitation may be in our use of patient-reported
could more effectively reduce BP. Whether hemodynamic medication compliance. Without using automatic counting
data led to a more tailored approach to selection and moni- procedures, our goal was to educate both arms equally and to
toring of antihypertensive agents or by other factors, it reinforce patient compliance with follow-up phone calls.
Downloaded from hyper.ahajournals.org by on June 11, 2011
84 of 158.
Smith et al Noninvasive Hemodynamics and BP Control 777
Lastly, treatment differences in the hemodynamic arm do not output determination in pulmonary arterial hypertension. Congest Heart
imply superiority of one medication over another, because the Fail. 2004;10(2 suppl 2):710.
18. Defined Daily Dose (DDD) developed by the World Health Organization
study was not designed to evaluate this question. for use in drug utilization studies. WHO Collaborating Centre for Drug
Statistics Methodology. Anatomical Therapeutic Chemical (ATC) Clas-
Perspectives sification Index With Defined Daily Doses (DDDs). Oslo, Norway: WHO
The results of this study indicate that ICG-guided antihyper- Collaborating Centre; 2000.
19. Dahlof B, Devereux RB, Kjeldsen SE, Julius S, Beevers G, de Faire U,
tensive therapy in uncontrolled hypertensive patients on 1 to Fyhrquist F, Ibsen H, Kristiansson K, Lederballe-Pedersen O, Lindholm
3 antihypertensive medications is more effective than stan- LH, Nieminen MS, Omvik P, Oparil S, Wedel H; LIFE Study Group.
dard care. This was evident by greater reductions in systolic Cardiovascular morbidity and mortality in the Losartan Intervention for
Endpoint reduction in hypertension study (LIFE): a randomised trial
and diastolic BP and by achieving a better level of BP control. against atenolol. Lancet. 2002;359:9951003.
Our study showed that, in clinical practice, inclusion of ICG 20. ALLHAT Officers and Coordinators for the ALLHAT Collaborative
hemodynamic assessment may improve BP control rates in Research Group. The Antihypertensive and Lipid-Lowering Treatment to
patients who are not controlled on initial therapy. Prevent Heart Attack Trial. Major outcomes in moderately hypercholes-
terolemic, hypertensive patients randomized to pravastatin vs usual care:
Acknowledgments The Antihypertensive and Lipid-Lowering Treatment to Prevent Heart
Attack Trial (ALLHAT-LLT). JAMA. 2002;288:2998 3007.
Consideration of Noninvasive Hemodynamic Monitoring to Target
21. Singh H, Johnson ML. Prescribing patterns of diuretics in multi-drug
Reduction of Blood Pressure Levels (CONTROL) Site Investigators:
antihypertensive regimens. J Clin Hypertens. 2005;7:81 87.
Fred E. Abbo, Douglas C. Beatty, Donald M. Brandon, Milan L. 22. Ubel PA, Jepson C, Asch DA. Misperceptions about -blockers and
Brandon, Anthony V. Dallas, Jr, Lawrence Dinenberg, Mazhar El diuretics: a national survey of primary care physicians. J Gen Intern Med.
Amir, Nigar Enayat, Neil W. Hirschenbein, Dorothy Lebeau, Ber- 2003;18:1060 1061.
nard A. Michlin, John Millspaugh, Nell Nestor, Melissa Noble, 23. Connors AF, Dawson NV, Shaw PK, Montenegro HD, Nara AR, Martin
Robin F. Spiering, Joseph Taylor, and Allen R. Walker. Statistical L. Hemodynamic status in critically ill patients with and without acute
support was provided by Gerard Smits and the study was sponsored heart disease. Chest. 1990;98:1200 1206.
by CardioDynamics (San Diego, CA). 24. Fagard RH, Pardaens K, Staessen JA, Thijs L. Prognostic value of
invasive hemodynamic measurements at rest and during exercise in
References hypertensive men. Hypertension. 1996;28:3136.
1. Fields LE, Burt VL, Cutler JA, Hughes J, Roccella EJ, Sorlie P. The 25. Messerli FH, Garavaglia GE, Schmieder RE, Sundgaard-Riise K, Nunez
burden of adult hypertension in the United States 1999 to 2000: a rising BD, Amodeo C. Disparate cardiovascular findings in men and women
tide. Hypertension. 2004;44:398 404. with essential hypertension. Ann Intern Med. 1987;107:158 161.
2. Kearney PM, Whelton M, Reynolds K, Muntner P, Whelton PK, He J. 26. Hinderliter A, Blumenthal J, Waugh R, Chilukuri M, Sherwood A. Ethnic
Global burden of hypertension: analysis of worldwide data. Lancet. 2005; differences in left ventricular structure: Relations to hemodynamics and
365:217223. diurnal blood pressure variation. Am J Hypertens. 2004;17:43 49.
3. Chery DK, Woodwell DA. National Ambulatory Care Medical Survey: 27. Alfie J, Galarza C, Waisman G. Noninvasive hemodynamic assessment of
2000 summary. Advance Data. 2002;328:132. the effect of mean arterial pressure on the amplitude of pulse pressure.
4. American Heart Association. Heart Disease and Stroke Statistics 2004 Am J Hypertens. 2005;18:60S 64S.
Update. Dallas, Texas: American Heart Association; 2003. 28. Blood Pressure Lowering Treatment Trialists Collaboration. Effects of
5. Wang Y, Wang QJ. The prevalence of prehypertension and hypertension different blood-pressure-lowering regimens on major cardiovascular
among US adults according to the new joint national committee guidelines: events: results of prospectively-designed overviews of randomized trials.
new challenges of the old problem. Arch Intern Med. 2004;164:21262134. Lancet. 2003;362:15271545.
6. Yakovlevitch M, Black HR. Resistant hypertension in a tertiary care 29. Williams B. Recent hypertension trials: implications and controversies.
clinic. Arch Intern Med. 1991;151:1786 1792. J Am Coll Cardiol. 2005;45:813 827.
7. Ferrario CM, Page IH. Current views concerning cardiac output in the 30. Abdelhammed AI, Smith RD, Levy P, Smits GJ, Ferrario CM. Nonin-
genesis of experimental hypertension. Circ Res. 1978;43:821 831. vasive hemodynamic profiles in hypertensive subjects. Am J Hypertens.
8. Davidson RC, Ahmad S. Hemodynamic profiles in essential and sec- 2005;18:51S59S.
ondary hypertension. In: Izzo JL, Black HR, eds. Hypertension Primer. 31. Bhalla V, Isakson S, Bhalla MA, Lin JP, Clopton P, Gardetto N, Maisel
3rd ed. Dallas, TX: Council on High Blood Pressure Research, American AS. Diagnostic ability of B-type natriuretic peptide and impedance car-
Heart Association; 2003:349 351. diography: testing to identify left ventricular dysfunction in hypertensive
9. Sullivan JM, Schoeneberger TE, Ratts ET, Palmer ET, Samaha JK, patients. Am J Hypertens. 2005;18:73S 81S.
Mance CJ, Muirhead EE. Short-term therapy of severe hypertension. 32. Mitchell A, Buhrmann S, Saez AO, Rushentsova U, Schafers RF, Philipp
Hemodynamic correlates of the antihypertensive response in man. Arch T, Nurnberger J. Clonidine lowers blood pressure by reducing vascular
Intern Med. 1979;139:12331239. resistance and cardiac output in young, healthy males. Cardiovasc Drugs
10. Egan B, Schmouder R. The importance of hemodynamic considerations Ther. 2005;19:49 55.
in essential hypertension. Am Heart J. 1988;116:594 599. 33. Breithaupt-Grogler K, Gerhardt G, Lehmann G, Notter T, Belz GG.
11. Northridge DB, Findlay IN, Wilson J, Henderson E, Dargie HJ. Non- Blood pressure and aortic elastic propertiesverapamil SR/trandolapril
invasive determination of cardiac output by Doppler echocardiography compared to a metoprolol/hydrochlorothiazide combination therapy. Int
and electrical bioimpedance. Br Heart J. 1990;63:9397. J Clin Pharmacol Ther. 1998;36:425 431.
12. Taler SJ, Textor SC, Augustine JE. Resistant hypertension: Comparing he- 34. Ashida T, Nishioeda Y, Kimura G, Kojima S, Kawamura M, Imanishi M,
modynamic management to specialist care. Hypertension. 2002;39:982988. Abe H, Kawano Y, Yoshimi H, Yoshida K, Kuramochi M, Omae T.
13. Ventura HO, Taler SJ, Strobeck JE. Hypertension as a hemodynamic Effects of salt, prostaglandin, and captopril on vascular responsiveness in
disease: the role of impedance cardiography in diagnostic, prognostic, and essential hypertension. Amer J Hypertens. 1989;2:640 642.
therapeutic decision making. Am J Hypertens. 2005;18:26S 43S. 35. Sharman DL, Gomes CP, Rutherford JP. Improvement in blood pressure
14. Van De Water JM, Mount BE, Chandra KM, Mitchell BP, Woodruff TA, control with impedance cardiograph-guided pharmacologic decision
Dalton ML. TFC (thoracic fluid content): a new parameter for assessment making. Congest Heart Fail. 2004;10:54 58.
of changes in chest fluid volume. Am Surg. 2005;71:81 86. 36. Sramek BB, Tichy JA, Hojerova M, Cervenka V. Normohemodynamic
15. Treister N, Wagner K, Jansen PR. Reproducibility of impedance cardi- goal-oriented antihypertensive therapy improves the outcome (abstract).
ography parameters in outpatients with clinically stable coronary artery Amer J Hypertens. 1996;9:141A.
disease. Am J Hypertens. 2005;18:44S50S. 37. Julius S, Kjeldsen SE, Weber M, Brunner HR, Ekman S, Hansson L, Hua
16. Van De Water JM, Miller TW. Impedance cardiography: the next vital T, Laragh J, McInnes GT, Mitchell L, Plat F, Schork A, Smith B,
sign technology? Chest. 2003;123:2028 2033. Zanchetti A. VALUE trial group. Outcomes in hypertensive patients at
17. Yung G, Fedullo P, Kinninger K, Johnson W, Channick R. Comparison high cardiovascular risk treated with regimens based on valsartan or
of impedance cardiography to direct fick and thermodilution cardiac amlodipine: the VALUE randomised trial. Lancet. 2004;363:20222031.
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.
1. Background and aims measure of systolic cardiac contractile reserve, better than
VO 2 and echocardiographic ejection fraction (EF).
Cardiac power index (Cpi) is the product of simulta- Recently, Samejima et al. [4] demonstrated that CO
neously measured cardiac index (CI) and mean arterial increase during stress using non-invasive CO determina-
blood pressure (MAP). Cpi increases cardiac power tion with bioimpedance was correlated with stress induced
reserve during stress [1,2] or dobutamine administration dyspnea.
[3] was shown in previous studies to be an important The use of hemodynamic measures such as increase in
vascular resistance for the detection of ischaemia was
* Corresponding author. Duke Clinical Research Institute, 2400 Pratt
suggested almost a decade ago [5]. However, this research
Street, Durham, 27715 NC, USA. avenue has not been pursued due to the lack of simple non-
E-mail address: cotte013@mc.duke.edu (G. Cotter). invasive devices for CO measurement. Recently Weiss et al.
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
86 of 158.
M. Leitman et al. / The European Journal of Heart Failure 8 (2006) 136 140 137
87 of 158.
138 M. Leitman et al. / The European Journal of Heart Failure 8 (2006) 136 140
-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.
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.
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 cardiac function strongly predictive of prognosis in chronic heart
conclusions, if these results are reproduced by additional, failure. Eur Heart J 2001;22:1496 503.
[2] Cohen-Solal A, Tabet JY, Logeart D, Bourgoin P, Tokmakova M,
larger studies, cardiac power and vascular resistance Dahan M. A non-invasively determined surrogate of cardiac power
changes could be used for non-invasive detection of left (Fcirculatory power_) at peak exercise is a powerful prognostic factor
ventricular dysfunction and ischaemia during simple stress in chronic heart failure. Eur Heart J 2002;23:806 14.
tests such as electrocardiographic exercise stress test or [3] Marmor A, Schneeweiss A. Prognostic value of noninvasively
obtained left ventricular contractile reserve in patients with severe
mental stress test. Moreover, serial changes in these
heart failure. J Am Coll Cardiol 1997;29:422 8.
measures may be useful for improving detection of [4] Samejima H, Omiya K, Uno M, Inoue K, Tamura M, Itoh K, et al.
ischaemia in patients at home using telemedicine. Relationship between impaired chronotropic response, cardiac output
during exercise, and exercise tolerance in patients with chronic heart
4.1. Study limitations 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
The study included a small, select group of patients resistometer: preliminary report of a new method tested during
referred for DSE due to various symptoms. Hence, the percutaneous transluminal coronary angioplasty. Circulation 1986;
results require further confirmation in a larger study 74:780 5.
including a non-selected group of outpatients. [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.
5. Conclusion [7] Cohen AJ, Arnaudov D, Zabeeda D, Schultheis L, Lashinger J,
Schachner A. Non-invasive measurement of cardiac output during
The results of the present study suggest that a simple coronary artery bypass grafting. Eur J Cardiothorac Surg 1998;14:
stress test using dobutamine infusion and non-invasive 64 9.
[8] Cotter G, Moshkovitz Y, Kaluski E, Cohen AJ, Miller M, Goor DA,
determination of CI by the NICaSi 2001 apparatus and et al. Accurate, non-invasive continuous monitoring of cardiac output
calculation of Cpi and TPRi can be used for easy out-patient by whole body electrical bio-impedance. Chest 2004;125:1431 40.
screening of patients for systolic LV dysfunction and [9] Torre-Amione G, Cotter G, Kaluski E, Salah A, Milo O, Richter C,
myocardial ischaemia. 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].
References [10] Bland JM, Altman DG. Statistical methods for assessing agreement
between two methods of clinical measurement. Lancet 1986;1:307 10.
[1] Williams SG, Cooke GA, Wright DJ, Parsons WJ, Riley RL, Marshall [11] Cotter G, Williams SG, Vered Z, Tan LB. The role of cardiac power in
P, et al. Peak exercise cardiac power output. A direct indicator of heart failure. Curr Opin Cardiol 2003;18:215 22.
90 of 158.
CLINICAL INVESTIGATIONS
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 uncertainty to diagnosis and treatment. In patients with
department (ED) symptoms in older patients.1 both cardiac and pulmonary disease, the initial assess-
Various conditions, including heart failure, ment and therapy in the ED are challenging.
chronic obstructive pulmonary disease, pneumonia, pul- Because cardiovascular disease, specifically decom-
monary embolus, and acute coronary syndromes, may pensated heart failure (HF), is a relatively common cause
occur alone or in combination in a given patient, adding of dyspnea in elders, an assessment of hemodynamics,
From the Department of Emergency Medicine, Cleveland Clinic participated in the creation of the educational process for the par-
(WFP), Cleveland, OH; Department of Emergency Medicine, ticipating physicians before study commencement. The sponsor
University of Mississippi (RLS), Jackson, MS; Wayne State had no role in data collection or statistical analyses, and the man-
University (JV), Detroit, MI; and Department of Emergency uscript is the sole responsibility of the authors. W.F.P. and R.L.S.
Medicine, Case Western Reserve University (CEE), Cleveland, OH. received honoraria in 2003 for speaking for CardioDynamics.
Received September 6, 2005; revision received November 7, Presented as a moderated poster at the American College of
2005; accepted November 15, 2005. Emergency Physicians Research Forum, October 2003.
Supported by GE Medical Systems (Milwaukee, WI), which Address for correspondence and reprints: W. Frank Peacock,
provided devices and disposables for this study, and by Cardio- MD, The Cleveland Clinic Foundation, Department of Emer-
Dynamics (San Diego, CA), which provided a study grant for gency Medicine, Desk E-19, 9500 Euclid Avenue, Cleveland,
support of research assistants. In addition, CardioDynamics OH 44195. Fax: 216-445-4552; e-mail: peacocw@ccf.org.
91 of 158.
366 Peacock et al. IMPEDANCE CARDIOGRAPHY IN EMERGENT DYSPNEA
including cardiac output, systemic vascular resistance, by the ED physician. Because ICG is not currently recom-
and fluid status, may provide important information mended (per U.S. Food and Drug Administration guide-
and aid decision making beyond what is possible from lines) for the diagnosis of acute coronary syndromes,
history and physical examination alone. Unfortunately, including acute myocardial infarction (MI), patients were
hemodynamic parameters cannot be accurately deter- excluded if electrocardiogram (ECG) or serum markers
mined by patient history or physical examination.25 Until were positive for acute MI. Additional exclusions in-
recently, hemodynamic data could only be obtained by cluded the following: if ICG monitoring was not possible
pulmonary artery catheterization. Because this invasive because of inability to place electrodes, if the patients
procedure is not practical in the ED, physicians typically weight was greater than 341 pounds, or if the patient
are left to make diagnosis and treatment decisions with- had an activated minute ventilation pacemaker (which
out reliable information about a patients hemodynamic uses an impedance signal). Also, although severe aortic
status. regurgitation that could give a falsely elevated ICG CO
Noninvasive hemodynamic monitoring by impedance is rare and generally evident on ED evaluation, we ex-
cardiography (ICG) has been used in more than four cluded those with aortic regurgitation by past history,
million patients. Cardiac output (CO) by ICG has been and those with the typical diastolic murmur. Last, because
shown to correlate well with CO obtained by invasive the treatment and disposition actions for patients needing
methods in hospitalized patient populations with correla- immediate intubation and mechanical ventilation are gen-
tion coefficients for CO by ICG and thermodilution rang- erally well defined from the emergency physician point
ing from 0.76 to 0.89.610 ICG also has been used as an of view, and because it was our intent to study the
alternative to invasive monitoring in the critical care diagnostically most challenging patients, those requiring
setting.11 In the ED setting, ICG has been studied for the urgent intubation and mechanical ventilation upon
differential diagnosis of dyspnea1214 and the identification presentation to the ED were excluded.
of pulmonary edema15,16 and provides prognostic infor-
mation about hospitalization costs and length of stay.17 Study Protocol
ICG results are available within a few minutes, allowing Project coordinators screened candidates and an inde-
more rapid patient evaluation than that afforded by pendent research nurse, not involved in the diagnosis
radiographic or laboratory studies. or treatment of the patient, obtained hemodynamic
Given the high rate of morbidity, mortality, and hospi- data. Hemodynamic data were collected by using the
tal readmissions for patients with dyspnea and acute de- BioZ ICG monitor (CardioDynamics, San Diego, CA), as
compensated HF, there is an urgent need to examine has been described elsewhere.22 ICG data are obtained
technologies that could lead to improvements in care in by the following technique: four dual sensors (each sen-
the ED. The present study examines an aspect of thera- sor consisting of two electrodes) are placed on the
peutic efficacy18 as it relates to ICG and the acutely dysp- patient, as shown in Figure 1, on opposite sides of the
neic emergency patient, and not simply the performance neck at a level between the ears and shoulders and on ei-
of ICG as a testing modality. Put into context, previous ther side of the chest in the mid-axillary line at the level of
studies of commonly used ED tools, such as pulse oxim- the xiphoid process. The outer electrodes in each sensor
etry19,20 and B-type natriuretic peptide (BNP) testing,21 transmit a low-amplitude, high-frequency current (2.5
suggest that a 5% to 11% rate of change in diagnosis, mA, 70 kHZ), and the inner electrodes detect thoracic
or 10% rate of change in therapy, is clinically relevant. voltage changes. Changes in voltage are used to calculate
The effect of ICG-derived hemodynamics on diagnosis changes in impedance (Z). Baseline, static impedance is
and treatment of dyspnea in the ED is not yet known. indicative of chest fluid volume, and dynamic impedance
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.
92 of 158.
ACAD EMERG MED April 2006, Vol. 13, No. 4 www.aemj.org 367
is affected by aortic blood volume and velocity. Beat- the diagnosis changed on the basis of ICG data, a com-
to-beat changes in thoracic impedance are processed to parison to the final diagnosis was made to determine
calculate blood flow per heartbeat (stroke volume) and whether the pre- or post-ICG diagnosis was more consis-
per minute (cardiac output). By using standard equations, tent with the final ED diagnosis. The final ED primary di-
other hemodynamic parameters, such as systemic vascu- agnosis was defined as the principal diagnosis at the end
lar resistance, are calculated. The reciprocal of baseline of the ED visit after all diagnostic testing was completed
thoracic impedance can provide an index of intrathoracic and reviewed by the ED physician responsible for dispo-
fluid and is termed thoracic fluid content (TFC). TFC has sition. A change in therapy was defined as the addition or
been used to identify intravascular and extravascular subtraction of a drug or procedure. Changing the dose of
fluid changes23,24 and to titrate diuretic therapy.25 a previously ordered drug was not considered a thera-
Before study initiation, participating physicians re- peutic change. Adverse events were defined as cardiac
ceived instruction regarding the interpretation of the arrest, intubation for respiratory failure, urgent cardio-
hemodynamic values obtained by the ICG device. version, or blood transfusion.
Attending physicians, all of whom were board-certified
or board-eligible in emergency medicine, were given a Data Analysis
description of ICG technology and hemodynamic param- The size of the study was prospectively determined on the
eters provided on the ICG report, including definitions basis of the number needed to detect a 5% rate of change
and normal values for cardiac index (CI), resistance, tho- in diagnosis or therapy. Given an alpha of 0.05 and a beta
racic fluid content, and measures reflecting left ventricu- of 0.20, a sample size of 100 was needed to detect a statis-
lar performance. This was performed at departmental tically significant change. Data were analyzed by an inde-
grand rounds and at the monthly attending-physician pendent statistician using SAS Software (Cary, NC).
staff meeting. Additional information was disseminated Demographic data are reported descriptively. Continu-
in hardcopy by mailing and was duplicated by e-mail. ous variables are reported as mean standard deviation
The pathophysiology of HF and hemodynamic findings (SD). Rates of change were calculated by dividing the
most suggestive of dyspnea caused by decompensated number of patients in whom diagnosis or therapeutic
HF (reduced CI, elevated systemic vascular resistance, plan changed by the total number of patients and were
and increased TFC) were described. The expected effects reported as percentages. An analysis of variance was
of various medications on hemodynamic parameters performed to assess for differences among vital signs
were discussed, including use of diuretics, vasodilators, and ICG parameters in the final diagnosis categories.
and drugs affecting contractility. Additionally, physicians
were provided personal reference cards for use at their
discretion that detailed normative values for all ICG RESULTS
data. Copies of the data card were also kept fixed to the
Eighty-nine patients, cared for by 31 ED staff physicians,
ICG device. These data were also shown at the time of
were enrolled from December 2001 through July 2003
ICG unblinding. For any given parameter, ICG data are
and are included in the analysis. No adverse event,
presented as a bar indicating the normal human range.
defined as cardiac arrest, intubation, cardioversion, or
The average result and the currently measured data point
blood transfusion, occurred during the course of the
then are indicated on this bar, such that variations from
ED observation during this study.
normal are readily apparent.
The patient characteristics and vital signs are summa-
All staff involved with patient care were blinded to the
rized in Table 1. The mean (SD) age of the subjects
ICG data until after the initial history and physical exam-
was 74.8 (7.0) years. Fifty-eight percent of the patients
ination by the attending physician. After the initial his-
were African American, and 61% had a history of HF,
tory and physical exam, but before initiation of therapy
(other than supplemental oxygen), and before obtaining
any central laboratory or radiographic data, the attend- Table 1
ing physician completed a case report form indicating Patient Characteristics and Vital Signs
his or her working diagnoses and short-term therapeutic
plans. The physician was then immediately shown the Patient Characteristic Value
ICG hemodynamic data and was asked to complete the Male, n (%) 42 (47)
case report form again, this time with consideration of Ethnic background, n (%)
the ICG data. All patient care then proceeded according African American 52 (58)
White 36 (41)
to usual ED routine. Blood tests, including electrolytes,
Other 1 (1)
blood urea nitrogen (BUN), serum creatinine (Cr), and Medical history, n (%)
BNP levels, were obtained in the majority of cases. History of heart failure 43 (48)
Although these data were not mandated as part of the History of chronic obstructive 34 (38)
protocol, they were used in most cases to determine final pulmonary disease or asthma
ED diagnosis. History of heart failure and COPD 11 (13)
Vital signs (mean SD)
Measures Temperature (C) 36.5 0.7
The two primary endpoints were 1) the rates of change Respiratory rate (min1) 22.2 5.1
Heart rate (min1) 84.6 18.8
in working diagnosis and 2) medical therapy after the
Systolic blood pressure (mm Hg) 145.6 29.3
addition of ICG data to the physicians initial clinical
N = 89.
assessment and therapeutic plan. In the cases in which
93 of 158.
368 Peacock et al. IMPEDANCE CARDIOGRAPHY IN EMERGENT DYSPNEA
94 of 158.
ACAD EMERG MED April 2006, Vol. 13, No. 4 www.aemj.org 369
catheter, performance of ICG is noninvasive and can be nosis and therapy in a manner that would be consistent
readily accomplished in the ED without specialized train- with clinical practice in the ED, where patients present-
ing and at minimal risk to the patient. ing with dyspnea might be evaluated with ICG either
The magnitude of the changes in diagnosis and treat- before or within minutes of the ED physicians initial
ment resulting from ICG-derived hemodynamic data assessment. Furthermore, as seen in our study, the find-
can be compared with that from other technologies ings of ECGs and chest radiographs are often normal or
that are currently the standard of care in most EDs. His- nonspecific and may not provide significant diagnostic
torically, changes in therapy on the order of 5% to 11% certainty. Because ICG is not part of the diagnostic crite-
appear to define utility of testing in the ED. In one study, ria for acute coronary syndromes, including acute MI, we
Summers et al.19 reported that the ED physician assess- did exclude patients with evidence of myocardial necro-
ment of patient severity of illness was changed by pulse sis from analysis. Therefore, the role of ICG in providing
oximetry in 3% of cases. Kosowky et al.,21 evaluating possible clues in the evaluation of patients with dyspnea
BNP testing in patients older than 40 years of age, found as a manifestation of MI cannot be assessed by the pres-
that BNP data changed the diagnosis in 10%, and treat- ent study.
ment in 11%, of cases. These trials suggest that the rate Our study was also limited by the use of the final ED
of change that resulted from ICG use in the present study diagnosis as the criterion standard for diagnostic cate-
would be clinically significant in the ED environment. gorization. Although it is possible that this diagnosis
Moreover, ICG can be performed concurrently with was incorrect or incomplete in some patients, this repre-
existing diagnostic and therapeutic strategies, such that sents the real-life diagnosis based on current evaluation
the information is incremental in the decision-making strategies during the patients ED visit. It is also possible
process. that a physician had the right diagnosis and treatment
The changes in ED decision making from esophageal plan before reviewing ICG results and that ICG data
Doppler results, a more invasive and less common form resulted in inappropriate therapies. A larger prospective
of cardiac output measurement, have been studied else- outcome-based study will be required to determine the
where.28 Those investigators found a change in man- potential for this to occur.
agement decisions in 31% of cases. Our results show a In our study, ICG data were available and likely con-
greater change in therapy alone, perhaps because of tributed to the final ED diagnosis, thereby introducing
the incremental information provided by systemic vascu- possible bias. However, the goal of this study was not
lar resistance and thoracic fluid content parameters. to assess technical accuracy of the technology, which
Although most ED physicians would not subject a patient has been evaluated in previous studies. In contrast, this
to esophageal monitoring to obtain hemodynamic mea- study was designed to assess whether physicians would
surements, it is likely that many would consider the incorporate early hemodynamic information into the
collection of ICG data, which requires little more time or process of formulating an initial working diagnosis and
inconvenience than obtaining an electrocardiogram. treatment plan. In addition, the study design does not al-
We did not measure the time required to obtain ICG low us to draw conclusions about the sensitivity or spec-
data; however, in routine use, these data can be obtained ificity of ICG criteria, or to compare diagnostic accuracy
in about 3 to 5 minutes and require 30 to 60 seconds to other measures, such as BNP or chest radiography.
to interpret. Because ICG provides early and accurate The accuracy of the post-ICG diagnosis based on these
data, there is a potential for significant clinical impact hemodynamic criteria could only be verified by a more
from its use. We did not specifically study the financial standardized diagnostic approach including cardiac
effects of ICG in this study or how it might have affected imaging studies, blinded reviews of subsequent hospital
length of ED stay or hospital admission rate. However, records with adjudication of discordant diagnoses, and
at a procedural cost for each test of less than $20 and long-term follow-up, which were not within the scope
with the cost of a day in intensive care at more than of the current study.
$1,000, the provision of ICG would be cost-effective
even if, for example, it reduced hospital length of stay CONCLUSIONS
by only one day for every 50 patients monitored.
Knowledge of ICG data early in the ED evaluation of pa-
LIMITATIONS tients older than 65 years of age presenting with dyspnea
results in significant changes in diagnosis and treatment
We acknowledge several limitations. This study evalu- plan. Whether changes in diagnosis, diagnostic certainty,
ated the effect of ICG on working diagnosis and initial or therapy from ICG improve outcomes or are cost-effec-
treatment plan before the results of chest radiograph, tive will require a prospective, randomized clinical trial
ECG, or BNP level. Thus, it is impossible to gauge the rel- with longer periods of clinical follow-up.
ative importance of the information obtained from ICG
to that obtained by these other tests or to judge the addi- The authors thank Gerard Smits, PhD, for his statistical assis-
tional contribution of ICG for cases in which the results tance.
of other tests were available before performing ICG. Al-
though blood work and various ancillary testing such as References
chest radiography are part of the complete ED evaluation
of such patients, the results are generally not available 1. McCraig LF, Burt CW. National hospital ambulatory
within the first few minutes of patient assessment. By de- medical care survey: 2001 emergency department
sign, this study evaluated ICGs effect on working diag- summary. Report from Centers for Disease Control
96 of 158.
ACAD EMERG MED April 2006, Vol. 13, No. 4 www.aemj.org 371
and Prevention, National Center for Health Statistics. tide levels and non-invasive hemodynamic parame-
Advance Data Vital Health Stat. 2003; 335:18. ters in diagnosing congestive heart failure in the
2. Eisenberg PR, Jaffe AS, Schuster DP. Clinical evalua- emergency department. Congest Heart Fail. 2004;
tion compared to pulmonary artery catheterization in 10:1716.
the hemodynamic assessment of critically ill patients. 15. Peacock WF, Albert NM, Kies P, White RD, Emerman
Crit Care Med. 1984; 12:54953. CL. Bioimpedance monitoring: better than chest
3. Speroff T, Connors AF Jr, Dawson NV. Lens model x-ray for predicting abnormal pulmonary fluid? Con-
analysis of hemodynamic status in the critically ill. gest Heart Fail. 2000; 6(2):325.
Med Decis Making. 1989; 9:24352. 16. Newman RB, Pierre H, Scardo J. Thoracic fluid con-
4. Neath SX, Lazio L, Guss DA. Utility of impedance car- ductivity in peripartum women with pulmonary
diography to improve physician estimation of hemo- edema. Obstet Gynecol. 1999; 94:4851.
dynamic parameters in the emergency department. 17. Milzman D, Morrisey J, Pugh C, Napoli A, Gerace T,
Congest Heart Fail. 2005; 11:1720. Fernandez E. Occult perfusion deficits in heart failure
5. Van De Water JM, Dalton ML, Parish DC, Vogel RL, patients: identification through noninvasive central
Beatty JC, Adeniyi SO. Cardiopulmonary assess- hemodynamic monitoring [abstract]. Crit Care Med.
ment: is improvement needed? World J Surg. 2005; 1999; 27:A88.
29(Suppl 1):S958. 18. Pearl WS. A hierarchical outcomes approach to test
6. Sageman WS, Riffenburgh RH, Spiess BD. Equiva- assessment. Ann Emerg Med. 1999; 33:7784.
lence of bioimpedance and thermodilution in mea- 19. Summers R, Anders R, Woodward L, Jenkins A, Galli
suring cardiac index after cardiac surgery. J R. Effect of routine pulse oximetry measurements on
Cardiothorac Vasc Anesth. 2002; 16:814. ED triage classification. J Emerg Med. 1999; 16:57.
7. Drazner M, Thompson B, Rosenberg P, Yancy C. 20. Mower WR, Sachs C, Nicklin EL, Safa P, Baraff LJ.
Comparison of impedance cardiography with inva- Effect of routine emergency department triage pulse
sive hemodynamic measurements in patients with oximetry screening on medical management. Chest.
heart failure secondary to ischemic or nonischemic 1995; 108:1297302.
cardiomyopathy. Am J Cardiol. 2002; 89:9935. 21. Kosowsky JM, Weiner C, Morrissey JH. Impact of
8. Van De Water JM, Miller TW. Impedance cardiogra- B-type natriuretic peptide testing on medical deci-
phy: the next vital sign technology? Chest. 2003; 123: sion-making for older patients with dyspnea. Ann
202833. Emerg Med. 2003; 42:S11.
9. Yung GL, Fedullo PF, Kinninger K, Johnson FW, 22. Summers R, Schoemaker W, Peacock WF, Ander D,
Channick RN. Comparison of impedance cardiogra- Coleman T. Bench to bedside series: impedance car-
phy to direct Fick and thermodilution cardiac output diography (ICG). Acad Emerg Med. 2003; 10:66980.
determination in pulmonary arterial hypertension. 23. Luepker R, Michael JR, Warbasse JR. Transthoracic
Congest Heart Fail. 2004; 10(2 Suppl 2):710. electrical impedance: quantitative evaluation of a
10. Albert NM, Hail MD, Li J, Young JB. Equivalence of noninvasive measure of thoracic fluid volume. Am
bioimpedance and thermodilution in measuring car- Heart J. 1973; 85:8393.
diac output in hospitalized patients with advanced, 24. Van de Water JM, Mount BE, Chandra KM, Mitchell
decompensated chronic heart failure. Am J Crit BP, Woodruff TA, Dalton ML. TFC (thoracic fluid
Care. 2004; 3:46979. content): a new parameter for assessment of changes
11. Silver M, Cianci P, Brennan S, Longeran-Thomas H, in chest fluid volume. Am Surg. 2005; 71:816.
Ahmad F. Evaluation of impedance cardiography as 25. Taler SJ, Textor SC, Augustine JE. Resistant hyper-
an alternative to pulmonary artery catheterization tension: comparing hemodynamic management to
in critically ill patients. Congest Heart Fail. 2004; 10(2 specialist care. Hypertension. 2002; 39:9828.
Suppl 2):1721. 26. Maisel AS, Krishnaswamy P, Nowak RM, et al. Rapid
12. Springfield C, Sebat F, Johnson D, Lengle S, Sebat C. measurement of B-type natriuretic peptide in the
Utility of impedance cardiography to determine emergency diagnosis of heart failure. N Engl J Med.
cardiac vs. noncardiac cause of dyspnea in the 2002; 347:1617.
emergency department. Congest Heart Fail. 2004; 27. Peacock WF, Roe MT, Chen AY, et al. Vein-to-brain
10(2 Suppl 2):146. time: an emergency department quality of care
13. Han J, Lindsell C, Tsurov B, Storrow A. The clinical marker for non-ST-segment elevation acute coro-
utility of impedance cardiography in diagnosing con- nary syndrome [abstract]. Acad Emerg Med. 2004;
gestive heart failure in dyspneic emergency depart- 11:569.
ment patients [abstract]. Acad Emerg Med. 2002; 9: 28. Urrunaga J, Rivers E, Mullen M, et al. Hemodynamic
43940. assessment of the critically ill: the clinician versus
14. Barcarse E, Kazanegra R, Chen A, Chiu A, Clopton esophageal Doppler monitoring (EDM) [abstract].
P, Maisel A. Combination of B-type natriuretic pep- Acad Emerg Med. 2000; 7:587b.
97 of 158.
AJH 2005; 18:26S 43S
Hypertension is the most common cardiovascular disease, namic parameters. Measurement of the various hemody-
affecting approximately 60 million Americans. Despite the namic components using ICG in those with hypertension
importance of this condition, only the minority of patients allows more complete characterization of the condition, a
are appropriately identifie and treated to reach recom- greater ability to identify those at highest risk, and allows
mended blood pressure (BP) goals. Although historically more effectively targeted drug management. This article
define as an elevation of BP alone, hypertension is char- reviews the importance of hemodynamic factors in hyper-
acterized by abnormalities of cardiac output, systemic tension and the evolving role of ICG technology in the
vascular resistance, and arterial compliance. These hemo- assessment and management of this important cardiovas-
dynamic aspects of hypertension have implications for cular condition. Am J Hypertens 2005;18:26S 43S
diagnosis, risk stratification and treatment. Impedance 2005 American Journal of Hypertension, Ltd.
cardiography (ICG) has emerged as a unique and highly Key Words: Hypertension, hemodynamics, impedance
accurate noninvasive tool that is used to assess hemody- cardiography.
hen functioning properly, the cardiovascular hypertension has intrigued scientists and clinicians since
Received October 29, 2004. Accepted November 4, 2004. Address correspondence and reprint requests to Dr. Hector O.
From the Ochsner Clinic Foundation (HOV), New Orleans,Loui- Ventura, Ochsner Clinic Foundation, 1514 Jefferson Highway, Room
siana; Mayo Clinic College of Medicine (SJT), Rochester, Minnesota; 3E419, New Orleans, LA 70121; e-mail: Hventura@ochsner.org
and Cabrini Medical Center (JES), New York, New York.
Table 1. JNC 7 Classification of blood pressure for gencies. The etiologies of hypertension (essential versus
adults 18 years of age secondary) and the various clinical presentations (benign,
malignant, unspecified and with or without associated
Systolic Diastolic co-morbidities) are reflecte in the World Health Organi-
BP BP
BP classification (mm Hg) (mm Hg)
zations International Classificatio of Diseases, Ninth
Revision (ICD-9), coding for hypertension (Table 2).
Normal 120 and 80
Prehypertension 120139 or 8089
Stage 1 hypertension 140159 or 9099 Hypertension: Magnitude of the
Stage 2 hypertension 160 or 100 Problem
JNC Joint National Committee on the Prevention, Detection, Eval- Hypertension affects up to 60 million Americans and as
uation, and Treatment of High Blood Pressure.
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
dence using noninvasive monitoring of hemodynamics significantl with advancing age (Fig. 1), such that a
with ICG. The specifi role of hemodynamics in diagnos- normotensive adult in the United States 55 years of age
tic, prognostic, and therapeutic decision making in the still has a 90% lifetime risk of developing hypertension.9
patient with hypertension is reviewed in detail. In fact, the most common group with hypertension is
comprised of elderly patients with systolic hypertension.12
Hypertension: Definition and Although controlling BP levels reduces the incidence of
stroke and other cardiovascular complications, BP control
Clinical Presentation in the US is well below stated goals. For an individual
Hypertension is most commonly define as a systolic BP patient, this may be due to the lack of recognition of the
(SBP) of 140 mm Hg or a diastolic BP (DBP) of 90 condition, failure to institute effective treatment, or the result
mm Hg. In patients at high risk for complications from of a suboptimal long-term medical regimen (Table 3). There
elevated BP levels, such as those with diabetes or chronic remains a substantial need for improvement in the effective-
renal disease, lower levels of BP (eg, 130/80 mm Hg) ness of hypertension treatment. As reported in JNC 7, only
are recommended. Between BP levels of 115/75 mm Hg 34% of adults aged 18 to 74 years with hypertension have
and 185/115 mm Hg, each 20 mm Hg increase in SBP or achieved BP control, despite a published goal of 50%. In
10 mm Hg increase in DBP doubles the risk of a cardio- the elderly population, BP control is even less successful:
vascular event.9 In recognition of this increase in risk from fewer than 20% of treated patients 70 years or more of age
levels as low as 115/75 mm Hg, the Seventh Report of the attain BP levels of 140/90 mm Hg.9
Joint National Committee on Prevention, Detection, Eval- Hypertension treatment commonly requires multiple
uation, and Treatment of High Blood Pressure (JNC 7)9 medications. Refractory hypertension has been define
recently identifie prehypertension (define as a BP of as hypertension that is not controlled on two or more
120 to 139 / 80 to 89 mm Hg) as a significan potential antihypertensive medications.13 Resistant hypertension
health problem associated with increased risk for pro- has been define by some as BP readings of 140/90 mm
gression to hypertension. The JNC 7 recommended life- Hg despite an optimal two-drug regimen that has had
style modificatio for the management of prehypertension. adequate time to work (at least 1 month since last drug or
Table 1 lists the stages of hypertension as define in the dosage adjustment).14 As define by Gifford, resistant
JNC 7 classification hypertension is the failure to reach goal BP in patients who
Secondary hypertension results from an identifiabl are adhering to full doses of an appropriate three-drug
cause such as renal, adrenal, or vascular pathology. In regimen that includes a diuretic.15 The JNC 7 recommends
contrast, 90% of patients have no specifi identifiabl a goal BP of 140/90 mm Hg for the general population,
cause of their BP elevation10 and are thus diagnosed with with the tighter goal of 130/80 mm Hg for persons with
essential hypertension. Most patients with hypertension chronic renal disease or diabetes mellitus.
are asymptomatic and do not show evidence of acute Hypertension substantially increases the incidence of
pathologic changes; their clinical presentation has been cardiovascular events, especially the risk of stroke. Wilk-
termed benign, although their long-term risk of cardio- ing et al,16 in data from the Framingham study, reported on
vascular complications is significantl greater than in nor- the prognostic significanc of systolic hypertension. They
motensive persons without so-called benign hypertension. found that for men and women, the relative risk of car-
Severe elevation of BP, associated with papilledema on diovascular disease event adjusted for age was approxi-
fundoscopic examination, is termed malignant hyperten- mately 2.5 times greater for persons with isolated systolic
sion. Similar levels of BP elevation with congestive heart hypertension compared with those with BP levels
failure, anginal symptoms, or other evidence for acceler- 140/95 mm Hg. Lower BP levels are thus associated
ated end-organ injury (but without papilledema) are with improved prognosis and decreased incidence of mor-
termed hypertensive urgencies or hypertensive emer- bidity and mortality. From pooled data of more than 60
99 of 158.
28S HYPERTENSION AS A HEMODYNAMIC DISEASE AJHFebruary 2005VOL. 18, NO. 2, Part 2
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 Importantly, the authors note that even a 2mm Hg reduc-
report that 10 mm Hg reductions in systolic BP would be tion in systolic BP is associated with a 10% lower death
expected to reduce stroke mortality by as much as 40%. 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
FIG. 1 Prevalence of hypertension by age. 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
expensive costing businesses an average of $392 per offic setting, CO is not generally reported by most phy-
eligible employee per year.19 Improving the efficienc and sicians interpreting echocardiograms in clinical practice.
effectiveness of drug management in the hypertensive Thus, until recently, CO and SVR were commonly mea-
population would likely reduce these costs in addition to sured only in the intensive care setting or catheterization
decreasing morbidity and mortality associated with the laboratory setting using invasive means such as a pulmo-
condition. nary artery catheter.
In recent years, ICG has emerged as an accurate, safe,
and inexpensive tool with which to measure hemodynamic
Hemodynamic Measurements parameters by noninvasive means. The procedure is most
Using ICG commonly performed in the physician offic setting by
The historical use of BP without CO or SVR is, in part, medical assistants or nurses, requiring about 5 min to
because it has been impractical to estimate or measure complete the test. Using four sets of paired sensors on the
these parameters in most clinical settings. The assessment neck and chest, ICG measures the instantaneous change of
of the hemodynamic components of hypertension from an electrical signal across the thoracic cavity (Fig. 2). As
clinical evaluation alone is unreliable. Even in patients the changes of thoracic impedance during the cardiac
with acute conditions such as those requiring the emer- cycle are most dependent on the changes in the size and
gency department or patients with decompensated conges- the blood volume of the thoracic aorta, ICG is able to
tive heart failure (in whom hemodynamic derangements calculate the amount of blood ejected from the left ven-
are greater than those in patients with essential hyperten- tricle (that is, the stroke volume [SV]). The product of
sion), clinicians are generally unable to estimate CO or heart rate (HR) and SV yields CO. In addition, ICG-
SVR with accuracy.20,21 derived parameters related to the changes of thoracic im-
Echocardiography has been used to measure cardiac pedance are indicative of aortic blood velocity and
output and in some studies has demonstrated acceptable acceleration, and they correlate with measures of inotropic
correlation with invasive techniques.22 However, in a state and cardiac performance. As flui is the best con-
comparison with ICG, echocardiography is considerably ductor of the electrical signal through the chest (when
more time consuming and technically demanding.23 In the
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
102 of 158.
AJHFebruary 2005VOL. 18, NO. 2, Part 2 HYPERTENSION AS A HEMODYNAMIC DISEASE 31S
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 atrial fibrillatio or frequent premature ventricular con-
related to the technology and patient factors. Although tractions, marked irregularity in heart rhythm can affect
ICG equations have demonstrated accuracy over a wide data collection and analysis of wave forms.
range of conditions and patient populations, ICG has
not been evaluated extensively in patients 66 pounds
or 342 pounds. Severe aortic insufficienc may affect Hemodynamic Parameters in
ICG reliability, but it has not been fully studied and Hypertension
validated in such patients. In addition, a few models of Hypertension is the result of complex cardiac, renal, neu-
permanent pacemakers use impedance technology to rohormonal, and vascular mechanisms that are modulated
measure minute ventilation. If the minute ventilation by both genetic and environmental factors.10,32 The inter-
function is activated, the paced rate may increase be- actions of these many factors result in endothelial dysfunc-
cause of ICG signals31; therefore, patients with such tion and hemodynamic derangements of arterial
pacemakers must have the minute ventilation sensor compliance, CO, and SVR. As noted earlier, MAP is the
function inactivated before ICG testing. In patients with product of CO and SVR, and elevations of BP can result
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 AJHFebruary 2005VOL. 18, NO. 2, Part 2
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 FIG. 6 Components of mean arterial pressure (MAP) and pulse
pressure (PP). CO cardiac output; SV stroke volume; SVR
systemic vascular resistance.
104 of 158.
AJHFebruary 2005VOL. 18, NO. 2, Part 2 HYPERTENSION AS A HEMODYNAMIC DISEASE 33S
105 of 158.
34S HYPERTENSION AS A HEMODYNAMIC DISEASE AJHFebruary 2005VOL. 18, NO. 2, Part 2
Table 8. Improved predictive power of pulse pres- average of 16.2 years of follow-up. In this study, exercise
sure (PP) to stroke volume index ratio (SVIR) com- SVR but not exercise BPadded prognostic value to
pared with pulse pressure alone* parameters measured at rest, suggesting that hemodynamic
variables other than BP might have greater prognostic
Hazard rate, Hazard rate,
CV event mortality
value.
A subsequent article56 reported the relationship between
PP 1.46 (0.912.35) 1.65 (0.833.30) another hemodynamic parameter, namely, the ratio of PP to
PP to SVIR 1.79 (1.152.80) 2.05 (1.153.65)
stroke index (PP-to-SVi ratio), and outcomes in patients
CV cardiovascular. followed for an average of 16.5 years. In this study, the
Values and 95% confidence intervals for 1-SD increase in the PP-to-SVi ratio, that is, the reciprocal of TAC index, was
variable of interest.
Adapted from Fagard et al.56
independently associated with cardiovascular events or death.
* Adjusted for age, gender, mean arterial pressure, and heart Each increase in PP-to-SVi ratio of 0.75 mm Hg/(mL/m2)
rate; P NS; P .01. 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
MAPwith affective state is further evidence of the het- increased hazard rates with PP-to-SVi ratio compared with
erogeneity of hemodynamic subsets within various groups PP alone demonstrates the additional predictive value when
clinically identified the hemodynamic parameter of flo (SVi) is added to the
Hinderliter53 reported that African American men and pressure measurement alone.
women had increased SVR, decreased CO, and associated De Simone et al57 studied the effects of TAC on car-
LV remodeling compared with Caucasian men and women diovascular events over a 10-year period. They found that
despite similar BP readings. In normotensive African risks of fatal and total cardiovascular events were inde-
Americans, Calhoun et al54 postulated that vasoconstrictor pendently correlated with age, LV mass, and lower levels
responses seen with mental stress and cold presser testing of arterial compliance, define as decreasing values of the
may contribute to elevated SVR and the development of measured ratio of echocardiographic SV to PP (SV/PP) to
hypertension. that predicted from previously developed equations (%
These and other studies using ICG technology show SV/PP). Moreover, consistent with the results of Fagard et
that within any given population SVR, CO, and TAC show al,56 the investigators found that systolic BP, mean BP, or
significan variation. The heterogeneity of hemodynamic PP alone (without including the flow-relate hemody-
finding within various cohorts is evidence that the spe- namic parameter of SV) were not independent predictors
cifi hemodynamic status of an individual patient cannot of prognosis. After adjustment for age and LVH there
reliably be predicted on the basis of age, gender, or ethnic remained an independent effect of % SV/PP on cardiovas-
background. Moreover, that hemodynamic values cannot cular endpoints at 10-year follow-up (Fig. 9). These in-
be identifie by BP levels or clinical assessment alone vestigators found that hemodynamic parameters such as
supports the need for measurement of hemodynamic pa- arterial compliance and percent predicted arterial compli-
rameters in individual patients with hypertension. ance correlated better with changes in cardiac structure (ie,
hypertrophy and remodeling) than did BP levels alone.
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.
106 of 158.
AJHFebruary 2005VOL. 18, NO. 2, Part 2 HYPERTENSION AS A HEMODYNAMIC DISEASE 35S
Gender has long been known to affect prognosis in was normal in individuals with normal BP levels. Others
patients with hypertension. In 1913, Janeway58 published have evaluated measures of arterial compliance (or, alter-
the observation that women with hypertension tended to natively, arterial stiffness) using the measure of pulse
have a better prognosis than men. More recent studies wave velocity.66,67
have suggested that this difference may be related to Additional structural changes occuring at the level of
different hemodynamic substrates in men compared to the heart and blood vessel have prognostic significanc in
women with elevated BP levels. Messerli et al44 suggested persons with hypertension, including vascular remodeling
that the disparate prognosis between men and women with changes in lumen to wall thickness.68,69 Apoptosis, or
might be explained on the basis of differing hemodynamic programmed cell death, contributes to the vascular
mechanisms: For any level of arterial pressure, total changes (ie, remodeling) in hypertension. Inflammatio
peripheral resistance (and therefore the risk of hyperten- and fibrosi similarly contribute with the accumulation of
sive cardiovascular disease) was lower in women than in various components in the extracellular matrix such as
men. collagen and fibronectin Intengan and Schiffrin69 re-
The mechanism of the adverse prognosis from hyper- viewed the factors that result in arterial remodeling and
tension is in part related to structural changes in the heart altered hemodynamic parameters in patients with hyper-
that result from elevated wall stress. Prolonged increases tension.
in wall stress lead to left ventricular (LV) structural Importantly, studies have shown that treatment with
changes with relative increases in wall thickness, overall antihypertensive agents may result in regression of the
LV mass or both.59 In concentric remodeling, there is a structural abnormalities caused by long-standing hyperten-
relative increase in LV thickness without increase in over- sion and may result in improved prognosis.70 72 Mathew
all LV mass. This structural change appears to be related et al,70 reporting on data from the Heart Outcomes Pre-
to increased pressure load but with relative decrease in vention Evaluation (HOPE) study, demonstrated that treat-
volume as evidenced by low normal CO (termed volume ment with the angiotensin-converting enzyme (ACE)
underload). Concentric left ventricular hypertrophy inhibitor ramipril was associated with regression of LVH
(LVH) is characterized by an increase in wall thickness by electrocardiographic criteria compared with placebo
with increase in LV mass or mass index and also results control. That BP showed minimal difference between the
from pressure overload caused by long-term hypertension. treatment and control groups is consistent with other stud-
Eccentric hypertrophy is define as increased LV mass ies demonstrating improvement in overall hemodynamics
index with preserved relative wall thickness and is asso- (as shown by significan decreases in SVR and parallel
ciated with both pressure and volume overload. This pat- increases in CO) that are not evident from BP levels alone.
tern, a common result of the afterload and volume Ofil et al,73 in an echocardiographic substudy of the
excesses in long-standing severe aortic insufficiency re- Systolic Hypertension in the Elderly Program (SHEP),
sults in spherical remodeling of the LV. Interestingly, in demonstrated partial regression of LVH in patients treated
this study of hypertensive individuals,59 both eccentric with a diuretic-based regimen for a minimum of 3 years. In
hypertrophy and concentric remodeling were more com- a meta-analysis of 1000 patients with serial echocardi-
mon than the classic pattern of hypertensive heart dis- ography during treatment of essential hypertension, Ver-
ease, namely, concentric LVH. decchia et al74 demonstrated that patients whose LVH
Nonetheless, LVH is a powerful predictor of cardiovas- regressed during treatment had significantl fewer cardio-
cular risk and is independently associated with mortality in vascular events compared with those in whom LV mass
patients with coronary artery disease.60 63 For example, increased, consistent with the hypothesis that improve-
Vakili et al63 reported on the pooled results of 20 pub- ments in hemodynamics correlate with improved progno-
lished studies of LVH as define by electrocardiographic sis in patients with appropriately treated hypertension.
or echocardiographic criteria. They demonstrated a
weighted mean relative risk of cardiovascular morbidity
Impedance Cardiographic Studies
from LVH of 2.3, independent of all covariates analyzed.
As reported by Ichkhan et al,64 LVH is associated with The ICG technique has been used to explore age-related
abnormalities of ventricular repolarization and at least a changes in hemodynamic variables and their correlation
twofold increase in the risk of serious ventricular arrhyth- with cardiovascular risk and the adverse prognosis. These
mias. studies support previous finding that future risk in pa-
Hypertension results in abnormalities of endothelial tients with hypertension may not be reflecte in BP levels
function, affecting hemodynamic factors such as arterial alone. Alfi et al50 demonstrated that despite similar ele-
compliance. Gomez-Cerezo et al65 demonstrated impaired vations in PP, younger men had preserved stroke index
brachial artery flowmediate dilation, a common test of (and arterial compliance) compared with older men. They
endothelial function, in patients with sustained or labile concluded that preserved arterial compliance and cardiac
hypertension. They found that flow-mediate dilation was pump function may explain the lack of prognostic signif-
abnormal to a similar degree in patients with sustained icance of elevated PP in younger men. These finding lend
essential hypertension or white-coat hypertension but further support to the value of the incremental information
107 of 158.
36S HYPERTENSION AS A HEMODYNAMIC DISEASE AJHFebruary 2005VOL. 18, NO. 2, Part 2
108 of 158.
AJHFebruary 2005VOL. 18, NO. 2, Part 2 HYPERTENSION AS A HEMODYNAMIC DISEASE 37S
-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
NC
2
2
1
2
1
1 or NC
1 or NC
NC or 2
as to optimize therapy.
1 or NC
LVH
CO
109 of 158.
38S HYPERTENSION AS A HEMODYNAMIC DISEASE AJHFebruary 2005VOL. 18, NO. 2, Part 2
impedance, is strongly correlated with amount of flui in with fall in diastolic BP. An increase in overall thoracic
the chest cavity, whether intravascular or extravascular. In impedance (the reciprocal of TFC) was consistent with a
patients undergoing thoracentesis, Petersen et al97 demon- decrease in extracellular flui volume. In addition, ICG
strated a strong correlation between the volume of pleural has been used to explore the mechanisms of responses to
flui removed and the change in total thoracic impedance ACE inhibitors and prostaglandin inhibitors in patients
(correlation coefficient 0.97). In studies using lower-body who are either salt sensitive or salt insensitive.102 These
negative pressure to create pooling of venous blood in the examples illustrate the use of ICG in assessing the mech-
lower extremities, Ebert et al98 found a nearly perfect anisms of BP elevation and the hemodynamic effects of
linear correlation with changes in central venous pressure nonpharmacologic interventions in hypertension.
and changes in thoracic impedance. Thus, TFC has been As noted above, antihypertensive medications ulti-
used to monitor changes in flui volume and guide diuretic mately act on one or more of the hemodynamic compo-
therapy in patients with hypertension. nents that determine BP.79,80 Once the baseline
Linb et al89 reported that BP reductions resulted from hemodynamic status is known, an appropriate medical
improvements in baseline hemodynamic abnormalities; regimen can be designed based on the expected hemody-
patients with elevated CO responded to targeted therapy namic effects of various medications. However, individual
with a -blocker (propranolol), whereas those with ele- patients vary in their responses to antihypertensive drugs
vated SVR responded to treatment with the vasodilating such that the actual hemodynamic effects and side effects
CCB (nifedipine). Mattar et al99 showed that an intensive cannot reliably be predicted. Therefore, empiric selection
regimen of diet and exercise resulted in improvements in of drug combinations based on their general hemodynamic
hemodynamic parameters with substantial increases in CO actions as a class may not be successful in managing a
and decreases in SVR despite only modest changes in specifi patient, even if the baseline hemodynamic status is
MAP. Moreover, the investigators speculated that failure known. The ICG technique is unique in that it can provide
of some hypertensive patients to show hemodynamic im- not only an accurate hemodynamic profil noninvasively
provement on serial ICG measures resulted from the in- but can guide therapy toward a drug regimen that is most
appropriate choice of medications that were not targeted appropriate for the specifi patient based on serial mea-
toward the underlying hemodynamic abnormalities. surements. Periodic measurements of hemodynamic status
Hemodynamic parameters derived from ICG have been allow the physician to monitor therapy when results are
used to evaluate the differential effects of medications in suboptimal or unexpected. It is for these reasons that ICG
patients with essential hypertension. In a study of the has emerged as a valuable tool in the evaluation and
effects of a cardioselective -blocker compared with a treatment of patients with hypertension.
-blocker with intrinsic sympathomimetic activity, Toth et
al100 studied 57 patients randomized to treatment with Outcome Studies Using ICG-
either atenolol or pindolol for 12 weeks. Pindolol therapy
was associated with a 12% decrease in SVR compared
Guided Therapy in Hypertension
with minimal change with atenolol. Atenolol-related im- The observation that hypertension is a hemodynamic dis-
provement in BP resulted from decrease in HR and cardiac ease implies that measurement of hemodynamic parame-
index. Breithaupt-Grogler et al67 reported on the differen- ters can be used to guide medication selection, to titrate
tial hemodynamic effects of combination therapy with dose, and to evaluate efficac of the medical regimen.
verapamil/trandolapril (Vera/Tran) compared with meto- Several studies have used ICG to evaluate hemodynamic
prolol/hydrochlorothiazide (Meto/HCTZ) in 26 patients parameters and demonstrated that ICG-guided therapy im-
after 6 months of therapy. In addition to ICG-derived CO proves BP control. Taler et al103 randomized 104 patients
and SVR, the authors measured carotidofemoral pulse with hypertension uncontrolled on two or more drugs to a
wave velocity as a measure of arterial stiffness. The com- 3-month trial of ICG-guided therapy or standard therapy
bination of CCB and ACE inhibitor (Vera/Tran) reduced directed by a hypertension specialist. In this study, BP
diastolic BP to a greater degree than Meto/HCTZ and control (define as achieving BP 140/90 mm Hg) oc-
lowered SVR by about 15% compared with minimal curred 70% more often in the ICG-guided group (Fig. 11).
change with the -blocker/diuretic combination. Treat- Use of ICG resulted in greater reductions in SVR index
ment with Meto/HCTZ was associated with a significan and more intensive use of diuretic therapy, guided by
reduction in CO compared with baseline, which was not levels of TFC. According to the study investigators, mea-
seen with Vera/Tran. However, pulse wave velocity de- surement of hemodynamic and impedance parameters was
creased with Vera/Tran but not with Meto/HCTZ, suggest- more effective than clinical judgment alone in guiding
ing an improvement in the elastic properties of the aorta selection of antihypertensive therapy patients resistant to
with the former drug regimen. empiric therapy.
The ICG technique has been used to assess the hemo- Sharman et al104 studied a cohort of patients in the
dynamic effects of sodium restriction in a small group of primary care offic setting with drug-resistant hyperten-
subjects with mild hypertension.101 During sodium restric- sion, define as systolic BP 140 mm Hg or diastolic BP
tion, ICG-derived measures of SV decreased in association 90 mm Hg during treatment with two antihypertensive
110 of 158.
AJHFebruary 2005VOL. 18, NO. 2, Part 2 HYPERTENSION AS A HEMODYNAMIC DISEASE 39S
111 of 158.
40S HYPERTENSION AS A HEMODYNAMIC DISEASE AJHFebruary 2005VOL. 18, NO. 2, Part 2
systolic time ratio, may be the initial signs of the devel- 4. Lund-Johansen P: Newer thinking on the hemodynamics of hyper-
opment of LV dysfunction.106,107 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
Implications of Hemodynamics Rep 2000;6:565569.
7. Messerli FH, DeCarvalho JG, Christie B, Frohlich ED: Essential
and Future Considerations hypertension in black and white subjects. Hemodynamic finding
In addition to improving the diagnosis and therapy of and flui volume state. Am J Med 1979;67:2731.
8. Ventura H, Messerli FH, Oigman W, Suarez DH, Dreslinski GR,
hypertension, hemodynamic measurements provide in- Dunn FG, Reisin E, Frohlich ED: Impaired systemic arterial com-
sights into other aspects of cardiovascular function. For pliance in borderline hypertension. Am Heart J 1984;108:132136.
example, studies have shown the importance of endothe- 9. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA,
lial function in the development and progression of car- Izzo JL, Jones DW, Materson BJ, Oparil S, Wright JT, Roccella EJ:
diovascular disease.108 Endothelial dysfunction, as The Seventh Report of the Joint National Committee on Prevention,
Detection, Evaluation, and Treatment of High Blood Pressure. The
measured by reduced flow-mediate arterial dilation, is
JNC 7 Report. J Am Med Assoc 2003;289:2560 2572.
associated with abnormal hemodynamic measures, includ- 10. Oparil S, Zaman MA, Calhoun DA: Pathogenesis of hypertension.
ing elevated SVR.109 In the HOPE study,110 an ACE Ann Intern Med 2003;139:761776.
inhibitora drug that both lowers SVR and improves 11. Schappert SM, Nelson C: National ambulatory medical care survey:
endothelial functionreduced mortality from cardiovas- 199596 summary. Vital and health statistics, series 13, no. 142.
Washington, DC: Government Printing Office November 1999.
cular disease despite only minor effects on BP. Future
DHHS publication no. (PHS) 2000-1713.
studies will likely examine the significanc of elevated 12. Hyman DJ, Pavlik VN: Characteristics of patients with uncontrolled
SVR and arterial compliance in individuals with hyperten- hypertension in the United States. N Engl J Med 2001;345:479
sion or prehypertension and will correlate ICG-derived 486.
hemodynamic parameters with other evolving markers of 13. Setaro JF, Black HR: Refractory hypertension. N Engl J Med
1992;327:543547.
increased cardiovascular risk such as C-reactive protein,
14. Oparil S, Calhoun DA: Managing the patient with hard-to-control
homocysteine level, and the metabolic syndrome. hypertension. Am Fam Phys 1998;57:10071014.
The studies included in this supplement of the journal add 15. Gifford RW Jr. Resistant hypertension: introduction and definitions
to the growing body of literature that supports the accuracy, Hypertension 1988;11(Suppl II):II-65II-66.
reliability, and clinical utility of ICG in diagnostic and prog- 16. Wilking SV, Belanger A, Kannel WB, DAgostino RB, Steel K:
nostic assessment and therapeutic management of patients Determinants of isolated systolic hypertension. J Am Med Assoc
1988;260:34513455.
with hypertension. The use of ICG has added significantl to 17. Lewington S, Clarke R, Qizilbash N, Peto R, Collins R: Age-
our understanding of hypertension as a disease with both specifi relevance of usual blood pressure to vascular mortality: a
hemodynamic causes and hemodynamic consequences. Just meta-analysis of individual data for one million adults in 61 pro-
as congestive heart failure reflect abnormal flo or inappro- spective studies. Lancet 2002;360:19031913.
priate ventricular fillin pressures, hypertension occurs when 18. Elliott WJ: The economic impact of hypertension. J Clin Hypertens
2003;5(3 Suppl 2):313.
there is abnormal flo or inappropriate vascular resistance or 19. Goetzel RZ, Long SR, Ozminkowski RJ, Hawkins K, Wang S,
compliance. When hypertension impairs LV performance Lynch W: Health, absence, disability, and presenteeism cost esti-
(either systolic or diastolic), heart failure ensues. Although mates of certain physical and mental health conditions affecting U.S.
these conditions often co-exist, in many cases hypertension is employers. J Occup Environ Med 2004;46:398 412.
a step in a hemodynamic continuum that leads to further 20. Steingrub JS, Celoria G, Vickers-Lahti M, Teres D, Bria W: Ther-
apeutic impact of pulmonary artery catheterization in a medical/
hemodynamic derangement and heart failure. It is believed surgical ICU. Chest 1991;99:14511455.
that future studies will confir recent finding that hemody- 21. Connors AF, Dawson NV, Shaw PK, Montenegro HD, Nara AR,
namic measurements in individual patients will improve di- Martin L: Hemodynamic status in critically ill patients with and
agnosis, risk assessment, and treatment for these patients. It is without acute heart disease. Chest 1990;98:1200 1206.
also possible that further exploration of the implications of 22. Hoit BD, Rashwan M, Watt C, Sahn DJ, Bhargava V: Calculating
cardiac output from transmitral volume flo using Doppler and
hypertension as a hemodynamic disease will lead to studies
M-mode echocardiography. Am J Cardiol 1988;62:131135.
demonstrating that earlier detection and treatment of the 23. Northridge DB, Findlay IN, Wilson J, Henderson E, Dargie HJ:
hemodynamic components of hypertension may change the Non-invasive determination of cardiac output by Doppler echocar-
natural history of this disease process. 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
References thermodilution in measuring cardiac output and index in patients
with advanced, decompensated chronic heart failure hospitalized in
1. Wiggers CJ: Hemodynamic changes in hypertension. Am Heart J critical care. Am J Crit Care 2004;13:469 479.
1938;16:515525. 26. Drazner M, Thompson B, Rosenberg P, Kaiser PA, Boehrer JD,
2. Freis ED: Hemodynamics of hypertension. Physiol Rev 1960;40: Baldwin BJ, Dries DL, Yancy CW: Comparison of impedance
2754. cardiography with invasive hemodynamic measurements in patients
3. Freis ED: Studies in hemodynamics and hypertension. Hypertension with heart failure secondary to ischemic or nonischemic cardiomy-
2001;38:15. opathy. Am J Cardiol 2002;89:993995.
112 of 158.
AJHFebruary 2005VOL. 18, NO. 2, Part 2 HYPERTENSION AS A HEMODYNAMIC DISEASE 41S
27. Ziegler D, Grotti L, Krucke G: Comparison of cardiac output Stegall M, Textor S: Changes in hemodynamic patterns with age
measurements by TEB vs. intermittent bolus thermodilution in in normotensive subjects (Abstract). Am J Hypertens 2004;17:
mechanical ventilated patients (Abstract). Chest 1999;116:281S. 373.
28. Sageman WS, Riffenburgh RH, Spiess BD: Equivalence of bioim- 48. Hinderliter AL, Sherwood A, Blumenthal JA, Light KC, Girdler SS,
pedance and thermodilution in measuring cardiac index after cardiac McFetridge J, Johnson K, Waugh R: Changes in hemodynamics and
surgery. J Cardiothoracic Vasc Anesth 2002;16:8 14. left ventricular structure after menopause. Am J Cardiol 2002;89:
29. Yung, G, Fedullo P, Kinninger K, Johnson W, Channick R: Com- 830 833.
parison of impedance cardiography to direct Fick and thermodilu- 49. Galarza CR, Alfi J, Waisman GD, Mayorga LM, Camera LA, del
tion cardiac output determination in pulmonary arterial Rio M, Vasvari F, Limansky R, Farias J, Tessler J, Camera MI:
hypertension. Congest Heart Fail 2004;10(Suppl 2):710. Diastolic pressure underestimates age-related hemodynamic impair-
30. Verhoeve PE, Cadwell CA, Tsadok S: Reproducibility of noninva- ment. Hypertension 1997;30:809 816.
sive bioimpedance measurements of cardiac function (Abstract). 50. Alfi J, Gabriel D, Waisman G, Camera MI: Contribution of stroke
J Card Fail 1998;4(Suppl):S3. volume to the change in pulse pressure pattern with age. Hyperten-
31. Aldrete AJ, Brown C, Daily J, Buerke V: Pacemaker malfunction
sion 1999;34:808 812.
due to microcurrent injection from a bioimpedance noninvasive
51. Hartley TR, Lovallo WR,Whitsett TL: Cardiovascular effects
cardiac output monitor. J Clin Monit 1995;11:131133.
of caffeine in men and women. Am J Cardiol 2004;93:
32. Frohlich ED: Local hemodynamic changes in hypertension. Insights
10221026.
for therapeutic preservation of target organs. Hypertension 2001;38:
52. Yu BH, Nelesen R, Ziegler MG, Dimsdale JE: Mood states and
1388 1394.
impedance cardiography-derived hemodynamics. Ann Behav Med
33. Segers P, Brimioulle S, Stergiopulos N, Westerhof N, Naeije R,
2001;23:2125.
Maggiorini M, Verdonck P: Pulmonary arterial compliance in dogs
and pigs: the three-element Windkessel model revisited. Am J 53. Hinderliter A, Blumenthal J, Waugh R, Chilukuri M, Sherwood A:
Physiol 199;277:H725H731. Ethnic differences in left ventricular structure: relations to hemody-
34. Randall OS, Westerhof N, van den Bos GC, Alexander B: Reliabil- namics and diurnal blood pressure variation. Am J Hypertens 2004;
ity of stroke volume to pulse pressure ratio for estimating and 17:43 49.
detecting changes in arterial compliance. J Hypertens 1986;4(Suppl 54. Calhoun DA, Mutinga ML, Collins AS, Wyss JM, Oparil S: Nor-
5):S293S296. motensive blacks have heightened sympathetic response to cold
35. Chemla D, Hebert J-L, Coirault C, Zamani K, Suard I, Colin P, pressor test. Hypertension 1993;22:801 805.
Lecarpentier Y: Total arterial compliance estimated by stroke vol- 55. Fagard RH, Pardaens K, Staessen JA, Thijs L: Prognostic value of
ume-to-aortic pulse pressure ratio in humans. Am J Physiol 1998; invasive hemodynamic measurements at rest and during exercise in
274: (Heart Circ Physiol 43) H500 H505. hypertensive men. Hypertension 1996;28:3136.
36. Egan B, Schmouder R: The importance of hemodynamic consider- 56. Fagard RH, Pardaens K, Staessen JA, Thijs L: The pulse pres-
ations in essential hypertension. Am Heart J 1988;116:594 599. sure-to-stroke index ratio predicts cardiovascular events and
37. Izzo J: Arterial stiffness and the systolic hypertension syndrome. death in uncomplicated hypertension. J Am Coll Cardiol 2001;
Curr Opin Cardiol 2004;19:341352. 38:22731.
38. Guyton AC: The relationship of cardiac output and arterial pressure 57. de Simone G, Roman MJ, Koren MJ, Mensah GA, Ganau A,
control. Circulation 1981;64:1079 1088. Devereux RB: Stroke volume/pulse pressure ratio and cardiovas-
39. Fouad-Tarazi F, Izzo JL: Hemodynamic profile and responses, in cular risk in arterial hypertension. Hypertension 1999;33:800
Izzo JL, Black HR (eds): Hypertension Primer, : The Essentials of 805.
High Blood Pressure. Council on High Blood Pressure Research 58. Janeway TC: A clinical study of hypertensive cardiovascular dis-
(American Heart Association), Dallas, 1999, pp 115117. Second ease. Arch Intern Med 1913;12:755798.
Edition 59. Ganau A, Devereux RB, Roman MJ, de Simone G, Pickering TG,
40. Julius S, Krause L, Schork NJ, Mejia AD, Jones KA, van de Ven C, Saba PS, Vargiu P, Simongini I, Laragh JH: Patterns of left ven-
Johnson EH, Sekkarie MA, Kjeldsen SE, Petrin J: Hyperkinetic tricular hypertrophy and geometric remodeling in essential hyper-
borderline hypertension in Tecumseh, Michigan. J Hypertens 1991; tension. J Am Coll Cardiol 1992;19:1550 1558.
9:77 84. 60. Kahan T: The importance of left ventricular hypertrophy in human
41. Lund-Johansen P: Hemodynamic patterns in the natural history of
hypertension. J Hypertens 1998;16(Suppl):S23S29.
borderline hypertension. J Cardiovasc Pharm 1986;8(Suppl 5):S8
61. Verdecchia P, Schillaci G, Borgioni C, Ciucci A, Gattobigio R,
S14.
Zampi I, Santucci A, Santucci C, Reboldi G, Porcellati C: Prognos-
42. Slotwiner DJ, Devereux RB, Schwartz JE, Pickering TG, de Simone
tic value of left ventricular mass and geometry in systemic hyper-
G, Roman MJ: Relation of age to left ventricular function and
tension with left ventricular hypertrophy. Am J Cardiol 1996;78:
systemic hemodynamics in uncomplicated mild hypertension. Hy-
197202.
pertension 2001;37:1404 1409.
62. Verdecchia P, Carini G, Circo A, Dovellini E, Giovannini E, Lom-
43. Franklin SS: Blood pressure and cardiovascular disease: what re-
mains to be achieved? J Hypertens 2001;19(Suppl 3):S3S8. bardo M, Solinas P, Gorini M, Maggioni AP: Left ventricular mass
44. Messerli FH, Garavaglia GE, Schmieder RE, Sundgaard-Riise K, and cardiovascular morbidity in essential hypertension: the MAVI
Nunez BD, Amodeo C: Disparate cardiovascular finding in men study. J Am Coll Cardiol 2001;38:1829 1835.
and women with essential hypertension. Ann Intern Med 1987;107: 63. Vakili BA, Okin PM, Devereux RB: Prognostic implications of left
158 161. ventricular hypertrophy. Am Heart J 2001;141:334 341.
45. Davidson RC, Ahmad S: Hemodynamic profile in essential and 64. Ichkhan K, Molnar J, Somberg J: Relation of left ventricular mass
secondary hypertension, in Izzo JL, Black HR (eds): Hypertension and QT dispersion in patients with systematic hypertension. Am J
Primer: The Essentials of High Blood Pressure. Council on High Cardiol 1997;79:508 511.
Blood Pressure Research (American Heart Association), Dallas, 65. Gomez-Cerezo J, Rios Blanco JJ, Suarez Garcia I, Moreno Anaya P,
2003, pp 349 351. Third Edition Garcia Raya P, Vazquez-Munoz E, Barbado Hernandez FJ: Nonin-
46. Romney JS, Lewanczuk RZ: Vascular compliance is reduced in vasive study of endothelial function in white coat hypertension.
the early stages of type 1 diabetes. Diabetes Care 2001;24:2102 Hypertension 2002;40:304 309.
2106. 66. London GM, Cohn JN: Prognostic application of arterial stiffness:
47. Taler S, Driscoll N, Tibor M, Sprau G, Augustine J, Larson T, task forces. Am J Hypertens 2002;15:754 758.
113 of 158.
42S HYPERTENSION AS A HEMODYNAMIC DISEASE AJHFebruary 2005VOL. 18, NO. 2, Part 2
67. Breithaupt-Grogler K, Gerhardt G, Lehmann G, Notter T, Belz GG: 87. Bailey JM, Levy JH, Kopel MA: Relationship between clinical
Blood pressure and aortic elastic propertiesverapamil SR/tran- evaluation of peripheral perfusion and global hemodynamics in
dolapril compared to a metoprolol/hydrochlorothiazide combination adults after cardiac surgery. Crit Care Med 1990;18:1353
therapy. Int J Clin Pharmacol Ther 1998;36:425 431. 1356.
68. Glasser SP: Hypertension syndrome and cardiovascular events. 88. Tibby SM, Hatherill M, Marsh MJ, Murdoch IA: Clinicians abili-
Postgrad Med 2001;110:29 35. ties to estimate cardiac index in ventilated children and infants. Arch
69. Intengan HD, Schiffrin EL: Vascular remodeling in hypertension: Dis Child 1997;77:516 518.
roles of apoptosis, inflammation and fibrosis Hypertension 2001; 89. Linb G, Eisenberg BM: Noninvasive techniques for evaluation of
38:581587. heart function and hemodynamics in arterial hypertension. Acta
70. Mathew J, Sleight P, Lonn E, Johnstone D, Pogue J, Yi Q, Bosch J, Cardiol 1990;45:133139.
Sussex B, Probstfiel J, Yusuf S: Reduction of cardiovascular risk 90. Abdelhamed AI, Levy P, Smith KS, Wilkins J, Smith R, Ferrario
by regression of electrocardiographic markers of left ventricular CM: Value of non-invasive hemodynamic measurements in hyper-
hypertrophy by the angiotensin-converting enzyme inhibitor tensive patients. Am J Hypertens 2002;15(Suppl 1):A89.
ramipril. Circulation 2001;104:16151621. 91. de Divitiis O, Di Somma S, Liguori V, Petitto M, Magnotta C,
71. Corea L, Bentivoglio M, Verdecchia P, Providenza M, Motolese M: Ausiello M, Natale N, Brignoli M, Galderisi M: Effort blood pres-
Left ventricular hypertrophy regression in hypertensive patients sure control in the course of antihypertensive treatment. Am J Med
treated with metoprolol. Int J Pharmacol Ther Toxicol 1984;22:365 1989;87:(Suppl 3C)46S56S.
370. 92. Eliot RS: The dynamics of hypertensionan overview: present
72. Yoshitomi Y, Nishikimi T, Abe H, Nagata S, Kuramochi M, Mat- practices, new possibilities, and new approaches. Am Heart J 1988;
suoka H, Omae T: Left ventricular systolic and diastolic function 116:583589.
and mass before and after antihypertensive treatment in patients 93. Taler SJ, Augustine J, Burnett JC, Textor SC: Hemodynamic and
with essential hypertension. Hypertens Res 1997;20:2328. volume changes during intensive treatment (Rx) for resistant hyper-
73. Ofil EO, Cohen JD, St Vrain JA, Pearson A, Martin TJ, Uy ND, tension (ResHTN) (Abstract). Am J Hypertens 2000;13:61A 62A.
Castello R, Labovitz AJ: Effect of treatment of isolated systolic 94. Lang M: Noninvasive hemodynamic-driven care management.
hypertension on left ventricular mass. J Am Med Assoc 1998;279: Achieving quality patient outcomes across a continuum of care.
778. Cardiovasc Dis Manage 2001;7173.
74. Verdecchia P, Angeli F, Borgioni C, Gattobigio R, de Simone G, 95. Vuurmans T, Boper P, Koomans H: Effects of endothelin-1 and
Devereux RB, Porcellati C: Changes in cardiovascular risk by
endothelin-1 receptor blockade on cardiac output, aortic pressure,
reduction of left ventricular mass in hypertension: a meta-analysis.
and pule wave velocity in humans. Hypertension 2003;41:1253
Am J Hypertens 2003;16:895 899.
1258.
75. Galarza C, Alfi J, Waisman G, Burgos M, Zaniello G, del Rio M,
96. Margulis F, Dalton R, Killinger C, Pisarello J, Alvarez C: Decreased
Serra L, Toselli P, Vasvari F, Camera M: Severe systemic hemo-
myocardial contractility and increased thoracic flui content de-
dynamic impairment in patients with stroke (Abstract). Am J Hy-
tected by impedance cardiography in untreated hypertensive patients
pertens 1996;9(Suppl 1):172A.
(Abstract). Am J Hypertens 1997;10(Suppl 1):55A.
76. Johns DW, Peach MJ: Factors that contribute to resistant forms of
97. Petersen JR, Jensen BV, Drabaek H, Viskum K, Mehlsen J: Elec-
hypertension. Pharmacological considerations. Hypertension 1988;
trical impedance measured changes in thoracic flui content during
11(Suppl II):II-88 II-95.
thoracentesis. Clin Physiol 1994;14:459 466.
77. Tsukiyama H: Choice of antihypertensive agents in hemodynamic
98. Ebert TJ, Smith JJ, Barney JA, Merrill DC, Smith GK: The use of
aspects to match pathophysiology and pharmacology in essential
thoracic impedance for determining thoracic blood volume changes
hypertension. Jpn J Med 1989;28:261264.
in man. Aviat Space Environ Med 1986;57:49 53.
78. Zusman RM: Left ventricular function in hypertension. Relevance
to the selection of antihypertensive therapy. Am J Hypertens 1989; 99. Mattar JA, Salas CE, Bernstein DP, Lehr D, Bauer R: Hemody-
2:200S206S. namic changes after an intensive short-term exercise and nutrition
79. Sullivan JM, Schoeneberger TE, Ratts ET, Palmer ET, Samaha JK, program in hypertensive and obese patients with and without cor-
Mance CJ, Muirhead EE: Short-term therapy of severe hyperten- onary disease. Arq Bras Cardiol 1990;54:307312.
sion. Hemodynamic correlates of the antihypertensive response in 100. Toth PD, Demeter RJ, Woods, JR, Nyhuis AW, Judy WV:
man. Arch Intern Med 1979;139:12331239. Comparison of the effects of pindolol and atenolol on hemody-
80. Easterling TR, Benedetti TJ, Schmucker BC, Carlson KL: Antihy- namic function in systemic hypertension. Am J Cardiol 1988;62:
pertensive therapy in pregnancy directed by noninvasive hemody- 413 418.
namic monitoring. Am J Perinatol 1989;6:86 89. 101. Koga Y, Gillum RF, Kubicek WG: An impedance cardiographic
81. Resnick LM, Lester MH: Differential effects of antihypertensive study of the mechanism of blood pressure fall after moderate dietary
drug therapy on arterial compliance. Am J Hypertens 2002;15: sodium restriction. Jpn Heart J 1985;26:197207.
1096 1100. 102. Ashida T, Nishioeda Y, Kimura G, Kojima S, Kawamura M,
82. Weber K, Bohmeke T, van der Does R, Taylor SH: Hemodynamic Imanishi M, Abe H, Kawano Y, Yoshimi H, Yoshida K, Kuramo-
differences between metoprolol and carvedilol in hypertensive pa- chi M, Omae T: Effects of salt, prostaglandin, and captopril on
tients. Am J Hypertens 1998;11:614 617. vascular responsiveness in essential hypertension. Am J Hyper-
83. ALLHAT Coordinators and Officers Major outcomes in high-risk tens 1989;2:640 642.
hypertensive patients randomized to angiotensin-converting enzyme 103. Taler SJ, Textor SC, Augustine JE: Resistant hypertension: com-
inhibitor or calcium channel blocker vs. diuretic. The Antihyperten- paring hemodynamic management to specialist care. Hypertension
sive and Lipid-Lowering Treatment to Prevent Heart Attack Trial 2002;39:982988.
(ALLHAT). J Am Med Assoc 2002;288:29812997. 104. Sharman DL, Gomes CP, Rutherford JP: Improvement in blood
84. Yakovlevitch M, Black HR: Resistant hypertension in a tertiary care pressure control with impedance cardiograph-guided pharmacologic
clinic. Arch Intern Med 1991;151:1786 1792. decision making. Congest Heart Fail 2004;10:54 58.
85. Houston MC: Hypertension strategies for therapeutic interven- 105. Sramek BB, Tichy JA, Hojerova M, Cervenka V: Normohemody-
tion and prevention of end-organ damage. Prim Care 1991;18: namic goal-oriented antihypertensive therapy improves the out-
713753. come. Am J Hypertens 1996;9:141A.
86. Taler SJ, Augustine J, Textor SC: A hemodynamic approach to 106. Parrott C, Burnham K, Quale C, Lewis D: Comparison of changes
resistant hypertension. Congest Heart Fail 2000;6:90 93. in ejection fraction to changes in impedance cardiography cardiac
114 of 158.
AJHFebruary 2005VOL. 18, NO. 2, Part 2 HYPERTENSION AS A HEMODYNAMIC DISEASE 43S
index and systolic time ratio. Congest Heart Fail 2004;10(2 Suppl 109. Panza JA: High-normal blood pressuremore high than nor-
2):1113. mal.N Engl J Med 2001;345:13371340.
107. Thomson B, Drazner M, Dries D, Yancy C: Is impedance cardiog- 110. Dagenais GR, Yusuf S, Bourassa MG, Yi Q, Bosch J, Lonn EM,
raphy-derived systolic time ratio a useful method to determine left Kouz S, Grover J: Effects of ramipril on coronary events in high-risk
ventricular systolic dysfunction in heart failure (Abstract)? J Card persons: results of the Heart Outcomes Prevention Evaluation
Failure 2004;10(Suppl 4):S38. Study. Circulation 2001;104:522526.
108. Ross R: Atherosclerosisan inflammator disease. N Engl J Med 111. Herrera-Acosta J, Perez-Grovas H, Fernandez M, Arriaga J: Enala-
1999;340:115126. pril in essential hypertension. Drugs 1985;30(Suppl 1):35 46.
115 of 158.
European Journal of Heart Failure, June 2004
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
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
Figure 1. Recordings of the interrelation between impedance variations and hemodynamic parame-
ters, according to Djordjevich and Sadove.16
are significantly higher than the NICO CI; when CI that the difference between stage 3 and stage 4 was
is 3 L/min/ m2, the thermodilution CI is lower statistically significant according to the two measure-
than the NICO CI (Table 3). When CI was between ments in subgroup A (NICO CI, p 0.0135; ther-
2 L/min/m2 and 3 L/min/m2, there was a slight modilution CI, p 0.002; Fig 5, top). In 17 of the
difference of only 3.28% between the two methods, CHF patients in group 3 who were treated with
with a borderline significance (Table 3). an IV vasodilator agent, and in whom the CI was
Twenty-five patients with CABG (subgroup A) measured simultaneously at seven time points during
had complete information of the NICO CI and the treatment, the total time trend was significant in
thermodilution CI results at five operative and post- the NICO (p 0.0056) but not significant in ther-
operative stages (Table 4, Fig 5, top). A further 26 modilution (Fig 5, bottom).
CABG patients (subgroup B) had incomplete data
( 5 paired measurements for each patient), yielding
additional 80 paired measurements (Table 4). There Discussion
were time-related changes in the CI, and the NICO
and the thermodilution followed these changes In recent years it has been suggested that CO and
by producing similar results at each time point MAP measurement and the calculation of CI, Cpi,
(p 0.0035 and p 0.0058, respectively; Fig 5, and SVRi might be instrumental in the diagnosis,
top). A contrast analysis, performed to compare the treatment titration, and risk stratification of cardiac
CO in successive stages of the measurements, found patients, especially those with CHF.710 However,
CI has been measured only during invasive right- ing by the literature,24 26 as long as the heart func-
heart catheterization, which requires intensive care tion is intact, TEB can be useful for monitoring the
admission and may be associated with complica- hemodynamic state in various clinical conditions
tions.2123 Hence, CI was measured only rarely, and such as trauma, massive surgery, sepsis, etc.24 But
in the sickest patients. Therefore, a simple, reliable, when it comes to monitoring and managing patho-
noninvasive, and continuous method for CI mea- logic cardiac conditions TEB requires further im-
surement has become necessary in order to enable provement.24 26
its application to cardiac patients with different Thus far, eight groups of patients who submitted
degrees of medical severity and in diverse settings. to CO measurements by WBEB have been reported
Currently there are four accepted methods for in six published articles. Kedrov,17 who was the first,
noninvasive CO measurement. The Doppler echo- compared the average CI measured by the WBEB
cardiogram obtained from the left ventricular out- in 57 subjects with normal hearts in published results
flow track and CO2 rebreathing techniques have of the Fick method, revealing 3.3 28% vs 3.31
been shown to be accurate in measuring CI. But L/min/m2 (range, 2.4 to 4.2 L/min/m2), respectively.
these methods are limited by the requirement for Tischenko27 compared the CI results measured by
expensive equipment and specialized personnel for WBEB in three groups of subjects with normal
their application and therefore are not simple to use, hearts vs three standard methods. There were 31
and moreover do not enable continuous measure- cases vs acetylen (r 0.84), 28 cases vs thermodilu-
ments. Thoracic bioimpedance has been used in the tion (r 0.95), and 28 cases vs Fick (r 0.99).
last decade for continuous CO measurement. Judg- Using a modified Tischenko algorithm vs thermodi-
Table 3Differences Between the NICO CI and Thermodilution CI Within the CI Ranges
lution, Koobi et al28 obtained simultaneous measure- well as in each diagnostic group, together with the
ments in 74 patients with coronary disease, reaching relatively large sample size, further endorses the
a bias between the two methods of 0.25 0.8 L/min significance of the results.
(SD), where the limits of agreement (2 SD) were The mean difference between the NICO and
1.37 L/min and 1.897 L/min, respectively. Using thermodilution in the entire sample range was
the NICaS apparatus, Cohen et al14 compared its 0.0009 L/min (Table 2, Fig 4), ranging from 0.0040
performance against thermodilution by measuring to 0.0271 L/min in the three diagnostic groups. This
the CO in patients undergoing CABG operations, disparity is smaller than the level of accuracy of
with a correlation of r 0.91. Thus, this six-clinical thermodilution, which is defined to a 15% range.31
series, which included 274 subjects, revealed similar Linear regression applied to the data revealed that
correlation coefficients between the compared the line slope was close to 1.00 in the entire sample
methods, just as in the present series. Moreover, in range and in each specific diagnostic group. There
none of these publications did the authors express were no significant differences between the slopes
reservations about the performance of the WBEB. and the intercepts of the three diagnostic groups.
There were, nonetheless, two publications in This indicates that the relation between the NICO
which no correlation was found between WBEB and and thermodilution is similar in all diagnostic groups.
thermodilution; in both instances, the underlying Following the recommendations of Bland and Alt-
clinical conditions are listed in the exclusion criteria man,20 the differences between the two measure-
of the NICO. Lamberts et al29 compared the original ments were plotted against their means. This plot
Tischenko equation with dye dilution CO in 10 demonstrates that the range of differences is similar
patients, 4 of whom had significant aortic regurgita- along the different values of the average.
tion and 1 had coarctation of the aorta. The NICO Although the main purpose of this work was to
apparatus cannot measure the CO in such condi- compare the performance of the NICO vs thermo-
tions. dilution, following the suggestion in previous stud-
Imhoff et al30 compared the NICO apparatus ies3135 that thermodilution tends to overestimate CI
against thermodilution in 22 postpancreatectomy or when low and underestimate it when high, we have
esophagectomy patients. They were all managed compared the NICO CI and thermodilution CI in
postoperatively by Swan-Ganz catheters for boosting the different CI ranges. The results of this analysis
the oxygen delivery to 600 mL/min/m2 and the CI to have shown that when the CI was 2 L/min/m2, the
4.5 L/min/m2. Hence, in these patients the radical thermodilution results were higher than the NICO
surgical procedures were followed by massive inter- results; when the CI was 3 L/min/m2, the ther-
compartmental volume shifts due to IV administra- modilution results were slightly lower than the
tion of up to 6 L per 24 h of crystalloids and plasma, NICO results (Table 3). As a consequence, the
often accompanied by massive peripheral edema. In differences of the hemodynamic responses to vaso-
such hemodynamic situations, the baseline imped- dilation therapy may be better depicted by the
ance should properly become distorted, preventing NICO when compared with thermodilution (Fig 5,
an accurate measurement of the systolic impedance bottom).
variation. The link between a low CI and a higher thermo-
In the present study, the agreement between dilution readout, and between a higher CI with a
NICO CI and thermodilution CI as tested by com- low thermodilution readout, is also expressed in the
parisons of the means is highly significant. The stability of the range of differences along the differ-
similarity between the means in the entire cohort as ent values of the average (Fig 4). Furthermore, the
Received 16 August 2002; received in revised form 6 January 2003; accepted 21 January 2003
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.
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
130 of 158.
G. Cotter et al. / The European Journal of Heart Failure 5 (2003) 443451 445
2.4. Assessment of hemodynamic variables best cut-off point for the various parameters, in terms
of highest sensitivity and specificity.
Prior to enrolment in this study all patients gave
written informed consent. The study protocol was 2.8. CpiySVRi diagnostic graph
approved by the local ethics review board. In all patients
the hemodynamic variables were obtained during right A classification rule was developed using second
heart catheterization using a SwanGanz catheter placed order discriminant analysis. The normality of the distri-
under fluoroscopic guidance. All measurements were bution of Cpi and SVRi was examined by the Wilk
obtained while patients were at least 30 s without IABP Shapiro test. Due to the skewness of the data in some
while on the same treatment used at the time the clinical groups, both variables (CPi and SVRi) were transformed
diagnosis was made. Measurement of hemodynamic into log scale for better approximation to normality.
variables was performed at least 6 h after the last intake Since the number of patients with HT crisis was small
of an oral drug and 2 h after intravenous drug therapy. they were considered together with the exacerbated
CI was measured by thermodilution, using the mean systolic CHF group. The classification used two steps.
of at least three consecutive measurements within a In the first step the rule separated three classes: septic
range of -15%. In normal subjects, right heart cathe- shock, cardiogenic shock and a combined group, which
terization was not performed for ethical reasons. The included the normal controls, compensated CHF and
values used in this cohort were obtained by standard pulmonary edema patients (NCP). If, after the first
noninvasive cuff blood pressure measurement and eval- step the patient was defined as NCP, the second
uation of CI by the FDA approved NICaS 2001, a classification was used to differentiate between the
noninvasive continuous cardiac output monitor w10x. normal, exacerbated systolic CHF and pulmonary edema
131 of 158.
446 G. Cotter et al. / The European Journal of Heart Failure 5 (2003) 443451
Table 1
Baseline characteristics of the five patient groups
Table 2
Baseline distribution of the various hemodynamic parameters in the six diagnosis groups presented as means and S.D.
132 of 158.
G. Cotter et al. / The European Journal of Heart Failure 5 (2003) 443451 447
Fig. 2. Pulmonary capillary wedge pressure (mmHg) Box-plots Fig. 4. Systemic vascular resistance index (SVRi) (wood m2) Box-
(median and 2575% percentile range) in patients with the different plots (median and 2575% percentile range) in patients with the dif-
syndromes of acute CHF and patients with septic shock and normal ferent syndromes of acute CHF and patients with septic shock and
controls. normal controls.
133 of 158.
448 G. Cotter et al. / The European Journal of Heart Failure 5 (2003) 443451
Table 3
Number of observations classified into the correct clinical group using log(Cpi) or log(SVRi) only
(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)
134 of 158.
G. Cotter et al. / The European Journal of Heart Failure 5 (2003) 443451 449
For our data Lambda (RNC)s0.95 (S.D. (Lambda)s of outcome in chronic CHF patients w79x, exacerbated
0.03) which corresponds to three errors of classification systolic CHF w1x and cardiogenic shock (unpublished
according to the classification rule, instead of 59 errors data).
of classification according to the prior probabilities. In the present study, we found that in patients with
exacerbated systolic CHF either Cpi was decreased or
4. Discussion SVRi was increased or both changes occurred. In a
previous study, we described the sequence of events
During the last decade the pathogenesis of CHF has leading to acute heart failure w1x. In most patients an
become clearer. The emphasis on cardiac performance acute CHF event starts with a progressive decrease in
as the sole pathogenic mechanism of CHF has changed cardiac contractility and power (Cpi). Thereafter, as Cpi
to a more comprehensive understanding of the impor- decreases, neurohormonal vascular control mechanisms
tance of the interaction between cardiac contractility, are activated and SVRi increases w1,17x. This increase
neurohormonal and inflammatory control mechanisms is a very important protective mechanism for two
and vascular resistance. We have recently studied the reasons:
treatment of the acute CHF syndromes of pulmonary
1. The increase in SVRi in the face of decreased
edema and cardiogenic shock and have shown that
contractility maintains blood pressure and the perfu-
treatment modalities with significant vascular effect are
sion of vital organs.
effective in improving the outcome of these patients
w1416x. These findings substantiated our theory that 2. This afterload increase (while within certain limits)
may improve contractility (possibly through the
the SVRi reaction to the decrease in Cpi determines the
Gregg phenomenon w18x), which may account for the
hemodynamic condition and clinical syndrome of
normal Cpi we observed in some patients with
135 of 158.
450 G. Cotter et al. / The European Journal of Heart Failure 5 (2003) 443451
response, related to neurohormonal, endothelial and tility reserve in chronic and acute CHF as well as in
perhaps inflammatory activation w20,21x. This remark- cardiogenic shock.
able increase in SVRi induces an afterload mismatch,
reducing CI and increasing intracardiac pressures, Appendix A: Classification rule
LVEDP and wedge pressure, resulting in the severe
congestive symptoms of pulmonary edema. Therefore,
as previously suggested w14,16,22,23x, vasodilator treat- Given a patient with measured values of SVRi and
ment is effective in the treatment of pulmonary edema. Cpi, the classification may be performed either (A)
through special calculations or (B) using the Graph for
4.3. Two-dimensional graphic representation of Cpiy classification of CHF patients (CpiySVRi graph).
SVRi and its use in the treatment of cardiogenic shock (A) Classification using calculations
and pulmonary edema Step 1. Calculate three values v1, v2, v3 according to
the formulas below.
In the present study, we found that when plotting on
a two-dimensional graph the results of Cpi and SVRi v1sLCPi2=21.54q2=LCPi=LSVRi=10.61
for individual patients, each clinical group of patients qLSVRi2=59.44yLCPi=305.24yLSVRi
could be segregated into a specific area on the graph =417.70q1408.89
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, v2sLCPi2=10.12q2=LCPi=LSVRi=5.67yLSVRi2
pulmonary edema, HTN crisis, cardiogenic shock and =4.99yLCPi=135.81yLSVRi=90.11q482.61
136 of 158.
G. Cotter et al. / The European Journal of Heart Failure 5 (2003) 443451 451
(B) Classification using the diagnostic graph w10x Cohen JA, Arnaudov D, Zabeeda D, Schlthes L, Lashinger J,
Put the point (CPi, SVRi) on the diagnostic graph Schachner A. Non-invasive measurement of cardiac output
during coronary artery bypass grafting. Eur J Card Thoracic
Fig. 5 or point (LCPi, LSVRi) and classify the patient Surg 1998;14:64 9.
according to the area of the graph, where the point is w11x Lange RA, Hillis LD. Cardiac catheterization and hemodyn-
located. amic assessment. In: Topol EJ, Textbook of Cardiovascular
Medicine. page 1653.
References w12x Sasieni PD. Statistical analysis of the performance of diagnostic
tests. Cytopathology 1999;10:73 8 (invited review).
w1x Cotter G, Moshkovitz Y, Milovanov O, et al. Acute congestive w13x Lijmer JG, Willen Mol B, Heisterkamp S, et al. Empirical
heart failure: a novel approach to its pathogenesis and treat- evidence of design related bias in studies of diagnostic tests. J
ment. Eur J Heart Fail 2002;4:227 34 (review). Am Med Assoc 1999;282:1061 6.
w2x Menon V, White H, LeJemtel T, Webb JG, Sleeper LA, w14x Cotter G, Metzkor E, Kaluski E, et al. Randomised trial of
Hochman JS. The clinical profile of patients with suspected high-dose isosorbide dinitrate plus low-dose furosemide versus
cardiogenic shock due to predominant left ventricular failure: high-dose furosemide plus low-dose isosorbide dinitrate in
a report from the SHOCK Trial Registry. Should we emergently severe pulmonary oedema. Lancet 1998;351:389 93.
revascularize occluded coronaries in cardiogenic shock? J Am w15x Cotter G, Kaluski E, Blatt A, et al. L-NMMA (a nitric oxide
Coll Cardiol 2000;36:1071 6. synthase inhibitor) is effective in the treatment of cardiogenic
w3x Kaluski E, Kobrin I, Zimlichman R, et al. RITZ-5: Randomized shock. Circulation 2000;101:1358 61.
Intravenous Tezosentan (an Endothelin ET-AyB Antagonist) w16x Sharon A, Shpirer I, Kaluski E, et al. High-dose intravenous
for the treatment of pulmonary edema. A prospective random- isosorbide-dinitrate is safer and better than Bi-PAP ventilation
ized, multicenter, double-blind placebo controlled study. J Am combined with conventional treatment for severe pulmonary
Coll Cardiol 2003;41:20410. edema. J Am Coll Cardiol 2000;36:832 7.
w4x Morley D, Brozena SC. Assessing risk by hemodynamic profile w17x Mohr R, Rath S, Meir O, et al. Changes in systemic vascular
in patients awaiting cardiac transplantation. Am J Cardiol resistance detected by the arterial resistometer: preliminary
137 of 158.
www.lejacq.com ID:3405
EDITORIAL
and sympathetic nervous systems and Damage Cold & Dry Cold & Wet
139 of 158.
141 of 158.
cant implications for the management of 1. Establish baseline cardiac index (CI), systemic vascular resistance index (SVRI) and thoracic fluid content (TFC)
patients with HF. 2. Monitor changes in CI to objectify perfusion assessment
3. Monitor changes in TFC to objectify congestion assessment
Incorporating this model of assess- 4. Monitor changes in SVRI to objectify vasoactive status
ment into a proposed therapeutic algo-
rithm is shown in Figure 10. Ideally, Figure 9. Model for clinical profiles in heart failure utilizing impedance cardiogra-
phy hemodynamic measurements
a baseline measurement of ICG in
addition to other standard clinical
variables would be collected and uti- of higher risk may lead to increased Note: This supplement to Congestive
lized in combination to more precisely clinical surveillance or a decision to Heart Failure contains articles dealing
assess a patients perfusion, conges- intervene to prevent a negative patient with ICG. Readers are reminded that
tion, and vasoactive status. This assess- outcome. In addition, ICG parameters positive statements about the clinical
ment would lead to a categorization may aid in the assessment of a stable, utility of ICG, and the BioZ ICG Monitor
of the patients absolute or relative low-risk hemodynamic profile toward in particular, are solely the opinions of
change in hemodynamic profile, facili- the initiation and up-titration of neuro- the authors and do not represent an offi-
tating assessment of short-term risk for hormonal agents that are often under- cial endorsement by Congestive Heart
adverse HF-related events. The change prescribed but are known to improve Failure, its Editors or Editorial Board, or
in hemodynamic status and assessment 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
Higher SVRI
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 8 Marik PE, Pendelton JE, Smith R. A compari- Care Clin. 1996;12:771776.
American College of Cardiology/American son of hemodynamic parameters derived from 13 Mueller C, Scholer A, Laule-Kilian K, et al.
Heart Association Task Force on Practice. J transthoracic electrical bioimpedance with Use of B-type natriuretic peptide in the evalu-
Am Coll Cardiol. 2001;38(7):21012113. those parameters obtained by thermodilution ation and management of acute dyspnea. N
3 OConnell JB. The economic burden of heart fail- and ventricular angiography. Crit Care Med. Engl J Med. 2004;350:647654.
ure. Clin Cardiol. 2000;23(3 suppl):III6III10. 1997;25(9):15451550. 14 Shah MR, OConnor CM, Sopko G, et
4 Heart Disease and Stroke Statistics2003 9 Greenberg BH, Hermann DD, Pranulis MF, al. Evaluation Study of Congestive Heart
Update. Dallas, TX: American Heart et al. Reproducibility of impedance cardiog- Failure and Pulmonary Artery Catheterization
Association; 2002. raphy hemodynamic measures in clinically Effectiveness (ESCAPE): design and rationale.
5 Nohria A, Tsang SW, Fang JC, et al. Clinical stable heart failure patients. Congest Heart Am Heart J. 2001;141(4):528535.
assessment identifies hemodynamic profiles Fail. 2000;6(2):7480. 15 Peacock F, Summers R, Emerman C. Emergent
that predict outcomes in patients admit- 10 Drazner M, Thompson B, Rosenberg P, et dyspnea impedance cardiography-aided
ted with heart failure. J Am Coll Cardiol. al. Comparison of impedance cardiography assessment changes therapy: The ED IMPACT
2003;41(10):17971804. with invasive hemodynamic measurements trial. Ann Emerg Med. 2003;42(4):S82.
6 Shah MR, Hasselblad V, Stinnett SS, et al. in patients with heart failure secondary to 16 Taler SJ, Textor SC, Augustine JE. Resistant
Hemodynamic profiles of advanced heart ischemic or nonischemic cardiomyopathy. Am hypertension: comparing hemodynamic man-
failure: association with clinical character- J Cardiol. 2002;89(8):993995. agement to specialist care. Hypertension.
istics and long-term outcomes. J Card Fail. 11 Kazuhiko N, Kawasaki M, Kunihiko T. 2002;39:982988.
2001;7(2):105113. Usefulness of thoracic electrical bioimpedance 17 Santos JF, Parreira L, Madeira J, et al.
7 Sageman WS, Amundson DE. Thoracic elec- cardiography: noninvasive monitoring of car- Noninvasive hemodynamic monitorization for
trical bioimpedance measurement of cardiac diac output. J Card Fail. 2003;9(suppl):S170. AV interval optimization in patients with ven-
output in postaortocoronary bypass patients. 12 Ginosar Y, Sprung CL. The Swan-Ganz cath- tricular resynchronization therapy. Rev Port
Crit Care Med. 1993;21(8):11391142. eter: twenty-five years of monitoring. Crit Cardiol. 2003;22(9):10911098.
143 of 158.
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
new technology for measuring cardiac output with thermo- f rmed by hemodynamic measurements and by post-
dilution measured CO in patients undergoing coronary operative chest roentgenograms. The pulmonary artery
artery bypass grafting (CABG). catheter was connected to a thermodilution cardiac output
monitor (Horizon 2000, Mennen Medical, Israel was used
at Wolfson Medical Center, and 7010 Series Marquette,
2. Patients and methods Madison, WI at Johns Hopkins Institutions). Ten milliliters
of room temperature 5% dextrose solution was injected
2.1. Patients manually at end-expiration by an experienced anes-
thesiologist who was unaware of the bioresistance CO
Thirty-one patients undergoing elective, primary CABG results.
were prospectively studied; 15 at Wolfson Medical Center
and 16 at Johns Hopkins University School of Medicine. 2.4. Whole body resistance method
Patients with aortic valve insufficien y, mitral and tricuspid
regurgitation or cardiac shunt were not included. The NICaSy 2001, a non-invasive cardiac system device
CABG was similar in both hospitals. All patients underwent (NICaSy 2001, Teledyne-NIM, LLC, North Andover,
a median sternotomy. Cardiopulmonary bypass was estab- MA, 510K, certificat number K942227), was utilized for
lished using an aortic and single venous cannula. Diastolic measuring bioresistance CO. Two proprietary designed,
arrest was achieved with cold sanguineous potassium cardi- NICaSy disposable electrodes (510K, certificat number
oplegia in both antegrade and retrograde fashion. The inter- K972002) were applied to each wrist, and were then con-
nal mammary artery was used to graft the left anterior nected to the non-invasive cardiac system device. The sys-
coronary artery (LAD) in all cases, and saphenous vein tem passed an AC current 30 kHz and 1.4 mA through the
grafts were used to bypass the other obstructed vessels. whole body. It then measured the resistive portion of the
Proximal anastomoses were performed using a partial bioimpedance. Since the electrodes were placed on the dis-
cross clamp. During the study, no patient had atrial flutte tal aspect of the extremities, the current passed through all
nor atrial fibrillation major arteries and veins of the systemic circulation. As
such, the system measured the bioresistance of the systemic
2.2. Protocol circulation and allowed calculation of the stroke volume.
In 1992, Tsoglin and Frinerman and developed a new
The study protocol was approved by Edith Wolfson and semi-empirical formula for calculating the stroke volume
Johns Hopkins Institutional Review Boards. For each (SV) [13].
patient demographic and clinical data was tabulated. In
Hctcorr H 2 DR a+b
each patient, CO measurements were taken at fiv different SV = kel Kweight IB corr
Ksex, age R b
time intervals during the procedure; after induction of
anesthesia but before incision (T1), after sternotomy (T2), where Kel is a coefficien related to blood electrolytes,
after creation of a pericardial pocket (T3), ten min after Kweight is a ratio of a patients weight to ideal weight, IB
weaning from bypass (T4), and immediately after arrival is the index balance equal to the ratio of the measured
in the intensive care unit (T5). extracellular flui (as measured from the baseline impe-
At each time interval, three adequate thermodilution mea- dance) to the expected extracellular flui volume, HCTcorr
surements were made. All measurements were made at end is a hematocrit correction factor proportional to measured
expiration during the respiratory cycle. Thirty-one patients hematocrit, Ksex,age is a coefficien depending upon a
underwent 155 thermodilution CO measurements. Nineteen patients sex and age, H is a patients height, DR is the
were excluded because of 10% variation between the mea- incremental change of the resistive portion of the bioimpe-
surements. Simultaneously, three bioresistance measure- dance, R is the baseline whole body bioresistance and
ments were taken at each interval, and their average was a + b/b is the ratio of the sum of the systole time plus
considered the bioresistance CO. diastole time divided by the diastole time derived from
Each bioresistance measurement took 20 s. Since these the fin structure of the varying portion of the bioresistance.
measurements were continuous, there was no time between Using whole body resistance measurements and the
measurements such that to achieve an average bioresistance above formula, SV was calculated and converted to CO.
CO measurement took 1 min.
2.4.1. Statistical analysis
2.3. Thermodilution method The correlation between methods at each time point was
evaluated by Pearsons correlation coeff cient and simple
At both hospitals seven French true size thermodilution linear regression. The degree of agreement between meth-
catheters (Baxter Healthcare, Irvine, CA) were inserted in ods at each time point was evaluated by calculation of the
the operating room after induction of anesthesia. The bias (mean between-method difference) and precision
proper location of the thermodilution catheter was con- (mean bias 2 SD) [14].
145 of 158.
66 A.J. Cohen et al. / European Journal of Cardio-thoracic Surgery 14 (1998) 6469
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
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 icant current flo and constant resistance. In each
current introduced into the body will travel primarily in this individual, the resistance in the tissue compartment is
compartment. The resistance changes in this compartment determined by the patients height, lean body weight (mus-
as the blood volume varies in the great arteries due to pul- cle) to fat ratio, sex, age, body build, extracellular flui
satile flow In the tissue compartment, there is less signif- 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 data. The device allows accurate, easy to obtain, non-inva-
pulsatile blood volume changes within the system. To mea- sive CO during cardiac surgery. Within the above stated
sure SV, the pulsatile changes within the blood compartment limitations, we have confir ed the validity of the new biore-
and baseline bioresistance must be measured accurately and sistance methodology in 31 patients who underwent CABG.
all other factors must be accounted for with appropriate Bioresistance measured cardiac output correlated well with
correction factors. The measurement of whole body biore- ThCO during the CABG procedure and the immediate post-
sistance allows for the accurate measurement of pulsatile operative period.
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 Acknowledgements
calculated.
The method utilized in this study has been proven accu- This paper was prepared in consultation with Diklah
rate in patients undergoing catheterization [12]. It was also Geva who prepared the statistical calculations, and with
shown to be accurate in a pilot study in patients under- the technical assistance of Sally Esakov.
going CABG [26]. Compared with previous attempts to
correlate bioresistance with SV, the equation has major
advantages. References
1. The equation does not depend on rapidly fluct ating
time derivatives of bioimpedance. [1] Mermel LA, Maki DG. Infectious complications of SwanGanz
2. The equation uses empirically derived coeffici nts that pulmonary artery catheters: pathogenesis, epidemiology, prevention,
and management. Am J Respir Crit Care Med 1994;149:1020
are obtained from laboratory values and simultaneously 1036.
measured bioimpedance values instead of the artifi ial [2] Choh JH, Khazei AH, Ihm HJ, Thatcher WC, Batty PR. Catheter
and difficu t to approximate VEPT. induced pulmonary arterial perforation during open heart surgery. J
3. The precise placement of the electrodes is not a critical Cardiovasc Surg (Torino) 1994;35:6164.
factor. [3] Denniston JC, Maher JT, Reeves JT, Cruz JC, Cymerman A, Grover
RF. Measurement of cardiac output by electrical impedance at rest
4. The NICaSy electrode arrangement is optimal to mea- and during exercise. J Appl Physiol 1976;42:9195.
sure the left ventricular SV. [4] Smith SA, Russell AE, West MJ, Chalmers J. Automated non-inva-
5. The respiration has almost no effect on the NICaSy CO sive measurement of cardiac output: comparison of electrical bioim-
measurements in real time. pedance and carbon dioxide rebreathing techniques. Br Heart J
1988;59:292298.
The new methodology showed good correlation between [5] Reybrouck T, Fagard R. Assessment of cardiac output at rest and
thermodilution and bioresistance CO during all phases of during exercise by a carbon dioxide rebreathing method. Eur Heart J
CABG and the immediate postoperative period in two inde- 1990;11 (suppl I ):2125.
[6] Belardinelli R, Ciampani N, Costantini C, Blandini A, Purcaro A.
pendent hospital populations. Correlation was good in all Comparison of impedance cardiography with thermodilution and
ranges, including low cardiac output. This fact is important direct Fick methods for non-invasive measurement of stroke volume
since patients undergoing CABG frequently have low car- and cardiac output during incremental exercise in patients with
diac output. This was even true in the immediate post car- ischemic cardiomyopathy. Am J Cardiol 1996;77:12931301.
diopulmonary bypass period where changes in the patients [7] White SW, Quail AW, DeLeeuw PW, Traugott FM, Brown WJ,
Porges WL, Cottee DB. Impedance cardiography for cardiac output
volume status, temperature, blood electrolytes and hemody- measurement: an evaluation of accuracy and limitation. Eur Heart J
namics are changing rapidly. Such a correlation would be 1990;11 (Suppl ):7992.
impossible using previous techniques. [8] Teo KK, Hetherington MD, Haennel R, Greenwood PV, Rossall RE,
The new device has certain limitations. First, it cannot Kappagoda T. Cardiac output measured by impedance cardiography
during maximal exercise tests. Cardiovasc Res 1985;19:737743.
measure CO while using diathermy. Second, the device is
[9] Sageman WS, Amundson DE. Thoracic electrical bioimpedance
sensitive to movement so that the patient cannot be manipu- measurement of cardiac output in postaortocoronary bypass
lated while the CO is being measured. These two factors patients. Crit Care Med 1993;21 (8 ):11391142.
require that the operation stop for the 2030 s required to [10] Clarke DE, Raffi TA. Thoracic electrical bioimpedance measure-
measure CO. Finally, the device measures SV and multiplies ment of cardiac output not ready for prime time. Crit Care Med
1993;21 (8 ):11111112.
it by heart rate to measure CO. Arrythmias in which the
[11] Thomas AN, Ryan J, Doran BR, Pollard BJ. Bioimpedance versus
heart rate varies signific ntly will result in a non-represen- thermodilution cardiac output measurement: the Bomed NCCOM3
tative measurement of CO. after coronary bypass surgery. Intensive Care Med 1991;17 (7):383
In summary, a new device using a semi-empirical equa- 386.
tion relating SV to changes in the resistive portion of the [12] Miller HI, Frinerman E, Tsoglin A, Frinerman A, Rosenschein U,
patients bioresistance has been developed. Information Keren G, Roth A, Laniado S. Non-invasive bioimpedance measure-
ment of cardiac output: a validation study (Abstr). Second Confer-
about specif c patient data, blood components and body ence of the Israel Heart Society, 1993:58.
habitus inserted into the analytical software has signific ntly [13] Tsoglin, A., Frinerman, E, N.I. Medical, Ltd. Non-invasive method
improved the ability to calculate SV from the bioresistance and device for collecting measurements representing body activity
148 of 158.
A.J. Cohen et al. / European Journal of Cardio-thoracic Surgery 14 (1998) 6469 69
and determining cardiorespiratory parameters of the human body [20] Wong DH, Onishi R, Tremper KK, Reeves C, Zaccari J, Wong AB,
based upon the measurements collected. US Patent No. 5,469,859, Miller JB, Cordero V, Davidson J. Thoracic bioimpedance and Dop-
Nov. 28 1995. pler cardiac output measurement: learning curve and interobserver
[14] Bland, J.M., Altman, D.G. Statistical methods for assessing agree- reproducibility. Crit Care Med 1989;17:11941198.
ment between two methods of clinical measurement. Lancet [21] Fuller HP. Evaluation of left ventricular function by impedance car-
1986;Feb 81(8476):307310. diography: a review. Prog Cardiovasc Dis 1994;36:267273.
[15] Kubicek WG, Patterson RP, Witsoe DA. Impedance cardiography as [22] Patterson RP. Sources of the thoracic cardiogenic electrical impe-
a non-invasive method of monitoring cardiac function and other dance signal as determined by a model. Med Biol Eng Comput
parameters of the cardiovascular system. Ann NY Acad Sci 1985;23:41417.
1970;170:724729. [23] Karnegis JN, Kubicek WG. Physiological correlates of the cardiac
[16] Bernstein DP. A new stroke volume equation for thoracic electrical thoracic impedance waveform. Am Heart J 1970;79:519523.
bioimpedance: theory and rationale. Crit Care Med 1986;14:904 [24] Sakamoto K, Muto K, Kanai H, Iizuka M. Problems of impedance
909. cardiography. Med Biol Eng Comput 1979;17:697709.
[17] Donovan KD, Dobb GJ, Woods WP, Hockings BE. Comparison of [25] Paul E, Marik MD, Judy E. A comparison of hemodynamic para-
transthoracic electrical impedance and thermodilution methods for meters divided from transthoracic electrical bioimpedance with those
measuring cardiac output. Crit Care Med 1986;14:10381044. parameters obtained by thermodilution and ventricular angiography.
[18] Salandin V, Zussa C, Risica G, Michielon P, Paccagnella A, Cipolotti Crit Care Med 1997;25:15451550.
G, Simini G. Comparison of cardiac output estimates by thoracic [26] Cohen A, Frinerman E, Katz M, Ezra S, Dotan A, Weissberg A,
electrical bioimpedance, thermodilution and Flick methods. Crit Schachner A. Validity of bioimpedance hemodynamic measurements
Care Med 1988;16:11571158. during coronary artery bypass grafting (CABG) (Abstr). J Cardiovasc
[19] Thomas SHL. Impedance cardiography using Sramek-Bernstein Diag Proc 1996;13:57.
method: accuracy and variability at rest and during exercise. Br J
Clin Pharmacacol 1992;34:467476.
149 of 158.
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.
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.
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.
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.
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.
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.
157 of 158.
Examples of theoretical usage of non-invasive measurement of trans-
thoracic impedance in the homebound patient:
158 of 158.