The European Journal of Heart Failure 4 (2002) 305309

The addition of pentoxifylline to conventional therapy improves outcome

in patients with peripartum cardiomyopathy
Karen Sliwa*, Daniel Skudicky, Geoffrey Candy, Anette Bergemann, Mark Hopley, Pinhas Sareli
Department of Cardiology, Baragwanath Hospital, University of the Witwatersrand, PO Bertsham 2013, Johannesburg, South Africa

Received 29 June 2001; received in revised form 3 August 2001; accepted 23 October 2001

Abstract

Downloaded from http://eurjhf.oxfordjournals.org/ by guest on June 8, 2013

We have reported previously that despite treatment with angiotensin-converting enzyme inhibitors and b blockers, the outcome
of patients with peripartum cardiomyopathy (PPC) remains unfavorable. Similar to other etiologies of left ventricular dysfunction,
we found elevated levels of tumor necrosis factor-a (TNF-a) in this group of patients. In the present study we sought to evaluate
the effects of pentoxifylline, a drug known to inhibit the production of TNF-a, on clinical status, left ventricular function, and
circulating plasma levels of TNF-a, in patients with PPC. We followed prospectively 59 consecutive women with PPC. The first
29 patients (group 1) were treated with diuretics, digoxin, enalapril and carvedilol. The next 30 consecutive patients (group 2)
received pentoxifylline 400 mg TID in addition to the previous therapy. Clinical evaluation, echocardiograms and TNF-a
determinations were performed at baseline and after 6 months of treatment. Patients in the pentoxifylline group were older and
had a higher EyA ratio. Nine patients died (eight in group 1, Ps0.009 between groups). A combined end-point of poor outcome
defined as either death, failure to improve the left ventricular ejection fraction )10 absolute points or functional class III or IV
at latest follow-up, occurred in 52% of patients in group 1 and 27% of patients in group 2 (Ps0.03). Treatment with pentoxifylline
(Ps0.04) was the only independent predictor of outcome. In conclusion, the results of this study suggest that the addition of
pentoxifylline to conventional treatment, improves outcome in patients with peripartum cardiomyopathy. 2002 European Society
of Cardiology. Published by Elsevier Science B.V. All rights reserved.

Keywords: Cardiomyopathy; Cytokines; Heart failure

1. Introduction factor-a (TNF-a), has been shown to improve functional

Peripartum cardiomyopathy (PPC) is a disorder of
unknown etiology in which left ventricular dysfunction
and symptoms of heart failure occur between the last
trimester of pregnancy and up to the first 6 months
postpartum. A high mortality rate and overall poor
clinical outcome has been reported in a high percentage
of these patients w14x. We have reported previously
the clinical outcome of patients with PPC receiving
current optimal treatment for heart failure, including
diuretics, digoxin, angiotensin-converting enzyme inhib-
itors and carvedilol w4x. Despite this treatment, mortality
at 6 months remained very high (28%).
Treatment with pentoxifylline, a xanthine derived
agent known to inhibit the production of tumor necrosis
*Corresponding author. Tel.: q27-11-933-8197; fax: q27-11-938-
8945.
E-mail address: hahnle@netactive.co.za (K. Sliwa).
class and left ventricular function in patients with
idiopathic dilated cardiomyopathy w5,6x. However,
whether similar results can be achieved in patients with
other etiologies of left ventricular dysfunction remains
to be established. Therefore, we designed the present
study to evaluate the effects of pentoxifylline, on clinical
status, left ventricular function, and circulating plasma
levels of TNF-a, in patients with PPC.
2. Methods
2.1. Study design and patient enrollment
The protocol was approved by the ethics committee
of the University of the Witwatersrand and the prescrip-
tion and therapeutic committee of Baragwanath Hospital.
All patients gave informed consent before study entry.
The investigation conforms with the principles outlined

1388-9842/02/$ - see front matter 2002 European Society of Cardiology. Published by Elsevier Science B.V. All rights reserved.
PII: S 1 3 8 8 - 9 8 4 2 0 2 . 0 0 0 0 8 - 9
306 K. Sliwa et al. / The European Journal of Heart Failure 4 (2002) 305309
Fig. 1. Flow diagram.
in the Declaration of Helsinki. We enrolled prospectively
59 consecutive black patients with newly diagnosed
PPC attending Baragwanath Hospital cardiac clinic. This
study population fulfilled the following inclusion crite-
ria: (1) age )16 years; (2) New York Heart Association
functional class IIIV; (3) symptoms of congestive heart
failure that developed in the last trimester of pregnancy
or in the first 6 months postpartum; (4) no other
identifiable cause of heart failure; (5) left ventricular
ejection fraction -40% by transthoracic echocardio-
graphy; (6) sinus rhythm; and (7) eligible patients in
whom high quality echocardiographic images could be
obtained. Exclusion criteria were: (1) chronic obstruc-
tive pulmonary disease; (2) significant organic valvular
heart disease; (3) systolic blood pressure )170 mmHg
andyor diastolic blood pressure )105 mmHg; (4) clin-
ical conditions other than cardiomyopathy that could
increase the cytokine levels, (i.e. rheumatoid arthritis,
sepsis, acquired immuno-deficiency syndrome; (7) sig-
nificant liver disease (defined as enzymes )two times
the upper limit of normal); and (8) severe anemia
(hemoglobin concentration -9.0 gydl).
The first 29 patients (group 1) were treated with
conventional therapy including diuretics, digoxin, ena-
lapril and carvedilol. The 6-month outcome of these
patients has been reported previously w4x. The following
30 consecutive patients that were entered into the study,
received pentoxifylline 400 mg TID for 6 months in
addition to conventional therapy (group 2). One month
Downloaded from http://eurjhf.oxfordjournals.org/ by guest on June 8, 2013
after enrolment the dose of drugs were evaluated. The
dose of enalapril and carvedilol remained unchanged
until the end of the trial. All patients were entered into
the trial at first presentation to the hospital. Patients
attended the cardiac clinic for clinical evaluation at least
once a month. The same physicians were in charge of
the evaluation and treatment of the 59 patients. Echo-
cardiograms and TNF-a determinations were performed
at baseline and after 6 months of treatment (Fig. 1).
2.2. TNF-a plasma levels
Fifteen ml of blood were withdrawn from an antecu-
bital vein and collected into prechilled evacuated tubes
containing ethylenediaminetetraacetic acid. Plasma was
separated by centrifugation at 2500 rpm for 12 min
within 15 min of collection and aliquots frozen at y
70 8C. TNF-a measurements were performed using a
commercially available enzyme-linked immunoassay
(Amersham, Maidstone). The average of triplicate undi-
luted determinations was calculated.
2.3. Echocardiographic studies
Two-dimensional targeted M-mode echocardiography
with Doppler color flow mapping was performed using
a Hewlett Packard Sonos 5500 echocardiograph attached
to a 2.5 or 3.5 Mhz transducer. All studies were recorded
on videotape and were done by the same operator. Left
 K. Sliwa et al. / The European Journal of Heart Failure 4 (2002) 305309 307

Table 1
Baseline characteristics of patients treated without (group 1) or with pentoxifylline (group 2)

Group 1 (ns29) Group 2 (ns30) P

Age (years) 29"7 33"5 0.009
Parity 2.6 3.2"1 NS
NYHA functional class (n, %)
II 9 (31%) 10 (33%) NS
III 12 (41%) 13 (44%) NS
IV 8 (28%) 7 (23%) NS
Beginning of symptoms
13 month post partum 17 14 NS
46 month post partum 12 16 NS
Systolic BP (mmHg) 108"16 114"18 NS
Diastolic BP (mmHg) 68"13 72"14 NS
Heart rate (beatsymin) 92"19 97"18 NS
Echocardiographic data
Left-ventricular EDD (mm) 61"13 62"6 NS
Left-ventricular ESD (mm) 54"12 56"6 NS

Downloaded from http://eurjhf.oxfordjournals.org/ by guest on June 8, 2013

Ejection fraction (%) 25"9 23"8 NS
Deceleration time (ms) 104"29 112"44 NS
EyA ratio 1.8"0.9 2.5"1.0 0.007
Data are mean"S.D. EDD, end-diastolic diameter; ESD, end-systolic diameter; NYHA, New York Heart Association.

ventricular dimensions were measured according to the
American Society of Echocardiography guidelines w7x.
For left ventricular measurements the average of )3
beats was obtained. The Left ventricular ejection fraction
was determined as previously described w8x. None of
the patients had paradoxical septal motion (no patient
had left bundle branch block). Diastolic mitral flow was
assessed by pulsed-wave Doppler echocardiography
from the apical four-chamber view. The E wave decel-
eration time was measured as the interval between the
peak early diastolic velocity and the point at which the
steepest deceleration slope was extrapolated to the zero
line.
2.4. Analysis of outcome
A pre-specified combined end-point of poor outcome
was defined as either death, functional class III or IV at
latest follow-up class, or failure to improve left ventric-
ular ejection fraction by 10 absolute units.
2.5. Statistical analysis
Significance was assumed at a two-tailed value of P-
0.05.
3. Results
Baseline characteristics of the study population are
shown in Table 1. Patients treated with pentoxifylline
were older and had a higher EyA ratio. There were no
other significant baseline differences between the two
groups. Only one patient had twin pregnancy. All
patients started with symptoms in the postpartum period.
There were eight deaths in the first group and only one
death among patients treated with pentoxifylline (Ps
0.009 between groups). Four patients were lost to
follow-up (two in each group), and two patients in the
first group relocated to remote areas and did not com-
plete the study. One month after inclusion into the trial
all patients received digoxin 0.25 mg daily, enalapril 20
mg BID and carvedilol. Mean dose of Carvedilol was
27"8 mg daily in group 1 and 26"12 mg daily in the
pentoxifylline group (PsNS). The mean daily dose of
furosemide was also similar (156"27 mg vs. 153"18
mg for groups 1 and 2, respectively, PsNS).

Data are presented as mean"standard deviation.
Group comparisons were made by use of MannWhit-
ney U test or binomial test as appropriate. Wilcoxon
matched pairs test was used for comparison of baseline
data and the results after 6 months. Multivariate analysis
of baseline characteristics was done using logistic regres-
sion analysis to establish predictors of outcome. Data
were analyzed on a personal computer by use of a
commercially available statistical program (Statistica).
3.1. Functional class and left ventricular function
The functional class was considered to improve, if
the patient increased the functional status by at least 1
grade of the New York Heart Association classification.
It was considered to deteriorate, if the patient decreased
the functional class by at least 1 grade or died. In group
1, 60% of patients improved the functional class, and in
40% it remained unchanged or deteriorated. In patients
308 K. Sliwa et al. / The European Journal of Heart Failure 4 (2002) 305309
treated with pentoxifylline, functional class improved in
93% and remained unchanged or deteriorated in 7%
(Ps0.006 between groups). Patients in group 1 that
survived and completed the 6 months follow up (ns
17), showed a non-significant reduction in left ventric-
ular end-diastolic (61"9 to 54"9 mm, Ps0.26) and
end-systolic diameter (53"9 to 43"11 mm, Ps0.54),
with a significant improvement in left ventricular ejec-
tion fraction (27"10% to 43"16%, Ps0.00003). In
the pentoxifylline group, 27 patients completed the
follow up period. In this group there was a significant
reduction in left ventricular end-diastolic (63"6 to
58"8 mm, Ps0.0005) and end-systolic diameter
(56"6 to 45"10 mm, P-0.000001) resulting in an
increment in ejection fraction from 23"7 to 44"13%,
P-0.000001.
3.2. Analysis of outcome
Composite end-point of poor outcome (either death,
functional class III or IV at latest follow-up class, and
or failure to improve left ventricular ejection fraction by
10 absolute units) occurred in 15 patients (52%) in
group 1, and in eight patients (27%) in group 2 (Ps
0.03 between groups). From all baseline characteristics
analyzed, treatment with pentoxifylline (Ps0.04) was
the only independent predictor of outcome using logistic
regression analysis.
3.3. TNF-a plasma levels
Baseline plasma levels of TNF-a were not different
between groups (6.3"4.7 pgyml (ns29) vs. 4.4q3.3
pgyml (ns30), Ps0.08). There was a significant
reduction in TNF-a concentration in the 27 patients
treated with pentoxifylline that completed the study
(from 4.5"3.5 to 3.0"1.1 pgyml, Ps0.03), with a
non-significant decline in the 17 patients in group 1 that
finished the trial (6.9"4.7 to 4.6"4.4 pgyml, Ps0.13).
4. Discussion
Treatment with angiotensin-converting enzyme inhib-
itors and b blockers have significantly improved the
outcome of patients with heart failure due to left ven-
tricular systolic dysfunction. However, despite the addi-
tion of these drugs, the outcome of patients with PPC
remains unfavorable. We have documented previously a
mortality rate of 28% after 6 months of treatment with
diuretics, digoxin, enalapril and carvedilol w4x in patients
with PPC. Similar to other etiologies of left ventricular
dysfunction, we found elevated levels of TNF-a in these
patients. Therefore, and based on our previous observa-
tion that pentoxifylline improves functional class and
left ventricular performance in patients with idiopathic
dilated cardiomyopathy w5,6x, we decided to evaluate
whether the addition of pentoxifylline to conventional
treatment for heart failure would improve outcome in
patients with PPC. Due to the fact that PPC is a
relatively uncommon entity, and the consequent diffi-
culties in recruiting an adequate number of patients to
conduct a randomized trial, we decided to treat the next
30 consecutive patients with PPC that presented to our
clinic and fulfilled the inclusion-exclusion criteria with
pentoxifylline and compared the results with the first 29
patients. Pentoxifylline has been shown to suppress or
reduce the production of TNF-a w911x. It was also
found to inhibit apoptosis in different human cell types
in vitro and in vivo w12,13x. In this registry of consec-
utive patients diagnosed with PPC, we observed a lower
mortality rate and better functional class in patients
treated with pentoxifylline in addition to conventional
therapy. Both groups of patients showed a significant
improvement in ejection fraction. However, only patients
Downloaded from http://eurjhf.oxfordjournals.org/ by guest on June 8, 2013
treated with pentoxifylline had a concomitant significant
reduction of the left ventricular diameters. Furthermore,
due to the high mortality rate, only 17 patients treated
with conventional therapy completed the study and
therefore had a second echocardiogram to evaluate
changes in left ventricular diameters and function from
baseline. Another interesting finding in this study was a
non-significant reduction in TNF-a levels among the 17
patients in group 1 that completed the trial. Therapy
with b blockers has been previously shown to reduce
the levels of circulating cytokines. Ontsuka et al. w14x
showed a significant decline in the TNF-a plasma levels
following 12 weeks of treatment with metoprolol or
bisoprolol in patients with dilated cardiomyopathy.
Moreover, Prabhu et al. w15x showed a significant
decline in the myocardial expression and protein pro-
duction of TNF-a with metoprolol. Therefore, it is
possible that treatment with carvedilol resulted in a
decrease in TNF-a levels in the group of patients not
treated with pentoxifylline.
In conclusion, the results of this non-randomized
study suggest that treatment with pentoxifylline
improves outcome of patients with PPC. These results
must be confirmed in a randomized trial.
4.1. Potential study limitations
The study included consecutive patients but was not
randomized. Thus, one cannot exclude an unintended
selection bias in the patients. Therefore erroneous con-
clusions may be drawn when comparing one population
group to another. The patients with peri-partum cardiom-
yopathy have poor prognosis w4x and are referred infre-
quently to our clinic. Given the favorable outcome in
patients with idiopathic dilated cardiomyopathy treated
with pentoxifylline w5,6x, we treated the consecutive
peri-partum patients with pentoxifylline. The baseline
characteristics were similar between groups. Moreover,
 K. Sliwa et al. / The European Journal of Heart Failure 4 (2002) 305309
the only two different parameters at baseline (older age
and worse diastolic function in the pentoxifylline group)
should have adversely affected the outcome of these
patients. A larger prospective randomized trial is
required to confirm these results.
Acknowledgments
This study was partially funded by the Helen Griffin
Trust and Iris and Ellen Hodges Trust, University of the
Witwatersrand, Johannesburg.
References
w1x Seftel H, Susser M. Maternity and myocardial failure in African
woman. Br Heart J 1961;23:43 52.
w2x Demakis JG, Rahimtoola SH, Sutton GC, Meadows R, Szanto
PB, Tobin JR, Gunnar RM. Natural course of peripartum
cardiomyopathy. Circulation 1971;44:1053 61.
w3x Meadows WR. Idiopathic myocardial failure in the last trimes-
ter of pregnancy and the puerperium. Circulation 1957;15:903
14.
w4x Sliwa K, Skudicky D, Bergemann A, Candy G, Sareli P.
Peripartum cardiomyopathy: analysis of clinical outcome, left
ventricular function, plasma levels of cytokines and FasyAPO-
1. J Am Coll Cardiol 2000;35:701 5.
w5x Sliwa K, Skudicky D, Candy G, Wisenbaugh T, Sareli P.
Effects of pentoxifylline on left ventricular functions in patients
with idiopathic dilated cardiomyopathy. Lancet
1998;351:1091 3.
w6x Skudicky D, Bergemann A, Sliwa K, Candy G, Sareli P. Heart
Failure Research Unit. Beneficial effects of pentoxifylline in
patients with idiopathic dilated cardiomyopathy treated with
The author has requested enhancement of the downloaded file. All in-text references underlined in blue are linked to publications on ResearchGate.
309
angiotensin converting enzyme inhibitors and carvedilol. Cir-
culation 2001;103:1083 8.
w7 x Sahn DJ, DeMaria A, Kisslo J, Weyman A. The committee on
M-mode standardization of the American Society of Echocar-
diography. Recommendations regarding quantitation in M-
mode echocardiography; results of a survey of
echocardiographic measurements. Circulation 1978;58:1072
83.
w8 x Quinones MA, Pickering E, Alexander JK. Percentage of
shortening of the echocardiographic left ventricular dimension:
its use in determining ejection fraction and stroke volume.
Chest 1978;745965.
w9 x Waage A, Sorensen M, Stordal B. Differential effects of
pentoxifylline in tumor necrosis factor and interleukin 6 pro-
duction. Lancet 1990;335:543.
w10x Zabel P, Schonhartig MM, Wolter DT, Schade UF. Oxpentoxi-
fylline in endotoxaemia. Lancet 1989;ii:147477.
w11x Kremsner PG, Grundmann HJ, Neifer S, Sliwa K, Sahlmuller
G, Hegenscheid B, Bienzle U. Pentoxifylline prevents murine
cerebral malaria. J Infect Dis 1991;164:605 8.
w12x Belloc F, Jaloustre C, Dumain P, Lacombe F, Lenoble M,
Downloaded from http://eurjhf.oxfordjournals.org/ by guest on June 8, 2013
Boisseau MR. Effect of pentoxifylline on apoptosis of cultured
cells. J Cardiovasc Pharmacol 1995;25(suppl 2):S71 S74.
w13x Schwarz A, Mahnke K, Luger TA, Schwarz T. Pentoxifylline
reduces formation of sunburned cells. Exp Dermatol 1997;6:1
5.
w14x Ontsuka T, Hamada M, Shigematsu Y, Hara Y, Suzuki M,
Ikeda S, Sasaki O, Hiasa G, Ogimoto A, Hiwada K. Beta
blockade modifies the abnormalities in levels of cytokines and
cytokine moderators in patients with dilated cardiomyopathy.
Circulation 1999;100:1 203.
w15x Prabhu SD, Chandrasekar B, Murray DR, Freeman GL. Adre-
nergic blockade in developing heart failure. Effects on myo-
cardial inflammatory cytokines, nitric oxide, and remodeling.
Circulation 2000;101:2103 9.