Brief Communication: Outcomes of Subsequent Pregnancy after Peripartum Cardiomyopathy: A Case Series from Haiti

Article
Brief Communication: Outcomes of Subsequent Pregnancy after
Peripartum Cardiomyopathy: A Case Series from Haiti
James D. Fett, MD; Len G. Christie, MD; and Joseph G. Murphy, MD
Background: Maternal risks with pregnancies after an index diag-
worsening heart failure; of these, 1 died and 1 regained normal left
nosis of peripartum cardiomyopathy (PPCM) are inadequately un-
ventricular systolic function. Seven patients tolerated pregnancy
derstood.
without worsening heart failure, and ventricular function recovered
in these patients within 30 months after the subsequent pregnancy.
Objective: To describe the clinical outcomes of subsequent preg-
nancy in Haitian women with PPCM.
Limitations: The results may not apply to non-Haitian women, and
power was insufficient to identify factors that might predict recov-
Design: Prospectively identified cases from a defined population
ery (n
15).
base, 2000 –2005.
Conclusions: Half of the women with subsequent pregnancy after
Setting: Hoˆpital Albert Schweitzer, Deschapelles, Haiti.
PPCM experienced worsening heart failure and long-term systolic
Patients: 15 patients with PPCM and subsequent pregnancies
dysfunction, while the other half experienced no deterioration and
among 99 prospectively identified patients with PPCM.
regained normal left ventricular systolic function.
Measurements: Clinical and echocardiographic parameters.
Results: Fifteen women with PPCM had 16 subsequent pregnan-
Ann Intern Med. 2006;145:30-34.
www.annals.org
cies after the index pregnancies. Eight of these patients experienced
For author affiliations, see end of text.
Peripartum cardiomyopathy (PPCM) is a poorly charac- cardiographic evidence of left ventricular systolic dysfunc-
terized cause of heart failure in previously healthy
tion with ejection fraction less than 0.45. We confined the
women, with onset in late pregnancy or early in the post-
study to HIV-negative patients.
partum period. We have reported (1–3) the high incidence
We defined the pregnancy associated with the initial
of PPCM in the Hoˆpital Albert Schweitzer (4) district of
diagnosis of PPCM as the index pregnancy, and we defined
Haiti: 1 case per 300 live births (10 times the incidence in
all later pregnancies as the subsequent pregnancies. We use
the United States). Other outcome features reported in this
the term worsening heart failure throughout the paper to
population (3) include a mortality rate of 15.3% over a
denote 1) the reappearance of clinical signs of heart failure
5-year observational period for 98 prospectively identified
or progression to New York Heart Association (NYHA) (9)
patients and a left ventricular systolic function recovery
functional class II or greater and 2) a decrease in left ven-
rate of 28% in 92 patients who were observed for at least 6
tricular ejection fraction by at least 0.10 points during or
months after diagnosis. Important additional concerns of
within 5 months after the subsequent pregnancy. We de-
clinical research in PPCM are maternal survival and toler-
fined recovery of cardiac function as 1) NYHA functional
ance of future pregnancies. In 1 of few articles that exam-
class I, 2) no clinical signs of heart failure, and 3) left
ine this, Elkayam and colleagues (5) reported a retrospec-
ventricular ejection fraction greater than 0.50.
tive multicenter case study from the United States that
Patients had a clinical and echocardiographic examina-
underlined the high level of risk in subsequent pregnancies.
tion at least every 6 months and were also counseled to
We report the outcome of subsequent pregnancies in a
avoid pregnancy until ventricular function returned to nor-
case-series study from Haiti, an island nation that is bur-
mal. We reviewed family-planning measures and made
dened by poverty, political instability, and massive health
them available to patients at no cost during each patient
problems.
visit. Standard treatment goals included an angiotensin-
converting enzyme inhibitor during the postpartum pe-
riod, replaced by nitrates plus hydralazine during pregnan-
METHODS
cy; diuretics for clinical fluid overload and digoxin for
Our study focuses on women with subsequent preg-
nancies among 99 prospectively identified patients with
PPCM who were enrolled in the Hoˆpital Albert Schweitzer
See also:
Peripartum Cardiomyopathy Registry from 1 February
Print
2000 to 31 January 2005. We included patients if their
Editors’ Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
disease met accepted definition criteria (6 – 8) for PPCM:
1) the onset of heart failure in the month before delivery to
Web-Only
5 months after delivery, 2) no preexisting heart disease, 3)
Conversion of figure and tables into slides
no other cause identified for the heart failure, and 4) echo-
30 © 2006 American College of Physicians

---

Subsequent Pregnancy after Peripartum Cardiomyopathy Article
patients who were not improving with initial treatment
(withdrawn if they achieved NYHA functional class I);
Context
prophylactic heparin for all hospitalized patients; warfarin
Peripartum cardiomyopathy occurs once in every 300 live
for selected nonpregnant outpatients on the basis of clini-
births in Haiti. The outcome of subsequent pregnancies in
cian assessment of risk and benefit and adherence; and
this impoverished population is not known.
-blockers (limited to atenolol) on the basis of clinician
Contribution
preference and availability.
We are volunteer medical staff of the Hoˆpital Albert
The authors prospectively identified 99 cases and followed
Schweitzer, and our study was part of the Peripartum Car-
them clinically and with echocardiography. Fifteen patients
became pregnant again; only 1 had already recovered nor-
diomyopathy Research Project, conducted with the ap-
mal ventricular function. Eight experienced worsening
proval of the Hoˆpital Albert Schweitzer Ethics Committee.
heart failure during the subsequent pregnancy; only 1 re-
Statistical Analysis
gained normal ventricular function. The remaining 7 re-
We performed statistical analyses with Epi Info 2000
gained normal ventricular function after the pregnancy.
software (U.S. Department of Health and Human Ser-
Cautions
vices, Public Health Service, Centers for Disease Control
and Prevention, Atlanta, Georgia). We considered a P
The study was too small to reliably identify predictors of
value (Mann–Whitney test, Wilcoxon 2-sample test,
worsening heart failure during a subsequent pregnancy.
mid-P exact test, and Fisher exact test) of 0.05 or less to be
Implications
statistically significant.
Some women in heart failure from peripartum cardiomy-
Role of the Funding Source
opathy tolerate subsequent pregnancy, but many get
The study was funded by charitable contributions
worse.
(Pierre Paulette Peripartum Cardiomyopathy Fund) that
are separate from the operating budget of the Hoˆpital Al-
—The Editors
bert Schweitzer. The funding source had no role in the
design, conduct, or reporting of the study or in the deci-
sion to submit the manuscript for publication.
subsequent pregnancy became pregnant before full recov-
ery of left ventricular systolic function and against medical
RESULTS
advice.
We identified 16 subsequent pregnancies in 15 pa-
Eight of 15 patients (53%) had worsening heart failure
tients with PPCM who had had an index pregnancy with a
during subsequent pregnancy, and only 1 of these patients
diagnosis of PPCM. Mean interval between delivery of in-
regained normal left ventricular systolic function after the
dex pregnancy and delivery of subsequent pregnancy was
subsequent pregnancy. One patient died of severe heart
26.7 months (range, 16 to 37 months). Mean follow-up
failure 10 months after the subsequent pregnancy. Seven
after the delivery of the subsequent pregnancy was 20.1
patients (47%) showed no worsening heart failure with
months (range, 6 to 30 months). All but 1 patient with
subsequent pregnancy, and they recovered normal left ven-
Table 1. Baseline Characteristics*
Characteristic
Patients with Subsequent Pregnancy
Patients without
(n
15)
Subsequent
Pregnancy (n
84)
Patients with
Patients without
Worsening Heart
Worsening Heart
Failure (n
8)
Failure (n
7)
Mean age at diagnosis (range), y
34 (21–37)
32.3 (21–41)
32.7 (17–45)
Median parity (range)
5 (2–9)
5 (2–7)
4 (1–11)
Mean interval between index and subsequent pregnancies (range), mo
17.6 (7–32)
20.3 (14–37)
NA
Median NYHA functional class at diagnosis (range)
IV (II–IV)
IV (II–IV)
IV (II–IV)
Eventual recovery of LV systolic function, n (%)
1 (11.1)
7 (100)
22 (26.1)
Survival, n (%)
7 (87.5)
7 (100)
70 (83.3)
Fetal loss in subsequent pregnancy, all spontaneous abortions, n (%)
2 (25)
1 (14)
NA
Toxemia of pregnancy, n (%)
0
0
11 (13.1)
Obstetric delivery, n (%)
Home
7 (87.5)
5 (71.4)
59 (70.2)
Hospital
1 (11.1)
2 (28.6)
21 (25)
No formal schooling, as measure of resources, n (%)
3 (37.5)
4 (57.1)
49 (58.3)
* Last point of data collection was 27 January 2006. LV
left ventricular; NA
not applicable; NYHA
New York Heart Association.
www.annals.org
4 July 2006 Annals of Internal Medicine Volume 145 • Number 1 31

---

Article Subsequent Pregnancy after Peripartum Cardiomyopathy
tricular systolic function during or after the subsequent
with the subsequent pregnancy; only 1 case-patient re-
pregnancy. One patient had 2 subsequent pregnancies after
gained normal clinical or echocardiographic cardiac func-
the index pregnancy, with worsening heart failure occur-
tion during the 6- to 30-month follow-up after the subse-
ring after the second subsequent pregnancy but not after
quent pregnancy. Further observations are necessary to
the first subsequent pregnancy.
determine whether their ventricular dysfunction will be
We could not identify any characteristic that distin-
permanent.
guished the group with subsequent pregnancy (n
15)
Second, the 7 patients who did not experience wors-
from the group without subsequent pregnancy (n
84)
ening heart failure associated with the subsequent preg-
(Table 1). Also, except for recovery of left ventricular sys-
nancy regained normal left ventricular function. They not
tolic function, we could not find a distinguishing feature
only showed improvement after the index pregnancy but
between the group that fully recovered (n
8) and the
also continued improvement during and after the subse-
group that continued to have abnormal heart function
quent pregnancy. Only 1 of these patients was still receiv-
(n
7). The Figure shows the mean echocardiographic
ing heart failure medication during the postpartum period
left ventricular ejection fractions over time. Table 2 shows
of the subsequent pregnancy (Table 2).
the treatment regimens for the 15 patients with PPCM and
Third, improvement in ventricular systolic function is
subsequent pregnancy.
not limited to the first 6 to 12 months after diagnosis, as is
Approximately 75% of the 99 patients with PPCM
commonly believed, but may continue for several years
accepted family-planning measures and did not become
after a diagnosis of PPCM in Haiti. The extended obser-
pregnant again. Among the 25 patients who did not use
vation period up to 30 months postpartum for the subse-
family-planning measures, 14 became pregnant again. One
quent pregnancy allowed us to identify both late recovery
patient became pregnant again despite having an intrauter-
and late deterioration.
ine device.
Fourth, toxemia of pregnancy, twin pregnancy, and
the use of tocolytics, which are all previously cited as risk
DISCUSSION
factors for PPCM, did not occur in any of the 15 patients
Our observational study of subsequent pregnancies in
with PPCM and subsequent pregnancy. Toxemia of preg-
women with well-documented PPCM reports several new
nancy and PPCM are both common in Haiti, but the
important findings. First, approximately half of the case-
absence of toxemia in the 15 patients and its low incidence
patients experienced worsening heart failure associated
in the other 84 patients suggest that toxemia is not a major
Figure. Left ventricular ejection fraction in 15 patients with peripartum cardiomyopathy (PPCM) who experienced a subsequent
pregnancy with (n
8) or without (n
7) recovery during follow-up.
0
at diagnosis, PPCM index pregnancy; 1
postpartum to index pregnancy; 2
postpartum to subsequent pregnancy. Solid lines connect individual
patients’ left ventricular ejection fractions. Dashed lines are smooth mean estimates from cubic smoothing splines. Last point of data collection was 27
January 2006.
32 4 July 2006 Annals of Internal Medicine Volume 145 • Number 1
www.annals.org

---

Subsequent Pregnancy after Peripartum Cardiomyopathy Article
Table 2. Treatment of Heart Failure for 15 Patients with Peripartum Cardiomyopathy and Subsequent Pregnancy*
Patient
Postpartum to Initial Diagnosis†
During Subsequent Pregnancy
Postpartum to Subsequent Pregnancy
Group 1‡
1
Furosemide, hydralazine, nitrates
Furosemide, hydralazine, nitrates
Furosemide, captopril
2
Furosemide, captopril
Furosemide, hydralazine, nitrates
Captopril
3
Furosemide, captopril, digoxin
None
Furosemide, captopril
4
Furosemide, captopril, digoxin
Furosemide, hydralazine, nitrates
Captopril, atenolol
5
Furosemide, captopril, digoxin
Furosemide, atenolol
None
6
Furosemide, captopril, digoxin
None
Furosemide, captopril
7
Furosemide, captopril, digoxin
Furosemide, hydralazine
Furosemide, captopril, hydralazine, digoxin
8
Furosemide, captopril, digoxin
None
Furosemide, captopril
Group 2‡
1
Furosemide, captopril, digoxin
None
None
2
Furosemide, captopril, digoxin
None
None
3
Captopril
Hydralazine, nitrates
None
4
Furosemide, captopril, digoxin
None
None
5
Furosemide, captopril, digoxin
Atenolol
None
6
Furosemide, captopril
Furosemide, atenolol
Atenolol
7
Furosemide, captopril, digoxin
None
None
* Last point of data collection was 27 January 2006.
† All patients received initial diagnosis during the postpartum period.
‡ Group 1: worsening heart failure associated with subsequent pregnancy; group 2: no worsening heart failure associated with subsequent pregnancy.
risk factor in the development of PPCM in this popula-
published studies have confirmed this observation. We do
tion.
not know the reason for the high incidence of PPCM in
Elkayam and colleagues (5), in a multicenter U.S.
Haiti, and we cannot yet explain the reasons for these dif-
study, retrospectively analyzed 44 patients with PPCM and
ferent responses in subsequent pregnancies. Some women
subsequent pregnancy and identified 2 groups of patients:
with subsequent pregnancy after PPCM experienced wors-
those who began the subsequent pregnancy with normal
ening heart failure and subsequent abnormal heart func-
left ventricular function and those who began the subse-
tion, while others experienced no deterioration and re-
quent pregnancy with persistent left ventricular dysfunc-
gained normal left ventricular systolic function.
tion. Heart failure occurred in 21% of the former patients
and 44% of the latter patients. All deaths (3 deaths [7%] in
From Hoˆpital Albert Schweitzer, Deschapelles, Haiti; Oregon Health
44 patients) occurred in patients with abnormal baseline
Sciences University, Portland, Oregon; and Mayo Clinic, Rochester,
left ventricular function.
Minnesota.
Sliwa and colleagues (10) reported that 5 of 6 patients
Acknowledgments: The authors thank the patients, volunteers, and staff
with PPCM in Soweto, South Africa, had deterioration of
at Hoˆpital Albert Schweitzer.
left ventricular systolic function with subsequent preg-
nancy, and 2 patients (33%) died of heart failure within 8
Grant Support: No grants were received. Funding was provided through
weeks postpartum. All patients began the subsequent preg-
charitable contributions (Pierre Paulette Peripartum Cardiomyopathy
nancy with impaired left ventricular systolic function.
Fund) that are separate from the operating budget of the Hoˆpital Albert
In the 15 Haitian patients with PPCM and subse-
Schweitzer.
quent pregnancy, the distribution of echocardiographic
measures at initial diagnosis and before the subsequent
Potential Financial Conflicts of Interest: None disclosed.
pregnancy was similar between the 8 patients with and the
Requests for Single Reprints: Joseph G. Murphy, MD, Division of
7 patients without worsening heart failure. This has impli-
Cardiovascular Diseases, Mayo Clinic, 200 First Street SW, Rochester,
cations for counseling in that clinicians may find it difficult
MN 55905; e-mail, murphy.joseph@mayo.edu.
to predict which patients will and will not recover, as well
as which patients will and will not experience worsening
Current author addresses and author contributions are available at www
heart failure, before the subsequent pregnancy.
.annals.org.
Our results may not apply to non-Haitian patients
with PPCM. Since only 1 of the 15 patients had regained
References
normal ventricular function before the subsequent preg-
1. Fett JD, Carraway RD, Dowell DL, King ME, Pierre R. Peripartum cardio-
nancy, results may not apply to patients with PPCM who
myopathy in the Hospital Albert Schweitzer District of Haiti. Am J Obstet
have recovered normal heart function. Our case series does
Gynecol. 2002;186:1005-10. [PMID: 12015528]
not permit identification of factors that might predict left
2. Fett JD, Carraway RD, Perry H, Dowell DL. Emerging insights into peri-
partum cardiomyopathy. J Health Popul Nutr. 2003;21:1-7. [PMID: 12751668]
ventricular recovery. Although anecdotal evidence suggests
3. Fett JD, Christie LG, Carraway RD, Murphy JG. Five-year prospective study
a high incidence of PPCM in other regions of Haiti, no
of the incidence and prognosis of peripartum cardiomyopathy at a single institu-
www.annals.org
4 July 2006 Annals of Internal Medicine Volume 145 • Number 1 33

---

Article Subsequent Pregnancy after Peripartum Cardiomyopathy
tion. Mayo Clin Proc. 2005;80:1602-6. [PMID: 16342653]
1971;44:964-8. [PMID: 4255967]
4. Barry M, Stansfield SK, Bia FJ. Haiti and the Hoˆpital Albert Schweitzer. Ann
8. Hibbard JU, Lindheimer M, Lang RM. A modified definition for peripartum
Intern Med. 1983;98:1018-20. [PMID: 6344707]
cardiomyopathy and prognosis based on echocardiography. Obstet Gynecol.
5. Elkayam U, Tummala PP, Rao K, Akhter MW, Karaalp IS, Wani OR, et al.
1999;94:311-6. [PMID: 10432149]
Maternal and fetal outcomes of subsequent pregnancies in women with peripar-
9. Konstam MA, Dracup K, Baker DW, Bottorff MB, Brooks NH, Dacey RA,
tum cardiomyopathy. N Engl J Med. 2001;344:1567-71. [PMID: 11372007]
et al. Heart Failure: Evaluation and Care of Patients with Left-Ventricular Sys-
6. Pearson GD, Veille JC, Rahimtoola S, Hsia J, Oakley CM, Hosenpud JD, et
tolic Dysfunction. Clinical Practice Guideline no. 11. Rockville, MD: Agency for
al. Peripartum cardiomyopathy: National Heart, Lung, and Blood Institute and
Health Care Policy and Research; 1994. AHCPR publication no. 94-0612.
Office of Rare Diseases (National Institutes of Health) workshop recommenda-
10. Sliwa K, Forster O, Zhanje F, Candy G, Kachope J, Essop R. Outcome of
tions and review. JAMA. 2000;283:1183-8. [PMID: 10703781]
subsequent pregnancy in patients with documented peripartum cardiomyopathy.
7. Demakis JG, Rahimtoola SH. Peripartum cardiomyopathy. Circulation.
Am J Cardiol. 2004;93:1441-3, A10. [PMID: 15165937]
DOWNLOAD IMPORTANT REFERENCES TO CITATION MANAGERS
At www.annals.org, article citations may be directly downloaded to any
of the following formats: EndNote, Reference Manager, ProCite, BibTeX,
or Medlar.
34 4 July 2006 Annals of Internal Medicine Volume 145 • Number 1
www.annals.org

---