ppcficTl2a

Journal of the American College of Cardiology Vol. 58, No. 7, 2011© 2011 by the American College of Cardiology Foundation ISSN 0735-1097/$36.00

STATE-OF-THE-ART PAPER

Clinical Characteristics ofPeripartum Cardiomyopathy in the United StatesDiagnosis, Prognosis, and Management

Uri Elkayam, MD

Los Angeles, California

Peripartum cardiomyopathy is a pregnancy-associated myocardial disease characterized by the development ofheart failure due to marked left ventricular systolic dysfunction. Although the disease is relatively uncommon, itsincidence is increasing, and it can be associated with important and lasting morbidity and with mortality. Peri-partum cardiomyopathy seems to affect women in different parts of the world but with considerable differ-ences in clinical presentation. The purposes of this review are to describe the clinical profile of peripartumcardiomyopathy in the United States and to provide recommendations for the diagnosis and the manage-ment of this disease. (J Am Coll Cardiol 2011;58:659–70) © 2011 by the American College of CardiologyFoundation

Published by Elsevier Inc. doi:10.1016/j.jacc.2011.03.047

Peripartum cardiomyopathy (PPCM) is a pregnancy-associated myocardial disease, reported to occur in differentparts of the world (1). The disease is heterogeneous andseems to have important phenotypic variations in differentgeographical regions; for this reason as well as differences inthe availability and delivery of care, it is difficult to formulateuniform recommendations throughout different parts of theworld. The purposes of this review are therefore to describethe clinical characteristics of PPCM in the United Statesand to provide recommendations for the diagnosis andtreatment of this condition.

Historical Perspective and Definition

Heart failure (HF) associated with pregnancy was firstdescribed as a definitive form of cardiomyopathy in 1937(2). In 1971, Demakis et al. (3) published data on 27

atients with pregnancy-associated cardiomyopathy whoresented in the peripartum period. These investigatorsoined the term “peripartum cardiomyopathy” and de-ned diagnostic criteria on the basis of their patients’haracteristics and available diagnostic tools at the time.hese criteria included: 1) the development of HF in the

ast month of pregnancy or within 5 months of delivery;) the absence of a determinable etiology for HF; and 3) thebsence of demonstrable heart disease before the last

From the Division of Cardiology, Department of Medicine and the Department ofObstetrics and Gynecology, University of Southern California Keck School ofMedicine, Los Angeles, California. Dr. Elkayam has reported that he has norelationships to disclose.

Manuscript received January 6, 2011; revised manuscript received March 2, 2011,accepted March 21, 2011.

month of pregnancy. A workshop organized by theNational Heart, Lung, and Blood Institute and the Officeof Rare Diseases Research in 1997 (4) added an addi-tional criterion proposed by Hibbard et al. (5) of leftventricular (LV) systolic dysfunction demonstrated byechocardiography with left ventricular ejection fraction(LVEF) �45%, fractional shortening �30%, or both.Additional information has indicated that although themajority of patients with PPCM are diagnosed in theperipartum period (Fig. 1), early presentation duringpregnancy is not uncommon (6,7). A recent study of 23cases with pregnancy-associated cardiomyopathy diag-nosed between the 17th and 36th weeks of gestationfound them to be indistinguishable from 100 womenmeeting classic criteria for PPCM (8). These findings,supported by numerous other reports (6,9 –15), clearlyindicate that PPCM and pregnancy-associated cardiomy-opathy represent a continuum of the same disease (7,8).A recent position statement from a European Society ofCardiology working group on PPCM has thereforeexpanded the definition of PPCM to “an idiopathiccardiomyopathy presenting with HF secondary to LVsystolic dysfunction towards the end of pregnancy or inthe months following delivery, where no other cause ofheart failure is found” (16). The majority of patients whoare diagnosed during pregnancy present in the thirdtrimester, with a few in the second trimester (8).

PPCM is a diagnosis of exclusion, and other causes ofcardiac dysfunction should be ruled out. At the same time,however, transient and unexpected depression of LV func-tion typical to PPCM has been described in women with

other forms of heart disease (6). These findings therefore

660 Elkayam JACC Vol. 58, No. 7, 2011Peripartum Cardiomyopathy August 9, 2011:659–70

suggest that the diagnosis ofPPCM should not be excluded inpatients with heart disease,which is otherwise not likely tocause LV dysfunction during orafter pregnancy.

Incidence

A number of recent studies haveprovided information regardingthe incidence of PPCM in theUnited States, ranging from 1 in1,149 to 1 in 4,350 live births(11,17–19), with a mean of 1 in3,186 live births (Table 1). Dif-ferences in incidence amongpublished reports are probablydue to variations in patient pop-ulations but also study design,sample size, and degree of under-reporting (20). Mielniczuk et al.

(18) reported a trend toward an increase in incidence overtime from 1 in 4,350 in 1990 to 1993 to 1 in 2,229 in 2000 to2002 (Fig. 2). This suggested increase in the incidence ofPPCM in the U.S. might be related to a rise in maternalage, a substantial increase in the rate of multifetal pregnan-cies due to contemporary reproductive techniques, andpossibly to increased recognition of the disease. A popula-tion study in Southern California (19) found the greatest

Abbreviationsand Acronyms

ACE � angiotensin-converting enzyme

ARB � angiotensinreceptor blocker

BNP � brain natriureticpeptide

DCM � dilatedcardiomyopathy

HF � heart failure

ICD � implantablecardioverter-defibrillator

LV � left ventricular

LVEF � left ventricularejection fraction

MRI � magnetic resonanceimaging

PPCM � peripartumcardiomyopathy

Figure 1 Time of Diagnosis of PPCM in 123 Patients

Red bars represent 23 patients with diagnosis before the last month of pregnancy. Gpostpartum. PP � postpartum; PPCM � peripartum cardiomyopathy. Adapted from Elk

incidence of PPCM in African Americans (1 in 1,421) andthe lowest in Hispanics (1 in 9,861). The incidence inCaucasians was 1 in 4,075 and in Asians was 1 in 2,675.Higher incidence in African American women has recentlybeen confirmed by Gentry et al. (21), who conducted acase-control study in Augusta, Georgia, and MemphisTennessee, and found almost a 16-fold higher incidence ofPPCM in African American compared with non–AfricanAmerican women.

In summary, the incidence of PPCM in the U.S. seems tobe increasing and is estimated to be approximately 1 in 3,200deliveries, with a significantly higher incidence in AfricanAmerican women and possibly lower incidence in Hispanicscompared with non-Hispanic whites. Because the number oflive births in the United States is �4,300,000 per year (22), theestimated annual number of new patients with PPCM in theU.S. is approximately 1,350.

Etiology

The etiology of PPCM is still unknown, and many potentialcauses have been proposed but not proven (1,16,23). Theseinclude viral myocarditis, abnormal immune response topregnancy, abnormal response to increased hemodynamicburden of pregnancy, hormonal abnormalities, malnutrition,inflammation, and apoptosis. Most recently, experimentalwork has suggested a novel and specific pathogenic mech-anism by demonstrating the development of PPCM infemale mice with a cardiomyocyte-specific deletion of thetranscription factor signal transducer and activator of tran-

rs represent 100 patients diagnosed in the last month of pregnancy or the 5-monthet al. (8).

reen baayam

p

RaHAc(Ho2(wdpPiddhmtddptafaphsPapmopMp

661JACC Vol. 58, No. 7, 2011 ElkayamAugust 9, 2011:659–70 Peripartum Cardiomyopathy

scription 3 (STAT3) protein (24). Absence of cardiomyo-cyte STAT3 in the postpartum heart resulted in increasedoxidative stress secondary to blunted induction of theantioxidant enzyme manganese superoxide dismutase, lead-ing to increased expression and proteolytic activity of cardiaccathepsin D and resulting in cleavage of the nursinghormone prolactin into an antiangiogenic and proapoptotic16-kDa form with a detrimental effect on the myocardialmicrovasculature resulting in myocardial hypoxemia andapoptosis and the development of PPCM. Preliminary datain human demonstrating a favorable effect of bromocriptine,a pharmacological inhibitor of prolactin in a limited numberof patients with PPCM, may support this mechanism ofPPCM (25).

Associated Conditions

Strong associations have been shown between PPCM andolder maternal age, history of hypertension, multiple preg-nancies, and African American background.Age. Although the disease has been reported in womenbetween the ages of 16 and 44 years, the mean age ofwomen with PPCM in the United States has ranged from27 to 33 years (8,11,17–19) (Table 1), with �60% of

atients reported in 1 study to be �30 years of age (8).

Incidence of PPCM in the United StatesTable 1 Incidence of PPCM in the United States

First Author(Year)

(Ref. #) Study Design n

Witlin et al.(1997) (11)

Prospective, single institution 28 of 67,369 deliveries 2

Chapa et al.(2005) (17)

Retrospective, single institution 35 of 40,200 deliveries 1

Brar et al.(2006) (19)

Retrospective review,Southern California KaiserPermanente

60 of 241,497 deliveries

Mielniczuk(2006) (18)

National hospital discharge 16,296 patients among51,966,561 live births

NA � not available; PPCM � peripartum cardiomyopathy.

Figure 2 Change in the Incidence of PPCM Over Time

PPCM � peripartum cardiomyopathy. Data derived from Mielniczuk et al. (18).

U

ace. PPCM in the United States has been reported toffect women of different ethnic groups, including non-ispanic whites, African Americans, Hispanics, andsians. However, the incidence of the disease seems to be

onsiderably higher among African American patients19,21).

ypertension. Hypertension—chronic, pregnancy induced,r preeclampsia—has been described in 15% to 68% (mean3%) of patients with PPCM in the United States3,8,11,17,18,26,27), with a similar incidence reported inomen diagnosed antepartum and postpartum (8). This inci-ence is considerably higher than the 8% reported in allregnant patients (28,29). Symptoms of HF in patients withPCM are often attributed to preeclampsia and hypertension

n patients with both conditions, resulting in a delay of PPCMiagnosis and treatment. Although one can argue that LVysfunction is not truly “idiopathic” in the setting of severeypertension, chronic hypertension is not likely to causearked LV systolic dysfunction in young women, and hyper-

ensive pulmonary edema is due mostly to exacerbation ofiastolic dysfunction by hypertension, not to transient systolicysfunction (31). In fact, assessment of systolic LV function inregnant women with hypertension by a number of investiga-ors has shown it to be preserved (32–34). Preeclampsia canlso present with signs and symptoms of HF, but systolicunction is usually preserved or even improved (33,35–37). Forll these reasons, as well as similar rates of LV recovery inatients with PPCM with and without histories of gestationalypertension (38), the latter does not seem to be a cause of LVystolic dysfunction but a strong associated condition toPCM. Because brain natriuretic peptide (BNP) levelsre only mildly elevated (36,39,40) in patients withreeclampsia, echocardiographic evaluation and measure-ent of BNP levels are advisable for the early diagnosis

f PPCM in patients with preeclampsia who are sus-ected of having HF.

ultifetal pregnancies. Multiple births have been re-orted in 7% to 14.5% of patients with PPCM in the

(yrs),(Mean) Ethnic Background

Period ofStudy Rate

(32) African American 75%;white 21%; Asian 0.5%

1988–1994 1 in 2,406

(27 � 6) African American 80%;white 20%

1988–2001 1 in 1,149

� 7 African American 28%;white 27%; Hispanic20%; Asian 17%;others 8%

1996–2005 All patients 1 in 4,025;whites 1 in 4,075;African Americans1 in 1,421;Hispanics 1 in 9,861;Asians 1 in 2,675

(29.7) African American 32%;white 42%

1990–2002 1 in 3,189

AgeRange

9–35

6–38

33

NA

nited States (3,17,26,27,38), compared with only 3% in

662 Elkayam JACC Vol. 58, No. 7, 2011Peripartum Cardiomyopathy August 9, 2011:659–70

the overall population (41), confirming a strong associationbetween multifetal pregnancies and the development ofPPCM.Parity. Multiparity has been traditionally considered to bea risk factor for PPCM (1). However, most studies in theUnited States have reported the development of PPCM inconjunction with the first or second pregnancy in �50% ofpatients (8,17,26,27). Therefore, these data do not supporta strong association between multiparity and PPCM in theUnited States.

Genetics of Peripartum Cardiomyopathy

PPCM has been classified as a nongenetic form of dilatedcardiomyopathy (DCM) (10). However, a number of stud-ies have reported familial clustering (42–45). Morales et al.(9) recently performed a systematic search of 110 womenfrom 520 families of patients with nonischemic DCM andidentified 45 patients with PPCM. Nineteen of the patientshad been sequenced for genes known to be associated withDCM. This observation was further supported by a Euro-pean study that found PPCM in 6% of 90 families withDCM (10). Screening of first-degree relatives of 3 patientswith PPCM with persistent LV dysfunction revealed undi-agnosed DCM in all 3 families. Furthermore, geneticanalyses showed a mutation in the gene encoding cardiactroponin C (TNNC1) in 1 DCM family with members withPPCM. These findings may suggest that in a proportion ofpatients, PPCM is due to genetic causes (7) or representscases of familial DCM that was unmasked or first recog-nized in pregnancy.

Clinical Presentation

Many of the signs and symptoms of normal pregnancy aresimilar to those of HF; for this reason, and because of thelow incidence of this condition, the diagnosis of PPCM isoften missed or delayed, allowing the development ofpreventable complications (15,30). Most patients presentwith typical signs and symptoms of HF, including dyspneaand orthopnea (6,17); in addition, cough, chest pain, andabdominal pain are frequently encountered and tend toconfuse the initial clinical evaluation (6). Physical examina-tion often reveals tachycardia and tachypnea, blood pressuremay be elevated or reduced, and patients are often not ableto lie down flat because of shortness of breath. There isusually increased jugular venous pressure, displaced apicalimpulse, right ventricular heave, murmurs of mitral andtricuspid regurgitation, third heart sound, pulmonary rales,and peripheral edema. Electrocardiography usually showssinus tachycardia with nonspecific ST-T wave changes. LVhypertrophy can be found as well as left atrial enlargementand, occasionally, conduction abnormalities including leftbundle brunch block (11). Chest radiography usually showscardiomegaly and pulmonary venous congestion or pulmo-nary edema, with or without pleural effusion (11,46). Echo-

cardiography shows variable degrees of LV dilatation, with

moderate to severe depression of systolic function. Rightventricular and biatrial dilatation as well as moderate tosevere mitral and tricuspid regurgitation are commonly seen,with increased pulmonary pressures and mild pulmonaryregurgitation (5,8,11,17,46). Cardiac magnetic resonanceimaging (MRI) has been used in a limited number ofpatients for the assessment of cardiac function and thedetection of mural thrombi or myocardial fibrosis (25,48–51).Although MRI is probably safe during pregnancy (52,53),intravenous gadolinium crosses the placenta, and the 2007American College of Radiology document on safe MRIpractices recommends that it be avoided during pregnancyand used only if absolutely essential (53). Although only0.04% of the maternal dose of gadolinium passes into thebreast milk, it has been recommended to discontinuebreast-feeding for 24 h after intravenous administration(53). In a group of 8 women with PPCM who were studiedwith MRI, none exhibited abnormal myocardial late en-hancement, and no difference was found in the MRIpatterns in 4 patients who recovered normal LV functioncompared with those who did not (47).Brain natiuretic peptide. Levels of BNP do not changesignificantly during normal pregnancy or in the postpartumperiod (36,40,54,55). An early measurement of BNP couldhelp in diagnosing PPCM, in which levels of BNP havebeen shown to be markedly elevated (56).

Prognosis

Recovery of LV function. Recent publications combiningclose to 300 U.S. patients have reported recovery of LVfunction (LVEF to �50%) at 6 months in 45% to 78% ofpatients, with a mean of 54% (26,27,30). Data from mygroup (8) in 40 patients with longitudinal follow-up of 30 �29 months showed that improvement usually occurredwithin the first 6 months after the diagnosis (Fig. 3). Amos

Figure 3 Pattern of Recovery of Left Ventricular Functionin 40 Patients With PPCM

There was a significant increase in left ventricular ejection fraction (LVEF)between time of diagnosis and 6 months (*p � 0.0001), with only a small andstatistically insignificant further increase after 6 months. F/U � follow-up.Adapted from Elkayam et al. (8).

663JACC Vol. 58, No. 7, 2011 ElkayamAugust 9, 2011:659–70 Peripartum Cardiomyopathy

et al. (26) demonstrated LV recovery in 45% of 55 women,mostly occurring within the first 2 months, with continuedimprovement over 1 year (Fig. 4). Most recently, a prelim-inary report from a Utah PPCM registry described LVrecovery in 62% of 58 patients, with an average time of 9months (57). In contrast, Modi et al. (58) reported recoveryof LV function in only 35% of 40 indigent patients, with amedian time to recovery of 54 months. Because 87.5% ofthe patients in this group were African Americans, theinvestigators suggested that race and ethnicity might beresponsible for poorer outcomes. This assumption is sup-ported by a recent analysis by my group demonstrating asignificantly lower rate of LV recovery in 52 AfricanAmerican patients compared with 104 Caucasians (40% vs.61%, p � 0.02; Elkayam et al., unpublished data, 2011). Insummary, the majority of available information in the U.S.demonstrates normalization of LV function in �50% ofwomen with PPCM, mostly occurring within 2 to 6 monthsafter diagnosis; later recovery, however, is possible andoccurs in some patients. The rate of LV recovery seems tobe significantly lower in African American patients com-pared with whites. More information will be needed todetermine potential genetic and environmental causes forthis difference.Predictors of LV recovery. A number of factors have beenshown to be associated with a higher likelihood of recovery,including LV diastolic dimension (�5.5 to 6.0 cm) andsystolic function (LVEF �30% to 35% and fractionalshortening �20%) at the time of diagnosis (5,11,17,27),

Figure 4 Trend in LVEF According to Final Outcome

EF � ejection fraction; LV � left ventricular.Adapted, with permission, from Amos et al. (26).

lack of troponin elevation (59), a lower level of plasma BNP

(56), absence of LV thrombus (26), breast-feeding (27),diagnosis after the delivery (27), and non-African Americanethnicity (30). Recent multivariate analysis by Goland et al.in 187 patients with PPCM (38) found LVEF �30% andLV end-diastolic dimension �55 mm to be significantlyrelated to LV recovery, suggesting a relationship betweenthe degree of initial myocardial insult and recovery. Theseparameters, however, have limited sensitivity in predictingrecovery in individual patients, as evidenced by full recoveryfound in 37% of patients with baseline LVEFs �20% and in51% of those with LVEFs �30%. Baseline parameters ofLV function should therefore not be used as an indicationfor the premature use of devices or heart transplantation.

Is LV recovery related to medical therapy? The relation-ship between standard HF therapy and recovery is notcompletely clear. The rates of recovery in early studies,before the era of contemporary HF therapy (3,60), weresimilar to rates reported in recent studies, and early recoveryoften occurred before up-titration of drugs to optimaltherapeutic doses (26). In addition, similar to nonischemicDCM (61), preliminary reports have shown no significantdifference in the use of beta-blockers in recovered comparedwith nonrecovered patients with PPCM (62,63).

Complications

PPCM can be associated with important and lasting com-plications, including severe HF, cardiogenic shock, cardio-pulmonary arrest secondary to HF or arrhythmias, throm-boembolic complications, and death. Goland et al. (30)recently described major adverse events in 25% of 182patients with PPCM, with 80% of these occurring duringthe first 6 months after the diagnosis and one-third of thesurvivors having residual brain damage secondary to cardio-pulmonary arrest or cerebral vascular events. Predictors ofcomplications were LVEF �25%, non-Caucasian ethnicbackground, and delay of diagnosis.Thromboembolism. PPCM is associated with increasedincidence of thromboembolism compared with DCM ofother etiologies, and LV thrombus has been found on initialechocardiography in 10% to 17% of patients (26,64). Severalreports have described severe thromboembolic events, in-cluding embolization to the coronary, pulmonary, periph-eral, and cerebral arteries (11,26,30,64–70). Increased inci-dence of thromboembolism is probably due to multiplereasons, including the hypercoagulable state of pregnancy(71), cardiac dilatation and dysfunction, endothelial injury,venous stasis, and prolonged bed rest after commonlyperformed instrumental deliveries and cesarean section inpatients diagnosed during pregnancy.Mortality and heart transplantation. Reported mortalityrates associated with PPCM in the United States havevaried widely between 0% and 19%, while rates of cardiactransplantation have ranged from 6% to 11% (11,17–19,26,

30,58,72) (Table 2). Substantial differences in the reported

tcom

esof

Pat

ient

sW

ith

PP

CM

inth

eU

nite

dS

tate

s*ab

le2

Out

com

esof

Pat

ient

sW

ith

PP

CM

inth

eU

nite

dSta

tes*

irst

Aut

hor

(Yea

r)(R

ef.

#)

Stu

dyD

esig

nn

Mea

nFo

llow

-Up

Per

iod

Rec

over

yof

LVFu

ncti

on(L

VEF

>50%

)M

orta

lity

Hea

rtTr

ansp

lant

atio

nP

redi

ctor

sof

LVR

ecov

ery

Pre

dict

ors

ofD

eath

and

Tran

spla

ntat

ion

Witl

inet

al.(

19

97

)(1

1)

Pro

spec

tive,

sing

lece

nter

28

NA

NA

18

%1

1%

NA

NA

Felk

eret

al.(

20

00

)(7

2)

Ret

rosp

ectiv

e,si

ngle

cent

er4

28

.6yr

sN

A7

%7

%N

ALV

stro

kevo

lum

ein

dex

Cha

paet

al.(

20

05

)(1

7)

Ret

rosp

ectiv

e,si

ngle

cent

er3

2N

A4

1%

at3

mon

ths

9.6

%6

.5%

LVED

D�

6.0

cm,L

VEF

�2

0%

NA

Am

oset

al.(

20

06

)(2

6)

Ret

rosp

ectiv

e,si

ngle

cent

er5

54

3�

43

mon

ths

45

%0

%1

0%

LVED

D�

5.6

cm,n

on–A

fric

anA

mer

ican

race

,no

LVth

rom

bus

NA

Mie

lnic

zuk

(20

06

)(1

8)

Nat

iona

lHos

pita

lDis

char

geSur

vey

16

,29

6N

AN

AIn

hosp

italt

otal

2.5

%N

AN

AN

A

Bra

ret

al.(

20

07

)(1

9)

Ret

rosp

ectiv

e,Sou

ther

nCal

iforn

iaK

aise

rP

erm

anen

te6

04

.7yr

sN

A3

.2%

NA

NA

NA

Gol

and

etal

.(2

00

9)(

30

)R

etro

spec

tive,

natio

nwid

e1

82

19

mon

ths

49

%at

6m

onth

s7

%6

%N

ALV

EF�

25

%,d

elay

ofdi

agno

sis

�1

wee

k

Mod

ieta

l.(2

00

9)(

58

)R

etro

spec

tive,

sing

lece

nter

40

54

mon

ths

35

%1

5.9

%0

%LV

EFN

A

dies

with

mor

talit

yre

port

s.�

left

vent

ricu

lar;

LVED

D�

left

vent

ricu

lar

end-

dias

tolic

diam

eter

;LVE

F�

left

vent

ricu

lar

ejec

tion

frac

tion;

othe

rab

brev

iatio

nsas

inTa

ble

1.

664 Elkayam JACC Vol. 58, No. 7, 2011Peripartum Cardiomyopathy August 9, 2011:659–70

incidence of these complications are probably due to varia-tions in patient populations, diagnostic criteria, and treat-ments, as well as reporting bias. Felker et al. (72), in aretrospective review of cardiomyopathies of various etiolo-gies, reported markedly lower mortality in PPCM comparedwith other forms of myocardial disease. At the same time,however, PPCM has become an increasingly recognizedcause of pregnancy-related maternal mortality (73,74).Timing and mode of death. Goland et al. (30) provideddetailed information regarding mortality in 13 patients,most of whom died either suddenly (38%) or of progressiveHF (45%) between the day of delivery and 8 years postpar-tum. Whitehead et al. (74) reported on 17 cases of deathdue to PPCM between 1991 and 1997. Mortality increasedwith maternal age, in women with live birth order of �4,and in black women, who were 6.4 times more likely to diecompared with whites. Eighteen percent of deaths occurredwithin 1 week and 87% within 6 months of diagnosis (Fig. 5),and mortality was due either to progressive HF or to suddencardiac death. Mortality was found by Goland et al. (30) tobe higher in women with baseline LVEFs �25% as well asin women in whom the diagnosis of PPCM was delayed.

Outcome of Subsequent Pregnancy

Habli et al. (75) reported on 21 patients with a meanLVEF �40% who had subsequent pregnancy, with wors-ening of HF in 29% and in none of 8 other patients whoterminated their subsequent pregnancies. Two patients withinitial LVEFs �25% (follow-up LVEFs not provided) whohad subsequent pregnancy and 5 of 8 women who termi-nated their pregnancies demonstrated clinical deteriorationrequiring referral for cardiac transplantation.

Modi et al. (58) described 44 indigent patients withPPCM and reported clinical worsening in 28% of those whohad subsequent pregnancies (number of patients not pro-vided) but no maternal death. Elkayam et al. (76) reportedon the outcomes of 60 subsequent pregnancies in 44women, 28 after recovery of LV function (group 1) and 16with LV dysfunction (group 2). Subsequent pregnancies

Figure 5 Timing of Mortality After Diagnosisin Patients With PPCM

PPCM � peripartum cardiomyopathy. Data derived from Whitehead et al. (74).

were associated with reductions in mean LVEF in the totalOu T F

*Stu LV

gfpaoDc

665JACC Vol. 58, No. 7, 2011 ElkayamAugust 9, 2011:659–70 Peripartum Cardiomyopathy

cohort from 49 � 12% to 42 � 13% (p � 0.001) and from56 � 7% and 36 � 9% to 49 � 10% and 32 � 11% ingroups 1 and 2, respectively. Reductions �20% in LVEFswere seen in 21% of group 1 and 44% of group 2, and therewas 0% mortality in group 1 women and 19% mortality ingroup 2 (Fig. 6). When pregnancies that ended by abortionwere excluded, the risk for unfavorable maternal and fetaloutcomes was even higher, especially in women with per-sistent LV dysfunction. One woman in group 1 whose LVfunction did not change during her first subsequent preg-nancy had a significant decrease from 55% to 40% duringher second subsequent pregnancy. A recent publication byFett et al. (77) reported on 61 post-PPCM pregnancies,with data mostly obtained from an Internet support groupin the United States, and described relapses of PPCM in29% of the entire group, with a significantly higher rate(46%) in women with LVEFs �55%. Nine of the patientswith recovered LV function underwent stress echocardiog-raphy that demonstrated normal contractile reserve, andthese patients did not experience relapse. Although it hasbeen suggested that normal contractile reserve in patientswith PPCM with recovered LV function may ensure goodoutcomes during subsequent pregnancies (78), this concepthas not been tested and should therefore not be used topredict the risk of subsequent pregnancies in women withhistories of PPCM.

In summary, subsequent pregnancies in women withhistories of PPCM in the U.S. are associated with a risk forrecurrent and persistent cardiac dysfunction and even mor-tality. The risk is substantially higher in patients withpersistent LV dysfunction before subsequent pregnancy. Atthe same time, however, recovery of LV systolic functiondoes not guarantee an uncomplicated subsequent pregnancy.Although mortality in such patients is rare, marked de-

Figure 6 Incidence of Maternal Complications AssociatedWith Subsequent Pregnancy in Women With PPCM

Red bars represent women with recovered left ventricular (LV) function beforesubsequent pregnancy; green bars represent women with persistent LV dys-function. HF � heart failure; LVEF � left ventricular ejection fraction. Dataderived from Elkayam et al. (76).

creases in LV function have been reported in approximately r

20% of patients, with persisting dysfunction after pregnancyin about one-half (76,79). Patients should be advised on therisk of subsequent pregnancy and on the safest and mosteffective contraceptive method by both their cardiologistsand obstetricians (80). Patients who decide to becomepregnant again should undergo baseline echocardiographybefore or early in pregnancy, as well as determination ofserum BNP level. In patients on HF medications, baselineLVEF prior to subsequent pregnancy should be determined 3months after the discontinuation of angiotensin-convertingenzyme (ACE) inhibitors or angiotensin receptor blockers(ARBs), which cannot be used during pregnancy. In patientswith LV dysfunction, ACE inhibitors or ARBs should besubstituted with isosorbide dinitrate-hydralazine combination.Because no information is available regarding the safety ofcarvedilol during pregnancy, the use of metoprolol may beconsidered instead. Patients should be followed with repeatechocardiography during the early second and third trimesters,during the last gestational month, and early after delivery andat any time if new symptoms of HF develop. Repeat determi-nation of BNP levels should be helpful in differentiatingbetween HF-like symptoms associated with normal pregnancyand hemodynamic deterioration (55,56). Early termination ofan unintentional pregnancy should be considered to preventworsening of LV function and potential maternal mortality,especially in patients with persistent LV dysfunction.

Breast-Feeding

Breast milk is associated with health, nutritional, immuno-logic, physical, cognitive, and emotional developmentalbenefits to the infant (81). The American Academy ofPediatrics recommends human milk for all infants in whombreast-feeding is not specifically contraindicated. As shownin the following paragraphs, most drugs used for themanagement of HF are compatible with breast-feeding; inaddition, a recent study reported breast-feeding in 67% of55 patients with PPCM without adverse effects to themothers (27). Moreover, the rate of recovery of LV functionwas significantly higher in lactating women. For all thesereasons, clinically stable women with PPCM should not bediscouraged from breast-feeding their infants.

Treatment

Drugs. Standard drug therapy for acute and chronic HFincludes the potential use of diuretic agents, intravenous andoral vasodilators, intravenous inotropes, ACE inhibitors orARBs, beta-blockers, spironolactone, and digoxin (82). Ineneral, the treatment of HF in patients with PPCM shouldollow recent guideline recommendations, except duringregnancy and lactation, when drug therapy may need to beltered because of potential detrimental effects on the fetusr the lactating infant.iuretic agents. Loop diuretic agents are indicated for the

orrection of volume overload and excessive and rapid

eductions in intravascular volume, but they should be used

rcOuthucciBefpfmhibsColmtnwSttfepid

666 Elkayam JACC Vol. 58, No. 7, 2011Peripartum Cardiomyopathy August 9, 2011:659–70

with caution during pregnancy to prevent hypotension anddecreased uterine perfusion. Furosemide (risk category C) isexcreted into breast milk, but no reports of adverse effects innursing infants have been found, and it is listed as probablycompatible with breast-feeding (83).Intravenous vasoactive medications. Vasodilators are rec-ommended in patients with decompensated HF for hemo-dynamic and symptomatic improvement (82). Among theavailable intravenous vasodilators, nitroglycerin (risk cate-gory B) is preferred during pregnancy because nitroprusside(risk category C) may be associated with thiocyanate toxic-ity, and no information is available regarding the safety ofnesiritide. There are only limited data regarding the use ofinotropic agents, including dopamine (risk category C),dobutamine (risk category B), and milrinone (risk category C),and these drugs should therefore be used as recommended(82) only in patients with advanced HF, low blood pressure,high filling pressure, and diminished peripheral perfusiondue to low-output syndrome and in patients who areunresponsive or intolerant to intravenous vasodilators.ACE inhibitors and ARBs. The use of these drugs (bothisk category C) is contraindicated during pregnancy be-ause of toxic effects, mostly on the developing fetal kidneys.ther potential side effects include oligohydramnios, intra-

terine growth retardation, prematurity, bony malforma-ion, limb contractures, patent ductus arteriosus, pulmonaryypoplasia, respiratory distress syndrome, hypotension, an-ria, and neonatal death (84). During pregnancy, theombination of organic nitrates and hydralazine (both riskategory C) should be used as a substitute for ACEnhibitors or ARBs.

eta-blockers. There is lack of human pregnancy experi-nce with the use of all 3 beta-blockers approved in the U.S.or the treatment of HF (carvedilol, bisoprolol, and meto-rolol succinate, all risk category C), and their effects on theetus are therefore unknown. Metoprolol tartrate has beenore commonly used in pregnancy for the management of

ypertension, arrhythmias, mitral stenosis, and myocardialschemia (85). In addition, the use of beta-1-selectiveeta-blockers is preferred during pregnancy, because non-elective beta-blockade could facilitate uterine activity (85).arvedilol and bisoprolol are excreted into the breast milkf lactating rats; no information is available on their use inactating women (83). Metoprolol is secreted into breast

ilk, with a milk/plasma concentration ratio of 2 to 4, buthe amount of drug estimated to be ingested by the infant isegligible, and the drug has been classified as compatibleith breast-feeding (86).pironolactone (risk category C). There is no report of a

eratogenic effect in humans, but there is concern regardinghe antiandrogenic effect of the drug in humans andeminization reported in male rat fetuses (83). The rate ofxcretion of spironolactone in breast milk is unknown. Itsrincipal metabolite, canrenone, is excreted into breast milkn a small amount (approximately 0.2% of the mother’s daily

ose), which seems to be insignificant (83). The American

Academy of Pediatrics (86) classifies spironolactone ascompatible with breast-feeding.Digoxin (risk category C). This drug has been used inpregnancy for both maternal and fetal indications withoutcausing fetal harm. It is excreted into breast milk, but noadverse effects have been reported, and the drug is compat-ible with breast-feeding (86).Anticoagulation. Because of the high incidence of throm-boembolism associated with PPCM (26,64–70), anticoag-ulation from the time of the diagnosis until LV functionrecovers (LVEF �35%) is advisable. Anticoagulation seemsparticularly important during pregnancy and the first 6 to 8weeks postpartum because of persistent hypercoagulablestate (71). In contrast to warfarin (risk factor D), bothunfractionated heparin and low-molecular-weight heparin(risk factor C) do not cross the placenta and are safe duringpregnancy (83). Because of a high prevalence of prematurelabor and a possible need for urgent delivery because ofmaternal or fetal instability, the use of unfractionatedheparin is preferred during pregnancy because of its shorterhalf-life and reversible effect. Neither warfarin nor heparinis secreted into breast milk, and both drugs are thereforecompatible with breast-feeding (83).Experimental drug therapy. IMMUNE GLOBULIN. Bozkurtet al. (87) added intravenous immune globulin to conven-tional HF therapy in 6 women with PPCM and reported asignificantly greater improvement in LVEFs compared with11 historical control patients who received conventionaltherapy alone. Although the results seemed encouraging, avery small number of patients and the lack of a blindlyrandomized, well-matched control group limited the study.

PENTOXIFYLLINE. Sliwa et al. (88) investigated the effect ofpentoxifylline, a xanthine agent known to inhibit the pro-duction of tumor necrosis factor and prevent apoptosis, in30 South African patients with PPCM. These patientsreceived the drug at a dose of 400 mg 3 times daily for 6months in addition to standard HF therapy and werecompared with 29 patients with PPCM who receivedstandard therapy alone. The results of the study demon-strated a significant improvement in a combined endpointincluding death, failure to improve LVEF by 10 absolutepoints, or persistence of New York Heart Associationfunctional class III to IV at the last follow-up (52% vs. 27%,p � 0.03). Despite these positive results, no further studieshave been conducted, and this therapy has not been widelyused. In addition, the safety of pentoxifylline during preg-nancy and lactation has not been established; it is excretedinto human milk and has been defined as probably compat-ible with breast-feeding (83).

BROMOCRIPTINE. On the basis of the concept of enhancedoxidative stress–mediated cleavage of the nursing hormoneprolactin into an antiangiogenic and proapoptotic 16-kDaform that may be responsible for the development of PPCM(24), Sliwa et al. (25) attempted the use of bromocriptine, a

prolactin blocker, in the treatment of 10 African patients

667JACC Vol. 58, No. 7, 2011 ElkayamAugust 9, 2011:659–70 Peripartum Cardiomyopathy

with PPCM. The drug was given after diagnosis at a dose of2.5 mg twice daily for 2 weeks, followed by 2.5 mg/day for6 weeks, in addition to standard HF therapy and resulted ina significantly larger rate of LV recovery at 6 monthscompared with a control group of 10 women with PPCMtreated with standard therapy alone (31% vs. 9%, p �0.012). In addition, there was a lower rate of mortality in thetreatment group (1 vs. 4 patients) and a lower rate of acombined endpoint of death, New York Heart Associationfunctional class III or IV, or LVEF �35% at 6 months.Although the results are intriguing, the study suffered fromimportant limitations, including a very small number ofpatients, excessive mortality and a lower recovery rate in thecontrol group compared with rates reported in the UnitedStates and even previously by the same investigators inSouth Africa (89). In addition, the use of bromocriptine isassociated with suppression of breast milk production andpotential complications to the mother (90), and its approvalfor the prevention of lactation was withdrawn by the U.S.Food and Drug Administration for safety concerns (91). Forall these reasons, further studies aimed at clearly establishingthe efficacy and safety of bromocriptine are needed before itcan be recommended for the treatment of PPCM.Should drug therapy be stopped in women with PPCMafter recovery? This is a commonly asked question bypatients with PPCM who are eager to stop taking medica-tions after recovery. Because only limited long-term, pro-spective data are available, no recommendations can bemade. Amos et al. (26) reported a lack of deterioration ofLV function during an average follow-up period of 29months in 15 patients with full recovery who stopped takingACE inhibitors, beta-blockers (n � 11), or both (n � 5).When discontinuation of drug therapy is desired, it shouldbe done gradually, with repeated echocardiographic evalu-ations of cardiac function. Because the spontaneous deteri-oration of LV function was reported by my group in 3patients after either complete recovery (n � 2) or partial(LVEF 45%) recovery (n � 1) 3 to 60 months afterdiagnosis (30), annual echocardiographic examinations areadvisable in all patients with histories of PPCM.Implantable cardioverter-defibrillators. Because earlysudden death is likely in high-risk patients (8,30,74) andarrhythmias are common in the postpartum period (92), it isoften tempting to consider the early implantation of im-plantable cardioverter-defibrillators (ICDs) in such cases.Recent guidelines, however, recommend consideration ofICDs only in patients with persistent LV dysfunctiondespite optimal drug therapy. These recommendations areespecially applicable to patients with PPCM, in whom theimprovement of LV function is likely, and failure toimprove cannot be predicted in individual patients on thebasis of initial LV function (38). For these reasons, andbecause recovery of LV function occurs in most patientswithin 2 to 6 months after the diagnosis (8,26), it may beadvisable to consider the temporary use of wearable external

defibrillators (93) or entirely subcutaneous ICDs (94) in

high-risk patients as a bridge to recovery or to ICDimplantation in patients with persistent LV dysfunctiondespite adequate trials of optimal medical therapy.Cardiac assist devices. In patients demonstrating rapiddeterioration not responding to medical therapy includingvasoactive medications, intra-aortic balloon pumps, extra-corporeal membrane oxygenation, and LV assist deviceshave been used successfully and should be considered(95–101). Because the rate of recovery in patients withPPCM is higher than in those with other forms of DCM,an attempt should be made to use such devices as a bridgeto recovery before referral for cardiac transplantation(98,100,101).Cardiac transplantation. This procedure has been per-formed successfully in patients with PPCM (102–104). Arecent multi-institutional study by Rasmusson et al. (105)using data from a cardiac transplantation research databasedescribed 69 women who underwent heart transplantationfor PPCM in 29 institutions in the United States. The riskfor rejection was somewhat higher in patients with PPCMcompared with men or women of similar age who did nothave history of pregnancies and similar to that of womenwith histories of pregnancy. The cumulative risk for infec-tions was lowest in patients with PPCM, while freedomfrom allograft vasculopathy and mortality was similar orhigher compared with the other groups. These data indi-cate, therefore, that the overall outcome of heart transplan-tation in women with PPCM is comparable with that oftransplantation for other reasons.

Labor and Delivery

The timing and mode of delivery in a patient diagnosedduring pregnancy should be determined by the clinicalstatus of the mother and the fetus. Termination of preg-nancy or early delivery may result in improvement of bothsymptoms and cardiac function and should be considered inpatients with deteriorating symptoms or cardiac function.Continuation of pregnancy can be allowed, with frequentmonitoring, to allow fetal maturity in patients who can bestabilized on medical therapy. The mode of delivery in astable patient with PPCM should be determined jointly bythe obstetrician and the cardiologist. Vaginal delivery pre-vents potential risks associated with anesthesia and surgicaldelivery that include hemodynamic fluctuations, largerblood loss, pain, infections, respiratory and thromboemboliccomplications, damage to pelvic organs, and potential un-favorable effects on future reproductive health (106). At thesame time, an elective cesarean section is more rapid andallows better planning of the time of delivery as well as thepresence of the most experienced medical team during thedelivery. Hemodynamic monitoring for labor and delivery isdesirable in a patient with PPCM who is diagnosed duringpregnancy and allows optimization of hemodynamic statusbefore delivery as well as monitoring of changes related to

fluid intake and blood loss during delivery and early hemo-

668 Elkayam JACC Vol. 58, No. 7, 2011Peripartum Cardiomyopathy August 9, 2011:659–70

dynamic changes as a result of increased venous return andperipheral vascular resistance after delivery. In case ofvaginal delivery, assisted second stage is recommended toreduce maternal efforts and shorten labor. Maternal vitalsigns as well as oxygen saturation, electrocardiogram, andfetal heart rate should be continuously monitored.

Ongoing Research

The IPAC (Investigation in Pregnancy Associated Cardio-myopathy) study is currently under way in the United States(107). This is a National Institutes of Health–sponsoredmulticenter study and the first prospective trial in theUnited States aiming to enroll 100 patients with newlydiagnosed PPCM and evaluate systemic immune activationas the etiology of this disease and the relationship betweenautoimmunity and LV dysfunction and recovery and inaddition investigate the frequency of myocardial injury orinflammation on cardiac MRI and the ability of tissuecharacteristics to predict subsequent recovery of LVEF.

Reprint requests and correspondence: Dr. Uri Elkayam, LAC/USC Medical Center, 2020 Zonal Avenue, Los Angeles, California90033. E-mail: elkayam@usc.edu.

REFERENCES

1. Sliwa, K, Fett J, Elkayam U. Peripartum cardiomyopathy. Lancet2006:368:687–93.

2. Gouley BA, McMillan TM, Bellet S. Idiopathic myocardial degen-eration associated with pregnancy and especially the puerperium.Am J Med Sci 1937;194:185–99.

3. Demakis JG, Rahimtoola SH, Sutton GC, et al. Natural course ofperipartum cardiomyopathy. Circulation 1971;44:1053–61.

4. Pearson GD, Veille JC, Rahimtoola S et al. Peripartum cardiomy-opathy: National Heart, Lung, and Blood Institute and Office of RareDiseases (National Institutes of Health) workshop recommendationsand review. JAMA 2000;283:1183–8.

5. Hibbard JU, Lindheimer M, Lang R. A modified definition forperipartum cardiomyopathy and prognosis based on echocardiogra-phy. Obstet Gynecol 1999;94:311–6.

6. Lang RM, Lampert MB, Poppas A, Hameed A, Elkayam U.Peripartal cardiomyopathy. In: Elkayam U, Gleicher N, editors.Cardiac Problems in Pregnancy. 3rd edition New York, NY: Wiley-Liss, 1998:87–100.

7. Anderson JL, Horne BD. Birthing the genetics of peripartumcardiomyopathy. Circulation 2010;121:2157–9.

8. Elkayam U, Akhter, MW, Singh H, et al. Pregnancy-associatedcardiomyopathy: clinical characteristics and a comparison betweenearly and late presentation. Circulation 2005;11:2050–5.

9. Morales A, Painter T, Li R et al. Rare variant mutations inpregnancy-associated or peripartum cardiomyopathy. Circulation2010;121:2176–82.

10. Van Spaendonck-Zwarts KY, van Tintelen JP, van Veldhuisen DJ, etal. Peripartum cardiomyopathy as a part of familial dilated cardio-myopathy. Circulation 2010;121:2169–75.

11. Witlin AG, Mabie WC, Sibai BM. Peripartum cardiomyopathy: anomnious diagnosis. Am J Obstet Gynecol 1997;176:182–8.

12. Jahns BG, Stein W, Hilfiker-Kleiner D, Pieske B, Emons G.Peripartum cardiomyopathy—a new treatment option by inhibitionof prolactin secretion. Am J Obstet Gynecol 2008;199:e5–6.

13. Kumari I, Kumar S, Gupta S. Sequential combined spinal epiduralanaesthesia for caesarean section in peripartum cardiomyopathy.

Indian J Anaesth 2007;51:137–9.

14. Velickovic IA, Leicht CH. Peripartum cardiomyopathy and cesareansection: report of two cases and literature review. Arch GynecolObstet 2004;270:307–10.

15. Groesdonk HV, Dinse- Lambracht A, Doblanzki W, Doblanzki U,Galm C, Muth C-M. Unrecognized peripartum cardiomyopathy,case series and comprehensive review of literature. Appl CardiopulmPathophysiol 2009;13:237–42.

16. Sliwa K, Hilfiker-Kleiner D, Petrie MC, et al. Current state ofknowledge on aetiology, diagnosis, management, and therapy ofperipartum cardiomyopathy: a position statement from the heartfailure association of European Society of Cardiology WorkingGroup on Peripartum Cardiomypathy. Eur J Heart Fail 2010;12:767–78.

17. Chapa JB, Heiberger HB, Weinert L, DeCara J, Lang R, HibbardJU. Prognostic value of echocardiography in peripartum cardiomy-opathy. Obstet Gynecol 2005;105:1303–8.

18. Mielniczuk LM, Williams L, Davis DR, et al. Frequency of peripar-tum cardiomyopathy. Am J Cardiol 2006;97:1765–8.

19. Brar SS, Khan SS, Sandhu GK, et al. Incidence, mortality, and racialdifferences in peripartum cardiomyopathy. Am J Cardiol 2007;100:302–4.

20. Deneux-Tharaux C, Berg C, Bouvier-Colle M-H, et al. Underre-porting of pregnancy-related mortality in the United States andEurope. Obstet Gynecol 2005;106:684–92.

21. Gentry MB, Dias JK, Luis A, Patel R, Thornton J, Reed JL.African-American women have a higher risk for developing peripar-tum cardiomyopathy. J Am Coll Cardiol 2010;55:654–9.

22. Hamilton BE, Martin JA, Ventura SJ, Sutton PD. Births: prelimi-nary data for 2007. Natl Vital Stat Rep 2009;57:1–23.

23. Ntusi N, Mayosi BM. Aetiology and risk factors of peripartumcardiomyopathy a systematic review. Int J Cardiol 2009;131:168–79.

24. Hilfiker-Kleiner D, Kaminski K, Podewski E, et al. A cathepsinD-cleaved 16 kDa form of prolactin mediates postpartum cardiomy-opathy. Cell 2007;128:589–600.

25. Sliwa K, Blauwet L, Tibazarwa K, et al. Evaluation of bromocriptinein the treatment of acute severe peripartum cardiomyopathy: a proofof concept pilot study. Circulation 2010;121:1465–73.

26. Amos A, Jaber WA, Russell SD. Improved outcomes in peripartumcardiomyopathy with contemporary. Am Heart J 2006;152:509–13.

27. Safirstein JG, Ro AS, Grandhi S, Wang L, Fett JD, Staniloae C.Predictors of left ventricular recovery in a cohort of peripartumcardiomyopathy patients recruited via the internet. Int J Cardiol 2010Sep 20 [E-pub ahead of print].

28. Robert UM, Pearson G, Cutler J, Lindheimer M. Summary of theNHLBI Working Group on Research on Hypertension DuringPregnancy. Hypertension 2003;41:437–45.

29. Podymow T, August P. Antihypertensive drugs in pregnancy. SeminNephrol 2011;31:70–85.

30. Goland S, Modi K, Bitar F, et al. Clinical profile and predictors ofcomplications in peripartum cardiomyopathy. J Card Fail 2009;15:645–50.

31. Bibbins-Domingo K, Pletcher MF, Lin F, Vittinghott E, Garden J,Arynchyn A, et al. Racial differences in incident heart failure amongyoung adults. N Engl J Med 2009;360:1179–90.

32. Sanchez RA, Glenny JE, Marco E, et al. Two-dimensional andM-mode echocardiographic findings in hypertensive pregnantwomen. Am J Obstet Gynecol 1986;154:910–3.

33. Thompson JA, Hays PM, Sagar KB, Cruikshank DP. Echocardio-graphic left ventricular mass to differentiate chronic hypertensionfrom preeclampsia during pregnancy. Am J Obstet Gynecol 1986;155:994–9.

34. Blanco MV, Roisinbilt J, Grosso O, et al. Left ventricular functionimpairment in pregnancy-induced hypertension. Am J Hypetens2001;14:271–5.

35. Simmons LA, Gillin AG, Jeremy RW. Structural and functionalchanges in left ventricle during normotensive and preeclampticpregnancy. Am J Physiol Heart Circ Physiol 2002;283:627–33.

36. Hamad RR, Larsson A, Pernow J, Bremme K, Eriksson MJ.Assessment of left ventricular structure and function in preeclampsiaby echocardiography and cardiovascular biomarkers. J Hypertens2009;27:2257–64.

37. Dennis AT. Cardiac function in women with preeclampsia. PhD thesis,Department of Pharmacology, Faculty of Medicine, Dentistry and

Health Sciences, University of Melbourne, Melbourne, Australia; 2010.

669JACC Vol. 58, No. 7, 2011 ElkayamAugust 9, 2011:659–70 Peripartum Cardiomyopathy

38. Goland S, Bitar F, Modi K, et al. Evaluation of the clinical relevanceof baseline left ventricular ejection fraction as a predictor of recoveryor persistence of severe dysfunction in women in the United Stateswith peripartum cardiomyopathy. J Cardiac Fail 2011;17:426–30.

39. Borghi C, Esposti DD, Immordino V, et al. Relationship of systemichemodynamic, left ventricular structure and function, and plasmanatriuretic peptide concentrations during pregnancy complicated bypreeclampsia. Am J Obstet Gynecol 2000;183:140–7.

40. Resnik JL, Hong C, Resnik R, et al. Evaluation of B-type natriureticpeptide (BNP) levels in normal and preeclamptic women. Am JObstet Gynencol 2005;193:450–4.

41. Russell RB, Petrini JR, Damus K, Mattison DR, Schwarz RH. Thechanging epidemiology of multiple births in the United States.Obstet Gynecol 2003;101:129–35.

42. Pearl W. Familial occurrence of peripartum cardiomyopathy. AmHeart J 1995;129:421–2.

43. Fett JD, Sundstrom BJ, Ansari AA, Etta King M. Mother-daughterperipartum cardiomyopathy. Int J Cardiol 2002;86:331–2.

44. Baruteau AE, Leurent G, Schleich JM, Gervais R, Daubert JC,Mabo P. Can peripartum cardiomyopathy be familial? Int J Cardiol2009;137:183–5.

45. Massad LS, Reiss CK, Mutch DG, Haskel EJ. Familial peripartumcardiomyopathy after molar pregnancy. Obstet Gynecol 1993;81:886–8.

46. Witlin AG, Mabie WC, Sibai BM. Peripartum cardiomyopathy: alongitudinal echocardiographic study. Am J Obstet Gynecol 1997;177:1129–32.

47. Mouquet F, Lions C, de Groote P, et al. Characterisation ofperipartum cardiomyopathy by cardiac magnetic resonance imaging.Eur Radiol 2008;18:2765–9.

48. Kawano H, Tsuneto A, Koide Y, et al. Magnetic resonance imagingin a patient with peripartum cardiomyopathy. Intern Med 2008;47:97–102.

49. Ntusi NB, Chin A. Characterisation of peripartum cardiomyopathyby cardiac magnetic resonance imaging. Eur Radiol 2009;19:1324–5.

50. Baruteau AE, Ollivier R, Boulmier D, et al. Contribution of cardiacMRI in the comprehension of peripartum cardiomyopathy pathogen-esis. Int J Cardiol 2009;132:e91–3.

51. De Wilde JP, Rivers AW, Price DL. A review of the current use ofmagnetic resonance imaging in pregnancy and safety implications forthe fetus. Prog Biophys Mol Biol 2005;87:335–53.

52. Kanal E, Barkovich AJ, Bell C, et al. ACR guidance document forsafe MR practices: 2007. AJR Am J Roentgenol 2007;188:1447–74.

53. Webb JA, Thomsen HS, Morcos SK. The use of iodinated andgadolinium contrast media during pregnancy and lactation. EurRadiol 2005;15:1234–40.

54. Hameed AB, Chan K, Ghamsary M, Elkayam U. Longitudinalchanges in the B-type natriuretic peptide levels in normal pregnancyand postpartum. Clin Cardiol 2009;32:E60–2.

55. Tanous D, Siu SC, Mason J, et al. B-type natriuretic peptide inpregnant women with heart disease. J Am Coll Cardiol 2010;56:1247–53.

56. Forster O, Hilfiker-Kleiner D, Ansari AA, et al. Reversal ofIFN-gamma, oxLDL and prolactin serum levels correlate withclinical improvement in patients with peripartum cardiomyopathy.Eur J Heart Fail 2008;10:861–8.

57. Rasmusson KD, Budge D, Alharethi R, et al. Long-term outcomesin patients with peripartum cardiomyopathy and no recovery of leftventricular function. J Card Fail 2010;16:S97.

58. Modi KA, Illum S, Jariatul K, Caldito G, Reddy PC. Poor outcomeof indigent patients with peripartum cardiomyopathy in the UnitedStates. Am J Obstet Gynecol 2009;201:171.e1–5.

59. Hu CL, Li YB, Zang JM, et al. Troponin T measurement can predictpersistent left ventricular dysfunction in peripartum cardiomyopathy.Heart 2007;93:488–90.

60. O’Connell JB, Costanzo-Nordin MR, Robinson JA et al. Peripartumcardiomyopathy: clinical, hemodynamic, histologic and prognosticcharacteristics. J Am Coll Cardiol 1986;8:52–6.

61. Starling R, Cooper L, Boehmer J, et al. Beta blockers, myocardialrecovery and outcomes in recent onset cardiomyopathy; results ofIMAC2. J Card Fail 2010;16:S97.

62. Palmer BA, Janosko KM, McTiernan C, Sherman F, McNamaraDM. Left ventricular recovery in peripartum cardiomyopathy: impact

of beta blockade. Circulation 2007;116:II551.

63. Rasmusson K, Alharethi R, Budge D, et al. Peripartum cardiomy-opathy: echocardiographic characteristics and medication use inpatients with versus without recovery of left ventricular function.J Card Fail 2010;16:S86.

64. Napporn AG, Kane A, Damorou JM, et al. Intraventricular throm-bosis complicating peripartum idiopathic-myocardiopathy. Ann Car-diol Angeiol 2000;49:309–14.

65. Agunanne E. Peripartum cardiomyopathy presenting with a pulmo-nary embolism: an unusual case. South Med J 2008;101:646–7.

66. Box LC, Hanak V, Arciniegas JG. Dual coronary emboli in peripar-tum cardiomyopathy. Tex Heart Inst J 2004;31:442–4.

67. Ibebuogu UN, Thornton JW, Reed GL. An unusual case of peripar-tum cardiomyopathy manifesting with multiple thrombo-embolicphenomena. Thromb J 2007:5:18.

68. Jha P, Jha S, Millane TA. Peripartum cardiomyopathy complicatedby pulmonary embolism and pulmonary hypertension. Eur J ObstetGynecol Reprod Biol 2005;123:121–3.

69. Quinn B, Doyle B, Mclnerney J. Postnatal pre-cordial pain, Pulmo-nary embolism or peripartum cardiomyopathy. Emerg Med J 2004;21:746–7.

70. Shimamoto T, Marui A, Oda M, et al. A case of peripartumcardiomyopathy with recurrent left ventricular apical thrombus. CircJ 2008;72:853–4.

71. Hellgren M. Hemostasis during normal pregnancy and puerperium.Semin Thromb Hemost 2003;29:125–30.

72. Felker GM, Thompson RE, Hare J, et al. Underlying causes andlong-term survival in patients with initially unexplained cardiomyop-athy. N Engl J Med 2000;342:1077–84.

73. Lang CT, King JC. Maternal mortality in the United States. BestPract Res Clin Obstet Gynaecol 2008;22:517–31.

74. Whitehead SJ, Berg CJ, Chang J. Pregnancy-related mortality due tocardiomyopathy: United States, 1991–1997. Obstet Gynecol 2003;102:1326–31.

75. Habli M, O’Brien T, Nowack E, Khoury S, Barton JR, Sibai B.Peripartum cardiomyopathy: prognostic factors for long-term mater-nal outcome. Am J Obstet Gyenecol 2008;199:415.e1–5.

76. Elkayam U, Tummala PP, Rao K et al. Maternal and fetal outcomesof subsequent pregnancies in women with peripartum cardiomyopa-thy. N Engl J Med 2001;334:1567–71.

77. Fett JD, Fristoe KL, Welsh SN. Risk of heart failure relapse insubsequent pregnancy among peripartum cardiomyopathy mothers.Int J Gynaecol Obstet 2010;109:34–6.

78. Fett J. Personal commentary: monitoring subsequent pregnancy inrecovered peripartum cardiomyopathy mothers. Crit Pathw Cardiol2010;8:172–4.

79. Elkayam U. Pregnant again after peripartum cardiomyopathy: to beor not to be? Eur Heart J 2002;23:753–6.

80. Thorne S, MacGregor A, Nelson-Piercy C. Risks of contraceptionand pregnancy in heart disease. Heart 2006;92:1520–5.

81. American Academy of Pediatrics. Policy statement: breastfeeding andthe use of human milk. Pediatrics 2005;115:496–506.

82. Lindenfeld J, Albert NM, Boehmer JP, et al. Executive summary:HFSA 2010 comprehensive heart failure practice guideline. J CardFail 2010;16:475–539.

83. Briggs GG, Freeman RK, Yatte SJ. Drugs in Pregnancy andLactation. 8th edition Philadelphia, PA: Wolters Kluwer, 2008.

84. Shotan A, Widerhorn J, Hurst A, Elkayam U. Risks of angiotensin-converting enzyme inhibition during pregnancy: experimental andclinical evidence, potential mechanisms and recommendations foruse. Am J Med 1994;96:451–6.

85. Hurst AK, Hoffman K, Frishman WH, Elkayam U. The use of betaadrenergic blocking agents in pregnancy and lactation. In: ElkayamU, Gleicher N, editors. Cardiac Problems in Pregnancy. 3rd edition.New York, NY: Wiley-Liss, 1998:357–72.

86. American Academy of Pediatrics, Committee on Drugs. The transferof drugs and other chemicals into human milk. Pediatrics 2001;108:776–89.

87. Bozkurt B, Villaneuva FS, Holubkov R, et al. Intravenous immuneglobulin in the therapy of peripartum cardiomyopathy. J Am CollCardiol 1999;34:177–80.

88. Sliwa K, Skudicky D, Candy G, Bergemann A, Hopley M, Sareli P.The addition of Pentoxifylline to conventional therapy improvesoutcome in patients with peripartum cardiomyopathy. Eur J Heart

Fail 2002;4:305–9.

1

1

1

1

1

1

1

1

670 Elkayam JACC Vol. 58, No. 7, 2011Peripartum Cardiomyopathy August 9, 2011:659–70

89. Sliwa K, Skudicky D, Bergemann A, Candy G, Puren A, Sareli P.Peripartum cardiomyopathy: analysis of clinical outcome, left ven-tricular function, plasma levels of cytokines and Fas/APO-1. J AmColl Cardiol 2000;35:701–5.

90. Fett JD. Caution in the use of bromocriptine in peripartum cardio-myopathy. J Am Coll Cardiol 2008;51:2083.

91. U.S. Food and Drug Administration. Sandoz Pharmaceutical Corp.;bromocriptine mesylate (Parlodel); withdrawal of approval of theindication for the prevention of physiological lactation. Fed Regist1995;60:3404–5.

92. Diao M, Diop IB, Kane A, et al. Electrocardiographic reading oflong duration (Holter) of 24 hours during idiopathic cardiomyopathyof the peripartum. Arch Mal Coeur Vaiss 2004;97:25–30.

93. Reek S, Geller JC, Meltendorf U, Wollenbrueck A, Szymkiewicz SJ,Klein HU. Clinical efficacy of wearable defibrillator in acutelyterminating episodes of ventricular fibrillation using biphasic shocks.Pacing Clin Electrophysiol 2003;26:2016–22.

94. Bardy GH, Smith WM, Hood MA, et al. An entirely subcutane-ous implantable cardioverter-defibrillator. N Engl J Med 2010;363:36 – 44.

95. Tandler R, Schmid C, Weyand M, Scheld HH. Novacor LVADbridge to transplantation in peripartum cardiomyopathy. EurJ Cardiothorac Surg 1997;11:394 – 6.

96. Lewis R, Mabie WC, Burlew B, Sibai BM. Biventricular assist deviceas a bridge to cardiac transplantation in the treatment of peripartumcardiomyopathy. South Med J 1997;90:955–8.

97. Yang HS, Hong YS, Rim SJ, Yu SH. Extracorporeal membraneoxygenation in a patient with peripartum cardiomyopathy. AnnThorac Surg 2007;84:262–4.

98. Oosterom L, de Jonge N, Kirkels JH, Klöpping C, Lahpor JR. Leftventricular assist device as a bridge to recovery in a young woman

admitted with peripartum cardiomyopathy. Neth Heart J2008;16:426–8.

99. Zimmerman H, Bose R, Smith R, Copeland JG. Treatment ofperipartum cardiomyopathy with mechanical assist devices and car-diac transplantation. Ann Thorac Surg 2010;89:1211–7.

00. Zimmerman H, Coelho-Anderson R, Smith R, Nolan P, CopelandJ. Bridge to recovery with a Thoratec biventricular assist device forpostpartum cardiomyopathy. ASAIO J 2010;56:479–80.

01. Emmert MY, Pretre R, Ruschitzka F, Krahenmann F, Falk V,Wilhelm MJ. Peripartum cardiomyopathy with a cardiogenic shock:recovery after prolactin inhibition and mechanical support. AnnThorac Surg 2011;91:274–6.

02. Keogh A, Macdonald P, Spratt P, Marshman D, Larbalestier R,Kaan A. Outcome in peripartum cardiomyopathy after heart trans-plantation. J Heart Lung Transplant 1994;13:202–7.

03. Rickenbacher PR, Rizeq MN, Hunt SA, Billingham ME, FowlerMB. Long-term outcome after heart transplantation for peripartumcardiomyopathy. Am Heart J 1994;127:1318–23.

04. Aziz TM, Burgess MI, Acladious NN, et al. Heart transplantationfor peripartum cardiomyopathy: a report of three cases and a literaturereview. Cardiovasc Surg 1999;7:565–7.

05. Rasmusson KD, Stehlik J, Brown R, et al. Long-term outcomes ofcardiac transplantation for peri-partum cardiomyopathy: a multiinstitutional analysis. J Heart Lung Transplant 2007;26:1097–104.

06. Ecker JL, Frigoletto FD Jr. Cesarean delivery and the risk-benefitcalculus. N Engl J Med 2007;356:885–8.

07. Investigation in Pregnancy Associated Cardiomyopathy (IPAC).Available at: http://clinicaltrials.gov/ct2/show/NCT01085955. Ac-cessed June 13, 2011.

Key Words: cardiomyopathy y heart failure y pregnancy.