[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 54.204.247.205. Please contact the publisher to request reinstatement.
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Download PDF
Figure.
Study Cohort Assembly and Exclusion Criteria
Study Cohort Assembly and Exclusion Criteria
Table 1.  
Characteristics of a Woman’s First Pregnancy in the Study Period Resulting in Live Birth or Stillbirth, Denmark, 1978-2012 (Unless Otherwise Specified)
Characteristics of a Woman’s First Pregnancy in the Study Period Resulting in Live Birth or Stillbirth, Denmark, 1978-2012 (Unless Otherwise Specified)
Table 2.  
Hazard Ratios for Cardiomyopathy More Than 5 Months After First Delivery and More Than 5 Years After the Latest Delivery, by History of Hypertensive Disorders of Pregnancy, Denmark, 1978-2012
Hazard Ratios for Cardiomyopathy More Than 5 Months After First Delivery and More Than 5 Years After the Latest Delivery, by History of Hypertensive Disorders of Pregnancy, Denmark, 1978-2012
Table 3.  
Hazard Ratios for Dilated Cardiomyopathy More Than 5 Months After First Delivery by History of Hypertensive Disorders of Pregnancy, Denmark, 1994-2012a,b
Hazard Ratios for Dilated Cardiomyopathy More Than 5 Months After First Delivery by History of Hypertensive Disorders of Pregnancy, Denmark, 1994-2012a,b
Table 4.  
Hazard Ratios for Cardiomyopathy More than 5 Months After First Delivery by History of Hypertensive Disorders of Pregnancy, With Censoring at Onset of Ischemic Heart Disease and Additional Adjustment for Diabetes, Denmark, 1978-2012a,b,c,d
Hazard Ratios for Cardiomyopathy More than 5 Months After First Delivery by History of Hypertensive Disorders of Pregnancy, With Censoring at Onset of Ischemic Heart Disease and Additional Adjustment for Diabetes, Denmark, 1978-2012a,b,c,d
1.
American College of Obstetricians and GynecologistsTask Force on Hypertension in Pregnancy.  Hypertension in pregnancy: report of the American College of Obstetricians and Gynecologists’ Task Force on Hypertension in Pregnancy. Obstet Gynecol. 2013;122(5):1122-1131.
PubMedArticle
2.
Abalos  E, Cuesta  C, Grosso  AL, Chou  D, Say  L.  Global and regional estimates of preeclampsia and eclampsia: a systematic review. Eur J Obstet Gynecol Reprod Biol. 2013;170(1):1-7.
PubMedArticle
3.
Steegers  EA, von Dadelszen  P, Duvekot  JJ, Pijnenborg  R.  Pre-eclampsia. Lancet. 2010;376(9741):631-644.
PubMedArticle
4.
Bello  N, Rendon  ISH, Arany  Z.  The relationship between pre-eclampsia and peripartum cardiomyopathy: a systematic review and meta-analysis. J Am Coll Cardiol. 2013;62(18):1715-1723.
PubMedArticle
5.
Elkayam  U.  Clinical characteristics of peripartum cardiomyopathy in the United States: diagnosis, prognosis, and management. J Am Coll Cardiol. 2011;58(7):659-670.
PubMedArticle
6.
Sliwa  K, Fett  J, Elkayam  U.  Peripartum cardiomyopathy. Lancet. 2006;368(9536):687-693.
PubMedArticle
7.
Brown  MC, Best  KE, Pearce  MS, Waugh  J, Robson  SC, Bell  R.  Cardiovascular disease risk in women with pre-eclampsia: systematic review and meta-analysis. Eur J Epidemiol. 2013;28(1):1-19.
PubMedArticle
8.
McDonald  SD, Malinowski  A, Zhou  Q, Yusuf  S, Devereaux  PJ.  Cardiovascular sequelae of preeclampsia/eclampsia: a systematic review and meta-analyses. Am Heart J. 2008;156(5):918-930.
PubMedArticle
9.
Lykke  JA, Langhoff-Roos  J, Sibai  BM, Funai  EF, Triche  EW, Paidas  MJ.  Hypertensive pregnancy disorders and subsequent cardiovascular morbidity and type 2 diabetes mellitus in the mother. Hypertension. 2009;53(6):944-951.
PubMedArticle
10.
Wikström  A-K, Haglund  B, Olovsson  M, Lindeberg  SN.  The risk of maternal ischaemic heart disease after gestational hypertensive disease. BJOG. 2005;112(11):1486-1491.
PubMedArticle
11.
Bellamy  L, Casas  J-P, Hingorani  AD, Williams  DJ.  Pre-eclampsia and risk of cardiovascular disease and cancer in later life: systematic review and meta-analysis. BMJ. 2007;335(7627):974.
PubMedArticle
12.
Melchiorre  K, Sutherland  GR, Liberati  M, Thilaganathan  B.  Preeclampsia is associated with persistent postpartum cardiovascular impairment. Hypertension. 2011;58(4):709-715.
PubMedArticle
13.
Melchiorre  K, Sutherland  GR, Baltabaeva  A, Liberati  M, Thilaganathan  B.  Maternal cardiac dysfunction and remodeling in women with preeclampsia at term. Hypertension. 2011;57(1):85-93.
PubMedArticle
14.
Powe  CE, Levine  RJ, Karumanchi  SA.  Preeclampsia, a disease of the maternal endothelium: the role of antiangiogenic factors and implications for later cardiovascular disease. Circulation. 2011;123(24):2856-2869.
PubMedArticle
15.
Patten  IS, Rana  S, Shahul  S,  et al.  Cardiac angiogenic imbalance leads to peripartum cardiomyopathy. Nature. 2012;485(7398):333-338.
PubMedArticle
16.
Tranquilli  AL, Dekker  G, Magee  L,  et al.  The classification, diagnosis and management of the hypertensive disorders of pregnancy: A revised statement from the ISSHP. Pregnancy Hypertens. 2014;4(2):97-104.
PubMed
17.
Duckitt  K, Harrington  D.  Risk factors for pre-eclampsia at antenatal booking: systematic review of controlled studies. BMJ. 2005;330(7491):565.
PubMedArticle
18.
Elliott  P, Andersson  B, Arbustini  E,  et al.  Classification of the cardiomyopathies: a position statement from the European Society of Cardiology Working Group on Myocardial and Pericardial Diseases. Eur Heart J. 2008;29(2):270-276.
PubMedArticle
19.
Lange  T, Vansteelandt  S, Bekaert  M.  A simple unified approach for estimating natural direct and indirect effects. Am J Epidemiol. 2012;176(3):190-195.
PubMedArticle
20.
Mosca  L, Benjamin  EJ, Berra  K,  et al; American Heart Association.  Effectiveness-based guidelines for the prevention of cardiovascular disease in women—2011 update: a guideline from the American Heart Association. J Am Coll Cardiol. 2011;57(12):1404-1423.
PubMedArticle
21.
Therneau  TM, Grambsch  PM, Fleming  TR.  Martingale-based residuals for survival models. Biometrika. 1990;77(1):147-160.Article
22.
Melchiorre  K, Sharma  R, Thilaganathan  B.  Cardiovascular implications in preeclampsia: an overview. Circulation. 2014;130(8):703-714.
PubMedArticle
23.
Melchiorre  K, Thilaganathan  B.  Maternal cardiac function in preeclampsia. Curr Opin Obstet Gynecol. 2011;23(6):440-447.
PubMedArticle
24.
Melchiorre  K, Sutherland  G, Sharma  R, Nanni  M, Thilaganathan  B.  Mid-gestational maternal cardiovascular profile in preterm and term pre-eclampsia: a prospective study. BJOG. 2013;120(4):496-504.
PubMedArticle
25.
Levine  RJ, Maynard  SE, Qian  C,  et al.  Circulating angiogenic factors and the risk of preeclampsia. N Engl J Med. 2004;350(7):672-683.
PubMedArticle
26.
Levine  RJ, Lam  C, Qian  C,  et al; CPEP Study Group.  Soluble endoglin and other circulating antiangiogenic factors in preeclampsia. N Engl J Med. 2006;355(10):992-1005.
PubMedArticle
27.
Chaiworapongsa  T, Romero  R, Kim  YM,  et al.  Plasma soluble vascular endothelial growth factor receptor-1 concentration is elevated prior to the clinical diagnosis of pre-eclampsia. J Matern Fetal Neonatal Med. 2005;17(1):3-18.
PubMedArticle
28.
Chaiworapongsa  T, Romero  R, Espinoza  J,  et al.  Evidence supporting a role for blockade of the vascular endothelial growth factor system in the pathophysiology of preeclampsia. Am J Obstet Gynecol. 2004;190(6):1541-1547.
PubMedArticle
29.
Noori  M, Donald  AE, Angelakopoulou  A, Hingorani  AD, Williams  DJ.  Prospective study of placental angiogenic factors and maternal vascular function before and after preeclampsia and gestational hypertension. Circulation. 2010;122(5):478-487.
PubMedArticle
30.
Rana  S, Powe  CE, Salahuddin  S,  et al.  Angiogenic factors and the risk of adverse outcomes in women with suspected preeclampsia. Circulation. 2012;125(7):911-919.
PubMedArticle
31.
ACOG Committee on Practice Bulletins—Obstetrics.  ACOG practice bulletin: diagnosis and management of preeclampsia and eclampsia. Obstet Gynecol. 2002;99(1):159-167.
PubMedArticle
32.
Klemmensen  AK, Olsen  SF, Osterdal  ML, Tabor  A.  Validity of preeclampsia-related diagnoses recorded in a national hospital registry and in a postpartum interview of the women. Am J Epidemiol. 2007;166(2):117-124.
PubMedArticle
33.
Wilkins-Haug  L, Celi  A, Thomas  A, Frolkis  J, Seely  EW.  Recognition by women’s health care providers of long-term cardiovascular disease risk after preeclampsia. Obstet Gynecol. 2015;125(6):1287-1292.
PubMedArticle
Original Investigation
March 8, 2016

Association Between Hypertensive Disorders of Pregnancy and Later Risk of Cardiomyopathy

Author Affiliations
  • 1Department of Epidemiology Research, Statens Serum Institut, Copenhagen, Denmark
  • 2Department of Obstetrics, Copenhagen University Hospital (Rigshospitalet), Copenhagen, Denmark
  • 3Unit for Inherited Cardiac Diseases, The Heart Centre, Copenhagen University Hospital (Rigshospitalet), Copenhagen, Denmark
  • 4Department of Clinical Medicine, Copenhagen University, Copenhagen, Denmark
  • 5Department of Medicine, Stanford University School of Medicine, Stanford, California
JAMA. 2016;315(10):1026-1033. doi:10.1001/jama.2016.1869
Abstract

Importance  Women with hypertensive disorders of pregnancy, preeclampsia in particular, have an increased risk of cardiomyopathy during the peripartum period. Whether hypertensive disorders of pregnancy are also associated with cardiomyopathy later in life is unknown.

Objective  To determine whether hypertensive disorders of pregnancy are associated with cardiomyopathy beyond the peripartum period.

Design, Setting, and Participants  Nationwide register–based cohort study using Cox regression to compare rates of cardiomyopathy in women with and without a history of hypertensive disorders of pregnancy in a cohort of 1 075 763 women with at least 1 pregnancy ending in live birth or stillbirth in Denmark, 1978-2012, with follow-up through December 31, 2012.

Exposures  A hypertensive disorder of pregnancy (severe or moderate preeclampsia or gestational hypertension) registered in the National Patient Register.

Main Outcomes and Measures  Cardiomyopathy more than 5 months after delivery (outside the peripartum period) up to 34 years 7 months.

Result  The women in the primary cohort had 2 067 633 eligible pregnancies during the study period, 76 108 of which were complicated by a hypertensive disorder of pregnancy. During follow-up, 1577 women (mean age, 48.5 years at cardiomyopathy diagnosis; 2.6% with multiple pregnancies) developed cardiomyopathy. Compared with women with normotensive pregnancies (18 211 603 person-years of follow-up; n = 1408 cardiomyopathy events, 7.7/100 000 person-years [95% CI, 7.3-8.2]), women with a history of hypertensive disorders of pregnancy had significantly increased rates of cardiomyopathy (in 173 062 person-years of follow-up among women with severe preeclampsia, n = 27 cardiomyopathy events; 15.6/100 000 person-years [95% CI, 10.7-22.7]; adjusted hazard ratio [HR], 2.20 [95% CI, 1.50-3.23]; in 697 447 person-years of follow-up among women with moderate preeclampsia, n = 102 cardiomyopathy events; 14.6/100 000 person-years [95% CI, 12.0-17.8]; adjusted HR, 1.89 [95% CI, 1.55-2.23]; in 213 197 person-years of follow-up among women with gestational hypertension, n = 40 cardiomyopathy events; 17.3/100 000 person-years [95% CI, 12.7-23.6]; adjusted HR, 2.06 [95% CI, 1.50-2.82]). These increases persisted more than 5 years after the latest pregnancy. Mediation analyses suggested that only about 50% of the association was an indirect association through postpregnancy chronic hypertension. In this cohort, 11% of all cardiomyopathy events occurred in women with a history of hypertensive disorders of pregnancy.

Conclusions and Relevance  Women with a history of hypertensive disorders of pregnancy, compared with women without such a history, had a small but statistically significant increased risk of cardiomyopathy more than 5 months after delivery. Further research is necessary to understand whether there is a causal mechanism behind this association.

Introduction

Hypertensive disorders of pregnancy (HDP), which include preeclampsia and gestational hypertension, are characterized by de novo hypertension, with or without abnormal biochemical findings, in the second half of pregnancy and occur in up to 10% of pregnancies worldwide.1,2 In severe cases, preeclampsia can lead to multiple organ failure, seizures (eclampsia), and fetal and maternal death.3

Women with preeclampsia have a greatly increased risk of cardiomyopathy in the peripartum period (the last month of pregnancy until 5 months after delivery).46 Peripartum cardiomyopathy is an idiopathic pregnancy-related cardiomyopathy often characterized by severely reduced myocardial contractility and symptoms of heart failure.6 Preeclampsia is also associated with cardiovascular disease later in life; women with a history of preeclampsia have 3 to 4 times the risk of hypertension and twice the risk of ischemic heart disease and cerebrovascular disease years after the affected pregnancy,711 risks that may increase with preeclampsia severity.810 Whether women with a history of preeclampsia also have an increased risk of cardiomyopathy outside the peripartum period is unknown, but recent findings of persistent cardiac dysfunction and remodeling after preeclampsia suggest that such an association is plausible.12,13 The etiology of preeclampsia and the link between preeclampsia and cardiovascular disease are poorly understood, but shared underlying mechanisms may be involved in the pathophysiology of preeclampsia and pregnancy-related cardiac/cardiovascular damage.14,15

It was hypothesized that women with a history of HDP are at increased risk of cardiomyopathy not only in the peripartum period but also years after pregnancy and that the magnitude of the association increases with the severity of the HDP. A nationwide register-based cohort study was conducted to investigate the association between HDP and cardiomyopathy after the peripartum period.

Methods
Study Cohort

Using the National Patient Register and the Medical Birth Register (see eMethods in the Supplement), all women in Denmark with at least 1 pregnancy ending in live birth or stillbirth between 1978 and 2012 were identified (Figure). Pregnancies with gestational length less than 20 weeks and women registered with any cardiovascular disease (including hypertension) or diabetes mellitus more than 1 month before their first registered delivery were excluded (Figure; eMethods in the Supplement). The study was approved by the Danish Data Protection Agency; neither informed consent nor ethics committee approval are required for strictly register-based research.

Hypertensive Disorders of Pregnancy (Exposure)

A woman was considered to have an HDP in a given pregnancy if she was registered with gestational hypertension, moderate preeclampsia or severe preeclampsia (including eclampsia and HELLP syndrome [hemolysis, elevated liver enzymes, low platelet count]) in the National Patient Register any time between 1 month before delivery and 7 days postpartum. By definition, gestational hypertension and preeclampsia involve de novo hypertension in a pregnant woman, with onset after 20 weeks’ gestation.1,16 As registered in the National Patient Register, gestational hypertension is defined as hypertension without accompanying proteinuria, whereas in moderate preeclampsia, mild to moderate hypertension is accompanied by proteinuria. Severe preeclampsia fulfills the criteria for moderate preeclampsia, with the addition of severe hypertension and/or severe proteinuria and/or signs of organ failure (which can include the HELLP syndrome) and/or generalized seizures (eclampsia) (see eMethods in the Supplement). To try to ensure that HDP diagnoses reflected true cases, women whose only HDP diagnoses were registered more than 1 month before or more than 7 days after delivery were not considered to have had an HDP.

HDP status was handled as a time-dependent variable. A woman could contribute person-time to several exposure groups, changing her exposure status if a given pregnancy was more complicated (in terms of HDP) than a previous pregnancy (but not if the pregnancy was less complicated—she could only become more severely affected) (eMethods and eFigure in the Supplement).

In sensitivity analyses, classification of preeclampsia based on gestational length at delivery was also investigated. Preeclampsia (any registered diagnosis code, regardless of severity) was classified as early preterm if delivery occurred at less than 34 completed weeks’ gestation and late preterm/term when delivery occurred at 34 or more completed weeks’ gestation.

Cardiomyopathy (Outcome)

A woman was considered to have cardiomyopathy after the peripartum period if she was registered in the National Patient Register or Causes of Death Register with cardiomyopathy (eMethods in the Supplement) more than 5 months after delivery, in the absence of peripartum cardiomyopathy, up to 34 years 7 months of follow-up. Subanalyses then focused specifically on dilated cardiomyopathy, the cardiomyopathy phenotype typically seen in the peripartum period. Furthermore, since some cases of cardiomyopathy could have been coded as heart failure, particularly before echocardiography became a widely used diagnostic tool (in the mid-1990s), parallel supplemental analyses for heart failure were conducted (eMethods in the Supplement).

Covariates

Two groups of covariates were considered: variables considered a priori to be potential confounders (birth year, age, smoking, parity, multiple pregnancy, stillbirth) and variables that might be confounders but might also be intermediates linking HDP and cardiomyopathy or competing causes of cardiomyopathy (postpregnancy diabetes, ischemic heart disease, hypertension, obesity) (eMethods in the Supplement). The latter are all associated with HDP,3,9,17 and associations with cardiomyopathy are also plausible. Women registered with cardiovascular disease or diabetes before their first pregnancy in the study period were excluded. Obesity and incident postpartum diabetes were judged to be true potential confounders. Although the reduced left ventricular systolic function often caused by ischemic heart disease is an entity separate from primary cardiomyopathy and should not be coded as cardiomyopathy,18 such miscoding may occur, making ischemic heart disease a competing cause of cardiomyopathy in practice. Because hypertension developing postpartum is likely either an intermediate linking HDP and cardiomyopathy or an underlying factor common to shared pathophysiologic processes, hypertension was treated as a potential mediator in mediation analyses.

Statistical Analyses

Women were followed up from 5 months after their first delivery in the study period to the first of the following events: (1) cardiomyopathy; (2) death; (3) emigration; (4) designated “missing” in the Civil Registration System; or (5) December 31, 2012 (the end of follow-up). Women with more than 1 pregnancy during the follow-up period contributed follow-up time during and after all pregnancies subsequent to the first, but the peripartum time associated with any subsequent pregnancies (1 month before delivery to 5 months after delivery) was excluded from the analyses. Women who developed peripartum cardiomyopathy (cardiomyopathy in the peripartum period) were censored and did not contribute further follow-up time.

Cox proportional hazards modeling with the woman’s age as the underlying time scale was used to estimate hazard ratios (HRs) comparing rates of cardiomyopathy for women with a history of HDP and women with normotensive pregnancies. All estimates were adjusted for maternal birth year (1-year intervals), parity (1, 2, ≥3), multiple pregnancy (yes/no) and stillbirth (yes/no), by stratifying the baseline hazards in the Cox models on these variables; in this way, women with a history of HDP were compared with normotensive women of the same age, birth year, and reproductive history. Subanalyses focusing on dilated cardiomyopathy were conducted in a subcohort with follow-up from 1994 (when registration of cardiomyopathy subtypes began).

Primary results are presented unadjusted for postpregnancy diabetes and ischemic heart disease and ignoring the possible mediating effect of hypertension. However, adjusting for diabetes diagnosed after the first registered pregnancy and stopping follow-up (censoring) at the time of any ischemic heart disease diagnosis, and adjusting for smoking in a subcohort with smoking information, were also explored. (Adjusting for body mass index proved impossible, as this information was only available for women delivering in or after 2004.) Furthermore, mediation analyses were performed to estimate the degree to which an association between HDP and cardiomyopathy might be related to postpregnancy hypertension, by extending the approach of Lange et al19 (eMethods in the Supplement). Hazard ratios for the association between HDP and cardiomyopathy in women with known or possible pregestational hypertension were also estimated (eMethods in the Supplement). To evaluate the potential for surveillance bias, a sensitivity analysis was performed in which follow-up ended in 2005, thereby excluding later years during which women with a history of HDP might have been more closely monitored for cardiovascular disease than other parous women because of increasing awareness of the link between preeclampsia and cardiovascular disease.711,20

Potential violations of the proportional hazards assumption were checked by plotting cumulative Martingale residuals against maternal age (the underlying time scale in the Cox models).21 All analyses were performed using SAS version 9.4 (SAS Institute Inc). P < .05 (2-sided) was considered statistically significant.

Results

The primary cohort consisted of 1 075 763 women with 2 067 633 pregnancies ending in live birth or stillbirth during the study period. Of these pregnancies, 12 974 were complicated by severe preeclampsia, 44 711 by moderate preeclampsia, and 18 423 by gestational hypertension. The cohort was followed up for 19.3 million person-years (mean, 17.9 years per woman). Of the 1577 women who developed cardiomyopathy during follow-up, 169 (10.7%) had a history of HDP. A total of 26 945 women (2.5%) were lost to follow-up; the majority of these women (26 399) moved abroad. Table 1 reports characteristics of each woman’s first pregnancy in the study period by HDP status; eTable 1 in the Supplement reports these characteristics for women lost to follow-up. During follow-up, 19 (0.20%) women with severe preeclampsia, 84 (0.26%) with moderate preeclampsia, 29 (0.25%) with gestational hypertension, and 1445 (0.14%) with no HDP in their first pregnancy developed cardiomyopathy.

Comparing rates of cardiomyopathy for women with a history of HDP and rates for women with a history of only normotensive pregnancies yielded HRs of 2.20 (95% CI, 1.50-3.23) for severe preeclampsia, 1.89 (95% CI, 1.55-2.32) for moderate preeclampsia, and 2.06 (95% CI, 1.50-2.82) for gestational hypertension (Table 2). Even more than 5 years after a woman’s latest pregnancy in the study period, HRs were 2.22 (95% CI, 1.47-3.36) for severe preeclampsia, 1.86 (95% CI, 1.50-2.30) for moderate preeclampsia, and 2.25 (95% CI, 1.63-3.09) for gestational hypertension (Table 2). When the analyses were stratified by current age younger than 45 years and 45 years or older, the HRs for the 2 age groups were not statistically significantly different (P = .09) (9.2 cardiomyopathy events/100 000 person-years [95% CI, 7.3-11.6] among women <45 years with a history of HDP and 3.6/100 000 person-years [95% CI, 3.3-4.0] among women <45 years with no history of HDP; HR, 2.35 [95% CI, 1.83-3.01]; 30.4 cardiomyopathy events/100 000 person-years [95% CI, 24.9-37.1] among women ≥45 years with a history of HDP and 17.8/100 000 person-years [95% CI, 16.7-19.0] among women ≥45 years with no history of HDP; HR, 1.77 [95% CI, 1.43-2.18]). Limiting the focus to dilated cardiomyopathy produced similar results (Table 3); wider confidence intervals reflect the reduced number of outcomes in this subcohort. When women with early preterm preeclampsia (16.3 cardiomyopathy events/100 000 person-years [95% CI, 8.2-32.6]) and late preterm/term preeclampsia (14.7 cardiomyopathy events/100 000 person-years [95% CI, 12.2-17.7]) were compared with women with no history of HDP (7.6 cardiomyopathy events/100 000 person-years [95% CI, 7.2-8.0]), the respective HRs were 2.29 (95% CI, 1.14-4.60) and 1.96 (95% CI, 1.61-2.38).

Adjusting for diabetes diagnosed during the follow-up period, and stopping follow-up if ischemic heart disease was diagnosed, did not change the magnitudes of the observed associations (Table 4). Further adjustment for smoking also did not affect the strength of the estimates (eTable 2 in the Supplement). Considering only HDP status in a woman’s first pregnancy in the study period also produced similar results (eTable 3 in the Supplement), as did ending follow-up in 2005 (eTable 4 in the Supplement) (although in the latter analysis, the association with severe preeclampsia was no longer statistically significant).

Mediation analyses performed to estimate how much of the observed association between HDP and cardiomyopathy was not related to an indirect association through postgestational hypertension showed a significant direct association with HDP, estimated to be 49% (95% CI, 29%-59%) of the total observed association (eResults in the Supplement). Among women with known or possible pregestational hypertension excluded from the mediation analyses (n = 23 810), 78 developed cardiomyopathy during 321 608 person-years of follow-up (35.2 [95% CI, 23.2-53.5] vs 21.6 [95% CI, 16.6-28.1] per 100 000 person-years), yielding an HR for cardiomyopathy by history of HDP of 1.72 (95% CI, 1.03-2.84).

Similar patterns were observed for heart failure (eTable 5 in the Supplement). Hazard ratios comparing heart failure rates for women with a history of HDP and rates for women with normotensive pregnancies were 2.23 (95% CI, 1.71-2.89) for severe preeclampsia, 1.86 (95% CI, 1.63-2.12) for moderate preeclampsia, and 2.07 (95% CI, 1.70-2.53) for gestational hypertension (eTable 5 in the Supplement).

Discussion

In this study, HDPs were associated with a lasting (>5 years postpartum) increase in risk of cardiomyopathy beyond the peripartum period, regardless of HDP severity. This increase in risk appeared to be independent of ischemic heart disease, the risk of which is increased among women with a history of preeclampsia,7,9 and approximately 50% of the risk was not associated with postgestational hypertension. The results suggest that links with HDP might be associated with a substantial proportion of idiopathic cardiomyopathy cases in women; in this cohort, 11% of all cardiomyopathy events in parous women occurred among women with a history of HDP.

However, cardiomyopathy events are rare, even among women in this study with a history of HDP (rates in women with a history of HDP, based on our data, were 14.6-17.3 cases/100 000 person-years). Accordingly, even though there was an association between HDP and increased risk of cardiomyopathy, the absolute risk was small. Although the American Heart Association20 recommends monitoring for ischemic heart disease following preeclampsia, a similar recommendation would not be justified for a more rare condition such as cardiomyopathy, regardless of the strength of the observed association with HDP. On the other hand, it seems prudent to suggest that physicians consider this association in the diagnostic workup of women with a history of HDP presenting with possible symptoms of heart failure.

Shared underlying mechanisms could explain the association between HDP and cardiomyopathy.21 The pathological processes implicated in preeclampsia—angiogenic imbalance, complement activation, inflammation, hemodynamic changes—likely also contribute directly to cardiac stress exceeding that of normal pregnancy, producing overt cardiac damage in some women.14,15,22 Recent evidence of altered cardiac structure and function (predominantly left ventricular remodeling and diastolic dysfunction) during preeclamptic pregnancies and in the years immediately thereafter12,13,2224 supports both this contention and the biological plausibility of the observed associations. Patten et al15 demonstrated that exposing mice to an excess of the antiangiogenic factor soluble fms-like tyrosine kinase-1 (sFlt-1), levels of which are higher in women with preeclampsia than in women with normotensive pregnancies,2529 caused cardiac dysfunction in wild-type mice and profound cardiomyopathy in knockout mice with hearts unable to withstand antiangiogenic insult. In women with preeclampsia, sFlt1 levels were correlated with cardiac diastolic dysfunction.15 In women whose genetic make-up or lifestyle already renders them susceptible to cardiac disease, HDP may therefore provide an additional cardiac stressor, at which time the most severely affected women develop overt peripartum cardiomyopathy (which is characterized not just by diastolic dysfunction but by systolic dysfunction), while in other women, asymptomatic diastolic dysfunction may progress to heart failure, cardiomyopathy, or both over time.15,22 In women without pregestational cardiac susceptibility, cardiac diastolic dysfunction subsequent to HDP may increase susceptibility to later insults.

Because Melchiorre et al12 also found the most severe and persistent cardiac dysfunction in women with preterm preeclampsia, and changes in antiangiogenic factor levels are more marked in early preeclampsia,25,26,29,30 a stronger association for early preterm preeclampsia than for late preterm/term preeclampsia was expected, but this was not the case. However, in women who do not immediately develop peripartum cardiomyopathy, the risk of later cardiomyopathy may depend more on the ability to remodel cardiac damage, other cardiac risk factors, or the nature of the later postpregnancy cardiac stressors than on the degree of cardiac dysfunction or susceptibility at the end of pregnancy. Alternatively, the late preeclampsia group included many women with severe preeclampsia, which would have reduced the differences between the 2 preeclampsia groups.

Cardiomyopathy occurring after the peripartum period might be suspected simply to be unrecognized peripartum cardiomyopathy, the symptoms of which were ascribed to pregnancy or the aftermath of delivery. However, more than 80% percent of peripartum cardiomyopathy events in this cohort were diagnosed within 1 month of birth. Furthermore, the associations with cardiomyopathy persisted more than 5 years after an affected pregnancy, which argues for a distinct, persistent risk of cardiomyopathy outside the peripartum period.

Although treated as such in the primary analyses, cardiomyopathy is not a single entity, and the risk associated with HDP may apply only to specific cardiomyopathy subtypes. However, when the analyses were restricted to dilated cardiomyopathy, the results did not differ substantially, suggesting that the associations may primarily be driven by dilated cardiomyopathy.

The degree to which the findings in this study were related to unrecognized pregestational hypertension and chronic hypertension subsequent to HDP is critical to the interpretation of the results. Women with known pregestational hypertension were excluded from the study, but some of the women included in the cohort might have had unrecognized pregestational hypertension. However, bias from this source is unlikely, because similar associations between HDP and cardiomyopathy were observed for women with known or possible pregestational hypertension. Assuming that initiation of medication use is a good proxy for the postgestational development of chronic hypertension, mediation analysis results suggested that approximately 50% of the association between HDP and cardiomyopathy was associated with postgestational hypertension. However, the remaining 50% was not associated with hypertension and could be directly attributable to HDP (or an underlying common cause).

This study has several strengths and several potential limitations. Klemmensen et al validated HDP codes in the National Patient Register against the American Congress of Obstetricians and Gynecologists’ 2002 definitions31 and found that while the sensitivity of the register is moderate (69%) for preeclampsia (all types combined) and low (10%) for gestational hypertension, its specificity for HDP is very high (>99%).32 Therefore, although not all HDP diagnoses are registered, those that are registered are likely correctly registered. Because of this high specificity, any bias attributable to misclassification of HDP-affected pregnancies as normotensive is most likely negligible. The register’s low sensitivity for gestational hypertension is likely a consequence of the management of the condition by general practitioners, who do not report to the National Patient Register, and underreporting by hospital obstetricians of cases that do not warrant clinical intervention; in this study, women with gestational hypertension may have resembled women with moderate preeclampsia more than average women with gestational hypertension.

Cardiomyopathy diagnoses in the National Patient Register have not been validated. Because these diagnoses are only assigned following clinical workup, their specificity should be high. However, the register’s sensitivity for cardiomyopathy may be low, because asymptomatic cardiomyopathy may go undetected or the condition may be mistaken for other conditions presenting with the same symptoms (eg, asthma, overweight, poor cardiovascular fitness). Parallel analyses of the associations between HDP and heart failure showed results consistent with those obtained for cardiomyopathy, suggesting that some misclassification of cardiomyopathy as heart failure probably did occur but that this misclassification did not affect the conclusions regarding the strength and direction of the observed associations.

Adjustment for diabetes and smoking did not affect the results, suggesting that diabetes and smoking were not important confounders of the association between HDP and cardiomyopathy. In sensitivity analyses in which a woman’s contribution to the study ended if she developed ischemic heart disease, the results did not change appreciably, suggesting that the relationship between HDP and cardiomyopathy was independent of ischemic heart disease, and any association of cardiovascular disease risk factors (eg, high cholesterol levels) with cardiomyopathy via ischemic heart disease likely did not explain the findings either. Adjusting for body mass index would have been valuable but was not possible, because only 9 years of follow-up including this variable were available.

Recent increases in awareness of the link between HDP and later ischemic heart disease7,20 could have resulted in heightened monitoring of, and increased detection of cardiomyopathy in, women with prior HDP. However, awareness of this link may not be common knowledge among physicians, as suggested by a single-institution study conducted in 2012-2013,33 despite the recent inclusion of previous preeclampsia as a risk factor for later ischemic heart disease in American Heart Association guidelines.20 Results of a sensitivity analysis that excluded years when surveillance was potentially heightened did not differ from the main results, making it unlikely that surveillance bias explained the results.

Conclusions

Women with a history of hypertensive disorders of pregnancy, compared with women without such a history, had a small but statistically significant increased risk of cardiomyopathy more than 5 months after delivery. Further research is necessary to understand whether there is a causal mechanism behind this association.

Back to top
Article Information

Corresponding Author: Ida Behrens, MD, Department of Epidemiology Research, Statens Serum Institut, Artillerivej 5, 2300 Copenhagen S, Denmark (idbe@ssi.dk).

Author Contributions: Drs Behrens and Boyd had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: All authors.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Behrens.

Critical revision of the manuscript for important intellectual content: Basit, Lykke, Ranthe, Wohlfahrt, Bundgaard, Melbye, Boyd.

Statistical analysis: Basit, Wohlfahrt.

Obtained funding: Behrens, Ranthe, Boyd.

Study supervision: Lykke, Ranthe, Wohlfahrt, Bundgaard, Melbye, Boyd.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Ranthe reported that at the time of manuscript submission he was an employee of Novo Nordisk A/S, Global Development, Medical and Science, Søborg, Denmark. Dr Bundgaard reported receiving lecture fees from AstraZeneca, Sanofi, Shire, Pfizer, and Merck Sharp & Dohme. None of the other authors reported disclosures.

Funding/Support: This study was funded by the Danish Heart Association and the Danish Council for Independent Research.

Role of Funders/Sponsors: The Danish Heart Association and the Danish Council for Independent Research had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or the decision to submit the manuscript for publication.

References
1.
American College of Obstetricians and GynecologistsTask Force on Hypertension in Pregnancy.  Hypertension in pregnancy: report of the American College of Obstetricians and Gynecologists’ Task Force on Hypertension in Pregnancy. Obstet Gynecol. 2013;122(5):1122-1131.
PubMedArticle
2.
Abalos  E, Cuesta  C, Grosso  AL, Chou  D, Say  L.  Global and regional estimates of preeclampsia and eclampsia: a systematic review. Eur J Obstet Gynecol Reprod Biol. 2013;170(1):1-7.
PubMedArticle
3.
Steegers  EA, von Dadelszen  P, Duvekot  JJ, Pijnenborg  R.  Pre-eclampsia. Lancet. 2010;376(9741):631-644.
PubMedArticle
4.
Bello  N, Rendon  ISH, Arany  Z.  The relationship between pre-eclampsia and peripartum cardiomyopathy: a systematic review and meta-analysis. J Am Coll Cardiol. 2013;62(18):1715-1723.
PubMedArticle
5.
Elkayam  U.  Clinical characteristics of peripartum cardiomyopathy in the United States: diagnosis, prognosis, and management. J Am Coll Cardiol. 2011;58(7):659-670.
PubMedArticle
6.
Sliwa  K, Fett  J, Elkayam  U.  Peripartum cardiomyopathy. Lancet. 2006;368(9536):687-693.
PubMedArticle
7.
Brown  MC, Best  KE, Pearce  MS, Waugh  J, Robson  SC, Bell  R.  Cardiovascular disease risk in women with pre-eclampsia: systematic review and meta-analysis. Eur J Epidemiol. 2013;28(1):1-19.
PubMedArticle
8.
McDonald  SD, Malinowski  A, Zhou  Q, Yusuf  S, Devereaux  PJ.  Cardiovascular sequelae of preeclampsia/eclampsia: a systematic review and meta-analyses. Am Heart J. 2008;156(5):918-930.
PubMedArticle
9.
Lykke  JA, Langhoff-Roos  J, Sibai  BM, Funai  EF, Triche  EW, Paidas  MJ.  Hypertensive pregnancy disorders and subsequent cardiovascular morbidity and type 2 diabetes mellitus in the mother. Hypertension. 2009;53(6):944-951.
PubMedArticle
10.
Wikström  A-K, Haglund  B, Olovsson  M, Lindeberg  SN.  The risk of maternal ischaemic heart disease after gestational hypertensive disease. BJOG. 2005;112(11):1486-1491.
PubMedArticle
11.
Bellamy  L, Casas  J-P, Hingorani  AD, Williams  DJ.  Pre-eclampsia and risk of cardiovascular disease and cancer in later life: systematic review and meta-analysis. BMJ. 2007;335(7627):974.
PubMedArticle
12.
Melchiorre  K, Sutherland  GR, Liberati  M, Thilaganathan  B.  Preeclampsia is associated with persistent postpartum cardiovascular impairment. Hypertension. 2011;58(4):709-715.
PubMedArticle
13.
Melchiorre  K, Sutherland  GR, Baltabaeva  A, Liberati  M, Thilaganathan  B.  Maternal cardiac dysfunction and remodeling in women with preeclampsia at term. Hypertension. 2011;57(1):85-93.
PubMedArticle
14.
Powe  CE, Levine  RJ, Karumanchi  SA.  Preeclampsia, a disease of the maternal endothelium: the role of antiangiogenic factors and implications for later cardiovascular disease. Circulation. 2011;123(24):2856-2869.
PubMedArticle
15.
Patten  IS, Rana  S, Shahul  S,  et al.  Cardiac angiogenic imbalance leads to peripartum cardiomyopathy. Nature. 2012;485(7398):333-338.
PubMedArticle
16.
Tranquilli  AL, Dekker  G, Magee  L,  et al.  The classification, diagnosis and management of the hypertensive disorders of pregnancy: A revised statement from the ISSHP. Pregnancy Hypertens. 2014;4(2):97-104.
PubMed
17.
Duckitt  K, Harrington  D.  Risk factors for pre-eclampsia at antenatal booking: systematic review of controlled studies. BMJ. 2005;330(7491):565.
PubMedArticle
18.
Elliott  P, Andersson  B, Arbustini  E,  et al.  Classification of the cardiomyopathies: a position statement from the European Society of Cardiology Working Group on Myocardial and Pericardial Diseases. Eur Heart J. 2008;29(2):270-276.
PubMedArticle
19.
Lange  T, Vansteelandt  S, Bekaert  M.  A simple unified approach for estimating natural direct and indirect effects. Am J Epidemiol. 2012;176(3):190-195.
PubMedArticle
20.
Mosca  L, Benjamin  EJ, Berra  K,  et al; American Heart Association.  Effectiveness-based guidelines for the prevention of cardiovascular disease in women—2011 update: a guideline from the American Heart Association. J Am Coll Cardiol. 2011;57(12):1404-1423.
PubMedArticle
21.
Therneau  TM, Grambsch  PM, Fleming  TR.  Martingale-based residuals for survival models. Biometrika. 1990;77(1):147-160.Article
22.
Melchiorre  K, Sharma  R, Thilaganathan  B.  Cardiovascular implications in preeclampsia: an overview. Circulation. 2014;130(8):703-714.
PubMedArticle
23.
Melchiorre  K, Thilaganathan  B.  Maternal cardiac function in preeclampsia. Curr Opin Obstet Gynecol. 2011;23(6):440-447.
PubMedArticle
24.
Melchiorre  K, Sutherland  G, Sharma  R, Nanni  M, Thilaganathan  B.  Mid-gestational maternal cardiovascular profile in preterm and term pre-eclampsia: a prospective study. BJOG. 2013;120(4):496-504.
PubMedArticle
25.
Levine  RJ, Maynard  SE, Qian  C,  et al.  Circulating angiogenic factors and the risk of preeclampsia. N Engl J Med. 2004;350(7):672-683.
PubMedArticle
26.
Levine  RJ, Lam  C, Qian  C,  et al; CPEP Study Group.  Soluble endoglin and other circulating antiangiogenic factors in preeclampsia. N Engl J Med. 2006;355(10):992-1005.
PubMedArticle
27.
Chaiworapongsa  T, Romero  R, Kim  YM,  et al.  Plasma soluble vascular endothelial growth factor receptor-1 concentration is elevated prior to the clinical diagnosis of pre-eclampsia. J Matern Fetal Neonatal Med. 2005;17(1):3-18.
PubMedArticle
28.
Chaiworapongsa  T, Romero  R, Espinoza  J,  et al.  Evidence supporting a role for blockade of the vascular endothelial growth factor system in the pathophysiology of preeclampsia. Am J Obstet Gynecol. 2004;190(6):1541-1547.
PubMedArticle
29.
Noori  M, Donald  AE, Angelakopoulou  A, Hingorani  AD, Williams  DJ.  Prospective study of placental angiogenic factors and maternal vascular function before and after preeclampsia and gestational hypertension. Circulation. 2010;122(5):478-487.
PubMedArticle
30.
Rana  S, Powe  CE, Salahuddin  S,  et al.  Angiogenic factors and the risk of adverse outcomes in women with suspected preeclampsia. Circulation. 2012;125(7):911-919.
PubMedArticle
31.
ACOG Committee on Practice Bulletins—Obstetrics.  ACOG practice bulletin: diagnosis and management of preeclampsia and eclampsia. Obstet Gynecol. 2002;99(1):159-167.
PubMedArticle
32.
Klemmensen  AK, Olsen  SF, Osterdal  ML, Tabor  A.  Validity of preeclampsia-related diagnoses recorded in a national hospital registry and in a postpartum interview of the women. Am J Epidemiol. 2007;166(2):117-124.
PubMedArticle
33.
Wilkins-Haug  L, Celi  A, Thomas  A, Frolkis  J, Seely  EW.  Recognition by women’s health care providers of long-term cardiovascular disease risk after preeclampsia. Obstet Gynecol. 2015;125(6):1287-1292.
PubMedArticle
×