Whitehead, Sara J. MD, MPH; Berg, Cynthia J. MD, MPH; Chang, Jeani MPH
Cardiomyopathy in pregnancy and the postpartum period, although rare, accounts for a rising proportion of reported pregnancy-related deaths in the United States. During 1979–1984, 3.0% of reported pregnancy-related deaths were caused by cardiomyopathy; this percentage increased to 7.7% in 1991–1997.1–3 Cardiomyopathy is one of the few causes of pregnancy-related death that has risen since 1979.3
Cardiomyopathy in pregnancy can be divided into two groups. The first group is peripartum cardiomyopathy, first described in 1971.4,5 This is a distinct syndrome of unknown etiology with onset during pregnancy or in the postpartum period. Peripartum cardiomyopathy, defined by a National Institutes of Health consensus conference, is the development of cardiac failure in the last month of pregnancy or within 5 months of delivery, in the absence of both an identifiable cause of cardiac failure and of recognizable heart disease before the last month of pregnancy.6 In addition, echocardiographic findings of left ventricular dysfunction should be present.6,7 The etiology, prevalence, and case-fatality rate for peripartum cardiomyopathy are poorly understood. Risk factors for peripartum cardiomyopathy reported in the literature include black race, older maternal age, high parity, twin and higher-order gestation, and the use of tocolytics during labor.6,8–12 However, none of these studies were population-based.
The second group of cardiomyopathies in pregnancy and the puerperium will be referred to as other cardiomyopathy. This category includes those cardiomyopathies with an onset before late pregnancy or with specific underlying causes. These other cardiomyopathies can have any of the three classically defined pathophysiologic presentations (dilated, restrictive, or hypertrophic) and can be either idiopathic or due to a specific cause (eg, ischemia, hypertension, alcohol, human immunodeficiency virus [HIV] disease, connective tissue diseases, or diabetes).
We used the Pregnancy Mortality Surveillance System from the Division of Reproductive Health at the Centers for Disease Control and Prevention (CDC) to describe the characteristics of and risk factors for pregnancy-related deaths due to cardiomyopathy during 1991–1997 and to examine possible reasons for the trend in deaths due to cardiomyopathy from 1979 through 1997.
MATERIALS AND METHODS
The Pregnancy Mortality Surveillance System includes data from 1979 through 1997 and has been described in detail.3 The offices of vital registration from the 50 states, New York City, and the District of Columbia provided de-identified copies of death certificates for all pregnancy-related deaths. For those deaths that followed a live birth or stillborn fetus, the matching birth or fetal death certificates were also requested. Beginning with deaths occurring in 1991, states were asked to send certificates of all deaths that occurred during or within 1 year of the end of pregnancy, regardless of the cause of death or causal relationship between pregnancy and the death. Because, beginning in 1991, states were asked to send all documentation on all deaths within 1 year of pregnancy, which increased the likelihood of a death due to cardiomyopathy being identified, we elected to focus our analysis of risk factors for these deaths to 1991–1997, the 7-year period with better and more uniform ascertainment.
Clinically experienced epidemiologists (SJW and CJB) reviewed available information on all deaths, including death certificates, handwritten notes, matching live birth or fetal death certificates, pre-existing conditions in the mother, date of delivery, and any other available information. Causes of death and associated conditions were coded according to the system developed by the American College of Obstetricians and Gynecologists/CDC Maternal Mortality Study Group, and International Classification of Diseases (ICD) codes were not used to classify deaths. Deaths were classified as pregnancy-related if they occurred during pregnancy or within 1 year after the end of a pregnancy and if they resulted from 1) complications of pregnancy itself, 2) a chain of events initiated by pregnancy, or 3) aggravation of an unrelated condition by the physiologic effects of pregnancy. This definition differs from the definition of maternal death and the maternal mortality ratio, as used by the National Center for Health Statistics, which is limited to deaths during pregnancy or within 42 days of its end.
Pregnancy-related mortality ratios were calculated as pregnancy-related deaths per 100,000 live births. Numerator data, the number of deaths occurring in the 50 states and the District of Columbia, were obtained from the Pregnancy Mortality Surveillance System. Denominator data, the number of live births that occurred in the 50 states and the District of Columbia, were obtained from public-use natality tapes from the CDC's and Prevention's National Center for Health Statistics. For both the numerator and denominator, race was defined as the race of the mother and categorized as white, black, or other (nonwhite, nonblack).
Death certificates and associated birth or fetal death certificates and any additional information for deaths from 1991 to 1997 for which cardiomyopathy was listed as the cause of death or as an associated condition were manually reviewed by the authors and were determined to be due to cardiomyopathy or another cause. We classified cases as peripartum cardiomyopathy when all three of the following criteria were present: 1) the cause of death was described on the death certificate as “peripartum cardiomyopathy,” “postpartum cardiomyopathy,” or “cardiomyopathy of pregnancy,” 2) the death occurred in the third trimester of pregnancy or in the year postpartum, and 3) there was no indication of specific underlying potentially causal factors, as listed below. Cardiomyopathy deaths were classified as “other cardiomyopathy” when any of the following factors were present: 1) the death occurred in the first or second trimester of pregnancy, 2) the death certificate documented idiopathic dilated cardiomyopathy with onset before pregnancy or with unknown time of onset, 3) the death certificate documented restrictive or hypertrophic cardiomyopathy, or 4) any of the following specific underlying factors were present: hypertension (specifying hypertensive heart disease), congenital cardiac anomaly (excluding mitral valve prolapse), primary pulmonary hypertension, sickle cell disease, thyroid disease, HIV, type I or II diabetes, connective tissue disease, or ischemic heart disease. After review of all available data in some cases, although it was clear the death was due to cardiomyopathy, we were unable to determine whether it was peripartum cardiomyopathy or other cardiomyopathy; these are referred to as cardiomyopathy of unknown cause.
Because parity is not recorded on vital records, we used live birth order as a proxy for parity. Information on live birth order was available only on live birth certificates, so analysis for this variable was limited to deaths occurring after a live birth.
Of the 245 cardiomyopathy deaths during 1991–1997, 171 (70%) were due to peripartum cardiomyopathy (Table 1). Among the 49 (20%) that were classified as other cardiomyopathy, only a small number (ranging from two to ten) were attributable to any specific pathophysiologic presentation or underlying factor. Twenty-five (10%) of cardiomyopathy cases were cardiomyopathy of unknown type and onset time.
The pregnancy outcome was documented in 85% of cases. Among these cases, 89% of deaths occurred after a live birth (96% of peripartum cardiomyopathy deaths and 72% of other cardiomyopathy deaths). The outcome was a stillborn fetus for two peripartum cardiomyopathy deaths (1.4%) and 6 other cardiomyopathy deaths (9.8%). In the remaining cases, 11 women died undelivered, three after an abortion, and one after ectopic pregnancy.
The pregnancy-related mortality ratio due to all cardiomyopathies was 0.88 per 100,000 live births. The pregnancy-related mortality ratio for both peripartum cardiomyopathy and other cardiomyopathy tended to increase with increasing maternal age (Table 2). Among white women, the pregnancy-related mortality ratio showed little change among women aged less than 20–34 years but rose for women aged 35 years and older. Among black women, the pregnancy-related mortality ratio rose steadily with increasing age and increased dramatically for those aged 35 years and older. The pregnancy-related mortality ratio for death from cardiomyopathy was markedly higher for black women than for white women in every age group; overall, a black woman was 6.4 times as likely as a white woman to die of cardiomyopathy. These patterns were found for both peripartum cardiomyopathy and other cardiomyopathy deaths.
The pregnancy-related mortality ratio due to peripartum cardiomyopathy was 0.4 per 100,000 live births among singleton pregnancies and 1.7 for multiple births; this gives a multiple/singleton ratio of 4.1. The pregnancy-related mortality ratio due to other cardiomyopathy was 0.1 for singletons and 0.8 for multiples, an eight-fold difference in risk.
We used live birth order as a proxy for parity. Among cases where the outcome was a live birth, the pregnancy-related mortality ratio for both peripartum cardiomyopathy and other cardiomyopathy showed little change from live birth orders one to three (pregnancy-related mortality ratio 0.5 to 0.4) but did increase with a live birth order of four or greater (pregnancy-related mortality ratio 1.0 for four live births, 1.4 for five or more live births).
We also examined associated medical conditions that were documented for women who died of cardiomyopathy. Among peripartum cardiomyopathy cases, 25 (15%) were noted to have preeclampsia or pregnancy-induced hypertension, and eight (5%) had a pulmonary or cerebrovascular embolism. Among women who died of other or unknown cardiomyopathy, five (7%) had preeclampsia or pregnancy-induced hypertension documented, and one (1%) had a pulmonary embolism.
The interval from the end of pregnancy to death was documented in 73% of cases. Among peripartum cardiomyopathy cases for which the interval was known, 2% died undelivered, 48% died within 42 days postpartum, and 50% died between 43 days and 1 year postpartum (Table 3). The corresponding values among women who died of other cardiomyopathy were 15%, 47%, and 37%. In contrast, of pregnancy-related deaths due to all other causes, only 8.1% of deaths occurred more than 42 days postpartum. Although the number of cardiomyopathy deaths decreased after 26 weeks (183 days), such deaths continued to occur through 1 year postpartum.
An average of ten pregnancy-related deaths due to cardiomyopathy per year were reported for the period 1979–1984, 18 per year for 1985–1990, and 35 per year for 1991–1997. Cardiomyopathy deaths accounted for 3%, 5%, and 8% of all reported pregnancy-related deaths in each of these periods, respectively (Figure 1). This increase over time was similar to the rise in the percentage of pregnancy-related death due to all other causes combined that occurred between 43 days and 1 year from the end of pregnancy, which after the 42-day period is usually considered the puerperium. Deaths from 43 days to 1 year due to all causes combined except cardiomyopathy accounted for 4%, 3%, and 8% of all pregnancy-related deaths in each of the three periods. The cause-specific pregnancy-related mortality ratio due to cardiomyopathy for each of the time periods was 0.29, 0.46, and 0.88 per 100,000 live births. The corresponding pregnancy-related mortality ratio for deaths from 43 to 365 days, due to all causes other than cardiomyopathy, was 0.24, 0.33, and 0.69 for each of the periods.
Pregnancy-related deaths due to cardiomyopathy are a heterogeneous group, with 70% due to peripartum cardiomyopathy and the remaining 30% due to cardiomyopathy associated with a wide range of underlying factors or of unknown cause and time of onset. The racial disparity in pregnancy-related mortality due to cardiomyopathy is striking. Black women are 6.4 times as likely to die from cardiomyopathy as are white women. This increased risk is significantly greater than the 3.7-fold difference in overall risk for pregnancy-related deaths.3 A clear age gradient in the risk of death is also seen; women aged 35 years and older have a risk of death 2.8 times greater than that of women aged 19 years and younger. Twin and higher-order gestations and parity greater than three were also risk factors noted in our work. All of these findings are consistent with previously published studies.6–12
The true prevalence of peripartum cardiomyopathy in the United States is unknown. Published estimates range from 1 in 1400 to 1 in 15,000 live births4,11–17; however, these estimates are drawn from case series at tertiary care hospitals and are not population-based. In addition, some of these studies did not use standardized definitions of peripartum cardiomyopathy. More accurate prevalence data are required to further clarify risk factors for the condition and for mortality.
Prevalence estimates for peripartum cardiomyopathy could be drawn from hospitalization data because the condition is usually severe enough to require admission to a hospital. Unfortunately, the use of ICD codes to identify either cases of or deaths due to cardiomyopathy presents significant challenges. Peripartum cardiomyopathy not assigned a specific code in ICD-9, but can be coded as 648.6, which includes other (ie, not otherwise specified) cardiovascular conditions in the mother complicating pregnancy, childbirth or the puerperium; or as 674.8, which includes other complications of the puerperium (eg, hepatorenal syndrome after delivery, postpartum cardiomyopathy, subinvolution of the uterus, and uterine hypertrophy). If the link to pregnancy was not noted when coding occurred, ICD-9 codes 425.4 (other primary cardiomyopathies), 425.9 (secondary cardiomyopathies, unspecified), and 428 (heart failure) might be used. Currently, databases involving hospital discharge data continue to use ICD-9 coding.
Mortality coding in the United States changed from ICD-9 to ICD-10 in 1999. The ICD-10 offers a slight improvement because it includes a specific code for postpartum cardiomyopathy (O90.3); however, no code exists for peripartum cardiomyopathy causing death before delivery, and no distinction is made between peripartum cardiomyopathy and other cardiomyopathy in the postpartum period. Because of these difficulties with ICD codes, meaningful estimates of the prevalence of peripartum cardiomyopathy and mortality from cardiomyopathy will require detailed chart review or other methods that are more comprehensive than hospital discharge summary data.
The increase in the percentage of pregnancy-related deaths due to cardiomyopathy within the CDC's Pregnancy Mortality Surveillance System since 1979 is most likely due to improvements in case ascertainment. There are several reasons why cardiomyopathy deaths have been historically underreported among pregnancy-related deaths; all are factors influencing whether the relationship of the death to the pregnancy is documented on the death certificate. For example, deaths due to cardiomyopathy are frequently “late” maternal deaths, occurring more than 42 days postpartum (50% versus 8% of pregnancy-related deaths due to other causes). Another reason is that when coding with ICD-9, deaths occurring more than 42 days postpartum are not considered maternal deaths; therefore, if no additional methods are used for identification of pregnancy-related deaths, these cases are missed. In addition, clinicians might be less likely to note the history of pregnancy on a death certificate when the death occurs months after the end of pregnancy. Further, cardiologists, rather than obstetricians, are more likely to be the main providers of care for women with cardiomyopathy and might be even less likely to document the history of pregnancy on the death certificate. To help resolve these problems, enhanced methods of case finding have been implemented to help identify these cases as pregnancy-related deaths, including pregnancy status check boxes on deaths certificates (currently used in 16 states)18 and computerized record linkages. The latter method links deaths of reproductive-aged women to live birth and fetal death records for the previous year, thus identifying virtually all deaths occurring within 1 year of a pregnancy that ended in a live birth or a fetal death.
The increasing use of record linkages by states over the study period is probably the most important factor in the increasing number of pregnancy-related cardiomyopathy deaths being reported. In some states that have performed linkages, cardiomyopathy emerged as the leading cause of pregnancy-related death. For example, in North Carolina from 1992 to 1998, 32 of 151 (21.2%) pregnancy-related deaths were due to cardiomyopathy.19 In Hawaii, from 1992 to 1996, seven of 27 (25.9%) pregnancy-related deaths were due to cardiomyopathy (personal communication, Cheryl Prince, Hawaii Department of Health June 8, 2002).
Additional reasons for improved case ascertainment might include heightened awareness of peripartum cardiomyopathy among obstetricians and cardiologists, improvements in diagnostic modalities, and the regionalization of perinatal care. In addition to improved case ascertainment, the increase in the proportion of reported pregnancy-related deaths due to cardiomyopathy might also be partially attributable to a decline in other causes, such as hemorrhage.3
According to our search of the literature (MEDLINE; English language; 1966–May 2003; search terms: [myocardial diseases/ OR cardiomyopathy.tw] AND [pregnancy complications/ OR perpartum.tw]; [myocardium/] AND [pregnancy complications/ OR perpartum.tw]), this study is the first use of a national population-based data set to assess mortality resulting from cardiomyopathy in pregnancy. Because these deaths are relatively rare, national data aggregated over several years are required to describe trends and major risk factors. Despite improvements in the past 10 years, however, the identification of pregnancy-related deaths is known to be incomplete.3 The unusually long duration of peripartum cardiomyopathy also means that a significant number of deaths from this condition are not included in the definition of a pregnancy-related death. Although we only requested that states send us records of deaths that occurred within a year of pregnancy, we received an additional 29 certificates with a cause of death of cardiomyopathy for the period 1991–1997 that occurred up to 10 years after delivery. We do not know how many additional deaths from cardiomyopathy in pregnancy occurring more than 1 year postpartum were not reported to the Pregnancy Mortality Surveillance System. Thus, the impact of this condition might be underestimated in unique ways. Because the Pregnancy Mortality Surveillance System is based on vital record data, very limited clinical, behavioral, and social information is available on each case; if this information were available, it would allow for a much more accurate distinction between peripartum cardiomyopathy and other cardiomyopathies and a more detailed assessment of risk factors. Given the limited information available to us, we tried to adhere as closely as possible to the National Institutes of Health definition of peripartum cardiomyopathy. However, in many cases the exact onset of disease was not given, nor was the history of other medical conditions or echocardiographic results available. Given these limitations, we believe we approximated the National Institutes of Health definition as closely and consistently as possible.
An additional limitation of our study is that small case numbers prevented an assessment of confounding variables. For example, small cell sizes prevented stratification and adjustment to assess the risks related to age and parity.
To improve our assessment of pregnancy-related deaths overall and of cardiomyopathy as a cause of pregnancy-related death, a continued increase in vital record linkages by states is essential. Because this is not done by all states at this time, universal linkage would allow us to identify a higher proportion of cardiomyopathy deaths, which could increase the proportion of pregnancy-related deaths attributable to this cause to a level closer to that found in North Carolina and Hawaii. In addition, a sound population-based estimate of the prevalence of peripartum cardiomyopathy is required to estimate the case-fatality rate and to clarify whether the risk factors we observed for pregnancy-related death due to cardiomyopathy are the same as the risk factors for occurrence of cardiomyopathy. The extreme racial disparity in risk of death for women who have cardiomyopathy, for example, could be due to a difference in risk for the condition, in risk of death once the condition occurs, or both.
As cardiomyopathy in the peripartum is increasingly recognized and documented, efforts to reduce maternal mortality in the United States will need to include this rare but potentially devastating complication of pregnancy.
1. Atrash HK, Koonin LM, Lawson HW, Franks AL, Smith JC. Maternal mortality in the United States, 1979–1986. Obstet Gynecol 1990;76:1055–60.
2. Berg CJ, Atrash HK, Koonin LM, Tucker M. Pregnancy-related mortality in the United States, 1987–1990. Obstet Gynecol 1996;88:161–7.
3. Berg CJ, Chang J, Callaghan WM, Whitehead SJ. Pregnancy-related mortality in the United States, 1991–1997. Obstet Gynecol 2003;101:289–96.
4. Demakis J, Rahimtoola S. Peripartum cardiomyopathy. Circulation 1971;44:964–8.
5. Demakis JG, Rahimtoola SH, Sutton GC, Meadows WR, Szanto PB, Tobin JT, et al. Natural course of peripartum cardiomyopathy. Circulation 1971;44:1053–61.
6. Pearson GD, Veille JC, Rahimtoola S, Hsia J, Oakley CM, Hosenpud JD, et al. Peripartum cardiomyopathy: National Heart, Lung, and Blood Institute and Office of Rare Diseases (National Institutes of Health) workshop recommendations and review. JAMA 2000;283:1183–8.
7. Hibbard JU, Lindheimer M, Lang RM. A modified definition for peripartum cardiomyopathy and prognosis based on echocardiography. Obstet Gynecol 1999;94:311–6.
8. Witlin AG, Mabie WC, Sibai BM. Peripartum cardiomyopathy: An ominous diagnosis. Am J Obstet Gynecol 1997;176:182–8.
9. Blickstein I, Zalel Y, Katz Z, Lancet M. Ritodrine-induced pulmonary edema unmasking underlying peripartum cardiomyopathy. Am J Obstet Gynecol 1988;159:332–2.
10. Edoute Y, Blumenfeld Z, Bronstein M, Aharoni L, Moskowitz M. Peripartum congestive cardiomyopathy and endocardial fibroelastosis associated with ritodrine treatment. J Reprod Med 1987;32:793–7.
11. Lampert MB, Lang RM. Peripartum cardiomyopathy. Am Heart J 1995;130:860–70.
12. Ford RF, Barton JR, O'Brian JM, Hollingsworth RW. Demographics, management, and outcomes of peripartum cardiomyopathy in a community hospital. Am J Obstet Gynecol 2000;182:1036–8.
13. Veille JC. Peripartum cardiomyopathies: A review. Am J Obstet Gynecol 1984;148:805–18.
14. Pierce JA, Price BO, Joyce JW. Familial occurrence of postpartal heart failure. Arch Intern Med 1963;111:651–5.
15. Woolford RM. Postpartum myocardiosis. Ohio State Med 1952;48:924–30.
16. Meadows WR. Idiopathic myocardial failure in the last trimester of pregnancy and the puerperium. Circulation 1957;15:904–14.
17. Cunningham FG, Pritchard JA, Hankins GC, Anderson PL, Lucas MJ, Armstrong KF. Peripartum heart failure: Idiopathic cardiomyopathy or compounding cardiovascular events? Obstet Gynecol 1986;67:157–67.
18. Berg C, Danel I, Atrash H, Zane S, Bartlett L, eds. Strategies to reduce pregnancy-related deaths: From identification and review to action. Atlanta: Centers for Disease Control and Prevention, 2001.
19. Buescher PA, Harper M, Meyer RE. Enhanced surveillance of maternal mortality in North Carolina. N C Med J 2002;63:76–9.