Dr EV Kuklina, Centers for Disease Control and Prevention, 4770 Buford Highway NE, Mailstop K-37, Atlanta, GA 30341-3724, USA. Email email@example.comThe findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
Please cite this paper as: Kuklina E, Callaghan W. Chronic heart disease and severe obstetric morbidity among hospitalisations for pregnancy in the USA: 1995–2006. BJOG 2011;118:345–352.
Objectives To describe changes in characteristics of delivery and postpartum hospitalisations with chronic heart disease from 1995 to 2006.
Design Cross-sectional study.
Setting USA, nationwide hospital discharge data.
Population A total of 47 882 817 delivery hospitalisations and 660 038 postpartum hospitalisations.
Methods Adjusted odds ratios describing the associations between chronic maternal heart disease and severe obstetric complications were obtained from multivariable logistic models. The contribution of chronic heart disease to severe morbidity was estimated using adjusted population-attributable fractions.
Main outcome measures Prevalence and trends in chronic heart disease, rate and risk of severe obstetric complications.
Results In 2004–2006, about 1.4% of delivery hospitalisations were complicated with chronic heart disease. No substantial changes in the overall prevalence of chronic heart disease among hospitalisations for delivery were observed from 1995–1997 to 2004–2006. Even so, a linear increase was found for specific congenital heart disease, cardiac dysrhythmias, and cardiomyopathy and congestive heart failure (P < 0.01). During this same period the rate of postpartum hospitalisations with chronic heart disease tripled (P < 0.01). Severe complications during hospitalisations for delivery among women with chronic heart disease were more common in 2004–2006 than in 1995–1997. In 2004–2006, 64.5% of the cases of acute myocardial infarction, 57.5% of the instances of cardiac arrest/ventricular fibrillation, 27.8% of in-hospital mortality and 26.0% of the cases of adult respiratory distress syndrome were associated with hospitalisations with chronic heart disease.
Conclusions In the USA chronic heart disease among women hospitalised during pregnancy may have increased in severity from 1995 to 2006.
According to the most recent data available, heart disease is the leading indirect cause of maternal mortality in the developed countries.1,2 Moreover, in some countries the number of maternal cardiac deaths has actually risen in recent years. For example, such deaths increased substantially in the UK between 1986–1988 and 2003–2005.3 Rationales proposed for the increase have included survival into adulthood of the majority of children born with congenital heart disease,4 growing rates of obesity and chronic conditions among women of reproductive age,5,6 and increasing age at birth.7 Yet, despite the growing recognition of the importance of chronic heart disease in the obstetric population, data on its epidemiology are scarce.8 Clearly, determining the prevalence (and trends) of chronic heart disease among pregnant women in the USA could be a useful first step toward understanding this problem.
Although the increased risk of maternal mortality among women with acute and chronic heart disease has been documented, maternal mortality remains a very rare event in developed countries.1,2 As a result, there is an interest in understanding the risk of severe maternal (obstetric) complications, events occurring at least 10 times as often as maternal death.9,10 Fortunately, large hospital discharge databases are available to monitor trends in rare conditions that increase maternal risk and to estimate the contribution of diseases that complicate pregnancy to severe maternal morbidity at the population level.6
The primary objectives of the present report were to: 1. describe prevalence and trends for chronic heart disease during delivery and postpartum hospitalisations; 2. investigate the associations of chronic heart disease with severe obstetric complications; and 3. estimate the contribution of chronic heart disease to the rate of severe obstetric complications among delivery hospitalisations, in each using the 1995–2006 Nationwide Inpatient Sample (NIS) of the Healthcare Cost and Utilization Project (HCUP).
The HCUP is a federal–state–industry partnership sponsored by the Agency for Healthcare Research and Quality,10 and the NIS is the largest all-payer database for inpatient care in the USA. All of the nonfederal community hospitals in the states that participate in the annual HCUP data collection are stratified by rural/urban location, number of beds, region, teaching status and ownership. Within each stratum, a systematic random sample of 20% of the hospitals is drawn; 100% of the discharges from these sampled hospitals are included. The methodology, which allows for making nationwide estimates, is described in full on the HCUP website.11 Because the NIS excludes data elements that could directly or indirectly identify individual persons, this research was considered exempt from review by the institutional review board of the Centers for Disease Control and Prevention.12 Our analysis included all delivery and postpartum hospitalisations in 1995–2006, which we identified hierarchically. First, we identified the delivery hospitalisations using the enhanced identification method, which identifies hospitalisations resulting in childbirth. It includes the ninth modification of the International Classification of Diseases (ICD-9-CM) diagnosis codes for an outcome of delivery, diagnosis-related group (DRG) delivery codes, and ICD-9-CM procedure codes for selected delivery-related procedures as documented in detail elsewhere.13 Second, we identified the postpartum hospitalisations using the fifth digit = 4 in ICD-9-CM codes for primary or secondary diagnosis or ICD-9-CM code = V24 for any listed diagnosis, or postpartum DRG codes = 376–377.
The delivery hospitalisations with severe obstetric complications were identified using condition-specific ICD-9-CM codes; the specific ICD-9-CM codes for severe complications and details of the method to identify delivery hospitalisations with these conditions can be found elsewhere.10 In-hospital mortality and transfer status were identified using the variables ‘died during hospitalisation’ and ‘disposition of patient’ or ‘admission source’ in the HCUP data system, respectively.
Among all hospitalisations with ICD-9-CM codes for selected severe complications, hospitalisations with in-hospital mortality and transfer were hierarchically classified as hospitalisations with severe complications regardless of the length of stay. The ICD-9-CM codes at discharge might be used to indicate an historical condition or a ‘ruled out’ condition. To use the most conservative approach in identifying the hospitalisations with severe obstetric complications, we reclassified the remaining hospitalisations that did not have a long length of stay (defined as hospitalisations with length of stay <90th percentiles calculated separately for vaginal, and primary and repeat caesarean deliveries) into hospitalisations without severe complications. In this study, in addition to the severe complications defined by the ICD-9-CM codes described above, we included acute myocardial infarction (ICD-9-CM = 410.x); disorders of fluid, electrolyte and acid–base balance (ICD-9-CM = 276.x) and cardiac arrest/ventricular fibrillation (ICD-9-CM = 427.5, 427.41 and 427.42) as severe complications. The hospitalisations with chronic heart disease were also identified hierarchically by the ICD-9-CM codes listed in Supporting Information Appendix S1. The estimates for the prevalence of specific types of congenital heart disease during postpartum hospitalisations were unreliable, because of the small sample, so we were able to report only the overall prevalence of congenital heart disease in the postpartum period.
The unit of analysis was a hospitalisation, not an individual, because our data did not allow us to account for multiple delivery and postpartum hospitalisations of the same women during the study period. We report rates of delivery and postpartum hospitalisations with chronic heart disease per 10 000 deliveries for four 3-year study periods: 1995–1997, 1998–2000, 2001–2003 and 2004–2006. Rates of severe complications per 10 000 delivery hospitalisations during the same 3-year study intervals were also estimated. We calculated rates of severe complications and linear trends for all four intervals. Orthogonal polynomial coefficients calculated recursively by the method of Fisher and Yates were used to test linear trend.
The likelihood of severe complications among hospitalisations for delivery with chronic heart disease using delivery hospitalisations without chronic heart disease as the referent was calculated using odds ratios (ORs) and their 95% confidence intervals (CI). This analysis was also performed for the four 3-year study periods: 1995–1997, 1998–2000, 2001–2003 and 2004–2006. All ORs were adjusted for maternal age, multiple births, mode of delivery, a variety of comorbidities (including diabetes, hypertensive disorders, chronic renal failure, rheumatic disorders, anaemia and nutritional disorders) and hospital characteristics (US region, location type, teaching status and bed number). The specific ICD-9-CM codes for these comorbidities can be found elsewhere.8 We calculated population-attributable fractions while accounting for adjustment of ORs and 95% CI using methods described elsewhere.14,15 Because some states do not report race/ethnicity to HCUP, we did not include this variable in the analysis.10 We used SAS software (version 9.1; SAS Institute Inc., Cary, NC) to manage the data and SAS-callable sudaan software (version 9.0; RTI International, Research Triangle, NC) to account for the multistage probability sampling design. Hence, all results are based on weighted estimates of hospitalisations for pregnancy in the USA during the study period.
Using the enhanced method, we identified 47 959 613 delivery hospitalisations that occurred during 1995–2006. After excluding 57 918 (0.12%) with nondelivery pregnancy outcomes such as hydatidiform mole, other abnormal product of conception, ectopic pregnancy or abortion, and 18 877 (0.04%) with missing data on age, we were left with 47 882 817 delivery hospitalisations available for analysis. Among these hospitalisations, we identified 658 843 with chronic heart disease: 167 552 in 1995–1997, 160 287 in 1998–2000, 156 556 in 2001–2003 and 174 448 in 2004–2006. Of 660 038 postpartum hospitalisations, we identified 44 888 with chronic heart disease: 5442 in 1995–1997, 8614 in 1998–2000, 12 581 in 2001–2003, and 18 251 in 2004–2006.
No substantial changes were observed in the overall prevalence of chronic heart disease among delivery hospitalisations from 1995–1997 to 2004–2006 (for linear trend P = 0.26). In 2004–2006, about 1.4% of delivery hospitalisations were complicated with chronic heart disease. A significant linear increase (P < 0.01) was observed for four specific categories of chronic heart disease: congenital heart disease (Figure 1), rheumatic valve disorders (Figure 1), cardiomyopathy and/or congestive heart failure (Figure 1), and cardiac dysrhythmias (Figure 2). The rates of non-rheumatic valve disorders and other heart disorders decreased significantly. The use of ICD-9-CM codes for specific congenital heart disorders 745 (septal defects), 747 (anomalies of circulatory system), and 746 (valve disorders) increased significantly from 2.0 to 2.7, from 0.9 to 1.5, and from 1.8 to 3.1 per 10 000 deliveries, respectively, during the period of study (data are not shown). In contrast, the use of ICD-9-CM code 648.5 (non-specific congenital heart disease) decreased significantly from 2.7 in 1995–1997 to 1.2 in 2004–2006 per 10 000 deliveries (data are not shown). Contrary to the findings for delivery hospitalisations, the overall prevalence of chronic heart disease among postpartum hospitalisations increased from 4.8 to 14.4 per 10 000 delivery hospitalisations (P < 0.01) during the period of study (data are not shown). A significant growth in the rate of postpartum hospitalisations with chronic heart disease was observed for all groups of chronic heart disease investigated in our study (P < 0.01) (Figure 3). The largest percentage increases were observed for congenital heart disease and cardiac dysrhythmias/conduction disorders, from 0.1 to 0.4 and from 1.1 to 4.0 per 10 000 delivery hospitalisations, respectively. The specific ICD-9-CM code for peripartum cardiomyopathy was only introduced in 2003 and in 2004–2006, the rate of peripartum cardiomyopathy was 0.6 and 1.4 per 10 000 deliveries for delivery and postpartum hospitalisations, respectively.
The rate of severe obstetric complications among hospitalisations with chronic heart disease increased significantly (P < 0.01 for linear trend) during the period of study (Table 1) except for in-hospital mortality (a slight drop from 15.5 to 15.1 per 10 000 delivery hospitalisations). Among delivery hospitalisations with chronic heart disease, the rates of cardiac arrest/ventricular fibrillation, sepsis and fluid, and electrolyte disorders increased from 15.1 to 29.1, from 18.0 to 27.3, and from 96.2 to. 167.8 per 10 000 delivery hospitalisations, respectively. A different pattern of changes was observed among hospitalisations without chronic heart disease. There, significant decreases from 1995–1997 to 2004–2006 were seen for sepsis (3.0 to. 2.3), and in-hospital mortality (0.7 to. 0.5), whereas there were no significant changes in fluid and electrolyte disorders (from 15.4 to. 16.7) and cardiac arrest/ventricular fibrillation (0.3 in both periods). The rates of other complications increased during the period of interest.
Table 1. Rates of severe obstetric complications by chronic heart disease status among delivery hospitalisations in 1995−1997 (n = 11 345 756) and 2004−2006 (n = 12 670 136); Nationwide Inpatient Sample
Rate (standard error) per 10 000 delivery hospitalisations
Chronic heart disease, yes
Chronic heart disease, no
ARDS, adult respiratory distress syndrome; DIC, disseminated intravascular coagulation syndrome; N/A, number of events is too small to estimate rate.
*P for linear trends is nonsignificant (>0.01); for all other conditions P < 0.01. Trend tested for four periods: 1995−1997, 1998−2000, 2001−2003 and 2004−2006.
Acute renal failure
Acute myocardial infarction
Cardiac arrest/ventricular fibrillation
Fluid and electrolyte disorders
In 1995–1997, the adjusted likelihood (OR) of severe complications for delivery hospitalisations with chronic heart disease using delivery hospitalisations without chronic heart disease as the referent ranged from 2.1 for disseminated intravascular coagulation syndrome to 212.7 for acute myocardial infarction (Table 2). Except for acute renal failure and acute myocardial infarction, the likelihood of the severe complications we investigated for delivery hospitalisations with chronic heart disease increased from 1995–1997 to 2004–2006. The largest increases were observed for cardiac arrest/ventricular fibrillation (OR increased from 35.7 to. 82.3). The pattern of increases in population-attributable fractions from 1995–1997 to 2004–2006 was similar to the pattern for ORs (Table 3). In 2004–2006, 64.5% of the cases of acute myocardial infarction, 57.5% of the instances of cardiac arrest/ventricular fibrillation, 27.8% of in-hospital mortality, and 26.0% of the cases of adult respiratory distress syndrome were associated with hospitalisations with chronic heart disease.
Table 2. Estimated odds ratios for severe obstetric complications for delivery hospitalisations with chronic heart disease compared with delivery hospitalisations without chronic heart disease in 1995−1997 (n = 11 345 756) and 2004−2006 (n = 12 670 136); Nationwide Inpatient Sample
*Odds ratios and 95% confidence intervals adjusted for mode of delivery, multiple birth, diabetes, hypertensive disorders, chronic renal failure, rheumatic disorders, anaemia, nutritional disorders, maternal age, payer, hospital location, hospital teaching status, and hospital region.
Acute renal failure
Acute myocardial infarction
Cardiac arrest/ventricular fibrillation
Fluid and electrolyte disorders
Table 3. Estimated population attributable fractions for severe obstetric complications for delivery hospitalisations with chronic heart disease compared with delivery hospitalisations without chronic heart disease in 1995−1997 (n = 11 345 756) and 2004−2006 (n = 12 670 136); Nationwide Inpatient Sample
*Odds ratios and 95% confidence intervals adjusted for mode of delivery, multiple birth, diabetes, hypertensive disorders, chronic renal failure, rheumatic disorders, anaemia, nutritional disorders, maternal age, payer, hospital location, hospital teaching status, and hospital region. Population-attributable fractions and 95% confidence intervals calculated using the following formula: pd(RR–1/RR), where pd = proportion of cases (hospitalisations with severe complications) exposed to risk factor (heart disease) and RR = relative risk for outcome for exposed compared with unexposed.
Acute renal failure
Acute myocardial infarction
Cardiac arrest/ventricular fibrillation
Fluid and electrolyte disorders
From 1995–1997 to 2004–2006, the mean length of stay increased among delivery hospitalisations regardless of chronic heart disease status (P < 0.01 for linear trend) (data are not shown). The mean length of stay increased from 2.21 (95% CI 2.19–2.23) to 2.61 (95% CI 2.59–2.64) and from 2.90 (95% CI 2.82–2.96) to 3.58 (95% CI 3.50–3.67) for delivery hospitalisations with and without chronic heart disease, respectively. The change in the mean length of stay for postpartum hospitalisations without chronic heart disease was not significant (from 3.29 [95% CI 3.04–3.53] to 3.09 [95% CI 3.03–3.15]; data are not shown). In contrast, a significant decrease (P < 0.01 for linear trend) in the mean length of stay was observed among postpartum hospitalisations with chronic heart disease during the period of study (from 4.86 [95% CI 4.53–5.17] to 4.31 [95% CI 4.07–4.56]).
Previous reports from developed countries indicated that the prevalence of chronic heart disease among pregnant women was in the range of 1% to 3%, with congenital heart disease being the most prevalent condition.16 However, for the USA and other developed countries, estimates of the nationwide prevalence of chronic heart disease that required some kind of treatment or intervention during a delivery hospitalisation have not been available. The present study is the first large USA population-based investigation to report the prevalence of chronic heart disease during delivery hospitalisations and to estimate the contribution of chronic heart disease to severe obstetric complications. During the period of our study, the overall prevalence of delivery hospitalisations with chronic heart disease stayed relatively stable, between 1.3% and 1.5%. However, several lines of evidence suggest that the severity of chronic heart disease during these hospitalisations increased during 1995–2006. First, the prevalence of hospitalisations with ICD-9-CM codes for certain conditions, including specific congenital heart disease, cardiac dysrhythmias, and cardiomyopathy or congestive heart failure, increased linearly from 1995–1997 to 2004–2006. Given that hospitalisations of persons with either congenital or acquired heart disease may include codes in the discharge records for cardiac dysrhythmia, congestive heart failure or cardiomyopathy,17 future studies based on medical records are needed to estimate what proportions of the increases we observed are the result of congenital compared with acquired heart disease. Second, our surveillance data demonstrated that the risk for the majority of severe complications among delivery hospitalisations with chronic heart disease, using hospitalisations without chronic heart disease for comparison, increased from 1995–1997 to 2004–2006. Finally, a tripling in the rate of postpartum hospitalisations with chronic heart disease was observed from 1995–1997 to 2004–2006. Severe obstetric complications have been shown to mirror maternal mortality in developed countries.18 According to the UK national reporting system, the substantial increase in maternal cardiac deaths associated with acquired disease, such as myocardial infarction, aortic dissection and cardiomyopathy, has been observed from 1986–1988 to 2003–2005 (the most recent data available).19 As a result of this, heart disease was the leading indirect (due to pre-existing conditions) cause of maternal deaths in 2003–2005.19
Even though significant progress has been made in recent years in understanding, diagnosing and treating heart disease, protocols for the management of pregnancies complicated by chronic heart disease have been difficult to establish,20–24 and clinical decisions for these women are often based on the experiences of single institutions.20 This is not surprising, as the prevalence of any specific heart condition is relatively low, as seen in our large nationwide study. It is widely known, however, that certain cardiac lesions, the presence of cyanosis, specific functional classes and the use of anticoagulation are indicators of severe heart disease and are strongly associated with severe complications during pregnancy.25
While regurgitant valve diseases are generally well tolerated during pregnancy, valvular stenosis is often associated with poor outcomes.26 Although anticoagulant therapy is indicated in pregnant women with a prosthetic valve replacement, the 2008 American College of Cardiology/American Heart Association guidelines acknowledge that currently there is not sufficient evidence to make definitive recommendations about optimal antithrombotic therapy in these women.20 Women with congenital valvular disease corrected in early life without requiring prosthetic valves are considered a low-risk group. However, residual defects may be present in up to 50% of these women, and cardiac assessment through a clinical evaluation and echocardiography is recommended every trimester for them.27 A national registry of chronic heart disease in pregnancy could serve as a valuable source of clinically detailed data. For example, the ongoing European Pregnancy and Heart Disease Registry28 collects information about disease characteristics and pregnancy outcomes from pregnant women with valvular defects, congenital heart disease, ischaemic disease or cardiomyopathy during pregnancy and 6 months after the pregnancy has ended.
The results of the present analysis should be interpreted in the light of the following limitations of the study. First, chronic heart disease was identified by using ICD-9-CM codes abstracted from discharge records for delivery and postpartum hospitalisations. Although chronic heart disease, as in the case of any pre-existing condition, should be coded if this condition interferes with patient management,17 low sensitivity (albeit high specificity) of the ICD-9-CM codes for chronic conditions among obstetric hospitalisations has been reported.29 Moreover, milder forms of chronic heart disease are likely to be less commonly reported in hospital records, as the management of these women is not affected. Indeed, antibiotic prophylaxis for vaginal delivery is considered optional for high- and moderate-risk women with pre-existing cardiac abnormalities and it is not recommended for patients in the negligible-risk category.21 As a result our estimates may be low and biased toward the reporting of more severe disease. Second, we were unable to track multiple hospitalisations per patient and so could not distinguish between the first time a woman diagnosed with chronic heart disease was admitted for a delivery and repeat hospitalisations for that woman in the postpartum period. We also could not track the repeat hospitalisations in the postpartum period for the same woman. However, our results confirmed that postpartum period is a vulnerable period for women with chronic heart disease and highlighted the importance of tracking chronic heart disease during postpartum hospitalisations using external datasets to fully understand the trends and burden of this condition. Future studies based on the more detailed clinical data are needed to examine early discharge and lack of appropriate care as well as other factors contributing to the fast-growing number of postpartum hospitalisations complicated with chronic heart disease in the USA. Finally, although our results showed strong associations between chronic heart disease and severe obstetric complications, causality cannot be established using our cross-sectional data. In addition, because of the nature of chronic heart disease in pregnancy, it is unrealistic to assume that all chronic heart disease that requires some kind of clinical management during hospitalisations for delivery can be eliminated. Hence, our population-attributable fractions can only be interpreted as the proportions of delivery hospitalisations with particularly severe complications that are associated with chronic heart disease.
In conclusion, we have shown the possibility that chronic heart disease among women hospitalised during pregnancy has increased in severity from 1995–1997 to 2004–2006 in the USA. We have also provided evidence that chronic heart disease in pregnant women is one of the major contributors to severe obstetric complications. Our results demonstrated the changing landscape of pregnant women with heart disease. They also indicated that clinicians who take care of pregnant women with heart disease should be particularly vigilant for signs and symptoms potentially indicative of severe obstetric complications and consider using a low threshold for initiating a diagnostic investigation to exclude these complications. Identification of women with chronic heart disease and collecting clinically detailed data through a registry or other means is essential for improving maternal care and pregnancy outcomes in women with heart disease.
Disclosure of interests
The authors have nothing to disclose.
Contribution to authorship
Elena V Kuklina and William M Callaghan contributed equally to study concept, design, interpretation of data, drafting the manuscript and its revision for important intellectual content. Elena V Kuklina was involved in acquisition of data and undertook the analysis.