In the United States, the maternal mortality ratio in 1915 was 607.9 deaths per 100,000 live births.1 The discovery and introduction of best practices including aseptic technique, improved surgical practices, and the use of antibiotics and blood transfusion all have contributed to a marked decrease in the rate of maternal mortality. In 2011, the maternal mortality ratio was approximately 17 deaths per 100,000 live births.2
Despite continued advances in various medical technologies, rates of maternal death in the United States have increased in the past 25 years.2 Some of the increase may be attributed to changes in definitions of mortality related to hospital discharge coding and revision of the United States death certificate.3 Yet, the rate of increase in maternal mortality in the United States is not consistent with trends in other developed countries.2 Most striking of the epidemiologic changes in maternal mortality in the United States are the causes of these deaths. As in the rest of the world, the most commonly cited causes of pregnancy-associated maternal death in the United States are complications of pregnancy-related hypertension, hemorrhage, and embolism.4,5 From 1987 to 2005 the percentage of deaths from hemorrhage, infection, and hypertensive disorders of pregnancy decreased, whereas death from pre-existing medical conditions, including cardiovascular, increased.3
Associations between specific underlying maternal diseases in cases of maternal mortality are pertinent to modern obstetric care for many reasons. More women are delaying childbearing until they are older, a time of increasing prevalence of chronic disease.6 Obesity is common in pregnancy,7 and comorbidities are more common in obese patients.8
The purpose of this study was to investigate the association between maternal disease and maternal death at the time of delivery hospitalization.
MATERIALS AND METHODS
Data regarding singleton births from 1995 to 2003 were obtained from a dataset linking birth certificate data from the New York City Department of Health and Mental Hygiene to hospital discharge data from the New York Statewide Planning and Research Cooperative System. Previous studies have indicated an improved positive predictive value for identifying maternal medical conditions when birth certificate data were combined with hospital discharge data.9,10 Our study population was limited to patients delivering liveborn singleton neonates at 22 or more completed weeks of gestation (by clinical estimate) with known maternal demographics and hospital discharge disposition (death or discharged home alive). Information regarding maternal events that occurred before neonatal viability (less than 22 weeks) or after discharge from the delivery hospitalization was not linked to this dataset.
Maternal demographic and pregnancy characteristics were obtained from birth certificate data and the Statewide Planning and Research Cooperative System database and were examined as potential confounders. Maternal demographic characteristics derived from birth certificates included age (younger than 20 years, between 20 years and 35 years, and older than 35 years), race or ethnicity, level of education (more than high school, high school or less), insurance status, employment status, nativity, parity, mode of delivery (vaginal including forceps or vacuum assistance, primary cesarean, repeat cesarean), and gestational age at delivery (less than 28 weeks, 28–36 weeks, 37–40 weeks, more than 40 weeks).11–13 All deliveries occurred in an urban setting. Information regarding delivering hospitals was deidentified in the dataset; therefore, location of delivery could not be examined.
Maternal diseases examined included chronic hypertension, pregestational or gestational diabetes mellitus, chronic cardiovascular disease, pulmonary hypertension, chronic lung disease, human immunodeficiency virus (HIV), and preeclampsia or eclampsia. The presence of one disease or more was ascertained using combined data from the International Classification of Diseases, 9th Revision, Clinical Modification codes (Table 1) from the mother's hospitalization record and using the checkbox information from the birth certificate. The morbidity was considered present if it was identified in the discharge diagnosis codes, if it was noted on the birth certificate, or if it was present on both when possible. Diabetes was classified as pregestational or gestational. Because of dataset limitations, differentiation between insulin-dependent diabetes mellitus and noninsulin-dependent diabetes mellitus was not possible. Both chronic hypertension and pregestational diabetes mellitus are common risk factors for the development of preeclampsia or eclampsia. Patients with superimposed preeclampsia or eclampsia in the setting of either of these underlying diseases may be at higher risk for development of more severe disease than patients with preeclampsia or eclampsia but no underlying disease. Therefore, we analyzed chronic hypertension and pregestational diabetes alone and combined with the diagnosis of preeclampsia or eclampsia. Because the definitions of chronic cardiovascular disease and chronic lung disease are not specific on the birth certificate, we chose to create these variables using International Classification of Diseases, 9th Revision, Clinical Modification codes only. Chronic cardiovascular disease included the diagnoses of chronic rheumatic heart disease, diseases of any of the four main cardiac valves, hypertensive heart disease with heart failure, previous myocardial infarction, and current systolic heart failure, diastolic heart failure, or both. Chronic lung disease diagnoses included sarcoidosis, chronic bronchitis, emphysema, asthma, and pulmonary fibrosis. Because pulmonary hypertension is known to have a high incidence of maternal mortality, we chose to analyze the condition separately.
Because the birth file does not contain data regarding the mother's height, body mass index (calculated as weight (kg)/[height (m)]2) could not be calculated. Instead, the mother's prepregnancy weight was used as a proxy to define obesity. Thus, we defined obesity as a prepregnancy weight of more than 114 kilograms (250 pounds). This cutoff has been validated in a previous study14 demonstrating sufficient sensitivity in identifying obese patients without compromising specificity. For example, a woman with a height of 5 feet 10 inches and a weight of 114 kilograms (250 pounds) has a calculated body mass index of 35.12–14
We considered as potential confounders variables that were plausibly related to maternal death and to the predictors of interest. Among these candidates, those with statistically significant associations with both predictors of interest as well as maternal death (P<.05) were included. Maternal mortality rates were calculated and 95% confidence intervals (CIs) were generated for mortality rates using the Poisson distribution.15 Logistic regression was used to estimate the odds of maternal death in the presence of various maternal comorbidities adjusting for maternal characteristics, pregnancy characteristics, and delivery outcomes. Analyses were conducted using SAS 9.2. This study was considered exempt by the Yale School of Medicine Human Investigation Committee.
Between January 1995 and December 2003, there were 1,084,862 live singleton births and 132 maternal deaths. The maternal mortality rate of this cohort is 12.2 per 100,000. Maternal age, race, primary payer, mode of delivery, and gestational age at delivery differed between comparison groups (P<.001). In particular, patients who were older, black, and those with Medicaid or self-pay were at increased risk for mortality at delivery hospitalization. Additionally, patients who died at delivery hospitalization were more likely to have a primary cesarean delivery and to be delivered before 37 weeks of gestation or after 40 weeks of gestation (Table 2).
The unadjusted and adjusted odds ratios (ORs) for maternal morbidities in relation to maternal mortality are presented in Table 3. Out of the entire study population, 4.9% had gestational diabetes (52,970/1,084,862), 4.1% had preeclampsia or eclampsia (44,004/1,084,862), 1.8% had chronic hypertension (19,647/1,084,862), 1.5% had chronic lung disease (16,319/1,084,862), 1.1% of patients were classified as obese (11,936/1,084,862), 0.8% had chronic cardiovascular disease (8,943/1,084,862), 0.7% had pregestational diabetes (7,474/1,084,862), 0.4% had HIV (4,665/1,084,862), 0.01% had pulmonary hypertension (166/1,084,862), 0.6% had a combined diagnosis of chronic hypertension and preeclampsia or eclampsia (6,309/1,084,862), and 0.08% had a combined diagnosis of pregestational diabetes and preeclampsia or eclampsia (906/1,084,862). Maternal mortality was associated with all of the morbidities analyzed, with the exception of chronic cardiovascular disease and chronic lung disease, after adjusting for potential confounders. Obese patients were 190% more likely to die at delivery hospitalization than nonobese patients (adjusted OR, 2.9; 95% CI 1.1–8.1). The number needed to harm for obesity is 9,091; therefore, for every 9,091 individuals exposed to obesity, one will experience death at delivery hospitalization. The pregnancy-related hypertensive disorders all were associated with maternal death. Pregestational diabetes mellitus was associated with maternal death at delivery hospitalization (adjusted OR, 3.3; 95% CI 1.3–8.1), but gestational diabetes mellitus was not. The presence of superimposed preeclampsia or eclampsia combined with chronic hypertension or pregestational diabetes was associated with greater increases in risk of death at delivery hospitalization, for example, patients with both pregestational diabetes and preeclampsia or eclampsia had an adjusted OR for risk of death of 8.4 (95% CI 2.1–34.4).
The presence of select maternal morbidities was significantly associated with increased odds of mortality at the time of delivery hospitalization of a liveborn neonate, even after adjustment for potential confounders. The association between pulmonary hypertension and maternal mortality is well-described in the literature, with mortality risk reported as high as 50%.16 However, the associations between the more common morbidities of obesity, chronic hypertension, and pregestational diabetes have been less frequently described. Further characterizing these associations is important because the prevalence of obesity and rates of medical conditions like chronic hypertension and diabetes have increased in the past two decades.17,18 Additionally, there is evidence that maternal medical comorbidities are increasingly connected with maternal deaths.19
The association between HIV and maternal death is less clear. Patients who are HIV-positive are more likely to come from socially disadvantaged communities with poor access to health care.20,21 Delayed diagnosis is common in patients, with one third either presenting with acquired immune deficiency syndrome (AIDS) or progressing to AIDS within 12 months of the HIV diagnosis.21 Thus, despite screening and access to highly active antiretroviral therapy during pregnancy, it is possible that the same social risk factors that place some patients at higher risk for acquiring HIV may add to an increased risk of mortality at delivery hospitalization.
Changing sociodemographic trends in childbearing may be partially responsible for the increased prevalence of maternal morbidities. With more patients delaying childbearing beyond the age of 35 years, there is more risk for maternal morbidities, and this may shift the mortality calculation in this group.6 Older age by itself is a risk factor for maternal mortality, regardless of parity, time of entry into prenatal care, and level of education.22 Additionally, there is an increased prevalence of chronic medical conditions in women with advanced maternal age.23,24 Obesity is also a risk factor for the development of chronic medical conditions, including pregestational diabetes and chronic hypertension, and may be, in part, driving the increase in the prevalence of these maternal morbidities.8
Maternal demographics in our cohort showed patterns similar to the most recently published national data. We found no association between nativity and maternal death. Thus, this dataset may be generalizable to other urban areas with ethnically and socioeconomically diverse populations within the United States.25 In our study, maternal mortality was significantly associated with increasing maternal age, and black non-Hispanic patients accounted for a disproportionate number of maternal deaths. Additionally, patients without insurance (self-pay) comprised a larger portion of the maternal deaths. No difference was found between groups when comparing level of maternal education, maternal employment status, maternal birth outside the United States, or increasing parity. Larger maternal mortality studies have identified a protective effect of increasing levels of education, especially when stratified by maternal age.22,25 Other studies have shown increasing parity to be associated with increased risk of maternal mortality, especially in patients with five or more previous deliveries.25,26
Our study has several strengths compared with previous studies. First, New York City provides a model study population because the cohort is large and economically and ethnically diverse. Additionally, although maternal medical conditions are consistently underreported on birth certificates,10 we were able to use a dataset that matched all birth certificate data with hospital discharge data for the mother from the Statewide Planning and Research Cooperative System database, significantly improving data quality.9
We acknowledge limitations of this study. First, this study is dependent on the accuracy and completeness of information from discharge and birth certificate data, so misclassification may have occurred. In cases of maternal death, there may be reporting bias and morbidities may have been more accurately reported in the discharge data compared with patients discharged after a successful delivery. This may lead to an overrepresentation of morbidity in the maternal death group. Death certificate records were not available for review; therefore, the chain of events that led to a maternal death could not be ascertained. Furthermore, in an attempt to extract missing data and further explain the causes of death, we were not able to examine the medical records from the cases of maternal death. We were only able to examine the effect of maternal diseases on mortality among patients whose pregnancies ended in a live birth beyond 22 completed weeks of gestation.
In summary, there is an increased risk of maternal death at delivery hospitalization when the pregnancy is complicated by common medical conditions. This finding is clinically relevant because of the increasing prevalence of underlying medical disease in females of reproductive age. Primary or secondary prevention of some of these morbidities is possible. One of the greatest accomplishments may be achieved by limiting the development of obesity in adolescent girls and targeting weight loss in obese women of reproductive age. Reduction of the burden of obesity will subsequently lead to fewer diagnoses of chronic hypertension and noninsulin-dependent diabetes mellitus and fewer pregnancies complicated by these diseases.27
Maternal mortality remains a significant burden, globally and in the United States, and is receiving renewed attention from local and international health authorities. Identification of these important contributors to maternal mortality is a critical step in the approach to prediction and prevention.
1. Hoyert DL. Maternal mortality and related concepts. Vital & Health Statistics. Series 3, Analytical and Epidemiological Studies. Washington, DC: U.S. Department of Health and Human Services, Public Health Service, National Center for Health Statistics. 2007:1–13.
2. Hogan M, Foreman K, Naghavi M, Ahn S, Wang M, Makela S, et al.. Maternal mortality for 181 countries, 1980-2008: a systematic analysis of progress towards Millennium Development Goal 5. Lancet 2010;375:1609–23.
3. Berg C, Callaghan W, Henderson Z, Syverson C. Pregnancy-related mortality in the united states, 1998 to 2005. Obstet Gynecol 2011;117:1230.
4. Clark S, Belfort M, Dildy G, Herbst M, Meyers J, Hankins G. Maternal death in the 21st century: causes, prevention, and relationship to cesarean delivery. Am J Obstet Gynecol 2008;199:36.e1–5.
5. MacKay A, Berg C, Duran C, Chang J, Rosenberg H. An assessment of pregnancy-related mortality in the United States. Paediatr Perinat Epidemiol 2005;19:206–14.
6. Martin J, Hamilton B, Ventura S, Osterman MJK, Kirmeyer S, Mathews TJ, et al.. Births: final data for 2009. Natl Vital Stat Rep 2011;60:1–70.
7. Flegal K, Carroll M, Kit B, Ogden C. Prevalence of obesity and trends in the distribution of body mass index among US adults, 1999-2010. JAMA 2012;307:491–7.
8. Willett WC, Dietz WH, Colditz GA. Guidelines for healthy weight. N Engl J Med 1999;341:427–34.
9. Lydon Rochelle M, Holt V, Crdenas V, Nelson J, Easterling T, Gardella C, et al.. The reporting of pre-existing maternal medical conditions and complications of pregnancy on birth certificates and in hospital discharge data. Am J Obstet Gynecol 2005;193:125–34.
10. Roohan P, Josberger R, Acar J, Dabir P, Feder H, Gagliano P. Validation of birth certificate data in New York State. J Community Health 2003;28:335–46.
11. Lydon Rochelle M, Holt V, Nelson J, Crdenas V, Gardella C, Easterling T, et al.. Accuracy of reporting maternal in-hospital diagnoses and intrapartum procedures in Washington State linked birth records. Paediatr Perinat Epidemiol 2005;19:460–71.
12. Lipkind H, Duzyj C, Rosenberg T, Funai E, Chavkin W, Chiasson M. Disparities in cesarean delivery rates and associated adverse neonatal outcomes in New York city hospitals. Obstet Gynecol 2009;113:1239–47.
13. Rosenberg T, Garbers S, Chavkin W, Chiasson M. Prepregnancy weight and adverse perinatal outcomes in an ethnically diverse population. Obstet Gynecol 2003;102:1022–7.
14. Rosenberg T, Garbers S, Lipkind H, Chiasson M. Maternal obesity and diabetes as risk factors for adverse pregnancy outcomes: differences among 4 racial/ethnic groups. Am J Public Health 2005;95:1545–51.
15. Breslow NE, Day NE. Statistical methods in cancer research. Volume II–the design and analysis of cohort studies. IARC Sci Publ 1987;1–406.
16. Kahn ML. Eisenmenger's syndrome in pregnancy. N Engl J Med 1993;329:887.
17. Kim S, Dietz P, England L, Morrow B, Callaghan W. Trends in pre-pregnancy obesity in nine states, 1993-2003. Obesity 2007;15:986–93.
18. Berg C, Mackay A, Qin C, Callaghan W. Overview of maternal morbidity during hospitalization for labor and delivery in the United States: 1993-1997 and 2001-2005. Obstet Gynecol 2009;113:1075–81.
19. Berg C, Callaghan W, Syverson C, Henderson Z. Pregnancy-related mortality in the United States, 1998 to 2005. Obstet Gynecol 2010;116:1302–9.
20. Melnick SL, Sherer R, Louis TA, Hillman D, Rodriguez EM, Lackman C, et al.. Survival and disease progression according to gender of patients with HIV infection. The Terry Beirn Community Programs for Clinical Research on AIDS. JAMA 1994;272:1915–21.
22. Callaghan W, Berg C. Pregnancy-related mortality among women aged 35 years and older, United States, 1991-1997. Obstet Gynecol 2003;102:1015–21.
23. Jacobsson B, Ladfors L, Milsom I. Advanced maternal age and adverse perinatal outcome. Obstet Gynecol 2004;104:727–33.
24. Bianco A, Stone J, Lynch L, Lapinski R, Berkowitz G, Berkowitz RL. Pregnancy outcome at age 40 and older. Obstet Gynecol 1996;87:917–22.
25. Chang J, Elam Evans L, Berg C, Herndon J, Flowers L, Seed K, et al.. Pregnancy-related mortality surveillance–United States, 1991–1999. MMWR Surveillance Summ 2003;52:1–8.
26. Berg C, Chang J, Callaghan W, Whitehead S. Pregnancy-related mortality in the United States, 1991-1997. Obstet Gynecol 2003;101:289–96.
27. Sjöström L, Lindroos A, Peltonen M, Torgerson J, Bouchard C, Carlsson B, et al.. Lifestyle, diabetes, and cardiovascular risk factors 10 years after bariatric surgery. N Engl J Med 2004;351:2683–93.