Objective
The objective of this study was to estimate the prevalence and temporal trends of medical conditions among women with stillbirth and to determine the effect of medical comorbidities on the trend of stillbirth.
Study Design
The Nationwide Inpatient Sample (NIS) for the years 2008–2010 was first queried for all delivery-related discharges. A multivariable logistic regression model was constructed with adjusted odds ratios (ORs) and 95% confidence intervals (CIs) calculated for medical conditions among women with stillbirth. The NIS was then queried for the years 2000–2010, and the effect of maternal medical conditions on the stillbirth rate was estimated.
Results
From 2008 to 2010, there were 51,080 deliveries to women with stillbirth, giving a rate of 4.08 per 1000 live births. Women with stillbirth were more likely to be African American (OR, 2.12; 95% CI, 2.07–2.17), with an age less than 25 years (OR, 1.19; 95% CI, 1.16–1.22) or older than 35 years (OR, 1.40; 95% CI, 1.37–1.44) compared with women without stillbirth. Medical conditions such as cardiac, rheumatological, and renal disorders; hypertension; diabetes; thrombophilia; and drug, alcohol and tobacco use, were independent predictors of fetal demise in multivariable logistic regression modeling. From 2000 to 2010, despite an increase in the total number of births to women with comorbidities, there was a significant decrease in the stillbirth rate, which was more pronounced among women with comorbidities compared with women without comorbidities ( P = .021).
Conclusion
From 2000 to 2010, there was a significantly greater decrease in the stillbirth rate among women with maternal medical conditions than there was among women without comorbidities. These findings occurred despite an overall increase in the number of pregnancies to women with medical comorbidities over the time period. Because the NIS does not include information on gestational age, birthweight, or whether subjects had antepartum testing, we are not able to determine the effect of these variables on the observed outcomes.
Stillbirth is defined as intrauterine fetal death after 20 weeks’ gestation or 350 g birthweight and the United States stillbirth rate in 2006 was 6.05 per 1000 live births. Over the last 20 years, there has been a slight decline in the rate of stillbirth in the United States; however, the rate is still nearly 50% higher than the Healthy People 2010 target goal of 4.1 stillbirths per 1000 live births. Additionally, there are still significant racial and ethnic disparities, with stillbirth rates among non-Hispanic blacks greater than 2 times the rate of non-Hispanic whites. This disparity is believed to be multifactorial in nature and largely remains unexplained.
Although nearly 25% of all stillbirths remain unexplained, of those that have a probable cause, maternal medical conditions contribute to nearly 1 in 5 stillbirths. The Stillbirth Collaborative Research Network recently collected detailed information on 614 cases of stillbirth from 59 hospitals across the United States and identified maternal sociodemographic and medical factors associated with stillbirth. Non-Hispanic black race/ethnicity, advanced maternal age, obesity, nulliparity, multiple gestation, diabetes, and a history of previous stillbirth were all associated with stillbirth. Most larger studies determining the association of risk factors with stillbirth are limited to vital statistics with relatively limited data. Furthermore, it is difficult to determine the effect of maternal medical conditions on the changing rate of stillbirth with vital statistics data.
The objectives of this study were to estimate the prevalence of maternal medical conditions associated with stillbirth using a large national administrative database and then to estimate the impact of the change in prevalence of maternal medical comorbidities on the change in the rate of stillbirth over the years 2000–2010.
Materials and Methods
Study design
The study was reviewed and approved by the Duke University Institutional Review Board as exempt research. The Nationwide Inpatient Sample (NIS) from the Healthcare Cost and Utilization Project of the Agency for Healthcare Research and Quality was queried for all pregnancy-related discharges.
The NIS is a national database that contains hospital discharge data from more than 1000 hospitals in 45 states (2010) and represents the largest all-payer database for US hospital admissions. The hospitals in the NIS are stratified based on ownership, bed size, teaching status, urban/rural location, and region. From each strata, the NIS contains approximately 20% of US hospitals and includes 100% of the discharges from those hospitals. From this sampling, the NIS provides weights to each entry that allows for national estimates to be produced.
Because previous studies have shown that the prevalence of severe comorbid conditions has been changing rapidly among delivery hospitalizations, we examined maternal characteristics and risk of medical and obstetric complications among hospitalizations with and without stillbirth for the last 3 years of NIS data analyzed (2008–2010). This 3 year data interval has been chosen to ensure an adequate sample size.
Next, to determine the association of medical complications on stillbirth, we analyzed data over the 11 year period from 2000 to 2010. Using the NIS for the years 2008–2010, all records containing a delivery discharge were identified. An admission for delivery was defined as any discharge record that included a delivery code ( International Classification of Diseases , ninth revision [ICD-9] codes 74.x for cesarean delivery [except 74.91]; V27, 72.x, 73.x, and 650 for general delivery codes [not utilized to specify vaginal or cesarean]).
Deliveries were also identified by diagnosis-related group codes. Diagnosis-related group codes 765 and 766 were utilized to identify cesarean deliveries and codes 767, 768, 774, and 775 for vaginal deliveries. The ICD-9 codes used to identify discharges with stillbirth were 656.4x. Women without a code for stillbirth were assumed to have had a live birth. For comorbidities, both the ICD-9 code for a particular condition in pregnancy and the general ICD-9 code for that condition were used. If the pregnancy-related code was not specific, it was not used (a list of ICD-9 codes utilized is presented in the Appendix , Supplementary Table ). Race/ethnicity status is available in the NIS and is listed as follows: white, black, Hispanic, Asian or Pacific Islander, Native American, other, or unknown. The sampling frame for events was limited to delivery admissions only.
Statistics
Data were weighted based on the NIS sampling design. Two-way χ 2 tests generated cell frequencies and SDs for demographics, medical conditions and events, and pregnancy-related complications. Logistic regression analyses were used to compute the odds ratios (ORs) with 95% confidence intervals for age, race, medical conditions and events, and pregnancy-related complications among women with stillbirth compared with women without stillbirth.
Our a priori analysis plan was to determine the associated independent odds of stillbirth for medical conditions while controlling for demographics and other present medical conditions. A multivariable logistic regression model was created to determine the adjusted odds of stillbirth for each studied medical and obstetric condition. To do this, we calculated the adjusted OR for each medical or obstetric condition found to be significantly different between the 2 groups in a univariate analysis while controlling for demographic variables also found to be significant between women with stillbirth and those without.
In addition, while determining the adjusted OR for each medical or obstetric condition, we controlled for all other medical and obstetric conditions that were tested to estimate the risk for stillbirth for each condition. Insurance status and the proportion of women residing in a ZIP code in the bottom quartile were found to be correlated, so only insurance status was included in the final model.
Using the NIS for the years 2000–2010, the total number of births and the number of births to women with and without comorbidities was then determined. Women were considered to have a comorbidity if they had at least 1 of the following: cardiomyopathy, congenital heart disease, valvular heart disease, cardiac conduction disorders, asthma, pulmonary hypertension, diabetes mellitus, thyroid disorders, systemic lupus erythematosus, rheumatoid arthritis/collagen vascular disease, human immunodeficiency virus (HIV), thrombophilia/antiphospholipid antibody syndrome, anemia, thrombocytopenia, sickle cell disease, thalassemia, drug or alcohol use, tobacco use, chronic hypertension, chronic renal failure, myocardial infarction, acute heart failure, cardiac arrest, pneumonia, pulmonary edema, acute respiratory distress syndrome, pulmonary embolism, deep vein thrombosis, sepsis, influenza, acute renal failure, multiple gestation, gestational diabetes, preeclampsia/hypertensive disorders of pregnancy, and fetal growth restriction.
A linear regression model was developed to estimate the linear change in the number of total births and number of births to women with and without comorbidities during this time period.
Next, the linear change in the overall stillbirth rate was calculated using data from the NIS for the years 2000–2010. Stillbirth rate was defined as the number of stillbirths per total live births per year. The change in the stillbirth rate among women with and without comorbidities was then determined. To determine whether the change in the rate of stillbirths to women with comorbidities differed from that with women without comorbidities, the slope of the 2 linear regression models were compared.
A statistically significant difference in the slopes from each model suggested that the rate in stillbirths between each group differed over the study time period. Statistical significance for all analyses was assigned as a value of P < .05. Analyses were performed using SAS version 9.3 (SAS Institute Inc, Cary, NC) and GraphPad Prism version 6.0 for Macintosh (GraphPad Software, San Diego, CA).
Results
During the period from 2008 to 2010, there were 12,524,119 delivery records in the NIS. Of these deliveries, there were 51,075 delivery records with an ICD-9 code for stillbirth (0.41%). Women with a stillbirth code were more likely to be African American (OR, 2.12; 95% confidence interval [CI], 2.07–2.17) or Hispanic (OR, 1.09; 95% CI, 1.06, 1.12) compared with those with a live birth. Race/ethnicity status was missing in 15.5% of the delivery records and did not differ by stillbirth status. Women with a stillbirth were more likely to be under age 24 years or aged 35 years and older. Women with a stillbirth were also more likely to reside within a ZIP code in which the median household income was in the lowest quartile (between $1 and $38,999 [OR, 1.39; 95% CI, 1.37–1.42]) and were less likely to have private insurance (OR, 0.69; 95% CI, 0.68–0.70) ( Table 1 ).
Variable | Stillbirth (n = 51,075) | Live birth (n = 12,473,044) | Unadjusted OR (95% CI) | P value |
---|---|---|---|---|
Race/ethnicity, n, % | ||||
Caucasian (referent) | 19,142 (37.5) | 5,514,822 (44.2) | 1.0 | — |
African American | 10,943 (21.4) | 1,479,890 (11.9) | 2.12 (2.07–2.17) | < .0001 |
Hispanic | 9040 (17.7) | 2,396,294 (19.2) | 1.09 (1.06–1.12) | < .0001 |
Asian/Pacific Islander | 1636 (3.2) | 554,846 (4.4) | 0.85 (0.81–0.90) | < .0001 |
Native American | 430 (0.84) | 94,803 (0.76) | 1.30 (1.18–1.43) | .0372 |
Other | 1947 (3.8) | 500,492 (4.0) | 1.12 (1.07–1.18) | .0046 |
Missing | 7938 (15.5) | 1,931,896 (15.5) | — | — |
Age, y | ||||
15-24 | 18,297 (35.8) | 4,195,809 (33.6) | 1.19 (1.16–1.22) | < .0001 |
25-29 (referent) | 12,845 (25.1) | 3,505,305 (28.1) | 1.0 | — |
30-34 | 10,320 (20.2) | 2,905,125 (23.3) | 0.97 (0.94–0.99) | < .0001 |
≥35 | 9485 (18.6) | 1,837,864 (14.7) | 1.40 (1.37–1.44) | < .0001 |
Missing | 129 (0.3) | 28,942 (0.2) | — | — |
Age, y a | 27.8 ± 15.1 | 27.6 ± 13.7 | — | .006 |
Private insurance, n, (%) | 20,740 (40.6) | 6,000,743 (48.1) | 0.69 (0.68–0.70) | < .0001 |
Low income ZIP code, n, % b | 16,986 (33.3) | 3,288,921 (26.4) | 1.39 (1.37–1.42) | < .0001 |
Length of stay, d c | 2 (1, 3) | 2 (2, 3) | — | < .0001 c |
Total charges, $ c | 11,254 (7331, 17,778) | 10,019 (6807, 15,102) | — | < .0001 c |
b Low-income ZIP code defined as median household income in subject’s ZIP code $1–38,999 (lowest quartile)
Table 2 describes medical conditions among women with stillbirth compared with women with a live birth. Women with a stillbirth had higher odds of cardiac disease such as cardiomyopathy, congenital heart disease, and cardiac conduction disorders with cardiomyopathy, demonstrating the greatest association (OR, 4.36; 95% CI, 3.56–5.32).
Condition, n, % | Stillbirth (n = 51,075) | Live birth (n = 12,473,044) | Unadjusted OR (95% CI) | P value |
---|---|---|---|---|
Heart disease | ||||
Cardiomyopathy | 97 (0.19) | 5462 (0.04) | 4.36 (3.56–5.32) | < .0001 |
Valvular heart disease | 189 (0.37) | 54,464 (0.44) | 0.85 (0.73–0.98) | .025 |
Congenital heart disease | 58 (0.11) | 10,602 (0.08) | 1.35 (1.04–1.74) | .023 |
Conduction disorders | 481 (0.94) | 80,338 (0.64) | 1.47 (1.34–1.61) | < .0001 |
Pulmonary disease | ||||
Asthma | 1724 (3.37) | 403,370 (3.23) | 1.05 (0.99–1.10) | .063 |
Pulmonary, circ/pulm HTN | 19 (0.037) | 2768 (0.022) | 1.66 (1.05–2.61) | .029 |
Endocrine disorders | ||||
Diabetes (nongestational) | 2056 (4.02) | 128,914 (1.03) | 4.02 (3.84–4.20) | < .0001 |
Thyroid disorder | 1236 (2.42) | 296,306 (2.37) | 1.02 (0.96–1.08) | .510 |
Autoimmune disorders | ||||
Systemic lupus erythem | 218 (0.43) | 14,132 (0.11) | 3.78 (3.30–4.32) | < .0001 |
Rheum arth/collagen vasc | 92 (0.18) | 15,309 (0.12) | 1.48 (1.20–1.81) | .0002 |
HIV | 35 (0.07) | 3581 (0.03) | 2.41 (1.73–3.36) | < .0001 |
Hematological disorders | ||||
Thrombophilia/APS | 533 (1.04) | 63,984 (0.51) | 2.05 (1.88–2.23) | < .0001 |
Anemia | 5580 (10.9) | 1,349,763 (10.8) | 1.01 (0.98–1.04) | .450 |
Thrombocytopenia | 824 (1.61) | 112,784 (0.90) | 1.80 (1.68–1.92) | < .0001 |
Sickle cell | 123 (0.24) | 18,871 (0.15) | 1.60 (1.34–1.91) | < .0001 |
Drugs/alcohol/tobacco | ||||
Drug use | 1629 (3.2) | 160,637 (1.3) | 2.52 (2.40–2.65) | < .0001 |
Alcohol use | 160 (0.31) | 13,491 (0.11) | 2.90 (2.48–3.39) | < .0001 |
Tobacco | 4176 (8.2) | 782,806 (6.3) | 1.33 (1.29, 1.37) | < .0001 |
Chronic hypertension | 2465 (4.8) | 242,486 (2.0) | 2.56 (2.46, 2.66) | < .0001 |
Chronic renal failure | 135 (0.26) | 4782 (0.04) | 6.90 (5.81, 8.19) | < .0001 |
Women with stillbirth were also more likely to have diabetes (OR, 4.02; 95% CI, 3.84–4.20), systemic lupus erythematosus (OR, 3.78; 95% CI, 3.30–4.32), HIV (OR, 2.41; 95% CI, 1.73–3.36), thrombophilia or antiphospholipid antibody syndrome (OR, 2.05; 95% CI, 1.88–2.23) and chronic hypertension (OR, 2.56; 95% CI, 2.46, 2.66), or chronic renal failure (OR, 6.90; 95% CI, 5.81–8.19). Drug, alcohol, and tobacco use were all more common among women with stillbirth compared with women without ( Table 2 ).
Medical complications present at delivery were common among women with stillbirth. Of the 51,075 women with stillbirths, 76 occurred to women with a maternal death during the admission for delivery. Because of limitations in the NIS, the timing of the maternal death in relation to the stillbirth is not known. There was an increased odds of cardiac events such as myocardial infarction (OR, 9.95; 95% CI, 5.92–16.71) or cardiac arrest (OR, 14.84; 95% CI, 10.97–20.07) among women with a stillbirth.
Pulmonary events such as acute respiratory distress syndrome (OR, 12.25; 95% CI, 10.30–14.57), pulmonary edema (OR, 7.66; 95% CI, 5.94–9.89), and pneumonia (OR, 3.92; 95% CI, 3.39–4.53) were all more common among women with a stillbirth compared with those with a live birth. Additionally, women with a stillbirth had increased odds of pulmonary embolism, deep venous thrombosis, sepsis, and acute renal failure ( Table 3 ).
Condition, n, % | Stillbirth (n = 51,075) | Live birth (n = 12,473,044) | Unadjusted OR (95% CI) | P value |
---|---|---|---|---|
Maternal death | 76 (0.1) | 835 (0.007) | 22.34 (17.67–28.25) | < .0001 |
Mechanical ventilation | 481 (0.94) | 8277 (0.07) | 14.33 (13.06–15.71) | < .0001 |
Transfusion | 2631 (5.2) | 127,035 (1.0) | 5.28 (5.07–5.49) | < .0001 |
Cardiac events | ||||
Myocardial infarction/ischemia | 15 (0.029) | 365 (0.003) | 9.95 (5.92–16.71) | < .0001 |
Acute heart failure | 78 (0.15) | 4359 (0.04) | 4.39 (3.51–5.49) | < .0001 |
Cardiac arrest/ventricular fibrillation | 44 (0.086) | 737 (0.006) | 14.84 (10.97–20.07) | < .0001 |
Pulmonary events | ||||
Pneumonia | 187 (0.37) | 11,703 (0.09) | 3.92 (3.39–4.53) | < .0001 |
Pulmonary edema | 61 (0.12) | 1946 (0.02) | 7.66 (5.94–9.89) | < .0001 |
Acute respiratory distress syndrome | 135 (0.26) | 2691 (0.02) | 12.25 (10.30–14.57) | < .0001 |
Thromboembolic events | ||||
Pulmonary embolism | 69 (0.13) | 3371 (0.03) | 5.06 (4.00–6.42) | < .0001 |
Deep vein thrombosis | 72 (0.14) | 6113 (0.05) | 2.89 (2.29–3.64) | < .0001 |
Infections | ||||
Sepsis | 300 (0.59) | 6004 (0.05) | 12.29 (10.94–13.80) | < .0001 |
Pyelonephritis | 78 (0.15) | 20,594 (0.16) | 0.93 (0.75–1.16) | .53 |
Influenza | 15 (0.029) | 4207 (0.034) | 0.90 (0.52–1.42) | .69 |
Renal event | ||||
Acute renal failure | 517 (1.01) | 6386 (0.05) | 20.00 (18.28–21.88) | < .0001 |
The odds of obstetric events at admission for delivery were influenced by stillbirth. Women with stillbirth had higher odds of multiple gestation, hypertensive disorders of pregnancy, placental abruptions, fetal growth restriction, postpartum hemorrhage, chorioamnionitis, and transfusion. Both cesarean delivery (OR, 0.47; 95% CI, 0.46–0.48) and gestational diabetes (OR, 0.74; 95% CI, 0.71–0.77) were less common among women with stillbirth compared with women with live births ( Table 4 ).
Condition, n, % | Stillbirth (n = 51,075) | Live birth (n = 12,473,044) | Unadjusted OR (95% CI) | P value |
---|---|---|---|---|
Obstetric events | ||||
Cesarean delivery | 9420 (18.4) | 4,031,937 (32.3) | 0.47 (0.46–0.48) | < .0001 |
Operative vaginal delivery | 998 (1.9) | 791,245 (6.3) | 0.29 (0.28–0.31) | < .0001 |
Multiple gestation | 1999 (3.9) | 264,753 (2.1) | 1.88 (1.80–1.96) | < .0001 |
Gestational diabetes | 2168 (4.2) | 706,904 (5.7) | 0.74 (0.71–0.77) | < .0001 |
Preeclampsia, eclampsia, gestational hypertension | 4872 (9.5) | 918,590 (7.4) | 1.33 (1.29–1.37) | < .0001 |
Fetal growth restriction | 1912 (3.7) | 269,566 (2.2) | 1.76 (1.68–1.84) | < .0001 |
Postpartum hemorrhage | 2106 (4.1) | 317,357 (2.5) | 1.65 (1.58–1.72) | < .0001 |
Chorioamnionitis | 3438 (6.7) | 320,038 (2.6) | 2.74 (2.65–2.84) | < .0001 |
Placental abruption | 5421 (10.6) | 129,348 (1.0) | 11.33 (11.01–11.66) | < .0001 |