Hypertension is a common consequence of coarctation of the aorta. The frequency of hypertensive complications of pregnancy in women with coarctation in the general population is undefined. In this study, we used the 1998 to 2007 Nationwide Inpatient Sample, a nationally representative data set, to identify patients admitted to an acute care hospital for delivery. The frequency of hypertensive complications of pregnancy was compared between women with and without coarctation. Secondary outcomes, including length of stay, hospital charges, Caesarean delivery, and adverse maternal outcomes, were also assessed. There were an estimated 697 deliveries among women with coarctation, compared to 42,601,409 deliveries by women without coarctation. The frequency of hypertensive complications of pregnancy was 24.1 ± 3.3% for women with coarctation compared to 8.0 ± 0.1% for women without coarctation (multivariate odds ratio [OR] 3.6, 95% confidence interval [CI] 2.5 to 5.2). Preexisting hypertension complicating pregnancy (10.2 ± 2.5% vs 1.0% ± 0.02%, multivariate OR 10.8, 95% CI 5.9 to 19.8) and pregnancy-induced hypertension (13.9 ± 3.0% vs 7.0% ± 0.1%, multivariate OR 2.1, 95% CI 1.3 to 3.3) were more common in women with coarctation. Women with coarctation were more likely to deliver by Caesarean section (41.6 ± 3.3% vs 26.4% ± 0.2%, multivariate OR 2.0, 95% CI 1.4 to 2.8), have adverse cardiovascular outcomes (4.8 ± 2.2% vs 0.3 ± 0.01%, multivariate OR 16.7, 95% CI 6.7 to 41.5), have longer hospital stays, and incur higher hospital charges (both p values <0.0001) than women without coarctation. In conclusion, women with coarctation are more likely to have hypertensive complications of pregnancy, deliver by Caesarean section, have adverse cardiovascular outcomes, have longer hospitalizations, and incur higher hospital charges than women without coarctation.
Pregnancy often brings women born with congenital cardiovascular disease, including those with coarctation of the aorta (CoA), to heightened medical attention. Normally during pregnancy systemic arterial blood pressure decreases through the second trimester. In women with CoA, however, hypertension can manifest for the first time and preexisting hypertension can be exacerbated during pregnancy. Published series suggest that CoA confers an elevated risk for hypertensive complications of pregnancy, but these reports focus on the experience of individual referral centers or lack comparisons to unaffected women. In this study we used a nationally representative database of hospitalizations to compare the frequency of hypertensive complications of pregnancy between women with and without CoA to establish the burden of hypertensive complications of pregnancy as well as their relation to resource utilization in this population.
Methods
We analyzed Nationwide Inpatient Sample (NIS) data from 1998 to 2007. The NIS, a subset of the Healthcare Cost and Utilization Project, is the largest publicly available all-payer inpatient database in the United States, containing data from approximately 1,000 hospitals and 7 million to 8 million hospital stays annually. The NIS approximates a 20% stratified sample from a sampling frame that comprises 90% of United States acute care hospital admissions. Because of its large sample size, the NIS is useful for the study of rare conditions such as CoA and has been used previously to study congenital heart disease as well as complications of pregnancy.
Our study cohort included women admitted to acute care hospitals with International Classification of Diseases, Ninth Revision (ICD-9), codes indicating vaginal delivery or Caesarean section, as described previously. The analysis was limited to delivery-associated hospitalizations to avoid including multiple hospitalizations for a given patient over the course of a pregnancy.
The primary exposure variable was CoA (ICD-9 code 747.1), and the primary outcome was a hypertensive disorder of pregnancy (ICD-9 code 642). This code encompasses preexisting hypertension complicating pregnancy (ICD-9 codes 642.0 to 642.2), as well as the spectrum of pregnancy-induced hypertension, which includes the nonproteinuric transient hypertension of pregnancy (ICD-9 code 642.3) as well as the proteinuric syndromes of mild preeclampsia (ICD-9 code 642.4), severe preeclampsia (ICD-9 code 642.5), preeclampsia in the setting of preexisting hypertension (ICD-9 code 642.7), and eclampsia (ICD-9 code 642.6) ( Figure 1 ). Because of the sample size, the spectrum of proteinuric syndromes were combined and analyzed as a single variable (referred to as preeclampsia given the very low frequency of eclampsia). Secondary outcomes for the study were length of stay (LOS), total hospital charges, proportion delivering by Caesarean section, and a composite cardiovascular outcome consisting of death, heart failure (ICD-9 code 428), arrhythmia (ICD-9 code 427), cerebrovascular accident (stroke or transient ischemic attack; ICD-9 codes 431 and 433 to 436), other embolic events (ICD-9 codes 415.1, 444, 445, and 673), and unspecified cardiovascular complications of pregnancy (ICD-9 code 668.1).
Covariates of interest included reported risk factors for hypertensive complications of pregnancy: maternal age, multiple gestation, obesity, rheumatologic disease, chronic kidney disease, diabetes mellitus (gestational or preexisting), and migraine headaches. We also adjusted for the presence of a maternal chromosomal abnormality, hospital teaching status, year of hospital admission, and number of medical co-morbidities defined by a comprehensive set of co-morbidities described by Elixhauser et al. We excluded cardiovascular co-morbidities, because these could be directly related to CoA. Because very few women had >1 co-morbidity, we present dichotomous data on the presence or absence of co-morbidities.
Continuous and categorical demographic and clinical variables are presented as mean or percentage ± SE, and univariate linear regression and the Rao-Scott chi-square test, respectively, were used to compare values by CoA status. We performed univariate and multivariate linear regression (adjusting for the covariates described above) to determine the relation of CoA to log-transformed LOS in days and log-transformed total hospital charges. Univariate and multivariate logistic regression models were performed to determine the relation between CoA and subsets of hypertensive complications of pregnancy, Caesarean section, long LOS (top quintile), high total hospital charges (top quintile), and the combined cardiovascular outcome. Statistical analyses were performed using SAS for Windows version 9.2 (SAS Institute Inc., Cary, North Carolina). All analyses were performed using the provided sample weights and accounting for the complex survey design and clustering by hospital.
Results
From 1998 to 2007, there were an estimated 697 ± 46 deliveries by women with CoA compared to 42,472,821 ± 1,017,967 deliveries in women without diagnoses of CoA. Demographic and clinical characteristics were similar between the 2 groups, including age, insurance status, proportion with co-morbidities, preexisting diabetes, gestational diabetes, maternal chromosomal abnormality, and multiple gestation. Women with CoA were more likely to deliver at a teaching hospital ( Table 1 ).
| Variable | CoA | No CoA | p Value |
|---|---|---|---|
| Number (unweighted) | 141 | 8,746,978 | |
| Number (weighted) | 697 ± 46 | 42,472,821 ± 1,017,967 | |
| Age (years) (mean ± SE) | 27.0 ± 0.5 | 27.4 ± 0.1 | 0.48 |
| Age group (years) | 0.37 | ||
| <18 | 8.1 ± 3.9 | 6.5 ± 0.1 | |
| 18–34 | 84.2 ± 3.1 | 82.3 ± 0.1 | |
| >34 | 7.8 ± 2.1 | 11.2 ± 0.2 | |
| Insurance ⁎ | 56.0 ± 4.4 | 55.0 ± 0.8 | 0.28 |
| Private | |||
| Public | 41.6 ± 4.4 | 38.7 ± 0.7 | |
| Other | 2.3 ± 1.6 | 6.2 ± 0.3 | |
| Preexisting diabetes mellitus | 2.1 ± 1.2 | 0.9 ± 0.02 | 0.11 |
| Gestational diabetes mellitus | 4.9 ± 1.8 | 4.4 ± 0.1 | 0.80 |
| Delivered at teaching hospital | 70.7 ± 2.4 | 46.2 ± 1.2 | <0.0001 |
| Co-morbidities ≥1 † | 19.2 ± 2.9 | 12.0 ± 0.2 | <0.01 |
| Multiple gestation | 1.4 ± 1.0 | 1.6 ± 0.03 | 0.86 |
⁎ Public insurance includes Medicaid and Medicare. Other insurance includes no insurance, self-pay, and miscellaneous.
† Indicates ≥1 medical co-morbidity as defined by Elixhauser et al’s comprehensive set of co-morbidities.
Admissions for delivery increased at a higher rate among women with CoA than in the general population from 1998 and 2007. Over that time period, the estimated number of admissions for delivery of women with CoA increased 2.7-fold. In contrast, the estimated number of admissions for delivery among women without diagnoses of CoA increased <1.2-fold (p <0.001; Figure 2 ).
Hypertensive complications of pregnancy were more common in women with CoA than in women without CoA, occurring in 24.1 ± 3.3% of pregnancies associated with CoA, compared to 8.0 ± 0.1% of pregnancies in women without CoA (univariate odds ratio [OR] 3.7, 95% confidence interval [CI] 2.5 to 5.3, multivariate OR 3.6, 95% CI 2.5 to 5.2; Table 2 ).
| Variable | Coarctation | Univariate OR (95% CI) | Multivariate OR (95% CI) | |
|---|---|---|---|---|
| Yes | No | |||
| Hypertensive complications of pregnancy | 24.1 ± 3.3 | 8.0 ± 0.1 | 3.7 (2.5–5.3) | 3.6 (2.5–5.2) |
| Preexisting hypertension | 10.2 ± 2.5 | 1.0 ± 0.02 | 10.8 (6.3–18.3) | 10.8 (5.9–19.8) |
| Pregnancy-induced hypertension | 13.9 ± 3.0 | 7.0 ± 0.1 | 2.2 (1.4–3.6) | 2.1 (1.2–3.3) |
| Preeclampsia | 4.3 ± 1.7 | 3.6 ± 0.1 | 1.2 (0.6–2.8) | 1.1 (0.5–2.4) |
| Other | 9.6 ± 2.6 | 3.4 ± 0.04 | 3.1 (1.7–5.4) | 3.0 (1.7–5.5) |
| Caesarean delivery | 41.6 ± 3.6 | 26.4 ± 0.2 | 2.0 (1.5–2.7) | 2.0 (1.4–2.8) |
| LOS, top quintile ⁎ (%) | 27.7 ± 18.9 | 13.4 ± 0.2 | 2.6 (1.8–3.7) | 1.7 (1.1–2.6) |
| Total charges, top quintile ⁎ (%) | 38.1 ± 4.8 | 19.3 ± 0.6 | 2.6 (1.8–3.8) | 2.0 (1.3–3.1) |
| Combined cardiovascular outcome † | 4.8 ± 2.2 | 0.3 ± 0.01 | 18.0 (7.2–45.2) | 16.7 (6.7–41.5) |
⁎ Top quintile does not constitute exactly 20% in the general population because of sample weights and complex survey design. The top quintile for total charges includes admissions with total charges ≥$10,470. The precision of LOS quintile is limited (38.1 ± 0.3%, 13.4 ± 0.2%, and 4.9 ± 0.1% of women stayed ≥3, 4, and 5 days, respectively).
† The combined cardiovascular outcome includes maternal death, heart failure, arrhythmia, cerebrovascular events, other embolic events, and unspecified cardiovascular complications of pregnancy.
Preexisting hypertension complicating pregnancy and pregnancy-induced hypertension were both more common in pregnancies associated with CoA: preexisting hypertension complicated 10.2 ± 2.5% of pregnancies associated with CoA, compared to 1.0 ± 0.02% of pregnancies not associated with CoA (univariate OR 10.8, 95% CI 6.3 to 18.3, multivariate OR 10.8, 95% CI 5.9 to 19.8). Pregnancy-induced hypertension occurred in 13.9 ± 3.0 of pregnancies associated with CoA compared to 7.0 ± 0.1% of pregnancies not associated with CoA (univariate OR 2.2, 95% CI 1.4 to 3.6, multivariate OR 2.1, 95% CI 1.2 to 3.3).
Pregnancy-induced hypertension may be nonproteinuric (transient hypertension of pregnancy) or be associated with proteinuria (preeclampsia and eclampsia). The frequency of nonproteinuric pregnancy-induced hypertension was higher in women with CoA (9.6 ± 2.6% vs 3.4 ± 0.04%, univariate OR 3.1, 95% CI 1.7 to 5.4, multivariate OR 3.0, 95% CI 1.7 to 5.5). The incidence of proteinuric pregnancy-induced hypertension (preeclampsia and eclampsia) was not significantly higher for women with CoA (4.3 ± 1.7% vs 3.6 ± 0.1%, univariate OR 1.2, 95% CI 0.6 to 2.8, multivariate OR 1.1, 95% CI 0.5 to 2.4).
CoA was associated with significantly longer LOS ( Figure 3 ). The mean LOS was 3.5 ± 0.3 days (median 3, interquartile range 2 to 4) for women with CoA compared to 2.6 ± 0.01 days (median 2, interquartile range 2 to 3) for women without CoA (multivariate p = 0.03). Women with CoA were significantly more likely to be in the top LOS quintile than women without CoA (27.7 ± 18.9% vs 13.4 ± 0.2%, multivariate OR 1.7, 95% CI 1.1 to 2.6; Table 2 ).
