Materials and Methods
The Nationwide Inpatient Sample (NIS) from the Healthcare Cost and Utilization Project of the Agency for Healthcare Research and Quality was queried for pregnancy-related discharges from January 2008 through December 2010. The NIS contains information from approximately 8 million hospital stays from over 1000 hospitals in 45 states and is the largest all-payer inpatient care database in the United States. The sampling frame uses 5 strata including ownership, hospital size, geographic region, teaching status, and location (urban or rural). The NIS is sampled to approximate a 20% stratified sample of US community hospitals. The data are weighted to generate nationwide estimates based on probabilities proportional to the number of hospitals in each stratum. This accounts for changes in the number of states included over the years and the different hospitals sampled over time. The sampling frame comprises over 96% of the US population.
Information contained in the NIS includes discharge diagnoses and procedure codes based on the International Classification of Diseases, 9th revision, Clinical Modifications (ICD-9-CM). The Agency for Healthcare Research and Quality receives data from individual states and insures data values are valid and internally consistent. Further information about the NIS can be found elsewhere. The information included in the NIS contains safeguards to protect the privacy of individual patients, physicians, and hospitals. Although the nature of the data is limited to discharge diagnoses and demographic information, the NIS allows for the study of relatively rare conditions. Because the NIS excludes data elements that could directly or indirectly identify individuals, the study was deemed exempt by the Duke University Health System Institutional Review Board (eIRB approval no. Pro00045222).
Using the NIS for each of the years 2008-2010, all delivery-related discharge records were identified. We focused only on admissions for delivery to eliminate the possibility of counting a woman more than once per pregnancy in the analysis. There is a chance that over the 3-year period, 1 woman may have had 2 pregnancies and therefore be represented twice in the study. Nonetheless, outcomes are unique for each pregnancy and therefore there is value in counting 1 woman with 2 different pregnancies during that period. An admission for delivery was defined as any discharge record that included a delivery code (ICD-9-CM codes 74.x [except 74.91] for cesarean delivery and V27, 72.x, 73.x, and 650-659 for general delivery codes). Deliveries were also identified by diagnosis-related group codes. Diagnosis-related group codes 765 and 766 were used to identify cesarean deliveries and codes 767, 768, 774, and 775 for vaginal deliveries.
Women with CF were identified by an ICD-9-CM code of 277.0x and were compared with women without CF. For comorbidities, both the ICD-9-CM for a particular condition in pregnancy (ie, 6xx code) and the general ICD-9-CM codes for that condition were used ( Appendix ; Supplementary Table for ICD-9-CM codes used). A composite CF outcome variable was created that included any one of the following: death, mechanical ventilation, sepsis, pneumonia, acute respiratory failure, acute respiratory distress syndrome, or acute renal failure. Data were weighted by the sampling weights provided by the NIS. Continuous variables were compared between pregnant women with CF and pregnant women without CF using Student t test or Wilcoxon sign rank tests where appropriate. Logistic regression was used to calculate odds ratios (ORs) and 95% confidence intervals (95% CIs) for medical conditions and medical and obstetric outcomes among women with CF at delivery compared with women without CF at delivery. Next, a multivariable logistic regression model was constructed for medical and obstetric outcomes among women with CF whereas controlling for age, race/ethnicity, diabetes, gestational diabetes, hypertension, preeclampsia, multiple gestation, and mode of delivery.
To determine the change in prevalence of CF at delivery, delivery admissions were identified using the NIS from the years 2000 through 2010. A period of 11 years was chosen to allow for assessment of change over a longer time frame. For each year, the number of women with CF per 100,000 deliveries was calculated. Linear regression was then used to determine whether there was a significant linear change in the number of women with CF at delivery over the 11-year time period.
Statistical significance was assigned as a P value of < .05. Analyses were performed using SAS version 9.3 (SAS Institute Inc, Cary, NC) and GraphPad Prism version 6.0 for Macintosh (GraphPad Software, La Jolla, CA).
Results
Over the years 2000 to 2010, there was a significant linear increase in the number of women with CF at delivery from 2.99 per 100,000 deliveries in 2000 to 9.84 per 100,000 in 2010 ( P < .0001; R 2 = 0.91; Figure ). The years 2008 to 2010 were then used to determine characteristics of women with CF at delivery. Between 2008 and 2010, there were 12,628,746 deliveries identified in the NIS. Of those deliveries, 1119 occurred in women with a diagnosis of CF. Demographic characteristics of women with and without CF are shown in Table 1 . Women with CF were more likely to be white and younger compared with women without CF. Race/ethnicity was missing for 15.2% of women with CF and 15.5% of women without CF. Women with CF also had a longer median length of hospital stay (3 vs 2 days; P < .0001) and greater median hospital charges ($13,727 vs $10,002; P < .0001) ( Table 1 ).
| Description | CF n = 1119 | No CF n = 12,627,627 | OR (95% CI) | P value |
|---|---|---|---|---|
| Race/Ethnicity, n (%) | ||||
| White | 794 (70.9) | 5,570,518 (44.1) | 1.0 | — |
| African American | 45 (4.0) | 1,511,168 (12.0) | 0.2 (0.2–0.3) | < .0001 |
| Hispanic | 70 (6.3) | 2,426,137 (19.2) | 0.2 (0.2–0.3) | < .0001 |
| Asian/Pacific Islander | 10 (0.9) | 559,837 (4.4) | 0.1 (0.1–0.2) | < .0001 |
| Other | 30 (2.7) | 603,467 (4.8) | 0.3 (0.2–0.5) | < .0001 |
| Missing | 170 (15.2) | 1,956,499 (15.5) | — | — |
| Age, y a | 26.5 ± 13.5 | 27.6 ± 13.7 | — | .006 |
| LOS, d b | 3 (2, 4) | 2 (2, 3) | — | < .0001 |
| Total charges, $ b | 13,727 (8471, 26,494) | 10,002 (6785, 15,096) | — | < .0001 |
Table 2 describes preexisting medical conditions among women with CF compared with women without CF. Women with CF were more likely to have significant medical comorbidities such as asthma (OR, 5.1; 95% CI, 4.3–6.1; P < .0001) and diabetes mellitus (OR, 14.0; 95% CI, 11.8–16.7; P < .0001). Other findings included increased incidence of cardiac conduction disorders (OR, 5.5; 95% CI, 4.0–7.6; P < .0001) and thrombophilia (OR, 5.7; 95% CI, 2.7–12.1; P < .0001).
| Condition, n (%) a | CF n = 1119 | No CF n = 12,627,627 | OR (95% CI) | P value |
|---|---|---|---|---|
| Cardiac conduction disorders | 39 (3.5) | 83,146 (0.7) | 5.5 (4.0–7.6) | < .0001 |
| Asthma | 165 (14.7) | 411,451 (3.3) | 5.1 (4.3–6.1) | < .0001 |
| Diabetes (nongestational) | 147 (13.1) | 134,209 (1.1) | 14.0 (11.8–16.7) | < .0001 |
| Thyroid disorder | 33 (2.9) | 299,896 (2.4) | 1.2 (0.9–1.7) | .24 |
| Thrombophilia/APS | 36 (3.2) | 71,418 (0.6) | 5.7 (2.7–12.1) | < .0001 |
| Anemia | 154 (13.8) | 1,363,611 (10.8) | 1.3 (1.1–1.6) | .001 |
| Thrombocytopenia | 11 (1.0) | 114,686 (0.9) | 1.1 (0.6–1.9) | .85 |
| Drug use | 14 (1.2) | 165,626 (1.3) | 0.9 (0.5–1.6) | .80 |
| Tobacco | 90 (8.0) | 794,831 (6.3) | 1.3 (1.04–1.6) | .02 |
a The NIS does not allow reporting the number of cases when the cell frequency is less than or equal to 10. There were 10 or fewer cases of chronic hypertension, systemic lupus erythematosus, rheumatoid arthritis, human immunodeficiency virus, alcohol use, and chronic renal failure among women with CF.
Medical complications were common among women with CF at the time of delivery ( Table 3 ). Women with CF had increased odds of death and blood transfusion compared with women without CF at delivery. With regard to respiratory complications, there was increased odds of pneumonia, need for mechanical ventilation, and acute respiratory failure. Women with CF were more likely to have the adverse CF outcome composite variable compared with women without CF (OR, 35.3; 95% CI, 28.6–43.5).
| Condition, n (%) a | CF n = 1119 | No CF n = 12,627,627 | OR (95% CI) | P value |
|---|---|---|---|---|
| Death | 11 (1.0) | 921 (0.007) | 125 (67–233) | < .0001 |
| Mechanical ventilation | 25 (2.2) | 9003 (0.07) | 31.9 (21.4–47.5) | < .0001 |
| Transfusion | 20 (1.8) | 131,684 (1.0) | 1.7 (1.1–2.7) | .01 |
| Pneumonia | 75 (6.7) | 13,150 (0.1) | 68.7 (54.3–86.9) | < .0001 |
| Acute respiratory failure | 14 (1.2) | 5450 (0.04) | 29.6 (16.7–48.0) | < .0001 |
| Acute renal failure | 11 (1.0) | 7075 (0.06) | 16.4 (8.9–30.4) | < .0001 |
| Composite CF outcome b | 95 (8.5) | 33,275 (0.26) | 35.3 (28.6–43.5) | < .0001 |
a The NIS does not allow reporting the number of cases when the cell frequency is less than or equal to 10. There were 10 or fewer cases of myocardial infarction, cardiac arrest, acute heart failure, pulmonary edema, acute respiratory distress syndrome, pulmonary embolism, deep vein thrombosis, stroke/cerebral vascular accident, sepsis, pyelonephritis and influenza among women with CF
b Composite CF outcome includes any of the following: death, mechanical ventilation, sepsis, pneumonia, acute respiratory failure, acute respiratory distress syndrome, or acute renal failure.
Next, we looked at obstetric events at time of delivery. Women with CF were more likely to have multiple gestation, operative vaginal delivery, gestational diabetes, and preterm labor ( Table 4 ). There was no difference in the odds of cesarean delivery, preeclampsia/eclampsia/gestational hypertension, placental abruption, fetal growth restriction, or chorioamnionitis.
| Condition, n (%) a | CF n = 1119 | No CF n = 12,627,627 | OR (95% CI) | P value |
|---|---|---|---|---|
| Cesarean delivery | 351 (31.4) | 4,041,005 (32.0) | 1.0 (0.9–1.1) | .67 |
| Operative vaginal delivery | 100 (8.9) | 792,143 (6.3) | 1.5 (1.2–1.8) | .0002 |
| Multiple gestation | 39 (3.5) | 267,193 (2.1) | 1.7 (1.2–2.3) | .0013 |
| GDM | 148 (13.2) | 714,940 (5.7) | 2.5 (2.1–3.0) | < .0001 |
| Preeclampsia, eclampsia, gest HTN | 76 (6.8) | 931,154 (7.4) | 0.9 (0.7–1.1) | .48 |
| Preterm labor | 209 (18.7) | 1,051,494 (8.3) | 2.5 (2.2–2.9) | < .0001 |
| Abruption | 16 (1.4) | 136,053 (1.1) | 1.3 (0.8–2.2) | .22 |
| Fetal growth restriction | 29 (2.6) | 271,882 (2.2) | 1.2 (0.8–1.8) | .26 |
| Postpartum hemorrhage | 15 (1.3) | 321,959 (2.5) | 0.5 (0.3–0.9) | .012 |
| Chorioamnionitis | 36 (3.2) | 323,531 (2.6) | 1.3 (0.9–1.8) | .17 |
a The NIS does not allow reporting the number of cases when the cell frequency is less than or equal to 10. There were 10 or fewer cases of fetal demise and placenta previa among women with CF.
Using multivariable logistic regression to control for age, race/ethnicity, multiple gestation, mode of delivery, diabetes (gestational or pregestational), chronic hypertension, and preeclampsia/eclampsia, an adjusted analysis for medical and obstetric complications was performed ( Table 5 ). Women with CF had an increased odds of death with an adjusted OR (aOR) of 76.0 (95% CI, 31.6–183; P < .0001) compared with women without CF. Increased odds of mechanical ventilation, transfusion, pneumonia, acute respiratory failure, and acute renal failure among women with CF remained significant in our adjusted model. The composite CF outcome variable also remained significantly associated with CF in the adjusted model (aOR, 28.1; 95% CI, 21.8–36.3). With regard to pregnancy outcomes, women with CF had increased odds of preterm labor (aOR, 2.2; 95% CI, 1.9–2.6; P < .0001) ( Table 5 ). Odds of cesarean delivery, preeclampsia/eclampsia/gestational hypertension, abruption, fetal growth restriction, postpartum hemorrhage, and chorioamnionitis were not significant.
| Maternal condition | Adjusted OR (95% CI) | P value |
|---|---|---|
| Death | 76.0 (31.6–183) | < .0001 |
| Mechanical ventilation | 18.3 (10.8–31.2) | < .0001 |
| Transfusion | 1.68 (1.01–2.81) | .045 |
| Pneumonia | 56.5 (43.2–74.1) | < .0001 |
| Acute respiratory failure | 20.3 (10.5–39.0) | < .0001 |
| Acute renal failure | 17.3 (9.1–32.6) | < .0001 |
| Composite CF outcome | 28.1 (21.8–36.3) | < .0001 |
| Preterm labor | 2.2 (1.9–2.6) | < .0001 |
Results
Over the years 2000 to 2010, there was a significant linear increase in the number of women with CF at delivery from 2.99 per 100,000 deliveries in 2000 to 9.84 per 100,000 in 2010 ( P < .0001; R 2 = 0.91; Figure ). The years 2008 to 2010 were then used to determine characteristics of women with CF at delivery. Between 2008 and 2010, there were 12,628,746 deliveries identified in the NIS. Of those deliveries, 1119 occurred in women with a diagnosis of CF. Demographic characteristics of women with and without CF are shown in Table 1 . Women with CF were more likely to be white and younger compared with women without CF. Race/ethnicity was missing for 15.2% of women with CF and 15.5% of women without CF. Women with CF also had a longer median length of hospital stay (3 vs 2 days; P < .0001) and greater median hospital charges ($13,727 vs $10,002; P < .0001) ( Table 1 ).
