Materials and Methods

From 2008 through 2011, investigators at 25 medical centers of the Eunice Kennedy Shriver National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network assembled an observational obstetric cohort (ie, the Assessment of Perinatal EXcellence [APEX] study) that included detailed information on patient characteristics, intrapartum events, and pregnancy outcomes that was collected by trained and certified nurses. Patients eligible for data collection were those who delivered within the institution, were at least 23 weeks of gestation, had a live fetus on admission, and delivered during the 24-hour period of randomly selected days representing one third of deliveries over this 3-year period (hence approximating the number of deliveries in 1 year in the maternal-fetal medicine units network). Days were chosen via computer-generated random selection, stratified by weekdays, weekends, and holidays and generated separately for each hospital. On selected days, the labor and delivery logbook at each participating center was screened to identify all eligible women. Data from maternal and neonatal charts for all eligible deliveries were abstracted as soon as the delivery and nursery medical records became available (n = 115,502). Institutional review board approval for the study and a waiver of informed consent was obtained at all centers. Full details of the technique of data collection have been described previously.

This secondary analysis included women delivering a nonanomalous singleton between 23 weeks 0 days and 33 weeks 6 days of gestation. Women with an antepartum stillbirth and women who were admitted with nonreassuring fetal status, cord prolapse, placenta previa, placenta accreta, placental abruption, and unstable and severe maternal condition (cardiopulmonary collapse, acute respiratory distress syndrome, seizures) were excluded from the main analysis due to the extremely high maternal morbidity associated with these indications for delivery by themselves and to understand the effect of the route of delivery on serious maternal postpartum complications. Serious maternal complications were defined as the occurrence of one of the following: hemorrhage (blood loss ≥1500 mL, blood transfusion, or hysterectomy for hemorrhage), infection (endometritis, wound dehiscence, or wound infection requiring antibiotics, reopening, or unexpected procedure), intensive care unit (ICU) admission, or death.

The delivery route was classified as vaginal delivery, CCD, LTCD, and LVCD. Uterine incisions that started as low segment incisions but were extended into the upper uterine segment (J or inverted T incisions) were included as part of the CCD group.

Univariate analyses used the χ ² or Fisher exact test for categorical variables and the Kruskal-Wallis test for continuous variables. The association of delivery route (CCD, LTCD, LVCD, vaginal delivery) with the occurrence of serious maternal complications was determined using multivariable log Poisson regression yielding adjusted relative risks (aRR) and 95% confidence intervals (CIs) after controlling for maternal age, race, body mass index (BMI) (kg/m ² ), hypertension (chronic, gestational, or preeclampsia), diabetes (pregestational or gestational), preterm premature rupture of membranes (PPROM), preterm labor, GA, and hospital. We also sought to describe the association between delivery route and maternal morbidity by categories of GA and conducted a post hoc analysis that included a multivariable model with an interaction term between delivery route and GA. Software (SAS; SAS Institute, Cary, NC), was used for the analyses. All tests were 2-tailed and P < .01 was used to define statistical significance to account for multiple univariate and multivariate analyses. No imputation for missing data was performed.

Results

Of 3390 women delivering a nonanomalous singleton between 23 weeks 0 days and 33 weeks 6 days of gestation, 712 women were excluded from the main analysis and evaluated separately due to the presence of the following: antepartum stillbirth (n = 11); nonreassuring fetal status or cord prolapse (n = 374); an adverse placental condition (previa, accreta, or abruption) (n = 324); or an unstable and severe maternal condition on admission (n = 3). An additional 19 were excluded because they were delivered by cesarean and had missing type of incision or were missing data regarding type of fetal presentation, leaving 2659 women as the subject of the main analysis. All 2659 women had complete data to define the outcomes of infection, ICU admission, and death; however, 63 women had missing data to define hemorrhage and 61 had missing data to define the composite of serious maternal complication. There were 272 women in the CCD group (248 with vertical incision in the upper uterine segment and 24 with T and J incisions), 904 women in the LTCD group, and 29 women in the LVCD group.

Patient characteristics differed significantly by route of delivery. Women in the CCD, LTCD, and LVCD groups had a higher BMI, were more likely to have had a prior cesarean delivery and hypertension, were less likely to have PPROM and preterm labor, and were more likely to have a nonvertex fetal presentation, a lower birthweight, and small-for-GA neonate than the vaginal delivery group: P < .01 ( Table 1 ). Women in the CCD group were more likely to be black, less likely to have private insurance, and more likely to have a nonvertex fetal presentation, have a lower birthweight neonate, and deliver at an earlier GA than the LTCD group: P < .01 ( Table 1 ).

Overall, 8.6% of women experienced a serious maternal complication. The composite of serious maternal complications also captured other serious outcomes such as disseminated intravascular coagulation (n = 5), idiopathic thrombocytopenic purpura (n = 1), cardiopulmonary arrest (n = 1), cardiac dysfunction (n = 8), adult respiratory distress (n = 6), renal failure (n = 2), sepsis (n = 6), and hysterectomy (n = 2), which were classified under ICU admission. The occurrence of serious maternal complications was associated with GA and highest in the earliest GA range (23-27 weeks of gestation). The rate of CCD also was associated with GA and highest in the earliest GA range ( Table 2 ). The frequency of complications was associated with route of delivery, with 23.0% of those undergoing CCD, 12.1% of those undergoing LTCD, and 10.3% of those undergoing LVCD experiencing serious maternal complications compared with 3.5% of women delivering vaginally ( P < .001) ( Table 3 ). Hemorrhage, infection, and ICU admission also were increased among women undergoing CCD or LTCD compared with vaginal delivery ( P < .001) ( Table 3 ). In multivariable analyses, CCD (aRR, 3.54; 95% CI, 2.29–5.48) and LTCD (aRR, 2.59; 95% CI, 1.77–3.77) were associated with significantly more serious maternal complications than vaginal delivery when adjusting for characteristics affecting maternal outcome ( Table 4 ). There was no significant difference in serious maternal complications between CCD and LTCD (aRR, 1.37; 95% CI, 0.95–1.97) or between CCD and LVCD (aRR, 1.56; 95% CI, 0.48–5.07).

Stay updated, free articles. Join our Telegram channel

Join