Multivariate analysis of risk factors for operative delivery in nulliparous women

OBJECTIVE: Our purpose was to develop predictive models of operative delivery in nulliparous women on the basis of sociodemographic, anthropometric, and obstetric risk factors. STUDY DESIGN: Data were obtained prospectively on 925 nulliparous women in spontaneous term labor with a single fetus in cephalic presentation. Operative delivery was defined as either a midforceps or a cesarean delivery. Variables were grouped into two categories: those that could be assessed at admission and those appearing during labor. Multiple logistic regression was used to identify variables predictive of operative delivery. RESULTS: Among variables that can be documented at admission, independent predictors of operative delivery were maternal age and height, pregnancy weight gain, smoking status, gestational age, and admission cervical dilatation. Of these, maternal age greater than or equal to 35 years was the most strongly related to operative delivery. When variables documented later during labor were added to this first model, variables retained in the second model were age and height, smoking status, presence of dystocia, epidural analgesia, and fetal heart rate tracing abnormalities. The adjusted odds ratio of operative delivery in the presence of epidural anesthesia was 3.4 (95% confidence interval 2.0 to 5.8). This association was similar in the presence or absence of dystocia. When the specificity was in the range of 85%, the first and second models have sensitivities of 34% and 48%, respectively, and positive predictive values of 39% and 46%, respectively, which is higher than the a priori risk of operative delivery in the study population (21%). CONCLUSIONS: The models, based on data easily available, may help to predict the need for midforceps or cesarean section in low-risk nulliparous women. Before application in a clinical setting, these statistical models require validation in a separate cohort. The observed association between epidural anesthesia and operative delivery deserves interest but clinical trials are required to determine whether this relation is causal.