A model for predicting unscheduled caesarean section in nulliparae.

OBJECTIVE The objective of the study is to analyze the predictors of unplanned cesarean section in nulliparae. DESIGN Prospective cohort study. SETTING Institute for the Care of Mother and Child in Prague. METHODS This study consisted of nulliparae giving birth between the 37th and 42nd weeks of singleton low-risk pregnancy, with the fetus in vertex position and without primary indication for CS. Selected prenatal and intranatal factors were analyzed in relation to acute CS due to a failure to progress in labor and/or fetal distress. Using logistic regression analysis (LR1-3) and the classification tree method (chi-square automatic interaction detector 1-2), five prediction models were tested. RESULTS Of 3,728 nulliparae, 908 (24.4%) had an acute CS. All logistic regression models were comparable (receiver operating characteristic (ROC) 0.837-0.0881) and identified the occiput posterior position (OPP) of the fetus, maternal age, and epidural analgesia as the most influential risk factors. Spontaneous onset of labor, oxytocin administration, and maternal body height decreased are likely indicated for acute CS. The ability to predict a vaginal delivery was 95.7-96.3% and CS was 58.5-61.8%. The classification tree method (ROC 0.860-0.861) identified similar risk factors such as the OPP, peridural analgesia, and spontaneous onset of labor. The prediction abilities were similar at 94.5-96.4% for vaginal delivery and 64.6-59.0% for CS. CONCLUSION OPP of the fetus was the strongest risk factor for the unsuccessful trial of vaginal labor.