Predicting operative delivery.

Objective: Unplanned operative delivery (vaginal or abdominal) is associated with maternal anxiety, maternal and neonatal morbidity and increased resource use. We aimed to identify potential predictors for emergency operative delivery.

Methods: This was a prospective observational study of 202 nulliparous women in a tertiary antenatal unit between 36 and 40 weeks' gestation. The assessment included an interview, a vaginal examination for Bishop score (optional), and a translabial ultrasound examination performed with the woman in a supine position and after voiding to determine cervical length, bladder position on Valsalva, and fetal head engagement. Clinical data were obtained from the institutional obstetric database and patient records.

Results: In the late third trimester, body mass index (P = 0.016), maternal age at due date (P < 0.0001), history of Cesarean section in first-degree relatives (P = 0.009), Bishop score (P = 0.0004), cervical length (P = 0.001), bladder position on Valsalva (P = 0.003) and head engagement (P < 0.0001) were significantly associated with delivery mode. On multivariate logistic regression analysis, the best model for predicting normal vaginal delivery contained maternal age, history of Cesarean section, Bishop score and bladder position on Valsalva and had excellent ability to discriminate between normal vaginal delivery and operative delivery (c = 0.85). The model with the best ability to discriminate between vaginal delivery and Cesarean section contained the same parameters plus body mass index; this model performed even better (c = 0.87).

Conclusions: Identification of women at increased risk of operative delivery appears feasible. A combination of clinical and ultrasound variables yielded a model that is likely to predict delivery mode accurately in up to 87% of cases. Such a model may become useful as an entry criterion for intervention trials in women at low or very high risk of operative delivery.