Outcome of induction of labour using maternal characteristics, ultrasound assessment and biochemical state of the cervix

Objective. To assess pre-induction sonographic, digital examination and biochemical changes in the cervix to predict induction outcome. Methods. Transvaginal and abdominal scans were performed in 460 women at 37-41 weeks of gestation to determine cervical length (CL), posterior cervical angle (PCA) and foetal occipital position. The Bishop Score (BS) and the absence/presence of phosphorylated form of insulin-like growth factor-binding protein-1 (phIGFBP-1) in cervical secretions were assessed. Independent parameters significantly associated with a vaginal delivery were identified. Results. A total of 340 (73.9%) women achieved a vaginal delivery following induction. Multivariate analysis indicated that significant independent predictors of vaginal delivery were CL (adjusted odds ratio [AOR]: 0.59, 95% confidence interval [CI]: 0.45-0.79), PCA (AOR: 1.89, 95% CI: 1.09-3.28) and multiparae (AOR: 10.02, 95% CI: 5.10-19.69). For a specificity of 75%, the sensitivity for prediction of vaginal delivery using the BS, the CL and the multivariate model using the identified significant independent predictors were 37.1, 46.8 and 68%, respectively. Conclusion. The combination of sonographic assessment of the cervix and maternal characteristics was superior to the either BS or CL alone in the prediction of the induction outcome. Inclusion of the absence/presence of phIGFBP-1 did not further improve induction outcome.