Objective: To prospectively compare digital cervical score with Bishop score as a predictor of spontaneous preterm delivery before 35 weeks of gestation.
Methods: Data from a cohort of 2,916 singleton pregnancies enrolled in a multicenter preterm prediction study were available. Patients underwent digital cervical examinations at 22-24 and 26-29 weeks of gestation for calculation of Bishop score and cervical score. Relationships between Bishop score, cervical score, and spontaneous preterm delivery were assessed with multivariable logistic regression analysis, McNemar test, and receiver operating characteristic (ROC) curves to identify appropriate diagnostic thresholds and predictive capability.
Results: One hundred twenty-seven of 2,916 patients (4.4%) undergoing cervical examination at 22-24 weeks had a spontaneous preterm delivery before 35 weeks. Eighty-four of the 2,538 (3.3%) reexamined at 26-29 weeks also had spontaneous preterm delivery. Receiver operating characteristic curves indicated that optimal diagnostic thresholds for Bishop score were at least 4 at 22-24 weeks, at least 5 at 26-29 weeks, and less than 1.5 at both examinations for cervical score. At 22-24 weeks, areas under the ROC curve favored Bishop score. At 26-29 weeks, there was no significant difference in areas under the ROC curve; however, a cervical score less than 1.5 (sensitivity 35.7%, false positive rate 4.8%) was superior to a Bishop score of 5 or more (P<.001).
Conclusion: Both cervical evaluations are associated with spontaneous preterm delivery in a singleton population; however, predictive capabilities for spontaneous preterm delivery were modest among women with low event prevalence. Although Bishop score performed better in the mid trimester, by 26-29 weeks a cervical score less than 1.5 was a better predictor of spontaneous preterm delivery before 35 weeks than a Bishop score of at least 5.