VP46.16: Second trimester cervical length and maternal characteristics in predicting preterm delivery among singleton pregnancies in a Filipino population

Abstract

This study aimed to identify second trimester cervical length measurements and maternal characteristics associated with preterm delivery among singleton pregnant Filipino women, to generate a model for the prediction of preterm delivery, and to determine its accuracy. This is a retrospective cohort study of Filipino pregnant patients with singleton pregnancies who underwent cervical length evaluation between 16–27 6/7 weeks AOG by transvaginal ultrasound at The Medical City Women's Healthcare Center and delivered in the same institution between 2015 to 2019. Maternal characteristics and CL measurements were obtained from medical records. Univariate analysis using chi-square test for the predictors was initially done. A receiver operating characteristic curve was obtained to determine the area under the curve and the cut-off values of the variables under investigation with the highest sensitivity and specificity. A p value of < 0.05 was considered statistically significant. A risk scoring index was developed for preterm birth using the beta coefficients/odds ratios of the significant clinical predictor variables derived from the logistic regression model. Age > 34 years (OR 1.82, 95% CI 1.04-3.20), prior spontaneous delivery < 37 weeks (OR 2.90, 95% CI 1.07-7.92), and short CL (OR 5.6, 95% CI 2.52-12.57) were associated with preterm delivery, with specificity of 83.5%, sensitivity 50%, NPV 95.7%, PPV 18.5%, and AUC = 0.571. Combined screening using cervical length, age, and prior spontaneous delivery shows a modest increase in sensitivity to 52.6%, and negative predictive value to 97.4%. Although the specificity was only 82.73% and the positive predictive value remained low at 12.3%, the combined model yielded an AUC = 0.677 compared to AUC = 0.571 when cervical length was used alone. Second trimester cervical length combined with age and prior spontaneous preterm delivery has the potential to predict preterm delivery in the current pregnancy.