473: Prediction of cesarean delivery after induction of labor in twin pregnant women

Abstract

Induction is one of the most common procedures in obstetrics. Although several factors have been suggested as risk factors for cesarean delivery (CD) in previous anecdotal studies with twin pregnancies, risk factors alone are not sufficient to estimate actual risk of CD after induction of labor (IOL). The objective of this study was to develop and validate a prediction model for CS after IOL in twin pregnancies. The study population consisted of twin pregnant women who were underwent IOL at ≥ 36 weeks of gestation after IOL from 2005 to 2018. The study population was randomly divided into training and test set with a ratio of 2:1, by three-fold cross validation with repletion of 100 times. In training set, the prediction model was developed by logistic regression analysis with clinical variables which were different between cases with vaginal delivery (VD) and those with CD in univariable analysis (p < 0.05). The developed model was evaluated in the test set. The model with highest average-test the area under the receiver operating characteristic (AUROC) was selected as the best model. Total 1,136 twin pregnant women were delivered after IOL, and 17.7% women (n=201) were delivered by CD. Clinical variables which were different between the two groups was put in the logistic regression analysis (Table), and the best prediction model was selected including 4 variables (maternal age, parity, height, and summated birthweight of both twin), with AUROC of 0.743 (95% CI 0.695-0.789) in the training set and 0.732 (95% CI, 0.664-0.800) in the test set. In addition, we developed a nomogram for predicting a risk of CD after IOL in twin pregnancies (Figure). The prediction model for CD after IOL was created in twin pregnancies. This model could be used to provide the information and evaluate the risk of CD after IOL in twin pregnant women.View Large Image Figure ViewerDownload Hi-res image Download (PPT)