311 Development of a prediction model for cesarean delivery among term women with presumed intraamniotic infection

Abstract

To identify factors associated with cesarean delivery (CD) among term laboring women who develop an intrapartum fever due to presumed intraamniotic infection and to develop a predictive model for CD to assist in patient counseling. We performed a multi-center, retrospective cohort study among febrile, term laboring women between January 1, 2015 and December 31, 2017. We excluded women who had a documented non-obstetric source of infection, received antibiotics on admission for other indications, used immunosuppressant medication chronically, or had an intrauterine fetal demise. The medical record was reviewed, and sociodemographic, labor, and clinical characteristics were extracted. We identified the factors associated with CD and built our predictive model using backward stepwise regression. The final model was shown through the area under the receiver operating characteristic curve (AUC). The cohort consisted of 850 febrile laboring women, of whom 295 (34.7%) had a CD. Indications for Cesarean delivery included 34.6% for failure to progress, 33.6% for non-reassuring fetal heart tracing, 29.8% for arrest of descent, and 2.0% for failed operative delivery due to maternal exhaustion. The final model included stage of labor stratified by cervical exam dilation (less than 6 cm of dilation in latent labor, 6 cm to 9.5 cm in active first stage of labor, and 10 cm dilation in second stage of labor), duration of rupture of membranes (in hours), and if labor was induced. Our final model had an AUC of 0.74 (95% CI 0.71-0.78) with a Hosmer-Lemeshow test P=0.07. Using a predicted probability cutoff of 0.47, the model’s sensitivity was 54.7%, specificity was 80.7% with a positive predictive value of 60.1% and negative predictive value of 77.0% to predict CD. Information available at the time of development of an intrapartum fever may be used to counsel patients regarding the likelihood of CD. Our study was limited by unavailable electronic fetal monitoring data which also heavily influences CD risk.View Large Image Figure ViewerDownload Hi-res image Download (PPT)