763: Predicting fetal acidemia with continuous fetal monitoring in extreme preterm birth

Abstract

Objective: To determine which characteristics of continuous fetal monitoring are predictive of fetal acidemia in extreme preterm birth (< 28 weeks). Methods: This was a retrospective cohort study of all extreme preterm births at an academic medical center from 2012 to 2017. The outcome was fetal pH and base excess by cord gas at time of birth. Fetal acidemia was defined as pH≤7.1 or BE≥12. Primary exposures were characteristics of fetal monitoring: baseline, variability, nonreassuring antenatal testing as indication for delivery, and total deceleration area (TDA) during the last hour and last thirty minutes of tracing before delivery. TDA was defined as area below the fetal baseline in units of 10 beats per 10 seconds. Patients with abnormal baseline or variability, missing outcome or exposure data, and patients in labor delivered by cesarean were excluded. Data was tested for normalcy, and log-linear analysis was utilized to determine associations. Chi-square analysis was used to determine if a significant difference exists between acidemia and the exposures. A receiver-operator characteristic (ROC) curve was utilized to determine what magnitude of decelerations was associated with fetal acidemia. Results: Of the 22,594 deliveries reviewed, 133 were periviable and 74 met inclusion and exclusion criteria. TDA was not associated with fetal acidemia within the last hour (r=0.20, p=0.09) or last thirty minutes (r=0.17, p=0.15) of tracing prior to delivery, and the rates of fetal acidemia in patients with nonreassuring antenatal testing as an indication for delivery were not different (p=0.06). When TDA was adjusted for amount of interpretable tracing present within the last hour (r=0.22, p=0.058) and last thirty minutes (r=0.14, p=0.25), there was no association with fetal acidemia. ROCs of exposures in the last hour and last thirty minute intervals versus acidemia were not significant (AUCs ≤ 0.558). Even stratifying at the ninety-fifth percentile led to insignificant results. This cut off is approximately equivalent to 5 one minute decelerations dropping 40 beats from the baseline within the last hour prior to delivery. Conclusion: In this cohort, neither the total deceleration area nor nonreassuring status as an indication for delivery was associated with fetal acidemia. Decelerations in the absence of a category III tracing should be interpreted with caution in the periviable period as they may not be related to fetal acidemia.