(908) Validation of External Cephalic Version Prediction Model in Population with Low Neuraxial Anesthesia Use

External cephalic version (ECV) is a key tool in managing malpresentation and reducing the need for cesarean delivery. To aid counseling, Dahl et al. (2021) developed a prediction model for ECV success based on BMI, parity, placental location, and fetal presentation, with an area under the receiver operating characteristic curve (AUC) of 0.667. The objective of this study was to externally validate this model at a separate institution.

Study Design:

We performed a retrospective cohort study of ECVs performed at an academic institution between 7/16-12/21. The primary outcome was fetus in vertex presentation after ECV attempt. We collected the same maternal and fetal characteristics as in the original derivation study and assessed their relationship with ECV success. Coefficients from the original prediction model were used to generate predictive probabilities of successful ECV. The model’s performance was evaluated using receiver operating characteristic curves and calibration plots.

Results:

434 ECV procedures were performed, with a success rate of 44.4%. The rate of neuraxial anesthesia usage was 10.8%. Rate of complications leading to delivery was 2.3%. Factors significantly associated with ECV success were older maternal age, multiparity, non-anterior placenta location, polyhydramnios, transverse or oblique position, and greater estimated fetal weight (all p ≤ 0.001). The AUC for the Dahl prediction model for ECV success was 0.6988 in this separate patient population (Figure 1), which is similar to the derivation cohort. The model was also well calibrated across predicted probabilities (Figure 2).

Conclusion:

In this analysis, we externally validated a prediction model for ECV success. Of note, patients in this study had a much lower rate of neuraxial use than the original model population (10.5% vs 83.5%), but with similar success rates and lower rates of complications leading to delivery (2.3% vs 6.7%). This data suggests that the Dahl model is generally applicable in predicting ECV success and has potential use outside of the original derivation cohort.