A prediction score to assess the risk of delivering a large for gestational age infant among obese women

Abstract

Objective: To create a prediction score incorporating both maternal clinical characteristics and sonographic measurements in an effort to more accurately determine the risk of a large for gestational age (LGA) infant in the obese gravida.Methods: We performed a retrospective cohort study of obese women with singleton pregnancies who had a fetal ultrasound performed between 32 and 36 weeks from 1/2008 to 12/2011. LGA was defined as birth weight (BW) ≥90%. Clinical characteristics associated with fetal overgrowth were included in a multivariable logistic model and stepwise backwards regression was performed to identify which risk factors generated the most parsimonious predictive model. Adjusted odds ratios of covariates in the final model were used to estimate weights for each risk factor that were summed to generate a predictive score.Results: Six-hundred and sixty-nine obese women were included. The incidence of LGA infants was 11.8%. Ultrasound estimation of fetal weight alone accurately predicted LGA in 17.7 % of cases (AUC = 0.58). The most parsimonious model to accurately predict LGA at birth included 3rd trimester ultrasound EFW >90th percentile, interval from scan to delivery, and maternal history of diabetes mellitus (DM) (AUC = 0.74). A positive prediction score test result was associated with 92% specificity and 89% negative predictive value.Conclusions: A clinical prediction rule was developed and internally validated to predict the risk of an LGA infant among obese women. The ability to calculate a prediction score at the time of delivery is appealing to the clinician in order to accurately counsel women regarding the risks surrounding the delivery.