304 Predicting severe adverse perinatal outcomes in pregnancies complicated by fetal growth restriction: a validation study

Abstract

Validate a model predictive of severe adverse perinatal outcomes in pregnancies complicated by fetal growth restriction using clinical factors widely available at time of diagnosis. We assessed the performance of a 7-item model previously developed from 336 pregnancies (2013-2019) meeting Delphi procedure-based consensus criteria [Gordijn SJ et al. Ultrasound Obstet Gynecol 2016;48(3):333-9] for fetal growth restriction (FGR). The model assesses risk of composite severe adverse perinatal outcome: perinatal demise, 5-minute Apgar < 7, cord pH ≤ 7.1, or cord base excess > 12. It includes factors available at diagnosis: hypertensive disorder of pregnancy (HDP), gestational age ≤ 30 weeks, isolated abdominal circumference < 3rd centile, absent/reversed end diastolic flow (A/REDF), chronic hypertension, maternal height > 160 cm, and pre-gravid weight 68.5 kg. To validate the model, we retrospectively analyzed data from non-anomalous, singleton gestations (2010-2020) not included in model development. Severe adverse perinatal outcomes complicated 22% (127) of 585 pregnancies. In addition, there were 257 (45%) NICU admissions and 169 (29%) cesareans for non-reassuring fetal status. Women with the composite outcome were more likely to have a higher body mass index, hypertensive disorders, and abnormal umbilical artery flow. Model performance in the validation cohort [bias-corrected bootstrapped AUC of 0.76 (95% CI 0.69-0.79)], was similar to the development cohort [AUC of 0.81 (95% CI 0.73-0.84)]; however only HDP, A/REDF, and gestational age ≤ 30 weeks were significant. Performance was better for early-onset [n = 316, AUC 0.76 (95% CI 0.71-0.82)] compared to late-onset (n = 269) growth restriction [AUC 0.69 (95% CI 0.55-0.66)]. Performance of a model predictive of severe adverse perinatal outcome in a larger validation cohort was relatively strong. However, lack of significance of four covariates indicates model redevelopment may improve performance and, in doing so, its utility in counseling and management of pregnancies complicated by growth restriction.View Large Image Figure ViewerDownload Hi-res image Download (PPT)