The use of the Edmonton Obesity Staging System compared to body mass index to predict adverse pregnancy outcomes

In nonpregnant populations, obesity staging systems such as the Edmonton Obesity Staging System (EOSS) improve risk stratification of patients with obesity by also accounting for medical comorbidities. While the EOSS has been shown to better predict morbidity and mortality compared to body mass index (BMI)-based obesity classification in the nonpregnant population, its utility in pregnancy and performance compared to the current standard remain unclear. Thus, the objective of this study was to compare a modified EOSS and BMI-based obesity classification in predicting specific adverse pregnancy outcomes.

We created a modified EOSS using available diagnoses within the Consortium on Safe Labor (CSL) dataset and applied it to all women with a prepregnancy BMI ≥ 30 kg/m². The primary outcome was a composite of maternal adverse outcomes including chorioamnionitis, endometritis, preeclampsia with severe features, eclampsia, venous thromboembolism, wound infection/separation, intensive care unit (ICU) admission, and maternal death. Prespecified secondary outcomes included a severe maternal morbidity composite and a neonatal adverse outcome composite. In order to compare EOSS with BMI classification, the association between EOSS stages and adverse outcomes was first examined. Bivariate analyses assessed maternal and obstetric characteristics by EOSS stage. Multivariable logistic regression estimated associations between EOSS stage and outcomes. Predictive performance of EOSS versus the World Health Organization (WHO) BMI classification was evaluated using receiver operating characteristic (ROC) curve analysis.

Among 228,668 subjects in the cohort, 139,126 pregnant individuals met inclusion criteria for analysis; 67,289 had BMI ≥ 30 kg/m². EOSS classification was associated with increased odds of the maternal composite outcome, secondary maternal severe composite outcome, and neonatal composite outcome (all p < 0.01). However, ROC analysis demonstrated that EOSS did not outperform the WHO BMI classification in predicting adverse outcomes, with both systems demonstrating poor-to-fair discrimination (AUC 0.57-0.67).

A modified EOSS identified increased risk of adverse maternal and neonatal outcomes when applied to a pregnant population but did not demonstrate improved predictive ability compared to traditional BMI classification. Both EOSS and WHO BMI systems exhibited only poor-to-fair predictive ability for select adverse outcomes.