RESUMEN
BACKGROUND: Little is known about the ability of the recently released Brazilian gestational weight gain (GWG) charts to predict the occurrence of adverse birth outcomes. OBJECTIVES: We compared the new Brazilian weight gain charts with 3 international charts and determined their ability to predict the occurrence of small-for-gestational-age (SGA) and large-for-gestational-age (LGA) births in Brazilian women. METHODS: A subsample of 6888 adult women (43,931 weight measurements) with singleton pregnancies from a nationwide, hospital-based cohort study conducted in 2011-2012 was analyzed. Selected percentiles from Brazilian GWG charts were compared with those from American, International Fetal and Newborn Growth Consortium for the 21st Century study, and Lifecycle consortium charts. Sensitivity, specificity, and AUROC values for SGA and LGA births were estimated with 95% CIs using the classification of GWG below or above selected percentiles of each chart. RESULTS: The weight gain corresponding to a given percentile varied among the charts, especially for women with pre-pregnancy overweight and obesity. The proportions of women with GWG classified below or above selected percentiles were closest to the expected values for all pre-pregnancy BMI categories in the Brazilian and Lifecycle charts. At the 10th percentile, the highest sensitivity for SGA births was observed for the American charts at midpregnancy (36.8%) and the highest specificity was observed using the Brazilian charts in the first trimester (93.4%). At the 90th percentile, the highest sensitivity for LGA births occurred in midpregnancy for the Lifecycle charts (26.8%) and the highest specificity occurred in the American charts using total GWG (97.1%). All the AUROCs were under 0.5 for SGA births and ranged from 0.55 (first trimester) to 0.62 (total GWG) for LGA births. CONCLUSIONS: The charts differ in GWG trajectories, especially for women with overweight and obesity. The 4 charts had low predictive ability of SGA and LGA births and should not be considered as isolated screening tools for those outcomes.
