Association between Gestational Weight Gain and Neonatal Hypoxic and Traumatic Events: Results from a Large Retrospective Cohort

Most of the published studies focus on the relationship between gestational weight gain (GWG) and maternal outcomes, and less so on the neonatal outcomes. Furthermore, optimal GWG ranges defined by the Institute of Medicine were challenged by research showing the benefits of smaller GWG for overweight and obese mothers. Therefore, we intended to assess whether there is an association between GWG and hypoxic or traumatic events in the neonate, especially among overweight and obese mothers, as we sought the best methodological approach to study GWG. We used the Consortium on Safe Labor data, a retrospective cohort study collecting pregnancy data from 2002 to 2008 in 12 U.S. centers. The association between GWG and neonatal morbidity was estimated in singleton gestations ≥ 37 weeks' gestation using binomial logistic regression models. GWG was analyzed as a continuous linear, continuous quadratic, categorical, and binary variable. The Hosmer–Lemeshow test (HL) assessed calibration of the models, the Brier score evaluated calibration and the overall model performance, whereas the area under the receiver-operating characteristic curve (AUC) measured discrimination. Among 104,599 mothers 24,609 (23.5%) were overweight and 19,324 (18.5%) obese; 7,974 (7.6%) infants experienced hypoxic, whereas 2,031 (1.9%) developed traumatic morbidity. The risk of infant morbidity increased with GWG, with thresholds of 6 to 23 kg for ischemic events, and 4 to 21 kg for traumatic events (p < 0.05). Multivariable models showed adequate calibration for traumatic, and for hypoxic events, the latter only when GWG was analyzed as a quadratic or categorical variable (HL tests, p > 0.01). In contrast, discrimination was poor for all models (AUCs ≤ 0.6), as their Brier score reached the cutoff value of 0.07 for ischemic and 0.02 for traumatic morbidity. In conclusion, while GWG was associated with neonatal morbidity, it was not predictive of neonatal ischemic or traumatic events.