Maternal Obesity and Neonatal Death in Preterm US Pacific Islander Neonates Using 2 Analytic Approaches.

Importance: Owing to a high prevalence of obesity, Pacific Islander individuals in the US are at higher risk for preterm birth (PTB), but outcomes after PTB remain understudied. Existing literature suggests that associations between prepregnancy obesity and neonatal mortality stratified by gestational age (GA) are modest or null, which may be a result of overlooked stratification bias. Objective: To estimate the association between prepregnancy body mass index (BMI) and neonatal death (NND) following PTB by different degrees of prematurity in US Pacific Islander neonates using birth-based and fetuses-at-risk (FAR) approaches. Design, Setting, and Participants: This population-based cohort study used 2014 to 2018 data files from the National Center for Health Statistics for Pacific Islander singletons born at 22 to 41 weeks' GA without congenital anomalies. The 2 at-risk populations were PTBs (for the birth-based approach) and all identified pregnancies resulting in live birth (for the FAR approach). Analyses were finished in March 2023. Main Outcomes and Measures: The primary outcome was NND following PTB stratified by GA intervals (22-27 weeks, 22-31 weeks, and 22-36 weeks). Cox proportional hazards models were used and stratified by GA at birth. Results: Among 55 975 mother-neonate dyads (27 320 [48.8%] female neonates), the mean (SD) maternal age was 27.8 (5.8) years, and the mean (SD) gestational age of all neonates was 38.5 (1.9) weeks. The PTB prevalence was 9.3% (5192 neonates), and the neonatal mortality rate was 20.4 deaths per 1000 PTBs by the birth-based approach and 1.9 deaths per 1000 live-born pregnancies by the FAR approach. Among extreme PTBs (22-27 weeks), using the FAR approach, associations between prepregnancy obesity and NND following PTB were evident for obesity class I (adjusted hazard ratio [aHR], 2.31; 95% CI, 1.12-4.79) and class II (aHR, 2.82; 95% CI, 1.24-6.41). These associations were attenuated using the birth-based approach (obesity class I aHR, 1.33; 95% CI, 0.61-2.87; class II aHR, 1.73; 95% CI, 0.71-4.28). Conclusions and Relevance: In this cohort study of US Pacific Islander individuals, according to the FAR approach, prepregnancy obesity class I and II were associated with an increased rate of NND following PTB, yet these associations may be overlooked when analyses use the birth-based approach. This is likely because prepregnancy BMI affects GA at birth, which, in turn, may have its own confounded association with NND, leading to stratification bias and attenuated associations between prepregnancy BMI and NND using the birth-based approach. The FAR approach offers an alternative that may highlight important risk factors for key perinatal outcomes.