Body mass index and breastfeeding behavior among women participants in the Million Veteran Program

Abstract Background Breastfeeding has established health benefits for infants and has been associated with postpartum improved maternal cardiometabolic health in the long term. However, breastfeeding prevalence is also inversely associated with prepartum body mass index (BMI), and both are linked to socioeconomic factors. We sought to clarify the relationship between prepartum BMI and breastfeeding prevalence in the Million Veteran Program (MVP), a large-scale genetic epidemiology study of US Veterans. Methods We included data from parous female participants with available breastfeeding information from the MVP cohort. BMI at enrollment as well as earliest BMI available were extracted from the electronic health record, and polygenic scores (PGS) for BMI were calculated for the subset of participants with genotype data. We modeled whether participants breastfed an infant for one month or more (BF ≥ 1 M) as a function of BMI at enrollment ( n = 20,293); earliest BMI where available pre-pregnancy ( n = 532); and PGS for BMI among genetically inferred European ancestry participants ( n = 11,568). We conducted Mendelian randomization for breastfeeding using PGS as an instrumental variable. Results A higher BMI predicted a lower likelihood of BF ≥ 1 M in all analyses. A + 5 kg/m 2 BMI pre-pregnancy was associated with a 24% reduced odds of BF ≥ 1 M, and a + 5 kg/m 2 genetically predicted BMI was associated with a 17% reduced odds of BF ≥ 1 M. Conclusions BMI predicts a lower likelihood of BF ≥ 1 M. Given the high success of breastfeeding initiation in supportive environments combined with potential health benefits to both infant and mother, pregnant Veterans with prepartum elevated BMI may benefit from additional postpartum breastfeeding support.