Abstract
OBJECTIVE: Maternal genotypes may be useful to customise foetal growth assessment, but generalisability across diverse racial and ancestral groups remains uncertain. We assessed the generalisability of a genetic risk score for birth weight (GRS(BW)), derived from participants of predominantly European ancestry, within a diverse U.S. DESIGN: Secondary analysis of a prospective observational cohort of nulliparous patients. SETTING: Eight U.S. recruitment centers. POPULATION OR SAMPLE: Participants in the parent study with available maternal DNA. METHODS: We used log-linear modelling to test the association of maternal GRS(BW) with infant birth weight. We then assessed the robustness of the association by self-identified race and genetically predicted continental ancestry (GPA) groups. MAIN OUTCOME MEASURES: Birth weight. RESULTS: Among 8147 eligible participants, GRS(BW) was positively associated with birth weight (p < 0.001). Across self-identified racial groups, the association was significant in White (n = 5394, mean ratio 1.04, 95% CI 0.97-1.11, p = 0.007) and multiracial (n = 508, mean ratio 1.10, CI 1.01-1.2, p = 0.03) groups but not in Black (n = 1139), Asian (n = 358), or unknown race groups (p > 0.09 for all). Among GPA groups, the association was significant among European (mean ratio 1.04, CI 1.02-1.07, p = 0.001) and American (mean ratio 1.08, CI 1.01-1.14, p = 0.02) ancestry groups but not African, East or South Asian, or unknown ancestry (p > 0.05 for all). CONCLUSIONS: This GRS(BW) is not generalisable across self-described racial identities or GPA groups, highlighting the need for globally representative genetic discovery cohorts as well as further investigation into the complex role of race, ethnicity, and epigenetic influences on foetal growth.