A sex- and site-specific relationship between body mass index and osteoarthritis: evidence from observational and genetic analyses

Objective: We primarily aimed to investigate whether there are phenotypic and genetic links underlying body mass index (BMI) and overall osteoarthritis (OA). We then intended to explore whether the re-lationships differ across sexes and sites.Method: We first evaluated the phenotypic association between BMI and overall OA using data from the UK Biobank. We then investigated the genetic relationship leveraging summary statistics of the hitherto largest genome-wide association studies performed for BMI and overall OA. Finally, we repeated all analyses in a sex-(female, male) and site-(knee, hip, spine) specific manner.Results: Observational analysis suggested an increased hazard of diagnosed OA per 5 kg/m2 increment in BMI (hazard ratio = 1.38, 95% confidence interval (CI) = 1.37-1.39). A positive overall genetic correlation was observed for BMI and OA (rg = 0.43, P = 4.72 x 10-133), corroborated by 11 significant local signals. Cross-trait meta-analysis identified 34 pleiotropic loci shared between BMI and OA, of which seven were novel. Transcriptome-wide association study revealed 29 shared gene-tissue pairs, targeting nervous, digestive, and exo/endocrine systems. Mendelian randomization demonstrated a robust BMI-OA causal relationship (odds ratio = 1.47, 95% CI = 1.42-1.52). A similar pattern of effects was observed in sex-and site-specific analyses, with BMI affecting OA comparably in both sexes and most strongly in the knee.Conclusion: Our work demonstrates an intrinsic relationship underlying BMI and overall OA, reflected by a pronounced phenotypic association, significant biological pleiotropy, and a putative causal link. Stratified analysis further reveals that the effects are distinct across sites and comparable across sexes.& COPY; 2023 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved.