Bidirectional causal relationship between obesity and osteoarthritis: Insights from a two-sample Mendelian randomization study

Abstract

Objective: Osteoarthritis (OA) is a prevalent chronic disease associated with disability worldwide, and obesity is a key modifiable risk factor for OA. The study's aim was to investigate the causal relationship between obesity and OA.

Method: This study employed a two-sample Mendelian randomization (MR) approach to investigate the bidirectional causal relationship between obesity, using body mass index (BMI) as its proxy, and OA of the knee, hip, and hand. Genetic instruments were derived from large-scale GWAS meta-analyses, including ∼681,000 individuals for BMI and ∼827,000 individuals (177,000 OA cases) for OA. Inverse variance weighted with multiplicative random effects analysis was performed as primary analysis, and in addition sensitivity analyses relying on different assumptions were performed.

Results: The MR analysis revealed that genetically predicted BMI had a causal effect on increased risk of knee (OR 1.91, 95 ​% CI 1.80-2.03), hip (OR 1.52, 95 ​% CI 1.41-1.64) and hand OA (OR 1.21, 95 ​% CI 1.04-1.23). Sensitivity analyses confirmed the robustness of these associations. However, there was no evidence for a causal effect from knee, hip or hand OA on BMI.

Conclusion: This study provides strong evidence supporting a causal effect of obesity (measured by BMI) on OA, with a more pronounced effect in weight-bearing knee & hip joints compared to non-weight-bearing hand joint. There was no causal evidence for the reverse direction. Future research could look more in depth into differences in the genetic variants that may represent different biological underlying mechanisms.

Fig. 1

A directed acyclic graph illustrating the bidirectional, two-sample MR design of the study. In the first analyses, one sample is used to instrument BMI as the exposure and a second sample to instrument knee, hip, and hand OA (studied separately) as the outcomes. For the other direction, the exposure and outcomes are reversed, so that OA becomes the exposure and BMI the outcome.