Higher Visceral Fat and Body Mass Index Increases Risk for Multiple Sclerosis

Background: Observational studies have demonstrated that obesity across the lifespan increases MS risk, but the precise mechanisms through which this proinflammatory health state causes MS is not clear. Mendelian randomization (MR) is genetic instrumental variable analysis that can facilitate our understanding of these mechanisms. To date, MR analyses have shown that higher BMI causes MS but variation in leptin or adiponectin did not. A concern with using BMI as a measure of obesity is that BMI captures non-specific variation across a wide range of body types.

Objectives: To investigate the causal relationships between multiple robust measures of obesity and MS risk.

Methods: We performed two-sample MR (2SMR) analyses to characterize causal relationships for multiple obesity traits. This approach relies on genome-wide association (GWA) summary statistics to infer causality. GWA summary statistics for obesity-related exposures were obtained for BMI (N=681,275), waist-hip ratio (WHR; N=694,649), visceral adipose tissue (VAT; N=325,123), and the ratio of arm, leg, and trunk fat to total body fat measured through segmental bioelectric impedance analysis (AFR, LFR, and TFR, respectively; N=362,499). GWA summary statistics for MS risk were also obtained (14,802 cases, 26703 controls). 2SMR was performed to obtain effect estimates for the relationship between each obesity trait and MS risk. Genetic instruments were clumped at a linkage disequilibrium threshold of r²<0.05 using the 1000 Genomes EUR Reference Panel. The effects of horizontal pleiotropy were investigated via the implementation of MR-PRESSO.

Results: Only BMI and VAT were significantly associated with MS risk in a 2SMR analysis of 14,802 MS cases and 26,703 controls (β_BMI=0.25, p_BMI<0.001; β_VAT=0.24, p_VAT=0.04). This association was further confirmed upon adjustment for horizontal pleiotropy with MR-PRESSO (β_BMI=0.25, p_BMI<0.001; β_VAT=0.35, p_VAT<0.001). In addition, under MR-PRESSO, leg to total body fat ratio (LFR) was also significant (β_LFR=-0.27, p_LFR=0.04), suggesting that genetically-driven LFR is affected by horizontal pleiotropy.

Conclusions: Significant causal associations were observed for BMI and VAT. There was suggestive evidence for LFR, but not AFR, TFR, or WHR. Future work will investigate potential horizontal pleiotropic effects across the different measures of obesity, and the role such variants play in risk of MS.