Walking pace is a protective factor for rheumatoid arthritis: a mendelian randomization study

Abstract

Walking pace is a simple and functional form of exercise and a strong predictor of health, but little is known about its causal association with rheumatoid arthritis. This study aimed to investigate the causal effect of WP on the developing RA using Mendelian randomization analysis. The genetic variation associated with WP was selected as an instrumental variable from the latest genome-wide association studies. Summary-level data for the outcomes were obtained from the corresponding GWAS. The inverse-variance weighted method was used as the primary MR analysis. The results were further tested using a multivariable MR approach based on Bayesian model averaging. Confounders (BMI, SMK, HBP, TD) with close associations with RA were included in the analysis. An observational study with individual data from UK Biobank was performed to reinforce our findings. The MR results indicated the significant inverse associations of WP with the risk of RA (odds ratio (OR), 0.31; 95% confidence interval (CI), 0.15, 0.62; p = 1.05 × 10 -3). After adjusting for the risk factors, the associations for WP and RA did not change substantially. Observational study results demonstrated the same effect of WP on reducing the risk of RA. The Mendelian randomization analysis and observational study provide evidence suggesting that walking pace is a protective factor for rheumatoid arthritis. Given its simple measurement, walking pace may be a pragmatic target for interventions.

Keywords: Causal association; MR-BMA; Mendelian randomization; Rheumatoid arthritis; Walking pace.

Fig. 1

Workflow of the Mendelian randomization and observational analysis.

Fig. 2

Mendelian randomization results.A. Forest plot summarising the overall Mendelian randomization estimates of SNP specificity and the causal effect on RA. B-E: Scatter plot of Mendelian randomization analysis for the associations of WP, BMI, SMK, SMK with the risk of RA.

Fig. 3

MR-BMA plots for outliers and influential genetic variants.(A) Diagnostic plots for invalid SNPs (outliers) (B) Diagnostic plots for influential genetic variants (C) Diagnostic plots for invalid SNPs after excluding invalid SNPs (rs2894446, rs3790604, rs4605363) (D) Diagnostic plots for influential genetic variants after excluding invalid SNPs (rs2894446, rs3790604, rs4605363). When Q-statistic > 10, the instrument for the corresponding gene is invalid and marked with a red box. When Cook’s distance > the threshold, the instrument for the corresponding gene is influential and marked with a red box.

Fig. 4

Result of multivariable MR anlysis and observational analysis. Crude: only adjusted for sex and age; Adjusted: adjusted for sex, age, BMI, smoking and high blood pressure; MR: Mendelian randomization; LR: logistic regression

Fig. 5

GO enrichment analysis and KEGG enrichment analysis of differential genes where the SNPs obtained by intersecting the summary data of WP and RA.(A) The top 10 GO enrichment analyses of biological processes, cellular components, and molecular functions. (B) KEGG enrichment analysis of differential genes where the SNPs obtained by intersecting the summary data of WP and RA.