A bidirectional Mendelian randomization study of glycemic and anthropometric traits and Parkinsons disease

Abstract

BackgroundImpaired glucose and obesity are frequently observed in patients with Parkinsons disease (PD), although it is unclear whether the impairment precedes or results from the neurodegeneration. ObjectiveWe aimed to assess whether glycemic and anthropometric traits can influence the risk of PD in 33,674 cases and 449,056 healthy controls using the Mendelian randomization (MR) framework. MethodsWe investigated causality with a two-sample MR approach in the European population to compute effect estimates with summary statistics from available discovery meta-analyses of genome-wide association studies (GWAS) on glycemic and anthropometric traits. ResultsWe considered a threshold of p-value=0.0038 as significant after accounting for multiple testing, and p-value<0.05 was considered to be a suggestive evidence for a potential association. We observed a protective effect of waist-hip ratio (WHR) on PD (Inverse variance-weighted (IVW): OR IVW=0.735; 95%CI= 0.622-0.868; p-value=0.0003; I2 index=22.0%; MR-Egger intercept p-value=0.1508; Cochran Q test p-value=0.0003). The association was further retained after the exclusion of overlapping UK biobank (UKB) samples between the WHR and PD datasets (ORIVW=0.791; 95%CI=0.659-0.950; p-value=0.012; I2 index=13.0%; MR-Egger intercept p-value=0.733; Cochran Q test p-value=0.035). The sensitivity analysis provided suggestive evidence of an increased risk of PD on fasting glucose (FG) ({beta} IVW=0.0188; 95%CI=0.0062-0.0313, p-value=0.0055; I2 index=0.0%; MR-Egger intercept p-value=0.0957; Cochran Q test p-value=0.4555) and protective effect of PD on T2D (Weighted median effect: ORWME=0.946; 95%CI=0.9290.983; p-value=0.0051; Weighted mode effect: ORMBE=0.943; 95%CI=0.904-0.983; p-value=0.0116). ConclusionsOur results showed that central or abdominal obesity may be protective against PD development, independent of glucose levels.