Hearing loss is not associated with risk of Parkinson's disease: A Mendelian randomization study

Abstract

Purpose: A few observational studies have indicated that Parkinson's disease (PD) risk may be higher in those with hearing loss, but the two's causal relationship is yet unknown. Using Mendelian randomization (MR) methods, this study sought to explore the causal link between hearing loss and the risk of PD.

Methods: We identified single nucleotide polymorphisms (SNPs) linked to hearing loss (P-value<5E-08) in a genome-wide association study (GWAS) included 323,978 people from the UK Biobank. The summary data for PD in the discovery group came from a GWAS meta-analysis of 33,647 cases and 449,056 healthy participants of European descent. Using summary data from the aforementioned GWAS of PD (N = 33,647) and hearing loss (N = 323,978), we carried out a two-sample MR study. As validation groups, two separate PD GWAS studies were used. Inverse variance weighting (IVW) was utilized in the principal MR analysis. For our findings to be reliable, further analyses were carried out with the Cochran's Q test, MR-Egger intercept, and leave-one-out analysis. In addition, we assessed the causal link between various forms of hearing loss and PD using the IVW approach.

Results: Twenty-two SNPs with genome-wide significance linked to hearing loss were used as instrumental factors. In the discovery dataset, we failed to detect a causal relationship between hearing loss and PD (OR = 1.297; 95 % CI = 0.420-4.007; P-value = 0.651). The findings of other methods agreed with the IVW method. The results were robust under sensitivity analyses. Furthermore, the above findings were confirmed in two validation PD datasets. Additionally, no causal correlation was found between genetic prediction of four different types of hearing loss and PD (conductive hearing loss, IVW: OR = 1.058, 95%CI = 0.988-1.133, P-value = 0.108; sudden idiopathic hearing loss, IVW: OR = 0.936, 95%CI = 0.863-1.016, P-value = 0.113; mixed conductive and sensorineural hearing loss, IVW: OR = 0.963, 95%CI = 0.878-1.058, P-value = 0.436; sensorineural hearing loss, IVW: OR = 1.050, 95%CI = 0.948-1.161, P-value = 0.354).

Conclusion: In those of European heritage, our investigation revealed no causal link between hearing loss and PD risk.

Keywords: Genome-wide association study; Hearing loss; Mendelian randomization study; Parkinson's disease.

Fig. 1

Assumptions in Mendelian randomization analysis.

Fig. 2

Causal effect results for hearing loss and risk of PD (Nalls et al.). (A) Forest plot showing Mendelian randomization analysis results to assess the causal relationship between hearing loss and PD. (B) A scatter plot showing the implications of genetic variations on hearing loss and PD. The line that runs vertically indicates the 95 % confidence interval for the impact size of hearing loss, whereas the line that runs horizontally represents the 95 % confidence interval for the impact size of PD. The trend of the colored line shows the estimation of causal effects using different methods. (C) The hearing loss funnel plot displays the estimates using each genetic variation as a tool, utilizing the inverse of the standard error of the causal estimate. The predicted causal impact determined by IVW and MR-Egger approaches is shown by the line running vertically.

Fig. 3

Forest plots for a leave-one-out sensitive evaluation of hearing loss in PD (Nalls et al.). When a genetic variation was eliminated one at a time, the dots and lines in black represent the causative value and 95 % confidence interval. The dot and line in red represent the total estimation and 95 % confidence interval obtained via the fixed-effect IVW approach. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)