Understanding Factors That Cause Benign Paroxysmal Positional Vertigo, Ménière Disease, and Vestibular Neuritis: A Two-Sample Mendelian Randomization Study.

Abstract

Vertigo is a prevalent clinical symptom, frequently associated with benign paroxysmal positional vertigo (BPPV), Ménière disease (MD), and vestibular neuritis (VN), which are three common peripheral vestibular disorders. However, there is a relative lack of research in epidemiology and etiology, with some existing studies presenting discrepancies in their conclusions. We conducted a two-sample Mendelian randomization (MR) analysis to explore potential risk and protective factors for these three peripheral vestibular disorders. Based on genome-wide association studies, we executed a univariable MR to investigate the potential associations between 38 phenotypes and MD, BPPV, and VN. We used the inverse variance weighted method as the primary MR result and conducted multiple sensitivity analyses. We used false discovery rate (FDR) correction to control for type I errors. For findings that were significant in the univariable MR, a multivariable MR analysis was implemented to ascertain direct effects. In addition, we replicated analyses of significant results from the univariable MR to enhance the robustness of our analyses. For BPPV, both alcohol consumption (odds ratio [OR] = 0.57, 95% confidence interval [CI] = 0.43 to 0.76, FDR Q = 0.004) and educational attainment (OR = 0.77, 95% CI = 0.68 to 0.88, FDR Q = 0.003) were found to decrease the risk. The genetic prediction analysis identified major depression (OR = 1.75, 95% CI = 1.28 to 2.39, FDR Q = 0.008) and anxiety (OR = 5.25, 95% CI = 1.79 to 15.42, FDR Q = 0.036) increased the risk of MD. However, the impact of major depression on MD could be influenced by potential horizontal pleiotropy. Systolic blood pressures (OR = 1.03, 95% CI = 1.02 to 1.04, FDR Q = 4.00 × 10 -7 ) and diastolic blood pressures (OR = 1.05, 95% CI = 1.03 to 1.07, FDR Q = 2.83 × 10 -6 ) were associated with an increased risk of VN, whereas high-density lipoprotein (OR = 0.77, 95% CI = 0.67 to 0.89, FDR Q = 0.009) and urate (OR = 0.75, 95% CI = 0.63 to 0.91, FDR Q = 0.041) reduces the risk of VN. Only the relationship between urate and VN was not replicated in the replication analysis. Multivariable MR showed that the protective effect of education on BPPV was independent of Townsend deprivation index. The protective effect of high-density lipoprotein against VN was independent of triglycerides and apolipoprotein A1. The risk impacts of systolic and diastolic blood pressures on VN exhibited collinearity, but both are independent of chronic kidney disease and estimated glomerular filtration rate. The impacts of anxiety and severe depression on MD demonstrated collinearity. Our study identified the risk association between systolic and diastolic blood pressure with VN and the protective influence of high-density lipoprotein on VN, which may support the vascular hypothesis underlying VN. Furthermore, we observed an elevated risk of MD associated with anxiety. The potential protective effects of education and alcohol consumption on BPPV need further exploration in subsequent studies to elucidate specific mechanistic pathways. In summary, our MR study offers novel insights into the etiology of three peripheral vestibular diseases from a genetic epidemiological standpoint.