Thyroid function in causal relation to MRI markers of cerebral small vessel disease: a Mendelian randomization analysis.

Abstract

Observational studies provided insufficient information on the association between thyroid function and the risk of cerebral small vessel disease (CSVD); moreover, the causality of this link was unclear. This study aimed to investigate whether genetically predicted variation within thyroid function was causally associated with the risk of CSVD using two-sample Mendelian randomization (MR) analysis. In this two-sample MR study with genome-wide association variants, we estimated the causal effects of genetically predicted thyrotropin (TSH; N = 54,288), free thyroxine (FT4; N = 49,269), hypo (N = 51,823), and hyperthyroidism (N = 51,823) on three neuroimaging markers of CSVD, including white matter hyperintensities (WMH; N= 42,310), mean diffusivity (MD; N = 17,467), and fractional anisotropy (FA, N = 17,663). The primary analysis used inverse-variance-weighted MR method, followed by sensitivity analyses using MR-PRESSO, MR-Egger, weighted median, and weighted mode methods. Genetically increased TSH was associated with increased MD (β = 0.311, 95% CI = [0.0763, 0.548], P = 0.01). Genetically increased FT4 was associated with increased FA (β = 0.540, 95% CI = [0.222, 0.858], P < 0.001). Sensitivity analyses using different MR methods showed similar directions but lower precision. No significant associations of hypo and hyperthyroidism with WMH, MD, or FA were found (all P > 0.05). This study indicated that genetically predicted increased TSH was associated with increased MD, as well as increased FT4 with increased FA, implying the causal effect of thyroid dysfunction on white matter microstructural damages. There was no evidence for causal relationships of hypo or hyperthyroidism with CSVD. Further investigations should verify these findings and clarify the underlying pathophysiological mechanisms.