Diabetes, glycemic profile and risk of vitiligo: A Mendelian randomization study.

ABSTRACT

BACKGROUD Previous observational studies have shown that vitiligo usually co-manifests with a variety of dysglycemic diseases, such as Type 1 diabetes mellitus (T1DM) and Type 2 diabetes mellitus (T2DM). Mendelian randomization (MR) analysis was performed to further evaluate the causal association between fasting plasma glucose, glycosylated hemoglobin (HbA1c), T1DM, T2DM and vitiligo. MATERIALS AND METHODS: We used aggregated genome-wide association data from the Integrative Epidemiology Unit (IEU) online database of European adults vitiligo; HbA1c data were from IEU. Fasting blood glucose data were obtained from the European Bioinformatics Institute (EBI). T1DM and T2DM data were from FinnGen. We used bidirectional two-sample and multivariate MR analyses to test whether dysglycemic measures (fasting blood glucose, HbA1c), diabetes-related measures (T1DM, T2DM) are causatively associated with vitiligo. Inverse variance weighting (IVW) method was used as the main test method, MR-Egger, Weighted mode and Weighted median were used as supplementary methods. RESULTS: We found no statistically significant evidence to support a causal association between dysglycemic traits and vitiligo, but in the correlation analysis of diabetic traits, our data supported a positive causal association between T1DM and vitiligo (p = 0.018). In the follow-up multivariate MR analysis, our results still supported this conclusion (p = 0.016), and suggested that HbA1c was not a mediator of T1DM affecting the pathogenesis of vitiligo. No reverse causality was found in any of the reverse MR Analyses of dysglycemic traits and diabetic traits. CONCLUSIONS: Our findings support that T1DM is a risk factor for the development of vitiligo, and this conclusion may explain why the co-presentation of T1DM and vitiligo is often seen in observational studies. Clinical use of measures related to T1DM may be a new idea for the prevention or treatment of vitiligo.