Mendelian Randomization Analysis Supports Causal Effect of Type II Diabetes Mellitus on Onychomycosis

Abstract Introduction: Onychomycosis is common among adults with diabetes mellitus (DM). We used two-sample Mendelian randomization to estimate the causal effect of genetic risk for DM on onychomycosis and tinea skin infections in the All of Us Research Program. Methods: Onychomycosis and tinea corporis, pedis, manus, and cruris cases were identified using electronic health record data, and genetic instrument variables and summary statistics were collected from a type II DM (T2DM) genome-wide association study (GWAS) meta-analysis. Results: Inverse variance weighted regression showed positive effect of T2DM genetic risk on onychomycosis (beta = 0.135, p = 1.86E−2), and weighted median regression produced a comparable estimate of effect size (beta = 0.148). There was no significant effect of T2DM on skin dermatophytosis. Conclusions: Our results suggest that T2DM has a positive causal effect on onychomycosis but not tinea skin infection risk. As onychomycosis may impair occupational function and increase risk for secondary soft tissue infections, patients with diabetes should be screened for onychomycosis and counseled on mitigating infection risk.


Introduction
Diabetes mellitus (DM) is a significant risk factor for onychomycosis.While the estimated worldwide prevalence of onychomycosis is 5-10%, approximately a third of adults with type II DM (T2DM) had onychomycosis in several cross-sectional studies [1][2][3][4].While a positive association between onychomycosis and diabetes is well established, prior analyses of these conditions have been correlative and may have been confounded by other factors including age, sex, and environmental exposures.To distinguish the causal role of T2DM on onychomycosis from reverse causation and confounding, we performed two-sample Mendelian randomization (MR) analysis.

Materials and Methods
Individuals recruited to the All of Us Research Program between May 6, 2018, and December 20, 2022 who had undergone microarray genomic profiling and were of predicted European genetic ancestry were identified.International Classification of Diseases (ICD) and Systematized Nomenclature of Medicine (SNOMED) diagnostic codes from electronic health record (EHR) data were used to identify cases of tinea unguium (ICD10 B351, SNOMED 402134005), tinea skin infections (tinea pedis: ICD10 B353, SNOMED 6020002; tinea corporis: ICD10 B354; SNOMED 84849002; tinea cruris: ICD10 B356; SNOMED 399029005; SNOMED 399172001; tinea manus: ICD10 B352, SNOMED 48971001), and T2DM (ICD10 E11.9, SNOMED 44054006).Each tinea and onychomycosis case was matched to three controls by age, sex, and genomic ancestry principal components using nearest neighbor exact matching.T2DM prevalence in fungal infections of skin and skin appendages was compared to matched controls through multivariate logistic regression, with age, sex, and genetic ancestry principal components as covariates.The Wald test was used to calculate statistical significance.
T2DM genomic instrument variables and exposure summary statistics for two-sample MR analysis were retrieved from a metaanalysis which included 62,892 cases and 596,424 controls [5].Variants profiled by microarray in the All of Us Research Program with genome-wide significant association (p <5E−8) and uncorrelated within a 10 Mbp window (linkage disequilibrium r 2 < 0.001) were selected, resulting in a total of 46 genomic instrument variables.Outcome variable summary statistics were computed with an additive multivariate logistic regression model with age, sex, and genomic ancestry principal components as covariates.
The primary method of assessing correlation between effect and outcome effect sizes was inverse variance weighted regression.Weighted median regression and MR-Egger regression, which produce consistent estimates of effect size in the presence of invalid instrument variables, were performed as complementary analyses.
Statistical analyses were performed in R (4.0.5) using the software packages MatchIt for nearest neighbor exact matching and TwoSampleMR for MR regression analyses.The software package PLINK (version 1.07) was used to compute outcome variable summary statistics.

Results
In total, 2,974 individuals had onychomycosis and 2,428 had skin dermatophytosis, with age and sex similar between cases and matched controls ( Inverse variance weighted regression identified significant positive effect of T2DM genetic risk on onychomycosis (beta = 0.135, standard error = 5.75E−2, p = 1.86E−2).Weighted median regression produced a comparable estimate of effect size (beta = 0.148).There was no significant effect of T2DM on skin dermatophytosis (Table 2).MR-Egger regression intercepts for nail and skin infections were −7.98E−3 and −2.07E−2 with standard errors of 8.60E−3 and 9.38E−2, respectively, suggesting minimal horizontal pleiotropy.

Discussion
This is the first study using MR to establish causality between diabetes and onychomycosis.Our finding that T2DM has a positive causal effect on onychomycosis risk is consistent with cross-sectional studies performed in Mexico [1], Taiwan [2], India [3], and at multiple sites in the UK, Canada, and the USA [4], which estimated 28-37% onychomycosis prevalence among diabetics, a rate much higher than the 5-10% estimated onychomycosis global prevalence.In addition, obesity, elevated glycated hemoglobin, and elevated triglyceride levels were associated with higher onychomycosis risk in a nested case-control analysis of diabetics, suggesting a correlation between infection and metabolic disease severity [2].
While several risk factors, including humidity, perspiration, excessively tight clothing/footwear, and exposure to fomites, are common to both onychomycosis and dermatophyte skin infections, there was no statistically significant causal relationship between T2DM and tinea skin infection, suggesting that distinct factors affect risk for nail versus skin dermatophytosis.For instance, unrecognized nail trauma due to diminished peripheral sensation and impaired wound healing due to vascular insufficiency and immune cell dysfunction may specifically mediate onychomycosis infection in diabetics [4].
While the identity of causal fungal species was not available for this study, future studies should also examine whether susceptibility to skin and nail infection in diabetes is subject to species-specific factors.
Limitations include small sample size, which reduced statistical power to detect causal relationship with modest effect sizes.EHR records might have missed some dermatophytosis cases and incorrectly included cases not confirmed mycologically.
In conclusion, our results suggest that T2DM has a statistically significant positive causal effect on onychomycosis.Onychomycosis may impair mobility and occupational function and cause cellulitis due to direct extension of infection or secondary infection following disruption in skin integrity.Hence, patients with diabetes should be screened for onychomycosis and counseled on infection risk.

Table 1 .
Demographics and prevalence of T2DM among tinea nail and skin infection cases and controls