Elucidating causal relationships of diet-derived circulating antioxidants and the risk of non-scarring alopecia: A Mendelian randomization study

Abstract

Previous observational studies revealed controversy about the effect of circulating antioxidants on risk of alopecia. In the present study, we investigated the causal relationships between diet-derived circulating antioxidants and 2 non-scarring alopecia using Mendelian randomization (MR). Instrumental variables for antioxidants (lycopene, retinol, ascorbate, β-carotene, α-tocopherol, and γ-tocopherol) were selected from published studies. Data for alopecia areata (AA) and androgenetic alopecia (AGA) was obtained from the FinnGen study project (R9 released in 2023), including 195 cases and 201,019 controls for AGA and 682 cases and 361,140 controls for AA. We used the inverse variance weighted method as the primary MR method. Three additional methods were used as sensitivity analysis to validate the robustness of the results. We found a causal relationship between absolute β-carotene levels and AGA risk (P = .039), but not with AA (P = .283). The results of Wald ratio showed a protective effect of absolute β-carotene levels against AGA, with per 0.1 ln-transformed β-carotene being associated with a 76% lower risk of AGA (OR: 0.24, 95% CI: 0.06-0.93). Based on the fixed effects inverse variance weighting results, we found that α-tocopherol was protective against both AGA (P = .026) and AA (P = .018). For each unit increase in α-tocopherol, the effects of change in AGA and AA were 0.02 (95% CI: 0.00-0.61) and 0.10 (95% CI: 0.01-0.67), respectively. The results did not reveal any other causal relationships. Our study identified 3 causal associations of antioxidants with the risk of non-scarring alopecia. These results provide new insights into the prevention of non-scarring alopecia through diet.

Figure 1.

Schematic overview of the MR study design. (A) The 3 main assumptions of MR analysis. (B) Schematic overview and framework of the present MR study design. MR = Mendelian randomization.

Figure 2.

The primary MR analyses results of the causal effects of absolute circulating antioxidant levels on AGA and AA. Significant and suggestive results highlighted in bold. “-” represents not applicable. AA = alopecia areata, AGA = androgenetic alopecia, MR = Mendelian randomization.

Figure 3.

The leave-one-out plots and funnel plots for significant results. (A) Leave-one-out plot for α-tocopherol on AA. (B) Leave-one-out plot for α-tocopherol on AGA. (C) Funnel plot for α-tocopherol on AA. (D) Funnel plot for α-tocopherol on AGA. AA = alopecia areata, AGA = androgenetic alopecia.

Figure 4.

The primary MR analyses results of the causal effects of circulating antioxidant metabolites on AGA and AA. Significant and suggestive results highlighted in bold. “-” represents not applicable. AA = alopecia areata, AGA = androgenetic alopecia, MR = Mendelian randomization.