OBJECTIVE: Previous studies have demonstrated an association between sex hormone-related traits and systemic lupus erythematosus (SLE). However, because of the difficulties in determining sequential temporality, the causal association remains elusive. In this study, we used two-sample Mendelian randomization (MR) to explore the genetic causal associations between sex hormone-related traits and SLE.
METHODS: We used a two-sample MR to explore the causal association between sex hormone-related traits and SLE. The summarized data for sex hormone-related traits (including testosterone, estradiol (E2), sex hormone-binding globulin (SHBG), and bioavailable testosterone (BT)) originated from large genome-wide association studies (GWASs) of European descent. Aggregated data for SLE were derived from the FinnGen consortium (835 cases and 300,162 controls). Random-effects inverse-variance weighted (IVW), MR-Egger, weighted median, simple mode, weighted mode, and fixed-effects IVW methods were used for the MR analysis. Random-effects IVW was the primary method used to analyze the genetic causal association between sex hormone-related traits and SLE. Heterogeneity of the MR results was detected using the IVW Cochran's Q estimates. The pleiotropy of MR results was detected using MR-Egger regression and the MR pleiotropy residual sum and outlier (MR-PRESSO) test. Finally, leave-one-out analysis was performed to determine whether MR results were affected by a single single-nucleotide polymorphism (SNP).
RESULTS: Random-effects IVW as the primary method showed that testosterone (odds ratio (OR), 0.87; 95% confidence interval (CI), 0.41-1.82; P = 0.705), E2 (OR, 0.95; 95% CI, 0.73-1.23; P = 0.693), SHBG (OR, 1.25; 95% CI, 0.74-2.13; P = 0.400), and BT (OR, 0.99; 95% CI, 0.67-1.47; P = 0.959) had no potential causal association with SLE. The MR-Egger, weighted median, simple mode, weighted mode, and fixed-effects IVW methods all indicated consistent results. The results of the MR-Egger regression showed that there was no pleiotropy in our MR analysis (P > 0.05). The IVW Cochran's Q estimates showed that the MR analysis results of E2, SHBG, and BT on SLE had no heterogeneity (P > 0.05), but testosterone and SLE had heterogeneity (P < 0.05). The leave-one-out analysis confirmed that a single SNP did not affect the MR results.
CONCLUSIONS: Our MR analysis demonstrated that genetically predicted testosterone, E2, SHBG, and BT levels were not associated with SLE risk, but the roles of other non-genetic pathways cannot be ruled out. Key Points • This is the first MR study to explore the causal association of sex hormone-related traits with SLE. • No evidence to support causal associations between sex hormone-related traits and SLE. • Our MR analysis may provide novel insights into the causal association between sex hormone-related traits and SLE risk.
