Causal association of sex hormones with osteoporosis: a two-sample Mendelian randomization study

Abstract

Background Epidemiologic studies have identified a correlation between sex hormones and osteoporosis(OP), yet the presence of a causal effect remains unverified. Confounding factors and reverse causal associations mean observational studies merely reflect this correlation, without clarifying the intricate causal relationships at play. This study, therefore, employed Mendelian randomization to thoroughly examine the causal link between sex hormones and osteoporosis. Methods Exposure variables for sex hormones, encompassing total testosterone (TT), bioavailable testosterone (BAT), sex hormone binding globulin (SHBG), and estradiol (E2), were derived from the IEU Open GWAS aggregated data. The outcome variable, osteoporosis, was sourced from the UK Biobank database. The Mendelian randomization analysis was conducted using the inverse variance weighted analysis (IVW) method, MR-Egger, and weighted median (WM) method to assess the causal relationship between sex hormones and osteoporosis. Outliers were identified via MR-PRESSO, the effect of individual SNPs on IVW analysis outcomes was investigated through leave-one-out sensitivity analysis, horizontal pleiotropy was detected using MR-Egger-intercept, heterogeneity among instrumental variables was evaluated with Cochran's Q-test, potential biases in findings were examined through funnel plots, and MR Steiger analysis was employed to explore any reverse causal associations. Results Mendelian randomization revealed that genetically predicted TT levels were inversely correlated with osteoporosis risk [IVW: OR (95% CI): 0.997 (0.995, 0.999), P = 0.004], while BAT exhibited a negative causal association with osteoporosis [IVW: OR (95% CI): 0.990 (0.987, 0.994), P = 4.83E-07]. Conversely, SHBG was positively correlated with osteoporosis [IVW: OR (95% CI): 1.008 (1.003, 1.012), P = 0.0004], and E2 demonstrated no causal effect [IVW: OR (95% CI): 0.981 (0.960, 1.002), P = 0.07129]. Cochran's Q-test indicated no heterogeneity in the instrumental variables (IVs) for BAT and E2(P > 0.05) and detected heterogeneity in the IVs for TT and SHBG (P < 0.05), thereby underscoring the relevance of the random-effects IVW model. Neither MR-Egger intercept nor MR- PRESSO analyses detected any level pleiotropy (P > 0.05). The IVW results remained robust after sequential omission of SNPs via the leave-one-out method.MR Steiger analysis confirmed the absence of reverse causality. Conclusion Our results indicate a causal relationship between sex hormones and osteoporosis, wherein TT and BAT decrease osteoporosis risk, whereas SHBG increases it. These findings suggest that sex hormones levels are significant in osteoporosis prevention and treatment strategies