Bayesian Mendelian randomization reveals a protective effect of later age at first sexual intercourse against erectile dysfunction.

Erectile dysfunction (ED) is a prevalent health condition with significant psychosocial impacts, yet the causal role of age at first sexual intercourse (AFS) remains unclear. This study investigated the causal effect of AFS on the risk of ED using Mendelian randomization (MR) and Bayesian methods. Five traditional 2-sample MR analyses and 5 Bayesian MR analyses were performed using genome-wide association studies summary statistics from European populations. Sensitivity analyses included MR Egger regression, MR-pleiotropy residual sum and outlier, and Cochran Q-test. In mixed-sex cohorts (Groups 1 and 2), inverse variance weighted results demonstrated significant protective effects: odds ratio (OR) = 0.626, θ = -0.469, P = 2.73 × 10-6 for Group 1 and OR = 0.617, θ = -0.483, P = 3.56 × 10-5 for Group 2. The analyses for male-specific cohorts (Groups 3-10) showed weaker but consistent effects. For Group 3, OR = 0.643, θ = -0.442, P = .010. For Group 4, some instrumental variables associated with confounders were removed. The result became statistically insignificant: OR = 0.680, θ = -0.385, P = .064. For Group 5, the instrument selection criteria were relaxed and significance was retained: OR = 0.695, θ = -0.364, P = .016. For Groups 6 to 10, Bayesian MR was used to strengthen the inferences. In particular, for Group 8, which has a strongly informed prior, a posterior mean θ = -0.358 and a 95% credible interval (-0.575, -0.136) were obtained. This study provides evidence supporting a causal protective effect of later AFS on ED risk. While traditional MR analyses in male-specific cohorts yielded suggestive results, Bayesian MR analyses, which allow for the integration of prior evidence, provided more precise estimates and strengthened the causal inference. These findings may inform future sexual health policies. Strengths include the use of male-specific cohorts and Bayesian enhancement for weak instruments. Limitations include reliance on European-ancestry data and inability to stratify ED subtypes.