To investigate the causal relationship between testosterone (BT) levels and Alzheimer’s disease (AD) risk and to quantify the role of obesity and lipid metabolism as potential mediators.

We used a two-sample, two-step MR to determine:1) the causal effect of BT levels on AD; 2) the causal effect of two lipid metabolites, obesity and LDLc on AD; and 3) the mediating effects of these metabolites. Pooled data for BT levels and lipid metabolism were obtained from the UK Biobank. AD data were obtained from the Alzheimer’s Disease Project International Genomics Consortium, FinnGen Consortium, and UK Biobank study. Effect estimates from external genome-wide association study (GWAS) pooled statistics were obtained using inverse variance-weighted (IVW) MR analysis.

Higher levels of BT were associated with a reduced risk of AD (odds ratio [OR] 0.9992, 95% CI 0.9985–0.9998, P = 0.019), and there was a negative correlation with LDLc (OR 0.9208, 95% CI 0.8569–0.9895, P = 0.024) and obesity class 2 (OC2) (OR 0.7445, 95% CI 0.5873–0.9437, P = 0.014). Conversely, there was a positive correlation between LDLc (OR 1.0014, 95% CI 1.0000–1.0029, P = 0.043) and OC2 (OR 1.0005, 95% CI 1.0001–1.0009, P = 0.003) and AD. Mediation analysis showed that the indirect effect of BT levels on AD was achieved through LDLc and OC2, which accounted for 17% and 17% of the total effect, respectively.

Our study identified a causal role of BT levels in LDLc and OC2. BT levels may affect AD through LDLc and OC2 metabolic processes.