Objective There is a high prevalence of low testosterone and dyslipidaemia in men with type 2 diabetes. The androgen receptor CAG repeat polymorphism (AR CAG) affects receptor transcriptional activity (the shorter repeats the more sensitive AR) and is associated with androgenic parameters and obesity. This study describes the relationships between testosterone, AR CAG and serum lipids in men with type 2 diabetes.
Design and Patients Cross-sectional study of men with type 2 diabetes in a District General Hospital Diabetes Centre.
Measurements Correlation between testosterone, AR CAG and serum lipids.
Results HDL cholesterol (HDL-C) correlated with total testosterone (TT) (r = 0·251, P < 0·001), bioavailable testosterone (BT) (r = 0·19, P = 0·001), free testosterone (FT) (r = 0·165, P = 0·005) and sex hormone-binding globulin (SHBG) (r = 0·147, P = 0·014). HDL-C did not correlate with oestradiol, but men with the lowest quartile of oestradiol had lower HDL-C compared to highest quartile (P = 0·046). Triglycerides correlated negatively with TT (r = −0·195, P = 0·001), BT (r = −0·148, P = 0·013) and SHBG (−0·14, P = 0·019) but not with FT or oestradiol. Total and LDL cholesterol (LDL-C) correlated negatively with oestradiol (r = −0·121, P = 0·05) but not with testosterone or SHBG. One-way anova testing across four quartiles of AR CAG showed a trend to alteration in HDL-C across groups of AR CAG (P = 0·08). HDL-C was significantly higher in men with the longest AR CAG compared with the shortest (1·19 vs 1·08 mmol/l, P = 0·02).
Conclusions Lower testosterone and oestradiol levels in men with diabetes are associated with an adverse lipid profile. Shorter AR CAG is associated with low HDL-C and testosterone. The paradox that HDL-C is associated with low testosterone levels and a more active AR may suggest divergent effect of testosterone on HDL-C via genomic vs nongenomic mechanisms.