ABSTRACT Background Klinefelter syndrome (KS) is characterized by marked phenotypic heterogeneity that might be influenced by genetic modifiers, including androgen receptor (AR) repeat length (CAGn and GGCn). The clinical relevance of these repeat lengths in patients with KS before testosterone replacement therapy (TRT) remains unclear. Objectives To investigate the association between AR repeat length and anthropometric, hormonal, metabolic, and reproductive parameters in a well‐characterized cohort of untreated adult patients with KS. Materials and Methods In this cross‐sectional single‐center study, 214 men with classical 47, XXY karyotype were evaluated prior to TRT. Clinical, biochemical, and reproductive parameters were analyzed according to AR CAGn and GGCn repeat length. Nonparametric tests, multivariable linear and logistic regression models, and interaction terms between CAGn and GGCn were tested. Standardized beta coefficients were used to compare the relative contribution of AR repeat length with major clinical determinants. Results In unadjusted analyses, CAG repeat length was associated with estradiol concentrations, whereas GGC repeat length showed associations with hematocrit, platelet count, and total cholesterol. However, most associations were characterized by small effect sizes and did not persist after multivariable adjustment for possible confounders (age, BMI, and total testosterone levels). Moreover, AR repeat length was not associated with sperm retrieval rate. Standardized beta analyses demonstrated that testosterone levels, BMI, and age accounted for the largest proportion of phenotypic variability, whereas CAGn and GGCn repeat length had minimal roles. Discussion and Conclusion In untreated patients with KS, AR repeat length (CAGn and GGCn) appears to have a limited clinical impact compared with classical endocrine and metabolic determinants. These findings suggest that phenotypic variability in KS might be primarily driven by chromosomal aneuploidy and primary testicular dysfunction rather than AR repeat length.
Graziani et al. (Fri,) studied this question.