The role of psychological symptoms in recognizing late-onset hypogonadism (LOH) is still controversial. The aim of the study is to evaluate the association between LOH and specific psychological symptoms, and to verify whether investigating intra-psychic domain improves the accuracy of a validated case-history tool (ANDROTEST) in detecting LOH. A consecutive series of 1009 subjects (mean age 49.23±13.34) consulting for sexual dysfunction was studied. Intra-psychic symptoms were investigated by Middlesex Hospital Questionnaire (MHQ), a self-reported questionnaire for screening of mental disorders. A minimum set of two MHQ items was identified through iterative receiver-operating characteristic analysis, with assessment of sensitivity and specificity for hypogonadism (calculated free testosterone <0.225 nmol l(-1)) in an exploratory sample of 462 patients. Sensitivity and specificity were verified in a validation sample of 547 subjects, in which the final two-item version showed an accuracy of 58.4±3.2% in detecting hypogonadism. The combination of the two-item score with ANDROTEST increased the accuracy in predicting hypogonadism (0.741±0.029; P<0.0001) when compared with ANDROTEST (0.696±0.018; P<0.0001) and the two-item score (P<0.05) alone. Hence, combining these two psychological symptoms with a physical scoring system improves its ability in detecting hypogonadism. The combination of the scores should be tested in other studies.