Elderly men with low testosterone (T) levels are increasingly diagnosed to have partial androgen deficiency of the aging male (PADAM). Frequently, magnetic resonance imaging is performed to exclude pituitary adenoma. The value of GnRH testing to differentiate PADAM from secondary hypogonadism is unknown. Serum levels of T as well as LH and FSH at baseline and after GnRH were evaluated in the following groups: 1) 24 elderly men with low serum T (<11.7 nmol/liter), 2) 25 elderly men with normal serum T levels (>11.7 nmol/liter), 3) 10 men with primary hypogonadism, 4) 24 men with secondary hypogonadism, and 5) 13 healthy young volunteers. In elderly men, T levels were lower (P < 0.001) and gonadotropin levels higher (P = 0.03) compared with younger controls. LH and FSH response to GnRH was higher in elderly men with low T levels (PADAM) compared with elderly men with normal T levels (P = 0.02 and P < 0.001) in the presence of similar basal gonadotropin levels. To differentiate secondary hypogonadism from PADAM with a sensitivity of 100%, a T less than 10 nmol/liter had a specificity of 50%. This specificity was improved to 75% by using a cutoff of less than or equal to 15 mU/liter increase of LH upon GnRH stimulation. Overall, decreased T levels and increased LH levels in elderly men suggest a primary Leydig cell dysfunction. In the subgroup of elderly men with low T levels, an increased LH response to GnRH with normal basal LH levels suggests additional, possibly hypothalamic changes. To exclude secondary hypogonadism in PADAM, diagnostic accuracy can be improved by using GnRH testing.