Testosterone (T) therapy in older men with low serum T levels increases lean body mass and decreases fat mass. These changes might improve physical performance and strength; however, it has not been established whether T therapy improves functional outcome in older men. Moreover, concerns exist about the impact of T therapy on the prostate in older men. The administration of finasteride (F), which partially blocks the conversion of T to the more potent androgen, dihydrotestosterone, attenuates the impact of T replacement on prostate size and prostate-specific antigen. We hypothesized that T replacement in older, hypogonadal men would improve physical function and that the addition of F to this regimen would continue to provide the T-induced improvements in physical performance, strength, and body composition. Seventy men with low serum T (<350 ng/dl), age 65 yr and older, were randomly assigned to receive one of three regimens for 36 months: 1) T enanthate, 200 mg im every 2 wk, with placebo pills daily (T-only); 2) T enanthate, 200 mg every 2 wk, with 5 mg F daily (T + F); or 3) placebo injections and pills (placebo). We obtained serial measurements of timed physical performance, grip strength, lower extremity strength, body composition (by dual-energy x-ray absorptiometry), fasting cholesterol profiles, and hormones. Fifty men completed the 36-month protocol. After 36 months, T therapy significantly improved performance in a timed functional test when compared with baseline and placebo [4.3 +/- 1.6% (mean +/- sem, T-only) and 3.8 +/- 1.0% (T + F) vs. -5.6 +/- 1.9% for placebo (P < 0.002 for both T and T + F vs. placebo)] and increased handgrip strength compared with baseline and placebo (P < 0.05). T therapy increased lean body mass [3.77 +/- 0.55 kg (T-only) and 3.64 +/- 0.56 kg (T + F) vs. -0.21 +/- 0.55 kg for placebo (P < 0.0001)], decreased fat mass, and significantly decreased total cholesterol, low-density lipoprotein, and leptin, without affecting high-density lipoprotein, adiponectin, or fasting insulin levels. These results demonstrate that T therapy in older men with low serum T improves physical performance and strength over 36 months, when administered alone or when combined with F, and suggest that high serum levels of dihydrotestosterone are not essential for these beneficial effects of T in men.