Testosterone (T) levels decline as men age, but it is unclear whether this has an effect on cognition. Some studies indicate that lower T levels are associated with memory loss; thus, maintaining a higher T level could have positive effects on aspects of cognitive function. Concerns exist, however, about the effect of T therapy on the prostate in older men. We hypothesized that T replacement in older men with low T levels would improve aspects of cognitive function and that the addition of finasteride would not affect the T-induced cognitive improvements. Healthy men, 65 to 83 years of age, with baseline total T below 350 ng/dL and no evidence of cognitive impairment were randomly assigned to 1 of 3 regimens: 200 mg of T every 2 weeks by intramuscular injection with placebo pill daily (T-only), 200 mg of T every 2 weeks by intramuscular injection with 5 mg of finasteride daily (T+F), or placebo injections and pills (placebo). Sixty-nine men completed baseline cognitive testing; 65 completed at least 4 months, and 46 completed all 36 months of the study. Participants were given a battery of cognitive evaluations at baseline, 4 months, and 36 months, along with measurement of serum hormone levels. Serum total T, bioavailable T, and estradiol levels in the T-only and T+F groups significantly increased throughout the treatment period, whereas these hormone levels did not change in the placebo group. Only minimally significant differences were seen among the 3 groups in any evaluation of cognitive performance, either in the short-term (4 months) or the long-term (36 months) analysis. These results indicate that T replacement, whether given alone or in combination with finasteride, for 36 months in healthy older men without cognitive impairment at baseline has no clinically significant effect on tests of cognitive function. Further studies are warranted to determine whether hormone replacement in men with preexisting cognitive impairment is beneficial.