Aging in men is accompanied by a progressive, but individually variable decline of serum testosterone production, more than 20% of healthy men over 60 yr of age presenting with serum levels below the range for young men. Albeit the clinical picture of aging in men is reminiscent of that of hypogonadism in young men and decreased testosterone production appears to play a role in part of these clinical changes in at least some elderly men, the clinical relevancy of the age-related decline in sex steroid levels in men has not been unequivocally established. In fact, minimal androgen requirements for elderly men remain poorly defined and are likely to vary between individuals. Consequently, borderline androgen deficiency cannot be reliably diagnosed in the elderly, and strict differentiation between "substitutive" and "pharmacological" androgen administration is not possible. To date, only a few hundred elderly men have received androgen therapy in the setting of a randomized, controlled study, and many of these men were not androgen deficient. Most consistent effects of treatment have been on body composition, but to date there is no evidence-based documentation of clinical benefits of androgen administration to elderly men with normal or moderately low serum testosterone in terms of diminished morbidity or of improved survival or quality of life. Until the long-term risk-benefit ratio for androgen administration to elderly is established in adequately powered trials of longer duration, androgen administration to elderly men should be reserved for the minority of elderly men who have both clear clinical symptoms of hypogonadism and frankly low serum testosterone levels.