The male aging process brings about declines in hormonal function including a gradual decline in bioavailable testosterone levels. Animal studies suggest that testosterone modulates cognitive function through enhancing acetylcholine release and up-modulation of nicotinic receptors. Tau protein deposition is also affected by androgen supplementation in animals. We hypothesize that testosterone replacement in elderly hypogonadal males may improve cognition, in particular the visual-spatial domain. Thirty-six male patients with a new diagnosis of Alzheimer's disease had their total and bioavailable testosterone levels measured. None of the patients had been on acetylcholinesterase inhibitors. Ten of the 36 patients (28%) were deemed biochemically hypogonadal (total testosterone < 240 ng/dl or 7 nmol/l). Five of the hypogonadal patients were randomized to testosterone and five to placebo. Initial Alzheimer's Disease Assessment Scale cognitive subscale (ADAScog) and Mini Mental Status Examination (MMSE) ranged from 31 to 19 and from 17 to 22, respectively. The clock drawing test (CDT) and the pentagon-tracing portion of the MMSE were used as measures of visual-spatial abilities. Normal prostate-specific antigen (PSA) levels were essential before treatment with intramuscular testosterone, 200 mg every 2 weeks. Measurement of testosterone, complete blood count, lipids, PSA and neuropsychological cognitive tests were repeated at 3, 6, 9 and 12 months of treatment. In the testosterone-treated group, levels of total testosterone increased from a mean of 126.4 ng/dl to 341 ng/dl or 3.6 nmol/l to 9.7 nmol/l (p = 0.11). Bioavailable testosterone also increased from a mean of 48.7 ng/dl to 142 ng/dl or 1.39 nmol/l to 4.05 nmol/l (p = 0.10). PSA levels were also elevated from a mean of 0.98 to 1.37 ng/ml (p = 0.07). ADAScog improved from a mean of 25 to 16.3 (p = 0.02); MMSE improved from a mean of 19.4 to 23.2 (p = 0.02), CDT also improved from 2.2 to 3.2 (p = 0.07). One patient stopped treatment because of hypersexual behavior. The placebo-treated group deteriorated gradually. This small pilot study performed in aging male patients suggests that testosterone could indeed improve cognition, including visual-spatial skills in mild to moderate Alzheimer's disease.