09 Apr 2003

The hormone dehydroepiandrosterone (DHEA) failed to significantly improve cognition in a small pilot trial of patients with Alzheimer's disease, according to a report in the April issue of Neurology.

Funded by the National Institute on Aging and Neuroscience Pharma of Montreal, Canada, the multicenter study was sparked in part by reports of reduced DHEA in the aged and also by in-vitro and animal research ascribing neurotrophic and memory-enhancing effects to DHEA. In the United States, DHEA is available over the counter as a dietary supplement. It has been touted as an "anti-aging hormone," with extravagant claims of benefit against every ill from cancer to decreased libido—and Alzheimer's disease.

Fifty-eight patients with mild to moderate AD were randomized to receive, in double-blind fashion, either 50 mg oral DHEA twice a day (n = 28) or placebo (n = 30) for six months. While there was a trend toward improved cognition in the experimental group at three months, this benefit did not last through six months, nor was it accompanied by any overall improvement, as reported by caregivers and physicians.

One factor that made statistical interpretation difficult was the high dropout rate; only 14 of 30 placebo group patients, and 19 of 28 treatment group patients completed the trial. While noting these limitations in their editorial accompanying the article, David Knopman of the Mayo Clinic in Rochester, Minnesota, and Victor Henderson of the University of Arkansas in Little Rock, also opine that the trial "provides a useful estimate of DHEA effect size on AD symptoms. DHEA alone seems unlikely to be superior to currently available anticholinesterases."

Knopman and Henderson also make the point that trials should be designed to allow for use of anticholinesterase drugs in the early stages of disease, since most AD patients wish to take them. The presence of these drugs, Knopman and others have previously argued, need not hinder interpretation of the value of an experimental drug such as DHEA (Knopman et al., 1998). Adopting this approach may prove useful for evaluating other experimental drugs such as estrogen, nonsteroidal antiinflammatory drugs, statins, and homocysteine-lowering agents.—Hakon Heimer