Science News logo

Science Service logo


Placebos for depression attract scrutiny

B. Bower

New antidepressant medications have gained widespread use in the past decade, and more await approval from the Food and Drug Administration following clinical trials. Much debate currently concerns whether it's ethical for physicians to give placebo pills to depressed volunteers in such studies, instead of providing either the drug being tested or an FDA-approved antidepressant.

An analysis of the FDA's clinical-trial database on recently approved antidepressants now promises to enliven the controversy further. It finds that depressed patients assigned to 4 to 8 weeks of placebo treatment—typical clinical-trial lengths—exhibited no increased risk for suicide or suicide attempts and showed substantial symptom relief. Still, depression subsided to an even greater extent among those receiving antidepressants.

Placebos pack enough therapeutic punch to calm ethical qualms about their use in antidepressant studies, conclude psychiatrist Arif Khan of the Northwest Clinical Research Center in Bellevue, Wash., and his coworkers.

"Patients who are assigned to placebo treatment in clinical trials are not untreated," Khan's team concludes in the April Archives of General Psychiatry. "The capsule they receive is pharmacologically inert, but hardly inert with respect to its symbolic value and its power as a conditioned stimulus."

Like other participants, placebo patients received physical examinations, attention and guidance from a physician, opportunities to talk about their condition, and other assistance that chipped away at their depression, the researchers assert. Psychotherapy of all theoretical stripes may work primarily because of a common emphasis on these types of interventions, they propose.

Khan's group obtained FDA clinical data for the seven antidepressants approved from 1987 through 1997. These drugs, which include fluoxetine (Prozac), sertraline (Zoloft), and paroxetine (Paxil), exert different effects on brain chemistry from those of antidepressants approved earlier.

Of the 45 studies, 23 compared a new drug with a placebo only. The rest compared a new drug with both an approved antidepressant and a placebo. In most cases, the 8,731 participants suffered from moderate depression and had no other serious physical or psychiatric conditions.

During the trials, 34 people killed themselves and 130 others tried to do so. Rates of suicide and attempted suicide were nearly the same for groups treated with placebo or with drugs, the team says.

Depression symptoms declined by 41 percent with new antidepressants, 42 percent with previously approved medications, and 31 percent with placebos.

Such findings will aid researchers and oversight boards struggling to determine the ethics of using placebos in antidepressant studies, Khan's group concludes. However, comments on the analysis, published in the same journal, offer varying interpretations of the results.

Nothing in the report raises concerns about placebo use, holds psychiatrist Paul Leber of Neuro-Pharm Group in Potomac, Md., a private research firm. Some depressed patients respond at least as well to placebos as to antidepressants, making randomized placebo-controlled trials essential, Leber argues.

Epidemiologist Karin B. Michels of the Harvard School of Public Health in Boston disagrees. Antidepressants relieved symptoms better than placebos in the FDA trials, even if no disparity in suicides emerged, she says. A physician who knowingly uses placebos thus unethically deprives patients of effective treatment, in her view.

The FDA data don't adequately assess placebo effects on depressed patients, argues psychologist Helena Chmura Kraemer of Stanford University. For instance, the studies didn't test whether improvement on placebo exceeded that of depressed people given no treatment. Also, more patients dropped out of the trials than completed them, making it difficult to draw conclusions.

Studies of depression treatment lasting at least 6 months have found much higher relapse rates for placebo responders than for those who improve on antidepressants, say Frederic M. Quitkin and Donald F. Klein, both psychiatrists at Columbia University. The FDA review and some other reports (SN: 8/24/96, p. 123) underestimate antidepressants' effectiveness, they argue.


From Science News, Volume 157, No. 18, April 29, 2000, p. 278.