Drugs are hardly better than a placebo for the mildly depressed, researchers find.

Mildly depressed Americans have spent billions of dollars over the past few decades hoping to relieve their blues by taking designer antidepressants. They might have done just as well by taking a placebo.

That’s the surprising implication of a new study by psychologists at the University of Pennsylvania, Vanderbilt University, the University of New Mexico and the University of Colorado. Many industry-sponsored studies have excluded patients with mild depression, even though these patients often end up in doctors’ offices looking for help. The new study, published in the Journal of the American Medical Association (JAMA), culled the results from six antidepressant trials that did include mild and moderate cases, totaling 718 people overall.

It found that antidepressants drugs were virtually no better than a placebo for people with mild or moderate depression. Only in people with very severe depression did the antidepressant effect become substantially greater than that produced by a dummy pill.

It is unknown exactly how many patients taking antidepressants have milder cases of depression. But one survey cited by the researchers found that 71% of all patients seeking treatment for depression fall in the milder category, where placebos are likely to do as well.

"The evidence we now have suggests there is very little benefit [from antidepressants] for people with less than very severe depression," says study co-leader Robert DeRubeis, a psychologist at the University of Pennsylvania. He says doctors may want to consider alternative treatments, including exercise, cognitive therapy and self-help books, before using drugs to treat milder cases.

The Penn-led study only looked at two drugs, Paxil from GlaxoSmithKline and a well-known older drug called imipramine. But it follows on the heels of a larger 2008 study from University of Hull psychologist Irving Kirsch that compared 35 trials of four different antidepressant medications, including Prozac from Eli Lilly, Effexor from Pfizer and Paxil. This study, published in PLoS Medicine, also found "virtually no difference" between drugs and placebos for the moderately depressed, and only a relatively small difference even in the very severely depressed.

If antidepressants aren’t much better than a placebo for many patients, how is it that researchers are just figuring this out now? Kirsch says it’s a common pattern in medicine to think a new drug has great benefits, only to discover years later that the apparent efficacy is mostly due to the placebo effect.

Another reason the public may have gotten a skewed view of antidepressant efficacy is that drug companies have tended to publicize only their most positive studies. Until recently, some failed antidepressant trials weren’t published in medical journals, or were buried deep in the literature where they were hard to find. "The clinical trials showing the least benefits for drugs compared to placebo were just not published," says Kirsch. "We had to go to the FDA and [submit a Freedom of Information Act request] to get the data."

Dug studies also may overestimate benefits of antidepressants because many drug studies have a placebo "washout" period, which attempts to exclude people likely to respond to a placebo. To get around this potential problem, the University of Pennsylvania study only examined studies without a washout period. This resulted in over 200 studies being excluded from the analysis.

In response to the paper, GlaxoSmithKline said, "The studies used for the analysis in the JAMA paper differ methodologically from studies used to support the approval of Paxil for major depressive disorder, so it is difficult to make direct comparisons between the results." It added, "Since its approval by the FDA in 1992, Paxil has helped millions of people battling mental illness lead more productive, happier lives."

So-called selective serotonin reuptake inhibitors such as Paxil, Prozac and Zoloft, gained widespread use on the theory that depressed people suffer from a deficit of the brain chemical serotonin. But little hard data supports the popular concept, Kirsch says. "This whole idea of serotonin deficiency is a myth."