Antidepressants of the “Prozac generation” have been hailed as miracle drugs and they're a multibillion-dollar boon to the pharmaceutical industry. But a controversial new study claims that the drugs, which largely replaced older medicines in the 1990s, do little good for the vast majority of patients who take them. Only in the most severely depressed people do these so-called selective serotonin reuptake inhibitors (SSRIs) really outperform placeboes, according to the paper, which analyzed both published and unpublished studies.
Many published papers have demonstrated that SSRI's help relieve depression. But clinical trials in which drugs fail sometimes don't appear in journals, a phenomenon called publication bias. These negative trials do get sent to the U.S. Food and Drug Administration (FDA), however, and the agency looks at all trials performed on a drug before approving it. Psychologist Irving Kirsch of the University of Hull, U.K., relied on FDA's requirement to better assess true benefits of antidepressants.
Using the Freedom of Information Act, Kirsch obtained from FDA the trial results for all SSRIs approved between 1987 and 1999. In a much-debated paper published in 2002, Kirsch showed that the drugs lowered patients' scores on the widely-used 52-point Hamilton Rating Scale of Depression only 1.8 points more than did placeboes--a difference that's statistically significant yet almost meaningless.
But after the 2002 report, critics charged that Kirsch's overall finding might mask improvements caused by the drugs in certain subgroups of patients. So Kirsch and colleagues at four institutes in the U.S. and Canada plowed through the same data again, this time hunting for a link between severity of depression and drug efficacy. Their analysis included 35 trials of fluoxetine, venlafaxine, nefazodone, and paroxetine (known under the brand names Prozac, Effexor, Serzone, and Paxil respectively.) And again, the results were unimpressive, Kirsch's team reported Monday in PLoS Medicine.
Compared to placeboes, the SSRIs showed no benefit in moderately depressed patients. Only those starting with the very highest depression rating--corresponding to s severely debilitating disease--reaped a clinically significant benefit. Because SSRIs, like all drugs, come with risks, most patients should try other treatments, such cognitive therapy, first, says Kirsch.
But others are not convinced that an entire generation of depression drugs does so little. In the studies Kirsch analyzed, a very high proportion of patients responded to placebos, says psychiatrist Peter Kramer of Brown Medical School in Providence, Rhode Island, author of the bestseller Listening to Prozac. That may have been because researchers, eager to enroll as many patients as they could, included many who were not that ill at all, he says. That bias could mask a strong SSRI response among those who took the drugs.
The design of clinical trials are designed before a depression drug is approved also tends to boost placebo rates, says David Nutt, head of the Psychopharmacology Unit at the University of Bristol. Volunteers get extra attention from study coordinators, which in itself may help them recover. "Anti-depressants work in clinical practice," adds Nutt. "Everybody knows they work."
Kirsch says he's not surprised that doctors and patients have a more favorable impression of SSRIs than he believes they deserve. Many depressed patients improve no matter what, he says, and it's easy to attribute improvement to a drug.