Why Antidepressants Don’t Live Up to the Hype
But in the past few years, researchers have challenged the effectiveness of Prozac and other SSRIs in several studies. For instance, a review published in the Journal of Affective Disorders in February attributed 68% of the benefit from antidepressants to the placebo effect. Likewise, another paper published in PLoS Medicine a year earlier suggested that widely used SSRIs, including Prozac, Effexor and Paxil, offer no clinically significant benefit over placebos for patients with moderate or severe depression. Meanwhile, pharmaceutical companies maintain that their own research shows that SSRIs are powerful weapons against depression. (Here’s a helpful blog post that summarizes the debate.)
Now a major new study suggests that both critics and proponents might be right about SSRIs: the drugs can work, but they appear to work best only for a subset of depressed patients, those with a limited range of psychological problems. People whose depression is compounded with, say, substance abuse or a personality disorder may not get much help from SSRIs — unfortunately for the 45% to 60% of patients in the U.S. who have been diagnosed with a common mental disorder like depression and also meet the criteria for at least one other disorder, such as substance abuse. (Such multiple diagnoses are known in medical parlance as comorbidities.)
The new study, published online in April by the American Journal of Psychiatry, was conducted using data from a large, government-funded trial called Sequenced Treatment Alternatives to Relieve Depression, which usually goes by the moniker STAR*D. The STAR*D project, which collected data from 2001 to 2004 at 41 U.S. psychiatric facilities, was one of the most ambitious efforts ever to understand how best to treat people with major depression. STAR*D participants comprise a powerful research sample because they are highly representative of all depressed Americans. Very few depressed people were excluded from STAR*D; only women who were pregnant, those with seizure disorders and a few others with acute conditions were kept out. All other psychiatric and medical comorbidities were allowed.
The authors of the new paper, a team of 11 researchers led by University of Pittsburgh professor of epidemiology Stephen Wisniewski, were curious how the STAR*D group would compare to a typical group of patients selected for a run-of-the-mill drug-company trial for a new antidepressant — the very trials on which the Food and Drug Administration bases its decisions regarding new drug approval. Drawing on their own experiences helping to conduct such clinical trials, which have far more stringent inclusion criteria than the STAR*D group, Wisniewski and his team divided the STAR*D patients into two groups — an “efficacy” sample of patients who would normally be included in a typical Phase III clinical trial for a new antidepressant and a “nonefficacy” sample of patients who would normally be rejected.
Depressed STAR*D patients who were classified for inclusion had no more than one general medical condition (like, say, heart disease) and no more than one additional primary psychiatric disorder besides depression. All patients with multiple comorbidities — along with anyone whose depression had lasted more than two years — were excluded. Once the authors crunched all the numbers, they found that only 22% of STAR*D patients met entry criteria for a conventional antidepressant trial.
All the STAR*D patients were taking citalopram, an SSRI marketed in North America as Celexa. Not surprisingly, those who met standard inclusion criteria for a clinical trial had significantly better outcomes on the drug. In the efficacy group, 52% responded to Celexa versus 40% of the nonefficacy group. Patients in the latter group also took longer to respond and had to be readmitted to psychiatric settings more often. “Thus,” the authors conclude, “current efficacy trials suggest a more optimistic outcome than is likely in practice, and the duration of adequate treatment suggested by data from efficacy trials may be too short.”
To bolster their findings, the authors cite a smaller 2002 study that arrived at similar results: in that paper, published in the American Journal of Psychiatry, Dr. Mark Zimmerman of Brown University and his colleagues found that of 315 patients with major depressive disorder who sought care, only 29, or 9.2%, met typical criteria for an efficacy trial. Similarly, psychologist Ronald Kessler of Harvard coauthored a 2003 paper in theJournal of the American Medical Association that concluded that most “real world” patients with major depression would be excluded from clinical trials because of comorbidities.
Such findings help explain why antidepressants haven’t quite lived up to their promise. But the University of Pittsburgh’s Wisniewski, the lead author of the new study, cautions against interpreting the results as an indictment against greedy drug companies eager to exclude difficult patients in order to show better results. “If the population in a [clinical] trial were more representative, that would come at a cost,” he says. Researchers expect a certain number of bad reactions during clinical trials; some of these reactions can cause serious medical problems. If patients enter a trial with multiple complications — if they are, say, not only depressed but cocaine-addicted, hypertensive and diabetic — you dramatically increase the chances of adverse side effects. “That’s why trials to determine efficacy are done on a relatively homogeneous population,” Wisniewski says.
That’s understandable, but the new study does shed light on the limitations of antidepressants. Conducting clinical trials with representative samples would undoubtedly be more complex — and expensive — since patients with multiple risk factors would have to be monitored more carefully. But for a future generation of antidepressants to be truly effective for most patients, more inclusive trials may be the best answer.