Antidepressants: can't live with them, can't live without them.

A new wave of meta-analyses suggests that antidepressants are no better than placebo for major depressive disorder (MDD), and therefore, antidepressants not only don't work, but even worse, they harm patients because of the risk of adverse effects. The authors analyzed data from all antidepressant studies submitted to the Food and Drug Administration for registration (including failed studies with inordinately high placebo reponses), or used a metanalysis filter and selected those few studies that meet those criteria. In aggregate, the data, at best, show a clinically trivial advantage of the antidepressants over placebo in acute randomized trials. However, their conclusions range from antidepressants don't work at all to the antidepressants should be reserved only for the most seriously depressed patients. Kirsch is capitalizing on this trend with his recently published book. I view this debate through my perspective of over 25 years of clinical experience, serving as a rater for clinical trials, planning and conducting National Institute of Mental Health and industry efficacy and effectiveness clinical trials, and consulting to the pharmaceutical industry. The real story, I believe, is a bit more complicated. Give everyone with fever penicillin and many will improve. Compare penicillin to placebo and on average, you would find no difference. Why? Most people with fever have viral infections or non-penicillin sensitive bacterial infections that are time limited (eg, common upper respiratory infections). One could reasonably conclude that penicillin doesn't work and we should all take chicken soup instead. Give everyone with MDD an antidepressant and many will improve. Compare any given antidepressant to placebo and on average, you should find no difference. Yet, a difference does exist and even if the trials have an overall small effect size in favor of antidepressants, it is not quite accurate to state that antidepressants are equal to placebo. Why? If that were true, then in a third of the trials placebo would beat antidepressants, in a third of the trials antidepressants would beat placebo, and the remaining third of the trials it would be a tie. MDD is analogous to fever. It is a heterogeneous, nonspecific syndrome that is the final common pathway of multiple dysregulated psychological and brain processes. Genetic epidemiological studies strongly suggest that stress is a causal factor and that the persistence or recur-