University of Pennsylvania researchers Jay Fournier, MA, and Robert DeRubeis, PhD, led an analytical team of 5 psychiatrists and psychologists. They analyzed 6 randomized placebo-controlled trials of the tricyclic antidepressant imipramine(Drug information on imipramine) (Tofranil and others) or the SSRI paroxetine(Drug information on paroxetine) (Paxil). The trials, involving 718 adult outpatients with mild to very severe depression, included an antidepressant/placebo comparison of at least 6 weeks’ duration using Hamilton Depression Rating Scale (HDRS) scores at intake and at the end of treatment, without a placebo washout period.
DeRubeis told Psychiatric Times that differences in treatment effect between antidepressant and placebo were significantly affected by initial symptom severity. His team concluded, “Whereas antidepressant medications can have a substantial effect with more severe depressions, there is little evidence to suggest that they produce specific pharmacological benefit for the majority of patients with less severe acute depressions.”
For patients with an initial HDRS score of less than 23, Cohen d effect sizes for the difference between medication and placebo were estimated to be less than 0.20 (a standard definition of a small effect). As baseline severity increased, so did the magnitude of the benefit of antidepressant compared with placebo. The threshold for a clinically significant difference of 3 HDRS points was not crossed until a baseline HDRS score of 25.
Most people with depression do not fall into the very severe (HDRS, 23 or higher) category, DeRubeis said. He points to research that found that 71% of patients assessed had HDRS scores below 22.2
More than 164 million prescriptions for antidepressants are dispensed each year in this country, making antidepressants among the most commonly prescribed medications. The efficacy of antidepressants is based primarily on studies of patients with more severe forms of depression. This fact, according to the analytical team, has not been reflected in marketing messages to clinicians and the public. Therefore, it is important, DeRubeis explained, to look at the effectiveness of antidepressants for mild to moderate depression.
DeRubeis warned against drastic responses to the team’s findings. No simple recommendations follow, he said. However, our data suggest that practitioners examine the findings, think about them, and in a wise and responsible way, factor them into their clinical decision making.
When asked whether insurance companies might deny reimbursement for depression treatment in many cases on the basis of the study findings, DeRubeis responded, “I would find that reaction foolish and unwise in the extreme, because our findings don’t say that you shouldn’t provide antidepressants. They say that placebo does as well or almost as well as full medicine treatment for mild to moderate depression.”
The question—for insurance companies and clinicians—is which treatment to provide in their place, DeRubeis said. Helpful alternatives include physical exercise; bibliotherapy; and empirically supported psychotherapies, such as cognitive-behavioral therapy (CBT).
DeRubeis also cautioned that the meta-analysis results applied only to acute treatment—and not to continuation or maintenance treatment. The findings are not relevant to treatment of patients with chronic depression, said DeRubeis. “Those who have been depressed for a long time are unlikely to respond to placebo, so clinicians should start them on medications and give them some hope,” he added. Study results have led to fierce debates on such issues as treatment and study design.
Treatment adequacy
In a January 9 New York Times article, “The Wrong Story About Depression,” which discussed the meta-analysis, former columnist Judith Warner focused on inadequate treatment by nonspecialists. “Most patients with depression are treated by general practitioners, not psychiatrists. Studies have shown that these primary care doctors don’t strenuously enough screen their patients for depression before prescribing drugs, or closely monitor their care afterward . . . the trouble is not that the drugs don’t work; it’s that the care is not very good.”
Warner cited a study that found that “despite the availability of highly effective [psychological] interventions, relatively few psychologists learn or practice these interventions.”3 One of the most studied therapies is CBT, for which there is valid evidence for relapse prevention in depression, said DeRubeis. But some problems exist: How many cognitive therapists are good at dealing with depression? Do they work close to where their patients live? How many patients are willing to talk with them?
Two antidepressants
Critics also questioned the team’s restriction of studies to 2 antidepressants. “Its conclusions are based on only 2 antidepressants when there are 25 or so on the market,” said Richard Friedman, MD, professor of psychiatry at Weill Cornell Medical College, in his January 12 New York Times column, “Before You Quit Antidepressants.” “Antidepressants are not interchangeable; studies show that a patient who fails to respond to one has about a 30% chance of responding to another,” said the Science Times writer.
“Yes, we only had data on 2 different drugs,” DeRubeis acknowledged, “but most experts know that imipramine is quite effective, despite its problems with side effects and lethality risk. Most would say that responses to paroxetine are a pretty good predictor of what you would find with most SSRIs.”
But in an interview with MedPage Today, David Hellerstein, MD, associate professor of clinical psychiatry at the Columbia University College of Physicians and Surgeons and director of the mood disorders research program at Columbia University Medical Center in New York, questioned whether the findings about the 2 drugs can be generalized to other antidepressants and other studies.4 Paroxetine and imipramine, he noted, are associated with more adverse effects than other antidepressants, which may have affected compliance and dropout rates. Moreover, in the 3 studies involving imipramine, the dose was “subtherapeutic” (100 mg/d) in one and less than optimal (100 to 200 mg/d) in another.
Placebo response
While conceding the dearth of studies that include large numbers of people with mild to moderate depression, Friedman criticized the analytical team for excluding a whole class of studies that tried to correct for the so-called placebo response. “An analysis that eliminates such studies is bound to show a comparatively small average difference between drug treatment and placebo treatment. Not surprisingly, this is just what happened in the recent analysis,” Friedman said. “But in randomized clinical trials that try to correct, or wash out, the placebo effect, patients with mild to moderate depression respond to antidepressants at rates nearly identical to those of patients with severe depression” (who tend to have a much lower response to placebos).
In their January 19 response letter to The New York Times, “Critiquing a Critique,” DeRubeis and Fournier argued that Friedman’s endorsement of studies that screen out placebo responders “misses the point that our study aimed to estimate the difference between the effects of medications and placebo.” They took Friedman to task for “his failure to acknowledge findings consistent with our own from 2 previous reviews.” Teams led by Kirsch5 and Khan6 found that antidepressant/placebo difference increases as initial severity increases. “We used individual patient data and included patients with less severe depression, whereas both the Kirsch and the Khan teams analyzed group means that largely excluded patients with HDRS scores below 20,” they said.
Assessment tools
In an upcoming editorial for the Journal of Clinical Psychopharmacology, Ronald Pies, MD,7 Psychiatric Times’ editor in chief and professor of psychiatry at SUNY Upstate Medical Center at Syracuse, said there may be reasons tangentially related to antidepressants that explain why studies of these agents have yielded disappointing results in recent decades.
If interviews producing HDRS scores are not performed skillfully, study results may be distorted. Kobak and colleagues8 pointed to several instances in which poor interviewing techniques led to outcomes that showed little difference between antidepressant and placebo. Conversely, good interviewing techniques led to a more robust improvement rate (effect size) for the antidepressant.
Assessment tools can certainly be improved, DeRubeis said, but while the HDRS results showed only a small difference between placebo and active drug in mild and moderate depression, the same Hamilton interviewers found big differences with more severe depression.
Multistage strategy
Pies called for a multistage strategy in dealing with depression. The kind of meta-analyses done by the Kirsch and Fournier groups involve “crunching numbers” on individual trials in which usually a single antidepressant was tested over a few weeks, he said. “To be sure, the effect sizes [compared with placebo] are often unimpressive in such single-stage trials. But when psychiatrists use a ‘full court press’ and treat depressed patients over many months, using various combination and augmentation strategies, we often see better results with antidepressant medication,” he said.
The STAR*D studies, he said, provide decent, though not conclusive, evidence that antidepressants are effective in major depressive disorder (MDD) when administered in a robust, appropriately dosed, multistaged strategy.9 Although the STAR*D studies precluded the use of a placebo group, Pies said after the fourth and final pharmacotherapy “hoop” was jumped through, the cumulative rate of remission (full resolution of the depressive syndrome) in patients with resistant nonpsychotic MDD was about 67%—much higher than historically reported rates of remission with placebo, which average around 25% in a single trial.
Reacting to criticisms of the mega-analysis, DeRubeis said, “We have been accused of somehow trying to bad-mouth antidepressants. That is by no means true. We just want to know what to expect from our treatments, and the majority of placebo-controlled findings have been obtained with antidepressants.”
DeRubeis sees the potential of antidepressants for treating other related disorders. He cites a study he coauthored that examined whether SSRI recipients with MDD report greater changes in neuroticism and extraversion than patients receiving inert placebo.10 Paroxetine appeared to have a specific pharmacological effect on personality distinct from its effect on depression. “The study,” he said, “is one of the first to look carefully at other potentially important effects of SSRIs on the psychology of the depressed patient.”
