The efficacy of antidepressant treatment over placebo for major depressive disorder “varies considerably,” depending upon symptom severity, a recent meta-analysis showed.
Only patients whose depression is classified as “very severe” appear to have a greater benefit from antidepressants than from placebo pills, according to the study by Jay C. Fournier of the University of Pennsylvania, Philadelphia, and his colleagues (JAMA 2010;303:47-53).
Most placebo-controlled studies of antidepressants specifically exclude individuals who score below 23 on the Hamilton Depression Rating Scale (HDRS). HDRS scores of 8-13 indicate mild depression, scores of 14-18 indicate moderate depression, scores of 19-22 indicate severe depression, and scores of 23 or above indicate very severe depression.
In addition, many antidepressant studies include a “placebo washout period” in which all patients receive placebo pills for several days to 2 weeks before randomization. Often, patients who improve by 20% or more in the HDRS are excluded from the trial. Removing known placebo responders at the outset is thought to enhance the statistical power of the antidepressant-placebo comparison.
“This design feature severely limits the ability to generate accurate estimates of the placebo response rate,” wrote Mr. Fournier and his colleagues, noting that “the true rate of placebo response may be underestimated in trials that use this feature.”
To determine the true placebo response rate across the entire range of depression severity, the investigators combed the literature for randomized, controlled studies of minor depressive disorder that did not include a placebo washout period.
Of the 2,146 randomized, controlled trials of depressants published in English from January 1980 to March 2009, only 23 studies met those criteria. But only six of those studies could provide patient-level data to the investigators. It was those six studies, comprising 434 patients in antidepressant groups and 284 patients in placebo groups, that were the subject of the meta-analysis.
Actually, most meta-analyses include only group-level data. A study such as this one that includes patient-level data is called a “mega-analysis.” This approach allows investigators to conduct a more fine-grained multivariate analysis.
Three of the six studies used imipramine, a tricyclic antidepressant, and the other three used paroxetine, a selective serotonin reuptake inhibitor. The mean baseline HDRS score in the studies ranged from 14 to 24.
As expected, the higher the patient's baseline HDRS score, the more improvement was seen with both the active drug and the placebo. A difference of 3 points or more on the HDRS is considered clinically significant. It was only at HDRS baseline levels of 25 and above that active drug was both statistically and clinically better than placebo. This finding was the same for imipramine and paroxetine.
The investigators wrote that in marketing antidepressants, pharmaceutical manufacturers rarely mention that most efficacy studies specifically exclude patients who derive little benefit from their medications.
The authors concluded that “efforts should be made to clarify to clinicians and prospective patients that 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 depression.”
Commenting on the meta-analysis, Dr. Eric G. Tangalos emphasized that conclusions based on studies published as early as 1980 might not be relevant to current medical practice.
“Papers published in the 1980s took their data from clinical practice in the 1970s, and papers in the 1990s took their information from the 1980s. SSRIs, which are the current mainstay of therapy, did not emerge until the 1990s. Prior to the advent of SSRIs, physicians were reluctant to even identify depression because the available treatments (MAO inhibitors and, later, tricyclics) carried so many serious side effects,” said Dr. Tangalos, a professor of medicine at the Mayo Clinic in Rochester, Minn. He reported no relevant conflicts of interest.
Disclosures: The National Institute of Mental Health funded the study. Mr. Fournier stated that he had no relevant financial conflicts of interest. Several of the other investigators did report financial relationships with several pharmaceutical companies.
My Take
Findings Need Careful Assessment
This is an important study for internists, but I am cautious about its message. In view of the diagnostic variability of mild depression, we need to examine the findings carefully before we implement them wholesale.
I have treated many patients with mild chronic depression or dysthymic conditions who have had excellent (family member validated) short- and long-term results with antidepressant medications—so many patients, in fact, that the responses do not seem to reflect a placebo effect.
I would hope that formulary committees would not jump to blunt conclusions about the value of antidepressants without nuanced review of the study and substantiation of its findings.