May 20, 2009 — New evidence suggests that clinical trials of antidepressants do not represent the majority of patients seeking treatment and that these studies may overestimate the efficacy of such drugs in clinical practice.
The latest findings, from the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study, show that less than a quarter of patients with depression would qualify for inclusion in a phase 3 clinical trial. The results also reveal that when treated with the selective serotonin-reuptake inhibitor (SSRI) citalopram (Celexa, Forest Pharmaceuticals), patients who would be ineligible for research studies had lower remission rates and higher rates of adverse effects than patients who would qualify for inclusion in these trials.
"This means that the population that was used for testing the efficacy of a drug now on the market is probably very different from the population doctors see on a regular basis in their practice. So don't expect the same results," said lead author Stephen R. Wisniewski, PhD, from the University of Pittsburgh, in Pennsylvania.
However, the study's findings do not necessarily mean changes in the design of clinical trials, because most "real-world" patients are much more likely to experience severe adverse events, said Dr. Wisniewski.
The study is published in the May issue of the American Journal of Psychiatry.
Real World Patients
The aim of the study is to prospectively define which of several treatments are most effective for outpatients with nonpsychotic major depressive disorder who have an unsatisfactory clinical outcome to an initial and, if necessary, subsequent treatment.
The study included 4041 subjects at more than 40 sites across the United States, including 18 primary-care and 23 psychiatric clinics. "These were not research-based clinics but clinics out in the real world, seeing real people," said Dr. Wisniewski.
Participants in this current analysis had to be 18 to 75 years old, meet criteria for major depressive disorder, and score 14 or higher on the 17-item Hamilton Depression Rating Scale (HAM-D).
This study compared participants who would meet typical inclusion criteria for a phase 3 trial (efficacy sample) with those who would not (nonefficacy sample).
The efficacy group, for example, had to have a baseline HAM-D score higher than 19, have no more than 1 concurrent medical condition, have no more than 1 additional concurrent psychiatric disorder, and have a current depression episode lasting less than 24 months. Those not meeting these criteria were included in the nonefficacy sample.
Of the 2855 patients included in the current analysis, only 635 (22.2%) could be classified into the efficacy sample, and 2220 (77.8%) were classified into the nonefficacy sample.
There were numerous baseline differences between the 2 groups. Among others, participants in the efficacy sample were more likely to be younger, more educated, white, employed, married, and privately insured and to have a higher income. They also had a shorter average duration of illness.
Patients completed the self-rated 16-item Quick Inventory of Depressive Symptomatology test during clinic visits at weeks 2, 4, 6, 9, and 12 (there was also an optional week 14 if needed). They received citalopram once a day, starting at 20 mg/day, which was raised to 40 mg/day by week 4 and to 60 mg/day by week 6. Flexibility in dosing was allowed.
Hybrid Design
The study was something of a hybrid of efficacy and effectiveness trials. For example, said Dr. Wisniewski, researchers combined the rigorous treatment-administration protocols characteristic of an efficacy trial with the broadness of an effectiveness trial.
This means, he said, the study ensured that patients received treatment in the appropriate dose for a sufficient length of time but allowed for inclusion of subjects who would not have met inclusion criteria of a phase 3 efficacy trial but who are the type of people who are actually taking the drug in the real world.
At the end of the study, the remission rate, which was defined as a score of 5 or less on the Quick Inventory of Depressive Symptomatology or equivalent to a score of 7 or less on the HAM-D, was 34.4% in the efficacy sample and 24.7% in the nonefficacy sample.
The response rate, defined as a reduction of at least 50% from baseline on the Quick Inventory of Depressive Symptomatology, was 51.6% in the efficacy group and 39.1% in the nonefficacy group.
Even after adjustment for potential baseline confounding characteristics, the efficacy sample had significantly better depression-symptom outcomes and shorter time to remission and response.
The efficacy sample was also less likely to have serious and frequent adverse effects or to have at least 1 serious adverse event or psychiatric serious adverse event, compared with the nonefficacy group.
Unlike Patients Seeking Treatment
These results demonstrate that patients in phase 3 trials are not, in general, like most people who are seeking treatment, said Dr. Wisniewski. "As well, the outcomes of those in the nonefficacy group, the people who would have been excluded from the phase 3 trial, were generally worse, in that they had lower remission rates and higher rates of side effects."
Given these results, should the medical establishment change the design of phase 3 clinical trials? "I actually don’t think so, because it might put some patients at risk," said Dr. Wisniewski. "If you could just increase the generalizability at no cost, I would say, yes, do it. But there is a cost, and that cost is that those who are excluded also are at increased risk for serious adverse events."
Dr. Wisniewski's next research steps include attempting to determine which patients respond to which treatment approaches. "We’re designing a study to look at clinical and genetic moderators of treatment in an effort to determine which treatments do or do not work in certain populations," he said.
Asked to comment on the study, American Psychiatric Association spokesperson Jeffrey Lieberman, MD, who is professor and chair of psychiatry at Columbia University, in New York City, New York, said it is "very significant."
"This paper is a very creative and convincing analysis of why treatments don’t work in clinical practice in the same way they’re reported to work when they’re approved by regulatory agencies."
According to Dr. Lieberman, future clinical trials might include a broader range or "cross section" of patients to get a better understanding of the "real degree of treatment efficacy."
Dr. Wisniewski has been a consultant for Bristol-Myers Squibb, Case Western University, Cyberonics, ImaRX Therapeutics, and Organon over the past 4 years. Disclosures for the other researchers are listed in the paper.
Am J Psychiatry. 2009;166:599-607. Abstract
Medscape Medical News © 2009 Medscape, LLC
Send press releases and comments to news@medscape.net.
Cite this: STAR*D: Antidepressant Trial Results May Overestimate Efficacy - Medscape - May 20, 2009.
Comments