Antidepressant Study Revives Efforts to Establish Public Clinical Trial Registry

The American Psychiatric Association (APA) and the American Academy of Child and Adolescent Psychiatry (AACAP) renewed their 2004 call for a mandatory public clinical trial registry to be established and overseen by the federal government. This is in response to a study that showed that research on antidepressants is published selectively-effectively suppressing trials deemed negative and overstating the benefits of antidepressant therapy. The study was published in the January 17 issue of the New England Journal of Medicine.

A research team led by Erick H. Turner, MD, Oregon Health and Science University and Portland Veterans Affairs Medical Center, analyzed all phase 2 and phase 3 clinical trials of 12 antidepressant agents approved by the FDA from 1987 through 2004. “According to the published literature, the results of nearly all of the trials of antidepressants were positive. In contrast, FDA analysis of the trial data showed that roughly half of the trials had positive results,” the researchers wrote.

The team found that 23 of the 74 studies (31%) were never published. Sample size was not a factor in publication. However, what was a factor, the investigators reported, was outcome. Positive studies were 11.7 times more likely to be published than studies with negative or questionable findings (P = .001). Of the 38 studies (51%) viewed by the FDA as having positive results, 37 were published. The remaining 36 studies (49%) deemed by the FDA as having negative results (24 studies) or questionable results (12 studies), with 3 exceptions, either were not published (22 studies) or were published in a way that conveyed a positive outcome (11 studies).

The publication bias resulted in 94% (95% confidence interval [CI], 84% to 99%) of published studies appearing to have positive results, whereas the FDA analysis showed that only 51% (95% CI, 39% to 63%) of trials conducted had positive results, the researchers found. The bias also had an impact on effect sizes, which ranged from 11% to 69% for individual antidepressants (P = .001) and was 32% overall (P = .012).

“By altering the apparent risk-benefit ratio of drugs, selective publication can lead doctors to make inappropriate prescribing decisions that may not be in the best interest of their patients and, thus, the public health,” the researchers noted.

“Our patients deserve the best health care available, and having full disclosure of research findings-both positive and negative-will help clinicians develop the most effective treatment plans. Open access to all clinical trial data is necessary to better understand the risks and benefits of treatments,” said Carolyn Robinowitz, MD, APA president. “A national registry will allow patients to have access to data on a complete range of treatment options, including medication, to discuss with their physician,” added Robert L. Hendren, MD, AACAP president.

Pharmaceutical companies indicated that they are moving to a policy of greater disclosure of clinical trial results regardless of outcome. Mary Anne Rhyne, director of US media relations at GlaxoSmithKline (GSK), told Drug Benefit Trends, “GSK agrees that public disclosure of clinical trial results for marketed medicines is essential and fully supports registration of all trials in progress. We have taken the following steps to demonstrate that commitment: GSK posts on the Internet results of all company-sponsored trials of all GSK-marketed medicines which we have completed since the formation of our company in December 2000-every trial, every phase, in every country.” She continued: “In many instances, when data from still earlier trials may inform medical judgment, those data have been posted as well. The data are presented in a consistent format without regard to whether they may be viewed as positive or negative.”

Lilly issued the following statement on January 18: “We are committed to publicly disclosing medical research results-whether favorable or unfavorable to a Lilly medicine-in an accurate, objective, and balanced manner in order for our customers to make more informed decisions about our products.”

“Pfizer is committed to the communication of results of all registered clinical studies, regardless of outcome,” Pfizer spokesman Jack Cox explained. “More specifically, we have committed to disclose clinical trial results within 1 year after study completion for all of our marketed products.”

Refined Criteria Predict Psychosis in High-Risk Youth

Youth at risk for psychosis can be identified before their illness becomes fully developed 68% to 80% of the time when the standard criteria are refined by combining 2 or 3 variables, according to findings of a study published in the January issue of Archives of General Psychiatry. These figures are significantly higher than the predictive accuracy of a single prodromal criterion (35%), reported lead researchers Tyrone D. Cannon, PhD, University of California at Los Angeles, and Robert Heinssen, PhD, National Institute of Mental Health (NIMH).

The researchers evaluated 291 adolescents (median age, 16 years; 42% female) at high risk for psychotic disorders, which is approximately 3 times the number of participants in any previous study of this kind. A separate group of 134 persons with no known risk factors for psychosis served as a control group.

The team found that 5 risk factors assessed at baseline contributed to the prediction of psychosis: a genetic risk for schizophrenia with recent deterioration in functioning, higher levels of unusual thought content, higher levels of suspicion and paranoia, greater social impairment, and a history of substance abuse. Of the 26 possible combinations of criteria, the two 2-factor combinations with the highest positive predictive power (PPP) were genetic risk for schizophrenia and unusual thought content or impaired social functioning (69% and 61%, respectively). Two 3-factor combinations (genetic risk, unusual thought content, and either suspicion and paranoia or impaired social functioning) led to a significant rise in PPP (74% and 81%, respectively).

Link Between BD and Substance Abuse Confirmed

Persons with manic symptoms and bipolar disorder (BD) type II are at significant risk for developing a substance abuse or dependence problem, reported researchers who had conducted one of the few prospective long-term studies demonstrating a clear association between mood disorders and substance abuse. The study was published in the January issue of Archives of General Psychiatry.

Researchers led by Kathleen R. Merikangas, PhD, NIMH, and colleagues followed 292 men and 299 women from 1979 to 1999. The team conducted 6 waves of direct diagnostic interviews to assess mood and substance use disorders.

By 1993, 9.7% of participants met Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) criteria for major depression. Although no participant met DSM-IV criteria for BD type I, 4.4% of participants met criteria for BD type II. In addition, 23.5% of participants had manic symptoms but did not meet specific criteria for BD. In 1999, 17.9% of subjects met DSM-IV criteria for alcohol abuse/dependence, 8% met criteria for cannabis abuse/dependence, and 3.4% met criteria for benzopdiazepine abuse/dependence.

The team found that participants who showed symptoms of mania but did not meet DSM-IV criteria for BD were twice as likely to have an alcohol dependence problem (odds ratio [OR], 4.44; 95% CI, 1.55 to 12.73) than an alcohol abuse problem (OR, 2.41; 95% CI, 1.22 to 4.76). The development of cannabis abuse/dependence (OR, 4.82; 95% CI, 2.0 to 11.60) and of benzodiazepine abuse/dependence (OR, 11.54; 95% CI, 4.80 to 27.78) was also significantly predicted by manic symptoms.

Participants with BD type II were at even greater risk for developing an alcohol dependence problem (OR, 21.05; 95% CI, 6.57 to 67.47) or alcohol abuse problem (OR, 9.11; 95% CI, 2.66 to 31.21) than those with manic symptoms alone. Subjects with BD type II were also at higher risk for developing a benzodiazepine abuse/dependence problem (OR, 14.10; 95% CI, 2.65 to 75.09).