A clinical trial of an anticonvulsant in the treatment of bipolar disorder
in children may be more notable for its publication than its outcomes.
The multicenter trial found that oxcarbazepine was "not significantly
superior to placebo" in a seven-week test of youngsters between the ages
of 7 and 18.
"Five or 10 years ago papers like this were routinely rejected
because they showed `negative results,' " Scott Zeger, Ph.D., professor
and chair of biostatistics at the Johns Hopkins Bloomberg School of Public
Health, told Psychiatric News. "This negative study is
important both because it teaches us about this particular trial, but also
because it reflects a trend in favor of publication of negative studies,
allowing us to overcome biases that have been rampant in the medical
The study sought to fill a largely empty space in the research on treatment
of bipolar disorder in young people, added Ellen Leibenluft, M.D., of the
National Institute of Mental Health, in an editorial accompanying the trial's
publication in the July American Journal of Psychiatry. The only
previously published placebo-controlled trial of the drug in adults with
bipolar disorder included only six patients.
"Parents and clinicians can only be appalled by the contrast between
the ample evidence documenting the severity of pediatric bipolar disorder and
the paucity of controlled trials to guide its treatment," wrote
To fill this gap, karen Dineen Wagner, M.D., Ph.D., professor and vice
chair in the Department of Psychiatry and Behavioral Sciences and director of
the Division of Child and Adolescent Psychiatry at the University of Texas
Medical Branch in Galveston, and colleagues recruited 116 outpatients who had
been diagnosed with bipolar I disorder, manic or mixed episode, using
DSM-IV criteria. The subjects also had a score of 20 or above on the
Young Mania Rating Scale (YMRS) at the baseline visit. Half the patients were
randomly assigned to get oxcarbazepine and half to get a placebo in the
Fifty-five patients in each group were available for efficacy analysis. A
large percentage of each group dropped out of the trial before its
completion—20 patients (34 percent) in the oxcarbazepine group and 24
(42 percent) in the placebo arm. Patients received an average dose of 1,515
mg/day of oxcarbazepine for a median 48 days.
The study was supported by funding from Novartis Pharmaceuticals.
Mean YMRS scores dropped among both groups of patients, and no
statistically significant differences were found at any point during the
trial, the authors noted. The adjusted average change was -10.90 for the
oxcarbazepine group and -9.79 for the placebo group.
Most of the patients had at least one adverse event—90 percent of
those taking oxcarbazepine and 79 percent of placebo patients. The
oxcarbazepine group was more likely to experience dizziness, nausea,
sleepiness, double vision, fatigue, or rash. The effects were mild to
moderate, of short duration, and likely to occur in the early part of the
About 42 percent of the patients in the oxcarbazepine arm and 26 percent in
the placebo arm achieved the secondary efficacy goal of at least a 50 percent
reduction in the YMRS score. There was no statistically significant difference
between these figures, the researchers noted.
Leibenluft wrote in the accompanying editorial, "Without the placebo
arm, in which 26 percent of patients responded, the authors might have
concluded that a response rate of 42 percent demonstrated that oxcarbazepine
is effective in treating pediatric bipolar disorder, when in fact there was no
significant difference in outcome between the two treatment groups."
Even nonsignificant trends toward an effect can guide clinicians, said
"Statistical significance is not the same as clinical
significance," he said. "Knowing how to quantify evidence in a
trial is not the same as making a treatment decision."
He continued, "The most important thing is that people are aware of
all trials, whether they have a positive or negative outcome. This is a shift
in position. Once, studies showing no difference in effect would have been
called `failed trials.' Often they were self-censored by authors, who figured
they weren't worth the effort to publish, or by journal editors who thought
they weren't interesting to readers. As a result, readers would tend to think
that treatments they did read about were better than they really
"Although the results of the paper do not quite reach
significance," said Robert Freedman, M.D., the editor in chief of the
American Journal of Psychiatry, "there are no comparable
studies that attempt to provide empirical evidence to clinicians about the
safety and efficacy of antiepileptic drugs for the treatment of bipolar
disorder in adolescence and childhood. Therefore, the journal decided to
publish these data with an accompanying editorial explaining their
One way to track all clinical trials, regardless of their outcomes, is to
register every trial when it begins in a recognized database, such as the
National Institutes of Health's
The World Health Organization now recommends registration of all medical
studies that test treatments on human beings. Editors of major medical
journals (including the American Journal of Psychiatry) have declared
that they will not publish trials that have not been registered and will
adhere to standards set by the International Committee of Medical Journal
Editors in 2004.
"A Double-Blind, Randomized, Placebo-Controlled Trial of
Oxcarbazepine in the Treatment of Bipolar Disorder in Children and
Adolescents" is posted at<http://ajp.psychiatryonline.org/cgi/content/full/163/7/1179>.▪