New analysis discredits Paxil study

A 2001 clinical trial that concluded Paxil was “well tolerated and effective” for treating major depression in adolescents helped clear the way for the drug to become the best-selling antidepressant in the world the following year.

But that clinical trial led by Brown University Professor Emeritus of Psychiatry and Human Behavior Martin Keller, M.D. got it all wrong, according to a new re-analysis published Sept. 16 in the BMJ (formerly the British Medical Journal).

The new paper found that paroxetine – the generic name for Paxil which is classified as selective serotonin reuptake inhibitor – is neither safe nor effective in adolescents with depression.

An independent team of seven authors documented clinically significant increases in harms, including suicidal ideation and behavior in subjects taking paroxetine. They concluded that paroxetine was no more effective than a placebo in the treatment of adolescents with unipolar major depression.

“The fact that we completed this paper is testament to something of ours, probably our bloody-mindedness,” said Jon Jureidini, Ph.D., MB BS, FRANZCP, one of the authors and a child psychiatrist at the University of Adelaide in Adelaide, South Australia.

“It took a year to write and everybody doing it in their unpaid time,” Jureidini explained in a Skype interview.

“Then it took a year to get it through the BMJ review process. The first round of reviewers, instead of the usual two or three reviewers, I had six and that got us 27 pages of comment on the first draft. Our paper has been through a level of peer review in order that I doubt any randomized control trial that’s published for the first time has ever been through.”

The team reviewed the original study, known as Study 329, reported in the Journal of the American Academy of Child and Adolescent Psychiatry.

Funded by SmithKline Beecham, which subsequently became GlaxoSmith-Kline (GSK), Study 329 enrolled a total of 275 subjects aged 12-18 between April 1994 and March 1997 at 10 U.S. study sites and two more in Canada affiliated with a university or hospital psychiatry department.

Participants had to meet the DSM-IV criteria for a current episode of major depression of at least eight weeks’ duration to participate in an eight-week double blind randomized control trial followed by a six-month continuation phase.

Subjects were randomly assigned paroxetine, imipramine (a tricyclic antidepressant sold as Tofranil) or placebo in a 1:1:1 ratio and instructed to take the study drug twice daily in flexible doses adjusted on the basis of clinical response and tolerability over the eight weeks.

A total of 190 subjects completed the original study with withdrawal rates of 40 percent in the imipramine group, 28 percent among the paroxetine, and 24 percent in the placebo.

Reported adverse effects such as headache, nausea, dizziness, dry mouth and drowsiness were frequently cited in all groups. Tachycardia was also cited as a common reading for imipramine group withdrawal.

Serious adverse effects were reported in 11 paroxetine patients, including five who experienced suicidal ideation and two with worsening depression.

The re-analysis team analyzed 34 percent of case report forms from the original study and found evidence of significant under-recording of adverse events and identified idiosyncrasies in the coding system used.

Keller has an active email account at Brown University but did not respond to emailed requests for interviews. He retired from Brown “a few years ago,” according to an email from the university’s Acting Director of News and Communications Mark Nickel. But Nickel would not comment on specifically when Keller left the university or the circumstances.

GSK issued a statement on the reanalysis of Study 329, noting that it provided the research team access to detailed data from the original trial: “This reflects our commitment to data transparency – we publish the results of all our studies regardless of whether they are positive or negative.”

“We have also led the way in giving external researchers access to the very detailed patient-level data behind our studies, granting access to more than 50 research teams around the world so they can independently use our data in their research,” the statement continued.

Jureidini said that the database of clinical record forms GSK released to his research team had restrictions that made their work extremely time-consuming.

“It makes it like they went out of their way to help us. They made things as difficult as they could for us,” Jureidini said. “We could only look at one page at a time. We couldn’t print it so imagine trying to look at 70,000 pages of documents.”

GSK’s statement noted: “Importantly, the findings from this team’s analysis appear to be in line with the longstanding view that there is an increased risk of suicidality in pediatric and adolescent patients given antidepressants like paroxetine. This is widely known and clear warnings have been in place on the product label for more than a decade. As such, we don’t believe this reanalysis affects patient safety.”

Jureidini is spokesman for Healthy Skepticism, an international non-profit dedicated to countering misleading drug promotion. He believes poor clinical reasoning often leads to the prescription of antidepressants.

“Doing things that are bad for people is worse than not doing anything, but practice wouldn’t seem to support that,” Jureidini explained.

“The argument is something like antidepressants don’t really work, try something else, but if that doesn’t work, use antidepressants. Or if things are really severe, use antidepressants, or if patients aren’t interested in having psychotherapy, give them antidepressants. So the decision on whether antidepressants are useful drugs relies on making a sound clinical judgment about whether the benefits outweigh the harms.”

In 2004, the U.S. FDA warned doctors not to prescribe the drug to children after British regulators issued a similar warning.

The actions came in response to new data showing that among 1,100 children enrolled in clinical trials of Paxil, those taking the drug were nearly three times as likely to consider or attempt suicide as children taking placebos.

Strong warnings were added to the labels of Paxil and other antidepressants, citing the potential suicide risk for children and young adults.