PharmExec Blog

Attacks on Pharma: Is the Quality Getting Better?

We read a lot of critiques of the pharma industry, and frankly many of them leave a lot to be desired in terms of logic, clarity, and thoughtfulness.

For an example of the opposite—a truly clear and scorching attack, let us recommend to you a document prepared by Dr. Joseph Glenmullen, a Clinical Instructor in Psychiatry at Harvard Medical School for the lawsuit O’Neal v. SmithKline Beecham d/b/a GlaxoSmithKline. Glenmullen’s report, which was recently unsealed and first brought to light in an article in New Scientist, mostly asks a couple of straightforward questions: When should GlaxoSmithKline have known that its Paxil created an elevated risk of suicide? and Why did it take so long for that information to get to patients and doctors?

Glenmullen’s conclusion: GSK’s data proved an elevated risk right from the start, a fact that was concealed by inappropriate data reporting.

Wash-out suicides

What went wrong? The problem has to do with the “wash-out” or “run-in” period. A week or so at the start of the clinical trial when both active and placebo arms receive placebo, which allows the effects of any previous treatments to clear. Several suicides took place during the Paxil wash-out. Because the controlled portion of the experiment had not yet begun, they weren’t supposed to be counted against either arm. But GSK counted them as suicides on placebo. That error—or deliberate deception—distorted results for years to come.
For example , in a 1991 safety report submitted to FDA, GSK stated that there had been five suicides among 2,963 patients receiving Paxil (or 0.17 percent) and two among the 554 patients receiving placebo (0.34 percent). The numbers looked great: placebo had double the suicide rate. The only problem was that in fact there were no suicides on placebo.


As Glenmullen tells it, the misattributed suicides were the main defect in GSK’s safety reporting, but he identifies others:

  • Suicidal behavior was coded as “emotional lability,” disguising the seriousness of the problem
  • The protocol called for no probing questions on suicidal feelings or behovior, just open-ended questions on how the patient was doing
  • When GSK finaly corrected its reporting of the wash-out suicides in 2002, it began lobbying FDA to include in its safety results two studies of patients at high risk of suicide. These studies accounted for only 5.5 percent of patients but 68.8 percent of suicidal behavior. Not surprisingly, when these patients were included, the evidence of Paxil’s effect on suicidal behavior mostly disappeared.

The Glenmullen report didn’t win O’Neal for the plaintiffs. (Read about it here.) But it’s triggering action elsewhere. Senator Charles Grassley of the Senate Finance committee has writtn to GSK demanding to know what the company knew and when. (You can obtain a PDF of Grassley’s letter and a redacted version of the Glenmulen report from the committee’s Web site.) But I suspect that the real impact will come as subsequent litigators and investigators learn from the report how to make an easily understood, high-impact case out of reams of numbers. Bravo to Dr. G. Everybody else, make sure this doesn’t happen to you.

Meanwhile, our friends at the Drug & Device Law blog make an interesting case about why suicide numbers for many dugs may be inflated by the inherent nature of most clinical trials. Read it here.

This entry was posted in Legal, Safety. Bookmark the permalink. Post a comment or leave a trackback: Trackback URL.

One Comment

  1. Carl Kinard MD
    Posted February 14, 2008 at 9:11 pm | Permalink

    What would the logic of excluding the 2 high risk studies with most of the suicidal behavior be?
    It is certainly not valid to include the washout suicides in the placebo data, but 5 suicides in 2963 Paxil patients and 0 in 554 placebo patients is not a significant difference. After all, the rate of suicide on placebo is certainly not zero.

Post a Comment

Your email is never published nor shared. Required fields are marked *


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

  • Categories

  • Meta