Paroxetine and adult patients


GlaxoSmithKline has been examining the issue of suicidality in adults with depression and other psychiatric disorders since before paroxetine was approved. Over the past several years, GSK has conducted analyses of adult suicidality: (1) the European Article 31 Referral (a European regulatory process involving a series of requests for analyses of clinical trial data), conducted during 2003 and 2004; (2) a 2004 report of the UK's General Practice Research Database (GPRD), an epidemiological database; (3) an analysis completed in 2006 that is similar to FDA's review of antidepressants and suicidality; (4) a 2002 analysis of results from a review of data regarding “suicide attempts” in paroxetine clinical trials originally submitted to FDA in 1991; and (5) a 2002 analysis of data submitted to FDA in 2001 regarding suicide attempts and suicidal thinking; the 2002 summary uses narrow and broad definition algorithms. Additionally, an excerpt from GlaxoSmithKline’s response on 15 February 2008 to a letter from Senator Grassley seeking information on Paxil and adult patients is provided.  The results of the 2006 analysis indicate that young adults, particularly those with Major Depressive Disorder, may be at increased risk for suicidal behavior with paroxetine. While the individual studies comprising these analyses have already been made available on the GSK Clinical Study Register, GSK is taking the additional step of posting to this web page the following documents.

2006 Analysis

2006 analysis

The US Food and Drug Administration (FDA) has requested data from antidepressant manufacturers for an analysis of adult suicidality data from short-term placebo-controlled trials. GSK elected to conduct an analysis of its own data from short- and long-term placebo-controlled paroxetine trials in adults, building upon what was done in previous analyses and the developing expertise in the field.

Please find below the descriptions of and links to documents related to the 2006 analysis.

Summary/Briefing Document: A summary of the results from the 2006 analysis has been submitted to regulatory agencies in the United States, Europe, and other countries around the world. This summary (also known as a Briefing Document - PDF 41.6KB) provides both safety data (related to suicidal behaviour and ideation) and efficacy data (how well symptoms of psychiatric disorders respond to paroxetine), as well as GSK's conclusions from the analysis.

Appendices to the Briefing Document: These Appendices provide the datasets and clinical narratives that are referenced in the Briefing Document. Note that Appendix I contains the data analysis plan. Download PDFs:

Information for healthcare providers

European healthcare providers can get more information by consulting the paroxetine SPC (Summary of Product Characteristics), sections 4.4 and 5.1.

Article 31 analysis

Article 31 analysis

In June 2003, a review of paroxetine clinical trial data was initiated by European regulators in a process known as an "Article 31 referral". As part of this process, GSK was asked to provide specific analyses of its clinical trial data to evaluate the risk of suicide, suicidal thoughts and behaviours, and self-harm, with particular attention to potential risk factors including age and gender. GSK submitted the first set of analyses to the initial Article 31 questions in September 2003, the second set in January 2004 and a third set in March 2004.

At that time, the European Committee for Human Medicinal Products (CHMP) performed a full risk-benefit review of paroxetine which resulted in a CHMP Opinion dated April 2004. In December 2004, the CHMP reaffirmed these conclusions following consideration of three new epidemiology studies which utilized the UK General Practice Research Database.

The documents provided below contain GSK's responses to questions that had a primary focus on suicidality prepared during the Article 31 review - both in adults and paediatrics. The responses are provided as they were submitted to European regulators and are based on the analyses conducted at that time.

Article 31 questions*

September 2003 Questions & Appendices

January 2004 Questions

January 2004 Appendices

March 2004 Question

* Some appendices did not generate additional data

** Appendix 1 relates to Questions 1, 4, 8, 9, 11, 13, 15, 27

GPRD report

General Practice Research Database (GPRD)

In response to the European regulatory review of paroxetine (the "Article 31 referral" process), GSK conducted an analysis of suicidality in patients receiving paroxetine and other antidepressants, based on data from the UK GPRD. The GPRD is a computerised database of anonymous medical records from primary care physicians in the UK that is publicly available for research purposes and is managed by the Medicines & Healthcare products Regulatory Agency (MHRA) in London.

Response to US Senator Grassley

Response to US Senator Grassley

The following is an excerpt from GlaxoSmithKline’s response on 15 February 2008 to a letter from Senator Grassley seeking information on Paxil and adult patients:
GlaxoSmithKline has always acted with the best interests of our patients in mind. We have strived for transparency when it comes to safety issues, through timely and complete submissions to FDA, and by posting important and helpful information for physicians and patients on our website.

  • The issue raised by Dr. Glenmullen about run-in events is old news. It was debated in the public domain as long ago as 2002. Further, the approach GSK used in 1989 and 1991 in counting placebo run-ins was also used by FDA in its Safety Review of Paxil in 1991.
  • Dr. Glenmullen’s report was created to support plaintiffs’ allegations in litigation against GSK. His methodology is scientifically flawed and unreliable; no one but plaintiffs’ experts in product liability litigation use it. A federal court rejected similar opinions when presented by another plaintiffs’ “expert."
  • In fact, that federal court excluded those opinions because they were not based on “a methodology for determining general causation which has been accepted in the scientific community.”1
  • GSK has made Paxil risk information, including information about clinical trials, readily accessible to the public and the scientific community through a clinical trial registry, the GSK website, letters to healthcare professionals and publications in peer-reviewed literature. On GSK’s corporate homepage alone, there is a wealth of information about Paxil, much of which lays out in great detail the various analyses of suicidal thinking and behavior the company has conducted over the years.
  • This information demonstrates that GSK detected no signal of any possible association between Paxil and suicidality in adult patients until late February 2006, and even then the finding related primarily to young adults, and to suicide attempts and not completed suicides. The numbers were small and the study was retrospective. GSK promptly contacted FDA, submitted a briefing document to the agency on 8 March 2006, discussed appropriate content for a labeling change, and communicated the findings to healthcare professionals on 5 May 2006.
  • In 2007, FDA directed that labeling be revised for the entire class of selective serotonin reuptake inhibitor antidepressants (or “SSRIs”, of which Paxil is one). FDA required that the label state that the number of suicides in the adult trials was not sufficient to reach any conclusion about any effect on suicide. FDA also directed that the label reflect that analyses had detected no increased risk of suicidality (attempts or ideation) in adults above 24 and that there appeared to be a protective effect in adults over 64.
  • FDA has systematically monitored the issue of a possible association of antidepressants with suicide and suicide attempts for almost two decades, periodically requesting and receiving from SSRI manufacturers updated, comprehensive information on these events. As early as 1991, FDA concluded that there was no association between SSRIs and suicide/suicide attempts.
  • GSK’s 1989 and 1991 analyses of the clinical trial data for Paxil, properly analyzed, revealed no increase in the risk of suicides or suicide attempts for adult patients on Paxil.

1. See Vanderwerf v. SmithKline Beecham Corp., Civil Action No. 05-2271-KHV, slip op. (D.Kan. 9 January 2008).