Social Audit Ltd
P.O. Box 111 London NW1 8XE
Telephone/Fax: 020 7586 7771

[email protected]



 Background to this FOI request
This Freedom of Information (FOI) request seeks to establish what steps have been taken by the UK drug regulatory authorities to establish the circumstances in which four patients committed suicide, during clinical trials comparing the antidepressant, paroxetine (Seroxat®), with placebo.

These patients were involved in the (paroxetine versus placebo) trials scrutinised by the Expert Working Group (EWG) of the Committee on Safety of Medicines (CSM), whose final report on the safety of selective serotonin reuptake inhibitors (SSRIs) was published in December 2004. A more detailed review of the same trials was later reported in the British Medical Journal by two members of the EWG (Gunnell et al, 2005) and this FOI request relates to their findings too.

This FOI request concerns matters of real public interest. The immediate point at issue is whether paroxetine, and other SSRI antidepressant drugs, can sometimes precipitate suicidal thoughts and acts in adults. The wider point is whether and to what extent the public may and should rely on regulatory assessments of drug benefit and risk.

The case for disclosure relates mainly to public health and safety, and to public confidence in the authorities responsible for assessing and giving advice about the risks and benefits of medicines. Nor is there any apparent reason why the requested information should not be disclosed: the question of commercial confidentiality should not arise because the authorities have already published their findings and conclusions. They reported that, in these clinical trials, three of the four people who committed suicide were on placebo, with only one on the active drug. The inference was that the risk of suicide for untreated patients (on placebo) was significantly lower than the risk for patients on paroxetine.

The purpose of this request is in effect to verify that conclusion (or not). The question is whether the authorities reliably established that the three ‘placebo suicides’ did in fact involve patients who were taking only dummy pills; did occur within the 30-day post-treatment period of the relevant clinical trials; and whether the circumstances in each case were properly investigated by the sponsoring company and the regulators and were/are sufficiently well understood.

These questions are central and critical to the long-running controversy about the propensity of paroxetine and other antidepressants to produce paradoxical reactions in a small minority of patients. Precisely because suicides in antidepressant drug trials are relatively rare, only one or two such events in a series of clinical trials might have a profound outcome on the overall study conclusions. This underlines the need to be certain that the three reported ‘placebo suicides’ were properly investigated and reliably reported. If they were not, it would signal the need for a fundamental reassessment of the risk of antidepressant drug-induced suicidality in adults – and there are several reasons for doubting the reported results:

The reported results (one paroxetine suicide and three placebo suicides) seem anomalous [a] in relation to the results reported for similar drugs; and [b] in relation to evidence of the propensity of paroxetine to cause (non fatal) suicidal behaviour.

The fact that three patients committed suicide, reportedly on placebo, raises worrying questions about the quality of the medical supervision in these trials. Patients at risk for suicide, actively monitored in the post-treatment phase of a clinical trial, would be obvious candidates for active drug treatment. It therefore seems all the more important to confirm that the three ‘placebo suicides’ were indeed receiving no drug treatment (e.g. had not been started on any other antidepressant) and that the suicides occurred in the 30-day post treatment phase of the trial, i.e. during the period in which patients would have been under close medical supervision.

The evidence suggests that the CSM Expert Working Group did not examine the relevant data about the four paroxetine/placebo suicides: "We did not have access to individual patients’ data" (Gunnell et al, 2005). The EWG Secretariat - staff of the Medicines & Healthcare products Regulatory Agency (MHRA) – appear to have relied on the summary data provided by the manufacturer (license holder/Market Authorisation Holder) of Seroxat®.

The EWG report gives no indication of what the three ‘placebo suicide’ patients were being treated for – but this must be relevant because the paroxetine/placebo adult clinical trials did not only involve patients treated for depression. Of the 5808 patients on placebo, 2117 (36.45%) were suffering from depression, but most (63.55%) were involved in trials for other disorders. It may well be that all three ‘placebo suicides’ were involved in depression trials, but it would not be safe to take this for granted.

FOI request
The following questions arise from these concerns. They are addressed principally to the Medicines and Healthcare products Regulatory Agency, whose staff acted as Secretariat for the Committee on Safety of Medicines Expert Working Group.

Did the MHRA request from the GlaxoSmithKline, the License Holder for paroxetine (Seroxat®) access to the detailed case reports of the four suicides (‘the data’) reported in all paroxetine and placebo trials? What was the date at which those data were first requested by the MHRA?

If so, did the MHRA obtain all the data requested? Was the MHRA refused any of the data requested? Was the MHRA fully satisfied with quality of data supplied? And if those data were supplied to the MHRA by the License Holder, please explain why they were not seen by members of the Expert Working Group.

Can the MHRA confirm that all three ‘placebo suicides’ involved patients involved in clinical trials for patients with depression? If not, please identify the treatment group (diagnosis) in which each of these placebo suicides occurred.

Please identify, in the case of each of the three placebo suicides, how many days had elapsed after the end of the treatment phase when each suicide occurred.

Can the MHRA confirm that none of the three people on placebo who committed suicide during the post-treatment phase had been taking an antidepressant drug at or around the time they died? If not, please identify which drugs were being taken by each of the three patients identified as ‘placebo suicides’, and how many days had elapsed between commencement of that drug treatment and the date of suicide

Is the MHRA entirely satisfied that the sponsoring company (MAH) fully and properly investigated the circumstances of each of the three ‘placebo suicides’? If not, please identify, in each case, what data were missing and why.

Will the MHRA now publish all (appropriately anonymised) information in its possession at the time of publication of the Expert Working Group report (December 2004) relating to each of the three ‘placebo suicides’ and if not, why not?

Charles Medawar
16 January 2006