Social Audit Ltd
P O Box 111 London NW1 8XG
Telephone/Fax 44 (0)171 586 7771


M. Patrick Deboyser, Head of Pharmaceutical
Products and Cosmetics Unit,
DG-III, European Commission
Rue de La Loi 200
Brussels, Belgium 1049

7 October 1998

Dear M. Deboyser,

Re: antidepressant drug warnings

Thank you for your letter faxed on 5 October, and for taking the trouble to spell out all the complexities of the different appeal mechanisms. I look forward to reading your article, but think I get the general picture: there are several legal doors but none would admit a natural person. Your letter told me several other things that came as a complete surprise to me and they suggest repeated failures of communication.

1. I was certainly not aware that CPMP had already finalised (adopted) the fluoxetine warning in December 1997. Professor Rolf wrote to me two months after this (18 February) saying my paper had been circulated to relevant experts and that "evaluation of the risk of dependence on SSRIs has been initiated".

2. Nor was I aware that the Commission adopted the CPMP proposal on 9 June - ie two weeks before Professor Rolf wrote: "this CPMP decision has been transmitted to the European Commission for decision".

3. In his letter of 23 June, Professor Rolf also wrote: "However the CPMP has considered appropriate to require a number of changes in the SPC … (but) … was of the opinion that no statement on dependency should be included". Naturally, I gained the impression that the CPMP had adopted that warning after considering the risks discussed in my paper.

In the last paragraph of your letter, you wrote: "It is not clear from the correspondence … whether the CPMP was in a position to consider the concerns outlined in your paper before adopting its opinion in December 1997". I agree, but I think it is now much clearer. I first sent my paper to EMEA (M. Sauer) on 2 December and the MCA's Alert to the CPMP Pharmacovigilance Working Party was dated 19 December. It therefore seems highly unlikely that the CPMP had considered the risks involved before approving the fluoxetine warning.

In my paper, since updated by much information on the Social Audit website, I have spelled out the risks involved. I shall not rehearse them again here, but would suggest they boil down to this central question for the CPMP: Is there, or is there not, a substantial risk that withdrawal reactions to SSRIs (among other antidepressants) will lead to something closely analogous to the 'normal dose dependence' problem that was first recognised in the 1980s as characteristic of the benzodiazepines in longer-term therapeutic use?

I cannot conceive of any competent authority denying that this risks exists. The only real uncertainties relate to subsidiary questions, such as: How should dependence be defined, and what do people think it means? Do antidepressants in fact go on working and/or do they tend to sensitise users to episodes of depression? Should patients be warned of risk when they may already be very distressed? To what extent should warnings be based on overwhelming evidence of risk, as opposed to clear evidence of harm?

Given the confusion generated in this correspondence over the past nine months, I now have no idea whether the CPMP Pharmacovigilance Working Party is still actively considering this matter, which was first referred to them late last year. All the indications have been that they are not - apart from the last sentence in your letter, which could be interpreted to mean that the CPMP is still working on it. I would welcome clarification of this point, including some indication of how long it usually takes to make such assessments of risk.

Thank you for your attention, and I shall look forward to hearing from you.

Yours sincerely,
Charles Medawar



Contents page
Correspondence with government