Department of Health
Market Towers 1 Nine Elms Lane London SW8 5NQ
Telephone 0171-273 0600
Facsimile 0171- 273 0737 .
28 January 1999

Dear Mr Medawar,

Thank you for your letters of 22 November and 3 December to Dr Jones to which I am replying. I am sorry I have not been able to reply sooner but I needed to seek the views of colleagues.

You suggested that the evidence presented by the 1998 Rosenbaum et al study is sufficient to warrant withdrawal of the 1996 Price et al article. We cannot agree and we would take the same view if the dates of publication for the two articles were reversed. This is because the studies use different - not superior - approaches and methodologies. Consequently, the results are different too. On the one hand, Rosenbaum et al report the results of clinical trials in which small numbers of patients are closely observed and questioned about symptoms on withdrawal. This approach may overstate the clinical relevance of such symptoms. On the other hand, adverse reactions which are recorded via spontaneous reporting are likely to be considered clinically important. However, such data can only give a limited indication of frequency because such symptoms may not be recognised as such and, even if they are recognised, may not be reported.

Any decision to withdraw the paper is one for the authors and not the MCA. However, we can see no reason for them to agree your suggestion for the reasons given above. We understand that Dr Mackay and Dr Waller share this view.

In the light of the Rosenbaum study, you ask the MCA to ensure withdrawal of the statement from the paroxetine data sheet that efficacy is maintained for at least a year. The study by Rosenbaum et al was designed to assess the effects of intermittent missed doses of antidepressant treatment. It did not examine the long term efficacy of paroxetine. We therefore cannot accept that this study discredits earlier studies nor that it provides any evidence to justify the action you suggest.

You ask what the MCA's reactions are to the suggestion that the paper points to a serious problem with fluoxetine and what our plans are to limit this risk. Our review of the available data indicates that withdrawal reactions occur with all SSRIs, although they are reported less frequently with fluoxetine than with the other SSRIs. This may be for a number of reasons, including the possibility that the onset of the withdrawal reactions after stopping fluoxetine may be delayed and therefore not recognised as such by patients or doctors. We do not consider that the current evidence points to a serious problem with fluoxetine.

Nonetheless we will bear this in mind as our review of the evidence of the issues raised by your paper The Antidepressant Web continues. You will know from previous replies we have sent you that we have been seeking a range of views on this matter and we will keep you informed of progress on this matter.

I have noted your comments about the current UK medicines control system. You will not be surprised to learn that I do not share your views.

Yours sincerely
Roy Alder
Head of Executive Support


Contents page
List of correspondence with MCA/CSM