In both of our meetings last week, you erupted about my criticism of conflicts of interest in the medicines control system. In particular, you said you take the greatest exception to questions I have raised about a former member of the CSM. On both occasions you protested before I had said anything, which indicates you were provoked by something I have written.

It may therefore help to review what I wrote and why, in particular about licensing decisions relating to paroxetine. The following are the main points in The Antidepressant Web. I believe they are matters of fact, but please let me know if you disagree with any of them and, of course, I'll think again.

Paroxetine was licensed in 1991. It has been sold in the UK under the brand name Seroxat® (SKB Pharmaceuticals) since then.

Sometime before 1991, the CSM/MCA realised that some patients experienced withdrawal symptoms when they tried to stop taking paroxetine. The MCA has refused to disclose how many cases there were but, since 1991, the data sheet for paroxetine has warned of this and advised gradual withdrawal.

Since 1991, the data sheet for paroxetine has also included the claim that: "Long-term treatment with Seroxat has shown that efficacy is maintained for periods of at least one year"

This claim was essentially based on the results of a study, subsequently published as: S.A. Montgomery, G. Dunbar, Paroxetine is better than placebo in relapse prevention and the prophylaxis of recurrent depression, Int. J. Clin. Psychopharmacol., 1993, 8, 189-195. Dr. Dunbar was then employed by SKB; Dr Montgomery was a member of the CSM (also Editor of the journal in which this study was published).

The design of this study altogether discounted the possibility of withdrawal symptoms following abrupt discontinuation of paroxetine after long-term use. Therefore its conclusions are invalid and misleading: the investigators would have interpreted paroxetine withdrawal symptoms as a relapse into depression - mistakenly counting this as evidence of drug effectiveness.

Notwithstanding early evidence of withdrawal symptoms with paroxetine, and subsequent evidence of a greater problem than previously envisaged by the licensing authorities, the claim for long-term efficacy still stands and the possibility of dependence on SSRIs is denied.

The MCA/CSM still maintain that paroxetine withdrawal symptoms are rare and generally mild. Their best evidence is that 20% of cases are mild and 20% severe, with most people experiencing something in between. If withdrawal symptoms were indeed rare, they would not have been seen in pre-marketing clinical trials in numbers sufficient to justify a warning - especially not when the trials involved limited numbers of patients, and when no systematic attempt was made to assess the effects of withdrawal. The CSM/MCA's estimate of the incidence of withdrawal problems is of the order of at least 100-times lower than other investigators have found.

Dr Montgomery was a member of the CSM between 1987 and 1992 and the Committee's leading authority on SSRIs. Four such drugs were licensed in this period and he declared personal financial interests with three of manufacturers involved, including SKB. He declared professional and personal interests with several other companies besides.

At about this time, the CSM would have been much concerned about the problem of dependence on benzodiazepines. The Committee knew that the earliest recognition of withdrawal symptoms with benzodiazepines - following longer-term use at normal doses - clearly signaled a major dependence problem (which had gone unnoticed for two decades).

The CSM did not require SSRI manufacturers to provide any substantial evidence of lack of risk of therapeutic dependence with antidepressants, to compare with that seen with benzodiazepines.

The number of yellow card reports about withdrawal problems with paroxetine now far exceeds the number received for any other drug.

Yes, this does worry me, but my main concern is not about one CSM member, but about the system in which he worked. It is particularly important that the medicines control system be seen to be independent because it is highly secretive, systematically excludes consumer representation, and deals with very personal issues including matters of life and death.

It adds to my concern that, initially, the MCA sought to brush aside my concerns (Press statement, 4 December 1997), and that now you seem to be defending the system essentially on the grounds that I am behaving outrageously.

You have defended CSM members accepting payments from companies on the grounds that companies like to contract with the best people in the field and that CSM members may need to work closely with companies to be well informed. To some extent I agree that the experience of working with companies may be useful, also that companies go for the best people. But they also tend to go for the best for them and, these days, companies also play a major part in deciding who gets to the top, and who does not. Accepting payment from companies is clearly not a pre-requisite for becoming well-informed, and getting close to them does carry some risk of being over-exposed to good news about drug products and to the "hospitality" that goes with it.

You have also referred to specific safeguards the CSM uses to prevent undue influence in the system and said that CSM members with possible conflicts must disclose all relevant interests, not take part in any vote; and leave the room during certain discussions. I could find no specific reference to either voting or leaving the room in the Code on Declaration of Interests, and I wonder how relevant they are. However, I note the following about disclosure of a "personal specific interest":

"The member shall take no part in the proceedings as they relate to the product, except, at the Chairman's discretion, to answer questions from other members." (section 11a)

To help me to understand the position better, please could you could arrange for me to be sent the following information, and for these to be treated as formal Code requests:

1. On how many occasions in meetings of the CSM held during both 1991 and 1997, was there recorded in the minutes or elsewhere: [a] any formal vote on any matter directly connected with the licensing of any particular drug; and [b] any instance when a member absented him/herself from the room to avoid some possible conflict?

2. Please supply the following from the record (referred to in section 14 of the Code of Declaration of Interests) of names of CSM members "who have declared interests at meetings … giving dates, names of relevant products and companies, details of the interest declared and whether the member took part in the proceedings":

[a] copies of the full entries made, from 1987 to date, relating to any personal specific interests disclosed during CSM meetings, when the product under consideration was fluvoxamine, fluoxetine, sertraline, paroxetine, venlafaxine, citalopram or nefazodone.

[b] copies of the full entries made, relating to any personal specific interests disclosed during the last five CSM meetings, for any "relevant product".

I confirm we will pay all reasonable charges. There is no need to check back with me, unless the cost is estimated to be over £500 (which I cannot conceive it would be).

I hope this approach seems reasonable to you. You have "challenged" me to refer the matter to the Committee on Standards in Public Life, and I may yet do so. In the meantime, my mind is not closed to the idea that the CSM did all that might reasonably be expected of it to avoid conflicts of interest - even if it is now at the centre of a real mess for which I'm afraid the public will pay dearly.

As a matter of (public) interest, has the CSM ever agreed any sort of policy on what to do when it fears or recognises it has made a serious mistake involving public health and safety? Or does the situation never arise?

With good wishes.

Yours sincerely,