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


Dr. Robert Kendell CBE, MD, FRCP, PRCPsych  
President, Royal College of Psychiatrists  
17 Belgrave Square  
London SW1X 8PG 2 December 1997

Dear Dr Kendell

Thank you for your letter of 28th November which I received this morning, together with the opinion from Professor Lader. It will take more time than I have at present to respond in detail, but I would want you to know in the meantime that I have not altered my view of the need for a fundamental reappraisal of some of the key messages and recommendations of the College's Defeat Depression Campaign.

I am not confusing dependence with withdrawal, and it seemed quite unnecessary to spell out the criteria for diagnosis of a withdrawal state when, as Malcolm Lader puts it: "there is no doubt whatever that antidepressants in general, the SSRIs in particular and most specifically paroxetine are associated with a withdrawal syndrome in a proportion of cases".

I remain convinced that the RCP is misleading the public when it talks about 'dependence'. The fact is that the benzodiazepines were defined as drugs of dependence essentially on the basis of observations of withdrawal reactions following normal dose use, in the days when "The diagnosis of all of the Substance Dependence categories requires only evidence of tolerance or withdrawal…" (DSM-III). As I said in my letter of 30th October, though you did not comment on it, that same point is underlined in the 1990 report of the APA's Task Force report on Benzodiazepine Dependency: "The presence of a predictable abstinence syndrome following abrupt discontinuance of benzodiazepines is evidence of the development of physiological dependence…" Under this definition, I believe the SSRIs would have to be counted as drugs of dependence. Please let me know if and why you would not agree.

The more recent ICD-10 definition of dependence which you spell out in your letter underlines my second point - that the goalposts have been dramatically moved. The definition of 'dependence' that the RCP refers to now is fundamentally different from the definition you employed a decade ago. But instead of explaining these nicer nosological points, the College has gone to considerable lengths to tell everyone there is no risk of dependence with SSRI or other antidepressants - and that they are quite different in this respect from benzodiazepines.

I think it almost self-evident that not only the public and the media but also many doctors will have been misled by this but, again, I invite you to explain the College's position if you do not agree. I agree with Malcolm Lader's point, "the question as to whether there is a risk of dependence is … a matter of semantics" but, given that it was the College that has sought to completely revise the meaning of the word 'dependence', it seems pretty cool to suggest (on the basis of its MORI surveys) that three-quarters of the public is confused (and that I am too).

You claimed that the College dealt even-handedly in assessing the relative merits of SSRIs and older antidepressants. My view would be that the timing and very existence of the Defeat Depression Campaign signalled rather the opposite, and I also think the College would be hard put to dissociate itself from the following view: "Many newer compounds are less toxic in overdose and have fewer side effects. They are therefore particularly useful when there is a clear suicidal risk or when side effects are likely to be a problem". Given the nature of the side effects problem with antidepressant drugs, and concerns about risk of suicide, this must be regarded as a positive recommendation. It comes from the 1992 consensus statement, copies of which were widely circulated by the RCP as part of the Campaign. I take your point about the consensus statement not being an expression of the formal views of the College, but it seems very academic.

I will comment only briefly on Malcolm Lader's opinion, which I thought very diplomatically expressed. I was especially reassured not to find anything in his opinion which contradicted his account of the Historical Development of the Concept of Tranquilliser Dependence, as follows:

"My own studies (with Hallström and Pétursson)… established unequivocally that normal dose dependence as manifested by a physical withdrawal syndrome was a definite entity and supervened even if the dosage was tapered off. Tolerance with escalation of dosage was not a prerequisite for physical dependence. Indeed our initial study compared withdrawal syndromes in small groups of patients withdrawing from high dose misuse or low dose use and found that the syndromes were identical"

Admittedly, Lader now emphasises the significance of new symptoms occurring on withdrawal, but given reports of neonatal withdrawal following maternal use of SSRIs, and the fact that it took over 20 years to define the BDZ withdrawal syndrome, I can feel sure he is keeping an open mind.

I would not, however, agree that SSRIs and benzodiazepines differ significantly in respect of either development of tolerance or escalation of dosage. Perhaps Malcolm Lader will reconsider his view that there is no evidence of escalation of dose with antidepressants in the light of the evidence in my paper, also taking account of the results of the ‘naturalistic’ study, involving analysis of 21,000 SSRI prescriptions for outpatients at a US urban teaching hospital. This Lilly-sponsored study reported that a mean 5% of patients on fluoxetine and 15% on sertraline "had their daily dose increased with each prescription refill during the first nine prescriptions". Indeed, the mean daily dose for all 460 patients on sertraline doubled during the same period. *** (Gregor et al., 1994)

Finally, I should like to emphasise that I have no problem whatever with the proposition that dependence would be a price well worth paying for the effective treatment of severe depression. My concerns are [a] that the Royal College has consistently and emphatically advised people there is no risk of dependence, when there clearly is; and [b] that because this risk has been denied, sweeping and I think quite mistaken assumptions have been made about the long-term effectiveness of antidepressant drug treatments.

Perhaps you would want to comment on these points before hearing from me further on the several other matters in your letter I might address. I shall look forward to hearing from you.

Yours sincerely,

Charles Medawar

Reference: K.J. Gregor, J.M. Overhage, S.J. Coons, R.C. McDonald
Selective serotonin reuptake inhibitor dose titration in the naturalistic setting.
Clinical Therapeutics 1994; 16(2): 306-15.


*** Correction. Please delete the words, "all 460" from this sentence: there was a very high attritrion rate in this study: only 38/460 patients completed the nine courses of sertraline treatment. Please refer to review of the paper by Gregor et al. CM, January 1998.  
Contents page
List of MCA/CSM Correspondence