Department of Health
Market Towers 1 Nine Elms Lane London SW8 5NQ
Telephone 0171-273 0270
Facsimile 0171- 273 0293 15 September 1997

Dear Mr Medawar


Thank you for your letter of 23 July 1997 and for taking an interest in our publication "A Comparison of Post-Marketing Safely of Four Selective Scrotonin Reuptake Inhibitors Including the Investigation of Symptoms Occurring on Withdrawal" (B.J Clin. Pharmacol., 1996, 42, 757-763) and the presentation at the International Conference on Pharmacoepidemiology on the same subject. Please accept my apologies for the delay in replying. Unfortunately your letter arrived while I was on annual leave and in formulating the reply it was necessary to consult with my colleagues.

Our response to your questions are as follows:

1 . Pre-licensing clinical trials on paroxetine revealed symptoms on abrupt withdrawal, similar in nature to those reported subsequently as adverse drug reactions through the spontaneous reporting scheme. There were no other features to suggest dependence. At our request, the applicant included advice against abrupt withdrawal of paroxetine in the data sheet.

2. "Dependence syndrome" is defined in the International Classification of Disease (ICD-10, World Health Organisation, 1992). A definite diagnosis of dependence should usually be made only if 3 or more of the following have been present together.

a) Strong desire or sense of compulsion to take the substance.

b) Difficulties in controlling substance-taking behaviour in terms of its onset, termination, or levels of use.

c) Physiological withdrawal state.

d) Evidence of tolerance, such that increased doses of the substance are' required in order to achieve effects originally produced by lower doses.

e) Progressive neglect of alternative pleasures or interests because of psychoactive substance use, increased amount of time to obtain or take the substance or to recover from its effects.

f) Persisting with substance use despite clear evidence of overtly harmful consequences. Only the third of these characteristics (a physiological withdrawal state) has been reported with paroxetine.

3. The paper did not attempt to make any comparison with benzodiazepines. A comparison would be inappropriate in view of different marketing lives, usage, publicity and other factors which influence the number of spontaneous reports received. However, reports of dependence for benzodiazepines include other features listed in paragraph 2, such as craviiig, drug seeking behaviour and dose escalation (tolerance). There are therefore qualitative differences to the reports with SSR1s.

4. Your question highlights an error on our part in the presentation of the results from the survey. Doctors were invited to categorise the reaction as "mild", "moderate" (not "moderately severe") or "severe". Paragraph 5 of the results of the survey should therefore read "21% of reactions were said to be mild, 58% were moderate and 21% severe. The conclusion is therefore consistent with the findings for the majority of patients.

5. Dependence has been defined above (paragraph 2). We consider benzodiazepines to be capable of causing true dependence.

6. The study did not attempt to make any comparison with tricyclic antidepressants. Such a comparison. would be inappropriate for the same reasons as it is inappropriate to draw any quantitative comparisons with benzodiazepines from these data. We have received a few reports of patients experiencing symptoms on cessation of tricyclic antidepressants. In one case the patient escalated the dose, but the product taken contained a combination with a benzodiazepine. The discussion section referred to the results of Prescription Event Monitoring (PEM, reference 11 of the paper). PEM has an advantage over spontaneous reporting data in that the denominator is known. In this study there was only one case of withdrawal symptoms among over 13,700 patients given paroxetine. This low frequency is consistent with the conclusion that withdrawal reactions are rare.

7. As already stated, quantitative comparisons with benzodiazepines or tricyclic 'depressants are inappropriate from the data presented. The difficulty detecting anti differences between each SSRIs in the relative risk of withdrawal symptoms refers to the failure of PEM to detect a difference. This is an important difference between the methodology of spontaneous adverse drug reaction reporting and PEM. PEM collect data on around 10,000 patients, and therefore cannot estimate the frequency of rare reactions. Spontaneous adverse drug reaction reporting collects data from the entire patient population treated, which in the case of SSRIs is now estimated from the number of prescriptions to exceed several million, and it can therefore detect much rarer reactions. The ability of spontaneous ADR reporting to detect rare reactions not identified by alternative methods is one of its strengths.

8. There has been no change in our understanding of dependence since advice was provided to doctors and pharmacists in Current Problems No. 21 (January 1988).

9. The proportion of patients in the survey who were still taking paroxetine 3 months after re-starting it (19%) is an important observation. However, on its own this observation does not constitute evidence of dependence (defined in paragraph 2) or habituation (which is characterised by reduced effectiveness and a craving for the drug). It is also possible that at least some of these 19% had continued with paroxetine because of a recurrence of the depressive illness, occurring coincidentally with, or soon after the doctor restarted paroxetine to treat withdrawal symptoms. The questionnaire did not specifically seek the reason why paroxetine was being used 3 months later, though 3 doctors volunteered that it was because depression had recurred.

10. The reasons why the symptoms seen on withdrawal of paroxetine are unlikely to represent recurrence of the original illness are stated in the paper: the speed of onset and the nature of the reactions are unlike those characterising depressive illness. Nevertheless it is important that doctors and patients are alert to the possibility of withdrawal symptoms and do not confuse them with disease recurrence. For this reason, the data sheet states the symptoms that can occur on drug cessation. We continue to monitor reports of adverse reactions for the SSR1s, scrutinising them in particular for the other features of dependence.

1 hope that these responses are helpful.

Yours sincerely

Dr J Price
Head of Cinical Evaluation Unit


Contents page
List of MCA/CSM Correspondence