WITHDRAWAL FROM ANTIDEPRESSANTS

COMMENTS BY PROFESSOR MALCOLM LADER

1. I have been asked by Dr R E Kendell to comment on the important issues raised by Mr Charles Medawar of Social Audit Limited in his letter to Dr Kendell in his capacity as President of the Royal College of Psychiatrists. The main question which Mr Medawar raises concerns the College's Defeat Depression Campaign and the dependence potential of the anti-depressants advocated in that campaign. I was not involved in the Defeat Depression Campaign and so cannot comment on the funding.

2. I must also declare an interest in that I have advised some of the companies marketing anti-depressants about the problem of withdrawal syndromes. In particular, I have advised SmithKline Beecham about Paroxetine which seems to be the compound producing the most problems.

3. It is important not to become too embroiled in the semantics of what we mean by relapse, rebound, withdrawal etc. The important aspect is to maintain the usual perspective on the risk/benefit ratio of any therapeutic manoeuvre. In the case of anti-depressants the salient features for our purposes are efficacy in long term use versus risk of clinically important problems when trying to withdraw from medication both short term and long term.

4. As is well known depression can become a chronic condition in a proportion of patients. Any therapeutic benefits from an anti-depressant will therefore need to be continued. If that anti-depressant is withdrawn then the depression will recur and the patient will feel the necessity to revert to medication which proved successful before.

5. As with all drugs which interact with receptors or enzymes, adaptive processors take place so that withdrawal of the medication can be followed by rebound which is an overshoot of that interaction. This is clearly seen with central drugs such as hypnotics with rebound insomnia but also with drugs acting peripherally such as blood pressure and beta blockers or gastric acidity and H2 antagonists.

6. However, with centrally acting drugs there is an additional problem that further interactions can take place which result in a central withdrawal reaction when that medication is stopped. The usual criteria with centrally acting drugs are that new symptoms occur which were not present when the medication was initially given. This is clearly seen with the benzodiazepines: withdrawal reactions with new symptoms, often perceptual, are seen when the drug is stopped abruptly or even tapered off. However, with the benzodiazepines there are other problems such as escalation of dosage with an undoubted physical dependence, and abuse potential where the drug is used and sought for its mind altering properties.

7. There is no doubt whatsoever that the anti-depressants in general, the SSRIs in particular and most specifically Paroxetine are associated with a withdrawal syndrome in a proportion of cases. That syndrome is usually fairly short-lived in some patient but in some patients can given rise to prolonged problems in which it is difficult to distinguish drug discontinuation phenomena from the natural progression of what may be a complex depressive illness. The question arises as to whether it is justified to use anti-depressants in this context. Unlike the benzodiazepines there is no evidence of escalation of dose and no evidence whatsoever that anti-depressants can be abused "on the street".

8. After that long prologemenon I will now attempt to address the various points raised by Charles Medawar. The question as to whether there is a risk of dependence is as I have pointed out a matter of semantics. I do assert that there is a difference between withdrawal from tranquillisers and withdrawal from anti-depressants where the former produce problems related to central dependence whereas the anti-depressants probably do not induce those changes. Rather there is a rebound of the changes in central biochemistry which the anti-depressants produce. Thus most of the withdrawal symptoms can be attributed to under-activity of neurotransmitters such as serotonin. As a clinical problem there is undoubtedly a proportion of patients who benefit from being tapered gradually and this should be normal practice, as it should be with the discontinuation of almost any drugs. The data would indicate that  relatively few patients compared with the benzodiazepines are disadvantaged clinically to any material extent although I do acknowledge that this can happen and I have attempted to cope with a few of these patients myself. I also agree that for some reason Paroxetine is more likely than the other SSRIs to produce problems. This may be related to its relatively short half life. In summary, therefore, I would state that there is a withdrawal reaction but this is not indicative of a central dependent state as in the true drugs of addiction.

 9. Charles Medawar also asks about evidence of long term effectiveness and takes Professor Montgomery to task for his studies on long term effectiveness. There have been several studies in the recent past, mostly involved in drug registration, which have allocated patients on anti-depressants randomly to continue either on the drug or to switch to placebo. By and large there is a useful reduction in relapse on such long term treatment with figures on placebo around the 50% to 60% per year mark and figures on active maintenance therapy around 70% to 90%. This is I think the basis for long term use and in my opinion does justify this. We do have to remember the dangers of depression not only those of suicide but also occupational handicap, disruption of interpersonal relationships and personal symptomatic distress. I do also accept that the efficacy of the anti-depressants both short term and long term is not particularly high. My estimates would be that anti-depressants help about 70% of acute depressives whereas placebo appears to help about 30% to 40%. However, there was some evidence that current placebo responders may relapse within the next three months. We also have to remember that in most patients depression is a remitting condition so that placebo response is essentially reflecting spontaneous remission.

M.H. Lader
11November 1997

CLICK HERE TO READ ON

Contents page
Correspondence: matters arising