Department of Health
Market Towers 1 Nine Elms Lane London SW8 5NQ
Telephone 020 7273 0763
Facsimile  020 7273 0675 .
Dr David Healy, Director
University of Wales College of Medicine
North Wales Department, Hergest Unit
Ysbyty Gwynedd, Bangor, Gwynedd, LL57 2PW 25 February 2002

Dear Dr Healy, 

I am writing further to my letter of 20 December. I apologise for the delay in responding.

First, I would like to update you on our current position in relation to SSRIs and suicidal behaviour. We consulted the Committee on Safety of Medicines (CSM) on the 12 December 2001 on a possible association between SSRIs and suicidal behaviour. The CSM considered all the material you have provided since June 2000, and data from volunteer studies provided by the marketing authorisation holders.

The Committee advised that these data did not provide evidence to alter their previous position that the evidence was not sufficient to confirm a causal association between SSRIs and suicidality, although an effect in a small high-risk population could not be ruled out. The Committee concluded that no amendment to SSRI product information in relation to suicidal behaviour was required.

In your letter of 17 December you raise a number of points and I will answer them individually.

On the issue of agitation and SSRIs, we consider agitation to be a recognised adverse effect of SSRIs and this is the reason for adding it to the product information. We also consider that agitation may be a feature of depression in some patients. The CSM has reviewed the evidence for an association between SSRI-induced agitation or akathisia and suicidal behaviour and considers that such a link remains surmise at present.

Suicidal behaviour
You have asked for our views on the further data you have provided in your letter. We have analysed this and consider that your analysis and your interpretation of your results all overstate the evidence for an effect of SSRIs on suicide rate and suicide attempts. You claim a relative risk of 2.09, and although the arithmetic is the correct and the result of analysing all the data in a single table gives a highly significant result (more significant than you imply), it is misleading to pool the data in this way. A correct analysis is only marginally significant, but should be based on the original trial data. It is likely that if the original data were subject to a correct meta-analysis, taking into account the time that patients were exposed to the risk of suicide, then the evidence would be even weaker. 

I would also draw your attention to a recent publication in the Journal Pharmacoepidemiology and Drug Safety (Carlsten et al, 2001; 10:525-530) where the authors found a significant reduction in suicide rates following the introduction of SSRIs in Sweden, corresponding to approximately 348 suicides in the seven year period following the introduction. While these data have limitations, they do not support an increased risk of suicidal behaviour with SSRIs.

Withdrawal reactions and dependence
With regard to dependence and SSRIs you make two points:

1. Withdrawal reactions and dependence are synonymous:
For our assessment of this issue in 1998 we used the definition of a dependence syndrome from the World Health Organisation International Classification of Diseases (ICD-10) is as follows:

'A cluster of behavioural, cognitive, and psychological phenomena that develop after repeated substance use and that typically include a strong desire to take the drug, difficulties in controlling its use, persisting in use despite harmful consequences, a higher priority given to drug use than to other activities and obligations, increased tolerance, and sometimes a physical withdrawal state.'

This definition is consistent with that in the American Psychiatric Association Diagnostic and Statistical Manual 4th Edition (DSM-IV)

It is generally accepted that withdrawal reactions on stopping a drug are not sufficient, or necessary, for a diagnosis of drug dependence. Other features such as tolerance (requiring increased doses of the drug to produce the same effect) and drug seeking behaviour are required for this diagnosis.

2) There are data on patients who have been taking SSRIs for months or years and cannot stop.
The sources of information used for our review of SSRIs, withdrawal reactions and dependence were: spontaneous ADR data; published literature; unpublished data; information from other regulatory authorities and usage data from the Prescription Pricing Authority and the MediPlus database. Detailed review of these data revealed evidence that SSRIs cause withdrawal reactions, however the was evidence did not suggest (sic) that SSRIs were drugs of dependence or that a large proportion of patients were taking SSRIs for months or years and were unable to stop. We would be very interested to receive any information that you may have in relation to such patients.

The product information for SSRIs warns that they can cause withdrawal reactions and contain appropriate warnings and advice on gradual discontinuation of treatment.

Interests of experts
As you have suggested in your letter we have written to our experts on psychiatric matters and asked them to outline their interests over the years.

Finally thank you for the offer of a meeting. We would be happy to meet with you again, I will write again next week with a possible date.

Yours sincerely
Sarah Wark
Team Leader - Pharmacovigilance

Copy: Dr J. Raine MCA/PL
Prof Stephen Evans MCA/PL
Dr Cheryl Key MCA/PL


or HERE for full index of this correspondence

Contents page
What's New?