Intended for healthcare professionals


Antidepressant discontinuation reactions

BMJ 1998; 316 doi: (Published 11 April 1998) Cite this as: BMJ 1998;316:1105

Antidepressant discontinuation reactions - dependence on antidepressants is significant

(Copy of letter sent a few weeks ago to - before online letters feature available - to which I have not yet had reply, nor am I sure whether is being considered for publication)

Haddad et al minimise the problems caused by discontinuing antidepressant treatment, claiming without evidence that antidepressants are not drugs of dependence.1 The dependence potential of drugs has regularly been denied in history, until eventually accepted for drugs such as opiates, barbiturates and most recently benzodiazepines.2 Discontinuation symptoms were regarded as evidence of benzodiazepine dependence, but this clinical condition would now hardly meet the modern DSM-IV criteria for substance dependence.3 Recognition of the withdrawal effects of benzodiazepines caused a collapse in their market. It may be suspicious that Haddad et al prefer the term discontinuation reaction rather than withdrawal reaction, but it is difficult to assess their bias as they do not declare a conflict of interest.

Semantic confusion about discontinuation, withdrawal and relapse can be traced to dissatisfaction with the definitions of addiction and habituation, leading to the introduction of the single term drug dependence by a WHO Expert Committee in 1964. Since then there have been varying shades of meaning of dependence. The Diagnostic and Statistical Manual of the American Psychiatric Association made tolerance or withdrawal a required criterion in DSM-III, but in DSM-IIIR dependence was redefined as the antisocial syndrome of clinically significant behaviours and symptoms indicating loss of control of substance use and continued use despite adverse consequences.

The former distinction between physical and psychological dependence may still have some relevance in clinical practice. Haddad et al leave open the question of the nature of antidepressant discontinuation reactions. They may be nonspecific effects. More worryingly though are the few reports of suspected neonatal withdrawal reactions resulting from maternal SSRI use in pregnancy.4 Nonetheless it is disingenuous to criticise the public for their commonsense belief that people can become dependent on medication which is regarded as improving mood.5 High placebo response rate of antidepressants is recognised and suggestion can play an important part in initial response to treatment, so expectations are as likely to play a role in withdrawal. Although discontinuation reactions may be minimal, placebo effects can be powerful. Evidence for the value of continuation treatment means patients are likely to remain on antidepressants for some time, increasing the risk of discontinuation reactions. This reliance on medication is significant and is present with other psychotropic medication such as neuroleptics and lithium. Simplistic, dismissive views such as Haddad et al will not help the recognition of these difficulties.

Yours faithfully

D B Double, Consultant Psychiatrist, Norfolk Mental Health Care, Hellesdon Hospital, Drayton High Road, Norwich NR2 2AE. (

1. Haddad P, Lejoyeux M, Young A. Antidepressant discontinuation reactions. Are preventable and simple to treat. BMJ 1998;316:1105-6 (11 April).

2. Medawar C. Power and dependence. Social audit on the safety of medicines. London: Social Audit, 1992

3. Medawar C. The antidepressant web. International Journal of Risk and Safety in Medicine 1997;10:75-126 and

4. Kent LSW and Laidlaw JDD. Suspected congenital sertraline dependence. British Journal of Psychiatry 1995;167:412-3

5. Double DB. Prescribing antidepressants in general practice. People may become psychologically dependent on antidepressants. [letter] BMJ 1997;314:829

Competing interests: No competing interests

01 May 1998
D B Double
Consultant Psychiatrist, Norfolk Mental Health Care NHS Trust
Hellesdon Hospital, Norwich NR6 5BE