Antidepressant discontinuation reactions

Discontinuation reactions from antidepressants have been recognised since the drugs were first introduced¹ and can occur with all the major classes of antidepressants.²,³ This phenomenon has important implications for antidepressant prescribing, particularly as these drugs are increasingly used in disorders other than depression. Nevertheless, antidepressant discontinuation reactions have received little systematic study and many clinicians are unaware of them.⁴

The incidence of discontinuance reactions is unclear owing to the lack of research and of an accepted definition of what constitutes a discontinuation reaction. Antidepressants vary in their propensity to cause reactions,⁵ and reactions are more common after abrupt termination and longer courses of treatment.⁶,⁷ Given this background, the reported incidence has varied from 0%⁶ to 100%.⁸ One of the few double blind placebo controlled studies found that in the two weeks after a 12 week treatment period adverse events, mostly mild or moderate, occurred in 35% of patients treated with paroxetine compared with 14% of controls.⁹ Although this study was carried out in patients with panic disorder, with certain antidepressants reactions probably occur in a significant minority of patients of all diagnostic categories when they stop treatment.

Discontinuation reactions are distinct from recurrence of the primary psychiatric disorder. They usually start abruptly within a few days of stopping the antidepressant (or, less commonly, of reducing its dose) and are short lived, resolving within one day to three weeks.⁵,⁶ In contrast, depressive relapse is uncommon in the first week after stopping an antidepressant: symptoms tend to build up gradually and become chronic. Discontinuation symptoms are varied and differ depending on the class of antidepressant. Common symptoms include gastrointestinal disturbance (nausea, abdominal pain, diarrhoea), sleep disturbance (insomnia, vivid dreams, nightmares), general somatic distress (sweating, lethargy, headaches), and affective symptoms (low mood, anxiety, irritability). Although there is some overlap with anxiety and depressive disorders,² many discontinuation symptoms are distinct. With the serotonin reuptake inhibitors the commonest symptom appears to be dizziness/light headedness, with sensory abnormalities—including numbness, paraesthesia, and electric shock-like sensations—also well recognised.⁶ Discontinuation reactions usually resolve within 24 hours of reinstating antidepressant treatment,⁶ whereas in depressive relapse the response is slower.

Discontinuation symptoms do not in themselves indicate drug dependence. Dependence is a syndrome,¹⁰ and diagnosis requires several other features, such as tolerance, inability to control drug use, primacy of drug taking behaviour, and continued use despite harmful consequences. Antidepressants are not associated with these other features and are not drugs of dependence. The common lay belief that antidepressants are addictive probably contributes to the significant undertreatment of depressive illness.¹¹ It is important not to foster this belief inadvertently—one reason that “discontinuation reaction” is a better term than “withdrawal reaction.”

Discontinuation reactions are clinically important for several reasons. Firstly, although most are mild and short lived, a minority are severe or chronic and cause considerable morbidity.² Secondly, if the reaction is misdiagnosed, inappropriate treatment may result. For example, a reaction after stopping antidepressants may be misdiagnosed as a relapse of the psychiatric illness, leading to unnecessary reinstatement of the antidepressant. A reaction after covert non-compliance may lead to the erroneous conclusion that a higher dose or a switch to another antidepressant is needed. Many discontinuation symptoms are physical and may prompt investigations to identify the cause. These scenarios waste money, put the patient at unnecessary risk, and lead to a more negative prognosis that may have social implications. Finally, if the phenomenon is not explained, the patient who recognises the association between the antidepressant and the discontinuation symptoms may comply poorly with further antidepressant treatment.

To reduce the likelihood of discontinuation reactions the British National Formulary recommends that antidepressants that have been continuously prescribed for eight weeks or more should not be stopped abruptly but gradually reduced over four weeks.¹² Given current knowledge this seems reasonable, though anecdotal reports suggest that tapering may be unnecessary when switching between serotonin reuptake inhibitors.⁷ Patients need to be educated that antidepressants are non-addictive, doses must not be omitted, and courses not stopped abruptly. If recognised, discontinuation reactions are not a serious problem. Clinicians should consider the diagnosis when faced with unexpected physical or psychological symptoms in patients who have just stopped taking antidepressants or are apparently still on treatment: only a few days’ medication needs to be missed to precipitate a reaction, and antidepressant non-compliance is common and often covert unless specifically inquired about.

If antidepressant treatment is still required, restarting the antidepressant will rapidly resolve the discontinuation symptoms. If antidepressants are no longer clinically indicated treatment depends on severity. Most cases are mild and require only reassurance. Symptomatic treatment, such as a short course of a benzodiazepine for insomnia, may help with more troublesome symptoms. If severe the antidepressant should be restarted and tapered down gradually—occasionally very gradually. In summary therefore, discontinuation reactions are a significant problem only when strategies for prevention and recognition are ignored.