[Discontinuation of psychotropic medication: a critical review of the evidence]

Pharmacotherapy is an effective treatment strategy for psychiatric disorders. Nevertheless, many patients discontinue their medication at some point. Evidence-based practices for patients, clinicians, and policymakers to discontinue psychotropic medication properly are therefore important. However, guidelines and consensus papers on this topic have been limited in numbers and scope.

To provide a summary and critical review of the scientific evidence on discontinuation of four classes of psychotropic medication: antidepressants, antipsychotics, benzodiazepines and mood stabilizers.

Experts on each medication class conducted a literature review and convened in consensus meetings to identify similarities and differences between the medication classes. For each class, the following questions are discussed: 1. who can discontinue (including risk of relapse, risk of withdrawal symptoms); 2. when to discontinue (including optimal duration of medication use); and 3. how to discontinue (including interventions to reduce the risk of relapse, dose reduction versus complete discontinuation). Similarities and differences between the classes are described.

Abrupt or rapid discontinuation of most psychotropic medications results in more withdrawal symptoms compared to gradual tapering and a doubled risk of relapse compared to continued treatment. Additionally, several patient characteristics are associated with successful discontinuation of psychotropic medications. Furthermore, marked differences exist in the optimal timing and speed of tapering for each medication class.

Integrating and comparing scientific knowledge on discontinuation of different classes of psychotropic medications provides patients and professionals with guidance to make informed decisions about discontinuation and identifies knowledge gaps. There is a need for research on the role of pharmacotherapeutic tapering schedules and supportive psychological interventions, focusing on patient-relevant outcome measures (such as social functioning).