Your 58-year-old patient started risperidone to augment lithium 2 years ago. It got her out of a severe mania, and she has stayed well since then. Now she’s worried about long-term risks and wondering if it’s time to come off.

Augmentation with an atypical antipsychotic may offer rapid relief from mania and depression, but antipsychotics’ potential side effects are substantial. They include hyperlipidemia, diabetes, obesity, and tardive dyskinesia, the latter increasing at a rate of 2.6% with each year of use. These side effects tend to build gradually over time, so it’s important to know when it’s clinically appropriate to try stopping the antipsychotic and possibly continuing solo with a mood stabilizer or antidepressant.

Around a dozen industry-supported trials have looked at this question in both bipolar and unipolar disorders. These studies all followed a similar design: They started with patients who recovered with antipsychotic augmentation and then tested what happens when the antipsychotic is withdrawn after 2–3 months of treatment. Typically half the patients continued with the antipsychotic and the other half were switched to placebo. The patients were then followed for a long stretch—from 6 months to 2 years—and the rate of relapse was carefully tracked (Lindström L et al, J Affect Disord 2017;213:138–150; Kato M et al, CNS Drugs 2013;27 Suppl 1:S11–19).

The consensus of these studies is that continuing the antipsychotic prevents relapse into new mood episodes, but there’s a catch. Most of these studies withdrew the antipsychotic within the first few months after recovery, which is a particularly vulnerable time for patients. What if the antipsychotic were stopped later, after 6 months of recovery?

We could find only a single study that looked at this question, and it was in bipolar I disorder. This non-industry-sponsored study enrolled 159 patients who had recovered from bipolar mania after their mood stabilizers were augmented with an antipsychotic, either risperidone or olanzapine. Patients were then randomized into 1 of 3 groups: 1) Continue the antipsychotic for a year; 2) Switch to a placebo after 2–6 weeks of recovery; or 3) Switch to a placebo after 6 months of recovery. All groups were kept on their original mood stabilizer (lithium or valproate). The result: Patients who stopped the antipsychotic after 6 months had the same risk of relapse as those who stayed on the antipsychotic (65%), but the risk was significantly higher when the medication was stopped after 2–6 weeks (87%). Those are high relapse rates across the board, but that’s not uncommon in mania (Yatham LN et al, Mol Psychiatry 2016;21(8):1050–1056).

That study suggests that antipsychotics can be safely discontinued after 6 months of recovery from mania. In depression, antipsychotics are overall less effective as preventative agents than they are for mania. Among the agents studied, only quetiapine has successfully prevented depression in controlled trials, 2 in bipolar disorder and 1 in unipolar (Liebowitz M et al, Depress Anxiety 2010;27:964–976; Lindström L et al, J Affect Disord 2017;213:138–150). Those trials tested discontinuation at 3 months. No one has tested whether 6 months of maintenance would be adequate in depression, but we have a hint that it might from the studies of antidepressant discontinuation. In those trials, most of the relapses were due to discontinuation of the antidepressant before 6 months of stabilization, according to a meta-analysis of 45 controlled trials (Baldessarini RJ et al, J Clin Psychopharmacol 2015;35(1):75–76).

In practice, I usually wait 6 months before trying antipsychotic discontinuation. The decision is collaborative and different for each patient. If we decide to come off, I’ll taper slowly, lowering the dose every 2 weeks. I’ll encourage the patient to build habits known to prevent mood episodes, such as mindfulness, exercise, the Mediterranean diet, and regular sleep and wake times. If the taper does not work, I may try again after 1–2 years. About half the time, the second attempt is a success.

TCPR Verdict: In mood disorders, wait at least 6 months after recovery before attempting to discontinue an adjunctive antipsychotic. Taper slowly, watch for relapse, and encourage lifestyle changes to prevent new episodes.