Edward had hypomania at 15. Now 17, he has been stable on lithium and aripiprazole for 12 months. But he tells you he doesn’t want to take medicine “forever” and has been hearing about “deprescribing” from podcasts.

Bipolar disorder (BD) is marked by functional impairment and recurrence. Should we ever reduce medications? If so, what is the safest approach? This update reviews maintenance, relapse prevention, adolescent antipsychotic exposure, and adjunctive psychotherapy.

FDA-approved treatments

Many medications are FDA approved for bipolar mania in youth (see table on page 8). Second-generation antipsychotics (SGAs) are often effective in acute pediatric mania. A 2025 network meta-analysis (18 RCTs, n=2,844) found that 6 SGAs outperformed placebo for manic symptoms, while no mood stabilizer except lithium (borderline significance) did (Vita G et al, J Am Acad Child Adolesc Psychiatry 2025;64(2):143–157).

AACAP’s most recent practice parameter recommends maintaining the acute stabilization regimen for 12–24 months (McClellan J et al, J Am Acad Child Adolesc Psychiatry 2007;46(1):107–125). The need for lifelong care is supported by more recent reviews (Nierenberg AA et al, JAMA 2023;330(14):1370–1380).

From the Clinical Update
“Medications for Pediatric Bipolar Disorder: Is Deprescribing an Option?”
by Anthony Charuvastra, MD
The Carlat Child Psychiatry Report, Volume 17, Number 7 & 8
October/November/December 2026
www.thecarlatreport.com

Why consider stopping?

The primary reasons to consider reducing or discontinuing medications for youths with BD include:

1. Reducing side effect burden in patients who are no longer acutely ill by lowering maintenance dosages and/or decreasing the number of medications.
2. Careful dosage reduction for youths who want to discontinue medication, in the hopes of keeping them engaged in treatment and demonstrating that the medication is helping.
3. Uncertainty about the BD diagnosis, which makes the ongoing need for medication unclear.

Relapse risk with deprescribing

We have little recent research to guide us. The 1990 Strober et al paper still stands as the source for stating that most patients who stop maintenance medication relapse, often within months. Abrupt lithium discontinuation is associated with relapse rates as high as 90%, and the Collaborative Lithium Trials found rates of 60%–80% even with careful tapering. Most who relapse stabilize again once medication is restarted (Strober M et al, Am J Psychiatry 1990;147(4):457–461; Findling RL et al, J Am Acad Child Adolesc Psychiatry 2006;45(2):142–148).

Maintenance treatment vs full taper

Maintenance treatment offers substantial protection against relapse. In a large meta-analysis (n=5,462, mostly BD I), continuing medication rather than stopping it prevented 1 mood episode of any type for every 5 patients treated (number needed to treat [NNT] = 5). The NNT was 13 for preventing a depressive episode and 6 for preventing a manic, hypomanic, or mixed episode (Kishi T et al, Psychol Med 2021;51:2721–2729).

Still, not every patient relapses. In the same study, nearly half (47%) who stopped medication remained relapse-free at six months. The pattern in adults is comparable: After gradual lithium taper, roughly two-thirds relapse. Notably, longer prior lithium treatment doesn’t protect against relapse once the drug is stopped; time to recurrence is similar regardless of treatment duration (Baldessarini RJ et al, Bipolar Disord 2022;24:720–725).

Bipolar II vs bipolar I?

Bipolar II isn’t “milder” than bipolar I. Suicide risk is equivalent, and patients spend 42% more time ill (Serra G et al, Eur Psychiatry 2017;43:35–43). But bipolar II carries lower relapse risk (odds ratio 0.47) than bipolar I, with the predominant risk being depressive relapse, typically more gradual and detectable earlier (Bai C et al, J Affect Disord 2026;394(Pt B):120570). Bipolar II patients also respond better to maintenance medicine, with lamotrigine possibly the most appropriate agent. A Cochrane review (11 studies, n=2,314) found it reduced manic recurrence (relative risk [RR] 0.67) and clinical worsening (RR 0.82), with an adverse-event profile similar to placebo (Hashimoto Y et al, Cochrane Database Syst Rev 2021;9(9):CD013575).

Other risk factors for relapse

The most common risk factors in pediatric BD patients include younger onset, rapid cycling, and comorbid anxiety. Discontinuation is most debatable after a single manic episode, with 1-year recurrence in 35% of adults and 48% of youth, rising to about 58% in adults and 65% in youth by 4 years (Kessing LV et al, Bipolar Disord 2018;20(1):9–17).

Minimizing SGA exposure

SGA side effects are troubling; however, relapse risk demands careful monitoring. Here is an approach to minimizing antipsychotic exposure in bipolar I youth:

• Use SGAs for acute stabilization. Withholding them risks prolonged mania with neurodevelopmental cost.
• Initiate a mood stabilizer (lithium or lamotrigine) concurrently to reach therapeutic levels before SGA taper.
• Track relapse signs throughout tapering; maintain a low threshold for restoring prior dosage if signs emerge.
• Consider season, stressors, development (eg, rising executive function), and later health if following patients into adulthood.
• Plan SGA taper at three to six months post-remission while maintaining the mood stabilizer.
• Aripiprazole has lower metabolic burden and prolactin-lowering effects, and fewer EPS compared with lurasidone.
• Maintain robust psychosocial interventions. Family-focused therapy (FFT) and dialectical behavior therapy (DBT) have strong evidence for reducing symptoms in pediatric BD.
• Patient selection matters most. Candidates for tapering should have infrequent episodes; no suicidality history, rapid cycling, or mixed features; strong psychosocial supports; and engagement in evidence-based psychotherapy (cognitive behavioral therapy, interpersonal and social rhythm therapy, FFT, DBT).
• Adjunctive psychotherapy should be in place before and during tapering.
• Gradual taper is essential regardless of subtype. Go no faster than 25%/week—and ideally go slower.
• Residual subsyndromal symptoms strongly predict relapse, and their presence should delay or halt tapering (Bai et al, 2026; Radua J et al, Psychother Psychosom 2017;86:90–98).

Edward tapers aripiprazole over eight weeks. He starts DBT and is interested in understanding his mood symptoms. After six months, he feels in control of his moods and wants to discontinue lithium. He worries about the laboratory studies and side effects. You discuss switching to lamotrigine.

Efficacy of adjunctive psychotherapy

Residual symptoms should be treated. Adding psychotherapy to medication cuts relapse rates by about 50% compared with medication alone (Miklowitz DJ et al, JAMA Psychiatry 2021;78(2):141–150). Effective psychotherapy includes prodromal symptom self-monitoring, family-format psychoeducation and skill building, cognitive restructuring, regulating daily rhythms, and supportive emotion-focused communication (Hautzinger M et al, JAMA Psychiatry 2024;81(9):855–862).

After discussing warning signs and possible risks to his studies, Edward decides to try lamotrigine and continue supportive therapy after completing DBT. Two years on, he’s managed several brief mood disruptions with his therapist and is doing well in college.

CARLAT VERDICT

Relapse is already common on maintenance and even more common when patients stop medication. Adjunctive psychotherapy reduces relapse risk—try to include it in your dosage reduction plan. For patients stabilized on an SGA, after around six months of stability, discuss switching to more tolerable mood stabilizers for long-term maintenance. Do not attempt tapering when residual mood symptoms are present.