Recovery from first-episode psychosis (FEP) often brings a sense of relief, but it also raises a challenging question. After a period of experiencing hallucinations, delusions, paranoia or disorganised thinking, medication can help restore stability. Sleep improves. Thinking becomes clearer. Many people return to school or work, rebuild relationships, and start feeling like themselves again. Then, just as symptoms subside, doctors and patients face an important decision: should antipsychotic medication be continued, reduced or stopped?
Traditional medical advice has favoured continuing antipsychotic medication as these drugs are highly effective at reducing symptoms of acute psychosis. Research consistently shows that maintaining pharmacological treatment lowers the risk of relapse in the short term (Kishi et al., 2019), and clinical guidelines recommend continuing treatment for at least 1-2 years after remission (National Institute for Health and Care Excellence, 2014).
However, antipsychotic side effects like weight gain, sedation and metabolic changes can impact energy, motivation, and physical health (Leucht et al., 2013). These concerns matter deeply to individuals rebuilding their lives after psychosis. Beyond focusing on relapse, there is a growing body of work examining long-term outcomes such as functioning, recovery, and quality of life.
Sommer et al. (2026) explored the short and long-term effects of reducing or stopping versus staying on medication following recovery from FEP. The findings challenge the assumption that symptom control alone defines a successful recovery.
Recovery after a first-episode psychosis can bring relief, but it also raises a difficult question: once symptoms improve, should antipsychotic medication be continued, reduced, or stopped?
Methods
From 2017 to 2023, researchers followed patients in remission from FEP who had been stable on antipsychotic medication for 3-6 months. Those showing dangerous behaviour or requiring involuntary treatment were excluded. The sample included patients with different psychosis diagnoses (e.g., schizophrenia, schizoaffective disorder), which vary in symptoms and recovery trajectory, making it harder to isolate true treatment effects
Participants were randomly assigned in a 1:1 ratio to one of two conditions for a 6-month intervention period in a single-blind pragmatic randomised controlled trial (RCT):
- Maintenance (MT) Group (n = 179): Continued antipsychotic treatment with ≤25% dose reduction.
- Dose Reduction or Discontinuation (DRD) Group (n = 168): Gradual tapering toward zero, with minimum dose reductions of 25, unless symptoms returned.
All participants remained in the study regardless of adherence, reflecting real-world clinical practice. Follow-up extended to 4 years to observe longer-term effects.
Primary outcome: Self-reported personal and social functioning (World Health Organization Disability Assessment Schedule 2.0, WHODAS-2; 36 items).
Secondary outcomes: Researcher-rated global assessment of functioning (GAF), symptom severity (Positive and Negative Syndrome Scale [PANSS]), and health-related quality of life (Visual Analog Scale of the EuroQol 5-Dimension Health-Related Quality of Life scale [EQ-5D-VAS]), relapse (clinician-reported), and safety and tolerability outcomes (adverse events, serious adverse events, self-harm, aggressive behaviour, police contact, adverse effects and body mass index [BMI]).
Results
The study included 347 individuals (mean age 27.9 years). By the end of the intervention period (6-month mark), Dose Reduction or Discontinuation (DRD) participants were on lower doses than Maintenance (MT) Group participants (3.8 mg vs. 6.9 mg), with protocol adherence at 65.5% (n = 133) versus 27.4% (n = 108).
Short term (12 Months): DRD participants experienced poorer outcomes, with higher relapse rates at 12 months (OR = 2.84, 95% CI 1.08 to 7.66) and lower quality of life (EQ-5D-VAS: β = −3.31; 95% CI, −6.34 to −0.29; P = .03) at 6 months. Safety outcomes were concerning, with higher mortality rates in the DRD group (n = 5 vs. 1), including more deaths by suicide (n = 3 vs. 1).
Long term (36-48 Months): No significant differences in health-related functioning and symptom severity emerged until 36-months, with effects increasing at 48 months. By this stage, DRD participants showed better researcher-rated functioning (GAF: β = 6.13; 95% CI, 2.03 to 10.22; P = .003) and lower symptom severity (PANSS: B = −3.02) with benefits emerging earlier in women than men. However, these improvements were not reflected in patient-reported functioning, and a steep increase in BMI was observed in the DRD group. Notably, by this point, antipsychotic dosage was comparable across both groups, as some DRD participants had resumed medication after relapse, while some MT participants had independently reduced their doses.
Reducing antipsychotic dosage after a first episode of psychosis could lead to better long-term outcomes, but not without some risks along the way.
Conclusions
Overall, there was no significant difference in patient-rated functioning between the two groups. In the 1st year, the DRD group experienced a higher relapse risk and lower quality of life. By the 3-year mark, however, researcher-rated functioning and symptom severity improved. Since doses were similar from 12 months onwards, improved functioning could not be attributed to lower medication alone. Instead, the researchers speculated that it reflected the “empowering and insightful” learning experience of navigating early dose reduction.
Short-term risks, long-term rewards? Dose Reduction or Discontinuation showed early setbacks, but longer-term functioning told a more hopeful story.
Strengths and limitations
The 1:1 randomisation helped reduce selection bias and minimise baseline differences between groups, strengthening internal validity. Participants who did not adhere to their allocated condition were retained and analysed using an intention-to-treat approach, reducing the risk of overestimating treatment effects and better reflecting real-world clinical practice, where non-adherence to tapering is common (~70%).
Another strength of the study was that the 48-month follow-up enabled the assessment of both short- and long-term outcomes, which was particularly valuable given that functional benefits in the DRD group only emerged at 36 and 48 months. Overall, the findings have direct clinical relevance, emphasising shared decision-making on antipsychotic tapering after FEP remission, a common and often contentious issue in psychiatric care.
The authors acknowledge their research has limitations. Let’s consider some together:
- The trial was single-blinded, with researchers being aware of group allocation. Significant differences in functioning were only observed in researcher-rated measures, possibly due to observer bias exaggerating treatment effects in the DRD group.
- Participant’s ethnicity was not reported, posing a potential confounder given the known variation in psychosis prevalence and presentation across ethnic groups.
- The sample was predominantly male (69.5%), and improvements in symptoms and overall functioning were observed earlier in women, suggesting potential sex differences in treatment response. A more balanced sample would have facilitated sex-specific analyses.
- Most people with psychosis have comorbidities, which this study did not control for, limiting generalisability. Participants may have had different baseline features (e.g. polypharmacy), which could have influenced responses to tapering.
- Heterogeneity in the antipsychotics was standardised using olanzapine-equivalent doses, an approach that assumes pharmacological equivalence despite important differences in pharmacodynamics, efficacy, and side-effect profiles. This limits internal validity and the ability to extrapolate findings to specific drugs.
- The findings are limited to FEP and cannot be generalised to multi-episode psychosis, where maintenance dosage may be more effective. Conversely, for someone experiencing FEP, prolonged medication use may not be necessary.
- Improvement in functioning appeared only in researcher-rated measures, with no significant difference in patient-rated outcomes, raising questions about how “functioning” is defined. Functioning extends beyond clinical outcomes like symptom reduction to include self-efficacy, community participation, and broader contextual factors, which are not fully captured by tools like the GAF (global assessment of functioning scale). Given the greater adverse events in the DRD group, it is unclear whether findings reflect meaningful patient-level improvement.
- An unexpected finding was that BMI increased in the DRD group, contrary to the expectation that dose reduction would lower metabolic side effects. Potential mechanisms or confounders were not explored, limiting confidence in conclusions about the metabolic benefits and overall tolerability of dose reduction or discontinuation.
A robust, real-world trial with long follow-up offers valuable clinical insights; potential biases, missing patient perspectives, and limited generalisability mean the findings should be interpreted with caution.
Implications for practice
As antipsychotic side effects can significantly impact daily life, including major life decisions (e.g., family planning), these findings are important for supporting patients to make informed choices. Antipsychotic use may also contribute to feelings of stigma and emotional burden. As these experiences vary between individuals, the findings support a personalised approach to treatment. Clinicians should engage in iterative formulation, risk assessment, and collaborative care planning, including routine medication reviews and proactive conversations about treatment preferences and potential side effects.
Future research should focus on identifying which individuals are most likely to benefit from dose reduction or discontinuation. NICE guidelines on antipsychotic monitoring in primary care are largely focused on physical health, so expanding these to support primary care in monitoring psychotic symptoms and identifying early signs of relapse would be beneficial. Additionally, NHS England could facilitate GP training in antipsychotic management and safe dose reduction in FEP.
The key message of this study is not “stop medication” or “stay on medication”. Instead, it highlights that care should be flexible, collaborative, and person-centred. Some people may need long-term medication to stay well, while others, with careful monitoring, may do better on lower doses or none. What matters most is that decisions are shared by patients and doctors, based on what works best for the individual.
Antipsychotic management in first-episode psychosis should be individualised and collaborative, with clinicians proactively discussing side effects, treatment preferences, and dose reduction options.
Statement of interests
As MSc Clinical Mental Health Sciences students at UCL, we want to clarify that while some individuals involved in the discussed research are UCL faculty, we have no direct involvement in the study, its review, or publication. This blog was created independently for our coursework and reflects our interpretation. We have no conflicts of interest beyond our academic relationship with the study.
Contributors
Thanks to the UCL Mental Health MSc students who wrote this blog from the Saleh Student Group: Sonali Garg, Thelma Kokroko, Valerie Muzraeva, Benedetta Rondelli, Umut Sadet, and supported by Connor Clarke.
UCL MSc in Mental Health Studies
This blog has been written by a group of students on the Clinical Mental Health Sciences MSc at University College London. A full list of blogs by UCL MSc students can be found here.
We regularly publish blogs written by individual students or groups of students studying at universities that subscribe to the National Elf Service. Contact us if you’d like to find out more about how this could work for your university.
Edited by
Dr Dafni Katsampa
Links
Primary paper
Iris Sommer, Franciska de Beer, Shiral Gangadin, Lieuwe de Haan, Wim Veling, Nico van Beveren, Nynke Boonstra, Bram-Sieben Rosema, Jim van Os, Martijn Kikkert, Sanne Koops, Jort Noorman, Frederick Thielen, Ben Wijnen & Marieke Begemann (2026). Early Dose Reduction or Discontinuation vs Maintenance Antipsychotics After First Psychotic Episode Remission: A Randomized Clinical Trial. JAMA Psychiatry,83(1), 68–73.
Other references
Kishi, T., Ikuta, T., Matsui, Y., Inada, K., Matsuda, Y., Mishima, K., & Iwata, N. (2019). Effect of discontinuation v. maintenance of antipsychotic medication on relapse rates in patients with remitted/stable first-episode psychosis: a meta-analysis. Psychological Medicine, 49(5), 772–779.
Leucht, S., Cipriani, A., Spineli, L., Mavridis, D., Örey, D., Richter, F., Samara, M., Barbui, C., Engel, R. R., Geddes, J. R., Kissling, W., Stapf, M. P., Lässig, B., Salanti, G., & Davis, J. M. (2013). Comparative efficacy and tolerability of 15 antipsychotic drugs in schizophrenia: a multiple-treatments meta-analysis. The Lancet, 382(9896), 951–962.
National Institute for Health and Care Excellence. (2014). Psychosis and schizophrenia in adults: Prevention and management(NICE guideline CG178). NICE.
