Antipsychotic treatment in schizophrenia: balancing relapse prevention and functional recovery

Relevance

Schizophrenia spectrum and other psychotic disorders remain among the most severe and costly psychiatric conditions worldwide, both in terms of individual suffering and societal expenditure. These disorders are heterogeneous in both presentation and course: some patients recover after a single psychotic episode, while others require repeated or even continuous hospitalization. Schizophrenia, the prototypical psychotic disorder, is characterized by positive symptoms (hallucinations, delusions), negative symptoms (amotivation, social withdrawal), and associated cognitive impairments.

Importantly, long-term prognosis is not determined solely by positive symptoms. Negative and cognitive symptoms strongly predict functional outcomes, such as independent living, employment, and social integration¹. While antipsychotics effectively reduce positive symptoms, their efficacy for negative and cognitive symptoms is limited². Moreover, side effects of prolonged antipsychotic treatment can hinder recovery, leading to patient ambivalence toward long-term use³.

The clinical challenge is therefore twofold: it involves not only managing relapse risk, but also redefining what counts as “successful treatment”, with functional and social recovery at the forefront. This tension has fueled debate about if, when, and how dose reduction or discontinuation of antipsychotics might be feasible. In this light, the comprehensive review by Moncrieff and Horowitz is highly relevant, as it compels us to reconsider where the balance should lie⁴.

What the evidence on dose reduction and relapse tells us

Meta-analyses and large-scale trials consistently show that dose reduction or discontinuation of antipsychotics substantially increases relapse risk. In a landmark meta-analysis, patients on maintenance antipsychotics were two to three times less likely to relapse compared with those withdrawn from medication⁵. Similar findings emerged in the MEFISOS study of first-episode patients⁶, in the pragmatic RADAR trial of patients with recurrent psychosis⁷, and in a study in treatment-resistant patients (TRS) in which a minimum dose of 5 mg haloperidol-equivalents was used⁸.

This underscores the protective effect of sustained pharmacotherapy across patient populations, and that slow and gradual dose reduction in patients with moderate or severe illness does not necessarily protect against relapse in the long term. The end-dose is possibly more pivotal. In meta-analyses, relapse risk in chronic patients rises sharply with dose reduction below 3–5 mg haloperidol-equivalents per day^9,10. This threshold appears robust in long-stay patients and may represent a pragmatic compromise: minimizing side effects and cognitive blunting while still providing a pharmacological safeguard against relapse. For first-episode patients, doses under 1–2 mg probably substantially increase relapse risk.

However, the timing of relapse may vary. During long follow-up, relapse rates in the MEFISOS study tended to equalize between dose-reduction and maintenance conditions after 3 years, raising the possibility that relapse risk diminishes over longer follow-up once the most fragile patients have relapsed¹¹. Still, in all studies, most patients tolerated dose reduction in the long run, and even a minority were able to discontinue antipsychotics, at least for a period of time.

Looking beyond relapse

Although dose reduction increases relapse risk, the long-term picture appears more nuanced. In the 7-year follow-up of the MEFISOS trial, relapse differences between groups had diminished after 3 years. Crucially, after 7 years, more than twice as many patients in the reduction condition achieved functional remission and recovery, despite no significant differences in symptomatic remission. This suggests that functional gains may only emerge after prolonged follow-up once dosing has stabilized and vulnerable individuals have already relapsed.

This delayed recovery trajectory aligns with Harrow’s 20-year naturalistic study, which found that some patients off antipsychotics experienced more or longer periods of functional recovery compared with those on continuous medication¹². While relapse was more common without medication, long-term recovery was feasible for a subset of patients with favorable prognostic factors such as good premorbid adjustment, low vulnerability to anxiety, and preserved neurocognition.

Thus, the evidence indicates a paradox: dose reduction or discontinuation raises short-term relapse risk but may confer advantages in functional recovery over the long term, at least for some patients.

Unfortunately, dose reduction in the RADAR study and in the TRS study did not lead to greater patient satisfaction, improved quality of life, or better social functioning. Doses were reduced from a median of 6 (IQR 4–9) to 4 (IQR 1.5–8) (RADAR) and from a mean of 18 to 5 mg HE/day (TRS). Was the end-dose too high, do recurrent-psychosis and TRS groups differ from first-episode patients, or is functional recovery only to be expected in the long term?

Key messages

From the evidence reviewed, several points stand out:

• Relapse risk substantially increases following dose reduction. Across first-episode, recurrent, and treatment-resistant patients, dose reduction approximately doubles or triples the chance of relapse^5–11.

• A small but substantial group can discontinue or reduce dosage. Especially among first-episode patients, some are able to stop or substantially reduce their dose. Therefore, dose reduction can be considered in patients with schizophrenia spectrum disorders as long as they showed no dangerous behavior during psychosis. If patients are worried about side effects or feel antipsychotics hamper their recovery, dose reduction might encourage treatment continuation, which still protects them against becoming (more) psychotic. It is important to note that patients may also relapse while on maintenance treatment. Moreover, patients can recover from relapse. Restarting antipsychotic treatment is possible, although some reports describe reduced responsiveness after reinstating therapy¹³.

• Relapse differences diminish over time. Long-term follow-up suggests that the relapse gap between dose-reduction and maintenance groups narrows after about three years, possibly because the most vulnerable patients relapse early in the reduction process¹¹.

• Functional recovery may occur later. While dose reduction does not immediately improve functioning, long-term evidence indicates that functional recovery may only become visible after years of stabilization, particularly in first-episode patients with favorable prognostic factors^11,12.

• Gradual dose reduction has practical advantages. Even if prolonged dose reduction does not reduce relapse risk, a gradual approach may still help mitigate withdrawal symptoms, allow for prompt dose reinstatement if needed, and help patients find a lower effective dose with a more favorable side-effect profile, potentially reducing dose-related complications.

• Patient preferences matter. Regardless of whether they view antipsychotic treatment positively or negatively, many patients wish to reduce or even discontinue their medication. Clinicians must inform patients about the risks and potential benefits and guide them carefully through this process.

Why this matters for practice

The data collectively suggest that a one-size-fits-all approach to antipsychotic dosing is inadequate. Maintenance treatment remains essential for many patients, but not for all. For a subset of stable patients, preferably with strong prognostic indicators, cautious and individualized dose reduction may support long-term recovery without unacceptable risk.

The early years of dose reduction appear to be the most precarious, requiring close monitoring, gradual tapering, and readiness to reinstate medication if relapse occurs. For many patients, a protective end-dose seems to exist that reduces the risk of relapse.

Future research should focus on identifying biomarkers or clinical predictors of who can tolerate dose reduction, optimizing tapering strategies, and clarifying the timeline of functional recovery beyond symptom control.

Concluding reflections

The treatment of schizophrenia remains a careful balancing act between preventing relapse and fostering recovery. The evidence is clear that antipsychotic dose reduction increases relapse risk—by a factor of two to three compared with maintenance—but that this risk may diminish over time, and that functional recovery may emerge only after prolonged stabilization. These dynamics underscore the need for individualized treatment, careful long-term follow-up, and a redefinition of success in psychosis treatment to encompass not just symptom management, but also the restoration of autonomy, quality of life, and social functioning.

Competing interests

The author declares no competing interests.