Rapid relapse after paliperidone discontinuation reflects severity, not withdrawal

Design

Individual participant data meta-analysis of five double-blind discontinuation RCTs

Population

417 adults with schizophrenia or schizoaffective disorder whose illness relapsed

Primary outcome

Relapse trajectory class by discontinuation versus continuation

Effect

Rapid relapse 20% vs 11% (LAI, p=0.12) and 27% vs 26% (oral, p=0.95)

An individual participant data meta-analysis (Lancet Psychiatry, 24 August 2026) used five double-blind discontinuation trials of oral and long-acting injectable paliperidone, analysing the 417 participants whose illness relapsed after stabilisation and randomisation to placebo or continued drug.

Modelling of symptom scores found two relapse trajectories: rapid and delayed. Rapid relapse was not more common after discontinuation than on continued treatment - in the injectable trials 20% versus 11% (p=0.12), in the oral trials 27% versus 26% (p=0.95). Symptom profiles at relapse did not differ by whether the drug had been stopped. Rapid relapse was linked to higher baseline PANSS scores.

This matters because a fast relapse after stopping is often read as a withdrawal or rebound effect of the drug itself. These data suggest it more often reflects how unwell the person was to begin with. The authors note that screening scores in trials may be shaded to meet entry thresholds, which could also contribute.

The practical consequence is to stratify. Patients with more severe illness at baseline are the ones to reduce most cautiously and watch most closely. This does not show that stopping is safe - relapse is more common after discontinuation overall - and it covers one drug.

• Use baseline illness severity to decide how closely to monitor during deprescribing
• Reserve the slowest reductions and most frequent reviews for patients with more severe illness
• Do not assume a fast relapse after stopping is a withdrawal effect that will settle
• Remember the finding is from paliperidone trials and may not hold for every antipsychotic

Why it matters

It challenges the idea that early relapse after stopping is mainly rebound, which shifts attention to who is at risk.

Don't overread it

Based only on paliperidone trials and only on those who relapsed - it does not make discontinuation any safer.

The statistics, in plain English

The comparisons of rapid relapse between stopped and continued treatment have p-values of 0.12 and 0.95: no detectable difference. In the injectable trials the numbers on continued treatment were small (8 of 74), so a modest difference could have been missed. The study only looked at those who relapsed; it says nothing about how many relapse overall.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.