- Design
- Individual participant data meta-analysis of 5 randomised discontinuation trials
- Population
- 417 people with schizophrenia who relapsed, on oral or injectable paliperidone
- Primary outcome
- Relapse trajectory class by discontinuation status
- Effect
- Rapid relapse 20% vs 11% (injectable, p = 0.12); 27% vs 26% (oral, p = 0.95)
Researchers pooled individual participant data from five double-blind paliperidone discontinuation trials (271 on long-acting injectable, 146 oral) and modelled symptom trajectories before relapse. Two patterns emerged: rapid and delayed.
Rapid relapse was not significantly more common after switching to placebo than during continued treatment — 20% vs 11% in the injectable trials (p = 0.12) and 27% vs 26% in the oral trials (p = 0.95). Symptom profiles at relapse did not differ by discontinuation. Those with rapid relapse had higher baseline PANSS scores.
This argues against the idea that rapid relapse after stopping is commonly a supersensitivity or withdrawal effect of the drug. It points instead to baseline severity as the thing to watch when deprescribing.
- Record baseline severity before any dose reduction; it predicted rapid relapse.
- Monitor more closely in the first weeks for those with more severe illness at stabilisation.
- Taper rather than stop abruptly — this study does not test tapering but does not weaken the case for it.
- Agree an early-warning plan with the patient and family before reducing.
Why it matters
It challenges the assumption that early relapse after stopping is mainly caused by the drug withdrawal itself.
Don't overread it
Only paliperidone trials were available, and the injectable comparison was underpowered (p = 0.12).
The statistics, in plain English
A p-value of 0.12 in the injectable trials means the difference (20% vs 11%) could be chance, but the study was small and a real difference is not ruled out. Latent class models find patterns in data; they describe groups rather than prove why people fall into them.
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.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for psychiatry, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free