What Happens to First-Episode Patients After 5 Years?

Following a first episode of psychosis, patients always want to know when they can stop taking medication. Adverse effect such as weight gain or sedation may play a role, but the need to take a daily pill or receive an injection may be inconvenient or stigmatizing no matter the side effects. Existing evidence suggests a high risk of relapse during the first 5 years if medication is discontinued, up to a 5-fold compared to continuous medication treatment, although longer-term outcomes remain uncertain. Two new studies provide further evidence about outcomes longer than 5 years.

In 2003, Drs. Eric Chen, Bill Honer and collaborators in Hong Kong initiated a randomized trial with 178 first-episode patients in several clinics. To be eligible, patients had to be free of positive symptoms during at least 12 months on medication; the mean was 21.9 months. They then received either quetiapine 400 mg daily or placebo for a year or until relapse. Following the RCT, the patients returned to the community for usual clinical care.

Ten years later, the research team followed up the 178 patients; they performed a chart review on all and interviewed 142 of them. A poor outcome was defined as death by suicide, need for clozapine treatment, or persistent positive symptoms measured by the Positive and Negative Syndrome Scale (PANSS). Of 138 patients with adequate follow-up data, 110 were taking antipsychotic medication; the mean dose was 355 chlorpromazine equivalents.

Of those subjects assigned to a year of quetiapine treatment in the RCT 10 years before follow-up, 21% had a poor outcome; of those assigned to placebo in the RCT, 39% had a poor outcome. The relative risk was 1.84 (P = 0.012). Six patients died by suicide and 11 required clozapine; the incidence of these outcomes individually did not significantly differ between the groups. The investigators also found that among subjects originally assigned to medication discontinuation, i.e. placebo, relapse of psychosis during the first 2-3 years of diagnosis seemed to mediate the elevated risk of poor outcomes.

In Finland, Dr. Jari Tiihonen and his group performed a 20-year follow up on all persons hospitalized for schizophrenia for the first time during 1972–2014. Given that everyone in Finland has health and pharmacy services recorded in a national data registry, it was possible to determine who was readmitted to hospital and who filled prescriptions for antipsychotics. Based on this data, the researchers looked at whether subjects were taking medication or had been rehospitalized at various time points and then classified them as either antipsychotic users or nonusers. I suggest readers go to the original article to gain a full understanding of the methods.

Treatment failure was defined as rehospitalization or death. The table below shows the interesting finding that patients who continued antipsychotic treatment throughout the follow-up period had the lowest risk of relapse or death, but those risks rose as a function of the duration of treatment preceding discontinuation. In other words, stopping medication after several years of stability may be more associated with poor outcome than stopping it very soon after the first episode. However, compared to those who were treated continuously, the groups that discontinued treatment at any interval had a higher rate of poor outcome.

Adjusted Hazard Ratio chart

FIGURE: Adjusted hazard ratios as a function of duration of antipsychotic use prior to discontinuation.

The number of deaths was relatively small, but available data allowed the calculation of hazard ratios in 3 matched groups: those who discontinued antipsychotic treatment within the first year, those who remained on antipsychotics throughout, and those who did not use antipsychotics.  In 3057 subjects, 91 deaths occurred; compared with continuous antipsychotic users, nonusers had a 214% higher risk of death (hazard ratio, 3.14; 95% CI, 1.29–7.68), and those who discontinued within a year had a 174% higher risk of death (hazard ratio, 2.74; 95% CI,1.09–6.89).

These studies have important limitations, given their retrospective nature and the lack of details about important outcomes related to function and comorbidities. Nonetheless, they bring new understandings to the role of antipsychotic therapy after the first few years of psychotic illness: patients who go untreated have a higher risk of remaining psychotic and of dying. But the finding that discontinuation after 5 or more years of antipsychotic therapy is highly associated with relapse suggests that stopping medication in chronic patients is risky.


Hui CLM, Honer WG, Lee EHM, Chang WC, et al. Long-term effects of discontinuation from antipsychotic maintenance following first-episode schizophrenia and related disorders: a 10 year follow-up of a randomised, double-blind trial. Lancet Psychiatry. 2018;5(5):432-442.

Tiihonen J, Tanskanen A, Taipale H. 20-Year nationwide follow-up study on discontinuation of antipsychotic treatment in first-episode schizophrenia. Am J Psychiatry. Published online Apr 6, 2018: