Abstract
We studied 61 patients with schizophrenia spectrum disorders who discontinued medication after achieving symptomatic remission. Over 3 years, relapse rates were significantly higher in those not meeting full recovery (p = 0.006) or remission (p < 0.001) criteria, with rates up to twice as high. Significant differences between relapsed and maintained groups included age at onset (p = 0.004), age at discontinuation (p = 0.009), and proportions meeting full recovery (p = 0.001) or remission (p = 0.003). Univariate Cox regression identified older age of onset (p = 0.038), lack of full recovery (p = 0.008) or remission (p = 0.001), and higher positive symptom score (p = 0.018) as predictors of relapse. In multivariate analysis, only full remission remained significant (p = 0.002). Our findings suggest that in making decisions about discontinuation, applying more strict approaches—specifically full recovery or remission criteria and detailed assessment of positive symptoms—is critical.
Introduction
For patients who have achieved symptomatic remission from a first episode of psychosis, the decision to continue antipsychotics indefinitely is complex. While long-term use carries significant metabolic and neurological risks, premature discontinuation often leads to relapse, which can be neurotoxic and socially devastating. Current clinical guidelines for tapering are often criticized for being vague. This study addresses a critical gap by identifying predictors of relapse measured specifically at the point of discontinuation, utilizing more rigorous definitions of "recovery" and "remission" than previous trials.
Methods and Participant Flow
Researchers conducted a prospective analysis of 61 individuals from two South Korean cohorts: the Korea Early Psychosis Study (KEPS) and the Recovery and Relapse in Early Psychosis (R&R) study. Patients who met symptomatic remission criteria and wished to stop were tapered off medication over 3–6 months. Within 1–2 months of complete discontinuation, patients underwent a battery of assessments, including the Positive and Negative Syndrome Scale (PANSS) and cognitive functioning tests. Patients were then monitored for an average of 56 months to track relapse.
The study distinguished between "Full" and "Partial" states:
- Full Remission: Scoring ≤ 2 on eight specific PANSS items for at least 6 months.
- Partial Remission: Scoring ≤ 3 on those same items.
- Full Recovery: Meeting remission criteria plus achieving high social/occupational functioning (SOFAS score ≤ 71) for at least one year.
Key Results and Predictors
The cumulative relapse rate for the entire group over three years was 67.9%, with a median time to relapse of 369 days. However, the risk was not distributed equally across the cohort.
The Importance of Strict Criteria
Patients in "partial" remission relapsed at nearly twice the rate of those in "full" remission. In multivariate Cox regression analysis, Full Remission was the only variable that remained a significant independent predictor of staying relapse-free (Hazard Ratio = 3.41, p = 0.002). Similarly, those meeting "Full Recovery" criteria had a 54.5% relapse rate at 3 years, compared to 85.0% for those in partial recovery.
Clinical Red Flags
- Age of Onset: Interestingly, an older age of onset predicted a higher relapse risk. While older onset is often a positive general prognostic factor, the authors suggest younger patients may possess higher neuroplasticity or resilience during medication withdrawal.
- Subtle Positive Symptoms: Higher scores on positive symptoms (such as mild suspiciousness or ideas of reference) were significant predictors of relapse in univariate models, even when total PANSS scores remained low.
- Cognition: While relapsed patients showed cognitive impairments compared to healthy controls, cognitive scores at the time of discontinuation did not reliably predict which individuals would eventually relapse.
Discussion and Clinical Implications
These findings suggest that the common clinical practice of using Andreasen’s remission criteria (scores of ≤ 3) as a "green light" for discontinuation may be insufficient. A more stringent threshold (scores of ≤ 2) provides a significantly better safety net.
Clinicians should be particularly cautious with patients who have an older age of onset or who exhibit even subtle positive symptoms. The study also highlights the necessity of long-term monitoring; the median relapse time of approximately one year suggests that the risk remains high well beyond the initial months of being drug-free. Future assessments might benefit from more sensitive tools, such as the Psychotic Symptom Rating Scales (PSYRATS), to detect the subtle symptoms that standard scales might miss before a patient stops medication.
Summary updated