Relapse predictors for discontinuing and continuing maintenance medication in remitted first-episode psychosis patients

Abstract

After a first-episode psychosis, patients inevitably face a clinical dilemma regarding whether to continue with maintenance medication after a period of remission. This decision has to be weighed against the long-term medication side effects and the risk of relapse. In this study, we aim to identify potential predictors for relapse among patients who have continued with maintenance medication, as well as those who have discontinued with maintenance medication. This is a double-blind randomized placebo-controlled study. Patients who were remitted from a first-episode of schizophrenia or non-affective psychosis (DSM-IV) and had remained well on maintenance medication for at least 1 year were recruited into this study. Eligible patients were randomized to either quetiapine (400mg/d) or placebo for 12 months to detect a relapse. Relapse was defined as re-emergence of definitive positive symptoms. Baseline clinical and cognitive variables were measured as potential predictors for relapse. Univariate and multivariate Cox-proportional hazards models were applied to analyze significant predictors in the maintenance medication and the placebo groups. A total of 178 patients were recruited. In the placebo group, significant multivariate relapse predictors were smoking (relative risk 4.317, 95% CI = 1.012–18.417), negative symptoms (relative risk 1.030, 95% CI = 1.008–1.053) and verbal fluency (relative risk 0.880, 95% CI = 0.807–0.960). In the medication group, significant multivariate relapse predictors were pre-morbid functioning (interest) (relative risk 1.641, 95% CI = 1.215–2.216), PANSS (general psychopathology) (relative risk 1.548, 95% CI = 1.089–2.200), logical memory at immediate recall (relative risk 0.879, 95% CI = 0.782–0.988) and neurological soft signs (disinhibition) (relative risk 3.423, 95% CI = 1.585–7.390). Importantly, predictors identified in the two groups help provide information about the characteristics of the patients who are more prone to relapse. Future intervention strategies for relapse prevention can be based on the current findings.