17.1 PREDICTORS OF RELAPSE IN FIRST EPISODE PSYCHOSIS PATIENTS IN REMISSION

Abstract

Background: Patients in their first psychotic episode often respond to antipsychotic medications, with rates of response up to 80%. The OPTIMISE project was a large three-phase multicenter study, which included 446 patients in their first psychotic episode, who received open-label amisulpride, and/or double-blind amisulpride or olanzapine for 10 weeks. Of the patients recruited, 63% went into remission, as defined by the Andreasen criteria. This present study attempted to identify predictors of later relapse in these initial remitters during a follow-up period of up to a year and a half. Method(s): Relapse was defined by changes of PANSS scores compared to time of remission at any time during the follow-up period using one of two different criteria: (1) >= 20% increase in total PANSS score from the time of remission, or (2) no longer meeting one or more of the Andreasen criteria. Potential predictors tested in the analyses were age, sex, duration of untreated psychosis, cannabis use, baseline PANSS scores, and insight, as measured by the drug attitude inventory (DAI) scale and the Knowledge of Psychotic Illness (KPI) Questionnaire. Univariate analyses examining the effect of each of these potential predictors on risk for relapse were performed. Those variables who were significant in the univariate analyses were included in a multivariate logistic regression. Result(s): When using the 20% increase of PANSS criterion for relapse, a total of 67 out of 446 patients (15%) were included in the analyses, using the criterion of no longer meeting Andreasen criteria identified 101 out of 446 (23%) patients. When using the relapse criterion of 20% increase in PANSS, univariate analyses showed a significant effect on relapse, with lower baseline score increasing risk of relapse: for baseline positive PANSS score (OR=0.918, 95% CI: 0.868-0.971, p=0.003), baseline negative PANSS score (OR=0.926, 95% CI: 0.883-0.972, p=0.002), baseline general psychopathology PANSS score (OR=0.943, 95% CI: 0.911-0.976, p=0.001), and cannabis use (OR=2.284, 95% CI: 1.116-4.678, p=0.024). Multivariate analyses showed significant effects on relapse only for cannabis use (OR=2.011, 95% CI:1.024-4.087, p=0.021). Both univariate and multivariate analyses were not significant when using the relapse criterion of not meeting one or more Andreasen criteria. Conclusion(s): Cannabis use was the only significant clinical factor that increased risk for relapse. These findings are similar to previous studies which showed a strong correlation between cannabis use and earlier and more frequent relapses. Patients and family members should be made aware of this finding and future studies should investigate the effect of cannabis cessation programs on relapse rates.