WHAT IS KNOWN ON THE SUBJECT?: Relapse rates are high among patients who have experienced first-episode psychosis (FEP). Psychotic relapses are associated with worse quality of life and poorer functionality of the FEP patient. The use of psychoactive substances, non-adherence to drug treatment, and high expressed emotion (EE) are notable predictors of relapse after the FEP. Although some studies have suggested that psychotic relapse may be associated with a family environment with high levels of emotional over-involvement (EOI), this finding is still inconsistent across different cultures. EE specific components must be evaluated and interpreted according to the context of cultural norms. There is a scarcity of studies on the role of depression in the occurrence of relapses after the FEP, and the results remain uncertain. WHAT THE PAPER ADDS TO EXISTING KNOWLEDGE?: This study explored the predictors of psychotic relapses in Brazilian patients who experienced FEP. Our results indicate that 29.2% of the patients relapsed after the FEP. Patients diagnosed with depression and high-EOI in the family environment were predictors of psychotic relapses in this population. This study expands knowledge about the cultural specificity of EOI and the role of depression in psychotic relapse. WHAT ARE THE IMPLICATIONS FOR PRACTICE?: Nursing professionals must consider the implications of the family environment and depression in the course of psychosis. Family interventions and the appropriate treatment of depression are important for improving the prognosis of FEP patients. ABSTRACT: Introduction Psychotic relapse may be associated with relatives' high emotional over-involvement (EOI) and with a diagnosis of major depressive episode (MDE) among first-episode psychosis (FEP) patients, but the results are still inconsistent across different cultures. Aim Evaluate the predictors of relapse in FEP patients. Method Prospective cohort study with 6-month follow-up conducted with 65 dyads of patients and relatives from an early intervention unit in Brazil. At the baseline interview, relatives answered to a sociodemographic data form and to the Family Questionnaire. Patients provided sociodemographic and clinical data and answered the Measurement of Treatment Adherence; the Alcohol, Smoking and Substance Involvement Screening Test; the Severity of Dependence Scale to assess cannabis dependence, and the MDE module of the Mini-International Neuropsychiatric Interview. Psychotic relapses were evaluated using items from the Brief Psychiatric Rating Scale. The data were analysed using multiple logistic regression. Results 29.2% of the patients presented at least one psychotic relapse. High-EOI and MDE were predictors of psychotic relapses. Discussion Our findings expand the knowledge about the cultural specificity of EOI and the role of depression in psychotic relapse. Implications for practice Family nursing interventions and the appropriate treatment of MDE must be considered in the care of FEP patients.