Prevalence and predictors of treatment-resistant schizophrenia in a tertiary hospital in Northeast Brazil

Objective: To investigate epidemiological factors related to treatment-resistant schizophrenia (TRS) in Northeast Brazil, a region where data about mental health are still scarce. Methods: This retrospective cross-sectional study included all patients with schizophrenia currently receiving treatment at the outpatient psychiatric clinic of a tertiary hospital in Northeast Brazil. They were divided into TRS and treatment-responsive groups, and epidemiological characteristics of both groups were compared. A logistic regression model investigated factors related to treatment resistance. Results: Two hundred and five patients were included, 155 treatment-resistant and 50 treatment-responsive. The TRS group had higher use of benzodiazepines (36.1 vs. 18%, p = 0.017) and antiepileptics (36.8 vs. 8.0%, p < 0.001), antipsychotic polypharmacy (28.6 vs. 8%, p = 0.003) and suicide attempts (35.6 vs. 20%, p = 0.04). Age at onset was younger (19.7 +/- 7.3 vs. 24.6 +/- 8.6 years, p = 0.001) and CGI was higher in TRS (3.72 +/- 1.00 vs. 3.16 +/- 1.00, p = 0.001). In logistic regression, being married was a protector (odds ratio [OR] = 0.248, 95% confidence interval [95%CI] 0.091-0.679, p = 0.007) and younger age at onset was a predictor (OR = 1.076, 95%CI 1.034-1.120, p < 0.001) of treatment resistance. Conclusion: Early onset of disease was associated with more treatment resistance, while being married with less resistance. Clinicians should identify early predictors of resistance in order to reduce unfavorable outcomes.