Treatment resistance in schizophrenia: a meta-analysis of prevalence and correlates

Objective: To define the prevalence and correlates of treatment-resistant schizophrenia (TRS) through a systematic review and meta-analysis.Methods: Following PRISMA criteria, an electronic search was performed in PubMed and EMBASE through May 17, 2022. All study designs that assessed a minimum of 20 schizophrenia-spectrum patients, providing data on TRS prevalence or allowing its calculation, were included. Estimates were produced using a random-effects model meta-analysis.Results: The TRS prevalence across 50 studies (n=29,390 subjects, mean age=38<middle dot>3 +/- 6<middle dot>2 years, males=64<middle dot>8 +/- 12<middle dot>1%, mean illness duration=14<middle dot>3 +/- 5<middle dot>3 years) was 36.7% (95%CI=33.1-40.5%; I2=97.3%, p<0.0001). The prevalence ranged from 22% (95%CI=18.4-25.8%) in first-episode to 39.5% (95%CI=32.2-47.0%) in multiple-episode samples (Q=18.27, p<0.0001). Primary treatment resistance was 23.6% (95%CI=20<middle dot>5-26<middle dot>8%) vs. 9.3% (95%CI=6<middle dot>8-12<middle dot>2) for later-onset/secondary (>= 6 months after initial treatment response). Longer illness duration and recruitment from long-term hospitals or clozapine clinics were associated with higher prevalence estimates. In meta-regression analyses, older age and poor functioning predicted greater TRS. When including only studies with lower bias risk, the TRS prevalence was 28.4%.Conclusions: Different study designs and recruitment strategies accounted for most of the observed heterogeneity in TRS prevalence rates. The results point to early-onset and later-onset TRS as two separate disease pathways requiring clinical attention.