Effects of Benzodiazepine Exposure on Real-World Clinical Outcomes in Individuals at Clinical High-Risk for Psychosis

BACKGROUND: Animal models indicate GABAergic dysfunction in the development of psychosis, and that benzodiazepine (BDZ) exposure can prevent the emergence of psychosis-relevant phenotypes. However, whether BDZ exposure influences the risk of psychosis in humans is unknown. METHODS: This observational-cohort study used electronic health record data from 818 individuals at clinical high-risk for psychosis (CHR-P) to investigate whether BDZ exposure (including hypnotics e.g., zopiclone) reduces the risk of developing psychosis and adverse clinical outcomes. Cox proportional-hazards models were employed in both the whole-unmatched sample, and a propensity score matched (PSM) subsample. RESULTS: 567 CHR-P individuals were included after data cleaning (105 BDZ-exposed, 462 BDZ-unexposed). 306 (54%) individuals were male, and the mean age was 22.3 years (SD 4.9). The BDZ-exposed and BDZ-unexposed groups differed on several demographic and clinical characteristics, including psychotic symptom severity. In the whole-unmatched sample, BDZ exposure was associated with increased risk of transition to psychosis (HR=1.61; 95%CI:1.03–2.52; P=0.037), psychiatric hospital admission (HR=1.93; 95%CI:1.13–3.29; P=0.017), home visit (HR=1.64; 95%CI:1.18–2.28; P=0.004), and A&E attendance (HR=1.88; 95%CI:1.31–2.72; P<0.001). However, after controlling for confounding-by-indication through PSM, BDZ exposure did not modulate the risk of any outcomes (all P>0.05). In analysis restricted to antipsychotic-naïve individuals, BDZ exposure reduced the risk of transition to psychosis at trend-level (HR=0.59; 95%CI:0.32–1.08; P=0.089). CONCLUSIONS: BDZ exposure in CHR-P individuals was not associated with a reduction in the risk of psychosis transition or other adverse clinical outcomes. Results in the whole-unmatched sample suggest BDZ prescription may be more likely in CHR-P individuals with higher symptom severity.