Austin-Zimmerman · Psychological medicine 2024 · case-control and observational cohort study · n=144698

The impact of schizophrenia genetic load and heavy cannabis use on the risk of psychotic disorder in the EU-GEI case-control and UK Biobank studies.

Level 4 - case-series / case-control

Case-control and population-based observational cohort study

PubMed 39637925 · doi:10.1017/S0033291724002058 · record verified 2026-08-26

What was done

Researchers evaluated the independent and combined effects of heavy cannabis use and schizophrenia polygenic risk score (PRS) on psychosis risk. They analyzed 1,098 participants from the EU-GEI case-control study and 143,600 participants from the UK Biobank. Polygenic risk scores for schizophrenia and cannabis use disorder were derived from genome-wide association study summary statistics and examined alongside self-reported cannabis use.

What was found

Schizophrenia PRS and cannabis use independently increased the risk of psychosis in both cohorts, with no evidence of an interaction between schizophrenia PRS and patterns of cannabis use. In the EU-GEI sample, daily use of high-potency cannabis remained strongly associated with psychotic disorder after adjusting for schizophrenia PRS (adjusted OR = 5.09, 95% CI 3.08-8.43, p = 3.21 x 10^-10). Schizophrenia PRS was not associated with cannabis use in EU-GEI cases or controls or UK Biobank cases, and its association with cannabis use in UK Biobank non-psychosis controls was substantially attenuated after accounting for cannabis use disorder PRS.

Why it matters

These findings suggest that the elevated risk of psychotic disorders linked to heavy, high-potency cannabis use is not simply a byproduct of shared genetic liability for schizophrenia.

Limits

Cannabis exposure relied on retrospective self-report subject to recall and reporting bias. Detailed high-potency cannabis data was restricted to the EU-GEI sample (n = 1,098). Polygenic risk scores explain only a portion of total schizophrenia heritability and unmeasured environmental confounders cannot be excluded.