Association between high-potency cannabis consumption and risk of offending behaviours in a two-year first-episode psychosis cohort

Brief introduction to the research question:

Cannabis use is highly prevalent among individuals with psychosis spectrum disorders. Both cannabis use and psychosis are independent risk factors for antisocial behaviours such as violence. However, existing studies report inconsistent findings regarding the association between cannabis use and risk of such behaviours in people with psychosis, and the nature of the relationship between specific patterns of cannabis use and antisocial behaviour outcomes remain poorly understood.

Aims

We aimed to examine the longitudinal effects of cannabis use patterns (based on continuity, potency and frequency) on antisocial outcomes in first-episode psychosis, and to explore the strength, temporality, and direction of this relationship.

Methods

This 2-year prospective cohort study followed 258 participants (aged 18–65 years) with first-episode psychosis recruited from the South London and Maudsley NHS Foundation Trust. Data were collected via structured interviews and medical records at baseline, 1-year, and 2-year follow-ups. Participants were grouped according to cannabis use patterns based on self-reported continuity, potency and frequency of use from psychosis onset to last follow-up. Antisocial outcomes were defined as self-reported offending behaviours.

ANOVA, t-tests, and chi-square or Fisher’s exact tests were used to examine group differences in clinical, sociodemographic, and antisocial characteristics. To investigate the relationship between cannabis use patterns and antisocial outcomes, we employed a three-model statistical framework comprising Cox proportional hazards regression (survival analysis), fixed-effects models (within-person effects), and cross-lagged panel analysis (temporal and directional effects), adjusting for relevant clinical and sociodemographic confounders.

Results

This sample (mean age 28.10 [SD 8.18]; 158 [61%] male; 231 [83%] with non-affective psychosis; attrition 33.16%) comprised four cannabis use patterns: former users (n=56), never users (n=103), intermittent/continuous hash users (n=44), and continuous-skunk users (n=55). Participants with different cannabis use patterns showed significant differences in the prevalence and type of offending behaviours, including interpersonal violence, theft, drug-related offence, and vandalism (all p<0.05).

Of the total sample, 214 participants were included in the longitudinal analyses. In survival analyses, adjusted Cox models showed a significantly higher hazard of offending in the intermittent/hash group (HR=3.69, 95% CI: 1.15–11.85, p=0.028) and continuous-skunk group (HR=4.63, 95% CI: 1.52–14.09, p=0.007), compared with former users. A dose-response relationship was observed in fixed-effects models, with increases in cannabis potency and frequency associated with higher offending risk (β=1.47, 95% CI: 0.67–2.25, p<0.001). Cross-lagged models indicated that offending in the second year was significantly predicted by high-potency cannabis use in the first year (β=0.039, p=0.003), while the reverse effect was not significant (β=-0.182, p=0.097).

Conclusions

These findings indicate a dose-response, temporal, and unidirectional association between high-potency cannabis use and subsequent offending behaviours in first-episode psychosis. Our results suggest that targeted interventions to reduce or withdraw cannabis use may mitigate antisocial or offending risk in the psychosis population. However, reliance on self-reported data warrants cautious interpretation. Further research is needed to replicate the findings in a larger population and explore the biological mechanisms underlying this association.

References

No references