Age at onset of psychosis (AAO) may be younger in patients with cannabis use disorders (CUD) compared to those without CUD (NCUD). Most previous studies did not control for potential confounders, did not report effect sizes and included mostly adult patients from non-representative samples.

Controlling for relevant confounders, differences in AAO between patients with and without lifetime CUD were analysed in a large epidemiologically based cohort of 606 first-episode psychosis (FEP) patients (age 14 to 29 years) admitted within three years to the Melbourne Early Psychosis Prevention and Intervention Centre. Data were collected from medical files using a standardized scale.

Overall, AAO was not significantly different in CUD (n = 449; 74.1%) compared to NCUD, neither univariate nor when controlling for gender and premorbid functioning. However, AAO was younger in those with early CUD (starting before age 14) compared to NCUD (F(1) = 11.3; p = 0.001; partial η2 = 0.042). When considering the subgroups of early versus late onset psychosis, AAO was even later in early onset psychosis patients with CUD compared to those with NCUD (F(1) = 8.4; p = 0.004; partial η2 = 0.072). These findings were consistent for patients with non-affective psychoses, in those with CUD without other substance use disorders and in those with CUD explicitly starting in the pre-psychotic phase. Notably, 89.1% started cannabis before the onset of psychotic symptoms.

CUD starting before age 14 was associated with an earlier AAO at a small effect size, but only in adult onset FEP patients.