Longitudinal Relationships Between Cannabis and Tobacco Use and Symptom Severity in Individuals at Clinical High Risk for Psychosis

Abstract Objective Tobacco and cannabis are the most used substances among individuals at clinical high risk for psychosis (CHR-P), but it remains controversial whether substance use drives symptom exacerbation and psychosis transition, or vice versa. We investigated longitudinal dose-response relationships of tobacco and cannabis use with clinical presentation in a CHR-P population. Methods Data was obtained from the North American Prodrome Longitudinal Study (NAPLS2) CHR-P cohort (n=764). Participants were assessed every 6 months over two years. Substance use frequency, psychiatric symptoms (psychosis, depression, anxiety, and social anxiety), global social and role functioning, and neurocognitive performance were measured. Linear mixed effect models were used to model the relationship between substance use and clinical measurements across visits, and that between baseline use and trajectory of symptoms, functioning, and cognition. Results Psychiatric symptoms, functioning, and cognitive performance improved, while tobacco and cannabis use frequency did not change over two years for CHR-P individuals in NAPLS2. Heavier tobacco and cannabis use at current visit predicted worse anxiety at next visit (tobacco: β=0.178, p=0.033; cannabis: β=0.162, p=0.018). Better social functioning predicted heavier tobacco (β=0.178, p<0.001) and cannabis: (β=0.162, p<0.001) use at next visit. We observed a significant baseline cannabis-by-time interaction, where heavier baseline cannabis use predicted slower improvement of negative symptoms (β=0.159, p=0.0017, FDRp=0.0067) and deterioration of role function (β=-0.046, p=0.018). Conclusions In CHR-R, current tobacco and cannabis use predicted worse anxiety at future visits. Baseline cannabis use frequency predicts worse clinical trajectory, especially for negative symptoms.