Cannabis use disorder showed a genetic link to schizophrenia even after accounting for tobacco

Genetic liability to CUD predicted schizophrenia (beta=0.29, p=0.001) even after accounting for cannabis ever-use, tobacco smoking, and nicotine dependence. Mendelian randomization found a small but significant causal effect of CUD on schizophrenia (beta=0.10, p=0.02), but latent causal variable analysis did not support causation. The key distinction was between CUD (disordered use) and ever-use (trying cannabis), with only CUD showing the association.

Used GWAS summary statistics (sample sizes 161,405-357,806 Europeans) for genomic structural equation modeling, latent causal variable analysis, and multivariable Mendelian randomization. Examined genetic relationships between CUD, cannabis ever-use, tobacco smoking, nicotine dependence, and schizophrenia.

This study separates CUD from merely trying cannabis, finding that only disordered use genetically predicts schizophrenia. It also controls for tobacco (a major confounder), strengthening the CUD-specific signal. The mixed causal evidence keeps the question open but points toward CUD as more concerning than casual use.

The distinction between CUD and ever-use is critical. Most people who try cannabis do not develop schizophrenia, and this study genetically confirms that ever-use does not carry the same risk signal as disordered use. This nuance is often lost in public debate.

European ancestry GWAS limits generalizability. Two causal methods gave conflicting results. GWAS summary statistics may not capture all genetic variation. Cannot determine which aspects of CUD (frequency, potency, age of onset) drive the association.