Most genetic risk for cannabis use disorder overlaps with genes affecting opportunity and frequency of use

Trivariate Cholesky decomposition in 3,303 Australian twins measuring age of onset of cannabis use opportunity, lifetime frequency of use, and lifetime DSM-IV cannabis abuse/dependence. Estimated additive genetic (A), shared environment (C), and unique environment (E) contributions and overlap.

This finding has major implications for genetic research: studies searching for genes specific to cannabis addiction will mostly find genes that affect whether someone uses cannabis at all. This means genetic studies of cannabis use disorder must carefully account for exposure to avoid misidentifying general use genes as disorder genes.

This study reframes cannabis addiction genetics. Rather than searching for genes that make someone vulnerable to addiction once they start using, most of the genetic action is in the earlier stages: who encounters cannabis, who decides to try it, and who uses it frequently. The progression from frequent use to disorder involves relatively few additional genetic factors.

Australian twin sample may not generalize to other populations. DSM-IV criteria used (not DSM-5). Self-reported measures of opportunity, frequency, and diagnosis. Twin designs assume equal environments for identical and fraternal twins. The Cholesky decomposition assumes a specific causal ordering (opportunity -> frequency -> disorder).