Age at onset of psychotic disorder: cannabis, BDNF Val66Met, and sex-specific models of gene-environment interaction.

Discovering modifiable predictors for age at onset may help to identify predictors of transition to psychotic disorder in the "at-risk mental state." Inconsistent effects of sex, BDNF Val66Met (rs6265), and cannabis use on age of onset were previously reported. BDNF Val66Met and cannabis use before illness onset were retrospectively assessed in a sample of 585 patients with schizophrenia and their association with age at onset was evaluated. Cannabis use was significantly associated with earlier age at onset of psychotic disorder (AOP; average difference 2.7 years, P < 0.001), showing dose-response effects with higher frequency and earlier age at first use. There was a weak association between BDNF Val66Met genotype and AOP (difference 1.2 years; P = 0.050). No evidence was found for BDNF × cannabis interaction (interaction χ(2) (1) = 0.65, P = 0.420). However, a significant BDNF × cannabis × sex interaction was found (interaction χ(2) (1) = 4.99, P = 0.026). In female patients, cannabis use was associated with earlier AOP in BDNF Met-carriers (difference 7 years), but not in Val/Val-genotypes. In male patients, cannabis use was associated with earlier AOP irrespective of BDNF Val66Met genotype (difference 1.3 years). BDNF Val66Met genotype in the absence of cannabis use did not influence AOP, neither in female or male patients with psychotic disorder. Complex interactions between cannabis and BDNF may shape age at onset in female individuals at risk of psychotic disorder. No compelling evidence was found that BDNF genotype is associated with age at onset of psychotic disorder in the absence of cannabis use.