Association between polygenic liability for schizophrenia and substance involvement: A nationwide population-based study in Taiwan.

Schizophrenia and substance involvement frequently co-occur in individuals, and a bidirectional relationship between the two has been proposed; shared underlying genetic factors could be an alternative explanation. This study investigated the genetic overlap between schizophrenia and substance involvement, including tobacco, alcohol and betel nut use. The study subjects were recruited from the Taiwan Biobank, and genome-wide genotyping data was available for 18 327 participants without schizophrenia. We calculated the Psychiatric Genomics Consortium-derived polygenic risk score (PRS) for schizophrenia in each participant. The significance of the schizophrenia PRS associated with substance involvement was evaluated using a regression model with adjustments for gender, age and population stratification components. The modified effect of gender or birth decade was also explored. The schizophrenia PRS was positively associated with lifetime tobacco smoking in women (OR in per SD increase in PRS = 1.12 with 95% CI 1.04-1.20, P = .002), but not in men (OR = 0.99 with 95% CI 0.95-1.04, P = .74), and the gender-PRS interaction reached significance (P = .006). The OR between PRS and lifetime tobacco smoking increased with the birth decade (P of birth decade-PRS interaction = .0002). In women, OR increased from 0.97 (P = .85) for subjects with a birth decade before 1950 to 1.21 (P = .04) for subjects with a birth decade after 1980; in men, the corresponding OR increased from 0.88 (P = .04) to 1.13 (P = .11). There was no association between schizophrenia PRS and alcohol/betel nut use phenotypes. This study provides evidence for the genetic overlap between schizophrenia and tobacco use in women, and this overlap was stronger in the younger population.