Correlates of regular cigarette smoking in a population-based sample of Australian twins


Arpana Agrawal, Washington University School of Medicine, Department of Psychiatry, 40 N. Kingshighway, Ste ♯2, St Louis, MO 63108, USA, Tel: 314 286 1778, Fax: 314 286 2213, E-mail:


Aims  To investigate the role of measured risk factors and the influence of genetic and environmental factors on regular cigarette smoking.

Design  Members of monozygotic and dizygotic, including unlike-sex twin pairs (n = 6257) from a young adult cohort from the Australian Twin Registry.

Methods  Cox proportional hazards models were used to determine whether putative risk factors were significantly associated with regular cigarette smoking. Risk factors were classified into four tiers: tier 1 (parental history, including parental education, alcoholism and cigarette smoking), tier 2 (early home and family influences), tier 3 (early life events, e.g. trauma) and tier 4 (psychiatric symptoms/disorders with onset prior to 14 years), after controlling for gender, zygosity and their interactions. Genetic models were fitted to examine the heritability of smoking behavior before and after controlling for significant covariates from the four tiers.

Findings  Parental history of cigarette smoking and alcoholism, parental closeness and home environment, as well as incidence of childhood sexual abuse or other trauma, a history of early onset panic attacks and conduct problems were associated with regular cigarette smoking. Important age interactions were found, particularly for family background risk factors. Regular cigarette smoking was moderately heritable, even after accounting for significant covariates.

Conclusions  Several measured risk factors are associated with regular smoking. While some of the genetic influences on regular smoking may be shared with these risk factors, a significant proportion of the genetic vulnerability to regular smoking is phenotype-specific.