Gender differences in determinants of smoking initiation and persistence in California twins

Objective: To determine the effects of genetic versus environmental influences on smoking initiation (SI) and smoking persistence (SP). Methods: Native California twins (32,359 pairs), who completed a questionnaire in 1992 or 1998 to 2001, were studied. Standard epidemiologic and genetic analyses were conducted using multiple logistic regression and biometric models to determine factors related to smoking phenotype. Results: The strongest influence on SI was having a co-twin who ever smoked; the adjusted odds ratio was 9.7 [95% confidence limits (CL), 8.8-10.6] among monozygotic twins and 5.7 (95% CL, 5.2-6.2) among dizygotic like-sex pairs. The risk of SP was also increased if the co-twin currently smoked [adjusted odds ratios, 3.5 (95% CL, 3.0-4.1) for monozygotic twins and 2.3 (95% CL, 2.0-2.7) for like-sex dizygotic pairs]. The proportions of variance due to genetic effects, shared environment, and individual environment for SI were 31.6% (24.2-39.1), 47.5% (41.1-53.7), and 20.9% (18.8-23.1) for females, and 71.2% (66.7-75.4), 12.0% (8.7-15.7), and 16.7% (15.0-18.7) for males. For SP, estimates were identical by gender: 54.6% (43.6-65.5), 8.6% (0-17.1), and 36.8% (32.9-40.9). Modification of SI by closeness between twins was found, but little difference was seen for SP by closeness, birth cohort, or age. Conclusions: Gender differences in the pattern of genetic and environmental determinants of SI indicate that gender-specific approaches may be needed for smoking prevention efforts. Modification of genetic effects by closeness between twins and birth cohort suggests that environmental interventions could reduce a heritable propensity to smoke. However, the apparently heritable tendency to continue smoking is unaffected by gender, age, birth cohort, or closeness between twins.