Genetic and environmental contributions to the diversity of substances used in adolescent twins: a longitudinal study of age and sex effects

Abstract

Aims: To determine how genetic and environmental contributions affecting the number of psychoactive substances used varies with age and gender over the course of adolescence.

Design: Estimates of genetic, shared environmental and non-shared environmental contributions to total variance in diversity of substances used at ages 11, 14 and 17 years were obtained by fitting a multivariate behavior genetic (Cholesky) model.

Participants: A total of 711 male and 675 female twins.

Measurements: Participants reported whether they had used each of 11 substances.

Findings: The average diversity of substances used increased over time for both males and females, and males generally reported a wider diversity of substances used than females. Influences of genetic factors increased with age and were greater for males than for females at ages 14 and 17 years. Genetic factors remained consistent (i.e. highly correlated) across ages for both males and females, as did shared environmental influences for males. Non-shared environmental factors for both sexes and females' shared environmental factors were age-specific.

Conclusions: Regardless of sex, the proportion of variance in substances used attributable to genetic factors increases during adolescence, although it is greater for males than females at later ages. These findings indicate that prevention interventions may be most effective if they target early adolescence when environmental factors account for the majority of variance in substance use. The high correlation of genetic factors across ages suggests that early use may sometimes signal an early expression of a developmentally stable genetic predisposition.

Figure 1

Steps (and subsequent ns) for participant inclusion in final data analyses at each time-point.

Figure 2

Proportion of population variance in substance use counts accounted for by each of the variance components, estimated separately by sex and age. Estimates of 95% confidence intervals are given in brackets.