Examining the Heritability of Tobacco and Alcohol Use Among a Sample of Adolescent Twins

Background: Despite modest decreases in recent years in the prevalence of adolescent substance use, national estimates suggest that in 2013 about 48% of youth had ever tried alcohol and 26% had ever smoked a cigarette. It is widely known that familial and peer factors predispose adolescents to use tobacco and alcohol, but familial factors can exert their influence through either genetic or shared environmental effects. Applying a twin modelling approach to understanding adolescent substance use behavior can help delineate the effects of biology and the environment in generating risk and determine the extent to which the behaviors we aim to modify among youth are open for intervention.

Method: The methodology used in this study is a novel one for our field. Twin modeling approaches compare how similar, and different, pairs of twins are for a particular behavior. For example, if a behavior is under genetic influence monozygotic (MZ) twins are expected to be more similar than dizygotic twins because MZ twins not only have the same genetic make-up but also have experienced the same rearing environment and experiences. Comparing the behavior of twins allows researchers to determine the degree to which individual differences in a behavior are under the influence of genetic or environmental factors. This study examines the factors that contribute to tobacco and alcohol use, and estimates the heritability of these behaviors among a sample of adolescent twins (n=740 twin pairs, mean age = 16.07, 49% female) from the National Longitudinal Study of Adolescent Health.

Results: Logistic regression analyses adjusted for the lack of independence in the observations indicate that having more friends who smoke (β=0.91, p<.0001) and older age (β=0.13, p<.05) are associated with lifetime smoking. A second logistic regression indicated that older age (β=0.30, p<.0001), maternal smoking (β=0.47, p<.05), and having more friends who drink (β=0.87, p<.0001) are associated with lifetime alcohol use. Next, the contribution of genetic factors on the initiation of tobacco use during adolescence was assessed through twin modeling using OpenMX. It was found that genetic factors explain 44% of the variance in adolescent smoking initiation, shared environmental factors explained 30%, and unique environmental factors were found to explain 25%. Lastly, a twin model in which the contributions to alcohol initiation were examined revealed that genetic factors accounted for 34% of the variance, while 33% came from shared environmental factors, and 34% from unique environmental influences.

Conclusion: Although genetic factors play a role in the initiation of tobacco and alcohol use among this nationally representative sample of adolescent twins, quite a large portion of the variance in these behaviors comes from environmental sources. The present findings have implications for the extent to which the behaviors that social workers target are malleable and should be an essential part of the person-in-environment approach that drives our understanding of individual differences in human behavior.