Abstract
Background and aims
Addiction-related behaviors, such as loss of control eating (LOC), cigarette smoking and alcohol consumption, have been associated with high body mass index (BMI). This study aimed to assess genetic and environmental contributions to these associations over time.
Design
A longitudinal twin study using data from waves 2 and 3 of the Center on Antisocial Drug Dependence study, employing additive genetic (A), shared environmental (C), nonshared environmental (E) influences and cross-lagged models.
Setting
Colorado, USA.
Participants
The sample included 764 male and 997 female same-sex twins.
Measurements
BMI was calculated using self-reported height and weight. LOC was self-reported. Cigarettes smoked per day (CPD) and drinks per week (DPW) were assessed during interviews.
Findings
We conducted three cross-lagged models: LOC and BMI in males, LOC and BMI in females and CPD and BMI in females, after excluding small phenotypic correlations (|r| < 0.10). Trait stability over time was largely attributable to genetic factors, accounting for 62% of the variance in BMI (both sexes), 11% in LOC (males), 18% in LOC (females) and 56% in CPD (females) at wave 3. Residual effects were mostly from nonshared environmental factors, accounting for 38% of the variance in BMI (both sexes), 76% of LOC (females), 71% of LOC (males) and 44% of CPD (females) at wave 3. A small but statistically significant cross-lagged effect occurred from wave 2 BMI to wave 3 LOC, explaining 12% (males) and 3% (females) of the variance in wave 3 LOC, with genetic factors accounting for most of this effect. No cross-lagged effects emerged from LOC or CPD to BMI.
Conclusions
Genetic factors contributing to higher body mass index at an earlier age may also increase the risk of developing loss of control eating later in life, highlighting the importance of early weight-related interventions to prevent the onset of disordered eating behaviors.