Bulimia nervosa (BN) and alcohol use disorder (AUD) frequently co-occur and may share genetic factors; however, the nature of their association is not fully understood. We assessed the extent to which the same genetic and environmental factors contribute to liability to BN and AUD. A bivariate structural equation model using a Cholesky decomposition was fit to data from 7241 women who participated in the Swedish Twin study of Adults: Genes and Environment. The proportion of variance accounted for by genetic and environmental factors for BN and AUD and the genetic and environmental correlations between these disorders were estimated. In the best-fitting model, the heritability estimates were 0.55 (95% CI: 0.37; 0.70) for BN and 0.62 (95% CI: 0.54; 0.70) for AUD. Unique environmental factors accounted for the remainder of variance for BN. The genetic correlation between BN and AUD was 0.23 (95% CI: 0.01; 0.44), and the correlation between the unique environmental factors for the two disorders was 0.35 (95% CI: 0.08; 0.61), suggesting moderate overlap in these factors. The findings from this investigation provide additional support that some of the same genetic factors may influence liability to both BN and AUD.
Bulimia nervosa (BN) and alcohol use disorder (AUD; alcohol abuse/dependence) frequently co-occur (Bulik, 1987, Garfinkel et al., 1995, Bulik et al., 1997, Lilenfeld et al., 1998, Herzog et al., 1999, Dansky et al., 2000, Wade et al., 2004, Pereyra et al., 2010 and Root et al., 2010). Estimates of the lifetime prevalence of AUD in women with BN have ranged from approximately 9% (Pyle et al., 1981) to 49% (Bulik, 1987), with the majority of investigations reporting between 20% and 25% (Mitchell et al., 1985, Holderness et al., 1994, Bulik et al., 2004, Baker et al., 2010 and Root et al., 2010). Yet, the nature of this association is incompletely understood.
One compelling hypothesis is that BN and AUD may share a familial diathesis. Results from family studies have been inconsistent (Bulik, 1987, Kassett et al., 1989 and Lilenfeld et al., 1997). For example, although studies uniformly reported that BN and AUD co-occurred in families, Lilenfeld et al. (1997) found that only BN probands with AUD had higher rates of AUD in family members, suggesting possible independent transmission of the disorders. One limitation of family studies is that they are unable to disaggregate genetic and common environmental effects. Twin studies, in contrast, allow the variance in liability to be partitioned into additive genetic (A; the cumulative impact of several genes of small to moderate effect), common environmental (C; environmental effects that increase similarity in twins and result from etiological factors to which both members of a twin pair are exposed regardless of zygosity such as childhood socioeconomic status), and unique environmental factors (E; factors that make twins dissimilar; the E term also includes measurement error). Bivariate models can extend that paradigm to determine the extent to which these factors contribute to the liability of both disorders.
Two investigations using twin methodology have examined the genetic and environmental association between BN and AUD (Kendler et al., 1995 and Baker et al., 2010). Kendler et al. (1995) applied multivariate pathway models to lifetime history of six major psychiatric disorders including BN and alcoholism. In the best-fitting model, a large proportion of the genetic liability to alcoholism was attributable to genetic factors that did not influence liability to the other five disorders; however, there was evidence of some overlap in genetic liability between BN and alcoholism (6%, as reported by Slane et al., 2012).
In a second investigation using the same population-based sample from the Virginia Twin Registry (Kendler and Prescott, 2006), which included additional data from subsequent occasions of measurement, Baker et al. (2010) examined the genetic association between BN symptom count (i.e., a positive score was given for each BN symptom) and AUD. Using a bivariate approach, they reported a genetic correlation of 0.53 (95% confidence interval (CI): 0.30; 0.80) in the best-fitting model, suggesting moderate overlap in genetic factors contributing to BN and AUD. However, the broad CI indicates a lack of statistical power to assess the strength of the association.
Another recent investigation (Slane et al., 2012) examined the nature of the association between individual bulimic behaviors (i.e., binge eating and the use of inappropriate compensatory behaviors such as self-induced vomiting or laxative use) and problematic alcohol use in a small sample (n=292) of female twins from the Michigan State University Twin Registry ( Klump and Burt, 2006). Parameter estimates for the best-fitting models indicated some overlap in genetic factors contributing to the liability of problematic alcohol use and binge eating (genetic correlation=0.31 (95% CI: 0.09; 0.53)) and also of problematic alcohol use and compensatory behaviors (genetic correlation=0.61 (95% CI: 0.34; 1.00)). These results provide additional evidence that some of the same genetic factors influence both bulimic behaviors and problematic alcohol use. However, the sample in this investigation was small, which might have decreased power and resulted in less precise estimates.
Thus, the results of these studies taken together suggest that BN and AUD may have shared genetic factors; however, additional large population-based twin studies of different samples are necessary to replicate and further elucidate the nature of this association. The purpose of this investigation was to evaluate the extent to which the same genetic and environmental factors contribute to the liability to both BN and AUD using a large population-based sample of twins from Sweden.
