Academic journal article Ethical Human Psychology and Psychiatry

Overestimates of the Genetic Contribution to Eating Disorders

Academic journal article Ethical Human Psychology and Psychiatry

Overestimates of the Genetic Contribution to Eating Disorders

Article excerpt

The purpose of this article is to analyze and critique repeated claims in the literature that there is a substantial genetic contribution to eating disorders. Data from the existing twin and family studies of eating disorders were tabulated and compared to heritability estimates resulting from complex statistical analyses of the same data. Overall, concordance in monozygotic twins is 26% for bulimia and 35% for anorexia nervosa. Among the relatives of probands with bulimia, 95.1% do not have bulimia, whereas among the relatives of probands with anorexia nervosa, 97.1% do not have the disorder. The raw data refute claims that the genetic heritability of eating disorders is as high as 80%. The erroneous conclusion that there is a substantial genetic contribution to eating disorders needs to be corrected by focusing on the raw data for twin concordance and prevalence in first-degree relatives.

Keywords: genetics; eating disorders

This article analyzes and critiques repeated statements in the literature that the eating disorders have a substantial genetic component. Kaye, Klump, Frank, and Strober (2000), for instance, stated that genetic studies have resulted in "heritability estimates in the range of 50% to 90% for AN [anorexia nervosa] and 35% to 83% for BN [bulimia nervosa]." Wade, Bulik, Neale, and Kendler (2000) stated that "anorexia nervosa was estimated to have a heritability of 58%." Similarly, Bulik, Sullivan, and Kendler (1998) stated that the heritability of bulimia was 83% in a sample of 1897 female twins.

These heritability estimates are derived from statistical analyses of twin concordance data, but are overestimates when one considers the raw data on twin concordance and prevalence among first-degree relatives of affected probands.

THE EMPIRICAL DATA ON GENETICS OF EATING DISORDERS

Pairwise and Probandwise Concordance

The concordance rate for a disorder or trait can be calculated in two ways. The pairwise concordance rate corresponds to the raw data. For instance, if there are 100 monozygotic (MZ) twin pairs and 100 dizygotic (DZ) twin pairs, interviewers might first interview one twin in each of the 200 sets of twins. If they found 50 cases of anorexia nervosa among the 100 MZ twins and 50 cases among the 100 DZ twins, they would then proceed to interview the co-twin of each of these affected twins. This would result in interviews of 50 MZ and 50 DZ twins in the second round.

If there were 30 cases of anorexia in the 50 MZ co-twins, the pairwise concordance would be 30/50 = 60% for the MZ twins. If there were 10 cases of anorexia in the 50 DZ co-twins, the pairwise concordance would he 10/50 = 20% for the DZ twins.

The formula for pairwise concordance is:

A = C/C + D

where C = the number of concordant pairs, and D = the number of discordant pairs.

Probandwise concordance is calculated differently. For probandwise concordance, the formula is:

B = 2C^sub 1^ + C^sub 2^/2C^sub 1^ + C^sub 2^ + D

where C^sub 1^ = the number of concordant pairs and where both twins were independently ascertained, C^sub 2^ = the number of twins where only one twin was ascertained, and D = the number of discordant pairs. "Ascertained" means interviewed diagnostically.

In the same study used for calculation of pairwise concordance above, all twins were ascertained so C^sub 1^ = 50 and C^sub 2^ = 0. Probandwise concordance for MZ twins is [2 × 30] + 0/[2 × 30] + 0 + 20. This reduces to 60/80 = 75%. Probandwise concordance for DZ twins is [2 × 10] + 0/[2 × 10] + 0 + 40. This reduces to 20/60 = 30%.

Thus, the pairwise concordance rates are 60% for MZ and 20% for DZ twins, but the probandwise concordance rates are 75% for MZ and 30% for DZ twins. Probandwise rates carry the risk of overestimating the difference, and pairwise rates carry the risk of underestimating the difference between MZ and DZ twins. There are arguments in the literature as to which method is more relevant for mental disorders. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.