Searching for Childhood Risk Factors for AN

Do certain risks during childhood make an individual more likely than others to develop anorexia nervosa (AN)? Two London clinicians found that most previous studies identifying childhood risk factors for eating disorders have been limited by the studies’ designs. According to the clinicians, most are retrospective or involve too few subjects to be helpful.

Thus, Drs. Dasha E. Nicholls and Russell M. Viner, of the Institute of Child Health at University College, London, designed a study to examine whether previously identified childhood risk factors for AN could predict self-reported lifetime occurrence of AN by 30 years of age (J Am Acad Child Adolesc Psychiatry 2009; 48:791). To do so, they used data from the 1970 British Cohort Study, a continuing longitudinal study of infants born between April 5-11, 1970, in England, Scotland, Wales, and Northern Ireland.

Drs. Nicholls and Viner identified young adults with eating disorders through an interview and computer-generated questionnaire. All subjects were 30 years of age at the time they were studied. All were asked if they had ever had AN, bulimia nervosa, or compulsive eating, problems with swallowing, or some other kind of eating disorder. Of the 406 individuals who reported an eating difficulty, 101 (0.9%) of the entire sample reported having AN alone.

Individuals with AN were broken down into a number of categories: 5 (5%) were male; 3 (3%) reported the onset of their illness occurred at or before 10 years of age; 9 (9%) at 12 to 13 years of age; 18 (18%) at 14 to 15 years of age; 21 (21%) at 16 to 17 years of age; and 20 (20%) at 18 to 19 years of age. Twenty-four (24%) reported developing their eating disorder within the last year; 15 of these young adults said they had seen seeing a physician for their disorder within the last year.

Few risk factors were identified

The researchers could identify only a few risk factors in childhood that could be tied to later development of AN. However, the risk of developing AN was significantly predicted by female sex, frequent feeding problems in the first 6 months of life, maternal psychological morbidity at 5 years, separations from the mother lasting 1 month or longer before 5 years of age, and under-eating at 10 years of age. An additional risk factor included higher maternal body mass index (BMI). Higher self-esteem was a protective factor. In contrast, they did not find any role for many factors that had previously been suggested as risks, such as perinatal events or gestational age/birthweight, separation from mothers, childhood emotional disorder, parenting style, sleep problems, childhood BMI, or early puberty.

The authors’ findings closely followed those of other studies, including the lifetime prevalence of AN; in their cohort, this was 0.9%, the same as that reported by a nationally representative U.S. survey based on a diagnostic interview (Biol Psychiatry 2007; 61:348). In addition, the mean age at onset of AN was 18 years, with a median of 17 years; this compares with estimates from a recent US national study (median: 18 years) (Biol Psychiatry, ibid.). The authors note that their findings support the need for prevention strategies in AN that focus on enhancing self-esteem and that support maternal well-being, especially in association with feeding problems.

In an editorial in the same journal (J Am Acad Child Adolesc Psychiatry 48:782), David Herzog, MD and Kamryn T. Eddy, PhD report that the authors’ research is important because it begins to clarify variables that lead to risk for AN specifically and for eating disorders in general. According to the editorial, the next wave of risk factor research in eating disorders should build on these studies. Much waits to be explored, including shared risk for the varying eating disorder groups, especially in light of the proposed transdiagnostic concept of eating disorders by Fairburn and colleagues (Behav Res Ther 2003; 41:509).