Diagnostic Crossover: Challenging Anorexia Nervosa Subtypes

An initial diagnosis of an eating disorder can change over time. In fact, several long-term studies have reported substantial diagnostic crossover among the eating disorders. For example, results of recent studies show that 20% to 50% of patients with anorexia nervosa (AN) will develop bulimia nervosa (BN) over time. The percentages are smaller for crossover from BN to AN: One study found that up to 27% of patients with an initial diagnosis of BN crossed over to AN.

The trend toward crossover is common between subtypes, and up to 62% of patients with restricting-type AN will eventually develop binge eating/purging-type AN. According to the results of a recent study, these high rates of crossover challenge some of the validity of the current diagnostic guidelines in the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-R).

To examine the effects of crossover, Kamryn T. Eddy, PhD and colleagues at Massachusetts General Hospital, Boston, followed a group of 216 women with AN or BN for 7 years, using the Eating Disorders Longitudinal Follow-up Evaluation (Am J Psychiatry 2008;165:245). This modified version of the Longitudinal Interval Follow-up Evaluation assesses symptoms at intake and assigns DSM-IV-TR diagnoses during follow-up. Diagnoses were re-computed weekly during the follow-up period.

Patterns of crossover

During the 7 years of follow-up, 64 women (73%) who had been initially diagnosed with AN had diagnostic crossover: almost half (49%) crossed over between AN subtypes, and 34% crossed over from AN to BN. Among patients with an intake diagnosis of restricting-type AN, more than half (57.5%; 23 women) crossed over to another diagnosis during the follow-up period. Fifty-five percent (22) crossed over from AN, and 10% crossed over to diagnoses of BN. Among women with an intake diagnosis of binge eating/purging-type AN, 85% (41 women) experienced crossover during follow-up: 44% crossed over from the binge eating/purging-type to the restricting subtype of AN, and 54% crossed over to BN.

AN subtypes had the greatest degree of crossover

Thus, the greatest degree of crossover occurred in the AN subtypes; this was bidirectional and recurred throughout the 7 years of follow-up. Crossover from AN to BN was directly preceded by a period of binge eating/purging-type AN for most women, and was never preceded by a period of restricting-type AN, even among those who had an intake diagnosis of binge eating/purging type AN at intake. For some women, crossover from AN to BN preceded a progression to partial and full recovery; however, for about half of the women, crossover to BN was followed by crossover back to AN. Thus, for those with an intake diagnosis of AN, of either subtype, crossover was probable all through the follow-up period.

Movement from AN to partial recovery was common, and reported in 78% of the women. Women who experienced partial recovery represented 82% of those with restricting-type AN and 75% of those with binge eating/purging-type AN at intake.

Crossover from BN to AN

During follow-up, only a few women with an initial diagnosis of BN (18 women, or 14%) had a diagnostic crossover to AN. All of these women crossed over to binge eating/purging-subtype AN, and a small number (5 women, or 4% also experienced crossover to restricting-type AN. While crossover to AN was uncommon among women with an initial diagnosis of BN, partial (82%) or full recovery (65%) were common.

The authors noted that while their longitudinal data generally support the distinctiveness of the diagnostic categories of AN and BN, they found less support for the current AN subtype system. They suggest that there is a possibility that the transition from AN (particularly the binge eating/purging-subtype) to BN may not be a change in the disorder but a change in stage of illness. In practice, the primary difference between binge eating/purging-type AN and BN is weight and the associated criterion of amenorrhea. The long-term risk of relapse into AN suggests that a lifetime history of AN may carry important prognostic informationthat is, even after crossing over to BN, these women are at risk of relapsing into AN.

The authors also suggest that the frequency of crossover between restricting-type and binge eating/purging-subtype AN suggests that these two subtypes may not be unique diagnostic groups. Instead, restricting-type AN may just be an earlier phase in the course of the illness than the binge eating/purging phase. Thus, the two subtypes may in reality be different phases in the course of AN.