Some of the burden of eating disorders is still hidden

In this issue of Lancet Psychiatry, Santomauro and colleagues make a stunning claim: The 2019 Global Burden of Disease Study (GBD) underestimated the prevalence of eating disorders by 41.9 million cases. By training its focus solely on anorexia nervosa and bulimia nervosa (the tip of the iceberg when it comes to eating disorders), the GBD overlooked 17.3 million individuals with binge-eating disorder and 24.6 million with other specified feeding or eating disorder, who together accounted for 3.6 million disability-adjusted life years in 2019.¹ This landmark paper highlights that eating disorders are four times as common as previously thought and associated with double the disability burden. The results demonstrate that binge-eating disorder and other specified feeding or eating disorder are especially common with increasing age, and also more fully describe the burden of eating disorders in males, shattering the inaccurate but entrenched stereotype that eating disorders affect only thin young white women. The impact of these findings is made more impressive via the use of innovative, rigorous, and reproducible methods that make the best use of global data to further challenge misconceptions about the rarity of eating disorders across countries represented in the GBD. That said, for two reasons, the global burden of eating disorders is almost certainly greater than their already staggering findings suggest.

First, the authors note that avoidant/restrictive food intake disorder, rumination disorder, and pica could not be included in their analysis because the prevalence of these disorders is unknown. It’s true that rigorous epidemiological studies on the full spectrum of DSM-5 Feeding and Eating Disorders are lacking, but growing evidence from school and community samples suggests that these disorders are common. In one study of 1,430 elementary school children in Switzerland ages 7–13 years, 3.2% endorsed symptoms consistent with avoidant/restrictive food intake disorder (characterized by limiting food intake due to sensory sensitivity, fear of aversive consequences, and/or lack of interest in eating or food) on a self-report questionnaire.² Similarly, in a questionnaire-based study of 797 adults ages 21–77 years in Singapore, 4.1% screened positive for avoidant/restrictive food intake disorder.³ In the same study of Swiss schoolchildren, 1.7% endorsed behaviors consistent with rumination disorder (regurgitating food and then chewing, re-swallowing, or spitting it out), 3.8% endorsed behaviors consistent with pica (consuming non-nutritive, non-food substances), and 1.1% endorsed behaviors consistent with both disorders.⁴ Avoidant/restrictive food intake disorder in particular carries risk for psychiatric (e.g., suicidality⁵) and medical (e.g., low bone density⁶) complications, suggesting that the burden in disability-adjusted life years could be substantial.

Second, in order to estimate disability-adjusted life years, Santomauro and colleagues make the assumption that subthreshold eating disorders (e.g., subthreshold bulimia nervosa) are only half as impairing as their full threshold counterparts (e.g., bulimia nervosa), and that atypical anorexia nervosa (in which individuals restrict their food intake but are not underweight) carries disability equal to that of anorexia nervosa minus the impact of being underweight. This approach makes sense mathematically and is arguably the only way that the authors could have estimated disability-adjusted life years in their sample. However, it almost certainly under-estimates burden. Individuals with subthreshold eating disorders typically do not differ from their full threshold counterparts in terms of eating pathology, general psychopathology, or physical health.⁷ Furthermore, regardless of behavioral symptom frequency, full threshold eating disorders and their subthreshold counterparts are united by the same underlying cognitive psychopathology, which is strongly correlated with clinical impairment.⁸ Just like those with anorexia nervosa, individuals with atypical anorexia nervosa have often experienced precipitous weight loss, so absolute body mass index is not necessarily a valid marker of disability. For example, adolescent girls with restrictive eating who do not reach commonly cited weight cutoffs for anorexia nervosa still show significantly lower bone mineral density compared to healthy girls, putting the former group at increased risk for fractures.⁹

In sum, Santomauro and colleagues’ trenchant re-analysis of the 2019 GBD clearly demonstrates that eating disorders are common and impairing. Their paper is a clarion call for binge-eating disorder and other specified feeding or eating disorder to be included in future versions of GBD. We wholeheartedly agree and wish to push the envelope even further: We recommend that GBD (1) include prevalence estimates for avoidant/restrictive food intake disorder, rumination disorder, and pica; and (2) obtain direct measures of the disability associated with all DSM-5 Feeding and Eating Disorders. If they do, the global burden will be even more stunning, underscoring the clear need for increased funding to study, prevent, and treat these debilitating illnesses.