Predictors of complications in anorexia nervosa and atypical anorexia nervosa in adolescents: degree of underweight or extent and recency of weight loss?

Abstract

Anorexia Nervosa (AN) is the eating disorder characterised by emaciation and underweight, commonly precipitated by dietary restriction that progresses to increasing restriction and self-starvation. AN predominantly develops during adolescence. It is uncommon, with an estimated lifetime prevalence of 0.3% in the general adolescent population. Since the 1800s, the literature has extensively described the significant morbidity and mortality associated with the severity of underweight in people with AN. Over the last 15 years a new ED has increasingly been identified in adolescents who have lost significant amounts of weight, but notably are not underweight. Severe complications from weight loss have been observed in this group of higher weight adolescents despite the absence of underweight. The clinical significance of this ‘non-underweight’ ED has recently been acknowledged with the establishment of distinct diagnostic criteria for Atypical Anorexia Nervosa (AAN). Limited data are available about AAN, including in relation to the prevalence, clinical trajectory or optimal treatment. However, beyond these aspects, the absence of underweight in AAN poses questions about the relative contribution of weight loss compared to weight as predictors of the clinical complications in AN and AAN, that have long been thought to result from underweight.

This thesis focusses on AAN to gain greater understanding and knowledge about the prevalence, clinical trajectory and outcomes of this recently described ED. The proportion of adolescents diagnosed with AAN and hospitalised at the Royal Children’s Hospital (RCH) over a nine year period is assessed. A case study describes an adolescent who, despite significant weight loss and physical complications associated with AN, was not initially considered for diagnosis of an ED as she was not underweight. A comparison between AN and AAN is undertaken to compare the characteristics of these two diagnoses. The result of this thesis questions the fundamental aspect of nutrition in restrictive AN that is, the validity of assessing malnutrition by the degree of underweight rather than weight loss.

Over the nine year study period AAN emerged as a significant ED with a dramatic five-fold increase in hospital admissions to the RCH. Adolescents diagnosed with AAN were found to experience a similar profile of life-threatening complications as adolescents who had AN, including early refeeding syndrome. The only differences detected between the diagnostic groups were higher premorbid weight and weight at admission in AAN, and in AN a lower systolic blood pressure nadir during hospitalisation. Notably, the duration of weight loss and amount of weight loss were found to be similar in the two diagnostic groups. Finally, the very important question of the contribution of weight loss compared to underweight as predictors of clinical complications is addressed. The critical finding was that weight loss, not underweight, predicted the potentially life-threatening complications of sinus bradycardia and lower pulse rate nadir. Similarly, recent weight loss was associated with greater severity of ED cognitions, with no association between underweight and ED cognitions. Several complications were predicted by both lower weight and greater weight loss, but no outcome was independently associated with weight itself. For the first time weight loss, rather than weight itself was found to be a stronger predictor of some of the serious complications following self-starvation in adolescents diagnosed with AN and AAN. No complication was predicted by underweight itself, suggesting that underweight as a marker of starvation and malnutrition is now out of step with current trends towards patients having higher weight. It is hoped that the outcomes of this thesis will raise awareness of the prevalence and severity of complications in AAN, and also inform future research directions.