Anorexia nervosa (AN) is an illness commonly diagnosed in adolescence with low recovery rates
and high healthcare costs. The major medical complication of AN is malnutrition. Caloric
restriction, purging and other weight control behaviors can lead to medical instability
(abnormal vital signs) requiring hospitalization. The primary goal of hospitalization is to
restore medical stability by reintroducing nutrition, or "refeeding". Within 12 mo of
discharge, 43% of patients will require medical rehospitalization. This results in a costly
course of recovery, given that eating disorders are the most expensive among primary mental
health diagnoses requiring hospitalization. Several lines of inquiry seeking to identify
characteristics or short-term outcomes that may predict better recovery in AN point to rapid
short-term weight gain as a strong predictor of long-term outcomes. Greater weight gain in
hospital predicts weight recovery at 12 moand, in psychotherapeutic trials, greater weight
gain during the first 3-4 wk (1.7-1.9 lb/wk) predicts full remission at 12 mo. Unfortunately,
the currently recommended approach, Lower Calorie Refeeding (LCR), is associated with slow
weight gain and prolonged hospital stay. Treatment is now moving sporadically toward Higher
Calorie Refeeding (HCR) in the hope of improved recovery. However, no study to date has
directly compared these two approaches.] Since the refeeding syndrome, characterized by rapid
electrolyte shifts, delirium and cardiac arrest in response to the influx of nutrients was
first described around WWII, refeeding has been approached with caution. Following
documentation of this syndrome in patients with AN, conservative, consensus-based
recommendations for LCR were developed to ensure safety. LCR typically begins around 1200
kilocalories (kcal) per day and advances by 200 kcal every other day. The investigators found
that patients initially lose weight on this "start low and go slow" approach and require
prolonged hospitalizations to achieve medical stability. This finding contributed to
recognition of the "underfeeding syndrome". In subsequent studies, the investigators
demonstrated that HCR produced faster weight gain and shorter hospitalization. While no
increased risk of refeeding syndrome has been reported using HCR, the variety of electrolyte
supplementation protocols being used to manage risk have not been examined.
Findings from these observational and retrospective studies have been rapidly accepted by
many clinicians and insurers and HCR is now being integrated into practice in many hospitals.
However, there are major gaps in the evidence necessary to adopt HCR as the new standard of
care. These gaps are: 1.) It is not known if HCR impacts clinical remission, which is
typically defined as the combination of weight and cognitive recovery at 12 mo. 2.) The
safety of HCR has not been confirmed. The hallmark electrolyte imbalances of refeeding
syndrome occur frequently and still have not been systematically examined on differing
refeeding protocols. 3.) The relative cost-effectiveness of the two approaches has not been
established.
The investigators propose to conduct a randomized controlled trial (RCT) at two sites to
directly compare HCR and LCR for refeeding in AN. To accomplish the following aims, 120
adolescents will be enrolled upon admission to hospital for malnutrition secondary to AN and
randomized 1:1 to HCR (beginning with 2000 kcal and advanced 200 kcal/d) or LCR (beginning at
1400 kcal/d and advanced 200 kcal every other day) until medical stability is restored.
Participants will be followed for 12 mo after randomization: Daily while in hospital and at
follow-up [10 dy],1 mo, 3 mo, 6 mo, and 12 mo after randomization.