19. Higher Calorie Refeeding in Atypical Anorexia Nervosa: Short-Term Outcomes From the Study of Refeeding to Optimize iNpatient Gains (StRONG)


      StRONG was a multicenter randomized clinical trial of refeeding in hospitalized adolescents and young adults with malnutrition secondary to anorexia nervosa (AN) and atypical anorexia nervosa (AAN) ( #NCT02488109). At end-of-treatment, higher calorie refeeding (HCR) was more efficacious and less costly than lower calorie refeeding (LCR). Here we compare efficacy of HCR in AAN vs. AN.


      Participants were N=120, 12-24 yr-olds with > 60% median Body Mass Index (%mBMI) and medical instability. AAN was defined as %mBMI >85% at baseline. Meal-based HCR began 2000 calories/day (kcal/d) and advanced 200 kcal/d; LCR began 1400 kcal/d and advanced 200 kcal every other day (no tube feeding). Efficacy was defined as time to restore a 6-point Medical Stability Index (MSI): heart rate (HR) ≥ 45 bpm; systolic blood pressure (SBP) ≥ 90 mmHg, temperature ≥ 35.6° C, orthostatic increase in HR ≤ 35 bpm and decrease SBP ≤ 20 mmHg, and ≥ 75% median BMI (%mBMI). Main outcome was days to restore MSI, compared between groups with unpaired t-test. Exploratory moderator analyses examined the interaction between refeeding treatment and diagnosis on key outcomes (time to recover heart rate and weight gain). Weight gain was defined as change in %mBMI.


      Modified intention to treat analyses included N=111. Mean age was 16.5 (2.5) yrs, 43% had AAN. Upon admission, %mBMI was 95.2 (9) in AAN vs. 76.5 (5.9) in AN, p<.001. Upon discharge, %mBMI was 98.3 (8.9) in AAN vs. 82.2 (5.3) in AN, p<.001. MSI was restored fastest in patients with AN refed by HCR [7.1 (5.4) days], whereas medical stability required three additional days to restore in patients with AAN [10.1 (5.3) days, p<0.01]. Diagnosis (AAN or AN) and treatment (HCR or LCR) interacted to weaken the effect of refeeding on HR recovery [B=3.76 (.572,6.95), p=0.021] and weight gain [B=039 (.006,072), p=0.021], which was 0.3% mBMI per day slower (p=0.005) and 2.6% mBMI less overall (p=0.009) in AAN than AN.


      While HCR is more efficacious than LCR for refeeding in AN, it may contribute to underfeeding in AAN.

      Sources of Support

      National Institute Child Health & Human Development #R01HD082166; Identifier NCT02488109.