Determining resting energy expenditure (REE) may be important in the nutritional assessment of adolescents with eating disorders (EDs). Calculated equations assessing REE, developed according to data from healthy people, may under- or overestimate REE in EDs. We have sought to compare the REE measured in clinical settings to that calculated using equations in actively ill adolescents with anorexia nervosa (AN) and bulimia nervosa (BN), and following stabilization of weight and disordered eating. Thirty-five female adolescents with AN and 25 with BN were assessed at admission to inpatient treatment and at discharge. REE was measured using indirect calorimetry (DELTATRAC Metabolic Monitor). Expected REE was calculated using the Harris-Benedict equation. An overestimation of expected versus measured REE was found for both patients with AN and BN, both at admission and discharge. Second, the differences between expected and measured REE were significantly less robust in BN versus AN. Third, REE before renourishing was lower in inpatients with AN versus BN. Fourth, the REE of patients with AN (both measured and expected) increased from admission to discharge, to a greater extent than expected solely from the increase in weight. The difference between admission and discharge expected and measured REE was significant also in patients with BN. Our findings suggest that predicted and measured REE are different in both AN and BN, and that both expected and measured REE are not useful in the planning of renourishing programs in patients with AN.