To determine whether critically ill children are hypermetabolic and to calculate whether predictive equations are appropriate for critically ill children. Prospective, clinical study. Pediatric intensive care unit. A total of 57 children (39 boys) aged 9 months to 15.8 yrs. None. The median resting energy expenditure measurement measured by indirect calorimetry was 37.2 (range, 11.9-66.6) kcal x kg(-1) x day(-1). This was significantly lower than would be predicted using either the Schofield (42.7 [26.9-65.4] kcal x kg(-1) x day(-1)) or Fleisch equations (42.8 [20.9-66.2] kcalx kg(-1)-1 x day(-1), p < .001) but significantly higher than the White equation developed specifically for pediatric intensive care units (26.2 [8.5-70.1] kcal x kg(-1),day(-1), p < .0001). Methods comparison analysis showed the limits of agreement were -484 to 300, -461 to 319, and -3.2 to 854 kcal/day, respectively. Multivariate analysis indicated the following factors contribute to hypometabolism and hypermetabolism: age (p = .006), sex (p = .034), time spent in the pediatric intensive care unit (p = .001), diagnosis (p = .015), weight (p = .009), temperature (p = .04), continuous infusion for sedation (p = .04), and neuromuscular blockade (p = .03). Children do not become hypermetabolic during critical illness. These data suggest that agreement between resting energy expenditure and the predictive equations are so broad that they are inappropriate for use in critically ill children.