Trajectories of resting energy expenditure and performance of predictive equations in children hospitalized with an acute illness and malnutrition: a longitudinal study

There is scarce data on energy expenditure in ill children with different degrees of malnutrition. This study aimed to determine resting energy expenditure (REE) trajectories in hospitalized malnourished children during and after hospitalization. We followed a cohort of children in Bangladesh and Malawi (2–23 months) with: no wasting (NW); moderate wasting (MW), severe wasting (SW), or edematous malnutrition (EM). REE was measured by indirect calorimetry at admission, discharge, 14-and-45-days post-discharge. 125 children (NW, n = 23; MW, n = 29; SW, n = 51; EM, n = 22), median age 9 (IQR 6, 14) months, provided 401 REE measurements. At admission, the REE of children with NW and MW was 67 (95% CI [58, 75]) and 70 (95% CI [63, 76]) kcal/kg/day, respectively, while REE in children with SW was higher, 79 kcal/kg/day (95% CI [74, 84], p = 0.018), than NW. REE in these groups was stable over time. In children with EM, REE increased from admission to discharge (65 kcal/kg/day, 95% CI [56, 73]) to 79 (95% CI [72, 86], p = 0.0014) and was stable hereafter. Predictive equations underestimated REE in 92% of participants at all time points. Recommended feeding targets during the acute phase of illness in severely malnourished children exceeded REE. Acutely ill malnourished children are at risk of being overfed when implementing current international guidelines.

Supplementary Table S5.Parameter estimates testing association of resting energy expenditure (kcal/kg/day) trajectories overtime in children hospitalized for an acute illness with differing nutritional status with clinical variables of sepsis, high SIRS score and anemia.Piecewise mixed models were fitted with a single knot positioned at discharge and random intercepts per participant allowing to evaluate the intercepts, slopes, and differences in slopes between groups during hospitalization (i.e., during admission) versus post-discharge.Time was coded as weeks since admission binned within time points.Final models were fit using restricted maximum likelihood.Children with NW-no wasting (reference group), MW-moderate wasting, SW-severe wasting, or EM-edematous malnutrition.
Supplementary Table S6.Parameter estimates testing association of resting energy expenditure (kcal/kg/day) trajectories overtime in children hospitalized for an acute illness with differing nutritional status with clinical variables of pneumonia, diarrhea, fever and WBC (10 9 L).Median (IQR) of predicted REE estimated using 3 equations (i.e., WHO, Schofield weight, and Schofield weight and height).Percent bias is calculated as [(predicted REEmeasured REE)/measured REE] *100.Number and percentage of children with REE over-or under-estimated by more than ±10%.REE-resting energy expenditure; NW, no wasting; MW, moderate wasting; SW, severe wasting; EM, edematous malnutrition.NA, not applicable.S8.Differences between the measured resting energy expenditure accounting for non-fasting state and predicted resting energy expenditure as estimated by three equations in children hospitalized with acute illness at each time point.

Over estimation
Under estimation Anthropometry and weight-corrected REE (kcal/kg/day) and RQ values measured in ill children with different nutritional status at each time point.

Table S3 .
Parameter estimates of resting energy expenditure (kcal/kg/day) trajectories overtime in children hospitalized for an acute illness with differing nutritional status derived from piecewise mixed effect models.

Table S4 .
Parameter estimates testing association of resting energy expenditure (kcal/kg/day) trajectories overtime in children hospitalized for an acute illness with differing nutritional status with clinical variables of stunting and dehydration. M0

Table S7 .
Piecewise mixed models were fitted with a single knot positioned at discharge and random intercepts per participant allowing to evaluate the intercepts, slopes, and differences in slopes between groups during hospitalization (i.e., during admission) versus post-discharge.Time was coded as weeks since admission binned within time points.Final models were fit using restricted maximum likelihood.Children with NW-no wasting (reference group), MW-moderate wasting, SW-severe wasting, or EM-edematous malnutrition Supplementary Bias between measured and predicted resting energy expenditure as estimated by 3 equations in children hospitalized with acute illness and different nutritional status at each time point.
Median (IQR) of predicted REE estimated using 3 equations (i.e., WHO, Schofield weight, and Schofield weight and height).Percent bias is calculated as [(predicted REEmeasured REE)/measured REE] *100.⸸The reduced measured resting energy expenditure (REE) was calculated by applying a 10% reduction to the REE measured by indirect calorimetry.⸶p-value is <0.05 for all equations except Schofield weight and height equation during admission (p= 0.104).NA, not applicable. of funding and the role of the funders for the present study and, if applicable, for the original study on which the present article is based