Predictive energy equations for spinal muscular atrophy type I children

Background Knowledge on resting energy expenditure (REE) in spinal muscular atrophy type I (SMAI) is still limited. The lack of a population-specific REE equation has led to poor nutritional support and impairment of nutritional status. Objective To identify the best predictors of measured REE (mREE) among simple bedside parameters, to include these predictors in population-specific equations, and to compare such models with the common predictive equations. Methods Demographic, clinical, anthropometric, and treatment variables were examined as potential predictors of mREE by indirect calorimetry (IC) in 122 SMAI children consecutively enrolled in an ongoing longitudinal observational study. Parameters predicting REE were identified, and prespecified linear regression models adjusted for nusinersen treatment (discrete: 0 = no; 1 = yes) were used to develop predictive equations, separately in spontaneously breathing and mechanically ventilated patients. Results In naive patients, the median (25th, 75th percentile) mREE was 480 (412, 575) compared with 394 (281, 554) kcal/d in spontaneously breathing and mechanically ventilated patients, respectively (P = 0.009).In nusinersen-treated patients, the median (25th, 75th percentile) mREE was 609 (592, 702) compared with 639 (479, 723) kcal/d in spontaneously breathing and mechanically ventilated patients, respectively (P = 0.949).Both in spontaneously breathing and mechanically ventilated patients, the best prediction of REE was obtained from 3 models, all using as predictors: 1 body size related measurement and nusinersen treatment status. Nusinersen treatment was correlated with higher REE both in spontaneously breathing and mechanically ventilated patients. The population-specific equations showed a lower interindividual variability of the bias than the other equation tested, however, they showed a high root mean squared error. Conclusions We demonstrated that ventilatory status, nusinersen treatment, demographic, and anthropometric characteristics determine energy requirements in SMAI. Our SMAI-specific equations include variables available in clinical practice and were generally more accurate than previously published equations. At the individual level, however, IC is strongly recommended for assessing energy requirements. Further research is needed to externally validate these predictive equations.