Artificial Neural Network Algorithms to Predict Resting Energy Expenditure in Critically Ill Children

Abstract

Introduction: Accurate assessment of resting energy expenditure (REE) can guide optimal nutritional prescription in critically ill children. Indirect calorimetry (IC) is the gold standard for REE measurement, but its use is limited. Alternatively, REE estimates by predictive equations/formulae are often inaccurate. Recently, predicting REE with artificial neural networks (ANN) was found to be accurate in healthy children. We aimed to investigate the role of ANN in predicting REE in critically ill children and to compare the accuracy with common equations/formulae. Study methods: We enrolled 257 critically ill children. Nutritional status/vital signs/biochemical values were recorded. We used IC to measure REE. Commonly employed equations/formulae and the VCO2-based Mehta equation were estimated. ANN analysis to predict REE was conducted, employing the TWIST system. Results: ANN considered demographic/anthropometric data to model REE. The predictive model was good (accuracy 75.6%; R2 = 0.71) but not better than Talbot tables for weight. After adding vital signs/biochemical values, the model became superior to all equations/formulae (accuracy 82.3%, R2 = 0.80) and comparable to the Mehta equation. Including IC-measured VCO2 increased the accuracy to 89.6%, superior to the Mehta equation. Conclusions: We described the accuracy of REE prediction using models that include demographic/anthropometric/clinical/metabolic variables. ANN may represent a reliable option for REE estimation, overcoming the inaccuracies of traditional predictive equations/formulae.

Determining energy and protein needs in critically ill pediatric patients: A scoping review

“…One study found a positive association between age and EE. 34 A recent study also observed a positive association between EE and the presence of multiorgan failure, but the causes of PICU admission and underlying disease were not correlated with EE. 35 The systematic review 5 and three recent studies 32,35,37 showed no correlation between the disease severity and EE.…”

“…Most studies 13,35–37,39 observed that the Schofield equation, which predicted REE within ±10% of the measured REE in 30%–38% of observations only, 13,37 was one of the least inaccurate. The direction of accuracy differed among studies, with underestimation and overestimation of REE, 13,37 mainly an underestimation, 35,46 or an overestimation 31,34,36 . One study 39 showed that the bias of the Schofield equation differed along the REE range of values, resulting in the underestimation of REE in young children and the overestimation in older children, which may explain these differences.…”

“…Several studies 13,33,34,[38][39][40] assessed the simplified metabolic equation, which was developed by Mehta et al in 2015, in a derivation group of critically ill children undergoing ventilation, followed by a validation study. 41 Around 50% of estimations fell within ±10% of measured REE, and the remaining observations were an underestimation of REE.…”

“…As expected, the available data showed positive correlations between weight and EE and age and EE. 5,32,34,35,44 These variables are included in most, if not all, of the standard predictive equations of REE for children and adults. Among the dynamic clinical factors, there was good agreement for the severity of illness among the included studies, concluding in the absence of an association with EE.…”

Challenges and advances in nutrition for the critically ill child

“…Current guidelines suggest evaluating patients with suspected hypometabolism or hypermetabolism such as patients with a neurologic or oncologic diagnosis [9]. However, given the current limited applicability in the use of indirect calorimetry in large numbers of patients, the exploration into novel methods of measuring energy expenditure remains a high priority [10,11].…”