A critical view of the use of predictive energy equations for the identification of hypermetabolism in motor neuron disease

Abstract

Background and Aims People living with motor neuron disease (MND) frequently struggle to consume an optimal caloric intake. Often compounded by hypermetabolism, this can lead to dysregulated energy homeostasis, prompting the onset of malnutrition and associated weight loss. This is associated with a poorer prognosis and reduced survival. It is therefore important to establish appropriate nutritional goals to ensure adequate energy intake. This is best done by measuring resting energy expenditure (mREE) using indirect calorimetry. However, indirect calorimetry is not widely available in clinical practice, thus dietitians caring for people living with MND frequently use energy equations to predict resting energy expenditure (pREE) and estimate caloric requirements. Energy prediction equations have previously been shown to underestimate resting energy expenditure in over two-thirds of people living with MND. Hypermetabolism has previously been identified using the metabolic index. The metabolic index is a ratio of mREE to pREE, whereby an increase of mREE by ≥ 110% indicates hypermetabolism. We propose that the use of energy prediction equations to inform a metabolic index to indicate hypermetabolism in people living with MND is inappropriate and results in a biased identification of hypermetabolism in lighter individuals. Methods mREE was derived using VO2and VCO2measurements from a GEMNutrition indirect calorimeter. pREE was estimated by Harris-Benedict (HB) (1919), Henry (2005) and kcal/kg/day predictive energy equations. The REE variation, described as the percentage difference between mREE and pREE, determined the accuracy of pREE ([pREE-mREE]/mREE) x 100), with accuracy defined as ± 10%. A metabolic index threshold of ≥ 110% was used to classify hypermetabolism. All resting energy expenditure data are presented as kcal/24hr. Results Sixteen people living with MND were included in the analysis. The mean mREE was 1642 kcal/24hr ranging between 1110 and 2015 kcal/24hr. When REE variation was analysed for the entire cohort, the HB, Henry and kcal/kg/day equations all overestimated REE, but remained within the accuracy threshold (mean values were 2.81% for HB, 4.51% for Henry and 8.00% for kcal/kg/day). Conversely, inter–individual REE variation within the cohort revealed HB and Henry equations both inaccurately reflected mREE for 68.7% of participants, with kcal/kg/day inaccurately reflecting 41.7% of participants. Whilst the overall cohort was not classified as hypermetabolic (mean values were 101.04% for HB, 98.62% for Henry and 95.64% for kcal/kg/day), the metabolic index ranges within the cohort were 70.75%–141.58% for HB, 72.82%–127.69% for Henry and 66.09%–131.58% for kcal/kg/day, indicating both over– and under–estimation of REE by these equations. We have shown that pREE correlates with body weight (kg), whereby the lighter the individual, the greater the underprediction of REE. When applied to the metabolic index, this underprediction biases towards the classification of hypermetabolism in lighter individuals. Conclusion Whilst predicting resting energy expenditure using the HB, Henry or kcal/kg/day equations accurately reflects derived mREE at group level, these equations are not suitable for informing resting energy expenditure and classification of hypermetabolism when applied to individuals in clinical practice.