Comparison of the Beacon and Quark indirect calorimetry devices to measure resting energy expenditure in ventilated ICU patients

Abstract

Critically ill patients in the Intensive Care Unit (ICU) should receive nutritional support matched to their metabolic needs as both under- and overfeeding energy has been shown to increase mortality. Critical illness can significantly affect metabolism. Consequently, resting energy expenditure (REE) can vary markedly during critical illness. Therefore, indirect calorimetry to estimate REE is recommended to determine energy requirements in individual ICU patients and to guide optimal nutritional support. Currently, the Quark metabolic monitor is considered the gold standard in our ICU, but novel mechanical support devices are also equipped with indirect calorimetry functionalities. This study aimed to evaluate the performance of a currently unevaluated device. A cross-sectional analysis in mechanically ventilated patients was conducted in a mixed medical-surgical ICU. The primary outcome was a numerical and visual comparison of the performance of the Beacon indirect calorimeter to calculate REE compared to the Quark device using Bland Altman plots. Performance was evaluated using bias, precision, accuracy, and reliability. Secondary analysis included a comparison with REE estimated by predictive equations. Seventy-one measurements were obtained in 27 mechanically ventilated subjects. An underestimation by the Beacon device in calculated REE of-96.2kcal/day (4.5%) was found. There was a bias towards higher VCO2 and lower VO2 values with Beacon as compared to Quark. The reliability of the Beacon was good, with an absolute intraclass correlation coefficient of 0.897 (95%CI 0.751-0.955; p=0.000). There was a poor correlation (<0.40) between the separate indirect calorimetry devices and most predictive equations. Only the Faisy predictive equations had good reliability (ICC 0.687, p=0.002). Beacon indirect calorimetry accurately determined REE in mechanically ventilated critically ill patients compared to the gold standard in our ICU (Quark indirect calorimeter), although confidence intervals were wide. There was low bias and good reliability. On the other hand, predictive equations performed poorly compared to both devices, underestimating the true metabolic needs of mechanically ventilated ICU patients.