Comparison of predictive equations and indirect calorimetry in critical care: Does the accuracy differ by body mass index classification?

Abstract

Background: Nutrition support professionals are tasked with estimating energy requirements for critically ill patients. Estimating energy leads to suboptimal feeding practices and adverse outcomes. Indirect calorimetry (IC) is the gold standard for determining energy expenditure. However, access is limited, so clinicians must rely on predictive equations. Methods: A retrospective chart review of critically ill patients who underwent IC in 2019 was conducted. The Mifflin–St Jeor equation (MSJ), Penn State University equation (PSU), and weight-based nomograms were calculated using admission weights. Demographic, anthropometric, and IC data were extracted from the medical record. Data were stratified by body mass index (BMI) classifications, and relationships between estimated energy requirements and IC were compared. Results: Participants (N = 326) were included. Median age was 59.2 years, and BMI was 30.1. The MSJ and PSU were positively correlated with IC in all BMI classes (all P < 0.001). Median measured energy expenditure was 2004 kcal/day, which was 1.1-fold greater than PSU, 1.2-fold greater than MSJ, and 1.3-fold greater than weight-based nomograms (all P < 0.001). Conclusion: Despite the significant relationships between measured and estimated energy requirements, the significant fold-differences suggest that using predictive equations leads to significant underfeeding, which may result in poor clinical outcomes. Clinicians should rely on IC when available, and increased training in the interpretation of IC is warranted. In the absence of IC, the use of admission weight in weight-based nomograms could serve as a surrogate, as these calculations provided the closest estimate to IC in participants with normal weight and overweight, but not obesity.