Predicting resting energy expenditure: a critical appraisal

Area of Science:

• Metabolic research
• Nutritional science
• Clinical assessment

Background:

• Commonly used resting energy expenditure (REE) prediction models include Harris-Benedict, Schofield, Owen, and Mifflin-St Jeor (based on height, weight, age, gender), and Cunningham (based on body composition).

Purpose of the Study:

• To compare the accuracy of five established prediction models for resting energy expenditure (REE) against measured reference data.
• To identify the most reliable model for predicting REE in diverse populations.

Main Methods:

• Comparison of five REE prediction models (Harris-Benedict, Schofield, Owen, Mifflin-St Jeor, Cunningham) with individual REE measurements from 353 participants across 14 studies.
• Analysis of participant characteristics to assess the range covered by the reference data.

Main Results:

• The Harris-Benedict model demonstrated the closest predictions to measured REE in white adults.
• Over 70% of white adult participants had REE estimates within 10% of measured values using the Harris-Benedict model.

Conclusions:

• Differences between predicted and measured REE may stem from measurement validity, conditions, and insufficient fasting periods (12-14 hours) failing to achieve post-absorptive states.
• The classic Harris-Benedict model is recommended for white adults, with further research needed to define post-absorptive conditions using indicators like respiratory exchange ratio for improved accuracy.