Comparing Steady State to Time Interval and Non-Steady State Measurements of Resting Metabolic Rate

The 2 most common methods to determine resting metabolic rate (RMR) with indirect calorimetry are steady state (SS) and time intervals. Studies have suggested SS more accurately reflects RMR, but further research is needed. Our objective was to compare the bias, precision, and accuracy of SS to time intervals and non-SS measurements in a healthy adult population. Seventy-seven participants were measured for 45 minutes using a Quark RMR. Inclusion criteria included healthy participants aged 18-65 years. Pregnant and lactating women were excluded. Paired t tests compared differences between measures. Bland-Altman plots were used to determine precision. Bias occurred if there was a significant difference between the means. Accuracy was determined by counting the number of absolute differences between SS compared with non-SS and time intervals that were <75 calories. Of 77 participants, 84% achieved SS, and 95% achieved SS by minute 30. Most differences between SS and time intervals were statistically but not practically significant. Bland-Altman plots showed SS measurements were generally lower than any time interval, suggesting SS is more indicative of RMR. Non-SS was significantly more biased ( P = .0005), less precise (spread of limits of agreement was 269 calories), and less accurate (65%) than SS. We conclude that non-SS is not equivalent to SS. We also conclude that using 5-minute SS is more indicative of RMR than any time interval that was tested in healthy populations. If SS cannot be achieved, we recommend repeating the measurement.