Appropriate interpretation of indirect calorimetry for determining energy expenditure of patients in intensive care units.

Energy expenditure varies with patient agitation. We asked whether energy expenditure determined from 5 minutes of measurement with less than 5% variation in minute-to-minute measurements (short mean) was similar to energy expenditure determined from a longer measurement period (long mean; average duration: 31 minutes). The difference between the short mean and the long mean was less than 5% for 42 of 47 patients in intensive care units; the difference was more than 10% for 2 patients. We also generated a statistical model simulating energy expenditure measurement variations of 5% to 20%, and calculated the necessary measurement durations required to achieve a 3% error in measurement. The statistical model showed that for energy expenditure variations of 5%, 10%, 15%, and 20%, the necessary durations of measurement to achieve a 3% error were 3, 9, 16, and 25 minutes, respectively. We conclude that energy expenditure may be determined using the mean of a 5-minute period of measurement if variation in that measurement is less than 5%. Larger variation requires longer periods of measurement.