Evaluation of an apparatus for continuous monitoring of gas exchange in mechanically ventilated patients.

Indirect calorimetry has been integrated with a commercially available ventilator, permitting continuous monitoring of energy expenditure and respiratory quotient. This equipment, the ventilator Erica, the Engstrom metabolic computer EMC and the carbon dioxide analyzer Eliza (Gambro Engstrom AB, Sweden) was evaluated in laboratory tests and in critically ill patients. In the laboratory evaluation, the variability in FIO2 was less than 0.05% at FIO2<0.47. The linearity error of the oxygen sensor and the CO2 analyzer was less than 0.06%. The variability of the oxygen sensor was 0.01% O2 and of the CO2 analyzer 0.04% CO2. The CO2 analyzer was sensitive to oxygen concentration and underestimated CO2 concentration by 3.3% at FIO2=0.5 if calibrated on room air. The pneumotachometer was oxygen dependent and overestimated inspired volume by 5.5% at 100% O2, if calibrated on room air, and was negligibly affected by changes in minute volume or PEEP within the physiological range for adults. In the patient evaluation, 15 measurements of gas exchange were performed in 9 critically ill patients and the results were compared with data obtained simultaneously by the Scholander technique or mass spectrometry. The standard deviation of the mean difference in comparison with the reference methods was 9.4% for O2 uptake, 6.6% for CO2 elimination, 12.5% for respiratory quotient and 3.9% for minute volume. Paired t-test analysis showed no systematic difference between the various methods.