• W Weyland, A Weyland, O Gefeller, S al-Soufi, M Sydow, and U Braun.
• Department of Anesthesiology, Emergency and Intensive Care Medicine, University of Göttingen, FRG.
• Crit. Care Med. 1994 May 1;22(5):864-71.

ObjectiveTo evaluate in the laboratory a new ventilator with a built-in monitoring function for gas exchange.DesignProspective laboratory evaluation of a new mechanical ventilator.SettingGas exchange was simulated by injecting nitrogen and CO2 at a respiratory exchange ratio close to 1.SubjectsDifferent settings of controlled-mode ventilation were simulated by the gas-injection model, which permitted an independent adjustment of variables.InterventionsThe influence of respiratory variables on oxygen consumption (VO2) and the CO2 production (VCO2) agreement with a reference method was investigated.Measurements And Main ResultsIndependent variables, which may influence the accuracy and precision of the tested device, were adjusted systematically within the following ranges and resulted in 71 different combinations: FIO2 0.21 to 0.59; FIO2-mixed FEO2 0.02 to 0.05; mixed FECO2 0.02 to 0.05; inspiratory flow rate 25 to 120 L/min; inspiratory minute volume 4200 to 11,200 mL/min; tidal volume 420 to 1120 mL; respiratory rate 6 to 16 breaths/min; airway pressure 5 to 30 cm H2O. The resulting range for both VO2 and VCO2 was 140 to 390 mL/min. For each test situation, two reference measurements were taken by mass spectrometry and wet gas spirometry. Twelve measurements were taken by the gas-exchange measurement function of the tested ventilator. The two-fold standard deviation (SD) for repeated measurement differences was chosen to quantify the repeatability within each method. This repeatability coefficient gave 5.9 (4.7) mL/min for VO2 (VCO2) within the gas-exchange measurement function, compared with 3.6 (3.1) mL/min within the reference method measurements. The mean intermethod difference (test monitor--reference method) for VCO2 over all test situations was 4.1 +/- 10.7 (2 SD)%. The respective mean VO2 difference was 7.8 +/- 12.2 (2 SD)%. Eliminating test situations with inspiratory flow rates < 50 L/min (remaining n = 46) reduced the mean VO2 difference and variability by one third to 5.4 +/- 8.0 (2 SD)%.ConclusionsAssuming that limits of agreement for intermethod differences of +/- 20% are clinically acceptable, the VCO2 measurement indicates an acceptable accuracy and precision under controlled ventilation. The respective agreement for the VO2 measurement is lower, but still within the acceptable range. The systematic difference of the VO2 and the VCO2 is mainly influenced by a +8% bias in the inspiratory minute volume measurement, which seems especially susceptible to ventilator settings with inspiratory flow rates of < 50 L/min. An improvement of the minute volume detection would be desirable.

20 keywords...

hide…