Objectives: We designed a series of experiments to determine whether expiratory water condensate (PconCO2) can be used as a proxy for mixed expired gas collection.
Methods: In 18 adult mechanically ventilated patients with ARDS (40 samples), simultaneous collections of arterial blood, expiratory water trap condensate, mixed expired gas, and minute ventilation were used to calculate VCO2 and VD/VT. To assess the effect of temperature, a constant gas flow (PCO2 10-30 mm Hg) was bubbled through water at temperatures of 19.5-37 degrees C. Gas and water samples were collected, immediately analyzed for PCO2, and a temperature correction factor was calculated. A lung model was constructed using a 5 L anesthesia bag connected to a mechanical ventilator with a heated humidifier. Temperature at the Y-piece was set to approximately 37 degrees C and CO2 was injected into the bag to establish an end-tidal PCO2 of 20-70 mm Hg. After equilibration, condensate was collected, PCO2 was measured, and the temperature-corrected PCO2 was compared to PECO2. The capnogram at points along the expiratory limb circuit was used to evaluate gas mixing.
Results: There was an over-estimation of PECO2 by PconCO2 (p < 0.001) for the patient data, resulting in an underestimation of VD/VT (p < 0.001) and an overestimation of VCO2 (p < 0.001). The temperature correction factor for PCO2 in water was -0.010 (about half of the factor used for whole blood). The bias between temperature-corrected PconCO2 and PECO2 was 0.3 +/- 3.2 mm Hg in the lung model. Mixing in the expiratory limb was poor, as evaluated by the capnogram.
Conclusions: Even with temperature correction, we failed to precisely predict PECO2 from PconCO2. For measurement of VD/VT and VCO2, we do not recommend methods that use PconCO2.