The Science Journal of the American Association for Respiratory Care

2011 OPEN FORUM Abstracts


Joseph Orr, Lara M. Brewer; Anesthesiology, bioengineering laboratory, University of Utah, Salt Lake City, UT

Background: Measurement of CO2 excretion (VCO2) during non-invasive positive pressure ventilation (NPPV) is difficult because much of the exhaled CO2 is leaked out at the mask seal and the mask port before it can be analyzed by the volumetric capnometry sensors. We have developed an algorithm to compensate for mask leaks and calculate VCO2 during NPPV. This algorithm estimates and compensates for leak, modifies the measured PetCO2 to compensate for mixing in the mask volume and calculates the compensated VCO2. Because the vented port allows so much of the expired gas to escape prior to reaching the CO2 sensor, an additional compensation step may be needed. Methods: We used a styrofoam mannequin head to which we glued a size large Performax mask (Philips-Respironics, Murraysville, PA) to simulate a mask-patient interface. Flow through the mannequin mouth was directed to a test lung (TTL, Michigan Instruments, MI) using a 6.5mm endotracheal tube. CO2 gas was infused into the test lung using a precision mass flow controller (Alicat 1-SLPM-D, Alicat Scientific, Tucson Arizona). The volumetric capnometry sensors (Flow/CO2) (NM3, Philips-Respironics, Wallingford, CT) were connected between the vented elbow connector and the exhalation port. The lung was ventilated using a V60 ventilator (Philips-Respironics, Carlsbad, CA) set at IPAP of 18 and EPAP of 4 cm H2O. Respiratory rate was set to 10 breaths per minute. Lung compliance was set to 50 ml/cm H2O. The exhalation port was connected between the sensors and the hose. CO2 was infused into the test lung at 100, 150, 200, 250, 300, 350 and 400 ml/minute for 10 minutes at each infusion rate. Flow and CO2 data were collected using the volumetric capnometer (NM3, Respironics/Philips, Wallingford, CT) interfaced to a computer. The exported waveforms were processed using a custom windows program written to implement the compensation algorithm. The resulting compensated VCO2 measurement was compared to the actual CO2 infusion rate. Results: The average measured inspired tidal volume was 2600 ml and the average measured expired tidal volume was 87 ml, which indicated that 96.7% of the inspired gas leaked out without being directly analyzed by the sensors. The table below shows the measured and infused CO2 rates. With the compensation applied, the average error was reduced to -3.1%. Conclusions: This data shows that the leak and mask mixing compensation algorithm does well in spite of extremely large leaks.
Sponsored Research - Philips/Respironics