Evaluation of the effects of gas volume and composition on accuracy of volume measurement by two flow sensors and delivery by a piston-driven large-animal ventilator.

ABSTRACT

OBJECTIVE To evaluate the effects of 4 gas compositions at various volumes (simulated tidal volumes [VTs]) on accuracy of measurements obtained with 2 types of flow sensors and accuracy of gas volume delivery by a piston-driven ventilator. SAMPLE 4 gas mixtures (medical air [21% O279% N2], > 95% O2, O2-enriched air [30% O270% N2], and heliox [30% O270% He]). PROCEDURES For each gas mixture, reference VTs of 1 to 8 L were delivered into an anesthetic breathing circuit via calibration syringe; measurements recorded by a Pitot tube-based flow sensor (PTFS) connected to a multiparameter host anesthesia monitor and by a thermal mass flow and volume meter (TMFVM) were compared with the reference values. Following leak and compliance testing, the ventilator was preset to deliver each gas at VTs of 1 to 8 L into the calibration syringe. Effects of gas volume and composition on accuracy of VT measurement and delivery were assessed by ANOVA. Agreements between delivered and flow sensor-measured VT and preset versus ventilator-delivered VT were determined by Bland-Altman analysis. RESULTS Flow sensor measurements were accurate and not influenced by gas composition. Mean measurement error ranges for the PTFS and TMFVM were -4.99% to 4.21% and -4.50% to 0.17%, respectively. There were no significant differences between ventilator-delivered and reference VTs regardless of gas volume or composition. Bland-Altman analysis yielded biases of -0.046 L, -0.007 L, -0.002 L, and 0.031 L for medical air, > 95% O2, O2-enriched air, and heliox, respectively. CONCLUSIONS AND CLINICAL RELEVANCE The PTFS and the TMFVM measured VTs and the piston-driven ventilator delivered VTs with error rates of < 5% for all gas compositions and volumes tested.