The Science Journal of the American Association for Respiratory Care

2008 OPEN FORUM Abstracts


Dave N. Crotwell1, Rob DiBlasi1,2, Don Foubare1, John W. Salyer1

Introduction: Reports from clinicians led us to believe there were significant problems with the use of Viasys hot wire sensors when Vt < 5mL. These included displayed Vt inaccuracy and discrepancy between inspiratory and expiratory Vt (Vti and Vte). These problems varied from ventilator to ventilator. We evaluated the Vti to Vte error and accuracy of Vte across a large number of ventilators. Round one of testing was reported to Viasys and resulted in the development of a zero-flow calibration software upgrade for the hot-wire sensor. We then conducted a 2nd test and report these results.

Methods: Testing was divided into pre and post zero flow calibration software upgrades (Version 3.7A vs 3.9A). We used a rigid test lung, with compliance (CTL)=0.315 ml/cmH20. For the pre-calibration study we tested 26 Aveas using the same new hot-wire sensor, without the zero calibration. For the post calibration study we tested 15 Aveas, using one new sensor. All ventilators were pulled from our working fleet. Test conditions were: EST passed, ETT 2.5 mm ID, gas comp ATPD, humidifier active Off, Mode Pressure AC, Delta P 15 cmH2O, TI 0.35 sec, PEEP 3 cmH2O, FIO2 0.21. Ventilators were equipped with a dry customary circuit. The "actual" Vt (Vtactual) was determined by the following equation: Vt = Delta P x CTL, or 15 cmH20 x 0.315 ml/cmH20, yielding Vtactual = 4.7 mL. Data obtained were Vti and Vte from the Avea display after a five minute warm up period. For both pre and post calibration tests, we calculated the following error rates: 1) Vti to Vte, 2) Vti to Vtactual, and 3) Vte to Vtactual. Mean differences in error rates were tested using Mann Whitney unpaired two group means test with significance established as p < 0.05.

Results: There were clinically important and statistically significant improvements in Vt accuracy after the zero flow calibration. Results are described in the figure below.

Discussion: There are important implications of this work. The Vti to Vte discrepancies on the Avea led to frequent nuisance alarms indicating excessive leak. The inaccuracy of displayed Vt < 5 mL before the zero calibration procedure often caused clinical confusion when clinical presentation conflicted with displayed Vt. Our impression now is that nuisance alarms have significantly decreased, although we have no way of quantifying this. Complaints from RT's and neonatologist about clinical presentation conflicting with displayed Vt have all but disappeared.