2005 OPEN FORUM Abstracts
ACCURACY OF TIDAL VOLUME (VT) DISPLAYED DURING VOLUME-TARGETED VENTILATION IN NEONATAL VENTILATORS.
John Salyer, RRT, Cary Jackson RRT. Respiratory Care Service, Children's Hospital and Regional Medical Center, Seattle Washington.
Introduction: There is a growing emphasis on VT -targeted ventilation in neonates. Owing to known problems with the accuracy of displayed VT in some widely used neonatal ventilators, we sought to test the accuracy of some ventilators that are marketed as "Neonatal", during simulated VT-targeted neonatal ventilation.
Methods: Ventilators tested included the Viasys Avea, Puritan Bennet PB 840, and the Maquet Servo-i. Two different versions of the Servo-i were tested. After the first test (Servo-i-1) the manufacturer approached us with new software they claimed improved the accuracy of their volume measurements in the neonatal range. Our true or actual VT was determined independently of the ventilator display with a pneumotachometer (Cosmo Plus, Novametrix Inc.) placed at the proximal airway using data acquisition software (Analysis Plus, Novametrix Inc). The Cosmo was operated & calibrated in accordance with manufacturer's recommendations. VT displayed from the ventilators was recorded
|Percent Error Rates|
|Total Error||High VT Error||Low VT Error|
|All data are mean values|
|VT data for each ventilator, all data combined|
|Avea||6.7 ± 2.9||6.2 ± 2.8|
|Galileo||10.0 ± 0.0||12.4 ± 0.1|
|PB 840||5.6 ± 2.2||3.7 ± 1.7|
|Servo-i||8.4 ± 3.0||6.8 ± 1.4|
|Servo-i-1||5.2 ± 1.9||9.2 ± 2.5|
|All data are mean ± SD|
manually and then transcribed to a spreadsheet. The ventilators were set to operate in volume-targeted, decelerating flow modes, e.g. AVEA in the pressure-AC mode with machine volume, the Gallileo in the adaptive pressure ventilation mode (APV), the Servo-i in pressure regulated volume control for both tests. Ventilators were set to deliver either 5 mL (low VT) or 10 mL (high VT) settings, PEEP = 5 cmH20 and f = 15/min. The Galileo was only tested in the high VT setting because 10 mL is the lowest volume setting on the device. The Avea was tested using the hot-wire sensor. The test lung used was a 1 L rigid plastic bottle to which water was added to obtain a compliance of 0.4 mL/cmH2O. Ventilators were equipped with a dual-heated wired disposable circuit and heated humidifier set at 39 ° C. Each ventilator was tested under each condition for @ 3 minutes. Percent error was calculated as [(Mean Displayed VT)-(Mean Actual VT)]¸[Mean Actual VT]. Mean error rates of different ventilators and conditions were calculated and compared using ANOVA repeated measures test with statistical significance established as P< 0.05.
Results: A total of 493 breaths were obtained. Mean and SD values for displayed and actual VT are in top table, while the bottom table has mean values for error rates (ANOVA P < 0.001).
Discussion: There are important differences in the displayed VT accuracy of these neonatal ventilators in volume-targeted modes. The improvements in software for the Servo-i did not appear to improve its performance in this mode of ventilation. The 2 ventilators that measure flow at the proximal airway appear to be more accurate (Avea & Galileo). We only tested one ventilator of each type, which will limit the inference that can be made about the performance all ventilators of each type. We have conducted testing in other modes and will report on these in the future.