2007 OPEN FORUM Abstracts
ACCURACY OF DISPLAYED VT DURING MECHANICAL VENTILATION OF A VERY LOW BIRTHWEIGHT INFANT (VLBW) LUNG MODEL.
R. M. Diblasi1, J. Zignego1, J. W. Salyer2, D. Crotwell2, P. Richardson1
Introduction: Leading neonatal ventilators employ different hardware and software in measuring and displaying VT. Accurate VT is an essential part of ventilator management. There is little published, independently verified performance data of these ventilators when VT<5 mL.
Methods: We tested the accuracy of displayed VT of 2 neonatal ventilators. Devices tested included 2 each of the Viasys Avea and Drager Babylog 8000 Plus. For the Avea, we used 5 recently redesigned sensors. For the Babylog, 2 sensors were tested. A rigid test lung with compliance (CTL)=0.32mL/cmH2O was filled with steel wool. Ventilators were equipped with a dual-heated wire disposable circuit and heated humidifier set at 39 degrees C. BTPS correction was activated for each ventilator. Mode was PC+A/C: PIP=10cmH2O, f=50/min, FIO2=0.21. The “known or actual” VT was determined by the following equation: VT=PIPxCTL. Our PIP=10cmH2O and CTL=0.32mL/cmH2O, yielding an actual VT= 3.2mL when the PIP=10cmH2O on the ventilator. For each sensor tested, the displayed exhaled VT was recorded manually for 20 breaths. Error was defined as: [(Displayed VT-Actual VT)/Actual VT] x100. Mean error rates were calculated for each ventilator tested and overall for each type of ventilator. Mean differences were tested for significance (P<0.05)using an unpaired t-test for ventilator type and ANOVA for differences between each ventilator tested.
Results: Mean(±SEM) error rates for all breaths on the Aveas = -1.2 ± 0.3% and Babylog = -16.8 ± 0.3%, p<0.05 (t-test). Mean values for each ventilator are shown in the figure. The differences between all ventilators had a p<0.001. Post-hoc Fisher’s PLSD revealed that differences within the Aveas and within the Babylogs were not significant (p=0.07, p=0.09 respectively). All other comparisons = P<0.001. Discussion: Displayed VT is a function of the signal from the flow sensor, as well as the software within the ventilator which extracts and processes these data into readout of VT on the ventilator screen. Under these conditions, the Avea equipped with the recently redesigned hot wire flow sensor is significantly more accurate than the Babylog in displaying VT. The Babylog significantly under reported the actual VT. We speculate that this could misguide clinicians that use displayed VT in determining ventilator settings, in turn leading to lung over-distension. Inaccurate measurements of VT can also lead to unreliable calculation and display of respiratory system mechanics.