PERFORMANCE COMPARISON OF 15 TRANSPORT VENTILATORS
as the amount of time the ventilator could function on a fully charged battery with the ventilator set to deliver a VT of 1,000 mL at an RR of 10 breaths/min and an FIO2 of 0.21.
A ventilator was considered easy to use if all the pa- rameters were clearly labeled and easily set to deliver a precise variable (eg, VT or RR). We assumed the manu- facturer’s published weight and dimensions to be accurate. Ability to ventilate without compressed gas was met if the ventilator could deliver the set minute volume (V˙ E) under each of the test conditions without a compressed gas source or an external compressor.
bilization and data collection were repeated. All other set- tings remained unchanged. These settings were repeated for each of the compliance and resistance combinations. The cardiopulmonary monitor was interfaced with a laptop computer, on which the flow, volume, and pressure data were collected and analyzed. The set ventilation parame- ters, ventilator-displayed values, and cardiopulmonary- monitor-measured values were simultaneously recorded. All measurements during the bench assessment were at atmospheric-temperature-and-pressure-dry conditions.
The ability to deliver set parameters was evaluated with a test lung (Training and Test Lung, Michigan Instru- ments, Grand Rapids, Michigan) under 3 different test conditions: high resistance with normal compliance; nor- mal resistance with normal compliance; and normal resis- tance with low compliance. High and normal resistance was achieved with resistors (Pneuflo Rp20 and Rp5, Mich- igan Instruments, Grand Rapids, Michigan). Normal and low compliance were set on the test lung (0.05 L/cm H2O and 0.02 L/cm H2O, respectively). For each condition the tested ventilator was set to deliver a VT of 500 mL at 15 breaths/min and 30 breaths/min, and a VT of 1 L at 10 breaths/min and 20 breaths/min.
VT, RR, peak inspiratory pressure (PIP), and positive end- expiratory pressure (PEEP) were measured and analyzed with a cardiopulmonary monitor (NICO, Respironics, Walling- ford, Connecticut) and its software (Analysis Plus, Respiron- ics, Wallingford, Connecticut). Ventilator performance was determined by comparing the set parameters to the measure- ments from the cardiopulmonary monitor.
Each ventilator was bench tested as follows. With re- sistance set at 20 cm H2O/L/s and compliance set at 0.05 L/cm H2O, the ventilator was connected to the test lung, and the cardiopulmonary monitor’s flow sensor was placed between the ventilator circuit and the flow resistor. VT was initially set at 500 mL, RR at 15 breaths/min, inspiratory time (TI) at 1.0 s (if setting the TI was possible on that ventilator), and PEEP at 5 cm H2O (if PEEP was available on the ventilator). The FIO2 was set at the lowest available setting, which may have been 0.21, air mix (entrainment), or 1.0, depending on the ventilator’s capabilities. Follow- ing a 10-breath stabilization period, we recorded PIP, mean airway pressure, PEEP, VT, V˙ E, RR, and the pressure, flow, and volume graphics. After that data collection, the RR was increased to 30 breaths/min and the TI was de- creased to 0.5 s. All other settings remained unchanged. Following another 10-breath stabilization period, we again recorded PIP, mean airway pressure, PEEP, VT, V˙ E, RR, and graphics. The VT was then increased to 1,000 mL, RR was decreased to 10 breaths/min, and TI was increased to 1 s. Following another stabilization period and data col- lection, the RR was increased to 20 breaths/min, and sta-
This protocol was approved by the animal care commit- tee of Massachusetts General Hospital, Boston, Massachu- setts.
Using 30-kg sheep, we evaluated each ventilator’s abil- ity to ventilate both healthy and saline-lavage lung-injured sheep. In both settings we evaluated the ventilator’s ability to maintain normal arterial blood gas values and cardio- pulmonary hemodynamics. We studied 12 sheep: 6 with normal lungs and 6 with saline-lavage lung injury. Five ventilators were evaluated on each sheep (healthy and in- jured), and each group of 5 ventilators was studied on 2 healthy and 2 injured sheep. Three groups of 5 ventilators were randomly selected.
Healthy Lung Evaluation
Each group of ventilators was randomly applied for a 60-min period to a healthy sheep. Initially, each ventilator was set at a VT of 9 mL/kg and an RR of 20 breaths/min, with a TI or peak flow setting to maintain a TI of 1.0 s. If the device was capable of applying PEEP, PEEP of 5 cm H2O was applied with 50% oxygen. The ventilator was attached to the animal’s airway, followed by a 15-min stabilization period. After stabilization we collected arte- rial and mixed venous blood samples, and measured sys- temic arterial pressure, pulmonary artery pressure, pulmo- nary capillary wedge pressure, and heart rate. Airway pressure and VT were measured at the endotracheal tube (ETT). Cardiac output was measured in triplicate, using the thermodilution technique. The ventilator was adjusted and oxygen added if needed to attempt to reach the target blood gas values (PaO2 60 –100 mm Hg, PaCO2 30 – 50 mm Hg, pH 7.30–7.50). Once we determined whether the targets could be met, the next ventilator was attached to the animal’s airway for evaluation.
Injured Lung Evaluation
During the lung-injury tests, the ventilator was initially s e t a t a V T o f 6 m L / k g , a n R R o f 3 0 b r e a t h s / m i n , P E E P o f 1 5 c m H 2 O ( i f a v a i l a b l e ) , a n d F I O 2 o f 0 . 5 0 . A g a i n , b l o o d
RESPIRATORY CARE JUNE 2007 VOL 52 NO 6