Design of the bench study
The experiment was performed under simulated conditions in the NIV laboratory. A breathing simulator (series 1101, Hans Rudolph, Inc., Shawnee, Kansas) was used to test the different ventilators, which were connected to the simulator through a single standard 2-m tube with a leak in the expiratory port (Model 5804000, Intersurgical España SL, Madrid) placed at the distal end of the ventilator. The leak port was maintained throughout the experiment to simulate the intentional leak used in clinical practice in NIV with a single-limb configuration.
For the introduction of the external gas, a T-piece with a 3.5-mm diameter side port (Model 2713000. Intersurgical España SL, Madrid) was placed proximally to the ventilator first and then distally to the leak.
Three different treatment modalities were added to the NIV set-up:
- In the first model, a compressor (Mini Plus, APRES MEDICAL SL, Spain) was connected to a tube in the circuit. The flow of the compressor was monitored prior to placement, and a value of 9 l/min was obtained (Figure 1).
- The previous model was reproduced in terms of the amount of gas introduced, but a continuous gas source with a pressure reducer and flowmeter (compressed air bottle, Nippon Gases, Madrid) was used.
- Finally, in the third model, a continuous gas source was used at 4 litres per minute. Medical air was used to avoid oxidation of the simulator parts due to the use of pure oxygen.
The simulator was programmed as follows: For lung mechanics variables, compliance was set at 30 ml/cm H2O and resistance was set at 5 cm H2O/L/s in a restrictive pattern and at 60 ml/cm H2O and 18 cm H2O/L/s respectively in an obstructive pattern (21). To simulate patient effort, the amplitude was programmed at 6 cmH2O for low effort and 14 cmH2O for high effort, depending on the condition simulated. Finally, the frequency of spontaneous breathing was established at 15 breaths/min.
Four commercial ventilators (one with two types of triggers) with the same pressure support were compared: (1) Vivo 50 (General Electric, Sweden), (2) Astral 150 (ResMed, Australia), (3) Trilogy 100 (Philips Respironics, Pennsylvania), and (4) Puritan Bennett (PB) 560 (Covidien, Massachusetts). The parameters selected for each ventilator are summarized in Table 1.
Table 1. Parameterization of the ventilators in the bench study model
Model
|
PS
|
Inspiratory trigger
|
Cycle
|
Ramp
|
RFX
|
S**
|
M**
|
NS**
|
Sǂ
|
Mǂ
|
Lǂ
|
O§
|
R§
|
Trilogy
|
10*
|
2 lm
|
5lm
|
9lm
|
60%
|
50%
|
30%
|
2
|
3
|
12
|
Trilogy ATrak
|
10*
|
AUTOMATIC
|
AUTOMATIC
|
2
|
3
|
12
|
Vivo 50
|
10*
|
2
|
4
|
7
|
7
|
4-5
|
2
|
2
|
4
|
12
|
Astral 150
|
10*
|
High
|
Medium
|
low
|
High
|
Medium
|
Low
|
150 ms
|
250 ms
|
12
|
PB 560
|
10*
|
2
|
4
|
6
|
60%
|
50%
|
30%
|
2
|
3
|
12
|
* IPAP: 15; EPAP: 5, **S: Sensitive; M: Moderately sensitive; NS: Not sensitive, ǂS: Short; M: Medium; L: Long, §O: Obstructive, R: Restrictive, X: Respiratory Frequency
Signal acquisition system: An external polygraph (16Sp Powerlab, AD Instruments, Australia) equipped with two pressure transducers (model 1050) and two pneumotachographs (S300, instrumental dead space=70 ml, resistance=0.0018 cm H2O/L/s) was used. The sampling frequency was adjusted to 200 Hz, and the polygraph was connected to a personal computer equipped with Chart 7.0 software for Windows.
The ventilation periods were 1 minute (without gas, gas, without gas). This sequence was performed twice in each group of established conditions. The gas-induced asynchronies were compared by interpreting the graphs of pressure and flow time according to the SOMNONIV algorithm18.
Design of the short clinical study
The study was conducted with patients from the Pneumology Service of the Corporació Sanitària Parc Taulí (Sabadell, Barcelona) who met the following criteria: age over 18, hospital admission for acute chronic respiratory failure, home NIV (single-limb system with intentional leakage) and more than six months of use with adequate compliance (5 hours/night or moe). Patients with underlying psychiatric disease were excluded. The study was conducted during the patient's predischarge phase (the same day or the day before).
Protocol: The procedure was performed in the patient's room during his or her hospital stay to avoid additional visits. The patient was placed in the supine position, and the usual ventilator interface was placed in a single-limb system, along with the commercial ventilator to be studied. The same ventilators tested in the bench model were evaluated at the bedside in random order. The parameters of the ventilator were the same as those that the patients used at home (that is, unlike the bench study, the sensitivity of the trigger was not modified). The monitoring system was essentially the same as previously described, with the incorporation of thoracic and abdominal bands and parasternal electromyography to better evaluate asynchronies, in addition to pulse oximetry control. The gas sources were the same as those used in the bench study model. All methods were performed in accordance with the relevant guidelines and regulations. The study was approved by the hospital ethics committee of “Comité de ética e investigación con medicamentos (CEIm) del Parc Taulí de Sabadell” (ref. 2019/511). Written informed consent was obtained.
Statistical analysis
Quantitative data are expressed as means and standard deviation (µ, sd) according to the normality of the variables, which was assessed using the Shapiro Wilk test. Qualitative data are expressed as absolute frequencies. The chi-square test or Fisher's exact test was used in the absence of normality to evaluate the presence or absence of gas-induced asynchrony by comparing the variables controlled in the bench study and the ventilator model, gas source, and interindividual variability in the clinical study. The level of significance was established at p<0.05. Accepting an alpha risk of 0.05 and a beta risk of 0.2, seven subjects were necessary to recognize a significant difference, consisting in the absence of asynchronies before gas and a high presence after gas, considering a drop-out rate of 20%.