Evaluation of a Volume Targeted NIV Device: Bench Evaluation of the Breathe Technologies Non-Invasive Open Ventilation System (NIOV™)

Abstract

Purpose: Early ambulation in the ventilated patient is gaining wider acceptance. We evaluated a new portable (1 lb), gas powered, volume ventilator designed for NIV via a proprietary nasal pillows interface (Breathe Technologies, CA). Methods: We developed a model to approximate a patient's nose, upper airway and trachea. The model was connected to a test lung (ASL5000, Ingmar Medical, Pittsburgh, PA) via 22 mm ID corrugated tubing. The nasal pillows were adjusted in the nares using a lanyard. The ASL was set to represent a normal patient, a COPD patient, and a patient with interstitial lung disease (ILD). The Breathe ventilator was set at delivered volumes of 100 mL, 150 mL, 200 mL, and 250 mL. Baseline data was also collected without the appliance connected. Delivered volume, inspired oxygen concentration (FIO₂), inspiratory flow (V), and peak inspiratory pressure (PIP) were recorded for each breath. Data for 10 breaths were used to calculate the mean at each condition (± SD). Results: The Breathe volume ventilator delivered an augmented simulated patient tidal volume of 362 to 823 mL, augmenting the simulated patient's spontaneous volume by up to 459 mL, depending on ventilator settings and ASL lung conditions. Delivered FIO₂ ranged from 0.36 to 0.45 and was also dependent on ventilator settings and ASL lung conditions. Conclusions: The PIP, delivered tidal volumes, and measured FIO₂ support the hypothesis that this system can augment minute ventilation and supply supplemental oxygen in spontaneously breathing patients with a simple, non-invasive interface.

Keywords: :

• Rehabilitation
• COPD
• portable ventilator
• respiratory insufficiency

Introduction

Non-invasive ventilation (NIV) has changed the paradigm for treatment of patients with acute exacerbations of chronic obstructive pulmonary disease (COPD) and the long term management of patients with chronic neuromuscular disease (1–5). As a result, research on the use of NIV in a number of other disease states continues to expand (6–8).

Early ambulation of ventilated patients has also undergone a ‘rebirth’ in last few years and may represent another area where NIV may play a role (9–13). NIV could also be used to facilitate ambulation in patients with chronic disease by reducing the respiratory workload and oxygen cost of breathing during rehabilitation, and those patients participating in activities of daily living. We evaluated a portable ventilation system for NIV designed to facilitate ambulation using a benchtop model of the upper airway and lungs.

Device description

The Non-Invasive Open Ventilation (NIOV™) system (Breathe Technologies, Irvine, CA) consists of a pneumatically powered, electronically controlled ventilator operating from a 50 psig gas source. The cannula style interface is held in place by a lanyard and nasal pillows are fitted into the nares and includes 2 venturis/air entrainment ports. The Breathe NIOV ventilator (3 1/8″(h) × 1¼″(d) × 7 ½″(w), weight 1 lb) (Figure 1) provides patient triggered, volume breaths, which are time cycled. The inspiratory time is adjusted using a ‘percent delivery time’ control which bases inspiratory time on the total cycle time. All breaths are patient triggered based on a change in airway pressure. Sensitivity is adjustable from a setting of 0 to 9, which corresponds to negative pressures of approximately −0.01 to −0.34 cm H₂O below baseline. Inspired oxygen (FIO₂) delivered from the device is based on the input, typically at an FIO₂ of 1.0. Gas from the ventilator is delivered to a nasal interface which includes two venturi ports for air entrainment (Figure 2). As gas from the ventilator travels though the nozzles, room air can be entrained and increase the delivered tidal volume. Theoretically this entrainment may also add a small amount of humidity.

Figure 1.  Breathe ventilator.

Figure 2.  Components of the nasal interface. With each of the two nasal pillows there is a sense port, entrainment port, and nozzle.

Although the ventilator produces a constant volume breath that is time-cycled, because of the entrainment ports, neither the volume nor pressure delivered to the patient is controlled. The ventilator has three password-protected volume settings that are titrated by the clinician to the patient's needs at varying activity levels. The patient can select from these three volume settings as needed to support their current activity.

Methods

A model of the nose and upper airway was constructed to approximate a patient's nose, upper airway, and trachea but was not physiologically precise (Figure 3). The model was constructed based on the “Airway Larry” adult airway management training head. (http://www.enasco.com/product/LF03667U). The airway resistance of the nasal model was measured from the pressure drop across the device via an RT 200 analyzer (Timeter Instrument Corp., Lancaster, PA). The model was attached to a computerized test lung (ASL 5000, Ingmar Medical, Pittsburgh, PA) via standard 22 mm internal diameter (ID) tubing equivalent to the normal anatomic deadspace (150 mL). The test lung was set to simulate spontaneous breathing with lung conditions corresponding to normal lungs, a patient with chronic obstructive pulmonary disease (COPD) and a patient with interstitial lung disease (ILD). These characteristics are shown in Table 1.

Figure 3.  Testing model.

Table 1.  Test lung settings in spontaneous mode

	Normal Lungs	COPD	ILD
Compliance (mL/cm H2O)	100	125	50
Resistance (cm H2O/L/s)	5	10	5
Tidal volume (mL)	400	400	300
Respiratory Rate (breaths/minute)	15	15	20
Inspiratory time (secs)	1.0	0.8	1.0
I:E ratio	1:3	1:4	1:2
Pmax (cm H2O)	−5	−10	−6

Lung characteristics were altered in each condition as was the level of inspiratory effort by the test lung. The test lung was set to create a sinusoidal inspiratory flow profile. The ventilator was set at volumes of 100 mL, 150 mL, 200 mL, and 250 mL.

The percentage delivery time was set at 21% (the default) and the trigger sensitivity was set at the 4 (default) value. Baseline data was also collected using each of the lung conditions without the nasal model connected. The ASL 5000 was utilized to measure airway pressure, volume, and flow at a frequency of 512 Hz. FIO₂ was measured continuously via the integral oxygen analyzer. The oxygen analyzer was calibrated to both room air and an FIO₂ of 1.0. Delivered volume was determined by averaging the measured volume over one minute based on piston displacement of the test lung and reported as mean ± SD. We also calculated the difference between the delivered volume and the set ventilator volume to quantify the degree of air entrainment, based on the assumption that the ventilator volume settings were accurate. Peak flow and peak airway pressure are also presented as mean ± SD. Data from a minimum 10 breaths were used for the calculations.

Statistical analysis

Entrainment was expected to change with changes in set ventilator volume, inspiratory effort, and modifications in lung mechanics. We employed one-way ANOVA models to measure the differences between successive volume settings treated as categorical variables. We fit a separate model for each of the three lung conditions.

In addition, we employed linear regression analysis to measure the relationship between volume and combinations of compliance, resistance, ventilator setting, the interaction between compliance and ventilator setting, and the interaction between resistance and ventilator setting.Finally, we used a linear regression analysis to study the relationship between ventilator setting and volume. Separate models were fitted for each lung condition in order to be able to rank the strength of the relationship between those two variables.

Results

For each test lung setting, the FIO₂ was lowest at the lowest set ventilator volume and increased with each subsequent ventilator volume setting. For all ventilator volume settings, FIO_­2 was greatest at the lung setting representing the ILD condition, and FIO₂ had the greatest setting-to-setting increase with the COPD condition (Table 2).

Table 2.  Mean baseline, ventilator, entrained, and total volume (mL) and mean FIO₂ and peak inspiratory pressure (PIP) delivered under each condition.

Baseline volume	Ventilator volume	Variance from predicted volume	Total volume	FIO2	PIP (cm H2O)
365 ± 4	100	−21 ± 16	443 ± 16	0.38 ± .001	1.4 ± 0.4
365 ± 4	150	34 ± 2	549 ± 2	0.41 ± .001	2.7 ± 0.1
365 ± 4	200	123 ± 2	688 ± 2	0.43 ± .001	4.6 ± 0.1
365 ± 4	250	208 ± 3	823 ± 3	0.44 ± .001	6.6 ± 0.1
388 ± 3	100	−75 ± 24	413 ± 24	0.36 ± .001	2.9 ± 1.1
388 ± 3	150	−59 ± 3	479 ± 3	0.40 ± .001	4.4 ± 0.1
388 ± 3	200	6 ± 4	594 ± 4	0.42 ± .001	7.0 ± 0.1
388 ± 3	250	68 ± 6	706 ± 6	0.44 ± .001	9.5 ± 0.1
282 ± 3	100	−20 ± 5	362 ± 5	0.40 ± .001	1.8 ± 0.1
282 ± 3	150	38 ± 5	470 ± 5	0.44 ± .001	4.2 ± 0.2
282 ± 3	200	108 ± 10	590 ± 10	0.45 ± .001	7.0 ± 0.2
282 ± 3	250	167 ± 6	699 ± 6	0.45 ± .001	9.2 ± 0.1

The volume delivered to the test lung by the ventilator was greater than the test lung's spontaneous volume baseline for all test conditions. With each successive volume setting on the ventilator the volume delivered to the test lung increased. Volume increases of the test lung over baseline ranged from 25–458 mL, depending on the ventilator setting and test lung settings (Table 2 and Figure 4). Increasing the set ventilator volume resulted in a delivered volume to the test lung that was different from the sum of lung and ventilator volumes. This additional volume results from air entrainment (Table 2). However, at the 100 mL setting there was no entrainment and in some cases the volume delivered to the test lung was less than the sum of set ventilator volume and spontaneous breathing volume.

Figure 4.  Total delivered volume comparison with each lung model.

Entrainment was significantly greater at the 250-mL setting, thereby augmenting the delivered volume. The total volume was lowest in the COPD condition as compared to either of the other conditions. At normal lung conditions the largest entrained volume was 82% of ventilator volume (25% of total volume) at the 250 mL setting vs. 26% of ventilator volume (9% of total volume) for the COPD model at the same setting. The mean FIO₂ was 0.42 ± 0.001 (range 0.36–0.45) and the mean PIP was 5.1 cm H₂O (range 1.3–9.5). No instances of auto-triggering was observed with any of the lung models or ventilator settings.

Table 2 shows the effects of successive increments of set ventilator volume. For all lung conditions, increasing set volume from 0 to 100 mL resulted in a statistically significant increase in delivered volume but did not demonstrate entrainment. In the COPD condition, a further 50 mL increase in set volume also resulted in decreased anticipated entrainment. However, all other increases in set volume (from 150 mL to 250 mL for the COPD condition and from 100 mL to 250 mL for the other conditions) led to increases in entrained volume. The largest entrainment volumes were associated with the highest set volume of 250 mL.

To further investigate the relationship among variables, we fitted a linear regression model in which volume (V) depends upon compliance (C), resistance (R), and ventilator setting (S) and the interactions between compliance and setting (C*S) and between resistance and setting (R*S).

After dropping the non-significant C*S effect, the fitted model was with R₂ = 0.92. This equation says that volume depends positively on C and S and negatively (inversely) on R. The negative effect of R is stronger at higher vent ­settings.

An alternative approach recognizes that each C/R combination is equivalent to a specific lung model. Each model results in a different relationship between ventilator setting (S) and volume (V). The healthy lung model exhibits the strongest relationship between S and V. Specifically, each 50 mL increase in S results in a 91.5 mL increase in V. For ILD, the relationship is weaker; a 50 mL increase in S is associated with an 85 mL increase in V. Finally, the COPD condition exhibits the weakest relationship; a 50-mL increase in S results in a 64.5 mL change in V (Figure 4).

Examining the volume entrained/lost in Table 2, the relationship between the baseline volume and ventilator volume does not appear linear with increases in ventilator settings. Figure 5 shows that in fact there is a strong linear correlation between the baseline and ventilator volumes (R² = 0.99). The weakest relationship between the two volumes was with the COPD lung model but there was still a strong linear correlation (R² = 0.94).

Figure 5.  Correlation between ventilator V_T setting and anticipated entrained volume. Negative values represent total delivered volumes that were less than the sum of the ventilator V_T setting and the lung model baseline V_T.

The measured resistance of the nasal model was 1.7 cm H₂O/L/s at a flow of 40 L/min and 6.1 cm H₂O/L/s at a flow of 80 L/min. These flow rates represent the lowest and highest flows generated by the ventilator at the lowest and highest V_T settings.

Discussion

We found that the Breathe NIOV system was capable of augmenting tidal volume in a model of spontaneous breathing. The degree of tidal volume augmentation was proportional to the ventilator volume and compliance and inversely proportional to the resistance. The entrainment feature in the interface has an improved efficiency at higher tidal volumes. The average delivered FIO₂ was approximately 0.43.

Positive pressure NIV dates back to the 1930s for use in the treatment of acute pulmonary edema (14). Intermittent positive pressure breathing (IPPB) was widely used from the 1940s–1980s to deliver bronchodilators to COPD patients (15) and was studied as a modality for treating respiratory failure (16). In the 1960s NIV was used to treat patients with neuromuscular disease and with the incorporation of nasal and face masks to treat sleep apnea and muscular dystrophy from the 1980s until present (17).

Numerous observational and randomized controlled trials have shown that NIV improves quality of life and exercise tolerance in chronic restrictive and ­obstructive pulmonary disease (18–24).Ventilators suitable for NIV use have gotten smaller, lighter, and more portable over the past decade but arguably remain too bulky for someone with pulmonary disease to carry. One study found that in COPD patients, dyspnea did not improve when walking and carrying an NIV ventilator, due to the added weight (25). The NIOV device in this bench study is small enough to clip onto a person's belt or be placed in a shoulder bag and provides supplemental ventilatory support and oxygen to the user.

The NIOV system was able to augment tidal volume delivered to the test lung at all settings tested. As set ventilator volume increased, the entrained volume increased further augmenting ventilation. The largest delivered volume (spontaneous volume + ventilator volume + entrained volume) was at the 250 mL setting.

The smaller volumes delivered with the COPD model compared to the Healthy and ILD models is not counterintuitive since downstream resistance has a negative effect on the rate of entrainment, resulting in lower than expected total volume delivery. With a set compliance of 125 cm H₂O/L/s one would expect the volume to increase more across all settings with this model than with the other lung models. This would likely be the case if the airway resistance was the same with all three models. With the COPD model, the airway resistance was double that of the other models. This demonstrates that lung model resistance and the operating characteristics of the ventilator are the major determinants of the entrained volume.

For the evaluation, the NIOV device had a set inspiratory time of 21% of the total breath period with all settings, therefore an increase in delivered volume increased the inspiratory flow. At the 100 mL setting the inspiratory flow was 24 L/min and was 60 L/min at the 250-mL setting. Our data suggest that the higher driving pressure and flow facilitates entrainment. In the COPD model with higher airway resistance, entrainment was reduced and the delivered volume was less than the other models. The increase in downstream resistance results in back pressure which reduces the entrainment effectiveness of the venturi apparatus.

The NIOV device is small and lightweight but requires a 50 psig gas source necessitating the concurrent use of an oxygen cylinder. This diminishes the portability feature due to having to either carry a cylinder or having one in tow in a rolling carrier. However, most patients who would use this technology already use oxygen cylinders. Albeit better than having to manage a heavy ventilator, the required compressed gas source may limit mobility. Additionally, oxygen cylinders present a projectile hazard if dropped or mishandled. There is no humidification with the system and the entrainment of room air to add humidity may be insufficient is some cases. In a study with 18 subjects wearing the NIOV for 6 consecutive hours, the mean score was 4/5 on a Likert scale indicating that nasal dryness did not present a significant problem during the usage time (26).

Clinical trials conducted using the NIOV system support the hypothesis that providing ambulatory ventilation in conjunction with supplemental oxygen can improve exercise tolerance and dyspnea. Hilling et al. (27) demonstrated that exercise distances using NIOV increased in 25 out of 30 subjects with a mean improvement in 6 minute walk distance (6MWD) of 57 m (±54). In a follow up study of 32 subjects, a mean improvement in 6MWD distance of 36 (±34.1) meters was observed in subjects using NIOV (28). In a third study, NIOV was evaluated in 18 subjects with severe COPD at 3 pulmonary rehabilitation centers (26). Subjects reported experiencing less dyspnea, greater mobility, and improved exercise endurance compared to their current oxygen systems.

Limitations

There are several limitations to our methods. We used a lung model and while volume and FIO₂ can be characterized, the model cannot sense a reduction in dyspnea or detect improved comfort. Patient studies are needed to determine utility of the device. We used a set of compliance and resistance characteristics based on disease states, which cannot represent the entire range in the actual target population. Additionally, linear resistance was used for the lung models and may not represent the true inspiratory and expiratory airway resistance of the different lung models. We did not measure humidity, which may be an important issue in patient use.

The nasal model used in this experiment was created to resemble the human nasal airway anatomy but was not physiologically precise. Nasal resistance varies from person to person but is reported to be approximately 3 cm H₂O/L/s (29). Determining actual resistance in the human nasal cavity and upper airway is difficult due to differences in individuals’ anatomy. The differences in airway resistance at various flow rates may have affected the rate of entrainment and therefore decreasing the total volume delivered to the test lung.

Conclusions

The addition of the Breathe NIOV device augmented tidal volume delivered to a spontaneously breathing model. The tidal volume is augmented both by the ventilator setting and the entrained volume. At higher volume settings and higher compliance/lower resistance, tidal volume is enhanced. The NIOV system results in ventilation of the model and stable FIO₂ in the range of 0.36 to 0.45. This device may be useful for ambulation at home and pulmonary rehabilitation. Studies evaluating walking distance and patient comfort during exercise or ambulation need to be done.

Declaration of Interest Statement

Breathe Technologies provided research support funding to our institution for this evaluation and bench study. The authors have no other conflicts of interest to disclose. The authors are responsible for the content and the writing of this paper.