Evaluation of Venturi Adapters for CPAP

Abstract

Background:

CPAP therapy has been shown to effectively reduce the risk of obstructive apnea and oxygen desaturation in procedural sedation. An economical approach to administering CPAP is through a Venturi adapter, which amplifies oxygen flows to produce pressures suitable for CPAP. The objective of this study was to evaluate the performance of four Venturi adapters across clinically relevant metrics including pressure-generation efficiency, inspired oxygen concentration (F_IO2), dilution of the end-tidal carbon dioxide (E_tCO2) sample for capnometry, and noise levels produced.

Methods:

A benchtop simulator was used to evaluate Venturi adapters. The simulator was composed of a test lung, corrugated tubing to simulate a trachea, a silicone-molded face, an air-cushioned mask strapped to the face, a high-pressure oxygen source, and several Venturi adapters. The oxygen flows required to produce specific airway pressures were measured to evaluate pressure-generation efficiency, and the noise levels produced by the Venturi adapters at these pressures were recorded. CO₂ was infused into the test lung while the lung was ventilated, and the E_tCO2 measured at the mask was compared with a reference measurement taken from the trachea. F_IO2 was evaluated by ventilating the test lung and observing the steady-state oxygen concentration in the lung.

Results:

Each of the Venturi adapter designs provides their unique strengths and weaknesses. There appears to be a direct trade-off between high pressure-generation efficiency and high F_IO2. There also appears to be a direct trade-off between high F_IO2 and the extent of dilution in the E_tCO2 sample. Some Venturi adapters produce high noise levels, which might result in patient discomfort and clinician unwillingness to use.

Conclusion:

Variations in Venturi adapter designs create different clinical benefits. Clinicians should choose the Venturi adapter to best suit their present clinical needs.

Introduction

It is well established that many drugs used in procedural sedation result in patients experiencing respiratory depression, which increases the risk for arterial oxygen desaturation.^1–4 An additional risk these sedatives pose to patients is inducing airway obstruction, where the tissues in the airway may partially or completely occlude the airway, further increasing the possibility of oxygen desaturation. The American Society of Anesthesiologists estimates that airway obstruction and oxygen desaturation are the leading causes of morbidity associated with sedation.⁵ To mitigate the risk of oxygen desaturation during sedation, the standard of care is to administer intraoperative supplemental oxygen.^5,6 This is traditionally done with a nasal cannula or face mask. However, when airway obstruction occurs, the benefits of supplemental oxygen may be negated when oxygen is unable to pass through the airway to the lungs.

The standard treatment for airway obstruction is CPAP or noninvasive ventilation, which involves the noninvasive application of pressure into a tight-fitting mask over the patient’s face to generate a positive pressure that opens the airway, allowing passage of oxygen to the lungs.⁷ Positive airway pressure (PAP) therapy has been shown to significantly reduce the occurrence of apnea and oxygen desaturation during procedural sedation when administered with a noninvasive ventilator.⁸ Another method of administering CPAP is through a Venturi adapter, which is a smaller, lower-cost device that amplifies oxygen flows through jet mixing and through ambient air entrainment.^9,10 As oxygen flows into the device, a narrow nozzle increases the velocity of the flow. The high velocity causes jet mixing, where viscous shearing occurs between the high-velocity oxygen jet and the low-velocity ambient gas, and the ambient gas is drawn into the oxygen jet, resulting in amplified flows (a blend of oxygen and room air) that are sufficient for administering CPAP.¹⁰ The small and simple hardware mechanism of a Venturi adapter provides a cost-effective and feasible alternative to a noninvasive ventilator.^9,11

A mathematical model of a CPAP system utilizing a Venturi adapter demonstrated that design features intrinsic to specific Venturi adapters, such as nozzle placement, nozzle resistance, and inner diameter, may influence performance metrics such as F_IO2, the amount of flow amplification, the expiratory resistance to patient exhalation, and pressure-generation efficiency.¹² Pressure-generation efficiency refers to the amount of oxygen flow required for a Venturi adapter to generate positive airway pressures, with a more efficient Venturi adapter requiring less oxygen flow to achieve a target pressure compared with a less efficient Venturi adapter. Some of these design features can also influence the audible noise produced by the high-velocity oxygen jet, as well as the amount of CO₂ rebreathing that may occur.^13,14 A CPAP system with a Venturi adapter is most desirable for clinical use if it produces low-amplitude noise and a high F_IO2 with high pressure-generation efficiency. High levels of audible noise can be irritating (and unhealthy) to the patient as well as the clinician.¹⁵ A high F_IO2 provides a longer duration for the patient to safely tolerate apnea before oxygen desaturation occurs.¹⁶ The alveolar F_IO2 may also be related to the rate at which oxygen mixes in the mask, as a significant portion of the patient’s tidal volume is composed of gas inhaled directly from the mask. High pressure-generation efficiency, where minimal pressurized oxygen flow is required to generate a positive pressure for CPAP, is desirable for the general conservation of oxygen, for example, when oxygen supplies are limited, such as when gas cylinders are the high-pressure oxygen source.¹⁷ A desirable Venturi adapter for CPAP also minimizes CO₂ rebreathing and creates low expiratory resistance so the patient can comfortably exhale if flow is discontinued. Research has demonstrated that the administration of supplemental oxygen dilutes the CO₂ sample during exhalation, and thus, the true end-tidal CO₂ (E_tCO2) is not available to be monitored.¹⁸ This also adversely impacts breathing frequency monitoring, which is generally monitored based on the CO₂ waveform. Therefore, minimizing dilution of the CO₂ sample to allow for better CO₂ monitoring is also desirable in a Venturi adapter CPAP system.

Detailed comparisons of the performance in these aspects among different Venturi adapters are needed so clinicians can make informed decisions regarding which device will be most beneficial for a specific patient. The purpose of this study was to examine the design differences of various commercial Venturi adapters for CPAP and evaluate their performance in relevant clinical aspects.

QUICK LOOK

Current knowledge

Drugs used during procedural sedation induce respiratory depression and can lead to airway obstruction and oxygen desaturation. CPAP therapy with a noninvasive ventilator has been shown to effectively reduce the incidence of apnea and oxygen desaturation during sedation. Using a Venturi adapter is an alternative for administering CPAP at a lower cost and smaller footprint.

What this paper contributes to our knowledge

The four CPAP Venturi adapters studied were found to vary in their design, which affected efficiency in generating pressure from oxygen flow, noise levels produced, delivered F_IO2, and dilution of end-tidal CO₂ readings. The noise levels of some of the adapters were found sufficiently high as to be potentially detrimental to patients or caregivers, depending on the duration of noise exposure.

Methods

This study examined four commercial Venturi adapters for CPAP. These adapters include the Boussignac CPAP device (Vygon, 5570.13, Ecouen, France), the o_two Single Use CPAP device (o_two controlled ventilation, 01CV0211-CS, Mississauga, Canada), the Flow-Safe II device (Mercury Medical, 1057203, Clearwater, Florida), and the SoPAP device (Dynasthetics, Salt Lake City, Utah). Details about each of these Venturi adapters are included in Supplementary Figure S1 in Supplementary Data. Primary features of clinical interest among these commercial Venturi adapters include flow amplification, expiratory resistance, oxygen mixing in the mask, F_IO2 characteristics, noise levels, pressure-generation efficiency, dilution of E_tCO2, and potential CO₂ rebreathing.

Lung simulator and venturi adapter CPAP system

The bench setup to evaluate the Venturi adapter CPAP systems closely follows the system described by Blackwell et al.¹² A test lung (Vent Aid TTL, Michigan Instruments, Grand Rapids, Michigan) was used to simulate lung behavior. It has adjustable compliance, and the manufacturer reports that it has a functional residual capacity similar to an adult human lung (∼1 L per lung). The lung compliance was set to 0.05 L/cm H₂O to represent normal lung compliance.¹⁹ The test lung was modified with a software-controlled servo motor that raised and lowered the test lung bellows to simulate spontaneous breathing. The motion of the motor could be adjusted to simulate various tidal volumes above the expiratory pressure, various inspiratory times, and various breathing frequencies. A 300 mm length of corrugated tubing with a volume of ∼150 mL was used to simulate a trachea and was connected in series with a pneumatic flow resistor (Rp 20 Pneuflo Resistor, Michigan Instruments) into the back of a silicone-molded face with a resin-printed nose embedded inside. The flow resistor was chosen to simulate an airway resistance slightly higher than a normal airway resistance since this fits the profile of patients who would benefit from CPAP therapy. An air-cushioned mask was strapped to the silicone face. The Venturi adapter being studied connects to the mask, and its inlet flow port is connected to an oxygen flow regulator that regulates flow out of a compressed oxygen source. A diagram is shown in Figure 1, along with a photograph of the silicone face.

Fig. 1.

Right: A diagram of the bench setup for a Venturi adapter CPAP system. The CPAP is administered by controlling oxygen flow into a Venturi adapter that is interfaced with a mask. The mask is attached to a simulated patient comprised of a silicone-molded face, corrugated tubing for a trachea, and a test lung. Left: A photograph of the silicone-molded face.

Flow amplification, expiratory resistance, and oxygen mixing in the mask

Flow amplification

The flow amplification factor (scale factor relating input oxygen flow to total output flow) of a Venturi adapter can be calculated based on the amount of oxygen dilution from the air in the unrestricted output flow.¹² A higher amplification factor means there is a higher ratio of air to oxygen, resulting in more oxygen dilution and a lower oxygen fraction measured from the amplified flow jet. The amplification factor for each adapter was calculated and recorded using the measured oxygen concentration of the amplified flow (Supplementary Fig. S2 in Supplementary Data) when a 10 L/min oxygen flow was input to each adapter. The oxygen concentration of the output flow was sampled using a fast oxygen analyzer (Capnomac Ultima, Datex, Helsinki, Finland).

Expiratory resistance

To test the expiratory resistance, a 30 L/min oxygen flow was passed through a gas flow analyzer (VT Plus, Fluke Biomedical, VT+HFUS120, Cleveland, Ohio) into the open air. The pressure in the flow analyzer was recorded. A Venturi adapter was then placed on the outlet of the flow analyzer, and the change in pressure was recorded. The expiratory resistance was then calculated in units of cm H₂O/L/s by dividing the pressure change by the flow in L/s. This was repeated for each Venturi adapter. This test was repeated with a 60 L/min flow through the gas flow analyzer.

Oxygen mixing

The characteristics for mask oxygen mixing for each adapter were evaluated by placing a Venturi adapter onto the mask such that there was no flow administered and the oxygen concentration in the test lung and mask was equal to that of room air (∼21%). Oxygen flow was then applied until 5 cm H₂O of steady-state pressure was generated (a normal CPAP level for clinical use) for a given adapter and applied to the system without any spontaneous breathing from the test lung. The oxygen waveform in the mask was observed on the fast oxygen analyzer until the oxygen level reached a steady state. The oxygen fraction was recorded as the mask F_IO2 and the time elapsed from the application of the flow to the settled oxygen level was recorded as the rise time. This process was repeated for each of the Venturi adapters.

Noise evaluation

The noise level produced by each of the Venturi adapters was measured in decibels measured on the a-weighted scale (dBA) using a sound level meter (Digital Sound Level Meter, Bafx Products, BAFX3370, Muskego, Wisconsin). The baseline noise level of the room was recorded for reference. Two separate tests were conducted to measure the noise when the adapter flow output was directed into the ambient air without being interfaced with the simulated patient, as well as when it was interfaced with the mask that was attached to a simulated patient’s face. For the first test, the decibel meter was placed 1 m away from the Venturi adapter, which was directed to the ambient air. Constant oxygen flows ranging from 5 to 30 L/min in increments of 5 L/min were applied to each of the Venturi adapters, and the steady-state noise levels were measured and recorded. For the second test, the flow into each Venturi adapter was adjusted to achieve steady-state pressures in the mask ranging from 5 to 20 cm H₂O in increments of 5 cm H₂O. Mask pressures were measured using the gas flow analyzer (VT Plus, Fluke Biomedical). The noise level in dBA was recorded from 1 m away for each adapter at each pressure.

Pressure-generation efficiency

During the second test for noise evaluation described previously, the input oxygen flow required to generate the desired mask pressure was measured and recorded. This was done for all tested flows for each tested Venturi adapter.

CO₂ rebreathing and E_tCO2 dilution

CO₂ was infused into the test lung with a mass flow controller (Alicat Scientific, Tucson, Arizona), whereas the test lung was set to breathe at 12 breaths/min with a 500 mL tidal volume and an inspiratory-expiratory ratio (I-E ratio) of 1:2 to simulate normal stable breathing. The CO₂ was monitored using a fast carbon dioxide analyzer (Capnomac Ultima, Datex), and CO₂ was sampled from the mask and the trachea. The mass flow controller was adjusted until an E_tCO2 of ∼35–40 mm Hg was observed in the mask when no Venturi adapter was attached. The Venturi adapter was then placed on the mask with no oxygen flow administered, and the changes in E_tCO2 and F_iCO2 (fraction of inspired CO₂) were recorded after 2 min from both the mask and the proximal end of the trachea. The input oxygen flow was then adjusted to achieve a mask pressure of 5 cm H₂O. The E_tCO2 and F_iCO2 were then recorded from the mask and the trachea. The trachea served as the reference for the true E_tCO2. This was then repeated with 10 cm H₂O of mask pressure. The entire process was then repeated with each of the Venturi adapters.

F_IO2

Each adapter was tested at steady-state mask pressures of 6 and 10 cm H₂O. The input oxygen flow was adjusted until the desired steady-state mask pressure was reached with a constant input oxygen flow. The test lung was set to simulate spontaneous breathing at various desired breathing frequencies, inspiratory times, and tidal volumes that may be seen in sedated patients. Oxygen levels in the lung were monitored using a fast oxygen analyzer (Capnomac Ultima, Datex). When a stable oxygen concentration was observed in the lung, the oxygen concentration of the gas in the test lung was sampled and recorded. Because the test lung does not consume any oxygen, the steady-state oxygen concentration in the lung is equal to the F_IO2. This process was repeated for all four Venturi adapters for all combinations of selected breathing frequencies, inspiratory times, and tidal volumes. Breathing frequencies tested were 6 and 12 breaths/min. Inspiratory times tested were 0.7, 1.0, and 1.3 s. Tidal volumes of 250 mL, 500 mL, and 750 mL were tested at a CPAP of 6 cm H₂O. Tidal volumes of 250 mL, 500 mL, and 600 mL were tested at a CPAP of 10 cm H₂O.

Results

Flow amplification, expiratory resistance, and oxygen mixing in the mask

The results for flow amplification, expiratory resistance, and oxygen mixing in the mask are reported in Supplementary Table S3 in Supplementary Data. The o_two CPAP device had the highest amplification factor of 13.2, and the Boussignac had the lowest of 7.9. The SoPAP adapter had a much higher expiratory resistance than the rest. The Boussignac and Flow-Safe II had equal expiratory resistances, and the expiratory resistance of the o_two device was slightly higher. At 5 cm H₂O of CPAP, the Boussignac achieved the highest maximum oxygen fraction in the mask and achieved it in only 7 s. The Flow-Safe II attained the lowest oxygen fraction in the mask (79%). The o_two device attained 90% oxygen in the mask, but it took 14 s for it to reach that level. The SoPAP adapter achieved almost as high of oxygen as the Boussignac, but in a longer time at 12 s as opposed to the Boussignac’s 7 s.

Noise and pressure-generation efficiency evaluation

The results of the noise generated by each Venturi adapter when the flow is directed into the open air are shown in Figure 2. The o_two CPAP device generated the most noise, reaching nearly 85 dBA. The Boussignac generated lower noise at lower flows, but the noise quickly increased with higher flows. The Flow-Safe II exhibited much lower noise levels than the Boussignac as flows increased, and the SoPAP adapter exhibited the lowest noise levels at all flows.

Fig. 2.

The sound level generated by each Venturi adapter for a given input oxygen flow while blowing into open air is shown. The baseline sound level in the room was measured to be 46.7 dBA. dBA, decibels measured on the a-weighted scale.

The pressure-generation efficiency and associated noise generation are shown in Figure 3. The Boussignac and Flow-Safe II both required much more oxygen flow to maintain desired pressures than the o_two CPAP device and the SoPAP adapter. However, the Flow-Safe II produced lower noise levels than the Boussignac, especially as pressure increases. The o_two device and SoPAP adapter exhibit much higher efficiency at generating pressure, with the o_two device being more efficient at lower pressures, and the SoPAP being more efficient at higher pressures. The SoPAP was quieter than any of the adapters, especially at lower pressures, and exhibits noise levels slightly lower than the Flow-Safe II at higher pressures.

Fig. 3.

(A) The input oxygen flow required to generate a given pressure is plotted on the Y axis versus the associated pressure. (B) The generated noise level versus the corresponding pressure. The baseline noise level with no oxygen flow given was 46.7 dBA.

CO₂ rebreathing and E_tCO2 dilution

The results from the CO₂ rebreathing evaluation and the E_tCO2 sample dilution are shown in Figure 4. Only the E_tCO2 results are shown because F_iCO2 remained near 0 under all circumstances. None of the Venturi adapters caused CO₂ rebreathing when no flow was present, with a 2 mm Hg increase in F_iCO2 or less when measured from the mask. Administering CPAP reduced the E_tCO2 observed in the trachea equally for all Venturi adapters. The E_tCO2 sample in the mask became more diluted as the CPAP level increased, although the SoPAP adapter diluted the E_tCO2 sample the least, and the Boussignac diluted the E_tCO2 sample the most. There also appears to be an inherent 2 mm Hg difference between the mask and trachea measurements, where the measurement from the mask reads slightly lower than the measurement from the trachea.

Fig. 4.

The E_tCO2 for each of the adapters, as sampled from the proximal end of the trachea and the mask, is shown for each tested CPAP level. The top of the bars reflects the E_tCO2 as measured from the trachea, which is the reference measurement, and the bars are filled to the level measured from the mask. The E_tCO2 measured in the trachea with no Venturi adapter was ∼39–40 mm Hg. The E_tCO2 measured in the mask with no Venturi adapter was ∼37–38 mm Hg. E_tCO2, end-tidal carbon dioxide.

F_IO2

The results of the F_IO2 evaluation are shown in Table 1. For all tested Venturi adapters at all tested ventilatory settings, the F_IO2 decreases as tidal volume increases. F_IO2 also appears to decrease as inspiratory time decreases. Increasing the breathing frequency also appears to decrease F_IO2. The Boussignac yielded the highest F_IO2 of all the Venturi adapters at every ventilatory setting, and the o_two CPAP device yielded the lowest F_IO2. The SoPAP adapter and the Flow-Safe II yielded F_IO2 between the other two adapters, where the SoPAP adapter produced a higher F_IO2 at the lower pressure, and the Flow-Safe II produced a higher F_IO2 at higher pressure.

Table 1.

F_IO2 achieved for each Venturi adapter at various combinations of tidal volume, breathing frequency, and inspiratory time

The observed trends suggest that for any Venturi adapter tested, F_IO2 decreases with higher frequency, higher tidal volume, and shorter inspiratory time.

Discussion

The objective of this study was to compare designs of four commercially available Venturi adapters for CPAP that have value in clinically relevant situations when a patient’s ventilation requires extra support. With increasing rates of comorbidities and sleep-disordered breathing, clinicians may need easier options to change from passive oxygen administration to positive pressure administration.^20,21 In an ideal world, a clinical CPAP system would monitor ventilation, be able to supply a range of oxygen percentages on an as-needed basis, keep CO₂ rebreathing to a minimum, be inexpensive, and require minimal space for both storage and utilization. Thus, this article set out to evaluate clinically relevant performance metrics.

The primary function of Venturi adapters is to amplify gas flows, and the detailed results reported in Supplementary Table S3 in Supplementary Data show that the adapter design impacts the flow amplification. A high flow amplification factor is advantageous because smaller adjustments to oxygen flow are needed to compensate for mask leaks than with Venturi adapters with lower amplification factors. For example, the o_two CPAP device had the highest flow amplification factor of 13.2, and the Boussignac had the lowest flow amplification factor of 7.9. If there is a 15 L/min mask leak, only an additional 1.1 L/min of oxygen flow is required to compensate for the leak when using the o_two device, whereas the Boussignac requires 1.9 L/min of additional oxygen flow to compensate for the same leak. The amount of flow amplification is likely impacted by the nozzle design of the Venturi adapter, where nozzles with narrower orifices will result in more jet mixing and thus higher amplified flows.^10,12

The expiratory resistance created by a Venturi adapter should be low enough to be comfortable for the patient during exhalation if oxygen flow is stopped. The SoPAP adapter exhibits a significantly higher expiratory resistance than the other Venturi adapters, with a resistance of 1 cm H₂O/L/s at 30 L/min and 1.3 cm H₂O/L/s at 60 L/min as shown in Table 1. However, this amount of resistance is less than many heat-and-moisture exchanger filters and endotracheal tubes and, therefore, likely not clinically important.^22,23 Expiratory resistance is likely related to the inner diameter of the Venturi adapter, where a narrower inner diameter creates a higher expiratory resistance.¹² In the case of the SoPAP, there is also the filter cloth that adds additional resistance. Minimal CO₂ rebreathing occurs (less than 2 mm Hg F_iCO2) for all tested Venturi adapters when no oxygen is being administered. No CO₂ rebreathing is observed when oxygen is flowing.

The ideal CPAP system would optimize oxygen used in generating adequate positive airway pressure. The pressure-generation efficiency demonstrated in Figure 3 shows the amount of oxygen flow required to generate a given pressure in the bench setup with no mask leak for each Venturi adapter. The flow-to-pressure relationship for the o_two device shown in Supplementary Figure S1 in Supplementary Data prescribes lower oxygen flows than were actually observed to generate the desired pressures. The SoPAP device flow-to-pressure relationship (Supplementary Fig. S1 in Supplementary Data) was similar to the results in Figure 3. The flows prescribed for the Boussignac device were very similar to the results in Figure 3 but did not provide prescribed flows for pressure above 10 cm H₂O. The same applies to the Flow-Safe II, where the flow-to-pressure relationship (Supplementary Fig. S1 in Supplementary Data) is reasonably close to that in Figure 3 but does not provide flows for attaining higher pressures. Although the Flow-Safe II is marketed²⁴ as being more efficient at generating pressures than the Boussignac, the results found in this study suggest that the efficiency of both adapters is almost the same. The results are likely also related to the inner diameter of the Venturi adapter, where a narrower inner diameter creates more resistance to flow exiting the mask, thus requiring less flow to maintain pressures. This is consistent with the data in Figure 3, as the o_two device and the SoPAP device have smaller inner diameters than the Boussignac and Flow-Safe II and are more efficient at generating pressures. Mask leak flow would increase the oxygen flow required to attain a desired mask pressure and would, therefore, make the flow-to-pressure tables less accurate and make it more difficult to administer desired pressures.¹² Thus, a system with automatic adjustment of oxygen flow to maintain the desired pressure, such as with the intended use of the SoPAP adapter, may be a desirable feature in a Venturi adapter CPAP system.

Guidelines from the National Institute on Deafness and Other Communication Disorders suggest that prolonged exposure to sound above 85 dBa may be damaging to hearing, but some experts are suggesting the safe level may be closer to 70 dBA.²⁵ Figure 3 demonstrates that the o_two CPAP device, the Boussignac, and the Flow-Safe II noise levels exceeded 70 dBA at relatively low pressures. The o_two device even reaches the higher 85 dBA threshold at 20 cm H₂O of pressure. These noise levels were measured from 1 m away, and patients would be much closer than this to the Venturi adapter, and the noise levels would likely be higher. The SoPAP noise level is near to the 70 dBA safe level at higher pressures near 15–20 cm H₂O but generates much lower noise levels than the other Venturi adapters at lower pressures. These noise levels would be elevated in the presence of mask leak for all Venturi adapters, as additional flow would be necessary to maintain the desired mask pressures. A high flow amplification would be advantageous in this regard, as less additional flow would be needed to compensate for the pressure lost to mask leak, resulting in less of an increase in audible noise. Figure 2 demonstrates that these Venturi adapters can produce very high noise levels that are likely irritating when the flow is directed into the open air. This could occur incidentally in a clinical environment if the oxygen flow is not reduced before the Venturi adapter is removed from the patient’s mask. Noise levels do not seem to be necessarily directly related to the amount of flow amplification or pressure-generation efficiency. It may be related to the nozzle design or amount of turbulent air flow, and further investigation into Venturi adapter design choices and their relationship to noise production is warranted. As discussed in the background section, a clinically desirable Venturi adapter CPAP system would provide a high F_IO2, efficiently generate pressures, produce low noise levels, and create minimal dilution of the exhaled sample of CO₂. The Venturi adapter that provides the highest F_IO2 under all tested ventilatory circumstances is the Boussignac adapter (Table 1). However, the Boussignac adapter is much less efficient at generating pressures than the SoPAP and the o_two CPAP devices and produces higher noise levels than the Flow-Safe II and the SoPAP (see Fig. 3). The o_two device, although very efficient at generating pressures, generates the highest level of noise among all the adapters and also produces the lowest F_IO2 (Table 1). The Flow-Safe II adapter is just as inefficient at generating pressures as the Boussignac and is quieter but produces a lower F_IO2. The SoPAP adapter has slightly lower pressure generation efficiency compared with the o_two device at lower pressures while exhibiting lower noise levels than any of the other tested Venturi adapters (Fig. 3). It also produces a higher F_IO2 than the o_two device (Table 1). Although the F_IO2 is lower than the Boussignac adapter, it is still relatively high, especially with low tidal volume and low breathing frequency. The SoPAP adapter also dilutes the exhaled sample of CO₂ the least of all tested Venturi adapters (Fig. 4). Clinicians should select the Venturi adapter based on specific clinical needs. For example, if minimizing oxygen use is a high priority, the clinician may consider using the o_two device or the SoPAP. Similarly, if maximizing the patient’s F_IO2 is the highest priority, the clinician may consider using the Boussignac. A summary of each Venturi adapter’s strengths is outlined in Table 2.

Table 2.

Summary of Venturi adapter performance

Green indicates that the corresponding Venturi adapter exhibited the highest performance in the corresponding clinical aspect. Red indicates the lowest performance, and yellow indicates intermediate performance.

ETCO₂, end-tidal carbon dioxide.

There also appears to be a relationship between F_IO2 and the oxygen-mixing properties in the mask. When the patient takes a breath, the first gas to enter the patient’s airway is the gas in the mask. Therefore, if the oxygen fraction in the mask is high during the end-expiratory pause, a greater portion of the tidal volume is composed of oxygen. Thus, a Venturi adapter that mixes the gas in the mask to a higher oxygen fraction at a faster rate results in a higher effective F_IO2. The Boussignac adapter mixes the oxygen in the mask to a very high fraction in the shortest amount of time and thus produces the highest F_IO2. The o_two CPAP device mixes oxygen the slowest, and although the attainable fraction is higher than the Flow-Safe II adapter, because of the slower mixing, the maximum oxygen fraction may not be attained before a breath is taken, thus producing the lowest F_IO2 of all the tested Venturi adapters. The oxygen mixing speed may be related to pressure-generation efficiency. Where more flow is required to maintain a desired pressure, flow circulates through the mask at a higher rate, mixing the gas more quickly. Therefore, there may be a direct trade-off between oxygen mixing speed in the mask and pressure-generation efficiency, where lower efficiency results in faster oxygen mixing. This also explains why decreased tidal volume results in increased F_IO2 because as the tidal volume gets smaller, the high-oxygen gas in the mask accounts for a greater portion of the tidal volume. A slower breathing frequency provides more time between breaths for the oxygen in the mask to mix to its maximum fraction, which would also increase the F_IO2. It is worth noting that actual F_IO2 values are dependent upon a variety of uncontrollable factors, such as patient anatomy, and may vary among patients and that the F_IO2 values reported in Table 1 are only representative of the bench setup that we used. Mask leak will also result in overall decreased F_IO2 values, as the oxygen concentration in the mask will not reach as high of an oxygen fraction between breaths. However, examining the relative differences in F_IO2 among Venturi adapters in Table 1 demonstrates that some Venturi adapters may have advantages when it comes to delivering a high oxygen fraction.

Further work regarding Venturi CPAP systems may include exploring methods for maximizing the oxygen mixing level and rate without sacrificing pressure-generation efficiency through different nozzle designs. Although there appears to be a relationship between pressure-generation efficiency and mixing speed, the SoPAP adapter had a very similar mixing speed to the Flow-Safe II device, suggesting that additional variables such as nozzle design may alter this relationship. Additional research on noise reduction for Venturi adapters would be beneficial, as the noise levels for Venturi adapters when the flow is directed to the open air can still be dangerous, even for the lower-noise adapters. There is also room for further research regarding methods for monitoring the true E_tCO2 and mitigating sample dilution. Researching Venturi CPAP systems that automatically adjust the oxygen flow to maintain the desired mask pressure in the presence of mask leak would also be beneficial.

Limitations

The scope of this study was limited to only a few commercially available Venturi CPAP adapters, and other Venturi adapters may produce a wider variety of results. There are also many other Venturi CPAP adapters not included in this study that provide additional features that may be of interest to clinical personnel, such as spring-loaded PEEP valves, or adjustable F_IO2 at a fixed oxygen flow. This study also did not evaluate the performance of these Venturi adapters with human subjects. Additional variables and effects may come into play when evaluating these adapters under use with actual patients. The noise levels were also not tested in a clinical environment, and other sounds in such an environment may shift the reported noise levels. An additional limitation in the F_IO2 testing is that metabolic oxygen consumption and CO₂ production were not accounted for. The reported F_IO2 numbers were steady-state oxygen concentrations sampled from a test lung from which no oxygen was consumed or removed. F_IO2 values will also vary among patients, and although our results demonstrate relative differences among Venturi adapters in their capability to deliver a high F_IO2, the numbers presented in Table 1 may differ from what would be seen in patients. Additionally, although we discussed the implications of leak flow, our study did not include any data for the tested Venturi adapters under various mask leak conditions.

Author Disclosure Statement

Mr. Blackwell has received funding from Dynasthetics. Prof. Orr and Dr. Sakata hold equity in Dynasthetics. Prof. Kuck declares no financial interests.

Funding Information

This study was funded by Dynasthetics and by NASA Grant 80NSSC20M0103.