Abstract
BACKGROUND: During continuous flow CPAP for noninvasive respiratory support, a high flow (eg, 60–90 L/min) of gas with FIO2 titratable up to 1.0 is provided within a helmet or face mask, while a PEEP valve maintains the set pressure. A large amount of oxygen is wasted, whereas only a minimal amount is consumed. We describe a recirculation circuit designed to reuse the exhaust gas and save oxygen.
METHODS: A standard Venturi-based continuous flow system delivering a flow of ≥ 60 L/min to a helmet, and a modified system designed to recirculate oxygen were tested on the bench during simulated active breathing. The proposed system recirculates the oxygen-enriched gas escaping from the PEEP valve; after CO2 removal, the gas is entrained by the Venturi flow generator and redirected to the helmet. We compared oxygen consumption, pneumatic performance, and gas conditioning of the standard and the recirculation systems.
RESULTS: The recirculation system reduced the oxygen consumption up to 80% as compared to the standard system. Oxygen sparing increased with increasing FIO2 and total flow delivered to the helmet. Exhaled CO2 was efficiently removed by a single soda lime canister for about 10 h. Pressure swings during a respiratory cycle slightly worsened when using the recirculation system as compared to the standard one (from 3.5–4.4 cm H2O). At FIO2 1.0, humidity was 6.1 mg/L and 17.3 mg/L with the standard and the recirculation system, respectively.
CONCLUSIONS: The recirculation system allowed a 80% reduction of oxygen consumption during simulated helmet CPAP therapy, whereas CO2 removal was effective for > 10 h. Recirculation minimally affected pneumatic performance of the CPAP continuous flow system, while improving gas conditioning as compared to the standard system.
- compressed oxygen recirculation
- continuous flow CPAP
- head helmet
- humidity and gas water content
- acute hypoxic respiratory failure
Introduction
Acute respiratory failure is a leading cause for ICU admission; several etiologies may represent the underlying cause leading to respiratory failure, but the supportive treatment is essentially based on invasive or noninvasive respiratory support.1 Continuous flow CPAP, often delivered by a helmet, is one of the available modalities: A mixture of air and oxygen titratable up to FIO2 = 1.0 is delivered into a device covering patient’s airways (either helmet or face mask) at high flows (eg, 60–90 L/min), while an adjustable PEEP valve maintains the set pressure within the circuit.2 Continuous flow helmet CPAP has proved effective in different clinical scenarios, such as community-acquired pneumonia, acute pulmonary edema, and patients with COVID-19.3
Continuous flow CPAP presents several advantages: A constant airway pressure adjustable on patient’s need (typically 5–12 cm H2O) is maintained with a simple and inexpensive equipment based on a Venturi flow generator; a humidifier is not needed when FIO2 is < 0.7; after appropriate staff training, a large number of patients can be effectively treated also outside the ICU without a ventilator, as happened during the pandemic; awake prone positioning is feasible.4,5 On the other side, the main limitations of continuous flow helmet CPAP are the lack of an active ventilatory support (such as during bi-level ventilation) and the large amount of O2 enriched gas wasted in the ambient after escaping from the PEEP valve, as compared to the small amount actually required to match patient’s O2 consumption. Whereas the first limitation is intrinsic to the concept of CPAP, the second may prevent the use of such device in settings where the availability of compressed oxygen is limited, such as during transportation, low-resource settings, and even in large hospitals burdened by a high amount of oxygen consumption, with its associated costs.6,7 As an example, > 99% of the oxygen is dispersed into the surrounding environment during continuous flow CPAP at 60 L/min and FIO2 = 1.0, whereas < 1% is actually consumed by the patient. In a standard continuous flow system, the 2 options to limit oxygen consumption are either decreasing FIO2 (feasible if tolerated by the patient) or to reduce the amount of gas flowing through the helmet. However, such strategy is minimally effective in saving oxygen: When patient’s gas exchange is impaired, high FIO2 is required; the total free flow should remain high enough to maintain helmet pressure during inspiration and always > 40 L/min to effectively wash the CO2 exhaled by the patient.8
Therefore, we hypothesized that a recirculation circuit could be used to recover the gas flowing outside the PEEP valve and otherwise dispersed into the ambient; after CO2 removal, the gas could be entrained by the Venturi and reused in the circuit, obtaining a substantial reduction of compressed oxygen consumption without reduction of the total flow and maintaining a high FIO2.
The purpose of this work was to characterize the performance of the recirculation system (as compared to a standard continuous flow system) under several aspects: O2 consumption, duration, pneumatic performance, gas temperature, and humidity.
QUICK LOOK
Current Knowledge
CPAP is currently used for noninvasive respiratory support. The large part of the compressed oxygen delivered through the interface (helmet or face mask) is dispersed into the surrounding environment, with only a small volume consumed by the patient, reducing the usability of this modality when compressed oxygen is limited.
What This Paper Contributes to Our Knowledge
A recirculation circuit designed to spare oxygen reduced the amount of required oxygen by 80% in bench tests, while maintaining optimal inspired gas humidity without the need for an active humidifier. Recirculation minimally affected the pneumatic performance of the recirculation system as compared to standard.
Methods
We performed bench tests to compare the performance of a standard continuous flow helmet system (standard system) (Fig. 1A) with a modified system designed to reduce compressed O2 consumption (recirculation system) (Fig. 1B). The standard system was based on a Venturi flow generator (EasyVEE, Flow-Meter, Levate, Italy)9 fed by pure oxygen, connected to a medium-size CPAP helmet (DimAIR, Dimar, Medolla, Italy) placed on a plastic mannequin head, and equipped either with a fixed 7.5 cm H2O or 12.5 cm H2O PEEP valve (Caradyne, Parkmore West, Galway, Ireland). Desired FIO2 was obtained by rotation of the variable area valve (Fig. 2) to regulate the amount of ambient air entrained by the Venturi; the lateral port was occluded when using the flow generator with the standard system.
Schematic representation of the standard (A) and the recirculation (B) circuits. Bench tests were performed comparing a standard Venturi-based helmet CPAP system and the recirculation system. In the bench model, CO2 and water vapor were added to simulate a patient’s exhaled gas; active breathing was replicated by a lung simulator.
Schematic representation of the Venturi flow generator. A Venturi flow generator connected to a 4-bar oxygen port was used to pressurize the helmet and wash CO2. Compressed oxygen escaping from a small-bore channel generates a negative pressure, leading to gas entrainment from the lateral channel (A) or both from the lateral channel and the variable area valve (B). Blue dots represent oxygen, while red dots represent air. Arrows indicate the direction of flow.
The recirculation system comprised an added expiratory circuit to direct the gas flowing outside the PEEP valve (with a high oxygen content) into a 1-L soda lime (Drägersorb 800+, Drägerwerk, Lübeck, Germany) canister; the gas, free of CO2, was then entrained by the Venturi flow generator and redirected to the helmet. A T-piece was placed after the PEEP valve, resulting either in escape of excess gas from the system or entrainment of supplemental room air in case of reduced recirculating flow, therefore avoiding overpressurization or depressurizations of the circuit; excess gas typically flows out of the T-piece; strong inspiratory efforts may lead to transient PEEP valve closing and subsequent entrainment of room air due to the negative pressure from the Venturi placed downstream.
When FIO2 > 0.7 was needed, a 50-cm-long breathing circuit tube extension was connected to the ambient side of the T-piece acting as an open reservoir to reduce the amount of gas dispersed in the ambient. Finally, an adjustable opening was placed just before the Venturi flow generator to titrate the amount of recovered gas relative to the entrained ambient air.
Measurements
A multi-parameter monitor (VT PLUS HF, Fluke Biomedical, Cleveland, Ohio) was used to record pressure signals, flow, and FIO2. A ventilator tester (VT-2, Agilent, Santa Clara, California) customized in order to generate active efforts was used to simulate subject’s breaths and connected to the plastic head “trachea.” A constant flow of CO2 (250 mL/min) was injected within the same trachea. A carbon dioxide monitor (ViaSensor G100, QED, Dexter, Michigan) was used to measure residual CO2 after the soda lime canister every 30 min. Active humidification was obtained by a dedicated humidifier (Lifeneb nebulizer coupled with Neptune heater; flow-meter, Levate, Italy) connected between the lung simulator and the helmet, whereas temperature and relative humidity data were collected by a portable sensor (Thermohygrometer BC21, Trotec, Heinsberg, Germany).
Experimental Design
We conducted a comprehensive evaluation of different aspects of the recirculation system performance, mimicking clinically realistic and meaningful conditions.
Oxygen Consumption.
To evaluate the amount of oxygen spared by the recirculation system as compared to the standard, we set the lung simulator at a moderate level of respiratory distress (tidal volume [VT] = 500 mL and breathing frequency 24 breaths/min). We then measured the amount of oxygen required to obtain a flow through the helmet of 60 L/min and 75 L/min, at different FIO2 levels, in comparison with a standard system. A continuous flow of 90 L/min and FIO2 = 1.0 was tested with the recirculation system only.
Pneumatic Performance.
To investigate the ability of the system to maintain a constant pressure, we compared the pressure levels within the helmet (CPAP; minimum and maximum pressure during simulated respiratory activity, Pmin and Pmax, respectively) during 3 different levels of respiratory distress (mild: VT = 450 mL, breathing frequency = 18 breaths/min; moderate: VT = 500 mL, breathing frequency = 24 breaths/min; and severe: VT = 600 mL, breathing frequency = 30 breaths/min) of the standard and the recirculation systems. The tests were performed both at 7.5 cm H2O and 12.5 cm H2O CPAP, at 60 L/min and 75 L/min helmet flow; the recirculation system was tested both with and without the reservoir, at maximum recirculation.
System Duration.
To verify the effectiveness of expired CO2 removal from the recirculating gas and the theoretical feasibility of a prolonged therapy using the recirculating system, we measured the concentration of CO2 during time at maximum recirculation with the reservoir connected. The helmet flow was set at 60 L/min, whereas the simulated respiratory distress was set at a moderate level (see above); we considered an inspired CO2 concentration of 0.1% as a sign of CO2 absorber exhaustion.
Gas Conditioning.
Temperature and relative humidity of the gas flowing out of the helmet were evaluated at FIO2 = 1.0 with the standard and the recirculation systems (reservoir connected). To verify the stability of the signal, measurements with maximum recirculation were performed during a prolonged test (up to 14 h). The helmet flow was set at 60 L/min, whereas the simulated respiratory distress was set at a moderate level (see above).
Given the bench nature of the study, and the lack of variability present among the collected data under same experimental conditions (ie, repeating controlled bench measurements would always result in the same values), we did not conduct a formal statistical analysis and focused on the clinical relevance of differences between means for the recorded variables.
Results
Compressed Oxygen Requirements
The recirculation system effectively reduced compressed oxygen required to feed a Venturi-based helmet for CPAP therapy; saving oxygen was feasible for FIO2 > 0.4, being maximum at 1.0 (Fig. 3). Oxygen requirement was reduced by the recirculation system to nearly one third (35%, 13 L/min vs 37 L/min) with 60 L/min of gas flow through the helmet at FIO2 = 0.7. When the reservoir was added and FIO2 reached 1.0, oxygen consumption was reduced to nearly one fifth (22% and 20% at 60 L/min and 75 L/min, respectively). To obtain a 100% oxygen free flow of 75 L/min with the standard system, the flow regulator was completely open and the Venturi system bypassed, reaching the maximum possible flow provided by a single 4-bar O2 wall port; the corresponding oxygen consumption with the recirculation system was 15 L/min (difference 60 L/min). A 100% oxygen free flow of 90 L/min flow required 18 L/min of compressed O2; the same free flow could not be obtained in a standard system fed by a single oxygen flow regulator.
Oxygen consumption during CPAP therapy. When using a standard continuous flow CPAP system (60 L/min), the amount of needed oxygen was proportional to the set FIO2 and increased at higher flows (75 L/min). The recirculation system provided CPAP with similar flows (triangles) with a constantly low oxygen requirement regardless of FIO2. To obtain FIO2 > 0.7 (squares), a reservoir was added to the recirculation system (see text for further details).
Carbon Dioxide Removal
Exhaled CO2 contained in the recirculating gas was efficiently removed by the soda lime, with an entrained gas CO2 concentration undetectable for about 10 h at FIO2 = 1.0 (maximum recirculation with reservoir). Carbon dioxide concentration progressively increased, reaching 0.2% (1.5 mm Hg) after another 2 h; soda lime exhaustion occurred at 14 h from the beginning of the test, corresponding to a CO2 concentration in the entrained gas of 0.4%. (Fig. 4). The amount of CO2 absorbed by the canister after 10 h was about 120 L; it can be calculated that in case of lower CO2 productions (eg, 200 mL/min) or lower FIO2 (eg, 0.6) CO2 adsorbent canister would last for > 17 h.
Performance of CO2 adsorption. Complete removal of CO2 from the recirculating gas was feasible for 10 h in case of maximal recirculation (CO2 injection 250 mL/min; FIO2 = 1.0). Performance of the CO2 adsorbent medium progressively decreased in the following 4 h.
Pneumatic Performance
Pressure levels within the helmet during a respiratory cycle are reported in Figure 5. Average differences between the nominal and the recorded CPAP levels were similar between the standard and the recirculation system (−0.15 cm H2O and −0.10 cm H2O, respectively); as expected, when the reservoir was added to the recirculation system, CPAP showed a negligible increase (+0.32 cm H2O) due to the flow resistance added after the PEEP valve. Stability of the CPAP system pressure, expressed as the pressure swings recorded during a respiratory cycle (Pmax-Pmin), slightly worsened when using the recirculation system as compared to the standard one (from 3.5 cm H2O to 4.4 cm H2O); when the reservoir was added, pressure swings were more pronounced (5.8 cm H2O).
Pressure measured within the helmet during a respiratory cycle. CPAP levels remained stable with the recirculation system as compared to the standard, regardless of simulated level of respiratory distress, both at 60 L/min (A and C) or 75 L/min (B and D). Pressure swings within the helmet (vertical lines), due to reduced pressure during inspiration and increased pressure during expiration, resulted higher with the recirculation system, particularly when the reservoir was connected; pressure swings were reduced by higher flow levels (B and D). Panels A and B report data collected with a 7.5 cm H2O mechanical PEEP valve, Panels C and D with a 12.5 cm H2O valve. Pmin = minimum pressure during simulated respiratory activity; Pmax = maximum pressure during simulated respiratory activity.
Temperature and Humidity
With the standard system, temperature of the gas flowing out of the helmet was 23°C; the use of recirculation stabilized the temperature at about 24°C. In the standard system, relative humidity within the helmet dropped from 55% (ambient air) to 30% at FIO2 = 1.0, corresponding to an absolute humidity of 6.1 mg/L. The recirculating system maintained an adequate level of relative humidity at about 80% (absolute humidity 17.3 mg/L) due to the recirculation of water vapor; no condensation was present within the helmet after several hours of maximum recirculation.
Prolonged measurements with maximum recirculation and reservoir showed a stabilization of temperature and humidity within the helmet during the first 30–60 min and no relevant changes thereafter (Fig. 6).
Temperature and relative humidity during recirculation. Recirculation of expired gas was associated with optimal helmet conditioning; at maximum recirculation and FIO2 = 1.0, temperature stabilized at 23–24°C and relative humidity at 80% (squares and cross symbols, respectively). No condensation was observed within the helmet.
Discussion
In the present study, we report the effectiveness of a recirculation system, combined to a continuous flow helmet CPAP, in sparing medical oxygen. At FIO2 = 1.0, compressed oxygen requirement was reduced by 75% at 60 L/min of helmet flow and 80% at 75 L/min. A single canister of CO2 adsorbent material lasted for > 10 h in maximal recirculation configuration. CPAP levels within the helmet remained stable independent of the recirculation circuit; pneumatic performance was substantially unaffected when the recirculation circuit was used without the reservoir, but increased helmet pressure swings were recorded during the respiratory cycle with the reservoir in place. Temperature and humidity of the inspired gas were more favorable than a standard system.
When using the recirculation system, oxygen needed to obtain CPAP in a helmet at FIO2 = 1.0 was comparable to a non–rebreather mask (ie, 13–15 L/min) and largely < a standard continuous flow CPAP system. The daily amount of compressed oxygen potentially saved for each subject treated uninterruptedly at FIO2 = 1.0 would be relevant: about 65 m3 and 86 m3 per day for a helmet free flow of 60 L/min and 75 L/min, respectively. Reducing oxygen consumption is a typical challenge in low-resource settings, where the extended use of continuous flow therapies is not feasible due to the shortage of compressed gas and the most diverse strategies are considered to reduce oxygen use.10 Similarly, limited oxygen is a relevant factor during in- and out-of-hospital transport of critically ill patients, where continuous flow therapy is feasible but requires careful monitoring of oxygen availability. During the recent COVID-19 pandemic, in-hospital transport of patients who were dependent from maximal continuous flow therapy (eg, directed to the computed tomography scan or to the ICU for urgent intubation) was also a challenge due to high oxygen consumption, requiring the use of multiple tanks to minimize the risk of therapy interruption in patients with marginal stability. Moreover, the large use of compressed oxygen resulted in unexpected episodic hospital system failures, showing that oxygen shortage might be an exceptional but possible issue also in-hospital when treating noninvasively a high number of respiratory failure patients.11,12 In Venturi-based systems, compressed oxygen is the driver of the flow feeding the helmet; additional saving of oxygen might be obtained by a further reduction of compressed oxygen flow (ie, below 13–15 L/min) at the price of lower helmet total flows. However, we suggest to avoid such a strategy due to the risk of reduced CO2 wash and lower pneumatic performance when helmet total flow is not adequate. Another advantage of the recirculation system is the possibility to reach high flows (eg, 90 L/min) with FIO2 = 1.0, a rare condition sometimes useful in severely ill patients but not feasible with a single oxygen wall port.
The extended duration of a single soda lime canister (10–12 h with maximum recirculation) suggests that the recirculation system has the potentiality to cover most of the aforementioned situations of low oxygen availability and to result feasible for prolonged CPAP therapy. We set a cutoff limit of 0.1% CO2 concentration in the inhaled gas as a sign of absorber exhaustion. We chose a conservative cutoff, corresponding to a partial pressure < 1 mm Hg, for 2 main reasons: first, real-life scenarios are less controlled than bench tests; second, absorber performance decreases in few hours, as shown in Figure 4, suggesting that an initial loss of CO2 removal performance needs to be recognized for a timely substitution of the absorber cartridge. However, we believe that if clinical tests will confirm the presented data a slightly higher cutoff might be chosen to define cartridge exhaustion.
In the presented tests, the recirculation system led to an optimal “conditioning” of temperature and humidity of the inspired gas. Humidity is progressively reduced when increasing FIO2 in standard continuous flow systems, and an active humidifier is recommended at FIO2 > 0.7.13 Instead, no active humidification is needed with recirculation, thanks to the small amount of fresh but dry gas relative to the water-containing recirculating gas: We recorded an absolute humidity of 17 mg/L, a value above the minimum recommended in the literature (10 mg/L) to minimize mucosal damage in the upper airways.13,14 Although a clinical test is necessary to assess comfort while using the recirculation circuit, we expect that humidity and temperature resulting from recirculation will be well tolerated. It is known that the use of active heating above 31°C, while necessary during invasive ventilation, might lead to excessive temperature within the helmet, resulting in discomfort.15 Temperature levels within the helmet measured in the present study were far below 31°C, although they might be slightly higher in vivo due to the presence of the subject’s head. The presence of moisture on helmet inner surface, which is another sign associated with patient discomfort, was never recorded during the bench tests despite an adequate level of absolute humidity, further suggesting good tolerability.14
To our knowledge, no reports are available in the literature regarding the use of recirculation applied to a Venturi-based helmet CPAP system. The concept of recirculating exhaled gas after CO2 removal is well known and applied in many fields, ranging from space walks to scuba diving; regarding medical devices, recirculation is commonly used in anesthesia machines to reduce anesthetic gas consumption.16 A recent report suggested the theoretical use of a CPAP sleep apnea machine coupled to a recirculation circuit to pressurize a helmet in the aim of providing CPAP therapy, an application similar to the present study but functioning with a helmet flow of about 10 L/min; since it is well known that at least 30 L/min, and possibly more, are required to obtain a good CO2 washout of the helmet dead space, the translation to the clinical scenario of the described system does not appear feasible as described.8,17 Moreover, a turbine-ventilator machine was used, increasing the complexity and the expenses as compared to a simple Venturi flow generator.
However, the recirculation system as described in our study has also some limitations. First, FIO2 is influenced by patient’s respiratory pattern when the reservoir is not used; as an example, FIO2 ranged from 0.74–0.58 at 60 L/min at 7.5 cm H2O CPAP depending on the level of simulated distress. Such a variability is typical of the open systems, as happens with oxygen masks; since the recirculation system is open to the ambient, more air is entrained in the circuit when patient’s distress increases resulting in lower FIO2. Repeated or continuous monitoring of FIO2 is, therefore, needed to match patient’s distress if the reservoir is not used; when the reservoir is present, FIO2 is stable irrespective of patient’s distress.
The second limitation is related to the CO2 adsorbent canister. First, monitoring of inspired CO2 will be mandatory once the recirculation system becomes a medical device applied on the patient to prevent increase of CO2 within the helmet and proceed with a timely substitution of the canister; a 2-bed system could be used for nearly 1 d at FIO2 = 1.0, and even more at lower FIO2 levels, without therapy interruption. Second, oxygen consumption is reduced at the expense of the need of a different consumable: the CO2 absorber. Whereas the shortage of CO2 absorber represents a clear limitation for the use of the recirculation circuit, the transport and storage of CO2 absorber cartridges are probably less challenging than production and transport of compressed oxygen cylinders in specific settings. A 1-L cartridge removes CO2 from the equivalent of 27 5-L compressed oxygen tanks when operating at 60 L/min free flow at FIO2 = 1.0, with a clear advantage in terms of volumes and weights.
The third limitation is the lower pneumatic performance of the recirculation systems as compared to the standard one when using the reservoir; whereas PEEP was substantially unchanged, the recirculation system led to higher pressure swings within the helmet, particularly when the reservoir was used. In the present study, we tested a circuit based on tubes and connections commonly used in the clinical practice, designed for ventilators; we hypothesize that a dedicated recirculation circuit specifically designed to minimize flow resistance could result in a better pneumatic performance. Increased swings were particularly relevant when simulating high distress at a relatively low flow (60 L/min), suggesting that an adequate match between patient’s effort and helmet free flow could be particularly relevant when using the recirculation system to maintain a stable pressure within the helmet, resulting in optimal CPAP therapy.
Lastly, a recirculation system assembled with standard components, although possibly effective, does not necessarily reach medical device safety standards for use on patients. In this proof-of-concept study, we did not evaluate all the possible safety issues related to the clinical use of the recirculation system. In case of cartridge exhaustion, CO2 rebreathing is a real risk as it is in anesthesia machines, and continuous CO2 monitoring should be warranted as discussed above. Another issue might be related to the use of inadequate circuit components, leading to increased circuit resistance and lower pneumatic performance as well as inadvertent occlusion of the T-piece or the reservoir leading to system overpressurization. The placement of the T-piece is also relevant: It should be placed near the helmet to minimize the circuitry resistance effect. In summary, we completely discourage the use of a recirculation system built with standard parts or without appropriate monitoring of recirculating gases. Further tests of a recirculation circuit engineered into a dedicated medical device are needed to confirm everyday use safety and feasibility.
Conclusions
The bench evaluation of the recirculation system showed an 80% reduction of oxygen consumption during simulated helmet CPAP therapy. Removal of CO2 was effective for > 10 h using a standard soda lime canister, suggesting a potential application for prolonged continuous treatment. Pneumatic performance of the recirculation system was acceptable, although further improvements might be obtained with a circuit specifically designed to reduce flow resistance. Despite maximal FIO2, recirculation prevented the phenomenon of gas drying within the helmet typically associated with standard continuous flow systems, eliminating the need for an adjunctive humidifier.
Footnotes
- Correspondence: Giacomo Bellani MD PhD, University of Milan-Bicocca, Department of Medicine and Surgery, Via Cadore 48, Monza (MB), Italy; E-mail: giacomo.bellani1{at}unimib.it
See the Related Editorial on Page 286
Drs Coppadoro and Bellani disclose a relationship with Flow-Meter and are coinventors of a pending patent relative to the presented device. Mr Paratico is employed by Flow-Meter. Dr Bellani discloses relationships with Dräger Medical, Getinge, and Siaretron.
- Copyright © 2023 by Daedalus Enterprises