KEY POINTS
- Question: Do anesthesia machines accurately detect and deliver prescribed breathing circuit pressures when they are altered for off-label use?
- Findings: The tested anesthesia machines accurately measure circuit pressures but, under certain conditions, can deliver significantly different pressures from what is prescribed.
- Meaning: Anesthesiologists should be aware of the potential clinically important increases in pressure that may be unintentionally delivered on some anesthesia machines when altered for nonanesthetic use.
The impact of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and its resulting illness coronavirus disease 2019 (COVID-19) has stressed the health care system in the United States as it has done in other countries. As of early April 2020, multiple prediction models of COVID-19’s incidence in the US project that the number of critically ill patients will outnumber available intensive care unit (ICU) beds within weeks.1–3 Severe cases of COVID-19 lead to acute respiratory failure and the need for urgent mechanical ventilation.
In an effort to the fill this emergent need for ventilators, the US Food and Drug Administration (FDA) has temporarily approved the use of anesthesia machines for use in the ICU.4 Repurposing anesthesia machines as ICU ventilators presents a particular challenge, as waste anesthesia gas suction (WAGS) is often unavailable in the ICU.5 Anecdotal reports have suggested as much as 18 cm H2O positive end-expiratory pressure (PEEP) may inadvertently be applied with WAGS disconnected, accompanied by inaccurate airway pressure readings by the anesthesia machine itself.6 In addition, previously reported cases of scavenging system obstructions caused clinically significant increases in PEEP.7,8 The impact of an altered or malfunctioning anesthesia gas scavenging system (AGSS) on airway pressures could lead to significant patient harm.9
Anesthesia machine manufacturers and national anesthesia guidance boards have recently published advisories for the use of these machines in nonoperating room settings.5,10,11 However, no published studies to date have quantified the amount of increased airway pressure in anesthesia machines that may be present if operated with the WAGS disconnected.
This study aimed to simulate an anesthesia machine operating in an ICU environment without WAGS availability. We hypothesized that resistance caused by disconnecting an anesthesia machine’s closed AGSS may contribute to inadvertently increased levels of continuous positive airway pressure (CPAP). We also hypothesized that anesthesia machines accurately detect and display delivered pressures during normal use as compared to a reference manometer.
METHODS
;)
Figure 1.: The experimental setup on the Dräger Apollo (A) includes a reference manometer (star) connected to the anesthesia circuit sampling line connector and a standard anesthesia circuit reservoir bag (arrow) connected to the breathing circuit to simulate a lung. Observations were recorded with the WAGS connected to the AGSS (B), the WAGS disconnected from the AGSS (C), and with medical vacuum suction connected to the AGSS (D) using a WAGS to vacuum hose (E, F). AGSS indicates anesthesia gas scavenging system; WAGS, waste anesthesia gas suction.
Figure 2.: On the GE Avance CS2, the protocol was additionally performed as shown, with the WAGS disconnected (A, arrow) and the 3 L AGSS visual indicator reservoir bag removed (B, arrow) according to the manufacturer’s recommendations. AGSS indicates anesthesia gas scavenging system; WAGS, waste anesthesia gas suction.
The following anesthesia machines were evaluated: Dräger Perseus A500, Dräger Apollo (Dräger, Inc, Telford, PA), and GE Avance CS2 (GE Healthcare, Chicago, IL). The 2 Dräger machines utilize an open AGSS, and the GE Avance CS2, a closed one. A reference manometer (Spec Scientific, Scottsdale, AZ, Model 9300990) was connected at the anesthesia circuit sampling line connector, and a standard anesthesia circuit reservoir bag (King Systems, DF3115-6121Z, Braunfels, Germany) was connected to simulate a lung as shown in Figure 1. Ventilation was then initiated with the following settings: volume control mode, tidal volume (TV) 500 mL, RR 12, fraction of inspired oxygen (Fio2) 1.0, fresh gas flow (FGF) rate 2.0 liters per minute (LPM), ratio of inspiration to expiration times (I:E) 1:1.9, inspiratory time 1.7 seconds, PEEP 0 cm H2O. After engaging the ventilator, peak inspiratory pressure (PIP) and PEEP measurements were made using the reference manometer and the anesthesia machine display simultaneously. This was repeated for a total of 5 respiratory cycles. Then, the process was repeated using prescribed PEEP levels of 5, 10, 15, and 20 cm H2O. Next, the WAGS tubing was disconnected from the AGSS, and the measurements were repeated. These steps were performed again at FGF rates of 4, 6, 8, 10, and 15 LPM. Therefore, for each pair of FGF and prescribed PEEP level, 5 PIP and 5 PEEP measurements from both the reference manometer and the anesthesia machine display were recorded. For the GE Avance CS2, the protocol was again repeated with the following modifications: WAGS disconnected and the 3 L AGSS visual indicator reservoir bag removed according to the manufacturer’s recommendations (Figure 2).12 Afterward, each ventilator AGSS was connected to the medical vacuum system, and measurements were repeated once more. We defined a priori a clinically significant difference of 3 cm H2O between the reference manometer and the machine pressure reading. This article adheres to the applicable Standards for Quality Improvement Reporting Excellence (SQUIRE) guidelines.
Statistical Analysis
Bland-Altman plots were used to assess the agreement between the reference manometer and anesthesia machine pressure. The normality of these differences was assessed using Shapiro-Wilk and Kolmogorov-Smirnov tests, as well as the Q-Q plot, all of which showed that the data were not normally distributed. A Bland-Altman plot assumes that the data are normally distributed. Therefore, we used a nonparametric Bland-Altman plot, reporting the median difference rather than the mean difference.13 In addition, the 2.5th percentile and the 97.5th percentile of the observed differences in the data were reported instead of the 95% confidence interval (CI) for the limits of agreement (LOA). Because nonparametric methods had to be used, there is no standard deviation estimate to use to calculate the 95% CIs around each LOA. Thus, to estimate the 95% CIs around each of the LOA, bootstrap methods were used. Using the differences in PEEP and PIP manual and machine measures, we generated 50,000 samples. Each bootstrap sample was equal to the size of our study sample, and generated by selecting from the distribution of differences with replacement. Sampling with replacement allows for each difference to be selected multiple times within a given sample, creating a variety of bootstrap samples across the 50,000 that were generated. The LOAs were saved from each bootstrap sample and were then used to estimate the 2.5th and 97.5th percentiles of each LOA.
Linear regression models were used to assess the difference in reference manometry measures and prescribed measures between WAGS connected and WAGS disconnected. The outcome of the model was the difference between the reference manometry measurement and the prescribed measurement, with the predictor being the interaction between the WAGS status and the prescribed measure. Statistical significance was set at P value <.05. To control the false discovery rate (FDR), the Benjamini-Hochberg method was applied to all P values.14 This method adjusts P values by sorting all P values in ascending order then adjusting them by their rank while accounting for the total number of tests performed and the desired FDR. The FDR was set to 0.05. After FDR adjustment, 268 tests yielded significant results; a maximum of 14 of these (5%) are expected to be false findings.
RESULTS
With WAGS disconnected (Figure 1C), the reference manometer corresponds to each machine’s PEEP and PIP readings by Bland-Altman analysis as shown in Figure 3. The median difference between PEEP measured by the machine and PEEP measured by the reference manometer was 0.01 cm H2O (95% LOA, −0.40 to 0.60) for the Dräger Perseus A500, −0.10 cm H2O (95% LOA, −0.90 to 1.00) for the Dräger Apollo, and 0.10 cm H2O (95% LOA, −0.35 to 0.70) for the GE Avance CS2. The median difference between PIP measured by the machine and PIP measured by the reference manometer was −0.40 (95% LOA, −1.10 to 0.41) for the Dräger Perseus A500 and −0.40 (95% LOA, −1.00 to 0.55) for the Dräger Apollo. However, the LOAs for the PIP readings on the GE Avance CS2, which are 0.8 for the 2.5th percentile and 3.0 for the 97.5th percentile, indicate that there is less agreement in the machine and reference readings. However, these differences are likely clinically insignificant. Refer to Supplemental Digital Content, Table 1, https://links.lww.com/AA/D395, for the complete data summary, including CIs for the LOA.
Figure 3.: Bland-Altman plots for repeated measurements of PEEP and PIP between anesthesia machine manometer and reference manometer at flow rates of 2, 4, 6, 8, 10, and 15 LPM for each of the following anesthesia mechanical ventilators: Dräger Perseus A500 (A), Dräger Apollo (B), and GE Avance CS2 (C). Continuous lines correspond to the median difference; the dashed lines correspond to the 95% LOA, and the highlighted areas correspond to the 95% confidence intervals of the upper and lower LOA. LOA indicates limits of agreement; LPM, liters per minute; PEEP, positive end-expiratory pressure; PIP, peak inspiratory pressure.
The differences in PEEP before and after the WAGS was disconnected (Figure 1B, C) on each anesthesia machine are displayed in Figure 4. Note that differences approaching 0 cm H2O indicate correctly delivered, nonelevated airway pressures. Statistical significance is indicated by P values<.05. The Benjamini-Hochberg method was applied to all P values to prevent the FDR from rising above 0.05. This method adjusts P values sequentially while accounting for the total number of tests performed and the desired FDR. All of the following reported P values reflect this FDR adjustment.
Figure 4.: Difference in PEEP when disconnecting the WAGS on the Dräger Perseus A500 (A), Dräger Apollo (B), and GE Avance CS2 (C) with 95% confidence intervals denoted by vertical bars. FGF indicates fresh gas flow; LPM, liters per minute; PEEP, positive end-expiratory pressure; WAGS Off, waste anesthesia gas suction disconnected; WAGS On, waste anesthesia gas suction connected.
At FGF of 2 LPM and PEEP 0 cm H2O with the WAGS disconnected (Figure 1C), the Dräger Apollo had a difference in PEEP of 0.02 cm H2O (95% CI, −0.04 to 0.08; FDR-adjusted P = .59), and the Dräger Perseus A500, <0.0001 cm H2O (95% CI, −0.11 to 0.11; FDR-adjusted P = 1.00). The Dräger Apollo and Perseus A500 had statistically significant differences in PEEP when the WAGS was disconnected at various fresh flow rates and prescribed PEEP levels, but they were infrequent and always clinically insignificant, <3 cm H2O. At FGF of 2 LPM and at PEEP 0 cm H2O, the GE Avance CS2 showed an increase in PEEP of 8.62 cm H2O (95% CI, 8.55-8.69; FDR-adjusted P < .0001) when the WAGS was disconnected. When prescribed PEEP was increased to 5 cm H2O, there was an increase of observed PEEP by 5.84 cm H2O (95% CI, 5.77-5.91; FDR-adjusted P < .0001). When the prescribed PEEP was increased to 10 cm H2O, measured PEEP increased by 0.58 cm H2O (95% CI, 0.51-0.65; FDR-adjusted P < .0001). At prescribed PEEP of 15 and 20 cm H2O, measured PEEP increased by 0.04 cm H2O (95% CI, −0.03 to 0.11; FDR-adjusted P = .33) and <0.0001 cm H2O (95% CI, −0.07 to 0.07; FDR-adjusted P = 1.00), respectively. Similar findings were seen as FGF was increased until FGF reached 15 LPM. At prescribed PEEP levels of 0 and 5 cm H2O at FGF 15 LPM, the increase in PEEP was 1.08 (95% CI, 0.95-1.21; FDR-adjusted P < .0001).
The differences in PIP after the WAGS was disconnected are shown in Figure 5. At FGF of 2 LPM and PEEP 0 cm H2O, the Dräger Perseus A500 had a median PIP difference of −0.04 cm H2O (95% CI, −0.42 to 0.34; FDR-adjusted P = .89) when the WAGS was disconnected, and the Dräger Apollo, −0.28 cm H2O (95% CI, −0.53 to −0.03; FDR-adjusted P = .05). At the same settings, the GE Avance CS2 showed an increase in PIP of 6.80 cm H2O (95% CI, 6.30-7.30; FDR-adjusted P < .0001). Increasing the prescribed PEEP to 5 cm H2O, the measured PIP increased by 4.56 cm H2O (95% CI, 4.06-5.06; FDR-adjusted P < .0001). When the PEEP was increased to 10 cm H2O, the GE Avance CS2 showed an increase in PEEP of 0.42 cm H2O (95% CI, −0.08 to 0.92; FDR-adjusted P = .14). At PEEP levels of 15 and 20 cm H2O, PIP changed by 0.00 (95% CI, −0.50 to 0.50; FDR-adjusted P = 1.00) and −0.24 (95% CI, −0.74 to 0.26; FDR-adjusted P = .42), respectively. Similar findings were seen as FGF was increased until FGF reached 15 LPM, where at set PEEP levels of 0 and 5 cm H2O, the increase in PIP was 1.36 cm H2O (95% CI, 0.86-1.86; FDR-adjusted P < .0001) and 0.22 cm H2O (95% CI, −0.28 to 0.72; FDR-adjusted P = .46), respectively. Refer to Supplemental Digital Content, Table 2, https://links.lww.com/AA/D395, for the data summary at prescribed PEEP levels of 0 and 5 cm H2O.
Figure 5.: Difference in PIP when disconnecting the WAGS on the Dräger Perseus A500 (A), Dräger Apollo (B), and GE Avance CS2 (C) with 95% confidence intervals denoted by vertical bars. FGF indicates fresh gas flow; LPM, liters per minute; PEEP, positive end-expiratory pressure; PIP, peak inspiratory pressure; WAGS Off, waste anesthesia gas suction disconnected; WAGS On, waste anesthesia gas suction connected.
After removing the hose connected to the AGSS and the visual indicator bag on the GE Avance CS2 according to the manufacturer’s recommendations for off-label use in the ICU (Figure 2),10 the differences in PEEP and PIP were small for FGFs 2, 4, 6, 8, and 10 LPM compared to when the WAGS was connected under normal operation, as shown in Figure 6. At FGF of 2 LPM and PEEP 0 cm H2O, the PEEP difference was 0.12 cm H2O (95% CI, 0.06-0.18; FDR-adjusted P = .0002). Increasing FGF and PEEP levels had similar findings until FGF reached 15 LPM. At FGF of 15 LPM with PEEP set at 0 and 5 cm H2O, the difference in PEEP was −7.98 cm H2O (95% CI, −8.11 to −7.85; FDR-adjusted P < .0001) and −6.70 cm H2O (95% CI, −6.83 to −6.57; FDR-adjusted P < .0001), respectively. Similarly, the differences in PIP were −6.04 cm H2O (95% CI, −6.53 to −5.55; FDR-adjusted P < .0001) and −5.66 cm H2O (95% CI, −6.15 to −5.17; FDR-adjusted P < .0001), respectively. Refer to Supplemental Digital Content, Table 3, https://links.lww.com/AA/D395, for the complete data summary.
Figure 6.: Difference in PEEP (A) and PIP (B) when disconnecting both the WAGS and the visual indicator bag from the AGSS on the GE Avance CS2 with 95% confidence intervals denoted by vertical bars. AGSS indicates anesthesia gas scavenging system; FGF, fresh gas flow; LPM, liters per minute; PEEP, positive end-expiratory pressure; PIP, peak inspiratory pressure; WAGS Off, waste anesthesia gas suction disconnected; WAGS On, waste anesthesia gas suction connected.
Each mechanical ventilator’s AGSS was then connected to medical vacuum, as shown in Figure 1D–F. At 15 LPM, the GE Avance CS2 at prescribed PEEP levels of 0 and 5 cm H2O had clinically significant differences in PEEP of −8.02 cm H2O (95% CI, −8.09 to −7.95; FDR-adjusted P < .0001) and −6.78 (95% CI, −6.85 to −6.71; FDR-adjusted P < .0001), respectively, when compared to the use of WAGS. Clinically significant differences in PIP of −6.26 cm H2O (95% CI, −6.78 to −5.74; FDR-adjusted P < .0001) and −6.02 cm H2O (95% CI, −6.54 to −5.50; FDR-adjusted P < .0001) were also seen at these settings respectively. The Dräger Apollo demonstrated statistically significant differences in PIP throughout the tested parameters and reached clinical significance 5 times; there was no clinically significant differences in PEEP. The Dräger Perseus A500 did not show any clinically significant difference in PEEP or PIP when the medical vacuum was connected to its AGSS. Refer to Supplemental Digital Content, Table 4, https://links.lww.com/AA/D395, for the data analysis at PEEP of 0 and 5 cm H2O.
DISCUSSION
The incidence of COVID-19 is increasing and is projected to infect 30% (96 million) of the US population with an estimated mortality rate of 1.8%–3.4%.15 Of those infected, roughly 5% (4.8 million) will require hospitalization, and multiple prediction models anticipate the number of critically ill patients will outnumber available ICU beds within weeks.1–3 About 20%–31% (0.96–1.49 million) of hospitalized patients will require mechanical ventilation for an undetermined period of time.16,17 A 2010 survey estimated that US acute care hospitals have a cumulative total of approximately 62,000 full-feature ventilators.18,19 As a result, an urgent need for ICU ventilators has arisen due to this unanticipated shortfall, and the use of anesthesia machines can help fulfill this need.4,20
To repurpose anesthesia machines for ICU use, scavenging system design should be considered. Open AGSS are designed to relieve internal pressure build-up by releasing this pressure into the surrounding environment. Therefore, this design is less likely to inadvertently increase airway pressures when disconnected from WAGS, but has a higher risk of atmospheric contamination. In contrast, closed AGSS are designed to work specifically with the WAGS connected, which utilize a pressure relief valve to limit the release of volatile anesthetic vapors to the surrounding environment. Specifically, the GE Avance CS2 with a closed AGSS complies with standards that allow a pressure increase of up to 15 cm H2O at the inlet of the AGSS. Once pressure builds up, the waste gas reservoir bag is filled until a built-in positive pressure relief valve allows release to the atmosphere at 10 cm H2O.6
Anecdotal accounts have reported possible inaccurate detection of airway pressures with a malfunctioning AGSS. However, our data show that there is generally clinically acceptable agreement in the measured pressures from the reference manometer and that of the tested anesthesia machines.
Both PEEP and PIP demonstrated clinically insignificant differences after the WAGS was disconnected from the open AGSS on the Dräger Perseus A500 and Dräger Apollo. However, on the GE Avance CS2, both PEEP and PIP had clinically significant increases after the WAGS was disconnected from the closed AGSS when the prescribed PEEP was set at 0 or 5 cm H2O. At a prescribed PEEP of 0 cm H2O, the PEEP could increase by a median of 8.62 cm H2O, and the PIP, by a median of 6.80 cm H2O. With the WAGS disconnected, the maximum difference we observed in prescribed PEEP and measured PEEP by the reference manometer was 12.4 cm H2O with FGF of 15 LPM and PEEP of 0 cm H2O. If a patient was being mechanically ventilated with similar settings, then this increase in airway pressures could induce barotrauma. These data support making modifications to the closed AGSS before safe utilization in the ICU without a WAGS connection.
Prior reports of AGSS obstruction have indicated that its malfunction or alteration could deliver higher airway pressures.7 GE Healthcare recommended if an anesthesia machine with a closed AGSS is utilized without the WAGS, then clinicians should remove the hose connected to the AGSS and the visual indicator bag.10 Our study has shown that this recommendation effectively resolved the potential pressure build-up in the anesthesia circuit for the GE Avance CS2. There were no clinically significant increases in airway pressures.
The differentiation between waste anesthesia gas disposal (WAGD) and medical vacuum had been made to prevent oxidizers, such as oxygen and nitrous oxide, from reaching the oil lubricated pumps commonly used in vacuum systems and, thereby, reduce the risk of fire and explosion.21 However, many hospital systems have dual use WAGD/medical vacuum systems, where WAGD gases will mix with vacuumed gases to dilute the oxidizers. Previous models of Dräger Apollo and Dräger Fabius AGSS’s are equipped with fittings to be connected to vacuum suction. Since WAGD connections are not readily available in many ICUs, vacuum suction can be easily attached to these machines to reduce contamination of the immediately surrounding environment with expiratory gases. Doing so would also obviate the need for modifications to ensure appropriate airway pressures. However, manufacturers do not recommend connecting the AGSS to medical vacuum suction due to the potential risk of fire with high fractions of oxidizers reaching oil lubricated pumps.
The primary limitation of this study is that not all types of anesthesia machines were tested. Representative samples of both open and closed AGSS were studied among 3 commonly used anesthesia machines from 2 leading manufacturers. Clinicians should also note that open AGSS options exist for the GE Aespire, Avance, and Aisys (GE Healthcare).7 Therefore, each institution must confirm the type of AGSS used in its anesthesia machines and make the necessary adjustments before implementation as ICU ventilators.5 In addition, if the AGSS is not utilized, then the use of viral filters and removal of vaporizers must be considered to reduce the risk of atmospheric contamination with both bioburden and volatile anesthetics.5 Anesthesiologists are most familiar with the use of their workstations, aware of the potentially increased risk of barotrauma, and best trained in addressing the limitations of using these machines as ICU ventilators, as compared to respiratory therapists who typically operate ICU-based ventilators.5,9 Note also that this study did not address a COVID-19 patient’s significantly reduced lung compliance and how it may change throughout recovery. The low variability of our simulated lung’s compliance may differ from clinical observations, which necessitates anesthesiologists’ vigilance.
Displayed airway pressure measurements are clinically accurate in the setting of disconnecting the WAGS for the tested anesthesia machines. Therefore, the variability in the differences between the machine and reference manometers are expected to be low in clinical observations, as well. The Dräger Perseus A500 and Apollo with open AGSS do not deliver inadvertent CPAP when the WAGS is disconnected, but the GE Avance CS2 with a closed AGSS does at low prescribed levels of PEEP. This increase in circuit pressure can be mitigated by instituting the manufacturer’s recommended alterations. Anesthesiologists should be aware of the potential clinically important increases in pressure that may be unintentionally delivered on some anesthesia machines, should the WAGS not be properly connected.
DISCLOSURES
Name: Vinh Pham, MD.
Contribution: This author helped design the study, analyze the statistics, and prepare, review, and approve the final manuscript.
Name: Le Nguyen, MD.
Contribution: This author helped design the study, collect the data, and prepare, review, and approve the final manuscript.
Name: Riley J. Hedin, DO, MPH.
Contribution: This author helped design the study, collect the data, and prepare, review, and approve the final manuscript.
Name: Courtney Shaver, MS.
Contribution: This author helped design the study, analyze the statistics, and prepare, review, and approve the final manuscript.
Name: Kendall A. P. Hammonds, MPH.
Contribution: This author helped design the study, analyze the study, and prepare, review, and approve the final manuscript.
Name: William C. Culp Jr, MD.
Contribution: This author helped design the study, collect the data, and prepare, review, and approve the final manuscript.
This manuscript was handled by: Thomas M. Hemmerling, MSc, MD, DEAA.
REFERENCES
1. Moghadas SM, Shoukat A, Fitzpatrick MC, et al. Projecting hospital utilization during the COVID-19 outbreaks in the United States. Proc Natl Acad Sci U S A. 2020;117:9122–9126.
2. COVID-19 Projections. Institute for health metrics and evaluation web site. 2020. Accessed April 4, 2020.
https://covid19.healthdata.org/united-states-of-america.
3. Henderson M, Kreiss-Tomkins J, Kofman I, Rosen Z, Shah N. COVID act now. 2020. Accessed April 4, 2020.
https://covidactnow.org/.
4. US Food & Drug Administration. Ventilator supply mitigation strategies: letter to health care providers.March 22, 2020. Accessed April 5, 2020.
https://www.fda.gov/medical-devices/letters-health-care-providers/ventilator-supply-mitigation-strategies-letter-health-care-providers.
5. American Society of Anesthesiologists, Anesthesia Patient Safety Foundation. APSF/ASA guidance on purposing anesthesia machines as ICU ventilators.Updated April 4, 2020. Accessed April 5, 2020.
https://www.asahq.org/-/media/files/spotlight/anesthesia-machines-as-icu-ventilators-4-4-20.pdf?la=en&hash=165C9C91FA07344DBF36ED360D9A13D15820B4E9.
6. Saxena S, Laykish M, Mangione M. Safety Issues with gas scavenging system in GE Avance and GE Aespire anesthesia machines. Anesthesia Patient Safety Foundation Newsletter. 2016;31:17.
7. Joyal JJ, Vannucci A, Kangrga I. High end-expiratory airway pressures caused by internal obstruction of the Draeger Apollo
® scavenger system that is not detected by the workstation self-test and visual inspection. Anesthesiology. 2012;116:1162–1164.
8. Cattano D, Henschel J. Obstruction to Dräger Apollo exhaust valve. Anesthesia Patient Safety Foundation Newsletter. 2013;28:36.
9. Ioannidis G, Lazaridis G, Baka S, et al. Barotrauma and pneumothorax. J Thorac Dis. 2015;7:S38–S43.
10. Lehtonen M. COVID-19 - requests for information regarding the off-label use of GE Healthcare anesthesia devices for ICU ventilation 2020.March 23, 2020. Accessed April 5, 2020.
https://www.gehealthcare.com/-/jssmedia/3c655c83bd6b427e9824994c12be0da5.pdf?la=en-us.
12. Healthcare GE. Avance CS
2 user’s reference manual, software revision 10.X. March 82018. Accessed April 5, 2020.
https://www.manualslib.com/manual/1210542/Ge-Auisys-Cs2.html.
13. Analyse-it Software L. Estimating nonparametric limits of agreement in non-normally distributed data.2020. Accessed April 1, 2020.
https://analyse-it.com/docs/tutorials/bland-altman/estimating-loa-percentile.
14. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B (Methodol). 1995;57:289–300.
15. CDC COVID-19 Response Team. Severe outcomes among patients with coronavirus disease 2019 (COVID-19) - United States, February 12–March 16, 2020. MMWR Morb Mortal Wkly Rep. 2020;69:343–346.
16. Zhou F, Yu T, Du R, et al. Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study. Lancet. 2020;395:1054–1062.
17. Fink S. Worst-case estimates for U.S. coronavirus deathsNew York Times. March 13, 2020. Accessed April 11, 2020.
https://www.nytimes.com/2020/03/13/us/coronavirus-deaths-estimate.html.
18. Rubinson L, Vaughn F, Nelson S, et al. Mechanical ventilators in US acute care hospitals. Disaster Med Public Health Prep. 2010;4:199–206.
19. Center for Health Security. Ventilator stockpiling and availability in the US. February 142020. Accessed April 11, 2020.
http://www.centerforhealthsecurity.org/resources/COVID-19/200214-VentilatorAvailability-factsheet.pdf.
20. Society of Critical Care Medicine. United States resource availability for COVID-19. March 192020. Accessed April 5, 2020.
https://www.sccm.org/Blog/March-2020/United-States-Resource-Availability-for-COVID-19.
21. Allen M. Waste anesthetic gas disposal (WAGD) systems. BeaconMedaes continuing education publication web site. 2004. Accessed June 7, 2020.
https://www.beaconmedaes.com/content/dam/brands/beacon-medaes/documents/brochures/nfpa-usa-/WAGD-paper.pdf.
