Comparison of partial pressure of oxygen between compression-only CPR and standard CPR in simulated adult cardiac arrest - A manikin-based clinical modelling

Abstract

Background and Aims:

Compression-only cardiopulmonary resuscitation (CPR) has been shown to be as effective as conventional CPR, and oxygen supplementation during compression-only CPR may be beneficial. The study aimed to compare the arterial oxygen levels achieved while supplementing oxygen through high flow nasal cannula (HFNC) during compression-only CPR and bag-mask ventilation (BMV) during conventional CPR in simulated cardiac arrest scenarios on a high-fidelity simulator.

Methods:

The study included a simulated cardiac arrest created on a human patient simulator (HPS). The simulation included two sets of scenarios. In Simulation A, cardiac arrest was simulated on HPS, and compression-only CPR was provided by AutoPulse, and oxygen supplementation was provided using HFNC. In Simulation B, chest compression was provided by AutoPulse, and BMV was supplemented with oxygen at 15 L/min at a compression-to-ventilation ratio of 30:2. Both simulation scenarios were evaluated for three different starting PaO₂ values: 100 mmHg, 80 mmHg, and 60 mmHg. The change in PaO₂ and PAO₂ values was recorded every minute for 6 minutes. Statistical analysis was conducted using SPSS Statistics (Version 24.0; IBM, Armonk, NY), and P < 0.05 was considered statistically significant.

Results:

In Simulation A, at a starting PaO₂ of 100 mmHg, there was an increase in the PaO₂ at the 2^nd minute, which was sustained till the 6^th minute. PaO₂ values were persistently higher at all time points as compared to Simulation B (P < 0.001). At a starting PaO₂ of 80 mmHg, there was no change in PaO₂ in Simulation A as compared to a sustained fall in Simulation B (P < 0.001). At the starting PaO₂ of 60 mmHg, a decrease in PaO₂ was observed in both Simulation A and Simulation B (P = 0.57).

Conclusion:

In a simulated setting, compression-only CPR with HFNC results in better PaO₂ levels compared to conventional CPR with BMV.

INTRODUCTION

Early and high-quality cardiopulmonary resuscitation (CPR) is vital for adequate oxygenation and good neurological recovery.[1] American Heart Association guidelines updated in 2020 emphasised the need to provide bag-mask ventilation (BMV) or placement of an advanced airway in any cardiac arrest situation while minimising interruptions in compression.[2] Concerns about the aerosolisation of the virus from BMV during the recent pandemic resulted in several suggestions for modifications, such as compression-only CPR.

Compression-only CPR is proven to be as effective as conventional CPR, and oxygen supplementation during compression-only CPR can be beneficial.[3] Apnoeic oxygenation is now an established technique for maintaining oxygenation during prolonged periods of apnoea, such as during the intubation of a patient with a difficult airway.[4] However, its use in cardiac arrest situations, particularly during compression-only CPR, is not evaluated. The primary objective of this study was to assess the arterial oxygen levels in various cardiac arrest scenarios when supplementing oxygen using high flow nasal cannula (HFNC) during compression-only CPR on a high-fidelity manikin. The secondary objective was to compare this with the arterial oxygen levels achieved while supplementing oxygen during conventional CPR with BMV.

METHODS

This study was a technical simulation of high-quality CPR achieved through automated chest compressions delivered by the AutoPulse Resuscitation System Model 100 (ZOLL Medical Corporation, Chelmsford, MA, USA) on a high-technology full-scale simulator, the human patient simulator (HPS) version 10 (CAE Healthcare, USA) [Figure 1]. No ethical approval was required for this study, as it did not involve any human participants.

Figure 1.

Schematic representation of the human patient simulator and its components. CPR = Cardiopulmonary resuscitation; HFNC = High-flow nasal cannula; CO₂ = Carbon dioxide; N₂ = Nitrogen; O₂ = Oxygen; PAO₂ = Partial pressure of oxygen in alveoli; PaO₂ = Partial pressure of oxygen in arterial blood

The simulator is engineered with lungs that function as near-real clinical representations of alveolar gas exchange, created with nitrogen, oxygen, carbon dioxide, and medical-grade air. The uptake of alveolar gases is recreated by the computer software triggered by the sensors that recognise gas flows through conducting airways, and proportional exhaled gases with carbon dioxide are generated, which can then be measured using clinical patient monitors with end-tidal gas measurements. The lung mechanics are simulated by two bellows within the chest cavity of the manikin, which are propelled by pneumatics electronically controlled by signals processed from the lab rack [Figure 1]. The user interface is a software application in a desktop computer that is connected to this simulator. Parameters such as this simulator. Parameters such as were monitored virtually in the user interface and displayed on the headboard (instructor desktop), with all advanced vital parameters continuously displayed [Figure 2a].

Figure 2.

(a) Human patient simulator user interface and (b) Autopulse mounted on human patient simulator

The AutoPulse^TM is an automated, portable, battery-powered, load-distributing band chest compression device that delivers consistent chest compressions without interruption.[5] It consisted of a board on which the HPS manikin was mounted, with a lifeband encircling the chest. The stroke was triggered by mechanical shortening of the band by an electric motor, resulting in anteroposterior chest compression [Figure 2b]. HPS is an advanced high-technology simulator that can precisely simulate various physiological and pathological conditions through input from computer-based software.[6] The fundamental components of the HPS include a manikin attached to a central control unit (lab rack) via an umbilical assembly, which is connected to a desktop instructor workstation. The instructor’s desktop consists of programmable patient clinical states that can be dynamically altered through a set of instructions with corresponding changes to ‘patient physiology’.

Two groups of simulated scenarios were conducted: Simulation A – Compression-only CPR and Simulation B – Conventional CPR. In both simulations, the same HPS manikin was used to simulate cardiac arrest, and AutoPulse initiated compressions at 85 compressions/min (factory preset rate). Baseline parameters were recorded at three different simulated baseline PaO₂ levels of 100 mmHg, 80 mmHg, and 60 mmHg that were created by preparation of the manikin’s simulated alveoli. For each baseline PaO₂, data were recorded for room air and enriched oxygenation through high-flow nasal oxygen at 30 L/min (Simulation A) or conventional CPR with a compression ventilation ratio of 30:2 with BMV provided using a bag-valve mask with a reservoir at 15 L/min by an experienced anaesthesiologist (Simulation B). For oxygen supplementation at high flows to the manikin, a respirator (inspired^TM O2FLO) with a high-flow nasal cannula (HFNC, inspired^TM O2FLO nasal cannula size L) was applied to the manikin’s nostrils. PaO₂ and PAO₂ values were recorded from the HPS monitor every minute for 6 minutes during CPR.

PaO₂ and PAO₂ values at varying time points were compared between Simulation A and Simulation B with and without oxygen supplementation. Statistical analysis was conducted using the Statistical Package for the Social Sciences (SPSS) statistics software, version 24.0 (International Business Machines Corporation (IBM Corp), Armonk, NY, USA). Intergroup differences in PaO₂ and PAO₂ were analysed by a two-way repeated-measure analysis of variance. A P value of <0.05 was considered statistically significant.

RESULTS

Our study demonstrated that, in Simulation A, at a starting PaO₂ of 100 mmHg, there was an increase in PaO₂ at the 2^nd minute, which was sustained until the 6^th minute. PaO₂ values remained high at all time points when compared to Simulation B (P < 0.001). When the starting PaO₂ was 80 mmHg in Simulation A, the PaO2 values remained unchanged, and a sustained value of 80 mmHg was maintained throughout the 6 minutes. However, in Simulation B, there was a sustained decline in PaO₂ over 6 minutes, with a considerable difference observed at the 6^th minute (P < 0.001) At the starting PaO₂ of 60 mmHg, a decrease in PaO₂ was observed in both Simulation A and Simulation B (P = 0.57) [Figure 3].

Figure 3.

Comparison of the partial pressure of oxygen in arterial blood (PaO₂) between Simulation A and Simulation B with oxygen supplementation for three different baseline PaO₂ values (100 mmHg, 80 mmHg, and 60 mmHg). *P < 0.05

When the entire experiment was repeated in room air, a decrease in PaO₂ was observed compared to the baseline in both scenarios at the end of 6 minutes. However, there was no difference between the scenarios at any time [Figure 4].

Figure 4.

Comparison of the partial pressure of oxygen in arterial blood (PaO₂) between Simulation A and Simulation B at room air for three different baseline PaO₂ values (100 mmHg, 80 mmHg, and 60 mmHg)

There was an increase in PAO₂ from baseline at 1 minute, which sustained till the 6^th minute in both scenarios, irrespective of the starting PaO₂. The increase in PAO₂ was higher at all time points in HFNC-enriched compression-only CPR when compared to conventional CPR.

DISCUSSION

Our simulation study found a higher level of PaO₂ in all scenarios with oxygen supplementation than with CPR performed with room air. The levels of PaO₂ achieved were higher with HFNC than with BMV_. The improvement in PaO₂ in either modality was observed only if the initial PaO₂ was greater than 60 mmHg. This highlights the vital consideration for advocating supplemental oxygen during CPR, irrespective of whether controlled ventilation was provided to facilitate transalveolar gas exchange.

Simulation of clinical effects of high-quality chest compressions with variations in airway management methods is technically challenging in low-fidelity simulators. Existing high-technology manikins are not complete in measuring all components of the effectiveness of high-quality CPR. Many provide good feedback on the adequacy of chest compressions, utilising both visual and numerical metrics, and simulate advanced parameters with alterations in lung mechanics, such as compliance and resistance. HPS is known to have precise simulated alveolar gas exchange measurements corresponding to the tracheal flux of inspired gases. It further reciprocates to changes in inspired oxygen concentrations and reliably simulates cardiovascular physiology in cardiac arrest scenarios. In HPS, correct hand placement, depth, and rate of compressions are reflected in physiological feedback rather than a virtual target on the instructor’s workstation. Adequate chest compressions in HPS result in simulated circulation, cardiac output, central and peripheral blood pressures, and return of carbon dioxide. Similarly, many studies have found that the use of Autopulse for chest compression in out-of-hospital cardiac arrest resulted in significantly higher rates of return of spontaneous circulation (ROSC), hospital admission, and hospital discharge as compared to manual compressions.[7,8,9] Compared to manual compressions, Autopulse produces more effective circumferential compression, which raises intrathoracic pressure and creates forward blood flow. Therefore, the same effect is obtained at a compression rate of 85/min as opposed to a compression rate of 100–120 manual chest compressions per minute. The combination of these two validated technologies rendered our study set-up a reliable contemporaneous model for measuring changes in gases during cardiac arrest and CPR.

Ventilation is essential to ensure adequate oxygenation and CO₂ washout. Several studies have demonstrated the value of supplemental oxygen during CPR and ventilation.[10] Supplemental oxygen improves blood oxygenation and, therefore, tissue oxygenation during CPR. Higher PaO₂ improves the likelihood of ROSC and better Glasgow outcome scales during CPR.[11,12] The results of our study endorse the importance of considering emergency oxygen sources (oxygen cylinders or battery operated oxygen concentrators) in the emergency ambulance system, while in transit and during ongoing resuscitation in out-of-hospital arrest, and to enhance inspired oxygen concentration (FiO₂) during CPR for in-hospital arrest and near collapse.

A systematic review and meta-analysis assessing the clinical efficacy of HFNC therapy as apnoeic oxygenation in critically ill patients requiring endotracheal intubation in the intensive care unit (ICU) found that HFNC was non-inferior to the standard of care during endotracheal intubation when assessed for incidence of severe hypoxemia, mean lowest oxygen saturation, and in-hospital mortality.[13] The risk ratio of severe hypoxemia decreased with increasing baseline PaO₂ to the FiO₂ ratio. However, no studies or reports show that HFNC was used to supplement oxygen during CPR. We demonstrated that HFNC can achieve higher PaO₂ levels and hence may contribute to high-quality CPR.

Supplementing oxygen through HFNC as a modality of apnoeic oxygenation during compression-only CPR may offer several advantages and potentially replace BMV. It displaces dead space gases with oxygen through the ‘bulk-flow’ principle while providing PEEP to prevent alveolar derecruitment during ventilatory pauses, resulting in increased functional residual capacity with enriched oxygen. Patients develop metabolic acidosis during cardiac arrest, and high-flow oxygenation may also aid in improving pH by facilitating carbon dioxide washout of the nasopharyngeal space.[14] Passive ventilation achieved during compression-only CPR is inadequate to achieve optimum alveolar ventilation,[15,16] and application of HFNC could facilitate a higher chance of oxygenation and alveolar gas exchange during compression-only CPR. Although our study found improved oxygenation with HFNC during CPR, these findings are based on simulated normal lung conditions. For patients with associated co-morbidities such as obesity and lung diseases, the effect of HFNC-assisted oxygenation may differ and needs to be studied. Thus, the utility of HFNC during CPR needs clinical validation in real-life scenarios.

This study has several strengths. High-technology full-scale simulation provided us with the environment to create several controlled conditions, as expected in real-life situations, by altering PaO₂ levels. The ability to preset the baseline PaO₂ to 100, 80, and 60 mmHg allowed us to explore the three states as a functional baseline and then add additional variables, such as ventilation or no ventilation, to explore the type of oxygenation, including BMV versus HFNC. This may be difficult to replicate in real-life settings as all these components cannot be controlled in the same setting of a cardiac arrest. The presence of a calibrated simulated alveolar milieu allowed for the understanding of time-based changes in arterial oxygenation (baseline, 1 minute, and 6 minutes). Furthermore, using a mechanical compression device ensured that chest compressions were standardised, uniform, and uninterrupted, eliminating provider fatigue that could affect the quality of CPR.

Our study has a few limitations. One of the limitations of our study is that the data were recorded for only 6 minutes of CPR, with the AutoPulse battery capacity limited to 30 minutes. We chose this 6-minute period, as it is a typical time interval for planning a definitive airway. Second, the baseline PaO₂ of 80 mmHg and 60 mmHg were simulated in the manikin by adjusting the lung mechanics (A-a gradient; the higher the gradient, the lower the present PaO₂ established). The clinical effect of using HFNC in patients with low PaO₂ may vary in real-life scenarios, as hypoxemia can be multifactorial, resulting from factors beyond alveolar gas exchange (A-a gradient) and the consistency of chest compressions (Autopulse usage), which we optimised in this study. Future clinical studies can measure PaO₂ during cardiac arrest and study the corresponding clinical outcomes of CPR and oxygenation. The study has various bearings in different settings that provide resuscitation: out-of-hospital, in-hospital, high-dependency care units, ICUs, and operating theatres, and the implications can be variable based on adherence to standard practice.

Our study also highlights the difficulties in precisely replicating clinical effectiveness in a simulated high-fidelity setting. The physiological capabilities of these modern simulators require improvements in oxygenation parameters to ensure they are valid and reliable measures of the clinical effectiveness of CPR. Additionally, the bulkiness of the mechanical components in high-technology simulators, as well as their fixity to predesigned simulation centres with gas supplies, makes it more complex. There is a need for the development of affordable simulators that can achieve the same level of functionality as demonstrated in the present study.

CONCLUSION

Compression-only cardiopulmonary resuscitation with enriched oxygenation through high flow nasal cannula results in better partial pressure of oxygen in arterial blood levels than conventional CPR with bag-mask ventilation in a simulated setting. The study findings need to be corroborated in clinical trials to establish clinical utility.

Conflicts of interest

The authors declare that they have no conflicts of interest. Dr. Pankaj Kundra, who is one of the co-authors of this manuscript, is an Editor of this journal. He was not involved in any decision-making process, and an independent editor handled this manuscript.

Study data availability

De-identified data may be requested with reasonable justification from the authors (email to the corresponding author) and shall be shared upon request.

Authors contributions

All authors were involved in Concepts, Design, Definition of intellectual content, Data acquisition, Data analysis, Manuscript preparation, Manuscript editing, Manuscript review.

Presentation at conferences/CMEs and abstract publication

None.

Disclosure of use of artificial intelligence (AI)-assistive or generative tools

The AI tools or language models (LLM) have not been utilised in the manuscript, except that software has been used for grammar corrections and references.

Declaration of Use of Permitted Tools

The scales, scores, figures, and tables are not copyrighted.

Funding Statement

Nil.