Ventilation During Cardiopulmonary Resuscitation: What Have We Learned From Models?

Chest Compression-Related Ventilation

In 1961, Safar et al¹⁰ reported that the substantial ventilation generated by chest compressions in healthy volunteers undergoing simulated CPR was reduced to almost nothing in cardiac arrest patients. This observation suggested that a change in respiratory system behavior could play an important role in the limitation of ventilation produced by the variation of intrathoracic pressure secondary to external chest compressions. Since the first report, other authors have demonstrated that chest compressions could generate variable tidal volumes that were often lower than dead space.^11,12 The mechanisms regulating the capability of chest compressions to produce alveolar ventilation had never been thoroughly investigated until recent observations. It is of great interest because animal models have shown that progressive hypoxemia, hypercapnia, and acidemia develop if adequate ventilation is not provided during prolonged CPR.^13–15 In contrast, some authors reported that continuous flow insufflation, providing a small amount of positive airway pressure, allowed adequate gas exchange during human CPR with minimal impact on circulation.^16,17

Thanks to a recent series of studies conducted on bench and human cadaver models, new hypotheses have been generated to explain these observations, based on a better understanding of respiratory system and airway response to chest compressions.

Thoracic Airway Closure and Lung Volume Reduction

In 5 out-of-hospital cardiac arrest subjects, Cordioli et al¹⁸ recorded extremely low ventilation associated with chest compressions, depicting a typical inspiratory flow-limited tracing during chest decompression at zero PEEP. The physiological explanation was that the lung volume reduction below the end-expiratory thoracic volume induced by chest compressions may provoke the closure of distal airways, leading to flow limitation. Similar results have also been indirectly suggested by other groups, but the exact mechanism has not been explicated and may be partially misinterpreted.^12,19 These findings are consistent, however, with classical physiological descriptions, which report that airways are likely to collapse as end-expiratory lung volume falls below the closing capacity.²⁰ Hence, intrathoracic airway closure is also documented during ARDS, which is characterized by lung volume loss, reduced functional residual capacity (FRC) due to alveolar flooding, and inflammatory edema.^21,22 In cardiac arrest patients, the lung volume reduction due to chest compressions is possibly enhanced by the loss of chest wall muscle tone.

These observations led to the hypothesis that lung volume is reduced during CPR and may result in intrathoracic airway closure, which may explain the low ventilation and gas exchange impairment often with during prolonged resuscitation. Results of experimental studies done in bench, animal, and human models supporting this theory are discussed in the following sections.

Bench Models of CPR Relative to FRC

An original thoracic lung model was designed to reproduce the mechanical properties of the thoracic compartment (ie, lung, airways, and thorax) during CPR, and this model was intentionally shaped to permit chest compressions (Fig. 1).¹⁸

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Mechanical lung model mounted on bellows and spring, designed to reproduce the physical properties of the thoracic compartment and to perform chest compressions. The artificial lung is connected to a compressible conducting airway connector to mimic airway collapse. Chest compressions are applied using a mechanical device (LUCAS 2). The precise description of the system and experiments are reported in Reference 18.

This model allows chest compressions to start from FRC and mimics the elastic recoil of the thorax during both the compression phase and the decompression phase of CPR. In addition, a compressible conducting connector applied at the airways creates a Starling resistor, which replicates airway collapse when the transmural pressure (ie, airway pressure − surrounding pressure) becomes < 0. A dedicated mechanical device is used to standardize chest compressions (LUCAS 2, Jolife AB/Physio-Control, Lund, Sweden). The settings allow the assessment of the effects of different ventilation strategies under comparable conditions of chest compressions.

In this study, tidal ventilation occurred at absolute lung volumes that were less than functional residual capacity. The results indicate that inspiration is expected to take place during thorax decompression because inspiratory flow occurs as a consequence of recoil pressure, while expiration occurs during chest compressions. Flow entering the thorax is limited by the collapsibility of the resistor that is mimicking airway closure, but this can be partially restored by the application of positive pressure at the airway opening; the latter also limits lung volume reduction, thereby facilitating alveolar ventilation.¹⁸

These observations, obtained in a model approximating physiology, were indirectly supported by the previously discussed clinical observations, showing significant lung volume reduction during CPR. The important limitation of this model is the extreme negative pressure induced by chest decompression in the system, which directly reflects the recoil forces. As discussed by Moore et al,²³ such negative pressure in the airways/thoracic compartment has never been reported in vivo. It is likely that the complex interaction between lungs and chest wall mitigates the negative pressure measured in the pleural space during chest decompression. Nevertheless, this artificial model described for the first time the theoretical behavior of the respiratory system during chest compressions. Of particular note is that ventilation during CPR was taking place below functional residual capacity, with inspiration occurring during negative airway pressure.

Ventilation and Interaction With Circulation in Animal Models

In animal models, the decline over time of ventilation produced by chest compressions alone has been reported in swine.¹⁵ Using dynamic computed tomography series in a pig model with ventricular fibrillation, Markstaller et al²⁴ showed that chest compressions alone led to large atelectatic areas and lower P_aO2 compared to animals actively ventilated early during CPR. Interestingly, in similar animal models, Brochard et al¹⁷ showed that positive airway pressure permitted the maintenance of efficient ventilation and adequate gas exchange with chest compressions only. These seemingly conflicting results may be explained by the major effect of positive airway pressure that could reverse airway closure and other impairments of the respiratory system during CPR.

Nevertheless, the debate on ventilation in animal models revolves around the effect of PEEP on perfusion, but with conflicting results. The potential negative effect of PEEP on blood flow during CPR comes from animal studies suggesting that it impairs venous return.²⁵ On the other hand, despite concerns regarding the circulatory effects of positive pressure, authors have observed a beneficial impact on survival, especially after asphyxial cardiac arrest.^13,26,27 This may be explained by the beneficial effect of PEEP to increase functional residual capacity and improve systemic oxygenation in a cardiac arrest model that directly result from severe respiratory failure.

To help reconcile the conflicting results obtained in animal models regarding circulatory and ventilatory effects of PEEP, the behavior of the animal respiratory systems during chest compressions should be better investigated. In fact, due to the differences between the shape of the thorax, distal bronchial tree architecture, and alveolar connections in swine compared to humans, one can't exclude that lung volume reduction and airway collapse may be less frequent, at least at the early stage of CPR. These distinctions are important because pigs are the most frequently used animal model for research on CPR.²⁸ To our knowledge, none of the animal studies have tested interventions after prolonged CPR, which could have allowed sufficient time for these phenomena to occur. Only one model of dogs receiving CPR, using external vest inflation and deflation, reported air trapping.²⁹

Finally, the most important concept regarding the impact of ventilation on circulation during CPR comes from animal studies describing the thoracic pump theory. During chest compressions, the chest recoil effect on intrathoracic pressure oscillation improves hemodynamics.³⁰ Indeed, incomplete chest wall decompression has a negative impact on perfusion pressure,³¹ and active compression-decompression improves it.³²

In an attempt to enhance the phenomenon of chest recoil during decompression with enhanced venous return through negative intrathoracic pressure, several investigators have used an active compression-decompression strategy with the use of a threshold valve to create negative pressure through the airways and thus improve circulation.³³ Interestingly, when active compression-decompression CPR with the use of an inspiratory impedance threshold valve was applied in a pig model, incremental PEEP improved oxygenation and left ventricular coronary perfusion pressure.³⁴ To explain these observations, one can hypothesize that lung volume loss expected with the threshold valve and the potential associated airway collapse may be partially relieved by PEEP. Another study in swine, which used constant positive pressure with continuous oxygen insufflation (15 L/min), did not show any difference in hemodynamics compared to chest compressions alone. Only the active compression-decompression CPR with the use of a threshold valve improved coronary and cerebral perfusion pressure.²³ Nevertheless, ventilation with continuous flow insufflation in that study was not sufficient to maintain adequate gas exchange and efficient CO₂ elimination. It is therefore of paramount importance to better understand the difference between animal and human models regarding ventilation during CPR.

From Clinical Observation to Human Model (Cadaver)

As previously discussed, the first observation of lung volume reduction during chest compressions (average of 335 mL) and associated intrathoracic airway closure was found in out-of-hospital cardiac arrest subjects.¹⁸ Recently, fresh and treated cadavers have been proposed as a model for studying CPR.³⁵ Among these, soft embalmed Thiel cadavers, which retain the body's natural look and feel, are reported to reliably reproduce human respiratory mechanics during cardiac arrest, thereby allowing realistic measurement during CPR.^36,37

One important advantage of the human cadaver model, besides its capacity to reflect human respiratory physiology, is the possibility to measure the intrathoracic pressure changes (ie, the pressure that drives the vascular pressure produced by chest compressions) with the help of esophageal manometry.³⁸ In addition, realistic simulation of CO₂ elimination and P_ETCO2 waveform analysis can be done with this model (Fig. 2).^37,38

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Experimental settings of the Thiel cadaver model. Flow and airway pressure measured at the airway opening were derived from a pneumotachometer. The CO₂ pattern was derived from capnography; CO₂ is delivered directly into the distal trachea (continuous flow: flat line). Continuous measures of intrathoracic pressure changes during chest compressions were made with the use of an esophageal catheter. P_aw = airway pressure; ITP = intrathoracic pressure.

Alveolar Ventilation Produced by Chest Compressions

As discussed at the beginning of this article, the tidal volume produced by chest compressions is extremely variable and is often lower than dead space, including after endotracheal intubation.^11,12 It is known that CPR yields significant alveolar de-recruitment and a reduction in compliance.^18,36,39 According to classic physiology, when lung volume decreases below a threshold value (ie, the closing capacity), small distal airways are likely to collapse; this could also be due to gas–liquid interfacing in small airways.²¹ Using a Thiel cadaver model, our group reported the existence of reproducible intrathoracic airway closure during CPR in humans.³⁶

In the presence of this intrathoracic airway closure, negative alveolar pressure produced by chest decompressions (recoil) is no longer transmitted to the airway opening, and no inspiratory flow is generated despite a significant pressure gradient; hence, no inspiratory tidal volume can be produced during the decompression phase of CPR (Fig. 3). Moreover, in the case of airway closure, the increase in airway pressure produced by positive-pressure ventilation is mostly required to re-open the airways. This means that intrathoracic airway closure may prevent ventilation, independent of the chest compressions strategy used.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Measures of airways occlusion in human cadaver. Typical patterns of pressure and flow transmission to airways by chest compression (measured with intrathoracic oscillations) at airways opening (A) and disappearance of transmission when airways are closed (B). PEEP is applied above the airway closing (dotted line) pressure (A), then no PEEP is applied (B). The oscillation of CO₂ induced by chest compression (A) disappear when airways are closed (B). P_aw = airway pressure.

In our model, we showed that the application of a small degree of PEEP to maintain airway patency permits the transmission of a pressure gradient produced by chest wall recoil to airway opening (Fig. 3), thereby enabling an inspired tidal volume and effective alveolar ventilation.³⁸ Regarding the impact of PEEP on the intrathoracic pressure levels and variations, further investigation is needed. As for the hemodynamic effect, no data exist in human subjects; the only hint comes from patients under anesthesia, where the application of PEEP does not change the venous return gradient in case of planned fibrillation.⁴⁰

Capnography During CPR

Because airway patency yields fresh gas flow in the system and airway closure impedes it, the existence of an intrathoracic airway closure can be noninvasively estimated by analysis of the CO₂/time waveform, using capnography systems during chest compressions.⁹

To reproduce the capnography distorted by chest compressions that we observed in real cardiac arrest patients, we set up an original bench and cadaver model incorporating CO₂ production.³⁸ The idea was to simulate CO₂ production to understand the physiological meaning of capnography patterns recorded at the airway opening.

When airways are patent, CO₂ concentration at airway opening falls during decompression because fresh gas flow enters the system, resulting in capnography oscillating steadily with the rate of chest compressions. When airways are closed, CO₂ at airway opening is stable, resulting in non-oscillating capnography (Fig. 3). Similar “distorted” capnography by chest compressions have been reported by Gutiérrez et al.⁴¹ Interestingly, non-oscillating capnography (suggesting airway closure) was observed in almost 60% of the subjects in this series, compared to < 30% in the series from Grieco et al,³⁸ where subjects were ventilated with 5 cm H₂O of PEEP. The airway opening index, based on capnographic analysis and assessing the degree of oscillations in the CO₂ signal, has been proposed by our group to detect airway closure during CPR. The use of PEEP after intubation during chest compressions consistently maintains airway patency, increases the airway opening index, and enhances alveolar ventilation produced by chest compressions. No effect is expected, neither harmful nor beneficial, when PEEP is set close to thoracic airway closure because the pressure in that case is transmitted minimally to the thorax. These findings open new opportunities to test ventilatory strategies that will impact the CO₂ pattern during CPR.

Overall, the development of a realistic human model to study respiratory physiology during CPR has revealed new concepts regarding the interaction between intrathoracic pressure and transmission to the airway opening.