Despite the inclusion of ventilation recommendations in the cardiopulmonary resuscitation guidelines, widespread monitoring of ventilation has yet to happen. Ventilation monitoring has been performed using airway pressure measurements,,  and algorithms have also been developed that use the capnogram signal and the thoracic impedance signal singly, or in combination, to monitor ventilation rate. The airway pressure measurements were more accurate, but airway pressure measurements are not commonplace in the pre-hospital and emergency department setting. Ventilation rate detected in the chest wall impedance or capnogram signal tended to underestimate the ventilation rate. The principle obstacle to ventilation rate measurement has been the removal of the chest compression artifact from the positive pressure ventilation signal.

In this issue of Resuscitation, Aramendi et al. report the development of a classifier algorithm that is able to determine the location of ventilations within the capnogram with a sensitivity above 99% and a positive predictive value of 97% during ongoing CPR. The classifier algorithm identifies each possible ventilation, and then analyzes five capnography waveform features (duration of inspiration, baseline CO2 value, expiratory plateau CO2 value, area under the first second of the expiratory plateau, and relative CO2 increase) for each possible breath to determine if the possible breath should be categorized as a breath or a chest compression artifact. The classifier is adaptive, and therefore can accommodate reasonable changes in ventilation rate, duty cycle, volume, and chest compression waveform. Interested readers should see the appendix for a complete description of the algorithm.

The development of this classifier algorithm is an important advance. The algorithm reported by Aramendi et al. has automated the detection of the beginning of the inspiratory phase of a ventilation during CPR in patients that were monitored by capnography. The information provided by that classifier is sufficient to report ventilation rate. Therefore, this classifier has enabled the widespread monitoring of a previously unmeasured CPR guideline recommendation. Their choice to use the capnogram for this purpose, as opposed to developing a new sensor, provides additional benefit because there are multiple public and corporate resuscitation databases containing capnography data on relatively large numbers of patients. Thus, it significantly expands our capacity to retrospectively analyze these databases to explore the effect of ventilation rate on outcomes from cardiac arrest.

The ability to monitor ventilation rate, both retrospectively and prospectively, will benefit the field of resuscitation because the literature is surprisingly unclear in this area. There is prospective observational human data that suggest that hyperventilation results in worse outcomes. Animal research designed to replicate the scenarios observed in the human data suggest that the negative effect of hyperventilation is due to increased intra-thoracic pressure, increased right atrial pressure during chest decompression, and lower coronary perfusion pressure during hyperventilation. However, a subsequent animal study suggests that CPR hemodynamics are relatively unaffected by changes in ventilation rate or ventilation volume. It has also been reported that positive pressure ventilation timed to coincide with chest compression (e.g. a ventilation rate of 100?min?1) to manipulate intra-thoracic pressure can improve hemodynamics. Finally, there is the research into compression-only CPR that suggests that omitting ventilations in favor of focusing on delivery of guideline compliant compressions during the earliest phase of resuscitation does not harm outcomes.

The guidelines cite this literature, and the language of the recommendations is purposefully weak. The AHA guidelines state that “it may be reasonable for the provider to deliver 1 breath every 6 seconds” and the ERC guidelines “suggest a ventilation rate of 10?min?1 during continuous chest compressions with an advanced airway based on very limited evidence.” Aramendi and colleagues have created a tool that will help us better understand the interaction between chest compression generated blood flow and ventilation rate.

There are limitations to the work that Aramedi and colleagues have reported. Ventilation rate is an incomplete metric of ventilation. Both the guidelines and the supporting articles provide a more complete definition of a ventilation, which includes descriptions of positive pressure inspiratory duration (1?s), a tidal volume sufficient to move the chest (500–600?ml) and duty cycle., , ,  The distillation of the guidelines into action items has tended not to emphasize this description and this classifier does not currently measure inspiratory duration or tidal volumes. It is easy to imagine that ventilation rates of 25?min?1 may result in smaller tidal volumes and shorter inspiratory durations. It is expected that variations in the ventilation duty cycle and tidal volume will lead to significant variations in the hemodynamic-ventilation interaction. Therefore, ventilation rate may not always be sufficient to differentiate beneficial ventilation from harmful ventilation.

Generally speaking, research into quality of CPR has been based on two significant and limiting assumptions: there is a single “best” description of CPR (e.g. all patients are the same), and the “best” description of CPR is the same 2?min into resuscitation as it is 20?min into resuscitation (e.g. the cardiac arrest pathology is time invariant). How these assumptions affect our efforts to improve outcomes from cardiac arrest is best exemplified by our experience with measuring chest compression depth. The advent of the depth measurement led to research showing that rescuers did a poor job following the CPR guidelines.,  The depth measurement made it possible to train people better, and to provide feedback when compressions did not meet guideline standards. Unfortunately, it remains unclear that this effort has directly improved outcomes. Meta-analyses suggest that adherence to guideline standards improves outcomes and retrospective analyses of large datasets have suggested the existence of an optimal chest compression rate and an optimal depth., However, reports of quality improvement have not independently resulted in improved outcomes. Given the difficulty we have experienced trying to translate CPR depth and rate guidelines into saved lives, and the fact that the ventilation guidelines are similarly rescuer centric, it is possible that we will also struggle to show that adherence to this guideline improves outcomes.

Despite the limitations mentioned above, human physiology allows for significant interactions between ventilation and cardiac output, and ventilation can be delivered in a manner that interferes with chest compression generated blood flow. Because we have struggled to measure ventilation during resuscitation, the interaction between ventilation and CPR generated blood flow remains unclear. Aramendi and colleagues have developed a new tool to study that interaction, and the research enabled by this classifier will move the field of resuscitation forward.

Conflict of interest statement

References