Capnography and chest-wall impedance algorithms for ventilation detection during cardiopulmonary resuscitation.

OBJECTIVE

Hyperventilation is both common and detrimental during cardiopulmonary resuscitation (CPR). Chest-wall impedance algorithms have been developed to detect ventilations during CPR. However, impedance signals are challenged by noise artifact from multiple sources, including chest compressions. Capnography has been proposed as an alternate method to measure ventilations. We sought to assess and compare the adequacy of these two approaches.

METHODS

Continuous chest-wall impedance and capnography were recorded during consecutive in-hospital cardiac arrests. Algorithms utilizing each of these data sources were compared to a manually determined "gold standard" reference ventilation rate. In addition, a combination algorithm, which utilized the highest of the impedance or capnography values in any given minute, was similarly evaluated.

RESULTS

Data were collected from 37 cardiac arrests, yielding 438min of data with continuous chest compressions and concurrent recording of impedance and capnography. The manually calculated mean ventilation rate was 13.3+/-4.3/min. In comparison, the defibrillator's impedance-based algorithm yielded an average rate of 11.3+/-4.4/min (p=0.0001) while the capnography rate was 11.7+/-3.7/min (p=0.0009). There was no significant difference in sensitivity and positive predictive value between the two methods. The combination algorithm rate was 12.4+/-3.5/min (p=0.02), which yielded the highest fraction of minutes with respiratory rates within 2/min of the reference. The impedance signal was uninterpretable 19.5% of the time, compared with 9.7% for capnography. However, the signals were only simultaneously non-interpretable 0.8% of the time.

CONCLUSIONS

Both the impedance and capnography-based algorithms underestimated the ventilation rate. Reliable ventilation rate determination may require a novel combination of multiple algorithms during resuscitation.