Logo: to the web site of Uppsala University

Background

A major barrier to the analysis of ventilation waveform data collected during CPR is the presence of artefacts caused by chest compressions. This study describes the development and evaluation of an algorithm to extract parameters regarding ventilation volume, pressure, and frequency from pneumotachography waveform data collected during ongoing simulated CPR.

Method

Ventilation waveform data was collected from a pneumotachograph connected to the respiratory circuit of a ventilator and a test lung. Both regular ventilation and ventilation during simulated CPR were used to develop the algorithm. A grid search was employed to optimize the algorithm parameters compared to the ventilator settings. The parameters were then manually tuned using clinical data from ventilation during CPR. The performance of the algorithm was described in terms of the median error vs. the known ventilator settings in the simulated data.

Results

Compared to the ventilator settings, the largest systematic errors of the algorithm was an overestimation of peak pressures during asynchronous CPR (median error of 3 (IQR 0.3–5.8) cmH₂O), and an underestimation of inspiratory volumes during synchronous CPR (median error 46 (IQR −76 to 10) ml).

Conclusion

In an experimental setting, the developed algorithm provides a novel solution to measure ventilation parameters during ongoing chest compressions. The algorithm is freely available under an open-source licence for use and further development. Further studies will be needed to validate the algorithm.

Ventilation during cardiopulmonary resuscitation remains understudied with current guideline recommendations for the treatment relying on low level evidence and expert opinion. The aim of this doctoral project was to explore ventilation during cardiopulmonary resuscitation, both in the experimental and clinical setting.

Study I investigated whether a suction cup on a mechanical chest compression device intended to assist chest recoil affected the haemodynamics and ventilation in an experimental porcine model. No difference in EtCO2, as a measurement of cardiac output, or ventilation could be found, although the suction cup increased the coronary perfusion pressure.

In study II, ventilation parameters, haemodynamics, blood gases and lung injuries were compared between ventilation during continuous chest compressions and ventilations given during a pause of the chest compressions (30:2) in an experimental porcine model. Continuous chest compressions were associated with higher peak inspiratory pressure, lower EtCO2 and PaCO2. No differences were found with regards to lung injuries between the groups.

Study III aimed to develop and test a novel algorithm designed to extract accurate ventilation parameters from ventilation waveform signals, gathered during experimental CPR, in the presence of chest compression artefacts in the signal, that otherwise interferes with the parameter extraction. The algorithm was tested with a pneumotacography device and with mechanical ventilators giving ventilation parameters with known values. The algorithm deviated only slightly from the ventilator settings and outperformed the standard software of the pneumotachograph.

Study IV was an observational multicentre study that aimed to describe ventilation parameters during cardiopulmonary resuscitation. Patients were included from five sites, four out of hospital and one in hospital. Included in the study were 241 patients and 28120 ventilations. The ventilations were heterogenous and varied with airway modality and ventilation mode. Bag-valve-mask ventilations were associated with large levels of leakage and asynchronous ventilations with endotracheal tubes with high airway pressures. No obvious signs of hyperventilation were found.

Future research on cardiopulmonary resuscitation should when possible include measurements of ventilation, in order to deduce if the varying ventilation parameters affects outcomes and to decide optimal ventilation strategies for survival.

BACKGROUND: Ventilation during cardiopulmonary resuscitation (CPR) has long been a part of the standard treatment during cardiac arrests. Ventilation is usually given either during continuous chest compressions (CCC) or during a short pause after every 30 chest compressions (30:2). There is limited knowledge of how ventilation is delivered if it effects the hemodynamics and if it plays a role in the occurrence of lung injuries. The aim of this study was to compare ventilation parameters, hemodynamics, blood gases and lung injuries during experimental CPR given with CCC and 30:2 in a porcine model.

METHODS: Sixteen pigs weighing approximately 33 kg were randomized to either receive CPR with CCC or 30:2. Ventricular fibrillation was induced by passing an electrical current through the heart. CPR was started after 3 min and given for 20 min. Chest compressions were provided mechanically with a chest compression device and ventilations were delivered manually with a self-inflating bag and 12 l/min of oxygen. During the experiment, ventilation parameters and hemodynamics were sampled continuously, and arterial blood gases were taken every five minutes. After euthanasia and cessation of CPR, the lungs and heart were removed in block and visually examined followed by sampling of lung tissue which were examined using microscopy.

RESULTS: In the CCC group and the 30:2 group, peak inspiratory pressure (PIP) was 58.6 and 35.1 cmH₂O (p < 0.001), minute volume (MV) 2189.6 and 1267.1 ml (p < 0.001), peak expired carbon dioxide (PECO₂) 28.6 and 39.4 mmHg (p = 0.020), partial pressure of carbon dioxide (PaCO₂) 50.2 and 61.1 mmHg (p = 0.013) and pH 7.3 and 7.2 (p = 0.029), respectively. Central venous pressure (CVP) decreased more over time in the 30:2 group (p = 0.023). All lungs were injured, but there were no differences between the groups.

CONCLUSIONS: Ventilation during CCC resulted in a higher PIP, MV and pH and lower PECO₂ and PaCO₂, showing that ventilation mode during CPR can affect ventilation parameters and blood gases.

Introduction: The presented study aimed to investigate whether a mechanical chest compression piston device with a suction cup assisting chest recoil could impact the hemodynamic status when compared to a bare piston during cardiopulmonary resuscitation.

Methods: 16 piglets were anesthetized and randomized into 2 groups. After 3 minutes of induced ventricular fibrillation, a LUCAS 3 device was used to perform chest compressions, in one group a suction cup was mounted on the device's piston, while in the other group, compressions were per -formed by the bare piston. The device was used in 30:2 mode and the animals were manually ventilated. Endpoints of the study were: end tidal carbon dioxide, coronary and cerebral perfusion pressures, and brain oxygenation (measured using near infrared spectroscopy). At the end of the protocol, the animals that got a return to spontaneous circulation were observed for 60 minutes, then euthanized.

Results: No difference was found in end tidal carbon dioxide or tidal volumes. Coronary perfusion pressure and cerebral oxygenation were higher in the Suction cup group over the entire experiment time, while cerebral perfusion pressure was higher only in the last 5 minutes of CPR. A passive tidal volume (air going in and out the airways during compressions) was detected and found correlated to end tidal carbon dioxide.

Conclusions: The use of a suction cup on a piston-based chest compression device did not increase end tidal carbon dioxide, but it was associated to a higher coronary perfusion pressure.

Background The COVID-19 pandemic has presented emergency medical services (EMS) worldwide with the difficult task of identifying patients with COVID-19 and predicting the severity of their illness. The aim of this study was to investigate whether physiological respiratory parameters in pre-hospital patients with COVID-19 differed from those without COVID-19 and if they could be used to aid EMS personnel in the prediction of illness severity. Methods Patients with suspected COVID-19 were included by EMS personnel in Uppsala, Sweden. A portable respiratory monitor based on pneumotachography was used to sample the included patient's physiological respiratory parameters. A questionnaire with information about present symptoms and background data was completed. COVID-19 diagnoses and hospital admissions were gathered from the electronic medical record system. The physiological respiratory parameters of patients with and without COVID-19 were then analyzed using descriptive statistical analysis and logistic regression. Results Between May 2020 and January 2021, 95 patients were included, and their physiological respiratory parameters analyzed. Of these patients, 53 had COVID-19. Using adjusted logistic regression, the odds of having COVID-19 increased with respiratory rate (95% CI 1.000-1.118), tidal volume (95% CI 0.996-0.999) and negative inspiratory pressure (95% CI 1.017-1.152). Patients admitted to hospital had higher respiratory rates (p<0.001) and lower tidal volume (p = 0.010) compared to the patients who were not admitted. Using adjusted logistic regression, the odds of hospital admission increased with respiratory rate (95% CI 1.081-1.324), rapid shallow breathing index (95% CI 1.006-1.040) and dead space percentage of tidal volume (95% CI 1.027-1.159). Conclusion Patients taking smaller, faster breaths with less pressure had higher odds of having COVID-19 in this study. Smaller, faster breaths and higher dead space percentage also increased the odds of hospital admission. Physiological respiratory parameters could be a useful tool in detecting COVID-19 and predicting hospital admissions, although more research is needed.