Outcomes and interventions in patients transported to hospital with ongoing CPR after out-of-hospital cardiac arrest - An observational study.

INTRODUCTION: The main objective was to present characteristics and outcome of patients without sustained field return of spontaneous circulation (ROSC) transported to hospital with ongoing cardiopulmonary resuscitation (CPR). Our secondary objectives were to investigate hospital-based interventions and the performance of the universal Termination of Resuscitation-rule (uTOR). METHODS: In this retrospective observational cohort study, out-of-hospital cardiac arrest (OHCA) patients arriving to the emergency department of a university hospital in Sweden during a six-year period (2010-2015) were identified using a prospectively recorded hospital-based registry. Additional data were retrieved from medical records and from the Swedish cardiopulmonary resuscitation registry. RESULTS: Among 409 patients transported with ongoing CPR, 7 survived to hospital discharge (1.7%). Hospital-based interventions against a suspected cause of arrest were attempted during ongoing resuscitation in 34 patients (8.3%), of whom 3 survived to hospital discharge. The remaining 4 survivors had spontaneous in-hospital ROSC. Survivors presented with either a shockable rhythm (n = 4) or pulseless electrical activity (n = 3). The uTOR identified non-survivors with a positive predictive value (PPV) of 98.4% and a specificity of 71.4% for termination. CONCLUSION: Survival after OHCA where sustained prehospital ROSC is not achieved is rare and available in-hospital interventions are rarely utilised. No patient with asystole as the first recorded rhythm survived. The uTOR identified non-survivors with a PPV of 98.4% but showed poor specificity.

Out-of-hospital cardiac arrest termination of resuscitation with ongoing CPR: An observational study.

INTRODUCTION: Termination of resuscitation guidelines for out-of-hospital cardiac arrest can identify patients in whom continuing resuscitation has little chance of success. This study examined the outcomes of patients transferred to hospital with ongoing CPR. It assessed outcomes for those who would have met the universal prehospital termination of resuscitation criteria (no shocks administered, unwitnessed by emergency medical services, no return of spontaneous circulation). METHODS: A retrospective cohort study of consecutive adult patients who were transported to hospital with ongoing CPR was conducted at three hospitals in the West Midlands, UK between September 2016 and November 2017. Patient characteristics, interventions and response to treatment (ROSC, survival to discharge) were identified. RESULTS: 227 (median age 69 years, 67.8% male) patients were identified. 89 (39.2%) met the universal prehospital termination of resuscitation criteria. Seven (3.1%) were identified with a potentially reversible cause of cardiac arrest. After hospital arrival, patients received few specialist interventions that were not available in the prehospital setting. Most (n = 210, 92.5%) died in the emergency department. 17 were admitted (14 to intensive care), of which 3 (1.3%) survived to hospital discharge. There were no survivors (0%) in those who met the criteria for universal prehospital termination of resuscitation. CONCLUSION: Overall survival amongst patients transported to hospital with ongoing CPR was very poor. Application of the universal prehospital termination of resuscitation rule, in patients without obvious reversible causes of cardiac arrest, would have allowed resuscitation to have been discontinued at the scene for 39.2% of patients who did not survive.

Machine learning "red dot": open-source, cloud, deep convolutional neural networks in chest radiograph binary normality classification.

AIM: To develop a machine learning-based model for the binary classification of chest radiography abnormalities, to serve as a retrospective tool in guiding clinician reporting prioritisation. MATERIALS AND METHODS: The open-source machine learning library, Tensorflow, was used to retrain a final layer of the deep convolutional neural network, Inception, to perform binary normality classification on two, anonymised, public image datasets. Re-training was performed on 47,644 images using commodity hardware, with validation testing on 5,505 previously unseen radiographs. Confusion matrix analysis was performed to derive diagnostic utility metrics. RESULTS: A final model accuracy of 94.6% (95% confidence interval [CI]: 94.3-94.7%) based on an unseen testing subset (n=5,505) was obtained, yielding a sensitivity of 94.6% (95% CI: 94.4-94.7%) and a specificity of 93.4% (95% CI: 87.2-96.9%) with a positive predictive value (PPV) of 99.8% (95% CI: 99.7-99.9%) and area under the curve (AUC) of 0.98 (95% CI: 0.97-0.99). CONCLUSION: This study demonstrates the application of a machine learning-based approach to classify chest radiographs as normal or abnormal. Its application to real-world datasets may be warranted in optimising clinician workload.