Abstract 237: Assessment of Airway Placement During Out-of-Hospital Cardiac Arrest Using an Automated Capnogram Analysis Algorithm
Background: Misplacement of airway devices inserted by Emergency Medical Services (EMS) personnel during out-of-hospital cardiac arrest (OHCA) remains a life-threatening problem. Although the manual use of colorimetric end-tidal carbon dioxide (etCO2) detectors or real time capnography has reduced the incidence of misplaced airways, the validation process is cumbersome.
Objective: To validate the performance of a defibrillator-based capnogram (CO2) analysis program to provide an automated airway placement check that can be initiated immediately after airway placement and following any patient transfer.
Methods: Recordings from 210 OHCA patients with capnogram waveforms were collected from 2008 to 2010 in two EMS agencies based in Oregon&Texas. The capnogram files were extracted from storage cards of Philips HeartStart MRx defibrillators. Five-hundred (n=500) representative 20-second segments were selected and annotated as either success (n=344) or failure (n=156) by two emergency physicians with capnography expertise. Included were waveform segments from initial airway placement, after transfer to stretcher or ambulance, CPR artifact, and documented improper placements. Questionable waveforms were resolved via consensus review. This database was then randomly split into equal training and testing sets to tune and assess the performance of the automated C02 analysis program. Measurements incorporated into the automated C02 analysis program include the respiratory rate and the etCO2 value on the respiratory capnogram. The analysis program applied special techniques to eliminate CPR artifacts observed in the capnogram. The minimum criteria for consideration of a successful airway placement is two ventilated breaths associated with an etCO2 level > 3 mm Hg within a 20 second timeframe; additional logic looks for decaying etCO2 values.
Results: On the development set, the automated program achieved Se 92.3%, Sp 92.4%, PPV 84.7%, and positive and negative likelihood ratio of 12.2 and 0.08. On the testing set, we obtained Se 91.0%, Sp 92.4%, PPV 84.5%, and likelihood ratios of 12.0 and 0.10.
Conclusion: An automated capnogram analysis program can help EMS personal detect problematic airway placement during OHCA.
- © 2010 by American Heart Association, Inc.