Background: Alternative cardiopulmonary resuscitation (CPR) algorithms, introduced to improve outcomes after cardiac arrest, have so far not been compared in randomized trials with established CPR guidelines.
Methods: 286 physician teams were confronted with simulated cardiac arrests and randomly allocated to one of three versions of a CPR algorithm: (1) current International Liaison Committee on Resuscitation (ILCOR) guidelines ("ILCOR"), (2) the cardiocerebral resuscitation ("CCR") protocol (3 cycles of 200 uninterrupted chest compressions with no ventilation), or (3) a local interpretation of the current guidelines ("Arnsberg", immediate insertion of a supraglottic airway and cycles of 200 uninterrupted chest compressions). The primary endpoint was percentage of hands-on time.
Results: Median percentage of hands-on time was 88 (interquartile range (IQR) 6) in "ILCOR" teams, 90 (IQR 5) in "CCR" teams (p = 0.001 vs. "ILCOR"), and 89 (IQR 4) in "Arnsberg" teams (p = 0.032 vs. "ILCOR"; p = 0.10 vs. "CCR"). "ILCOR" teams delivered fewer chest compressions and deviated more from allocated targets than "CCR" and "Arnsberg" teams. "CCR" teams demonstrated the least within-team and between-team variance.
Conclusions: Compared to current ILCOR guidelines, two alternative CPR algorithms advocating cycles of uninterrupted chest compressions resulted in very similar hands-on times, fewer deviations from targets, and less within-team and between-team variance in execution.
Keywords: adherence; cardiocerebral resuscitation; cardiopulmonary resuscitation (CPR); guidelines; randomized trial; simulation.