ObjectiveTo determine whether continuous semiquantitative assessment of end-tidal CO2could provide a highly sensitive predictor of return of spontaneous circulation during cardiopulmonary resuscitation (CPR).DesignProspective, clinical study.SettingPrehospital CPR.PatientsOne hundred twenty patients, during nontraumatic cardiac arrest.InterventionsEnd-tidal CO2values were measured continuously after tracheal intubation, and were categorized as the initial value, and as minimal and maximal values during the first 20 mins.Measurements and Main ResultsPresenting rhythm was asystole in 22 of the first 24 patients. Return of spontaneous circulation occurred in eight patients. Initial, minimal, and maximal end-tidal CO2values were significantly (p less than .01) higher in these patients than in the patients without return of spontaneous circulation. Cutoff values providing a 100% sensitivity and the highest specificity in predicting return of spontaneous circulation were found to be 10 torr for initial and maximal end-tidal CO2values, and 2 torr for the minimal end-tidal CO2value. The number of patients required to reject (with a risk error of less than .05) the hypothesis of an actual sensitivity of less than equals 90% for an observed sensitivity of 100% was found to be 95. In the second part of the study, this hypothesis was prospectively tested for initial and maximal end-tidal CO2values in the subsequent 96 patients. Presenting cardiac rhythm was asystole in 87 patients. Return of spontaneous circulation was obtained in 30 patients. The cutoff value of 10 torr for maximal end-tidal CO2during the first 20 mins after tracheal intubation provided an observed sensitivity of 100% in predicting return of spontaneous circulation with a specificity of 67%. This result allows rejection of the hypothesis of an actual sensitivity of less than equals 90% (p equals .042). By contrast, the observed sensitivity of initial end-tidal CO2was only 87%.ConclusionsEnd-tidal CO2represents a valuable tool for monitoring patients presenting with asystole during prehospital CPR. Fluctuations in end-tidal CO2during CPR and the utility of end-tidal CO2in detecting return of spontaneous circulation justify its continuous measurement. In addition, a high sensitivity (more than 90%) in predicting return of spontaneous circulation is prospectively demonstrated using the maximal end-tidal CO2during the first 20 mins after tracheal intubation, with a cutoff value of 10 torr. Such a prognostic indicator could be used for a more rational approach to prolonged CPR.(Crit Care Med 1996; 24:791-796)