The effectiveness of ongoing cardiopulmonary resuscitation efforts is difficult to evaluate. Recent studies suggest that carbon dioxide excretion may be a useful noninvasive indicator of resuscitation from cardiac arrest. A prospective clinical study was done to determine whether end-tidal carbon dioxide monitoring during cardiopulmonary resuscitation could be used as a prognostic indicator of resuscitation and survival. Thirty-five cardiac arrests in 34 patients were monitored with capnometry during cardiopulmonary resuscitation during a 1-year period. Nine patients who were successfully resuscitated had higher average end-tidal carbon dioxide partial pressures during cardiopulmonary resuscitation than 26 patients who could not be resuscitated (15 +/- 4 vs 7 +/- 5 mm Hg). The 3 patients who survived to leave the hospital had a higher average end-tidal carbon dioxide partial pressure than the 32 nonsurvivors (17 +/- 6 vs 8 +/- 5 mm Hg). All 9 patients who were successfully resuscitated had an average end-tidal carbon dioxide partial pressure of 10 mm Hg or greater. No patient with an average end-tidal carbon dioxide partial pressure of less than 10 mm Hg was resuscitated. Data from this prospective clinical trial indicate that findings from end-tidal carbon dioxide monitoring during cardiopulmonary resuscitation are correlated with resuscitation from and survival of cardiac arrest.