A prospective, cross-over trial was performed comparing two different rates of precordial compression using end-tidal carbon dioxide as an indicator of the efficacy of cardiopulmonary resuscitation in 23 adult patients. A second purpose of this study was to determine the effect of audio-prompted, rate-directed chest compressions on the end-tidal carbon dioxide concentrations during cardiopulmonary resuscitation. Patients with cardiac arrest received external chest compressions, initially in the usual fashion without rate direction and then with rhythmic audiotones for rate direction at either 80 compressions per minute or 120 compressions per minute. Nineteen of 23 patients had higher end-tidal carbon dioxide levels at the compression rate of 120 per minute. The mean end-tidal carbon dioxide level during compressions of 120 per minute was 15.0 +/- 1.8 mm Hg, slightly but significantly higher than the mean level of 13.0 +/- 1.8 mm Hg at a compression rate of 80 per minute. However, end-tidal carbon dioxide levels increased rather dramatically when audiotones were used to guide the rate of chest compressions. Mean end-tidal carbon dioxide concentration was 8.7 +/- 1.2 mm Hg during standard cardiopulmonary resuscitation immediately before audio-prompted, rate-directed chest compression and increased to 14.0 +/- 1.3 mm Hg after the first 60 seconds of audible tones directing compressions. Using end-tidal carbon dioxide as an indicator of cardiopulmonary resuscitation efficacy, we conclude that audible rate guidance during chest compressions may improve cardiopulmonary resuscitation performance.