Recent studies of thoracic pressure changes during external cardiopulmonary resuscitation (CPR) suggest that there may be a significant difference in the rate of delivery of intravenous drugs when they are administered through the extrathoracic inferior vena cava (IVC) rather than the intrathoracic superior vena cava (SVC). Comparison of delivery of a radionuclide given using superior and inferior vena caval access sites was made during normal blood flow and during CPR. Mean times from injection to peak emission count in each ventricle were determined. There were no significant differences between mean peak times for SVC or IVC routes during normal flow or CPR. When peak times were corrected for variations in cardiac output, there were no significant differences between IVC and SVC peak times during normal flow. During CPR, however, mean left ventricular peak time, when corrected for cardiac output, was significantly shorter (P less than .05) when the SVC route was used. The mean time for the counts to reach half the ventricular peak was statistically shorter (P less than .05) in both ventricles with the SVC route during the low flow of CPR. This suggests that during CPR, increased drug dispersion may occur when drugs are infused by the IVC route and thus may modify the anticipated effect of the drug bolus. These results suggest that during CPR, both the cardiac output and the choice of venous access are important variables for drug delivery.