Comparison of central venous and arterial pH and PCO2 during open-chest CPR in the canine model.

Arterial blood gases are difficult to obtain during cardiopulmonary resuscitation (CPR) in human beings, and the possibility of venous sampling is raised frequently. The reliability of central venous gases as a substitute for arterial blood gases in assessing acid base status, however, has not been investigated adequately under conditions of CPR. Therefore, femoral arterial and central venous catheters were placed in 24 mongrel dogs, and ventricular fibrillation was electrically induced. After varying predetermined downtimes from five to 60 minutes, open-chest CPR was begun, and arterial and central venous blood gases were simultaneously drawn every five minutes during a 30-minute period. Arterial pH (pHa) was consistently higher than central venous pH (pHcv) by an average of .048 units. A significant correlation existed between the pHa and pHcv at all times during CPR, with an overall r = .9771 (P less than .0001). The difference between central venous PCO2 (PcvCO2) and arterial PCO2 (PaCO2) was 5.17 mm Hg prior to cardiac arrest, but it increased 300% to a mean of 15.51 mm Hg during CPR. Correction of pHcv using conventional methods to account for this respiratory component decreased the correlation between pHa and pHcv to r = .6905. The ability of pHcv to substitute for pHa was assessed, and showed a sensitivity of 100% when pHa of 7.2 was used as a criterion for treatment. In this model, pHcv is a sensitive indicator of pHa and it may be used to guide bicarbonate therapy. The increased PcvCO2 during CPR probably results from the marked tissue lactic acid production and subsequent shift of the bicarbonate buffer into free carbon dioxide.