The clinical role of central venous pressure measurements

Central venous pressure (CVP) is commonly measured, but its clinical use is still not clear. We argue that the interpretation of the CVP needs to be considered in conjunction with an assessment of cardiac output. The objective of this study was to define an elevated CVP as one in which there is a low probability for cardiac output to increase with a volume infusion through a Starling mechanism by relating the initial CVP (measured relative to a reference point 5 cm below the sternal angle) to the response in cardiac output with volume infusion. The authors studied consecutive patients who had pulmonary artery catheters in place and who had a volume challenge as part of routine management as ordered by the treating physician. To ensure an adequate test of the Starling mechanism, data were included only if the volume infusion increased CVP by > or = 2 mm Hg. Responders were defined a priori as those with an increase in cardiac index > or = 300 and nonresponders as < 300 mL/min/m2. Patients failed to respond to volume infusion at all CVP values, and even 25% of those with CVP < 5 mm Hg were nonresponders. However, when CVP was > 10 mm Hg, physicians prescribed less fluid challenges, and when they did, a positive response was much less likely. Change in blood pressure or changes in urine output with volume infusion correlated poorly with change in cardiac index. A CVP of > 10 mm Hg should be considered high, and the probability of an increase in cardiac output with volume infusion is low. This value is a reasonable upper limit for algorithms for empiric fluid challenges.