Variability in central venous pressure measurements and the potential impact on fluid management

In the intensive care unit (ICU) of our tertiary care university medical center, central venous pressure (CVP) measurements derived from bedside monitors differ considerably from measurements by trained intensivists using paper tracings. To quantify these differences, printed CVP tracings and concurrent respiratory waveforms were collected from 100 consecutive critically ill patients along with the corresponding monitor-displayed CVP. Four blinded intensivists interpreted the tracings. The mean difference between the intensivists and the monitor was -0.26 mmHg (95% confidence interval, +7.19 to -7.71 mmHg). Seventy-six percent of the paired measurements were within 2 mmHg, whereas 7% differed by more than 5 mmHg. To determine the potential clinical impact of these differences, we used the original Surviving Sepsis Campaign Guidelines for fluid administration based upon the measurement of CVP. For individual physicians, protocol-driven fluid management strategy would have differed in 19.2% to 25.3% of cases, dependent upon which measured value was chosen. Although protocol-driven strategies to direct fluid infusion therapy may improve outcomes, these interventions in a specific patient are dependent upon the method by which the CVP is measured.