Reliability of continuous pulse contour cardiac output measurement during hemodynamic instability

Arterial pulse contour analysis is gaining widespread acceptance as a monitor of continuous cardiac output (CO). While this type of CO measurement is thought to provide acceptable continuous measurements, only a few studies have tested its accuracy and repeatability under unstable hemodynamic conditions. We compared continuous CO measurement using the pulse contour method (PCCO) before and after calibration with intermittent transpulmonary thermodilution cardiac output (TpCO).

We compared the two methods of CO measurements in 15 Landrace pigs weighing 20-25 kg in an experimental model of sepsis. Nine pigs were given an infusion of E. coli lipopolysacchride (LPS), and six pigs acted as controls. PCCO values before and after calibration (PCCO1 and PCCO2 respectively) were registered, and their errors relative to TpCO measurements were compared.

The mean coefficient of variation for repeated PCCO measurements was 6.85% for the control group, and 13.99% for the endotoxin group. The range of TpCO was 1.01-3.15 L/min. In the control group the bias +/-2SD was 0.11 +/- 0.53 L/min (TpCO vs PCCO1) and -0.02 +/- 0.38 L/min (TpCO vs PCCO2). In the endotoxin group, the agreement was poor between TpCO and PCCO1, 0.08 +/- 1.02 L/min. This improved after calibration (TpCO vs PCCO2) to 0.01 +/- 0.31 L/min.

In hemodynamically stable pigs, both pre- and post-calibration PCCO measurements agreed well with the intermittent transpulmonary thermodilution technique. However, during hemodynamic instability, and pre-calibration PCCO values had wide limits of agreement compared with TpCO. This was reflected by larger coefficients of variation for PCCO in hemodynamic instability. The error of PCCO measurement improved markedly after calibration, with bias and limits of agreement within clinically acceptable limits.