Pulse contour-derived cardiac output for continuous hemodynamic monitoring is becoming popular in critical care. However, the data regarding its reliability during acute hemodynamic instability are inconsistent. This study was conducted to determine whether pulse contour- derived cardiac output truly reflects rapid hemodynamic changes. Methods: Hemorrhagic shock was created in seven anesthetized piglets by continuous blood withdrawal at a rate of 1 mL . kg(-1) . min(-1) for 20 minutes. Volume expansion with 10% hydroxyethyl starch 8 mL . kg(-1) was then administered for 5 minutes. Pulse contour-derived and thermodilution- derived hemodynamic parameters were compared. Results: Baseline thermodilution-derived cardiac index was 3.2 +/- 0.4 L.min(- 1).M-2. After exsanguination, it decreased to 2.1 +/- 0.3 L . min(-1).M-2 while pulse contour-derived cardiac index increased to 4.4 +/- 0.4 L.min(-1).M-2 ( p value