Thermodilution cardiac output--an in vitro model of low flow states.

The accuracy and reproducibility of thermodilution cardiac output measurements were examined in vitro at low flows, using the Edwards cardiac output computer. For each of 18 different volumetrically measured flows between 130 and 1035 ml/min, three cardiac outputs were determined for each of four different injectate volumes (1,2,3, and 5 ml) at two different temperatures (0 degrees C and room temperature). There was a significant (p less than .001) correlation between measured flow and cardiac output for all injectate volumes at both temperatures. The slopes of the regression lines ranged between 0.97 and 1.25, and the y-intercepts were all greater than 0. Although this thermodilution technique overestimated cardiac output, it was a reproducible means of measuring cardiac output in this low-flow in vitro model.