Comparison of pulsed Doppler and thermodilution methods for measuring cardiac output in critically ill patients

We obtained 145 consecutive cardiac output measurements in 38 critically ill patients, using the invasive thermodilution and the noninvasive pulsed Doppler methods. The mean thermodilution cardiac output (TDCO) was 5.7 +/- 1.87 L/min and the mean pulsed Doppler cardiac output (PDCO) was 5.16 +/- 1.66 L/min. The mean difference between the two measurements was 0.51 L/min with an SD greater than 1.6 L/min, reflecting the scattering of results. The overall correlation coefficient was .58. The intercepts were large and the regression equation some way from the line of equal values (TDCO = 2.28 + 0.66 PDCO). When the results were analyzed according to diagnosis or by group experience, there were some differences in the bias of the estimate; however, the SD of the difference between methods was greater than one liter/min in all groups. Thus, the pulsed Doppler method failed to estimate accurately TDCO in critically ill patients.