Transtracheal Doppler cardiac output monitoring: comparison to thermodilution during noncardiac surgery.

Abstract

The validity of transtracheal Doppler (TTD) cardiac output (CO) monitoring during noncardiac surgery has not been established. A prospective evaluation was undertaken in 30 patients undergoing noncardiac surgery to assess the agreement between TTD and thermodilution measurements of CO. Linear regression, Bland-Altman analysis, and receiver operator characteristic (ROC) techniques were employed to evaluate the accuracy, reliability, and trending capability of TTD monitoring. A total of 250 simultaneous TTD and thermodilution CO values were compared. TTD and thermodilution CO measurements were highly correlated (P < 0.005, r = 0.84) and Bland-Altman analysis revealed a small systematic underestimation of thermodilution CO (mean bias = -0.25 L/min) with a SD of the bias of 0.88 L/min and a mean percent error of 12.4%. TTD performed particularly well in patients in whom the Doppler signal was stable throughout surgery and required minimal manipulation. In these patients, linear regression yielded the relation TTD CO = 0.96 thermodilution CO + 0.15 with a correlation coefficient r = 0.92. Mean percent error was 10.0% with a mean bias of -0.02 L/min and a SD of the bias of 0.58 L/min. The ability of TTD to track directional changes in thermodilution CO was evaluated by regression analysis and a ROC plot. Changes in TTD CO were highly correlated to changes in thermodilution CO (r = 0.81). ROC plots showed that changes in TTD CO reliably identified large (greater than 15%) changes in thermodilution CO with a sensitivity of 92% and a specificity of 87%. Clinical experience with the TTD device is needed to obtain accurate measurements.(ABSTRACT TRUNCATED AT 250 WORDS)