Echo is a good, not perfect, measure of cardiac output in critically ill surgical patients.

Abstract

Compared with a pulmonary artery catheter (PAC), transthoracic echocardiography (TTE) has been shown to have good agreement in cardiac output (CO) measurement in nonsurgical populations. Our hypothesis is that the feasibility and accuracy of CO measured by TTE (CO-TTE), relative to CO measured by PAC thermodilution (CO-PAC), is different in surgical intensive care unit patients (SP) and nonsurgical patients (NSP). Surgical patients with PAC for hemodynamic monitoring and NSP undergoing right heart catheterization were prospectively enrolled. Cardiac output was measured by CO-PAC and CO-TTE. Pearson coefficients were used to assess correlation. Bland-Altman analysis was used to determine agreement. Over 18 months, 84 patients were enrolled (51 SP, 33 NSP). Cardiac output TTE could be measured in 65% (33/51) of SP versus 79% (26/33) of NSP; p = 0.17. Inability to measure the left ventricular outflow tract diameter was the primary reason for failure in both groups; 94% (17/18) in SP versus 86% (6/7) NSP; p = 0.47. Velocity time integral could be measured in all patients. In both groups, correlation between PAC and TTE measurement was strong; SP (r = 0.76; p < 0.0001), NSP (r = 0.86; p < 0.0001). Bland-Altman analysis demonstrated bias of -0.1 L/min, limits of agreement of -2.5 and +2.3 L/min, percentage error (PE) of 40% for SP, and bias of +0.4 L/min, limits of agreement of -1.8 and +2.5 L/min, and PE of 40% for NSP. There was strong correlation and moderate agreement between TTE and PAC in both SP and NSP. In both patient populations, inability to measure the left ventricular outflow tract diameter was a limiting factor. Diagnostic tests or criteria, level III.