RESUMO
OBJECTIVE: To investigate the accuracy, precision, and trending ability of noninvasive bioreactance-based Starling SV and the mini invasive pulse-power device LiDCOrapid as compared to thermodilution cardiac output (TDCO) as measured by pulmonary artery catheter when assessing cardiac index (CIx) in the setting of elective open abdominal aortic (AA) surgery. DESIGN: A prospective method-comparison study. SETTING: Oulu University Hospital, Finland. PARTICIPANTS: Forty patients undergoing elective open abdominal aortic surgery. INTERVENTIONS: Intraoperative CI measurements were obtained simultaneously with TDCO and the study monitors, resulting in 627 measurement pairs with Starling SV and 497 with LiDCOrapid. MEASUREMENTS AND MAIN RESULTS: The Bland-Altman method was used to investigate the agreement among the devices, and four-quadrant plots with error grids were used to assess trending ability. The agreement between TDCO and Starling SV was associated with a bias of 0.18 L/min/m2 (95% confidence interval [CI] = 0.13 to 0.23), wide limits of agreement (LOA = -1.12 to 1.47 L/min/m2), and a percentage error (PE) of 63.7 (95% CI = 52.4-71.0). The agreement between TDCO and LiDCOrapid was associated with a bias of -0.15 L/min/m2 (95% CI = -0.21 to -0.09), wide LOA (-1.56 to 1.37), and a PE of 68.7 (95% CI = 54.9-79.6). The trending ability of neither device was sufficient. CONCLUSION: The CI measurements achieved with Starling SV and LiDCOrapid were not interchangeable with TDCO, and the ability to track changes in CI was poor. These results do not support the use of either study device in monitoring CI during open AA surgery.
