Abdominal surgery alters the calibration of bioimpedance cardiac output measurement

The performance of impedance cardiography (TEBco), using the BoMed NCCOM3-R7S, and thermodilution (TDco) were compared in eight patients during major abdominal surgery. An opioid, volatile and relaxant anaesthetic technique was employed. This was supplemented with an epidural in five cases. Sets of three cardiac output readings, for both methods, were made at 10-20 min intervals throughout surgery. Data were compared using the Bland and Altman method, regression analysis and a nested model to measure variance components at different stages of surgery. Data from 157 sets of readings are presented. Agreement between the two devices was poor, with a ratio of TDco/TEBco of 115% and limits of agreement of 51-193%. The regression line was TDco=(0.98) x TEBco-0.95 with r=0.60. A more detailed analysis, using nested data, showed good repeatability with coefficients of variation of 5.4% for TDco and 4.8% for TEBco. During surgery shifts in the bias between the two devices occurred, which were related to changes in surgical conditions. Between shifts both devices showed good repeatability over time. Variance components were 0.27 within nested data and 0.082 between bias shifts, with a significantly greater overall component of 1.2 (ANOVA; P=0.0001). Shifts could be explained by deficiencies in the algorithm used to calculate TEBco. Current TEBco technology is too inaccurate for intra-operative use. However, under stable operating conditions TEBco and TDco showed good repeatability.