Assessment of stroke volumeindex with three different bioimpedance algorithms: lack of agreement compared to thermodilution

OBJECTIVE: The accuracy of bioimpedance stroke volumeindex (SVI) is questionable as studies report inconsistent results. It remains unclear whether the algorithms alone are responsible for these findings. We analyzed the raw impedance data with three algorithms and compared bioimpedance SVI to transpulmonary thermodilution (SVI(TD)). DESIGN AND SETTING: Prospective observational clinical study in a university hospital. PATIENTS: Twenty adult patients scheduled for coronary artery bypass grafting (CABG). INTERVENTIONS: SVI(TD) and bioimpedance parameters were simultaneously obtained before surgery (t(1)), after bypass (t(2)), after sternal closure (t(3)), at the intensive care unit (t(4)), at normothermia (t(5)), after extubation (t(6)) and before discharge (t(7)). Bioimpedance data were analyzed off-line using cylinder (Kubicek: SVI(K); Wang: SVI(W)) and truncated cone based algorithms (Sramek–Bernstein: SVI(SB)). MEASUREMENTS AND RESULTS: Bias and precision between the SVI(TD) and SVI(K), SVI(SB), and SVI(W) was 1.0 ± 10.8, 9.8 ± 11.4, and −15.7 ± 8.2 ml/m(2) respectively, while the mean error was abundantly above 30%. Analysis of data per time moment resulted in a mean error above 30%, except for SVI(W) at t(2) (28%). CONCLUSIONS: Estimation of SVI by cylinder or truncated cone based algorithms is not reliable for clinical decision making in patients undergoing CABG surgery. A more robust approach for estimating bioimpedance based SVI may exclude inconsistencies in the underlying algorithms in existing thoracic bioimpedance cardiography devices. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (doi:10.1007/s00134-007-0938-y) contains supplementary material, which is available to authorized users.