The influence of weight on stroke volume determination by means of impedance cardiography in cardiac surgery patients

OBJECTIVES: Obesity is thought to be one of the conditions in which the impedance cardiographic method is less reliable for estimating stroke volume (SV). This led to the introduction of a weight correction factor, sigma, into the equation according to Sramek and Bernstein. However, no scientific evidence has been published to support the use of this factor. The objectives of the present study are to evaluate the influence of body weight on the accuracy of impedance cardiography and to validate Bernstein's weight correction factor by comparison with thermodilution in patients after coronary bypass surgery.

DESIGN: Prospective clinical study.

SETTING: A surgical intensive care unit in a university hospital.

PATIENTS: 37 consecutive patients 24-36 h after coronary bypass surgery, sub-divided into a normal-weight group (n = 24), patients whose weight deviated less than 15% from their ideal weight, and an obese group (n = 13), patients whose weight deviated more than 15% from their ideal weight.

MEASUREMENTS: Kubicek's impedance cardiographic method and Sramek and Bernstein's method to assess SV are applied and compared to thermodilution. In order to study the validity of sigma, the results are compared between 24 patients with normal weight and 13 obese patients.

RESULTS: A significant correlation between miscalculation of SV by impedance cardiography and the degree of obesity for Sramek and Bernstein's method is found when sigma is not included in the equation (r = -0.55, p < 0.05). This relation, however, remained significant when sigma was included in the equation (r = -0.40, p < 0.05). Kubicek's method shows no significant correlation for this relation (r = -0.30). Besides this, Sramek and Bernstein's method underestimates SV significantly in the obese group, independent of the use of sigma in the equation. These results are explained as being intrinsic to the equation, according to Sramek and Bernstein. In the whole group the impedance-derived SV did not significantly differ from SV as measured by means of thermodilution, independent of the method used to calculate SV. However, a considerably better correlation and agreement (mean difference +/- 2 standard deviations is found when Kubicek's method is applied (r = 0.90, 0.5 +/- 17.1 ml vs 0.64, -4.9 +/- 31.8 ml for Sramek and Bernstein's method).

CONCLUSIONS: Weight significantly influences Sramek and Bernstein's method of impedance cardiography, whereas Kubicek's method is not biased by this factor.