Objectives: Impedance cardiography has been used to detect early hemorrhagic shock, but, to the best of the authors' knowledge, no previous study has looked at the test characteristics of bioimpedance-derived hemodynamic parameters for the detection of a measured loss of blood volume. The authors set out to investigate the effect of controlled hemorrhage on stroke index (SI) using impedance cardiography, and to evaluate the performance of this test for moderate blood loss in comparison with the performance of traditional vital signs.
Methods: Supine and standing measurements of SI were made using a thoracic electrical bioimpedance monitor in 197 healthy, adult volunteers before and after donation of a standard unit (500 mL) of whole blood. Pre- and post-donation means for supine SI, heart rate (HR), systolic blood pressure (SBP), and diastolic blood pressure (DBP), and for the postural changes in each of these measurements ( triangle up SI, triangle up HR, triangle up SBP, triangle up DBP), were compared using a paired-sample, two-tailed t-test. For each term, receiver operating characteristic (ROC) curves were generated, treating pre- and post-donation populations as independent. Areas under the ROC curves were compared using a method for correlated test result data.
Results: Mean +/- SD pre- and post-donation values for SI were 47.0 +/- 6.9 mL/m(2) and 43.9 +/- 7.3 mL/m(2), respectively (p < 0.0001). Mean pre- and post-donation values for triangle up SI were -7.7 +/- 5.8 mL/m(2) and -9.0 +/- 8.2 mL/m(2), respectively (p = 0.02). For discriminating between pre- and post-donation status, the areas under the ROC curves for SI and triangle up SI were 0.62 and 0.58, respectively. In comparison, the areas under the ROC curves for HR, SBP, DBP, triangle up HR, triangle up SBP, and triangle up DBP were 0.49, 0.61, 0.56, 0.82, 0.53, and 0.50, respectively. The area under the ROC curve for triangle up HR was significantly greater than the area under any of the other curves (p < 0.0001).
Conclusions: Moderate blood loss was associated with a decline in mean SI that was clearly detectable with impedance cardiography. However, as a test for moderate blood loss in this controlled setting, neither SI nor triangle up SI performed better than traditional vital signs.