Determination of cardiac output requires measurement of both heart rate and stroke volume. Techniques for measuring heart rate are widespread, and 1 technique for bedside monitoring of stroke volume is electrical impedance cardiography.
To determine the accuracy and precision of stroke volume measured via impedance cardiography and whether the technique can be used to detect trends.
Eleven healthy research participants (22–52 years old) were examined with simultaneous impedance cardiography and phase-contrast magnetic resonance imaging at rest and during exercise. Bland-Altman analysis with repeated-measures correction was used to compare stroke volumes determined with the 2 methods. The suitability of impedance cardiography for detecting trends in stroke volume was analyzed by using the Critchley radial limits of agreement method.
Phase-contrast magnetic resonance imaging indicated a mean stroke volume of 87 (SD, 16) mL at rest; in 9 volunteers, it changed during exercise (P = .04 to P < .001); in 2 volunteers, it did not (P = .32, P = .06). For the range of stroke-volume measurements (60–122 mL), impedance cardiography yielded underestimates of stroke volumes at the low end (bias, −17 mL) and overestimates at the high end (bias, +17 mL; P < .001). Corresponding 95% limits of agreement were 64 mL, a 73% overestimate or underestimate of stroke volume at rest. Critchley radial limits of agreement indicated poor concordance of stroke-volume trends.
Impedance cardiography had low accuracy and precision in measuring absolute stroke volume and was a poor detector of stroke-volume trends.