BACKGROUND: Currently, the monitoring of cardiac output (CO) and stroke volume (SV) is mainly performed using invasive techniques. Therefore, performing CO monitoring noninvasively by means of electrical impedance tomography (EIT) would be advantageous for intensive care. Our hypothesis was that, by means of EIT, it is possible to assess heart rate (HR) and to quantify changes in SV due to changes in ventilator settings.
METHODS: CO (HR and SV) of 14 pigs (32–40 kg body weight) was changed by incremental increases in positive end-expiratory pressure levels (0, 5, 10, 15, and 20 cm·H2O; ramp maneuver). This ramp maneuver was applied 4 times in each animal, yielding 43 evaluable single experiments. At each positive end-expiratory pressure level, SV was assessed by transpulmonary thermodilution using a PiCCO device. EIT data were acquired using a Dräger EIT Evaluation Kit 2.
RESULTS: The EIT-based SV-related signal, ZSV (in [AU]), showed only a weak correlation (after excluding 2 measurements) with SVTTD of r = 0.58 (95% confidence interval, 0.43–0.71). If ZSV is calibrated by the reference 1 time for each experiment (defined as SVEIT), the correlation is approximately 0.85 (95% confidence interval, 0.78–0.90). A possible reason for the moderate correlation is the unexpected scaling pattern, leading to amplification of the cardiac impedance signal, found in some animals. The scaling is probably due to the imperfect reconstruction (i.e., a change of sensitivity) of the EIT images or to a change in the position of the heart.
CONCLUSIONS: The hypothesis that EIT can be used to monitor CO and SV was confirmed, but further studies are required before this technique can be applied in clinical practice. HR was determined robustly and accurately. For SV monitoring, promising results were obtained in 80% of the experiments. However, unexpected scaling of the cardiac EIT signal causing inaccurate estimation of SV remains an issue. Before robust assessment of SV by EIT is suitable for clinical practice, the cause of and compensation for undesired scaling effects need to be investigated.