OBJECTIVES: To compare the performances of two impedance cardiographs, the RheoCardioMonitor (RCM) and the BoMed NCCOM3, using trend analysis. This involved a series of head-up tilts, a simulation of the stroke volume (SV) and cardiac output (CO) response, calculation of prediction errors and cumulative sums (Cusum). METHODS: Eight healthy male volunteers, aged 27-37 years, were tilted on four occasions to angles of 55 degrees, 15 degrees, 30 degrees and 55 degrees, whilst recording SV and CO every 10-sec. Baseline and percentage changes with tilting were calculated. A simulation of the tilt response was constructed, and from this residuals (observed-predicted) and prediction errors ((observed-predicted)/predicted) x 100% were calculated at 10-sec intervals. Trend analysis was performed by multiple analysis of the variance of serial measurements and graphically assessing changes in serial SV prediction errors, using Cusums. Results are presented as mean (range or SD). RESULTS: Baseline values for RCM-SV were 76 (35-94) ml and for CO 4.7 (2.8-6.1) litre x min(-1). For BoMed-SV they were 113 (90-164) ml and for CO 7.2 (5.5-11.9) litre x min(-1). Head-up tilting to 55 degrees resulted in a 32 (12)% decrease in RCM-SV and a 21 (11)% decrease in BoMed-SV (p < 0.01). Prediction errors for RCM-SV were (6.5 (4.9)%) and for BoMed-SV (6.8 (5.2)%) (p = 0.048). Cusum analysis showed that in 84% of tests, impedance measurements remained within +/- 10% of the initial calibration. There was no difference between devices (chi2 = 0.92). CONCLUSIONS: Simulation of a physiological response, such as that to head-up tilting, and using a trend analysis based on prediction errors and Cusum, is a useful technique. The trending abilities of the RCM and BoMed were similar.
OBJECTIVES: To compare the performances of two impedance cardiographs, the RheoCardioMonitor (RCM) and the BoMed NCCOM3, using trend analysis. This involved a series of head-up tilts, a simulation of the stroke volume (SV) and cardiac output (CO) response, calculation of prediction errors and cumulative sums (Cusum). METHODS: Eight healthy male volunteers, aged 27-37 years, were tilted on four occasions to angles of 55 degrees, 15 degrees, 30 degrees and 55 degrees, whilst recording SV and CO every 10-sec. Baseline and percentage changes with tilting were calculated. A simulation of the tilt response was constructed, and from this residuals (observed-predicted) and prediction errors ((observed-predicted)/predicted) x 100% were calculated at 10-sec intervals. Trend analysis was performed by multiple analysis of the variance of serial measurements and graphically assessing changes in serial SV prediction errors, using Cusums. Results are presented as mean (range or SD). RESULTS: Baseline values for RCM-SV were 76 (35-94) ml and for CO 4.7 (2.8-6.1) litre x min(-1). For BoMed-SV they were 113 (90-164) ml and for CO 7.2 (5.5-11.9) litre x min(-1). Head-up tilting to 55 degrees resulted in a 32 (12)% decrease in RCM-SV and a 21 (11)% decrease in BoMed-SV (p < 0.01). Prediction errors for RCM-SV were (6.5 (4.9)%) and for BoMed-SV (6.8 (5.2)%) (p = 0.048). Cusum analysis showed that in 84% of tests, impedance measurements remained within +/- 10% of the initial calibration. There was no difference between devices (chi2 = 0.92). CONCLUSIONS: Simulation of a physiological response, such as that to head-up tilting, and using a trend analysis based on prediction errors and Cusum, is a useful technique. The trending abilities of the RCM and BoMed were similar.