Comparison of aortic pulse wave contour analysis and thermodilution methods of measuring cardiac output during anesthesia in the dog

Abstract

Accuracy and reliability of cardiac output (Qt measurement utilizing computer analysis of the aortic pressure wave (Warner method) were assessed by comparing this technique with a standard thermodilution method in 20 halothane-anesthetized dogs over a wide range of Qt and systemic vascular resistance (SVR) values. Five of the dogs also had electromagnetic flow (EMF) probes placed around the ascending aorta. More than 800 pairs of simultaneous Qt determinations were performed using pulse-wave and thermodilution methods, and more than 150 triplicate measurements made using all three techniques during changes in inspired halothane concentrations (0.5-2.0 per cent) or during infusions of sodium nitroprusside (SNP) and phenylephrine (PE). Cardiac outputs ranged from 0.7 to 5.3 l/min, mean aortic blood pressures (BP) ranged from 30 to 200 torr, and SVR varied from 70 per cent below to 200 per cent above control values. Correlation of pulse-wave-computed and thermodilution-calculated Qt values was high (r = .91 or better), irrespective of SVR during changes in halothane concentration. Correlation of pulse-wave and EMF methods was equally good during this period (r = .92 or better). Correlation was lower, but still good, with alterations in SVR of 30 per cent or less induced by SNP or PE (r = .90 or better for thermodilution and r = .91 or better for EMF). When SVR was changed from 30 to 50 per cent of control with SNP or PE, correlation between pulse-wave computed and EMF Qt values remained good (r = .85), but correlation between pulse-wave-computed and thermodilution-calculated Qt values was only fair (r = .79-.77). Correlation of thermodilution-calculated or EMF and pulse-wave-computed Qt values deteriorated significantly with SNP- and PE-induced changes of SVR greater than 50 per cent (r =