Continuous cardiac output monitoring with an uncalibrated pulse contour method in patients supported with mechanical pulsatile assist device

We evaluated the accuracy of an uncalibrated pulse contour method called Pressure Recording Analytical Method (PRAM) compared with continuous thermodilution for cardiac output (CO) monitoring in patients implanted with a pulsatile left ventricular assist device (LVAD). Twelve adult patients implanted with the HeartMate I-XVE device were studied. CO was simultaneously evaluated by PRAM and by continuous thermodilution. Blood flow values displayed by the LVAD's console were also used for the comparison. Bland-Altman and linear regression analyses were applied. A total of 72 CO measurements (range 3.8-6.7 l/min) were obtained. Mean CO was 5.23±0.70 l/min for the 'hot' pulmonary thermodilution (ThD-CCO) method, 5.28±0.63 l/min for PRAM and 4.83±0.67 l/min for LVAD-CO. A high correlation (r=0.90), a good agreement (mean bias -0.04 l/min, precision ±0.38 l/min) and a low percentage of error (7.3%) were observed between PRAM-CO and ThD-CCO. A good correlation was found between LVAD-CO and either ThD-CCO (r=0.88) or PRAM-CO (r=0.86), but an overestimation of 10% was observed for both PRAM-CO (mean bias -0.44 l/min) and ThD-CCO (mean bias -0.40 l/min). Our results demonstrated good agreements between PRAM-CO, ThD-CCO and LVAD-CO. PRAM derives CO from a peripheral artery without calibration and may be a complementary tool in the hemodynamic assessment of patients supported with a VAD.