Uncalibrated arterial pulse contour analysis versus continuous thermodilution technique: effects of alterations in arterial waveform.

Objective: To compare an arterial pressure-derived cardiac output (APCO) (Vigileo software version 1.07; Edwards Lifesciences, Irvine, CA) and a thermodilution cardiac output (CCO) as methods for measuring cardiac output under different pathologic and experimental conditions that induce changes in arterial waveform morphology.
Design: A prospective study.
Setting: A university hospital, single institutional.
Participants: Fifty-two patients undergoing elective cardiac surgery.
Interventions: Simultaneous APCO and CCO were compared in low-risk patients undergoing elective coronary artery surgery (without valvular disease) (control, n = 20), patients with aortic stenosis (AS, n = 10), aortic insufficiency (AI, n = 10), and intra-aortic balloon pump (IABP, n = 12). In the control group, additional data were registered before and after median sternotomy and phenylephrine administration.
Measurements and Main Results: In the control group, Bland-Altman showed a bias of -3% (95% limits of agreement: -59% to +53%) before cardiopulmonary bypass (CPB) and of -1% (95% limits of agreement: -51% to +50%) after CPB. In the AS group, the bias was -5% (95% limits of agreement: -34% to +24%) before CPB and 1% (95% limits of agreement: -28 to +30%) after CPB. In the AI group bias was +32% (95% limits of agreement: -4% to +68%) before CPB and -2% (95% limits of agreement: -35% to +32%) after CPB. Median sternotomy decreased CCO by 10% +/- 10%, whereas it increased APCO by 56% +/- 28%. Phenylephrine administration decreased CCO by 11% +/- 16%, whereas it increased APCO by 55% +/- 34%.
Conclusions: Cardiac output measurement based on uncalibrated pulse contour analysis is able to reflect cardiac output measured with the continuous thermodilution method in patients undergoing uncomplicated coronary artery surgery. However, in situations in which the arterial pressure waveform is changed, agreement between techniques may be altered and data obtained with uncalibrated pulse contour analysis may become less reliable.