Oxygen consumption (VO2) can be obtained by the Fick method as the product of cardiac index (CI) by the arteriovenous oxygen content difference (D[a-v]O2) or by indirect calorimetry (IC) based on gas exchange measurements. IC is considered the "gold standard" but is not widely available in clinical practice. Our objective was to test the reliability of the reversed Fick method compared with IC under conditions of hemodynamic changes. For this purpose we chose aortic clamping and unclamping as a model. Twelve patients undergoing abdominal aortic surgery were monitored with a fiberoptic pulmonary artery catheter (Opticath®). Calculated oxygen consumption (cVO) was obtained from hemodynamic and analytic data according to the Fick method at six stages: postinduction, before cross-clamping (Pre-C), postclamping (Post-C), before unclamping (Pre-U), after unclamping (Post-U), and the end of surgery. cVO2 was compared with measured oxygen consumption (mVO2) obtained by IC. CI and mixed venous oxygen saturation (Svc2) varied significantly (P < 0.0001) during the interventions. Although mVO2 did not vary, cVO2 varied significantly (P = 0.0001), reaching a maximum at Post-U coinciding with a sudden decrease in Svo2. There was good concordance between mVO2 and cVO2 except at Post-U (P = 0.005). We conclude that cVO2 is a reliable indicator of VO2 in this hemodynamic model except at Post-U. This lack of correlation seems to be due to a mathematical artifact, because the low value of Svo2 registered at this stage and introduced into the Fick formula reflects a sudden venous return of desaturated blood and not an increase in VO2.