Study objectives: During isovolemic hemodilution, healthy individuals maintain oxygen consumption ((V) over dot o(2)) by identical increases in cardiac index (CI) and oxygen extraction ratio (O2ER). In critically ill patients, the relationship between CI and O2ER may be different. Patients with an altered cardiac function may have a decreased CI/O2ER ratio, whereas patients with sepsis may have an increased CI/O2ER ratio. We hypothesized that the analysis of the CI-O2ER relationship could help us to assess the adequacy of cardiac function in critically ill patients with anemia, Design: Prospective, observational study, Setting: Thirty-one-bed medicosurgical ICU of a university hospital. Patients: Sixty patients equipped with arterial and Swan-Ganz catheters presenting with anemia, which was defined as a hemoglobin level less than or equal to 10 g/dL in the absence of active bleeding. Patients were classified into those with compromised cardiac function (group 1; n = 40), and those with normal cardiac function (group 2; n = 20). Measurements and results: In addition to the pertinent clinical data, initial hemodynamic measurements, including pulmonary artery occlusion pressure (PAOP), CI, and O2ER, were collected in all patients at the onset of anemia. As anticipated, group 1 patients (n = 40) had lower CIs, higher O2ER levels, and lower CI/O2ER ratios than group 2 patients. However, there was no significant difference in PAOP values between the groups. The CI/O2ER ratio was < 10 in 27 of 40 group 1 patients but only in 4 of 20 group 2 patients. Of these latter four patients, three were found to be hypovolemic, and one patient with sepsis had severe myocardial depression. There was no statistically significant difference in PAOP in group 2 patients;vith or without hypovolemia ([mean +/- SD] 12.3 +/- 2.1 mm Hg) vs 13.7 +/- 4.3 mm Hg; p = 0.21). In group 1, survivors had a higher CI and CI/O2ER ratio than nonsurvivors. In group 2, however, such a relationship did not reach statistical significance. Conclusions: The relationship between CI and O2ER level can help interpret the CI in anemic patients. In anemic patients with no cardiac history, a low CI/O2ER ratio (< 10) suggests hypovolemia even when CI is not depressed.