Using arterial-venous oxygen difference to guide red blood cell transfusion strategy

Background: Guidelines recommend a restrictive red blood cell transfusion strategy based on hemoglobin (Hb) concentrations in critically ill patients. We hypothesized that the arterial-venous oxygen difference (A-V O2diff), a surrogate for the oxygen delivery to consumption ratio, could provide a more personalized approach to identify patients who may benefit from transfusion. Methods: A prospective observational study including 177 non-bleeding adult patients with a Hb concentration of 7.0-10.0 g/dL within 72 h after ICU admission. The A-V O2diff, central venous oxygen saturation (ScvO2), and oxygen extraction ratio (O2ER) were noted when a patient's Hb was first within this range. Transfusion decisions were made by the treating physician according to institutional policy. We used the median A-V O2diff value in the study cohort (3.7 mL) to classify the transfusion strategy in each patient as "appropriate" (patient transfused when the A-V O2diff > 3.7 mL or not transfused when the A-V O2diff ≤ 3.7 mL) or "inappropriate" (patient transfused when the A-V O2diff ≤ 3.7 mL or not transfused when the A-V O2diff > 3.7 mL). The primary outcome was 90-day mortality. Results: Patients managed with an "appropriate" strategy had lower mortality rates (23/96 [24%] vs. 36/81 [44%]; p = 0.004), and an "appropriate" strategy was independently associated with reduced mortality (hazard ratio [HR] 0.51 [95% CI 0.30-0.89], p = 0.01). There was a trend to less acute kidney injury with the "appropriate" than with the "inappropriate" strategy (13% vs. 26%, p = 0.06), and the Sequential Organ Failure Assessment (SOFA) score decreased more rapidly (p = 0.01). The A-V O2diff, but not the ScvO2, predicted 90-day mortality in transfused (AUROC = 0.656) and non-transfused (AUROC = 0.630) patients with moderate accuracy. Using the ROC curve analysis, the best A-V O2diff cutoffs for predicting mortality were 3.6 mL in transfused and 3.5 mL in non-transfused patients. Conclusions: In anemic, non-bleeding critically ill patients, transfusion may be associated with lower 90-day mortality and morbidity in patients with higher A-V O2diff.