Objective: To determine whether transfusion of red cells either ≤5 days or ≥20 days from donation alters tonometric indexes of gastric mucosal oxygenation or global oxygenation parameters in euvolemic anemic critically ill patients without ongoing hemorrhage. The a priori hypothesis was that stored red cells worsen gastric oxygenation.
Design: Prospective, double-blind, randomized study.
Setting: A 12-bed general medical/surgical intensive care unit in a Scottish teaching hospital.
Patients: Ventilated euvolemic anemic (mean ± sd hemoglobin, 85.8 ± 8.4 g/L) critically ill patients with significant organ failure, but no evidence of hemorrhage.
Interventions: After baseline measurements, patients were randomized to receive two units of leukodepleted red cells that were either ≤5 days (ten patients) or ≥20 days (12 patients) after donation according to a standardized protocol.
Measurements and Main Results: Changes in gastric to arterial Pco2 gap (Pg-Paco2 gap), gastric intramucosal pH, arterial pH, arterial base excess, and arterial lactate concentrations were measured during baseline (2.5 hrs), during transfusion (3 hrs), and for 5 hrs after transfusion. Mean age of red cells stored ≤5 days was 2 days (first and third quartile, 2, 2.25; range, 2-3); red cells stored ≥20 days had a mean age of 28 days (first and third quartile, 27, 31; range, 22-32). Hemoglobin concentration increased by 15.0 g/L and 16.6 g/L, respectively, in the fresh and stored groups (p = .62). There were no significant differences between the groups either using treatment-by-time analysis or comparing the pre- and posttransfusion periods either for Pg-Paco2 gap (mean difference, 0.03 kPa; 95% confidence limits, -1.66, 1.72) or gastric intramucosal pH (mean difference, 0.015 pH units; 95% confidence limits, -0.054, 0.084). The mean change within each group from the pre- to posttransfusion period for Pg-Paco2 gap and gastric intramucosal pH, respectively, was 0.56 kPa (95% confidence limits, -0.68, 1.79) and -0.018 pH units (95% confidence limits, -0.069, 0.032) for fresh red cells and 0.52 kPa (95% confidence limits, -0.6, 1.64) and -0.033 pH units (95% confidence limits, -0.080, 0.129) for stored red cells. There was no statistically or clinically significant improvement in any other oxygenation index during the measurement period for either group compared to baseline values.
Conclusions: Transfusion of stored leukodepleted red cells to euvolemic, anemic, critically ill patients has no clinically significant adverse effects on gastric tonometry or global indexes of tissue oxygenation. These findings do not support the use of fresh red cells in critically ill patients.
Our understanding of the best way to optimize oxygen delivery in the critically ill has changed significantly in recent years. The traditional view that supranormal levels of oxygen delivery should be maintained using inotropic drugs and red cell transfusions has been replaced by randomized, controlled trial evidence that supranormal goal-directed therapy does not improve outcome in patients with established critical illness (1) and might have adverse effects (2). Another randomized, controlled trial (the Transfusion Requirements in Critical Care study) comparing a liberal transfusion strategy (hemoglobin transfusion trigger >100 g/L) with a restrictive strategy (hemoglobin transfusion trigger 70, maintaining level at 70-90 g/L) further supported this finding by showing a trend toward improved outcome in the restrictive group, particularly in patients who were younger and less severely ill (3). This study raised the possibility that anemia is beneficial during critical illness and/or that transfusion of stored red cells has adverse effects (4). In contrast, a recent randomized, controlled trial of early goal-directed therapy in patients with early sepsis found improvements in outcomes, including mortality, when a protocol that included liberal transfusion practice to keep hematocrit ≥30% was used (5). Recent large cohort studies in European (6), Australian (7), U.S. (8), Scottish (9), and London (UK) (10) intensive care units (ICUs) indicate that mean hemoglobin transfusion thresholds are typically 85 g/L since publication of these studies.
If red cell transfusion has adverse effects, these might be due to transfusion of donor leukocytes or because of depletion of red cell 2, 3 diphosphoglyceric acid (DPG) and ATP and changes to the red cell membrane that reduce cell deformability (11). This storage lesion could impair oxygen delivery to tissues by reducing both capillary flow and oxygen unloading from hemoglobin. There are remarkably few data in humans regarding the importance of the red cell storage lesion, but many clinicians believe that fresh red cells are more likely than stored cells to benefit patients (12, 13). One study supporting this view is an oxygen kinetics study that retrospectively found a correlation between the transfusion of red cells stored for >15 days and a deterioration in gastric intramucosal pH (pHi), an index of gastric oxygenation status (14). This study is frequently cited as evidence for a detrimental effect of stored red cell transfusion, despite controversy surrounding the clinical relevance of tonometry-derived oxygenation indexes. Another study in oxygen supply-dependent rats found that red cells stored for 28 days, in contrast to fresh red cells, did not reverse oxygen supply dependency (15), although this may have been because of low rat red cell survival after storage (16). There are no prospective, randomized studies in humans that compare the effect of fresh and stored red cells on indexes of tissue hypoxia in critically ill humans. This is, in part, because a protocol randomizing patients to receive fresh or stored red cells is complex to organize and requires close collaboration between clinicians and the blood bank.
Many blood transfusion services have recently added universal leukodepletion to the routine processing of all blood components. The previously cited studies all used nonleukodepleted red cells or whole blood. These may have different effects on patients from the current product that is leukofiltered, plasma depleted, and resuspended in additive solution. There are no data regarding the effect of transfusing leukodepleted red cells on indexes of tissue hypoxia in critically ill patients. We therefore carried out an exploratory double-blind, prospective, randomized study to determine whether the age of transfused leukodepleted red cells influenced indexes of regional (gastric mucosal) oxygenation or clinically relevant global indexes of tissue hypoxia.