Revisiting transfusion practices in critically ill patients

Objective: The objective of this study was to characterize contemporary red cell transfusion practice in the critically ill and to examine changes in practice over time. Study Design: The authors conducted a scenario-based national survey. Study Population: This study consisted of Canadian critical care practitioners. Outcome Measure: The authors evaluated transfusion thresholds in four hypothetical scenarios. Results: Of 343 eligible Canadian critical care physicians, 235 (68.5%) responded to the survey. Most respondents were general internists (57%) who had been in practice for an average of 11.1 (±7.1) yrs and worked most often in combined medical/surgical intensive care units. Transfusion thresholds differed significantly among the four scenarios (p < .0001). The proportion of respondents adopting a threshold of 70 g/L was 63% and 70% in the hypothetical scenarios of trauma and septic shock compared with 16% and 3% who adopted the same threshold for scenarios involving patients with stable gastrointestinal hemorrhage and postoperative myocardial infarction, respectively. Fifteen percent of respondents identified transfusion thresholds exceeding 100 g/L for the postoperative myocardial infarction scenario, and 7% identified this threshold for the gastrointestinal hemorrhage scenario. Only 0.4% of respondents adopted a 100-g/L threshold for the two remaining scenarios. There was a significant decrease in transfusion thresholds in all four scenarios (p < .001) since the administration of a previous survey in 1993. The reported use of single-unit transfusions was 56% in 2002 vs. 10% in 1993. Eighty-five percent of physicians stated that they had modified their approach to transfusion, primarily in response to the publication of a major Canadian clinical trial and institutional guidelines. Conclusions: Canadian physicians appear to have adopted lower transfusion triggers and an increase in the use of single-unit red cell transfusion.