We assessed the transfusion trigger for patients admitted to the medical intensive care unit (MICU), the impact of acute coronary syndromes or active bleeding on transfusion threshold, and ICU and hospital outcomes.
We retrospectively reviewed patients seen between July 1, 2000 and Dec 31, 2000. Data included demographics, hemoglobin (HB) values at the time of admission, discharge and transfusion, history of coronary artery disease(CAD), number of units of blood transfused, evidence of active bleeding or acute CAD, APACHE III scores and outcomes.We defined two groups based on clinical characteristics and HB trigger: “appropriate” as the group which had evidence of acute CAD and/or clinically significant bleeding and had a transfusion at HB less than or equal to 10 gm/dl and “inappropriate” as the group which had no evidence of acute CAD or bleeding and had transfusions at HB greater than or equal to 7 gm/dl.
A total of 677 patients were admitted during this period of whom 231 (29.31%) received blood transfusions. The mean age at admission was 66.7±15years; 134 (58.00%) were male and the mean APACHE III score was 68.9±32. Predicted hospital mortality was 27.9% and predicted ICU mortality was 17.5%. The mean HB at admission was 9.3±2.1 and HB at discharge was 10.4±1.4. Mean HB at transfusion was 8.2 ±1.3 and patients received a median of 3 units. There were no differences for the predicted hospital mortality between the appropriate and inappropriate group. Overall mortality was18.6% in the ICU and 26.4% in the hospital.
Despite evidence recommending lower HB transfusion trigger,we observed that nearly 37% of patients received transfusions outside these guidelines. However, when adjusted for severity of illness, inappropriate transfusion was not associated with increased mortality.
Common clinical practice may not match evidence based guidelines for red cell transfusions. Transfusion practices should be monitored and clearly defined indications should be specified.
Patient CharacteristicsAppropriate Mean ±SDInappropriate Mean ±SDFrequency distributionN=146(63.20%)N=85(36.80%)Age(year)68.41 ± 1563.77 ± 16Female- sex40.4%(59/146)44.7%(38/85)APACHE 11168.45 ± 3469.56 ± 28Pred ICU mortality17.5 ± 23%17.4 ± 2% Observed ICU mortality17.1%21.2% p=0.44Pred hospital mortality27.9±28%27.9 ± 25% Observed hospital mortality24.7%29.4% p=0.43Number of units transfused3 (minimum–1 and maximum of 17)4 (minimum –1 and maximum –26) (p<.04)HB transfusion7.8 ± 1.38.7 ± 1.1
J. Chandrasekhar, None.