Abstract
Although restrictive red cell transfusion practice has become a standard of care in the critically ill, data on the use of fresh frozen plasma (FFP) are limited. We hypothesized that the practice of FFP transfusion in the medical intensive care unit is variable and that liberal use may not be associated with improved outcome.
Retrospective cohort study.
A 24-bed medical intensive care unit in a tertiary referral center.
All patients admitted to a medical intensive care unit during a 5-month period who had abnormal coagulation defined as international normalized ratio (INR) of > or = 1.5-times normal.
None.
We collected data on demographics, severity of illness as measured by Acute Physiology and Chronic Health Evaluation (APACHE) III scores, INR, bleeding episodes, and transfusion complications. We identified 115 patients with coagulopathy (INR of > or = 1.5) but without active bleeding. A total of 44 patients (38.3%) received FFP transfusion. INR was corrected in 16 of 44 patients (36%) who received transfusion. Median dose of FFP was 17 mL/kg in patients who had INR corrected vs. 10 mL/kg in those who did not (p = .018). There was no difference in age, sex, APACHE III scores, liver disease, Coumadin treatment, or INR level between those who did and did not receive FFP. Invasive procedures (68.2% vs. 40.8%, p = .004) and history of recent gastrointestinal bleeding (41% vs. 7%, p < .001) were more frequent in the group with transfusion. Although there was no difference in new bleeding episodes (6.8% in transfused vs. 2.8% in nontransfused group, p = .369), new onset acute lung injury was more frequent in the transfused group (18% vs. 4%, p = .021). Adjusted for severity of illness, hospital mortality and intensive care unit length of stay among survivors were not different between the two groups.
The risk-benefit ratio of FFP transfusion in critically ill medical patients with coagulopathy may not be favorable. Randomized controlled trials evaluating restrictive vs. liberal FFP transfusion strategies are warranted.