Transfusion of fresh frozen plasma in critically ill surgical patients is associated with an increased risk of infection

Abstract

Objective: To determine whether there is an association between transfusion of fresh frozen plasma and infection in critically ill surgical patients.

Design: Retrospective study.

Setting: A 24-bed surgical intensive care unit in a university hospital.

Patients: A total of 380 non-trauma patients who received fresh frozen plasma from 2004 to 2005 were compared with 2,058 nontrauma patients who did not receive fresh frozen plasma.

Interventions: None.

Measurements and main results: We calculated the relative risk of infectious complication for patients receiving and not receiving fresh frozen plasma. T-test allowed comparison of average units of fresh frozen plasma transfused to patients with and without infectious complications to describe a dose-response relationship. We used multivariate logistic regression analysis to evaluate the association between fresh frozen plasma and infectious complication, controlling for the effect of red blood cell transfusion, Acute Physiology and Chronic Health Evaluation II, and patient age. A significant association was found between transfusion of fresh frozen plasma and ventilator-associated pneumonia with shock (relative risk 5.42, 2.73-10.74), ventilator-associated pneumonia without shock (relative risk 1.97, 1.03-3.78), bloodstream infection with shock (relative risk 3.35, 1.69-6.64), and undifferentiated septic shock (relative risk 3.22, 1.84-5.61). The relative risk for transfusion of fresh frozen plasma and all infections was 2.99 (2.28-3.93). The t-test revealed a significant dose-response relationship between fresh frozen plasma and infectious complications (p = .02). Chi-square analysis showed a significant association between infection and transfusion of fresh frozen plasma in patients who did not receive concomitant red blood cell transfusion (p < .01), but this association was not significant in those who did receive red blood cells in addition to fresh frozen plasma. The association between fresh frozen plasma and infectious complications remained significant in the multivariate model, with an odds ratio of infection per unit of fresh frozen plasma transfused equal to 1.039 (1.013-1.067). This odds ratio resembled that noted for each unit of packed red blood cells, 1.074 (1.043-1.106).

Conclusions: Transfusion of fresh frozen plasma is associated with an increased risk of infection in critically ill patients.