Blood transfusion and the lung: first do no harm?

Marked variability in transfusion practice exists in cardiac surgical patients, with consumption of approximately 20% of the worldwide allogeneic blood supply. Observational studies have reported an association between red blood cell transfusion and adverse outcome, including pulmonary complications, in cardiac surgery. Tuinman and colleagues report that transfusions were associated with activation of pulmonary inflammation and coagulation by measurement of biomarkers in bronchoalveolar lavage fluid, and suggest that transfusion may be a mediator of acute lung injury. This study provides interesting preliminary data, but is limited by multiple confounding variables (plasma transfusion, use of anticoagulants and heparin antagonists) and the small sample size. A large multicenter, prospective, randomized clinical trial regarding the safety (inclusive of pulmonary complications) and efficacy of red blood cell transfusion in cardiac surgery is needed.

Tuinman and colleagues report the results of a singleinstitution study in cardiac surgery patients (n = 45), docu menting that bronchoalveolar lavage fl uid cytokine and coagulation markers were signifi cantly increased in a dose-dependent manner with transfusions [1]. Th ey conclude that transfusion was associated with activation of pulmonary infl ammation/coagulation and systemic coagu lation derangement.
Th e data are interesting and provocative, and provide a biologic basis for observational studies that reported an association between red blood cell (RBC) transfusions and increased pulmonary complications, including acute lung injury and acute respiratory distress syndrome [2,3]. Translational research studies like this are needed to move this fi eld forward.
Although increased infl ammatory/coagulation markers were identifi ed in the multiple transfusion cohort, no clinically relevant diff erence in pulmonary function (includ ing the PaO 2 /FiO 2 ratio) was identifi ed [1]. Duration of mechanical ventilation was longer in the multiple transfusion cohort, but all were of less than 1 day duration.
A major study limitation is that the three cohorts were not defi ned on the basis of RBC transfusions alone, as the multiple transfusion cohort received plasma/platelets in addition to RBC transfusions. Th is factor signifi cantly confounds the issue, as plasma administration is associated with signifi cantly increased risk for pulmonary complications [4][5][6][7]. In a systematic review/meta-analysis of 37 studies, plasma transfusion was associated with signifi cantly increased acute lung injury risk (odds ratio, 2.92; 95% confi dence interval, 1.99 to 4.29) [8].
Most importantly, the eff ects of anticoagulants, heparin antagonists (for example, protamine) or blood-saving strategies (for example, cell-saver technique) and the degree of intraoperative blood loss and hypoperfusion/ shock were not considered. Additional study limitations include the small sample size, and increased EuroSCORE in the multiple transfusion cohort, and an inability to assess age of blood as a variable.
Th e fundamental question of whether RBC transfusion is safe/eff ective in cardiac surgery is important, and this current study highlights signifi cant concerns particularly with regard to pulmonary complications. Signifi cant variability in transfusion use in cardiac surgery persists, ranging from 7.8 to 92.8% for RBC transfusion [9]. Th e Transfusion Requirements in Critical Care trial excluded cardiac surgical patients and patients who received transfusions before admission to the ICU [10]. Th e recent Transfusion Requirements After Cardiac Surgery trial -a single-center prospective, randomized clinical trial with patients (n = 502) randomized to a liberal strategy (maintain hematocrit ≥30%) or to a restrictive strategy (maintain hematocrit ≥24%) -reported that for each transfused RBC unit, the risk of respiratory complications increased (odds ratio, 1.27; 95% confi dence interval, 1.12 to 1.45; P <0.001) with no diff erence in 30-day all-cause mortality [11].
To resolve this issue regarding transfusion, a large multi center, prospective, randomized clinical trial

Abstract
Marked variability in transfusion practice exists in cardiac surgical patients, with consumption of approximately 20% of the worldwide allogeneic blood supply. Observational studies have reported an association between red blood cell transfusion and adverse outcome, including pulmonary complications, in cardiac surgery. Tuinman and colleagues report that transfusions were associated with activation of pulmonary infl ammation and coagulation by measurement of biomarkers in bronchoalveolar lavage fl uid, and suggest that transfusion may be a mediator of acute lung injury. This study provides interesting preliminary data, but is limited by multiple confounding variables (plasma transfusion, use of anticoagulants and heparin antagonists) and the small sample size. A large multicenter, prospective, randomized clinical trial regarding the safety (inclusive of pulmonary complications) and effi cacy of red blood cell transfusion in cardiac surgery is needed.
regarding the safety (including pulmonary complications) and effi cacy of RBC transfusion in cardiac surgery is needed. Th e National Heart, Lung, and Blood Institute established a State-of-the-Science Symposium on Transfusion Medi cine to identify important clinical trial research issues in this fi eld, and a trial in cardiac surgery was strongly recommended [12].

Abbreviations
FiO 2 , fraction of inspired oxygen; ICU, intensive care unit; PaO 2 , partial pressure of arterial oxygen; RBC, red blood cell.