Background and Aim

Studies have reported association between red cell (RBC) transfusion (tx) and increased mortality and morbidity in cardiac surgery (CS). Conflicting data exist on whether universal RBC leucodepletion (ULD) reduces mortality and morbidity, particularly infection and intensive care unit (ICU) length of stay (LOS), in CS patients. RBC tx has also been independently associated with renal failure (RF) in CS, however mechanisms for this effect are unclear. An inflammatory basis has been postulated. No large studies have explored the potential benefit of ULD in reducing RF. We investigated whether introduction of ULD in Australia (July 2008) was associated with reduced in-hospital mortality, infection, new RF (NRF) and ICU LOS in a large cohort of CS patients.


All consecutive CS patients 18y and older at 6 hospitals between 2005 and 2010 were included. Clinical and outcome data were obtained from the Australian and New Zealand Society for Cardiac and Thoracic Surgeons Cardiac Surgery Database, which prospectively collects data on all CS patients. Laboratory results (pre-operative hemoglobin, platelet, coagulation profile and creatinine) and transfusion data were obtained from hospital laboratory information systems (LIS).

Statistical analysis

Tx was defined as ≥ 1 RBC within 48h of CS. Patients were categorized as having surgery pre- or post-ULD and as having received either exclusively LD, non-LD or mixed RBC. Logistic regression modeled the association between LD and mortality, infection and NRF. The association with LD was analyzed in two ways: whether surgery occurred pre- or post-ULD, and whether patients received LD or non-LD RBCs. To examine for trends over time, an equal number of non-tx patients were selected as ‘controls’, based on a propensity score for RBC tx. The relationship between LD and ICU LOS among survivors was explored using linear regression.


16,253 patients underwent CS in the study period. LIS data were available for 14,980 (92%), and of these, 8857 (59%) had surgery pre-ULD. RBC tx was given in 3799 (43%) pre-ULD, and 2525 (41%) post-ULD. Of the 6324 patients who were tx, 2794 (44%) received LD RBC, 2702 (43%) non-LD RBC and 828 (13%) received mixed LD and non-LD RBCs.

In tx patients, there was no significant difference in mortality (136 [5.4%] vs. 240 [6.3%], p=0.125) or infection (310 [11.9%] vs. 510 [13.4%], p=0.080) post-ULD compared with pre-ULD, however there was a significant reduction in NRF (215 [8.5%] vs. 398 [10.5%], p=0.010). After adjusting for hospital, age, sex, co-morbidities, ejection fraction, pre-operative laboratory parameters, surgical history, type and urgency of procedure, pre-operative shock, medications, perfusion time, drain output in first 4 hours and number of RBC, there was no difference in mortality (OR 0.92, 95% CI 0.71-1.20, p=0.534) or infection (OR 0.96, 95% CI 0.81-1.15, p=0.671), however there was a difference in NRF post-ULD compared with pre-ULD (OR 0.77, 95% CI 0.64-0.94, p=0.011). When the analysis was repeated comparing those who received LD vs. non-LD RBC the results were similar, with no difference in mortality or infection, but a reduction in NRF (OR 0.80, 95% CI 0.65-0.98, p=0.035). In the non-tx controls, there was no significant difference in mortality (30 [1.2%] vs. 39 [1.0%], p=0.546), infection (139 [5.5%] vs. 231 [6.1%], p=0.338) or NRF (110 (4.4%] vs. 181 [4.8%], p=0.445) post-ULD compared with pre-ULD. After adjusting as above, there was no significant difference in mortality (OR 1.18, 95% CI 0.72-1.9, p=0.512), infection (OR 0.82, 95% CI 0.66-1.03, p=0.094) or NRF (OR 0.83, 95%CI 0.64-1.07, p=0.148) post-ULD compared with pre-ULD in non-tx pts. In the unadjusted linear regression, post-LD was associated with a small reduction in ICU LOS in transfused (41 vs. 44 h, p=0.034) and with a small increase in non-transfused (29 vs. 31 h, p<0.001). After adjustment there was no reduction in ICU LOS post-ULD.


Introduction of ULD was associated with a reduction in NRF but not mortality, infection or ICU LOS in transfused CS pts. These findings support the potential role of inflammation as a contributor to acute RF in tx CS pts. ULD may be worth exploring prospectively as a possible strategy to reduce the incidence of NRF in CS, given its substantial morbidity and mortality.


Phillips:CSL Behring: Research Funding. Wood:CSL Behring: Research Funding.

Author notes


Asterisk with author names denotes non-ASH members.