A Preliminary Analysis of the Incidence of Transfusion-Related Acute Lung Injury (TRALI) in a Sheep Model Following Transfusion with the Supernatant from Leukodepleted Red Cell Units

Introduction

Despite the introduction of risk reduction strategies, transfusion-related acute lung injury (TRALI) continues to be a significant cause of morbidity and mortality. TRALI development may be associated with the transfusion of anti-leucocyte antibodies or biological response modifiers (BRMs). Our group has previously developed a model in which the development of TRALI in lipopolysaccharide (LPS) treated sheep was precipitated by the transfusion of pooled supernatant from date-of-expiry (day (d) 42) packed red blood cell (PRBC) units. This work pre-dated the introduction of pre-storage leukodepletion of PRBCs in Australia, therefore we investigated whether TRALI would still develop in LPS-treated sheep transfused with supernatant from d42 leukodepleted PRBCs.

Methods

On either d2 (n=75 units) or d42 (n=113 units) of storage, leukodepleted PRBCs underwent dual centrifugation to obtain acellular supernatants. Two supernatant pools (d2 and d42) were then prepared by pooling and heat-treating (56°C for 30min) the supernatants. Levels of potential BRMs in the PRBC supernatant pools were characterised using cytometric bead array and ELISA. Instrumented sheep (n=14) were infused with LPS then transfused (10% v/v) with either d2 (n=7) or d42 (n=7) pooled PRBC supernatant. Two hours later sheep were euthanized and post-mortem lung samples were collected. Physiological data were recorded continuously and then averaged in 30 minute blocks for these preliminary analyses. Blood and broncoalveolar lavage (BAL) fluid samples were collected at specific intervals. Blood samples were used for arterial blood gas analyses, coagulation tests (ROTEM), platelet function tests (MultiPlate) and ELISAs. TRALI was defined by hypoxemia (PaO₂/FiO₂ < 300 on arterial blood gas) and histological evidence of pulmonary edema (by two blinded histopathological assessments of hematoxylin and eosin stained lung sections). Data were compared by group (sheep transfused with d2 PRBC supernatant vs. d42 PRBC supernatant) and by outcome (sheep who developed TRALI vs. those who did not) with either t-tests or 2-way ANOVAs as appropriate.

Results and Discussion

Storage duration of leukodepleted PRBC was associated with increased levels of the potential BRMs 5-HETE, 12-HETE, 15-HETE and IL-8, but not soluble CD40 ligand. Only 3 sheep developed TRALI: one transfused with d2 leukodepleted PRBC supernatant and two transfused with d42 leukodepleted PRBC supernatant corresponding to an incidence of 14% and 29% respectively. This indicated a reduced incidence of TRALI in the sheep model compared to previous data using supernatant from d42 non-leukodepleted PRBC which resulted in an incidence of 75% (Tung et al. Vox Sanguinis. 2011). Preliminary analyses of the physiological data revealed a number of differences. Heart rate (P<0.001), blood pressure (P=0.003), pulmonary artery pressure (P<0.001), cardiac output (P<0.001) and PaO₂/FiO₂ (P<0.001) were different between sheep transfused with either d2 or d42 PRBC supernatant. Heart rate (P<0.001), pulmonary artery pressure (P<0.001), pulse oximeter oxygen saturation (P=0.003), cardiac output (P<0.001) and PaO₂/FiO₂ (P<0.001) were different between sheep that developed TRALI and sheep that did not. However, given the complexity of these data, further analyses using mixed effects modelling are required to better understand these differences. Sheep transfused with d42 PRBC supernatant had reduced IL-1β lung gene expression and reduced ADP-induced platelet aggregation (both total and area under curve) compared to sheep transfused with d2 PRBC supernatant (P=0.047, 0.012 and 0.028 respectively). In addition, sheep that developed TRALI had worse lung histology scores as well as reduced collagen-induced platelet aggregation (both total and area under curve) compared to sheep who did not develop TRALI (P=0.024, 0.049 and 0.018 respectively).

Conclusions

Comparison of these results to previously published data from the sheep model is suggestive that pre-storage leukodepletion of PRBCs is associated with a reduced incidence of TRALI. Unsurprisingly, the development of TRALI was associated with worsened respiratory function (oxygen saturation and PaO₂/FiO₂) and lung histology scores, while changes in hemodynamics and platelet function were also observed.