Effect of red blood cell transfusion on inflammation, endothelial cell activation and coagulation in the critically ill

Abstract Background and Objectives Red blood cell (RBC) transfusion is a frequently applied intervention in an intensive care unit. However, transfusion is associated with adverse outcomes including organ failure and thrombo‐embolic events. Mechanisms of these effects are not known but may be related to activation of the endothelium or of the coagulation or inflammatory system. We hypothesized that a RBC transfusion in the critically ill would result in further activation of these systems. Materials and Methods In 74 non‐bleeding critically ill patients receiving one RBC unit, markers of inflammation, endothelial cell activation and coagulation were measured before transfusion, at 1 h after transfusion and 24 h after transfusion. The impact of disease severity of the recipient on these changes was assessed by comparing septic and non‐septic patients (according to sepsis‐3 definition) and by correlation of biomarkers with the sequential organ failure assessment (SOFA) score. Results Levels of von Willebrand Factor (vWF), soluble ICAM‐1, soluble thrombomodulin, fibrinogen and d‐dimer were already high at baseline, whereas ADAMTS13 levels were low. VWF levels increased significantly 24 h after RBC transfusion (median 478% (338–597) vs. 526% (395–623), p = 0.009). The other biomarkers did not change significantly. Post transfusion change was not dependent on the presence of sepsis and was not correlated with SOFA score. Conclusion RBC transfusion in critically ill patients was associated with an increase in circulating vWF levels, suggesting a further increase in activation of the endothelium, a finding that was independent of the presence of sepsis or organ injury level.


INTRODUCTION
More than 1 out of four patients in an intensive care unit (ICU) receive a red blood cell (RBC) transfusion during their admission, rendering RBC transfusion one of the most frequently applied interventions at the ICU [1]. However, RBC transfusion is associated with adverse outcomes including organ failure and thrombo-embolic events, in particular in the critically ill [2][3][4][5]. The mechanisms responsible for these adverse events are not fully understood.
Critically ill patients are often in an inflammatory state in which the endothelium and coagulation system are already activated before transfusion [6]. The presence of an inflammatory state in the recipient has been shown to be a risk factor for the development of transfusion related acute lung injury (TRALI) [5]. Also, in patients with an inflammatory state that do not develop full-blown TRALI, an extra 'hit' by the RBC transfusion may exacerbate inflammation, potentially resulting in transfusion related adverse events [7][8][9][10]. Since the vascular endothelium and glycocalyx are among the first that interact with the donor RBCs after transfusion, these structures might also play a role in the pathophysiology. RBC transfusion is associated with increased biomarker levels of endothelial cell activation in haematological and paediatric patients [11,12]. Endothelial cell activation can lead to increased endothelial permeability with neutrophil extravasation and capillary leakage, resulting in organ injury [13]. Activation of the endothelium also activates the coagulation system, resulting in (micro)thrombus formation [14,15]. Therefore, the vascular endothelium and activation of the coagulation system might be involved in the adverse events of RBC transfusion.
The aim of this study is to investigate the effect of RBC transfusion on several biomarkers of inflammation, endothelial cell activation and coagulation in adult critically ill patients (Table 1). We hypothesized that a RBC transfusion in the critically ill results in further activation of the vascular endothelium and also in activation of the coagulation system. Since sepsis is a risk factor for developing transfusion-associated adverse events, we expected a greater effect in septic patient compared to non-septic patients [5].  Human ADAMTS13 (a disintegrin and metalloproteinase with a thrombospondin type 1 motif, member 13) antigen levels were determined using a monoclonal antibody-based human ADAMTS13 antigen ELISA, as previously described [17][18][19]. Microtiter plates were coated with the monoclonal mouse anti-human ADAMTS13 antibody 3H9 [17,18,20]

RESULTS
Seventy-four patients (55% male) were included with an median age of 63 years old (IQR 57-73). In four patients, 24 h timepoint was missed because indwelling arterial catheter was removed. Patients had a SOFA score of 8.5 (IQR 7-11). Most patients came from surgical departments (62%). Forty-one patients (55%) fulfilled the sepsis criteria. Baseline characteristics are given in Table 2. Median haemoglobin level at inclusion was low (6.8 g/dl) and increased after transfusion (7.9 g/dl, p < 0.001). Septic patients were more often female and had a longer ICU admission duration at time of inclusion compared to non-septic patients. and soluble TM were high at baseline but did not change significantly after transfusion. Levels of soluble syndecan-1 were low and did not increase significantly after transfusion. Concentration of IL-6 and TNF-alpha did also not significantly change (Table 3).

RBC transfusion resulted in an increase in vWF antigen, but not in other markers
D-dimer and fibrinogen levels were elevated at baseline, but did not increase after transfusion. Platelet count, PT and APTT were in between reference values and also did not significantly change after transfusion ( Table 4).
The inflammatory state of the recipient did not impact the effect of a RBC transfusion So far, only one study has studied the effect of RBC transfusion on vWF antigen levels. In cardiology patients, an increase in vWF antigen levels was not observed directly following RBC transfusion [23].
However, as vWF release takes some time following a stimulus this time point may have been too short after transfusion [24]. We think it is unlikely the increase of vWF is caused by vWF that was present in RBC units, since increase was not observed immediately after transfusion but after 24 h. Most likely, RBC transfusion led to shedding of vWF from endothelial cells.
vWF antigen plays an important role in arterial and venous thrombus formation [15,25,26]. The increase in circulating vWF antigen levels after RBC transfusion can therefore potentially explain the increase in thrombo-embolic events found after transfusion [4,27].
However, we did not find an effect on markers of disseminated intravascular coagulation (DIC), such as a decreased platelet count, which is in line with earlier research [28]. Furthermore, the vWF/ADAMTS13 ratio did not increase. Therefore, the impact of 1 RBC unit on the development of thrombosis seems minimal. However, patients often receive multiple transfusions over time during their ICU stay. Whether multiple transfusions result in a more pronounced increase of vWF, followed by changes in DIC markers or the vWF/ADAMTS13 ratio should be investigated in a future study.
The occurrence of thrombosis in critically ill patients is associated with worse outcomes [29], therefore more knowledge on the mechanism behind the association between RBC transfusion and thrombosis can be of great importance. By unravelling the mechanism responsible for transfusion related adverse events we may identify the responsible compounds in the RBC unit [30], enabling improvement of the transfusion product or protocols.  SOFA score Delta Fibrinogen (mg/mL) (e) 9 1   F I G U R E 1 (a-h) correlation plots between post transfusion change in biomarker level and sequential organ failure assessment (SOFA) score. Rho = spearman rank correlation. TM = thrombomodulin. vWF = von Willebrand factor