Dr Annemarie Docherty, University of Edinburgh, Edinburgh, UK.
What is known?
Death from haemorrhage is the second most common cause of death in the trauma population and a high proportion of severely injured patients receive red blood cell transfusions. Evidence from randomised controlled trials in critically ill patients support a restrictive transfusion threshold, however the effect of transfusion on outcomes in trauma may differ due to the timing and amount of transfusion required. Trauma patients may be unstable or actively bleeding, as opposed to the slow decline in haemoglobin often seen in critical care. Evidence in the trauma population is based primarily on small observational studies.
What did this paper set out to examine?
This systematic review and meta-analysis set out to assess the association between red blood cell transfusion and mortality in the trauma population. Secondary outcomes included acute respiratory distress syndrome or acute lung injury (ARDS/ALI) and multiorgan failure. Comparative observational and interventional studies were eligible for inclusion.
What did they show?
The authors included 40 studies in the qualitative review. No randomised controlled trials addressed the study question. All studies were observational cohort studies, which increased the risk of selection bias and confounding. Particularly relevant confounders were injury severity and other measures of shock which were strongly associated with the study outcomes. The authors assessed the quality of the studies using the Newcastle-Ottawa Scale, and the quality of the meta-analysis using the GRADE guidelines.
There was significant heterogeneity. Study size varied from 29 to 25,299. Timing of red blood cell transfusion varied considerably from studies that included transfusion within 24-48 hours only, total in-hospital transfusion, to studies that excluded patients transfused within 48 hours of admission. There were also marked differences in the categorisation of red blood cell transfusion: continuous variable (per unit change), binary variable (transfused/not transfused) and a categorical variable. In addition to this, patient populations also varied: multiply injured patients, patients with only one system injured, massively transfused patients, patients only admitted to the intensive care unit, surgical patients only, intubated patients only, and various injury severity score cutoffs.
Seventeen studies attempted to determine the effect of transfusion on mortality after adjusting for important confounders, and nine of these had enough information to be pooled in the meta-analysis. Eight studies found that red blood cell transfusion was associated with increased odds of mortality, and the adjusted pooled analysis showed an increase in the odds of mortality with each additional unit transfused (OR 1.07, 95%CI 1.04-1.10, p<0.001, I^2=94.6%). The authors graded this evidence as low.
Six studies attempted to determine the adjusted association with multiorgan failure. The odds of multiorgan failure increased with each additional unit of blood (OR 1.08, 95%CI 1.02-1.14, p=0.012, I^2=95.9%). The grade of evidence was moderate.
Six studies assessed the adjusted association between transfusion and ARDS, but only two had enough information to be included in the meta-analysis (transfused vs not transfused: OR 2.04, 95%CI 1.47-2.83, p<0.001, I^2=0%). The authors graded this evidence as very low.
What are the implications for practice and for future work?
The observational studies all showed an association between transfusion and mortality and other negative outcomes. However, there was considerable heterogeneity between the studies, and as the authors acknowledge, it is likely that significant confounding persisted even after attempts to adjust for injury and illness severity. The authors have graded the evidence as very low to moderate, and it is not possible to refine red blood cell transfusion practice in trauma on the basis of these observational studies.
This systematic review highlights the lack of evidence for red blood cell transfusion in trauma, and the need for a robust randomised controlled trial in this population. This would minimise confounding and bias, and give a definitive answer regarding the effect of red blood cell transfusion on mortality.