EJRC Article Review

Effect of Red Cell Storage Duration on Mortality in Critically Ill Patients

The standard transfusion laboratory policy of issuing the oldest blood for transfusion first in order to protect available supplies has raised the question of whether this exposes sick patients to potential harm. Despite being stored refrigerated in preservative solutions licensed for storage up to 49 days in some countries, red cells develop storage lesions due to a combination of oxidative, metabolic and mechanical stress, which affect the deformability and integrity of the cells and cause haemolysis. In animal models, the increased free plasma oxyhaemoglobin triggers NO scavenging and pulmonary vasoconstriction with direct and indirect cardiorespiratory compromise resulting¹. Furthermore, longer storage times are associated with increased risk of bacterial contamination, particularly of slow-growing organisms². Conversely, damage associated molecular pathogens are derived from mitochondria within the first 2 weeks of storage and are implicated in the development of transfusion-associated acute lung injury³. Additionally, utilising the oldest stock first better manages this precious resource.

A number of clinical studies have sought to evaluate the effect of storage duration on efficacy of red cell transfusion. The limited evidence specific to the intensive care population has thus far demonstrated no significant differences. The ABLE trial was relatively small and has been criticised for the time between randomisation and first transfusion⁴, and the intensive care subgroup of the INFORM trial had a low mortality rate suggesting the group is not representative of the general ICU cohort⁵. This coupled with the theory that critically ill patients may be more susceptible to changes in red cells associated with storage led to the Standard Issue Transfusion versus Fresher Red-Cell Use in Intensive Care (TRANSFUSE) trial⁶.

The authors of the TRANSFUSE study hypothesised that transfusion with fresher red cells would lead to lower 90-day mortality rates in critically ill patients. 4994 patients over the age of 18 admitted to intensive care units predominantly in Australia and New Zealand were recruited between 2012 and 2016. Exclusion criteria included prior transfusion, cardiac surgery, haematological malignancy, organ transplantation, pregnancy or expected death within 24 hours. The mean patient age in the short-term storage group was 62.5 years compared to 61.4 years in the long-term storage group, but otherwise the baseline characteristics of the two arms were well matched with a range of diagnoses across all major organ groups including sepsis and trauma, a median SOFA score of 7 in both arms and APACHE III risk of death at hospital discharge of 21.5% across both arms.

The need for transfusion was determined by managing physicians with a mean haemoglobin at randomisation of 77.3g/L and no transfusion protocol mandated. Around 97.5% went on to receive a transfusion in both arms with an average delay of 90 minutes between randomisation and transfusion, an average of 2 units of red cells were transfused with an interquartile range of 1-4 units, and similar transfusion rates of other blood products. The mean age of red cells in the short-term storage arm was 11.8 days versus 22.4 days in the long-term storage arm. There was no significant difference in the primary outcome of 90-day all-cause mortality – 24.8% in the short-term storage arm versus 24.1% in the long-term storage arm – nor in any secondary measures including 28-day mortality, ICU length of stay, rates of bloodstream infection, and requirement for mechanical ventilation or renal replacement therapy.

The ages of blood transfused in each arm resulted from the pragmatic methodology utilising the blood stocks available in each institution at the time, designed to ensure both that blood products were not wasted and that patients randomised were not excluded due to a lack of available blood of an appropriate age, as had happened in previous studies. Therefore in order to be able to compare the effect of transfusing the youngest compared to the oldest blood, the authors prespecified a subgroup comparing patients receiving blood stored for fewer than 8 days versus that stored for longer than 35 days. There was no difference in 90-day mortality in either of these subgroups or in an analysis considering storage duration as a continuous variable. These subgroups were however smaller and not powered to detect a significant difference. There was a signal of difference in a further prespecified subgroup analysis suggesting that in those with an APACHE III risk of death above average, transfusion with long-term storage red cells had a favourable impact on 90-day mortality. A similar effect was not seen with SOFA.

This is a well-conducted multicentre multinational trial, which improves on the data previously available, albeit confirming a lack of evidence that red cells stored for a particular length of time confer a mortality or morbidity benefit in intensive care patients. Whilst the trial achieved a significant difference of over 10 days between the storage durations of red cells transfused in the two arms, it could be argued that comparing red cells transfused at 11 and 22 days of storage might mask a potential difference between red cells at the extremes of storage duration. However as discussed the logistics of a randomised controlled trial focused on these groups are all but prohibitive. Additionally, although 10% of patients studied were admitted with trauma, the upper quartile of 4 units transfused means that patients requiring massive transfusion do not represent a significant proportion of the study population, nor were they separately studied. It is possible that the impact of red cell age may be more pronounced where there is a greater transfusion requirement. Similarly, a study sufficiently powered to do so may have been able to analyse the impact in other subgroups. For example, the detrimental effect of mitochondrial-derived damage-associated molecular patterns may be more marked in patients with sepsis where there is already an exaggerated and deleterious inflammatory response. This could potentially underline the interesting signal noted in the subgroup with a higher than average APACHE III predicted mortality.

In other aspects the study design is robust – the results are likely reflective of a general ICU population given the range of admitting diagnoses and the mortality figures observed, approximating those predicted; whilst there was no protocol determining transfusion practice, the average transfusion threshold recorded would be in keeping with the restrictive practice commonly followed; and there was no significant delay to transfusion resulting from randomisation.

Although the outcome does not support the authors’ hypothesis, the TRANSFUSE study is reassuring in that aside from a small increase in the rate of febrile non-haemolytic transfusion reactions in the short-term storage arm, where transfusion is indicated, there is no significant clinically-relevant harm directly attributable to use of red cells stored for either longer or shorter durations. This vindicates current transfusion laboratory practices. Future research may now be better focused on elucidating more objective triggers for and alternatives to transfusion in the setting of critically ill patients without major haemorrhage.

Article review prepared and submitted by David Woods (Department of Critical Care, St Thomas’ Hospital, London) on behalf of the EJRC.

References

1. Flegel WA, Natanson C, Klein HG. Does prolonged storage of red blood cells cause harm? Br J Haematol 2014;165(1):3-16
2. Korsak J (2012). Transfusion-Associated Bacterial Sepsis, Severe Sepsis and Septic Shock – Understanding a Serious Killer, Fernandez R (Ed.), ISBN: 978-953-307-950-9, InTech, Available from: http://www.intechopen.com/books/severe-sepsis-and-septic-shock-understanding-a-seriouskiller/transfusion-associated-sepsis
3. Simmons JD, Lee YL, Pastukh VM et al. Potential contribution of mitochondrial DNA damage associated molecular patterns in transfusion products to the development of acute respiratory distress syndrome after multiple transfusions. J Trauma Acute Care Surg 2017;82(6):1023-29
4. Lacroix J, Hébert PC, Fergusson DA et al. Age of Transfused Blood in Critically Ill Adults. N Engl J Med 2015;372:1410-18
5. Heddle NM, Cook RJ, Arnold DM et al. Effect of Short-Term vs. Long-Term Blood Storage on Mortality after Transfusion. N Engl J Med 2016;375:1937-45
6. Cooper DJ, McQuilten ZK, Nichol A et al. Age of Red Cells for Transfusion and Outcomes in Critically Ill Adults. N Engl J Med 2017;377:1858-67