Red blood cell storage age – what we know from clinical trials

KEYWORDS:

• Red blood cell
• transfusion
• storage lesion
• clinical trials
• oxygen delivery

1. Introduction

Red blood cell (RBC) transfusion has become one of the most commonly utilized life-saving therapies in medicine with approximately 12–16 million units transfused in the US per year [1,2]. Since storage solutions were developed 100 years ago, they have been modified to allow for increased storage duration, primarily to increase availability, especially in remote areas, such as during combat operations [3]. Currently in the US and Europe, RBCs can be stored up to 42–56 days based upon at least 75% of radiochromium-labeled cells remaining in the circulation at 24 h after infusion and hemolysis less than 1% [4,5]. As RBCs are stored over time, the metabolic and biochemical changes that occur have been coined ‘the storage lesion.’

To further improve the availability of RBCs, it is standard for blood distributors to transfer RBCs that are close to expiration at community hospitals to medical centers where utilization rates are higher, and for hospital blood banks to issue the ‘first-in, first-out’ inventory to reduce RBC outdating and waste [2,6]. A consequence of these policies focused on inventory management is that patients with increased severity of illness are transfused RBCs of increased storage duration compared to less acutely ill patients. It is important to recognize that as storage duration for RBC units has increased for logistical purposes, there has not been a commensurate evaluation of clinical efficacy or safety.

As a result of the preferential allocation of older RBCs to sicker patients, there has been concern regarding the quality of RBCs after prolonged storage related to both efficacy and safety. This has led to a significant amount of research into the clinical relevance of the storage lesion with laboratory, preclinical, and clinical studies [4,7–9]. The results of these studies suggested biologic plausibility that the RBC storage lesion was potentially able to affect outcomes, especially in critically ill populations. These studies though are limited due to potential bias and higher-quality studies were needed to more accurately determine if RBC storage age affects outcomes. Consequently, a few randomized controlled trials (RCTs) have been completed, and a few are still underway. Dissecting out what each of these studies may mean for patients who are transfused RBCs based upon the population studied and their design is controversial.

2. ‘Storage lesion’

As RBCs age in storage solution, changes that occur due to impaired cellular energetics include alterations in RBC shape and deformity, oxygen affinity, rheologic properties, and loss of membrane proteins and carbohydrates [4,10–12]. These changes result in an early loss of nitric oxide bioactivity, accumulation of lactic acid and potassium, a decrease in 2, 3-diphosphoglycerate, an increase in NO scavenging by cell-free hemoglobin, increased levels of iron moieties, and an accumulation of shed bioactive proteins, lipids, and microparticles [4,13–15]. The composite effect of each aspect of the storage lesion has the potential to reduce perfusion and oxygen delivery, which is the opposite of intended effect of the RBC transfusion [11]. While a number of animal studies in sepsis that have demonstrated an increased mortality with transfusion of longer stored RBCs [4,7,14,16–18], human data have been less consistent.

3. Clinical studies

Over 80 different studies in humans have demonstrated conflicting results when comparing ‘older, less fresh’ RBCs with ‘younger, fresh’ transfused products. These include 8 RCTs, 32 observational studies, and multiple meta-analyses [7,8,19,20]. Many of these studies have included patients with critical illness, cardiac surgery, trauma, and cancer with only a few pediatric studies. Most studies have compared mortality as their primary end point; however, other outcomes have included alterations in nosocomial infections, duration of mechanical ventilation, development of deep venous thrombosis, hospital or intensive care unit (ICU) length of stay (LOS), or increased rates of sepsis, renal failure, or multi-organ dysfunction (MODS) [4,10,19,21].

Presently, the impact of the ‘storage lesion’ on clinical outcomes varies by the type of study and outcome variable. Most published studies are retrospective, observational studies conducted in single centers which have been helpful in generating hypotheses for RCTs. Completed RCTs have compared RBC transfusions aged less than 10 days to between a mean/median of 14- to 32-day-old RBCs. In these studies, there was no survival advantage in transfusing ‘fresher’ blood in different clinical populations including premature infants, children primarily with malaria, adults with complex cardiac surgery, and critically ill adult patients [7,8,22–25]. Only one of these studies in children primarily with malaria was able to randomize RBCs < 10 days to 28–35 days of storage. None of the RCTs randomized to ‘very old’ blood at the end of storage (>35 days). Thus, the RCTs published have primarily been designed to ask whether fresh RBCs improve outcomes compared to standard issue or slightly older RBCs. None of the RCTs have been designed to determine if older RBCs (>35 days of storage) worsen outcomes. This is important to determine since in some regions of the US, more than 20% of RBCs transfused are older than 35 days [26].

To address these issues, numerous meta-analyses and systematic reviews have attempted to tackle the question of the efficacy and safety of longer stored RBCs by combining studies. In two separate systematic reviews by Lelubre et al. (55 studies) and Aubron et al., (32 studies), 26 and 18 studies, respectively, reported harmful effects while 29 and 14, respectively, did not [23,27]. Two separate meta-analyses arrived at conflicting results. In a large meta-analysis of 21 studies enrolled between 1991 and 2009 with no heterogeneity in effects size (I² = 0%) adjusted mortality was substantially increased (OR 1.16; 95% CI 1.07–1.24; P = 0.0001) for patients receiving RBCs stored longer than 21 days [19]. In this analysis, pneumonia (OR 1.17; 95% CI: 1.08–1.27; P = 0.0001) and MODS (2.26; 95% CI: 1.56–3.25; P < 0.001) were worsened with RBCs stored longer than 21 days [19]. However, in another large meta-analysis (27 studies) by Vamvakas et al., there was no associated increased mortality or morbidity with stored older RBCs [28]. Finally, in the two largest and most recent meta-analyses by Remy et al. and Alexander et al., the evidence is still divided [7,8]. Alexander et al. found in a recent meta-analysis of 12 RCTs that compared two defined ages per study (‘fresh’ vs. ‘older or standard issue.’ no difference in mortality (RR 1.04; 95% CI 0.94–1.14; P = 0.45 I² = 0) or on adverse events [8]. Remy et al. detected similar findings of RCTs in their recent meta-analysis finding an increased risk of death (OR 1.13; 95% CI 1.03–1.24, P = 0.01) and a different mortality effect than RCTs (p = 0.02) [7].

So, why the differences between RCTs and observational studies (and related meta-analyses)? Could these discrepancies be related to the overall differences in patient populations, quality of study design, definition of end points, preparation of RBCs, and definitions of ‘old’ and ‘fresh’ blood? The latter question was answered in part by Remy et al. who found in their analysis a striking difference in that RBC storage age in RCTs versus observational studies was significantly lower (p = 0.01) [7]. Likewise, other confounders such as differences in transfusion volume, blood type, mixed blood ages, and patient case mix make direct comparisons between studies impossible. Finally, lacking a mutually agreed definition of what constitutes ‘fresh’ versus ‘older’ will significantly limit comparisons between RCTs. Another explanation for why the RCTs are not indicating that RBC storage age effects outcomes is that due to donor variation in the rate of ‘RBC aging’ or development of the storage lesion, RBC age itself may not be an accurate surrogate for RBC quality. As a result, randomizing according to RBC age may not separate RBCs of increased versus decreased quality. The RCTs completed have also not examined patients that require a large volume of RBCs such as those with traumatic injury. It is possible that the clinical relevance of the RBC storage lesion may require a larger dose of RBCs that has been evaluated in currently published RCTs. The trauma population has been identified as a group of patients that is vulnerable to the storage lesion [29,30]. High-quality trials in this group of patients are required. Perhaps comparative effectiveness studies are the optimal trial design in this population and should be considered as another type of high-quality trial design.

4. Where do we go from here?

Presently, we have learned that

1. Currently published RCTs indicate that fresher RBCs do not improve outcomes compared to moderately aged RBCs in severely anemic populations and those that require a small volume transfusion of RBCs.

2. It has been difficult to design and execute RCTs to determine if old RBCs (>35 days) worsen outcomes especially in patients that require a high volume of RBCs such as those with hemorrhagic shock.

3. Perhaps RBC storage age is not an accurate surrogate for RBC quality and that there is a need to establish which parameters do reflect optimal efficacy and safety.

Provocatively, as it would be deemed unethical in most countries to design an RCT specifically testing the superiority of fresh versus very old RBCs (35 days), the precautionary principle would suggest that we should at least question whether we should keep giving these very same ‘older’ units we are unwilling to test prospectively to our recipients.

RBC quality metrics that accurately reflect efficacy and safety are needed. The US FDA is hosting a working group meeting in October of 2016 that will survey key opinion leaders on candidate measures for RBC quality. These discussions may incorporate the use of cell deformability, oxidative injury, nitric oxide metabolism, or free iron into consideration as RBC quality metrics. Studies will then be needed to determine how to define RBC quality and if changes in these metabolic or biochemical profiles are more accurate surrogates for RBC quality and if they are clinically relevant. RBC storage age is still a candidate quality metric that requires further examination. It is likely that not just one parameter will reflect quality. Instead, a panel of metrics may more accurately reflect quality, and in the age of precision medicine perhaps we will eventually align RBC quality metrics with transfusion recipient characteristics for optimal efficacy and safety. Likewise, it is paramount to the discussion to understand how to design carefully, future studies that can decipher subtle nuances in RCT and that reflect actual clinical practice. Likewise, a better understanding of what the key elements that need to be studied in order to better to understand the storage lesion influence on specific patient population outcomes is needed. Presently, these questions remain elusive and open-ended.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Additional information

Funding

This paper was not funded.