Share this article on:

A Pilot Trial Evaluating the Clinical Effects of Prolonged Storage of Red Cells

Section Editor(s): Takala, JukkaHébert, Paul C. MD*; Chin-Yee, Ian MD; Fergusson, Dean PhD*; Blajchman, Morris MD; Martineau, Raymond MD§; Clinch, Jennifer BSc, MA*; Olberg, Bernhard MD

doi: 10.1213/01.ANE.0000148690.48803.27
Critical Care and Trauma: Research Report
Chinese Language Editions

The clinical consequences of prolonged storage of red cells have not been established. In this pilot study, we evaluated whether it would be feasible to provide a continuous supply of red cells stored <8 days. In addition, we examined the potential benefits attributed to “fresh” as compared to standard red cells in 66 critically ill and cardiac surgical patients. Nine patients were issued red cells but were not transfused. From the 57 remaining patients, the number of units transfused averaged 5.5 ± 8.43 red cell units in the experimental group compared to 3.3 ± 3.27 red cell units in the standard group (P = 0.25). The median storage time was 4 days in the experimental group compared to 19 days in the standard group (difference of 15 days; interquartile range of 12–16 days; P < 0.001). Overall, 73% of patients received red cells with storage times that corresponded to the treatment allocation more than 90% of the time. The group receiving red cells <8 days old tended to be older on average (68 ± 8.54 yr versus 63 ± 15.30 yr; P = 0.13) and have more comorbid illnesses (85% versus 65%; P = 0.09). In total, 27% of patients in the experimental group died or had a life-threatening complication as compared to 13% in the standard group (P = 0.31). There were no differences in prolonged respiratory, cardiovascular, or renal support after randomization (P > 0.05). A large clinical trial comparing red cell storage times is feasible and warranted given the limited available evidence.

IMPLICATIONS: Despite concerns regarding old blood, there is little evidence of adverse clinical consequences.

*University of Ottawa Centre for Transfusion Research, Ottawa Hospital Research Institute, Clinical Epidemiology Unit; †Division of Hematology, London Health Sciences Centre, Ontario; ‡Department of Laboratory Medicine, McMaster University Health Centre; §Division of Cardiovascular Anaesthesiology, Institut de Cardiologie de Montreal; and ∥Department of Pathology, Ottawa Hospital, Canada

Accepted for publication October 11, 2004.

Address correspondence and reprint requests to Paul C. Hébert, MD, Division of Hematology, London Health Sciences Centre, London, Ontario, Canada. Address e-mail to

By increasing hemoglobin concentration and improving oxygen (O2) delivery to tissues, red blood cell (RBC) transfusions serve as an important supportive measure for critical care patients and patients who undergo operative interventions with significant blood loss. The ability of RBC transfusions to deliver and release O2 has been evaluated using crude measures of global O2 delivery and utilization. The effect of prolonged storage of transfused RBC on tissues, specific organs, and globally on critically ill patients has never been investigated in vivo, primarily because of difficulties in accurately measuring tissue oxygenation. As a result, the shelf life for RBC was established using the proportion of cells that survive at least 24 h and median survival times, rather than an evaluation of RBC ability to deliver O2. Based on this standard, most regulatory agencies have approved a shelf life of 42 days for RBC.

However, there are well-defined biochemical and corpuscular changes to RBC during storage, collectively referred to as “the storage lesion” (1–3). These changes include a depletion of adenosine triphosphate and 2,3-DPG, membrane vesiculation (4–6), lipid peroxidation of cell membranes (7), and loss of deformability (8,9).1 During RBC storage, changes to the storage medium occur, including a progressive decrease in pH, an increase in plasma potassium, release of free hemoglobin from lysed RBC (10), and the generation of cytokines and other bioreactive substances (11–15). The clinical consequences of transfusing modified corpuscle and storage by-products are unknown (16–19). If prolonged storage truly does render RBC ineffective, then critically ill patients and patients undergoing cardiac surgical procedures may well experience negative clinical consequences from prolonged RBC storage. Despite these concerns, regional and hospital blood banks have developed inventory management strategies based upon minimizing blood wastage because of outdating of products because the benefits of fresh blood remain theoretical. As a consequence, patients in major centers often receive older products to ensure its use.

Given that we were unsure if blood banks could meet transfusion requirements for fresh RBC, we evaluated compliance with a strategy that would provide RBC stored <8 days. We also undertook a preliminary evaluation of major clinical consequences in cardiovascular surgical and critically ill patients who receive at least one RBC unit in the course of their care.

Back to Top | Article Outline


We designed a double-blind, multicenter, randomized, controlled pilot study. Patients who required at least one unit of allogeneic RBC, accepted to participate in the study, and who were either undergoing elective or urgent cardiac surgical procedures or admitted to an intensive care unit (ICU) at one of four Canadian hospitals between the dates of October 28, 1999 and May 31, 2001 were considered for enrollment. We excluded patients <16 yr of age, enrolled in another interventional study, previously enrolled in this study, with a terminal illness with a life expectancy <3 mo, not expected to survive more than 12 h post-ICU admission, not being provided with all required therapeutic interventions to sustain life, unable to receive blood products because of difficulties with cross matching or a history of unexplained severe transfusion reaction, who received a blood transfusion in the 6 mo preceding enrollment, who had undergone bone marrow transplantation, who had a malignant hematological illness treated within a year of hospital admission, or were a heart transplant or artificial heart recipient. Each day, before screening, study personnel confirmed the adequacy of supply of RBC stored for <8 days (experimental treatment) and standard RBC. Consent was obtained either before scheduled cardiac surgical procedures or when the hemoglobin concentration decreased to less than 100 g/L for critically ill patients. The Research Ethics Committees of all participating institutions approved this study. Informed consent was obtained from the patient or surrogate decision maker, as well as the attending staff.

Patients were randomly allocated to receive either RBC stored for <8 days or standard issue RBC. The standard therapy group received RBC issued from the hospital blood bank with the longest storage times first in accordance to standard blood bank procedure. To ensure the maximum age separation between groups, we only allocated patients on days when the average age of RBC in the blood bank exceeded 15 days. The experimental group received RBC that were stored <8 days from the day of collection. This time interval represented a compromise between feasibility and having a product that would still be considered fresh. Once randomized, age-specific RBC were transfused during the patients’ entire hospital stay up to 30 days. All clinical management decisions, including the decision to administer transfusions, remained at the discretion of the attending team. The use of other blood products and alternatives to RBC transfusions, such as cell salvage and medications, were not controlled but recorded.

Randomization was initiated by a telephone request from the attending team for a RBC transfusion. Upon receipt of a request for RBC, a blood bank technologist who was not involved in clinical care opened a sequentially numbered sealed opaque envelope. As a consequence, concealment of randomization and blinding of the study interventions were ensured. The randomization schedule consisted of a computer-generated random listing of the two treatment allocations blocked into groups of four and six and stratified by center and categories of cardiac surgery or critical care. An independent statistician prepared the listing. A patient was considered randomized once requested RBC were issued by the local blood banks.

Canadian Blood Services or Héma Québec supplied all RBC products administered in this study. Both organizations produced RBC by collecting whole blood into CPD-2 anticoagulant. The platelet-rich plasma was removed, and then 100 mL of AS-3 (Nutricel™, ProNature Laboratories) was added. The blood providers reported that RBC units had the following characteristics: an average volume of 300 mL; a hematocrit of 65%; and an average storage time of 14 days, ranging from region to region and according to use and blood procurement. All blood products had undergone universal prestorage leukoreduction using either the Leukotrap-RC or the Leukotrap WB in-line systems (Pall Medical, Blood Processing Group, East Hills, NY). Leukofiltration using these systems reduced white blood cell content of a unit of RBC from an average of 3.0 × 109/unit to 2.5 × 105/unit, a decrease of 4 logs.

We evaluated compliance to treatment strategies by determining the availability and administration of predefined RBC products. Patients were considered compliant to protocol if 90% or more of transfused RBC were stored 8 days or less in the experimental group or more than 15 days in the control group. To maximize our ability to detect a difference in a small trial, we grouped hospital mortality, serious nosocomial infections, and thrombotic events including myocardial infarctions and acute ischemic strokes into a predefined composite outcome. Myocardial infarction was defined as new q waves in two consecutive leads on electrocardiography or a fivefold increase in serum troponin or the MB fraction of serum creatine kinase (20), whereas cerebrovascular accidents were defined as new focal neurological deficit lasting more than 24 h confirmed by a neurologist or new deficit on an imaging study reference (21,22). Specific nosocomial infections considered in the definition were pneumonia, surgical wound infections including mediastinitis and osteomyelitis, and bacteremia from organisms not considered normal skin flora. Definitions for all infections were based upon the Centers for Disease Control criteria (23). As secondary clinical outcomes, we examined the number of organ failures combined with death assessed using the score by Hebert et al. (24,25). Finally, we examined the intensity of interventions including respiratory support, defined as the number of days on mechanical ventilation; intensity of cardiovascular support, defined as the number of days requiring vasoactive drugs, an aortic balloon pump, or ventricular assist device; and the intensity of renal support by measuring the number of days of renal replacement therapy.

The sample size was arbitrarily established at 15–20 patients per center in this pilot. A total sample size approximating 70 patients would allow us to estimate adherence rates within each group with 95% confidence intervals (CI) of ±3.5% if the compliance rate approximated 90%. For a composite clinical outcome, a difference between groups of 50% would have a 95% CI of ±12%.

Our primary analysis was based upon treatment received given that most of the reasons for not receiving RBC were related to inadequate implementation of study procedures at a single site. Compliance rates were established for each group followed by calculations of 95% CI and compared using a χ2 statistic. RBC storage times were described by calculating median values and interquartile ranges for all RBC products transfused within each group once randomized during the entire hospital stay. To demonstrate a significant separation in storage times using standard RBC as a control, we compared median differences in storage time using a Wilcoxon rank sum test statistic. The composite measure, including major adverse clinical consequences, was contrasted using a χ2 statistic. We explored the influence of comorbid illnesses (presence or absence), major diagnostic grouping (cardiac surgery and critical care), and center (four centers) using stratified analyses and logistic regression models with the composite measure as the dependent variable. Outcomes related to the intensity and type of care received such as lengths of hospital and ICU stay, the number of days of mechanical ventilation, vasoactive drug use, and dialysis were compared using the Wilcoxon rank sum test. We reported all P values as two-tailed, without corrections for multiple comparisons.

Back to Top | Article Outline


Sixty-six patients were allocated to one of two treatment groups. However, seven patients (five in the group allocated to receive RBC stored for <8 days and two in the standard group) were issued RBC by the blood bank at the request of the attending staff but did not receive any transfusions. Two additional patients in the experimental group were considered crossovers because they did not receive RBC stored for <8 days. From the remaining 57 patients in the analysis based upon treatments received, 26 were assigned to the experimental group, and 31 were assigned to receive RBC considered standard blood bank issue. Forty-two cardiovascular surgical patients were equally allocated to each treatment group, and the remaining 15 critically ill patients were also equally divided between groups (Table 1). At baseline, the two treatment groups seemed similar in most respects. However, the group receiving RBC <8 days of age tended to be older on average (68 ± 8.54 yr versus 63 ± 15.30 yr) and had more comorbid illnesses (85% versus 65%). The median number of units transfused was three (interquartile range, 2–5) RBC units in the experimental group as compared with two (interquartile range, 2–4) RBC units in the standard group (P = 0.25).

Table 1

Table 1

Overall, 73% of patients received RBC with storage times that corresponded to the treatment allocation more than 90% of the time. The compliance target of 90% was attained by 91% of patients allocated to the group receiving RBC stored 8 days or less versus 59% of patients in the standard group (P = 0.008). The median storage time was 4 days in the experimental group compared to 19 days in the standard group, a median difference of 15 days (interquartile range, 12–16 days; P < 0.001)(Fig. 1).

Figure 1

Figure 1

For all randomized patients (n = 66), 24% (n = 7) of patients in the experimental group either died or had a life-threatening complication, the composite outcome, as compared to 15% (n = 5) in the standard group (risk difference of 0.09; 95% CI, −0.10–0.28; P = 0.35). If we considered all patients transfused (including two crossovers), then the proportion of patients with the composite outcome was 27% (n = 7) in the experimental group compared to 12% in the control group (risk difference of 0.15; 95% CI, −0.06–0.35; P = 0.19). In considering only the 57 patients who received age-appropriate RBC, the proportion of patients with the composite outcome was 27% versus 12% in the experimental versus control groups, respectively (risk difference of 0.14; 95% CI, −0.07–0.35; P = 0.31) (Table 2). While in the hospital, there were five deaths observed in the experimental group and three in the standard group (P = 0.45). There were no reported bacteremia, deep incisional infections, endocarditis, or mediastinitis reported in either group. No patients had a cerebral hemorrhage, whereas three patients in the experimental group had a postoperative myocardial infarction, and one patient in the same group had an acute ischemic neurologic event. The stratified analysis or logistic regression did not detect any clinically or statistically important changes in the odds of having an adverse clinical outcome as a composite measure.

Table 2

Table 2

Back to Top | Article Outline


In this study, we demonstrated that it is possible to maintain a supply of RBC stored for less than eight days more than 90% of the time while ensuring a separation in storage times exceeding 15 days using standard therapy as a control. In addition, we were unable to detect a difference in major clinical consequences after critical illness and cardiac surgery.

A number of considerations related to feasibility arose before and during this study. Before the enrollment of patients, we developed a detailed description of blood bank procedures at each institution focused on ensuring a supply of RBC stored less than eight days (inventory management considerations), distribution of the appropriate product (allocation and distribution considerations), and considerations related to the identification of patients with a very high or definitive probability of requiring RBC. Despite these many procedures, seven patients who were considered randomized were not transfused RBC. In this instance, patients were issued RBC by the blood bank at the request of the attending anesthesiologist in response to anticipated blood loss during surgical interventions. Fortunately, the decision to withhold a transfusion was not based upon group assignment. This concern arose early in the trial and was remedied by providing assurances to clinicians that RBC would be delivered within minutes of the request. The blood bank use of rapid intraoperative randomization based on a transfusion decision creates a number of logistical challenges. Study maneuvers such as pretesting of interventions at each site, a rapid call back of blood products issued by local blood banks, and availability of point-of-care hemoglobin measurements may all enhance the decision process thereby minimizing the number of RBC requests but not transfused. Early attention to study procedures, such as ensuring the availability of cross-matched fresh RBC, would also have potentially avoided the two crossovers, a rate still considered acceptable in a large clinical trial.

Although we were unable to detect meaningful clinical benefits from the transfusion of RBC stored for less than eight days, we believe that prolonged RBC storage should be further evaluated in laboratory and clinical studies. This view was shared by the authors (26) of an editorial accompanying the only other randomized controlled trial (RCT) evaluating storage times. In the only other published RCT, Walsh et al. (27) evaluated changes in gastric intramucosal pH (pHi), a measure of gastric perfusion, in 22 mechanically ventilated critically ill patients who required a RBC transfusion. In this study, the authors were not able to detect any adverse consequences on pHi and changes in the arterial-gastric mucosal CO2 gap with a storage time exceeding 20 days compared with patients receiving RBC less than five days old. These results contradicted earlier observations in a before and after study conducted by Marik and Sibbald (28) who documented an inverse relationship between the age of transfused RBC and gastric pHi (r = −0.71; P < 0.001).

A number of retrospective clinical studies have documented an association between prolonged storage times and adverse clinical outcomes including mortality (29), pneumonia (30), serious infections (31), and length of stay (32) in many patient populations including multiple trauma victims, critically ill patients, and patients undergoing cardiac surgical procedures. Unfortunately, all retrospective studies evaluating prolonged RBC storage will invariably be subject to the confounding influences of factors such as the number of RBC units transfused, the mixture of storage times from the multiple units transfused throughout a hospital stay, and patient factors, including severity of illness. Inferences related to the clinical consequences of transfusing RBC with a storage time less than eight days were limited by a small sample size and imbalances and clinically important imbalances in baseline characteristics.

In summary, we documented that a larger trial would be feasible but would require significant attention to the adequacy of the RBC supply during the trial, randomization concerns, and crossovers. In addition, given the costs of adapting the collection and distribution process within blood systems, absolute differences in major outcomes such as mortality, organ failure, and infections less than 3%–4% between RBC with short storage times and standard issue RBC may not be worth pursuing. At this time, there is insufficient evidence to make any recommendations regarding the transfusion of RBC stored for eight days or less in patients considered at high risk of adverse consequences from blood transfusions.

Back to Top | Article Outline


1. Chin-Yee I, Arya N, d’Almeida MS. The red cell storage lesion and its implication for transfusion. Transfus Sci 1997;18:447–58.
2. Card RT. Red cell membrane changes during storage. Transfus Med Rev 1988;2:40–7.
3. Wolfe LC. The membrane and the lesions of storage in preserved red cells. Transfusion 1985;25:185–203.
4. Greenwalt TJ, Bryan DJ, Dumaswala UJ. Erythrocyte membrane vesiculation and changes in membrane composition during storage in citrate-phosphate-dextrose-adenine-1. Vox Sang 1984;47:261–70.
5. Snyder LM, Fairbanks G, Trainor J, et al. Properties and characterization of vesicles released by young and old human red cells. Br J Haematol 1985;59:513–22.
6. Wagner GM, Chiu DTY, Yee MC, Lubin BH. Red cell vesiculation-a common membrane physiologic event. J Lab Clin Med 1986;108:315–24.
7. Knight JA, Voorhees RP, Martin L, Anstall H. Lipid peroxidation in stored red cells. Transfusion 1992;32:354–7.
8. Card RT, Mohandas N, Perkins HA, Shohet SB. Deformability of stored red blood cells: relationship to degree of packing. Transfusion 1982;22:96–101.
9. Card RT, Mohandas N, Mollison PL. Relationship of post-transfusion viability to deformability of stored red cells. Br J Haematol 1983;53:237–40.
10. Latham JT Jr, Bove JR, Weirich FL. Chemical and hematologic changes in stored CPDA-1 blood. Transfusion 1982;22:158–9.
11. Heddle NM. Febrile nonhemolytic transfusion reactions to platelets. Curr Opin Hematol 1995;3:478–83.
12. Smith KJ, Sierra ER, Nelson EJ. Histamine, IL-1β, and IL-8 increase in packed RBC stored for 42 days but not in RBC leukodepleted pre-storage. Transfusion 1993;33:53S.
13. Miletic VD, Popovic O. Complement activation in stored platelet concentrates. Transfusion 1993;33:150–4.
14. Schleuning M, Bock M, Mempel W. Complement activation during storage of single-donor platelet concentrates. Vox Sang 1994;67:144–8.
15. Silliman CC, Clay KL, Thurman GW, et al. Partial characterization of lipids that develop during the routine storage of blood and prime the neutrophil NADPH oxidase. J Lab Clin Med 1994;124:684–94.
16. Chaplin H Jr, Beutler E, Collins JA, et al. Current status of red-cell preservation and availability in relation to the developing National Blood Policy. N Engl J Med 1974;291:68–74.
17. Bunn HF, May MH, Kocholaty WF, Shields CE. Hemoglobin function in stored blood. J Clin Invest 1969;48:311–21.
18. Sugerman HJ, Davidson DT, Vibul S, et al. The basis of defective oxygen delivery from stored blood. Surg Gynecol Obstet 1970;137:733–41.
19. Valeri CR, Collins FB. The physiologic effect of transfusing preserved red cells with low 2,3-diphosphoglycerate and high affinity for oxygen. Vox Sang 1971;20:397–403.
20. Karski JM, Dowd NP, Joiner R, et al. The effect of three different doses of tranexamic acid on blood loss after cardiac surgery with mild systemic hypothermia. J Cardiothorac Vasc Anesth 1998;12:642–6.
21. Randomised trial of endarterectomy for recently symptomatic carotid stenosis: final results of the MRC European Carotid Surgery Trial (ECST). Lancet 1998;351:1379–87.
22. Taylor DW, Barnett HJM, Haynes RB, et al. Low-dose and high-dose acetylsalicylic acid patients undergoing carotid endarterectomy: a randomised controlled trial. Lancet 1999;353:2179–84.
23. Garner JS, Jarvis WR, Emori TG, et al. CDC definitions for nosocomial infections, 1988. Am J Infect Control 1988;16:128–40.
24. Hebert PC, Drummond AJ, Singer J, et al. A simple multiple system organ failure scoring system predicts mortality of patients who have sepsis syndrome. Chest 1993;104:230–5.
25. Bihari D, Smithies M, Gimson A, Tinker J. The effects of vasodilation with prostacyclin on oxygen delivery and uptake in critically ill patients. N Engl J Med 1987;317:397–403.
26. Napolitano LM, Corwin HL. Efficacy of blood transfusion in the critically ill: does age of blood make a difference [comment]? Crit Care Med 2004;32:594–5.
27. Walsh TS, McArdle F, McLellan SA, et al. Does the storage time of transfused red blood cells influence regional or global indexes of tissue oxygenation in anemic critically ill patients? Crit Care Med 2004;32:364–71.
28. Marik PE, Sibbald WJ. Effect of stored-blood transfusion on oxygen delivery in patients with sepsis. JAMA 1993;269:3024–9.
29. Purdy FR, Tweeddale MG, Merrick PM. Association of mortality with age of blood transfused in septic ICU patients. Can J Anaesth 1997;44:1256–61.
30. Vamvakas EC, Carven JH. Transfusion and postoperative pneumonia in coronary artery bypass graft surgery: effect of the length of storage of transfused red cells. Transfusion 1999;39:701–10.
31. Offner PJ, Moore EE, Biffl WL, et al. Increased rate of infection associated with transfusion of old blood after severe injury. Arch Surg 2002;137:711–6.
32. Keller ME, Jean R, LaMorte WW, et al. Effects of age of transfused blood on length of stay in trauma patients: a preliminary report [see comment]. J Trauma 2002;53:1023–5.

1 1. Card RT, Fergusson DJ. Relationship of stored red cell deformability to survivability following transfusion: in vitro prediction of in vivo viability. Blood 1987;70:327a.
Cited Here...

© 2005 International Anesthesia Research Society