Mortality outcomes in patients transfused with fresher versus older red blood cells: a meta-analysis

Transcription

1 ORIGINAL PAPER DOI: /vox Mortality outcomes in patients transfused with fresher versus older red blood cells: a meta-analysis C. Chai-Adisaksopha, 1,2 P. E. Alexander, 1,2 G. Guyatt, 1,2 M. A. Crowther, 1,2,3 N. M. Heddle, 1,2,3,4,5 P. J. Devereaux, 1,2,6 M. Ellis, 7 D. Roxby, 8 D. I. Sessler 9 & J. W. Eikelboom 1,6,10 1 Department of Medicine, McMaster University, Hamilton, ON, Canada 2 Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, ON, Canada 3 Department of Pathology and Molecular Medicine, McMaster University, Hamilton, ON, Canada 4 McMaster Centre for Transfusion Research, McMaster University, Hamilton, ON, Canada 5 Centre for Innovation Canadian Blood Services, Hamilton, ON, Canada 6 Population Health Research Institute, Hamilton, ON, Canada 7 Meir Medical Centre Kfar Saba and Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel 8 SA Pathology Transfusion Service, Flinders Medical Centre and Flinders University, Adelaide, SA, Australia 9 Department of Outcomes Research, Anesthesiology Institute, Cleveland Clinic, Cleveland, OH, USA 10 Thrombosis & Atherosclerosis Research Institute, Hamilton, ON, Canada Received: 11 November 2016, revised 3 January 2017, accepted 6 January 2017 Background Among transfused patients, the effect of the duration of red blood cell storage on mortality remains unclear. This study aims to compare the mortality of patients who were transfused with fresher versus older red blood cells. Methods We performed an updated systematic search in the CENTRAL, MED- LINE, EMBASE and CINAHL databases, from January 2015 to October RCTs of hospitalized patients of any age comparing transfusion of fresher versus older red blood cells were eligible. We used a random-effects model to calculate pooled risk ratios (RRs) with corresponding 95% confidence interval (CI). Results We identified 14 randomized trials that enrolled participants. Allcause mortality occurred in 1219 of 9531 (128%) patients who received a transfusion of fresher red blood cells and 1810 of (107%) in those who received older red blood cells (RR: 104, 95% CI: , P = 090, I 2 = 0%, high certainty for ruling out benefit of fresh blood, moderate certainty for ruling out harm of fresh blood). In six studies, in-hospital death occurred in 691 of 7479 (92%) patients receiving fresher red cells and 1291 of (88%) receiving older red cells (RR: 106, 95% CI: , P = 081, I 2 = 0%, high certainty for ruling out benefit of fresh blood, moderate certainty for ruling out harm of fresh blood). Conclusion Transfusion of fresher red blood cells does not reduce overall or in-hospital mortality when compared with older red blood cells. Our results support the practice of transfusing patients with the oldest red blood cells available in the blood bank. Key words: blood processing, clinical trial, red cell component, transfusion therapy. Background Correspondence: John W. Eikelboom, 237 Barton St. E., Hamilton, ON L8L 2X2, Canada eikelbj@mcmaster.ca Transfusion of red blood cells aims to replace blood volume and restore oxygen-carrying capacity in anaemic patients and those with acute haemorrhage. Red blood cells are the most frequently transfused blood product [1]. 1

2 2 C. Chai-Adisaksopha et al. Each year, approximately 85 million red blood cells are transfused worldwide [2]. Red blood cells are stored at a temperature of 4 C for up to 42 days [3]. Blood storage is associated with biochemical, structural and functional changes that may impair their ability to deliver oxygen to tissues [4 8]. The effect of prolonged red blood cell storage prior to transfusion on clinical outcomes remains controversial. Many observational studies have suggested a survival benefit of fresher compared with older red blood cell transfusions [9 15]. Wang et al. [16] conducted a metaanalysis (18 observational studies and three randomized controlled trials) comparing between new blood and old blood (definitions of new and old blood were set by each study). This study concluded that older blood was associated with higher risk of mortality (odd ratio [OR]: 116, 95% confidence interval [CI]: [16]). However, observational studies on the effects of red blood cells storage are subject to be bias due to definition of storage time, number of transfused red blood cells, confounders and publication bias [17]. To solve the conflicting results of observational studies, randomized controlled trials were conducted. A meta-analysis of 12 randomized controlled trials involving 5229 transfused participants [18 29] reported no significant impact of fresher compared with older red blood cell transfusion on mortality (relative risk [RR]: 104, 95% CI: , P = 045) [30]. We recently reported the results of a large pragmatic randomized trial of more than patients in which we found no difference in survival between patients receiving fresher versus older red blood cells [31]. The availability of new data from more than patients provides an opportunity to further reduce residual uncertainty regarding the effect of transfusion of fresher compared with older red blood cell transfusion on mortality. Accordingly, we performed an updated systematic review and meta-analysis to obtain best estimates of the effect of red blood cell storage duration prior to transfusion on mortality. Methods We conducted a systematic review and meta-analysis of randomized controlled trials (RCTs) comparing fresher versus older red blood cell transfusion, adhering to standards of the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) statement [32]. Data source and search strategy We performed a systematic search in the Cochrane Library Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE and CINAHL. The reference lists of included RCTs, systematic reviews and relevant reviews were searched to identify further relevant trials. We did not apply language restrictions. Our search strategy replicated a prior systematic review/meta-analysis [30]. The search was implemented from January 2015 to October The search used the following terms (appendix): blood or erythrocyte* or red cell*, red blood cell*, RBC or whole blood, transfuse*, infuse*, retransfus*, erythrocyte transfusion, age, aged, aging, fresh*, old, older, oldest, new, newer, newest, young, younger, youngest or store*, storage, storing, preserv*, MeSH term blood transfusion, blood component transfusion, erythrocytes, *time factors. Eligibility criteria RCTs of hospitalized patients of any ages comparing transfusion of fresher versus older red blood cells were eligible. We included studies investigating both surgical and nonsurgical participants. We applied no exclusion criterion based on the indications for the blood transfusion. Studies were excluded if they did not report mortality outcomes. Study selection and data extraction Two investigators (CC and PEA) independently screened the titles and abstracts and retrieved the full text of any articles that either reviewer deemed potentially eligible studies. The investigators resolved disagreements on full-text eligibility review through discussion. Two investigators (CC and PEA) independently performed data extraction including study design, setting, sample size, inclusion and exclusion criteria, patient characteristics, details of transfusion, age of transfused red blood cells, indication for transfusion and outcome measures including number of events in each treatment group. End-point The primary outcome of this study was all-cause mortality at any time-point. The secondary outcome was in-hospital mortality. Risk of bias assessment Two investigators (CC and PEA) independently assessed risk of bias using the Cochrane Collaboration s tool for assessing risk of bias [33, 34]. We assessed risk of bias in nine domains: selection bias (random sequence generation and allocation), performance bias (blinding of participants and personnel), detection bias (blinding outcome assessors), attrition bias (incomplete outcome data), reporting

3 Fresher vs older red cell transfusion 3 bias (selective outcome report), bias due to severe baseline imbalance, bias due to early stoppage or discontinuation, and other bias. Each domain was judged to be either high risk of bias, probably high risk of bias, low risk of bias and probably low risk of bias [35]. The reviewers resolved disagreements by discussion. Statistical analysis Primary analysis We calculated the pooled estimate of the treatment effect using a random-effects model (DerSimonian and Laird random-effects methods) [36]. We present risk ratios (RRs) with their corresponding 95% confidence intervals (CIs), along with associated forest plots. P-values were calculated using a two-sided test with a threshold P-value of <005. Heterogeneity was assessed using the I 2 statistic.[37] An I 2 of 0 40% indicated minor heterogeneity, 30 60% moderate heterogeneity, 50 90% substantial heterogeneity and % considerable heterogeneity [37]. We assessed publication bias by constructing a funnel plot of standard error of the effect estimate vs effect size and considered publication bias present if inspection of the funnel plot showed substantial asymmetry [38]. Statistical tests for publication bias used both Egger s regression test [39] and Peters test [40]. The main analyses were conducted using Stata version 13.0 (StataCorp Stata Statistical Software: Release 13. College Station, TX: StataCorp LP). The funnel plot was generated using Review Manager [(RevMan) version 5.3. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014]. Subgroup analysis We performed a priori subgroup analyses based on age of red blood cells in the old red cells group (<20 or 20 days) and study participants (paediatric or adult patients). Subgroup interactions were investigated using a chi-square statistic [37]. Certainty of evidence assessment We assessed the certainty of evidence (also referred to as confidence or quality) for all-cause mortality and in-hospital mortality according to the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) [41] evaluating the following domains: risk of bias, inconsistency of the results, indirectness of the evidence, publication bias, imprecision and other considerations. The certainty of the evidence was judged as high, moderate, low or very low. Results Study selection We screened 2220 citations from the updated search and 12 RCTs from a previous systematic review [30]. After screening of titles and abstracts, we retrieved 19 full-text articles for eligibility assessment. Of these, 14 RCTs proved eligible (Fig. 1) [18 29, 42]. Table 1 presents the characteristics of the included trials. Five trials [18, 25, 27, 28, 31, 42] investigated paediatric patients, eight trials investigated adult patients, [19 22, 24, 26, 29, 31], and one trial did not report the age of participants [23]. Seven studies [18, 20, 25 28, 42] enrolled patients in intensive care units (ICUs) or non-surgical units, four [19, 21, 23, 24] enrolled surgical patients, and three [22, 29, 31] enrolled those who had either medical or surgical conditions. The mean or median duration of storage of transfused red blood cells ranged from 16 to 13 days in patients who were allocated to receive fresher red blood cells and days in patients allocated to receive older red blood cells. Risk of bias assessment Figure 1S appendix presents the risk of bias assessment. All of the studies except one [19] were judged to be low risk of bias. Most of the included trials did not blind the outcome assessors; however, mortality can be ascertained without risk of bias. Trials demonstrated concealment of randomization was satisfactory and loss to follow-up was infrequent. Main outcome In the 14 eligible trials, 1219 of 9531 (1279%) patients who received transfusion of fresher red blood cells died compared with 1810 of (1074%) of those who received older red blood cells (RR: 104, 95% CI: , P = 090, I 2 = 0%, Fig. 2). P-values for the test of interaction for the two subgroup analyses did not approach statistical significance (P = 077 for duration of red blood cells storage, P = 039 for patient population, Appendix Figs S2, S3). Although the inspection of the funnel plot demonstrated minimally asymmetrical appearance (Fig. 3), the Egger s and Peters tests did not suggest publication bias (P-value 049 and 027, respectively). In six studies [26 29, 31, 42] that reported in-hospital mortality, death occurred in 691 of 7479 (924%) patients who received transfusion of fresher red blood cells and 1291 of (875%) who received older red blood cells (RR: 106, 95% CI: , P = 081, I 2 = 0%, Fig. 4). We did not assess publication bias for this

4 4 C. Chai-Adisaksopha et al. Included Eligibility Screening Identificatio n Final meta-analytical review inclusion of n = 12 RCTs from prior published manuscript one newly completed RCT (INFORM) Records screened (n = 2232) Articles for full-text review (n = 19) from updated search Final meta-analytical review inclusion n = 14 RCTs Citations from updated search (January 2015 to September 25th 2016) of MEDLINE, EMBASE, CINAHL and CDSRs n = 2220 citations Exclusion on the basis of title and abstract screening and duplicates (n = 2213) n = 18 duplicates n = 2195 irrelevant Patients, Interventions, Comparators, Outcomes (PICOs), non-comparative or did not compare age of blood Exclusion on the basis of fulltext screening (n = 6); ineligible PICOs Fig. 1 Flow diagram of study selection. subgroup due to small number of included studies (less than 10) [43]. Certainty of evidence With respect to ruling out the possibility of benefit from transfusion of fresher red blood cells, the certainty of the evidence was rated as high for overall mortality and in-hospital mortality (Table 2). We judged all domains to have no serious concerns. With respect to ruling out the possibility of harm of fresh red blood cell transfusion, the certainty of the evidence was rated as moderate for overall mortality and in-hospital mortality (Table 2). We judged all domains to have not serious concerns except for imprecision. The confidence intervals of the pooled estimates in the primary analysis included both a potential for very small benefit and appreciable harm of fresher red blood cell transfusion. In a population with a risk of dying of over 10% (the average in the included studies), fresher blood could lead to as many as 13 deaths in 1000 patients (Table 2). We therefore rated down the certainty of evidence for imprecision. Discussion This systematic review/meta-analysis established that fresher red blood cell transfusions do not reduce overall mortality and in-hospital mortality when compared with transfusions of older red blood cells. The certainty of evidence was high for ruling out benefit of fresher blood. The certainty of evidence was only moderate for ruling out the possibility of harm because the upper limits of the 95% confidence intervals suggested that mortality could be increased by as much as 12% with fresher red blood cells (RR: 104, 95% CI: ). Heterogeneity was trivial, and subgroup analyses failed to support any effect modification. Strengths and limitations Strengths of this systematic review/meta-analysis include the comprehensive search including recently published RCTs, duplicate assessment of eligibility and risk of bias, and application of the GRADE approach to rating certainty of evidence. We determined no serious problems in risk of bias, consistency and directness and excluded publication bias based on inspection and statistical tests. Although findings were not definitive with regard to harm of fresher red blood cells, the confidence interval was sufficiently narrow to exclude all but the smallest of benefits. We included both paediatric and adult patients and found results consistent across these groups, testifying to the generalizability of the studies to patients who require a transfusion of red blood cells.

5 Fresher vs older red cell transfusion 5 Table 1 Characteristics of the randomized controlled trials Study Setting Patient population Inclusion criteria Exclusion criteria Intervention (n) Age of blood (days) Fresh RBCs Old RBCs Fresh RBCs Old RBCs Strauss (1996) [25] Single centre, US Paediatric Very low-birthweight infant in NICU (weighing kg) received blood transfusion Exclusion Schulman (2002) [23] Single centre, US NR All patients admitted to trauma centre Fernandes da Cunha (2005) [28] Single centre, Brazil Paediatric Very low-birthweight infant receiving blood transfusion Hebert (2005) [19] Multicentre, Canada Adult Patients requiring RBC transfusion, undergoing cardiac surgery or admitted to ICU Bennett-Guerrero (2009) [21] Single centre, US Adult 18 years old, high morbidity and mortality risk, undergoing CABG or heart valve surgery Aubron (2012) [26] Multicentre, Australia Adult 18 years old, hospitalized in ICU, requiring at least 1 RBC unit Fergusson (2012) [18] Multicentre, Canada [27, 42] Paediatric Premature infant with a birthweight of less than 1250 g, requiring 1 RBC transfusion Heddle (2012) [29] Single centre, Canada Adult 17 years old hospitalized patient who required a blood transfusion Kor (2012) [22] Single centre, US Adult 18 years old, admitted to medical and surgical ICU, endotracheally intubated, mechanically ventilated, required RBC transfusion Unlikely to survive (e.g. severe congenital anomalies) 7 days (21) 42 days (19) NR NR NR <11 days (8) >20 days (9) NR NR rapid infusion of RBC for haemorrhage, required >1 RBC transfusion within 24 h <16 years old, terminal illness, received blood transfusion in 6 months Haemodialysis or peritoneal dialysis within 7 days, history of RBC disease, blood type O negative Palliative care, history of organ transplantation, haematologic disease Already received RBC transfusion, undergoing an exchange transfusion, rare blood types, moribund Had a medical indication for fresh blood, scheduled for autologous donation, expected to receive massive blood transfusion or outpatient Concurrent transfusion with other blood product, emergency transfusion, hemodynamically instable 3 days (26) 28 days (26) 16 (06) 90 (89) < 8 days (26) <42 days (31) Median, 4 Median, days (12) 21 4 days (11) 6 (2) Freshest blood (25) Standard of care (26) 121 (38) 7 days (188) 2 42 days (189) 51 (20) Freshest blood (309) Standard of care (oldest in the inventory) (601) 5 days (50) Standard of care (50) 12 (68) Median (Q1, Q3), 4 (3, 5) 18 (7) 23 (84) 146 (83) 266 (78) Median (Q1, Q3), 265 (21, 36)

6 6 C. Chai-Adisaksopha et al. Table 1 (Continued) Study Setting Patient population Inclusion criteria Exclusion criteria Intervention (n) Age of blood (days) Fresh RBCs Old RBCs Fresh RBCs Old RBCs Dhabangi (2013) [27] Single centre, Uganda Paediatric years old, positive blood smear for malaria, Hb 5 g/dl and blood lactate 5 mmol/l Dhabangi (2015) [42] Single centre, Uganda Paediatric 05 5 years old, Hb 5 g/dl and blood lactate 5 mmol/l Lacroix (2015) [20] Multicentre, Canada, UK, France, the Netherlands and Belgium Adult 18 years old, admitted to ICU, required RBC transfusion within 7 days after admission to ICU or expected to require invasive or non-invasive mechanical ventilation Steiner (2015) [24] Multicentre, US Adult 12 years old, body weight 40 kg, scheduled for complex cardiac surgery with median sternotomy, score 3 on TRUST score, Heddle (2016) [31] Multicentre, Canada, Australia, Israel and US Adult 18 years old, RBC transfusion was requested, admission to hospital Patients with current cardiac disease, undergoing transfusion with other blood products Severe acute malnutrition, cardiac disease, transfusion within 48 h Already received RBC transfusion, elective cardiac surgery, terminal illness, required 1 unit of packed RBC Planned to use autologous transfusion, severe renal dysfunction, use of intraaortic balloon pump, previous RBC transfusion Expected to received massive transfusion, required uncrossmatched blood, required autologous or directed transfusions, had an indication for fresh RBC transfusion 1 10 days (37) days (37) Mean (SD), 78 (18) 1 10 days (145) days (145) <8 days (1211) Standard of care (1219) Median (Q1, Q3), 8 (7, 9) 61 (49) 10 days (538) 21 days (560) 78 (48) Freshest RBCs (6937) Oldest RBCs (13 922) Median (Q1, Q3), 110 (80, 160) 272 (39) Median (Q1, Q3), 32 (30, 34) 22 (84) 283 (67) Median (Q1, Q3), 230 (160, 310) NICU, neonatal intensive care unit; NR, not reported; U, United States; UK, United Kingdom; SD, standard deviation; RBC, red blood cell; ICUl, intensive care unit; CABG, coronary artery bypass graft; Q1, quartile 1; Q3, quartile 3.

7 Fresher vs older red cell transfusion 7 Authors, year RR (95% CI) RR (95% CI) Events, Treatment Events, Control % Weight Aubron et al (0 55, 12 19) 5/25 2/ Bennett-Guerrero et al (0 12, 61 65) 1/12 0/ Dhabangi et al (0 13, 71 34) 1/37 0/ Dhabangi et al (0 45, 4 31) 7/143 5/ Fergusson et al (0 61, 1 54) 30/188 31/ Fernandes da Cunha et al (0 44, 1 85) 9/26 10/ Hebert et al (0 45, 4 98) 5/26 4/ Heddle et al (0 96, 1 15) 634/ / Heddle et al (0 75, 1 65) 35/309 61/ Kor et al (0 47, 1 27) 17/50 22/ Lacroix et al (0 94, 1 17) 448/ / Schulman at al (0 55, 9 17) 4/8 2/ Steiner et al (0 48, 1 41) 23/538 29/ Strauss et al (0 01, 7 02) 0/21 1/ Overall (I-squared = 0 0%, P = 0 901) 1 04 (0 98, 1 12) 1219/ / NOTE: Weights are from random-effects analysis Favours Fresh blood Favours Old blood Fig. 2 Forest plot of overall mortality comparing fresh versus old red blood cell transfusion. Fig. 3 Funnel plot of the overall mortality. Limitations of our review include the failure of included studies to specifically investigate red blood cells stored for days. Goel et al. conducted a retrospective study using hospital billing database for patients discharged from John Hopkins Hospital. Although there was no significant difference among patients who received fresher red blood cells (storage 21 days) and older (either storage 28 days or 35 days), higher mortality was observed in subgroup of patients who were admitted in intensive care unit and who were transfused with older red blood cells (storage 35 days). It is important to note that patients who received older red blood cells in Goel et al. s study were older and had more hypertension, renal disease and anaemia at baseline when compared to those who received fresher red blood cells. Prospective studies on the effect of oldest red blood cells (storage 35 days) on mortality are needed. Secondly, we did not investigate the impact of duration of red blood cell storage on nosocomial infections and other adverse effects due to the lack of reporting across included studies. Furthermore, there was variability of definitions of fresher and older blood, as well as the methods of blood processing and storage solution across the included studies. Relation to prior work Over the past 4 years, several meta-analyses have addressed the mortality impact of fresh versus older red blood cell transfusion [16, 30, 44]. In 2012, Wang et al.

8 8 C. Chai-Adisaksopha et al. Authors, Events, Events, % year RR (95% CI) RR (95% CI) Treatment Control Weight Aubron et al (0 55, 12 19) 5/25 2/ Dhabangi et al (0 13, 71 34) 1/37 0/ Dhabangi et al (0 45, 4 31) 7/145 5/ Fernandes da Cunha et al (0 44, 1 85) 9/26 10/ Heddle et al (0 96, 1 15) 634/ / Heddle et al (0 75, 1 65) 35/309 61/ Overall (I-squared = 0 0%, P = 0 813) 1 06 (0 97, 1 15) 691/ / NOTE: Weights are from random-effects analysis Favours Fresh blood Favours Old blood Fig. 4 Forest plot of in-hospital mortality comparing fresh versus old red blood cell transfusion. [16] published a systematic review and meta-analysis that combined observational studies and randomized controlled trials. The main finding was that transfusion of older blood significantly increased the risk of death (odds ratio: 116, 95% CI: ) when compared to fresher blood. A 2016 update of this meta-analysis analysed the reported data from RCTs and observational studies separately [44]. The 31 observational studies showed an increased risk of death when older blood was transfused (OR: 113, 95% CI: ) (P = 001), whereas the results from six RCTs did not show a difference (OR: 091, 95% CI: ) [44]. The results provide an example of how bias may influence results in observational studies compared to RCTs. In 2016, our group published a systematic review of 12 RCTs comparing fresher versus older blood transfusion.[30] Mortality proved similar in fresher or older red blood cell transfusion, but the confidence interval was considerably wider than in the current review, still including possible benefit from fresher red blood cells (RR: 104, 95% CI: , P = 045, I 2 = 0%). The results of our updated review/meta-analysis are similar, but the narrower confidence interval excludes all but a trivial benefit from fresher red blood cells. Implications Our results essentially exclude a benefit from transfusion of fresher compared with older red blood cells; however, they do not exclude the possibility that transfusion of fresh red blood cells is associated with an increase in mortality. This possibility is relevant for certain populations, such as neonates in the NICU and patients with inherited disorders of haemoglobin production requiring lifelong transfusion support, who often receive fresher blood. Fresh red blood cells can have higher concentrations of extracellular vesicles and cell-free DNA, possibly reflecting differences in the method of red cell collection (apheresis, whole blood) and the method by which whole blood is separated into a fresh red cell product [45]. Extracellular vesicles and cellfree DNA may play a role in inflammation, immune modulation and risk of thrombosis [46]. One large retrospective study has shown an association between the method of processing whole blood donations and in-hospital mortality in transfused adult patients [47]. The American Association of Blood Banks (AABB) has recommended in their revised red cell transfusion guidelines that patients should receive red blood cell units selected at any point within their licensed dating period (standard issue) rather than limiting patients to transfusion of only fresh units [48]. Our results are consistent with this position and provide no support for the selective use of fresh red blood cells. Conclusion This systematic review and meta-analysis included the most recent randomized controlled trials and established that the transfusion of fresher red blood cell does not reduce mortality when compared with older red blood cells.

9 Fresher vs older red cell transfusion 9 Table 2 GRADE evidence profile Quality assessment No of patients Effect No of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Fresher blood Older blood Relative (95% CI) Absolute (95% CI) Mortality outcome 14 RCTs sample: n = events: n 3029 In-hospital mortality outcome 6 RCTs sample: n = events: n 1982 Not serious a Not serious b Not serious c Not serious d None e 1219/9532 (128%) Not serious a Not serious b Not serious c Not serious d None e 691/7479 (92%) 1810/ (107%) 1291/ (87%) RR 104 ( ) RR 106 ( ) Five more deaths per 1000 (2 fewer deaths to 13 more per 1000) Six more deaths per 1000 (2 fewer deaths to 14 more per 1000) CI, confidence interval; RR, risk ratio; RCT, randomized controlled trails. a Studies were at low risk of bias (Fig. S1). b For all outcomes, visual inspection of the forest plot, the test for heterogeneity (non-significant) and the I 2 results of 0% all suggested consistent relative effects. c All studies enrolled relevant representative populations. d Confidence intervals demonstrate adequate precision. e We identified no reason to suspect publication bias or any other biases of material impact on the analysis. f High certainty for benefit and moderate certainty for harm of fresh red blood cells on mortality. Quality Importance High f Critical High f Critical

10 10 C. Chai-Adisaksopha et al. Role of the funding source There was no funding source. CC and PEA had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Contributors CC, NMH, MAC and PEA conceived designed. JWE and GG supervised the study. CC and PEA performed the analysis and drafted the report. GG and JWE supervised data interpretation. All authors critically revised the report. Conflict of interests MAC discloses having sat on advisory boards for Janssen, Leo Pharma, Portola and AKP America. MAC holds the Leo Pharma Chair in Thromboembolism Research at McMaster University. MAC s institution has received funding for research projects from Leo Pharma and Bayer. MAC has received funding for presentations from Leo Pharma, Bayer, Celgene, Shire and CSL Behring. None of the other authors declares any competing interests. Reference 1 Hess JR, Biomedical Excellence For Safer Transfusion C: Scientific problems in the regulation of red blood cell products. Transfusion 2012; 52: Takei T, Amin NA, Schmid G, et al.: Progress in global blood safety for HIV. J Acquir Immune Defic Syndr 2009; 52(Suppl 2):S127 S131 3 Zimring JC: Fresh versus old blood: are there differences and do they matter? Hematology Am Soc Hematol Educ Program 2013; 2013: Stapley R, Owusu BY, Brandon A, et al.: Erythrocyte storage increases rates of NO and nitrite scavenging: implications for transfusion-related toxicity. Biochem J 2012; 446: D Alessandro A, Liumbruno G, Grazzini G, et al.: Red blood cell storage: the story so far. Blood Transfus 2010; 8: Hess JR: Red cell changes during storage. Transfus Apher Sci 2010; 43: Khan SY, Kelher MR, Heal JM, et al.: Soluble CD40 ligand accumulates in stored blood components, primes neutrophils through CD40, and is a potential cofactor in the development of transfusion-related acute lung injury. Blood 2006; 108: Kucukakin B, Kocak V, Lykkesfeldt J, et al.: Storage-induced increase in biomarkers of oxidative stress and inflammation in red blood cell components. Scand J Clin Lab Invest 2011; 71: Pettila V, Westbrook AJ, Nichol AD, et al.: Age of red blood cells and mortality in the critically ill. Crit Care 2011; 15:R Robinson SD, Janssen C, Fretz EB, et al.: Red blood cell storage duration and mortality in patients undergoing percutaneous coronary intervention. Am Heart J 2010; 159: Van. Straten AH, Soliman Hamad MA, Martens EJ, et al.: Effect of storage time of transfused plasma on early and late mortality after coronary artery bypass grafting. J Thorac Cardiovasc Surg 2011; 141: e Weinberg JA, McGwin G Jr, Vandromme MJ, et al.: Duration of red cell storage influences mortality after trauma. J Trauma 2010; 69: ; discussion Edgren G, Kamper-Jorgensen M, Eloranta S, et al.: Duration of red blood cell storage and survival of transfused patients (CME). Transfusion 2010; 50: Eikelboom JW, Cook RJ, Liu Y, et al.: Duration of red cell storage before transfusion and in-hospital mortality. Am Heart J 2010; 159: e1 15 Middelburg RA, van de Watering LM, Briet E, et al.: Storage time of red blood cells and mortality of transfusion recipients. Transfus Med Rev 2013; 27: Wang D, Sun J, Solomon SB, et al.: Transfusion of older stored blood and risk of death: a meta-analysis. Transfusion 2012; 52: van de Watering L, Biomedical Excellence for Safer Transfusion C: Pitfalls in the current published observational literature on the effects of red blood cell storage. Transfusion 2011; 51: Fergusson DA, Hebert P, Hogan DL, et al.: Effect of fresh red blood cell transfusions on clinical outcomes in premature, very low-birth-weight infants: the ARIPI randomized trial. JAMA 2012; 308: Hebert PC, Chin-Yee I, Fergusson D, et al.: A pilot trial evaluating the clinical effects of prolonged storage of red cells. Anesth Analg 2005; 100: , table of contents 20 Lacroix J, Hebert PC, Fergusson DA, et al.: Age of transfused blood in critically ill adults. N Engl J Med 2015; 372: Bennett-Guerrero E, Stafford-Smith M, Waweru PM, et al.: A prospective, double-blind, randomized clinical feasibility trial of controlling the storage age of red blood cells for transfusion in cardiac surgical patients. Transfusion 2009; 49: Kor DJ, Kashyap R, Weiskopf RB, et al.: Fresh red blood cell transfusion and short-term pulmonary, immunologic, and coagulation status: a randomized clinical trial. Am J Respir Crit Care Med 2012; 185: Schulman CI, Nathe K, Brown M, et al.: Impact of age of transfused blood in the trauma patient. J Trauma 2002; 52: Steiner ME, Ness PM, Assmann SF, et al.: Effects of red-cell storage duration on patients undergoing cardiac surgery. N Engl J Med 2015; 372:

11 Fresher vs older red cell transfusion Strauss RG, Burmeister LF, Johnson K, et al.: AS-1 red cells for neonatal transfusions: a randomized trial assessing donor exposure and safety. Transfusion 1996; 36: Aubron C, Syres G, Nichol A, et al.: A pilot feasibility trial of allocation of freshest available red blood cells versus standard care in critically ill patients. Transfusion 2012; 52: Dhabangi A, Mworozi E, Lubega IR, et al.: The effect of blood storage age on treatment of lactic acidosis by transfusion in children with severe malarial anaemia: a pilot, randomized, controlled trial. Malar J 2013; 12:55 28 Fernandes da Cunha DH, Nunes Dos Santos AM, Kopelman BI, et al.: Transfusions of CPDA-1 red blood cells stored for up to 28 days decrease donor exposures in very low-birthweight premature infants. Transfus Med 2005; 15: Heddle NM, Cook RJ, Arnold DM, et al.: The effect of blood storage duration on in-hospital mortality: a randomized controlled pilot feasibility trial. Transfusion 2012; 52: Alexander PE, Barty R, Fei Y, et al.: Transfusion of fresher vs older red blood cells in hospitalized patients: a systematic review and meta-analysis. Blood 2016; 127: Heddle NM, Cook RJ, Arnold DM, et al.: Effect of short-term vs. longterm blood storage on mortality after transfusion. N Engl J Med 2016; 375: Liberati A, Altman DG, Tetzlaff J, et al.: The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ 2009; 339:b Higgins JP, Altman DG, Gotzsche PC, et al.: The Cochrane Collaboration s tool for assessing risk of bias in randomised trials. BMJ 2011; 343:d Higgins JP, Altman DG: Assessing risk of bias in included studies; in Higgins JP, Green S (eds): Cochrane Handbook for Systematic Reviews of Interventions. Chichester, The Cochrane Collaboration, 2008: Guyatt G, Busse J: Methods commentary: risk of bias in randomized trials methodological-resources/risk-of-biascommentary/ [Last accessed 1 October 2016] 36 Deeks JJ, Altman DG, Bradburn MJ: Statistical methods for examining heterogeneity and combining results from several studies in meta-analysis; in Egger M, Smith GD, Altman DG (eds): Systematic Reviews in Health Care: meta-analysis in context. London, BMJ Publishing Group, 2001: Deeks JJ, Higgins JP, Altman DG, et al.: Analysing data and undertaking meta-analysis; in Higgins JP (ed.): Cochrane Handbook for systematic Reviews of Interventions. Chichester, The Cochrane Collaboration 2008: Sterne JA, Egger M: Funnel plots for detecting bias in meta-analysis: guidelines on choice of axis. J Clin Epidemiol 2001; 54: Egger M, Davey Smith G, Schneider M, et al.: Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315: Peters JL, Sutton AJ, Jones DR, et al.: Comparison of two methods to detect publication bias in meta-analysis. JAMA 2006; 295: Guyatt GH, Oxman AD, Vist GE, et al.: GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ 2008; 336: Dhabangi A, Ainomugisha B, Cserti- Gazdewich C, et al.: Effect of transfusion of red blood cells with longer vs shorter storage duration on elevated blood lactate levels in children with severe anemia: the TOTAL randomized clinical trial. JAMA 2015; 314: Sterne JA, Egger M, Moher D, et al.: Addressing Reporting biases; in Higgins JP (ed.): Cochrane Handbook for systematic Reviews of Interventions. Chichester, The Cochrane Collaboration, 2008: Remy KE, Sun J, Wang D, et al.: Transfusion of recently donated (fresh) red blood cells (RBCs) does not improve survival in comparison with current practice, while safety of the oldest stored units is yet to be established: a meta-analysis. Vox Sang 2016; 111: Acker JP, Hansen AL, Kurach JD, et al.: A quality monitoring program for red blood cell components: in vitro quality indicators before and after implementation of semiautomated processing. Transfusion 2014; 54: Radwanski K, Cognasse F, Garraud O, et al.: Comparison of apheresis and 24 h RT held red cell concentrates by measurement of storage lesion parameters and neutrophil activating factors during 42-day storage. Transfus Apher Sci 2013; 48: Heddle NM, Arnold DM, Acker JP, et al.: Red blood cell processing methods and in-hospital mortality: a transfusion registry cohort study. Lancet Haematol 2016; 3:e246 e Carson JL, Guyatt G, Heddle NM, et al.: Clinical practice guidelines from the AABB: red blood cell transfusion thresholds and storage. JAMA 2016; 316: Supporting information Additional Supporting Information may be found online in the supporting information tab for this article: Fig. S1 Summary of risk of bias assessment. Fig. S2 Forest plot of overall mortality comparing fresh versus old red blood cells transfusion, subgroup analysis according to duration of red blood cells storage. Fig. S3 Forest plot of overall mortality comparing fresh versus old red blood cells transfusion, subgroup analysis according to age of participants.