| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
* From the Clinical Epidemiology Program, Ottawa Health Research Institute; and
The University of Ottawa Centre for Transfusion Research, Ottawa, Ontario, Canada.
Address correspondence to: Dr. Dean Fergusson, University of Ottawa Centre for Transfusion Research, Ottawa Hospital, 501 Smyth Road, Box 201, Ottawa, Ontario KIH 8L6, Canada. E-mail: dafergusson{at}ohri.ca
 |
Abstract |
|---|
 TOP Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Methods: A systematic search of the scientific literature was conducted. The pooled relative risk ratio (RR) of developing an adverse postoperative outcome with either leukoreduced or non-leukoreduced blood was calculated using a random effects model. To better estimate the efficacy of leukoreduction, a second analysis of transfused patients only was conducted.
Results: Ten RCTs met inclusion criteria and eight provided separate data for patients randomized and transfused. The mean percentage of patients randomized but not transfused was 34%. For postoperative infection, the overall pooled RR was 0.76 [(95% confidence interval (CI): 0.54–1.08] for the "all patients randomized" analysis. For the "only patients transfused" analysis, the pooled RR became clinically and statistically significant (RR = 0.60 (95% CI: 0.38–0.93). For mortality, the pooled RR for the "all patients randomized" analysis was 0.71 (95% CI: 0.45–1.13) and 0.61 (95% CI: 0.36–1.04) for the "only patients transfused" analysis. When analyzing either all patients randomized or all patients transfused, there was no statistically significant difference in cancer recurrence rates (one study only).
Conclusion: We demonstrated that patients who were transfused leukoreduced red blood cells might benefit from a decrease in postoperative infections. A decrease in mortality may have been realized if more patients had been enrolled in the various randomized trials. Including all patients randomized, regardless of whether or not they were actually transfused diluted the observed clinical benefit of leukoreduction.
|
Introduction |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
However, two systematic reviews concluded that the effectiveness of leukoreduction in preventing postoperative infections, cancer recurrence and postoperative mortality in patients undergoing curative cancer surgery remained largely unproven.3,4 There were a number of methodological concerns identified in the meta-analyses and in subsequent narrative reviews.5,6 Both meta-analyses analyzed all patients randomized regardless of whether they received an intervention. This is commonly referred to as an intention-to-treat analysis. While considered the most conservative analytical approach, it may not reflect the true efficacy of leukoreduction. This is of significant concern in these particular trials as many investigators opted to prematurely randomize patients. As a consequence, a significant proportion of patients never received a blood transfusion, either leukoreduced or non-leukoreduced, and in some cases did not undergo surgery. When the lack of exposure to interventions is not considered in the planning of the study, then the relative risk ratio (RR) of adverse postoperative outcomes does not reflect the efficacy of treatment but rather an inadequate attempt to assess its effectiveness.7 Efficacy is a measure of the benefit resulting from a treatment evaluated under ideal conditions whereas effectiveness refers to whether the observed benefit transfers well to the real-world population. Given that many countries have implemented universal leukoreduction and countries such as the United States and Japan are contemplating such a decision, an accurate reflection of both its efficacy and effectiveness has substantial ramifications.
In this updated meta-analysis, we therefore chose to present an analysis of only transfused patients, which best reflects the evidence of efficacy of leukoreduction in decreasing postoperative infections, cancer recurrence and death while also presenting an effectiveness analysis that included all randomized patients.
|
Methods |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
All citations were manually reviewed and included if they consisted of a randomized, controlled clinical trial comparing leukoreduced allogeneic red blood cell (RBC) transfusions to non-leukoreduced allogeneic RBC transfusions in patients undergoing surgery. Full journal articles, letters, abstracts, and monographs were all included. At least one study arm of the trial must have consisted of patients randomized to receive leukoreduced allogeneic RBCs with the control arm consisting of patients receiving standard non-leukofiltrated allogeneic RBCs. Patients randomized to receive only autologous blood in either study arm were excluded. The outcomes evaluated in each study were clinical evidence of postoperative infection, cancer recurrence, and death. Studies that evaluated non-clinical outcome measures such as laboratory markers without clinical assessment were also excluded. The bibliographies of the selected citations and all systematic and narrative review articles were examined to identify any relevant references not identified in the original search.
Data abstraction
Three investigators (M.P.K., D.F., P.C.H.) abstracted the following information on standardized data abstraction forms: study design, randomization and blinding protocols, study population characteristics, sample size, withdrawals, type of filter, patients transfused, postoperative infections, cancer recurrence, deaths, amount of RBC units transfused, and mean preoperative and postoperative hemoglobin concentrations. Any interobserver discrepancies were resolved by consensus.
Data analysis
The RR of developing an adverse postoperative outcome with either leukoreduced or non-leukoreduced RBC transfusions was calculated for each study. Where clinically appropriate, data from studies were combined to estimate the pooled RR and 95% confidence intervals (CIs) using a DerSimonian and Laird random effects model.8
Two different analyses were performed: one evaluating all patients randomized to receive either leukoreduced or non-leukoreduced blood and the other evaluating only patients that received a transfusion. The second analysis was conducted to better estimate the true effect of leukoreduction. As the hypothesized benefits of leukoreduction can only be realized in patients that receive a transfusion, it seemed reasonable to exclude patients from analysis that do not undergo surgery or do not receive a transfusion. Excluding these patients from analysis is predicated on the assumption that the decision not to transfuse or operate was independent of treatment arm allocation. A summary RR was calculated for each analysis. A RR less than one indicates that more adverse events occurred in patients that received non-leukoreduced than leukoreduced blood and vice-versa for a RR greater than one. The same analytic methods were performed for the sub-group analyses of type of filter and type of surgery.
|
Results |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Eleven randomized controlled trials were identified that compared leukoreduced allogeneic RBC transfusions to non-leukoreduced allogeneic RBC transfusions with respect to the risks of postoperative infection, cancer recurrence, or mortality in patients undergoing surgery.9–19 One study evaluated patients undergoing surgery for burn trauma.9 It was excluded a posteriori as it was determined to be a distinct patient group from the other surgical populations. Thus, ten studies remained for analysis10–19 (Table
). Five studies provided postoperative infection data for all patients randomized and all patients transfused10–13,18 while two studies provided data only for patients transfused.14,15 Four studies provided mortality data for all patients randomized as well as all patients transfused.11–13,18 Three studies presented data only for patients randomized16,17,19 and one study presented mortality data only for patients transfused.10 In the eight trials that provided separate data for patients randomized and transfused, the mean percentage of patients randomized but not transfused was 34% (range 2%–73%).
|
Postoperative infection
In the "all patients randomized" analysis, a total of 3073 patients were analyzed (Figure 1). The overall pooled RR was 0.76 (95% CI: 0.54–1.08). In the subgroup analyses, the use of a bedside filter was found to be both clinically and statistically significant (RR = 0.38, 95% CI: 0.26–0.54). The pre- and post-storage filters were found to be less effective and did not reach statistical significance (RR = 0.91, 95% CI: 0.71–1.17 and RR = 0.82, 95% CI: 0.63–1.06 respectively). No difference in effects was seen between colorectal and cardiac surgery, however, the effect in the cardiac subgroup was found to be statistically significant (RR = 0.82, 95% CI: 0.63–1.00).
|
).
Cancer recurrence
One randomized controlled trial evaluated cancer recurrence in patients who received standard allogeneic RBCs vs leukoreduced blood.10 Supplemental information was provided in a separate publication of the same study subjects.20 When analyzing all patients randomized, there was no statistically significant difference in cancer recurrence rates between the patient group that received leukoreduced RBCs and the group that received standard buffy-coat depleted packed RBCs at the five-year follow-up point (27.9% vs 27.8%, P > 0.05.) Likewise, when analyzing only the patients who were transfused, there was no statistically significant difference in recurrence rates between the two groups (28.4% vs 31.2%, P > 0.05.)
Mortality
Perioperative mortality data were provided in eight clinical trials.10–13,16–19 Operative mortality was defined as death within 30 days of surgery in three studies,10,11,13 within 60 days for one study,12 within three months for one study,18 and one study defined in-hospital mortality up to a specific study cut-off date, thus, patient follow-up time was dependent upon date of trial entry.19 Two studies did not report a definition of operative mortality.16,17 For mortality, the pooled RR for all patients randomized was 0.71 (95% CI: 0.45–1.13). The results were similar for the "only patients transfused" analysis (RR = 0.61 (95% CI: 0.36–1.04). Mortality decreased with pre- and post-storage filters but not with bedside filters in the "only patients transfused" group (Figure 2).
|
|
Discussion |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The protective effects of leukoreduced blood were most evident in the cardiac surgery subgroup, where a decreased risk of infection and mortality was seen in both the "only patients transfused" and "all patients randomized" groups. The reason for increased effectiveness in cardiac surgery may be related to the higher volume of RBCs transfused per patient compared to colorectal and gastrointestinal surgery as well as the fact that the leukocytes are transfused to an already activated inflammatory system caused by cardiopulmonary bypass. By transfusing leukocytes to an activated inflammatory response system, the release of cytokines and free radicals may induce organ damage and, consequently, death. Thus, in addition to an immunosuppressive effect of non-leukoreduced RBCs there is a potential pro-inflammatory effect. Given the many limitations of the reported trials and this analysis, caution would suggest that further studies in a variety of patient populations would ensure that the results were accurate and generalizable.
It is interesting to note that although the bedside filter decreased the risk of infection, it also appeared to be associated with an increased number of deaths compared to the pre- and post-storage filters. Since the results regarding the bedside filter originated from a single study, it is plausible that the trend towards increased mortality or decreased infection was related to the small sample size rather than a true harmful effect in the case of mortality or a beneficial effect with respect to postoperative infection.11 The effectiveness of bedside filters compared to pre-storage filters has not been evaluated in head-to-head clinical trials. Animal studies have shown that pre-storage filtration is superior to bedside filters for the prevention of tumour growth and alloimmunization.21,22 In addition, laboratory evidence shows that pre-storage leukoreduction removes leukocytes before they breakdown and release potentially harmful substances such as cytokines.23,24 Thus, it is reasonable to postulate that pre-storage filters should confer a greater clinical benefit than bedside filters.
Only one study examined cancer recurrence in patients who received either leukoreduced or standard RBCs. No difference in recurrence rates was noted in either group, although more studies are required before making a definitive statement regarding this issue.
With premature randomization employed in the included trials, any beneficial effect of leukoreduction was diluted by the number of patients who were not transfused. Given this study design limitation, the most accurate measure of the intervention’s efficacy would require limiting the analysis to patients that actually received a transfusion. However, the primary concern in counting events only in patients who were transfused would be a potential bias introduced due to the un-blinded design of the studies. A biased result would occur if clinicians consider the type of product that will be administered when deciding to transfuse a given patient. For instance, transfusing leukoreduced products to patients not requiring transfusion could result in a greater beneficial effect, as "healthier" patients would be included in the leukoreduction study arm. Our report was not able to detect such subtle but plausible biases. Although highly unlikely, this should be considered a limitation of this meta-analysis. Furthermore, although both randomization groups were more or less matched for demographic characteristics in all the trials, such information was not available for the transfused groups in all trials. However, of the eight trials that provided separate data on all patients randomized and transfused, four provided baseline differences in the "transfused only" groups.11,13–15 For the four trials that did not provide baseline data for those transfused, this information is largely irrelevant for two as the proportion not transfused was very low.12,19 In the four trials that provided data, there were no large differences between treatment arms in the vast majority of baseline measurements. This supports our belief that patients were properly randomized to each arm of the trial and were not differentially excluded from receiving a transfusion based on treatment allocation.
The other limitations of this meta-analysis arise from limitations of the individual studies including an inconsistency in definition of outcome measures among the reported clinical trials such as what constitutes a postoperative infection and the time frame outcomes are reported. Inappropriate cross-over as well as exclusion of patients when the wrong blood was received was also identified as a concern in two studies.12,13 Because this limitation in study analysis was attributable only to 23 patients and unlikely to affect the aggregate results of this meta-analysis, we treated the patients as randomized to the blood products they received.
Large double-blinded clinical trials randomizing patients at the point of transfusion would best address the issue of whether leukoreduced blood transfusions are superior to standard, non-filtered blood. Ironically, in countries that adopted a universal leukoreduction program, randomized trials are no longer feasible. In such countries, effectiveness is best determined by conducting large before and after studies.25–27
|
Footnotes |
|---|
Accepted for publication October 16, 2003. Revision accepted February 13, 2004.
|
References |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
2 Miller JP, Mintz PD. The use of leukocyte-reduced blood components. Hematol Oncol Clin North Am 1995; 9: 69–90.[Medline]
3 McAlister FA, Clark HD, Wells PS, Laupacis A. Perioperative allogeneic blood transfusion does not cause adverse sequelae in patients with cancer: a meta-analysis of unconfounded studies. Br J Surg 1998; 85: 171–8.[Medline]
4 Vamvakas E. Transfusion-associated cancer recurrence and postoperative infection: meta-analysis of randomized, controlled clinical trials. Transfusion 1996; 36: 175–86.[Medline]
5 Vamvakas EC, Blajchman MA. Universal WBC reduction: the case for and against. Transfusion 2001; 41: 691–712.[Medline]
6 Vamvakas EC, Blajchman MA. Deleterious clinical effects of transfusion-associated immunomodulation: fact or fiction? Blood 2001; 97: 1180–95.
7 Fergusson D, Aaron SD, Guyatt G, Hebert P. Post-randomisation exclusions: the intention to treat principle and excluding patients from analysis. BMJ 2002; 325: 652–4.
8 Lau J. Meta-Analyst977. Boston, MA: New England Medical Center; 1995.
9 Nielsen HJ, Hammer JH, Krarup AL, et al. Prestorage leukocyte filtration may reduce leukocyte-derived bioactive substance accumulation in patients operated for burn trauma. Burns 1999; 25: 162–70.[Medline]
10 Houbiers JG, Brand A, van de Watering LM, et al. Randomised controlled trial comparing transfusion of leucocyte-depleted or buffy-coat-depleted blood in surgery for colorectal cancer. Lancet 1994; 344: 573–8.[Medline]
11 Jensen LS, Kissmeyer-Nielsen P, Wolff B, Qvist N. Randomised comparison of leucocyte-depleted versus buffy-coat-poor blood transfusion and complications after colorectal surgery. Lancet 1996; 348: 841–5.[Medline]
12 van de Watering LM, Hermans J, Houbiers JG, et al. Beneficial effects of leukocyte depletion of transfused blood on postoperative complications in patients undergoing cardiac surgery. A randomized clinical trial. Circulation 1998; 97: 562–8.
13 Titlestad IL, Ebbesen LS, Ainsworth AP, Lillevang ST, Ivist N, Georgsen J. Leukocyte-depletion of blood components does not significantly reduce the risk of infectious complications. Results of a double-blinded, randomized study. Int J Colorectal Dis 2001; 16: 147–53.[Medline]
14 Jensen LS, Andersen AJ, Christiansen PM, et al. Postoperative infection and natural killer cell function following blood transfusion in patients undergoing elective colorectal surgery. Br J Surg 1992; 79: 513–6.[Medline]
15 Tartter PI, Mohandas K, Azar P, Endres J, Kaplan J, Spivack M. Randomized trial comparing packed red cell blood transfusion with and without leukocyte depletion for gastrointestinal surgery. Am J Surg 1998; 176: 462–6.[Medline]
16 van Hilten JA, Brand A; TACTICS Reasearch Group. A multi-center prospective randomized trial of buffy coat depleted- and leukocyte filtered erythrocyte transfusions in vascular- and gastrointestinal oncologic surgery. Vox Sang 2002; 83(Suppl 1): 453–6.
17 Bracey AW, Radovancevic R, Nussmeier NA, et al. Leukocyte-reduced blood in open heart surgery patients: effects on outcome. Transfusion 2002; 42(Suppl): 5S (abstract).
18 Wallis JP, Chapman CE, Orr KE, Clark SC, Forty JR. Effect of WBC reduction of transfused RBCs on postoperative infection rates in cardiac surgery. Transfusion 2002; 42: 1127–34.[Medline]
19 Dzik WH, Anderson JK, O’Neill EM, Assmann SF, Kalish LA, Stowell CP. A prospective, randomized clinical trial of universal WBC reduction. Transfusion 2002; 42: 1114–22.[Medline]
20 van de Watering LM, Brand A, Houbiers JG, et al. Perioperative blood transfusions, with or without allogeneic leucocytes, relate to survival, not to cancer recurrence. Br J Surg 2001; 88: 267–72.[Medline]
21 Bordin JO, Bardossy L, Blajchman MA. Growth enhancement of established tumors by allogeneic blood transfusion in experimental animals and its amelioration by leukodepletion: the importance of the timing of the leukodepletion. Blood 1994; 84: 344–8.
22 Blajchman MA, Bardossy L, Carmen RA, Goldman M, Heddle NM, Singal DP. An animal model of allogeneic donor platelet refractoriness: the effect of the time of leukodepletion. Blood 1992; 79: 1371–5.
23 Shanwell A, Kristiansson M, Remberger M, Ringden O. Generation of cytokines in red cell concentrates during storage is prevented by prestorage white cell reduction. Transfusion 1997; 37: 678–84.[Medline]
24 Kristiansson M, Soop M, Shanwell A, Sundqvist KG. Prestorage versus bedside white blood cell filtration of red blood cell concentrates. Effects on the content of cytokines and soluble tumor necrosis factor receptors. J Trauma 1996; 40: 379–83.[Medline]
25 Hébert PC, Fergusson D, Blajchman MA, et al. Clinical outcomes following institution of the Canadian universal leukoreduction program for red blood cell transfusions. JAMA 2003; 289: 1941–9.
26 Fergusson D, Hebert PC, Lee SK, et al. Clinical outcomes following institution of universal leukoreduction of blood transfusions for premature infants. JAMA 2003; 289: 1950–6.
27 Baron JF, Gourdin M, Bertrand M, et al. The effect of universal leukodepletion of packed red blood cells on postoperative infections in high-risk patients undergoing abdominal aortic surgery. Anesth Analg 2002; 94: 529–37.
This article has been cited by other articles:
 |
|
 |
|
  R. S. Weber, N. Jabbour, and R. C. G. Martin II Anemia and Transfusions in Patients Undergoing Surgery for Cancer Ann. Surg. Oncol., January 1, 2008; 15(1): 34 - 45. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. T. Tinmouth MD MSc, L. A. McIntyre MD MSc, and R. A. Fowler MDCM MS Blood conservation strategies to reduce the need for red blood cell transfusion in critically ill patients Can. Med. Assoc. J., January 1, 2008; 178(1): 49 - 57. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. J. Murphy and G. D. Angelini Indications for Blood Transfusion in Cardiac Surgery Ann. Thorac. Surg., December 1, 2006; 82(6): 2323 - 2334. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. C. Frake, H. E. Smith, L.-F. Chen, and W. L. Biffl Prestorage Leukoreduction Prevents Accumulation of Matrix Metalloproteinase 9 in Stored Blood Arch Surg, April 1, 2006; 141(4): 396 - 400. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Couture, A.R. Turner, B. Melosky, L. Xiu, R.K. Plante, C.Y. Lau, and I. Quirt Prior Red Blood Cell Transfusions in Cancer Patients Increase the Risk of Subsequent Transfusions With or Without Recombinant Human Erythropoietin Management Oncologist, January 1, 2005; 10(1): 63 - 71. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. L. Carson and J. A. Berlin Will we ever know if leukoreduction of red blood cells should be performed?/Saurons-nous un jour s'il faut proceder a la reduction leucocytaire des culots globulaires ? Can J Anesth, May 1, 2004; 51(5): 407 - 410. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
