Deleterious clinical effects of transfusion-associated immunomodulation: fact or fiction?

doi:10.1182/blood.V97.5.11802001 97: 1180-1195

Eleftherios C. Vamvakas and Morris A. Blajchman immunomodulation: fact or fiction?Deleterious clinical effects of transfusion-associated

http://bloodjournal.hematologylibrary.org/content/97/5/1180.full.htmlUpdated information and services can be found at:

(238 articles)Transfusion Medicine � (449 articles)Review Articles � (5022 articles)Immunobiology �

(3722 articles)Clinical Trials and Observations �Articles on similar topics can be found in the following Blood collections

http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#repub_requestsInformation about reproducing this article in parts or in its entirety may be found online at:

http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#reprintsInformation about ordering reprints may be found online at:

http://bloodjournal.hematologylibrary.org/site/subscriptions/index.xhtmlInformation about subscriptions and ASH membership may be found online at:

Copyright 2011 by The American Society of Hematology; all rights reserved.Washington DC 20036.by the American Society of Hematology, 2021 L St, NW, Suite 900, Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly

For personal use only. by guest on June 12, 2013. bloodjournal.hematologylibrary.orgFrom

http://bloodjournal.hematologylibrary.org/content/97/5/1180.full.html

http://bloodjournal.hematologylibrary.org/cgi/collection/clinical_trials_and_observations

http://bloodjournal.hematologylibrary.org/cgi/collection/immunobiology

http://bloodjournal.hematologylibrary.org/cgi/collection/review_articles

http://bloodjournal.hematologylibrary.org/cgi/collection/transfusion_medicine

http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#repub_requests

http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#reprints

http://bloodjournal.hematologylibrary.org/site/subscriptions/index.xhtml

http://bloodjournal.hematologylibrary.org/subscriptions/ToS.dtl

http://bloodjournal.hematologylibrary.org/


Review article

Deleterious clinical effects of transfusion-associated immunomodulation:fact or fiction?Eleftherios C. Vamvakas and Morris A. Blajchman

Allogeneic blood transfusion results in the infusion into the recipient oflarge amounts of foreign antigens in both soluble and cell-associatedforms. The persistence of these antigens in thecirculation of therecipient may create conditions that allow the development ofimmune down-regulation. Evidence from a variety of sourcesindicates that allogeneic blood transfusion enhances the survival ofrenal allografts1 and may increase the recurrence rate of resectedmalignancies2 and the incidence of postoperative bacterial infec-tions,3-7 as well as reduce the recurrence rate of Crohn disease8

and/or activate infections with cytomegalovirus9 or human immu-nodeficiency virus.10 This clinical syndrome, the mechanisms ofwhich remain to be defined, has been referred to in the transfusionmedicine literature as allogeneic blood transfusion–associatedimmunomodulation (TRIM).11

Clinical evidence for the existence of TRIM has been availablesince 1973. In their seminal study, Opelz et al1 provided evidence,counterintuitive at the time, that recipients of allogeneic bloodtransfusion had improved renal allograft survival. Subsequentclinical studies and studies in experimental animals corroboratedthe results of Opelz et al,1 and allogeneic blood transfusions wereused deliberately in the early 1980s to prevent rejection of renalallografts.12 The beneficial effect of TRIM was obscured followingthe introduction of immunosuppression with cyclosporine, but itwas recently reported to persist even in renal allograft recipientsreceiving modern immunosuppressive therapy.13

On the basis of the immunomodulatory effect of allogeneicblood transfusion in renal allograft recipients, Gantt14 raised thequestion in 1981 whether the TRIM effect might also be associatedwith an increased risk of cancer recurrence in patients undergoingresection of a malignancy. Gantt’s hypothesis was based on thepremise that, if allogeneic blood transfusion down-regulated thehost’s immune surveillance mechanism that targets malignant cells,the receipt of allogeneic blood transfusion could enhance tumorgrowth. A subsequent hypothesis proposed that, if allogeneic bloodtransfusion causes immunosuppression, then recipients of perioper-ative allogeneic blood transfusion could be at increased risk forpostoperative bacterial infection.

Since 1981, more than 150 clinical studies have examined theassociation of perioperative allogeneic blood transfusion withcancer recurrence and/or postoperative bacterial infection. Most ofthese are observational cohort studies comparing patients who hador did not have transfusion.15-20 In addition, 7 randomized con-trolled trials (RCTs) have compared the risk of cancer recur-rence2,21,22 and/or postoperative infection3-7,21-23 between a treat-ment arm receiving standard7 or buffy-coat–reduced red blood cells(RBCs)2,4-6,21,22 or whole blood3 and a control arm receivingautologous or white blood cell (WBC)-reduced RBCs or whole

blood.24,25 These studies are based on the assumption that thetransfusion of autologous2,4,21,23or WBC-reduced3,5-7,22 RBCs, orwhole blood, is immunologically neutral. Both the earlier observa-tional cohort studies and the recent RCTs have produced contradic-tory findings, and—because of the discrepancies among thepublished studies—the long-standing hypothesis of the potentiallydeleterious immunomodulatory effect of perioperative allogeneicblood transfusion remains unresolved.11,26,27

The mechanism(s) of the TRIM effect(s) also remains elusive,and it is possible that a large number of biologic mechanisms mayunderlie the effect(s).28-31The infusion of foreign antigen in either asoluble31-36 or cell-associated37-43 form has been shown to induceimmune suppression, anergy, and clonal deletion in studies inexperimental animals. However, most studies evaluating proposedmechanisms have been done in rodents, and these findings may notbe applicable to the human immune system.44 Caution should beexercised, therefore, when findings from experimental animals areextrapolated to humans. Moreover, different biologic mechanismsmay be involved in each reported clinical manifestation of TRIM,1-10

and the clinical evidence supporting each of the aforementionedhypotheses1-10 should be examined on its own merits.

The specific constituent(s) of allogeneic blood that mediates theTRIM effect1-10 remains unclear. Allogeneic plasma,31-36allogeneicWBCs,30,37-43and substances that accumulate in blood componentsduring storage39 have been implicated in the pathogenesis ofTRIM. However, both the animal and human data suggest that theTRIM effects are most likely mediated by transfused allogeneicWBCs.45,46 Lee et al47 reported persistence of donor WBCs inhumans for up to 1.5 years after an allogeneic blood transfusion. Inmurine and rabbit experimental models, Blajchman et al37,38 andBordin et al39 demonstrated a tumor growth–promoting effect ofallogeneic blood transfusion that appeared to be associated withtransfusion of allogeneic WBCs. The findings of these experimentssupport the hypothesis that allogeneic WBCs actively induceimmune suppression in transfusion recipients. In another relevantstudy, Kao40 induced immune suppression in mice receiving donorWBCs free of plasma and platelets. These and several otherinvestigators41,42 also attributed an induction of TH2 cells intransfusion recipients to the allogeneic donor WBCs, showing thatTH2 cells can produce immunosuppression in transfusion recipientsby down-regulating the activity of TH1 cells. Mincheff et al43

implicated the antigen-presenting cells of the allogeneic donor inthe induction of a state of anergy in the recipient and proposed that,during refrigeration, antigen-presenting cells lose their ability todeliver costimulation. These investigators hypothesized that afteran allogeneic transfusion, the recipient’s T cells are stimulated by

From the Department of Pathology, New York University Medical Center, NewYork, NY; the Department of Pathology and Molecular Medicine, McMasterUniversity, London, ON, Canada; and Hamilton Blood Center, Canadian BloodServices, Hamilton, ON, Canada.

Submitted July 3, 2000; accepted November 2, 2000.

Reprints: Eleftherios C. Vamvakas, Blood Bank, RRG 17, New York Uni-versity Medical Center, 400 East 34th St, New York, NY 10016; e-mail:[email protected].

© 2001 by The American Society of Hematology

1180 BLOOD, 1 MARCH 2001 z VOLUME 97, NUMBER 5




allogeneic donor WBCs in the absence of costimulation and thatthis interaction induces a state of anergy in the recipient’s T cells.

More than 100 observational cohort studies comparing patientswith and without transfusion, as well as 2 RCTs comparingrecipients of allogeneic and autologous RBCs,2,4,21,23 did notassociate the TRIM effect with any particular blood constituent.Five recent RCTs3,5-7,22were designed on the basis of the assump-tion that the allogeneic transfusion effect responsible for cancerrecurrence and/or postoperative infection is mediated specificallyby allogeneic WBCs. These studies produced contradictory find-ings, and the issue as to whether any deleterious TRIM effects arelessened or abrogated by WBC reduction of the transfusedallogeneic cellular blood components remains unresolved.11,26,27,48

Jensen et al attributed an 86%3 and a 71%5 reduction in theincidence of postoperative infection to the use of WBC-reducedallogeneic whole blood3 or RBCs,5 with the use of poststorageWBC reduction by filtration. In contrast, Houbiers et al22 found nodifference in the incidence of postoperative infection between therecipients of WBC-reduced and buffy-coat–reduced RBCs with theuse of prestorage WBC reduction by filtration. In the RCT by vande Watering et al,6 the use of WBC-reduced RBCs decreased theincidence of postoperative infection by 30% compared with thatseen in patients having transfusion with buffy-coat–reduced RBCs.Moreover, there was no difference in the incidence of postoperativeinfection between the recipients of RBCs that were WBC reducedby filtration before or after storage.6

Recently, the United Kingdom, Ireland, and Portugal imple-mented universal WBC reduction of all transfused cellular bloodcomponents, based on the hypothesis that this practice wouldprevent the theoretical risk of transmission by transfusion of theagent of variant Creutzfeldt-Jakob disease (vCJD).49,50France andCanada also implemented universal WBC reduction, primarily toenhance overall transfusion safety.50 Following these develop-ments, public debate began regarding the appropriateness ofintroducing universal WBC reduction in the United States andelsewhere. The risk of transmission of vCJD by transfusion isconsidered only a theoretical possibility because an epidemic ofbovine spongiform encephalopathy and cases of vCJD haveoccurred almost exclusively in the United Kingdom. Therefore, if adecision were made to implement universal WBC reduction in theUnited States and elsewhere, such a policy would probably not beintroduced with an intent to prevent the transmission of vCJD, butto reap other potential benefits of WBC reduction. Most of thepublished evidence regarding such potential benefits pertains to theprevention of the deleterious TRIM effects, and the public debateregarding the implementation of universal WBC reduction is likelyto focus mainly (although not exclusively) on an examination of theefficacy of WBC reduction in preventing these potential adverseeffects of allogeneic blood transfusion.27,51 Furthermore, the re-cently described relation between WBC-containing allogeneicblood transfusion and increased postoperative mortality fromcauses other than postoperative infection,6 and the report by Hebertet al52 that a restrictive strategy of RBC transfusion may be superiorto a liberal transfusion strategy when mortality is evaluated as anoutcome in critically ill patients, may affect the debate overwhether to implement universal WBC reduction.

Thirty-one physicians from academic blood banks in the UnitedStates recently voiced their strong opposition to the intent of theFood and Drug Administration (FDA) to mandate the implementa-tion of universal WBC reduction in the United States.53 In a letter tothe editor ofTransfusion,53 they stated that “It is our view thatpublished reports fail to document a substantial health benefit thatwould serve to justify WBC reduction of cellular blood compo-

nents transfused toall patients. Accordingly, we feel that thecurrently available evidence regarding the deleterious effects ofallogeneic blood transfusion is not sufficiently compelling towarrant universal WBC reduction for the prevention of theseeffects.” This present review examines the available evidencefrom: (1) the observational cohort studies that investigated thehypothesis that allogeneic blood transfusion provokes cancerrecurrence and/or postoperative infection; (2) the findings of theRCTs that examined the possible deleterious TRIM effects associ-ated with allogeneic blood transfusion in general or with transfu-sion of allogeneic WBCs in particular; and (3) the results of therecent RCTs that reported an association between allogeneic bloodtransfusion and increased mortality.6,52To help readers process andunderstand the contradictory information, we include a section onthe rationale, design, and analysis of the clinical studies thatinvestigated these possible deleterious TRIM effects. Also, afterreviewing the available evidence regarding the association ofallogeneic transfusion with cancer recurrence or postoperativeinfection, we discuss the various arguments for and against apotential decision to implement universal WBC reduction of alltransfused cellular blood components to prevent the deleteriouseffects of allogeneic transfusion-associated immunomodulation.

Design of clinical studies

Evidence in support of deleterious TRIM effects has been obtainedfrom observational cohort studies, RCTs, meta-analyses of theobservational cohort studies, and meta-analyses of the RCTs.54 Theobservational studies were either retrospective or prospective, andcompared the risk of an adverse clinical outcome between cohortsof patients who did or did not receive transfusion and who differedwith respect to baseline determinants of the need for transfusionand the risk of cancer recurrence or postoperative infection. Thesestudies relied on multivariate statistical analyses to adjust for theeffects of possible confounding factors. The RCTs were prospec-tive clinical experiments comparing the risk of an adverse outcomebetween patients randomly allocated to receive different bloodproducts, thus relying on randomization (ie, chance) to distributeall possible confounding factors equally between the treatment andcontrol arms. The RCTs presented univariate (ie, unadjusted)analyses calculating the odds ratio (OR) of an adverse outcome in atreatment arm compared with that in a control arm, as shown inTable 1. Such analyses are valid if the counts in each of the 4 cellsof the 23 2 contingency table (Table 1) are free of the effects ofconfounding factors—a condition usually satisfied when largenumbers of patients are randomized. Meta-analyses of RCTsintegrate the ORs from individual RCTs into a summary OR, whichis also free of the effects of confounders and represents a measureof the “average” TRIM effect across the combined RCTs. Meta-analyses of observational studies extract 23 2 contingency tablecounts from each study, use these 4 counts to calculate an OR foreach corresponding original report, and then also integrate the ORsfrom the individual studies into a summary OR. Unlike thesummary ORs calculated by combining the results of RCTs, thesummary ORs reported from meta-analyses of observational stud-ies incorporate the effects of all known and unknown confoundersof the association of transfusion with cancer recurrence and/orpostoperative infection. Therefore, the summary ORs calculated bymeta-analyses of the observational studies of the adverse TRIMeffects should not be interpreted as a measure of the “average”TRIM effect across the combined studies. In only one of thesemeta-analyses18 was meta-regression used to adjust for the effects

TRANSFUSION IMMUNOMODULATION: FACT OR FICTION? 1181BLOOD, 1 MARCH 2001 z VOLUME 97, NUMBER 5




of 3 confounders on which information had been reported by mostof the published observational studies.

When the results of studies retrieved for a meta-analysis arediscrepant, or if the variation among the reported findings is greaterthan can be attributed to chance (a situation called “heterogeneityof effects”), a meta-analysis can offer insight into the reasons forthe disagreements among the available studies. Provided that theresults from the available studies are concordant, meta-analysis canbe used to derive, through the application of a number ofquantitative techniques, a measure of the effect of allogeneic bloodtransfusion based on a combination of the available data. This“average” TRIM effect is more precise (and more likely to attainstatistical significance) than the TRIM effects reported fromindividual studies because it is based on a much larger sample (ie,the study population accrued when the study populations from allavailable reports are combined). Before the results of studies areintegrated, however, the degree of agreement among reports mustbe assessed both conceptually and statistically. The Q test statisticquantifies the probability that the variation in reported results mighthave arisen by chance. Most analysts hold that there is sufficientagreement (or “homogeneity”) among studies to permit the under-taking of a meta-analytic synthesis of the results ifP . .05 for theQ test statistic (ie, if there is a greater than 5% probability that thevariation in reported results might have arisen by chance).55 In thisreview, we use meta-analysis both to investigate reasons fordisagreements among studies (ifP , .05 for the Q test statistic)and to integrate the results of the available studies for thecalculation of an “average” TRIM effect across combined investi-gations (ifP . .05 for the Q test statistic). Where meta-analysis isused in this review, results of homogeneous studies are integratedusing the random-effects method of DerSimonian and Laird.56

Patients allocated to the treatment arm of the reviewedRCTs3-7,22,23 were exposed to all the constituents of allogeneicblood and were at risk for developing adverse TRIM effects.Control-arm patients were presumed to be at reduced risk becausethey were not exposed to allogeneic WBCs3,5-7,22 or allogeneicWBCs and allogeneic plasma.4,23 Five4-6,22,23 of 6 RCTs3-6,22,23

conducted in Europe transfused buffy-coat–reduced RBCs to thetreatment arm. In one RCT,3 whole blood was given, and in oneRCT7 conducted in the United States, standard RBCs (ie, not buffycoat reduced) were transfused. The buffy-coat reduction method isused widely in western Europe, but not elsewhere, for the

preparation of blood components from whole blood. Componentsproduced by the buffy-coat reduction method contain 60% to 80%fewer WBCs compared with cellular blood components producedby North American methods.57 Thus, the buffy-coat–reduced RBCstransfused by Houbiers et al22 and van de Watering et al6 containeda mean of 0.83 109 WBCs per RBC unit; those administered byJensen et al5 contained a mean of 1.23 109 WBCs per unit. Thenumber of allogeneic WBCs contained in the average RBCtransfusion dose given in 4 of the European RCTs4,5,22,23(eg, in amedian dose of 3 U of buffy-coat–reduced RBCs administered inthe study of Houbiers et al22) was equivalent to the number ofallogeneic WBCs contained in only 1 U of allogeneic RBCsproduced in North America. Therefore, if the TRIM effect ismediated by allogeneic WBCs, the WBC dose given to patients inthe treatment arm of most European RCTs4,5,22,23may have beeninsufficient to provoke clinically significant TRIM effects. In astudy of New Zealand White rabbits with established tumors,Blajchman30 reported a significant (P, .0001) decrease in thetumor growth–promoting effect of allogeneic blood transfusion inassociation with buffy-coat reduction of whole blood (Table 2). Theameliorative effect of buffy-coat reduction was not complete,however, and the median number of pulmonary nodules seen inrabbits receiving buffy-coat–reduced whole blood was significantly(P 5 .0004) greater than that seen in control animals receiving notransfusion or in animals receiving WBC-reduced blood.30

All 5 RCTs administering WBC-reduced RBCs5-7,22 or whole

Table 2. Effect of buffy-coat reduction in decreasing allogeneic bloodtransfusion–induced tumor growth in New Zealand White rabbitswith established tumors*

Red cell transfusion receivedSample

sizeNumber of pulmonary tumor

nodules (median 6 SE)

Allogeneic, unmodified (stored) 20 74.0 6 8.7

Allogeneic, buffy coat reduced

before storage† 19 34.0 6 7.0

Allogeneic, WBC reduced before

storage‡ 20 23.5 6 1.8

Control group (no transfusion) 16 21.5 6 2.5

WBC indicates white blood cell.*Modified from Blajchman.30

†Mean decrease in WBC dose of 76% (similar to the buffy-coat reduction methodused in the RCTs4-6,22,23 reported from western Europe).

‡Mean decrease in WBC dose of 99.7%.

Table 1. Calculation of the odds ratio for studies investigating the relationship between allogeneic blood transfusion andcancer recurrence or postoperative infection

Group

Adverse outcome: cancer recurrence orpostoperative infection

TotalsOccurring Not occurring

Exposed to allogeneic transfusion* a b a 1 b

Not exposed to allogeneic transfusion† c d c 1 d

Totals a 1 c b 1 d n‡

Odds of an adverse outcome in those exposed to allogeneic transfusion: [a / (a 1 b)] / [b / (a 1 b)] 5 a / b

Odds of an adverse outcome in those not exposed to allogeneic transfusion: [c / (c 1 d)] / [d / (c 1 d)] 5 c / d

Odds ratio 5Odds of an adverse outcome in those exposed to allogeneic transfusion

Odds of an adverse outcome in those not exposed to allogeneic transfusion5

a

b

c

d

5ad

bc

*Treatment arm (patients prospectively randomized to receive transfusion of allogeneic WBCs) in RCTs, or treatment group (patients having transfusion) in observationalcohort studies.

†Control arm (patients prospectively randomized to receive transfusion of autologous or WBC-reduced allogeneic RBCs) in RCTs, or control group (patients not havingtransfusion) in observational cohort studies.

‡n 5 a 1 b 1 c 1 d; a, b, c, and d are the 4 counts of a 2 3 2 contingency table.54

1182 VAMVAKAS and BLAJCHMAN BLOOD, 1 MARCH 2001 z VOLUME 97, NUMBER 5




blood3 to the control arm transfused RBCs prepared by leukocytefiltration. Prestorage-filtered RBCs were prepared by passing,within 246 or 4822 hours of collection, a buffy-coat–reduced RBCunit through a leukocyte-reduction filter. Poststorage-filtered RBCswere prepared by passing a unit of buffy-coat–reduced RBCs5,6 orwhole blood,3 stored for 6 or more days, through a leukocyte-reduction filter. The average WBC content of transfused WBC-reduced RBCs was less than 53 106 per unit, and was often as lowas 13 106 per unit. The timing of the filtration procedure isimportant because biologic response modifiers released from WBCgranules in a time-dependent manner during storage may mediatethe adverse TRIM effects,39,48,56-60 and—if such TRIM effectsreally exist—only WBC-reduced RBCs filtered before storage canbe expected to abrogate the adverse effects. Two RCTs adminis-tered autologous RBCs4 or whole blood23 to the control arm.Autologous blood can be expected to abrogate the TRIM effectsmediated by allogeneic plasma,31-36 or by intact, immunologicallycompetent allogeneic WBCs, but not the TRIM effects mediatedby biologic response modifiers released from WBC granulesduring storage.59

Observational studies

Observational studies have compared the incidence of cancerrecurrence, death due to cancer recurrence, and/or overall mortal-ity15-18 between patients undergoing cancer resection who did ordid not receive transfusion; or the incidence of postoperativebacterial infection with or without transfusion in patients undergo-ing gastrointestinal surgery, orthopedic operations, cardiac surgery,or various other procedures. These studies tended to indicate thatpatients having transfusion (compared with those not havingtransfusion) had a higher incidence of cancer recurrence or deathdue to cancer recurrence as well as a shorter overall survival after acancer resection operation; and almost always had a higherincidence of postoperative bacterial infection.16 These studies alsoindicated that patients having transfusion generally differed fromthose without transfusion in several prognostic factors, includingclinical stage of the malignancy; size, histologic grade, and type oftumor; age; preoperative hemoglobin; duration and extent ofsurgery; amount of perioperative blood loss; the frequency ofchronic systemic illness, such as congestive heart disease, lungdisease, liver disease, kidney failure, or diabetes mellitus; and thepresence of risk factors for postoperative urinary tract infection(UTI), pneumonia, or wound infection.18,19,61

These 2 sets of observations have led to different interpretationsof the results of the studies that compared patients who did or didnot receive transfusion. Some authors concluded that perioperativeallogeneic blood transfusion had a direct deleterious effect on therecipient, causing an increase in the incidence of cancer recurrenceand/or postoperative bacterial infection.16 Other investigators con-cluded that allogeneic blood transfusions were a surrogate markerfor a variety of adverse prognostic factors and that the othervariables that generated the need for perioperative transfusionsdetermined the subsequent clinical outcome.19 The latter school ofthought reasoned that the adverse prognostic factors that wereassociated with both the need for perioperative transfusion and thecancer recurrence and/or postoperative bacterial infection wereconfounders of the relation between transfusion and these out-comes, and engendered a spurious association between allogeneictransfusion and cancer recurrence and/or postoperative infection.19

In many of the reported observational studies, the authors usedmultivariate regression analysis to adjust the effect of transfusion

for the effects of confounding factors. Regression analysis cancalculate an allogeneic blood transfusion effect that is independentof the effects of known and measurable confounders, provided that:(1) the investigators measure all known potentially confoundingfactors; (2) the variables that are associated with both transfusionand cancer recurrence and/or postoperative infection in a particularset of data are identified as true confounders; and (3) all trueconfounders are included in the final regression model on which theconclusions are based.19 In most published observational studies,however, important potential confounding factors were not consid-ered by the investigators61; and the multivariate regression modelon which the conclusions were based was built by a stepwise (asopposed to forced-entry) method, which did not ensure theinclusion of several measured confounders in the final model.19 Forexample, although postoperative UTI was the most frequent type ofpostoperative infection in most observational studies that reportedan association between transfusion and postoperative infection, thenumber of patient days with an indwelling urinary catheter—thepredominant risk factor for nosocomial UTI in the United States—was not considered as a potential confounder and was not measuredin most studies.61

As a result, TRIM effects reported as “independent” by manyteams of investigators were not truly free of the effects of knownand measurable confounding factors, and the calculated TRIMeffect decreased in magnitude (or became statistically insignificant)depending on which confounding factors were included in the finalregression model.61 These caveats notwithstanding, allogeneicblood transfusion often emerged as the leading predictor of cancerrecurrence and/or postoperative bacterial infection from multivari-ate analyses, and was reported to increase the risk of postoperativeinfection in patients having transfusion (compared with those nothaving transfusion) by up to 10-fold.16

Cancer recurrence

Chung et al,62 Vamvakas,17 and Brand and Houbiers18 usedmeta-analysis to integrate the findings from the univariate (ie,unadjusted; Table 1) analyses of the observational studies of theassociation of perioperative allogeneic blood transfusion withcancer recurrence, death due to cancer recurrence, or overallmortality. The most recent meta-analysis18 also included studies notpublished in English and considered 7 cancer sites; 100 observa-tional studies were identified as eligible for analysis. There wasagreement among the 3 meta-analyses with regard to the magnitudeand statistical significance of the calculated summary ORs of therisk of an adverse clinical outcome in patients having transfusion(compared with those not having transfusion). Figure 1 shows theunadjusted summary OR of an adverse outcome for 7 cancer sites.As shown in Figure 1, when the unadjusted results were integratedfrom 28 reports on colorectal cancer, 14 reports on head and neckcancer, 10 reports on breast cancer, 8 reports on gastric cancer, 8reports on lung cancer, 6 reports on cervical cancer, and 6 reportson prostate cancer, the summary OR of an adverse clinical outcomewith transfusion (versus without transfusion) was statisticallysignificant (P# .05) for all sites except the cervix (P. .05). Theseunadjusted results contrast with the conclusions of most of thestudies that presented a multivariate analysis. A statisticallysignificant (P# .05) TRIM effect adjusted for the effects ofconfounding factors was reported from only 11 studies of colorec-tal cancer,17 4 studies of head and neck cancer, 1 study of breastcancer, 2 studies of gastric cancer, 4 studies of lung cancer, and 2studies of prostate cancer.18

On the basis of such a statistical combination of the unadjustedfindings of the retrieved observational studies, Chung et al62





concluded that allogeneic blood transfusion given for a cancerresection operation had an adverse effect on subsequent prognosis;Vamvakas17 and Brand and Houbiers18 reached the oppositeconclusion. In the case of colorectal cancer studies, there wasmarked variation among the findings of the available reports, andthe probability that such variation might have arisen by chance wassmaller than 1 per 1000 (P , .001 for the Q test statistic). Asalready discussed, this situation precludes the calculation of asummary estimate of the allogeneic transfusion effect across theavailable studies by the techniques of meta-analysis.55

Brand and Houbiers18 separately analyzed the studies thatreported on colorectal cancer recurrence (27 studies), death due tocolorectal cancer recurrence (15 studies), or overall mortalityfollowing colorectal cancer resection (30 studies). When theunadjusted findings from these reports were integrated, there wasstill unexplained heterogeneity (P, .001 for the Q test statistic) inall 3 analyses. Meta-regression was used to adjust the summary ORof cancer recurrence for the effects of tumor location and clinical(Dukes) stage of tumor; the summary OR of death due to cancerrecurrence for the effect of tumor location; and the summary OR ofdeath from any cause for the effects of tumor location, clinical(Dukes) stage of tumor, and mean patient age. After adjustment forthe effects of these confounding factors, no association was foundbetween perioperative transfusion and cancer recurrence (P 5 .55)or death due to cancer recurrence (P5 .19), but the relationshipbetween perioperative transfusion and overall mortality persisted(P , .001). Brand and Houbiers18 therefore reasoned that, becausemeta-regression could attribute the association of transfusion withcancer recurrence (or death due to cancer recurrence) to the effectsof confounding factors, and because only a minority of thepublished observational studies had reported a significant adverseTRIM effect based on multivariate analyses, a link betweenperioperative transfusion and colorectal cancer recurrence couldnot be established. These investigators concluded that the observedeffect of allogeneic transfusion on overall mortality might be due to

an association of perioperative transfusion with causes of deathother than cancer recurrence.18

It is also possible to ascribe the observed effect of allogeneictransfusion on mortality to a greater physiologic severity of illnessin patients having transfusion than in those not having transfusion.Most observational studies presented data on the severity of diseasein the included subjects; that is, they reported the clinical (Dukes)stage of tumor. In contrast, severity of illness incorporates both theseverity of disease and the prevalence and severity of comorbidi-ties, such as diabetes mellitus, congestive heart disease, lungdisease, liver disease, or kidney failure. Such chronic systemicillnesses may have as much of an impact on overall mortality as theclinical (Dukes) stage of tumor, but the available observationalstudies did not present data on chronic systemic illness, making itimpossible to adjust for the effect of illness severity in themeta-analysis. It is thus possible that the patients having periopera-tive transfusion survived for a shorter period after the cancerresection operation simply because they were sicker at the time ofsurgery than the patients who did not require transfusion.

The findings of the available observational studies on theassociation of perioperative transfusion with cancer recurrencemay no longer be relevant to the management of contemporarycancer patients. Many of these studies (eg, all of the studies thatreported on patients with gastric cancer) examined only overallsurvival; others did not report a multivariate analysis; still othersincluded patients operated on before 1980 and over an extendedperiod. For example, only 3 of the 14 studies of head and neckcancer retrieved for the meta-analysis of Brand and Houbiers18

fulfilled the selection criteria of reporting on cancer-specificcomplications rather than overall mortality, adjusting for the effectsof confounding factors by multivariate analysis, and includingpatients operated on after 1980 and over a period of less than 5years. Only 1 of these 3 studies reported a statistically significantadverse effect of allogeneic blood transfusion.18

Furthermore, the findings of the available observational studiesof the association of perioperative transfusion with cancer recur-rence may reflect the effect of publication bias,63-65a considerationthat may also apply to the discussion of the reported observationalstudies of the relationship between perioperative transfusion andpostoperative infection. Studies reporting null results are less likelyto be published than studies reporting statistically significantfindings. Easterbrook et al63 documented a statistically significant3.8-fold increase in the odds of publication for observationalstudies reporting significant findings, compared with studies withnull results. Multivariate analysis showed that the better odds ofpublication could not be explained by the quality of the studydesign. On the contrary, there was a trend toward a greater numberof significant results with poorer quality studies.63

Rather than comparing patients with or without transfusion,Ness et al66 prospectively compared the recurrence rate of prostatecancer in 309 recipients of autologous or allogeneic blood transfu-sion, and observed no clinical benefit from the use of autologousblood. The hazard ratio of allogeneic transfusion in a univariateCox proportional hazard model of time to tumor recurrence was0.87 (P. .05), and Ness et al66 commented that a beneficial TRIMeffect could not be excluded. It is noteworthy that allogeneic bloodtransfusion is not associated with an adverse prognosis in patientswith cervical cancer, even when all available univariate (unad-justed) results are combined (Figure 1). Because human papilloma-virus (HPV) is implicated in the pathogenesis of cervical cancer,cytotoxic T cells directed against HPV antigens expressed on thesurface of tumor cells could contribute to immunologic control ofthe growth of residual tumor cells following cancer resection.

Figure 1. Summary odds ratio of an adverse clinical outcome (ie, cancerrecurrence, death due to cancer recurrence, or overall mortality) acrosspublished observational studies comparing patients having or not havingtransfusion. The univariate (unadjusted) results of the available studies areintegrated separately for each cancer site. The data on colorectal and prostate cancer(representing 28 and 6 studies, respectively) are shown as reported by Vamvakas.17

The data on head and neck (14 studies), breast (10 studies), gastric (8 studies), lung(8 studies), and cervical cancer (6 studies) are shown as reported by Brand andHoubiers.18 The references to the included primary studies can be found in thereports of these 2 meta-analyses.17,18 Each summary OR is surrounded by its 95%CI. A summary OR of the null value (1) indicates that the risk of an adverse clinicaloutcome is the same, on average, with or without transfusion across the combinedstudies. When the 95% CI of the summary OR extends on both sides of the null value,the calculated summary TRIM effect is not statistically significant.





Accordingly, if allogeneic blood transfusion induced immunesuppression, it could down-regulate these cytotoxic T-cell re-sponses, promoting tumor cell growth. Transfusion would thus beexpected to have a larger adverse effect in patients with cervicalcancer than in patients with cancer from other sites (Figure 1)where the tumors are not virus induced and are probably subject toweaker immunologic control mechanisms. However, it has beenobserved that early in the development of cervical intraepithelialneoplasia, the expression of class I HLA molecules on malignantcells is specifically down-regulated. HPV-specific, HLA-restrictedcytotoxic T cells (whose function may be suppressed by allogeneicblood transfusion) may not be effective against cervical cancercells because of the reduction in class I HLA molecule expressionon the tumor cells.67

Postoperative infection

Except for a handful of studies,61,68-70observational investigationsof the association of allogeneic blood transfusion with postopera-tive bacterial infection almost uniformly have reported a relation-ship between perioperative transfusion and infection, which per-sisted after statistical adjustment for the effects of the confoundingfactors considered by the authors.16,19Postoperative infection oftendevelops, however, because of a higher physiologic severity ofillness and a higher prevalence of risk factors for postoperativeinfection at specific sites in patients having transfusion comparedwith those not having transfusion. As already discussed in thecontext of cancer recurrence, in most published observationalstudies of transfusion and postoperative infection, the reportedallogeneic transfusion effect was adjusted for the effect of severityof disease (ie, the severity of the principal diagnosis), but theseverity of the principal diagnosis differs from the severity of apatient’s overall illness. In a study that compares the frequency ofpostoperative infection with or without transfusion in patientsundergoing colorectal cancer resection, the presence and severityof comorbidities (eg, diabetes mellitus, congestive heart disease,lung disease, liver disease, or kidney failure) may be a moreimportant determinant of postoperative infection than the severityof the principal diagnosis according to the clinical (Dukes) stage oftumor. Furthermore, the number of patient days with an indwellingurinary catheter may be the most important determinant ofpostoperative UTI; the number of days of endotracheal intubationand/or impaired consciousness may be a cardinal determinant ofpostoperative pneumonia; and so on.61

Until recently, observational studies reporting an associationbetween allogeneic blood transfusion and postoperative infectiondid not adjust for the effects of severity of illness and/or risk factorsfor postoperative infection at specific sites. Some teams of investi-gators secured partial control for the effects of these variables byexcluding UTIs from the definition of postoperative infection69,71; bylimiting the outcome variable to postoperative wound infection72-74; orby adjusting for the effects of serum albumin,75,76 insertion of aurinary catheter,77,78 or presence of chronic systemic illness,70 ordiabetes mellitus.79 However, adjustment for the effects of all thesefactors in combination has rarely been presented in the literature.

In a study of 492 patients undergoing colorectal cancer resec-tion, Vamvakas et al61 calculated the probability of infection inassociation with perioperative blood transfusion with and withoutadjustment for the effects of chronic systemic illness, number ofdays with an indwelling urinary catheter, endotracheal intubation,and impaired consciousness. In an analysis that adjusted only forthe effects of 18 confounding variables considered by previousauthors and that adjusted for insertion of a urinary catheter (asopposed to number of patient days with an indwelling urinary

catheter), these investigators61 detected a highly significant(P , .001) transfusion effect on the risk of postoperative infectionat any site. However, when adjustment was also made for theeffects of the aforementioned variables, the association betweentransfusion and postoperative infection at any site disappeared(P 5 .407); the only significant predictors of postoperative infec-tion were the number of patient days with an indwelling urinarycatheter, the presence of chronic systemic illness, the number ofdays of impaired consciousness, and the duration of anesthesia.Except for the duration of anesthesia, the 17 other confoundingvariables considered by previous authors proved to be insignificantpredictors of postoperative infection in this analysis.

Observational studies of the association of allogeneic bloodtransfusion with postoperative infection in patients having orthope-dic surgery secured partial adjustment for the effect of greaterillness severity in the transfusion group by comparing subjects whoreceived autologous or allogeneic blood transfusion. At least tosome extent, patients who made preoperative autologous blooddonations did so because they were in better health than subjectswho did not make autologous donations. Also, autologous bloodmay have been transfused more liberally than allogeneic blood,after less surgical blood loss, and it is difficult to attribute anyincrease in the infection rate seen in the allogeneic (compared withthe autologous) transfusion group to the allogeneic transfusion perse if adjustment is not made for the effects of chronic systemicillness and risk factors for postoperative infection at specific sites.Duffy and Neal20 conducted a meta-analysis of the univariate (ie,unadjusted) results of 5 observational studies68,70,80-82and 2 RCTs4,83

comparing the postoperative infection rates of patients havingtransfusion with similar volumes of autologous or allogeneicblood. Patients receiving no transfusion or a mixture of autologousand allogeneic RBCs in these studies4,68,70,80-83were excluded fromthe meta-analysis. Because it is uncommon for more than 3predonated units of autologous blood to be available, subjectsreceiving 4 or more units of allogeneic RBCs were also excluded.The studies of Busch et al21,23 and Triulzi et al84 were excludedbecause of insufficient published information. The summary ORacross the 7 studies was 2.4 (95% confidence interval [CI], 1.6-3.6;P , .0001). Figure 2 updates the meta-analysis of Duffy and Neal20

by including the study of Innerhofer et al.78 The 8 available studieswere homogeneous (P5 .50 for the Q test statistic), and thesummary OR of postoperative infection in the allogeneic (com-pared with the autologous) transfusion group was 2.1 (95% CI,1.3-3.4;P , .001).

Recently, Carson et al71 conducted a retrospective cohort studyof 9598 consecutive patients with hip fracture who underwentsurgical repair between 1983 and 1993 at 20 hospitals across theUnited States. The primary outcome variable was serious bacterialinfection, defined as bacteremia, pneumonia, deep wound infec-tion, or septic arthritis/osteomyelitis. Information was collected onnumerous variables, including comorbid conditions such as thedeterminants of the Charlson Comorbidity Index, but the methodused for building the statistical models was not described. Theadjusted relative risk of serious postoperative infection withtransfusion (versus without transfusion) was 1.43 (95% CI, 1.16-1.78;P 5 .001). Chang et al74 analyzed a database of 1349 patientsundergoing elective colorectal surgery for any disease of the colonor rectum at 11 university hospitals across Canada. To better adjustfor the effects of factors confounding the association of transfusionwith postoperative infection, these investigators limited the out-come variable to postoperative wound infection. Ten prognosticvariables were found to be associated with both transfusion andpostoperative wound infection, and the final regression model





adjusted for 4 of these identified confounders. Allogeneic bloodtransfusion was reported to be a significant independent predictorof postoperative wound infection (OR5 1.18; 95% CI, 1.05-1.33;P 5 .007). Vamvakas and Carven60 reported a retrospective cohortstudy of 416 consecutive patients admitted to one hospital forcoronary artery bypass graft (CABG) operations. The outcomevariable was limited to postoperative wound infection or pneumo-nia, and adjustment was made for the effects of chronic systemicillness and specific risk factors for wound infection or pneumonia.Statistical models were built by the forced-entry method, and theadjusted risk of postoperative wound infection or pneumoniaincreased by 6% per unit of allogeneic RBCs and/or plateletstransfused (P5 .0284), or by 43% for a patient receiving the meantransfusion dose of 7.2 U of RBCs and/or platelets.

Randomized controlled trials

Cancer recurrence

The 3 RCTs that compared the incidence of cancer recurrencebetween recipients of buffy-coat–reduced allogeneic RBCs andrecipients of autologous whole blood21 or RBCs2 or WBC-reduced,buffy-coat–reduced allogeneic RBCs22 were medically and statisti-cally homogeneous. All 3 studies enrolled patients undergoingcolorectal cancer resection. The proportion of patients havingtransfusion varied from 58%21 to 64%22 among the studies, and theproportion of patients developing recurrent cancer varied from23%2 to 25.5%.22 There was also agreement among the findings ofthese RCTs (Figure 3), as the noted variation in reported results wassufficiently modest to be attributed to chance (P. .10 for the Qtest statistic).55 Accordingly, the findings of the studies werecombined in 2 meta-analyses,24,25 and the summary OR of cancerrecurrence in the allogeneic transfusion (compared with thecontrol) group across the 3 studies was 1.04 (95% CI, 0.81-1.35;

P . .05) (Figure 3). The summary OR of death due to cancerrecurrence was 0.98 (95% CI, 0.76-1.26;P . .05).24

Busch et al21 reported on 423 of 510 randomized patients;Houbiers et al22 reported on 697 of 1021 randomized subjects; andHeiss et al2 reported on 100 of 120 randomized patients. In additionto the patients who did not have transfusion (42%, 36%, and 40%of the study samples, respectively), 28%, 10%, and 33%, respec-tively, of the subjects randomized to receive autologous or WBC-reduced allogeneic RBCs also received buffy-coat–reduced alloge-neic RBCs. Many violations of the experimental protocol occurredin the studies of Busch et al21 and Heiss et al2 because of a designproblem that needs to be discussed. Patients randomly allocated tothe autologous transfusion arm in these studies donated 2 U ofwhole blood before surgery, and, if they needed transfusion of morethan 2 U of RBCs perioperatively, they were given buffy-coat–reduced RBCs. In addition, because of the preoperative autologousblood donations, these patients presented to the operating roomwith a lower hematocrit than patients from the allogeneic transfu-sion arm. Therefore, autologous transfusion recipients probablyreceived transfusions sooner, after less surgical blood loss, than didpatients allocated to the allogeneic transfusion arm. It is thuspossible that patients from the autologous arm having transfusionwith a particular number of RBC units may have had less invasivesurgery than patients from the allogeneic transfusion arm given thesame number of RBCs; this design problem may have led tooverestimation of any adverse effect of allogeneic blood transfu-sion, as discussed by Heiss et al.2,4

It is impossible, for ethical reasons, to perform RCTs in whichpatients are randomly allocated not to receive blood transfusion orto always receive transfusion; only patients prospectively random-ized to receive different blood products—should the need fortransfusion arise perioperatively—can be compared in RCTs.Ideally, the blood components administered to patients enrolled inRCTs should differ in only one factor, which reflects a biologicmechanism presumed to underlie the immunosuppressive effect ofallogeneic blood transfusion. If the TRIM effect is assumed to bemediated by allogeneic WBCs, recipients of standard or WBC-reduced RBCs should be compared. However, only one RCT22

investigating the association between allogeneic transfusion and

Figure 2. Summary odds ratio of postoperative bacterial infection derived fromthe univariate (unadjusted) results of observational 68,70,78,80,82 and experimen-tal 4,83 clinical studies comparing recipients of similar volumes of allogeneic orautologous blood. 20 For each study, the figure shows the OR of postoperativebacterial infection in recipients of allogeneic (compared with autologous) blood. Thedata from the experimental studies4,83 depicted here represent observational compari-sons restricted to patients receiving similar volumes of allogeneic or autologous blood(see text). Each OR is surrounded by its 95% CI. If the 95% CI of the OR includes thenull value of 1, the TRIM effect is not statistically significant. The data are plotted on alogarithmic scale extending from 0.1 to 1000.

Figure 3. Summary odds ratio of cancer recurrence 24 derived from randomizedcontrolled trials investigating the association of perioperative allogeneic bloodtransfusion with cancer recurrence in patients undergoing elective colorectalcancer resection. 2,21,22 For each RCT, the figure shows the OR of cancer recurrencein recipients of buffy-coat–reduced allogeneic RBCs as compared with recipients ofautologous whole blood21 or RBCs2 or recipients of WBC-reduced allogeneicRBCs.22 Each OR is surrounded by its 95% CI. If the 95% CI of the OR includes thenull value of 1, the TRIM effect is not statistically significant. The data are plotted on alogarithmic scale extending from 0.1 to 10.





cancer recurrence has transfused WBC-reduced RBCs to patientsfrom the control arm.

Because of the high frequency of patients who did not havetransfusion and the protocol violations in the 3 RCTs investigatingthe association of allogeneic transfusion with cancer recur-rence,2,21,22 a deleterious TRIM effect on colorectal cancer recur-rence could be evaluated in an effective sample of only 696 patientsacross the 3 studies.18 Therefore, the meta-analysis of the 3 RCTs24

did not have adequate statistical power to rule out the possibility ofan adverse TRIM effect smaller than a 33% increase in the risk ofcancer recurrence among the recipients of buffy-coat–reducedallogeneic RBCs, compared with the recipients of autologous orWBC-reduced allogeneic RBCs. Moreover, as already discussed, ifallogeneic WBCs are assumed to mediate the deleterious TRIMeffect(s), then the problem of limited statistical power is com-pounded by the problem of limited exposure to allogeneic WBCs inthese RCTs in which buffy-coat–reduced RBCs were transfused tothe treatment arm.2,21,22 The question that is pertinent to clinicaltransfusion practice in North America—that is, whether transfusionof standard allogeneic RBCs increases the risk of cancer recur-rence—has not yet been addressed by RCTs.

Furthermore, the setting of colorectal cancer resection may beinappropriate for the detection of a deleterious TRIM effect oncancer recurrence.Allogeneic transfusion–associated immunomodu-lation can be expected to increase the recurrence rate of a resectedmalignancy if the growth of residual cancer cells is indeedcontrolled by immunologic mechanisms. The existence of aspecific immune response to colorectal cancer cells has not beenproven.18 Although it is possible to generate cytotoxic T cells invitro that recognize antigens expressed by colorectal cancer cells,the relevance of these cytotoxic cells in tumor growth may belimited because of loss of the expression of HLA molecules andadhesion molecules on colorectal cancer cells.85,86 If further RCTsare to be conducted, it may be preferable to concentrate on tumorsknown or presumed to be virus-induced and/or tumors that occurwith increased frequency in patients receiving high doses ofimmunosuppressive drugs for the prevention of organ allograftrejection (ie, skin cancers, lymphoma, cervical carcinoma, andKaposi sarcoma, as well as vulvar, perineal, and renal tumors).87

The reports of the 3 RCTs2,21,22 also presented observationalcomparisons of the incidence of cancer recurrence between patientshaving or not having transfusion in each study. The results of theseanalyses were conflicting.18 In the study of Houbiers et al,22 bloodtransfusion was not associated with cancer recurrence, and therewas a statistically significant association between blood transfusionand mortality from causes other than cancer. Busch et al21 observeda 10% increase in the risk of cancer recurrence with transfusion(versus without transfusion) that could be attributed to local tumorrecurrences (as opposed to distant metastases); the authors ascribedthis increase in risk to the factors that generate a need fortransfusion, as opposed to a TRIM effect.88 Heiss et al2 observed anincreased risk of cancer recurrence in recipients of allogeneic (orautologous and allogeneic) RBCs compared with patients nothaving transfusion or recipients of autologous RBCs. However,their study sample was small (n5 100), and these 4 groups ofpatients were unbalanced with regard to the levels of potentialconfounding factors.

Postoperative infection

The 7 RCTs that compared the incidence of postoperative infectionbetween recipients of buffy-coat–reduced4-6,22,23or standard7 allo-geneic RBCs or whole blood3 and recipients of autologous orWBC-reduced, buffy-coat–reduced allogeneic RBCs or whole

blood were medically and statistically heterogeneous (Table 3).When all 7 studies were considered together, the probability thatthe disagreements among the findings of these studies might havearisen by chance was smaller than 1 per 10 000 (P , .0001 for theQ test statistic) (Figure 4). Two studies3,5 reported a significant(P , .05) TRIM effect, 2 studies4,6 reported a marginally signifi-cant (P, .10) effect, and 3 studies7,22,23 did not detect an effect.More important, the variation in reported results ranged from a7.3-fold increase in the risk of infection3 (Figure 4) to no TRIMeffect.7,22,23Table 4 stratifies the 7 RCTs according to various studycharacteristics and then combines the results of RCTs that shareeach particular design attribute. The purpose of this meta-analysisis to examine whether specific differences in the design of thestudies may have been responsible for the disagreements amongthe results. A study design attribute is considered to account for thedisagreements among the studies if the integrated results fromRCTs that have the same level of that attribute (eg, multicenterstudy organization) are statistically homogeneous and do not showan association between allogeneic transfusion and postoperativeinfection; and if the combined findings from RCTs that have otherlevels of the attribute (eg, single-center study organization) are alsostatistically homogeneous and demonstrate a statistically signifi-cant relationship between allogeneic transfusion and postopera-tive infection.

Five study characteristics could account individually for thedisagreements among the 7 RCTs.3-7,22,23 These are: (1) studyorganization (ie, single center versus multicenter), (2) homogeneityof enrolled patients having elective abdominal surgery, (3) postop-erative infection rate recorded in the entire study population, (4)proportion of patients having transfusion, and (5) study authorship.The RBC product transfused to the treatment or control arm couldnot account for the disagreements among the studies (Table 4).Single-center RCTs,3-7 studies enrolling a mixed patient populationwith any disease of the colorectum or gastrointestinal tract,3,5,7

studies recording an overall infection rate of 20% or less,3-7 trials inwhich less than 50% or more than 67% of the enrolled patientsreceived perioperative transfusion,4-7 and RCTs reported by theteam of Jensen et al3,5 were all homogeneous when consideredtogether (P. .05 for the Q test statistic) and indicated a beneficialeffect of the transfusion of autologous or WBC-reduced allogeneicRBCs. Multicenter RCTs,22,23 studies enrolling a homogeneouspatient population having colorectal cancer resection,4,22,23studiesrecording an overall infection rate of greater than 20%,22,23trials inwhich 50% to 67% of the patients had transfusion,3,22,23and RCTsreported by teams of investigators other than Jensen et al4,6,7,22,23

were also homogeneous when considered together (P . .05 for theQ test statistic) and indicated no difference in the incidence ofpostoperative infection between the treatment and control arms(Table 4).

Only the 2 multicenter RCTs22,23 reported an overall infectionrate exceeding 20%; the overall infection rate was 20% or less inthe 5 single-center studies. The results of the 2 multicenter RCTs ofBusch et al23 and Houbiers et al22 may thus differ from the findingsof the other studies because of the lower statistical power ofmulticenter (compared with single-center) studies to detect atreatment effect.89-91 In addition, Busch et al23 and Houbiers et al22

may have used more lenient criteria for the diagnosis of postopera-tive infection (or may have included more patients with risk factorsfor postoperative infection) compared with the other studies.Because the reports of all 7 RCTs presented only very limitedinformation about the prevalence of risk factors for postoperativeinfection in the treatment and control arms of patients, the latterpossibility cannot be investigated without access to the raw data.





With regard to the former possibility, multicenter RCTs often havereduced statistical power compared with single-center studiesbecause of the influence of the “center effect” and their greatersusceptibility to the effects of random error.89-92The “center effect”refers to differences among the participating medical centers invariables that can influence the postoperative infection rate, but are

not captured (or standardized) by the common research protocol.The “center effect” can be removed from the results of a multi-center RCT if a statistical analysis stratified by participatinghospital is conducted, but such an analysis was not presented in thereports of Busch et al23 and Houbiers et al.22 Random error refers tomissing data, incorrect data, interobserver variability, and the like;both random error and the “center effect” bias the estimate of atreatment effect toward the null.92 The “center effect” may beespecially influential when the rates of postoperative infection inabdominal surgery are studied; a survey of 20 surgical departmentsacross Israel reported a variation in postoperative infection rates of0% to 65%.93

Jensen et al3,5 reported an implausibly large allogeneic bloodtransfusion effect and observed an extremely low postoperativeinfection rate among the patients who received WBC-reducedRBCs. In their 1996 study, Jensen et al5 detected 0 postoperativewound infections and intra-abdominal abscesses among 118 pa-tients sick enough to need perioperative transfusion with WBC-reduced RBCs. If this finding were not due to the effect(s) of biasand confounding, it would implicate the TRIM effect as the solecause of postoperative wound infections, because the incidence ofpostoperative wound infections and intra-abdominal abscessesamong the 142 patients who had transfusion with buffy-coat–reduced RBCs in that trial was 12%.5 Jensen et al3,5 did not presentinformation on some important potential confounding variables(eg, Dukes stage of tumor and tumor fixation to adjacent organs),allowing one to speculate that the observed large TRIM effect(Figure 4) could be due to the effects of uncontrolled confoundingfactors.24,25 In addition, selection bias and observation bias arepossible in RCTs despite the randomized design and the blinding ofinvestigators, respectively.94,95In the RCTs of Jensen et al,3,5 sickerpatients may have been systematically allocated to the treatmentarm, and/or investigators who became aware of the treatmentallocations may have been inclined to diagnose postoperativeinfection more often in patients from the treatment arm. Both

Figure 4. Randomized controlled trials investigating the association of periop-erative allogeneic blood transfusion with postoperative bacterial infection inpatients undergoing abdominal 3-5,7,22,23 or open heart 6 surgery. For each RCT, thefigure shows the OR of postoperative infection in recipients of buffy-coat–reduced4-6,22,23

or standard7 allogeneic RBCs or whole blood3 as compared with recipients ofautologous or WBC-reduced allogeneic RBCs or whole blood (Table 3). For thestudies of Jensen et al3,5 and Tartter et al,7 the OR is shown not as reported by theauthors, but as recalculated according to an intention-to-treat analysis.54 For thestudy of van de Watering et al,6 the depicted OR represents a comparison between 2groups, that is, the recipients of buffy-coat–reduced RBCs and the recipients ofWBC-reduced RBCs filtered before or after storage. Each OR is surrounded by its95% CI. If the 95% CI of the OR includes the null value of 1, the TRIM effect is notstatistically significant. The data are plotted on a logarithmic scale extending from 0.1 to 100.

Table 3. Randomized controlled trials investigating the association between perioperative allogeneic blood transfusion and postoperative bacterial infection

Study

Totalsamplesize* Design

RBC product given tothe treatment arm

RBC product given to thecontrol arm

Overallpercentage

havingtransfusion†

Overallpercentagedeveloping

postoperativeinfection†

Jensen et al3 197 ?Single blind Allogeneic whole

blood

WBC-reduced allogeneic

whole blood (filtered after

storage)

52.8 8.1

Heiss et al4 120 Unblinded BC-reduced

allogeneic RBCs

Autologous RBCs 75.0 20.0

Busch et al23 470 Unblinded BC-reduced

allogeneic RBCs

Autologous whole blood 59.0 26.0

Houbiers et al22 697 Single blind‡ BC-reduced

allogeneic RBCs

WBC-reduced, BC-reduced

allogeneic RBCs (filtered

before storage)

64.0 33.4

Jensen et al5 589 Single blind‡ BC-reduced

allogeneic RBCs



after storage)

44.1 17.5

van de Watering

et al6909 Single blind‡ BC-reduced

allogeneic RBCs



before [n 5 302] or after

[n 5 302] storage)

94.8 19.3

Tartter et al7 221 Unblinded Allogeneic RBCs WBC-reduced allogeneic

RBCs

26.7 16.7

RBC indicates red blood cell; WBC, white blood cell; BC, buffy coat.*Used in the analysis of postoperative bacterial infection after patient exclusions and withdrawals.†In both the treatment and control arms combined.‡Investigators making diagnoses of postoperative infection were blinded to the treatment allocations.





Vamvakas24 and McAlister et al25 excluded the RCTs of Jensen etal3,5 from their meta-analyses24,25 because of concerns about biasand residual, uncontrolled confounding factors in those studies.3,5

Jensen et al3,5 published 2 of the 3 studies3,5,7 that enrolled amixed population of patients having abdominal surgery for eitherbenign or malignant disease. In addition, 222,23of the 34,22,23studiesthat enrolled a homogeneous population of patients having colorec-tal cancer resection were multicenter. Furthermore, 222,23 of the33,22,23 studies in which 50% to 67% of the enrolled patients had

transfusion involved multiple centers. The influence of these 2study attributes (ie, homogeneity of the enrolled patient populationand overall percentage of transfusion) on the reported results isprobably due to the distribution of the 2 multicenter RCTs22,23andthe 2 trials by Jensen et al3,5 among the categories formed whenthese 2 study attributes are considered (Table 4).

To explain the disagreements among the 7 RCTs (Figure 4),Blajchman11 and Vamvakas and Blajchman96 proposed a meta-analysis of the 7 studies using individual patient data (IPD).97 The

Table 4. Possible sources of variation in the findings of the available randomized controlled trials investigating the associationbetween perioperative allogeneic blood transfusion and postoperative bacterial infection: integration of the findingsof studies sharing a particular design characteristic by a meta-analysis

Study attributeCategories of study

attributeNumber of

studiesQ test statistic:

P value

Summary odds ratio*

Pointestimate

95% confidenceinterval

Study organization Single center3-7 5 . .05 2.29† 1.35-3.87†

Multicenter22,23 2 .95 0.88 0.69-1.12

Surgical setting Open heart surgery6 1 N/A‡ 1.42§ 1.00-1.99§

Elective abdominal

surgery3-5,7,22,23

6 , .001 N/A\ N/A\

Homogeneity of

patients having

Colorectal

cancer4,22,23

3 . .10 0.96 0.67-1.37

elective

abdominal

surgery: diagnosis

Any disease of the

colorectum3,5 or

gastrointestinal

tract7

3 . .25 2.95† 1.73-5.05†

RBC product

transfused to the

Allogeneic whole

blood3

1 N/A‡ 7.32¶ 1.62-33.12¶

treatment arm Allogeneic RBCs7 1 N/A‡ 1.85¶ 0.89-3.85¶

Buffy-coat–reduced

RBCs4-6,22,23

5 , .001 N/A\ N/A\

RBC product

transfused to the

Autologous RBCs4

or whole blood23

2 . .05 1.35 0.45-4.08

control arm WBC-reduced

allogeneic

RBCs5-7,22 or

whole blood3

5 , .001 N/A\ N/A\

WBC-reduced

allogeneic

RBCs5-7,22

4 , .001 N/A\ N/A\

Fresh-filtered,

WBC-reduced

allogeneic

RBCs6,22

2 , .05 N/A\ N/A\

WBC-reduced

allogeneic RBCs

or whole blood

filtered after

storage3,5,6

3 , .01 N/A\ N/A\

Overall Exceeding 20%22,23 2 .95 0.88 0.69-1.12

postoperative

infection rate#

20% or lower3-7 5 . .05 2.29† 1.35-3.87†

Overall percentage Less than 50%5,7 2 . .10 2.73† 1.48-5.04†

of patients having 50% to 67%3,22,23 3 . .10 0.96 0.62-1.49

transfusion# More than 67%4,6 2 . .25 1.53† 1.00-2.33†

Reported by Jensen Jensen et al3,5 2 .50 3.68 2.11-6.43

et al or by other

teams of

investigators

Other teams4,6,7,22,23 5 . .05 1.17 0.84-1.62

RBC(s) indicates red blood cell(s); WBC, white blood cell; N/A, not applicable.*Calculated by the method of DerSimonian and Laird.56

†Statistically significant (P # .05) summary odds ratio.‡Not applicable because only one study was available for meta-analysis.§Odds ratio of postoperative infection in patients receiving buffy-coat–reduced RBCs, as compared with recipients of WBC-reduced, buffy-coat–reduced RBCs filtered

before or after storage.\P , .05 for the Q test statistic; accordingly, the results of the studies could not be integrated by a meta-analysis, and N/A is shown under the summary odds ratio.¶Odds ratio shown not as reported by the authors, but as recalculated according to an intention-to-treat analysis.#In the entire study population regardless of randomization arm.





proposed project96 would require recoding of the raw data collectedprospectively by the authors of the 7 RCTs with a common patientdata form, as well as the collection of additional data through aretrospective review of the medical records of the patients whowere enrolled in the 7 studies.3-7,22,23The additional informationneeded to explain the disagreements among the studies relatesprimarily to the prevalence of risk factors for postoperativeinfection, as well as the criteria used to reach each diagnosis ofpostoperative infection in the 7 RCTs. If information on thesevariables were obtained, it would highlight any differences inseverity of illness or application of the diagnostic criteria forpostoperative infection between the treatment and control arms ofthe studies, thereby allowing meta-analysts to assess the possibleeffects of bias and confounding in the 7 RCTs. In addition, thevariation in the postoperative infection rate among the studies(Table 1) could be removed, or attributed to a differential distribu-tion of risk factors for postoperative infection among the RCTs; andthe data from the 2 multicenter RCTs22,23 could be stratified byparticipating hospital to remove the influence of the “center effect.”If these postulated sources of variation accounted for the disagree-ments among the studies, a summary estimate of the TRIM effect(or of the benefit from the use of autologous and/or WBC-reducedRBCs) across the 7 RCTs could be calculated by the techniques ofmeta-analysis.97

Because the results of an IPD meta-analysis of all publishedRCTs3-7,22,23 are not available, the results of the study of van deWatering et al6 may be the most reliable existing data regarding thepotential benefit from the use of WBC-reduced blood componentsfor the prevention of postoperative infection.27 The study of van deWatering et al6 was well designed and analyzed and enrolled a largenumber of patients, almost all of whom received transfusion.Furthermore, observers who were unaware of patient assignmentmade the diagnoses of postoperative infection on the basis of thedetailed criteria of the Centers for Disease Control and Preven-tion.98 All of these conditions together were not met by any of theother RCTs shown in Table 3.

Van de Watering et al6 randomized 914 patients scheduled tohave CABG surgery, cardiac valve surgery, or a combination of the2 to one of 3 treatment arms: buffy-coat–reduced allogeneic RBCs;WBC-reduced, buffy-coat–reduced allogeneic RBCs filtered be-fore storage; and WBC-reduced, buffy-coat–reduced allogeneicRBCs filtered after storage. The intention-to-treat analysis com-pared the incidence of postoperative infection among these 3 armsand found no significant difference (P5 .13); the incidence ofpostoperative infection was 23%, 16.9%, and 17.9%, respectively,in each arm. However, when the prestorage and poststorageWBC-reduced arms were combined, recipients of buffy-coat–reduced RBCs had a higher incidence of postoperative infectionthan patients receiving WBC-reduced, buffy-coat–reduced RBCs,and this difference was marginally significant (P 5 .06; Figure 4).

Postoperative mortality

In addition to a possible association between allogeneic transfusionand postoperative infection (Figure 4), van de Watering et al6

detected an unexpected association between WBC-containingallogeneic blood transfusion and postoperative mortality fromcauses other than postoperative infection. Twenty-four of 306patients (7.8%) having transfusion with buffy-coat–reduced RBCsdied, compared with 11 of 305 patients (3.6%) receiving buffy-coat–reduced RBCs that were WBC reduced before storage and 10 of303 patients (3.3%) receiving buffy-coat–reduced RBCs that wereWBC reduced after storage (P5 .015). This overall difference in60-day mortality was due to a highly significant (P 5 .001)

difference among the 3 randomization arms in the proportion ofpatients who died of noncardiac causes, that is, multiorgan failureor dehiscence of their aortic bypass anastomosis (3.3% of patientsreceiving buffy-coat–reduced RBCs, as compared with 0.3% ofpatients receiving buffy-coat–reduced RBCs that were WBCreduced either before or after storage). The proportions of patientswho died of cardiac causes (ie, myocardial infarction, heart failure,or arrhythmia) did not differ among the 3 arms (P5 .53).

Furthermore, in an analysis of the effect of the volume of bloodinfused among the 866 patients who had transfusion, no differencein mortality between recipients of buffy-coat–reduced or WBC-reduced RBCs was observed in the subgroup of subjects receiving1 to 3 U of blood (P5 .82). In this subgroup of 355 patients,mortality was 1.6% for the recipients of buffy-coat–reduced units,as compared with 1.8% for the recipients of WBC-reduced unitsfiltered after storage and 0.9% for the recipients of WBC-reducedunits filtered before storage. In contrast, in the subgroup of 511patients who received 4 U or more of RBCs, there was astatistically significant (P5 .014) association between mortalityand WBC-containing allogeneic blood transfusion across the 3arms. When the recipients of WBC-reduced RBCs filtered before orafter storage were combined, the postoperative mortality rate ofrecipients of buffy-coat–reduced RBCs was 12.5%, as comparedwith 5.1% for the recipients of WBC-reduced, buffy-coat–reducedRBCs (P5 .005).

Van de Watering et al6,99 pointed out that their study had notbeen designed to investigate postoperative mortality as an outcomevariable. For this reason, prognostic factors for mortality had notbeen measured during the trial, and randomization might (or mightnot) have distributed such prognostic variables equally among the 3arms. Accordingly, using multivariate regression analysis, van deWatering et al6 adjusted the observed difference across the random-ization arms for the effects of measured prognostic factors formortality. All variables that demonstrated a univariate association(P , .05) with postoperative mortality were entered into thismultivariate regression model. Randomization to the buffy-coat–reduced RBC arm was associated with significantly increasedmortality (P5 .012) after adjustment for the effects of type ofoperation, age, history of myocardial infarction, previous openheart surgery, preoperative platelet count, and gender. When thenumber of RBC units transfused was entered into the regressionmodel, it was the most significant (P, .001) predictor of post-operative mortality, while randomization to the buffy-coat–reduced RBC arm still exercised a significant (P 5 .009) effecton mortality.

Van de Watering et al6 concluded that when 4 U or more ofallogeneic buffy-coat–reduced RBCs are transfused in cardiacsurgery, WBC reduction of the transfused units results in signifi-cantly decreased mortality. However, recognizing that this findingwas novel and unexpected, and that a recommendation for usingonly WBC-reduced blood components in cardiac surgery wouldsignificantly increase the cost of transfusion, these investigatorsurged that further research be conducted to replicate their findingbefore such a recommendation was made.6

Van de Watering et al postulated that their study6 was morelikely than previous RCTs3-5,7,22,23to detect differences among therandomization arms because the transfusion dose administered totheir cardiac surgery patients was substantially higher than thatgiven to patients having colorectal surgery in the previouslypublished trials. Only 48 (5.2%) of these cardiac surgery patients6

received no RBC transfusion; 777 (85.0%) received 1 to 10 U ofRBCs each (mean, 5.5 U; median, 4.0 U) and 89 (9.7%) receivedmore than 10 U each (mean, 16.9 U; median, 14.0 U). To further





investigate a possible relation between WBC-containing allogeneicblood transfusion and an increased risk of death or multiorganfailure in patients receiving high transfusion doses perioperatively,investigators from the Leiden University Medical Center initiatedan RCT enrolling subjects admitted for resection of aortic aneu-rysm or gastrointestinal malignancy. The purpose of the new studyis to determine whether the previously detected6 associationsbetween WBC-containing allogeneic blood transfusion and in-creased risk of death or multiorgan failure represent transfusion-associated complications (TAC) or transfusion-induced complica-tions (TICS); thus, the acronym TACTICS is used to refer to thisnew study.99 To optimize the design of this large, double-blind,multicenter trial, Brand et al100conducted a pilot study enrolling 73patients with aortic aneurysm and 84 patients with abdominalmalignancy. No statistically significant differences were observedin the pilot study. Ten percent of the recipients of buffy-coat–reduced RBCs died, as compared with 7.8% of the recipients ofWBC-reduced RBCs; approximately 20% of the patients in eitherrandomization arm developed multiorgan failure.100

Information on 30-day mortality was available for 1806 patientsenrolled in the 3 RCTs conducted at the Leiden University MedicalCenter (ie, the RCTs of Houbiers et al22 and van de Watering et al6

and the pilot component of TACTICS), and an interim analysisbased on the accumulated experience from these patients has beenpresented.99,100 In the univariate analysis, an increased mortalityrisk in recipients of buffy-coat–reduced (compared with WBC-reduced) RBCs was observed only among recipients of 4 U or moreof RBCs. In a multivariate regression analysis, randomization tothe WBC-reduced (as compared with the buffy-coat–reduced) armwas associated with a statistically significant decrease in postopera-tive mortality after adjustment for the effects of age, gender, andtransfusion dose. However, because prognostic factors for mortal-ity had not been considered in the design of the 3 combined studies,there may exist other prognostic variables that could be unequallydistributed between the buffy-coat–reduced and WBC-reducedarms and that could account for the observed difference inmortality between the arms.

The association between WBC-containing allogeneic bloodtransfusion and increased mortality6 may have limited applicabilitybecause data obtained from patients undergoing cardiac surgerymay not be generalizable to other clinical settings. The extracorpo-real circuit used in cardiac surgery induces a diffuse inflammatoryresponse that may predispose to postoperative infection or othersurgical complications.101 It is possible that WBC reduction oftransfused blood components may be capable of suppressing thediffuse inflammatory response to the extracorporeal circuit andmay thus indirectly reduce the incidence of postoperative infectionor the incidence of other postoperative complications.101 If thiswere the mechanism of the beneficial effect observed by van deWatering et al,6 the same benefit from the use of WBC-reducedblood components would probably not be expected to occur inother surgical settings.

The mechanism of the reported association6 between WBC-containing allogeneic blood transfusion and increased postopera-tive mortality from causes other than postoperative infection isunclear. If such a relationship truly exists, its mechanism may (ormay not) be related to transfusion-induced immunomodulation. Itis possible that, among recipients of multiple transfusions, alloge-neic transfusion may exercise other deleterious effects that have notyet been recognized or defined, probably because of the confound-ing effect of the association between transfusion and illnessseverity (which becomes extremely strong in patients receivingmultiple transfusions). Associations between allogeneic transfu-

sion and various adverse clinical outcomes (eg, prolonged mechani-cal ventilation,102,103impaired wound healing,104or multiorgan fail-ure105-107) have been reported from retrospective cohort studies, butit remains uncertain whether these associations reflect true relation-ships because it has been impossible to separate the effects of theallogeneic transfusions from the effects of confounding factors.

Hebert et al52 reported that a restrictive strategy of (non–WBC-reduced) RBC transfusion may be superior to a liberal transfusionstrategy in critically ill patients with normovolemia. Thirty-daymortality was 18.7% in patients having transfusion when theirhemoglobin concentration fell to 7 g/dL, as compared with 23.3%in subjects having transfusion when their hemoglobin concentra-tion fell to 10 g/dL (P5 .11). In subgroup analyses, patients whowere less acutely ill (with an Acute Physiology and Chronic HealthEvaluation II score of# 20) and patients who were younger than55 years experienced a statistically significant survival benefit byhaving been allocated to the restrictive-strategy arm of the study.Among the less acutely ill patients, 30-day mortality was 8.7% inthe restrictive-strategy arm and 16.1% in the liberal-strategy arm(P 5 .03). Among the patients who were younger than 55 years,30-day mortality was 5.7% with the restrictive strategy and 13.0%with the liberal strategy (P5 .02).

As with the finding of a reduced 60-day mortality rate amongthe recipients of WBC-reduced RBCs reported by van de Wateringet al,6 the finding of a trend (P5 .11) toward reduced 30-daymortality in the restrictive-strategy arm of the RCT of Hebert et al52

was unexpected. There were no differences between the arms in therates of cardiac events, infectious complications, multiorgan failureof more than 3 organs, or multiorgan failure in the 48 hours beforedeath. The harm caused by the liberal use of transfusions incritically ill patients was thus difficult to explain.108 It is possiblethat the overall trend observed by Hebert et al52 was due to chancebecause a statistically significant difference between the restrictive-and liberal-strategy arms could be demonstrated only in subgroupanalyses. Alternatively, it is possible that allogeneic transfusionmay exercise other deleterious effects that have not yet beenrecognized or defined. More research is needed to elucidate thepossible biologic mechanism(s) that may underlie such otheradverse effects of allogeneic transfusion.27

Conclusions

A causal relationship between allogeneic blood transfusion andcancer recurrence or postoperative infection has not been proven bydouble-blind RCTs in accordance with the tenets of evidence-basedmedicine.109 The available RCTs2,21,22 provide no indication thatperioperative allogeneic blood transfusion causes an increase incancer-related complications, whether cancer recurrence or deathdue to cancer recurrence is considered the outcome variable. TheOR of an adverse outcome in the allogeneic blood transfusion(compared with the control) arm is approximately 1. In the case ofpostoperative infection, the results of the available RCTs3-7,22,23arehighly contradictory, and they cannot be combined by the tech-niques of meta-analysis to generate an estimate of the TRIM effectbased on all available data.

Observational studies comparing recipients of similar volumesof autologous or allogeneic blood indicate an increased risk ofpostoperative infection in association with allogeneic transfusion(Figure 2). However, it is impossible to determine from thepublished results whether some portion of this excess risk seen inassociation with allogeneic transfusion would persist if the effectsof selection bias, observation bias, and confounding factors were to





be removed completely. Observational studies comparing patientswith and without transfusion, and using multivariate analysis toadjust for the effects of confounding factors, almost uniformlysupport the hypothesis of an increased risk of postoperativeinfection in the transfusion group.71,72,74-77,79However, in most ofthese studies, the reported allogeneic transfusion effects were notadjusted for the effects of illness severity and risk factors forpostoperative infection at specific sites.19,61

Reported RCTs investigating the association of perioperativeallogeneic blood transfusion with postoperative infection wereeither unblinded or single-blind (Table 1), and no double-blindRCTs have been reported. Because the diagnosis of nosocomialinfections is subjective, observation bias is an important concern,and it may represent an alternative explanation for the findings ofsome studies and for disagreements among the reported RCTs.110

Furthermore, the reports of the available RCTs3-7,22,23 did notpresent sufficient information about illness severity or the distribu-tion of risk factors for postoperative infection in the treatmentversus control arms.96 In the absence of this information, thepossible contributions of selection bias, observation bias, andconfounding factors to the reported results cannot be evaluated, andtherefore cannot be excluded.

A possible explanation for the disagreements among the 7RCTs3-7,22,23 may be that the adverse TRIM effect is small: forexample, a 5% to 10% increase in the risk of postoperativeinfection in association with allogeneic blood transfusion. Such asmall effect would not be detected consistently, and its detectionwould be highly dependent on the size and the design of each study.To detect a 10% difference in the risk of postoperative infectionbetween treatment and control arms of equal size, 3000 patientswould have to be randomized to each arm if the overall infectionrate in the entire study population were 35%.24 A study populationof 20 000 patients would be required if the overall infection ratewere approximately 20% and if almost half of the enrolled patientsdid not receive transfusion. Therefore, the available data from the3203 patients enrolled in the 7 published RCTs3-7,21,22are grosslyinsufficient for the investigation of an adverse TRIM effect of suchsmall magnitude. In fact, it is unlikely that a definitive RCT capableof detecting such a small effect will ever be conducted. To reach aconclusion about the existence and magnitude of any increase inthe risk of postoperative infection that is caused by allogeneictransfusion, decision makers will probably have to rely on cumula-tive meta-analyses of the data from all available RCTs, and theexcessive variation in the findings of the published reports3-7,21,22

will have to be accounted for to conduct such meta-analyses.11,96

As shown in Table 4, it is impossible to reconcile the findings ofall the available RCTs3-7,22,23 by relying only on the publishedinformation. The crux of the disagreements involves the radicallydifferent results reported by the team of Jensen et al3,5 and the 2multicenter RCTs.22,23 To permit a meta-analytic synthesis of thepublished findings, either the studies of Jensen et al3,5 or the 2multicenter RCTs22,23 have to be excluded from the analysis.92 Inthe absence of objective evidence of bias or confounding, anyexclusions are arbitrary, and they bias the conclusions of themeta-analysis either in the direction of a null result or in thedirection of a large (and likely spurious) deleterious TRIM effect(Table 4). A reanalysis of the raw data on the individual patientsenrolled in the 7 RCTs3-7,22,23and collection of additional informa-tion on these subjects are needed11,96 to reconcile the publishedresults and to reach a conclusion about the existence of adeleterious TRIM effect based on all available data from clinicalexperimental studies.

The question currently debated at the national level is the issueof universal WBC reduction. If TRIM really causes deleteriouseffects in recipients of blood products, these likely affect many (orpotentially all) patients having transfusion. Therefore, it would belogical to implement measures to prevent these immunomodulatoryeffects of allogeneic blood transfusion. Because most evidencefrom studies in experimental animals implicates allogeneic WBCsin the production of the deleterious TRIM effects,30,37-43 andbecause universal WBC reduction of all transfused cellular bloodcomponents has been implemented in some western Europeancountries and in Canada,50,51policy makers in the United States andelsewhere are currently considering the question of implementinguniversal WBC reduction for the purpose of preventing the variouspossible deleterious TRIM effects.

As shown in Table 4, when the 5 RCTs that transfusedWBC-reduced allogeneic RBCs5-7,22or whole blood3 to the controlarm were considered together, the results were heterogeneous(P , .001 for the Q test statistic) and could not be combined.Therefore, the existing evidence regarding the efficacy of WBCreduction in preventing the deleterious TRIM effects is conflicting.If standards similar to those used by the FDA’s new-drug evalua-tion process were applied to transfusion safety, the only appropriaterecommendation at this time would be that further double-blindRCTs be conducted to establish the existence of any adverse TRIMeffects mediated by transfused allogeneic WBCs, as well as theefficacy of WBC reduction in preventing these effects.

Decisions made in the 1990s,111-113however, indicate a willing-ness on the part of policy makers to use a substantially differentthreshold for judging the efficacy of measures that affect transfu-sion safety, compared with the criteria used for the evaluation ofother therapeutic interventions. Because of these precedents,111-113

the question now is whether universal WBC reduction should beimplemented in the United States and elsewhere at approximatelythe same time that this practice is introduced in western Europeand Canada.

Should universal WBC reduction be introduced now, or shouldwe wait for definitive, large, double-blind RCTs to establish boththe existence of deleterious TRIM effects and the efficacy of WBCreduction in abrogating these effects? Table 5 looks at this questionfrom a public policy (as opposed to an evidence-based medicine109)perspective and summarizes the arguments for and against a policyto implement universal WBC reduction for the prevention of theadverse TRIM effects, based on the current state of knowledge.26,27

All “pro” arguments are logical, and they are countered by equallyreasonable “con” arguments. Therefore, any decision at this time toimplement (or not to implement) universal WBC reduction in theUnited States for the purpose of preventing the purported deleteri-ous TRIM effects will have to be arbitrary. One can argue that apolicy decision should be made on the basis of the existingevidence because better evidence is unlikely to be forthcoming.26

Alternatively, one can argue that no policy decision should be madeuntil it is possible to make such a decision based on sound scien-tific evidence.27

Regardless of the content of the public policy debate, shoulduniversal WBC reduction be implemented, the question as towhether WBC-containing allogeneic blood transfusion causesdeleterious TRIM effects will remain. Therefore, should a decisionto introduce universal WBC reduction be made in the near future,research into the purported deleterious TRIM effects shouldcontinue. Double-blind RCTs should be conducted in countries thatdo not implement universal WBC reduction; observational studiescomparing the incidence of postoperative infection and/or cancer





recurrence before and after the implementation of universal WBCreduction should be conducted in jurisdictions that convert to a100% WBC-reduced blood supply.26,27

In summary, definitive evidence regarding the existence of adeleterious immunomodulatory effect of allogeneic blood transfu-sion, causing an increased incidence of cancer recurrence and/orpostoperative bacterial infection in humans, has not been pre-sented. However, the existing data from the various clinical studiesand studies in experimental animals, as well as the data on possiblebeneficial immunomodulatory allogeneic transfusion effects,12,114

justify a high degree of suspicion that an adverse TRIM effectprobably does exist. This effect may be small, perhaps representingless than a 10% increase in the risk of postoperative infection.Because of the small magnitude of the effect and the strongrelationship between the need for transfusion and illness severity, it

may have been difficult to document the existence of the TRIMeffect after adjusting appropriately for the effects of confoundingfactors. However, a risk as small as a 10% increase in the risk ofpostoperative infection in association with WBC-containing alloge-neic blood transfusion—if it really exists—represents a clinicallyimportant complication of transfusion that ought to be prevented, ifpossible, in all patients by introducing universal WBC reduction. Adecision whether to implement such a policy in the United Stateson the basis of the currently available evidence will probably bemade soon, and will likely be based, in part, on political andsocietal considerations.27,50,51Depending on which policy decisionis made, appropriately designed clinical studies should be con-ducted to either establish the existence of any deleterious immuno-modulatory effect(s) of allogeneic blood transfusion or documentthat the TRIM effect(s) is not clinically significant.

References

1. Opelz G, Sengar DP, Mickey MR, et al. Effect ofblood transfusions on subsequent kidney trans-plants. Transplant Proc. 1973;5:253-259.

2. Heiss MN, Mempel W, Delanoff C, et al. Bloodtransfusion-modulated tumor recurrence: first re-sults of a randomized study of autologous versusallogeneic blood transfusion in colorectal cancersurgery. J Clin Oncol. 1994;12:1859-1867.

3. Jensen LS, Andersen AJ, Christiansen PM, et al.Postoperative infection and natural killer cell func-tion following blood transfusion in patients under-going elective colorectal surgery. Br J Surg. 1992;79:513-516.

4. Heiss MM, Mempel W, Jausch K-W, et al. Benefi-cial effect of autologous blood transfusion on in-

fectious complications after colorectal cancer sur-gery. Lancet. 1993;342:1328-1333.

5. Jensen LS, Kissmeyer-Nielsen P, Wolff B, et al.Randomized comparison of leucocyte-depletedversus buffy-coat-poor blood transfusion andcomplications after colorectal surgery. Lancet.1996;348:841-845.

6. van de Watering LMG, Hermans J, HoubiersJGA, et al. Beneficial effect of leukocyte depletionof transfused blood on post-operative complica-tions in patients undergoing cardiac surgery: arandomized clinical trial. Circulation. 1998;97:562-568.

7. Tartter PI, Mohandas K, Azar P, et al. Random-ized trial comparing packed red cell blood transfu-sion with and without leukocyte depletion for gas-

trointestinal surgery. Am J Surg. 1998;176:462-466.

8. Peters WR, Fry RD, Fleshman JW, et al. Multipleblood transfusions reduce the recurrence rate ofCrohn’s disease. Dis Colon Rectum. 1989;32:749-753.

9. Tegtmeier GE. Post transfusion cytomegalovirusinfection. Arch Pathol Lab Med. 1989;113:236-245.

10. Vamvakas EC, Kaplan HS. Early transfusion andlength of survival in acquired immune deficiencysyndrome: experience with a population receivingmedical care at a public hospital. Transfusion.1993;33:111-118.

11. Blajchman MA. Allogeneic blood transfusions,immunomodulation and postoperative bacterial

Table 5. Arguments for (pro) and against (con) a policy to implement universal white cell reduction of all transfused cellular blood components for the purposeof preventing postoperative infection caused by the purported deleterious immunomodulatory effects of allogeneic blood transfusion

Pro26,27 Con27

1. An increase in the incidence of postoperative infection attributable to an

immunomodulatory effect of allogeneic WBCs is supported by at least one

single-blind and correctly designed and analyzed RCT conducted by van de

Watering et al.6

1a. The RCT of van de Watering et al6 did not detect a TRIM effect in the

intention-to-treat analysis; only when the 2 control (WBC-reduced) arms were

combined was a marginally significant (P 5 .06) difference demonstrated

between the recipients of buffy-coat–reduced and WBC-reduced RBCs.

1b. The RCT of van de Watering et al6 enrolled patients having open heart

surgery, and an allogeneic blood transfusion effect detected in the setting of

cardiac surgery may not apply to other surgical settings.

2. Four3,5-7 of 5 RCTs3,5-7,22 investigating the relationship between

WBC-containing (versus non–WBC-containing) allogeneic blood transfusion

and postoperative infection show at least a trend toward an increase in the

incidence of postoperative bacterial infection in association with the

transfusion of allogeneic WBCs.

2. The 2 RCTs by Jensen et al3,5 reported an implausibly large allogeneic blood

transfusion effect, and, if these 2 studies3,5 were excluded from the analysis,

the OR of postoperative infection in the allogeneic transfusion (compared with

the control) arm across the remaining RCTs6,7,22 would no longer be

statistically significant and would be sufficiently modest to be attributed to

chance.*

3. It is unlikely that further RCTs examining the relationship between allogeneic

blood transfusion and postoperative infection will be reported in the

foreseeable future. Even if more studies are conducted in the future, it is

unlikely that such future RCTs will be large enough to be definitive.

3. If a policy decision is made to implement universal WBC reduction based on

the evidence that is currently available, it will be impossible to rescind this

policy in the future if further studies establish that universal WBC reduction

does not decrease the incidence of postoperative infection.

4. In the future, research into the purported deleterious immunomodulatory

effects of allogeneic blood transfusion can continue, by means of observational

studies comparing the incidence of postoperative infection between patients

having transfusion before or after the implementation of universal WBC

reduction.

4. Such future observational studies are bound to have at least some of the

flaws of their predecessors,19 making it impossible to attribute any observed

difference between patients having transfusion before or after the

implementation of universal WBC reduction to the receipt of WBC-containing

allogeneic blood transfusion.

5. The receipt of WBC-containing allogeneic blood transfusion is associated with

increased postoperative mortality in the RCT of van de Watering et al.65a. The RCT of van de Watering et al6 was not designed to investigate mortality

as an outcome variable.

5b. The association between WBC-containing allogeneic blood transfusion and

increased postoperative mortality reported by van de Watering et al6 has not

been corroborated by another RCT.

WBC indicates white blood cell; RCT, randomized controlled trial; TRIM, allogeneic blood transfusion–associated immunomodulation; RBC, red blood cell; OR, odds ratio.*The OR of postoperative infection in recipients of buffy-coat–reduced6,22 or standard7 RBCs, as compared with recipients of WBC-reduced RBCs, was 1.20 (95%

confidence interval for the OR, 0.78-1.85; P . .05). These 3 studies6,7,22 are sufficiently homogeneous to permit meta-analysis (P . .05 for the Q test statistic).





infection: do we have the answers yet? Transfu-sion. 1997;37:121-125.

12. Cicciarelli JC, Iwaki Y. Direct transfusion effects inorgan transplantation. In: Singal DP, ed. Immuno-logical Effects of Blood Transfusion. Boca Raton,FL: CRC Press; 1994:43-61.

13. Opelz G, Vanrentergehm Y, Kirste G, et al. Pro-spective evaluation of pretransplant blood trans-fusions in cadaver kidney recipients. Transplanta-tion. 1997;63:964-967.

14. Gantt CL. Red blood cells for cancer patients.Lancet. 1981;2:363.

15. Francis DM. Relationship between blood transfu-sion and tumor behavior. Br J Surg. 1991;78:1420-1428.

16. Blumberg N, Heal JM. Effects of transfusion onimmune function: cancer recurrence and infec-tion. Arch Pathol Lab Med. 1994;118:371-379.

17. Vamvakas E. Perioperative blood transfusion andcancer recurrence: meta-analysis for explanation.Transfusion. 1995;35:760-768.

18. Brand A, Houbiers JGA. Clinical studies of bloodtransfusion and cancer. In: Vamvakas EC, Blajch-man MA, eds. Immunomodulatory Effects ofBlood Transfusion. Bethesda, MD: AABB Press;1999:145-190.

19. Vamvakas E, Moore SB. Blood transfusions andpostoperative septic complications. Transfusion.1994;34:714-727.

20. Duffy G, Neal KR. Differences in postoperativeinfection rates between patients receiving autolo-gous and allogeneic blood transfusion: a meta-analysis of published randomized and nonran-domized studies. Transfusion Med. 1996;6:325-328.

21. Busch ORC, Hop WCJ, van Papendrecht MAWH,et al. Blood transfusions and prognosis in colo-rectal cancer. N Engl J Med. 1993;328:1372-1376.

22. Houbiers JGA, Brand A, van de Watering LMG, etal. Randomized controlled trial comparing trans-fusion of leucocyte-depleted or buffy-coat-depleted blood in surgery for colorectal cancer.Lancet. 1994;344:573-578.

23. Busch ORC, Hop WCJ, Marquet RL, et al. Autolo-gous blood and infections after colorectal surgery.Lancet. 1994;343:668-669.

24. Vamvakas E. Transfusion-associated cancer re-currence and infection: meta-analysis of the ran-domized controlled clinical trials. Transfusion.1996;36:175-186.

25. McAlister FA, Clark HD, Wells PS, Laupacis A.Perioperative allogeneic blood transfusion doesnot cause adverse sequelae in patients with can-cer: a meta-analysis of unconfounded studies.Br J Surg. 1998;85:171-178.

26. Blajchman MA. Transfusion-associated immuno-modulation and universal white cell reduction: arewe putting the cart before the horse? Transfusion.1999;39:665-670.

27. Vamvakas EC, Dzik WH, Blajchman MA. Delete-rious effects of transfusion-associated immuno-modulation: appraisal of the evidence and recom-mendations for prevention. In: Vamvakas EC,Blajchman MA, eds. Immunomodulatory Effectsof Blood Transfusion. Bethesda, MD: AABBPress; 1999:253-285.

28. Brunson ME, Alexander JW. Mechanisms oftransfusion-induced immunosuppression. Trans-fusion. 1990;30:651-658.

29. Blajchman MA, Bordin JO. Mechanisms of trans-fusion-related immunosuppression. Curr OpinHematol. 1994;1:457-461.

30. Blajchman MA. Immunomodulatory effects of al-logeneic blood transfusions: clinical manifesta-tions and mechanisms. Vox Sang. 1998;74(suppl2):315S-319S.

31. Magee CC, Sayegh MH. Peptide-mediated im-munosuppression. Curr Opin Immunol. 1997;9:669-675.

32. Krensky AM, Clayberger C. Structure of HLA mol-

ecules and immunosuppressive effects of HLAderived peptides. Intern Rev Immunol. 1996;13:173-185.

33. Buelow R, Burlingham WJ, Clayberger C. Immu-nomodulation by soluble HLA class I. Transplan-tation. 1995;59:649-654.

34. Chueh SC, Tian L, Wang M, Wang ME, Step-kowski SM, Kahan BD. Induction of tolerance to-ward rat cardiac allografts by treatment with allo-chimeric class I MHC antigen and FTY720.Transplantation. 1997;64:1407-1414.

35. Naji A. Induction of tolerance by intrathymic in-oculation of alloantigen. Curr Opin Immunol.1996;8:704-709.

36. Shirwan H, Mhoyan A, Leamer M, Wange C, Cra-mer DV. The role of donor class I major histocom-patibility complex peptides in the induction of allo-graft tolerance. Transplant Proc. 1997;2:1134-1135.

37. Blajchman MA, Bardossy L, Carmen R, Sastry A,Singal D. Allogeneic blood transfusion-inducedenhancement of tumor growth: two animal mod-els showing amelioration of leukodepletion andpassive transfer using spleen cells. Blood. 1993;81:1880-1882.

38. Blajchman MA, Bordin JO. Tumor growth-promot-ing effect of allogeneic blood transfusion. Immu-nol Invest. 1995;24:311-317.

39. Bordin JO, Bardossy L, Blajchman MA. Growthenhancement of established tumors by allogeneicblood transfusion in experimental animals and itsamelioration by leukodepletion: the importance oftiming of the leukodepletion. Blood. 1994;84:344-348.

40. Kao KJ. Induction of humoral immune toleranceto major histocompatibility complex antigens bytransfusions of UV-B irradiated leukocytes. Blood.1996;88:4375-4382.

41. Kirkley SA, Cowles J, Pellegrini VD, Harris CM,Boyd AD, Blumberg N. Cytokine secretion afterallogeneic or autologous blood transfusion. Lan-cet. 1995;345:527.

42. Babcock GF, Alexander JW. The effects of bloodtransfusion on cytokine production by Th1 andTh2 lymphocytes in the mouse. Transplantation.1996;61:465-468.

43. Mincheff MS, Meryman HT, Kapoor V, Alsop P,Wotzel M. Blood transfusion and immunomodula-tion: a possible mechanism. Vox Sang. 1993;65:18-24.

44. Goodarzi MO, Lee TH, Pallavicini MG, DoneganEA, Busch MP. Unusual kinetics of white cellclearance in transfused mice. Transfusion. 1995;35:145-149.

45. Bordin JO, Heddle NM, Blajchman MA. Biologiceffects of leukocytes present in transfused cellu-lar blood products. Blood. 1994;84:1705-1721.

46. Bordin JO, Blajchman MA. Blood transfusion andtumor growth in animal models. In: VamvakasEC, Blajchman MA, eds. Immunomodulatory Ef-fects of Blood Transfusion. Bethesda, MD: AABBPress; 1997:29-42.

47. Lee TH, Paglieroni T, Ohto H, Holland PV, BuschMP. Survival of donor leukocyte subpopulations inimmunocompetent transfusion recipients: fre-quent long-term microchimerism in severe traumapatients. Blood. 1999;93:3127-3139.

48. Vamvakas EC, Blajchman MA. Prestorage versuspoststorage white cell reduction for the preven-tion of the deleterious immunomodulatory effectsof allogeneic blood transfusion. Transfus MedRev. 2000;14:23-33.

49. Brown P. Can Creutzfeldt-Jakob disease betransmitted by transfusion? Curr Opin Hematol.1995;2:472-477.

50. Murphy M. New variant Creutzfeldt-Jakob dis-ease (nvCJD): the risk of transmission by bloodtransfusion and the potential benefit of leukocyte-reduction of blood components. Transfus MedRev. 1999;13:75-83.

51. Dzik S, AuBuchon JC, Jeffries L, et al. Leukocytereduction of blood components: public policy and

new technology. Transfus Med Rev. 2000;14:34-52.

52. Hebert PC, Wells G, Blajchman MA, et al. A multi-center randomized controlled trial of transfusionrequirements in critical care. N Engl J Med. 1999;340:409-417.

53. Thurer RL, Luban NLC, AuBuchon JC, et al. Uni-versal WBC reduction. Transfusion. 2000;40:751-752.

54. Vamvakas EC. Rationale, objectives, and inter-pretation of the findings of clinical studies of theimmunomodulatory effects of blood transfusion.In: Vamvakas EC, Blajchman MA. Immunomodu-latory Effects of Blood Transfusion. Bethesda,MD: AABB Press; 1999:95-128.

55. L’Abbe KA, Detsky AS, O’Rourke K. Meta-analy-sis in clinical research. Ann Intern Med.1987;107:224-233.

56. DerSimonian R, Laird N. Meta-analysis in clinicaltrials. Control Clin Trials. 1986;7:177-188.

57. Murphy S, Heaton WA, Rebulla P. Platelet pro-duction in the Old World—and the New. Transfu-sion. 1996;36:751-754.

58. Nielsen HJ, Reimert CM, Pedersen AN, et al.Time-dependent, spontaneous release of whitecell- and platelet-derived bioactive substancesfrom stored human blood. Transfusion. 1996;36:960-965.

59. Nielsen HJ. Detrimental effects of perioperativeblood transfusion. Br J Surg. 1995;82:582-587.

60. Vamvakas EC, Carven JH. Transfusion and post-operative pneumonia in coronary artery bypassgraft surgery: effect of the length of storage oftransfused red cells. Transfusion. 1999;39:701-710.

61. Vamvakas EC, Carven JH, Hibberd PL. Bloodtransfusion and infection after colorectal cancersurgery. Transfusion. 1996;36:1000-1008.

62. Chung M, Steinmetz OK, Gordon PH. Periopera-tive blood transfusion and outcome after resec-tion for colorectal carcinoma. Br J Surg. 1993;80:427-432.

63. Easterbrook PJ, Berlin JA, Gopalan R, MatthewsDR. Publication bias in clinical research. Lancet.1991;337:867-872.

64. Begg CB, Berlin JA. Publication bias and dis-semination of clinical research. J Natl CancerInst. 1989;81:107-115.

65. Dickersin K, Min YI, Meinert CL. Factors influenc-ing publication of research results: follow-up ofapplications submitted to two institutional reviewboards. JAMA. 1992;267:374-378.

66. Ness PM, Walsh PC, Zahurak M, Baldwin ML,Piantadosi S. Prostate cancer recurrence in radi-cal surgery patients receiving autologous or ho-mologous blood. Transfusion. 1992;32:31-36.

67. Bonktes HJ, Walboomers JMM, Meijer JLM, et al.Specific HLA class I downregulation is an earlyevent in cervical dysplasia associated with clinicalprogression. Lancet. 1988;351:187-188.

68. Fernandez MC, Gottlieb M, Menitove JE. Bloodtransfusion and postoperative infection in ortho-pedic patients. Transfusion. 1992;32:318-322.

69. Johnson S, Andersson M. Adverse effects of peri-operative blood transfusion in patients with colo-rectal cancer. Eur J Surg. 1992;158:419-425.

70. Vamvakas EC, Moore SB, Cabanela M. Bloodtransfusion and septic complications after hip re-placement surgery. Transfusion. 1995;35:150-156.

71. Carson JL, Altman DG, Duff A, et al. Risk of bac-terial infection associated with allogeneic bloodtransfusion among patients undergoing hip frac-ture repair. Transfusion. 1999;39:694-700.

72. Ford CD, Vanmoorleghem G, Menlove RL. Bloodtransfusions and postoperative wound infection.Surgery. 1993;113:603-607.

73. Vamvakas EC, Carven JH. Transfusion of whitecell-containing allogeneic blood components andpostoperative wound infections: effect of con-founding factors. Transfus Med. 1998;8:29-36.





74. Chang H, Hall GA, Geerts WH, et al. Allogeneicred blood cell transfusion is an independent riskfactor for the development of postoperative bac-terial infection. Vox Sang. 2000;78:13-18.

75. Tartter PI, Driefuss RM, Malon AM, Heimann TM,Aufses AH. Relationship of postoperative septiccomplications and blood transfusions in patientswith Crohn’s disease. Am J Surg. 1988;155:43-48.

76. Wobbes T, Bemelmans BL, Kuypers JH, Beer-thuizen GI, Theeuwes AG. Risk of postoperativeseptic complications after abdominal surgicaltreatment in relation to perioperative blood trans-fusion. Surg Gynecol Obstet. 1990;71:59-62.

77. Pinto V, Baldonedo R, Nicolas C, Barez A, PerezA, Aza J. Relationship of transfusion and infec-tious complications after gastric carcinoma op-erations. Transfusion. 1991;31:114-118.

78. Innerhofer P, Walleczek C, Luz G. Transfusion ofbuffy-coat-depleted blood components and risk ofpostoperative infection in orthopedic patients.Transfusion. 1999;39:625-632.

79. Jensen LS, Andersen A, Fristrup S, et al. Com-parison of one dose versus three doses of pro-phylactic antibiotics, and the influence of bloodtransfusion, on infectious complications in acuteand elective colorectal surgery. Br J Surg. 1990;77:513-518.

80. Murphy P, Heal JM, Blumberg N. Infection or sus-pected infection after hip replacement surgerywith autologous or homologous blood transfu-sions. Transfusion. 1991;31:212-217.

81. Howard HL, Rushambuza FG, Martlew VJ,Bolton-Maggs B. Clinical benefits of autologousblood transfusion: an objective assessment. ClinLab Haematol. 1993;15:165-171.

82. Mezrow CK, Bergstein I, Tartter PI. Postoperativeinfections following autologous and homologousblood transfusion. Transfusion. 1992;32:27-30.

83. Newman JH, Bowers M, Murphy J. Reduction ofpostoperative infection by the use of autologousreinfusion. Transfusion. 1996;36:657.

84. Triulzi DJ, Vanek K, Ryan DH, Blumberg N. Aclinical and immunologic study of blood transfu-sion and postoperative bacterial infection in spi-nal surgery. Transfusion. 1992;32:517-524.

85. Smith MEF, Marsh SGE, Bodmer JG, et al. Lossof HLA-A, B, C allele products and lymphocytefunction-associated antigen 3 in colorectal neo-plasia. Proc Natl Acad Sci U S A. 1989;86:5557-5561.

86. Moller P, Momburg F, Koretz K, Moldenhauer G,Hersfarth C, Otto HF. Influence of major histo-compatibility complex class I and II antigens onsurvival in colorectal carcinoma. Cancer Res.1991;51:729-736.

87. Penn I. The occurrence of malignant tumors inimmunosuppressed states. Prog Allergy. 1986;37:259-300.

88. Busch ORC, Hop WCJ, Marquet RL, et al. Bloodtransfusions and local tumor recurrence in colo-rectal cancer: evidence of a noncausal relation-ship. Ann Surg. 1994;220:791-797.

89. O’Gorman TW, Woolson RF, Jones MP. A com-parison of two methods of estimating a commonrisk difference in a stratified analysis of a multi-center clinical trial. Control Clin Trials. 1994;15:135-153.

90. Howard VJ, Grizzle J, Diener HC, et al. Compari-son of multicenter study designs for investigationof carotid endarterectomy efficacy. Stroke. 1992;23:583-593.

91. Fleiss JL. Multicenter clinical trials: Bradford Hill’scontributions and some subsequent develop-ments. Stat Med. 1982;1:353-359.

92. Vamvakas EC. Multicenter randomized controlledtrials in transfusion medicine. Transfus Med Rev.2000;14:137-150.

93. Simchem E, Zucker D, Siegman IY, Calai N.Method for separating patient and procedural fac-tors while analyzing interdepartmental differencesin rates of surgical infections: the Israeli study ofsurgical infection in abdominal operations. J ClinEpidemiol. 1996;49:1003-1007.

94. Schultz KF, Chalmers I, Hayes RJ, Altman DG.Empirical evidence of bias: dimensions of meth-odological quality associated with estimates oftreatment effects in controlled trials. JAMA. 1995;273:408-412.

95. Schultz KF. Subverting randomization in con-trolled trials. JAMA. 1995;274:1456-1458.

96. Vamvakas EC, Blajchman MA. A proposal for anindividual patient data based meta-analysis ofrandomized controlled trials of allogeneic transfu-sion and postoperative bacterial infection. Trans-fus Med Rev. 1997;11:180-194.

97. Stewart LA, Parmar MKB. Meta-analysis of theliterature or individual patient data: is there a dif-ference? Lancet. 1993;341:418-422.

98. Garner JS, Jarvis WR, Emori TG, et al. CDC defi-nitions for nosocomial infections, 1988. Am J In-fect Control. 1988;16:128-140.

99. Van de Watering LMG, Brand A. Mortality afterallogeneic blood transfusion. In: Blajchman MA,Dzik S, Vamvakas E, Sweeney J, Snyder E, eds.Clinical and Molecular Basis of Transfusion-In-duced Immunomodulation: Summary of the Pro-ceedings of a State-of-the-Art Conference. Trans-fus Med Rev. In press.

100. Brand A, van de Watering LMG, Claas FHJ. Clini-cal significance of leukoreduction of blood com-

ponents. Vox Sang. 2000;78(suppl 2):227S-229S.

101. Gott JP, Cooper WA, Schmidt FE, et al. Modifyingrisk for extracorporeal circulation: trial of four anti-inflammatory strategies. Ann Thorac Surg. 1998;66:747-754.

102. Habib RH, Zacharias A, Engoren M. Determi-nants of prolonged mechanical ventilation aftercoronary artery bypass grafting. Ann ThoracSurg. 1996;62:1164-1171.

103. Fransen E, Maessen J, Dentener M, Senden N,Buurman W. Impact of blood transfusions on in-flammatory mediator release in patients undergo-ing cardiac surgery. Chest. 1999;116:1233-1239.

104. Chmell MJ, Schwartz HS. Analysis of variablesaffecting wound healing after musculoskeletalsarcoma resections. J Surg Oncol. 1996;61:185-189.

105. Tran DD, van Onselen EB, Wensink AJ, CuestaMA. Factors related to multiple organ system fail-ure and mortality in a surgical intensive care unit.Nephrol Dial Transplant. 1994;9(suppl 4):172S-178S.

106. Maetani S, Nishikawa T, Hirakawa A, Tobe T. Roleof blood transfusion in organ system failure fol-lowing major abdominal surgery. Ann Surg. 1986;203:275-281.

107. Peerless JR, Alexander J, Pinchak AC, PiotrowskiJ, Malangoni MA. Oxygen delivery is an importantpredictor of outcome in patients with rupturedaortic aneurysms. Ann Surg. 1998;227:726-732.

108. Ely EW, Bernard GR. Transfusions in critically illpatients. N Engl J Med. 1999;340:467-468.

109. Evidence-Based Medicine Working Group. Evi-dence-based medicine: a new approach to teach-ing the practice of medicine. JAMA. 1992;268:2420-2425.

110. Huskisson EC, Scott J. How blind is double-blind? And does it matter? Br J Clin Pharmacol.1976;3:331-332.

111. AuBuchon JP, Birkmeyer JD, Busch MP. Cost-effectiveness of expanded human immunodefi-ciency virus testing protocols for donated blood.Transfusion. 1997;37:45-51.

112. Etchason J, Petz L, Keeler E, et al. The cost-effectiveness of preoperative autologous blooddonations. N Engl J Med. 1995;332:719-724.

113. American Association of Blood Banks. NATImplementation (Association Bulletin no. 99-3,February 8, 1999). Bethesda, MD: American As-sociation of Blood Banks; 1999.

114. Recurrent Miscarriage Immunotherapy TrialistsGroup (RMITG). Worldwide collaborative obser-vational study and meta-analysis on allogeneicleukocyte immunotherapy for recurrent spontane-ous abortion. Am J Reprod Immunol. 1994;32:55-72.





Deleterious clinical effects of transfusion-associated immunomodulation: fact or fiction?

Documents

Transcript of Deleterious clinical effects of transfusion-associated immunomodulation: fact or fiction?