BRECHER ET AL.

794 TRANSFUSION Volume 41, June 2001 www.transfusion.org

White cell reduction of blood components iswidely used in North America, Europe, Asia,and Australia, and several countries, includ-ing Canada, the United Kingdom, Portugal,

France, Ireland, and Norway, have announced plans toswitch to a completely WBC-reduced blood supply (so-called universal leukocyte [WBC] reduction, or ULR).1,2

Similarly, the FDA is currently encouraging voluntaryimplementation of ULR whenever feasible, and it plans ei-ther to create a regulatory requirement for ULR or to enforceULR as a component standard under current Good Manu-facturing Practices regulations.3

However, the implementation of ULR is not withoutcontroversy: objections to ULR have principally centered onthe question of degree of its benefit to all patients and theadditional financial cost associated with ULR.4 One aspectof WBC reduction that has been largely overlooked is theadditional impact of the associated RBC loss. It is knownthat the use of filtration to reduce WBCs in RBC units leadsto a 6- to 15-percent loss of RBCs.5-8 However, it is not knownif the use of WBC-reduced RBCs results in an increasedneed for RBC units or in an increased number of units trans-fused per year in patients with chronic anemia who aretransfusion dependent. In an attempt to answer this ques-

Number of RBC units and rate of transfusionto anemic HIV-positive patients assigned to receive

WBC-reduced or non-WBC-reduced RBCs:the Viral Activation Transfusion Study experience

Mark E. Brecher, Darrell J. Triulzi, and Susan F. Assmann for the Viral Activation Transfusion Study

BACKGROUND: It is known that the use of filtration to re-duce WBCs in RBC units is associated with a 6- to 15-percent loss of RBCs. It is not known if the use of suchWBC-reduced RBCs results in an increased need forRBC units or in the transfusion of more units per year topatients with anemia.STUDY DESIGN AND METHODS: In the multicenter ViralActivation Transfusion Study (VATS), anemic HIV-positivepatients were randomly assigned to receive either WBC-re-duced or non-WBC-reduced RBCs. The number of RBCunits transfused per patient and the rate of RBC use werestudied. All RBC units given after the enrollment transfusionwere counted, until the end of follow-up or the occurrenceof bleeding (receiving >5 RBCs within 2 consecutive days).RESULTS: As expected, the WBC-reduced RBC units inVATS were lighter in weight than the non-WBC-reducedunits (median weight: WBC-reduced, 300 g; non-WBC-re-duced, 330 g; p<0.0001). After the enrollment transfusion,258 WBC-reduced arm patients received 1279 units ofRBCs (average, 5.0 units/patient, median, 2 units) while262 patients in the non-WBC-reduced arm received 1111RBCs (4.2 units/patient; median, 2 units). The number ofunits transfused for anemia was slightly greater in theWBC-reduced arm, but the difference was not significant(p = 0.41). Similarly, the rate of RBC use was somewhathigher in the WBC-reduced arm, but the difference wasnot significant (p = 0.14). The median was 2.3 units perpatient per year of follow-up in the WBC-reduced arm; themedian in the non-WBC-reduced arm was 1.2 units.CONCLUSION: This study confirms that WBC-reducedRBC units are significantly lighter in weight than non-WBC-reduced RBCs. However, in the setting of a large,randomized, blinded study of transfusion for anemia, thesmaller size of the WBC-reduced RBC units had no sig-nificant effect on the number of RBC units transfused oron the rate at which RBC units were used. In this study,the frequency of blood transfusion may have had agreater relationship to the frequency of routine, scheduledappointments or transfusion orders for a specified Hb trig-ger than to the actual Hb content of the unit.

ABBREVIATIONS: T1 = initial transfusion; ULR = universal leu-

kocyte reduction; VATS = Viral Activation Transfusion Study.

From the Transfusion Medicine Service, University of North

Carolina, Chapel Hill, North Carolina; the Institute for Transfu-

sion Medicine, Pittsburgh, Pennsylvania; and the New England

Research Institutes, Watertown, Massachusetts.

Address reprint requests to: Mark E. Brecher, MD, Transfu-

sion Medicine Service, Campus Box 7600 UNC Hospitals, 101

Manning Drive, Chapel Hill, NC 27514; e-mail:

brecher@med.unc.edu.

Supported by the National Heart, Lung, and Blood Insti-

tute.

Received for publication October 13, 2000; revision re-

ceived December 16, 2000, and accepted December 21, 2000.

TRANSFUSION 2001;41:794-798.

T R A N S F U S I O N P R A C T I C E

RBC TRANSFUSION RATE AND WBC REDUCTION

Volume 41, June 2001 TRANSFUSION 795www.transfusion.org

tion, we reviewed the experience of the 11 clinical sites ofthe Viral Activation Transfusion Study (VATS).

MATERIALS AND METHODS

The multicenter VATS was a prospective, randomized,double-blind study comparing the use of WBC-reducedRBCs and of non-WBC-reduced RBCs in persons who wereseropositive for both HIV and CMV and who required trans-fusions for anemia. Primary endpoints were survival andthe change in the plasma HIV RNA level after transfusion.Details of the design and primary results of this study havebeen reported previously.9,10

We counted follow-up time in these analyses as start-ing on the day the initial transfusion (T1) was given, andending on the earliest of these dates: 1) the date of death(for patients who died during the study); 2) the last date thepatient was known to be alive (for patients who did not dieduring the study); and 3) the date 1 day before the first of 2consecutive days during which 5 or more total RBC unitswere given. The units that were counted in these analyseswere RBC units given at transfusions other than T1 and theday after T1, but during the follow-up period defined above.

The 5-unit exclusion was arrived at by a consensus ofthe Transfusion Subcommittee of the VATS Steering Com-mittee. This exclusion was based on the pragmatic limita-tion of transfusing 2 units of RBCs in 24 hours (or 4 in 48hours) in a nonbleeding patient, due to the fear of possiblecirculatory overload. It was felt that transfusion in excess of4 RBC units in 48 hours likely represented an urgent needfor transfusion in a hypovolemic (i.e., bleeding) patient.This exclusion was chosen before the actual analysis.

Units given at T1 (including any units given the follow-ing day) were not counted, because there is no possibilitythat the number of units ordered at T1 could be affected bywhether the patient was assigned to receive WBC-reducedunits. For example, let us suppose that a patient receivedT1 on January 1, 1996 (2 RBC units), a second transfusionon January 15, 1996 (2 RBC units), a third transfusion onFebruary 1, 1996 (4 RBC units), and a fourth transfusion onFebruary 2, 1996 (2 RBC units) and died on February 29,1996. The patient would contribute 30 days of follow-uptime to the analysis (from 1/1/96 to 1/31/96), because Feb-ruary 1 was the start of a 2-day period during which at least5 RBC units were transfused. This patient would contribute2 units to the analysis. As another example, a patient whoreceived T1 on January 1, 1996 (2 RBC units), had no moreRBC transfusions, and was last known to be alive on Janu-ary 1, 1998, would contribute 731 days of follow-up and 0units to the analysis.

The number of units transfused, the rate of RBC use(units/patient/year), and the weight of the units in theWBC-reduced and non-WBC-reduced RBC arms of thestudy were compared. Statistical analysis employed the

nonparametric Wilcoxon’s rank sum test; statistical signifi-cance was taken at p<0.05. Analyses were presented graphi-cally with box-and-whisker plots. The nonparametric boot-strap method was used to calculate CIs. Similar analyseswere performed with regard to the total number of trans-fusion days (24-hour period during which a patient receivedRBCs) and the number of transfusion days over time.

RESULTSThere were 265 patients randomly assigned to the WBC-reduced group and 266 to the non-WBC-reduced group.The analyses presented here are restricted to the 520 pa-tients who actually received at least the baseline transfusion(T1) (258 WBC-reduced, 262 non-WBC-reduced). Overall,the WBC-reduced group patients received 2038 units ofRBCs (7.90 units/patient) and the non-WBC-reduced grouppatients received 1827 units of RBCs (6.97 units/patient).As expected, the WBC-reduced RBC units in VATS werelighter than the non-WBC-reduced units, as illustrated inFig. 1 (median weight: WBC-reduced, 300 g; non-WBC-re-duced group, 330 g; p<0.0001).

After the enrollment transfusion, the WBC-reducedgroup patients received 1279 units of RBCs (4.96/patient;median, 2 units) for anemia, while the non-WBC-reduced

Fig. 1. Box-and-whisker plot of the weight of the RBC units

(median: WBC-reduced, 300 g; non-WBC-reduced, 330 g). The

center horizontal lines represent the median, and the boxes

encompass the middle 50 percent of cases (the middle two

quartiles). The vertical lines extending from the boxes (the

“whiskers”) reach the 10th and 90th percentiles. Single dots (•)

indicate outlier points. There was a significant difference in

the weight of the units (p<0.0001).

BRECHER ET AL.

796 TRANSFUSION Volume 41, June 2001 www.transfusion.org

group patients received 1111 RBCs (4.24/patient; median,2 units). This represents 16.9 percent more units per patientin the WBC-reduced group than in the non-WBC-reducedgroup (95% CI, –13.1% to +56.7%). The slightly higher num-ber of units transfused for anemia in the WBC-reducedgroup (p = 0.41) is illustrated in Fig. 2.

Similarly, the RBC use per year of follow-up was some-what greater in the WBC-reduced arm (Fig. 3), but the dif-ference was not significant (p = 0.14). The mean and me-dian numbers of units per patient-year were 9.25 and 2.28,respectively, in the WBC-reduced group and 8.54 and 1.24,respectively, in the non-WBC-reduced group. This repre-sents an 8.3-percent higher mean transfusion rate in theWBC-reduced group (95% CI, –22.4% to +53.3%).

There were 21 patients (12 WBC-reduced, 9 non-WBC-reduced) who met the criterion for bleeding. The resultswere similar if units and follow-up time after the start ofbleeding episodes were not excluded. Patients in both studyarms received a median of 2 units after T1 (p = 0.54). Themedian use was 2.7 units per year of follow-up in the WBC-reduced arm and 1.4 units per year of follow-up in the non-WBC-reduced arm (p = 0.16).

Patients in the WBC-reduced arm had a median of 1transfusion day after T1 and a mean of 2.4 transfusion days,while in the non-WBC-reduced arm, the median was 1transfusion day and the mean was 2.0 transfusion days (p= 0.31). For the rate of transfusion days after T1, the medianin the WBC-reduced arm was 1.1 transfusion days per year

of follow-up and the mean was 4.6, while in the non-WBC-reduced arm, the median was 0.6 transfusion days per yearof follow-up and the mean was 4.0 (p = 0.13).

DISCUSSIONIt is known that the filtration of RBC-containing blood com-ponents is associated with the loss of RBC content.5-8 Suchloss can exceed 20 percent of the original component. Thisis particularly true when (as in Europe) the buffy coat isremoved before filtration, because that step is associatedwith a median loss of 12.4 percent of Hb from the originaldonation (range, 7.5-18%).8 This loss, coupled with the lossdue to filtration, can result in an overall RBC recovery be-low 80 percent.8 In the United States, FDA guidance11 andAABB standards12 mandate that, after WBC filtration, theRBC components retain a minimum of 85 percent of theoriginal component. Most currently available filters reliablyprovide a mean RBC recovery in excess of 85 percent.5-8 AsRBCs are not specifically retained by current WBC-reduc-tion filters, RBC loss is a function of the dead-space volumeof the filter. Thus, the 9-percent difference in the weight ofWBC-reduced and non-WBC-reduced RBC componentsobserved in this study is consistent with a 9-percent loss ofRBC content.5

It is possible that the 6- to 15-percent loss of RBC con-tent due to WBC reduction of units by filtration would trans-late into a need for 6 to 15 percent more RBC units to sup-

Fig. 2. Box-and-whisker plot of the number of RBC units trans-

fused by WBC-reduction status after the T1. The number of

units transfused for anemia was slightly higher in the WBC-

reduced group, but the difference was not significant (p =

0.41).

Fig. 3. Box-and-whisker plot of the RBC units per year of fol-

low-up. The RBC use per year of follow-up was somewhat

higher in the WBC-reduced arm, but the difference was not

significant (p = 0.14): median: WBC-reduced, 2.3 units per pa-

tient per year of follow-up; non-WBC-reduced, 1.2 units per

patient per year of follow-up.

RBC TRANSFUSION RATE AND WBC REDUCTION

Volume 41, June 2001 TRANSFUSION 797www.transfusion.org

port patients. This potentially could translate into an addi-tional 6- to 15-percent expense for the transfusion of WBC-reduced RBCs above and beyond the cost of the filtrationprocess and thus may represent a hidden cost of ULR. Inaddition, a 6- to 15-percent increase in the demand forRBCs could exacerbate existing blood shortages.

Patients receiving RBC transfusions generally do so inone of two settings: 1) rapid massive transfusion for acuteblood loss or 2) repeated RBC transfusion support forchronic anemia. Intuitively, one would expect that rapidmassive transfusion would translate into a need for an ad-ditional 6- to 15-percent RBC mass to achieve a given Hctor Hb. However, in the case of chronic transfusions, myriadvariables come into play, such as the patient’s subjectiveperception of weakness, the frequency of clinic appoint-ments, and the range of Hb levels that may result in thesame transfusion prescription. All of these might mitigatethe need for additional transfusions.

Our study was subject to several shortcomings. We es-tablished by consensus that the need for more than 4 unitsof RBCs in 48 hours was likely to indicate that the patientwas bleeding, but we did not have information on actualbleeding episodes for each patient assigned to this category.We had too few patients who received large-volume trans-fusions to know if the smaller volume of WBC-reduced RBCswould have necessitated a greater number of transfusionsfor those experiencing major hemorrhage.

In conclusion, the present randomized, blinded studyof chronically transfused HIV-positive patients, the largeststudy of this type, confirms that WBC-reduced RBC unitsare significantly lighter in weight than non-WBC-reducedRBCs. In this setting, the smaller size of the RBC units hadno significant effect on the number of RBC units transfusedor on the rate at which RBC units were used. Other variables(e.g., the routine scheduling of clinic appointments andstandard transfusion orders for a specific Hb or Hct range)may be more important in chronic transfusion therapy thanan approximate 6- to 15-percent loss of RBCs per unit. How-ever, given the wide CIs observed and on the basis of thesedata, one cannot conclude with absolute certainty as towhether there was a “hidden cost” associated with the useof WBC-reduced RBCs. These results have important impli-cations for the ongoing controversy regarding the imple-mentation of ULR.

REFERENCES01. Dzik S, AuBuchon J, Jeffries L, et al. Leukocyte reduction of

blood components: public policy and new technology.

Transfus Med Rev 2000;14:34-52.

02. Leukocyte reduction. Association Bulletin 99-7. Bethesda:

American Association of Blood Banks, August 2, 1999.

03. Transcript of the Food and Drug Administration Blood

Product Advisory Committee meeting, June 16, 2000. Avail-

able at: http://www.fda.gov/ohrms/dockets/ac/00/tran-

scripts/3620t2.pdf.

04. Thurer RL, Luban NL, AuBuchon JP, et al. Universal WBC

reduction (letter). Transfusion 2000;40:751-2.

05. Van der Meer PF, Pietersz RN, Nelis JT, et al. Six filters for

the removal of white cells from red cell concentrates, evalu-

ated at 4°C and/or at room temperature. Transfusion

1999;39:265-70.

06. Bontadini A, Fruet F, Tazzari PL, et al. Comparative analysis

of six different white cell-reduction filters for packed red

cells. Transfusion 1994;34:531-5.

07. Pietersz RN, Steneker I, Reesink HW, et al. Comparison of

five different filters for the removal of leukocytes from red

cell concentrates. Vox Sang 1992;62:76-81.

08. Rebulla P, Porretti L, Bertolini F, et al. White cell-reduced

red blood cells prepared by filtration: a critical evaluation

of current filters and methods for counting residual white

cells. Transfusion 1993;33:128-33.

09. Busch MP, Collier A, Gernsheimer T, et al. The Viral Activa-

tion Transfusion Study (VATS): rationale, objectives, and

design overview. Transfusion 1996;36:854-9.

10. Collier AC, Kalish LA, Busch MP, et al. Leukocyte-reduced

of red blood cell transfusions in patient's with anemia and

human immunodeficiency virus infection: the Viral Activa-

tion Transfusion Study a randomized controlled trial. JAMA

2001;285:1592-601.

11. Recommendation and licensure requirements for leuko-

cyte-reduced blood products. Guidance document.

Rockville, MD: FDA, May 29, 1996. Available at http://

www.fda.gov/cber/bldmem/mem52996.pdf.

12. Menitove JE, ed. Standards for blood banks and transfusion

services. 19th ed. Bethesda: AABB, 1999.

In addition to the authors, the VATS is the responsibility ofthe following persons and is funded by the National Heart,Lung, and Blood Institute (contract numbers for each siteare listed in parentheses below):Clinical Sites:

Case Western Reserve University (N01-HB-57115):Michael Lederman, Roslyn Yomtovian, Michael Chance,Donna Hendrix

Georgetown University (N01-HB-57116): Princy Kumar,S. Gerald Sandler, Dawn Rumsey, Karyn Hawkins

Miriam Hospital/Brown University (N01-HB-57117):Timothy P. Flanigan, Joseph D. Sweeney, Maria Mileno,Melissa Di Spigno, Michelle Dupuis

Mt. Sinai Medical Center (N01-HB-57118): Henry S.Sacks, Kala Mohandas, Frances R. Wallach, Letty Mintz

Ohio State University (N01-HB-57119): Michael F. Para,Melanie S. Kennedy, Jane Russell, David Krugh

University of California, San Diego (N01-HB-57120):Thomas A. Lane, W. Christopher Mathews, Peggy Mollen-Rabwin

BRECHER ET AL.

798 TRANSFUSION Volume 41, June 2001 www.transfusion.org

University of California, San Francisco (N01-HB-57121): Edward L. Murphy, Steven G. Deeks, Maurene Viele,Chaolun Han, Joanne Moore

University of North Carolina (N01-HB-57122): MeeraKelley, Charles van der Horst, Linh Ngo, Laura McClannan(also supported by RR00046)

University of Pittsburgh (N01-HB-57123): John W.Mellors, Deborah K. McMahon, Sharon Riddler

University of Texas Medical Branch, Galveston (N01-HB-57124): David M. Asmuth, Richard B. Pollard, JaniceCurry, Gerald Shulman

University of Washington/Puget Sound Blood Center(N01-HB-57125): Ann C. Collier, Terry Gernsheimer, DeeTownsend-McCall, Jill CorsonCentral Laboratory:

Blood Centers of the Pacific-Irwin Center (N01-HB-57126): Michael P. Busch, Tzong-Hae Lee, W. Lawrence Drew

(UCSF Mt. Zion Medical Center, San Francisco), MeganLaycockCoordinating Center:

New England Research Institutes (N01-HB-57127):Leslie A. Kalish, Jane D. Carrington, Margot S. Kruskall (BethIsrael Deaconess Medical Center), Ruth EisenbudSponsoring Agency:

National Heart, Lung, and Blood Institute: George J.Nemo (project officer), Paul R. McCurdy, Dean FollmannData Safety Monitoring Board:

Jeffrey McCullough (chair) (University of Minnesota),Victor DeGruttola, Peter Frame, Janice G. McFarland,Ronald T. Mitsuyasu, Elizabeth J. Read, Dorothy E. VawterSteering Committee Chair:

Paul V. Holland (Sacramento Medical FoundationBlood Centers)