residual wbc subsets in filtered prestorage rbcs

B L O O D C O M P O N E N T S

Residual WBC subsets in filtered prestorage RBCs

Hiromichi Ariga, Tzong-Hae Lee, Megan E. Laycock, Beth A. Mohr, Leslie A. Kalish,

Roslyn Yomtovian, Terry Gernsheimer, and Michael P. Busch, for the Viral Activation

Transfusion Study

BACKGROUND: New-generation RBC filters reduceWBC concentrations by 4 to 5 logs and may prevent ordecrease transfusion complications such as HLA alloim-munization, nonhemolytic febrile reaction, and transfu-sion-transmitted infections. The residual level of WBCsubsets may influence efficacy of WBC reduction forpreventing various complications. This study analyzedsubsets of residual WBCs in WBC-reduced RBC com-ponents prepared for a large, multicenter prospectivestudy.STUDY DESIGN AND METHODS: The Viral ActivationTransfusion Study (VATS) assessed the impact of WBCreduction in HIV-1-infected patients undergoing RBCtransfusion. WBC-reduced RBC from 11 clinical siteswith variable filtration practices were sorted into “low,”“middle,” and “high” groups based on residual WBCconcentration. Subsets were isolated from units by im-munocapture (anti-CD4-, anti-CD8-, anti-CD15-, andanti-CD19-coated magnetic beads) and quantified byPCR amplification.RESULTS: After validation studies confirming test meth-odology, 105 VATS WBC-reduced RBC samples wereanalyzed. Concentrations of subsets in low and middleresidual WBC groups were very low in contrast to rela-tively high concentrations in the high group. Althoughhighly significant differences were identified between themiddle and high groups for total WBCs and all subsets,no single subset predominated.CONCLUSION: These results suggest that overall effi-cacy of WBC filtration correlates with removal of WBCsubsets.

Universal WBC reduction of blood componentssoon after collection and processing is in-creasingly being adopted internationally as anew standard in blood component manufac-

turing.1-3 In the US, the proportion of blood that is WBCreduced has increased from approximately 20 percent to70 percent over the past several years. In 1998, the FDABlood Products Advisory Committee recommended tran-sition to universal WBC reduction, and the AdvisoryCommittee on Blood Safety and Availability recom-mended implementation at its January 2001 meeting.

New-generation polyester filters are capable of re-ducing WBCs by up to 4 to 5 logs. The FDA currentlymandates that RBC products labeled as WBC reduced beprepared by a method known to achieve a residual WBCcount of less than 5 � 106 WBCs per unit.4 The Council ofEurope has established a standard of 90 percent of WBC-reduced RBC products containing less 1 � 106 WBCs.5 A

ABBREVIATION: VATS = Viral Activation Transfusion Study.

From the Department of Medicine, Neonatal Intensive Care

Unit, Fukushima Medical University, Fukushima, Japan; Blood

Centers of the Pacific, San Francisco, California; New England

Research Institutes, Watertown, Massachusetts; Blood Bank

Transfusion Medicine Service, University Hospitals of Cleve-

land, Cleveland, Ohio; the Department of Pathology, Case

Western Reserve University, Cleveland, Ohio; University of

Washington School of Medicine, Seattle, Washington; Puget

Sound Blood Center, Seattle, Washington; the Department of

Laboratory Medicine, University of California, San Francisco,

California; and Blood Systems, Inc., Scottsdale, Arizona.

Address reprint requests to: Michael P. Busch, MD, PhD,

Vice President, Research, Blood Centers of the Pacific,

270 Masonic Ave., San Francisco, CA 94118; e-mail:

[email protected].

Supported by contracts N01-HB-57126, N01-HB-57127,

N01-HB-57115, and N01-57125 from the National Heart Lung

and Blood Institute (National Institutes of Health), Bethesda,

Maryland.

Received for publication May 2, 2002; revision received

July 25, 2002, and accepted August 13, 2002.

TRANSFUSION 2003;43:98-106.

98 TRANSFUSION Volume 43, January 2003

recent draft of an FDA guidance document also proposesthis level, as well as detailed guidelines for implementa-tion and QC of WBC-reduction programs.6 Despite theserequirements, proposals, and recommendations, the 11Viral Activation Transfusion Study (VATS) sites demon-strated significant differences in the extent of WBC re-moval with 0.8 percent exceeding 5 � 106 WBCs per unitand 8.3 percent exceeding 1 � 106 WBCs per unit.7

Although WBC reduction clearly reduces transfusioncomplications,8,9 residual WBCs in WBC-reduced filteredRBC concentrates may cause complications in recipients,including HLA alloimmunization in presensitized recipi-ents,10 GVHD,11 and transfusion-transmitted infec-tions.12 The types of residual WBCs in WBC-reducedblood influence the potential for adverse reactions. Manytransfusion-associated complications may in fact be at-tributable to specific WBC subsets. For example, CMVis predominantly harbored in monocytes and granulo-cytes (CD15+ cells),12 B cells are the predominant reser-voir of EBV and human herpes virus-8 (HHV-8), and Tlymphocytes are associated with HIV and HTLV-I and -IItransmission.13 T cells are also responsible for GVHD.Peripheral blood progenitor cells appear to mediate mi-crochimerism and may play a role in transfusion-inducedimmunomodulation.14 Recently, B cells and dendriticcells have been implicated as physical carriers of prionsfrom peripheral inoculation sites to the central nervoussystem in diseases such as variant CJD.15 Establishing theactual distribution of residual WBCs in filtered units willcontribute to an understanding of the clinical risk ofWBC-reduced blood components. This study was under-taken to analyze the residual donor WBC subsets presentin filtered RBC components utilized in a large, multi-center clinical study.

MATERIALS AND METHODSStudy samplesThe participating clinical facilities in this study were the11 sites of the VATS, a randomized, double-blinded pro-spective study that assessed the impact of WBC reductionof transfused RBCs on survival, occurrence of opportu-nistic infections, and HIV viral load in anemic patientswith HIV infection. Details of the design, goals, and pri-mary outcomes of this study have been reported else-where.16,17 The VATS protocol and procedures were ap-proved by Committees on Human Research from allparticipating institutions. Patients randomized to WBCreduction were required to receive RBCs stored for nomore than 14 days that had undergone prestorage WBCreduction by a method known to achieve a postfiltrationWBC count of less than 5 � 106 per unit. No specificrestrictions were placed on filter type or filtration meth-odology. Study sites filtered units within approximately72 hours of collection using commercial filters (Baxter

Sepacell R500, Deerfield, IL; Pall Leukotrap and BPF4,Pall Biomedical, East Hills, NY) filters.7 All platelet trans-fusions provided to patients in both treatment groupswere WBC reduced and, preferably, comprised of apher-esis components. Residual WBC quantitation was per-formed in the VATS Central Laboratory (Blood Centers ofthe Pacific, San Francisco, CA) using aliquots of bloodcomponents collected from each WBC-reduced unit atthe time of transfusion from tubing used to transfer theunit to a blinded study bag. These aliquots were thenfrozen within approximately 24 hours after transfusionand shipped to the Central Laboratory, where they werebatch tested for residual WBC count by quantitative PCR.Any unit with a residual WBC count greater than 1.67 �

104 per mL (i.e., >5 � 106/unit) was considered to be astudy filter failure; of these, any unit with an off-scaleresidual WBC count (i.e., >8 � 104 copies/mL) was rou-tinely retested in dilution.

Residual WBC subset studyDonor units, stratified by the concentration of residualWBCs present, were randomly selected for WBC subsettesting. The units were first categorized into three re-sidual WBC groups (“low,” “middle,” “high”) as follows.All WBC-reduced RBC donor units for which residualWBC data were available as of October 1999 were rankedfrom lowest to highest residual WBC concentration.Twenty-five units were chosen randomly from those withthe 50 lowest residual counts (<5000 cells/unit) to formthe low group. For the high group, we considered the 50units with the highest residual WBC counts (>1,930,000cells/unit). Of these 50, 15 were filter failures and 35 werenot. We selected as many of the 15 filter failures as wereavailable (n = 5) plus a random subset of 20 from theremaining 35, for a total of 25. The middle residual WBCgroup (n = 55) consisted of 5 units per site randomlyselected from the remaining 1664 units. In January 2000,when final VATS data became available, there were threemore units in the high range and 100 more in the middlerange. For purposes of this analysis, these were includedin our summary of “unselected” units.

Assay validation studiesBefore selecting study samples for WBC subset analysis,several special studies were conducted to evaluate andvalidate our methodology. Routine, nonstudy donationsused for these experiments were either quarantined or ofinsufficient volume for transfusion. Blood samples werecollected as part of the routine donation process fromhealthy whole-blood donors after giving informed con-sent including permission to use residual or nontransfus-able blood for research purposes. The informed consentprocedures and wording were approved by the Commit-tee on Human Research of the University of California,

WBC SUBSETS IN FILTERED RBCs

Volume 43, January 2003 TRANSFUSION 99

San Francisco. All testing was performed on unlinked do-nor specimens, so that separate Committee on HumanResearch approval for this study protocol was not required.

Carrier cell study. This study was conducted to testthe effect of increasing carrier cells on the recovery oftarget WBCs. Units from two male donors were filtered.WBCs from a female donor were counted (Z-seriesCoulter Counter, Beckman Coulter, Fullerton, CA). Thefiltered male donor blood was mixed thoroughly andthree 2-mL aliquots were prepared per donor. The firstaliquot was left unspiked, to serve as the control. Thesecond and third tubes were spiked, with 2 � 104 and 2� 105 female WBCs, respectively. All tubes were mixed bygently rocking for 15 minutes before separating each2-mL aliquot of blood into four 500-�L aliquots. CD4,CD8, CD15, and CD19 subpopulations were purified andextracted separately from each 500-�L aliquot followingthe Dynal magnetic bead procedure described below.Each sample was amplified in duplicate using a humanY-chromosome-specific primer pair.

Spiking study. This study was undertaken to deter-mine the efficacy of the methodology to recover WBCsusing known quantities. Three units of nontransfusablewhole blood were collected and mixed thoroughly bygentle rocking. Before filtration, 5 mL was collected fromeach unit. After filtration of the remaining volume, theunfiltered WBCs were counted (Coulter counter, Beck-man Coulter) and then used to spike the filtered material(1 � 104 WBCs/1 mL of filtered blood). We prepared four500-�L aliquots of spiked filtrate per sample and four500-�L aliquots of unspiked filtered filtrate as controls.We used the Dynal magnetic bead protocol to performseparate extractions of CD4+, CD8+, CD15+, and CD19+cells separately from the spiked and nonspiked filtrates.

Comparison of fresh and frozen filtered blood. Thisstudy had two goals: 1) to verify the methodology usingfrozen samples because the samples to be tested wouldbe frozen; and 2) to check for a differential filter effect ondifferent subsets. In this study, 35 nontransfusable donorunits were collected for filtration using one of three com-mercial filters (Sepacell 500 filter, Baxter Healthcare [n =10]; RCXL filter [n = 10] and BPF4 filter [n = 15] PallBiomedical). The filtration procedures were performedaccording to the manufacturers’ instructions. Eight ali-quots containing 0.5 mL of filtrate were prepared persample immediately after filtration. Four of the aliquotsfrom each of the 35 filtrates were frozen at –80�C, and theother tubes were left unfrozen to serve as the source fordata from fresh whole blood. Each sample was mixedthoroughly before isolating the WBC subpopulations.

WBC subset isolation and DNA preparationBlood samples were either processed fresh or, if frozen,thawed at room temperature and well mixed. Magnetic

beads (Dynal, Lake Success, NY) coated with differentmonoclonal antibodies including anti-CD4, anti-CD8,anti-CD15, and anti-CD19 were added into 125 �L ofwhole blood and incubated separately, and each mono-clonal antibody-coated bead was rotated for 20 minutesat 4�C according to manufacturer’s instructions. After cal-culating the volume of beads being used, the beads werewashed twice, first with 1 mL PBS and then with 1 mLPCR solution A (100 mM KCl; 10 mM Tris HCl, pH 8.3; 2.5mM MgCl2; Cell Culture Facility, University of California,San Francisco, CA). Then the beads with bound specificWBC subsets were resuspended in 50 �L of solution Aand 50 �L of PCR lysis solution B (10 mM Tris HCl, pH8.3; 2.5 mM MgCl2; 1% Tween-20; 1% Nonidet P-40; and0.4 mg/mL proteinase K; Cell Culture Facility) and incu-bated at 60�C for 90 minutes with vortexing every 20 min-utes, followed by incubation at 95�C for 2 hours.

PCR for quantitation of human Y-chromosome andHLA DQ-ALPHA-specific alleleQuantitative allele-specific amplifications of the humanY-chromosome and HLA DQ-alpha sequence were per-formed using a thermal cycler (PE 9600 thermal cycler,Applied Biosystems, Foster City, CA), as described previ-ously.18 Briefly, the PCR reaction mixture consisted of 100mM KCl, 20 mM Tris HCl (pH 8.3), 2 mM Mg Cl2, 1 pmolper �L of each primer and 0.04 U per �L of Thermalase(IBI, New Haven, CT). Fifty �L of the PCR reaction mix-ture was added to 25 �L of each DNA sample and ampli-fied for 35 cycles. Each cycle was initiated at 95�C for 30seconds followed by amplification at 56�C for 1 minuteand then at 72�C for 2 minutes. After completion of the 35cycles, samples were amplified for an additional 10 min-utes at 72�C. All samples were processed and tested induplicate with results averaged. Duplicate standardcurves comprised of 10-fold serial dilutions of humanDNA concentrates (103,102,101,100) were analyzed in par-allel with samples.

Liquid hybridization and quantitation ofPCR productsSpecific amplified products were detected using liquidhybridization with 32P-labeled probes. The probe wasend-labeled at 37�C for 1 hour in a 40-�L solution con-sisting of 7 mmol Tris HCl, pH 7.6; 10 mmol MgCl2; 15mmol dithiothreitol; 10 U T4 polynucleotide kinase (NewEngland Biologicals, Cambridge, MA); 40 uCi 32P-adenosine triphosphate (32P-ATP; Dupont, Wilmington,DE); and 30 pmol probe. For hybridization, 10 �L ofprobe mixture was added to 20 �L of each postamplifi-cation specimen; this step was followed by denaturationfor 5 minutes at 95�C and hybridization for 5 minutes at59�C. Ten �L of loading buffer (0.25% bromphenol blue,0.25% xylene cyanol, and 30% glycerol) were added to

ARIGA ET AL.


each hybridized sample, which then underwent 6-per-cent PAGE at 12.5 V per cm. The gel was exposed toautoradiographic film (XAR-5, Kodak, Rochester, NY) us-ing an enhancing screen at room temperature for 30 min-utes, 2 hours, or overnight. Selected autoradiographswere analyzed using an image analyzer and software(Millipore BioImage Analyzer and Whole Band Analyzerapplication software, Millipore, Ann Arbor, MI).

Statistical analysesValidation studies. For the comparison of paired

fresh and frozen samples, the differences between copynumbers (frozen minus fresh) were tested for a system-atic shift from zero using Wilcoxon’s signed-rank test.Samples for which both the frozen and fresh counts werezero were excluded from this analysis.

Residual WBC subset study. The Wilcoxon’s rank-sum test was used to compare WBC subset distributionsof the middle versus low and the high versus middlegroups, and for comparing the distribution of residualWBCs in donor units selected for this study versus non-selected units.

RESULTS

Validation studiesDue to the very low concentration of total WBCs andWBC subsets in filtered blood, it was necessary to firstverify the immunocapture PCR methods that we had pre-viously validated for characterization of microchime-rism.19 To address the potential problem of low-efficiency recovery, we developed a strategy based on theaddition of carrier cells to filtered blood, similar to amethod previously used to enhance recovery of totalWBCs in filtered units.20 In the carrier cell study, we com-pared cell recovery after WBC reduction in blood spikedwith 104 and 105 carrier cells (female) per mL of filtered

male blood against recovery in the unspiked preparation.A comparison of recovery among subpopulations (CD4,CD8, CD15, and CD19) using the Dynal magnetic beadprotocol showed no improvement by addition of carriercells (data not shown).

A second experiment focused on recovery of spikedWBCs from filtered blood, in which unspiked blood wascompared to spiked filtered blood for each subpopula-tion extracted. Figure 1 displays the overall results fromthis spiking study, in which each panel shows data froma different donor. The recovery of WBC subsets from theunspiked filtrate ranged from 0 to 48 cells per mL (CD15+being the highest). In the spiked preparation, the recov-ery of each subset was consistent with the total numberof spiked cells and distribution of subsets in normal do-nor blood. Recovery of total WBCs (sum of all 4 subsets)was 6.89 � 103, 9.34 � 103, and 6.72 � 103 for the threedonor samples, yielding a mean of 7.65 � 103 or 76.5percent recovery of the 104 cells that were spiked intoeach of the filtrates. Based on this high recovery percent-age, we determined that our study of clinical sampleswould be feasible.

A third validation study centered on an evaluation offresh versus frozen WBC-reduced preparations. This wasimportant because the VATS clinical samples to be stud-ied subsequently were frozen whole-blood samples. Thefresh-frozen study included comparison of 35 pairedfresh and frozen blood preparations derived from wholeblood filtered using three widely used WBC-reduction fil-ters. As summarized in Table 1 and Fig. 2, no significantdifference among WBC subsets was identified betweenfresh and frozen preparations of filtered RBCs using anyof the three filters (Sepacell-500, RCXL, or BPF4 filters).Although the four WBC subsets evaluated were detectedin filtered blood, there was a predominance of CD15+cells, and CD8+ cells were least represented and almostundetectable in BPF4 filtrates. Overall, these data vali-dated our plan to test frozen samples.

Fig. 1. Results of spiking study designed to determine recovery of spiked WBCs into filtered blood. A side-by-side comparison of

unspiked (light gray) and spiked filtered blood (dark gray) for each subpopulation extracted for three independent experiments

is shown. The recovery of WBCs for the unspiked filtrate ranges from 0 to 48 cells (CD15+). From the 104 cells spiked into each

of the filtrates, recovery ranged from 6.72 � 103 to 9.34 � 103, yielding a mean of 7.65 � 103 or 76.5 percent for all three

samples.



Residual WBC subset studyA total of 105 WBC-reduced donor RBC specimens wereselected for our primary study from the 1764 availableVATS units. Residual WBC counts from an additional 103units became available for analysis after selection. Table2 displays the number of units in each group for thisstudy and for the full VATS sample of 1867 units. Therewere no significant differences in residual WBC distribu-tions by group between selected and nonselected units,providing evidence that the study sample is representa-tive of the full VATS sample (data not shown).

Table 3 and Fig. 3 present the WBC subset distribu-

tions for the selected specimens, by residual WBC group.The percentiles of each subset for the low group were lessthan or equal to that of the middle group, but there wasbroad overlap in the distributions. All p values were mar-ginally significant (p values ranged from 0.03 to 0.07).When subsets for the high group were compared withthose of the middle group, highly significant differencesemerged for all subsets, as well as for the sum of the foursubsets (p values ranged from 0.0001 to 0.001). Thesedifferences were highly significant even when the filterfailures were excluded from the high group. Figure 3, agraph of the subset distributions by group, illustrates thatthere is practically no overlap in the interquartile rangesbetween the middle and high groups for each subset.

We next addressed the issue of differential represen-tation of subsets within each group. Figure 3 shows thatthe four subsets are represented in approximately thesame proportions within each group, indicating that nosubset preferentially escapes filtration, regardless of theefficiency of filtration.

DISCUSSIONAs WBC-reduction implementation has expanded, manystudies have focused on relative efficacy of different com-mercial filters for depleting WBCs.21 Very few groups,however, have investigated the characteristics of WBCs infiltered blood. Because particular WBC subsets remainingin filtrates may influence the rate of residual transfusioncomplications from WBC-reduced components, it is im-portant to analyze the relative depletion of WBC subsetsin filtered blood units.

Rider et al.22 compared WBC subset depletion after

TABLE 1. Comparison between fresh and frozen preparations of residual WBC subsets in filtered RBCs

Filter Number

CD4 CD8 CD15 CD19

Fresh Frozen Fresh Frozen Fresh Frozen Fresh Frozen

Sepacell, mean (range) 10 4.5 (0-27) 3.2 (0-16) 3.8 (0-21) 8.3 (0-18) 19.3 (0-32) 37.7 (0-104) 3.2 (0-32) 5.2 (0-18)RCXL 10 2 (0-8) 4.4 (0-8) 2.8 (0-8) 2.4 (0-8) 22 (4-132) 6 (0-12) 2.8 (0-8) 3.6 (0-8)BPF4 15 14.4 (0-72) 17.1 (0-152) 0 (0) 0 (0) 27.2 (0-144) 6.4 (0-40) 12.3 (0-136) 8.5 (0-80)

Fig. 2. Combined results from Sepacell (n = 10), RCXL

(n = 10), and BPF4 (n = 15) filters. Data compare fresh and

frozen preparations from the three different filters. All four

subsets were detected at a very low number, and no signifi-

cant difference between fresh and frozen preparations was

observed. The horizontal lines that form the bottom and top

of the boxes represent the 25th and 75th percentiles, respec-

tively, and the horizontal line in between is the median. (If

no separate median line is shown, the median is zero. For

CD8, the 25th, 50th, and 75th percentiles are all zero.) The

vertical lines connected to the boxes extend to the 10th and

90th percentiles. Outlying values are depicted as individual

dots. Twelve extreme values less than –40 or greater than 40

are not shown but are included in the analysis.

TABLE 2. Number of units per residual WBC groupfor the residual WBC subset study (n = 105) and for

all available VATS results (n = 1867)

ResidualWBC group

Residual WBCdistribution study All available VATS results

n Percent n Percent

Low 25 23.8 50 2.7Middle 55 52.4 1764 94.5High 25* 23.8 53** 2.8Total 105 100 1867 100

* Includes five filter failures.** Includes 15 filter failures.

ARIGA ET AL.


filtration after brief storage at room temperature in 20units from 26 healthy volunteer donors. This study foundthat WBC levels in all units after filtration by eitherWBF-1 or RS2000 filters were within acceptable WBC-reduction ranges, with granulocytes and monocytesequally dominant followed by lymphocytes. This wasconsistent with previous observations that lymphocyteswere most susceptible to physical removal.

A study by Wegener et al.23 evaluated 10 fresh RBCconcentrates filtered (Leucoflex LCR-4) within 24 hoursafter donation. In this study, the distribution of WBC sub-sets in the filtrate favored granulocytes (91%) and mono-cytes (5.8%), with T lymphocytes composing only 1.9 per-cent and B lymphocytes 0.83 percent of the WBC filtratesubsets. It was speculated that B and T lymphocytes werepreferentially depleted by the combination of buffy-coatremoval and filtration.

In a study by Johnson et al.,24 which quantified WBCsubsets in RBC units filtered using the RS2000 filter at 4�Cand at 22�C, granulocytes again predominated at bothtemperatures. At 4�C, however, B cells were the mostplentiful after granulocytes, followed by monocytes and Tcells, respectively; at 22�C, granulocytes were followed bymonocytes, B cells, and T cells.

In each of these studies, CD15+ cells (granulocyteswith or without monocytes) predominated despite a widerange of filters and WBC-reduction techniques. Similarly,in our preliminary study of fresh versus frozen bloodsamples prepared using several of the most commonlyused filters (i.e., Sepacell-500, RCXL, and BPF4), we foundthat granulocytes (CD15+ cells) remained the predomi-nant subset after WBC reduction.

In contrast, a recent study conducted by Penningtonet al.,25 where real-time PCR was used to amplify themRNA of four different WBC subsets (CD3, CD19, CD14,

and CD66), found no significant difference in efficiencyin the removal of WBC subsets between the two whole-blood filters used (Pall WBF2 and Baxter RZ2000). Bothfilters followed the same pattern of subset removal (CD14> CD3 > CD66 > CD19), but because Pennington et al.25

amplified mRNA, it was not possible for them to performabsolute quantification; instead they showed only rela-tive reduction of mRNA by comparing samples beforeand after filtration. Given that our study measured abso-lute quantity of residual WBC subsets after filtration andthe fact that only one antigen (CD19) was included inboth studies, it is difficult to draw clear conclusions froma comparison of these studies.

The previous studies and our validation study wereall small, highly controlled laboratory-based studies offreshly filtered blood. To our knowledge, ours is the onlystudy to quantitate residual WBC subsets on a represen-tative sample of WBC-reduced RBC units from a large,multicenter clinical study. Our study, employing a sensi-tive and validated assay, shows that all subsets areequally represented after prestorage filtration and thatCD15+ cells do not predominate. The most significantdifference identified is between filtered units with a me-dium residual concentration and those with a high re-sidual concentration.

This lack of CD15+ cell predominance in the VATSsamples may relate to the fact that the tested samples hadbeen stored for a median of 9 days and up to approxi-mately 14 days before sampling at the time of transfu-sion. Storage has been shown by Fiebig et al.26 to causecell degradation, beginning almost immediately after 4�Cstorage with the most pronounced effect being loss ofgranulocytes. The findings reported by Fiebig et al.26 andour results suggest that the granulocytes reported as pre-dominant after filtration of fresh blood by other groups

TABLE 3. Percentiles for residual WBC subsets by residual WBC group, residual WBC subset study (n = 105)

Residual WBC subset(copies/mL)

ResidualWBC group

Percentile p value†

10th 50th 90th Middle vs. low High* vs. middle

CD4 Low 0 0 32 0.06 0.0001Middle 0 8 40High* 0 76 389.8

CD8 Low 0 0 34.5 0.07 0.0001Middle 0 16 127High* 18.5 250 10,078.1

CD15 Low 0 16 46 0.07 0.0001Middle 0 32 91High* 32 127.5 3235.5

CD19 Low 0 16 128.5 0.03 0.002Middle 0 32 588.3High* 0 644 11,847.5

Total‡ Low 0 48 221.5 0.03 0.0001Middle 0 111.5 1179.5High* 113.8 2353.2 22,623.5

* Includes five filter failures.† Wilcoxon’s rank-sum test comparing middle to low or high to middle.‡ Total is calculated as the sum of CD4, CD8, CD15, and CD19.



were likely nonviable and degraded upon subsequentstorage.

A study by Roback et al.27 that evaluated the differ-ential effect of storage on filtered and unfiltered WBCsubsets partially supports this storage effect explanation.In the Roback study, 14 packed RBC units were eitherfiltered by Sepacell R-500(II) and stored at 4�C for 42 daysor were stored without filtration; all units were pheno-typed before filtration, directly after filtration, and seriallythereafter. Although granulocytes were predominant af-ter filtration in this study, the granulocytes and otherWBC subsets all showed ongoing loss during storage. Thefact that the CD15+ predominance was not fully elimi-nated by the storage effect in the Roback study may bedue to the study’s more controlled nature and smallersample size.

Based on samples collected for the VATS study andfiltered according to participating study site methodolo-gies, our primary findings lead us to conclude that re-moval of WBC subsets correlates with overall efficacy ofWBC filtration and that no subset preferably escapes fil-tration. This implies that filter failures result in failure toadequately remove multiple WBC subsets, each of whichmay carry its own risk profile for transfusion recipients.As filtration methodology advances, evaluations of subsetremoval in addition to total residual WBC content shouldnonetheless be considered in filter validation studies.

ACKNOWLEDGMENTS

The Viral Activation Transfusion Study is the responsibility of

the following individuals. We thank the patients who partici-

pated in the study for their efforts, and the transfusion service

personnel and nurses at each medical center, without whose

assistance the study could not have been accomplished.

Clinical sitesCase Western Reserve University, Cleveland, OH (N01-

HB-57115): Michael Lederman, MD; Roslyn Yomtovian,MD; Michael Chance, RN; Donna Hendrix, RN.

Georgetown University, Washington, DC (N01-HB-57116): Princy N. Kumar, MD; S. Gerald Sandler, MD;Karyn Hawkins, RN.

Miriam Hospital/Brown University, Providence, RI(N01-HB-57117): Timothy P. Flanigan, MD; Joseph Swee-ney, MD; Maria D. Mileno, MD; Melissa Di Spigno, RN;Michelle Dupuis, MT(SSB).

Mt. Sinai School of Medicine, New York, NY (N01-HB-57118): Henry S. Sacks, PhD, MD; Kala Mohandas,MD; Frances R. Wallach, MD; Letty Mintz, ANP.

Ohio State University, Columbus, OH (N01-HB-57119): Michael F. Para, MD; Melanie S. Kennedy, MD;Jane Russell, RN; Dave Krugh, MT.

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

Fig. 3. Residual WBC subset distributions by residual WBC group. The horizontal lines that form the bottom and top of the

boxes represent the 25th and 75th percentiles, respectively, and the horizontal line in between is the median. The vertical lines

connnected to the boxes extend to the 10th and 90th percentiles. Outlying values are depicted as individual dots. Note that the

median for CD4 and CD8 in the low group is 0 copies per mL. The high group includes five filter failures.

ARIGA ET AL.


University of California, San Francisco, CA (N01-HB-57121): Edward L. Murphy, MD, MPH; Steven G. Deeks,MD; Maurene Viele, MD; Chaolun Han, MD; JoanneMoore, MT(ASCP) SBB.

University of North Carolina, Chapel Hill, NC (N01-HB-57122): Charles van der Horst, MD; Meera Kelley,MD; Mark E. Brecher, MD; Ling Ngo, FNP.

University of Pittsburgh, Pittsburgh, PA (N01-HB-57123): John W. Mellors, MD; Darrell J. Triulzi, MD;Deborah K. McMahon, MD; Sharon Riddler, MD.

University of Texas Medical Branch, Galveston, TX(N01-HB-57124): David M. Asmuth, MD; Richard B. Pol-lard, MD; Janice Curry, PAC; Gerald Shulman, MD.

University of Washington/Puget Sound Blood Cen-ter, Seattle, WA (N01-HB-57125): Ann Collier, MD; TerryGernsheimer, MD; Dee Townsend-McCall, RN; Jill Cor-son, RN.

Central laboratoryBlood Centers of the Pacific, San Francisco, CA (N01-HB-57126): Michael P. Busch, MD, PhD; Tzong-Hae Lee, MD,PhD; W. Lawrence Drew, MD, PhD (UCSF Mt Zion Medi-cal Center, San Francisco); Megan Laycock.

Coordinating centerNew England Research Institutes, Watertown, MA (N01-HB-57127): Leslie A. Kalish, ScD; Susan F. Assmann, PhD;Jane D. Carrington, RN, BS; Margot S. Kruskall, MD (BethIsrael Deaconess Medical Center, Harvard MedicalSchool, Boston, MA); Ruth Eisenbud, BA.

Sponsoring agencyNational Heart, Lung and Blood Institute, National Insti-tutes of Health, Bethesda, MD: George J. Nemo, PhD,project officer; Paul R. McCurdy, MD; Dean Follmann,PhD.

Steering committee chairPaul V. Holland, MD, Sacramento Medical FoundationBlood Centers, Sacramento, CA.

Data Safety Monitoring BoardJeffrey McCullough, MD (chair), University of Minnesota,Minneapolis, MN; Victor DeGruttola, ScD; Peter Frame,MD; Janice G. McFarland, MD; Ronald T. Mitsuyasu, MD;Elizabeth J. Read, MD; Dorothy E. Vawter, PhD.

REFERENCES

1. Guidelines on the clinical use of leucocyte-depleted blood

components. British Committee for Standards in Haema-

tology, Blood Transfusion Task Force. Transfus Med 1998;

8:59-71.

2. Danish recommendations for the transfusion of leuko-

cyte-depleted blood components. The Danish Society of

Clinical Immunology. Vox Sang 1996;70:185-6.

3. American Association of Blood Banks Association Bulletin,

Leukocyte Reduction, 99-7, 1999.

4. Food and Drug Administration. Recommendations and

Licensure Requirements for Leukocyte-Reduced Blood

Products. http://www.fda.gov/cber/bldmem/

mem52996.txt. 1996.

5. Europe Council. Guide to the use, preparation and quality

control of blood components. 3rd ed. Council of Europe

Press, 1992.

6. Food and Drug Administration. Pre-storage leukocyte re-

duction of whole blood and blood components intended

for transfusion. http://www.fda.gov/cber/gdlns/preleuk

.htm. 2001.

7. Yomtovian R, Gernsheimer T, Assmann SF, et al. WBC

reduction in RBC concentrates by prestorage filtration:

multicenter experience. Transfusion 2001;41:1030-6.

8. Federowicz I, Barrett BB, Andersen JW, et al. Character-

ization of reactions after transfusion of cellular blood

components that are white cell reduced before storage.

Transfusion 1996;36:21-8.

9. Heddle NM. Febrile nonhemolytic transfusion reactions

to platelets. Curr Opin Hematol 1995;2:478-83.

10. Leukocyte reduction and ultraviolet B, irradiation of

platelets to prevent alloimmunization and refractoriness

to platelet transfusions. The Trial to Reduce Alloimmuni-

zation to Platelets Study Group. N Engl J Med 1997;337:

1861-9.

11. Busch M, Lee TH, Heitman J. Allogeneic leukocytes but

not therapeutic blood elements induce reactivation and

dissemination of latent human immunodeficiency virus

type 1 infection: implications for the transfusion support

of infected patients. Blood 1992;80:2128-35.

12. Bowden RA, Slichter SJ, Sayers M, et al. A comparison of

filtered leukocyte-reduced and cytomegalovirus (CMV)

seronegative blood products for the prevention of transfu-

sion-associated CMV infection after marrow transplant.

Blood 1995;86:3598-603.

13. Lal RB, Owen SM, Rudolph DL, et al. In vivo cellular tro-

pism of human T-lymphotropic virus type II is not re-

stricted to CD8+ cells. Virology 1995;210:441-7.

14. Blajchman MA, Dzik S, Vamvakas EC, et al. Clinical and

molecular basis of transfusion-induced immunomodula-

tion: summary of the proceedings of a state-of-the-art

conference. Transfus Med Rev 2001;15:108-35.

15. Murphy MF. New variant Creutzfeldt-Jakob disease

(nvCJD): the risk of transmission by blood transfusion

and the potential benefit of leukocyte-reduction of blood

components. Transfus Med Rev 1999;13:75-83.

16. Busch MP, Collier A, Gernsheimer T, et al. The Viral Acti-

vation Transfusion Study (VATS): rationale, objectives,

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

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



red blood cell transfusions in patients with anemia and

human immunodeficiency virus infection: the Viral Acti-

vation Transfusion Study: a randomized controlled trial.

JAMA 2001;285:1592-601.

18. Lee TH, Donegan E, Slichter S, Busch MP. Transient in-

crease in circulating donor leukocytes after allogeneic

transfusions in immunocompetent recipients compatible

with donor cell proliferation. Blood 1995;85:1207-14.

19. Lee TH, Paglieroni T, Ohto H, et al. Survival of donor leu-

kocyte subpopulations in immunocompetent transfusion

recipients: frequent long-term microchimerism in severe

trauma patients. Blood 1999;93:3127-39.

20. Lee TH, Stromberg RR, Heitman J, et al. Quantitation of

residual white cells in filtered blood components by poly-

merase chain reaction amplification of HLA DQ-A DNA.


21. 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.

22. Rider JR, Want EJ, Winter MA, et al. Differential leucocyte

subpopulation analysis of leucodepleted red cell products.

Transfus Med 2000;10:49-58.

23. Wegener S, Marschall M, Schnabl J, et al. White cell sub-

sets in filtered red blood cell concentrates. Transfus Sci

2000;23:29-32.

24. Johnson J, Mijovic V, Brozovic B. Evaluation of a new fil-

ter for leucocyte depletion of blood. J Clin Pathol 1983;36:

1200-1.

25. Pennington J, Garner SF, Sutherland J, Williamson LM.

Residual subset population analysis in WBC-reduced

blood components using real-time PCR quantitation of

specific mRNA. Transfusion 2001;41:1591-600.

26. Fiebig E, Lane TA. Effect of storage and ultraviolet B irra-

diation on CD14-bearing antigen-presenting cells (mono-

cytes) in platelet concentrates. Transfusion 1994;34:846-

51.

27. Roback JD, Bray RA, Hillyer CD. Longitudinal monitoring

of WBC subsets in packed RBC units after filtration: im-

plications for transfusion transmission of infections.


ARIGA ET AL.


residual wbc subsets in filtered prestorage rbcs

Documents