32391.immuno 17 3 - red cross blood...transfusion history was also obtained, which included the...

ImmunohematologyJ O U R N A L O F B L O O D G R O U P S E R O L O G Y A N D E D U C A T I O N

V O L U M E 1 7, N U M B E R 3, 2 0 0 1

This Issue ofImmunohematology

Is Supported bya Contribution From

Dedicated toEducation in the Field of

Blood Banking

A

Comprehensive

Laboratory

Manual

From the publishers of Immunohematology

Featuring—• Over 100 methods—

just about every method used in a reference lab.

• Eleven chapters discussing problems faced by blood group serologists and the procedures and methods that can be used to solve them.

• An extra set of the methods to use at the bench,printed on durable waterproof paper.

• See business reply order card enclosed in this issue or order on the Web at redcross.org/immunohematology

ImmunohematologyMethods andProcedures

ImmunohematologyJ O U R N A L O F B L O O D G R O U P S E R O L O G Y A N D E D U C A T I O N

V O L U M E 1 7, N U M B E R 3, 2 0 0 1

C O N T E N T S63

Development of anti-Bw6 reactivity in patients receiving r-GCSF: a preliminary reportM. F. LEACH AND J. P. AUBUCHON

70Detection of granulocyte antibodies by flow cytometry without the use of pure granulocyte isolates

K. M. KIEKHAEFER, K. M. CIPOLONE, J. L. PROCTER, K. MATSUO,AND D. F. STRONCEK

76Low-incidence MNS antigens associated with single amino acid changes and their susceptibility

to enzyme treatmentM. E. REID AND J. R. STORRY

82PEG adsorption of autoantibodies causes loss of concomitant alloantibody

W. J. JUDD AND L. DAKE

86Selecting an acceptable and safe antibody detection test can present a dilemma

M. R. COMBS AND S. J. BREDEHOEFT

90C O M M U N I C A T I O N S

Letter to the EditorsA hemolytic transfusion reaction due to anti-K undetected

by a LISS antibody screenM. T. FRIEDMAN AND A. P. CARIOTI

91 91Letter From the Editors Errata

Ortho Dedication Vol. 17, No. 2, 2001

92 93 94I N M E M O R I U M I N M E M O R I U M I N M E M O R I U M

WILLIAM C. SHERWOOD, MD JOHN CASE, FIBMS, FIMLS FRED STRATTON, BSC, MD

95 95A N N O U N C E M E N T S A D V E R T I S E M E N T S

97I N S T R U C T I O N S F O R A U T H O R S

Patricia Arndt, MT(ASCP)SBBLos Angeles, California

James P. AuBuchon, MDLebanon, New Hampshire

Malcolm L. Beck, FIBMS, MIBiolKansas City, Missouri

Richard Davey, MDRosslyn, Virginia

Geoffrey Daniels, PhDBristol, United Kingdom

Sandra Ellisor, MS, MT(ASCP)SBBRaritan, New Jersey

George Garratty, PhD, FRCPathLos Angeles, California

Brenda J. Grossman, MDSt. Louis, Missouri

Christine Lomas-Francis, MScAustin, Texas

Gary Moroff, PhDRockville, Maryland

Ruth Mougey, MT(ASCP)SBBCarrollton, Kentucky

John J. Moulds, MT(ASCP)SBBRaritan, New Jersey

Marilyn K. Moulds, MT(ASCP)SBBHouston, Texas

Scott Murphy, MDPhiladelphia, Pennsylvania

Paul M. Ness, MDBaltimore, Maryland

Mark Popovsky, MDBraintree, Massachusetts

Marion E. Reid, PhD, FIBMSNew York, New York

Susan Rolih, MS, MT(ASCP)SBBCincinnati, Ohio

David F. Stroncek, M.D.Bethesda, Maryland

Marilyn J. Telen, MDDurham, North Carolina

PRODUCTION ASSISTANT

Linda Berenato

COPY EDITOR

Sonia E. Connor

ELECTRONIC PUBLISHER

Jane Moore

Immunohematology is published quarterly (March, June, September, and December)by the American Red Cross, National Headquarters, Washington, DC 20006.

The contents are cited in the EBASE/Excerpta Medica and Elsevier BIOBASE/CurrentAwareness in Biological Sciences (CABS) databases.

The subscription price is$30.00 (U.S.) and $35.00 (foreign) per year.

Subscriptions, Change of Address, and Extra Copies:Immunohematology, P.O. Box 40325

Philadelphia, PA 19106Or call (215) 451–4902

Web site: redcross.org/immunohematology

Copyright 2001 by The American National Red CrossISSN 0894-203X

EDITORIAL BOARD

EDITOR-IN-CHIEF

Delores Mallory, MT(ASCP)SBBRockville, Maryland

TECHNICAL EDITOR

Christine Lomas-Francis, MScAustin, Texas

MANAGING EDITOR

Mary H. McGinniss, AB, (ASCP)SBBBethesda, Maryland

MEDICAL EDITOR

S. Gerald Sandler, MDWashington, District of Columbia

I M M U N O H E M A T O L O G Y , V O L U M E 1 7, N U M B E R 3, 2 0 0 1 63

Development of anti-Bw6 reactivityin patients receiving r-GCSF:a preliminary reportM. F. LEACH AND J. P. AUBUCHON

Recombinant granulocyte colony-stimulating factor (r-GCSF) is usedin autologous bone marrow/peripheral blood progenitor celltransplantation (ABMT/PBPC) to shorten the period of neutropenia.As these patients require platelet transfusions, their sera may bemonitored for the presence of platelet/human leukocyte antigen(HLA) antibodies. Sera of some patients on r-GCSF (16 µg/kg/day)became difficult to evaluate in vitro for the presence of HLA andplatelet antibodies because of apparently anomalous reactions insolid phase red cell adherence (SPRCA) assays. SPRCA tests werepositive only when platelets were adhered to the polystyrene platesin the presence of glucose; when other simple sugars were used,the patients’ sera failed to react. HLA Bw6-positive platelets weremore likely than HLA Bw6-negative platelets to be reactive (p <.001). The transfusion of HLA Bw6-positive platelets to patientsdisplaying this in vitro reactivity (positive patients) resulted in a 50percent lower corrected count increment (CCI) than those given topatients without it (negative patients; p = <.001). When alltransfusions were considered, the CCI for those of the positivepatients was decreased 73 percent when compared to the controlpatients (p = .0005). The presence of the antibody also wasassociated with a twofold increase in the number of platelettransfusions given (p = .0005). ABMT/PBPC patients receivingr-GCSF may develop unexpected reactions on SPRCA antibodyscreens and poor responses to transfusion of Bw-6-positiveplatelets. Immunohematology 2001;17:63–69.

Key Words: r-GCSF, marrow/progenitor cell transplant,anti-Bw6

Potent hematopoietic stimulants offer importantclinical benefits in chemotherapy and marrowtransplant patients. Analogs of naturally occurringhormones and cytokines have reduced morbidity andthese patients have shortened hospital stays, but theymay elicit undesirable immunologic side effects. Thisstudy describes the discovery, investigation, andpresumptive resolution of an unexpected serologicfinding associated with use of a common hematopoieticstimulant.

Because of the frequency and clinical significance ofrefractoriness to platelet transfusion associated with thedevelopment of human leukocyte antigen (HLA) and/or platelet-specific alloantibodies, some institutionsregularly screen patients receiving platelet transfusions

to detect the formation of these antibodies and redirecthemotherapy. We encountered a group of patientsundergoing autologous bone marrow/peripheral bloodprogenitor cell transplantation (ABMT/PBPC) whoappeared to develop blunted responses to platelettransfusion while receiving recombinant granulocytecolony-stimulating factor (r-GCSF); all demonstratedanomalous serologic activity in HLA/platelet antibodyscreening associated with development of a novelglucose-dependent antibody. The investigation of thisphenomenon and its resolution is the subject of thisreport.

All hematology-oncology patients (including thoseundergoing ABMT/PBPC) treated at this institutionare routinely tested by a solid phase red cell adherence(SPRCA) assay for the development of HLA- and/orplatelet-specific antibodies. Platelet crossmatching isperformed when positive HLA/platelet antibody screensare found or when the patient has an unsatisfactorycorrected count increment (CCI)1 following twoconsecutive platelet transfusions. During this routineevaluation, anomalous positive reactions wereencountered in the HLA/platelet antibody screen ofsome patients. These tests were positive only for a shortperiod of time, beginning after r-GCSF therapy wasinitiated and continuing until approximately 7 daysafter r-GCSF therapy was discontinued. As thesereactions disappeared from the SPRCA assay followingchloroquine diphosphate treatment,2 the presenceof platelet-specific antibodies was ruled out.Concomitantly, an increase in ABO antibody titers wasnoted in group O patients, unexpectedly poorresponses were seen to some platelet transfusions,and there was an apparent requirement for moreplatelet transfusions. Following these observations,3 aprospective study was initiated.

64 I M M U N O H E M A T O L O G Y , V O L U M E 1 7, N U M B E R 3, 2 0 0 1

Materials and Methods

Patient selection and data collectionA prospective study involving 92 consecutive

patients undergoing ABMT/PBPC from 7/3/91 to7/13/94 was designed. All patients received r-GCSF(Amgen, Thousand Oaks, CA; 16 µg/kg/day) from 2 to 3days before marrow reinfusion until their absoluteneutrophil count rose above 5000/µL or until they hadreceived r-GCSF for 21 days. Sixteen patients wereexcluded due to the presence of factors that mightinfluence their response to platelet transfusions, such asthe presence of drug-associated platelet antibodies (2),the receipt of intravenous gammaglobulin (5),multispecific HLA antibodies for which crossmatch-compatible platelets were not available for transfusion(2), platelet-specific antibodies requiring transfusion ofplatelets collected from family members (4), renaldialysis (1), or transfusion of plasma-depleted plateletsdue to repeated severe allergic transfusion reactions (2).Data collected included timing of r-GCSF administrationfor PBPC collections and pregnancy history. A platelettransfusion history was also obtained, which includedthe following information: number and dates oftransfusions, the donor ABO group, donor HLA type(including Bw4 and Bw6 type), and the resultant CCI. Inaddition, aliquots of all available patient sera werefrozen for future testing. The day on which the patientwas reinfused with previously harvested bone marrowand/or PBPCs was designated as Day 0.

The presence of anomalous reactivity (that is,reactivity not attributable to reactivity with one or moreHLA/platelet antigens) in HLA/platelet antibody screenswas used to assign patients to either the study (n = 25)or control (n = 48) group. All patients evaluated had atleast two PBPC collections prior to transplant usingstandard ablative regimens of radiation and/orchemotherapy. r-GCSF therapy was initiated 4 days priorto the first PBPC collection and continued until the daybefore the last PBPC collection. Generally, 1 to 3 weekselapsed between the last PBPC collection andtransplantation. The negative patients included 7 maleand 41 female patients with the following diagnoses:breast cancer (n = 36), Hodgkin’s disease (n = 7), non-Hodgkin’s lymphoma (n = 4), and lung cancer (n = 1).These control patients received a total of 308 platelettransfusions, with a CCI being calculated for 246 (80%)of the transfusions. The positive patients included 3male and 22 female patients with the following

diagnoses: breast cancer (n = 18), non-Hodgkin’slymphoma (n = 2), osteosarcoma (n = 1), ovariancarcinoma (n = 1), Hodgkin’s disease (n = 1),rhabdosarcoma (n = 1), and germ cell cancer (n = 1).The positive patients received a total of 341 platelettransfusions, with a CCI being calculated for 269(79%)of the transfusions.

HemotherapyAll patients received one daily prophylactic platelet

transfusion using single donor platelet units when theirplatelet counts were less than 20,000/µL. Plateletcounts were kept above 50,000/µL in the event ofhemorrhage. A 10-minute posttransfusion platelet countwas obtained whenever possible, and the CCI wascalculated. All platelet and red cell units released fortransfusion were gamma irradiated with 25Gy (2500rad) using a 137Cs Irradiator (CIS-US, Bedford, MA) andleukocyte reduced by bedside adsorption filtration(PL100K and RC50K filters; Pall BiomedicalCorporation, Glen Cove, NY). As we had previouslydemonstrated that there was no effect of storage timeon platelet count increment in bone marrow transplantpatients,4 units were selected from inventory accordingto outdate, with the oldest units being transfused first.ABO compatibility was considered only when therecipient demonstrated in vitro or in vivo incom-patibility with out-of-group platelet units. (ABO incom-patibility was defined as the presence on platelets of Aor B substance lacking on the patient’s red cells; plasmaincompatibility was not considered.)

SPRCA HLA/ platelet antibody screensSPRCA HLA/platelet antibody screens were

performed at least weekly (Capture-P® Ready Screen,Immucor, Inc., Norcross, GA); manufacturer’s directionswere followed for all testing. In chloroquine diphos-phate treatment to remove HLA antigens from reagentplatelets (CDP; Gamma Biologicals, Houston TX), twodrops of CDP were added to the test strip, andmanufacturer’s directions were followed to completethe testing. The incubation time, after CDP addition, wasincreased to 1 hour to allow for sufficient time for theCDP to react with the reagent platelets. The CDPpretreatment was performed twice on each reagent teststrip to maximize the effect on the HLA antigens. If atransfusion in a control patient did not result in a CCI <7500/µL without demonstrable clinical cause, HLA andplatelet antibodies were assumed to be absent.

M. F. LEACH AND J. P. AUBUCHON


Anti-A and Anti-B titersAnti-A and anti-B titers were performed on all

available serum/plasma samples (stored frozen)collected from group O patients (n = 35).5 The resultswere reported as the reciprocal of the highest dilutionthat produced at least a 1+ reaction.

Platelets immobilized with various sugarsPlatelets in the Capture-P® Ready Screen kits are

immobilized by the manufacturer to a polystyrenemicrotiter plate in the presence of glucose. To evaluatethe effect of various sugars on the positive patients sera,plates on which platelets had been immobilized withvarious simple sugars were supplied by themanufacturer. These sugars included arabinose,fructose, galactose, glucose, lactose, maltose, mannose,melibiose, trebalose, and sucrose (Sigma ChemicalCompany, St. Louis, MO). Platelets from the same donorwere immobilized with different sugars so comparisonsof the results could be made. Antibody screening usingthe platelets that had been immobilized with thevarious sugars was performed as for routine SPRCA-HLAscreening.

Antibody neutralizationWhen positive HLA/platelet antibody screens were

identified in routine SPRCA techniques, the testing wasrepeated using serum or plasma treated with varioussubstances (Immucor, Inc., Norcross, GA), to determineif the observed reactivity could be neutralized.Solutions (0.1M) of glucose, sucrose, fucose, mannose,fructose, galactose, mannose, or galactosamine (allobtained from Sigma Chemical Company) wereprepared in deionized water. Undiluted r-GCSF (300 µg/mL) was also used as a neutralizing substance. Onehundred µL of each solution were added to 500 µL ofserum/plasma. The mixture was allowed to incubate atroom temperature (22 to 24oC) for 5 minutes. Thesetreated samples were used to repeat SPRCA HLA/platelet antibody screens; results were compared tothose obtained with untreated samples.

Ion exchange chromatography procedureIon exchange chromatography, using diethylamino-

ethyl Affi-Gel Blue agarose gels (Biorad Labs, Hercules,CA) was performed on 18 individual sera collected frompositive patients; samples from three control patientswere also included. The manufacturer’s instructionswere followed for use of small volumes of sera to allowseparation in a conical centrifuge tube rather than alarge column.

The effluent was collected in 1 to 5 mL fractions. Thefraction size was determined by the amount of bufferadded to the gel during the various steps in thechromatography procedure. The estimated total proteinconcentration of each fraction was calculated bymultiplying the optical density (obtained by spectro-photometric analysis) by the volume and the dilutionrequired to obtain an optical density reading that was <2. Following total protein determinations, the proteincomposition of each fraction was evaluated by SDS-PAGE gel electrophoresis following manu-facturer’sdirections (Biorad Labs). Fraction #1 was found tocontain the majority of IgG proteins (data not shown).This fraction had a total volume of 1 to 5 mL and wasexpected to contain the unbound (IgG) protein fromthe sample. All fractions were placed in a –70oC freezerfor use in future testing.

Additional testingAn enzyme-linked immunosorbent assay (ELISA)

HLA antibody screening kit (GTI, Brookfield, WI) wasused to perform HLA antibody screens on selectedfractions obtained from ion exchange chromatographyprocedures according to the manufacturer’s directions(Table 1).

Table 1. Results of ELISA-HLA antibody screening

Days from Fraction SPRCA screen withPatient* Transplant contains† Bw6+ Platelets

1 –51 IgG neg–51 Other neg–21 Other neg–15 Other neg–12 IgG neg–12 Other neg

–2 IgG neg–2 Other neg0 IgG neg0 Other neg

+8 IgG pos+8 Other neg+9 Other neg

+13 IgG pos+15 IgG pos+15 Other neg+16 Other neg

+850 Other neg2 +18 Other neg

+18 IgG pos+18 IgG(concentrated) pos

3 Pooled serum IgG negPooled serum Other neg

4 Pooled serum IgG negPooled serum Other neg

* = all samples except those from Patients 3 and 4 are from chromato-graphy fractions from positive patients. Patients 3 and 4 are fromnegative patients (serum pooled to aid in processing).

† = content of fractions determined by SDS-PAGE electrophoresis

Apparent anti-Bw6 in r-GCSF recipients



StatisticsThe results of the study were analyzed by logistic

regression to identify predictors for the presence of theantibody. Gender, blood type (A vs. B vs. O), diagnosis(breast cancer vs. other), HLA Bw type (Bw4 vs. Bw6 vs.other), radiation, days on r-GCSF, days on total parentalnutrition (TPN), and increase in anti-A and anti-B titer(group O patients only) were included in the model. Ananalysis of variance (ANOVA), using natural logarithmictransformation, was performed on the total number oftransfusions. Factors in this model included presence ofthe antibody, patients’ HLA Bw6 types, antibody by Bw6interaction (Does the patient’s Bw6 type influence thedevelopment of the antibody?), diagnosis, and bloodtype. A t-test, using log scale to meet normalityassumptions, was performed to compare average CCIper transfusion between the control and the positivepatients. Chi-square analysis was used to determine ifthe observed frequency of HLA Bw6-positive plateletsreacting with the sera under study were significantlydifferent than what would be expected based onchance. The observed frequency of patient ABO groupand Rh types was analyzed by Yates Chi-square analysisto determine if the incidence observed was significantlydifferent than what would be expected by chance.

Results

Anomalous results in SPRCA HLA/platelet antibodyscreening

All HLA/platelet antibody screens were evaluated forthe presence of unexplained (anomalous) reactivity.A total of 204 antibody screen procedures wereperformed on the 48 patients in the negative (control)group (M = 4.25); 130 procedures were completed onthe 25 patients in the positive group (M = 5.20).Specific HLA antibodies were identified in eightpatients in the control group and in six patients in thepositive group (p > .05). Eighty-five percent of all thescreens done on control patients were negativecompared with 58 percent in the positive patients (p <.001). Thirty-two percent (42/130) of the screenscompleted on patients in the positive group were foundto have anomalous reactivity. All samples in whichspecific HLA antibodies and/or unexplained reactivitywere identified were further tested to determine thecause of the anomalous reactivity. This reactivity wasfirst observed 5 to 16 days (M = 9.14) after receivingthe first r-GCSF after admission for transplant; this

reactivity was no longer detected 1 to 5 days (M = 2.17)after r-GCSF had been discontinued.

Relationship of anomalous reactivity to glucoseimmobilization of platelets

All samples displaying the presence of unexpectedreactivity, as well as randomly selected samples fromfive patients with a negative HLA antibody screen,were tested with platelets immobilized with arabinose,fructose, galactose, glucose, lactose, maltose, mannose,melibiose, trebalose, or sucrose. Samples known tocontain only HLA-specific antibodies were also tested.In all cases (i.e., samples from the “positive” studygroup) in which the original HLA/platelet antibodyscreen had exhibited the anomalous reactivity, negativeresults were also obtained when the sample was testedwith platelets immobilized with any sugar other thanglucose. Samples that had not exhibited this reactivitycontinued to show a negative antibody screen whenthe other sugars were utilized for plateletimmobilization. However, the samples from patientswhose samples had not exhibited the anomalousreactivity but had defined HLA specificities continuedto react in the presence of the other sugars as they hadwhen the reagent platelets had been immobilized withglucose. Similarly, samples from patients demonstratingthe anomalous reactivity as well as defined anti-HLAreactivity continued to show the specific anti-HLAreactivity when the reagent platelets were immobilizedwith other sugars. No reactivity was observed whensera from the positive study group were reacted withglucose in plates that contained no platelets.

The reagent platelets that showed reactivity onroutine antibody screening with positive patients’ serabut were negative following neutralization with glucosewere evaluated further to determine if there was anyspecific HLA antigen that appeared more frequentlythan would be expected. A total of 152 reagent plateletswere evaluated. Reactivity was seen with 140 (92%) ofHLA-Bw6+ reagent platelets but only 6 (4%) of Bw6–platelets (p < .001).6 When the patients in the studywere phenotyped for HLA-Bw6, 10/48 (21%) of thenegative patients were found to be Bw6– comparedwith 4/25 (16%) of patients in the positive group. Thiscompares with a population distribution of 10 percentHLA-Bw6 negativity (difference = p > .230).7

Changes in anti-A titer with r-GCSF administrationAll patients in the study were evaluated to determine

the number of ABO-incompatible platelet transfusions



received. Patients in the negative group received33 ABO-incompatible transfusions compared to 30received by patients in the positive group.

Anti-A and anti-B titers were performed on 182samples from patients in the negative group and on 167samples from patients in the positive group. Themedian number of titers performed and the median forthe lowest anti-A and anti-B titers were similar in bothgroups. However, the positive patients showed themedian highest anti-A titer of 512 compared to 64 forthe negative patients. The median for the highest anti-Btiter was slightly higher in the positive patients (Mdn =32) than in the negative patients (Mdn = 16). Themedian change in anti-A titer for the negative patientswas 44 (range 0 to 480) compared with 464 (range 14to 2016) for the positive patients. The median change inthe anti-B titer in the negative patients was 12 (range 0to 120) compared with 22 (range 2 to 480) in thepositive patients. The changes in titer were evaluated bythe z test for rejection of the null hypothesis of nodifference. The z value for changes in anti-A titer in thepositive group was > 7 standard deviations from themean for the negative patients, a significant differencebetween the two groups. When the changes in anti-Btiter were evaluated, the z value was found to be < 1standard deviation from the control mean; the change inanti-B titer between the two groups was notsignificantly different.

Antibodies in eluted fractionsAfter separation using ion exchange chromatography,

a total of 144 fractions (each fraction contained 1 to 5mL) collected from 18 samples from patients with theanomalous reactivity were evaluated. Each fraction wastested with glucose-immobilized donor platelets todetermine which fraction contained the anomalousreactivity. Twenty-one of these fractions, includingreactive and nonreactive fractions, underwentadditional testing; these fractions were the firstcollected and contained the IgG component ofimmunoglobulins as determined by SDS-PAGEelectropheresis. These fractions were selected forevaluation as the reactivity undergoing investigationappeared to be IgG (IgG-sensitive indicator red cells inSPRCA). In addition, six study serum samples, showingnegative results with glucose-immobilized platelets,were also included in the additional testing forcomparative purposes.

IgG-containing chromatography fractions fromsamples taken from positive patients about the time

when they exhibited the anomalous reactivity clinicallyshowed the same characteristic SPRCA anomalousreactivity. Two of these samples also contained HLA-antigen specific reactivity as previously documented;these specificities were confirmed in ELISA testing.Chromatography fractions containing IgG from negativepatient samples shown to contain HLA antibodiesdisplayed antigen-specific reactivity in both SPRCA andELISA test systems.

When all SPRCA testing was completed, all remainingpositive and negative fractions were pooled to allowsufficient volume for additional testing. First, thepresence of ABO antibodies was sought by standardtechniques. When testing fractions derived frompatients who had exhibited the anomalous reactivity,ABO antibodies were not present in the pooledfractions that exhibited the anomolous reactivity (thatis, the fractions presumed to contain IgG) but wereidentified in the pooled fractions that did not show thisreactivity (that is, fractions presumed to be of otherimmunoglobulin classes). Second, glucose neutralizationwas performed. The anomolous reactivity showed thesame neutralization characteristics in these pooledfractions as in the original samples.

Clinical effect of “anomalous” antibodyA logistic regression was fitted to identify predictors

of the presence of the antibody from among thefollowing factors: gender, blood type (A vs. B vs. O),diagnosis (breast cancer vs. other), HLA-Bw type (Bw4vs. Bw6 vs. other) of the patient, radiation (yes or no),days on r-GCSF, days on TPN, increase in anti-A or anti-Btiter (group O only).

An ANOVA was performed on the total number ofplatelet transfusions required by patients. Increasedlikelihood of the antibody’s presence was associatedonly with the number of platelet transfusions received(p = .0005). The geometric mean number oftransfusions (with 95% confidence interval) for controlpatients was 5.01 (4.08, 6.16), whereas that for studypatients was 10.49 (7.67, 14.36). No significantdifferences between the control and positive patientswere identified.

The significance of changes in anti-A titer betweenthe control and positive patients was evaluated byANOVA. Increases in titer in the positive patients wasshown to have borderline significance (p = 0.068), withan increase in titer of >100 being associated with anincrease in the probability of developing theunexpected reactivity. The type of ablative therapy used


had no association with the development of theanomolous reactivity.

A t test was performed to compare average CCI pertransfusion (on the log scale, to meet normalityassumptions) for the positive vs. the negative patients.The difference in CCI between the two groups wasstatistically significant (p = .0005). The geometric meanCCI (with 95% confidence) for the negative patientswas 10,822 (9,888;11,844) compared to 7,951(6,956;9,090) for the positive patients, a 73 percentdifference.

Chi-square analyses were used to determine if CCIsfollowing receipt of Bw–6 or Bw–4 positive plateletswere significantly different between groups. Nostatistically significant differences were found whenBw4+6–, ABO-compatible platelets were transfused.However, when Bw4+6+ or Bw4–6+ ABO-compatibleunits were transfused (168/216 [77%] transfusions inthe negative group vs. 205/244 [84%] transfusions inthe study group), the CCIs in the study group weresignificantly lower than those in the control group (p =<.001). No significant differences were seen when ABO-incompatible platelet units (n = 33 [negative group]compared to n = 30 [positive group]) were received.

Yates corrected Chi-square analyses were used todetermine if the incidence of ABO groups in both studyand negative patients was significantly differentbetween groups and from the expected incidence inthe general population. Half (24/48) of the patients inthe control group were group O, whereas 64 percent(16/25) of patients with the anomolous reactivity weregroup O (p > .1). The negative patients included 17group A patients compared to 8 in the positive group(p > .1). There were too few patients of other groups toevaluate.

DiscussionThis study documented the presence of an IgG

antibody giving anomalous reactions in SPRCA. Thiswas seen in all blood groups, was temporarilyassociated with r-GCSF administration, could be blockedin vitro with addition of glucose to the sera, andreduced platelet transfusion response. This reactivitywas also associated with the presence of the HLA Bw6antigen on the reagent platelets and an increase in ABOtiter (in group O patients only).

The unusual serologic findings in the preliminaryobservations3 prompted a closer, prospectiveinvestigation of the apparent phenomenon reportedhere. The source of the unexpected reactivity was

ultimately traced to an IgG antibody (reactivityobserved with IgG-coated indicator cells) that reactedto platelets bound to a polystyrene solid phase in thepresence of glucose. The strong association of thereactivity with the presence of the Bw6 antigen on thereagent platelets clearly suggests a role for this epitopein the antibody’s specificity. The very specific effect ofthe use of just one sugar might suggest the former asthe more plausible mechanism, but ascribingimmunogenicity to such a small and ubiquitousmolecule, even through haptenic combination, seemsunlikely.

The identification of this reactivity in the SPRCAsystem but its failure to react in ELISA andlymphocytotoxicity testing systems might initiallysuggest that this is solely a phenomenon restricted toone particular test system. In itself, such an occurrencewould still be important to note because anomalousresults in SPRCA antibody screening can be confusingand could delay selection of platelet units fortransfusion. However, our finding that the presence ofthis reactivity was associated with 73 percent lowerCCIs demonstrates that the in vitro reactivity may havean in vivo counterpart that is clinically significant. Theeffect may be associated with the need to administertwice as many platelet transfusions during ABMT/PBPC,thus increasing recipient risk and the procedure’s cost.This effect appeared independent of ABO compatibility;however, HLA Bw6+ platelet units were more likely toyield a poor transfusion response in patients with theunexpected reactivity, suggesting a clinical correlationof the increased in vitro reactivity of the positive serawith the presence of HLA Bw6 on platelets.

Red cell antibodies that react only in the presence ofcertain sugars have been reported.8 The presence ofthis unexpected reactivity with red cells suspended ina variety of sugars, including glucose, galactose,mannose, fructose, lactose, or dextran has been shownto occur in 2 to 22 percent of sera tested.9 However, areview of the literature failed to identify prior reports ofsimilar reactivity when testing reactivity with platelets.

Not all transplant centers utilize the SPRCAtechnique and thus may not have serologic findings inpatients such as these that correlate with poor platelettransfusion responses in affected patients. If a patient iscurrently receiving TPN and r-GCSF and has CCIs< 7500, the patient history should be carefully reviewedand evaluated. As group O patients were 11.4 timesmore likely to develop the antibody if there was agreater than 100-fold increase in their anti-A titer, this



simple titration test may be important in inferring thepossible presence of the antibody. As pretransplanttiters are required for this evaluation, it is recommendedthat a pretransplant sample be frozen and used forisoagglutinin titration when the presence of theanomalous antibody is suspected. If responses toplatelet transfusions and patient history indicate thatthis glucose-dependent reactivity might be present, HLAantibody screen results and determination of the HLA-Bw6 type of transfused platelet units and their CCIsmay provide useful information. If the CCIs for HLA-Bw6 negative units are significantly higher than thosefor HLA-Bw6 positive units without serologic evidenceof alloimmunization, the presence of this anomalousantibody should be considered.

It is important to note that all patients included inthis study received r-GCSF at a dose of 16µg/kg/day.Shortly after the completion of this study, the r-GCSFdose for ABMT/PBPC patients at our hospital wasdecreased to 5µg/kg/day; only two additional examplesof this unexpected reactivity have been identified sincethe r-GCSF dose was decreased. This suggests that thedose of r-GCSF received may contribute to thedevelopment of this unexpected reactivity.

Similar anomalous results in SPRCA testing onpatients who received r-GCSF have recently beenreported.10 These researchers identified the anomalousreactivity described in this report; however, as thereactivity was observed only in SPRCA, they reported itspresence as “false-positivity” and attributed the poorposttransfusion platelet response to the presence ofinfections and other clinical factors. Our researchindicates that r-GCSF recipients should be monitoredfor the development of a glucose-dependent HLA-Bw6antibody that may have clinical implications. SPRCAtechniques are easy to perform and currently are theonly techniques that will detect these significantantibodies. For patients displaying this unexpectedreactivity, all platelet units should either be ABOidentical or group O, and should be negative for HLA-Bw6 as well, of course, as any HLA antigens to whichthe patient has been sensitized. This protocol shouldbe followed until at least 7 days after r-GCSFdiscontinuation.

References1. Daly PA, Schiffer CA, Aisner J, Wiernik PH. Platelet

transfusion therapy. One hour posttransfusionincrements are valuable in predicting the need forHLA-matched preparations. JAMA 1980; 243:435–8.

2. Sinor LT. Preparation of chloroquine treatedplatelets to differentiate HLA and non-HLA plateletantibodies on a solid phase red cell adherence assay.Immucorrespondence, 1988.

3. Leach MF, AuBuchon JP, Sinor LT. Serologic effects ofG-CSF administration during autologous marrowtransplant (abstract). Transfusion 1993; 33:78S.

4. Leach MF, AuBuchon JP. Effect of storage time onclinical efficacy of single-donor platelet units.Transfusion 1993; 33:661–4.

5. Technical manual. Virginia Vengelen-Tyler, ed. 12thed. Bethesda, MD: American Association of BloodBanks, 1996; 376–7, 646–7.

6. Rodney G. Class I antigens: HLA-A, -B, -C andcrossreactive groups. In: Moulds JM, Fawcett KJ,Garner RJ, eds. Scientific and technical aspects ofthe major histocompatibility complex. Arlington, VA:American Association of Blood Banks, 1989:23–46.

7. Rodney G. Prevention of alloimmunization inthrombocytopenia patients. In: Smith DM Jr,Summers SH, eds. Platelets. Arlington, VA: AmericanAssoication of Blood Banks, 1988:93–113.

8. Garratty, G. In vitro reactions with red blood cellsthat are not due to blood group antibodies: a review.Immunohematology 1998; 14:49–60.

9. Lalezari P, Jiang AF, Kumar M, Lalezari I. Carbo-hydrate-specific antibodies in normal human sera. I.ß-D-glucose. Vox Sang 1984;47:133–45.

10. Martinez C, Muniz-Diaz E, Sola C, et al. Incidence ofimmune platelet refractoriness in breast cancerpatients treated with hematopoietic stem cellstransplantation (HSCT) (abstract). Vox Sang 1988;74:51a.

Miriam Fogg Leach, MS,MT(ASCP)SBB, (corres-ponding author), Special Projects Coordinator andInformation System Specialist, Transfusion ServiceLaboratory, Department of Pathology, Dartmouth-Hitchcock Medical Center, One Medical Center Drive,Lebanon, NH 03756; and James P. AuBuchon, MD,Department of Pathology, Dartmouth-HitchcockMedical Center, Lebanon, NH.


Notice to Readers: Immunohematology,Journal of Blood Group Serology and Education,is printed on acid-free paper.


Detection of granulocyte antibodiesby flow cytometry without the useof pure granulocyte isolatesK. M. KIEKHAEFER, K. M. CIPOLONE, J. L. PROCTER, K. MATSUO, AND D. F. STRONCEK

Established methods used to detect serum antibodies togranulocytes require the isolation of granulocytes. Flow cytometricanalysis of granulocytes with monoclonal antibodies eliminates theneed for granulocyte isolation. The purpose of this study wasto develop a method to evaluate reactions of antibodies togranulocytes without separating granulocytes from otherleukocytes. Three screening cell samples for granulocyte antibodydetection were prepared from whole-blood samples in which thered blood cells (RBCs) were lysed and remaining leukocytes testedagainst sera at 4oC. Binding of human alloantibodies to thescreening cells was determined by flow cytometric analysis usingphycoerythrin-conjugated antibody to human immunoglobulin.Forward and side scatter were used to analyze granulocytesseparately from other leukocytes. The assay was validated by testinggranulocytes with reference alloantibodies directed to NA1, NA2,5b, and Mart antigens. Samples from 32 patients were tested, andthe results of the assays were compared with the results of testingthe samples in a granulocyte immunofluorescence (GIF) assayperformed by a reference laboratory. In the whole-blood flowcytometric (WBFC) assay the mean fluorescence intensities ofreference antisera with antigen-positive cells, expressed in arbitraryunits, were anti-NA1 = 48 to 221, anti-NA2 = 24 to 69, anti-5b = 13to 57, and anti-Mart = 42 to 72. In contrast, the mean fluorescenceintensity of type AB-negative control sera ranged from 3 to 11. Ofthe 32 patient sera tested, 23 were positive (range = 12 to 56) and9 were negative (range = 3 to 10). When compared with the resultsobtained by the reference laboratory, 27 sera were concordantbetween the WBFC and the GIF assays. Four of the samples werepositive in WBFC (range = 11 to 31) and negative in GIF and onesample was negative in WBFC (range = 5 to 6) and positive in GIF.Leukocytes prepared from whole blood after lysis of RBCs can beused in flow cytometric analysis to detect granulocyte allo-antibodies. The results of testing for granulocyte antibodies withthis assay were similar to results of testing sera in GIF. Furthercomparative studies are indicated to confirm findings and explainthe discordant results. Immunohematology 2001;17:70–75.

Key Words: granulocytes, flow cytometry, granulocyteagglutination assay, granulocyte immunofluorescenceassay, monoclonal antibody immobilization of granu-locyte antigens

Granulocyte antibodies are important in thediagnosis of autoimmune neutropenia, neonatalalloimmune neutropenia, and transfusion reactions.1

Transfusion recipients who have antibodies to donor

leukocytes may experience febrile nonhemolyticreactions if they are transfused with blood componentscontaining leukocytes. Transfusion of blood compo-nents containing granulocyte antibodies can causetransfusion-related acute lung injury (TRALI).2 Inaddition, patients who are transfused with granulocyteconcentrates often develop antibodies to granulocyteantigens that render future transfusions ineffective.3

To diagnose and treat these conditions, it is imperativethat an easy-to-perform, rapid, accurate, and sensitivemethod is available to detect and identify granulocyteantibodies.

Methods for the detection of granulocyte antibodiesinclude the granulocyte agglutination (GA) assay,granulocyte immunofluorescence (GIF) assay,1 and themonoclonal antibody immobilization of granulocyteantigens (MAIGA) assay.4 These methods use granu-locytes that have been isolated from whole blood byremoving red blood cells (RBCs) by sedimentation andlysis and then separating granulocytes from otherleukocytes by density gradient solutions. The isolationprocedure is time-consuming and technique-dependent.

The purpose of this study was to develop andvalidate a method to detect granulocyte antibodiesusing whole blood without removing RBCs by sedimen-tation and without separating white blood cells (WBCs)using density gradient solutions.

Materials and Methods

Study designTo avoid isolating granulocytes from the samples to

be used as screening cells, whole blood was collectedand the RBCs were lysed. The granulocytes were testedand analyzed for fluorescence using flow cytometricmethods. This procedure was initially tested in a study


that optimized the variables of temperature ofincubation, volume of reactants, and the use offormaldehyde as a fixative. The initial study alsoassessed the efficacy of the method using high-titerantibodies specific to well-characterized antigens. Next,to validate the whole-blood flow cytometry (WBFC)antibody detection method, granulocytes of knownphenotype were tested with reference antisera. Then ablinded study compared the test results of samples thathad been previously tested for granulocyte antibodiesby traditional methods with test results on the samesamples using the WBFC assay.

Antisera and patient samplesThe reference antisera for neutrophil antigens NA1,

NA2, Mart, and 5b were provided by the American RedCross (North Central Blood Services, St Paul, MN).The group AB sera were obtained from healthy,nontransfused, nonpregnant donors. The patientsamples were from individuals who were suspected ofhaving autoimmune neutropenia or who weretransfused with granulocyte concentrates and seen atthe Warren G. Magnuson Clinical Center (NationalInstitutes of Health, Bethesda, MD).

Testing serum for granulocyte antibodiesInitially, three variables were optimized: the temper-

ature of incubation of antibodies with granulocytes, thevolumes of cell suspensions and sera, and the step tobest fix granulocytes with formaldehyde. For theseinitial studies, sera containing well-characterized anti-NA1 and anti-NA2 were used as positive controls and ABsera were used as a negative control. The sera weretested against granulocytes from donors whose NA1and NA2 phenotypes were previously determined usingmonoclonal antibodies and whose genotypes wereconfirmed by polymerase chain reaction and restrictionenzymes.5,6 Varying quantities of granulocytes and serawere tested at either 24oC or 4oC. Incubation at 4oC wasbetter than incubation at 24oC in that nonspecificfluorescence and decreased viability of granulocytesoccurred at 24oC (data not shown). In addition, it wasfound that 10 µL of sera should be incubated with 200µL of cells to decrease the time needed for the flowcytometer to acquire 10,000 events.

The treatment of granulocytes with fixativebroadened the window of time in which testing couldbe completed. In some methods, granulocytes werefixed with formaldehyde before incubating withantibodies,1,5 whereas in other methods, the granulo-

cytes were fixed after they had been incubated withantibodies.5,6 We found that formaldehyde fixation didnot affect the test result when used at the end of theprocedure, which allowed data acquisition by flowcytometry up to 24 hours after the testing procedurewas completed.

On the basis of these results, the following procedurewas used to test granulocyte antibodies. After obtaininginformed consent, whole blood from each granulocytedonor was collected into two 7 mL vials containingEDTA and stored at room temperature for no more than24 hours before testing. Approximately 10mL of wholeblood was combined with 40 mL of ACK Lysing Buffer(Quality Biologic, Inc., Gaithersburg, MD) in a 50 mLpoly-propylene conical tube (Becton Dickinson,Mountain View, CA) and placed on a mechanical rockerfor 6 minutes or until lysis of RBCs was complete. Thetube was then centrifuged at 1900 x g for 5 minutes.The hemolysate was decanted and the remaining WBCsediment was resuspended and washed x 2 in cold(4oC) Hanks buffered saline solution (HBSS) withoutcalcium (BioWhittaker, Walkersville, MD). After the lastwash, each WBC pellet was resuspended in 4 mL ofHBSS and 200 µL aliquots of this cell suspension wereplaced in 5 mL polypropylene test tubes (BectonDickinson). Ten microliters of test or control sera wereadded to the 200 µL aliquot of the donor cell sus-pension and incubated for 30 minutes at 4oC. Followingincubation, the WBCs were washed once with coldHBSS. Five microliters of secondary antibody, aphycoerythrin (PE)-conjugated goat anti-human IgG[F(ab' )

2-PE] (Jackson ImmunoResearch Laboratories,

Inc., West Grove, PA), were added to each tube andincubated at 4oC for 30 minutes, protected from light.After 30 minutes, the suspension was washed once withcold HBSS and resuspended to a final volume ofbetween 200 and 400 µL with HBSS or an HBSS 1%formaldehyde solution (Polysciences, Warrington, PA).The samples were then analyzed by flow cytometricmethods.

Flow cytometry analysisWhen WBCs were analyzed by flow cytometry

(Facsort, Becton Dickinson), forward and side scatterwere used to differentiate the granulocyte populationfrom the lymphocyte and monocyte populations.Software (Becton Dickinson CellQuest) was used tofurther analyze the fluorescence of the granulocytepopulation. The results of the analysis of granulocytesfrom each donor were expressed as the peak cell

Flow cytometry for granulocyte antibodies


fluorescence intensity (the fluorescence intensity thatcorresponded to the greatest number of granulocytesfrom each donor).

Establishing cutoffs for positive resultsTo differentiate between positive and negative test

results, a cutoff value was established by calculating themean plus 2 standard deviations using the values of themean peak channel fluorescence of known negativesamples. Both group AB sera and reference sera thatwere tested against cells that did not possess thecorresponding antigen were used to determine thecutoff values. This was done for each of the three WBCcell populations tested. When patient samples weretested, any value greater than the cutoff was consideredpositive.

Testing for granulocyte antibodies by a referencelaboratory

All patient samples were tested by GA and GIF by aneutrophil serology reference laboratory (AmericanRed Cross, North Central Region Blood Services, St. Paul,MN).1,7 The interpretations of the GA and GIF assayswere based on phase contrast microscopy or fluor-escent microscopy readings of reactions of the serawith granulocytes from five different donors. Sampleswith discordant results in the WBFC and GIF assayswere tested by the reference laboratory in the MAIGAassay against CD16 (Fc-gamma receptor IIIb), CD18 (ß

2

integrin), and CD177 (NB1) monoclonal antibodies.

Results

Testing reference antisera and sera withoutantibodies

Sera known to contain well-characterized allo-antibodies to granulocyte antigens were used to demon-strate that antibodies to well-described alloantigenscould be detected by the assay. The sera were testedagainst granulocytes from three different donors (Table1). Results of testing the control group AB sera againstgranulocytes from each of the three donors as well asanti-NA1 against antigen-negative granulocytes (DonorC) and anti-NA2 against antigen-negative cells (Donor A)are shown in Table 2. There was no difference in thereactions of group AB sera and anti-NA1 or anti-NA2with antigen-negative cells. On the basis of thesereactions, the cutoff for reactions of patient sera withgranulocytes from Donor A was determined to be 10.53arbitrary units. For granulocytes from Donor B, the

cutoff was 9.12. For granulocytes from Donor C, thecutoff was 10.32. The results of testing referencealloantibodies against granulocytes from the threedonors are shown in Table 3. Reactions of anti-NA1 andanti-NA2 with antigen-positive cells were much greaterthan with antigen-negative cells. Reactions of anti-Martand anti-5b with granulocytes from all three donorswere significantly greater than reactions of anti-NA2 andanti-NA1 with antigen-negative granulocytes.

K. M. KIEKHAEFER ET AL.

Table 1. Genotype and phenotype results of three granulocyte donors

Genotyping* Monoclonal Humanantisera† antisera‡

NA1 NA2 NA1 NA2 5b Mart

Donor A + – + – + +Donor B + + + + + +Donor C – + – + + +

*Determined by polymerase chain reaction and restriction enzymes8

†Determined by flow cytometry using NA1- and NA2-specific monoclonal antibodies7

‡Determined in-house using isolated granulocytes, alloantisera, and flow cytometry5

Table 2. Determination of negative cutoff for three granulocyte donorcells; analysis of reactions of nonreactive sera againstgranulocytes in the whole-blood flow cytometry assay

Donor A Donor B Donor C

Peak Peak Peakfluorescence fluorescence fluorescence

Sera intensity* Sera intensity Sera intensity

AB 4.22 AB 3.73 AB 4.62AB 3.67 AB 3.80 AB 4.25

AB 4.40 AB 4.99 AB 4.60

AB 7.43 AB 5.29 AB 4.68

AB 5.92 AB 3.69 AB 4.66

AB 4.31 AB 3.44 AB 4.34

AB 5.03 AB 3.95 AB 6.39

AB 5.06 AB 5.11 AB 4.64

AB 5.08 AB 4.26 AB 3.88

AB 5.90 AB 5.80 AB 9.56

AB 7.20 AB 11.32 AB 7.56

AB 7.10 AB 5.33 AB 7.83

Anti-NA2 11.39 Anti-NA1 6.51








Mean 5.99 5.06 6.36

StandardDeviation 2.27 2.03 1.98

NegativeCutoff 10.53 9.12 10.32

* Peak fluorescence intensity is expressed in arbitrary units


Testing patient samplesSera from 32 patients were tested and 23 were

reactive with granulocytes. However, the peak cellfluorescence intensities of the reactions of the 23 posi-tive patient sera were less than the peak fluorescenceintensities of reference antibodies.

The results of testing the 32 patient samples in theWBFC assay were compared with those obtained usingthe GIF and GA assays. Twenty of the 32 patientsamples were positive in the GIF assay and five werepositive in the GA assay. The results of the WBFC assaydiffered from the GIF assay for five of the 32 assays.Four of the five samples with discordant results werepositive in the WBCF assay and negative in the GIF and

GA assays (Table 4) and one was negative in the WBFCassay, but positive in the GIF assay (Table 5). Of the foursamples that were reactive in the WBFC assay but nega-tive in the GIF and GA assays, three samples werereactive in the WBFC assay with granulocytes from allthree donors tested and one was reactive with granu-locytes from one donor. All four of these samples weretested in another assay, MAIGA, and were negative. Thesample that was negative in the WBFC assay andpositive in the GIF assay was negative in the GA andMAIGA assays. All five samples that were positive in theGA assay were also positive in both the WBFC and GIFassays.

DiscussionA new assay to test for granulocyte antibodies was

developed and tested. This method has the advantageover other assays, including GIF, in that it avoids thelaborious density gradient separation step. Another ad-vantage is that flow cytometry rather than fluorescencemicroscopy is used to detect the binding of antibodiesto granulocytes. Flow cytometry permits more precisescoring of reactions of antibodies to granulocytes andproduces a permanent record of test results.


Table 3. Representative positive results of testing reference antiseraagainst granulocytes from three donors in the whole-bloodflow cytometry assay

Peak fluorescence intensity

Sera Donor A Donor B Donor C

Anti-NA1 112.15* 133.26 6.51

Anti-NA2 7.10 57.93 39.97

Anti-Mart 57.74 51.08 44.73

Anti-5b 31.96 18.59 19.34

*Bold = Peak fluorescence intensity greater than the negative cutoff

Table 4. Comparison of test results of samples positive in the whole-blood flow cytometry assay (WBFC) with the granulocyte agglutination (GA) andgranulocyte immunofluorescence (GIF) test results

Whole-blood flow cytometry results Reference lab(peak fluorescence intensity) test results

Patient Donor A Donor B Donor CPatient age (yrs) cutoff:10.53 cutoff:9.12 cutoff:10.32 Interpretation GA GIF

3 27 56.51 17.51 18.56 Pos Neg Pos

11 6 29.08 11.84 10.36 Pos Neg Pos

31 22 27.51 30.74 23.39 Pos Neg Neg

7 9 25.40 18.36 16.55 Pos Pos Pos

12 7 22.79 15.91 14.49 Pos Neg Pos

26 43 19.92 16.42 10.03 Pos Neg Pos

32 NA* 14.86 11.26 10.65 Pos Pos Pos

17 19 14.56 15.00 11.98 Pos Pos Pos

8 13 14.15 11.86 9.63 Pos Neg Pos

21 13 13.89 19.82 15.84 Pos Neg Neg

16 9 13.63 12.29 15.58 Pos Neg Pos

36 17 12.26 13.31 15.88 Pos Neg Pos

23 73 10.87 15.58 16.93 Pos Neg Neg

33 12 10.83 8.15 6.71 Pos Neg Pos

4 NA* 5.54 21.12 22.92 Pos Neg Pos

35 8 9.53 15.21 8.73 Pos Neg Pos

14 22 7.16 12.66 9.89 Pos Neg Neg

29 50 8.62 12.60 9.81 Pos Pos Pos

18 15 9.45 12.29 10.48 Pos Neg Pos

2 5 8.72 11.42 5.45 Pos Neg Pos

34 17 9.94 9.20 9.10 Pos Neg Pos

20 40 9.27 9.27 9.16 Pos Neg Pos

30 34 8.23 9.37 10.83 Pos Pos Pos

*Not available


Table 5. Comparison of test results of samples negative in the whole-blood flow cytometry assay (WBFC) with the granulocyte agglutination (GA ) andgranulocyte immunofluorescence (GIF) test results

Whole-blood flow cytometry Reference Lab(peak fluorescence intensity) test results

Patient Donor A Donor B Donor CPatient Age (yrs) cutoff:10.53 cutoff:9.12 cutoff:10.32 Interpretation GA GIF

10 6 4.67 3.26 4.00 Neg Neg Neg

19 9 3.25 4.23 4.25 Neg Neg Neg

5 12 8.46 6.23 6.99 Neg Neg Neg

6 33 7.54 7.96 6.72 Neg Neg Neg

9 51 4.05 3.61 4.79 Neg Neg Neg

13 26 3.96 4.20 5.17 Neg Neg Neg

1 44 10.07 6.72 7.75 Neg Neg Neg

24 NA* 5.91 6.12 6.76 Neg Neg Pos

37 NA* 5.05 4.14 7.25 Neg Neg Neg

*Not available

The results of testing patients’ samples in the WBFCassay and in the GIF assay were similar, but somediscordant results occurred. The reason for thesediscordant results is not certain, nor is it clear whichassay was correct. The differences could be due tovariable test results of weak antibodies. Alternatively,slight differences in the test methods could account forthe discordant results. In the WBFC assay, granulocytesand serum were incubated at 4oC and granulocyteswere treated with formaldehyde after incubation withthe serum and secondary antibody. In the GIF assay,granulocytes were fixed with formaldehyde prior to37oC incubation and subsequent incubation withsecondary antibody.

To detect all granulocyte antibodies, a combination ofassays must be used. However, the results of this studyshowed that the WBFC method is most similar to theGIF in its ability to detect granulocyte antibodies and ifa laboratory can use only a single assay, the WBFC orGIF assay is preferable to the GA assay to screen forgranulocyte antibodies. The GA assay was the mostspecific assay, but it lacked sensitivity.

The WBFC assay is a valid method to screen patientssuspected of having granulocyte antibodies. If thenumber and phenotype of granulocyte donors permit,this method could also be used to determine specificityof granulocyte antibodies. With the use of well-characterized granulocyte antibodies, this method canalso be used to phenotype granulocyte antigens.

The WBFC assay described is a relatively quick andsensitive method that could aid in the diagnosis andtreatment of clinical conditions in which granulocyte

antibodies are implicated. It may be feasible to use thisassay to screen whole blood donors for granulocyteantibodies and discard antibody-containing componentsin order to prevent the transfusion of granulocyteantibodies that could result in TRALI.

References1. McCullough J, Clay ME, Press C, Kline W.

Granulocyte serology: a clinical and laboratoryguide. Chicago: ASCP Press, 1988.

2. Popovsky MA, Abel MD, Moore SB. Transfusion-related acute lung injury associated with passivetransfer of antileukocyte antibodies. Am Rev RespirDis 1983;128:185–9.

3. Stroncek DF, Leonard K, Eiber G, Malech HL, GallinJI, Leitman SF. Alloimmunization following granu-locyte transfusions. Transfusion 1996;36:1009–15.

4. Metcalfe P, Waters AH. Location of the granulocyte-specific antigen LAN on the Fcγ-receptor III.Transfus Med 1992;2:283–7.

5. Chun H, Cipolone K, Procter J, Stroncek DF.Granulocyte storage and antigen stability.Transfusion 1999;39:983–90.

6. Matsuo K, Lin A, Procter JL, Clement L, Stroncek DF.Variations in the expression of granulocyte antigenNB1. Transfusion 2000;40:654–62.

7. Stroncek DF, Ramsey G, Herr GP, Eiber G, Clay ME,Ketyer EC. Identification of a new white cellantigen. Transfusion 1994;34:706–11.

8. Matsuo K, Procter J, Stroncek DF. Variations in genesencoding human neutrophil antigens NA1 andNA2.Transfusion 2000;40:645–53.

K. M. KIEKHAEFER ET AL.


Karen M. Kiekhaefer, MT(ASCP)SBB, TechnicalSpecialist/Customer Technical Services — BloodBank, Ortho-Clinical Diagnostics, Raritan, NJ;Karen M. Cipolone, MT(ASCP)SBB, Department ofTransfusion Medicine, Warren G. Magnuson ClinicalCenter, NIH, Bethesda, MD; Jo L. Procter, MEd,MT(ASCP)SBB, Department of Transfusion Medicine,


Warren G. Magnuson Clinical Center, NIH, Bethesda,MD; Kazuhiko Matsuo, MD, PhD, First InternalMedicine Kurume University School of Medicine,Fukuoka, Japan; and David F. Stroncek, MD(corresponding author), Department of TransfusionMedicine, Warren G. Magnuson Clinical Center, NIH/CC/DTM, 10/1C711, Bethesda, MD 20892–1184.

REMEMBER: THE PASSWORD IS “2000”for redcross.org/immunohematology

Now, as a subscriber, you can enter the password, 2000, to access complete journal contents forthe past 5 years. That means cover to cover! You will receive every article in total, every letter to the editor,every review, every ad, every notice, and every literature review! All of the other services will continue to beavailable on the Web page, including sending a letter to the editor, subscribing with a credit card on the secureorder site, performing a literature search, sending a message to the editor-in-chief, and linking to otherimportant sites. Log on now to see this new service!


Low-incidence MNS antigensassociated with single amino acidchanges and their susceptibility toenzyme treatmentM. E. REID AND J. R. STORRY

MNS antigens are carried on glycophorin A (GPA), glycophorin B(GPB), or their variants. Antigens at the N-terminus of GPA aresensitive to cleavage by ficin, papain, and trypsin but are resistant toα-chymotrypsin. Antigens at the N-terminus of GPB are sensitive tocleavage by ficin, papain, and α-chymotrypsin but are resistant totrypsin treatment. These characteristics have been used to aid in theidentification of blood group alloantibodies. Recent molecularanalyses have identified changes in amino acids that are associatedwith several low-incidence antigens in the MNS blood groupsystem. This review relates the molecular studies with thesusceptibility or resistance of these antigens to treatment of intactred blood cells by proteolytic enzymes. Immunohematology2001;17:76–81.

Key Words: MNS blood group system, low-incidenceantigens, molecular basis, enzymes, papain, trypsin,α-chymotrypsin, ficin

The many low-incidence antigens of the MNS bloodgroup system are carried on variant forms of glyco-phorin A (GPA) and glycophorin B (GPB). GPA and GPBare single pass, type I integral membrane glyco-proteinsof the red blood cell (RBC) membrane. The proteins areencoded by adjacent genes (GYPA and GYPB,respectively) on chromosome 4. Because these genesshare 95 percent identity, hybrid genes caused bysimple crossing over and gene rearrangements lead tonumerous hybrid proteins.1,2

M, N, S, and s are the polymorphic antigens afterwhich the MNS blood group system was named. The Mor N antigen is located at the N-terminus of GPA. GPBcarries the ‘N’ antigen at the N-terminus and the S or santigen at residue 29, where methionine is specific forthe S antigen and threonine is specific for the santigen.3,4 Other antigens in the MNS blood groupsystem arise from single amino acid substitutions, novelamino acid sequences at the junction of the hybridproteins, or within encoded regions of the GPBpseudogene.2

The purpose of this review is to summarize themolecular basis of low-incidence MNS antigens that arecarried on GPA or GPB and are associated with a singlepoint mutation. The molecular structure is consistentwith laboratory evidence of known resistance orsensitivity of these antigens to enzyme treatment ofintact antigen-positive RBCs. Low-incidence antigens inthe MNS system are well-expressed on cord RBCs5 andmany of the corresponding antibodies have beenimplicated in hemolytic disease of the newborn. Thesignificance of these antibodies in blood transfusion islargely unknown because crossmatch procedures thatinclude an antihuman globulin test will detectincompatibility in the absence of a positive antibodyscreen and finding compatible RBCs is not difficult.

Low-incidence antigens on GPATable 1 lists the low-incidence antigens that have

been shown to be associated with a single amino acidsubstitution in GPA.5 Table 1 also includes the effect ofpapain, ficin, trypsin, or α-chymotrypsin treatment ofantigen-positive RBCs. GPA on intact RBCs does nothave an α-chymotrypsin cleavage site. M and N antigensat the N-terminus of GPA are sensitive to treatment ofantigen-positive RBCs with trypsin, ficin, or papain.Similarly, Vw, Hut, and Nya antigens, which are locatedon the N-terminal side of the trypsin cleavage site onGPA,6-8 have the same characteristics (Fig. 1).

The Or antigen has tryptophan, instead of arginine, atamino acid residue 31,9,10 which ablates one of thetrypsin cleavage sites. Thus, depending on the extent oftrypsin treatment of antigen-positive RBCs, the Orantigen may be sensitive or weakened. The extent ofweakened reactivity of anti-Or with trypsin-treated


associated with the ERIK antigen, that other examplesof anti-ERIK may react differently with ficin- or papain-treated RBCs as seen with some examples of anti-Mta

and anti-Ria with different enzyme susceptibilities.14,15

The single amino acid changes associated with MARSand HAG antigens reside on GPA between the ficin/papain cleavage site and the RBC lipid bilayer.19,20 Bothantigens are, as expected, resistant to trypsin, α-chymotrypsin, ficin, or papain treatment (Fig. 1). Theamino acid substitution associated with the MARSantigen might be expected to introduce a novel enzymesite. However, because the antigen is resistant totreatment of RBCs with the proteolytic enzymes, it ispresumed that the site (close to the lipid bilayer) is notaccessible to the enzyme. The presence of either HAGor MARS antigens alters the expression of Wrb, a high-incidence antigen located on Band 3, but requiring an

MNS antigens and single amino acid changes

Or-positive RBCs may also be dependent on the precisespecificity of the antibody. The change of Arg31Trpon Or-positive RBCs provides an explanation for theobservation that the M antigen on Or-positive RBCs ismore resistant to trypsin treatment when comparedwith the M antigen on Or-negative (normal) RBCs.11

Vr and Osa antigens reside on GPA between the sitesfor trypsin and ficin/papain cleavage, and both areresistant to trypsin and sensitive to ficin or papaintreatment of antigen-positive RBCs (Fig. 1).7,12 Theamino acid substitution associated with the Vr antigen(tyrosine at residue 47) introduces a novel α-chymo-trypsin cleavage site. Thus, the Vr antigen is on GPA butsensitive to α-chymotrypsin treatment.

The Mta antigen is associated with a point mutationon GPA at residue 5812 and the Ria antigen is associatedwith a point mutation on GPA at residue 57.13 Theseamino acid positions are close to the cleavage sites forficin and papain and provide an explanation forvariation in reactivity with different examples of anti-Mta 14 and anti-Ria.15 The amino acid substitution asso-ciated with the Ria antigen introduces a novel trypsincleavage site, which explains the trypsin sensitivity ofthis antigen.15

The amino acid change associated with ERIK(Gly59Arg) is predicted to ablate a ficin cleavage siteand introduce a possible cleavage site for papain andtrypsin.16,17 The ERIK antigen is reported to be sensitiveto papain and partially sensitive to trypsin.18 However, itis possible, due to the position (close to the RBCmembrane) and nature of the substitution (Gly59Arg)

Table 1. MNS antigens associated with single amino acid changes inGlycophorin A

Antigen Amino acids Ficin/ α-chymo-Antigen ISBT # involved Papain Trypsin trypsin

Vw 002009 Thr28Met* Sensitive Sensitive Resistant

Vr 002012 Ser47Tyr Sensitive Resistant Sensitive

Mta 002014 Thr58Ile Partially Resistant Resistantsensitive

Ria 002016 Glu57Lys Partially Sensitive Resistantsensitive

Nya 002018 Asp27Glu Sensitive Sensitive Resistant

Hut 002020 Thr28Lys* Sensitive Sensitive Resistant

Or 002031 Arg31Trp Sensitive Partially Resistantsensitive

ERIK 002037 Gly59Arg Papain Partially Resistantsensitive sensitive

Ficinunknown

Osa 002038 Pro54Ser Sensitive Resistant Resistant

HAG 002041 Ala65Pro Resistant Resistant Resistant

MARS 002043 Gln63Lys Resistant Resistant Resistant

* Mechanism could be a single nucleotide substitution, or a GP(A-B-A) hybrid2

Fig. 1. GPA: Location of low-incidence antigens that are associatedwith a single amino acid change. The black vertical linerepresents glycophorin A. The location of M/N antigens andenzyme cleavage sites on intact RBCs are indicated to the left.The low-incidence antigens and the pertinent residue arenoted to the right of GPA.


between the α-chymotrypsin site and the lipid bilayer,22

which explains the resistance of Mit27,28 and of sD to thisenzyme (Fig. 2). The Mit antigen is partially sensitive topapain and resistant to ficin or trypsin treatment28

(Carole Green, personal communication). The resis-tance of the Mit antigen to ficin treatment is un-expected. There are only limited data regarding theprotease sensitivity of sD. The single report of anti-sD

notes that the antibody reacts “weaker with enzyme-treated cells,” but the enzyme is not named.29 The aminoacid substitution of Pro39Arg associated with the sD

antigen is C-terminal of the α-chymotrypsin, papain, andficin cleavage sites (and, thus, would be expected to beresistant to treatment of antigen-positive RBCs withthese enzymes) but creates a novel papain and trypsincleavage site. This may explain the weakened reactivityof anti-sD with enzyme treated sD-positive RBCs,although its close proximity to the lipid bilayer maymake the site relatively inaccessible to the enzyme andthereby allow only partial cleavage of the altered GPB.Our recent testing confirmed these presumtions (seeTable 2).

Table 2. MNS antigens associated with single amino acid changes inGlycophorin B

Antigen Amino acids Ficin/ α-chymo-Antigen ISBT # involved Papain Trypsin trypsin

S 002003 Met29 Sensitive Resistant Sensitives 002004 Thr29 Partially Resistant Sensitive

sensitiveMv 002021 Thr3Ser Sensitive Resistant SensitivesD 002023 Pro39Arg Partially Resistant Resistant

sensitiveMit 002024 Arg35His Papain Resistant Resistant

partiallysensitive

FicinResistant

M. E. REID AND J. R. STORRY

interaction with amino acids 59 to 76 of GPA for fullexpression.21

Low-incidence antigens on GPBThe low-incidence antigens, Mv, Mit, and sD, are each

associated with a single amino acid change on GPB22

(Table 2).5 The effect of papain, ficin, trypsin, orα-chymotrypsin treatment on antigen-positive RBCs isalso given in Table 2. The N-terminal 26 amino acids ofGPB usually are identical to the amino acid sequence ofGPA carrying the N antigen, but occasionally GPBcarries the He antigen.2 The N antigen on GPB is oftencalled ‘N’ (N quotes) in order to differentiate it from theN antigen on GPA. Diagnostic anti-N reagents areformulated to detect the N antigen on GPA, but not thaton GPB. Wild-type GPB does not have a trypsin cleavagesite on intact RBCs. The N or He antigens on GPB aresensitive to treatment of antigen-positive RBCs withficin, papain, or α-chymotrypsin but resistant to trypsintreatment. The Mv antigen, which is also located at theN-terminus of GPB is, as expected, sensitive to papain,ficin, or α-chymotrypsin.5,22–24 The presence of Mv

ablates the ‘N’ antigen.23,25,26

S and s antigens are carried on GPB at amino acidresidue 29 (methionine and threonine, respectively) andare sensitive to treatment of antigen-positive RBCs withα-chymotrypsin, resistant to trypsin treatment, andbehave variably after ficin or papain treatment. The Santigen is usually sensitive to ficin or papain treatmentof antigen-positive RBCs, but the s antigen is onlypartially sensitive as determined by most anti-s. Becauseneither methionine nor threonine are enzyme-cleavagesites, one can only presume that the cleavage sitesare more accessible on GPB carrying methionine (Santigen) than on GPB carrying threonine (s antigen) atposition 29. The Mit and sD antigens are located

Fig. 2. GPB: Location of low-incidence antigens that are associatedwith a single amino acid change.The black vertical linerepresents glycophorin B. The location of N/He antigens andenzyme cleavage sites on intact RBCs are indicated to the left.The low-incidence antigens and the pertinent residue arenoted to the right of GPB.


Weakened expression of S and s antigensThe presence of Mv, Mit, or sD on GPB weakens the

expression of S and s antigens.24,29–31 Dahr et al.3 showedthat, for expression of S and s antigens, the amino acidat position 29 is critical and that 25Thr (and/or theoligosaccharide attached to this residue), 28Glu, 34His,and 35Arg are also important (at least as defined bysome anti-S and anti-s).32 Thus, the change of Arg35 toHis associated with the Mit antigen would be expectedto alter the expression of the S or s antigen. Thereplacement of proline with arginine at residue 39 inGPB carrying the sD antigen would be expected to havea profound effect on the local conformation of GPB. Inaddition such a change could alter the expression of theS or s antigen due to the proximity of the amino acidsubstitution to the S or s antigenic determinant. Basedon analysis of the GPB amino acid sequence, thehydrophobic sequence from residues 36 to 66 formsthe α-helical membrane-spanning domain.32–34 Proline ismost commonly found in the N-terminal sequenceflanking the transmembrane region of type I membraneproteins and it is thought to be a helix initiator.35

Arginine is a highly positively charged amino acid, andstudies have shown that introduction of a positivelycharged amino acid at the extracellular boundary willforce the transmembrane helix out of the membrane.36

Therefore, it is likely that the amino acid substitution ofproline by arginine at position 39 (sD+) will alter theconformation of the glycophorin, thereby weakeningthe s antigen and, potentially, the insertion of GPB in themembrane. The substitution with arginine couldintroduce a trypsin and papain cleavage site, but it ispossible that residues this close to the lipid bilayer maynot be accessible to enzymes.

RBCs that are s positive, Mv positive, have a weakenedexpression of s,24,30 but there is no apparent effect on Santigen expression when it is carried in cis to Mv.24

Because the amino acid substitution associated with theMv antigen is a considerable distance from thatassociated with the S antigen, it is hard to understandhow Mv exerts its effect. Immunochemical analysis ofs-positive, Mv-positive RBCs by periodic acid-Schiff’s(PAS) staining of polyacrylamide gels showed that GPBassociated with the Mv haplotype is present atapproximately 25 percent of normal levels of the s formof GPB.23,25 In contrast, PAS staining of S-positive,s-positive, Mv-positive RBC membranes had grosslynormal amounts of GPB.22 It is possible that anydecrease in the amount of s-positive GPB was maskedby the S-positive GPB, as GPB carrying the S antigen is

present at 1.5 times the amount of GPB carrying the santigen.37 Thus, the weakening of s in Mv-positive RBCsis most likely due to the reduced GPB copy number inthese cells rather than to the amino acid substitutionhaving a direct effect on the s antigen.

Suppression of S or s antigen was more readilydemonstrated in early studies with RBCs carrying theMv, sD, or Mit antigens than is likely today. The morepotent contemporary reagents are less likely to showweakening of S and s antigens. As genotyping for bloodgroup alleles becomes more widespread, it is importantto be aware of variants that may affect oligonucleotideprimer design and amplification of alleles.

AcknowledgmentsWe thank Christine Lomas-Francis for critical review

of the manuscript and Robert Ratner for his assistancewith the manuscript and figures. This work wassupported in part by a National Institutes of HealthSpecialized Center of Research grant in transfusionmedicine and biology HL54459.

References1. Siebert PD, Fukuda M. Molecular cloning of a human

glycophorin B cDNA: nucleotide sequence andgenomic relationship to glycophorin A. Proc NatlAcad Sci USA 1987;84:6735–9.

2. Huang C-H, Blumenfeld OO. MNSs blood groups andmajor glycophorins: molecular basis for allelicvariation. In: Cartron J-P, Rouger P, eds. Molecularbasis of major human blood group antigens. NewYork: Plenum Press, 1995;153–83.

3. Dahr W, Beyreuther K, Steinbach H, et al. Structureof the Ss blood group antigens. II. A methionine/threonine polymorphism within the N-terminalsequence of the Ss glycoprotein. Hoppe-Seylers ZPhysiol Chem 1980;361:895–906.

4. Dahr W, Gielen W, Beyreuther K, Krüger J. Structureof the Ss blood group antigens. I. Isolation of Ss-active glycopeptides and differentiation of theantigens by modification of methionine. Hoppe-Seylers Z Physiol Chem 1980;361:145–52.

5. Reid ME, Lomas-Francis C. The blood group antigenfacts book. San Diego: Academic Press, 1996.

6. Giles CM. Serological activity of low frequencyantigens of the MNSs system and reappraisal of theMiltenberger complex. Vox Sang 1982;42:256–61.

7. Daniels GL, Bruce LJ, Mawby WJ, et al. The low-frequency MNS blood group antigens Nya (MNS18)and Osa (MNS38) are associated with GPA aminoacid substitutions. Transfusion 2000;40:555–9.



8. Dahr W, Newman RA, Contreras M, et al. Structuresof Miltenberger class I and II specific major humanerythrocyte membrane sialoglycoproteins. Eur JBiochem 1984;138:259–65.

9. Tsuneyama H, Uchikawa M, Matsubara M, et al.Molecular basis of Or in the MNS blood groupsystem (abstract). Vox Sang 1998;74 (Suppl 1):1446.

10. Reid ME, Sausais L, Øyen R, et al. First example ofhemolytic disease of the newborn caused by anti-Orand confirmation of the molecular basis of Or.Vox Sang 2000;79:180–2.

11. Bacon JM, Macdonald EB, Young SG, Connell T.Evidence that the low-frequency antigen Orriss ispart of the MN blood group system. Vox Sang1987;52:330–4.

12. Storry JR, Coghlan G, Poole J, et al. The MNS bloodgroup antigens, Vr (MNS 12) and Mta (MNS 14) eacharise from an amino acid substitution onglycophorin A. Vox Sang 2000;78:52–6.

13. Storry JR, Reid ME. A point mutation in GYPA exon3 encodes the low incidence antigen MNS16(abstract). Vox Sang 2000;78 (Suppl 1):P0025.

14. Holliman SM. MN blood group system: distribution,serology and genetics. In: Unger PJ, Laird-Fryer B,eds. Blood group systems: MN and Gerbich.Arlington, VA: American Association of Blood Banks,1989;1–29.

15. Contreras M, Armitage SE, Stebbing B. The MNSsantigen Ridley (Ria). Vox Sang 1984;46:360–5.

16. Huang C-H, Reid M, Daniels G, Blumenfeld OO.Alteration of splice site selection by an exonmutation in the human glycophorin A gene. J BiolChem 1993;268:25902–8.

17. Huang C-H, Reid M, Daniels G, Blumenfeld OO. Geneconversion between glycophorins A and B results inSta glycophorin in a family exhibiting the ERIK/Sta

blood group phenotype (abstract). Blood 1994;84(Suppl 1):238a.

18. Daniels GL, Green CA, Poole J, et al. ERIK, a low-frequency red cell antigen of the MNS blood groupsystem associated with Sta. Transfus Med1993;3:129–35.

19. Jarolim P, Moulds JM, Moulds JJ, et al. MARS and AVISblood group antigens: polymorphism of glycophorinA affects band 3-glycophorin A interaction(abstract). Blood 1996;88 (Suppl 1):182a.

20. Poole J, Banks J, Bruce LJ, et al. Glycophorin Amutation Ala65 —> Pro gives rise to a novel pair ofMNS alleles ENEP (MNS39) and HAG (MNS41) andaltered Wrb expression: direct evidence for GPA/

band 3 interaction necessary for normal Wrb

expression. Transfus Med 1999;9:167–74.21. Reid ME. Contribution of MNS to the study of

glycophorin A and glycophorin B. Immuno-hematology 1999;15:5–9.

22. Storry JR, Reid ME, MacLennan S, et al. The lowincidence MNS antigens, Mv, sD, and Mit arise fromsingle amino acid substitutions on glycophorin B.Transfusion 2001;41:269–75.

23. Dahr W, Longster G. Studies of Mv red cells. II.Immunochemical investigations. Blut 1984;49:299–306.

24. Gershowitz H, Fried K. Anti-M, a new antibody of theMNS blood group system. I. Mv, a new inheritedvariant of the M gene. Am J Hum Genet1966;18:264–81.

25. Dahr W, Longster G, Uhlenbruck G, Schumacher K.Studies on Miltenberger class III, V, Mv and Mk redcells. I. Sodium-dodecylsulfate polyacrylamide gelelectrophoretic investigations. Blut 1978;37:129–38.

26. Dahr W, Issitt P, Moulds J, Pavone B. Further studieson the membrane glycoprotein defects of S–s– andEn(a–) erythrocytes. Hoppe-Seylers Z Physiol Chem1978;359:1217–24.

27. Skradski KJ, McCreary J, Zweber M, Sabo B. Furtherinvestigation of the effect of Mitchell (Mit) antigenon S antigen expression (abstract). Transfusion1983;23:409.

28. Lubenko A, Savage JL, Gee SW, Cullen EM, Burslem SJ.Serology and genetics of the Mit antigen in NorthLondon blood donors (abstract). Proceedings of the20th Congress of the International Society of BloodTransfusion; 1988:116.

29. Shapiro M, Le Roux ME. Serology and genetics of a“new” red cell antigens: sD (the Dreyer antigen)(abstract). Transfusion 1981;21:614.

30. Beck ML, Hardman JT. Anti-s reagents (letter).Transfusion 1980;20:479.

31. Eichhorn M, Cross DE, Moulds M. Suppression of theS antigen by the MIT antigen: potential source oferror in red cell typing (abstract). Transfusion1981;21:614.

32. Dahr W. Immunochemistry of sialoglycoproteins inhuman red blood cell membranes. In: Vengelen-TylerV, Judd WJ, eds. Recent advances in blood groupbiochemistry. Arlington, VA: American Association ofBlood Banks, 1986;23–65.

33. Petrache HI, Grossfield A, MacKenzie KR, et al.Modulation of glycophorin A transmembrane helix

M. E. REID AND J. R. STORRY


interactions by lipid bilayers: molecular dynamicscalculations. J Mol Biol 2000;302:727–46.

34. Welsh EJ, Thom D, Morris ER, Rees DA. Molecularorganization of glycophorin A: implications formembrane interactions. Biopolymers 1985;24:2301–32.

35. Landolt-Marticorena C, Williams KA, Deber CM,Reithmeier RA. Non-random distribution of aminoacids in the transmembrane segments of humantype I single span membrane proteins. J Mol Biol1993;229:602–8.


36. von Heijne G. Structural aspects of transmembraneα-helices. Acta Physiol Scand Suppl 1998;643:17–9.

37. Dahr W, Uhlenbruck G, Schmalisch R, Janssen E. Ssblood group associated PAS-staining polymorphismof glycoprotein 3 from human erythrocytemembranes. Hum Genet 1976;32:121–32.

Marion E. Reid, PhD (corresponding author),Director, Immunohematology Laboratory, New YorkBlood Center, 310 East 67th Street, New York, NY10021; and Jill Storry, PhD, New York Blood Center,310 East 67th Street, New York, NY.

Attention SBB and BB Students: You areeligible for a free 1-year subscription toImmunohematology. Ask your education super-visor to submit the name and complete addressfor each student and the inclusive dates of thetraining period to Immunohematology, P.O. Box40325, Philadelphia, PA 19106.

Notice to Readers: All articles published,including communications and book reviews,reflect the opinions of the authors and do notnecessarily reflect the official policy of theAmerican Red Cross.

Attention: State Blood Bank Meeting Organizers: If you are planning a state meeting and wouldlike copies of Immunohematology for distribution, please contact Mary McGinniss, Managing Editor,3 months in advance, by phone or fax at (301) 299–7443.

IMPORTANT NOTICE ABOUT MANUSCRIPTS FOR IMMUNOHEMATOLOGYSubmit all manuscripts (original and 2 copies) to Mary H. McGinniss, Managing Editor, 10262 Arizona Circle,Bethesda, MD 20817. Please include the manuscript on a 3 1/2 inch disk, in Microsoft Word 97/6.0 orWordPerfect 6.0/5.1 or e-mail a copy to [email protected]

Manuscripts: The editorial staff of Immuno-hematology welcomes manuscripts pertaining toblood group serology and education forconsideration for publication. We are especiallyinterested in case reports, papers on platelet andwhite cell serology, scientific articles coveringoriginal investigations, and papers on the use ofcomputers in the blood bank. Deadlines for receipt ofmanuscripts for the March, June, September, andDecember issues are the first weeks in November,February, May, and August, respectively. Instructionsfor scientific articles and case reports can beobtained by phoning or faxing a request to Mary H.McGinnniss, Managing Editor, Immunohematology,at (301) 299–7443, or see “Instructions for Authors” inevery issue of Immunohematology or on the Web.


PEG adsorption of autoantibodiescauses loss of concomitantalloantibodyW. J. JUDD AND L. DAKE

Use of polyethylene glycol (PEG) to promote adsorption ofautoantibodies is reported to give good recovery of concomitantalloantibodies. In initial experiments, PEG and ZZAP (Ficin andDTT) adsorption procedures were compared for removal ofautoantibody and recovery of alloantibody. Postadsorption studies(n = 11) were performed and hemagglutination scores compared.In subsequent studies, equal volumes of alloantibody containingsera, PEG, and antigen-negative red blood cells (RBCs) were used intwofold adsorption experiments. Saline was substituted for PEG forcontrol purposes. Postadsorption titers and immunoglobulin levelswere determined. Autoantibodies were completely removed byboth methods (n = 5); better by PEG (n = 3); better by ZZAP (n =1); and not adsorbed (n = 1), and partially adsorbed by both (n = 1).Alloantibody recovery was comparable in three cases (E, K, Jka) butweaker by at least one reaction grade in four (K,E, Jka, and antibodyto low-frequency antigen). The latter anti-Jka reacted 1+ withJk(a+b+) RBCs after ZZAP adsorption but was nonreactive with thesame RBCs following PEG adsorption. Titers of six alloantibodiesadsorbed with antigen-negative RBCs in PEG were markedly weaker(range 2 to 8) compared to saline controls (range 4 to 32). IgGlevels for PEG adsorbed (range 128 to 243 mg/dL) were 50% lowerthan controls (range 265 to 505 mg/dL). Although PEG adsorptionis effective in removing autoantibody, the precipitation ofimmunoglobulin by PEG may result in failure to detect underlyingalloantibody. Immunohematology 2001;17:82–85.

Key Words: polyethylene glycol (PEG); PEG adsorption;autoantibody removal; detection of concomitantalloantibody; immunoglobulin precipitation by PEG

The promoting effect of polyethylene glycol (PEG)on red cell antigen-antibody interactions was firstdescribed in 1987 by Nance and Garratty.1 Since then,many transfusion service laboratories have incorporateda PEG method into their antibody detection and/oridentification testing protocols. In a 1995 letter to theeditor of Transfusion, Liew and Duncan2 suggested thatPEG be used in autoadsorption procedures to decreaseadsorption incubation times. They further suggestedthat the presence of PEG in the adsorbed serum wouldpromote detection of concomitant alloantibodies, ifpresent.

Champagne and Moulds3 questioned the use of PEGfor autoadsorption. They reported that, after PEG

adsorption, a weak alloanti-K was not detected and thereactivity of an alloanti-E was markedly weaker. In 1997,an evaluation of the PEG adsorption procedure waspublished in Immunohematology.4 The authorsreported no problems with the detection of con-comitant alloantibodies. Cheng and colleagues5 recentlyreported similar results.

Leger and colleagues6 have compared the efficacy ofa variety of methods available for autoantibody removaland the detection of underlying alloantibodies. Theyfound PEG adsorption to give similar results to thoseobtained by ZZAP adsorption,7 but PEG had theadvantages of eliminating the prior treatment of theadsorbing cells with ZZAP reagent (a mixture ofdithiothreitol and cysteine-activated papain), and a 50percent or more reduction in required adsorption time.They did mention the possibility of losing alloantibodyreactivity due to immunoglobulin precipitation. In aletter to Transfusion,8 they cautioned technologists totest PEG-adsorbed serum on the day of preparationbecause immunoglobulins precipitated upon storage.They also stated that if the adsorbed serum must bestored, it should be remixed and not centrifuged priorto testing.

We have used a modified ZZAP procedure9 since theoriginal method was described by Branch and Petz in1982.7 We simply substitute ficin for the cysteine-activated papain. Given the favorable reports of the PEGadsorption method, we performed crossover validationstudies, with a view to switching to the PEG adsorptionprocedure should our data support such a change. Ourfindings are presented in this report.

Materials and MethodsUnless otherwise noted, we used commercially

available PEG (PeG; Gamma Biologicals, Houston, TX).Twofold adsorptions were performed using 1 mL each


of PEG (added only to first adsorption), serum, andadsorbing red blood cells (RBCs). Each adsorption wasfor 15 minutes at 37oC. Following the secondadsorption, the sera were tested by the indirectantiglobulin test (4 volumes of adsorbed serum, 1volume of 3 to 4% RBCs, 30 minutes incubation at 37oC,anti-IgG). Reactions were graded and scored asdescribed by Marsh.10 Saline was substituted for PEG incontrol adsorption experiments.

ZZAP adsorptions were performed as describedelsewhere.9 Adsorption at 37oC was for 30 minutes.After the second adsorption, the sera were tested by asaline indirect antiglobulin technique (3 volumes ofadsorbed serum, 1 volume of 3 to 4% RBCs, 60 minutesat 37oC, polyspecific antihuman globulin reagent).9

Throughout this investigation, samples containingalloantibodies were adsorbed with RBCs that lacked thecorresponding antigen(s). Serum samples containedeither autoantibodies, autoantibodies spiked withknown alloantibodies, or only alloantibodies.

All other serologic procedures were performed asdescribed elsewhere.9 Immunoglobulin (IgG) levelswere measured by a standard nephelometric procedure(Beckman Array; Beckman Coulter, Inc., Fullerton, CA).

ResultsFigure 1 summarizes the results of adsorption studies

to remove warm reactive autoantibodies. In all of the 11cases, the unadsorbed sera reacted equal to or greaterthan 2+ by our routine antibody detection method(Löw and Messeter low-ionic-strength saline).11 Theassigned score values are the total of the reaction scoresobserved with all RBCs tested, divided by the number ofRBC samples. PEG and ZZAP autoadsorptions were noteffective in removing autoantibodies from one sample.Both methods completely removed autoantibody fromthree samples and reduced reactivity to about 1+ in twosamples; trace reactivity with some cells was seen infive samples, in three PEG-adsorbed samples, and in fourZZAP-adsorbed samples. These data confirm theobservations of Leger and Garratty6 that PEG and ZZAPmethods are essentially comparable for autoantibodyremoval.

We looked at a total of 12 alloantibodies and studiedthe effect of both PEG and ZZAP adsorption withantigen-negative RBCs. Figure 2 shows the results fromseven cases in which the reactivity of the PEG-adsorbedserum against antigen-positive RBCs was weaker thanthe reactivity of the ZZAP-adsorbed serum by one

Fig. 1. Reactivity of sera containing warm-reactive autoantibodiesfollowing ZZAP and PEG adsorptions. Scores represent averagereactivity of all red blood cells tested.

Loss of alloantibody with PEG adsorption

Fig. 2. Reactivity of alloantibodies following ZZAP and PEGadsorptions using red blood cells lacking the correspondingantigen. Scores represent average reactivity of all antigen-positive red blood cells tested.

reaction grade or more. Except for the antibody to anundefined low-frequency antigen, at least three antigen-positive RBCs were used to test the adsorbed sera; theassigned value represents the average score.

It is of note that an anti-Fya that gave a strong 1+(score 6) reaction with the ZZAP-adsorbed serum wasnot detected in the PEG-adsorbed sample. Also, an anti-c reacting 1+ strong after ZZAP adsorption gave onlytrace reactions (score 2) with PEG-adsorbed serum.


An anti-Jka reacting 2+ with ZZAP-adsorbed serum wasequally difficult to detect with PEG-adsorbed serum.Table 1 illustrates the stronger reactivity of this anti-Jka

after ZZAP adsorption compared to the PEG-adsorbedsample. To exclude the possibility that the enhancedanti-Jka reactions with the ZZAP-adsorbed serum weredue to bound complement (C3), as polyspecificantihuman globulin (containing anti-C3) was used fortesting, we repeated tests on the adsorbed serum withanti-IgG and obtained similar results.

Additional twofold PEG adsorptions of alloantibodieswith antigen-negative RBCs were performed. Forparallel control adsorptions, PEG was substituted withnormal saline. Titration studies were performed on theadsorbed sera using saline as a diluent, 60 minuteincubations at 37oC, and anti-IgG. The results aresummarized in Figure 3. With five of sevenalloantibodies, the titers after PEG adsorption withantigen-negative RBCs were twofold lower than thecontrol titers. With the other two antibodies, the titersafter PEG adsorption were onefold lower than thecontrol titers. Similar findings were obtained using a20% solution of 3350 MW PEG in isotonic saline (datanot shown).

Because PEG is known to precipitate proteins, welooked at the IgG levels of the PEG-adsorbed andcontrol (saline)-adsorbed sera. The PEG-adsorbedsamples had significantly lower IgG levels than thecontrols; on average, there was a 50 percent reductionin total IgG following PEG adsorption (Fig. 4).

To demonstrate that loss of antibody activity in thePEG-adsorption procedure is not solely an effect ofstorage of the postadsorbed samples, we incubatedequal volumes of serum and PEG in the absence ofRBCs for 15 minutes at 37oC, centrifuged the sample,and immediately measured the IgG levels in thesupernatants. Saline was substituted for PEG in controlexperiments. (Of note: As soon as PEG was added to thesera, a visible, cloudy precipitate formed, and a gela-tinous deposit was observed after centrifugation.) Welooked at six serum samples and measured the IgGcontent of the PEG-treated and control samples. Again,the IgG levels in the PEG-treated sera were on average50 percent lower than saline controls (Fig. 5).

DiscussionBased on our data, PEG adsorption appears to be less

effective than ZZAP adsorption for detectingconcomitant alloantibodies in patients with warmreactive autoantibodies. This is evident from our find-ing of one anti-Fya that was nonreactive after PEG

Table 1. Results of ZZAP and PEG adsorptions with Jk(a-b+) red blood cells on a serum containing alloanti-Jka and a warm reactive autoantibody.

Jk phenotype: a–b+ a+b– a–b+ a+b– a+b+ a+b+ a+b– a+b+ a–b+ a+b– a+b–

Unadsorbed 1+s 2+ 2+ 2+ 2+ 3+ 2+s 2+ 2+ 2+s 2+s

PEG-adsorbed 0 ± 0 w ± 1+w 1+ ± 0 1+w 1+w

ZZAP-adsorbed 0 1+S 0 1+ 2+ 2+ 2+ 2+ 0 2+s 2+s

W. J. JUDD AND L. DAKE

Fig. 3. Titers of alloantibodies after PEG and saline (control)adsorptions with red blood cells lacking the correspondingantigen.

Fig. 4. IgG levels in PEG and saline (control) adsorbed sera.


Fig. 5. IgG levels in supernatants from sera mixed with an equalvolume of PEG or saline.

Loss of alloantibody with PEG adsorption

adsorption with Fy(a+) RBCs, six antibodies that hadreduced reactivity by at least one reaction grade afterPEG adsorption, and a further seven alloantibodies thathad a one- to twofold decrease in titer after PEGadsorption. This loss of alloantibody reactivity isthe result of immunoglobulin precipitation by PEG.Autoantibody removal is also likely facilitated by thisPEG precipitation of immunoglobulin.

Our data are in conflict with other reports2, 4–6 on theuse of PEG-adsorbed sera for detecting concomitantalloantibodies in sera containing autoantibodies. In lightof these observations, we have chosen not toimplement PEG adsorption.

References1. Nance SJ, Garratty G. A new potentiator of red blood

cell antigen-antibody reactions. Am J Clin Pathol1987;87:633–5.

2. Liew YW, Duncan L. Polyethylene glycol in auto-adsorption of serum for detection of alloantibodies(letter). Transfusion 1995;35:713.

3. Champagne K, Moulds M. Autoadsorption fordetection of alloantibodies — should polyethyleneglycol be used (letter)? Transfusion 1996;36:384.

4. Barron CL, Brown MB. The use of polyethyleneglycol (PEG) to enhance adsorption of auto-antibodies. Immunohematology 1997;13:119–22.

5. Cheng C-K, Wong ML, Lee AW. PEG adsorption ofautoantibodies and detection of alloantibodies inwarm autoimmune hemolytic anemia. Transfusion2001;41:13–7.

6. Leger RM, Garratty G. Evaluation of methods fordetecting alloantibodies underlying warm auto-antibodies. Transfusion 1999;39:11–6.

7. Branch DR, Petz LD. A new reagent (ZZAP) havingmultiple applications in immunohematology. Am JClin Pathol 1982;78:161–7.

8. Leger RM, Ciesielski D, Garratty G. Effect of storageon antibody reactivity after adsorption in thepresence of polyethylene glycol (letter). Transfusion1999;39:1272–3.

9. Judd WJ. Methods in immunohematology. 2nd ed.Durham, NC: Montgomery Scientific Publications,1994.

10. Marsh WL. Scoring of hemagglutination reactions.Transfusion 1972;12:352–3.

11. Löw B, Messeter L. Antiglobulin tests in low-ionicstrength salt solutions for rapid antibody screeningand crossmatching. Vox Sang 1974;26:53–61.

W. John Judd, FIBMS, MIBiol, Professor ofImmunohematology, Department of Pathology, UH-2G332, University of Michigan Hospitals, 1500 EastMedical Center Drive, Ann Arbor, MI 48109–0054;and Louann Dake, MA, MT(ASCP)SBB, Coordinator,Reference Laboratory, Blood Bank and TransfusionService, Department of Pathology, UH-2G332,University of Michigan Hospitals, Ann Arbor, MI.

Phone, Fax, and Internet Information: If you have any questions concerning Immunohematology,Journal of Blood Group Serology and Education, or the Immunohematology Methods and Proceduresmanual, contact us by e-mail at [email protected]. For information concerning the NationalReference Laboratory for Blood Group Serology, including the American Rare Donor Program, please contactSandra Nance by phone at (215) 451–4362, by fax at (215) 451–2538, or by e-mail at [email protected]


Selecting an acceptable and safeantibody detection test can presenta dilemmaM. R. COMBS AND S. J. BREDEHOEFT

vacancies. Also, increasing clinical activity began tothreaten our capacity to maintain appropriatelaboratory support. Although satisfied with theperformance of PEG for antibody detection, we feltcompelled to evaluate new technologies that wouldeventually allow for automation. Our previouslyreported evaluation of the gel test (Ortho-ClinicalDiagnostics, Raritan, NJ),8,9 the solid phase test(Immucor, Norcross GA), 10,11 and the reports of otherinvestigators12 resulted in our continued use of PEG.When compared with gel, we found that PEG detectedmore potentially clinically significant antibodies thangel but without loss of specificity. The solid phase test,although exquisitely sensitive, yielded too manyunwanted positive results to be suitable for routineantibody detection in our laboratory.

Throughout 1998, staffing and workload issuescontinued to worsen, prompting reconsideration oftransition to automated methods. We chose toimplement gel testing paired with the Tecan MegaFlex-ID instrument (Micro Typing Systems Inc., PompanoBeach, FL), which automates pipetting the gel test. Wefound the gel test to be standardized and simple,with fewer steps and manipulations, and it providedconsistent and reproducible test results. As weconsidered transitioning from PEG to gel, we ques-tioned whether we might be sacrificing clinicallysignificant sensitivity. These preliminary concerns werelessened considerably due to the absence of reports ofincreased hemolytic reactions by gel users since the testhad been in use in the United States. We also reasonedthat perhaps some antibodies detected only in PEGmight not be clinically significant. For example, in aprevious report on the lack of significance of “enzyme-only antibodies,”13 we found that most of these enzyme-only antibodies that lacked clinical significance werealso detected by PEG.

The Transfusion Service at Duke University Hospital has changedantibody detection methods from the use of albumin in indirectantiglobulin tests to low-ionic-strength solution (LISS), and fromLISS to polyethylene glycol (PEG) in an effort to enhance the rapiddetection of clinically significant antibodies. In 1996, staffing issuesrequired the consideration of automation. Although previousstudies indicated that the gel test was not as sensitive as PEG fordetection of clinically significant antibodies, we chose to imple-ment the gel test to be used with the Tecan MegaFlex-ID. Weperformed a retrospective analysis of identified antibodies andtransfusion reactions to compare the outcomes of one year’sexperience with gel and PEG. We found comparable detection ofpotentially clinically significant antibodies by both methods andsignificantly fewer unwanted or clinically insignificant antibodiesdetected with the use of gel. Fewer delayed serologic transfusionreactions and no transfusion-associated hemolytic events occurredin the year that gel was used. Although we initially found theselection of the gel test to be a dilemma, our ultimate decisionappears to have successfully protected patient safety and balancedsensitivity with specificity. Immunohematology 2001;17:86–89

Key Words: antibody detection, gel, PEG

Since 1981, the Transfusion Service at DukeUniversity Hospital has eliminated many nonessentialtests and streamlined protocols while maintainingpatient safety. Our course of change has been verysimilar to that of the University of Michigan describedby Judd.1 Examples of changes include elimination ofa routine direct antiglobulin test on all patients,implementation of immediate spin compatibility testfor patients with no unexpected antibodies, andelimination of the test for weak D on patients. Inaddition to elimination of specific tests, we havechanged our antibody detection method from use ofalbumin in the indirect antiglobulin test to low-ionic-strength solution (LISS)2 in 1981 and from LISS topolyethylene glycol (PEG)3 10 years later. Each changein antibody detection method was made to enhancerapid detection of clinically significant antibodies.2–7

In 1996, we faced new challenges. Increasing scarcityof well-trained or adequately experienced technologistsbegan to create recurring difficulty in filling staff


showed that while using PEG, 71 percent of antibodiesidentified were potentially clinically significant and 29percent were considered unwanted or insignificant.While using gel, 80 percent of antibodies identifiedwere potentially clinically significant and 20 percentwere insignificant.

In 1998, we transfused 40,083 red cell units.Transfusion-related hemolytic events were reported infour patients. One reaction was due to mechanical orthermal effects, one was due to a delayed reactionbecause of anti-S in a sickle cell patient, and in twocases, there were no immunological causes of hemolysisdetected. Nineteen delayed reactions were discoveredas part of routine pretransfusion testing followingrecent transfusion and not as a result of reportedclinical hemolysis, thus fitting the definition of delayedserological transfusion reaction (DSTR) as described byNess et al.14 (Table 3). This was compared to the use of

We performed a retrospective analysis comparing theoutcome of one year’s experience with gel and PEG.The specificities of identified antibodies as well astransfusion reactions were evaluated.

Materials and MethodsIn 1998, PEG (3350 mw; Sigma, St. Louis, MO) was

used in antibody detection and identification tests andantiglobulin compatibility tests. Our method was thesame as described by Nance and Garratty3 except that15% PEG was used instead of 20%. With 20% PEG, wefound that there was inadequate removal of PEG-precipitated protein with the use of automated cellwashers and, as a result, IgG-coated control cells oftenwere not agglutinated. Reducing the concentration ofPEG alleviated much of this problem while stillmaintaining sensitivity. Two drops of patients’ serum, 4drops of 15% PEG, and 1 drop of 3% reagent or donorred blood cells (RBCs) were incubated at 37oC for 15minutes, followed by the indirect antiglobulin test withanti-IgG.

During 1999, we implemented gel testing (Ortho-Clinical Diagnostics) following parallel testing with PEGthat confirmed comparability in the test sample. In2000, we were using the gel test for antibody detection,identification, and antiglobulin compatibility tests.Twenty-five µL of patients’ serum and 50µL of 0.8% LISS-suspended reagent or donor RBCs were incubated for15 minutes at 37oC in gel cards containing anti-IgG(Micro Typing Systems). Following incubation, cardswere centrifuged for 10 minutes according tomanufacturer’s directions. Pipetting of reagents andsamples for antibody detection tests was performedeither manually or on the Tecan MegaFlex-ID.

To evaluate our experience for calendar years 1998and 2000, we conducted retrospective review ofrecords of antibody specificities and transfusionreactions.

ResultsIn 1998, while using PEG for antibody detection and

identification, 3085 antibodies were detected in 37,832samples (8%) submitted for pretransfusion testing. In2000, while using the gel test, 2715 antibodies weredetected in 43,405 samples (6%). Table 1 comparesthe percentages of potentially clinically significantor wanted specificities identified. Table 2 comparesinsignificant, or unwanted, antibodies identified. Overall,comparison of potentially clinically significant anti-bodies and clinically insignificant antibodies identified

Table 1. Comparative distribution of potentially clinically significantantibodies detected

Antibody PEG (1998) Gel (2000)

Rh 44.4% 51.9%

Duffy 5.4% 6.2%

Kidd 6.6% 4.0%

Kell 11.3% 15.0%

S/s 2.4% 2.7%

Total 3085 2715

Table 2. Comparative distribution of unwanted antibodies detected

Antibody PEG (1998) Gel (2000)

M/N 2.8% 2.6%

Lewis 6.3% 1.7%

P1

1.3% 0.1%

Inconclusive 7.7% 8.6%

Cold auto 2.2% 0.9%

Warm auto 5.8% 4.0%

HTLA* 2.4% 1.0%

Total 3085 2715

*High-titer–low-avidity

Table 3. Comparison of DSTR*

PEG (1998) Gel (2000)

4 anti-E 3 anti-K

3 anti-c 2 anti-c

3 anti-Fya 2 anti-E, -Fya

3 anti-Jka 1 anti-D

2 anti-Jkb 1 anti-Jka

2 anti-D

1 anti-c, -E

1 anti-E, -Jkb

Total: 19 DSTR Total: 9 DSTR

*Delayed serological transfusion reactions

Selecting an antibody detection test can present a dilemma


the gel test in 2000 when 42,204 RBC units weretransfused; no hemolytic events were reported and nineDSTRs were detected. During both years, in all cases,new antibody/ies that were detected in the eluate werealso detectable in the plasma.

DiscussionOur previous observations9 suggested the possibility

that gel might miss significant numbers of potentiallysignificant antibodies. However, analysis of antibodytesting and transfusion reaction experience using gelfailed to confirm this possibility. Gel detected slightlyhigher numbers of Rh and Kell antibodies and modestlylower numbers of anti-Jka (Table 1). During the year inwhich gel was being used, only one example of a DSTRwas detected due to anti-Jka, whereas three examples ofDSTRs due to anti-Jka were detected when PEG wasbeing used (Table 3). However, total clinically signifi-cant antibodies detected by the two methods arecomparable.

Comparison of unwanted antibodies agreed with ourperceptions as well as our previous reports; fewerunwanted antibodies are detected with the use of gel.This level of specificity is valuable, as additional testingsuch as cold antibody screens, prewarm testing, warmand cold adsorptions, and titrations are required toinvestigate unwanted antibodies, such as warm and coldautoantibodies and insignificant antibodies with high-titer–low-avidity characteristics. In addition to laborsavings, unnecessary delays in providing blood to thepatient are avoided when fewer unwanted positive testsare detected.

The comparison of transfusion reactions occurringwhile PEG and gel were used shows that no hemolyticreactions were reported and significantly fewer DSTRswere detected when gel was used. This agrees withthe commonly held belief that many of the antibodiesdetected only in PEG may not be clinically significant.

This report describes our experience of knowinglychanging to a method that appeared to have lesssensitivity than our current method. In our desire todetect as many clinically significant antibodies aspossible, our reluctance to change to the gel test in1996 suggests that perhaps we too were guilty ofoverkill as described by Issitt.15 As Issitt discussed,selection of an antibody detection method in the newmillenium must be based on prevailing circumstances.Our prevailing conditions of inadequate staffingrequired that we consider an alternative antibody

detection method that could be automated. Althoughwe initially found this to be a dilemma, our ultimatedecision appears to have successfully protected patientsafety and balanced sensitivity with specificity. Thesuccessful, safe transfusion of nearly 90,000 units of redcells since implementing gel technology reinforces thisjudgement.

References1. Judd WJ. Modern approaches to pretransfusion

testing. Immunohematology 1999;15:41–52.2. Löw B, Messeter L. Antiglobulin test in low-ionic-

strength salt solution for rapid antibody screeningand cross-matching. Vox Sang 1974;26:53–61.

3. Nance SJ, Garratty G. Polyethylene glycol. A newpotentiator of red blood cell antigen-antibodyreactions. Am J Clin Pathol 1987;87:633–5.

4. Wenz B, Apuzzo J, Shah DP. Evaluation of thepolyethylene glycol-potentiated indirect anti-globulin test. Transfusion 1990; 30:318–21.

5. deMan AJ, Overbeeke MA. Evaluation of the poly-ethylene glycol antiglobulin test for detection of redblood cell antibodies. Vox Sang 1990; 58: 207–10.

6. Slater JL, Griswold DJ, Wojtyniak LS, Reisling MJ.Evaluation of the polyethylene glycol-indirectantiglobulin test for routine compatibility testing.Transfusion 1989; 29:686–8.

7. Shirey RS, Boyd JS, Ness PM. Polyethylene glycolversus low-ionic-strength solution in pretransfusiontesting: a blinded comparison study. Transfusion1994;34:368–70.

8. Lapierre Y, Rigal D, Adam J, et al. The gel test: a newway to detect red cell antigen-antibody reactions.Transfusion 1990;30:109–13.

9. Issitt PD, Combs MR, Booth K. Comparison of theMTS-gel and polyethylene glycol (PEG) IAT methods(abstract). Transfusion 1997;37:64S.

10. Plapp FV, Sinor LT, Rachel JM, et al. A solid phaseantibody screen. Am J Clin Pathol 1984;82:719–21.

11. Issitt PD, Combs MR, Bradshaw T. Comparison of asolid phase and polyethylene glycol (PEG) IATmethod in a large transfusion service (abstract).Transfusion 1997;37:28S.

12. Reilly B, DePalma H, Cosgrove P, Parsi M, Grima K.Comparison between routine tube test (LISS), PEGtube test, gel technology and solid-phase red celladherence assays (abstract). Transfusion 1997;37:64S.

M. R. COMBS AND S. J. BREDEHOEFT


13. Issitt PD, Combs MR, Bredehoeft SJ, et al. Lack ofclinical significance of “enzyme-only” red cellantibodies. Transfusion 1993;33:284–93.

14. Ness PM, Shirey RS, Thoman SK, Buck SA. Thedifferentiation of delayed serologic and delayedserologic and delayed hemolytic transfusionreactions: incidence, long-term serologic findings,and clinical significance. Transfusion 1990; 30:688–93.

15. Issitt PD. From kill to overkill: 100 years of (perhapstoo much) progress. Immunohematology 2000;16:18–25.

Martha Rae Combs, MT (ASCP), SBB, AnalyticalSpecialist, Duke University Hospital, TransfusionService, Box 2928, Durham, NC 27710; Steven J.Bredehoeft, MD, Medical Director, Duke University,Transfusion Service, Durham, NC.

Selecting an antibody detection test can present a dilemma

Free Classified Ads and Announcements: Immunohematology will publish classified ads andannouncements (SBB schools, meetings, symposia, etc.) without charge. Deadlines for receipt of theseitems and the projected mailing weeks of the journal are as follows:

Deadlines

1st week in January for the March issue1st week in April for the June issue

1st week in July for the September issue1st week in October for the December issue

Fax these items to Mary H. McGinniss, Managing Editor, (301) 299–7443.

Attention: Presidents of State Blood BankAssociations—In order to increase the numberof subscribers to Immunohematology, we aresoliciting membership lists of your organizations.Upon receipt of such a list, each person willreceive a complimentary copy of Immunohema-tology, and, if desired, a personal letter from theassociation president. For further information,contact: Mary H. McGinniss, Managing Editor, byphone or fax at (301) 299–7443.

IMMUNOHEMATOLOGYIS ON THE WEB!

redcross.org/immunohematologyPassword “2000”

For more information or to send an e-mail messageto the editor: [email protected]


C O M M U N I C A T I O N S

Letter to the Editors

A hemolytic transfusion reaction due to anti-Kundetected by a LISS antibody screen

Significant hemolytic transfusion reactions caused byanti-K have been reported, including a case in whichthe antibody failed to react using various low-ionic-strength solution (LISS) reagents.1,2 We identified ananti-K that was initially undetectable by LISS techniqueand that resulted in an acute hemolytic transfusionreaction.

An 84-year-old, group O– female with a history ofatrial fibrillation, permanent pacemaker insertion,osteoarthritis, and Alzheimer’s dementia, underwentopen reduction-internal fixation for a right hip fracture.Her transfusion history was unknown although priortransfusion was likely due to hip surgery many yearsago. The preoperative antibody screen was found to benegative using LISS antiglobulin technique (N-Hance;Gamma Biologicals, Inc., Houston, TX) with a 15-minuteincubation time at 37oC. Two group O– red blood cell(RBC) units were transfused intraoperatively based oncompatible immediate-spin saline crossmatch results. Asper our blood bank protocol, antiglobulin crossmatcheswere completed for the two RBC units, although a35-minute delay was noted due to heavy work flowat the time. One of the two units was found to be1+ incompatible by the antiglobulin crossmatch, butthe unit had been transfused prior to completion ofthe test.

Clinically, 6 hours after surgery and intraoperativetransfusion of the two RBC units, the patient suddenlybecame hypoxic and unresponsive after receivingnarcotic pain medication. She was given oxygen bymask and treated with narcan and sodium bicarbonate,after which she became responsive although agitated.Shortly after, scanty, dark urine was noted along witha drop in hemoglobin to 8 g/dL from 11 g/dLpreoperatively, and a hemolytic transfusion reaction wassuspected owing to transfusion of the incompatibleRBC unit. One hour later, she became hypoxic andhypotensive (blood pressure: 70/50 mmHg), requiringintubation as well as IV fluid and renaldose dopaminesupportive therapy. Two additional units of RBCs werealso given. Four units of fresh frozen plasma weretransfused due to disseminated intravascular coagu-lation, but the patient later developed ecchymoses and

a left groin hematoma. The next morning, she went intocardiac arrest in the surgical intensive care unit and waspronounced dead. An autopsy was not performed.

On follow-up testing using LISS antiglobulintechnique, anti-K was identified in the preoperativesample (titer of 8) with a 30-minute incubation at 37oCbut was not detectable with a 15-minute incubation,confirming our suspicion that prolonged incubationtime was required for detection. Testing of thehemolyzed posttransfusion sample with either 15- or30-minute incubation time was negative, whereas thedirect antiglobulin test was 1+ positive for C3d andnegative for IgG: the eluate was also negative. Theincompatible unit was confirmed to be K+. Increasingthe serum-to-reagent RBC drop ratio from 2:1 to 3:1 andrepeating the tests on the pretransfusion sample with15- to 30-minute incubations once again yieldednegative and positive results, respectively.

Although it is likely that a number of factors wereinvolved in this patient’s demise, including herunderlying medical condition as well as the possibleadverse effect of narcotic medication, we believe thehemolytic reaction contributed significantly. Wesubsequently eliminated the antiglobulin crossmatchprocedure for patients with no antibody history and anegative antibody detection test. We conclude that thiswas indeed an example of an anti-K that wasnonreactive in LISS using our routine testing procedure;however, it is possible that the initial inability to detectthe antibody was due to factors unrelated to LISS, suchas weakened antigen expression on the reagentscreening cells after storage. Because anti-K as well asother RBC antibodies had been detected in many othercases using the same screening cell reagents maintainedunder the same conditions, we feel that the latter ismuch less likely. This case, and others like it, shouldserve to humble us even as we take confidence in ourtechnologic breakthroughs and abbreviated testingprocedures.

Mark T. Friedman, DOAlan P. Carioti, MT(ASCP)SBB

Lutheran Medical CenterBlood Bank

150 55th StreetBrooklyn, NY 11220


Letter From the Editors

Ortho DedicationThis is the 11th year that Ortho-Clinical Diagnostics

has sponsored the September issue of Immuno-hematology. The editors and readers would like tothank the management of Ortho-Clinical Diagnostics forsupporting the publication and distribution of theJournal. In addition to their donation to support thepublication of each September issue, they distributemany free copies of each issue to their customers.

Please tell your Ortho sales representative how muchyou appreciate their generosity and concern foreducation.

Delores MalloryEditor-in-Chief

Mary McGinnissManaging Editor

References1. Sazama K. Reports of 355 transfusion-associated

deaths: 1976 through 1985. Transfusion 1990;30:583–90.

C O M M U N I C A T I O N S C O N T ’ D

ERRATA

Vol. 17, No. 2, 2001

Page 58

The first paragraph of the response letter

from Martha Combs, et al. is missing several

words. The last two sentences should read:

“Three patients received one CPDA-1 unit

each and no other antibody-containing unit.

One patient received one CPDA-1 unit and

one additive unit containing alloantibodies.”

2. Molthan L, Strohm PL. Hemolytic transfusionreaction due to anti-Kell undetectable in low-ionic-strength solutions. Am J Clin Pathol 1981;75:629–31.


I N M E M O R I U M

William C. Sherwood, MD1934 – 2001

Dr. William Sherwood, medicaleditor of Immunohematology for 5years and a nationally respectedleader and authority on transfusionmedicine and blood centeroperations, died at his home at age67.

Early in his career, Dr. Sherwoodwas a captain in the U.S. Army andserved as chief of hematology atSecond General Hospital, Landstahl,Germany. He was chief of JeffersonHematology Service at PhiladelphiaGeneral Hospital.

In 1971 he became the medicaldirector and then director of thePenn-Jersey region of the AmericanRed Cross. He coordinated the effortsof the Penn-Jersey region to create asingle system of blood supply in thearea.

Dr. Sherwood served as seniorvice president of the American RedCross Biomedical Services. Among his many successesin the blood transfusion field, his interest in and skillswith computers directed the establishment of the firstAmerican Red Cross blood center computer program inthe Penn-Jersey region in 1975. While at NationalHeadquarters as senior vice president, he helped begin

the work of creating a singlecomputer system to serve allAmerican Red Cross blood centers.

Dr. Sherwood was known locallyand nationally as an innovativeexpert in transfusion medicine andwas honored in 1992 with theAmerican Red Cross Charles R.Drew Award and by thePennsylvania Association of BloodBanks with the Lyndall Molthan, MD,Memorial Lecture Award.

Dr. Sherwood was on the editorialboard as medical editor forImmunohematology from 1996until his retirement in 2000. He wasan advisor to Immunohematologyon the construction and content ofthe Immunohematology Web page,one of the first to appear on theAmerican Red Cross Web site.

Dr. Sherwood was an inspired andinspiring physician and a person

who worked tirelessly on behalf of the patients anddonors.




I N M E M O R I U M

John Case, FIBMS, FIMLS1926 – 2001

John Case was born December 29,1926 in Poole, Dorset, England. Hebegan his career in medicallaboratory technology serving in theRoyal Army Medical Corps from1945 to 1948. He worked atWembley and Whipscross Hospitalsin London and the South LondonBlood Transfusion Center in Sutton,Surrey. He and his family immigratedto New Zealand in 1959, where hewas in a charge position at the bloodbank at Dunedin Hospital. Moving toAustralia in 1971, he was at theCommonwealth Serum Laboratoriesin Melbourne. In 1976 he came toHouston, Texas to become vicepresident of Regulatory Affairs atGamma Biologicals, Inc., where heremained until his retirement inJanuary 1999.

John was an expert in blood group serology wellbefore he became involved with writing of productapplications, package inserts, and FDA responses. Hisearly work was with low-frequency antigens. He was anexpert on the blood group systems and his own redcells were discovered to be the rare Rg–. He was afaithful panel donor and his red cells have been used inserologic problem solving around the world.

He was known worldwide for his influence onkeeping English “proper” in publications.

He was fearless in hiscommitment to his views andhonest to his ideals and his friends.John was a member of committeesin the World Health Organizationand the American Association ofBlood Banks (AABB). Heparticipated in many technicallectures for yearly tutorials atGamma Biologicals, Inc. and for theAABB. For his contributions to thefield, he was greatly honored. Fromthe Australian Society of BloodTransfusion, he received the RuthSanger Oratorical Award. The KayBeattie Award was given from theMichigan Association of BloodBanks. The L. Jean Stubbins Awardwas presented from the SouthCentral Association of Blood Banks.

The AABB presented him with both the Sally FrankAward and the Ivor Dunsford Memorial Award. John wasa great communicator both in writing and orally.

John loved his roses and was a talented grower. Hegrew champion roses and was a long-time member ofand an author for an Australian rose association. He willbe greatly missed by his many friends and admirers.




I N M E M O R I U M

Fred Stratton, BSc, MD1913 – 2001

Dr. Fred Stratton was born October 18, 1913 andspent his childhood in a Quaker home in Levenshulme,Manchester, England.

He received a BSc in 1934, an MB ChB in 1937, adiploma in public health in 1939, an MD withcommendation in 1944, and a DSc in 1957. He was afounding Fellow of the Royal College of Pathologistsand was granted a Fellowship of the Royal College ofPhysicians for his outstanding work in transfusionmedicine.

Dr. Stratton began his career in blood transfusionmedicine in 1940 when he joined the “Blood Depot” inManchester as part of the plans for blood collection inWWII. He continued with various positions until hebecame regional transfusion director in 1949 and heldthis post until his retirement March 31, 1980.

In addition to his excellent work in blood transfusionmedicine at the Manchester Center, he becameinterested in blood group serology, particularlycomplement and blood group antibodies, and was thefirst to recognize the weaker form of the Rh antigen,

which became known as Du. His knowledge oflaboratory tests was profound, and in 1958, with hisassociate Dr. Peter Renton, he published the importanttext Practical Blood Groupings.

His tremendous accomplishments were honored in1963 with the Oliver Memorial Fund Award foroutstanding contributions to science and practice ofblood transfusion. In 1977, he was awarded a PersonalChair in Human Serology at the University ofManchester. In 1978, he was awarded the KarlLandsteiner Award, the highest award of the AmericanAssociation of Blood Banks. He was a founder of, andin 1981, the inaugural president for the British BloodTransfusion Society. In 1987, the BBTS awarded him itshighest award, the James Blundell Award.

Dr. Stratton was truly one of the pioneers of bloodtransfusion medicine and blood group serology.




A N N O U N C E M E N T S

International Society of Blood Transfusion. The11th Regional Western Pacific Congress of theInternational Society of Blood Transfusion will be heldin Shanghai, China, from November 10–13, 2001 at theExhibition and Conference Hall located in the ShanghaiRainbow Hotel. The program will offer plenary lectures,scientific symposiums, and poster sessions on the latestadvances in blood banking and transfusion medicine.There also will be planned social activities to promoteunderstanding and camaraderie among the differentnationalities and cultural backgrounds.

Important Dates• Early registration deadline: July 31, 2001• Hotel registration deadline: August 30, 2001; phone

(81–62) 62753388; fax (86–21) 62753736; e-mail:[email protected]

• Preregistration deadline: October 31, 2001

For further information, Contact: Congress Secretary,Prof. ZHU Youg Ming, Shanghai Blood Center, #1191,Hong Qiao Road, Shanghai 200051, China; phone:(81–62) 62780789; fax: (86–21) 62958414; e-mail:[email protected]

Monoclonal antibodies available. The New YorkBlood Center has developed murine monoclonalantibodies that are useful for donor screening and fortyping red cells with a positive direct antiglobulin test.Anti-Rh:17 is a directly agglutinating monoclonalantibody. Anti-Fya, anti-K, anti-Jsb, and anti-Kpa areindirect agglutinating antibodies that require anti-mouseIgG for detection. These antibodies are available inlimited quantities at no charge to anyone who requeststhem. Contact: Marion Reid, New York Blood Center,310 E. 67th Street, New York, NY 10021; e-mail:[email protected]

A D V E R T I S E M E N T S

Granulocyte Antibody Detection and Typing

• Specializing in granulocyte antibodydetection and granulocyte antigen typing

• Patients with granulocytopenia can beclassified through the following tests forproper therapy and monitoring:

Granulocyte agglutination (GA)Granulocyte immunofluorescence (GIF)Monoclonal Antibody Immobilization of

Granulocyte Antigens (MAIGA)

For information regarding services, call Gail Eiberat (651) 291–6797, e-mail: [email protected]

or write to:

Neutrophil Serology Reference LaboratoryAmerican Red Cross

St. Paul Regional Blood Services100 South Robert Street

St. Paul, MN 55107

CLIA licensed

NATIONAL REFERENCE LABORATORYFOR BLOOD GROUP SEROLOGY

Immunohematology Reference LaboratoryAABB, ARC, New York State, and CLIA licensed

(215) 451–4901 – 24 hr. phone number(215) 451–2538 – Fax

American Rare Donor Program

(215) 451–4900 – 24 hr. phone number(215) 451–2538 – [email protected]

Immunohematology

(215) 451–4902 – Phone, business hours(215) 451–2538 – Fax

[email protected]

Quality Control of Cryoprecipitated-AHF

(215) 451–4903 – Phone, business hours(215) 451–2538 – Fax


A D V E R T I S E M E N T S C O N T ’ D

Reference and Consultation Services

Red cell antibody identification and problemresolution

HLA-A, B, C and DR typingHLA-disease association typing

Paternity testing/DNA

For information regarding our services, contactZahra Mehdizadehkashi at (503) 280–0210,

or write to:

Pacific Northwest Regional Blood ServicesATTENTION: Tissue Typing Laboratory

American Red Cross3131 North Vancouver

Portland, OR 97227

CLIA licensed, ASHI accredited

National Platelet SerologyReference Laboratory

Diagnostic testing for:• Neonatal alloimmune thrombocytopenia

(NAIT)• Posttransfusion purpura (PTP)• Refractoriness to platelet transfusion• Heparin-induced thrombocytopenia (HIT)• Alloimmune idiopathic thrombocytopenia

purpura (AITP)• Medical consultation available

Test methods:• GTI systems tests

- detection of glycoprotein-specific plateletantibodies

- detection of heparin-induced antibodies(PF4 ELISA)

• Platelet suspension immunofluorescence test(PSIFT)

• Solid-phase red cell adherence (SPRCA) assay• Monoclonal antibody immobilization of

platelet antigens (MAIPA)

For information, e-mail:[email protected]

or call:

Maryann Keashen-Schnell(215) 451–4041 office

(215) 451–4205 laboratory

Sandra Nance(215) 451–4362

Scott Murphy, MD(215) 451–4877

American Red Cross Blood Services

Musser Blood Center

700 Spring Garden Street

Philadelphia, PA 19123–3594

CLIA licensed

IgA/Anti-IgA Testing

IgA and anti-IgA testing is available to—

• Monitor known IgA-deficient patients.• Investigate anaphylactic reactions.• Confirm IgA-deficient donors.

Our ELISA assay for IgA detects antigento 0.05 mg/dL.

For information on charges and samplerequirements, call (215) 451–4351, e-mail:[email protected]

or write to:

American Red Cross Blood ServicesMusser Blood Center

700 Spring Garden StreetPhiladelphia, PA 19123–3594

Attn: Ann Church

CLIA licensed


SCIENTIFIC ARTICLES, REVIEWS, AND CASE REPORTSBefore submitting a manuscript, consult current issues of

Immunohematology for style. Type the manuscript on white bondpaper (8.5" x 11") and double-space throughout. Number the pagesconsecutively in the upper right-hand corner, beginning with thetitle page. Each component of the manuscript must start on a newpage in the following order:1. Title page2. Abstract3. Text4. Acknowledgments5. References6. Author information7. Tables—see 6 under Preparation8. Figures—see 7 under Preparation

Preparation1. Title page

A. Full title of manuscript with only first letter of first wordcapitalized

B. Initials and last name of each author (no degrees; all CAPS),e.g., M.T. JONES

C. Running title of ≤ 40 characters, including spacesD. 3 to 10 key words

2. AbstractA. 1 paragraph, no longer than 200 wordsB. Purpose, methods, findings, and conclusions of studyC. Abstracts not required for reviews

3. Text (serial pages)Most manuscripts can usually, but not necessarily, be dividedinto sections (as described below). Results of surveys andreview papers are examples that may need individualizedsections.A. Introduction

Purpose and rationale for study, including pertinentbackground references.

B. Case Report (if study calls for one)Clinical and/or hematologic data and background serology.

C. Materials and MethodsSelection and number of subjects, samples, items, etc. studiedand description of appropriate controls, procedures,methods, equipment, reagents, etc. Equipment and reagentsshould be identified in parentheses by model or lot andmanufacturer’s name, city, and state. Do not use patients’names or hospital numbers.

D. ResultsPresentation of concise and sequential results, referring topertinent tables and/or figures, if applicable.

E. DiscussionImplications and limitations of the study, links to otherstudies; if appropriate, link conclusions to purpose of studyas stated in introduction.

4. AcknowledgmentsAcknowledge those who have made substantial contributions tothe study, including secretarial assistance.

5. ReferencesA. In text, use superscript, arabic numbers.B. Number references consecutively in the order they occur in

the text.C. Use inclusive pages of cited references, e.g., 1431–7.D. Refer to current issues of Immunohematology for style.

6. TablesA. Number consecutively, head each with a brief title, capitalize

first letter of first word (e.g., Table 1. Results of ...), and use nopunctuation at the end.

B. Use short headings for each column, and capitalize first letterof first word.

C. Place explanations in footnotes (sequence: *, †, ‡, §, ¶, **, ††).7. Figures

A. Figures can be submitted either drawn or photographed

(5" x 7" glossy).B. Place caption for a figure on a separate page (e.g., Fig. 1.

Results of ...), ending with a period. If figure is submitted asa glossy, put title of paper and figure number on back of eachglossy submitted.

C. When plotting points on a figure, use the following symbolswhen possible: �� .

8. Author informationA. List first name, middle initial, last name, highest academic

degree, position held, institution and department, andcomplete address (including zip code) for all authors. Listcountry when applicable.

SCIENTIFIC ARTICLES AND CASE REPORTS SUBMITTEDAS LETTERS TO THE EDITOR

Preparation1. Heading—To the Editor:2. Under heading—title with first letter capitalized.3. Text—write in letter format (paragraphs).4. Author(s)—type flush right; for first author: name, degree,

institution, address (including city, state, and zip code); for otherauthors: name, degree, institution, city, and state.

5. References—limited to ten.6. One table and/or figure allowed.

ImmunohematologyJOURNAL OF BLOOD GROUP SEROLOGY AND EDUCATION

Instructions for Authors

Send all submissions (original and two copies) to:Mary H. McGinniss, Managing Editor, Immunohematology,10262 Arizona Circle, Bethesda, MD 20817.Include your manuscript on a 3 1/2" disk, in MicrosoftWord 97/6.0 or Word Perfect 6.0/5.1 or e-mail yourmanuscript to [email protected]

Now available from Montgomery Scientific Publications

APPLIED BLOOD GROUPSEROLOGY, 4th EDITION

by Peter D. Issitt and David J. Anstee

A totally revised, mostly rewritten, fully up-to-date edition of one of the most popular books about the blood groups and blood transfusion ever published.

� 46 chapters, an increase of 16 over the third edition• 1,208 plus xxiv 81/2″× 11″ pages, hardbound, fully indexed, over 1,500 entries

� 260 tables and 112 figures, an increase of more than 60% over the third edition• Over 13,500 references, more than 5,000 are papers written since 1985

Prices; each includes shipping: USA $125.00; Canada/International $130.00 (surface mail); International $170.00 (air mail).

ALL ORDERS MUST BE PREPAID(Check or Credit Card) in U.S. DOLLARS

International orders by check drawn on a bank in the USAor by credit card please.

Order from:Montgomery Scientific Publications, P.O. Box 2704, Durham, NC 27715, U.S.A.

Credit card orders accepted by fax at (919) 489-1235(No phone orders, please.)

We accept VISA, MasterCard, and Discover Card.

Musser Blood Center700 Spring Garden StreetPhiladelphia, PA 19123-3594

(Place Label Here)

32391.immuno 17 3 - red cross blood...transfusion history was also obtained, which included the...

Documents