tm12 - the positive dat and immune-mediated red …...c3d/red cell, although newer reagents and more...

18 The Positive DAT andImmune-Mediated RedCell Destruction

18

Significance of a PositiveDATThe direct antiglobulin test (DAT) is gen-erally used to determine if red cells havebeen coated, in vivo, with immuno-globulin, complement, or both. A posi-tive DAT, with or without shortened redcell survival, may result from:

1. Autoantibodies to intrinsic red cellantigens.

2. Alloantibodies in a recipient’s circu-lation, reacting with antigens on re-cently transfused donor red cells.

3. Alloantibodies in donor plasma,plasma derivatives, or blood frac-tions that react with antigens on thered cells of a transfusion recipient.

4. Alloantibodies in maternal circula-tion that cross the placenta and coatfetal red cells.

5. Antibodies directed against certaindrugs that bind to red cell mem-branes (eg, penicillin).

6. Adsorbed proteins, including immu-noglobulins, that attach to abnor-mal membranes or red cells modi-

Copyright © 2002 by the A

fied by therapy with certain drugs,notably those of the cephalosporingroup.

7. Complement components or, rarely,IgG bound to red cells after the ad-min istrat ion o f drugs such asquinidine and phenacetin has in-duced a drug/anti-drug interaction.

8. Nonantibody immunoglobulins as-sociated with red cells in patientswith hypergammaglobulinemia orrecipients of high-dose intravenousgammaglobulin.1,2

9. Antibodies produced by passengerlymphocytes in transplanted or-gans.3

A positive DAT does not necessarilymean that a person’s red cells have ashortened survival. Small amounts ofboth IgG and complement appear to bepresent on all red cells. A range of 5-90molecules of IgG/red cell4 and 5-97 C3dmolecules/red cell5 (possibly up to 500or more) appears to be normal on thecells of healthy individuals.

The DAT, as routinely performed, candetect a level of 100-500 molecules ofIgG/red cell and 400-1100 molecules of

379ABB. All rights reserved.

380 AABB Technical Manual

C3d/red cell, although newer reagentsand more sensitive test methods maydetect lower levels of these proteins. Asmany as 15% of hospital patients,6 andbetween 1 in 1000 and 1 in 14,000 blooddonors, have a positive DAT withoutclinical manifestations of immune-me-diated red cell destruction.

Most blood donors with positive DATsappear to be perfectly healthy, althoughsome may show signs of increased celldestruction or develop autoimmunehemolytic anemia (AIHA) over time.7

Elevated levels of IgG or complementhave been noted on the red cells of pa-tients with sickle cell disease, β-thalas-semia, renal disease, multiple myeloma,autoimmune disorders (including sys-temic lupus erythematosus), AIDS, andother diseases with elevated serumglobulin or blood urea nitrogen (BUN)levels with no clear correlation betweenpositive DAT and anemia.2,8 Interpreta-tion of positive DATs must include thepatient’s history, clinical data, and theresults of other laboratory tests.

The Direct AntiglobulinTestThe principles of the DAT are discussedin Chapter 11. Although any red cellsmay be tested, anticoagulated samplesare preferred to avoid in-vitro uptake ofcomplement; EDTA is the anticoagulantmost frequently recommended. If redcells from a clotted blood sample have apositive DAT, the results should be con-firmed on cells from a freshly collectedEDTA-anticoagulated specimen.

Most DATs are initially performedwith a polyspecific reagent capable ofdetecting both IgG and C3d (see Method3.6). If positive, tests with more specificanti-IgG and anti-complement reagentsmay be appropriate. Occasionally, poly-specific antiglobulin reagents react with


cell-bound proteins other than IgG orC3d (eg, IgM, IgA or other complementcomponents); specific reagents to distin-guish these proteins are not readilyavailable in most laboratories.

The Pretransfusion DAT and theAutologous ControlNeither the AABB Standards for BloodBanks and Transfusion Services9 nor anyother accrediting agency requires a DATor an autologous control (autocontrol)as part of pretransfusion testing, butmany workers advocate this practice.Several studies, however, have shownthat eliminating the DAT/autocontrolportion of routine pretransfusion test-ing carries minimal risk.6,10

Evaluation of a Positive DAT

Extent of TestingClinical considerations should dictatethe extent to which a positive DAT isevaluated. Dialogue with the attendingphysician is important. Interpretation ofthe significance of serologic findings re-quires knowledge of the patient’s diag-nosis; recent drug, pregnancy, and trans-fusion history; and information on thepresence of acquired or unexplainedhemolytic anemia. The results of sero-logic tests are not diagnostic; their sig-nificance must be assessed in conjunc-tion with clinical information and suchlaboratory data as hematocrit, bilirubin,haptoglobin, and reticulocyte count.

Answers to the following questionsmay help decide what investigations areappropriate:

1. Is there any evidence of in-vivo redcell destruction?

Reticulocytosis, hemoglobinemia,hemoglobinuria, decreased serum hap-toglobin, and elevated levels of serumunconjugated bilirubin or lactate dehy-drogenase (LD), especially LD1, may beassociated with increased red cell de-

ABB. All rights reserved.

Chapter 18: Positive DAT and Immune-Mediated Hemolysis 381

struction. If an anemic patient with apositive DAT does show evidence ofhemolysis, a workup to evaluate a possi-ble immune etiology is appropriate. IFTHERE IS NO EVIDENCE OF IN-CREASED RED CELL DESTRUCTION,NO FURTHER STUDIES ARE NECES-SARY, unless the patient needs transfu-sion and the serum contains incom-pletely identified unexpected antibodiesto red cell antigens.

2. Has the patient been recently trans-fused?

Many workers routinely attempt todetermine the cause of a positive DATwhen the patient has received transfu-sions within the previous 3 months, be-cause the first indication of a developingimmune response may be the attach-ment of antibody to recently transfusedred cells. Mixed-field reactivity is theclassical observation but may be difficultto observe. Antibody may appear as earlyas 7-10 days after transfusion in primaryimmunization and within 2-7 days in asecondary response; these alloantibodiescould shorten the survival of cells al-ready transfused or given in later trans-fusions.

Antibodies coating the patient’s redcells and/or the transfused red cells mayalso be present in serum. Elution doneto evaluate a positive DAT often concen-trates antibody activity and may facili-tate identification of weakly reactive se-rum antibodies. If the serum containsmultiple antibodies, the eluate may con-tain only one or a few, and this mayclarify antibody identification.

If a non-group O patient has receivedplasma containing anti-A or anti-B (as intransfusion of group O platelets), ABOantibodies may be detected on the recipi-ent’s red cells without necessarily imply-ing significantly accelerated red cell de-struction. If, however, the recipient ofincompatible plasma does appear to haveimmune hemolysis, the eluate can, in-itially, be tested against A and/or B cells;

Copyright © 2002 by the

if predicted ABO antibodies are not de-tected, other causes of the positive DATshould be sought.

Intravenous immunoglobulin (IVIG)may contain ABO antibodies, anti-D or,sometimes, other antibodies. Signifi-cant red cell destruction due to passivetransfer of antibodies from IVIG is rare.

3. Is the patient receiving any drugs,such as procainamide, α-methyl-dopa, or intravenous penicillin?

Approximately 3% of patients receiv-ing intravenous penicillin, usually atvery high doses, and 15-20% of patientsreceiving α-methyldopa will develop apositive DAT, but fewer than 1% of thosepatients who develop a positive DAT havehemolytic anemia. Procainamide ther-apy is also associated with positiveDATs.11 Cephalosporins also are associ-ated with positive DATs but less oftenwith immune red cell destruction. Posi-tive DATs associated with other drugs arerare. If a positive DAT is found in a pa-tient receiving such drugs, the attendingphysician should be alerted so that ap-propriate surveillance for red cell de-struction can be maintained. If red cellsurvival is not shortened, no furtherstudies are necessary.

4. Has the patient received a bonemarrow or organ transplant?

Passenger lymphocytes of donor ori-gin produce antibodies directed againstABO or other antigens on the recipient’scells, causing a positive DAT.3

5. Is the patient receiving antilympho-cyte globulin (ALG) or antithymo-cyte globulin (ATG)?

Patients receiving ALG or ATG of ani-mal origin may develop a positive DATwithin a few days, apparently related tohigh-titer heterophile antibodies inthese products and the presence of cor-responding antibodies in animal-derivedantihuman globulin sera.12 Free, un-bound ALG/ATG in the circulation maycause positive results when the patient’sserum is used in indirect antiglobulin

AABB. All rights reserved.


testing. Problems due to heterophile an-tibodies can usually be circumvented byneutralization or by use of the low ionicpolycation technique (see Method 3.2.5).

Serologic Studies

Three investigative approaches are help-ful in the evaluation of a positive DAT:

1. Test the DAT-positive cells with anti-IgG and anti-C3d antiglobulin re-agents to characterize the types ofproteins coating the red cells.

2. Test the serum/plasma to detect andidentify clinically significant anti-bodies to red cell antigens.

3. Test an eluate (see Methods 5.1-5.6)prepared from the coated red cellsagainst a panel of antigen-typed red

Table 18-1. Antibody Elution TechniquMethod AdvantagesHeat (56 C) Good for ABO-HDN; quic

easy methodFreeze-Thaw Good for ABO-HDN; quic

easy method; requires smvolume of cells

Cold Acid Quick and easy method;sensitivity comparable todigitonin-acid

Digitonin Acid Nonhazardous; good recmost antibodies; availabcommercial kit

Dichloromethane(DCM)

Good for anti-K; nonflam

Chloroform Rapid method; nonflammrecovers most blood groantibodies

Xylene Recovers most blood groantibodies

Ether Excellent recovery of mogroup antibodies

*Modified from Judd13 and Sigmund.14


cells to define whether the coatingprotein has alloantibody activity.When the only coating protein iscomplement, eluates are frequentlynonreactive.

Results of these tests, combined withthe patient’s history and clinical data,should assist in classification of theproblems involved.

Elution

Elution frees antibody from sensitizedred cells and recovers it in a usable form.Details of eluate preparation are given inChapter 16 and in Methods 5.1-5.6. Table18-1 lists the advantages and disadvan-tages of several common elution meth-ods; no single elution method is ideal in

es*

Disadvantagesk and Poor recovery of other blood

group allo- and autoantibodiesk andall

Poor recovery of other bloodgroup allo- and autoantibodies

Less sensitive for warm auto-and alloantibodies

overy ofle in

Time-consuming washing ofstroma; less sensitive for Kiddantibodies

mable Toxic; poor recovery of anti-Fyb

able;up

Carcinogenic, toxic, narcotic;less sensitive than xylene orether

up Flammable, carcinogenic, toxic,narcotic; hemolysis of test cellsif residual xylene not removed

st blood Highly flammable, explosive, toxic,narcotic; requires special storage;poor recovery of anti-S, -s



all situations. Although many elutionmethods damage or destroy the red cells,certain techniques (see Methods 2.11,2.13, 2.14) remove antibody but leavethe cells sufficiently intact that they canbe typed for various antigens or used foradsorption purposes. Some antigensmay be altered by elution, however, andappropriate controls are essential.

When the cause of the positive DAT isunknown, an eluate may be prepared andtested against a panel of red cells.

In cases of hemolytic disease of thenewborn (HDN) or hemolytic transfu-sion reactions, specific antibody (or an-tibodies) is usually detected in the elu-ate. Usually the same specificity can bedetected in the patient’s (or, in HDN, themother’s) serum, although eluates mayhelp in antibody identification when se-rum reactions are weak. When the eluatereacts with all cells tested, autoantibodyis the most likely explanation, especiallyif the patient has not been recentlytransfused. WHEN NO UNEXPECTEDANTIBODIES ARE PRESENT IN THESERUM, AND IF THE PATIENT HAS NOTBEEN RECENTLY TRANSFUSED, NOFURTHER SEROLOGIC TESTING OF ANISOLATED AUTOANTIBODY IS NECES-SARY.

Sometimes no reactivity is detected inthe eluate, despite reactivity of the cellswith specific anti-IgG. The cause may bethat the eluate was not tested againstcells positive for the corresponding anti-gen, notably group A or group B cells.Antigens of low incidence are also absentfrom most reagent cell panels. It may beappropriate to test the eluate against redcells from recently transfused donorunits, which could have stimulated analloantibody to a rare antigen, or, inHDN, against cells from the father, fromwhom the infant may have inherited arare gene.

Reactivity of eluates can be enhancedby testing against enzyme-treated cellsor by the use of solid-phase or other


enhancement techniques, such as poly-ethylene glycol (PEG). Antibody reactiv-ity can be increased by the use of a con-centrated eluate, either by alteration ofthe fluid-to-cell ratio or by use of com-mercial concentration devices. Washingthe red cells with low ionic strength sa-line (LISS), instead of normal saline,may prevent the loss of antibody whilethe cells are being prepared for elution.

Certain elution methods give poor re-sults with certain antibodies. When elu-ates are nonreactive yet clinical signs ofred cell destruction are present, elutionby a different method may be helpful. Ifboth serum and eluate are nonreactive atall test phases, and if the patient hasreceived high-dose intravenous penicil-lin or other drug therapy, testing todemonstrate drug-related antibodiesshould be considered. Patients may havea positive DAT and nonreactive eluatewith no evidence of hemolysis, and ex-haustive pursuit of an explanation is notusually indicated.

Immune-MediatedHemolysisImmune-mediated hemolysis (immunehemolysis) is the shortening of red cellsurvival by the product(s) of an immuneresponse. If bone marrow compensationis adequate, the reduced red cell survivalmay not result in anemia. Immunehemolysis is only one cause of anemia,and many causes of hemolysis are unre-lated to immune reactions. The sero-logic investigations carried out in theblood bank do not determine whether apatient has a “hemolytic” anemia. Thediagnosis of hemolytic anemia rests onclinical findings and such laboratorydata as hemoglobin or hematocrit val-ues; reticulocyte count; red cell mor-phology; bilirubin, haptoglobin, and LDlevels; and, sometimes, red cell survival



studies. The serologic findings help de-termine whether the hemolysis has animmune basis and, if so, what type ofimmune hemolytic anemia is present.This is important since the treatment foreach type is different.

The terms hemolysis and hemolytic,frequently used to indicate both in-travascular and extravascular red cell de-struction, may be misleading. As a de-scription of in-vitro antibody reactivity,hemolysis—lysis of the red cells withrelease of free hemoglobin to the sur-rounding media—is both obvious andrare. In-vivo lysis of cells and release of

Table 18-2. Classification of Immune

Autoimmune Hemolytic Anemia1. Warm Autoimmune Hemolytic Anem

a. primary (idiopathic)b. secondary (to such conditions a

erythematosus (SLE), carcinom2. Cold Agglutinin Syndrome


infectious mononucleosis)3. Mixed-Type AIHA


4. Paroxysmal Cold Hemoglobinuriaa. primary (idiopathic)b. secondary (to such conditions a

5. DAT-Negative AIHAa. primary (idiopathic)b. secondary (to such conditions a

Drug-Induced Hemolytic Anemia

Alloimmune Hemolytic Anemia1. Hemolytic disease of the newborn2. Hemolytic transfusion reaction


free hemoglobin within the intravascu-lar compartment is uncommon and,often, dramatic. Most immune red celldestruction occurs extravascularly, withlittle or no escaping hemoglobin, even inthe immune “hemolytic anemias.”

Immune hemolytic anemias can beclassified in various ways. One classifica-tion system is shown in Table 18-2.Autoimmune hemolytic anemias (AI-HAs) are subdivided into four majortypes: warm antibody AIHA (WAIHA),cold agglutinin syndrome (CAS), mixed-type AIHA, and paroxysmal cold hemo-globinuria (PCH). Not all cases fit neatly

Hemolytic Anemias

ia

s lymphoma, systemic lupusa, or to drug therapy)

s lymphoma, mycoplasma pneumonia,

s SLE, lymphoma)

s syphilis, viral infections)

s lymphoma, SLE)



into these categories. Immune hemo-lysis may also be induced by drugs (dis-cussed in a later section of this chapter).The prevalence of each type can vary de-pending on the patient population stud-ied.15-18 (See Table 18-3.)

Findings in a DAT performed withanti-IgG and anti-C3d may reduce thediagnostic possibilities; eg, the DAT inCAS is almost always positive using anti-C3d and negative using anti-IgG. Table18-4 summarizes the expected DAT find-ings in the various autoimmune hemo-lytic anemias.

DATs performed with IgG- and C3-specific antiglobulin reagents, as well asthe serum and eluate studies describedearlier, can be used to classify AIHAs intoone of the four major types (WAIHA,CAS, mixed-type AIHA, PCH). Two addi-tional procedures may be useful: a diag-nostic cold agglutinin titer (Method 6.2)and the Donath-Landsteiner test forPCH (Method 6.7). Table 18-4 summa-rizes the expected serologic results withthe major types of AIHA.

The binding of antibody to red cellsdoes not, in itself, damage the cells. It isthe phenomena that the boundantibody-antigen complex promotesthat may eventually damage cells. Theseinclude complement binding, adherenceto Fc receptors on macrophages andmonocytes, phagocytosis, and cytotoxiclysis. The IgG subclass of bound anti-body may be significant. IgG1 is the sub-class most commonly found, sometimesalone but often in combination withother subclasses. The other IgG sub-classes occur more often in combinationwith other subclasses than alone. In gen-eral, IgG3 antibodies have the most dis-tinctive effects, followed by IgG1. IgG2antibodies are associated with less de-struction and IgG4 with no or little de-struction. Some monoclonal antiglobu-lin reagents may not detect IgG4.

The number of antibody moleculesper cell also plays a role. The number of


antibody molecules on the cells of appar-ently healthy blood donors with positiveDATs (<1000, weakly positive) is far lessthan that usually seen in patients withAIHA.4,5 Some patients with apparentimmune hemolysis may have negativefindings on a routine DAT.

Warm Antibody AutoimmuneHemolytic AnemiaThe most common type of AIHA is asso-ciated with warm-reactive (37 C) anti-bodies. Typical serologic findings are de-scribed below.

DATWhen IgG-specific and complement-speci f ic antig lobul in reagents areused, three patterns of reactivity maybe found: coating with IgG alone, withcomplement alone, or with both. Inapproximately 1% of cases the DAT willbe positive with a polyspecific antihu-man globulin reagent but negativewith IgG- and complement-specific an-tiglobulin reagents. Some of these maybe due to attachment of IgM or IgAalone.16,18

SerumAutoantibody in the serum typically isIgG and reacts by indirect antiglobulintesting against all cells tested. If theautoantibody has a high binding con-stant and has been adsorbed by the pa-tient’s red cells in vivo, the serum maycontain very little free antibody. The se-rum will contain antibody after all thespecific antigen sites on the red cellshave been occupied and no more anti-body can be bound in vivo. The DAT insuch cases is usually strongly positive.Approximately 50% of patients withWAIHA have serum antibodies that reactwith untreated red cells. With suchmethods as tests with PEG, enzyme-treated red cells, or solid phase methods,


Table 18-3. Serologic Findings in DAT-Positive AIHA

WAIHA CASMixed-Type

AIHA PCHDrug-Induced

AIHAPercentof cases

48%18 to 70%19 16%19 to 32%18 7%18 to 8%20 Rare inadults19

32% inchildren25

12%19 to18%18

DAT 20%19 to 66%18

IgG24%18 to 63%19

IgG+C37%18 to 14%19

C3

91%18 to 98%19

C3 only71%18 to 100%19

IgG + C394%18 to100%19

C3 only

94%18 IgG6%18 IgG + C3

Immuno-globulintype

IgG (sometimesIgA or IgM,rarely alone)

IgM IgG, IgM IgG IgG

Eluate IgG antibody Nonreactive IgG antibody Nonreactive IgG antibodySerum 57% react by IAT;

13% hemolyzeenzyme-treatedred cells at 37 C;90% agglutinateenzyme-treatedred cells at 37 C;30% agglutinateuntreated redcells at 20 C;rarely agglutinateuntreated redcells at 37 C19

IgMhemagglutinatingantibody; titerusually >1000 at4 C; usually reactat 30 C inalbumin;monoclonalantibody inchronicdisease19,23

IgG IAT-reactiveantibody; IgMhemagglutinatingantibody, usuallyreact at 30-37 Cin saline; titer at4 C often<6418,20,21

IgGbiphasichemolysin(Donath-Landsteinerantibody)19,23

IgG antibodysimilar toWAIHA19,23

Specificity 80% Type I,20% Type II22

specificity; Rhspecificity; otherspecificitieshave beenreported*

Usually anti-I butcan be anti-i;rarely anti-Pr24

Usuallyspecificityunclear18,20,21,26;can be anti-I, -i,or other coldagglutininspecificities

Anti-P(nonreactivewith p andPk redcells19,24

Specificityoften Rhrelated2,19

*See text.


over 90% of these sera can be shown tocontain autoantibody. Autoantibodiesthat hemolyze untreated red cells in vi-tro at 37 C are rare but, if present, maybe associated with significant in-vivo de-struction of autologous and transfusedcells. Approximately 15% of sera containantibodies that hemolyze enzyme-


treated red cells at 37 C. Approximatelyone-third of patients with warm-anti-body AIHA have cold-reactive autoagglu-tinins demonstrable in tests at 20 C, butcold agglutinin titers at 4 C are normal;this does not necessarily mean the pa-tient has CAS in addition to WAIHA (seemixed-type AIHA below).


Table 18-4. DAT Results in AIHA Using Anti-IgG and Anti-C3 Reagents18

WAIHA CASMixed-Type

AIHA* PCHDrug-

Induced

No. of patients 355 275 61 17 157IgG only† 67% 1% 18% 0 94%IgG + C3† 24% 1% 71% 0 6%C3 only† 7% 91% 9.8% 94% 0Polyspecific-positive,monospecific-negative‡

1% 0.7% 0 0 0

DAT-negative 1% 6.3% 1.2% 6% 0

*Patients had both cold hemagglutinating autoantibodies reactive to 30 C or higher andwarm-reactive autoantibodies.†With or without IgM and/or IgA.‡Polyspecific antihuman globulin gives a positive result but monospecific anti-IgG and anti-C3reagents are negative.


EluateThe presence of the IgG autoantibody onthe red cells can be confirmed by elution(see Methods 5.1-5.6). Typically the elu-ate reacts with virtually all cells tested,with reactivity enhanced in tests againstenzyme-treated cells or when PEG isused, but occasionally IgG is present atsuch low levels that the eluate is nonre-active. The eluate will usually have noserologic activity if the only proteincoating the red cells are complementcomponents. Occasionally antibody notdetected by the DAT will be detected inthe eluate, possibly due to the concen-trating effect of eluate preparation.

Specificity of AutoantibodyThe specificity of autoantibodies associ-ated with WAIHA appears very complex.In routine tests, all cells tested are usu-ally reactive. Some autoantibodies thathave weaker or negative reactivity withcells of rare Rh phenotypes such as –D–,or Rhnull appear to have broad specificityin the Rh system. Apparent specificityfor simple Rh antigens (D, C, E, c, e) isoccasionally seen, either as the sole


autoantibody or as a predominant por-tion, based on stronger reactivity withcells of certain phenotypes. Such reac-tivity is often termed a “relative” speci-ficity. Such relative specificity in a se-rum may be mistaken for alloantibody,but cells negative for the apparent targetantigen can adsorb and remove the“mimicking” specificity.

Unusual Specificities. Apart from Rhspecificity, warm autoantibodies withmany other specificities have been re-ported. When autoantibody specificity isto antigens in the Kell system (and per-haps others), the patient’s red cells mayexhibit depressed expression of the anti-gen and the DAT may be negative (seebelow). Dilution and selective adsorp-tion of eluates may uncover specificityor relative specificity of autoantibodies.

It has also been suggested that warm-reactive autoantibodies may be dividedinto those that react preferentially witholder cells and those that react with botholder and reticulocyte-enriched popula-tions.23 This theory has not been proven,but, if true, suggests that antibody tosenescent red cell antigen may be in-volved.



Practical Significance. Tests againstcells of rare phenotype and by specialtechniques have limited clinical applica-tion. In rare instances of WAIHA involv-ing IgM agglut inins , determiningautoantibody specificity may help differ-entiate such cases from typical CAS.27 Itis rarely, if ever, necessary to ascertainautoantibody specificity in order to se-lect antigen-negative blood for transfu-sion. If apparent specificity is directed toa high-incidence antigen (eg, anti-U), orwhen the autoantibody reacts with allred cells except those of a rare Rh phe-notype (eg, –D–, Rhnull), compatible do-nor blood is unlikely to be available andthere is little point in determining speci-ficity. Such blood, if available, should bereserved for alloimmunized patients ofthat uncommon phenotype. In rare casesin which patients respond very poorly toinfusion of “incompatible” red cells, itmay be appropriate to consider using redcells lacking the Rh or high-incidenceantigens.

Transfusion-Stimulated Autoanti-bodies. Transfusion itself may lead to theproduction of autoantibodies, which maypersist and cause positive DATs for sometime after transfusion yet not cause obvi-ous red cell destruction. Such cell-boundautoantibodies sometimes display bloodgroup specificity (eg, E, K, Jka), but persistlong after transfused red cells should havedisappeared from the circulation, appar-ently adsorbed to the patient’s own anti-gen-negative red cells.28

Transfusion of Patients withWarm-Reactive AutoantibodiesInherent Risks. Patients with warm-re-active autoantibodies range from thosewith no apparent shortening of red celllife span to those with life-threateninganemia. Patients with little or no evi-dence of significant in-vivo red cell de-struction often tolerate transfusionquite well, if required for reasons unre-


lated to the autoantibody. In general,however, transfusion of patients withWAIHA carries a greater-than-normalrisk; even when autoantibody is not de-monstrable in the serum, transfusedblood is not truly compatible and effectsof the transfusion are likely to be rela-tively short-lived.

When autoantibody is active in se-rum, it will be difficult to exclude thepresence of alloantibodies, which in-creases the risk of adverse reaction.Transfusion may stimulate alloantibodyproduction, complicating subsequenttransfusions. Transfusion may intensifythe autoantibody, inducing or increas-ing hemolysis and making serologictesting more difficult. Transfusion maydepress compensatory erythropoiesis;destruction of transfused cells may in-crease hemoglobinemia and hemoglo-b inur ia . In pat ients wi th act ivehemolysis, transfused red cells may bedestroyed more rapidly than the pa-tient’s own red cells. In rare cases thismay promote hypercoagulability anddisseminated intravascular coagulation.Transfusion reactions, if they occur, maybe difficult to investigate.

Alternatives to Transfusion. Giventhe inherent risks of transfusion, alter-native treatments should be considered.Corticosteroids, which inhibit antibodyproduction and may also inhibit releaseof lysosomal enzymes, are generally con-sidered a primary therapy. Marked in-vivo hemolysis may require intravenousadministration of high doses of steroid.Intravenous immunoglobulin or immu-nosuppressive drugs such as cyclophos-phamide may benefit some patients withautoantibody.14 Splenectomy is some-times effective, but is usually consideredonly when other therapies are ineffectiveor contraindicated. In cases of severehemolysis, plasma exchange or treat-ment with complement-inhibiting sub-stances (eg, heparin) have been sug-gested, although without supportive



documentation. Filtration of the plasmathrough immunoadsorption columns(eg, containing protein A) may tempo-rarily reduce the level of plasma IgG.

Transfusion in WAIHA. Transfusion isespecially problematic for patients withrapid in-vivo hemolysis, who may presentwith a very low hemoglobin level and hy-potension. Reticulocytopenia may accom-pany a rapidly falling hematocrit, and thepatient may exhibit coronary insuffi-ciency, congestive heart failure, cardiacdecompensation, or neurologic impair-ment. Under these circumstances transfu-sion is usually required as a life savingmeasure. The transfused cells may supportoxygen-carrying capacity until the acutehemolysis diminishes or other therapiescan effect a more lasting benefit. Thesepatients represent a significant challengebecause serologic testing may be complexwhile clinical needs are acute.

Transfusion should not be withheldsolely because of serologic incompatibil-ity. The volume transfused should usu-ally be the smallest amount required tomaintain adequate oxygen delivery, notnecessarily to reach an arbitrary hemo-globin level. Volumes of about 100 mLmay be appropriate. The patient shouldbe carefully monitored throughout thetransfusion. The dangers of transfusionmay be enhanced if the patient hasmixed-type AIHA.

Transfusion in Chronic WAIHA. Mostpatients with WAIHA have a chronic sta-ble anemia, often at relatively low hemo-globin levels. Those with hemoglobinlevels above 8 g/dL rarely require trans-fusion, and many patients with levels of5 g/dL (or even lower) can be managedwith bed rest, without transfusion.Transfusion will be required if the ane-mia progresses or is accompanied bysuch symptoms as severe angina, cardiacdecompensation, respiratory distress,and cerebral ischemia.

Most patients with chronic anemiadue to WAIHA tolerate transfusion with-


out overt reactions, even though thetransfused cells may not survive any bet-ter than their own. Transfusion volumewill be significantly greater than for thepatient with acute hemolysis. Sincetransfusion may lead to circulatory over-load or to increased red cell destruction,the decision to transfuse should be care-fully considered. A few patients withoutacute hemolys i s have had severehemolytic reactions after transfusion,sometimes with no exceptional serologicevidence of autoantibody. This may bedue to the sudden availability of largevolumes of donor cells and the exponen-tial curve of decay, by which the numberof cells hemolyzed is proportional to thenumber of cells present.7 Reports of re-actions in patients with HIV suggestthese patients may be more susceptibleto transfusion-induced DIC and hyper-coagulability.25,26

Selecting Blood by Specificity. If thedecision is made to transfuse, selectionof appropriate donor blood is essential.It is important to determine the patient’sABO type and, if time permits, to detectpotentially clinically significant alloan-tibodies. Adsorption and other specialtechniques described later in this chap-ter can greatly reduce the risk of unde-tected alloantibodies but may be time-consuming. If clinically significantalloantibodies are present, the trans-fused cells should lack the correspond-ing antigen(s).

If the autoantibody has apparent andrelatively clear-cut specificity for a sin-gle antigen (eg, anti-e) and there is ac-tive ongoing hemolysis, blood lackingthat antigen may be selected. There isevidence that in some patients such redcells survive better than the patient’sown red cells.13,14 In the absence ofhemolysis, autoantibody specificity canusually be ignored, although donorunits negative for the antigen may bechosen because this is a simple way tocircumvent the autoantibody and detect



potential alloantibodies. If the autoanti-body shows broader reactivity, reactingwith all cells but showing some relativespecificity (eg, reacts preferentially withe-positive red cells), the use of bloodlacking the corresponding antigen is de-batable. It may be undesirable to exposethe patient to Rh antigens absent fromautologous cells, especially D and espe-cially in women who may later bear chil-dren, simply to improve serologic com-patibility testing with the autoantibody(eg, when a D-negative patient hasautoanti-e).

Selecting Blood by Random Testing.In many cases of WAIHA no autoantibodyspecificity is apparent.The patient’s se-rum reacts with all red-cell samples tothe same degree, or reacts with red cellsfrom different donors to varying degreesfor reasons seemingly unrelated to Rhphenotypes. Even if specificity is identi-fied, the exotic cells used for such iden-tification are not available for transfu-sion. Some workers test a large numberof donor blood samples and select thoseunits that give the weakest reactions invitro, although no data exist to indicatebetter in-vivo survival of these least-in-compatible red cells. Other workers rec-ommend ignoring any specificity of theautoantibody and, if the presence of al-loantibody has been excluded, advocategiving blood of the patient’s Rh pheno-type to avoid subsequent development ofRh alloantibodies. However, in cases ofsevere and progressive anemia, it may beessential to transfuse blood that does notreact with the patient’s autoantibody.

Suitable Components. Red bloodcells are preferable to whole blood iftransfusion is undertaken. Use of leuko-cyte-reduced components can lessen thechance of febrile reactions and make theinterpretation of such reactions, shouldthey occur, more readily attributable tored cell antibodies.

Frequency of Testing. If the patienthas not been transfused or pregnant


since a previous investigation, mostworkers consider repeat testing to beunnecessary. Each transfusion makespretransfusion testing for subsequenttransfusions more complicated. Exten-sive adsorption tests are impractical ifrepeated transfusion is required within ashort period of time. Although not allimmunologists agree, Judd29 has sug-gested that repeated extensive serologictesting is not needed within a week of theprevious investigation, unless results ofroutine tests for unexpected antibodiesare significantly different from those ofprevious tests. Repeat testing may some-times be streamlined by testing adsorbedserum against cells from recently trans-fused donor units, because these are theones likely to have stimulated any cur-rent alloantibody production. This ap-proach cannot be used if the patient hasreceived multiple transfusions over anumber of days.

Cold Agglutinin SyndromeCold agglutinin syndrome (also calledcold hemagglutinin disease, CHD) is thehemolytic anemia most commonly asso-ciated with cold-reactive autoantibodiesand accounts for approximately 16-32%of all cases of immune hemolysis.16,17

(See Table 18-3.) It occurs as an acute orchronic condition. The acute form isoften secondary to lymphoproliferativedisorders (eg, lymphoma) or Myco-plasma pneumoniae in fect ion. Thechronic form, which causes mild to mod-erate hemolysis, is often seen in elderlypatients, sometimes associated withlymphoma, chromic lymphocytic leuke-mia, or Waldenstrom’s macroglobu-linemia; Raynaud’s phenomenon andhemoglobinuria may occur in coldweather. CAS is often characterized byrapid sedimentation, at room tempera-ture, of red cells in an EDTA specimen;clumping of red cells may be obvious insuch a sample, sometimes so strong that



the cells appear to be clotted. Problemswith ABO and Rh typing and other testsare not uncommon.

DATComplement is the only protein detectedon the red cells in almost all cases. Thecold-reactive autoagglutinin is usuallyIgM, which binds to red cells in the com-paratively low temperature of the pe-ripheral circulation, and causes comple-ment components (C3 and C4 inparticular) to attach to the red cells. Asthe red cells circulate to warmer areas,the IgM dissociates, but the complementremains. Regulatory proteins convertthe bound C3 and C4 to C3d and C4d, andit is the anti-C3d (-C4d) component ofpolyspecific antiglobulin reagents thataccounts for the positive DAT. The pres-ence of C3d, per se, does not shorten redcell survival.

In rare cases the cold-reactive autoan-tibody is IgG, which may or may not bindcomplement. The presence of this IgG maybe missed by the DAT unless all phases oftesting are done in the cold, using coldsaline and antiglobulin reagent.

SerumIgM cold-reactive autoagglutinins asso-ciated with immune hemolysis usuallyhave a titer above 1000 when tested at 4C; in vitro, they rarely react with saline-suspended red cells above 32 C. If 30%bovine albumin is included in the reac-tion medium, 70% of clinically signifi-cant examples will react at 37 C.30

Hemolytic activity against untreated redcells can sometimes be demonstrated at20-25 C and, except in rare cases with Prspecificity, enzyme-treated red cells arehemolyzed in the presence of adequatecomplement.

Determination of the true thermalamplitude or titer of the cold autoagglu-tinin requires that the specimen be col-


lected and maintained strictly at 37 Cuntil the serum and cells are separated,to avoid in-vitro autoadsorption. Alter-natively, plasma can be used from anEDTA-anticoagulated specimen that hasbeen warmed for at least 15 minutes at37 C (with repeated mixing) and thenseparated from the cells, ideally at 37 C.This should release autoadsorbed anti-body back into the plasma.23

In chronic CAS the IgM autoagglu-tinin is usually a monoclonal proteinwith kappa light chains. In the acuteform, the autoreactive antibody is poly-clonal IgM, with normal kappa andlambda light-chain distribution. Rareexamples of IgA and IgG cold-reactiveautoagglutinins have also been de-scribed.

EluateElution is seldom necessary in obviouscases of CAS. If the red cells have beencollected properly and washed at 37 C,there will be no immunoglobulin on thecells and no reactivity will be found inthe eluate. Because the serum antibodylevel is often very high, it is essential totest the last saline wash before elution toensure that unbound autoagglutininshave been removed.

Specificity of AutoantibodyThe autoantibody specificity in CAS isusually of only academic interest. CAS ismost often associated with antibodieswith I specificity; less commonly, i speci-ficity is found, usually associated withinfectious mononucleosis. On rare occa-sions, cold-reactive autoagglutininswith Pr or other specificities, are seen(see Method 6.3). Dilution of the serummay be necessary to demonstrate speci-ficity of very high-titer antibodies.

Autoantibody specificity is not diag-nostic for CAS. Autoanti-I may be seenin healthy subjects as well as patients



with CAS. The nonpathologic forms ofautoanti-I, however, rarely react to titersabove 64 at 4 C, and are usually nonre-active with I-negative (icord and iadult) redcells at room temperature. In contrast,the autoanti-I of CAS may react quitestrongly with I-negative red cells in testsat room temperature, while equal oreven stronger reactions are observedwith I-positive red cells. Auto-anti-i an-tibodies react in the opposite manner;they give much stronger reactions withI-negative red cells than with red cellsthat are I-positive. The cold-reactiveautoantibody found in mixed-type AIHAmay not show specificity, may react onlyweakly at 4 C, but usually reacts at up to30 C.21,23 Procedures to determine thespecificities of cold-reactive autoanti-bodies are given in Methods 6.2 and 6.3.

Transfusion in CAS

Patients suffering from CAS can often bemanaged without transfusion. Acute celldestruction may be reduced by keepingthe patient in a warm room (about 40 C),to prevent complement binding by thecold agglutinin. Steroid therapy is usu-ally not successful; however, if the anti-body is low-titered but reacts at a highthermal range, steroids may be helpful.15

Plasma exchange has been helpful insome cases, but the effects are usuallynot lasting. Transient hemolysis associ-ated with infectious diseases requirestreatment of the infectious disease.

If transfusion is undertaken, up to50% of the cells transfused initially maybe destroyed, until cells acquire someresistance to complement destructionby binding of C3d.7 If transfusion is re-quired, use of a blood warmer is oftenrecommended, but the need for this iscontroversial.13

Pretransfusion Testing. Antibody de-tection tests should be performed inways that minimize co ld -react iveautoantibody activity yet still permit de-


tection of clinically significant alloanti-bodies. The use of albumin and otherpotentiators increases the reactivity ofthe autoant ibodies and should beavoided. To avoid the detection of boundcomplement, some workers use an IgG-specific reagent, rather than a polyspeci-fic antiglobulin serum. Some workersperform compatibility tests strictly at 37C, but this may miss some potentiallysignificant alloantibodies.

Adsorption Procedures. When cold-reactive autoantibody reactivity persists,autoadsorption studies (see Method 6.1)can be performed if the patient has notbeen transfused within the past 3months. This should make it possible todetect alloantibodies masked by thecold-reactive autoantibody. If the pa-tient has been recently transfused, rab-bit red cells may be used to removeautoanti-I and -IH from sera,31 but thismay remove clinically significant alloan-tibodies, notably anti-B, -D, -E, and oth-ers.32 A preparation of rabbit red cellstroma is commercially available. Alter-natively, allogeneic adsorption studiescan be performed as for WAIHA (see be-low).

Mixed-Type AIHAAlthough some patients with WAIHAmay also have IgM antibodies that reactto high titer at low temperature, a sepa-rate mixed-type AIHA accounts for ap-proximately 7-8% of all AIHAs.18,20,21

These patients have “cold” agglutininsthat have low titers at 4 C but high ther-mal amplitudes, reacting at 30 C orabove. Mixed-type AIHA often presentsas an extremely acute condition charac-terized by very low hemoglobin concen-trations and complex serum reactivitypresent in all phases of testing. Transfu-sion should be carefully considered, es-pecially since prompt corticosteroidtherapy is frequently successful. How-ever, even with steroid therapy, this type



of AIHA may be persistent, with inter-mittent periods of hemolysis.

Mixed-type AIHA can be idiopathic orsecondary, often associated with sys-temic lupus erythematosus. Typical se-rologic findings are described below.

DATBoth IgG and C3d are usually detectableon patient’s red cells. Presumably, theIgG is due to the warm autoantibody,while the C3d is bound by the effects ofIgM autoantibody.

SerumBoth warm-reactive IgG autoantibodiesand cold-reactive, hemagglutinatingIgM autoantibodies are present in theserum. These usually result in reactivityat all phases of testing, with virtually allcells tested. The IgM hemagglutinatingautoantibody(ies), unlike those in CAS,usually have titers less than 64 at 4 C butreact at 30 C or above.21 If the presenceof cold hemagglutinin leads to misclas-sification as CAS, the option of cortico-steroid therapy may be dismissed, butsteroids are often effective in mixed-typeAIHA. Mixed-type AIHA should also bedistinguished from WAIHA in a patientwith a strongly reactive, but normal,cold agglutinin. Care should be taken toavoid loss of antibody reactivity byautoadsorption of the cold agglutinincomponent.

Albumin and other potentiators of ag-glutination increase the reactivity of thecold autoagglutinin; anti-IgG, ratherthan a polyspecific antihuman globulinreagent, should be used for the IAT. Ifadsorption studies are done to detect al-loantibodies, it may be necessary to per-form adsorptions at both warm and coldtemperatures. The cold (IgM) agglutininmay also be circumvented by treatingthe serum with sulfhydryl reagents, leav-ing intact the warm-reactive (IgG) auto-


agglutinin and most clinically signifi-cant alloantibodies.

EluateA suitably prepared eluate will contain awarm-reactive IgG autoantibody.

Specificity of AutoantibodiesThe unusual cold-reactive IgM hemag-glutinating autoantibody can have specifi-cities typical of CAS (ie, I or i) but oftenhas no apparent specificity.18,20,21 Thewarm-reactive IgG autoantibody oftenappears serologically indistinguishablefrom specificities encountered in typicalWAIHA, but may have atypical features.33

Transfusion in Mixed-Type AIHAPatients with mixed-type AIHA oftenpresent with severe anemia and markedintravascular hemolysis. Transfusion insuch cases may result in increasedhemolysis, which may be life-threaten-ing. Fortunately, these patients usuallyshow a dramatic initial response to cor-ticosteroid therapy and often do notneed to be transfused.20 Transfusion maybe less problematic if hemolysis is lessacute.

If blood transfusions are necessary,the considerations in the selection ofblood for transfusion are identical tothose described for patients with acutehemolysis due to warm antibody typeAIHA (see above). Considerations aboutblood warmers are the same as for CAS.

Paroxysmal Cold HemoglobinuriaThe rarest form of DAT-positive AIHA isPCH. In the past, it was character-istically associated with syphilis, butthis association is now unusual. Morecommonly, PCH presents as an acutetransient condition secondary to viralinfections, particularly in young chil-dren. It can also occur as an idiopathic



chronic disease in older people. Onelarge study found that none of 531 adultshaving well-defined immune hemolyticanemias had PCH, while 22 of 68 (32%)children were shown to have PCH.15

DATThe autoantibody in PCH is IgG, but aswith IgM cold-reactive autoagglutinins,it reacts with red cells in colder areas ofthe body (usually the extremities),causes C3 and C4 to bind irreversibly tored cells, and then dissociates from thered cells at warmer temperatures. Redcells washed in a routine manner for theDAT are coated only with complementcomponents, but IgG may be detectableon cells that have been washed with coldsaline and tested with cold anti-IgG re-agent. Keeping the system nearer its op-timal binding temperature allows thecold-reactive IgG autoantibody to re-main attached to its antigen.

SerumThe IgG autoantibody in PCH is classi-cally described as a biphasic hemolysin,since binding to red cells occurs at lowtemperatures but hemolysis does not oc-cur until the coated red cells are warmedto 37 C. This is the basis of the diagnostictest for the disease, the Donath-Land-steiner test (see Method 6.7). Theautoantibody may agglutinate normalred cells at 4 C, but rarely to titersgreater than 64. Because the antibodyrarely reacts above 4 C, the serum isusually compatible with random donorcells by routine crossmatch proceduresand pretransfusion antibody detectiontests are usually nonreactive.

EluateBecause complement components areusually the only globulins present oncirculating red cells, eluates prepared


from red cells of patients with PCH arealmost invariably nonreactive.

Specificity of AutoantibodyThe autoantibody of PCH has most fre-quently been shown to have P specificity,reacting with all red cells (including thepatient’s own red cells) except those ofthe very rare p or Pk phenotypes. Excep-tional examples with other specificitieshave been described.7

Transfusion in PCHTransfusion is rarely necessary for adultpatients with PCH, unless hemolysis issevere. In children, especially under age6, the thermal amplitude of the antibodytends to be much wider than in adultsand hemolysis more brisk, and transfu-sion may be required as a lifesavingmeasure. Normal bank blood can usuallybe used. While there is some evidencethat p red cells survive better thanP-positive (P1-positive or P1-negative)red cells,7 the prevalence of p blood isapproximately 1 in 200,000 and the ur-gent need for transfusion usually pre-cludes attempts to obtain this rareblood. Transfusion of random donorblood should not be withheld from PCHpatients whose need is urgent. Red cellsnegative for the P antigen should be con-sidered only for those patients who donot respond adequately to random donorblood.

DAT-Negative AIHAClinical evidence of hemolytic anemia ispresent in some patients whose DAT isnonreactive. Frequently, autoantibodycannot be detected in either eluate orserum. There may be several reasonswhy the DAT is negative. Antibodies withlow binding affinity may dissociate fromthe red cells during saline washing of thecells for the DAT. Washing with low ionicstrength saline or saline at 37 C or 4 C



may help retain antibody on the cells.There may be too few antibody mole-cules on the cell for detection by routinemethods, but may be demonstrable bymethods such as flow cytometry, en-zyme-linked antiglobulin tests, solidphase, or direct Polybrene.®

Nonroutine Reagents

The causative antibody may be IgM orIgA not detected by routine antiglobulinreagents. Anti-IgG, anti-C3d, and thecombined anti-C3b,-C3d reagents arethe only licensed products available foruse with human red cells. Antiglobulinreagents that react with IgA, IgM, or C4are available commercially but havebeen prepared for use with endpointsother than agglutination. These must beused cautiously and their hemaggluti-nating reactivity carefully standardizedby the user. Quality control must be rig-orous. Because agglutination with an-tiglobulin reagents is more sensitivethan precipitation, a serum that appearsto be monospecific by precipitation testsmay react with several different proteinswhen used in agglutination tests.

Antigen Dispersion

The patient with autoantibodies withspecificity in the Kell system may havedepressed red cell expression of the Kellantigens. When this occurs, antibodymay be detected in the serum and eluate,but the DAT may be negative or veryweakly positive. This may provide in-vivo protection of autologous cells. Do-nor cells of common Kell type may bedestroyed, but cells lacking the corre-sponding antigen (usually high-inci-dence) may survive well. When theautoantibody subsides, autologous cellsagain express normal amounts of anti-gen.7


Serologic Problems withAutoantibodiesIn pretransfusion tests on patients withautoantibodies, the following problemsmay arise:

1. Cold-reactive autoantibodies cancause autoagglutination, resultingin erroneous determinations of ABOand Rh type.

2. Red cel ls strongly coated withglobulins may undergo spontaneousagglutination with high-proteinanti-Rh blood grouping reagents,and occasionally even with low-pro-tein reagents.34

3. The presence of free autoantibody inthe serum may make antibody detec-tion and crossmatching tests diffi-cult to interpret. If time permits, thepresence or absence of alloantibodyshould be determined (see Methods6.4-6.6) before blood is transfused.

Although resolving these serologicproblems is important, delaying transfu-sion in the hope of finding serologicallycompatible blood may, in some cases,cause greater danger to the patient. Onlyclinical judgment can resolve this di-lemma. Dialogue with the patient’s phy-sician is important.

Resolution of ABO ProblemsThere are several approaches to the reso-lution of ABO typing problems associ-ated with cold-reactive autoagglutinins.Often, it is only necessary to maintainthe blood sample at 37 C immediatelyafter collection and to wash the red cellswith warm (37-45 C) saline before test-ing. It is helpful to perform a parallelcontrol test, using 6% bovine albumin insaline, to determine if autoagglutinationpersists. If the control test is nonreac-tive, the results obtained with anti-A andanti-B are usually valid. If autoaggluti-nation still occurs, interpretation of the



results can be difficult, but comparingthe strength of the observed reactionsmay be informative. If the blood samplehas been kept at room temperature, or ifcold-reactive autoagglutinins are par-ticularly potent, it may be necessary totreat the red cells with sulfhydryl re-agents.

Dispersing Autoagglutination

Because cold-reactive autoagglutininsare almost always IgM, and sulfhydrylreagents denature IgM molecules, re-agents such as 2-mercaptoethanol (2-ME) or dithiothreitol (DTT) can be usedto abo l i sh autoagglut inat ion (seeMethod 2.11). Other means of dispersingautoagglutination use the ZZAP re-agent35 or glycine-HCl/EDTA36 ( seeMethods 2.14 and 6.4). Some reagentsera should not be used with chemicallymodified red cells, eg, monoclonal ABOgrouping reagents with enzyme-treatedred cells. Appropriate controls are essen-tial for all tests.

An alternative but less effectivemeans of dispersing autoagglutinationdue to cold-reactive autoantibodies is toincubate the red cells at 45 C for 10minutes and wash them several times insaline at that temperature.

Problems with ABO Serum Tests

When the serum agglutinates group Oreagent red cells, the results of serumtests may be unreliable. Repeating thetests using prewarmed serum and A, B,and O red cells at 37 C will often resolveany discrepancy, but anti-A and/or -B insome patients’ sera do not react at 37 C.Alternatively, adsorbed serum (eitherautoadsorbed or adsorbed with allo-geneic group O red cells) can be used.Because rabbit red cells express a B-likeantigen, sera adsorbed with rabbit redcells or stroma may not contain anti-B,


and sera adsorbed in this manner shouldnot be used for ABO serum tests.

Resolution of Rh ProblemsSpontaneous agglutination of red cellsby cold- or warm-reactive autoanti-bodies may also cause discrepant Rh typ-ing. The same procedures described forthe resolution of ABO problems may beuseful, but most Rh typing discrepanciesoccur when IgG-coated red cells un-dergo spontaneous aggregation in thehigh-protein reagents designated for useby slide or rapid tube tests (sometimesreferred to as modified-tube anti-Rh).The use of low-protein reagents, such asmonoclonal/polyclonal blends, may benecessary to avoid erroneous results (seeChapter 13). IgG antibody can be disso-ciated from the cells by treatment withchloroquine diphosphate (Method 2.13),by glycine-HCl/EDTA (Method 2.14), orby other elution methods that leave redcells intact for subsequent typing. IgM-coated cells can be treated with sulfhy-dryl reagents (such as 2-ME or DTT,Method 2.11) to circumvent spontane-ous agglutination.

Detection of Alloantibodies in thePresence of Warm-ReactiveAutoantibodiesIf the patient who has warm-reactiveautoantibodies in the serum needs trans-fusion, it is important to evaluate thepossible simultaneous presence of al-loantibodies to red cell antigens. Somealloantibodies may make their presenceknown by reacting more strongly or atdifferent phases than the autoantibody,but initial studies may not suggest theexistence of masked alloantibodies. It ishelpful to know which of the commonred cell antigens are lacking on the pa-tient’s red cells, to predict which clini-cally significant alloantibodies the pa-tient might produce. Antigens absent



from autologous cells could well be thetarget of present or future alloanti-bodies. When the red cells are coatedwith IgG, antiglobulin-reactive re-agents, such as most anti-Fya or anti-Jka,cannot be used to test IgG-coated cellsunless the IgG is first removed (seeMethod 2.14). Monoclonal reagents orother “saline-reactive” antisera that donot require an antiglobulin test may behelpful in typing the DAT-positive redcells. Cell separation procedures (seeMethods 2.15 and 2.16) may be necessaryif the patient has been transfused re-cently.

Methods to detect alloantibodies inthe presence of warm-reactive autoanti-bodies attempt to remove, reduce, or cir-cumvent the autoantibody. Methods thatuse PEG, LISS, or enzymes generallyenhance autoantibodies; avoiding thesereagents may allow detection of mostsignificant alloantibodies. Other proce-dures involve adsorption, the principlesof which are discussed in Chapter 17.Three widely used approaches are dis-cussed below.

Autologous Adsorption

In a patient who has not been recentlytransfused, autologous adsorption (seeMethod 6.4) is the best way to detectalloantibodies in the presence of warm-reactive autoantibodies. The adsorbedserum can be used in the routine anti-body detection procedure.

Autoadsorption generally requiressome initial preparation of the patient’sred cells. At 37 C, in-vivo adsorption willhave occurred and all antigen sites onthe patient’s own red cells may beblocked. It may be necessary, therefore,to remove autoantibody from the redcells to make the sites available. Treat-ment of the autologous red cells withproteolytic enzymes also increases theircapacity to adsorb autoantibody. ZZAP, amixture of papain and dithiothreitol,35 is


sometimes used to treat the cells (seeMethod 6.5); the sulfhydryl componentmakes the IgG molecules more suscepti-ble to the protease and dissociates theantibody molecule from the cell. Multi-ple, sequential autoadsorptions may benecessary if the serum contains high lev-els of autoantibody. Once autoantibodyhas been removed, the adsorbed serumis examined for alloantibody activity.

Recent Transfusion. Autologous ad-sorption is most applicable to patientswho have not been recently transfused,so will not have an admixture of trans-fused red cells that might adsorb alloan-tibody. Recent, in this context, usuallymeans within the normal life span of redcells, or 3-4 months. Autologous adsorp-tion studies may be informative even af-ter transfusion as recent as one week,especially if there is evidence of mark-edly shortened red cell survival. Failureto detect alloantibody in “autoadsorbed”serum from a recently transfused patientshould never be considered conclusiveproof that no alloantibody exists.

Sources of Autologous Cells. The pa-tient’s own red cells, a portion of whichwill be younger and less dense thantransfused red cells, can sometimes beseparated by simple (Method 2.15) orgradient centri fugation for use inautoadsorption experiments. However,these techniques are highly variable andmay require expensive equipment and/orconsiderable expertise in interpretation.

If the patient is to be transfused, itcan be advantageous to collect and saveadditional aliquots of pretransfusioncells, to be used for later adsorptions.

Allogeneic Adsorption

The use of allogeneic red cells for ad-sorption may be helpful when the pa-tient has been recently transfused orwhen insufficient autologous red cellsare available. The goal is to removeautoantibody and leave the alloantibody



in the adsorbed serum. The adsorbingcells must be negative for the antigensagainst which the alloantibodies react.Because alloantibody specificity is un-known, cells of several different pheno-types will usually be used to adsorb sev-eral aliquots.

Given the number of potential alloan-tibodies, the task of selecting cells mayappear formidable. In actuality, however,the selected cells need only demonstratethose few alloantibodies of clinical sig-nificance likely to be present. These in-clude the common Rh antigens (D, C, E,c, and e), K, Fya and Fyb, Jka and Jkb, andS and s. Cell selection is made easier bythe fact that cells can be rendered nega-tive for some of these antigens by appro-priate treatment prior to adsorption. An-tibodies to other antigens are rare,detect low-incidence antigens, or havelittle clinical significance even if pre-sent. Antibodies to high-incidence anti-gens cannot be excluded by allogeneicadsorptions, because the adsorbing cellswill almost inevitably express the anti-gen and adsorb the alloantibody alongwith autoantibody.

Patient’s Phenotype Unknown. Whenthe patient’s Rh phenotype is not known,group O red cell samples of three differ-ent Rh phenotypes (R1, R2, and r) shouldbe selected. One should lack Jka and an-other Jkb. If treated with ZZAP, thesecells would lack all antigens of the Kellsystem and M, N, S, s, Fya, and Fyb.

Each aliquot may need to be adsorbedtwo or three times. The fully adsorbedaliquots are tested against reagent redcells known either to lack or to carrycommon antigens of the Rh, MNS, Kidd,Kell, and Duffy blood group systems. Ifan adsorbed aliquot is reactive, that ali-quot (or an additional specimen simi-larly adsorbed) should be tested to iden-tify the antibody. Adsorbing severalaliquots with different red cell samplesprovides a battery of potentially informa-tive specimens. For example, if the ali-


quot adsorbed with Jk(a–) red cells sub-sequently reacts only with Jk(a+) redcells, the presence of alloanti-Jka canconfidently be inferred.

The adsorbed serum can also be usedfor crossmatching. If the quantity of ad-sorbed serum is limited, crossmatchingmay be a more advantageous use thanantibody identification; if necessary, ad-sorbed aliquots may be pooled for com-patibility tests.

If ZZAP is not available, cells treatedonly with proteolytic enzyme can beused, but they must be K-negative be-cause Kell system antigens will not bedestroyed. Untreated cells may be used,but antibody will be more difficult toremove and the adsorbing cells must, ata minimum, include at least one nega-tive for the S, s, Fya, Fyb, and K antigens,in addition to the Rh and Kidd require-ments above.

Patient’s Phenotype Known. Thenumber of different red cell phenotypesrequired may be reduced if the patient’sphenotype is known, or can be deter-mined by separation of autologous fromtransfused red cells (see Method 2.15). Ifthe patient’s Rh and Kidd phenotypes areknown or can be determined, adsorptioncan be performed with a single sample ofallogeneic ZZAP-treated red cells of thesame Rh and Kidd phenotypes as thepatient.

Problems Encountered. Occasionallyautoantibody will not be removed bythree sequential adsorptions. Furtheradsorptions can be done, but multipleadsorptions are likely to dilute the se-rum. If the adsorbing cells do not appearto remove the antibody, the autoanti-body may have an unusual specificitythat does not react with the cells used foradsorption. For example, autoantibodieswith Kell specificity would not be re-moved by ZZAP-treated cells. Some anti-Ena, -Wrb, or other antibodies may notreact with cells treated with ZZAP orproteolytic enzymes.



Autoantibodies MimickingAlloantibodiesAutoantibodies sometimes have patternsof reactivity that are easily mistaken foralloantibody. For example, the serum of aD-negative patient may have apparentanti-C and -e reactivity. The anti-C reactiv-ity may reflect warm-reactive autoanti-body, even if the patient’s cells lack C. Theautoantibody nature of the reactivity canbe demonstrated by autologous and allo-geneic adsorption studies. The apparentalloanti-C would, in this case, be adsorbedby C-negative red cells, both autologousand allogeneic. This is quite unlike thebehavior of a true alloanti-C, which wouldbe adsorbed only by C-positive red cells.

Detection of Alloantibodies in thePresence of Cold-ReactiveAutoantibodiesCold-reactive autoagglutinins rarelymask clinically significant alloanti-bodies if serum tests are conducted at37 C and if IgG-specific reagents areused for the antiglobulin phase. In rareinstances it may be necessary to performautoadsorption at 4 C (see Method 6.1).Achieving complete removal of potentcold-reactive autoagglutinins is verytime-consuming, and may be facilitatedby treating the patient’s cells with ZZAPor enzymes before adsorption.

If autologous adsorption studies are in-appropriate because the patient has beenrecently transfused, or if complete re-moval of the autoantibody is not possible,tests for alloantibody activity may be car-ried out strictly at 37 C. Both the patient’sserum and the reagent red cells should bewarmed to 37 C before mixing. The testsshould be centrifuged at 37 C. If centrifu-gation at 37 C is impossible, and if timepermits, tests can be incubated at 37 C for1-2 hours to allow the red cells to settle,and they can then be examined for agglu-tination without centrifugation. Because


bovine albumin and other enhancementmedia increase the reactivity of cold-re-active autoantibodies, their use shouldbe avoided. The red cells should bewashed in saline at 37 C and then testedwith anti-IgG to avoid the effects of in-vitro complement binding by the cold-reactive autoantibody. EDTA-anticoagu-lated plasma, or serum to which EDTAhas been added (2 mL serum plus 0.25mL 4.45% K2EDTA),37 can also be used toavoid interference by complement-bind-ing autoantibodies.

Drug-Induced ImmuneHemolytic AnemiasDrugs sometimes induce the formation ofantibodies, either against the drug itself oragainst intrinsic red cell antigens thatmay result in a positive DAT, immune redcell destruction, or both. Some of the an-tibodies produced appear to be dependenton the presence of the drug for their de-tection or destructive capability while oth-ers do not. In some instances, a reactiveDAT may result from nonimmunologic ef-fects of the drugs.

The patient’s medications may in-clude over-the-counter medications andothers not prescribed by a physician. En-vironmental chemicals can cause simi-lar serologic and clinical effects. Theseinclude chlor inated hydrocarbons(found in insecticides) and chemicalsused in dyes and manufacturing proc-esses. The bite of the brown recluse spi-der may be associated with severe in-travascular hemolysis, generally with anegative DAT, although examples withpositive DATs have been reported.38

Theories of the Immune Responseand Drug-Dependent AntibodiesNumerous theories have been suggestedto explain how drugs induce immune



responses and what relation such re-sponses may have to the positive DAT andimmune-mediated cell destruction ob-served in some patients. For many years,drug-associated positive DATs were classi-fied into four mechanisms: drug adsorp-tion, immune complex formation, non-specific adsorption, and autoantibodyproduction. Such classification has beenuseful, but many aspects lacked definitiveproof or detailed mechanisms. These theo-ries were further complicated by evidencethat some drugs created immune prob-lems involving aspects of more than onemechanism. More recent theories, stillunproven, tend toward a more compre-hensive approach applicable to all drugs,allowing for differing manifestations.

Most drugs have a molecular weightsubstantially below the 5-kDa levelusually considered the threshold foreffective immunogenicity. Drugs mayact as haptens, eliciting antibody onlyafter they bind firmly to or become partof a protein carrier. Most drugs associ-ated with an immune response reactvery weakly with cellular proteins, butdrug-induced immune responses oftenshow extraordinary specificity to par-ticular cellular targets, such as redcells or platelets. This exquisite speci-ficity renders less plausible the classi-cal explanation of induction by thehapten carrier.

Current theories39-43 propose thatdrugs elicit antibodies based on their abil-ity to interact with specific cell membranecomponents, thus altering the normalcomponents so that they are no longerrecognized as self. A new configuration or“neoantigen” is produced, consisting ofboth drug and cell component, and thehost’s immune system perceives thisneoantigen as foreign. The antibody isproduced directly against the neoantigen,not against a haptenic constituent.

This theory assumes that drugs bindwith variable affinity to particular mem-brane components present on red cells,


granulocytes, or platelets. The degree ofassociation would be expected to reflectsuch factors as the chemical structure ofthe drug or its metabolite(s), the con-centration of the drug, the structure ofcellular membrane proteins, and the ca-pacity of the host proteins to associatewith the drug (possibly dependent onpH, temperature, ionic strength and,most probably, host genetic factors). Be-cause conditions must be just right tocreate a sufficiently high affinity to pro-voke an immune response, instances ofdrug-induced immunity are relativelyrare.

The polyclonal immune response pro-voked by such a configuration consists ofantibodies that can recognize at leastthree categories of epitopes. Some reactessentially with the drug portion itself,some with combinations of the drug andcellular component, and others with es-sentially cellular membrane components(see Fig 18-1). In some cases red cells orother specific cell types are the sole orprimary targets, while at other times avariety of cell types may be affected. Whycertain drugs repeatedly target single ormultiple cell lines is not always clear.

Serologic and Clinical ClassificationDrug-induced antibodies can be classifiedinto three groups according to their clini-cal and serologic characteristics. In onegroup the drug binds firmly to the cellmembrane and antibody is apparentlylargely directed against the drug itself.This is known as the drug adsorptionmechanism, and antibodies to penicillinare the best described of this group.

The second group of drug-dependentantibodies reacts with drugs that do notbind well to the cell membrane. The re-active mechanism of these antibodieswas previously described as immunecomplex formation, but this appears notto be quite the case. Antibodies in thisgroup cause acute intravascular hemolysis,


Figure 18-1. Proposed unifying theory of drug-induced antibody reactions (based on a cartoonby Habibi as cited by Garratty24). The thicker, darker lines represent antigen-binding sites onthe F(ab) region of the drug-induced antibody. Drugs (haptens) bind loosely, or firmly, to cellmembranes and antibodies may be made to: a) the drug [producing in-vitro reactions typicalof a drug adsorption (penicillin-type) reaction]; b) membrane components, or mainly mem-brane components (producing in vitro reactions typical of autoantibody); or c) part-drug,part-membrane components (producing an in-vitro reaction typical of the so-called immunecomplex mechanism).24(p55)


but they may be difficult to demonstrateserologically.

Antibodies of the third group have se-rologic reactivity independent of thedrug, despite the fact that it was the drugthat originally induced the immune re-sponse. Serologically, they behave asautoantibodies.

Some drugs are associated with a posi-tive DAT due to nonimmune mecha-nisms. These drugs, especially first-gen-eration cephalosporins, alter the red cellmembrane in a way that induces nonspe-cific adsorption of proteins, includingbut not limited to immunoglobulins.Hemolytic anemia is not associated withthis mechanism.

Drug-Dependent Antibodies Reactive withCell-Bound Drug: Penicillin-Type AntibodiesDrugs with a high binding affinity mayinduce a seemingly drug-specific im-mune response. The high degree of asso-


ciation with the cellular component(s)may cause the drug to act as a true hap-ten. These drugs with very high bindingaffinity may be more immunogenic thandrugs that bind to cell membrane com-ponents with lower affinity. Because theantibodies produced are directed pre-dominantly to the drug (hapten) itself,they can be expected to be inhibitable bya pure form of the drug (ie, hapten inhi-bition), and may require that the drug bebound to a solid matrix detection system(ie, red cells) for in-vitro detection.

Such drugs have previously been classi-fied as reacting by the drug adsorptionmechanism, with penicillin antibodies theprimary example (see Fig 18-2). Hemolyticexamples of anti-penicillin are not alwayscompletely inhibited with pure penicillin.This may simply reflect the need of thesevery high titers of antibody for a very highhapten concentration to cause complete in-hibition. Alternatively, this may indicatethat at least a portion of the polyclonal anti-



penicillin response recognizes someshared component of the membrane-drug neoantigen. If this occurs, the in-vitro serologic reactions resemble boththe drug adsorption and “immune com-plex” mechanisms.

General Observations. The clinicaland laboratory features of drug-inducedimmune hemolytic anemia operatingthrough this mechanism are:

1. The DAT is strongly positive due toIgG coating. Rarely, complement

Figure 18-2. The drug-adsorption mechanism. Tproteins. If a patient develops a potent anti-drdrug. Such rbcs will yield a positive result in thusually not activated and lysis is primarilyprototype drug. (Reprinted with permission from


coating may also be present, butweakly.

2. Tests for unexpected serum antibod-ies are nonreactive unless the pa-tient also has alloantibodies to redcell antigens.

3. Antibody eluted from the red cellsreacts with drug-coated red cells butnot with uncoated red cells.

4. The serum contains a high-titer IgGantibody, at least when the target ispenicillin.

he drug binds tightly to the red cell membraneug antibody, it will react with the cell-bounde DAT using anti-IgG reagents. Complement isextravascular in nature. Penicillin-G is the

Petz and Branch.23)



5. Hemolysis occurs only in patientsreceiving very large intravenousdoses of drug; for penicillin, thehemolysis-inducing dose is millionsof units daily for a week or more.

6. Hemolysis develops gradually, butmay be life-threatening if the etiol-ogy is unrecognized and drug ad-ministration is continued.

7. Discontinuation of the drug is usuallyfollowed by increased cell survival, al-though hemolysis of decreasing sever-ity may persist for several weeks.

Penicillin Antibodies. Approximately3% of patients receiving large doses ofpenicillin intravenously (ie, millions ofunits per day) will develop a positive DAT,but less than 5% of these will develophemolyt ic anemia.7 Intravascularhemolysis is rare. A possible mechanismfor the positive DAT is given in Fig 18-2.The penicillin becomes covalently linkedto the red cells in vivo, and if the patienthas antibodies to penicillin, they bind tothe penicillin bound to the red cells. Theresult is that the penicillin-coated redcells become coated with IgG. Comple-ment is not usually involved. If cell de-struction occurs, it takes place extravas-cularly, probably in the same way that redcells coated with IgG alloantibodies aredestroyed.

Drugs such as penicillin have severaldistinct haptenic groups, of which themost important is the benzyl-penicilloyl(BPO) group. With sufficiently sensitivetechniques, most adult human sera canbe shown to contain BPO antibodies.These are usually IgM and of low titer;however, some sera contain an IgG com-ponent.7 The high prevalence of penicil-lin antibodies in the normal populationprobably reflects the widespread expo-sure to this drug. Circulating IgM andIgG antibodies to penicillin are not in-volved in allergic reactions to the drug,which are due to IgE antibodies.

Cephalosporins. Cephalosporin drugs,which are related to penicillins, may be-


have in a similar manner. The drugs bindfirmly to red cells, which then interactwith the specific cephalosporin anti-body. Anti-cephalosporins may cross-re-act with penicillin-treated red cells, butcells prepared with the specific cepha-losporin may be necessary to demon-strate reactivity.

The cephalosporins are generallyclassified by “generations,” based ontheir effectiveness against gram-nega-tive organisms (see Table 18-5). Approxi-mately 4% of patients receiving first- orsecond-generation cephalosporins de-velop a positive DAT,44 although someseries have reported a higher prevalence,possibly due to drug dosages or the dif-ferent antiglobulin reagents used. Mostpositive DAT results, especially thoseseen with first-generation cephalospor-ins, reflect nonimmunologic adsorptionof protein (see below), a mechanism notassociated with reduced red cell sur-vival. There have been occasional re-ports of red cell destruction resultingfrom therapy with first- and second-gen-eration cephalosporins, probably result-ing from effects of specific anti-cepha-lospor in ant ibod ies . Even moredramatically reduced red cell survivalhas been associated with newer second-and third-generation cephalosporin, andin some cases a drug-independentautoantibody mechanism may also be in-volved.45,46 The prevalence and severityof cephalosporin-induced immune redcell destruction appears to be increas-ing.

Other Drug-Dependent Antibodies:“Immune Complex” MechanismThis mechanism is the least frequentfinding in drug-induced immune-medi-ated red cell destruction. Quinidine andphenacetin are prototype drugs. The fol-lowing observations are characteristic.

1. Acute intravascular hemolysis withhemoglobinemia and hemoglo-


Table 18-5. Some CephalosporinsGeneric Name Trade Name*

First GenerationCefadroxil DuricefCefazolin Ancef, KefzolCephalexin KeflexCepalothin KeflinCephapirin CefadylCephradine Anspor

Second GenerationCefaclor CeclorCefamandole MandolCefmetazole ZefazoneCefonicid MonocidCefotetan CefotanCefoxitin MefoxinCefprozil CefzilCefuroxime Zinacef, KefuroxCefuroxime axetil Ceftin

IntermediateCefixime Suprax

Third GenerationCefoperazone CefobidCefotaxime ClaforanCeftazidime Fortaz, Ceptaz,

Pentacef, Tazicef,Tazidime

Ceftizoxime CefizoxCeftriaxone RocephinMoxalactam Moxam*Several forms are also marketed underother trade names. This list is intended tobe informational, not inclusive.


binuria is the usual presentation.Renal failure occurs in approxi-mately 50% of cases.

2. Once antibody has been formed, se-vere hemolytic episodes may recurafter exposure to very small quanti-ties of the drug.


3. The antibody can be either IgM or IgG.4. Complement is usually the only

globulin found on the red cells.5. Drug must be present in vitro for

demonstration of the antibody inthe patient’s serum.

Many drugs have been implicated ascausing immune hemolysis, yet most arerepresented by only one or a few exam-ples. The immune hemolysis was pre-viously attributed to the formation ofdrug/anti-drug immune complexes thatinteracted loosely and nonspecificallywith the cell, resulting in the binding ofcomplement and lysis of the cell.

The neoantigen concept proposesthat these drugs do, in fact, associatewith the cell membrane, though notwith high affinity. The neoantigen, com-posed partly of drug, partly of membranecomponents, constitutes the stimulus toantibody formation and the target of an-tibodies once formed (see Fig 18-1).

The distinction between these “im-mune complex” drug-dependent reac-tions and those of the “drug-adsorption”mechanism may be more apparent thanreal.47 Classification depends in largepart on the ability to bind the implicateddrug to red cells in vitro for testing.

Drug-Independent Antibodies:Autoantibody ProductionSome drugs induce autoantibodies thatappear serologically indistinguishablefrom those of WAIHA: red cells arecoated with IgG, and the eluate as wellas the serum reacts with virtually allcells tested, in the absence of the drug.At times blood group specificity has beendemonstrated, similar to that seen inAIHA. The antibody has no in-vitro activ-ity with the drug, directly, or indirectly.

The best studied of such cases arethose induced by α-methyldopa. Aclosely related drug, L-dopa, has alsobeen implicated, as have several drugsunrelated to α-methyldopa, including



procainamide and nonsteroidal anti-in-flammatory drugs (eg, mefenamic acidand sulindac). (See Table 18-6.) In somecases drug-dependent antibodies arealso present.

Proof that a drug causes autoantibodyproduction is difficult to obtain. Suffi-cient evidence would include: demon-stration that autoantibody productionbegan after drug administration; resolu-tion of the immune process after with-drawal of the drug; and recurrence ofhemolytic anemia or autoantibodies ifthe drug is readministered. The last re-quirement is crucial and the most diffi-cult to demonstrate.

Possible Mechanisms. Previous theo-ries to explain autoantibody formationinclude:

1. The drug alters an intrinsic red cellantigen so that it is no longer recog-nized as self by the immune system.

2. The drug interferes with suppressorT-cell function, allowing overpro-duction of autoantibody by B cells.48

Studies attempting to confirm thistheory have given mixed results.43,49

3. The hemolyt ic autoantibody isqualitatively different, or recognizesa slightly different epitope, from thenonhemolytic autoantibody foundin the majority of DAT-positive pa-tients receiving α-methyldopa.

Table 18-6. Some Drugs That May Indu(Autoantibodies)

α-Methyldopa Cianidanol*Azapropazone* ChlorpromazinCarbimazole* CyclofenilCatergen Diclofenac*Cefotetan* FenfluramineCefoxitin* Glafenine*Chaparral IbuprofenChlorinated Hydrocarbons* Latamoxef**These drugs may also induce drug-dependent a


Current theories suggest that thedrugs bind to the cell membrane and/orsubtly alter membrane structures, form-ing a neoantigen. The autoantibodiesproduced are those directed againstmembrane components of the neoan-tigen, which are crossreactive with thedrug and do not require it for sub-sequent reactivity. Blood group specific-ity may occur if antigens present on thecell are involved with the neoantigen.45

With some drugs, drug-dependent anti-bodies have been shown in addition todrug-independent autoantibodies. Suchdrug-dependent antibodies have notbeen demonstrated with α-methyldopa,procainamide, or mefenamic acid, sug-gesting that possibly at least two differ-ent mechanisms may be involved.43

α-Methyldopa. The prototype drug re-sponsible for this mechanism, α-methyldopa, is now rarely prescribed, asmore effective antihypertensive drugswith fewer side effects have becomeavailable. The clinical and laboratoryfeatures associated with α-methyldopa-induced autoantibodies are as fol-lows47,50:

1. Approximately 15% of patients re-ceiving α-methyldopa develop apositive DAT. Only 0.5-1.0% of pa-tients taking α-methyldopa develophemolytic anemia.

ce Drug-Independent Antibodies

Levodopae Mefenamic Acid

Nomifensine*Phenacetin*ProcainamideStreptomycin*Teniposide*Tolmetin*

ntibodies.



2. The DAT becomes positive only after3-6 months of α-methyldopa ther-apy.

3. Red cells are usually coated onlywith IgG; weak complement coatinghas been reported occasionally.

4. Development of a positive DAT isdose dependent; approximately 36%of patients taking 3 g of the drugdaily develop a positive DAT, com-pared with 11% of patients receiving1 g per day.

5. Antibodies in the serum and on thered cells are indistinguishable fromthose found in WAIHA.

6. The strength of the positive DAT be-comes progressively weaker once α-methyldopa therapy is discontinued.It may take from 1 month to 2 yearsfor the DAT to revert, but in patientswith clinical hemolysis , hema-tologic values usual ly improvewithin a week or so after the drugtherapy is discontinued.

Nonimmunologic Protein AdsorptionThe positive DAT associated with somedrugs is due to a mechanism independentof antibody production. Hemolytic anemiaassociated with this mechanism occurs ex-tremely rarely, if at all.

Cephalosporins (primarily cepha-lothin) are the drugs with which this wasoriginally associated. Red cells coatedwith cephalothin (Keflin) and incubatedwith normal plasma will adsorb albu-min, IgA, IgG, IgM, and β and α (ie,complement) globulins in a nonimmu-nologic manner. If this occurs, a positiveDAT will be seen with antihuman re-agents to many serum proteins, not justanti-IgG or anti-C3.

Although it was once thought thatthese drugs altered the red cell mem-brane in some way, making the cells ad-sorb proteins nonspecifically, a differentmechanism has been proposed.51 Cepha-lothin can bind firmly, in vitro, to red


cells at acid pH by a mechanism unlikethat of other β-lactam antibiotics. Thisalternative binding method leaves ex-posed a β-lactam moiety of the molecule,to which several proteins can then be-come bound.

In addition to this nonimmune ad-sorption of proteins, the cephalosporinscan also induce a positive DAT by thedrug adsorption mechanism describedfor penicillin by the so-called immunecomplex mechanism, and even by theproduction of drug-independent autoan-tibodies (see above). Newer generationsof cephalosporins (see Table 18-5) havebeen associated with severe immunehemolysis.47

Other drugs that may cause nonim-munologic adsorption of proteins and apositive DAT include diglycoaldehyde,suramin, and cisplatin.

Laboratory Investigation ofDrug-Induced AntibodiesThe drug-related problems most com-monly encountered in the blood bankare those associated with a positive DAT.IgG-and complement-speci f ic an-tiglobulin reagents are useful in class-ifying cases of drug-related hemolysis.Typical DAT results are shown in Table18-3. Drug-dependent antibodies char-acteristically bind only complement tothe cells. Drugs, such as penicillin, thatbind firmly to the cells and induce IgGantibodies are usually associated withIgG-only DAT results. Drug-independentantibodies induce serologic phenomenasimilar to those in AIHA of other etiolo-gies. For investigation of suspectedhemolysis due to drug-dependent anti-bodies, special methods are often re-quired.

Adding Drug to the Test SystemThe patient’s serum should be tested forunexpected antibodies by routine proce-



dures. If the serum does not react withuntreated red cells, the tests should berepeated against ABO-compatible redcells in the presence of the drug(s) sus-pected of causing the problem.

If the drug is one already reported asbeing immunogenic, testing methodsmay be available in the case reports. Ifsuch information is not available, an in-itial screening test can be performedwith a solution of the drug at a concen-tration of approximately 1 mg/mL inphosphate-buffered saline at a pH opti-mal for solubility of the drug. Tech-niques are given in Methods 6.9 and6.10. For some drugs, a solvent otherthan saline may be required. The physi-cal properties of drugs, such as solubilityand stability, may be found in severaldrug reference books,52 through con-sultation with the hospital pharmacist,or by contacting the pharmaceuticalcompany. Detection of some drug-re-lated antibodies can be enhanced if theimplicated drug is added to wash solu-tions. The gel test has also been shownto be very sensitive in detecting drug-de-pendent antibodies.53

If these tests are not informative, at-tempts can be made to coat normal redcells with the drug, and the patient’s se-rum and an eluate from the patient’s redcells tested against the drug-coated redcells. This is the method of choice whenpenicillin or cephalosporins are thoughtto be implicated (see Method 6.8). Resultsdefinitive for penicillin-induced positiveDAT are reactivity of the eluate againstpenicillin-coated red cells and absence ofreaction between the eluate and uncoatedred cells. If the positive DAT is due tocomplement binding, the eluate is likelyto be nonreactive, even when the drug isadded to the test system.

Other Observations

The actual molecule that induces theimmune response may be a metabolite of


a drug, rather than the drug itself. It maybe helpful to test drug metabolites forreactivity by one of the above methods,having obtained the metabolites by col-lecting serum or urine from other indi-viduals taking the implicated drugs.

Drug-induced immune hemolysis hasbeen associated with apparent bloodgroup specificity. In some reported casesthe drug apparently bound to cells inassociation with specific antigen recep-tors (eg, Jka, e, I).47 Drugs involved haveincluded chlorpropamide, glafenine, ri-fampicin, and nomifensine. Most otherexamples of immune hemolysis associ-ated with these same drugs have notshown blood group specificity.

Drugs that have been reported tocause a positive DAT and hemolytic ane-mia are listed in Appendix 18-1.

References1. Heddle N, Kelton JG, Turchyn KL, Ali MAM.

Hypergammaglobulinemia can be associatedwith a positive direct antiglobulin test, a non-reactive eluate, and no evidence of hemolysis.Transfusion 1988;28:29-33.

2. Toy PT, Chin CA, Reid ME, Burns MA. Factorsassociated with positive direct antiglobulintests in pretransfusion patients: A case controlstudy. Vox Sang 1985;49:215-20.

3. Ramsey G. Red cell antibodies arising fromsolid organ transplants. Transfusion 1991;31:76-86.

4. Garratty G. The significance of IgG on the redcell surface. Transfus Med Rev 1987;1:47-57.

5. Freedman J. The significance of complementon the red cell surface. Transfus Med Rev 1987;1:58-70.

6. Judd WJ, Barnes BA, Steiner EA, et al. Theevaluation of a positive direct antiglobulintest (autocontrol) in pretransfusion testingrevisited. Transfusion 1986;26:220-4.

7. Mollison PL, Engelfriet CP, Contreras M.Blood transfusion in clinical medicine, 9th ed.Oxford: Blackwell Scientific Publications,1993.

8. Clark JA, Tanley PC, Wallas CH. Evaluation ofpatients with positive direct antiglobulin testsand nonreactive eluates discovered duringpretransfusion testing. Immunohematology1992;8:9-12.



9. Klein HG, ed. Standards for blood banks andtransfusion services, 17th ed. Bethesda, MD:American Association of Blood Banks, 1996.

10. Johnson MFM, Belota MK. Determination ofneed for elution studies for positive directantiglobulin tests in pretransfusion testing.Am J Clin Pathol 1988;90:58-62.

11. Kleinman S, Nelson R, Smith L, Goldfinger D.Positive direct antiglobulin tests and immunehemolytic anemia in patients receiving pro-cainamide. N Engl J Med 1984;311:809-12.

12. Shulman IA, Petz LD. Red cell compatibilitytesting: Clinical significance and laboratorymethods. In: Petz LD, Swisher SN, KleinmanS, et al, eds. Clinical practice of transfusionmedicine, 3rd ed. New York: Churchill Living-stone, 1996:199-244.

13. Judd WJ. Antibody elution from red cells. In:Bell CA, ed. A seminar on antigen-antibodyreactions revisited. Arlington, VA: AmericanAssociation of Blood Banks, 1982:175-221.

14. Sigmund KE. Direct antiglobulin testing. In:Pierce SR, Wilson JK, eds. Approaches to sero-logical problems in the hospital transfusionservice. Arlington VA: American Association ofBlood Banks, 1985:101-33.

15. Petz LD. Blood transfusion in acquiredhemolytic anemias. In: Petz LD, Swisher SN,Kleinman S, et al, eds. Clinical practice oftransfusion medicine, 3rd ed. New York: Chur-chill Livingstone, 1996:469-99.

16. Packman CH, Leddy JP. Acquired hemolyticanemia due to warm-reacting autoantibodies.In: Beutler E, Lichtman MA, Coller BS, KippsTJ, eds. Williams’ hematology, 5th ed. NewYork: McGraw-Hill, 1995:677-85.

17. Packman CH, Leddy JP. Cryopathic hemolyticsyndromes. In: Beutler E, Lichtman MA, Col-ler BS, Kipps TJ, eds. Williams’ hematology,5th ed. New York: McGraw-Hill, 1995:685-91.

18. Sokol RJ, Hewitt S, Stamps BK. Autoimmunehaemolysis: An 18 year study of 865 casesreferred to a regional transfusion centre. BrMed J 1981;282:2023-7.

19. Petz LD, Garratty G. Acquired immunehemolytic anemias. New York: Churchill Liv-ingstone, 1980.

20. Shulman IA, Branch DR, Nelson JM, et al.Autoimmune hemolytic anemia with bothcold and warm autoantibodies. JAMA 1985;253:1746-8.

21. Sokol RJ, Hewitt S, Stamps BK. Autoimmunehaemolysis. Mixed warm and cold antibodytype. Acta Haematol 1983;69:266-74.

22. Branch DR, Shulman IA, Sy Siok Hian AL, PetzLD. Two distinct categories of warm autoanti-body reactivity with age-fractionated red cells.Blood 1984;63:177-80.

23. Petz LD, Branch DR. Serological tests for thediagnosis of immune hemolytic anemias. In:McMillan R, ed. Methods in hematology: Im-

Copyright © 2002 by the AA

mune cytopenias. New York: Churchill Living-stone, 1983:9-48.

24. Garratty G. Target antigens for red cell-boundautoantibodies. In: Nance SJ, ed. Clinical andbasic science aspects of immunohematology.Arlington, VA: American Association of BloodBanks, 1991:33-72.

25. Bilgrami S, Cable R, Pisciotto P, et al. Fataldisseminated intravascular coagulation andpulmonary thrombosis following blood trans-fusion in a patient with severe autoimmunehemolytic anemia and human immunodefi-ciency virus infection. Transfusion 1994;34:248-52.

26. Telen MJ, Roberts KB, Bartlett JA. HIV-associ-ated autoimmune hemolytic anemia: Reportof a case and review of the literature. J AcquirImmune Defic Syndr 1990;3:933-7.

27. Freedman J, Wright J, Lim FC, Garvey MB.Hemolytic warm IgM autoagglutinins inautoimmune hemolytic anemia. Transfusion1986; 26:464-7.

28. Ness PM, Shirey RS, Thomas SK, Buck SA. Thedifferentiation of delayed serologic and de-layed hemolytic transfusion reactions: Inci-dence, long-term serologic findings, and clini-cal significance. Transfusion 1990;30:688-93.

29. Judd WJ. Investigation and management ofimmune hemolysis: autoantibodies and drugs.In: Wallace ME, Levitt JS, eds. Current appli-cations and interpretations of the direct an-tiglobulin test. Arlington, VA: American Asso-ciation of Blood Banks, 1988:47-103.

30. Garratty G, Petz LD, Hoops JK. The correla-tion of cold agglutinin titrations in saline andalbumin with haemolytic anemia. Br J Haema-tol 1975;35:587-95.

31. Marks MR, Reid ME, Ellisor SS. Adsorption ofunwanted cold autoagglutinins by formalde-hyde-treated rabbit erythrocytes (abstract).Transfusion 1980;20:629.

32. Dzik W, Yang R, Blank J. Rabbit erythrocytestroma treatment of serum interferes withrecognition of delayed hemolytic transfusionreactions (letter). Transfusion 1986;26:303-4.

33. Kaji E, Miura Y, Ikemoto S. Characterizationof autoantibodies in mixed-type autoimmunehemolytic anemia. Vox Sang 1991;60:45-52.

34. Garratty G, Postoway N, Nance SJ, Brunt DJ.Spontaneous agglutination of red cells with apositive direct antiglobulin test in various me-dia. Transfusion 1984;24:214-7.

35. Branch DR, Petz LD. A new reagent (ZZAP)having multiple applications in immunohe-matology. Am J Clin Pathol 1982; 78:161-7.

36. Byrne PC. Use of a modified acid/EDTA elutiontechnique. Immunohematology 1991;7:46-7.

37. Issitt PD, Smith TR. Evaluation of antiglobu-lin reagents. In: Myhre BA, ed. A seminar onperformance evaluation. Washington, DC:

BB. All rights reserved.


American Assoc ia t ion of Blood Banks ,1976:25-73.

38. Hardman JT, Beck ML, Hardman PK, Stout LC.Incompatibility associated with the bite of abrown recluse spider (Loxosceles reclusa).Transfusion 1983;23:233-6.

39. Salama A, Müeller-Eckhardt C. Immune-me-diated blood cell dyscrasias related to drugs.Semin Hematol 1992;29:54-63.

40. Petz LD, Müeller-Eckhardt C. Drug-inducedimmune hemolytic anemia. Transfusion 1992;32:202-4.

41. Shulman NR, Reid DM. Mechanisms of drug-induced immunologica l ly mediated cy-topenias. Transfus Med Rev 1993;7:215-29.

42. Christie DJ. Specificity of drug-induced im-mune cytopen i as . Transfus Med Rev1993;7:230-41.

43. Garratty G. Drug-induced immune hemolyticanemia. In: Garratty G, ed. Immunobiology oftransfusion medicine. New York: Marcel Dek-ker, 1994:523-51.

44. Spath P, Garratty G, Petz LD. Studies on theimmune response to penicillin and cepha-lothin in humans: 1. Optimal conditions fortitration of hemagglutinating penicillin andcephalothin antibodies. J Immunol 1971;107:845-58.

45. Gallagher NI, Achergen ML, Sokol-AndersonML, et al. Severe immune-mediated hemolyticanemia secondary to treatment with ce-fotetan. Transfusion 1992;32:266-8.

46. Garratty G, Nance S, Lloyd M, Domen R. Fatalimmune hemolytic anemia due to cefotetan.Transfusion 1992;32:269-71.

47. Packman CH, Leddy JP. Drug-related immunehemolytic anemia. In: Beutler E, LichtmanMA, Coller BS, Kipps TJ, eds. Williams’ hema-tology, 5th ed. New York: McGraw-Hill,1995:691-7.

48. Kirtland HH III, Mohler DN, Horwitz DA.Methyldopa inhibition of suppressor lympho-cyte function. A proposed cause of autoim-mune hemolytic anemia. N Engl J Med1980;302:825-32.

49. Klohe EP, Myers TA. Methyldopa inhibits T cellactivation and proliferation. Transfusion 1994(suppl);34:78S.


50. Gilliland BC. Autoimmune hemolytic anemia.In: Rossi EC, Simon TL, Moss GS, Gould SA,eds. Principles of transfusion medicine, 2nded. Baltimore, MD: Williams and Wilkins,1996:101-20.

51. Petz LD, Branch DR. Drug-induced immunehemolytic anemia. In: Chaplin H, ed. Methodsin hematology: Immune hemolytic anemias.New York: Churchill Livingstone, 1985:47-94.

52. Budvari S, ed. The Merck index. 11th ed.Rahway, NJ: Merck & Co, 1989.

53. Salama A, Berghöfer H, Müeller-Eckhardt C.Detection of cell-drug (hapten)-antibody com-plexes by the gel test. Transfusion 1992;32:554-6.

Suggested ReadingDacie JV. Autoimmune anemia. Arch Intern Med1975;135:1293-300.

Dacie JV. The haemolytic anemias. Congenital andacquired. II. The auto-immune anemias. 3rd ed.London: J & A Churchill Ltd, 1985.

Engelfriet CP, Overbeeke MAM, von dem BorneAEGKr. Autoimmune hemolytic anemia. SeminHematol 1992;29:3-12.

Garratty G. Immune cytopenias associated withantibiotics. Transfus Med Rev 1993;7:255-67.

Pirofsky B. Autoimmunization and the autoim-mune hemolytic anemias. Baltimore: Williams andWilkins, 1969.

Wallace ME, Levitt JS, eds. Current applicationsand interpretations of the direct antiglobulin test.Arlington, VA: American Association of BloodBanks, 1988.

Worlledge SM, Blajchman MA. The autoimmunehaemolyt ic anemias . Br J Haemotol 1972;23(suppl):61-9.


Appendix 18-1. Some Drugs Associated with Immune Hemolysis and/orPositive DATs Due to Drug-Induced Antibodies

Drug Therapeutic Action Possible MechanismAcetaminophen Analgesic, antipyretic DD-ICAminopyrine Analgesic, antipyretic DD-ICAmphotericin B Antifungal, antibiotic DD-ICAmpicillin Antibacterial DD-ICAntazoline Antihistamine DD-ICApazone (azapropazone) Anti-inflammatory, analgesic DI, DD-DAButhiazide (butazide) Diuretic, antihypertensive DD-ICCarbenicillin Antibacterial DD-DACarbimazole Thyroid inhibitor DD-ICCatergen Diarrheal astringent, treatment of

hepatic diseaseDI

Cephalosporins AntibacterialsFirst generation NIASecond generation DD-IC, DD-DA, DIThird generation DD-IC, DD-DA, DI

Chaparral DIChlorpropamide Antidiabetic DD-ICChlorpromazine Antipsychotic DI, DD-ICCianidanol DI, DD-DA, DD-ICCisplatin Antineoplastic NIACyclofenil Gonad-stimulating principle DIDiclofenac Anti-inflammatory DI, DD-ICDiglycoaldehyde Antineoplastic NIADipyrone Analgesic, antipyretic DD-IC, DD-DAErythromycin Antibacterial DD-DAFenfluramine Anorectic ?Fenoprofen Anti-inflammatory, analgesic DI, DD-ICFluorescein Injectable dye DD-DA, DD-ICFluorouracil Antineoplastic DD-ICGlafenine Analgesic DI, DD-ICHydralazine Antihypertensive DD-ICHydrochlorothiazide Diuretic DD-ICElliptinium acetate Antineoplastic DD-ICIbuprofen Anti-inflammatory DIInsulin Antidiabetic DD-DA?, DD-ICIsoniazid Antibacterial, tuberculostatic DD-DA?, DD-ICLevodopa Antiparkinsonian, anticholinergic DIMefenamic acid Anti-inflammatory DIMelphalan Antineoplastic DD-ICMethadone Narcotic analgesic ?Methicillin Antibacterial DD-DAMethotrexate Antineoplastic, antimetabolite DD-IC

(continued)


Copyright © 2002 by the AABB. All rights reserved.

Drug Therapeutic Action Possible MechanismMethyldopa Antihypertensive DINaficillin Antibacterial DD-DANomifensine Antidepressant DI, DD-ICp-Aminosalicylic acid Antitubercular DD-ICPenicillin G Antibacterial DD-DAPhenacetin Analgesic, antipyretic DI, DD-ICPodophyllotoxin Antineoplastic, cathartic ?Probenecid Uricosuric DD-ICProcainamide Cardiac depressant, antiarrhythmic DIPyramidon DD-ICQuinidine Cardiac depressant, antiarrhythmic DD-DA, DD-ICQuinine Antimalarial DD-ICRanitidine Antagonist (to histamine H2 receptors)Rifampin (rifampicin) Antibacterial, antitubercular DD-ICSodium Pentothal Anesthetic DD-ICStibophen Antischistosomal DD-ICStreptomycin Antibacterial, tuberculostatic DI, DD-DA, DD-ICSulfonamides Antibiotics DD-ICSulfonylurea derivatives Antidiabetic DD-ICSuramin Antitrypanosomal, antifilarial NIATeniposide Antineoplastic DI, DD-ICTetracycline Antibacterial, antirickettsial, antiamebic DD-DA?, DD-ICThiopental Anesthetic DD-ICTolbutamide Antidiabetic DD-DATolmetin Anti-inflammatory DI, DD-ICTriamterene Diuretic DD-ICTrimellitic anhydride Used in preparation of dyes, resins, etc ?Zomepirac Analgesic, anti-inflammatory DD-DA, DD-IC, DI

Mechanisms listed are based on descriptions in the literature.39-43,47,50

DAT = Direct antiglobulin test.DI = Drug-independent. Associated with autoantibodies similar to those in AIHA. Drug notrequired for in-vitro demonstration. Mechanisms of autoantibody production may vary.DD-DA = Drug-dependent. Drug adsorbed onto red cells, antibody reacts with drug on cells.DD-IC = Drug-dependent. “Immune complex mechanism.” Requires drug, serum, and redcells for serologic demonstration. For most of these drugs there are only single or very fewcase reports.NIA = Nonimmunologic adsorption of proteins.? = Mechanism unclear or unknown.



tm12 - the positive dat and immune-mediated red …...c3d/red cell, although newer reagents and more...

Documents