ESTIMATION OF CELL SURVIVAL AFTERTRANSFUSION BY SELECTIVEAGGLUTINATION

Dorothy E. Osborne, Orville. F. Denstedt

J Clin Invest. 1947;26(4):655-666. https://doi.org/10.1172/JCI101849.

Research Article

Find the latest version:

http://jci.me/101849/pdf

Pdf

ESTIMATION OF CELL SURVIVAL AFTER TRANSFUSIONBYSELECTIVE AGGLUTINATION

By DOROTHYE. OSBORNEAND ORVILLE. F. DENSTEDT

(From the Department of Biochemistry, McGill University, Montreal, Canada)

(Received for publication August 31, 1946)

Selective agglutination, as a means of estimatingthe number of donor's erythrocytes in the circula-tion of a recipient after transfusion, was used firstby Ashby (1) in 1919. In the procedure, group-0 blood was administered to a subject of group A(or B), and in blood specimens taken periodicallythereafter, the number of surviving donor-cells wasestimated by selectively agglutinating those of therecipient with anti-A (or anti-B) serum, andcounting the unagglutinated group-O cells in ahemocytometer. Wiener and Schaefer (2, 3), in1939, obtained the same result by using the Mand N blood types. More recently still Wiener(16), and Mollison (4) have extended the tech-nique to the Rh blood types using Rh-negativedonors and Rh-positive recipients with anti-Rhas the agglutinating serum. Wiener and Peters(5), and Mollison and Young (6) have demon-strated the adaptability of the method by follow-ing the survival of 2 donors' cells simultaneouslyin the same recipient. The results obtained bythese techniques are in good agreement, and showthat if fresh blood is used for transfusion the do-nor's erythrocytes disappear from the recipient'scirculation at the rate of about 1 per cent per day,which is the normal rate of red cell destruction. Inthe past 8 years the method has proved to be ofgreat value in studies on blood preservation, par-ticularly in testing the effects of various methodsof preservation on the viability of erythrocytes dur-ing storage. From the practical standpoint, the 0and A blood groups offer the advantage that theyoccur frequently in the population and that highlypotent anti-A serum is readily obtainable.

The original agglutination procedure of Ashbyhas been modified in various laboratories with re-gard to the amounts of cell suspension and anti-serum used, temperature conditions, and technique.Numerous workers (1, 3, 7, 8, 9, 10) perform thereaction in small serological test tubes with orwithout agitation. Maizels and Paterson (11)use small beakers, and Loutit (12), small bottles

with a relatively small amount of antiserum. Thai-himer (13) has reduced the number of manipula-tions to a minimum by carrying out the reaction ina red-cell pipette. The present authors have modi-fied their former method by using light centrifuga-tion to accelerate agglutination. This treatment isalso a feature of the method of Dacie and Mollisonexcept in the use of anti-M and anti-N sera (14)which are said to give some non-specific agglutina-tion with centrifugation. All these procedures arebased on extensive experimentation and, in thehands of experienced workers, give comparableresults.

In determining the number of the donor's eryth-rocytes in the recipient's circulation by selectiveagglutination, it is necessary beforehand to knowthe recipient's "free" or "unagglutinable" cellcount. In other words since complete agglutina-tion of erythrocytes is impossible by ordinary meth-ods the number of unagglutinable cells per mm.3must be ascertained prior to transfusion. Thenumber of donor's cells in the post-transfusionsamples may be determined by subtracting the re-cipient's free-cell count from the total numberper mm.8 Experience has shown that the free-cellcount of the individual is fairly constant with agiven potent antiserum, but it may vary withdifferent antisera depending largely on the ag-glutinin potency, or titre. Authorities are agreedthat strong antiserum is essential for accurate re-sults. The consensus is that an antiserum for se-lective agglutination tests should be specific forthe recipient's red cells and should give a free-cellcount not exceeding 50,000 per mm.- of blood.

From time to time the reliability of the selectiveagglutination method has been queried. The vari-ous criticisms and their refutation have been dis-cussed by Maizels and Paterson ( 11 ) and by Mol-lison and Young (15).

The study reported in the present paper arosefrom certain anomalies observed in an earlier in-vestigation on red cell survival carried on during

655

DOROTHYE. OSBORNEAND ORVILLE F. DENSTEDT

1940 to 1943. At that time we were using themethod of Wiener and Schaefer (3) with anti-Mand anti-N sera prepared in our laboratory accord-ing to the methods of Wiener (16) and of David-sohn and Rosenfeld ( 17). The antisera, which wereprepared aseptically, were potent and highly spe-cific. Prior to use, the serum for a complete ex-periment was given an additional absorption witha light suspension of saline-washed red cells of theprospective recipient to ensure the absence of non-specific agglutinins. Frequently, the serum, ifstrong, was diluted further before dispensing it intosmall serological tubes. The latter then werestored at 0° C. These antisera were of high avid-ity, producing visible and compact agglutinationwithin a few seconds. The free-cell count usuallywas less than 30,000 per mm.3 but up to 50,000was considered satisfactory.

The above mentioned anomalies in our cell-survival results were periodic increases in thenumber of unagglutinable cells, a condition whichshould not have existed provided that the recipi-ent's free-cell count remained constant and thetransfused cells were being destroyed continuously.That the increases were not attributable to errorsin counting or to alteration in the antiserum wasruled out by repetition of the counts with severallots of the antiserum. Some of the tests were re-peated as many as 12 times with consistent results;others were checked by both authors. In severalinstances the deviations in the survival counts fol-lowed a steady course over a period of 2 or 3 weeksand exhibited a periodicity which was inexplicable.That some erythrocytes were being sequesteredfrom the circulation and released again periodicallywas hardly tenable. As other investigators (3, 15)in similar studies had not encountered the sametype or degree of irregularity, it was decided toinvestigate further into the cause of the phen-omenon. Attention was directed first to ascertain-ing the constancy of the free-cell-count and sec-ondly to the influence of the titre and avidity of theantiserum on the completeness of agglutination.

Discussion of the more theoretical aspects of ag-glutination will be reserved for a later publication.For a survey of the literature relevant to the pres-ent study the reader is referred to recent reviews(16, 18, 19) and to papers by Eagle (20, 21),Duncan (22) and Kendall (23).

Whatever the mechanism of agglutination maybe, authorities are agreed that the phenomenoninvolves 2 main stages, namely; (1) the combina-tion of the antigen (agglutinogen) with the anti-body (agglutinin) and (2) aggregation of thesensitized cells. Sensitization is known to occurvery rapidly while aggregation is a slower process.Eagle (21) and others have shown that mechani-cal agitation accelerates the rate of agglutinationby increasing the incidence of collision of the cells.The process is very rapid at first but decreaseslogarithmically with time. It may be broughtabout more rapidly and firmly with the aid of cen-trifugation. Thus the titre of an antiserum, whendetermined with centrifugation is from 6 to 10times that without the treatment. Up to a pointagglutination increases with the concentration ofagglutinin. Very potent antisera may not produceagglutination or may not give a maximum reaction'until diluted. This upper zone of inhibition isusually called the "prozone" to distinguish it fromthe lower zone, the "postzone" where aggregationof cells is diminished due to dilution of the anti-body. The mechanism of prozone inhibition isnot fully understood although considerable lighthas been thrown upon the subject by the studiesof Jones and Orcutt (24).

The inadequacy of the ordinary titre method forcharacterizing the potency of an antiserum hasbeen stressed by numerous authors. Obviously thetitre, expressed as a number, merely gives a roughand comparative idea of the agglutinin concentra-tion of the original material. For the purposes ofthe clinical laboratory, however, the method isvaluable particularly for determining the point ofoptimal proportion of antigen and antibody, i.e.,the dilution which produces most rapid agglutina-tion. With moderately strong or weak serum, themaximum rate of agglutination is obtained withthe undiluted material. As erythrocytes of differ-ent individuals of the same blood group vary con-siderably in properties, the titre of a given serumand the concentrations of optimum proportion mayvary with the blood tested.

In the present study the authors have sought ananswer to the following questions:

(1) Can the free-cell count be abolished by care-ful adjustment of the antiserum concentration andby control of conditions of the reaction?

656

CELL SURVIVAL AFTER TRANSFUSIONBY SELECTIVE AGGLUTINATION

(2) What is the normal variation in the free-cell count ?

(3) To what degree is incomplete agglutinationof red cells influenced by the following factors:

(a) Too rapid agglutination and insuffi-cient time for completion of the reaction.

(b) Insufficient antibody.(c) Under-sensitization of cells due to de-

ficiency in antigen content, or to age.

EXPERIMENTAL

In 1944 the present authors modified their ag-

glutination technique by introducing centrifuga-tion, thus reducing the time of the procedure toabout one-third. Also, with the increased avail-ability of highly potent anti-A and anti-B sera a

change was made to the 0, A, and B blood groups

in preference to the Mand N types.

METHODS

1. Selection of antiserumt

In selecting an antiserum for an experiment an ordinarytitre series is set up for each serum to be tested usingthe saline-washed red cells of the prospective recipient.With the aid of a microscope the particular dilution inthe series which gives the most complete agglutination isascertained. This concentration may not be the one whichproduces the most rapid agglutination or the most compactcell sediment. Having selected the optimum serum con-

centration for completeness of reaction, a larger quantityof the particular dilution is prepared and dispensed asepti-cally into small tubes which then are stoppered andstored at 00 C. Prior to use, each tube is permitted tocome to room temperature and is inverted several timesto ensure uniformity of the contents.

Very potent anti-A or anti-B sera, particularly thoseprepared by concentrating the globulin fraction, fre-quently develop a surface precipitate of lipid when storedat 00 C. or after having been frozen. Unless removed,this material may prove a source of much annoyance andconfusion during cell counting since many of the par-

ticles are about the size of the erythrocytes. One of theauthors (D. E. 0.) has found that the lipid can easilybe removed by keeping the serum in a centrifuge tubein the cold for a few days and then cautiously overlayingthe surface with physiological saline followed by centri-fugation at high speed. If the saline is applied slowlywith a fine capillary pipette, no mixing with the serum

will occur and, after centrifuging, the saline layer, on

the surface of which the lipid will have collected, can beremoved by aspiration through a capillary tube.

2. Collection of blood samples

Prior to transfusion, and periodically thereafter, bloodsamples are taken from an antecubital vein of the re-

cipient. Approximately 10 ml. are delivered into a gradu-ated centrifuge tube containing 1 ml. of sterile sodiumcitrate solution (3.2 per cent). The tube is stoppered andinverted gently several times. By emptying the syringeagainst the wall of the tube and by using gentle mixingmethods excessive frothing of the sample can be avoided.Foam may be collapsed by touching it with a fine wire orneedle the mere tip of which has been dipped into caprylicalcohol. The total volume of the specimen is recorded toobtain the dilution factor. The hemoglobin concentra-tion of the specimen is determined with a photoelectriccolorimeter, and the hematocrit, using the Wintrobe tube.A total red-cell count is performed on the pretransfusionspecimen.

3. Agglutination procedureBlood samples, if refrigerated, are permitted to come to

room temperature. The red cells are suspended by gentlyand repeatedly inverting the tube for about 1 minute.Another light mixing is performed immediately beforepipetting.

Blood is drawn up to the 0.5 mark in a white-cellpipette which is then filled to the 11 mark with sodiumchloride solution (0.9 per cent). After the pipette isshaken for a few seconds, it is emptied into a serologicaltube (10 X75 mm.).

The tube is shaken a few times, and the cell suspensionis drawn up to the 1.0 mark in a dry white-cell pipette.The latter is then filled with the agglutinating serum.After the pipette is shaken vigorously for a few seconds,the contents are transferred to a serological tube andcentrifuged in an electric table centrifuge at 1500 r.p.m.,the machine being turned off automatically at the end of2 minutes.

The sedimented red cells are dispersed by flicking thetube sharply several times by hand and shaking for an ad-ditional 45 minutes on an electric shaker 1 at 176 vibra-tions per minute. (The table of the shaker in our labora-tory is covered with a 2-inch slab of fiberglass to insulatethe specimens from the heat of the motor.)

Finally, the tube is flicked sharply to suspend the cellsediment, the larger clumps are permitted to settle for acouple of seconds, and a sample of the supernatant fluidis transferred to the hemocytometer by means of a white-cell pipette filled to the 1.0 mark. Counts are made induplicate and usually agree within 10 per cent.

The free-cell counts are calculated to the number percubic millimeter of uncitrated blood. With experimentalmixtures of 0 and A erythrocytes or OMand ONin theproportion of 1:30, the lesser component can be determinedby selective agglutination within 5 per cent.

4. Transfusion procedureIn the majority of our red-cell survival experiments

since 1944, the practice has been to determine the bloodvolume of the recipient prior to transfusion and to de-termine accurately the quantity of blood administered by

1 No. 3623 rotating apparatus, obtainable from A. H.Thomas Company, Philadelphia.

657

DOROTHYE. OSBORNEAND ORVILLE F. DENSTEDT

weighing the container before and after the transfusion.The hemoglobin, hematocrit and total red-cell count alsoare determined on a sample taken from the bottle im-mediately before the experiment. The number of donorcells per mm.3 of the recipient's blood immediately aftertransfusion, therefore, can be quite accurately calculated.

If the number of cells found is considerably less thanthe theoretical number, as is often the case when preservedblood is administered, the discrepancy is assumed torepresent the number of the donor's erythrocytes elimi-nated by the recipient during the transfusion.

The volume of citrated blood administered in ourexperiments formerly was 450 ml., but since 1944 hasusually been about 200 ml.

RESULTS

1. Influence of agitation on the rate of agglutina-tion

The rate of combination of antigen with antibodyis known to be extremely rapid. Hence, as nu-merous workers have shown, the accelerating ef-fect of various modes of agitation and sedimenta-tion on the rate of agglutination is due chiefly toincreasing the incidence of contact between thesensitized cells. The size and shape of the con-tainer also may influence the rate especially ifshaking is the mode of agitation.

Figure 1 illustrates the influence of the rate ofhorizontal shaking on the completeness of ag-glutination of erythrocytes at the end of 90 minuteswith our apparatus. The unagglutinated-cellcount is expressed as the number of free cells ob-served in the entire field of the hemocytometer,i.e., in 0.9 mm.3 of cell suspension. The effect ofagitation above 190 vibrations per minute could

not be determined with our facilities, but pre-sumably a maximum eventually would be reachedbeyond which agglutination would be less com-plete and the damaging effect of the shaking wouldbecome a serious factor. Experiments with higher,but unknown, rates of shaking invariably have re-vealed considerable hemolysis. The results ofshaking also vary somewhat with the blood, theantiserum, and the mode of agitation.

Agglutination may be accelerated greatly by aninitial light packing of the cells by centrifugation.Figure 2 gives a comparison of the degrees of ag-glutination with and without centrifuging and at2 rates of shaking. It will be observed from curveD that centrifugation, followed by shaking for 30minutes, produces as complete agglutination asdoes shaking alone at 176 vibrations per minuteat the end of 6 hours. Attempts to reduce the un-agglutinated-cell count by agitation for more than6 hours invariably resulted in visible hemolysis.Centrifuging also tends to damage red cells, and,if too prolonged or too rapid, will result in earlyhemolysis on subsequent shaking.

2. Influence of strength of antiserumIn determining the agglutinating titre of a strong

antiserum by the ordinary method of progressivedilution, one tube in the series usually shows morerapid agglutination than those preceding or follow-ing it. This particular dilution represents, or ap-proximates, the concentration of optimum propor-tion of the antigen and antibody as far as rate ofagglutination is concerned. If aggregation of the

,oo Aoo soo 400

UNAGGLIUrINATE'D CELLSSoo Soo

FIG. 1. INFLUENCEOF THE RATEOF SHAKINGONTHE COMPLETENESSOF

AGGLUTINATION

Wjj 1.90I..

i170to2 /60

It 1Scc,

658

l2wr

0

CELL SURVIVAL AFTER TRANSFUSION BY SELECTIVE AGGLUTINATION

A

HOouss

FIG. 2. PROGRESSOF AGGLUTINATIONAT Two RATESOFSHAKING WITH AND WITHOUTAN INITIAL LIGHT CEN-TRIFUGATION

Curves A and C, 146 vibrations per minute; curves Band D, 176 per minute.

red cells takes the form of a single, compact clump,one easily may be misled into assuming that ag-glutination is complete. On examination under themicroscope, however, a surprisingly large numberof unagglutinated cells may be observed. The con-centration which affords the most rapid agglutina-tion, therefore, may not produce the most com-plete reaction. This is true especially of antiseraof marked avidity, notably anti-M and anti-N sera,which, even when diluted many times, may pro-duce visible and compact agglutination within afew seconds. In preparing an antiserum for selec-tive agglutination, therefore, it is essential to as-certain the relative concentration which producesthe most complete aggregation of cells. This con-centration usually is in the neighborhood of thatwhich produces most rapid agglutination. Withserum of only moderate potency, the maximum isobtained with the undiluted material.

The relation, between the strength of the anti-serum and the completeness of agglutination or thefree-cell count is illustrated in Figures 3 to 9 in-clusive. To facilitate comparison, the data are pre-sented in the form of graphs. In doing so theauthors wish to direct attention to the inconsistencyin the results with different dilutions of serumrather than to imply that the fluctuations are due

to variation in the properties of the erythrocytesor the antiserum, although either, or both the latterconceivably may occur.

(a) CONSTANCYOF THE UNAGGLUTINABLE-CELLCOUNT

Figures 3 to 6 inclusive illustrate the fluctua-tions in the free-cell count with various dilutionsof antiserum and in different individuals, over aperiod'of 2 months. The results are expressedas thousands of cells per cubic millimeter of therecipient's blood. Attention is directed to themagnitude of the fluctuations in Figure 6. Thecounts were done in duplicate.

The variability of the free-cell count with thelower concentrations of antiserum is pronounced.With serum of marked avidity, e.g., anti-M andanti-N, the rapidity, firmness, and compactness ofagglutination may easily mislead one into thinkingthat agglutination is complete.

The stronger concentrations, on the other hand,'produce a low free-cell count, which, in the case ofthe optimum concentration, remains practically con-stant within the counting error. Occasionally wehave found that the recipient's free-cell count, afterthe donor's cells have disappeared from the circula-tion, is slightly higher than before the transfusion.In these cases the higher count has been used incalculating the results. For the purpose of the red-cell survival test, provided a strong serum is used,the unagglutinable-cell count of the recipient maybe assumed to remain constant.

1:10O

D4YSFIG. 3. VARIATION IN THE FREE-CELL COUNT OF A

GROUP-B SUBJECT WITH POTENTANTI-B SERUM(TITRE5,000) DILUTED 5, 10, AND 20 TIMES, RESPECTIVELY

659

DOROTHYE. /OSBORNE AND ORVILLE F. DENSTEDT

1: 10). It is noteworthy also that under optimumconditions of preservation the agglutinability ofthe stored erythrocytes does not differ from thatof the fresh blood. The counts are expressed asthe number of free cells in the hemocytometer fieldin order to accentuate the fluctuations. (The fac-tor for converting the count to cells per mm.8 isapproximately 270.) Since many of the eryth-rocytes, by the fiftieth day of storage, become non-functional (at least 50 per cent of the cells), it ap-pears that agglutinability is not greatly impairedby the loss of viability provided conditions of pres-ervation are favorable.

The curve in Figure 8 illustrates the behavior ofthe free-cell count in a sample of preserved bloodof group OMRhusing a potent anti-M serumdiluted 20 times. The variation is typical of thatobserved with diluted antisera even in fresh blood.There is no evidence in this and in similar experi-ments that the reactivity of the M and N agglu-

DAYSFIG. 4. VARIATION IN THE FREE-CELL COUNT OF A

GROUP-A SUBJECT WITH A POTENT ANTI-A SERUM(TITRE 5,000) DILUTED 5, 10, 20, 30, AND 40 TIMES, RE-SPECTIVELY

(b) AGGLUTINABILITY OF PRESERVED ERYTH-

ROCYTES

It was anticipated that if the fluctuations ob-served in the agglutinability of erythrocytes withweak sera are due to changes in the agglutinogenson the cell membrane, the agglutinability of thecells might become increasingly erratic withstorage.

In Figure 7, the curves "A" represent periodicfree-cell counts on blood (group A1MRhRh2)preserved in citrate-dextrose solution at 50 C.over a period of 50 days; "B", the counts on freshblood from the same donor. Here again thestronger serum (anti-A diluted 1: 5) gave more

constant results than the weaker lot (diluted

DAYSFIG. 5. FLUCTUATION IN FREE-CELL COUNT OF A

GROUP-A INDIVIDUAL WITH POTENT ANTI-A SERUM(TITRE 5,000) DILUTED 5, 20, 30, AND 40 TIMES, RE-SPECTIVELY

a'

660

CELL SURVIVAL AFTER TRANSFUSIONBY SELECTIVE AGGLUTINATION

tinogens is impaired during storage under optimumconditions at least up to 6 weeks.

3. Estimation of transfused erythrocytes

In view of certain anomalies observed up to 1944in our experiments on mental patients it wasthought advisable to perform a few cell-survivaltests on normal subjects. Accordingly 200 ml. offresh citrated blood of a donor group ONRh, wasadministered to each of 2 recipients and countswere followed daily for about 60 days. Figure 9illustrates the disappearance of the donor's cellsfrom the circulation of a recipient, groupOMNRh2,as determined with anti-M serum (titre64), and Figure 10, the curve for the second re-cipient, group AMRhRh2, as determined simul-taneously with anti-A (dotted curve) and anti-Msera. The gaps in the curves were occasioned bya holiday. It is noteworthy that despite the ir-

5am

00

z zoo

Im10-~~~~

/0 to Jo 40 SoDAYS

FIG. 6. FLUCTUATION IN FREE-CExL COUNT OF AGROUP-A INDIVIDUAL WITH A POTENT ANTI-A SERUM(TITRE 5,000) DILUTED 5, 10, AND 20 TIMES, RESPEC-TIVELY

tos

J30tat

t ato

10 0o Jo

DAYS40 50

/0

/o 20 30 #0 50DAYs

FIG. 7. COMPARATIVE VARIATION IN THE FREE-CELLCOUNTIN PRESERVEDAND FRESH BLOOD OF A GROUP-ASUBJECT USING POTENT ANTI-A SERUM (TITRE 5,000)DILUTED 5 AND 10 TIMES, RESPECTIVELY

Curves A represent preserved blood and curves B, freshblood.

regularities due to the use of diluted sera, thecourse of the red-cell elimination is quite accuratelyindicated. With diluted antisera, cell countsshould be performed at least weekly in order tobe able to correct for irregularities.

Figure 11 represents a transfusion series, per-formed in June 1944, in which 150 ml. of bloodfrom 1 donor of group ON, was administered toeach of 4 recipients. The first received a freshspecimen and the others, in turn, samples that hadbeen preserved in citrate-dextrose at 50 C. for 12,22, and 40 days, respectively. The free-cell countsof the recipients are included in Table I.

The rapid elimination of non-viable red cellswithin 3 to 6 hours after transfusion is indicatedin curves B, C, D. In our experience, with bloodstored for more than 3 weeks, there usually is in-dication of a partial elimination of the non-viablecells even during the period of the transfusion(200 ml.). The 100 per cent count in the post-transfusion sample in recipient D, therefore, is ex-ceptional. It is noteworthy that a rapid elimina-tion of erythrocytes always occurs with preservedblood within a few hours after transfusion, the lossamounting roughly to 1 per cent of the cells perday of storage (or more if conditions of preserva-

661

o A

cw--oo .5/.15

62DOROTHYE. OSBORNEAND ORVILLE F. DENSTEDT

(I)-.mJ

1o teo 30 #0DAYS &TO/A GE

FIG. 8. FLUCTUATION IN THE FREE-CELL COUNTOF APRESERVEDBLOOD SAMPLE, GROUP-AM, USING AN AVID,BUT DILUTED, ANTI-M SERUM

tion are not optimum). With strong antisera thesurvival curves, for the most part, are free fromgross irregularities.

One type of anomaly which the writers have en-countered occasionally in exaggerated form withpotent anti-M and anti-N, and even with

100

, 80>a:V)

60-J

0w 40

zLoA 200-

O4011200

-J

20 40 60 80 100 120DAYS

FIG. 10. ELIMINATION OF THE ERYTHROCYTESOF AGROUP-ONDONORFROMTHE CIRCULATION OF A GRoup-AM RECIPIENT, COUNTSBEING MADESIMULTANEOUSLYWITH ANTI-A (DOTTED CURVE) AND ANTI-M SERA

strong anti-A serum, is illustrated by the large dis-continuities in Figures 12, 13, and 14. -The par-ticulars of these experiments are given in Table II.

It will be observed that the free-cell counts of oneof the recipients of type M differs considerablyfrom the others. The same anti-M serum was usedin each case. It is noteworthy also that the largedeviation in Figure 13 was given by the recipientwith the lower free-cell count. 'The irregularitiesin the 3 figures exhibit a remarkably smoothcourse, a circumstance which makes explanationdifficult on the basis of technical errors.

DISCUSSION

Of the various- factors that may influence theaccuracy of selective agglutination, the strength ofthe antiserum is of major concern. As yet thereis no practical method for characterizing a serum

quantitatively as to its avidity and agglutinin con-tent. The titre method of estimating potency givesonly comparative values. Nevertheless, it is valu-able for testing the strength of the se-rum for typ-

TABLE I

20 40 60 80 100 120DAYS

,IMINATION OF THE ERYTHROCYTESOF A)ONOR FROMTHE CIRCULATION OF A GROUP-LENT USING DILUTED ANTI-M AGGLUTINAT-

Recipient Group Antiserum Free-cell count

per mm.'A OMN anti-M 4,000B AMN anti-M 4 000C OMN anti-M 5,000D AMN anti-M 7,000

FIG. 9. ELGROUP-ONDOMNRECIPIING SERUM

662

CELL SURVIVAL AFTER TRANSFUSIONBY SELECTIVE AGGLUTINATION

100" 4

~80

V)60

zLu~20Lii

20 40. 60 80 100 120DAYS

FIG. 11. ELIMINATION OF THE ERYTHROCYTESOF A

DONOR, BLOOD GROUP-ON, FROM THE CIRCULATION OF4 RECIPIENTS, USING A STRONGANTI-M AGGLUTINATINGSERUM. (SEE TABLE I.)

ing purposes and for estimating the optimum con-centration of red cells and antiserum.

The term "optimum," as applied to agglutina-tion, usually implies rapidity of reaction ratherthan completeness. The point of optimum propor-tion in the titre series, for example, is the concen-tration of cells and serum which produces the mostrapid or most compact agglutination. As men-tioned previously, this concentration does not nec-essarily produce maximum agglutination and,with sera of marked avidity, e.g., anti-M and anti-N, the number of unagglutinable cells may be rela-tively large. Rapidity, compactness and firmnessof agglutination, therefore, are not reliable criteriaof the completeness of the reaction.

As shown in the foregoing graphs (Figures 3

TABLE II

Recipi-Figure Donor Recipient Storage Anti- went'speriod serum free-cell

count

days per mm.3OMRh1 fresh 4,000

12 ON anti-MOMRh2 25 3,000

Pooled OMRhRh2 22 18,00013 ONRh+ anti-M

ONRhj OMRh1 22 5,000

14 OMRh1 ONRh2 21 anti-N 42,300

100

-J< 80

> 60C-)

as, 40

zLu

) 20cL

20 40 60 80 100 120DAYS

FIG. 12. ELIMINATION OF THE ERYTHROCYTESOF AFRESH AND A PRESERVEDBLOODSAMPLEFROMA GROUP-ONDONOR,FROMTHE CIRCULATION OF 2 GROUP-OMRE-CIPIENTS. (SEE TABLE II.)

to 9) antisera, particularly diluted sera, which givea high free-cell count, usually display markedvariability. It is noteworthy, however, that thefree-cell count even with undiluted sera may varywith different bloods. Thus, in Table II, the free-cell count with one individual was 5,000 and withanother, was 18,000. Occasionaly also, one meetswith a potent serum which gives a constant free-cell count with one person and a variable countwith another. There are instances when dilutionof the serum reduces the free-cell count, e.g., whenthe serum concentration is above the optimum.In general, however, one can say that a high con-centration of agglutinin is necessary for maximumand consistent agglutination, and that dilution ofthe serum tends to increase the number of unag-glutinated cells and the variation in the counts.

The reason for the high free-cell count withmarkedly avid antisera is not clear. Anti-M andanti-N sera, like many heterologous immune sera,produce a very rapid and firm type of agglutina-tion. It is significant that the unagglutinated cellsapparently do not adhere to the agglutinates al-ready formed nor are they occluded to any con-siderable extent among the rapidly aggregatingmasses. The failure of many of the cells to becomeagglutinated cannot be attributed to a lack of anti-body. Centrifugation will cause further aggrega-tion but not to the extent one would expect withfully sensitized cells.

663

664

100

DOROTHYE. OSBORNEAND ORVILLE F. DENSTEDT

680

2606cL

LI~~~~~~~~~~~~~~~~~~~~~~I~~~~~:e

20 40 60 80 100 120DAYS

FIG. 13. ELIMINATION OF PRESERVEDERYTHROCYTESOF 2 GROUP-ONDo-NORS (POOLED SAMPLE) FROMTHE CIRCULATION OF 2 GROUP-OMRECIPIENTS.(SEE TABLE II.)

100

-I

* 80

_s 60Lii

o 40

z

x 200c

20 40 60 80 100DAYS

FIG. 14. ELIMINATION OF PRESERVEDERYTHROCYTESOFNORFROMTHE CIRCULATION OF A GROUP-ONRECIPIENT.

On first thought, the failure of some cells to beagglutinated with strong antiserum suggests thatthe red-cell population possibly may be made upof cells ranging in agglutinability from extremelyto feebly reactive, and that, while the latter may beloosely agglutinated by centrifuging, like weak Rhagglutinates, they may be dispersed again by agita-tion. Such a hypothesis would imply that the so-

called "unagglutinable" cells may be deficient inthe content, distribution, or quality of their ag-

A GROUP-OMDo-(SEE TABLE II.)

glutinogens due perhaps to imperfect formation.In opposition to this reasoning, one may point

out that the addition of a strong antiserum to theresidual free cells after agglutination with a weakantiserum, does not effect as complete agglutina-tion as when the strong serum is applied directlyto fresh cells. Weak serum may give rise to ag-glutinates in the form of doublets, or chains whichcan be broken by shaking. Addition of strongerantiserum does not consolidate the aggregates into

CELL SURVIVAL AFTER TRANSFUSIONBY SELECTIVE AGGLUTINATION

a compact mass. It is significant also that 2 vol-umes of half-strength antiserum are not necessarilyequivalent in agglutinating strength and effect to 1volume of undiluted serum for a given volume ofcell suspension even when centrifugation is em-

ployed. On the basis of these observations; oneis inclined to attribute the differences in free-cellcount and agglutinability to a reversible changein the state of the agglutinins or other componentsof the system with dilution and changes in tem-perature.

The variability in the counts with diluted anti-serum is of considerable interest since the fluctu-ations are not due to systematic errors. In thewriters' experience the free-cell counts are likelyto vary from day to day but they may be quiteconsistent on any day.

While the use of diluted, though avid, anti-Mand anti-N sera explains some of the sharp ir-regularities observed in our earlier study, thelarger and more regular deviations of the type il-lustrated in Figures 12, 13, and 14 are not so

easily explained. Our former interpretation, thatperhaps some of the donor's erythrocytes are

sequestered from the recipient's circulation and re-

leased later, has been abandoned. The deviationsdoubtless are connected in some way with re-

versible changes in the properties of the antiserum,or possibly to reversible changes in the reactivityof the recipient's cells. It is evident from TableII that they are not attributable to the use of weakantiserum, nor are they confined to the use of theMand N blood types. Other factors such as agi-tation, temperature, salt concentration, and pHhave been ruled out. The study of this phenom-enon is being continued.

Highly potent anti-A and anti-B sera being easy

to prepare, or being readily obtainable the writersrecommend for selective agglutination, sera whichgive a free-cell count not exceeding 10,000 per

mm.3 The lowest free counts obtained in our

studies. were about 2,500. With strong serum thevariation in the free-cell count is usually within5,000 per mm.8 Irregularities in the cell-survivalcurve become noticeable when the free-count ex-

ceeds 15,000, and, above 40,000, fluctuations may

occur ranging up to 200,000 from day to day. Ob-viously it is desirable to keep the variation withinthe limits of error of the technique. With small

transfusions, e.g., 200 ml., it is especially impor-tant to select antisera which will reduce the free-count to a minimum.

CONCLUSIONS

From the evidence submitted and additionalwork not reported herein, it appears that there isno routine method for achieving complete ag-glutination of erythrocytes and that such an at-tainment is not possible when agitation is involved.In some experiments, however, the unagglutinable-cell count has been reduced to as low as 2,500 permm.3 With strong agglutinating serum the free-cell count of the individual remains practicallyconstant for the majority of persons, especially ofgroup A,. The sharp irregularities observed inthe cell-survival curves in an earlier study areattributable to the use of diluted anti-M and anti-Nsera. The larger and more regular type of devia-tion, however, still remains unexplained; it isdoubtless connected with some obscure change inthe properties of the anti-serum or of the recipi-ent's red cells. These deviations and irregularitiesdo not necessarily vitiate the results of the cell-survival method. A high degree of accuracy canbe obtained in the estimation of survival if theblood volume of the recipient and the number oferythrocytes transfused are accurately known.

SUMMARY

The presentation concerns the factors that affectthe accuracy of the selective-agglutination pro-cedure for estimating the survival of erythrocytesafter transfusion. The influence of agitation, cen-trifugation, and the agglutinin titre of the anti-serum on the rate and completeness of agglutina-tion of red cells is discussed. The unagglutinable-cell count cannot be eliminated, but it can bereduced to a minimum and practically constantquantity by careful selection of the antiserum.Some of the irregularities observed in an earlierstudy undoubtedly were due to the use of diluted,though avid, anti-M and anti-N sera. The largerand more regular deviations, however, are not soeasily explained; they appear to be related to re-versible changes in the properties of the antiserumor to reversible changes in the properties of therecipient's red cells. If the blood volume of therecipient and the number of red cells transfused

665

DOROTHYE. OSBORNEAND ORVILLE F. DENSTEDT

are accurately known, the progressive eliminationof the donor's cells from the circulation of the re-cipient can be followed with a fairly high degreeof accuracy.

ACKNOWLEDGMENTS

This study is part of an investigation on blood preserva-tion carried on with the financial support of the NationalResearch Council of Canada.

The authors are indebted to Dr. L. K. Diamond for asupply of high-titred anti-A and anti-B serum, and toDr. A. S. Weiner for anti-M and anti-N sera, and for Rh-typing.

Thanks are due also to Dr. H. Lehmann of the VerdunProtestant Hospital who performed most of the transfu-sions And collected the blood specimens.

BIBLIOGRAPHY

1. Ashby, W., The determination of the length of life oftransfused blood corpuscles in man. J. Exper. Med.,1919, 29, 267.

2. Schaefer, G., and Wiener, A. S., Limitations in useof preserved blood for transfusions; preliminaryreport. Quart. Bull. Sea View Hosp., 1939, 5, 17.

3. Wiener, A. S., and Schaefer, G., Limitations in useof preserved blood for transfusion; study of fateof transfused erythrocytes in recipient's circulation.M. Clin. North America, 1940, 24, 705.

4. Mollison, P. L., The survival of transfused erythro-cytes in hemolytic disease of the newborn. Arch.Dis. Childh., 1943, 18, 161.

5. Wiener, A. S., and Peters, H. R., Hemolytic reactionsfollowing transfusions of blood of the homologousgroup, with 3 cases in which the same agglutinogenwas responsible. Ann. Int. Med., 1940, 13, 2306.

6. Mollison, P. L., and Young, I. M., Survival of trans-fused erythrocytes of stored blood. Lancet, 1940,2, 420.

7. Bushby, S. R. M., Kekwick, A., Marriott, H. L., andWhitby, L. E. H., Survival of stored red cellsafter transfusion. Lancet, 1940, 2, 414.

8. Belk, W. P., and Barnes, B. C., Survival time aftertransfusion of erythrocytes of citrated humanblood stored at 40 to 60 centigrade. Am. J. M. Sc.,1941, 201, 83&

9. Denstedt, 0. F., Osborne, D. E., Stansfield, H., andRochlin, I., Survival of preserved erythrocytes aftertransfusion. Canad. M. A. J., 1943, 48, 477.

10. Wearn, J. T., Warren, S., and Ames, O., Length oflife of transfused erythrocytes. Arch. Int. Med.,1922, 29, 527.

11. Maizels, M., and Paterson, J. H., Survival of storedblood after transfusion. Lancet, 1940, 2, 417.

12. Loutit, J. F., Factors influencing preservation ofstored red cells. J. Path. & Bact., 1945, 57, 325.

13. Thalhimer, W., and Taylor, E. S., Transfusion ofcentrifuged human type 0 cells resuspended andstored in 10 per cent corn syrup. J. A. M. A., 1945,127, 1096.

14. Dacie, J. V., and Mollison, P. L., Survival of normalerythrocytes after transfusion to patients withfamilial haemolytic anaemia (acholuric jaundice).Lancet, 1943, 1, 550.

15. Mollison, P. L., and Young, I. M., On survival oftransfused erythrocytes of stored blood. Quart. J.Exper. Physiol., 1941, 30, 313.

16. Wiener, A. S., Blood Groups and Transfusion. ThirdEdition. Charles C. Thomas, Springfield, Illinois,1943.

17. Davidsohn, I., and Rosenfeld, I., Preparation of anti-M and anti-N testing fluids. Am. J. Clin. Path.,1939, 9, 397.

18. Landsteiner, K., The Specificity of Serological Reac-tions. Harvard University Press, Cambridge,Mass., 1945.

19. Boyd, W. C., Fundamentals of Immunology. Inter-science Publishers, Inc., N. Y., 1943.

20. Eagle, H., Specific agglutination and precipitation;mechanism of reactions. J. Immunol., 1930, 18,393.

21. Eagle, H., Specific agglutination and precipitation;velocity of reactions. J. Immunol., 1932, 23, 153.

22. Duncan, J. T., Specific character of stage of aggrega-tion in agglutination and precipitation of antibody-antigen compounds. Brit. J. Exper. Path., 1938,19, 328.

23. Kendall, F. E., Quantitative relationship between anti-gen and antibody in precipitin reaction. Ann. NewYork Acad. Sc., 1942, 43, 85.

24. Jones, F. S., and Orcutt, M., Prozone phenomenon inspecific bacterial agglutination. J. Immunol., 1934,27, 215.

666