Evaluation of methods for detecting alloantibodies underlying warm autoantibodies.

Abstract

In pretransfusion testing of patients whose sera contain autoantibodies reacting optimally at 37 degrees C, it must be determined whether alloantibodies are also present. Two approaches, testing a 1-in-5 dilution of patients' sera and the adsorption of sera in the presence of polyethylene glycol (PEG), have been proposed as alternatives to the time-consuming approach of adsorbing sera with ficin- or ZZAP-treated red cells (RBCs). The three approaches were compared.
Patients' sera containing warm autoantibodies, with and without alloantibodies, were retested 1) after dilution (1-in-5) and 2) after adsorption with allogeneic RBCs in the presence of PEG. Results were compared to those after adsorption with ZZAP-treated allogeneic RBCs.
Dilution (1-in-5): Twenty-seven of 119 sera (7/26 [27%] with and 20/93 [22%] without alloantibodies) did not react; one example each of alloanti-D, -E, -e, -Fy(a), and -Jk(a), and two examples of anti-Jk(b) were not detected at a dilution of 1 in 5. Alloantibodies were identified in 5 (19%) of 26 1-in-5 diluted sera containing alloantibodies; 87 (73%) of 119 sera still reacted with all cells and would have required further workup. PEG adsorption: Thirty-nine sera were tested after parallel PEG and ZZAP adsorptions. The PEG adsorptions required a total of 55 aliquots of adsorbing cells and 13.75 hours, whereas ZZAP adsorptions required 61 aliquots and 30.5 hours. All alloantibodies (anti-D [3], -C [2], -c [1], -E [4], -K [2], -Fy(a) [1], -Jk(a) [2], -Jk(b) [1]) reacted in the adsorbed serum-PEG mixtures at a strength equal to or greater than that in the ZZAP-adsorbed sera.
Although the 1-in-5 dilution approach is convenient, only 22 percent of warm autoantibodies without alloantibodies were nonreactive, and 27 percent of alloantibodies of potential clinical significance were not detected. PEG adsorption appears to give similar results to those of ZZAP adsorption, but it has the advantages of eliminating the cost and time of prior treatment of the allogeneic adsorbing cells and of a reduction of at least a 50 percent in adsorption time.

Full text

Page 1

I M M U N O H E M A T O L O G Y
Evaluation of methods for detecting
alloantibodies underlying warm autoantibodies
R.M. Leger and G. Garratty
BACKGROUND: In pretransfusion testing o f patients
whose sera contain autoantibodies reacting optimally at
37”C, it must be determined whether alloantibodies are
also present. Two approaches, testing a 14-5 dilution o f
patients’ sera and the adsorption o f sera in the presence
of polyethylene glycol (PEG), have been proposed as al-
ternatives to the time-consuming approach o f adsorbing
sera with ficin- or ZAP-treated red cells (RBCs). The
three approaches were compared.
STUDY DESIGN AND METHODS: Patients’ sera con-
taining warm autoantibodies, with and without alloanti-
bodies, were retested 1) after dilution (1-in-5) and 2) af-
ter adsorption with allogeneic RBCs in the presence o f
PEG. Results were compared to those after adsorption
with ZAP-treated allogeneic RBCs.
RESULTS: Dilution (1-in-5): Twenty-seven of 119 sera
(7/26 [27%] with and 20/93 [22%0] without alloantibodies)
did not react; one example each o f alloanti-D, -E, -e, -
Fya, and -Jka , and two examples o f anti-Jkb were not de-
tected at a dilution o f 1 in 5. Alloantibodies were identi-
fied in 5 (19%) o f 26 1-in-5 diluted sera containing
alloantibodies; 87 (73%) of 119 sera still reacted with all
cells and would have required further workup. PEG ad-
sorption: Thirty-nine sera were tested after parallel PEG
and ZAP adsorptions. The PEG adsorptions required a
total o f 55 aliquots of adsorbing cells and 13.75 hours,
whereas ZAP adsorptions required 61 aliquots and
30.5 hours. All alloantibodies (anti-D [3], -C [2], -c [l], -E
[4], -K [2], -Fya [l], -Jka[2], -Jkb[l]) reacted in the
adsorbed serum-PEG mixtures at a strength equal to or
greater than that in the ZZAP-adsorbed sera.
CONCLUSION: Although the 1 -in-5 dilution approach is
convenient, only 22 percent of warm autoantibodies
without alloantibodies were nonreactive, and 27 percent
o f alloantibodies o f potential clinical significance were
not detected. PEG adsorption appears to give similar re-
sults to those o f ZZAP adsorption, but it has the advan-
tages o f eliminating the cost and time of prior treatment
of the allogeneic adsorbing cells and o f a reduction o f at
least a 50 percent in adsorption time.
atients with autoantibodies in their sera that re-
act optimally at 37°C (warm autoantibodies)
present a challenge to serologists who must de-
P
present. Warm autoantibody-adsorption procedures are
tedious and time-consuming. Laboratories often treat al-
logeneic adsorbing red cells (RBCs) with enzymes or ZZAP’
(a combination of dithiothreitol and enzyme) reagent be-
fore adsorption to increase the efficiency of the procedure.
Recently, two techniques were proposed as alternatives: 1)
testing a 1-in-5 dilution of patients’ sera2 and 2) adsorption
in the presence of polyethylene glycol
The 1-in-5 dilution technique reported by 0yen and
Angeles2 appears to be a modification of a method pro-
posed earlier by Petz and Garratty.’ Petz and Garratty pro-
posed testing the patient’s serum, at a dilution at which the
autoantibody reactivity is weak, against a panel of RBCs;
this dilution was selected by titrating the patient’s serum
against a pool of two antibody-detection RBCs. This method
was proposed for laboratories that do not have the capa-
bility to perform adsorptions: however, it was emphasized
that only alloantibodies whose titer is higher than the au-
toantibodies’ titer will be detected.
Liew and Duncan3 used PEG to enhance adsorption
with ficin-treated autologous RBCs. Subsequently, Miller et
al.4 compared adsorptions of ficin-treated RBCs with and
without the addition of PEG (though they did not specify
whether the RBCs were autologous or allogeneic), and
Ylagan et al.5 compared adsorptions of 10 sera with ficin-
treated allogeneic RBCs to adsorptions with untreated RBCs
in the presence of PEG. Barron and Brown6 recently re-
termine if underlying alloantibodies are also
ABBREVIATIONS: AIHA = autoimmune hemolytic anemia; LISS
= low-ionic-strength saline (solution); PEG = polyethylene gly-
col; RBC(s) = red cell(s).
From the American Red Cross Blood Services, Southern Califor-
nia Region, Los Angeles, California.
Received for publication.May 30, 1998; revision received
June 29,1998, and accepted July 15, 1998.
TRANSFUSION 1999;39: 1 1 - 16.
Volume 39, Januaty 1999 TRANSFUSION 11

Page 2

LEGER AND GARRATIT
strength medium and are used at a
different PEG:serum ratio (1:l for
commercial reagents rather than 2:l
ported their results from comparing adsorptions of 19 sera
with ficin-treated allogeneic RBCs to adsorptions with un-
treated allogeneic RBCs in the presence of PEG. We com-
pared the 1-in-5 dilution technique (dilution technique)
and adsorption with untreated allogeneic RBCs in the pres-
ence of PEG (PEG adsorption) to our routine adsorption
procedure using ZZAP-treated allogeneic RBCs (ZZAP ad-
sorption).8
Nonreactive
20 (22%)
7 (27%)
identiied '
reactive
73 (78%)
14 (54%)
Warm autoantibodies only
Warm autoantibodies + alloantibodies
5 (19%)
MATERIALS AND METHODS
A 12-month audit of reference laboratory cases was per-
formed to determine the number of samples that contained
warm autoantibodies that reacted in the low-ionic-strength
saline solution (LISS) or PEG antiglobulin test and the num-
ber of these samples that also contained alloantibodies.
Subsequent samples from repeat patients within the 12-
month period were noted.
Serologic techniques were performed as previously
de~cribed.~ Reagent RBCs were obtained from various
manufacturers (Gamma Biologicals, Houston, TX;
Immucor, Norcross, GA; Organon Teknika Corp., Durham,
NC; Ortho Diagnostics Systems, Raritan, NJ). LISS was pre-
pared according to the method of Low and Messeter.'O The
20-percent PEG was prepared as previously reported";
commercial PEG reagents were obtained from Gamma
(Gamma PeG), OrganonTeknika (PEG+), and GTI (GTI PER
Brookfield, WI). Anti-IgG (American National Red Cross,
Washington, DC) was used for all antiglobulin tests.
Dilution technique
Sera containing warm autoantibodies, undiluted and di-
luted 1-in-5 in pH 7.4 phosphate-buffered saline, were
tested against LISS-suspended RBCs.
ZZAP adsorption
ZZAP reagent was prepared from dithiothreitol (Sigma
Chemical Co., St. Louis, MO) and ficin (Sigma), and RBCs
were treated for 30 mintues as previously described.' Equal
volumes of packed ZZAP-treated allogeneic RBCs and se-
rum were incubated at 37°C for 30 minutes. After centrifu-
gation, 2 drops of the harvested adsorbed serum were
added to 2 drops of a 2-percent LISS suspension of RBCs
and incubated for 15 minutes at 37°C before the antiglobu-
lin test. Subsequent adsorptions with additional aliquots of
ZZAP-treated RBCs were performed as needed.
PEG adsorption
Unlike the 20-percent PEG prepared
in our laboratory, the three commer-
cia1 PEG reagents are in a low-ionic-
for 20% PEG) for serologic testing. In our preliminary stud-
ies, we found that, if we used 2 volumes of 20-percent PEG
for the adsorption, as in serologic tests, the warm autoan-
tibody reactivity remaining in the adsorbed sera appeared
to be enhanced. Thus, 1 volume of PEG was used for the
study.
Equal volumes of untreated allogeneic RBCs, serum,
and PEG were incubated at 37°C for 15 minutes. Four drops
of the harvested serum-PEG mixture (to account for the
dilution of the serum by the PEG) were added to 1 drop of
3- to 5-percent RBCs and incubated for 15 minutes at 37°C
before the antiglobulin test. When subsequent adsorptions
were needed to remove antibody, additional PEG was not
added.
RESULTS
In one year, our reference laboratory evaluated 694 samples
with warm autoantibodies demonstrable by a routine
screening procedure using the PEG antiglobulin test; 457
(66%) of these sera reacted in the LISS antiglobulin test and
required adsorptions with ZZAP- or ficin-treated RBCs to
detect underlying alloantibodies. Forty-seven percent (194/
41 1) of sera containing PEG-reactive warm autoantibodies
and 40 percent (105/263) of sera with LISS-reactive warm
autoantibodies, not including repeat samples in the 12-
month period, contained alloantibodies. Previous studies
reported alloantibody detection rates of 15 to 41 per~ent'~,'~
after adsorptions with allogeneic RBCs, 27 to 40 percentI4-I6
after adsorptions with autologous RBCs, and 14 percent
after adsorptions with a combination of autologous (most
cases) and allogeneic RBCs"; 92 percent of our adsorptions
used allogeneic RBCs.
Dilution technique
A total of 119 sera were retested at a 1-in-5 dilution; when
tested without dilution, 108 of these showed reactivity of the
same strength with all cells tested ("panagglutinin"), while
11 sera demonstrated variable reactivity from cell to cell.
Results for the diluted sera are shown in Table 1. Initial ad-
sorptions with ZZAP-treated RBCs showed that underlying
alloantibodies were present in 26 of the 119 sera (see Table
2 for specificities). Alloantibodies were not detected in 7 of
these 26 sera. Although alloantibodies were identified in
five of the 26 sera when tested at a 1-in-5 dilution, all five
of these antibodies had shown variable reactivity when the
sera were tested without dilution. Thus, the presence of one
or more antibodies was already suspected in those 5 sera.
Antibodv(ies)
All cells
I
TABLE 1. Results after 1-in-5 dilution of 119 sera
12 TRANSFUSION Volume 39, January 1999

Page 3

REMOVAL OF WARM AUTOANTIBODIES
~ ~~ ~~
TABLE 2. Alloantibody specificities underlying warm
autoantibodies tested by the 1-in-5 dilution
techniaue’
Antibodies
Not detected
Identified
D
D
E
E (2)
e
K
FYa
Jkb
J ka
Jkb (2)
Suggested
D
D+C (2)
e
Masked
C
E (3)
C+e
E+c
S
Jka (2)
Kn/McC
Alloantibody specificities were initially determined after adsorp-
tions with ZAP-treated allogeneic RBCs.
Warm autoantibodies continued to mask the presence
of alloantibodies in the 1-in-5 dilutions of the remaining 14
sera, although, in 4 sera, the presence of anti-D, -D+C, and
-e was suggested (i.e., reactivitywith antigen-negative RBCs
was weaker than that with the respective antigen-positive
RBCs). Most of the sera, with or without underlying alloan-
tibodies that were negative when tested at a 1-in-5 dilution,
reacted only 1+ (a few reacted 2+) when tested without di-
lution; the majority of the sera that still reacted after dilu-
tion at 1 in 5 were reactive 2+ to 4+ when tested without di-
lution.
PEG adsorption
Parallel PEG and ZZAP allogeneic adsorptions were per-
formed on 28 sera, 8 ofwhich contained alloantibodies. For
the 28 sera, a total of 36 RBC aliquots and 9 hours were re-
quired for the PEG adsorptions, compared to 37 aliquots
and 18.5 hours for the ZZAP adsorptions. For 23 of the 28
sera, an equal number of PEG and ZZAP adsorptions were
required; 3 sera needed one less PEG adsorption, and 2 sera
needed one more PEG adsorption. All alloantibodies de-
tected (Table 3) reacted in the serum-PEG mixture at a
strength equal to or greater than that in the ZZAP-adsorbed
serum, presumably because of enhancement by the PEG
that remained in the adsorbed serum. One previously iden-
tified, very weak anti-E did not react after either PEG or
ZZAP adsorption.
TABLE 3. Alloantlbody reactivity in adsorbed sera
Anti-
D
D
E
E
E
C
C
PEG-adsorbed
2+
2+
3%+
2%+
3+
4+
2%+
3+
1+
1+
ZZAP-adsorbed
2+
2+
3+
2+
3+
4+
2+
3+
1+
Y 2 +
Reactivity
K
J ka
Jkb
Additional parallel adsorptions of 10 sera were per-
formed “blind.” A total of 15 PEG adsorptions (3.75 hours)
were required, compared to 20 ZZAP adsorptions (10
hours). Identical results were obtained with both methods,
except for one autoanti-e that was identified in the PEG-
adsorbed serum but not in the ZZAP-adsorbed serum. An-
tibody specificities included in this group were anti-D, -C,
-E, -K, -Fp, -Jka, and autoanti-e.
Table 4 shows a comparison of ZZAP and PEG adsorp-
tions, using different sources of PEG, of a serum contain-
ing relative autoanti-E and autoanti-c specificities plus a
panagglutinin. Four adsorptions with R,R, RBCs were re-
quired for each reagent to remove the panagglutinin. Upon
testing of the adsorbed sera with E+ or c+ RBCs, one of the
commercial PEG reagents (Gamma PeG) appeared to ad-
sorb the autoantibodies (panagglutinin plus relative
autoanti-E and autoanti-c specificities) more efficiently
than the other PEG reagents. The autoanti-E and autoanti-
c were still detected in the other three PEG-adsorbed sera
and ZZAP-adsorbed serum. If the more complete adsorp-
tion of all the autoantibody with the Gamma PeG was due
to the low-ionic-strength medium, we would have expected
the other two commercial reagents to have performed simi-
larly. None of the reagent inserts indicate the concentration
of PEG or the final ionic strength of the reagent.
4°C adsorption with PEG for removal of cold
autoant i body
A mixture of a strong autoanti-I and alloanti-E, after three
15-minute adsorptions with PEG at 4°C or 37”C, reacted 1+
and 4+, respectively, with E- RBCs at room temperature. At
antiglobulin test, the 4°C adsorbed serum did not react with
E- RBCs and reacted 1%+ with E+ RBCs; the 37°C adsorbed
serum agglutinated the E- RBCs (1+) and E+ RBCs (3+) be-
fore the addition of anti-IgG and reacted %+ and 1%+, re-
spectively, after the addition of anti-IgG. Thus, the 4°C PEG
adsorption was more efficient than the 37°C PEG adsorp-
tion in removing the cold antibody, and the anti-E was still
demonstrable.
In a comparison of PEG and ZZAP adsorptions at 4°C
(15 min each), three PEG adsorptions and five ZZAP adsorp-
tions were required to remove a strong autoanti-I (4°C ti-
TABLE 4. Reactivity of autoanti-E and autoanti-c plus
panagglutinin after adsorption (x4) with ZAP-treated
RIRl RBCs versus untreated R,R, RBCs in the
presence of four sources of PEG
Test
20%
RBCs ZZAP PEG
0 0
1+
1+
1+ 1 Y2+
1+
1 %+
Reactivity observed microscopically only.
Gamma
PeG
0
0
0
0
GTI
PEP
0
1+
1+
1+
PEG+
0
%+
rr
R,R,
Y2 +
RlR,
RZRl
Volume 39, January 1999 TRANSFUSION 13

Page 4

LEGER AND GARRATTY
ter, 2048). Thus, PEG was more efficient than ZZAP in re-
moving this high-titer cold antibody.
Is weak antibody activity lost through adsorption
with PEG?
We observed that, after one or two adsorptions in the pres-
ence of PEG (20% [prepared in our laboratory] or commer-
cial reagent), plasma and sera became clear, both in
comparison to unadsorbed plasma and sera and to
ZZAP-adsorbed plasma and sera. In addition, it was noted
that a precipitate formed in serum-PEG mixtures after stor-
age at 4°C. Several experiments attempted to determine if
these phenomena represented loss of specific antibody re-
activity after adsorption with PEG.
A weak anti-Jk” (1+ by LISS antiglobulin test) was used
for sham adsorptions with three aliquots of Jk(a-) RBCs in
the presence of 20-percent PEG or Gamma PeG. The anti-
Jk“ was 1+ reactive with Jk(a+b-) RBCs after each of the three
adsorptions with either of the PEG reagents.
The same weak anti-Jk” diluted with an equal volume
of a warm autoantibody and adsorbed twice with PEG- or
ZZAP-treated Jk(a-) RBCs reacted 1+ with Jk(a+b-) RBCs in
both adsorbed sera after removal of the autoantibody. An
anti-E reacted 1 %+ in the serum-PEG mixture after each of
two sham adsorptions with E- RBCs in the presence of 20-
percent PEG; before the adsorption, the anti-E reacted 1 %+
with and without the addition of an equivalent volume (2
drops) of PEG. The anti-E-PEG mixture from the second
sham adsorption still reacted (1%+) after storage at 4°C
overnight (and centrifugation to remove the precipitate). An
equal volume of PEG was added to master dilutions of an
anti-E. Aliquots of the anti-E-PEG mixtures were tested
immediately and after 24 hours’ storage at 4°C. The stored
dilutions were centrifuged to remove the precipitate that
had formed. The anti-E titer was 8 before and after 4°C stor-
age.
DISCUSSION
The alloimmunization rate (40%) we observed in our pa-
tients with LISS-reactive warm autoantibodies is higher
than that in previous reports, especially those in which
adsorptions were performed primarily with allogeneic as
opposed to autologous RBCS.’~-” Wallhermfechtel et al.”
and Issitt et al.,I3 who performed adsorptions with alloge-
neic RBCs, reported detection of alloantibodies in 15 per-
cent (19/125) and 41 percent (14/34) of sera, respectively.
In the studies of adsorptions with autologous RBCs, Issitt
et al.,I3 James et al.,14 Laine and Beattie,I5 and Morel et a1.I6
detected alloantibodies in 27 percent (1 1 /41), 32 percent
(13/41), 38 percent (41/109), and 40 percent (8/20) of sera,
respectively. Sokol et al.” detected alloantibodies in 14 per-
cent (294/2149) of their patients, but that series included
samples from patients with warm- and cold-reactive au-
toantibodies and did not distinguish how many patients
with warm autoantibodies had alloantibodies. In addition,
Sokol et al.” state that they issued K-, Rh phenotype-
matched blood to decrease the incidence of alloimmuni-
zation. Issitt et al.13 found that 69 and 19 percent of the
suspected alloantibodies detected after adsorptions with
autologous and allogeneic RBCs, respectively, had autoan-
tibodies that mimicked alloantibodies. Although adsorp-
tions with autologous RBCs were not separated out in our
audit, our data primarily (92%) represent adsorptions with
allogeneic RBCs. Adsorptions with autologous RBCs were
not the focus of this investigation. Although most serolo-
gists would treat sensitized autologous RBCs (e.g., with
enzyme, ZW or chloroquine) to remove immunoglobu-
lins before adsorption, Combs and Issitt, as sited in Issitt
and Anstee,18 used untreated autologous RBCs for PEG ad-
sorptions without loss of efficiency.
We should emphasize that we studied more patients
thandidother investigators (i.e., 263 vs. 20-138),12-16
ing Sokol et al.,” for the reason stated above. The patient
population that is referred to our reference laboratory is
from a broad geographic base but may be biased toward
patients requiring transfusion and/or patients with stron-
ger antibodies. Another major factor is the number of trans-
fusions the patients have received previously. James et aI.l4
andwallhermfechtel et a1.12 found that 75 percent (3/4) and
32 percent (61 19), respectively, of patients who had received
>5 transfusions had alloantibodies present. Such a high rate
of alloimmunization in patients with warm autoantibodies
underscores the need for efficient techniques to evaluate
all sera from patients with autoimmune hemolytic anemia
(AIHA).’9
Although the 1-in-5 dilution technique, as described by
Byen and Angeles,2 would appear to be a convenient ap-
proach to screening for underlying alloantibodies, it proved
to be inefficient, compared to ZZAP and PEG adsorbtions,
in our hands. Only 22 percent (20/93) of sera containing
warm autoantibodies without alloantibodies were
nonreactive after dilution to 1-in-5, in contrast to the 42 per-
cent (501 119) of nonreactive samples reported by0yen and
Angeles.* More important, we did not detect 7 (27%) of 26
alloantibodies that were of potential clinical significance.
The dilution technique as used by Byen and Angeles is ap-
plied, by them, to study sera that show ~ 1 +
activity with all cells. Of the seven samples we tested con-
taining alloantibodies that were not detected after dilution,
only one (anti-D) showed variable reactivity (1-2+) when
tested without dilution; the other six met the criteria set
forth by Byen and Angeles. As discussed by Petz and
Garratty7 a dilution method might be appropriate if the
patient cannot wait for the length of time it takes for adsorp-
tions to be performed. Results from adsorption procedures
may take 4 to 8 hours, especially if the sample has to be sent
to a reference laboratory. Using the simple dilution ap-
exclud-
equivalent re-
14 TRANSFUSION Volume 39, January 1999

Page 5

REMOVAL OF WARM AUTOANTIBODIES
proach is better than transfusing incompatible units, as
approximately 35 percent of alloantibodies were detected
using this approach. If the dilution technique is used, it may
be appropriate to perform adsorption studies retrospec-
tively to detect any alloantibodies that may have been
missed by the dilution approach.
Adsorptions with untreated allogeneic RBCs in the
presence of PEG appeared to give similar results to adsorp-
tions with =-treated allogeneic RBCs. We were able to
detect the same alloantibody specificities after adsorption
with untreated RBCs in the presence of PEG that were found
in parallel ZZAP adsorptions. PEG adsorptions at 4°C ap-
peared to be more efficient than those with ZZAP at remov-
ing cold autoantibody in one serum tested, but more sera
need to be tested to compare this apparent efficiency. Be-
cause untreated RBCs do not pack as efficiently as enzyme-
or ZZAP-treated RBCs, care must be taken to adequately
pack the adsorbing cells before adsorption so that the se-
rum-PEG mixture is not diluted with excess saline. In ad-
dition, although we were unable to demonstrate a loss of
specific antibody reactivity due to of the precipitation of
protein that occurswith PEG, the potential still exists; test-
ing of the serum-PEG mixture at the time of adsorption
(and not after storage at 4°C) may decrease the likelihood
of missing antibodies. In the recent report by Barron and
Brown,‘j one anti-K and one anti-Jkh that were weakly and
microscopically reactive, respectively, in LISS antiglobulin
test after adsorption with papain-treated allogeneic RBCs,
were not detected after adsorption in the presence of PEG.
All the methods published to date for adsorption in the
presence of PEG, including our own, have varied somewhat
in technique or technique comparison (e.g., 30- to 60-min
adsorption times, comparison to ficin or ZZAP adsorption,
source of PEG).
The primary disadvantage of using PEG rather than
ZZAP adsorptions is that there is no visible clue as to the
efficiency of the adsorption, such as the agglutination that
is seen with ficin- or ZZAP-treated cells that indicates anti-
bodies have adsorbed onto the cells. The second disadvan-
tage is that, because ficin- or dithiothreitol-sensitive anti-
gens will not be destroyed, the interpretation of results is
affected, and alloantibodies to antigens of high-frequency,
such as anti-Kph and anti-LW, will be adsorbed (although it
must be remembered that antibodies to other high-fre-
quency antigens can be adsorbed even with ZZAP-treated
RBCs). The third disadvantage is that, although PEG ap-
pears to be more efficient in adsorbing autoantibodies,
warm andlor cold autoantibody reactivity may be enhanced
in some cases by the PEG remaining in the adsorbed sera.
The major advantage in using PEG adsorptions rather
than ZZAP is that no treatment of allogeneic RBCs is re-
quired before PEG adsorption, which saves reagents (e.g.,
enzyme, ZZAP) and time (30-45 additional minutes for
treatment and washing). The second advantage is a 150-
percent reduction in the time required for adsorptions (15
vs. 30 min) when ZZAP adsorption is performed as recom-
mended in the original method and the Technical Manual,y
of the American Association of Blood Banks, which results
in a shorter turnaround time for serologic resolution. We did
not test a shorter adsorption time (e.g., 15 min) with ZZAP-
treated RBCs, so the efficacy of the two adsorption meth-
ods may be similar. The third advantage is that weak alloan-
tibodies, such as anti-Jkaand anti-Jkb, may be enhanced by
the PEG that remains in the adsorbed sera.
Finally, our large series emphasizes that patients with
AlHA become alloimmunized much more commonly than
other patients (i.e., 40% vs. 25% of sickle cell disease pa-
tients, 5% of thalassemia patients, 5% of multiply transfused
patients*O). IfAIHA patients are to receive the same protec-
tion as other patients, then we believe efficient procedures
for detecting “hidden” alloantibodies, although labor-in-
tensive, must be used in pretransfusion testing, and if
American Association of Blood Banks standards are to be
satisfied, then these procedures must be repeated every 3
days after transfusion.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Branch DR, Petz LD. A new reagent (ZZAP) having multiple
applications in immunohematology. Am J Clin Pathol
1982;78:161-7.
0yen R, Angeles ML. A simple screening method to evalu-
ate the presence of alloantibodies with concomitant warm
autoantibodies. Immunohematology 1995; 11235-7.
Liew YW, Duncan N. Polyethylene glycol in autoadsorption
of serum for detection of alloantibodies (letter). Transfu-
sion 1995;35:713.
Miller R, Unger P, Shapiro A. Rapid adsorption of warm au-
toantibodies using polyethylene glycol (abstract). Transfu-
sion 1996;36(Suppl):2OS.
Ylagan ES, O’Gorman F, Roller V, Moon M. Differential ad-
sorptions using polyethylene glycol (PeG) (abstract). Trans-
fusion 1996;36(Suppl):2OS.
Barron CL, Brown MB. The use of polyethylene glycol (PEG)
to enhance the adsorption of autoantibodies. Immunohe-
rnatology 1997;13:119-22.
Petz LD, Garratty G. Acquired immune hemolytic anemias.
New York: Churchill Livingstone, 1980.
Leger R, Garratty G. Evaluation of methods for detecting
alloantibodies underlying warm autoantibodies (abstract).
Transfusion 1997;37(Suppl):65S.
Vengelen-Tyler V, ed. Technical manual. 12th ed. Bethesda:
American Association of Blood Banks, 1996.
Low B, Messeter L. Antiglobulin test in low-ionic strength
salt solution for rapid antibody screening and cross-match-
ing. Vox Sang 1974;26:53-61.
Volume 39, January 1999 TRANSFUSION 15