I M M U N O H E M AT O L O G Y
The role of the elution in antibody investigations
Mark H. Yazer and Darrell J. Triulzi
BACKGROUND: The direct antiglobulin test (DAT)
commonly detects immunoglobulin G (IgG) molecules
or complement fragments on the red blood cell (RBC)
surface. If IgG antibodies are present then elution pro-
cedures can be performed to identify the specificity of
these antibodies. Our reference laboratory performs
elutions on the RBCs of those patients who have
received cellular blood products in the past 30 days
and have either a newly identified positive DAT with
anti-IgG or the agglutination strength is increased over
a previous DAT and if ordered by a clinician regard-
less of transfusion history. This study questioned how
frequently elutions contributed novel serologic informa-
tion under our reference laboratory’s current policy or
whether elutions should be performed in more selec-
tive serologic conditions.
STUDY DESIGN AND METHODS: Recipients whose
RBCs underwent eluate testing were identified from
the blood bank’s database and information about the
antecedent DAT and antibody detection test and
eluate was recorded.
RESULTS: In total 648 eluates were evaluated and 82
of 648 (12.7%) revealed a novel antibody not present
in the serum (an informative eluate). In 2 of 82 infor-
mative eluates non–anti-A/B alloantibodies that were
not present in the serum were detected: one example
each of anti-D and anti-E. Both were associated with
a microscopically positive antecedent DAT. The rate of
an informative eluate was higher when the antibody
detection test was negative.
CONCLUSION: The strength of the DAT does not
indicate the likelihood of an informative eluate. Per-
forming an eluate when the antibody detection test is
positive has limited value.
_02304 2395..2399
T
he direct antiglobulin test (DAT) is performed
in the evaluation of patients with suspected
immune hemolysis. A positive DAT with anti-
immunoglobulin G (IgG) indicates that IgG
antibodies are bound to the red blood cell (RBC). These
antibodies can be eluted from the RBCs and their speci-
ficity determined. Identifying an alloantibody in the
eluate indicates that the recipient has become sensitized
to an alloantigen on transfused RBCs and might be expe-
riencing a delayed hemolytic transfusion reaction. If the
recovered antibody agglutinates all reagent RBCs then the
patient has likely produced an autoantibody and might be
experiencing hemolysis secondary to this antibody (warm
autoimmune hemolysis). In both cases, the antibody
recovered in the eluate might also be detectable in the
recipient’s serum; if so then the eluate would not contrib-
ute additional serologic information. A nonreactive eluate
can indicate the presence of drug-dependent antibodies,
nonimmunologic uptake of IgG onto the RBCs, or rarely
an antibody to a low-prevalence antigen not present on
the panel cells. In our reference laboratory a DAT is per-
formed on all specimens referred for antibody investiga-
tion, as part of a serologic investigation of a transfusion
reaction or if one is ordered by a clinician. The indications
for performing an elution include the identification of a
newly positive DAT with anti-IgG of any agglutination
strength or if the strength of the DAT with anti-IgG was
increased compared to a previous DAT and the patient
had received cellular blood products in the preceding 30
days. These criteria are also applied during the serologic
evaluation of transfusion reactions. Elutions are also per-
formed when ordered by a clinician in the course of diag-
nosing immune-mediated hemolysis. In light of evidence
that between 47%1 and 91%2 of DAT-positive RBCs can
From the Department of Pathology, University of Pittsburgh,
and The Institute for Transfusion Medicine, Pittsburgh,
Pennsylvania.
Address reprint requests to: Mark Yazer, MD, 3636 Boulevard
of the Allies, Pittsburgh, PA 15213; e-mail: myazer@itxm.org.
Received for publication February 21, 2009; revision
received April 17, 2009; and accepted May 6, 2009.
doi: 10.1111/j.1537-2995.2009.02304.x
TRANSFUSION 2009;49:2395-2399.
Volume 49, November 2009 TRANSFUSION 2395
YAZER AND TRIULZI

feature a nonreactive eluate, and that eluates prepared
from microscopically or weakly positive DATs are typically
nonreactive,3 we questioned how frequently eluates con-
tributed novel serologic information under our reference
laboratory’s current policy or whether they should be per-
formed in more selective serologic conditions.
or a warm autoantibody, and the recipient’s serum may or
may not have contained other antibodies at the time the
elution was performed. The significance of the differences
between dichotomous variables was assessed by a two-
tailed Fisher’s exact test (GraphPad software). This study
protocol was approved by the University of Pittsburgh’s
Quality Improvement institutional review board.

MATERIALS AND METHODS

The database of the Centralized Transfusion Service in
RESULTS
Pittsburgh, Pennsylvania, was reviewed to identify all A total of 648 eluates were analyzed during the two time
recipients on whose RBCs elutions had been performed periods. Table 1A demonstrates the types of antibodies
during a nonconsecutive 11.5-month period. The strength present in the recipient’s serum at the time the elution
of the antecedent DAT with anti-IgG (negative, micro- was performed. The type of antibody(-ies) recovered in
scopic, weak, 1+ through 4+) that prompted the elution the eluate is shown in Table 1B. There were in total 90
was recorded along with the specificity of any antibodies non-anti-A/B alloantibodies recovered among the 648
detected in the ensuing eluate and/or in the serum at that eluates.
time. Elutions performed on neonates or on recipients of Table 2 demonstrates the agglutination strengths of
Rh immune globulin were not included in this analysis. the 648 antecedent DATs along with the corresponding
Although a small number of recipients had multiple elu- number of times that an informative eluate was present.
tions performed, each elution was treated as a distinct There were 16 cases in which an eluate was prepared
event for statistical purposes. despite the fact that the DAT was negative. Although the
The plasma from all recipients underwent a two-cell indication for performing the elutions in these cases was
antibody detection test (Immucor/Gamma, Norcross, GA) not documented, it is likely that they were performed due
and if positive, antibody identification was generally per- to a clinical suspicion of hemolysis. Ten of the 16 elutions
formed using the low-ionic-strength saline (Immucor/ were nonreactive, in one case anti-Jka was present in both
Gamma) or polyethylene glycol (Immucor/Gamma) tube the eluate and the plasma, in three cases an antibody that
techniques and a panel of reagent RBCs (Immucor/ agglutinated all reagent RBCs in both the eluate and the
Gamma), although if the recipient had a history of an anti- serum (microscopically) was identified, and in two cases
body that was not a warm autoimmune antibody or a an antibody that agglutinated all RBCs was present in the
high-titer, low-avidity antibody and a sufficient quantity eluate only.
of their plasma was available then the identification was In total there were 82 of 648 eluates that were infor-
made using an automated solid-phase system (Galileo, mative for an antibody which was not detected in the
Immucor). The DAT was performed in tube by first using a serum at the time the elution was performed (12.7%,
polyspecific reagent (Immucor/Gamma) and if positive,
monospecific anti-IgG (Ortho Diagnostics, Raritan, NJ)
TABLE 1. Types of antibodies in the recipient’s
and anti-C3 (Immucor/Gamma) reagents were subse- serum at the time the eluate was performed and
quently employed. Elutions were performed using the types of antibodies recovered in the eluates
rapid acid kit (Gamma) and the antibody’s specificity was Percentage
determined using the same reagent RBC panel used for Antibody(-ies) Number of total
testing the serum. Reagent A1 and B cells (Immucor/ A. In serum
Alloantibody 219 33.7
Gamma) were routinely included amongst the eluate Warm autoantibody 180 27.7
panel cells. All tests were performed in accordance with Warm autoantibody plus 138 21.2
the manufacturer’s directions. alloantibody
No antibodies (antibody detection 111 17.1
To diagnose a delayed hemolytic transfusion reaction, test negative)*
the recipient’s biochemical and/or clinical markers of Total 648
hemolysis, potentially also including a decrease in hemo- B. Recovered in eluate
globin concentration relative to the posttransfusion level Warm autoantibody 284 43.8
Nonreactive 244 37.7
if available, would have become positive days to weeks
Alloantibody† 80 12.4
after receiving a cellular blood product and a new anti- Anti-A or anti-B 30 4.6
body detected in their eluate and/or serum. An informa- Warm autoantibody plus 10 1.5
alloantibody†
tive eluate was defined as one in which an antibody was
Total 648
detected in the eluate but not in the serum at the time the
* Includes the 30 anti-A and anti-B detected only in the eluate.
elution was performed. The recovered antibody could † Refers to any alloantibody other than anti-A or anti-B.
have been an alloantibody (including anti-A and anti-B)

2396 TRANSFUSION Volume 49, November 2009

 THE ELUTION IN ANTIBODY INVESTIGATIONS

Table 2). The majority (18/30, 60.0%) of the 30 eluates

eluates for panagglutinating

informative for anti-A and anti-B were prepared from a

antibodies in the eluate

Number of informative
microscopically positive DAT with anti-IgG. Of the

2 (12.5)
remaining 52 informative eluates, six recovered an anti-

15 (6.8)
13 (8.2)
8 (8.4)
4 (5.2)
1 (2.1)
1 (2.9)
44/648 (6.8)
only (%)
body that agglutinated all reagent RBCs; a panagglutinin
was later identified in the serum of all six of these recipi-
ents on average 32 days (median, 15.5 days; range,
11-115 days) after it was first detected in the eluate. The
strengths of the antecedent DATs with anti-IgG in these six
recipients ranged from weak to 2+. There were 44 elutions
that recovered an antibody that agglutinated all the
Number of informative
eluates for non-A/B
alloantibodies (%)

reagent panel RBCs, yet none of these recipients devel-

2 (0.9)

2/648 (0.3)
oped a panagglutinin in their serum, and the antecedent
TABLE 2. Characteristics of the DAT strengths and the informative eluates in this study

DAT with anti-IgG strengths ranged from 0 to 4+.

0

0
0
0
0
0

The remaining two informative eluates revealed

non-A/B alloantibodies, one example each of anti-D and
anti-E. The recipient with the anti-E was lost to follow-up
immediately after the antibody was identified in the
eluate and as antibody detection test was negative at the
Number of informative
eluates for anti-A

time the anti-E was identified in the eluate, it is unknown

or anti-B (%)

18 (8.2)
8 (5.1)
1 (1.1)
2 (2.6)
1 (2.1)

30/648 (4.6)

if it ever appeared in the serum. In the other recipient,

anti-D appeared in the serum 29 days after it was iden-
0

tified in the eluate. A warm autoantibody and a high-

titer, low-avidity antibody were identifiable in this
recipient’s serum at the time the elution was performed.
Both of these recipients had microscopically positive
antecedent DATs with anti-IgG, and neither demon-
detectable in the serum (%)
eluates for panagglutinating

strated any reactivity with anti-C3. In the 88 other eluates

Number of informative

antibodies eventually

in which a non-A/B alloantibody was recovered, the

3 (1.9)
2 (2.1)
1 (1.3)

6/648 (0.9)

same antibody was concomitantly identifiable in the

serum.
0
0

0
0

There were two delayed hemolytic transfusion reac-

tions reported among the 90 patients whose eluates recov-
ered a non-A/B alloantibody; in one case the patient had
received 6 units of RBCs during surgery and approxi-
mately 10 days later anti-C, -E, and -K were eluted from
of informative

2 (12.5)
35 (15.9)
24 (15.2)

82/648 (12.7)
11 (11.6)
Total number

the recipient’s RBCs. The antecedent anti-IgG DAT

eluates (%)

7 (9.1)
2 (4.2)
1 (2.9)

strength was 3+ (anti-C3 was not tested on the day the

eluate was performed but was microscopically positive 3
days later). In the second case the patient had received 4
units of RBCs over 2 weeks before anti-Fya was eluted from
the RBCs. The antecedent DAT with both anti-IgG and
Total number

(% of total)
16 (2.50)
220 (34.0)
158 (24.4)
95 (14.7)
77 (11.9)
of eluates

anti-C3 were both microscopically positive. In both recipi-

48 (7.4)
34 (5.2)

ents the alloantibodies were identified in the serum and

648

eluate concomitantly.
Table 3 demonstrates the rate of detecting an infor-
mative eluate based on the results of the antibody detec-
Antecedent DAT strength

tion test performed at the same time as the elution.

Overall the rate of informative eluates was higher when
the antibody detection test was negative (p < 0.0001). In
approximately half of the cases the antibody detection test
Microscopic

was positive (45/82, 54.9%), and in these 45 informative

Total (%)

eluates with positive antibody detection tests, the recipi-

Weak

ents had different antibodies detected in their serum com-

1+
2+
3+
4+
0

pared to the antibody recovered in their eluate.

Volume 49, November 2009 TRANSFUSION 2397

YAZER AND TRIULZI

specimens in patients with recent exposure to RBCs but

TABLE 3. Rate of occurrence of informative with negative pretransfusion DATs, Johnston and Belota7
eluates based on the presence or absence of a
positive antibody detection test at the time the found only two (2.4%) eluates that revealed a new alloan-
eluate was performed tibody in the eluate which was not in the serum. Both of
Total number Number of these patients suffered a delayed hemolytic reaction. Fur-
Antibody of elutions informative Percentage thermore, in patients with positive pretransfusion DATs
detection test performed eluates of total elutions
and in those with no recent RBC exposure, no new anti-
Negative 111 37 33.3†
Positive* 537 45 8.4 bodies were identifiable in the eluate alone.7 Perkins and
Total 648 82 colleagues8 found seven (1.1%) confirmed alloantibodies
* Indicates positivity for any alloantibody or warm autoantibody.
present only in the eluate among the 684 specimens tested
† p < 0.0001. from recipients with positive autocontrols but negative
antibody detection tests. Similarly, Judd and coworkers9
found new alloantibodies in the eluate but not in the
serum in approximately 3% of 778 samples studied
TABLE 4. Overall rates of reactive and although not all were clinically significant. As the sensitiv-
nonreactive eluates stratified by the strength of
the antecedent DAT ity of routine pretransfusion testing for detecting antibod-
Total* ies in the serum has increased since the time that many of
DAT strength/eluate reactivity % these reports were published, some of the antibodies that
Microscopic DAT/reactive eluate 93 (42.3) were detected only in the eluate would probably have
Microscopic DAT/nonreactive eluate 127 (57.7) been concomitantly detected in the serum had they been
Weak DAT/reactive eluate 87 (55.1)
Weak DAT/nonreactive eluate 71 (44.9) studied using modern reagents thus potentially lowering
1-4+ DAT/reactive eluate 218 (85.8) the rate of discovering antibodies in the eluate alone.
1-4+ DAT/nonreactive eluate 36 (14.2) In this study, 82 of 648 (12.7%) eluates were informa-
* In addition to the eluates listed in this table there were 16 elu- tive insofar as an antibody that was not detectable in the
tions performed on samples in which the antecedent DATs
with anti-IgG did not demonstrate agglutination; 10 eluates plasma was recovered in the eluate. Our rate of informa-
were nonreactive and six were reactive. See text for statistical tive eluates might have been higher had we included
analysis. nonreactive eluates in the setting of drug-dependent
antibodies in the definition. We included in our definition
of an informative eluate those which revealed an anti-
Table 4 demonstrates the overall recovery rate of an body that agglutinated all reagent RBCs yet a panaggluti-
identifiable antibody (warm autoantibody, alloantibody, nin was never detected in the serum. Forty-four such
or anti-A/B) in the eluate, regardless of its presence in the antibodies were recovered from informative eluates in
serum, stratified by the strength of the antecedent DAT. this study and these antibodies might have been clini-
There was a significant difference in the antibody cally significant; perhaps the recipient was started on a
recovery rates in eluates produced from DATs with anti- course of steroids or other treatment once this type of
IgG that were between 1+ and 4+ compared to those antibody was discovered in the eluate thereby preventing
produced from DATs with anti-IgG that were both the appearance of a panagglutinin in the serum. Similarly
microscopically and weakly positive (p < 0.0001). The dif- the anti-A and anti-B recovered in the eluate might have
ference in the antibody recovery rate between eluates pro- been clinically significant in terms of causing hemolysis;
duced from weakly reactive DATs with anti-IgG compared these antibodies were recovered in more than half of the
to microscopically reactive DATs with anti-IgG was also informative eluates with a negative antibody detection
significant (p = 0.016). test (21/37, 56.8% data not shown). Because we did not
differentiate between recipients whose RBCs underwent
elutions as a result of suspected immune hemolysis
DISCUSSION
versus those in which immune hemolysis was not sus-
Several previous studies described a very low incidence of pected, it is possible that neither of these types of anti-
detecting a novel antibody in the eluate when it was not bodies were clinically significant. Thus if an alternative
detectable in the serum. In these studies the eluate was analysis is performed excluding the 30 anti-A/B and the
informative for an antibody not identifiable by the routine 44 agglutinating antibodies where a panagglutinin did
pretransfusion testing of the serum in less than 1% of all not develop in the serum, then the rate of informative
eluates performed.2,4-6 In one of these reports, the lone eluates in the presence of a positive or negative antibody
antibody detectable only in the eluate, anti-Cw, was only detection test is no longer significant (p = 0.29), and the
detected in the eluate because a Cw-positive cell was only rate of informative eluates prepared from a microscopi-
present in the eluate antibody identification panel.4 Of 84 cally positive DAT is reduced from 15.9% (35/220) to 0.9%
elutions performed on 220 DATs with anti-IgG–positive (2/220).

2398 TRANSFUSION Volume 49, November 2009


Only a few other studies have evaluated the utility of
performing an eluate based on the strength of the ante-
cedent DAT. In six eluates yielding alloantibodies not
present in the serum, Judd and colleagues9 found the reac-
tion strengths of the antecedent DATs to range from 1+ to
2+. Richa and coworkers3 found that for discovery of either
a new alloantibody or a warm autoantibody that was not
present in the serum, microscopically positive DATs
yielded significantly fewer new findings compared to
macroscopically positive DATs of any agglutination
strength. In total they detected 10 new alloantibodies and
18 autoantibodies among the 310 eluates performed
(9.0%), with only one new alloantibody detected in an
eluate from a microscopically positive antecedent DAT.3
Our rate of informative eluates was similar at 12.7%,
and overall we likewise found that eluates prepared from
microscopically positive DATs were more frequently non-
reactive compared to eluates prepared from either weakly
positive or 1+ to 4+ positive antecedent DATs (Table 4).
In some of the earlier studies, the rates of nonreactive
eluates ranged from 47% to 91%. Our rate of nonreactive
eluates was lower at 37.7% and this might relate to the
elution reagent itself; in the earlier studies a variety of
reagents were employed including dichloromethane,4,8
ether,2,5,9 xylene,6 ethanol,1 or various preparations using
acidic pH to effect the elution including some commer-
cially available kits.3,4,6-8 In our study we exclusively used a
commercially available rapid acid elution kit, which might
be more effective in recovering antibodies than the
organic compound based methods, hence our lower rate
of nonreactive eluates.
In light of our findings we recommend not routinely
performing elutions when the antibody detection test is
positive and eliminating the microscopic reading of tube
DATs as eluates prepared from such RBCs are usually non-
reactive, unless immune hemolysis is strongly suspected.
Adopting the latter criteria alone would have reduced our
elution workload by approximately 35%. Although both of
the eluates informative for non-A/B alloantibodies in this
study were prepared from microscopically positive ante-
cedent DATs, we do not feel that this low rate of informa-
tive eluates justifies the considerable time and expense of
preparing eluates from these RBCs.
THE ELUTION IN ANTIBODY INVESTIGATIONS
ACKNOWLEDGMENT
MHY thanks Dr Jill Storry for thoughtful discussion and critical
review of the manuscript.
CONFLICT OF INTEREST
Neither of the authors has any conflicts to disclose.
REFERENCES
1. Lau P, Haesler WE, Wurzel HA. Positive direct antiglobulin
reaction in a patient population. Am J Clin Pathol 1976;65:
368-75.
2. Judd WJ, Butch SH, Oberman HA, Steiner EA, Bauer RC.
The evaluation of a positive direct antiglobulin test in pre-
transfusion testing. Transfusion 1980. 20:17-23.
3. Richa E, Benidt G, Tauscher C, Stowers R, Bryant S, Stubbs
J. Eluate testing following microscopically positive direct
antiglobulin tests with anti-IgG. Ann Clin Lab Sci 2007;37:
167-9.
4. Huh YO, Liu FJ, Rogge K, Chakrabarty L, Lichtiger B. Posi-
tive direct antiglobulin test and high serum immunoglobu-
lin G values. Am J Clin Pathol 1988;90:197-200.
5. Stec N, Shirey RS, Smith B, Kickler TS, Ness PM. The effi-
cacy of performing red cell elution studies in the pretrans-
fusion testing of patients with positive direct antiglobulin
tests. Transfusion 1986;26:225-6.
6. Domen RE, Grattan J. Efficacy of performing red-cell anti-
body elutions in patients with a positive direct antiglobulin
test. Vox Sang 1986;51:324-6.
7. Johnston MF, Belota MK. Determination of need for
elution studies for positive direct antiglobulin tests in pre-
transfusion testing. Am J Clin Pathol 1988;90:58-62.
8. Perkins JT, Arruza M, Fong K, Sosler SD, Saporito C. The
relative utility of the autologous control and the antiglobu-
lin test phase of the crossmatch. Transfusion 1990;30:
503-7.
9. Judd WJ, Barnes BA, Steiner EA, Oberman HA, Averill DB,
Butch SH. The evaluation of a positive direct antiglobulin
test (autocontrol) in pretransfusion testing revisited. Trans-
fusion 1986;26:220-4.
Volume 49, November 2009 TRANSFUSION 2399