Immunopathology / MULTIPLEX EBV ANTIBODY ASSAY
Evaluation of a Multiplex Fluorescent Microsphere
Immunoassay for the Determination of Epstein-Barr Virus
Serologic Status
Thomas B. Martins, MS,1 Christine M. Litwin, MD,1,2 and Harry R. Hill, MD1-4
Key Words: Epstein-Barr virus; EBV; Multiplex assay; Antibody response; Recombinant antigens
DOI: 10.1309/65VKWVNAQ38PHMGQ
Abstract
Epstein-Barr virus (EBV), a human herpesvirus,
affects up to 95% of adults. Diagnosis of acute EBV
infection can be challenging and often relies on the
serologic antibody pattern to 3 distinct antigens, most
often determined by indirect fluorescent antibody (IFA),
enzyme-linked immunosorbent assays (ELISAs), and,
more recently, multiplex assays. We compared a
multiplex assay for the simultaneous detection of
antibodies to viral capsid (VCA), nuclear (EBNA), and
early (EA) EBV antigens with ELISAs using IFA for
discrepancy resolution. Concordance of the multiplex
assay was good for all 4 antigens: VCA IgM, 86.6% vs
ELISA and 92.9% vs IFA; VCA IgG, 92.8% vs ELISA
and 98.0% vs IFA; EBNA IgG, 90.3% vs ELISA and
98.1% vs IFA; and EA IgG, 83.8% vs ELISA and 92.8%
vs IFA. After IFA resolution, correlation between the
multiplex assay and ELISA for serologic disease stage,
based on the antibody profile of all 4 analytes, was
90%. The multiplex assay showed good correlation with
an established ELISA and even better correlation with
the “gold standard” IFA. Advantages of the multiplex
assay over traditional methods include multiple results
per assay, inclusion of internal controls for each assay,
and well-to-well monitoring of assay drift.
34 Am J Clin Pathol 2008;129:34-41
34 DOI: 10.1309/65VKWVNAQ38PHMGQ
Epstein-Barr virus (EBV) is a member of the
Herpesviridae family with morphologic and structural similar-
ities to other human herpesviruses. By adulthood, 95% to 99%
of most of the world’s population has demonstrable EBV anti-
bodies, with seroconversion occurring before the age of 5
years in 50% of children in the United States.1 EBV is the
causative agent of infectious mononucleosis (IM) and has
been implicated in other diseases, including Burkitt lym-
phoma and nasopharyngeal carcinoma and lymphoprolifera-
tive disorders in immunocompromised patients. The virus is
generally acquired by oral transmission, with the initial infec-
tion occurring in the epithelial cells of the oropharynx.2 The
virus then selectively infects B lymphocytes migrating
through the oropharyngeal lymphoid tissue, thus establishing
a lifelong latent infection in the peripheral blood and other
reticuloendothelial tissues.
The diagnosis of IM is based on clinical manifestations,
which generally include sore throat, fever, lymphadenopathy,
and malaise, in conjunction with laboratory test results.
Owing to the high prevalence of EBV in the population, direct
detection methods of the virus, such as culture and poly-
merase chain reaction, are time-consuming and expensive and
generally not practical. For these reasons, most laboratory
tests are based on the detection of antibodies to distinct EBV-
associated antigens.
Several EBV antigens have been classified based on the
phase of the viral replication cycle during which they are
expressed. These include the viral capsid antigen (VCA), the
EBV nuclear antigen (EBNA), and the early antigen (EA).
The EAs are further classified as diffuse or restricted based on
the different patterns of immunofluorescence staining exhibit-
ed by the 2 components.3,4
© American Society for Clinical Pathology

Antibody responses and their associated titers to specific
EBV antigens correlate with different stages of IM.1,5-7 IgM
antibodies to VCA are usually produced within 4 to 7 days
after primary EBV infection and peak by 3 to 4 weeks, after
which they decline rapidly and are usually undetectable after
12 weeks. VCA IgG antibodies are initially produced concur-
rently with IgM but decline slowly after peaking and last
indefinitely. Antibodies to EA appear transiently for up to 3
months during the acute phase of IM in 85% of patients.8,9
There is a moderate rise in levels of antibodies to EA during
reactivation, and they may remain elevated in chronic infec-
tion. The antibody response to EA in patients with IM is usu-
ally to the diffuse component, whereas silent seroconversion
to EBV in children produces antibodies to the restricted com-
ponent.1,6 Antibodies to EBNA, characteristically a late-onset
response, are produced within 4 to 6 weeks of exposure and,
like antibodies to VCA, persist indefinitely.6,7
In this study, we evaluated a multiplex Luminex-based
(Luminex, Austin, TX) assay for the simultaneous detection of
VCA IgG and IgM, EBNA IgG, and EA IgG antibodies.
Results from this multiplex assay were compared with those of
single-result enzyme-linked immunosorbent assays (ELISAs)
for the same 3 antigens. Discrepant results were resolved by an
indirect fluorescent antibody (IFA) assay, which is still consid-
ered the “gold standard” reference method.
Materials and Methods
Clinical Samples
This study included 158 serum samples submitted to our
reference laboratory for EBV antibody testing. Patient study
subjects consisted of 88 females ranging in age from younger
than 1 year to 88 years, with a mean age of 31.6 years, and 70
males with an age range of 1 to 69 years and a mean age of
28.4 years. We also included 20 samples containing potential-
ly interfering factors10 in this study: 5 each of IgM+ samples
for herpes simplex virus, cytomegalovirus, varicella zoster
virus, and Toxoplasma gondii. In addition, 23 samples con-
taining elevated concentrations of rheumatoid factor (RF) IgM
were included. All patient samples included in this study were
deidentified according to the University of Utah Institutional
Review Board–approved protocol (No. 7275) to meet the
Health Information Portability and Accountability Act patient
confidentiality guidelines.
EBV Assays
Single-antigen ELISAs for EBV VCA IgM, VCA IgG,
EBNA IgG, and diffuse EA IgG were purchased from DiaSorin
(Stillwater, MN). The multiplex AtheNA Multi-Lyte EBV Test
System and the IFA assays were provided free of charge by
Inverness Medical Professional Diagnostics (Princeton, NJ). All
© American Society for Clinical Pathology
Immunopathology / ORIGINAL ARTICLE
assays used in this study have received US Food and Drug
Administration clearance and are labeled for in vitro diagnostic
use. The antigens used in the ELISAs were primarily synthetic
peptides. A 56-amino-acid p18 protein containing immuno-
dominant epitopes of a major component of the VCA com-
plex11-13 was used for the VCA IgG and IgM assays, The EBNA
IgG assay uses a 59-amino-acid synthetic peptide containing
several domains of the EBNA-1. The Gly-Ala repeat was
excluded owing to its known cross-reactivity with autoantigens.
The diffuse EA consists of a 47-kd recombinant polypeptide14
expressed in Escherichia coli and then purified.
The antigens used in the multiplex assay were primarily
affinity purified from virally infected cell lines, with each anti-
gen coupled to microspheres containing specific amounts of 2
fluorescent fluorophobes.15 Affinity-purified EBV VCA gly-
coprotein 125 was used in the IgG and IgM assays. The EA
was also affinity purified with roughly equal parts of diffuse
and restricted coupled to the same microsphere set. A recom-
binant protein of the EBNA-1 domain was used in the multi-
plex assay. In addition to the 3 antigen-coupled microspheres,
5 additional microsphere sets were included for quality con-
trol purposes. One was designed to detect binding of nonspe-
cific and RF IgM antibodies in the patient sample, and the
other 4 bead sets were used for intra-assay calibration.
All assays were performed with strict adherence to the
manufacturers’ instructions as detailed in the product inserts.
For the ELISA IgG assays, patient samples were diluted 1:101
in the supplied serum diluent. We added 100 µL of the calibra-
tors and diluted sample and control samples to the antigen-
coated microtiter plates. This step was followed by incubation
at 37°C ± 2°C for 60 ± 5 minutes. After 4 wash cycles using
an automated washer, 100 µL of horseradish peroxidase–con-
jugated antihuman IgG was added to each well. The microtiter
plate was again incubated and washed as described before the
addition of 100 µL of tetramethylbenzidine/hydrogen perox-
ide chromogen/substrate to each well, followed by incubation
at room temperature (22°C-27°C) for 30 ± 2 minutes. This
reaction was then stopped by the addition of 200 µL of 0.4N
sulfuric acid, and the absorbance was read at 450/630 nm. The
IgM ELISA used IgM capture via a surface-bound antihuman
IgM (µ chain specific) monoclonal antibody. The patient VCA
IgM antibodies were then detected by VCA p18 peptide anti-
gen, which is linked to an anti-p18 monoclonal antibody con-
jugated to horseradish peroxidase.
For the multiplex assay, patient samples, control samples,
and calibrator were diluted 1:21 in sample diluent. Next, 50
µL of the microsphere suspension was dispensed into each
well of a 96-well filtration plate, to which 10 µL of the dilut-
ed samples, control samples, and calibrator were added. The
microtiter plate was then thoroughly mixed using an orbital
plate shaker and then incubated at room temperature (20°C-
25°C) for 30 ± 10 minutes without shaking. The filter plate was
Am J Clin Pathol 2008;129:34-41 35
35 DOI: 10.1309/65VKWVNAQ38PHMGQ 35
Martins et al / MULTIPLEX EBV ANTIBODY ASSAY
then washed 3 times by vacuum filtration, followed by the addi-
tion of 150 µL of goat antihuman IgG (Fc chain specific) phy-
coerythrin conjugate to each well. The microtiter plate was then
thoroughly mixed and incubated at room temperature for 30 ±
10 minutes. After the final incubation, the microtiter plate was
placed on the Luminex 100 instrument, in which the micro-
spheres were identified based on the fluorescent intensities of 2
internal fluorophores. The amount of patient IgG or IgM bound
to the EBV antigen-coupled microspheres was determined by
the intensity of the fluorescence generated by the antihuman
IgG or IgM phycoerythrin-conjugated reporter antibody. The
amount of fluorescence is proportional to the amount of patient
antibodies bound to the microsphere. By using Intra-Well
Calibration Technology, we used internal calibration bead sets
to convert raw fluorescence into arbitrary units.
Data Analysis
Reference ranges for determining semiquantitative patient
results based on the amount of antibody measured were taken
from the package inserts and were different for the 2 assays.
Reference ranges for the multiplex assay were as follows: less
than 100 U/mL, negative, indicating no detectable IgG or IgM
antibody to the particular marker; 100 to 120 U/mL, inconclu-
sive, with the recommendation of the testing of a second sam-
ple at a later date; and more than 120 U/mL, positive, indicat-
ing that the specimen is positive for IgG or IgM antibody to the
marker. Reference ranges for the ELISAs were slightly modi-
fied from those supplied by the manufacturer to include an
inconclusive range and were as follows: 0.90 index value (IV)
or less, negative, no significant level of detectable antibody;
0.91 to 0.99 IV, inconclusive, repeated testing in 10 to 14 days
may be helpful; and 1.00 IV or greater, positive, indicating the
presence of IgG or IgM antibodies to the marker.
To determine agreement, sensitivity, and specificity
between the multiplex assay and the individual ELISAs,
results were analyzed using 2 × 2 contingency tables.16 All
samples with discrepant results were repeated in duplicate on
both assays. IFA testing was then performed on samples with
discrepant results that were not resolved by repeated testing,
with this result used in calculations as the “IFA-resolved”
result. Samples with inconclusive results were not included in
the calculations of clinical agreement, sensitivity, and speci-
ficity. The serologic status of EBV infection was determined
by using the criteria outlined in ❚Table 1❚, which were derived
from the literature.1,5-9
Results
Comparison of Multiplex Assay With ELISA
Initial analysis of results compared the multiplex assay
with ELISA for each individual EBV marker ❚Table 2❚. The
36 Am J Clin Pathol 2008;129:34-41
36 DOI: 10.1309/65VKWVNAQ38PHMGQ
❚Table 1❚
Serologic Status Based on EBV Antibody Profiles1,5-9
Clinical Stage VCA IgM EA IgG VCA IgG EBNA IgG
Susceptible – – – –
Early + – – –
Early – – + –
Acute + – + –
Acute ± + + –
Recent (or ± + + +
subacute)
Past – – + +
EA, early antigen; EBV, Epstein-Barr virus; EBNA, EBV nuclear antigen; VCA, viral
capsid antigen; +, positive antibody response; –, negative antibody response; ±,
positive or negative antibody response.
multiplex assay and the ELISA were then compared individu-
ally with the IFA-resolved results ❚Table 3❚ and ❚Table 4❚,
respectively.
Agreement for EBV VCA IgM between the multiplex
and ELISA assay was an acceptable 86.6%. Sensitivity
between the 2 assays, however, was poor (56.3%). This was
due to a proportionally large number of samples (14) with
positive results by ELISA but negative results with the multi-
plex assay. When compared with the IFA result, sensitivities
improved for the ELISA (81.3%) and the multiplex assay
(71.9%). Specificity compared with the IFA result was excel-
lent for both assays, 96.7% and 98.4%, respectively.
The multiplex assay showed good agreement (92.8%)
and sensitivity (99.2%) for EBV VCA IgG when compared
with the ELISA assay, but specificity was poor (54.5%). The
low specificity was due to 10 samples with positive results by
the multiplex assay but negative results by ELISA. Compared
with the IFA-resolved result, the specificity was 85.7% for the
multiplex assay and 100.0% for the ELISA assay. The multi-
plex assay demonstrated greater sensitivity (99.3%) than the
ELISA (94.9%) compared with the IFA result.
Comparing the 2 assays for EBNA IgG antibody responses
resulted in an agreement of 90.3%, sensitivity of 100.0%, and
specificity of 73.2%. When compared with the IFA result, the
multiplex assay had greater sensitivity than the ELISA (100.0%
vs 89.2%) but lower specificity (93.2% vs 100.0%), respectively.
Overall agreement between the multiplex and ELISA
assays for EA IgG was 83.8%. There were 17 samples with
positive results by ELISA but negative by the multiplex assay,
yielding a sensitivity of 73.0%. Agreement compared with the
IFA result was similar between the multiplex assay (92.8%)
and ELISA (89.9%), but the specificity vs the IFA result was
greater with the multiplex assay (96.3%) compared with
ELISA vs the IFA assay (88.6%).
Evaluation of Serologic Status
The evaluation of a patient’s EBV status is usually not
based on results of an individual marker but on clinical
© American Society for Clinical Pathology
 Immunopathology / ORIGINAL ARTICLE

❚Table 2❚
Correlation of the Multiplex Assay and ELISA Results*

ELISA Result

Multiplex Positive Negative Inconclusive Agreement (%) Sensitivity (%) Specificity (%)

VCA IgM 86.6 56.3 94.4

Positive 18 7 0
Negative 14 118 1
VCA IgG 92.8 99.2 54.5
Positive 130 10 1
Negative 1 12 0
Inconclusive 0 4 0
EBNA IgG 90.3 100.0 73.2
Positive 99 15 2
Negative 0 41 0
Inconclusive 0 1 0
EA IgG 83.8 73.0 93.2
Positive 46 5 2
Negative 17 68 3
Inconclusive 5 11 1

EA, early antigen; EBNA, Epstein-Barr virus nuclear antigen; ELISA, enzyme-linked immunosorbent assay; VCA, viral capsid antigen.
* Data are given as number of samples unless otherwise indicated.

❚Table 3❚
Correlation of the Multiplex Assay and IFA-Resolved Results*

IFA-Resolved Result

Multiplex Positive Negative Inconclusive Agreement (%) Sensitivity (%) Specificity (%)

VCA IgM 92.9 71.9 98.4

Positive 23 2 0
Negative 9 121 3
VCA IgG 98.0 99.3 85.7
Positive 138 2 1
Negative 1 12 0
Inconclusive 0 4 0
EBNA IgG 98.1 100.0 93.2
Positive 112 3 1
Negative 0 41 0
Inconclusive 0 1 0
EA IgG 92.8 87.7 96.3
Positive 50 3 0
Negative 7 79 1
Inconclusive 6 10 1

EA, early antigen; EBNA, Epstein-Barr virus nuclear antigen; IFA, indirect fluorescent antibody; VCA, viral capsid antigen.
* Data are given as number of samples unless otherwise indicated.

manifestations and the antibody pattern of EBV results
obtained. A patient’s serologic status can be determined
based on the presence and relative quantity of antibodies to
the different antigens. A serologic profile based on these
responses is given in Table 1. Agreement of serologic sta-
tus between the multiplex assay and ELISA was 76.6%
(121/158) ❚Table 5❚. There were 37 discrepant results, 18
(11.4%) of which were categorized as major and 19
(12.0%) of which were categorized as minor. When sero-
logic status based on IFA results was used to resolve the
discrepant results, overall agreement increased to 89.9%
(142/158). Of the 16 remaining discrepant results, 10 (6.3%)
were considered major and 6 (3.8%) were considered minor
❚Table 6❚. A further breakdown of the 18 major discrepant
results between the multiplex assay and ELISA is shown in
❚Table 7❚.
Frequency of Uninterpretable Results
The inconclusive range was set at up to 20% more than
the cutoff value for the multiplex assay and at 90% of the cut-
off for the ELISA. For the most part, neither assay had many
test interpretations with inconclusive results, fewer than 1%
for VCA IgG and IgM and for EBNA. For EA, however, the
multiplex assay reported 10.8% (17/158) of all results as
inconclusive, with the ELISA having an inconclusive rate of
3.8% (6/158). In regard to serologic status, the multiplex

© American Society for Clinical Pathology Am J Clin Pathol 2008;129:34-41 37

37 DOI: 10.1309/65VKWVNAQ38PHMGQ 37
Martins et al / MULTIPLEX EBV ANTIBODY ASSAY

❚Table 4❚
Correlation of ELISA and IFA-Resolved Results

IFA-Resolved Result

ELISA Positive Negative Inconclusive Agreement (%) Sensitivity (%) Specificity (%)

VCA IgM 93.5 81.3 96.7

Positive 26 4 2
Negative 6 119 0
Inconclusive 0 0 1
VCA IgG 95.5 94.9 100.0
Positive 131 0 0
Negative 7 18 1
Inconclusive 1 0 0
EBNA IgG 92.3 89.2 100.0
Positive 99 0 0
Negative 12 44 1
Inconclusive 1 1 0
EA IgG 89.9 91.8 88.6
Positive 56 10 1
Negative 5 78 1
Inconclusive 2 4 0

EA, early antigen; EBNA, Epstein-Barr virus nuclear antigen; ELISA, enzyme-linked immunosorbent assay; IFA, indirect fluorescent antibody; VCA, viral capsid antigen.

❚Table 5❚
Comparison of Serologic Status in 158 Samples by the Multiplex Assay With the ELISA*

ELISA Result

Early Acute Recent or Past

Multiplex Result Susceptible Infection Infection Subacute Infection Infection NAC Total

Susceptible 16 0 0 0 0 0 16
Early infection 4† 0 0 0 0 0 4
Acute infection 0 2‡ 19 0 0 1‡ 22
Recent or subacute infection 1† 1‡ 7‡ 42 2† 1‡ 54
Past infection 2† 3† 2† 4† 44 5‡ 60
NAC 0 0 1‡ 0 1‡ 0 2
Total 23 6 29 46 47 7 158

ELISA, enzyme-linked immunosorbent assay; NAC, no applicable category.

* Agreement, 76.6% (n = 121).
† Major discrepancy, 11.4% (n = 18).
‡ Minor discrepancy, 12.0% (n = 19).

❚Table 6❚
Comparison of Serologic Status in 158 Samples by the Multiplex Assay With ELISA With IFA Resolution*

ELISA Result With IFA Resolution

Early Acute Recent or Past

Multiplex Result Susceptible Infection Infection Subacute Infection Infection NAC Total

Susceptible 16 0 0 0 0 0 16
Early infection 2† 3 0 0 0 0 5
Acute infection 0 0 22 0 0 0 22
Recent or subacute infection 0 0 1‡ 50 2† 0 53
Past infection 0 2† 0 4† 50 4‡ 60
NAC 0 0 0 0 1‡ 1 2
Total 18 5 23 54 53 5 158

ELISA, enzyme-linked immunosorbent assay; IFA, indirect fluorescent antibody; NAC, no applicable category.
* Agreement, 89.9% (n = 142).
† Major discrepancy, 6.3% (n = 10).
‡ Minor discrepancy, 3.8% (n = 6).

38 Am J Clin Pathol 2008;129:34-41 © American Society for Clinical Pathology

38 DOI: 10.1309/65VKWVNAQ38PHMGQ
 Immunopathology / ORIGINAL ARTICLE

assay only had 2 (1.3%) of 158 interpretations of “no appli- ❚Table 7❚

cable category,” and the ELISA had 7 (4.4%). Breakdown of the 18 Major Discrepant Results Between the
Multiplex Assay and ELISA
Effects of RFs and Other Interferences Serologic Infection
Status by Method
Of the 20 samples positive for IgM antibodies to herpes
simplex virus, cytomegalovirus, varicella zoster virus, or T Multiplex No. of
ELISA Assay Samples IFA
gondii, 6 (30%) gave positive results for the multiplex EBV
VCA IgM assay. All 6 of these samples also contained elevat- Susceptible Early 4 Susceptible, 1; early, 3
ed concentrations of RF IgM and were, therefore, analyzed Susceptible Recent or 1 Early, 1
subacute
with the additional RF samples as described next. Susceptible Past 2 Early, 2
Of the 23 samples that contained elevated concentrations Early Past 3 Past, 3
Acute Past 2 NAC, 2
of RF IgM, 9 (39%) were reported as invalid results by the Recent or Past 4 Recent or subacute,
multiplexed EBV VCA IgM assay owing to elevated signals subacute 3; NAC, 1
Past Recent or 2 Past, 1; recent or
from the rheumatoid factor control microsphere ❚Table 8❚. An subacute subacute, 1
additional 9 (39%) were positive for the multiplex VCA IgM,
ELISA, enzyme-linked immunosorbent assay; IFA, indirect fluorescent antibody;
and 5 (22%) were negative. Only 3 (13%) of the same 23 sam- NAC, no applicable category.
ples were positive by the IgM capture ELISA (Table 8). When
the 9 multiplex-positive samples were retested using an anti-
human IgG polyclonal antibody in the serum diluent,17 5 sam-
ples (marked with a double dagger in Table 8) changed from ELISA and affinity-purified VCA glycoprotein 125 for the
a positive to negative result, which agreed with the IgM cap- multiplex assay.
ture ELISA result. Of the 4 samples that remained multiplex Conversely, the multiplex assay reported an additional 10
VCA IgM positive after IgG absorption, 2 were also positive VCA IgG positive results (141/158 [89.2%]) compared with
by the IgM capture ELISA. the ELISA (131/158 [82.9%]), even though both ELISA
assays use the same antigens for VCA IgG determination. The

❚Table 8❚
Discussion Determination of RF IgM Interference on the EBV VCA IgM
Although IFA assay is still considered the gold standard Assays
for EBV serologic diagnosis, this technique has been widely IgM RF Multiplex ELISA Capture
replaced by ELISA methods for routine clinical laboratory use Sample No. (IU/mL)* VCA IgM VCA IgM VCA IgG†
owing to its reliability and high-throughput analysis. 1 28 Positive‡ Negative Positive
Numerous studies have shown good concordance between the 2 29 Positive Positive Positive
ELISA and IFA assay.18,19 The new generation of multiplex 3 32 Positive Negative Positive
4 33 Positive‡ Negative Positive
EBV assays offers additional advantages over ELISA, but 5 34 Negative Negative Positive
their performance has not been established in the literature 6 35 Positive‡ Negative Positive
7 47 Negative Negative Positive
using the IFA assay for discrepancy resolution.20 In this study, 8 53 INV Positive Positive
we evaluated the first Food and Drug Administration–cleared 9 57 Positive‡ Negative Positive
10 57 INV Negative Positive
multiplex EBV assay in comparison with a well-established 11 57 Negative Negative Positive
ELISA method and traditional IFA assay. 12 58 Positive Negative Positive
13 63 Positive‡ Negative Positive
The ELISA gave a greater percentage of VCA IgM–pos- 14 69 INV Negative Positive
itive results (32/158 [20.3%]) compared with the multiplex 15 96 Negative Negative Positive
assay (25/158 [15.8%]). The ELISA is based on the IgM- 16 96 Negative Negative Positive
17 111 Positive Positive Positive
capture method, which generally results in a more specific 18 118 INV Negative Positive
but less sensitive assay compared with the method used by 19 124 INV Negative Positive
20 137 INV Negative Positive
the multiplex assay in which the antigen is directly bound to 21 244 INV Negative Positive
the solid phase (microsphere) to which specific antibodies in 22 253 INV Negative Positive
23 541 INV Negative Positive
the patient’s serum sample bind. In our study, however, the
ELISA method reported 7 additional positive results com- EBV, Epstein-Barr virus; ELISA, enzyme-linked immunosorbent assay; INV, invalid
result owing to high signal from RF control microsphere; RF, rheumatoid factor;
pared with the multiplex method. It could be assumed, there- VCA, viral capsid antigen.
* More than 26 IU/mL considered clinically significant.
fore, that the discrepancies are more likely due to differences † Positive by multiplex and ELISA.

in the antigen used by the 2 assays, synthetic p18 peptide for ‡ Result was negative after IgG absorption.

© American Society for Clinical Pathology Am J Clin Pathol 2008;129:34-41 39

39 DOI: 10.1309/65VKWVNAQ38PHMGQ 39
Martins et al / MULTIPLEX EBV ANTIBODY ASSAY
positivity rate for EBNA IgG was also higher for the multiplex
assay (116/158 [73.4%]) compared with the ELISA (99/158
[62.7%]). The ELISA antigen is a 59-amino-acid synthetic
peptide derived from the EBNA-1 domain, whereas an
EBNA-1 recombinant protein is used in the multiplex assay.
For the EA, the ELISA assay used a 47-kd recombinant
polypeptide for the diffuse component only but had a higher
positivity rate, 43.0% (68/158), than the multiplex assay,
33.5% (53/158), which contained affinity-purified antigens for
the restricted and diffuse antigen components.
Based on the pattern of antibody responses to the different
viral antigens produced during the EBV replication cycle, the
serologic profile of a patient can be classified as susceptible to
infection or as having an early, acute, recent or subacute, or
past infection (Table 1). There was good concordance for the
serologic profiles between the ELISA and multiplex methods
(89.9%) when the discrepant results were resolved by IFA
assay, the gold standard. Of the 18 major discrepant results
(Table 7), the multiplex assay result was in agreement with the
IFA result in 7 cases, ELISA and IFA results were in agreement
in 5 cases, and in 6 cases, the IFA result agreed with neither the
multiplex nor ELISA result. The most frequent major discrep-
ancy between the assays was profiles regarded as susceptible
(negative for all antigens) by the ELISA but classified as a past
infection by the multiplex assay owing to VCA IgG– and
EBNA-positive responses. Overall, the ELISA had lower pos-
itivity rates for VCA IgG and EBNA, leading to a greater num-
ber of susceptible profiles than the multiplex assay, but a high-
er number of VCA IgM–positive results giving rise to more
acute or recent infection profiles than the multiplex assay.
With the exception of EA, both assays reported minimal
inconclusive results, which may lead to less repeated testing
and confusion in the interpretation of serologic status. The
multiplex assay had 5 fewer patient antibody profiles giving
an uninterpretable status compared with the ELISA.
RFs are autoantibodies of any isotype with specificity for
the Fc portion of IgG. In IgM serologic assays in which the
antigen is bound to the solid phase, such as the multiplex VCA
assay, specific IgG antibodies present in the serum bind to the
antigen, presenting a site for RF to bind. RF of the IgM class
may then be recognized by the anti-IgM conjugate, giving rise
to a false-positive IgM reaction.17 The RF control microsphere
contained in the multiplex VCA IgM assay was effective in
identifying samples with highly elevated concentrations of RF
IgM by flagging these samples as invalid, but it did not iden-
tify samples with moderate concentrations of RF IgM, which
seemed to cause false-positive results in the multiplex assay.
IgM RF concentrations of more than 26 IU/mL are considered
clinically significant, but the amount of RF required to cause
assay interference has not been defined and would most like-
ly vary between assays. Seven of multiplex IgM-positive
results shown in Table 8 were most likely falsely positive
40 Am J Clin Pathol 2008;129:34-41
40 DOI: 10.1309/65VKWVNAQ38PHMGQ
owing to IgM RF because they were negative by the IgM-cap-
ture ELISA, which is not prone to RF interference. This
assumption is further supported as 5 of those samples had neg-
ative results after treatment with an IgG-absorbent serum dilu-
ent. Therefore, at least 5 of the VCA IgM discrepant results,
which were positive by the multiplex assay and negative by the
ELISA (Table 2), could be attributed to IgM RF interference.
The multiplex test system uses the Intra-Well Calibration
Technology that uses a multipoint standard curve contained
within the bead mix. In addition to the assay controls, this
technology generates a 4-point standard curve in each reaction
to internally calibrate each well of the assay. Regression
analysis of the internal standards is performed, allowing the
software to adjust the calculated unit values based on the
unique characteristics of the patient serum sample. The soft-
ware also minimizes front-to-back intra-assay variation by
using these internal controls.
The multiplex assay also has the advantage of reporting
IgG results for all 3 antigens from a single reaction, result-
ing in reagent, labor, sample volume, and storage space sav-
ings. The VCA IgM assay is performed separately, but the
same sample dilutions can be used for both IgG and IgM
assays. There are several options for automating the entire
multiplex procedure, but these were not evaluated in the
present study. The multiplex system provides an accurate
and cost-effective system for the detection of antibodies to
the standard EBV markers.
From the 1Associated Regional and University Pathologists
(ARUP) Institute for Clinical and Experimental Pathology, and the
Departments of 2Pathology, 3Pediatrics, and 4Medicine, University
of Utah School of Medicine, Salt Lake City.
Address reprint requests to Mr Martins: ARUP Institute for
Clinical and Experimental Pathology, 500 Chipeta Way, Salt Lake
City, UT 84108.
References
1. Fleisher G, Henle W, Henle G, et al. Primary infection with
Epstein-Barr virus in infants in the United States: clinical and
serologic observations. J Infect Dis. 1979;139:553-558.
2. Sixbey JW, Nedrud JG, Raab-Traub N, et al. Epstein-Barr virus
replication in oropharyngeal epithelial cells. N Engl J Med.
1984;310:1225-1230.
3. Henle W, Henle G, Niederman JC, et al. Antibodies to early
antigens induced by Epstein-Barr virus in infectious
mononucleosis. J Infect Dis. 1971;124:58-67.
4. Henle G, Henle W, Klein G. Demonstration of two distinct
components in the early antigen complex of Epstein-Barr
virus–infected cells. Int J Cancer. 1971;8:272-282.
5. Evans AS, Niederman JC, Cenabre LC, et al. A prospective
evaluation of heterophile and Epstein-Barr virus–specific IgM
antibody tests in clinical and subclinical infectious
mononucleosis: specificity and sensitivity of the tests and
persistence of antibody. J Infect Dis. 1975;132:546-554.
© American Society for Clinical Pathology

6. Henle W, Henle GE, Horwitz CA. Epstein-Barr virus specific
diagnostic tests in infectious mononucleosis. Hum Pathol.
1974;5:551-565.
7. Rea TD, Ashley RL, Russo JE, et al. A systematic study of
Epstein-Barr virus serologic assays following acute infection.
Am J Clin Pathol. 2002;117:156-161.
8. Horwitz CA, Henle W, Henle G, et al. Long-term serological
follow-up of patients for Epstein-Barr virus after recovery from
infectious mononucleosis. J Infect Dis. 1985;151:1150-1153.
9. Horwitz CA, Henle W, Henle G, et al. Clinical evaluation of
patients with infectious mononucleosis and development of
antibodies to the R component of the Epstein-Barr
virus–induced early antigen complex. Am J Med. 1975;58:330-
338.
10. Baetens DG, Van Renterghem LM. Coupled particle light
scattering: a new technique for serodiagnosis of Epstein-Barr
virus infection. J Med Virol. 2001;64:519-525.
11. van Grunsven WM, Nabbe A, Middeldorp JM. Identification
and molecular characterization of two diagnostically relevant
marker proteins of the Epstein-Barr virus capsid antigen
complex. J Med Virol. 1993;40:161-169.
12. van Grunsven WM, Spaan WJ, Middeldorp JM. Localization
and diagnostic application of immunodominant domains of
the BFRF3-encoded Epstein-Barr virus capsid protein. J Infect
Dis. 1994;170:13-19.
13. van Grunsven WM, van Heerde EC, de Haard HJ, et al. Gene
mapping and expression of two immunodominant Epstein-Barr
virus capsid proteins. J Virol. 1993;67:3908-3916.
© American Society for Clinical Pathology
Immunopathology / ORIGINAL ARTICLE
14. Halprin J, Scott AL, Jacobson L, et al. Enzyme-linked
immunosorbent assay of antibodies to Epstein-Barr virus
nuclear and early antigens in patients with infectious
mononucleosis and nasopharyngeal carcinoma. Ann Intern
Med. 1986;104:331-337.
15. Fulton RJ, McDade RL, Smith PL, et al. Advanced
multiplexed analysis with the FlowMetrix system. Clin Chem.
1997;43:1749-1756.
16. Stites DP, Rodgers RP. Basic and Clinical Immunology. 7th ed.
Norwalk, CT: Appleton & Lange; 1991.
17. Martins TB, Jaskowski TD, Mouritsen CL, et al. An
evaluation of the effectiveness of three immunoglobulin G
(IgG) removal procedures for routine IgM serological testing.
Clin Diagn Lab Immunol. 1995;2:98-103.
18. Gerber MA, Shapiro ED, Ryan RW, et al. Evaluations of
enzyme-linked immunosorbent assay procedure for
determining specific Epstein-Barr virus serology and of rapid
test kits for diagnosis for infectious mononucleosis. J Clin
Microbiol. 1996;34:3240-3241.
19. Gartner BC, Hess RD, Bandt D, et al. Evaluation of four
commercially available Epstein-Barr virus enzyme
immunoassays with an immunofluorescence assay as the
reference method. Clin Diagn Lab Immunol. 2003;10:78-82.
20. Klutts JS, Liao RS, Dunne WM Jr, et al. Evaluation of a
multiplexed bead assay for assessment of Epstein-Barr virus
immunologic status. J Clin Microbiol. 2004;42:4996-5000.
Am J Clin Pathol 2008;129:34-41 41
41 DOI: 10.1309/65VKWVNAQ38PHMGQ 41