Article in press - uncorrected proof

Clin Chem Lab Med 2011;49(12):2001–2006
2011 by Walter de Gruyter • Berlin • Boston. DOI 10.1515/CCLM.2011.702

Heterophilic antibody interference in commercial

immunoassays; a screening study using paired native
and pre-blocked sera

Nils Bolstad1,*, David J. Warren1, Johan Bjerner2, Introduction

Gunnhild Kravdal3, Lutz Schwettmann4,
Kari H. Olsen1, Pål Rustad2 and Kjell Nustad1 Immunoassay techniques have revolutionized the determi-
1
Department of Medical Biochemistry, Oslo University nation of clinically relevant protein and peptide analytes.
Hospital, Radiumhospitalet, Oslo, Norway However, these methods do not always give the ‘‘correct’’
2
Fürst Medical Laboratory, Oslo, Norway result (1–4), and extreme caution is needed when clinical
3
Unit of Medical Biochemistry, Division for Diagnostics findings and assay results are discordant. The unnecessary
and Technology, Akershus University Hospital, Lørenskog, diagnostic and therapeutic interventions that often follow
Norway such discordance can be costly to both patients and hospitals
4
Department of Medical Biochemistry, Ålesund Hospital, (5–11). In these situations troubleshooting is complicated by
Ålesund, Norway the proprietary nature of most information relating to the
assay kits.
Immunometric assays are particularly sensitive to interfer-
Abstract ence by multivalent antibody-binding moieties that can
bridge the reagent antibodies. Such cross-linking results in
Background: Heterophilic antibodies are still an important the generation of positive assay signals in the absence of
source of interference in immunoassays. We have conducted analyte. Heterophilic or human anti-mouse antibodies
a screening study for interference in a panel of commercially (HAMAs) present in patient sera are the usual culprits (1, 3,
available assays using two sera known to contain high titer 4). Several approaches can be effective in limiting hetero-
Fc-reactive heterophilic antibodies. philic antibody interference including sample pretreatment
Methods: The sera were distributed to laboratories partici- with heterophilic blocking tubes (HBT) (12), polyethylene
pating in the Nordic External Quality Assessment coopera- glycol (PEG) precipitation (13), affinity chromatography on
tion (EQANord). Duplicate samples pre-blocked with protein A (14) or size exclusion chromatography (15). These
aggregated murine monoclonal MAK33 were also supplied. methods are however not well suited to the high-throughput
Discrepancies ()50%) between the results for native and assays used in clinical laboratories. Indeed, optimal reduc-
blocked samples were used to classify the tested assays as tions in assay interference can most probably only be
susceptible to interference. A total of 170 different assay kits achieved by focusing on this problem during the assay design
covering 91 analytes were tested. phase (16).
Results: We found that 21 assays, covering 19 different ana- In a previous study, we found that the frequency of inter-
lytes, were susceptible to interference from the heterophilic ference in our in-house immunometric assay for carcinoem-
antibodies in the two sera. Many of these are clinically and bryonic antigen (CEA) was 4.0% (1). The addition of a
commercially important assays. Some of the false results heat-aggregated irrelevant murine monoclonal antibody,
were grossly elevated and could have been detrimental to MAK33, to the assay buffer reduced the frequency to 0.86%.
patient care in a clinical setting. Significantly, the use of F(ab’)2 fragments as assay solid
Conclusions: Heterophilic antibodies with Fc-reactivity phases was found to reduce the frequency of interference to
remain a threat. A more widespread use of antibody frag- 0.1% even in the absence of irrelevant mouse immuno-
ments and aggregated immunoglobulin could potentially globulin. This is in agreement with previous studies which
improve the heterophilic antibody resistance of assays indicated that most interfering immunoglobulin target the Fc
intended for clinical use. portion of assay antibodies (17, 18).
Most commercial immunoassays have irrelevant animal
Keywords: HAMAs; heterophile antibody; heterophilic; immunoglobulin added to the assay reagents in order to limit
interference; immunoassay. interference. Some manufacturers aggregate the immuno-
globulin by heat or chemicals, but this is rarely detailed in
*Corresponding author: Nils Bolstad, Department of Medical package inserts. To our knowledge, only a limited number
Biochemistry, Oslo University Hospital, Radiumhospitalet, 0310 of commercial assays are constructed using F(ab’)2 or Fab’
Oslo, Norway fragments.
Phone: q47 22935107, Fax: q47 22730725,
E-mail: [email protected] Herein we describe the results from a screening study of
Received June 20, 2011; accepted August 5, 2011; heterophilic antibody interference in commercially available
previously published online September 8, 2011 immunoassays using two high titer sera originally referred to

2011/0391
 Article in press - uncorrected proof
2002 Bolstad et al.: Heterophilic antibody interference in immunoassays

our laboratory for testing. Both sera, shown to contain Fc- Selection of tested assays and participating
reactive heterophilic antibodies, were distributed to 18 clin- laboratories
ical laboratories through the Nordic External Quality
EQAnord provided invaluable assistance in selecting representative
Assessment cooperation (EQAnord). The sera were supplied
laboratories and methods. Particular focus was directed at including
in both native form and pre-blocked by the addition of heat-
the assays most widely used in the Nordic countries. Where possi-
aggregated MAK33 non-specific immunoglobulin. ble, assays performed on different instrument models from the same
Our aims were firstly to investigate if selected sera, sup- manufacturer were tested. Of the 19 laboratories invited to partici-
plied as paired native and pre-blocked specimens, can be pate in our study, only one laboratory declined, citing reorganization
used as screening tools for assay interference and secondly, of laboratory services. The participating laboratories were invited to
how well a panel of commercial assays were protected include immunoassays at will if they had surplus test sera. For this
against Fc-reactive heterophilic antibodies. reason, some in-house assays and non-immunometric (competitive)
assays were included in the study. A total of 170 commercial immu-
noassay kits were tested.
Materials and methods
Heat treatment of MAK33
Human test sera used for assay screening
Murine monoclonal IgG1k antibody MAK33 (Roche Molecular
Serum 1 was from a man in his fifties. Interference was suspected Biochemicals, Mannheim, Germany) was stored at a concentration
when analysis of soluble transferrin receptor (Tina-quant sTfR on of 2 g/L in 0.15 mol/L NaCl, 0.01 mol/L Na2HPO4, pH 7.4, at
the Cobas Integra 800; Roche Diagnostics, Mannheim, Germany) –30oC. Aliquots were heat-treated by incubation in a 60oC water
gave an elevated result without corresponding laboratory results or bath for 10 min (1). The change in absorption at 595 nm from
clinical symptoms. The result was normalized after addition of aggre- approximately 0.02 to approximately 1.0 was used to monitor
gated MAK33 to the sample. Approximately 6 months after he donated aggregation.
serum to our study, he was diagnosed with rheumatoid arthritis.
Serum 2 was from a woman in her thirties. Interference was sus- Screening of commercial immunometric assays
pected when analysis of b-hCG (total b-hCG on the Architect
i2000 SR; Abbott Diagnostics, Abbott Park, IL, USA) gave an ele- Aliquots of the two patient sera with known interference from hete-
vated result without concomitant pregnancy or malignancy. The rophilic antibodies were distributed to participating laboratories.
result was normalized after addition of MAK33. Re-analysis of b- Duplicate samples pre-blocked with 180 mg/mL aggregated murine
hCG with a different method (hCGqb on the Cobas e601; Roche) monoclonal MAK33 were also supplied. The participating labora-
confirmed the normal result. Tragically, prior to our identification tories were informed about the purpose and design of the study, but
of heterophilic antibody interference, this donor endured unneces- were not informed about which aliquots were blocked with MAK33.
sary chemotherapy and three inappropriate surgical procedures, Laboratories were instructed to perform analyses and report results
including the laparoscopic removal of a fallopian tube. as for routine samples.
Both sera were obtained with informed consent following nation- Prior to the study, we set a cut-off limit of 50% for the difference
al and institutional guidelines. The study has been approved by the between the results from native and pre-blocked sera to indicate if
Oslo university hospital privacy office. the method tested is vulnerable to heterophilic antibody interference.
As assays are subject to analytical variation, such differences may
Characterization of heterophilic antibodies occur by chance. The probabilities for an observed 50% difference
between native and pre-blocked sera by chance are ps0.001 at an
All assays used to characterize the sera were manual 3-step methods analytical CV of 11.4%, ps0.01 at a CV of 15.1% and ps0.05 at
performed using streptavidin-coated DELFIA microtitration strips.
a CV of 21.4%. The three smallest differences considered to be
The wells were coated with biotinylated antibodies, washed, and
significant in the study were in the SHBG and BNP assays from
then incubated with the serum samples. Following additional wash-
Abbott laboratories, with observed differences of 60% and 107%,
ing, the assays were developed using europium-labeled tracer anti-
and analytical CV of 10% and 12% as stated by the manufacturer,
bodies. Methodological details are given in a supplemental file (web
and the AutoDELFIA TSH assay from Perkin Elmer Life Sciences,
only) and in Bjerner et al. (19).
with an observed difference of 60%, where analytical CV has been
To detect heterophilic antibodies, and characterize their reactivity
reported as 2.8% (20). This corresponds to p-0.001 for all the three
to murine immunoglobulin, assays were established using intact
assays. Thus, all differences reported in the study correspond to
IgG, F(ab’)2, and Fc fragments of the IgG1 monoclonal antibodies
p-0.001. Our screening method might not be suitable for some of
K57 (anti-a-fetoprotein) and T84.66 (anti-carcinoembryonic anti-
the included assays, e.g., competitive assays or assays using non-
gen) as solid phase reagents and K57-IgG as tracer. These are non-
murine antibodies. As we have little experience with interference in
sense immunometric assays since they use non-complementary
competitive assays, we have not classified these assays based on
assay pairs. A positive signal indicates the cross-linking of the solid
our test. However, results from all tested assays are presented in
phase and tracer antibodies in the absence of analyte.
Table 2 in the electronic supplement that accompanies this paper.
Species specificity of the heterophilic antibodies was determined
using polyclonal murine, rabbit, ovine, equine, bovine, and human
IgG as solid phase antibodies and K57-IgG as tracer antibody.
The size of the interfering antibodies was estimated by gel- Results
permeation chromatography on a pre-calibrated Superdex S200 col-
umn. Column fractions were assayed using a non-sense method Characterization of sera
(solid phase K57-IgG, tracer K57). Isotyping was performed using
K57-IgG as solid phase antibody and commercially available rabbit Both sera gave grossly elevated responses in non-sense
F(ab’)2 antibodies to human heavy and light chains as tracers. assays indicating the presence of heterophilic antibodies with
 Article in press - uncorrected proof
Bolstad et al.: Heterophilic antibody interference in immunoassays 2003

affinity for whole IgG and the Fc fragments of IgG1 anti-

bodies. They also displayed high titers and could be diluted
1:300 (serum 2) and 1:3,000 (serum 1) before a positive
assay signal was lost. Responses were normalized after the
addition of heat-aggregated MAK33 immunoglobulin to the
sera. Very little binding to F(ab’)2 fragments was observed
(Figure 1).
In addition to their strong reactivity to murine IgG1-anti-
bodies, the heterophilic antibodies in serum 1 showed some
cross-reactivity to rabbit IgG, but minimal cross-reactivity to
human, bovine, and equine IgG, while the heterophilic anti-
bodies in serum 2 showed some cross-reactivity to bovine
IgG (Figure 2).
Gel-permeation chromatography indicated that the size of
the heterophilic antibodies was )650 kDa suggesting that
they are most likely IgMs (data not shown). Isotyping using
a modified non-sense assay gave strong signals for m and k
but comparatively low signals for l light chain. However,
we were unable to detect monoclonal components using the
routine methods available at our hospital: capillary zone
electrophoresis with immunotyping and immunofixation
electrophoresis.

Screening of commercial immunoassays

Figure 2 Reactivity of heterophilic antibodies to IgG from differ-
Analysis of one or both sera showed interference in 21 assay ent animal species.
kits covering 19 different analytes (Table 1). As expected,
interference was not limited to assays from one or a few
manufacturers. False results were seen for both test sera in results for the pre-blocked samples in vulnerable assays were
13 assays, while 8 assays gave false results for either serum comparable to corresponding results for the native samples
1 (6 assays) or serum 2 (2 assays). The degree of false ele- in resistant assays (Table 2, electronic supplement).
vation varied between assays. In 15 assays, results for one For one assay, CA125 (CA125II, Abbott) on the Architect
or both native samples were increased at least five-fold com- platform, our results could indicate a significant variation
pared to corresponding blocked samples. In 11 assays, the between lots with respect to vulnerability from interference
difference between native and blocked sera was )10-fold. (lot numbers are reported in the electronic supplement). For
No assays displayed negative interference in this study. The the other interference-positive assays where different lots
were tested, such as b-hCG (total b-hCG, Abbott) and sol-
uble transferrin receptor (Tina-quant sTfR, Roche), inter-
batch variation was not observed in this study.
CA125 on Abbott Architect models i2000 and ci8200, and
b-hCG on Abbott Architect models i2000SR, ci8200 and
ci16200 showed falsely elevated results for both sera tested,
with little or no difference between instruments. All results
were normalized when adding heat-treated MAK33. The
results obtained for CA125 and b-hCG on the Architect
instruments using pre-blocked sera are comparable to results
from assays negative for interference.
D-dimer (STA-LIATEST D-DI) on STA-R Evolution
and STA Compact; (Diagnostica Stago, Gennevilliers,
France, www.stago.com) showed grossly elevated results for
both sera tested. Although significantly lower than in the
native sample, results after addition of heat treated MAK33
in serum 1 differed from results obtained with assays nega-
tive for interference. This is most likely due to the addition
of inadequate amounts of aggregated immunoglobulin to
completely block interference in this sample in this particular
assay. Results for serum 2 with MAK33 are comparable to
Figure 1 Reactivity of heterophilic antibodies to murine IgG1 those from assays negative for interference. The values
fragments. obtained in serum samples are higher than in a plasma sam-
 Article in press - uncorrected proof
2004 Bolstad et al.: Heterophilic antibody interference in immunoassays

Table 1 Assays vulnerable to inference in the present studya.

Analyte, unit Manufacturer Instrument Serum 1 Serum 1 Serum 2 Serum 2

blocked blocked
AFP, mg/L Diasorin Liaison 2.6 4.2 0.6 2.8
BNP, pmol/L Abbott Architect i2000SR -2.9 6.0 -2.9 -2.9
Architect ci16200 -2.9 3.4 -2.9 -2.9
CA125, kU/L Abbott Architect i2000 8 256 6 96
Architect ci8200 7 278 6 84
Architect i2000 7 20 6.7 20
Diasorin Liaison 5.5 7 4.2 74
CA19-9, kU/L Diasorin Liaison 17 18 7.3 89
CEA, mg/L Diasorin Liaison -1 1.9 -1 3.6
Crosslaps, ng/L Roche Modular E 170 75.6 1838 11.5 32.8
CRPb, mg/L Roche Cobas Integra 800 1.1 3.4 -0.6 0.7
Modular P -0.6 4.3 -0.6 -0.6
D-dimerb, mg/L Diagnostica Stago STA-R Evolution 1.8 )4.0 0.6 )4.0
STA Compact 2.5 )20 0.5 2.3
STA Compact 2.1 )20 0.6 2.2
GH, mg/L PerkinElmer AutoDELFIA 0.1 1.5 0.1 0.5
hCG, U/L Abbott Architect ci16200 -1.0 42 -1.0 113
(Total b) Architect i2000 SR -1.2 59 -1.2 147
Architect ci8200 -2 36 -2 117
Insulin, pmol/L Siemens Immulite 2000 351 1191 22 131
Interleukin 1b, ng/L Bio-Rad Bio-Plex 200 -0.13 134 -0.13 2.8
Interleukin 2, ng/L Bio-Rad Bio-Plex 200 -0.13 685 -0.13 4.8
Interleukin 6, ng/L Bio-Rad Bio-Plex 200 18.8 232 1.6 11
PAPP-A, U/L PerkinElmer AutoDELFIA 0.003 0.16 0.002 0.21
SHBG, nmol/L Abbott Architect i2000 32.8 52.6 97.3 98
Beckman Coulter UniCel DxI 800 29 99.5 73.7 85
sTfRb, mg/L Roche Modular P 3.5 26.5 1.7 1.8
Cobas Integra 800 4.4 19.2 1.9 3.1
Hitachi 917 2.7 25.7 1.7 1.8
TNF-a, ng/L Bio-Rad Bio-Plex 200 1.2 95.5 4.8 5.3
TSH, mU/L PerkinElmer AutoDELFIA 1.5 2.3 1.6 2.1
a
Results indicating interference (difference )50% between native and blocked sample) in bold. bImmunoturbidimetric assays. Assays
resistant to interference in our study (listed by manufacturer): Abbott: AFP, anti-CCP, anti-CMV IgG, anti-HAV IgG, anti-HAV IgM, anti-
HBcII, anti-HBs, anti-HCV, anti-HTLV I/II, anti-rubella IgG, anti-toxo IgG, anti-toxo IgM, anti-TPO, B12, CA15-3, CA19-9, CEA, CRP,
ferritin, FSH, FT3, FT4, HBsAG, HIV Ag/Ab combo, LH, prolactin, PSA, PSA(free), PTH, transferrin, tTroponin I, TSH. Beckman Coulter:
anti-TPO, BNP, CA125, CEA, estradiol, ferritin, FSH, hCG (total b), LH, prolactin, PSA, PTH, testosterone, troponin I, TSH. Biomerieux:
anti-CMV IgM. Bio-RAD: anti-toxo IgG, anti-toxo IgM, anti-mycoplasma IgM. Biotest: anti-EBNA IgG. Biotrin: anti-parvo B19 IgG, anti-
parvo B19 IgM. Brahms: AFP, CA19-9, procalcitonin, TgAB. Dade Behring (Siemens): anti-borrelia IgG, anti-borrelia IgM, anti-HSV IgG,
anti-syphilis, anti-varicella IgG. Diasorin: CA15-3, NSE. Instrumentation Lab: d-dimer. Novitec (Orion): anti-EBV VCA IgG, anti-EBV
VCA IgM. Ortho: BNP, CRP, troponin I. PerkinElmer: a-1-antitrypsin, LH, FSH, b-hCG (free), PSA, PSA (free). Roche: AFP, ApoA1,
ApoB, CA15-3, CA19-9, CA72-4, CA125, CEA, cortisol, C-peptide, cyfra 21-1, D-dimer, DHEAS, estradiol, ferritin, FSH, FT3, FT4,
haptoglobin, hCG (hCGqb), IgA, IgG, IgM, insulin, LH, myoglobin, NSE, NT-proBNP, progesterone, prolactin, PSA, PTH, S-100, SHBG,
testosterone, transferrin, troponin T, troponin T hs, TSH, TSH-receptor-Ab. Siemens (Advia-platforms): anti-TPO, B12, CRP, ferritin, FSH,
hCG (ThCG), LH, prolactin, PSA, PTH, TSH. Siemens (Immulite-platforms): ACTH, calcitonin, CDT (%), C-peptide, GH, hCG, IGF1,
interleukin 1b, interleukin 6, interleukin 8, interleukin 10, NT-proBNP, sIL-2R, TgAB, Tg (thyroglobulin), TNF-a.

ple due to the expected formation of some d-dimer fragments analyte. All results are given in the electronic supplement,
during the coagulation phase prior to sample centrifugation. while only the first result for each analyte is given in the
Insulin on Immulite2000 (Siemens Healthcare Diagnostics, printed table. It should be noted that for serum 1 (both native
Erlangen, Germany, www.medical.siemens.com) showed false- and blocked), the Bio-Plex instrument gave a warning signal
ly elevated results for both sera tested. Results after addition with the result, possibly due to bead aggregation.
of heat-treated MAK33 are very similar to the results from
another insulin-assay (Roche).
Interleukins 1b, 2, and 6 analyzed on Bio-Plex 200 (Bio- Discussion
Rad Laboratories, Hercules, CA, USA, www.bio-rad.com) all
gave falsely elevated results for both sera tested. TNF-a gave Characterization of the sera used in this study indicated that
a falsely elevated result for the native serum from donor 1. they contained heterophilic antibodies belonging to the IgM
For all four samples, duplicate results were reported for each class with principle reactivity to the Fc domain of murine
 Article in press - uncorrected proof
Bolstad et al.: Heterophilic antibody interference in immunoassays 2005

IgG1. In our non-sense assays this reactivity was effectively trol, but also it is easy to use and interpretation is relatively
blocked by the addition of heat-aggregated MAK33 immu- simple. It should, however, only be used with assays con-
noglobulin. These observations permitted the use of the sera, taining murine antibodies and, since MAK33 is an antibody
as paired native and pre-blocked samples, to test if commer- to CK-MB, its use is inappropriate with assays for this par-
cial immunoassays are sufficiently protected against inter- ticular analyte.
ference from Fc-reactive heterophilic antibodies. Herein we show that a surprising number of immunoassay
The fact that such a large number of the tested assays are kits (21 out of 170 tested) are vulnerable to Fc-reactive hete-
vulnerable to heterophilic antibodies with classic Fc-reacti- rophilic antibodies. Had more sera with heterophilic antibod-
vity is a cause for concern. We have previously shown that ies been included, or more assays been tested, it is likely that
this interference could probably have been avoided by additional vulnerable assays would have been identified. The
removing the Fc fragment from the solid phase assay anti- fact that 149 assays proved resistant to Fc-reactive hetero-
bodies (2). It is therefore surprising that this approach is not philic antibodies in our study should not lead to a false sense
used more frequently. of security when using these particular assays. As with other
In the rare event that heterophilic antibodies bind the interference tests, negative results do not exclude the possi-
F(ab’)2-region of the assay antibodies, the inclusion of block- bility of interference.
ing immunoglobulin in the assay buffers provides a final, but Based on this study and our previous findings (1, 30), we
important, line of defense. As demonstrated herein, and in a argue that some immunoassay kits need to be better protected
number of previous studies, aggregated antibodies are potent against Fc-reactive heterophilic antibodies. This could be
blockers of interference (1, 21, 22). This efficacy is most accomplished using either F(ab’)2, Fab’ or scFv assay anti-
probably related to the stable binding of low-affinity inter- bodies and adequate concentrations of aggregated irrelevant
fering IgMs (23) to the reiterative epitopes displayed on IgG in the assay reagents. The added blocking immunoglob-
aggregated immunoglobulin. MAK33 is a good choice for ulin should be similar to the assay antibodies with respect to
blocking reagent since IgG1 monoclonal antibodies are com- species and subclass.
monly chosen as capture antibodies to prevent consumption A potential limitation of this study, given the marked het-
of the solid phase through complement activation (24). erogeneity of heterophilic antibodies, is the small number of
As long as immunoassays are vulnerable, it is important sera used. However, in a prior investigation, using 198 inter-
that clinical laboratories implement strategies for identifying ference-positive sera (selected from 11,261 tested specimens)
samples with a high probability of interference. To identify we observed that 194 demonstrated Fc-reactivity (2). Thus,
samples, Ismail et al. (25) suggest a probabilistic approach, the two sera we used in this study contained heterophilic
i.e., elevated results in assays known to have a low rate of antibodies with the reactivity most commonly associated
true positive results should be retested for interference. We with antibody interference.
agree with this probabilistic approach, but we also think that In conclusion, this article describes a simple way of
the impact of the assay result should guide which samples screening immunoassays for interference using small panels
to retest for interference. A false-positive HIV-1, hCG or of native and pre-blocked sera. Using this method we dem-
troponin I result probably has more impact than a falsely onstrate that some commercial assays are poorly protected
elevated interleukin 6, although interference may be equally against heterophilic antibodies with Fc-reactivity.
probable in all these assays. An optimal strategy would be
based on detailed knowledge of (and experience with) the
assay, analyte, and interference tests in question (26, 27). We
stress this because interpretation of interference tests is rarely Acknowledgments
as simple and straightforward as we would like. An extensive
discussion on this subject has been published previously The authors would like to thank the two serum donors for their
(28). A general rule is that a negative interference test does admirable generosity. We would also like to acknowledge invaluable
not exclude heterophilic antibody interference. A positive contributions from Catarina Andersson, Tone Berge, Leila Zamani
test, given appropriate controls and correct interpretation, can Fekri, Laila Gjerdalen, Emma G. Haare, Eyjólfur Har>arson, Bente
normally be trusted as a proof of interference. Heesch, Sara Locke, Britt Marie Loo, Lena Malmstedt, Kristine
Solem, Astrid Steiro, Hildur Tannum, Trude Torsnes, Solfrid Elise
In this study, we relied on the ability of the commercially
Tungesvik, Hans Wallinder, Elisabeth Paus, Ole Børmer, Trine Bjøro
available immunoglobulin MAK33 to block heterophilic and EQANord.
antibodies when added to sera prior to assay. This approach
was chosen because in a previous study, using a panel of
11,261 sera, we demonstrated that this reagent was able to
reduce the level of interference to -1% (1). We believe that Conflict of interest statement
re-assay after blocking with aggregated MAK33, or other
Authors’ conflict of interest disclosure: The authors stated that
commercially available heterophilic blocking reagents
there are no conflicts of interest regarding the publication of this
(HBRs) (29) for that matter, may prove a good testing alter- article.
native when interference is suspected in the routine diagnos- Research funding: None declared.
tic laboratory. Not only is aggregated MAK33 commercially Employment or leadership: None declared.
available in a form that has undergone stringent quality con- Honorarium: None declared.
 Article in press - uncorrected proof
2006 Bolstad et al.: Heterophilic antibody interference in immunoassays

References 17. Csako G, Weintraub BD, Zweig MH. The potency of immu-
noglobulin G fragments for inhibition of interference caused by
anti-immunoglobulin antibodies in a monoclonal immunoradio-
1. Bjerner J, Nustad K, Norum LF, Olsen KH, Bormer OP. Immu- metric assay for thyrotropin. Clin Chem 1988;34:1481–3.
nometric assay interference: incidence and prevention. Clin 18. Vaidya HC, Beatty BG. Eliminating interference from hetero-
Chem 2002;48:613–21. philic antibodies in a two-site immunoassay for creatine kinase
2. Boscato LM, Stuart MC. Incidence and specificity of interfer- MB by using F(ab’)2 conjugate and polyclonal mouse IgG. Clin
ence in two-site immunoassays. Clin Chem 1986;32:1491–5. Chem 1992;38:1737–42.
3. Boscato LM, Stuart MC. Heterophilic antibodies: a problem for 19. Bjerner J, Olsen KH, Bormer OP, Nustad K. Human hetero-
all immunoassays. Clin Chem 1988;34:27–33. philic antibodies display specificity for murine IgG subclasses.
4. Kricka LJ. Human anti-animal antibody interferences in immu- Clin Biochem 2005;38:465–72.
nological assays. Clin Chem 1999;45:942–56. 20. Schmiegelow M, Feldt-Rasmussen U, Rasmussen AK, Poulsen
5. Bertholf RL, Johannsen L, Guy B. False elevation of serum HS, Muller J. A population-based study of thyroid function
CA-125 level caused by human anti-mouse antibodies. Ann after radiotherapy and chemotherapy for a childhood brain
Clin Lab Sci 2002;32:414–8. tumor. J Clin Endocrinol Metab 2003;88:136–40.
6. Ismail AA, Walker PL, Barth JH, Lewandowski KC, Jones R, 21. Mössner E, Lenz H, Bienhaus G. Elimination of heterophilic
Burr WA. Wrong biochemistry results: two case reports and antibody interference in monoclonal sandwich tests wAbstractx.
observational study in 5310 patients on potentially misleading Clin Chem 1990;36:1093.
thyroid-stimulating hormone and gonadotropin immunoassay 22. Tatarewicz S, Miller JM, Swanson SJ, Moxness MS. Rheu-
results. Clin Chem 2002;48:2023–9. matoid factor interference in immunogenicity assays for human
7. Morgan BR, Tarter TH. Serum heterophile antibodies interfere monoclonal antibody therapeutics. J Immunol Methods 2010;
with prostate specific antigen test and result in over treatment 357:10–6.
in a patient with prostate cancer. J Urol 2001;166:2311–2. 23. Bjerner J, Olsen KH, Bormer OP, Nustad K. Testing and vali-
8. Rotmensch S, Cole LA. False diagnosis and needless therapy dating a homogeneous immunometric assay for interference.
of presumed malignant disease in women with false-positive Clin Chem Lab Med 2004;42:208–14.
human chorionic gonadotropin concentrations. Lancet 2000; 24. Bormer OP. Interference of complement with the binding of
355:712–5. carcinoembryonic antigen to solid-phase monoclonal antibod-
ies. J Immunol Methods 1989;121:85–93.
9. Henry N, Sebe P, Cussenot O. Inappropriate treatment of pros-
25. Ismail AA, Ismail AA, Ismail Y. Probabilistic Bayesian reason-
tate cancer caused by heterophilic antibody interference. Nat
ing can help identifying potentially wrong immunoassays
Clin Pract Urol 2009;6:164–7.
results in clinical practice: even when they appear’not-unrea-
10. Spencer DV, Nolte FS, Zhu Y. Heterophilic antibody interfer-
sonable’. Ann Clin Biochem 2011;48(Pt 1):65–71.
ence causing false-positive rapid human immunodeficiency
26. Ellis MJ, Livesey JH. Techniques for identifying heterophile
virus antibody testing. Clin Chim Acta 2009;399:121–2.
antibody interference are assay specific: study of seven analytes
11. Pernet P, eteau-Burnat B, Hermand C, Vaubourdolle M. Point-
on two automated immunoassay analyzers. Clin Chem 2005;
of-care testing: false elevation of cardiac troponin I assayed in 51:639–41.
the emergency department. Am J Emerg Med 2008;26:969–72. 27. Ismail AA. Interference from endogenous antibodies in auto-
12. Giovanella L, Ceriani L. Spurious increase in serum chromo- mated immunoassays: what laboratorians need to know. J Clin
granin A: the role of heterophilic antibodies. Clin Chem Lab Pathol 2009;62:673–8.
Med 2010;48:1497–9. 28. Ismail AA. On the interpretation of affirmative follow-up tests
13. Despres N, Grant AM. Antibody interference in thyroid assays: in immunoassays: what must not be done? Ann Clin Biochem
a potential for clinical misinformation. Clin Chem 1998;44: 2006;43(Pt 4):249–51.
440–54. 29. DeForge LE, Loyet KM, Delarosa D, Chinn J, Zamanian F,
14. Hunter WM, Budd PS. Circulating antibodies to ovine and Chuntharapai A, et al. Evaluation of heterophilic antibody
bovine immunoglobulin in healthy subjects: a hazard for immu- blocking agents in reducing false-positive interference in immu-
noassays. Lancet 1980;2:1136. noassays for IL-17AA, IL-17FF, and IL-17AF. J Immunol
15. Webster R, Fahie-Wilson M, Barker P, Chatterjee VK, Halsall Methods 2010;362:70–81.
DJ. Immunoglobulin interference in serum follicle-stimulating 30. Warren DJ, Bjerner J, Paus E, Bormer OP, Nustad K. Use of
hormone assays: autoimmune and heterophilic antibody inter- an in vivo biotinylated single-chain antibody as capture reagent
ference. Ann Clin Biochem 2010;47(Pt 4):386–9. in an immunometric assay to decrease the incidence of inter-
16. Stenman UH. Improving immunoassay performance by anti- ference from heterophilic antibodies. Clin Chem 2005;51:
body engineering. Clin Chem 2005;51:801–2. 830–8.
Copyright of Clinical Chemistry & Laboratory Medicine is the property of De Gruyter and its content may not
be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written
permission. However, users may print, download, or email articles for individual use.