1682 Bizzaro et al.: Methods for anti-Scl70 Antibodies Compared
the type of laboratory involved and the reagents used. In antibody determinations for a total of 2389 results. The
addition, most of these studies involved selected reference number of series exceeds that of the participants because
laboratories that used home-made methods, whereas several laboratories used more than one method: 42
most clinical laboratories use only commercial reagents. laboratories used ELISA methods (61.8% of the tests
In recent years, the use of antigenic substrates obtained conducted); 21 used IB (30.9%), 3 used counterimmuno-
with molecular biology techniques and the optimization electrophoresis (CIE; 4.4%), and 2 used the dot-blot
of methods for the extraction and purification of native method (2.9%). The reagents were purchased from 23
antigens have contributed to a more precise characteriza- different manufacturers (listed in Table 1). Only one
tion and identification of these and other autoantibodies laboratory used an in-house CIE.
indicative of autoimmune diseases. Moreover, the contin- Statistical analyses were conducted by calculating the
uous development of analytical techniques has made sensitivity of each method and reagent set in detecting
possible the use of new commercial tests for the determi- anti-Scl70 antibodies in the 16 diffuse cutaneous SSc
nation of the various autoantibody specificities in an ever samples, and the specificity in the other 20 samples in
increasing number of laboratories, thus requiring contin- which this antibody was absent. We also evaluated ana-
uous verification of quality. For this reason, the Study lytical specificity, i.e., the number of false positives that
Group on the Diagnosis of Autoimmune Diseases of the were recorded for other anti-ENA antibodies (e.g., anti-
Italian Society of Laboratory Medicine has performed an RNP, anti-Sm, anti-Ro, anti-La, and anti-Jo1) in sera in
extensive multicenter study involving hospital clinical which such positivities were not expected. Moreover, in
laboratories to analyze serum samples from patients af- reference to the IB methods, data were elaborated by
fected by systemic sclerosis, using numerous commercial separately evaluating the two different types of commer-
methods and reagent set to define present-day reliability cially available reagents: the simplified IB (for which we
in determining anti-topoisomerase I autoantibodies. prefer to use the terms line-blot or bar-blot) (16 ), in which
Materials and Methods
Fifty clinical laboratories in 14 Italian regions participated Table 1. Manufacturers of anti-ENA reagents and number of
in this study that included the determination of anti- participating laboratories using the systems.
extractable nuclear antigen (anti-ENA)10 antibodies in 36 No. of labs
serum samples, 27 of which were obtained from patients ELISA
with a clinical diagnosis of scleroderma/systemic sclero- Bio-Rad, Hercules, CA 1
sis (SSc), according to the criteria of the American College Chematil, Scafati, Italy 1
of Rheumatology (15 ). Sixteen of these samples were Clarck, Wicklow, UK 2
obtained from subjects with diffuse cutaneous SSc (anti- CLS, Ely-Cambridgeshire, UK 2
Scl70 positive and anti-centromere negative), and 11 were Cogent, Peniculk, UK 3
from patients with limited cutaneous SSc (anti-Scl70 neg- Diamedix, Miami, FL 5
ative and anti-centromere positive). Two diffuse cutane- Euroimmun, Lubeck, Germany 2
ous SSc sera also displayed a simultaneous reactivity for Fenning, Freiburg I.Br., Germany 2
Ro/SSA. The expected antibody specificity was defined Helix, West Sacramento, CA 1
not only in reference to the clinical diagnosis, but also on Immuno Concepts, Sacramento, CA 1
the basis of preliminary results obtained with two differ- Imtec, Zepernick, Germany 1
ent methods [immunoenzymatic (ELISA) and immuno- Incstar, Stillwater, MN 1
blot (IB) assays] in the four laboratories that collected the Medic, Pavone Canavese, Italy 3
Pharmacia & UpJohn, Elias 6
test samples; only the samples that were positive with Division, Freiburg I.Br., Germany
both methods were included in the study. The control Shield, Dundee, UK 9
group consisted of sera from three patients with systemic Zeus, Raritan, NJ 1
lupus erythematosus (one with anti-RNP and anti-Sm, IB
one with anti-RNP, and one with anti-Ro/SSA), three Autoimmune Diagnostika, 4
patients with Epstein-Barr virus infection, and three Strassburg, Germany
healthy subjects. The aliquoted sera were stored at 85 °C Euroimmun, Lubeck, Germany 6
until they were shipped. The participating laboratories Imtex, Zepernick, Germany 3
were not given any clinical information and were asked to Innogenetics, Temse, Belgium 4
use their usual reagents and analytical systems to deter- MarDx, Carlsbad, CA 2
mine anti-ENA antibodies. Nurex, Sassari, Italy 1
The 50 participating laboratories conducted 68 series of Scimdex, Denville, NJ 1
Binding Site, Birmingham, UK 2
Nonstandard abbreviations: ENA, extractable nuclear antigen; SSc,
Alphadia, Wavre, Belgium 2
systemic sclerosis; IB, immunoblot; and CIE, counterimmunoelectrophoresis.
Clinical Chemistry 46, No. 10, 2000 1683
optimal quantities of ENAs are deposited on the nitrocel-
Table 3. False positives and false negatives for anti-Scl70
lulose supports and in predefined positions to facilitate
antibody, according to commercial ELISA method.
the reading and interpretation of the results, and the
False positives False negatives
classical Western blot, in which the entire electrophoretic
course is instead present, with all the extractable antigens Manufacturer % n % n
from the cell material used. Bio-Rad 0.0 (0/20) 0.0 (0/16)
Chematil 0.0 (0/20) 0.0 (0/16)
Results Clarck 0.0 (0/40) 0.0 (0/32)
For all of the methods used, the sensitivity and specificity CLS 5.0 (2/40) 6.2 (2/32)
for anti-Scl70 were 96.9% and 98.6%, respectively (Table Cogent 0.0 (0/60) 8.3 (4/48)
2). Detailed analyses of the performances of the single Diamedix 0.0 (0/100) 1.2 (1/80)
reagent set, subdivided according to method, are shown Pharmacia/Elias 0.0 (0/107) 6.2 (5/80)
Euroimmun 2.5 (1/40) 0.0 (0/32)
in Tables 3-T4-T5, 4, and 5 for ELISA, IB, and CIE
Fenning 0.0 (0/39) 12.9 (4/31)
Helix 2.5 (1/40) 0.0 (0/32)
Immuno Concepts 0.0 (0/20) 6.2 (1/16)
elisa method (1481 results) Imtec 5.0 (1/20) 0.0 (0/16)
For anti-Scl70 detection, the specificity was excellent Incstar 0.0 (0/20) 0.0 (0/16)
(99.2%) and the sensitivity was good (97.2%). Although it Medic 0.0 (0/60) 3.1 (1/32)
was reported that the systems using recombinant antigens Shield 1.1 (2/180) 0.0 (0/144)
have a better sensitivity than those using native antigens Zeus 0.0 (0/20) 0.0 (0/16)
(17 ), this observation was not confirmed in this study; the
mean sensitivity of the 30 laboratories that used native
antigens was better than that obtained by the 12 labora-
with in-house reagents whose sensitivity was seen to be
tories that used recombinant antigens (98.3% vs 89.4%).
entirely insufficient (68.7% false negatives; Table 5). When
Specificity data obtained with the two types of antigen
this in-house method was excluded and only the commer-
preparations overlapped (99.4% vs 98.3%).
cial ones considered, sensitivity and specificity (93.8% and
95%, respectively) were comparable to other methods.
ib method (729 results)
The performance of the reagents utilized in this method dot-blot method (71 results)
was slightly lower than that of the ELISA methods:
The best performance was obtained with this method,
specificity was 97.6%, and sensitivity was 96.1% for anti-
which showed 100% sensitivity and specificity. However,
Scl70. Bar-blot and Western-blot methods did not demon-
this finding should be considered with caution because of
strate different specificities (false positives, 2.4% vs 2.5%)
the small number of laboratories and observations in-
or sensitivities (false negatives, 4.3% vs 3.5%) for anti-
Scl70 (Table 4); however, remarkable differences emerged
regarding analytical specificity because of the high num-
ber of false positives for anti-ENA antibodies other than
Anti-topoisomerase I antibody is specific for patients
anti-Scl70 obtained with the Western-blot method (23%)
affected by scleroderma (18 ). The prevalence is 40%, but
compared with the bar-blot method (7.8%; Table 6).
the range is very wide (3–75%) (19, 20 ), and highest
cie method (108 results)
Although specificity was quite good (96.7%), sensitivity
was poor (72.9%) because of the single test conducted Table 4. False positives and false negatives for anti-Scl70
antibody, according to IB method.
False positives False negatives
Table 2. False positives and false negatives for anti-Scl70
antibody, according to the method used by the Method Manufacturer % n % n
participating laboratories. BBa Euroimmun 0.9 (1/106) 4.8 (4/83)
Method Innogenetics 5.0 (4/80) 4.6 (3/64)
Scimdex 0.0 (0/20) 0.0 (0/16)
ELISA IB Dot CIEa Total Total 2.4 (43/441) 4.3 (7/163)
False positives, 0.8 2.4 0.0 5.0 1.4 WB Autoimmune 3.7 (3/80) 7.8 (5/64)
% (n) (7/826) (10/406) (0/39) (2/40) (19/1311) Diagnostika
False negatives, 2.8 3.9 0.0 6.2 3.1 Imtec 3.3 (2/60) 0.0 (0/48)
% (n) (18/639) (12/307) (0/32) (2/32) (32/1010) MarDx 0.0 (0/40) 0.0 (0/16)
Specificity, % 99.2 97.6 100 95.0 98.6 Nurex 0.0 (0/20) 0.0 (0/16)
Sensitivity, % 97.2 96.1 100 93.8 96.9 Total 2.5 (63/252) 3.5 (5/144)
Excludes results for assay with homemade reagents. BB, bar blot; WB, Western blot.
1684 Bizzaro et al.: Methods for anti-Scl70 Antibodies Compared
preparations using proteins expressed by recombinant
Table 5. False positives and false negatives for anti-Scl70
antibody, according to CIE method.
Concerning analytical specificity, good performance
False positives False negatives
was obtained with all of the methods evaluated with the
Manufacturer % n % n exception of Western blot, which demonstrated poor
Binding site 5.0 (2/40) 6.2 (2/32) specificity, suggesting that interpretative difficulties re-
In-house 0.0 (0/20) 68.7 (11/16) main the main problem in assigning antibody positivity.
Total 3.3 (2/60) 27.1 (13/48) That this difficulty in interpretation is attributable not
only to the capacity and experience of the operator, but
also to an intrinsic problem in the method appears evident
values are observed in patients with the diffuse form for from the following observation. One of the four producers
which it is the diagnostic antibody (21, 22 ). Its presence of Western-blot methods who furnished the material to
has been associated with a poorer prognosis and in- the laboratories has a computerized system for reading
creased pulmonary involvement (23, 24 ). The early and the strips by means of a scanner and a software applica-
correct identification of this antibody is thus crucial in tion for the interpretation of the strips. Of the four
both the diagnostic and prognostic phases of the disease laboratories that used this reagent, two read the strips
and requires very accurate data from the autoimmunol- with a manual method and two used an automated
ogy laboratory. The objective of this study, therefore, was system, and no substantial differences in terms of speci-
to evaluate the sensitivities and specificities of the various
ficity emerged; indeed, in both situations there was a high
methods/reagent sets now on the market and the vari-
number of false positives (27%; data not shown). This
ability among autoimmunology laboratories in assaying
means that even the use of automated systems cannot
the anti-topoisomerase I antibody.
resolve the problem inherent in this methodology, which
Unlike some previous studies that had difficulty in
undoubtedly is an extremely valid tool for the identifica-
identifying this type of antibody (6, 10, 12, 13 ), the excel-
tion of rare or atypical antibody patterns but has a lower
lent results in terms of both sensitivity and specificity
specificity than other methods.
obtained with all the methods/reagent sets used in this
The data for the CIE and dot-blot methods do not
study allow us to state that, in general, the commercial
systems for determining anti-Scl70 antibodies are reliable. permit a reliable judgment given the small number of
The immunoenzymatic methods, which constituted 60% laboratories that used these methods. Although CIE to-
of the methods/reagent sets used, were determinant in gether with double immunodiffusion has represented the
reaching such high efficiency and produced the best most widespread method for years, its further use seems
results. unlikely mostly because of its poor sensitivity for some
It is worth mentioning that the reagents made up of autoantibodies; on the other hand, the dot-blot method
native antigens (obtained from either HEp-2 human cells offers ease of performance as well as direct interpretation,
or bovine or rabbit thymus) yielded better results in terms and produced excellent results, features that predict its
of sensitivity than reagents containing recombinant anti- much wider use in clinical laboratories.
gens (obtained from both eukaryotic and prokaryotic In agreement with the recommendations of the Euro-
systems), whereas no substantial differences in terms of pean Consensus Group (6 ) and the guidelines proposed
specificity were observed. This means that the consistent by our group (25 ), which support the need to use two
improvement in the techniques of purification of the different methods to identify anti-ENA antibodies, the use
native antigens, at least as far as Scl70 antigen is con- of two methods in series should have increased the
cerned, has allowed the production of sufficiently pure sensitivity, reducing by one-half the number of false
preparations, devoid of significant antigenic contamina- negatives. Indeed, almost one-half of the false negatives
tion, while maintaining the characteristics of molecular recorded in this study occurred with only one of the four
identity unchanged. The sensitivity and specificity thus methods used.
obtained are comparable, if not superior, to those of From a careful examination of the results of this study,
Table 6. Analytical specificity (false-positive reactions for anti-ENA antibodies other than anti-Scl70) according to the
methods used by the participating laboratories.
ELISA BB WB Dot CIE Total
False positives, % (n) 3.7 7.8 23.0 7.0 2.7 7.4
(55/1481) (29/369) (83/360) (5/71) (2/72) (174/2353)
Analytical specificity, % 96.3 92.2 77.0 93.0 97.3 92.6
BB, bar blot; WB, Western blot.
Clinical Chemistry 46, No. 10, 2000 1685
it is also possible to draw other conclusions that are van Venrooij WJ, Maini RN, eds. Manual of biological markers of
helpful for a correct interpretation of the findings, as well disease. Dordrecht: Kluwer Academic Publishers, 1996:C5.1:1–7.
as applicable to all external quality assessment studies of 8. Gonzalez C, Martin T, Arroyo T, Garcia-Isidoro M, Navajo JA,
Gonzalez-Buitrago JM. Comparison and variation of different meth-
this type. First, the expected unreliability of an isolated
odologies for the detection of autoantibodies to nuclear antigens
laboratory that furnished results that were clearly anom- (ANA). J Clin Lab Anal 1997;11:388 –92.
alous compared with the general mean, and especially 9. Bridges AJ, Lorden TE, Havighurst TC. Autoantibody testing for
with other laboratories that used the same reagent, con- connective tissue diseases. Comparison of immunodiffusion,
firms that the variance related to the operator is an immunoblot, and enzyme immunoassay. Am J Clin Pathol 1997;
important part of the total variability. Second, despite the 108:406 –10.
high number of observations, the data are not distributed 10. Bizzaro N, Tozzoli R, Tonutti E, Piazza A, Manoni F, Ghirardello A,
et al. Variability between methods to determine ANA, anti-dsDNA
homogeneously among the various reagent sets, and great
and anti-ENA autoantibodies: a collaborative study with the bio-
caution must be taken in interpreting results obtained medical industry. J Immunol Methods 1998;219:99 –107.
with reagent sets for which few data are available. Finally, 11. Jaekel HP, Mueller D, Grobe N. Detectability of autoantibodies to
the lack of a reference method for the anti-ENA antibodies ENA: comparative investigation of different ELISA reagent set,
that would allow assignment of a reliable value is one of indirect immunofluorescence and immunoblot. Clin Lab 1998;44:
the main problems that must be faced in the planning and 673–9.
interpretation of results of external quality assessment 12. Bizzaro N, Bassetti D, Manoni F, Piazza A, Tozzoli R, Tonutti E, et
al. Sensitivity of commercial reagents for the detection of anti-ENA
studies; only the use of sure positive samples with a
antibodies. Clin Chem Lab Med 1999;37:480.
medium-high antibody titer can obviate this important 13. Tan EM, Smolen JS, McDougal JS, Butcher BT, Conn D, Dawkins R,
methodological bias, even if the use of test sera with low et al. A critical evaluation of enzyme immunoassays for detection of
antibody titer and values distributed near the threshold of antinuclear autoantibodies of defined specificity. I. Precision, sensi-
positivity of the methods is more advantageous for eval- tivity, and specificity. Arthritis Rheum 1999;42:455– 64.
uating the accuracy of the analytical systems. 14. Feltkamp TEW. Standards and reference preparations. In: van Ven-
rooij WJ, Maini RN, eds. Manual of biological markers of disease.
Dordrecht: Kluwer Academic Publishers, 1993:A11:1–12.
In conclusion, this study demonstrated that the commer-
15. Subcommittee for Scleroderma Criteria of the American Rheuma-
cial reagents used in clinical laboratories for the determi- tism Association Diagnostic and Therapeutic Criteria Committee.
nation of anti-topoisomerase I antibody have acceptable Preliminary criteria for the classification of systemic sclerosis
reliability. In particular, specificity was excellent (98.6%) (scleroderma). Arthritis Rheum 1980;23:581–90.
and sensitivity was good (96.9%). Among the various 16. Manoni F, Valverde S, Antico F, Giacomini A, Gessoni GL. Autoan-
methods, ELISA presents the best characteristics of accu- ticorpi anti-antigeni nucleari estraibili (ENA): metodi di studio. In:
racy; among the blot methods, Western blot lacks suffi- Tozzoli R, Bizzaro N, eds. La diagnostica di Laboratorio delle
Malattie Autoimmuni Sistemiche. Padova: Piccin Nuova Libraria;
cient analytical specificity. It is likely that the realization
of national external quality assessment programs will lead 17. Vazquez-Abad D, Rothfield NF. Topoisomerase-I (Scl-70) autoanti-
to a further improvement in analytical quality, evidencing bodies. In: Peter JB and Shoenfeld Y, eds. Autoantibodies.
and correcting the few cases of unreliability related to Amsterdam: Elsevier Science BV, 1996:830 –5.
reagents and/or laboratories. 18. Rothfield NF. Autoantibodies in scleroderma. Rheum Dis Clin
North Am 1992;18:483–98.
19. van Venrooij WJ, Stapel SO, Houben H, Kallenberg CGM, Penner E,
References van de Putte LB. Scl-86: a marker antigen for diffuse scleroderma.
1. Shero JH, Bordwell B, Rothfield NF, Earnshaw WC. High titers of J Clin Invest 1985;75:1053– 60.
autoantibodies to topoisomerase I (Scl-70) in sera from sclero- 20. Spencer-Green G, Alter D, Welch HG. Test performance in sys-
derma patients. Science 1986;231:737– 40. temic sclerosis: anti-centromere and anti-Scl70 antibodies. Am J
2. Weiner ES, Earnshaw WC, Senecal JL, Bordwell B, Johnson P, Med 1997;103:242– 8.
Rothfield NF. Clinical associations of anticentromere antibodies 21. Tsay GJ, Fann RH, Hwang J. Specificity of anti-Scl70 antibodies in
and antibodies to topoisomerase I. Arthritis Rheum 1988;31: scleroderma: increased sensitivity of detection using purified DNA
378 – 85. topoisomerase I from calf thymus. J Rheumatol 1990;17:1314 –9.
3. LeRoy EC, Black C, Fleischmajer R, Jablonska S, Krieg T, Medsger 22. Verheijen R, van den Hoogen F, Beijer R, Richter A, Penner E,
TA, et al. Scleroderma (systemic sclerosis): classification, sub- Habets WJ, et al. A recombinant topoisomerase I used for
sets, and pathogenesis. J Rheumatol 1988;15:202–5. antibody detection in sera from patients with systemic sclerosis.
Clin Exp Immunol 1990;80:38 – 43.
4. Wilson L. Cost-of-illness of scleroderma: the case for rare dis-
23. Jacobsen S, Halberg P, Ullman S, Høier-Madsen M, Petersen J,
eases. Semin Arthritis Rheum 1997;27:73– 84.
Mortensen J, et al. A longitudinal study of pulmonary function in
5. Hildebrandt S, Weiner ES, Senecal JL, Noell GS, Earnshaw WC, Danish patients with systemic sclerosis. Clin Rheumatol 1997;
Rothfield NF. Autoantibodies to topoisomerase I (Scl70): analysis 16:384 –90.
by gel diffusion, immunoblot, and enzyme-linked immunosorbent 24. Steen VD, Medsger TA. Survival of patients with systemic and
assay. Clin Immunol Immunopathol 1990;57:399 – 410. scleroderma specific autoantibodies. Arthritis Rheum 1999;42:
6. van Venrooij WJ, Charles P, Maini RN. The consensus workshop S189.
for the detection of autoantibodies to intracellular antigens in 25. Tozzoli R, Bizzaro N, Bassetti D, Manoni F, Piazza A, Pradella M, et
rheumatic diseases. J Immunol Methods 1991;140:181–9. al. Linee guida per la diagnosi e il monitoraggio delle malattie
7. Rothfield NF. Autoantibodies to scleroderma-associated antigens. In: reumatiche autoimmuni. Med Lab 1999;7:124 –32.