Clinical Chemistry
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Clinical Chemistry 47: 1649-1659, 2001;
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (15)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Hayashi, N.
Right arrow Articles by Kumagai, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hayashi, N.
Right arrow Articles by Kumagai, S.
Related Collections
Right arrow Clinical Immunology
Right arrow Evidence Based Laboratory Medicine and Test Utilization
(Clinical Chemistry. 2001;47:1649-1659.)
© 2001 American Association for Clinical Chemistry, Inc.

Articles

Detection of Antinuclear Antibodies by Use of an Enzyme Immunoassay with Nuclear HEp-2 Cell Extract and Recombinant Antigens: Comparison with Immunofluorescence Assay in 307 Patients

Nobuhide Hayashi1,2, Tomoko Kawamoto2, Masahiko Mukai2, Akio Morinobu1, Masahiro Koshiba1, Shinichi Kondo1,2, Soichiro Maekawa1 and Shunichi Kumagai1,2a

1 Department of Clinical and Laboratory Medicine, Kobe University School of Medicine, 7-5-1 Kusunoki-cyo, Chuo-ku, Kobe, Hyogo 650-0017, Japan

2 Department of Clinical Laboratory, Kobe University Hospital, 7-5-2 Kusunoki-cyo, Chuo-ku, Kobe, Hyogo 650-0017, Japan.

aAddress correspondence to this author at: Department of Clinical and Laboratory Medicine, Kobe University School of Medicine, 7-5-1 Kusunoki-cyo, Chuo-ku, Kobe, Hyogo 650-0017, Japan. Fax 81-78-382-6209; e-mail kumagais{at}kobe-u.ac.jp.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Background: A new enzyme immunoassay (EIA) for automated detection of antinuclear antibodies (ANAs) uses a mixture of HEp-2 cell extracts and multiple recombinant nuclear antigens immobilized on beads. We compared this EIA and an immunofluorescence (IF) assay in a large group of patients and controls.

Methods: We studied 492 healthy individuals and 307 patients with connective tissue diseases (CTDs). Sera were tested by an automated EIA (COBAS® Core HEp2 ANA EIA; Roche Diagnostics) and IF. Samples were also tested for eight disease-specific antibodies, including antibodies against U1RNP, Sm, SSA/Ro, SSB/La, Scl-70, Jo-1, dsDNA, and centromere.

Results: Areas under ROC curves for the EIA were greater than (P = 0.008–0.012) or numerically identical to areas for the IF method for each of six CTDs studied. ROC areas for EIA were 0.98 (95% confidence interval, 0.95–0.99), 0.99 (0.96–1.00), and 0.99 (0.98–1.00) in systemic lupus erythematosus (n = 111), systemic sclerosis (n = 39), and mixed connective tissue disease (n = 33), respectively. For all 258 CTD patients with conditions other than rheumatoid arthritis (RA), the sensitivity and specificity of the IF method at a cutoff dilution of 1:40 were 92% and 65%, respectively, vs 93% and 79% for the EIA at a cutoff of 0.6. For the IF method at a cutoff dilution of 1:160, sensitivity and specificity were 81% and 87%, respectively, vs 84% and 94%, respectively, for the EIA at a cutoff of 0.9. For 207 sera containing at least one of eight disease-specific ANAs, positivities for the EIA and the IF method were 97.1% and 97.6%, respectively, at cutoffs of 0.6 and 1:40 (P = 0.76).

Conclusions: An EIA that can be performed by a fully automated instrument distinguishes CTDs (except RA) from healthy individuals with both higher sensitivity and specificity than the IF method when the cutoff index was set at 0.9. Moreover, it can be used to exclude the presence of disease-specific ANAs by setting the cutoff index at 0.6 with almost the same efficacy as the IF method.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Detection of antinuclear antibodies (ANAs) 1 is important to the diagnosis of connective tissue diseases (CTDs), such as systemic lupus erythematosus (SLE). In 1957, Friou (1) developed an immunofluorescence-based method (IF-ANA), and it has been widely used for detection of ANAs in general (a screening test for ANAs as a whole). Later, HEp-2 cells were used as nuclear substrates that allowed detection of anti-centromere antibodies much more sensitively, thus contributing to improvement of the IF-ANA (2)(3). In spite of this advantage, this method gives high (false) positive responses among healthy individuals and variability in results between institutions (4). In addition, the procedure is complicated and requires training, and the results are subjective because they are judged visually (5).

Enzyme immunoassay (EIA) methods that allow detection of individual disease-specific ANAs have recently been developed by use of antigens extracted from HEp-2 cells and/or specific purified or recombinant antigens (6)(7). However, large differences in sensitivity and response to disease-specific ANAs have been reported, depending on the antigens used (8). A new EIA method uses beads (rather than microtiter plates) on which antigens extracted from HEp-2 cells and multiple recombinant antigens are immobilized (9). In this study, we evaluated the usefulness of this new EIA method as a screening test for ANAs, in comparison with the conventional IF-ANA, using several hundred sera from patients diagnosed with CTDs and strictly selected healthy individuals.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
measurement of ANAs
The COBAS® Core HEp2 ANA EIA (COBAS-ANA; Roche Diagnostics, Mannheim, Germany), an EIA reagent set for ANA detection, was used to detect ANAs with a COBAS Core II (Roche Diagnostics), an automated analyzer for EIAs. In principle, this is a two-step EIA in which specimens are incubated with polystyrene beads coated with a mixture of a soluble HEp2 cell extract and seven recombinant proteins (60-kDa SSA/Ro, 52-kDa SSA/Ro, SSB/La, Scl-70, Jo-1, dsDNA, and CENP-B, which is a major autoantigen corresponding to anti-centromere antibodies) (9). After removal of unbound antibody, the antigen-coated beads are allowed to react with peroxidase-labeled anti-human IgG mouse monoclonal antibody, and sandwich complexes are formed. H2O2–3,3',5,5'-tetramethylbenzidine is added to initiate enzymatic reactions after washing, and the absorbance is measured. The results are shown as an index calculated using the following formula: index = (absorbance of a test sample)/(mean absorbance of negative control + 0.1). In accordance with the operating manual, the assessment criteria were set as follows: positive if index >=1.1, negative if index <0.9, and equivocal if >=0.9 but <1.1. All serum samples were analyzed in duplicate. The intra- (n = 20) and interassay (20 days) CVs of the COBAS-ANA were 4.5–5.9% and 3.2–5.6%, respectively.

The IF-ANA method used was the FANAwell (DIA-IATRON, Tokyo, Japan), a reagent set that uses HEp-2 cells as substrates and a mixture of fluorescein isothiocyanate-labeled mouse anti-human immunoglobulins (IgG, IgA, and IgM) as secondary antibody. Specimens were diluted 20- and 40-fold with the buffer solution included in the reagent set, and if 40-fold-diluted specimens were still positive in the IF-ANA, samples were further diluted up to 1:80–1:5120 for reanalysis. The antibody titer was expressed as the final dilution, regardless of the type of staining. Slides were stained according to the instructions in the operating manual, and they were examined with a BV-excitation type, epifluorescent microscope at x40 (dry) magnification. All slides were examined under the microscope by two laboratory technologists. For quality control, negative and positive controls with known antibody titers were used for each assay to check accuracy.

We analyzed eight disease-specific ANAs, using EIA-based reagent sets with the indicated antigens: anti-U1RNP antibody [MESACUP® RNP-II TEST; Medical & Biological Laboratories (MBL), Nagoya, Japan; antigens were recombinant 70-kDa A-peptide and C-peptide]; anti-Sm antibody (MESACUP Sm TEST; MBL; purified Sm antigens); anti-SSA/Ro antibody (MESACUP SS-A/Ro TEST; MBL; recombinant 60-kDa SSA/Ro antigen and purified SSA/Ro protein); anti-SSB/La antibody (MESACUP SS-B/La TEST; MBL; recombinant SSB/La protein); anti-Scl-70 antibody (MESACUP Scl-70 TEST; MBL; recombinant Scl-70 protein); anti-centromere antibody (MESACUP CENP-B TEST; MBL; recombinant CENP-B protein); anti-Jo-1 antibody (MESACUP Jo-1 TEST; MBL; recombinant Jo-1 protein); and anti-dsDNA antibody (MESACUP DNA-II TEST; MBL, {lambda}-phage-derived purified double-stranded DNA).

selection of patients and controls
It is important to establish the inclusion criteria for healthy individuals carefully so that the reference interval for ANAs can be set correctly. In other words, the reference interval may not be determined accurately if all ANA-positive individuals are excluded from the analysis because some healthy individuals are known to be positive for ANAs (4). The healthy individuals in this study were selected from volunteers by the following procedures. (a) Individuals who did not exhibit any abnormal finding on any of the test items during a routine physical examination were selected from the hospital staff and the residents of a certain town in Hyogo Prefecture, Japan. The test items were urinalysis, routine hematologic study, blood chemistry tests, and a plain chest film, but ANAs were not included. (b) The results obtained by COBAS-ANA were analyzed in these individuals. For individuals positive in the COBAS-ANA, although most of them were also positive in the IF-ANA, a survey was conducted by questionnaire, and further examinations were performed, if needed. (c) After persons who were diagnosis as having a CTD or were suspected of having CTD were excluded, we assumed that all other individuals were healthy. These healthy individuals were divided into 12 groups based on gender and age. To standardize the groups, whenever >45 individuals were enrolled in a group, 45 were selected at random. As a result, a total of 492 healthy individuals were included in this study, and the genders and ages of the healthy individuals were as follows: 43 women and 44 men 21–30 years of age, 45 women and 44 men 31–40 years of age, 45 women and 45 men 41–50 years of age, 45 women and 36 men 51–60 years of age, 45 women and 41 men 61–70 years of age, and 34 women and 25 men 71–80 years of age. The ratio of women to men among the healthy individuals was 1.1:1.

We examined sera from 307 CTD patients who were followed at the outpatient clinic of Kobe University Hospital. Serum samples were collected without conscious bias mostly at patients’ first visits during 1996 to 1999. The patients’ diseases had been defined by the established criteria for each disease (10)(11)(12)(13)(14)(15)(16). The numbers and genders of the patients were as follows: 111 patients with SLE (median age, 35 years; range, 17–72 years; 104 women, 7 men); 36 patients with Sjögren syndrome (SS; median age, 59 years; range, 21–77 years; 35 women, 1 man); 39 patients with systemic sclerosis (SSc; median age, 52 years; range, 35–73 years; 35 women, 4 men); 39 patients with dermatomyositis and polymyositis (DM/PM; median age, 54 years; range, 16–79 years; 27 women, 12 men); 33 patients with mixed connective tissue disease (MCTD; median age, 39 years; range, 16–65 years; 30 women, 3 men); and 49 patients with rheumatoid arthritis (RA; median age, 50 years; range, 25–76 years; 44 women, 5 men).

Each sample was obtained with informed consent. After separation of the serum by centrifugation, the serum samples were stored frozen at below -40 °C. All samples collected from CTD patients and healthy individuals were randomly numbered and were blindly examined for ANAs by both the COBAS-ANA and the IF-ANA and were also analyzed for the eight disease-specific ANAs by each EIA-based reagent set.

data analysis
ROC analyses were performed for each patient group, using healthy women as the reference group (17)(18). The sensitivity and specificity of the IF-ANA were determined at cutoff dilutions of 1:40 and 1:160, and those of the COBAS-ANA at cutoff indexes of 0.6 and 0.9. For computation and analysis of ROC curves, we used the software program MedCalc, Ver. 6.01 (MedCalc Software).

The Mann–Whitney test was used to assess differences in COBAS-ANA antibody titers among CTD patients, whereas gender-based differences in titers among healthy individuals were tested by the two-sample t-test with the Welch correction. In addition, age-related differences in the titers were tested by the Kruskal–Wallis test. Correlations between COBAS-ANA and IF-ANA positivity among CTD patients were tested by the {chi}2 test. Concordance between the two methods was determined by the {kappa} statistic (19).

The present report followed the guidelines proposed by Bruns et al. (20) for reporting of studies of diagnostic accuracy of medical tests.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
cobas-ana and if-ana positivity in healthy individuals
COBAS-ANA antibody indexes were 0.47 ± 0.45 (mean ± SD) in the 492 healthy individuals, 0.51 ± 0.58 in the women, and 0.43 ± 0.25 in the men, demonstrating significantly higher antibody titers in women than men (P = 0.042, t-test with Welch correction). Stratification of the COBAS-ANA antibody titers by age (every 10 years) showed significant differences among the six age groups in women but not in men (P = 0.016 for women and 0.38 for men, Kruskal–Wallis test).

The IF-ANA was positive in 34.6% of the women and 18.3% of the men on the basis of a cutoff dilution of 1:40, showing a significant difference between women and men (P <0.0001, comparison of two proportions). The COBAS-ANA was positive in 6.2% of the women and 3.8% of the men on the basis of a cutoff index of 0.9, showing no significant difference in positivity between women and men (P = 0.23, comparison of two proportions; Table 1 ).


View this table:
[in this window]
[in a new window]
  		Table 1. Comparison of positivity of the COBAS-ANA and the IF-ANA in healthy individuals by gender.

Thus, there were gender-related differences in COBAS-ANA antibody titers and in IF-ANA positivity among healthy individuals. Because most of the patients in the CTD group were women, healthy women were used as controls in the subsequent studies.

cobas-ana antibody titers and positivity in ctd patients
The COBAS-ANA antibody titers of the six CTD patient groups are shown in Fig. 1 . The median indexes (interquartile range) were 5.20 (1.93–11.83) in 111 SLE patients, 2.50 (1.10–7.80) in 36 SS patients, 4.00 (1.53–7.28) in 39 SSc patients, 0.80 (0.53–1.75) in 39 DM/PM patients, 4.50 (3.25–6.43) in 33 MCTD patients, and 0.60 (0.50–0.83) in 49 RA patients. The indexes were significantly higher in all patient groups than in healthy women (P <0.0001, Mann–Whitney test).



View larger version (15K):
[in this window]
[in a new window]
  		Figure 1. Antibody titers in the COBAS-ANA for six groups of CTD patients and for healthy individuals (Normals).

Filled circles, CTD patients positive for at least one of eight disease-specific ANAs, including anti-U1RNP antibody, anti-Sm antibody, anti-SSA/Ro antibody, anti-SSB/La antibody, anti-Scl-70 antibody, anti-Jo-1 antibody, anti-centromere antibody, and anti-dsDNA antibody; open circles, individuals negative for all of these ANAs (the small and large open circles show 1 case and 50 cases, respectively). Solid lines indicate the respective medians. Dashed line indicates the cutoff index of 0.9. P values indicate significant differences in COBAS-ANA antibody titers between healthy individuals and CTD patients by the Mann–Whitney test.

roc analyses for cobas-ana and if-ana
ROC analyses were performed for the COBAS-ANA and IF-ANA in the six groups of CTD patients (patient group) and healthy women (reference group; Fig. 2 and Table 2 ). The ROC curves for the COBAS-ANA in the patients with SLE, SS, or SSc in Fig. 1Up are to the left and above those for the IF-ANA, and the areas under the ROC curves (AUCs) for the COBAS-ANA and IF-ANA were 0.98 [95% confidence interval (95% CI), 0.95–0.99] and 0.93 (95% CI, 0.90–0.95), respectively, in SLE patients and 0.93 (95% CI, 0.89–0.96) and 0.82 (95% CI, 0.77–0.86), respectively, in the SS patients, showing significant differences in AUC (P = 0.008 for SLE; P = 0.012 for SS). The AUCs for the COBAS-ANA and IF-ANA were 0.99 (95% CI, 0.96–1.00) and 0.96 (95% CI, 0.93–0.98), respectively, in the SSc patients, showing no significant differences in AUC (P = 0.31). In the patients with DM/PM, MCTD, or RA, both ROC curves almost overlapped, with no significant difference in AUC. In the RA patients, both ROC curves were situated almost diagonally and linear, suggesting that both the COBAS-ANA and the IF-ANA were hardly able to distinguish RA patients from healthy individuals.



View larger version (50K):
[in this window]
[in a new window]
  		Figure 2. ROC analyses for the COBAS-ANA and IF-ANA.

ROC analyses were performed for six groups of CTD patients (each CTD as a different patient group) and healthy women [reference group (Control)]: (A), SLE patients vs controls; (B), SS patients vs controls; (C), SSc patients vs controls; (D), DM/PM patients vs controls; (E), MCTD patients vs controls; (F), RA patients vs controls. In addition, another ROC analysis was performed for all CTD patients excluding RA patients (All disease except RA) and healthy women (Control; G). Solid lines indicate COBAS-ANA with cutoff indexes of 0.6 (•) and 0.9 ({circ}); dotted lines indicate IF-ANA with cutoff dilutions of 1:40 ({blacksquare}) and 1:160 ({square}). The calculated areas under the ROC curves (AUCs) and P values are presented in Table 2Up .


View this table:
[in this window]
[in a new window]
  		Table 2. AUCs for ROC curves shown in Fig. 2Up : Comparison of AUCs of COBAS-ANA and IF-ANA.

ROC analysis was then performed for the COBAS-ANA and IF-ANA in all of the CTD patients except those with RA. The ROC curve for the COBAS-ANA was situated above the curve for the IF-ANA, and the AUCs for the COBAS-ANA and IF-ANA were 0.94 (95% CI, 0.92–0.96) and 0.90 (95% CI, 0.87–0.93), respectively, showing a significant difference (P = 0.005). As evidenced by a sensitivity of 92% (95% CI, 87–95%) and a specificity of 65% (95% CI, 59–71%), the IF-ANA had high sensitivity but poor specificity at the cutoff dilution of 1:40 that has been widely used to date (Table 3 ). However, the sensitivity and specificity were better balanced at a cutoff dilution of 1:160 (sensitivity, 81%; 95% CI, 76–86%; specificity, 87%; 95% CI, 82–91%).


View this table:
[in this window]
[in a new window]
  		Table 3. Test characteristics of the COBAS-ANA and the IF-ANA for cutoff points.

The COBAS-ANA had well-balanced sensitivity and specificity at cutoff indexes between 0.6 and 0.9, and its sensitivity and specificity were 84% (95% CI, 79–88%) and 94% (95% CI, 90–96%), respectively, at a cutoff index of 0.9. A cutoff index of 0.8 was found to be suitable for patients with SLE, SS, SSc, or MCTD, whereas a cutoff index of 0.7 was found to be most suitable for patients with DM/PM. The COBAS-ANA yielded a sensitivity of 93% (95% CI, 89–95%) and a specificity of 79% (95% CI, 74–84%) at a cutoff index of 0.6, thus showing that the COBAS-ANA with a cutoff index of 0.6 was more sensitive and more specific than the IF-ANA at a cutoff dilution of 1:40. The COBAS-ANA at a cutoff index of 0.9 yielded a likelihood ratio of 13.5 (95% CI, 10.2–18.0) for the CTD patients, excluding those with RA, whereas the IF-ANA at a cutoff dilution of 1:160 yielded a likelihood ratio of 5.9 (95% CI, 4.6–7.7) for CTD patients, excluding those with RA.

correlation between cobas-ana and if-ana titers in ctd patients
A significant correlation between COBAS-ANA and IF-ANA titers was found in the sera of CTD patients (rS = 0.647; P <0.0001, Spearman correlation coefficient; not shown). Comparisons between IF-ANA and COBAS-ANA titers were performed in the CTD patient groups with the COBAS-ANA cutoff indexes set at 0.6 and 0.9 and the IF-ANA cutoff dilutions set at 1:40 and 1:160 (Table 4 ). All patients with MCTD showed high antibody titers by both the COBAS-ANA and the IF-ANA. Among the patients with SLE, SS, or SSc, there was significant correlation between IF-ANA and COBAS-ANA titers (P = 0.003 for SLE; P = 0.002 for SS; P = 0.021 for SSc), but the {kappa} statistic of agreement was <0.4, except in SS, thus showing relatively poor agreement. Among the patients with DM/PM or RA, there was no significant agreement between the COBAS-ANA and IF-ANA (P = 0.20 for DM/PM; P = 0.25 for RA).


View this table:
[in this window]
[in a new window]
  		Table 4. Comparison of the COBAS-ANA and the IF-ANA for measurement of ANAs in patients with CTD.

Several samples showed extreme discrepancies between the COBAS-ANA index and IF-ANA titer. In 11 of the 307 patients with CTD, the IF-ANA titer was >=1:160 but the COBAS-ANA index was <0.6, whereas in another 10 patients, the IF-ANA titer was <1:40 but the COBAS-ANA index was >=0.9. The characteristics of the former 11 CTD patients are listed in Table 5 ; of these patients, 5 (45%) had RA and 4 (36%) had DM/PM. Four of these patients were positive for at least one disease-specific ANA, including one patient with RA who was positive for both anti-SSA/Ro and anti-dsDNA antibodies. The characteristics of the latter 10 CTD patients are shown in Table 6 ; of these patients, 3 (30%) had DM/PM and 3 (30%) had SS. Three of these patients were positive for at least one disease-specific ANA, including one patient with DM/PM who was positive for anti-Jo-1 antibody with an index of 85.9.


View this table:
[in this window]
[in a new window]
  		Table 5. Characteristics of the patients with CTD who exhibited discrepancies of COBAS-ANA titer and IF-ANA titer (COBAS-ANA <0.6; IF-ANA >=1:160).


View this table:
[in this window]
[in a new window]
  		Table 6. Characteristics of patients with CTD who exhibited discrepancies of COBAS-ANA titer and IF-ANA titer (COBAS-ANA >=0.9; IF-ANA <1:40).

To determine the correlation between COBAS-ANA antibody titers and IF-ANA staining pattern, we selected 140 sera from CTD patients whose IF-ANA antibody titers were >1:40 and with a single staining pattern, including 36 cases with a homogeneous pattern, 88 cases with a speckled pattern, 6 cases with a nucleolar pattern, and 10 cases with a discrete speckled pattern. At a cutoff index of 0.6, the COBAS-ANA was positive in 78% (28 of 36), 95% (84 of 88), 50% (3 of 6), and 90% (9 of 10) of the cases with the homogeneous, speckled, nucleolar, and discrete speckled patterns, respectively.

sensitivity of cobas-ana and if-ana for detection of disease-specific ANAs
Among 492 healthy individuals, 16 samples were demonstrated to be positive for at least one of the eight disease-specific ANAs by each EIA-based reagent set. The COBAS-ANA antibody titers were significantly higher in these 16 individuals than in the 476 specific-ANA-negative healthy individuals (P <0.0001, Mann–Whitney test). In addition, none of the healthy individuals negative for disease-specific ANAs exhibited COBAS-ANA antibody indexes >2.1 (Fig. 1Up ). Among the 307 patients with CTD, a total of 207 patients were positive for at least one of the eight disease-specific ANAs by each EIA-based reagent set. Using the sera from these 207 patients, we determined the sensitivity of the COBAS-ANA and the IF-ANA for detection of disease-specific ANAs. At cutoff dilutions of 1:40 and 1:160, the IF-ANA was positive for 97.6% (202 of 207) and 90.8% (188 of 207), respectively, whereas the COBAS-ANA was positive for 97.1% (201 of 207) and 91.8% (190 of 207) at cutoff indexes of 0.6 and 0.9, respectively.

To investigate the selectivity for the disease-specific ANAs that could be detected with certainty by the COBAS-ANA and IF-ANA, we analyzed CTD patients whose sera were positive for only one disease-specific ANA, including 15 patients with anti-U1RNP antibody, 18 patients with anti-SSA/Ro antibody, 14 patients with anti-centromere antibody, 5 patients with anti-Jo-1 antibody, and 19 patients with anti-dsDNA antibody. At a cutoff dilution of 1:40, the IF-ANA was positive in 100% of the patients with anti-U1RNP antibody, 100% of those with anti-SSA/Ro antibody, 100% of those with anti-centromere antibody, 60% of those with anti-Jo-1 antibody, and 100% of those with anti-dsDNA antibody. On the other hand, at a cutoff index of 0.6, the COBAS-ANA was positive in 100% of the patients with anti-U1RNP antibody, 89% of those with anti-SSA/Ro antibody, 86% of those with anti-centromere antibody, 80% of those with anti-Jo-1 antibody, and 100% of those with anti-dsDNA antibody (Table 7 ).


View this table:
[in this window]
[in a new window]
  		Table 7. Comparison of COBAS-ANA and IF-ANA, using monospecific antisera.1


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
The purpose of this study was to evaluate the usefulness of the COBAS-ANA, a new EIA-based screening method for detection of ANAs in comparison with the IF-ANA in patients who had been diagnosed as having CTD and in healthy individuals who had been selected based on strict criteria. This new method has two special features: (a) enhancement of overall reactions to multiple ANAs by the use of both HEp-2 cell-derived native antigens and multiple recombinant antigens; and (b) objective data are provided in the form of an index by use of an automated instrument.

It has been documented that the IF-ANA has poor specificity as evidenced by high false-positive rates in healthy individuals regardless of its high sensitivity for CTD patients (5). In 1997, Tan et al. (4) investigated the reference intervals for the IF-ANA in healthy individuals at 15 institutions worldwide and reported that there was variability within and between laboratories, with average CVs of 0.165 and 0.507, respectively. Because the IF-ANA was positive in 31.7% and 5.0% of healthy individuals at dilutions of 1:40 and 1:160, respectively, in their study, they recommended reporting IF-ANA results with the positivity in healthy individuals at two dilutions, 1:40 and 1:160. Laboratories must make a great effort to follow these recommendations and to keep their ANA data reliable and accurate. For these reasons, the development of a simple method for detecting ANAs without requiring training in laboratory procedures has been anticipated as a substitute for IF-ANA. Intraassay CVs of 4.3–12.4% for the COBAS-ANA have been reported (9). The intra- and interassay CVs for the COBAS-ANA in our study were 4.5–5.9% and 3.2–5.6%, respectively.

Two types of EIA-based ANA screening tests are commercially available at present (6)(8). In one type, whole nuclear extracts from HEp-2 cells are immobilized on 96-well plates as the antigens, whereas mixtures of purified or recombinant specific proteins are immobilized as the antigens in the other type. EIA-based methods are computerized and easier to operate than the IF-ANA, which requires complicated manipulations, but still seem to have several disadvantages (6)(8). In the former method, in which whole nuclear extracts are coated onto the EIA plates, the amounts of nuclear antigens may not be constant because the affinity of each antigen for the solid phase is different. The latter method is useful for screening disease-specific ANAs coated on the plates, but ANAs against nuclear antigens other than the specific antigens used in the method may not be detectable. By contrast, the COBAS-ANA investigated in this study is a hybrid between these two EIA-based assay systems and can react with various types of ANAs, including unknown ANAs, because native antigens are used for this assay, and responses are standardized by the addition of recombinant antigens.

The manufacturer used 50 IF-ANA-positive and 30 IF-ANA-negative sera to verify that the COBAS-ANA detects only the disease-specific ANAs and does not react against the control sera, but its clinical usefulness has not been evaluated. In the present study, we used hundreds of sera preselected by diagnosis to evaluate the performance of the COBAS-ANA for clinical use in comparison with the conventional IF-ANA. The internal validity of this study was guaranteed by the use of sera from CTD patients who had been diagnosed according to the established criteria for each disease and by blind comparisons of the COBAS-ANA with the conventional IF-ANA. It is important to identify a proper threshold for ANA determination to be able to distinguish healthy individuals from CTD patients. Using adequate numbers of samples, we therefore performed ROC analysis to evaluate the ability of the COBAS-ANA to discriminate CTD patients from healthy individuals, by which the diagnostic accuracy could be properly assessed. On the basis of the ROC analysis, a cutoff index of 0.8 was found to be most suitable for discriminating CTD patients, except those with RA, from healthy women, although the criteria for evaluation (positive if >=1.1, negative if <0.9) are listed in the operating manual for COBAS-ANA. The sensitivity and specificity of COBAS-ANA at a cutoff index of 0.8 were 88% and 93%, respectively, and were superior to those of the IF-ANA at a cutoff dilution of 1:160 (sensitivity, 81%; specificity, 87%).

In the ROC analysis of the COBAS-ANA performed in six groups of CTD patients (each CTD as a patient group) and healthy women (as reference group), the AUCs were 0.93–0.99 in patients with SLE, SS, SSc, or MCTD, indicating a highly accurate diagnosis. In addition, the AUCs of the COBAS-ANA and IF-ANA for SLE and SS were significantly different (P = 0.008 and 0.012, respectively). These findings indicate that the COBAS-ANA is superior to the IF-ANA for the detection of SLE and SS in terms of the sensitivity and specificity of the assays. There were no significant differences between the detection accuracies of these two methods in patients with SSc, DM/PM, or RA.

As indicated earlier, the {kappa} statistic for agreement (between the COBAS-ANA and IF-ANA titers) was <0.4, showing poor agreement between the COBAS-ANA and IF-ANA titers in CTD patients as a whole. This is primarily because the COBAS-ANA and IF-ANA titers were not concordant for many patients with RA or DM/PM, in whom positivity in the IF-ANA was higher than in the COBAS-ANA (Table 4Up ). Tan et al. (4) reported that no cutoff dilution was useful for differentiating patients with RA or soft tissue rheumatism from healthy individuals, and there was no suitable cutoff point for the ROC curves of the COBAS-ANA and IF-ANA in RA patients in the present study as well. At a cutoff index of 0.9, the sensitivity of the COBAS-ANA was 20% in RA patients, suggesting that the COBAS-ANA should not be used to diagnose RA. In the ROC analysis performed for SLE patients (as patient group) and RA patients (as reference group), the AUCs were 0.93 (95% CI, 0.88–0.96) and 0.84 (95% CI, 0.77–0.89) for the COBAS-ANA and IF-ANA, respectively, showing a significantly larger AUC for the COBAS-ANA than for the IF-ANA (P = 0.004). Therefore, the COBAS-ANA may be more useful than the IF-ANA for differential diagnosis between SLE and RA, although it is not useful for differentiating RA patients from healthy individuals. This is an advantage for the COBAS-ANA because the ANA test is frequently requested to screen CTD patients (excluding patients with RA) from patients with RA and soft tissue rheumatism.

The COBAS-ANA assessed in this study seems superior to most commercially available ANA-screening tests with regard to diagnostic accuracy. In the ROC analysis performed by Gniewek et al. (18), the AUC for an EIA-based ANA was not significantly different from that for an IF-ANA. According to the report published by Emlen and O’Neill (8), the positivity of six different EIA methods was 62–90% in SLE patients when the positivity of the IF-ANA at a cutoff dilution of 1:64 was 88%. In the present study, the COBAS-ANA could detect 102 of 111 SLE patients (91.9%) when 90 of the 111 patients (81.1%) were positive in the IF-ANA at cutoff dilution of 1:160 (Table 4Up ). It was difficult to compare the specificity of the six EIA methods and our COBAS-ANA because the false-positive rates (specificity) in the study by Emlen and O’Neill (8) were determined on the basis of reevaluation of sera from consecutive patients for whom ANA testing had been ordered. However, the EIA methods with higher sensitivity seemed to have lower specificity.

We also compared the ability of the COBAS-ANA and IF-ANA to detect the eight disease-specific ANAs. When the positivity (sensitivity for detection of disease-specific ANAs) of the COBAS-ANA was determined in CTD patients who were positive for at least one of eight disease-specific ANAs, the positivity was 91.8% at a cutoff index of 0.9, but rose to 97.1% at a cutoff index of 0.6, which was almost the same as the 97.6% positivity for the IF-ANA at cutoff dilution of 1:40. We therefore concluded that a cutoff index of 0.6 could be used for suspicion of the presence of disease-specific ANAs. In the experiments using specimens from CTD patients that were positive for only one of the eight disease-specific antigens (monospecific antisera), the positivity of the COBAS-ANA at a cutoff index of 0.6 was 100%, 89%, 86%, 80%, and 100% for anti-U1RNP antibody, anti-SSA/Ro antibody, anti-centromere antibody, anti-Jo-1 antibody, and anti-dsDNA antibody, respectively. These results were almost equal to the positivity of 60–100% for the IF-ANA at a cutoff dilution of 1:40. However, the COBAS-ANA positivity was as low as 50% for specimens that exhibited the nucleolar pattern of IF-ANA staining in the same manner as reported previously (9).

Finally, because performing the IF-ANA on a regular basis is labor-intensive and demands advanced skills in the operator, it is desirable to reduce the labor and costs involved in IF-ANA testing through the introduction of EIA methods (6)(21). In a study by Roche Diagnostics to evaluate time consumption and costs for ANA testing, the total costs per determination for the COBAS-ANA were calculated to be 7.04–7.26 DM (US $3.03-3.12), whereas those for the IF-ANA were 5.54–5.57 DM (US $2.38-2.40; Roche Diagnostics, unpublished data). Of course, the cost for additional titration of ANAs in the IF-ANA should be multiplied, whereas the COBAS-ANA provides quantitative data. Introduction of the COBAS-ANA would be of great help for this purpose because (a) the COBAS-ANA is at least as good as the IF-ANA with regard to sensitivity and specificity: (b) COBAS-ANA analysis can be performed easily with a fully automated instrument, and it takes only ~60 min to obtain the data; and (c) clinicians can evaluate the results with various cutoff indexes because quantitative data are printed out. For example, the result will be judged negative for the disease-specific ANAs if the index is <0.6 and will suggest the presence of CTDs (except RA) if the index is >=0.9.

The COBAS-ANA cannot tell the ANA staining pattern and may not detect some important antibodies to nuclear antigens, such as proliferating cell nuclear antigens and nucleolar antigens, that can be easily detected by the IF-ANA. Although there is no perfect assay available for screening ANAs, the COBAS-ANA is attractive as a new screening test for ANAs from the standpoints of diagnostic accuracy and effectiveness of ANA tests.


   Acknowledgments
 
This work was supported in part by a Grant-in-Aid for Scientific Research (B) (10922078) from the Ministry of Education, Science, Sports, and Culture of Japan. We are grateful to Prof. Eng M. Tan (Scripps Research Institute) for helpful suggestions and critical reading of the manuscript. We also thank Dr. Hans-Peter Lehmann, David Panneton, and Syuichi Hayashi (Roche Diagnostics), Hiroshi Suno (MBL), and Tomoyoshi Terakawa (DIA-IATRON) for providing the ANA reagent sets and for helpful discussions.


   Footnotes
 
1 Nonstandard abbreviations: ANA, antinuclear antibody; CTD, connective tissue disease; SLE, systemic lupus erythematosus; IF-ANA, immunofluorescence-based method for antinuclear antibodies; EIA, enzyme immunoassay; COBAS-ANA, COBAS Core HEp2 ANA EIA; MBL, Medical & Biological Laboratories; dsDNA, double-stranded DNA; SS, Sjögren syndrome; SSc, systemic sclerosis; DM/PM, dermatomyositis and polymyositis; MCTD, mixed connective tissue disease; RA, rheumatoid arthritis; 95% CI, 95% confidence interval; and AUC, area under the curve.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Friou GJ. Clinical application of lupus serum-nucleoprotein reaction using fluorescent antibody technique. J Clin Invest 1957;36:890-898.
  2. Kozin F, Fowler M, Koethe SM. A comparison of the sensitivities and specificities of different substrates for the fluorescent antinuclear antibody test. Am J Clin Pathol 1980;74:785-790.[Medline] [Order article via Infotrieve]
  3. Tan EM, Rodnan GP, Garcia I, Moroi Y, Fritzler MJ, Peebles C. Diversity of antinuclear antibodies in progressive systemic sclerosis: anti-centromere antibody and its relationship to CREST syndrome. Arthritis Rheum 1980;23:617-625.[Web of Science][Medline] [Order article via Infotrieve]
  4. Tan EM, Feltkamp TEW, Smolen JS, Butcher B, Dawkins R, Fritzler MJ, et al. Range of antinuclear antibodies in "healthy" individuals. Arthritis Rheum 1997;40:1601-1611.[Web of Science][Medline] [Order article via Infotrieve]
  5. Feltkamp TEW. Antinuclear antibody determination in a routine laboratory. Ann Rheum Dis 1996;55:723-727.[Abstract/Free Full Text]
  6. Jaskowski TD, Schroder C, Martins TB, Mouristsen CL, Litwin CM, Hill HR. Screening for antinuclear antibodies by enzyme immunoassay. Am J Clin Pathol 1996;105:468-473.[Medline] [Order article via Infotrieve]
  7. Hombburger HA, Cahen YD, Griffiths J, Jacob GL. Detection of antinuclear antibodies: comparative evaluation of enzyme immunoassay and indirect immunofluorescence methods. Arch Pathol Lab Med 1998;122:993-999.[Web of Science][Medline] [Order article via Infotrieve]
  8. Emlen W, O’Neill L. Clinical significance of antinuclear antibodies: comparison of detection with immunofluorescence and enzyme-linked immunosorbent assays. Arthritis Rheum 1997;40:1612-1618.[Web of Science][Medline] [Order article via Infotrieve]
  9. Bayer PM, Bauerfeind S, Bienvenu J, Fabien N, Frei PC, Gilburd B, et al. Multicenter evaluation study on a new HEp2 ANA screening enzyme Immune assay. J Autoimmun 1999;13:89-93.[Web of Science][Medline] [Order article via Infotrieve]
  10. Tan EM, Cohen AS, Fries JF, Masi AT, McShane DJ, Rothfield NF, et al. The 1982 revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 1982;25:1271-1277.[Web of Science][Medline] [Order article via Infotrieve]
  11. Vitali C, Bombardieri S, Moutsopoulos HM, Balestrieri G, Benci velli W, Bernstein RM, et al. Preliminary criteria for the classification of Sjögren’s syndrome. Results of a prospective concerted action supported by the European Community. Arthritis Rheum 1993;36:340-347.[Web of Science][Medline] [Order article via Infotrieve]
  12. . Subcommittee for Scleroderma Criteria of the American Rheumatism Association Diagnostic and Therapeutic Criteria Committee. Preliminary criteria for the classification of systemic sclerosis (scleroderma). Arthritis Rheum 1980;23:581-590.[Web of Science][Medline] [Order article via Infotrieve]
  13. Bohan A, Peter JB. Polymyositis and dermatomyositis (first of two parts). N Engl J Med 1975;292:344-347.[Web of Science][Medline] [Order article via Infotrieve]
  14. Bohan A, Peter JB. Polymyositis and dermatomyositis (second of two parts). N Engl J Med 1975;292:403-407.[Web of Science][Medline] [Order article via Infotrieve]
  15. Porter JF, Kingsland LC, III, Lindberg DAB, Shah I, Benge JM, Hazelwood SE, et al. The AI/RHEUM knowledge-based computer consultant system in rheumatology: performance in the diagnosis of 59 connective tissue disease patients from Japan. Arthritis Rheum 1988;31:219-226.[Web of Science][Medline] [Order article via Infotrieve]
  16. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 1988;31:315-324.[Web of Science][Medline] [Order article via Infotrieve]
  17. Hanley JA. Receiver operating characteristic (ROC) methodology: the state of the art. Crit Rev Diagn Imaging 1989;29:307-335.[Web of Science][Medline] [Order article via Infotrieve]
  18. Gniewek RA, Sandbutle C, Fox PC. Comparison of antinuclear antibody testing methods by ROC analysis with reference to disease diagnosis [Technical Briefs]. Clin Chem 1997;43:1987-1989.[Free Full Text]
  19. Koch GG, Landis JR, Freeman DH, Jr, Lehnen RC. A general methodology for the analysis of experiments with repeated measurement of categorical data. Biometrics 1977;33:133-158.[Web of Science][Medline] [Order article via Infotrieve]
  20. Bruns DE, Huth EJ, Magid E, Young DS. Toward a checklist for reporting of studies of diagnostic accuracy of medical tests. Clin Chem 2000;46:893-895.[Abstract/Free Full Text]
  21. Reisner BS, DiBlasi J, Goel N. Comparison of an enzyme immunoassay to an indirect fluorescent immunoassay for the detection of antinuclear antibodies. Am J Clin Pathol 1999;111:503-506.[Web of Science][Medline] [Order article via Infotrieve]



The following articles in journals at HighWire Press have cited this article:


Home page
 Clin. Chem.Home page
B. Lumbreras-Lacarra, J. M. Ramos-Rincon, and I. Hernandez-Aguado
Methodology in Diagnostic Laboratory Test Research in Clinical Chemistry and Clinical Chemistry and Laboratory Medicine
Clin. Chem., March 1, 2004; 50(3): 530 - 536.
[Abstract] [Full Text] [PDF]

Home page
 LupusHome page
S Kanagawa, A Morinobu, M Koshiba, G Kageyama, N Hayashi, S Yoshino, Y Tokano, H Hashimoto, and S Kumagai
Association of the TAP2*Bky2 allele with presence of SS-A/Ro and other autoantibodies in Japanese patients with systemic lupus erythematosus
Lupus, April 1, 2003; 12(4): 258 - 265.
[Abstract] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (15)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Hayashi, N.
Right arrow Articles by Kumagai, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hayashi, N.
Right arrow Articles by Kumagai, S.
Related Collections
Right arrow Clinical Immunology
Right arrow Evidence Based Laboratory Medicine and Test Utilization

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS