- Email: [email protected]
Performance evaluation of three assays for the detection of PR3-ANCA in granulomatosis with polyangiitis in daily practice
Autoimmunity Reviews 12 (2013) 1118–1122
Contents lists available at ScienceDirect
Autoimmunity Reviews journal homepage: www.elsevier.com/locate/autrev
Review
Performance evaluation of three assays for the detection of PR3-ANCA in granulomatosis with polyangiitis in daily practice Nicolas Noel a,1, Chantal André b,c, Djaouida Bengoufa b, Catherine Dehoulle b, Michael Mahler e, Nicolas Limal a, Bertrand Godeau a,c, Sophie Hüe b,c,d,⁎ a
AP-HP, Centre de référence labellisé pour la prise en charge des cytopénies auto immunes de l'adulte, Service de Médecine Interne, CHU Henri Mondor, Créteil, France AP-HP, Service d'Immunologie Biologique, CHU Henri Mondor, Créteil, France Université Paris-Est-Créteil (Paris XII), Créteil, France d Institut Mondor de Recherche Biomédicale, Equipe 16, INSERM U955, France e INOVA Diagnostics, Inc., San Diego, USA b c
a r t i c l e
i n f o
Article history: Received 19 January 2013 Accepted 21 January 2013 Available online 22 June 2013 Keywords: Granulomatosis with Polyangiitis PR3 ANCA Detection methods
a b s t r a c t Background: Anti-neutrophil cytoplasmic antibodies (ANCA) directed against proteinase 3 (PR3-ANCA) are a serological hallmark of small vessel vasculitis, particularly granulomatosis with polyangiitis (GPA). To increase their sensitivity, some ELISA employ the human native PR3 combined with a recombinant protein. Their specificity in daily practice is still to be defined. Our objective was to compare the performance for GPA diagnosis of three PR3-ANCA assays in daily practice. Patients and methods: Seventy-eight consecutive patients' sera with suggestive IIF were included. All sera were tested with a routine Enzyme Linked Immuno adsorbant Assay (ELISA) employing a mixture of human native and human recombinant (hn + hr) PR3 (EUROIMMUN™) compared to two assays using immobilized purified human PR3 (QUANTA Lite® ELISA and QUANTA Flash® Chemiluminescence assay (CIA), INOVA Diagnostics). Clinical data including BVAS score were collected retrospectively. Results: Nineteen out of the 78 patients had GPA. The hn + hr PR3 ELISA had a good sensitivity (100%) but a lower specificity for the diagnosis of GPA (61.0%) than the assays using the sole native protein (hn ELISA: 81.4%, hn CIA: 69.5%). False positive results mainly consisted of patients with inflammatory bowel disease, who had a specific PR3-ANCA positivity assembly when coupling the assays. The antibody titers by human native PR3 assays, but not hn + hr assay, positively correlated with BVAS score. Conclusion: These results highlight the need of a close collaboration between physicians and immunologists. Combining assays including last generation CIA employing human native antigens should improve the performance of GPA's diagnosis. © 2013 Elsevier B.V. All rights reserved.
Contents 1. 2. 3.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Patients and methods . . . . . . . . . . . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Patients' characteristics . . . . . . . . . . . . . . . . . . . . . 3.2. ROC curve analysis of the three assays for the diagnosis of GPA . . 3.3. Conditions resulting in “false positive” PR3-ANCA detection . . . . 3.4. Relation between anti-PR3 levels and diagnosis accuracy for GPA and 4. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conflict of interests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Take-home messages . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . GPA activity . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
1119 1119 1120 1120 1120 1120 1120 1120 1121 1122 1122
⁎ Corresponding author at: Service d'Immunologie biologique, APHP, CHU Henri Mondor, 51 avenue du Gal de Lattre de Tassigny, 94010 CRETEIL cedex, France. Tel.: +33 1 49 81 48 86; fax: +33 1 49 81 28 97. E-mail address: [email protected] (S. Hüe). 1 Current affiliation: AP-HP, Service de Médecine Interne et Immunologie Clinique, CHU Bicêtre, Le Kremlin-Bicêtre, France. 1568-9972/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.autrev.2013.06.009
N. Noel et al. / Autoimmunity Reviews 12 (2013) 1118–1122
1. Introduction Anti-neutrophil cytoplasm antibodies (ANCA)-associated vasculitis (AAV) are severe pathologies in which the need for a rapid and reliable diagnostic assay is important and has a significant impact on the treatment decision [1,2]. According to the current consensus statement, screening for ANCA begins with indirect immunofluorescence assay (IIF) on fixed human neutrophils followed by methods determining the antigen specificities (proteinase 3 (PR3) or myeloperoxidase (MPO)) of the autoantibodies [3]. Cytoplasmic (c-)ANCA fluorescence patterns are mainly found in granulomatosis with polyangiitis (GPA, formerly Wegener's granulomatosis), and are frequently associated with anti-PR3 reactivity [4]. As a major concern is to avoid false negative results, a broad spectrum of anti-PR3 detection methods has been developed in order to increase their sensitivity [5–12]. ELISA using the human native PR3 combined with the human recombinant protein (hn + hr) yield a very high sensitivity [7,11], but their specificity may vary [13,14]. Indeed, false positive results may occur in daily practice, as these assays offer numerous immunogenic epitopes which can be recognized by other auto-antibodies directed against granulocyte components. Such observations can for example occur in the context of other inflammatory conditions such as inflammatory bowel disease (IBD), in a similar manner to the classical “x-ANCA” pattern found in IIF [4,1]. Moreover, it is not clear whether these tests employing human native and recombinant proteins can reliably predict disease severity, compared with recent assays [7,10,15]. The purpose of this study was to evaluate the performance of an ELISA using a mixture of human native and human recombinant PR3 compared to two assays (one ELISA and one Chemiluminescence Assay, CIA) employing immobilized purified human native protein for the diagnosis of GPA in daily practice. 2. Patients and methods Between January 1st 2009 and January 1st 2011, all consecutive sera of patients with positive IIF on ethanol-fixed human neutrophils suggestive of anti-PR3 specificity (n = 78) were tested in our
1119
laboratory of biological immunology by three detection methods for anti-PR3 antibodies (EUROIMMUN® PR3 ELISA, QUANTA Lite® PR3 ELISA, and QUANTA Flash® PR3, CIA). All sera were addressed to our immunology laboratory by physicians for suspicion of inflammatory disorders. Each patient was followed in the Henri Mondor hospital, Assistance Publique — Hopitaux de Paris, Créteil, FRANCE. Patient's clinical data were retrospectively collected and anonymously computed with respect to the latest version of the Helsinki declaration of human research ethics. The local ethics committee approved the study. All patients diagnosed with granulomatosis with polyangiitis (GPA, N = 19) were classified according to the international nomenclature [16,17]. For these patients, disease activity score was measured using the last updated Birmingham Vasculitis Activity Score (BVAS) [18]. In order to make the flow easier to read, the three assays were referred to as Test A, Test B and Test C. Test A (Euroimmun PR3 ELISA) uses human recombinant (based on human cDNA, expressed in human cells) and human native purified proteinase 3 in pre-coated 96-microwell plates. This assay employs a semi-quantitative calibration method with a cut-off ratio set at 1. For Test B (QUANTA Lite® PR3 ELISA), purified human PR3 antigen is bound to the wells of a polystyrene microwell plate under conditions that will preserve the antigen in its native state. For Test C (QUANTA Flash® PR3 CIA), native purified PR3 is coated on to paramagnetic beads, stored in the reagent cartridge under conditions that preserve the antigen in its reactive state. Cut-off titer was defined as 20 AU/mL and 20 CU/mL, respectively, for Test B and Test C. For all sera, the tests were performed in duplicate, according to manufacturers' recommendations. For statistical analysis, data were stored in the GraphPad PRISM 5.01 package (GraphPad software Inc., La Jolla, San Diego, CA). Results are presented as the median (interquartile range) or n (%). Receiver operating characteristics (ROC) analyses were carried out for comparison between the assays' performances. Mann–Whitney U test, Spearman's correlation and Chi-square test were performed for analysis of non-parametric continuous and categorical variables, when applicable. A p value of less than 0.05 was considered statistically significant.
Table 1 Repartition of final retained diagnoses, depending on assays positivity. Final diagnoses
Positive IIF (n = 78)
Positive Test A (n = 42)
Positive Test B (n = 30)
Positive Test C (n = 36)
Age (years)
59.4 (39.1–67.0) 39 (50) 19 (24.4) 11 (14.1) 6 (7.7) 2 (2.6) 1 (1.3) 3 (3.9) 3 (3.9) 3 (3.9) 1 (1.3) 1 (1.3) 1 (1.3) 1 (1.3) 22 (28.2) 4 (5.1) 78 (100)
58.0 (34.2–65.3) 18 (41.9) 19 (45.2) 9 (21.4) 1 (2.4) 0 (0) 1 (2.4) 1 (2.4) 1 (2.4) 1 (2.4) 1 (2.4) 1 (2.4) 0 (0) 1 (2.4) 4 (9.5) 2 (4.8) 42 (100)
58.0 (34.2–65.8) 13 (43.3) 19 (65.3) 3 (10) 1 (3.3) 0 (0) 1 (3.3) 0 (0) 1 (3.3) 1 (3.3) 0 (0) 0 (0) 0 (0) 1 (3.3) 2 (6.7) 1 (3.3) 30 (100)
55.8 (32.7–64.5) 14 (38.9) 19 (52.8) 8 (22.2) 0 (0) 0 (0) 1 (2.8) 0 (0) 1 (2.8) 1 (2.8) 1 (2.8) 0 (0) 0 (0) 1 (2.8) 3 (8.3) 1 (2.8) 36 (100)
Women (n, %) GPA (n, %) IBD/PSC (n, %) MPA (n, %) Churg–Strauss syndrome (n, %) PAN (n, %) AI Hepatitis (n, %) SLE (n, %) GCA (n, %) DRESS (n, %) DM (n, %) Cryoglobulinemia (n, %) Crystal cholesterol embolism (n, %) Micellaneousa (n, %) Unknown (n, %) TOTAL (n, %)
Results are presented as median (IQR) or n (%). GPA = granulomatosis polyangiitis, IBD = inflammatory bowel disease, MPA: microscopic polyangiitis, PSC = primary sclerosing cholangitis, PAN = periarteritis nodosa, SLE = systemic lupus erythematosus, GCA = giant cell arteritis, DRESS = drug reaction with eosinophilia and systemic symptoms, DM = dermatomyositis. Test A: human native + human recombinant PR3 IgG ELISA (EUROIMMUN™), Test B: human native PR3 IgG ELISA (QUANTA Lite™), Test C: human native PR3 IgG CIA (QUANTA Flash™). a “Miscellaneous diagnosis” consisted of: polyarthralgia, breast cancer, Behçet's syndrome, iatrogenic cholestasis, rhabdomyolysis, idiopathic hypereosinophilic syndrome, ischemic colitis, tuberculosis, non-alcoholic steatosis, asthma, thrombotic microangiopathy following bone marrow transplantation, Hamman–Rich syndrome, post-infectious vasculitis (Mycoplasma pneumoniae), acute interstitial nephritis, nasal polyposis, sciatica, idiopathic axonal neuropathy, viral hepatitis, acute renal failure, idiopathic polyradiculitis, chronic kidney disease due to nephroangiosclerosis, and renal transplant rejection. Of note, AAV was not the cause for renal transplantation.
1120
N. Noel et al. / Autoimmunity Reviews 12 (2013) 1118–1122
Table 2 ROC curve analysis of the 3 assays (All IIF + patients, n = 78). Diagnostics method
Sensitivity
Specificity
Test A Test B Test C
100 (82.35–100) 61.0 (47.4–73.5) 2.57 100 (82.4–100) 81.4 (69.1–90.3) 5.36 100 (82.4–100) 69.5 (56.1–80.8) 3.28
Table 3 Final diagnoses of patients according to the PR3-ANCA positivity pattern.
Likelihood AUC (ROC) ratio 0.92 (0.86–0.98) 0.94 (0.89–1.00) 0.94 (0.89–1.00)
AUC: area under curve. IIF: indirect immunofluorescence. ROC: Receiver operating characteristics. Test A: human native + human recombinant PR3 IgG ELISA (EUROIMMUN®), Test B: human native PR3 IgG ELISA (QUANTA Lite®), Test C: human native PR3 IgG CIA (QUANTA Flash®).
3. Results 3.1. Patients' characteristics The main clinical characteristics of the patients are described in Table 1. Briefly, the median age was 59.4 years (25th–75th IQR: 39.1–67.0) and the sex ratio was 1 (F/M = 39/39). The main final diagnosis were GPA (n = 19, 24.4%), inflammatory bowel disease (IBD) (n = 11, 14.1%) and MPA (n = 6, 7.7%). Among the 78 patients, Tests A, B and C were positive in 42/78 (53.8%), 30/78 (38.5%), and 36/78 (46.2%) patients, respectively. Each detection method strongly correlated with each other (Test A vs. Test B: Spearman's r = 0.788, p b 0.0001; Test A vs. Test C: r = 0.79, p b 0.0001; and Test B vs. Test C: r = 0.72, p b 0.0001). 3.2. ROC curve analysis of the three assays for the diagnosis of GPA The first analysis of the entire sample (N = 78) revealed a sensitivity of 100% for GPA diagnosis for each of the three assays. The specificity of each method varied from 61% (Test A) to 81.4% (Test B) (Table 2), and the positive likelihood ratio was better for the two assays using human purified native protein (Tests B and C) than for Test A, although the differences between AUCs did not reach statistical significance. 23 out of 42 (54.8%) patients with positive hn + hr anti-PR3 detection (test A) who did not have GPA were diagnosed with a large panel of inflammatory diseases (essentially IBD or sclerosing cholangitis (n = 9), Table 1). 3.3. Conditions resulting in “false positive” PR3-ANCA detection As all anti-PR3 assays were positive for the 19 GPA patients but had varying specificity, we investigated whether the clinical diagnosis could also be predicted by the anti-PR3 positivity results according to the three different methods (Table 3). GPA patients represented 19 out of 29 (65.5%) of the patients with all positive testings. The repartition of other diagnosis in this subset was IBD (n = 3), periarteritis nodosa (n = 1), systemic lupus erythematosus (n = 1), giant cell arteritis (n = 1), multiple sclerosis (n = 1), hypereosinophilic syndrome (n = 1), cholesterol crystal embolism (n = 1), unknown (n = 1). When considering patients with apparent discordant results, we interestingly observed that IBD patients were overrepresented when Test A and Test C were positive, but not Test B (n = 5/ 6). Of note, when combining the assays' results, the diagnosis of GPA could be excluded if one of the three assays was negative, reminiscent with the excellent observed negative predictive value (100%). 3.4. Relation between anti-PR3 levels and diagnosis accuracy for GPA and GPA activity Lastly, we questioned whether anti-PR3 titers could help to increase GPA specificity when the anti-PR3 antibodies were positive (i.e., higher than manufacturer's cut-off value). The results are depicted in Fig. 1.
Test A
Test B
Test C
Number of patients
Final diagnosis
Positive
Positive
Positive
29
Positive Positive Positive
Positive Negative Negative
Negative Positive Negative
1 6 6
GPA (n = 19), IBD/PSC (n = 3), PAN (n = 1), SLE (n = 1), GCA (n = 1), MS (n = 1), HES (n = 1), CCE (n = 1), unknown (n = 1) Non-GPA vasculitis (n = 1) IBD (n = 5), DRESS (n = 1) IBD (n = 1), AIH (n = 1), DM (n = 1), TMA (n = 1), Hamman–Rich (n = 1), unknown (n = 1)
AIH: autoimmune hepatitis, CCE: cholesterol crystal embolism, DRESS: drug reaction with eosinophilia and systemic symptoms, DM dermatomyositis, GCA: giant cell arteritis, GPA: granulomatosis with polyangiitis, HES: hypereosinophilic syndrome, IBD: inflammatory bowel disease, MS: multiple sclerosis, PSC: primary sclerosing cholangitis, PAN: periarteritis nodosa, SLE: systemic lupus erythematosus, TMA: thrombotic microangiopathy. Test A: human native + human recombinant PR3 IgG ELISA (EUROIMMUN®), Test B: human native PR3 IgG ELISA (QUANTA Lite®), Test C: human native PR3 IgG CIA (QUANTA Flash®).
When the Test A method was positive (N 1, n = 42), GPA patients had higher anti-PR3 titers than patients without GPA (median (IQR): 5.5 (3.2–7.6) vs. 2.3 (1.3–3.9) AU/mL, respectively, p = 0.0015). This difference was also found significant with respect to positive Test results (threshold = 20 CU/mL, n = 30, median (IQR) titer: 264.1 (139.9–525.0) vs. 83.3 (46.7–185) CU/mL, respectively, p = 0.0024) and a trend was observed when considering positive Test B (threshold = 20 AU/mL, n = 36, median (IQR) titer 88.0 (63–118) vs. 60 (44–99) AU/mL respectively, p = 0.078) (Fig. 1). Among the subset of patients with GPA, the median disease activity index, measured by the BVAS, was 18 (IQR: 10–24, range: 5–36). As shown in Fig. 2, Test B and Test C titers significantly correlated with BVAS score (Spearman's r = 0.48, p = 0.04 and r = 0.55, p = 0.01, respectively), but not those obtained by Test A (hn + hr PR3 ELISA). 4. Discussion PR3-ANCA positivity represents a serological hallmark for the diagnosis of granulomatosis with polyangiitis, in combination with clinical criteria [16]. In order to optimize their detection, a large variety of assays have been developed, thus differing in their sensitivity and/or specificity. In this study, we compared the performance of three PR3-ANCA detection methods for the diagnosis of GPA in a population of 78 consecutive patients with suggestive IIF. Our first question was to determine whether increasing sensitivity by using a mixture of native and recombinant PR3 antigens in routine detection could affect the diagnostic specificity. Our study was a “real life” study, in which only samples of patient with positive indirect immunofluorescence were examined, as recommended [3]. All tests had an excellent sensitivity when using manufacturer's cut-off values (100%). This very high sensitivity confirms the aim of the hn + hr PR3 ELISA assays (here referred to as Test A) [7,11] and has to be compared with studies evaluating the performance of novel generation ELISA or CIA assays [10,19]. However, a better specificity was achieved when using purified human native PR3 antigens. Test B (QUANTA Lite®) and Test C (QUANTA Flash®) employ human purified PR3 antigens that are immobilized on the surface of ELISA plates by direct adsorption or on paramagnetic beads, respectively [10]. The conformational epitopes might then better presented to autoantibodies, with a better sensitivity than prior generation ELISAs. Our results highlight the observation that optimizing sensitivity by multiplying numbers of antigenic targets in assays combining human native + human
N. Noel et al. / Autoimmunity Reviews 12 (2013) 1118–1122
100 50
20 0 (N
=1 1)
+, G Te st C
Te st B
+, N
Te st B
+, G
on -G
PA
PA
PA
(N
(N =1 9)
=2 3) Te st A +, N on -G PA (N
(N =1 9)
=1 9)
20 0
1 0
(N =1 7)
5
150
4000 3500 3000 2500 2000 1500 1000 500
Te st C +, N on -G PA
Test B (AU/mL)
10
Test C (CU/mL)
200
15
Test A (ratio)
0.0024
0.078
0.0015
Te st A +, G PA
1121
Fig. 1. Comparison of PR3-ANCA titers among the three assays with respect to the patient's final diagnosis.
recombinant PR3 may induce a substantial loss of specificity due to the recognition of other non-conformational epitopes in various inflammatory disorders, especially IBD [20]. The assays' specificities found in our study were lower than those reported in the literature [7,10,13,14,19]. Of note, many of these published studies were designed to detect the optimal cut-off value on panels of selected patients with already known ANCA-associated vasculitis compared to control [7,13] at a pre-defined high specificity. The optimization of the pre-test probability of GPA diagnosis, e.g. with a good use of ACR criteria, would certainly improve the assay's specificity in our study. A close collaboration between physicians and biologists may limit unnecessary testings and optimize their diagnostic performance. In various inflammatory conditions, ANCA may be detected by IIF with an x-ANCA fluorescence pattern. This is due to the recognition of other granulocyte components (i.e., elastase, BPI, …) with sometimes unknown specificity [4,21]. We therefore looked whether such pathologies could result in "false" positive PR3-ANCA detection and explain part of the low specificities of our assays. These samples derived from patients with IBD. Moreover, when combining the test results, a specific pattern of anti-PR3 assays (i.e. positive Test A + C (hn + hr ELISA and human native CIA, respectively) and negative Test B (human native ELISA)) distinguished IBD patients from the other. Although the small number of IBD patients (n=11) in our study does not allow for general extrapolations, this observation suggests the presence of a shared epitope between those tests, which could eventually be specifically recognized by sera of IBD patients. Recently, it has been discussed wether PR3-ANCA might be useful to differentiate ulcerative colitis from Crohn's disease patients. As a consequence of the low cut-off value and the number of positive results in non-GPA patients, we found that GPA patients had
r=0.27 p=0.26
Conflict of interests M. Mahler is employed at INOVA Diagnostics, a company that manufactures and markets autoantibody assays.
r=0.48 p=0.04
10
5
0
150 100 50 0
0
10
20
BVAS
30
40
r=0.55 p=0.01
2500
Test C (CU/mL)
200
Test B (AU/mL)
Test A (ratio)
15
higher titers of PR3 antibodies than in other pathological conditions. Of note, these results relate to our patients and could not be extrapolated to other cohorts, as the link between anti-PR3 rates and diagnosis of AAV remains controversial. The PR3-ANCA titers correlated with disease activity, evaluated by the BVAS score, when detected by both assays employing human native PR3 (Tests B and C), but not the hn + hr PR3 ELISA (Test A). This assay-dependent correlation to the disease activity is in line with previously published results [10,22] and should be explained by the linearity and the dynamic range of the assays [10]. These results suggest that binding specificities of PR3-ANCA could account for a different pathogenic potential and may contribute to the variable severity of disease manifestations. Due to the retrospective design of our study, we were unable to conclude whether the results could be affected by the treatment or predict relapses of the disease. In conclusion, the quest of a high sensitivity in PR3-ANCA detection in daily practice leads to a subsequent decrease in the specificity, which can be rectified by a clinical pre-test probability. The use of novel assays improves the diagnostic performance but raises the question of the epitope recognition in other pathological conditions such as IBD. The combination of multiple assays may also enhance the negative predictive value, and could be proposed in confusing clinical situations. Although further studies are needed to confirm these results, these quantitative assays may also be used to predict disease severity in patients with granulomatosis and polyangiitis.
2000 1500 1000 500 0
0
10
20
BVAS
30
40
0
10
20
BVAS
Fig. 2. Correlation (Spearman's rho) between PR3-ANCA titers and BVAS score in GPA patients (n = 19) according to the three assays.
30
40
1122
N. Noel et al. / Autoimmunity Reviews 12 (2013) 1118–1122
Take-home messages • In suggestive situations, a rapid and reliable diagnostic confirmation of granulomatosis with polyangiitis is an urgent need. • Recent highly sensitive PR3-ANCA assays employing the human native and a recombinant protein expose to the risk of false positive diagnoses. • False positive results mainly consist of patients with inflammatory bowel disease, due to possible shared PR3 epitopes. • A close collaboration between physicians and immunologists is needed.
References [1] Radice A, Bianchi L, Sinico RA. Anti-neutrophil cytoplasmic autoantibodies: methodological aspects and clinical significance in systemic vasculitis. Autoimmun Rev 2013;12(4):487–95. [2] Holle JU, Gross WL. Treatment of ANCA-associated vasculitides (AAV). Autoimmun Rev 2013;12(4):483–6. [3] Savige J, Gillis D, Benson E, Davies D, Esnault V, Falk RJ, et al. International consensus statement on testing and reporting of antineutrophil cytoplasmic antibodies (ANCA). Am J Clin Pathol 1999;111(4):507–13. [4] Lochman I, Král V, Lochmanová A, Lupač J, Cebecauer L. ANCA in the diagnosis of neutrophil-mediated inflammation. Autoimmun Rev 2011;10(6):295–8. [5] Csernok E, Ahlquist D, Ullrich S, Gross WL. A critical evaluation of commercial immunoassays for antineutrophil cytoplasmic antibodies directed against proteinase 3 and myeloperoxidase in Wegener's granulomatosis and microscopic polyangiitis. Rheumatology (Oxford) 2002;41(11):1313–7. [6] Csernok E, Holle J, Hellmich B, Willem J, Tervaert C, Kallenberg CGM, et al. Evaluation of capture ELISA for detection of antineutrophil cytoplasmic antibodies directed against proteinase 3 in Wegener's granulomatosis: first results from a multicentre study. Rheumatology (Oxford) 2004;43(2):174–80. [7] Damoiseaux J, Dähnrich C, Rosemann A, Probst C, Komorowski L, Stegeman CA, et al. A novel enzyme-linked immunosorbent assay using a mixture of human native and recombinant proteinase-3 significantly improves the diagnostic potential for antineutrophil cytoplasmic antibody-associated vasculitis. Ann Rheum Dis 2009;68(2):228–33. [8] Damoiseaux J, Vaessen M, Knapen Y, Csernok E, Stegeman CA, Van Paassen P, et al. Evaluation of the FIDIS vasculitis multiplex immunoassay for diagnosis and follow-up of ANCA-associated vasculitis and Goodpasture's disease. Ann N Y Acad Sci 2007;1109:454–63.
[9] Hellmich B, Csernok E, Fredenhagen G, Gross WL. A novel high sensitivity ELISA for detection of antineutrophil cytoplasm antibodies against proteinase-3. Clin Exp Rheumatol 2007;25(1 Suppl. 44):S1–5. [10] Mahler M, Radice A, Yang W, Bentow C, Seaman A, Bianchi L, et al. Development and performance evaluation of novel chemiluminescence assays for detection of anti-PR3 and anti-MPO antibodies. Clin Chim Acta 2012;413(7–8):719–26. [11] Rarok AA, Huitema MG, Van der Leij MJ, Van der Geld YM, Berthold H, Schmitt J, et al. Recombinant protein to analyze autoantibodies to proteinase 3 in systemic vasculitis. Am J Clin Pathol 2003;120(4):586–95. [12] Kaul R, Johnson K, Scholz H, Marr G. Performance of the BioPlex 2200 Autoimmune Vasculitis kit. Autoimmun Rev 2009;8(3):224–7. [13] Holle JU, Hellmich B, Backes M, Gross WL, Csernok E. Variations in performance characteristics of commercial enzyme immunoassay kits for detection of antineutrophil cytoplasmic antibodies: what is the optimal cut off? Ann Rheum Dis 2005;64(12):1773–9. [14] Holle JU, Herrmann K, Gross WL, Csernok E. Comparative analysis of different commercial ELISA systems for the detection of anti-neutrophil cytoplasm antibodies in ANCA-associated vasculitides. Clin Exp Rheumatol 2012;30(1 Suppl. 70):S66–69. [15] Segelmark M, Phillips BD, Hogan SL, Falk RJ, Jennette JC. Monitoring proteinase 3 antineutrophil cytoplasmic antibodies for detection of relapses in small vessel vasculitis. Clin Diagn Lab Immunol 2003;10(5):769–74. [16] Jennette JC, Falk RJ, Andrassy K, Bacon PA, Churg J, Gross WL, et al. Nomenclature of systemic vasculitides. Proposal of an international consensus conference. Arthritis Rheum 1994;37(2):187–92. [17] Jennette J, Falk R, Bacon P, Basu N, Cid M, Ferrario F, et al. 2012 revised international Chapel Hill consensus conference nomenclature of vasculitides. Arthritis Rheum 2013;65(1):1–11. [18] Mahr AD, Neogi T, Lavalley MP, Davis JC, Hoffman GS, McCune WJ, et al. Assessment of the item selection and weighting in the Birmingham vasculitis activity score for Wegener's granulomatosis. Arthritis Rheum 2008;59(6):884–91. [19] Roggenbuck D, Buettner T, Hoffmann L, Schmechta H, Reinhold D, Conrad K. High-sensitivity detection of autoantibodies against proteinase-3 by a novel third-generation enzyme-linked immunosorbent assay. Ann N Y Acad Sci 2009;1173:41–6. [20] Trevisin M, Pollock W, Dimech W, Savige J. Evaluation of a multiplex flow cytometric immunoassay to detect PR3- and MPO-ANCA in active and treated vasculitis, and in inflammatory bowel disease (IBD). J Immunol Methods 2008;336(2):104–12. [21] Radice A, Vecchi M, Bianchi MB, Sinico RA. Contribution of immunofluorescence to the identification and characterization of anti-neutrophil cytoplasmic autoantibodies. The role of different fixatives. Clin Exp Rheumatol 2000;18(6): 707–12. [22] Trevisin M, Pollock W, Dimech W, Melny J, Paspaliaris B, Gillis D, et al. Antigen-specific ANCA ELISAs have different sensitivities for active and treated vasculitis and for nonvasculitic disease. Am J Clin Pathol 2008;129(1):42–53.
Contents lists available at ScienceDirect
Autoimmunity Reviews journal homepage: www.elsevier.com/locate/autrev
Review
Performance evaluation of three assays for the detection of PR3-ANCA in granulomatosis with polyangiitis in daily practice Nicolas Noel a,1, Chantal André b,c, Djaouida Bengoufa b, Catherine Dehoulle b, Michael Mahler e, Nicolas Limal a, Bertrand Godeau a,c, Sophie Hüe b,c,d,⁎ a
AP-HP, Centre de référence labellisé pour la prise en charge des cytopénies auto immunes de l'adulte, Service de Médecine Interne, CHU Henri Mondor, Créteil, France AP-HP, Service d'Immunologie Biologique, CHU Henri Mondor, Créteil, France Université Paris-Est-Créteil (Paris XII), Créteil, France d Institut Mondor de Recherche Biomédicale, Equipe 16, INSERM U955, France e INOVA Diagnostics, Inc., San Diego, USA b c
a r t i c l e
i n f o
Article history: Received 19 January 2013 Accepted 21 January 2013 Available online 22 June 2013 Keywords: Granulomatosis with Polyangiitis PR3 ANCA Detection methods
a b s t r a c t Background: Anti-neutrophil cytoplasmic antibodies (ANCA) directed against proteinase 3 (PR3-ANCA) are a serological hallmark of small vessel vasculitis, particularly granulomatosis with polyangiitis (GPA). To increase their sensitivity, some ELISA employ the human native PR3 combined with a recombinant protein. Their specificity in daily practice is still to be defined. Our objective was to compare the performance for GPA diagnosis of three PR3-ANCA assays in daily practice. Patients and methods: Seventy-eight consecutive patients' sera with suggestive IIF were included. All sera were tested with a routine Enzyme Linked Immuno adsorbant Assay (ELISA) employing a mixture of human native and human recombinant (hn + hr) PR3 (EUROIMMUN™) compared to two assays using immobilized purified human PR3 (QUANTA Lite® ELISA and QUANTA Flash® Chemiluminescence assay (CIA), INOVA Diagnostics). Clinical data including BVAS score were collected retrospectively. Results: Nineteen out of the 78 patients had GPA. The hn + hr PR3 ELISA had a good sensitivity (100%) but a lower specificity for the diagnosis of GPA (61.0%) than the assays using the sole native protein (hn ELISA: 81.4%, hn CIA: 69.5%). False positive results mainly consisted of patients with inflammatory bowel disease, who had a specific PR3-ANCA positivity assembly when coupling the assays. The antibody titers by human native PR3 assays, but not hn + hr assay, positively correlated with BVAS score. Conclusion: These results highlight the need of a close collaboration between physicians and immunologists. Combining assays including last generation CIA employing human native antigens should improve the performance of GPA's diagnosis. © 2013 Elsevier B.V. All rights reserved.
Contents 1. 2. 3.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Patients and methods . . . . . . . . . . . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Patients' characteristics . . . . . . . . . . . . . . . . . . . . . 3.2. ROC curve analysis of the three assays for the diagnosis of GPA . . 3.3. Conditions resulting in “false positive” PR3-ANCA detection . . . . 3.4. Relation between anti-PR3 levels and diagnosis accuracy for GPA and 4. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conflict of interests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Take-home messages . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . GPA activity . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
1119 1119 1120 1120 1120 1120 1120 1120 1121 1122 1122
⁎ Corresponding author at: Service d'Immunologie biologique, APHP, CHU Henri Mondor, 51 avenue du Gal de Lattre de Tassigny, 94010 CRETEIL cedex, France. Tel.: +33 1 49 81 48 86; fax: +33 1 49 81 28 97. E-mail address: [email protected] (S. Hüe). 1 Current affiliation: AP-HP, Service de Médecine Interne et Immunologie Clinique, CHU Bicêtre, Le Kremlin-Bicêtre, France. 1568-9972/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.autrev.2013.06.009
N. Noel et al. / Autoimmunity Reviews 12 (2013) 1118–1122
1. Introduction Anti-neutrophil cytoplasm antibodies (ANCA)-associated vasculitis (AAV) are severe pathologies in which the need for a rapid and reliable diagnostic assay is important and has a significant impact on the treatment decision [1,2]. According to the current consensus statement, screening for ANCA begins with indirect immunofluorescence assay (IIF) on fixed human neutrophils followed by methods determining the antigen specificities (proteinase 3 (PR3) or myeloperoxidase (MPO)) of the autoantibodies [3]. Cytoplasmic (c-)ANCA fluorescence patterns are mainly found in granulomatosis with polyangiitis (GPA, formerly Wegener's granulomatosis), and are frequently associated with anti-PR3 reactivity [4]. As a major concern is to avoid false negative results, a broad spectrum of anti-PR3 detection methods has been developed in order to increase their sensitivity [5–12]. ELISA using the human native PR3 combined with the human recombinant protein (hn + hr) yield a very high sensitivity [7,11], but their specificity may vary [13,14]. Indeed, false positive results may occur in daily practice, as these assays offer numerous immunogenic epitopes which can be recognized by other auto-antibodies directed against granulocyte components. Such observations can for example occur in the context of other inflammatory conditions such as inflammatory bowel disease (IBD), in a similar manner to the classical “x-ANCA” pattern found in IIF [4,1]. Moreover, it is not clear whether these tests employing human native and recombinant proteins can reliably predict disease severity, compared with recent assays [7,10,15]. The purpose of this study was to evaluate the performance of an ELISA using a mixture of human native and human recombinant PR3 compared to two assays (one ELISA and one Chemiluminescence Assay, CIA) employing immobilized purified human native protein for the diagnosis of GPA in daily practice. 2. Patients and methods Between January 1st 2009 and January 1st 2011, all consecutive sera of patients with positive IIF on ethanol-fixed human neutrophils suggestive of anti-PR3 specificity (n = 78) were tested in our
1119
laboratory of biological immunology by three detection methods for anti-PR3 antibodies (EUROIMMUN® PR3 ELISA, QUANTA Lite® PR3 ELISA, and QUANTA Flash® PR3, CIA). All sera were addressed to our immunology laboratory by physicians for suspicion of inflammatory disorders. Each patient was followed in the Henri Mondor hospital, Assistance Publique — Hopitaux de Paris, Créteil, FRANCE. Patient's clinical data were retrospectively collected and anonymously computed with respect to the latest version of the Helsinki declaration of human research ethics. The local ethics committee approved the study. All patients diagnosed with granulomatosis with polyangiitis (GPA, N = 19) were classified according to the international nomenclature [16,17]. For these patients, disease activity score was measured using the last updated Birmingham Vasculitis Activity Score (BVAS) [18]. In order to make the flow easier to read, the three assays were referred to as Test A, Test B and Test C. Test A (Euroimmun PR3 ELISA) uses human recombinant (based on human cDNA, expressed in human cells) and human native purified proteinase 3 in pre-coated 96-microwell plates. This assay employs a semi-quantitative calibration method with a cut-off ratio set at 1. For Test B (QUANTA Lite® PR3 ELISA), purified human PR3 antigen is bound to the wells of a polystyrene microwell plate under conditions that will preserve the antigen in its native state. For Test C (QUANTA Flash® PR3 CIA), native purified PR3 is coated on to paramagnetic beads, stored in the reagent cartridge under conditions that preserve the antigen in its reactive state. Cut-off titer was defined as 20 AU/mL and 20 CU/mL, respectively, for Test B and Test C. For all sera, the tests were performed in duplicate, according to manufacturers' recommendations. For statistical analysis, data were stored in the GraphPad PRISM 5.01 package (GraphPad software Inc., La Jolla, San Diego, CA). Results are presented as the median (interquartile range) or n (%). Receiver operating characteristics (ROC) analyses were carried out for comparison between the assays' performances. Mann–Whitney U test, Spearman's correlation and Chi-square test were performed for analysis of non-parametric continuous and categorical variables, when applicable. A p value of less than 0.05 was considered statistically significant.
Table 1 Repartition of final retained diagnoses, depending on assays positivity. Final diagnoses
Positive IIF (n = 78)
Positive Test A (n = 42)
Positive Test B (n = 30)
Positive Test C (n = 36)
Age (years)
59.4 (39.1–67.0) 39 (50) 19 (24.4) 11 (14.1) 6 (7.7) 2 (2.6) 1 (1.3) 3 (3.9) 3 (3.9) 3 (3.9) 1 (1.3) 1 (1.3) 1 (1.3) 1 (1.3) 22 (28.2) 4 (5.1) 78 (100)
58.0 (34.2–65.3) 18 (41.9) 19 (45.2) 9 (21.4) 1 (2.4) 0 (0) 1 (2.4) 1 (2.4) 1 (2.4) 1 (2.4) 1 (2.4) 1 (2.4) 0 (0) 1 (2.4) 4 (9.5) 2 (4.8) 42 (100)
58.0 (34.2–65.8) 13 (43.3) 19 (65.3) 3 (10) 1 (3.3) 0 (0) 1 (3.3) 0 (0) 1 (3.3) 1 (3.3) 0 (0) 0 (0) 0 (0) 1 (3.3) 2 (6.7) 1 (3.3) 30 (100)
55.8 (32.7–64.5) 14 (38.9) 19 (52.8) 8 (22.2) 0 (0) 0 (0) 1 (2.8) 0 (0) 1 (2.8) 1 (2.8) 1 (2.8) 0 (0) 0 (0) 1 (2.8) 3 (8.3) 1 (2.8) 36 (100)
Women (n, %) GPA (n, %) IBD/PSC (n, %) MPA (n, %) Churg–Strauss syndrome (n, %) PAN (n, %) AI Hepatitis (n, %) SLE (n, %) GCA (n, %) DRESS (n, %) DM (n, %) Cryoglobulinemia (n, %) Crystal cholesterol embolism (n, %) Micellaneousa (n, %) Unknown (n, %) TOTAL (n, %)
Results are presented as median (IQR) or n (%). GPA = granulomatosis polyangiitis, IBD = inflammatory bowel disease, MPA: microscopic polyangiitis, PSC = primary sclerosing cholangitis, PAN = periarteritis nodosa, SLE = systemic lupus erythematosus, GCA = giant cell arteritis, DRESS = drug reaction with eosinophilia and systemic symptoms, DM = dermatomyositis. Test A: human native + human recombinant PR3 IgG ELISA (EUROIMMUN™), Test B: human native PR3 IgG ELISA (QUANTA Lite™), Test C: human native PR3 IgG CIA (QUANTA Flash™). a “Miscellaneous diagnosis” consisted of: polyarthralgia, breast cancer, Behçet's syndrome, iatrogenic cholestasis, rhabdomyolysis, idiopathic hypereosinophilic syndrome, ischemic colitis, tuberculosis, non-alcoholic steatosis, asthma, thrombotic microangiopathy following bone marrow transplantation, Hamman–Rich syndrome, post-infectious vasculitis (Mycoplasma pneumoniae), acute interstitial nephritis, nasal polyposis, sciatica, idiopathic axonal neuropathy, viral hepatitis, acute renal failure, idiopathic polyradiculitis, chronic kidney disease due to nephroangiosclerosis, and renal transplant rejection. Of note, AAV was not the cause for renal transplantation.
1120
N. Noel et al. / Autoimmunity Reviews 12 (2013) 1118–1122
Table 2 ROC curve analysis of the 3 assays (All IIF + patients, n = 78). Diagnostics method
Sensitivity
Specificity
Test A Test B Test C
100 (82.35–100) 61.0 (47.4–73.5) 2.57 100 (82.4–100) 81.4 (69.1–90.3) 5.36 100 (82.4–100) 69.5 (56.1–80.8) 3.28
Table 3 Final diagnoses of patients according to the PR3-ANCA positivity pattern.
Likelihood AUC (ROC) ratio 0.92 (0.86–0.98) 0.94 (0.89–1.00) 0.94 (0.89–1.00)
AUC: area under curve. IIF: indirect immunofluorescence. ROC: Receiver operating characteristics. Test A: human native + human recombinant PR3 IgG ELISA (EUROIMMUN®), Test B: human native PR3 IgG ELISA (QUANTA Lite®), Test C: human native PR3 IgG CIA (QUANTA Flash®).
3. Results 3.1. Patients' characteristics The main clinical characteristics of the patients are described in Table 1. Briefly, the median age was 59.4 years (25th–75th IQR: 39.1–67.0) and the sex ratio was 1 (F/M = 39/39). The main final diagnosis were GPA (n = 19, 24.4%), inflammatory bowel disease (IBD) (n = 11, 14.1%) and MPA (n = 6, 7.7%). Among the 78 patients, Tests A, B and C were positive in 42/78 (53.8%), 30/78 (38.5%), and 36/78 (46.2%) patients, respectively. Each detection method strongly correlated with each other (Test A vs. Test B: Spearman's r = 0.788, p b 0.0001; Test A vs. Test C: r = 0.79, p b 0.0001; and Test B vs. Test C: r = 0.72, p b 0.0001). 3.2. ROC curve analysis of the three assays for the diagnosis of GPA The first analysis of the entire sample (N = 78) revealed a sensitivity of 100% for GPA diagnosis for each of the three assays. The specificity of each method varied from 61% (Test A) to 81.4% (Test B) (Table 2), and the positive likelihood ratio was better for the two assays using human purified native protein (Tests B and C) than for Test A, although the differences between AUCs did not reach statistical significance. 23 out of 42 (54.8%) patients with positive hn + hr anti-PR3 detection (test A) who did not have GPA were diagnosed with a large panel of inflammatory diseases (essentially IBD or sclerosing cholangitis (n = 9), Table 1). 3.3. Conditions resulting in “false positive” PR3-ANCA detection As all anti-PR3 assays were positive for the 19 GPA patients but had varying specificity, we investigated whether the clinical diagnosis could also be predicted by the anti-PR3 positivity results according to the three different methods (Table 3). GPA patients represented 19 out of 29 (65.5%) of the patients with all positive testings. The repartition of other diagnosis in this subset was IBD (n = 3), periarteritis nodosa (n = 1), systemic lupus erythematosus (n = 1), giant cell arteritis (n = 1), multiple sclerosis (n = 1), hypereosinophilic syndrome (n = 1), cholesterol crystal embolism (n = 1), unknown (n = 1). When considering patients with apparent discordant results, we interestingly observed that IBD patients were overrepresented when Test A and Test C were positive, but not Test B (n = 5/ 6). Of note, when combining the assays' results, the diagnosis of GPA could be excluded if one of the three assays was negative, reminiscent with the excellent observed negative predictive value (100%). 3.4. Relation between anti-PR3 levels and diagnosis accuracy for GPA and GPA activity Lastly, we questioned whether anti-PR3 titers could help to increase GPA specificity when the anti-PR3 antibodies were positive (i.e., higher than manufacturer's cut-off value). The results are depicted in Fig. 1.
Test A
Test B
Test C
Number of patients
Final diagnosis
Positive
Positive
Positive
29
Positive Positive Positive
Positive Negative Negative
Negative Positive Negative
1 6 6
GPA (n = 19), IBD/PSC (n = 3), PAN (n = 1), SLE (n = 1), GCA (n = 1), MS (n = 1), HES (n = 1), CCE (n = 1), unknown (n = 1) Non-GPA vasculitis (n = 1) IBD (n = 5), DRESS (n = 1) IBD (n = 1), AIH (n = 1), DM (n = 1), TMA (n = 1), Hamman–Rich (n = 1), unknown (n = 1)
AIH: autoimmune hepatitis, CCE: cholesterol crystal embolism, DRESS: drug reaction with eosinophilia and systemic symptoms, DM dermatomyositis, GCA: giant cell arteritis, GPA: granulomatosis with polyangiitis, HES: hypereosinophilic syndrome, IBD: inflammatory bowel disease, MS: multiple sclerosis, PSC: primary sclerosing cholangitis, PAN: periarteritis nodosa, SLE: systemic lupus erythematosus, TMA: thrombotic microangiopathy. Test A: human native + human recombinant PR3 IgG ELISA (EUROIMMUN®), Test B: human native PR3 IgG ELISA (QUANTA Lite®), Test C: human native PR3 IgG CIA (QUANTA Flash®).
When the Test A method was positive (N 1, n = 42), GPA patients had higher anti-PR3 titers than patients without GPA (median (IQR): 5.5 (3.2–7.6) vs. 2.3 (1.3–3.9) AU/mL, respectively, p = 0.0015). This difference was also found significant with respect to positive Test results (threshold = 20 CU/mL, n = 30, median (IQR) titer: 264.1 (139.9–525.0) vs. 83.3 (46.7–185) CU/mL, respectively, p = 0.0024) and a trend was observed when considering positive Test B (threshold = 20 AU/mL, n = 36, median (IQR) titer 88.0 (63–118) vs. 60 (44–99) AU/mL respectively, p = 0.078) (Fig. 1). Among the subset of patients with GPA, the median disease activity index, measured by the BVAS, was 18 (IQR: 10–24, range: 5–36). As shown in Fig. 2, Test B and Test C titers significantly correlated with BVAS score (Spearman's r = 0.48, p = 0.04 and r = 0.55, p = 0.01, respectively), but not those obtained by Test A (hn + hr PR3 ELISA). 4. Discussion PR3-ANCA positivity represents a serological hallmark for the diagnosis of granulomatosis with polyangiitis, in combination with clinical criteria [16]. In order to optimize their detection, a large variety of assays have been developed, thus differing in their sensitivity and/or specificity. In this study, we compared the performance of three PR3-ANCA detection methods for the diagnosis of GPA in a population of 78 consecutive patients with suggestive IIF. Our first question was to determine whether increasing sensitivity by using a mixture of native and recombinant PR3 antigens in routine detection could affect the diagnostic specificity. Our study was a “real life” study, in which only samples of patient with positive indirect immunofluorescence were examined, as recommended [3]. All tests had an excellent sensitivity when using manufacturer's cut-off values (100%). This very high sensitivity confirms the aim of the hn + hr PR3 ELISA assays (here referred to as Test A) [7,11] and has to be compared with studies evaluating the performance of novel generation ELISA or CIA assays [10,19]. However, a better specificity was achieved when using purified human native PR3 antigens. Test B (QUANTA Lite®) and Test C (QUANTA Flash®) employ human purified PR3 antigens that are immobilized on the surface of ELISA plates by direct adsorption or on paramagnetic beads, respectively [10]. The conformational epitopes might then better presented to autoantibodies, with a better sensitivity than prior generation ELISAs. Our results highlight the observation that optimizing sensitivity by multiplying numbers of antigenic targets in assays combining human native + human
N. Noel et al. / Autoimmunity Reviews 12 (2013) 1118–1122
100 50
20 0 (N
=1 1)
+, G Te st C
Te st B
+, N
Te st B
+, G
on -G
PA
PA
PA
(N
(N =1 9)
=2 3) Te st A +, N on -G PA (N
(N =1 9)
=1 9)
20 0
1 0
(N =1 7)
5
150
4000 3500 3000 2500 2000 1500 1000 500
Te st C +, N on -G PA
Test B (AU/mL)
10
Test C (CU/mL)
200
15
Test A (ratio)
0.0024
0.078
0.0015
Te st A +, G PA
1121
Fig. 1. Comparison of PR3-ANCA titers among the three assays with respect to the patient's final diagnosis.
recombinant PR3 may induce a substantial loss of specificity due to the recognition of other non-conformational epitopes in various inflammatory disorders, especially IBD [20]. The assays' specificities found in our study were lower than those reported in the literature [7,10,13,14,19]. Of note, many of these published studies were designed to detect the optimal cut-off value on panels of selected patients with already known ANCA-associated vasculitis compared to control [7,13] at a pre-defined high specificity. The optimization of the pre-test probability of GPA diagnosis, e.g. with a good use of ACR criteria, would certainly improve the assay's specificity in our study. A close collaboration between physicians and biologists may limit unnecessary testings and optimize their diagnostic performance. In various inflammatory conditions, ANCA may be detected by IIF with an x-ANCA fluorescence pattern. This is due to the recognition of other granulocyte components (i.e., elastase, BPI, …) with sometimes unknown specificity [4,21]. We therefore looked whether such pathologies could result in "false" positive PR3-ANCA detection and explain part of the low specificities of our assays. These samples derived from patients with IBD. Moreover, when combining the test results, a specific pattern of anti-PR3 assays (i.e. positive Test A + C (hn + hr ELISA and human native CIA, respectively) and negative Test B (human native ELISA)) distinguished IBD patients from the other. Although the small number of IBD patients (n=11) in our study does not allow for general extrapolations, this observation suggests the presence of a shared epitope between those tests, which could eventually be specifically recognized by sera of IBD patients. Recently, it has been discussed wether PR3-ANCA might be useful to differentiate ulcerative colitis from Crohn's disease patients. As a consequence of the low cut-off value and the number of positive results in non-GPA patients, we found that GPA patients had
r=0.27 p=0.26
Conflict of interests M. Mahler is employed at INOVA Diagnostics, a company that manufactures and markets autoantibody assays.
r=0.48 p=0.04
10
5
0
150 100 50 0
0
10
20
BVAS
30
40
r=0.55 p=0.01
2500
Test C (CU/mL)
200
Test B (AU/mL)
Test A (ratio)
15
higher titers of PR3 antibodies than in other pathological conditions. Of note, these results relate to our patients and could not be extrapolated to other cohorts, as the link between anti-PR3 rates and diagnosis of AAV remains controversial. The PR3-ANCA titers correlated with disease activity, evaluated by the BVAS score, when detected by both assays employing human native PR3 (Tests B and C), but not the hn + hr PR3 ELISA (Test A). This assay-dependent correlation to the disease activity is in line with previously published results [10,22] and should be explained by the linearity and the dynamic range of the assays [10]. These results suggest that binding specificities of PR3-ANCA could account for a different pathogenic potential and may contribute to the variable severity of disease manifestations. Due to the retrospective design of our study, we were unable to conclude whether the results could be affected by the treatment or predict relapses of the disease. In conclusion, the quest of a high sensitivity in PR3-ANCA detection in daily practice leads to a subsequent decrease in the specificity, which can be rectified by a clinical pre-test probability. The use of novel assays improves the diagnostic performance but raises the question of the epitope recognition in other pathological conditions such as IBD. The combination of multiple assays may also enhance the negative predictive value, and could be proposed in confusing clinical situations. Although further studies are needed to confirm these results, these quantitative assays may also be used to predict disease severity in patients with granulomatosis and polyangiitis.
2000 1500 1000 500 0
0
10
20
BVAS
30
40
0
10
20
BVAS
Fig. 2. Correlation (Spearman's rho) between PR3-ANCA titers and BVAS score in GPA patients (n = 19) according to the three assays.
30
40
1122
N. Noel et al. / Autoimmunity Reviews 12 (2013) 1118–1122
Take-home messages • In suggestive situations, a rapid and reliable diagnostic confirmation of granulomatosis with polyangiitis is an urgent need. • Recent highly sensitive PR3-ANCA assays employing the human native and a recombinant protein expose to the risk of false positive diagnoses. • False positive results mainly consist of patients with inflammatory bowel disease, due to possible shared PR3 epitopes. • A close collaboration between physicians and immunologists is needed.
References [1] Radice A, Bianchi L, Sinico RA. Anti-neutrophil cytoplasmic autoantibodies: methodological aspects and clinical significance in systemic vasculitis. Autoimmun Rev 2013;12(4):487–95. [2] Holle JU, Gross WL. Treatment of ANCA-associated vasculitides (AAV). Autoimmun Rev 2013;12(4):483–6. [3] Savige J, Gillis D, Benson E, Davies D, Esnault V, Falk RJ, et al. International consensus statement on testing and reporting of antineutrophil cytoplasmic antibodies (ANCA). Am J Clin Pathol 1999;111(4):507–13. [4] Lochman I, Král V, Lochmanová A, Lupač J, Cebecauer L. ANCA in the diagnosis of neutrophil-mediated inflammation. Autoimmun Rev 2011;10(6):295–8. [5] Csernok E, Ahlquist D, Ullrich S, Gross WL. A critical evaluation of commercial immunoassays for antineutrophil cytoplasmic antibodies directed against proteinase 3 and myeloperoxidase in Wegener's granulomatosis and microscopic polyangiitis. Rheumatology (Oxford) 2002;41(11):1313–7. [6] Csernok E, Holle J, Hellmich B, Willem J, Tervaert C, Kallenberg CGM, et al. Evaluation of capture ELISA for detection of antineutrophil cytoplasmic antibodies directed against proteinase 3 in Wegener's granulomatosis: first results from a multicentre study. Rheumatology (Oxford) 2004;43(2):174–80. [7] Damoiseaux J, Dähnrich C, Rosemann A, Probst C, Komorowski L, Stegeman CA, et al. A novel enzyme-linked immunosorbent assay using a mixture of human native and recombinant proteinase-3 significantly improves the diagnostic potential for antineutrophil cytoplasmic antibody-associated vasculitis. Ann Rheum Dis 2009;68(2):228–33. [8] Damoiseaux J, Vaessen M, Knapen Y, Csernok E, Stegeman CA, Van Paassen P, et al. Evaluation of the FIDIS vasculitis multiplex immunoassay for diagnosis and follow-up of ANCA-associated vasculitis and Goodpasture's disease. Ann N Y Acad Sci 2007;1109:454–63.
[9] Hellmich B, Csernok E, Fredenhagen G, Gross WL. A novel high sensitivity ELISA for detection of antineutrophil cytoplasm antibodies against proteinase-3. Clin Exp Rheumatol 2007;25(1 Suppl. 44):S1–5. [10] Mahler M, Radice A, Yang W, Bentow C, Seaman A, Bianchi L, et al. Development and performance evaluation of novel chemiluminescence assays for detection of anti-PR3 and anti-MPO antibodies. Clin Chim Acta 2012;413(7–8):719–26. [11] Rarok AA, Huitema MG, Van der Leij MJ, Van der Geld YM, Berthold H, Schmitt J, et al. Recombinant protein to analyze autoantibodies to proteinase 3 in systemic vasculitis. Am J Clin Pathol 2003;120(4):586–95. [12] Kaul R, Johnson K, Scholz H, Marr G. Performance of the BioPlex 2200 Autoimmune Vasculitis kit. Autoimmun Rev 2009;8(3):224–7. [13] Holle JU, Hellmich B, Backes M, Gross WL, Csernok E. Variations in performance characteristics of commercial enzyme immunoassay kits for detection of antineutrophil cytoplasmic antibodies: what is the optimal cut off? Ann Rheum Dis 2005;64(12):1773–9. [14] Holle JU, Herrmann K, Gross WL, Csernok E. Comparative analysis of different commercial ELISA systems for the detection of anti-neutrophil cytoplasm antibodies in ANCA-associated vasculitides. Clin Exp Rheumatol 2012;30(1 Suppl. 70):S66–69. [15] Segelmark M, Phillips BD, Hogan SL, Falk RJ, Jennette JC. Monitoring proteinase 3 antineutrophil cytoplasmic antibodies for detection of relapses in small vessel vasculitis. Clin Diagn Lab Immunol 2003;10(5):769–74. [16] Jennette JC, Falk RJ, Andrassy K, Bacon PA, Churg J, Gross WL, et al. Nomenclature of systemic vasculitides. Proposal of an international consensus conference. Arthritis Rheum 1994;37(2):187–92. [17] Jennette J, Falk R, Bacon P, Basu N, Cid M, Ferrario F, et al. 2012 revised international Chapel Hill consensus conference nomenclature of vasculitides. Arthritis Rheum 2013;65(1):1–11. [18] Mahr AD, Neogi T, Lavalley MP, Davis JC, Hoffman GS, McCune WJ, et al. Assessment of the item selection and weighting in the Birmingham vasculitis activity score for Wegener's granulomatosis. Arthritis Rheum 2008;59(6):884–91. [19] Roggenbuck D, Buettner T, Hoffmann L, Schmechta H, Reinhold D, Conrad K. High-sensitivity detection of autoantibodies against proteinase-3 by a novel third-generation enzyme-linked immunosorbent assay. Ann N Y Acad Sci 2009;1173:41–6. [20] Trevisin M, Pollock W, Dimech W, Savige J. Evaluation of a multiplex flow cytometric immunoassay to detect PR3- and MPO-ANCA in active and treated vasculitis, and in inflammatory bowel disease (IBD). J Immunol Methods 2008;336(2):104–12. [21] Radice A, Vecchi M, Bianchi MB, Sinico RA. Contribution of immunofluorescence to the identification and characterization of anti-neutrophil cytoplasmic autoantibodies. The role of different fixatives. Clin Exp Rheumatol 2000;18(6): 707–12. [22] Trevisin M, Pollock W, Dimech W, Melny J, Paspaliaris B, Gillis D, et al. Antigen-specific ANCA ELISAs have different sensitivities for active and treated vasculitis and for nonvasculitic disease. Am J Clin Pathol 2008;129(1):42–53.