Q J Med 2001; 94: 615-621
© 2001 Association of Physicians
The diagnostic value of anti-neutrophil cytoplasmic antibody testing in a routine clinical setting
From the Rheumatic Diseases Unit and 1 Department of Clinical Microbiology and Immunology, Western General Hospital, Edinburgh, UK
Received 10 July 2001 and in revised form 5 September 2001
 |
Summary |
|---|
 Top Summary IntroductionMethodsResultsDiscussionReferences |
|---|
Anti-neutrophil cytoplasmic antibody (ANCA) tests are a routine clinical assay in most UK hospitals. We examined the role of routine ANCA testing in achieving a diagnosis of systemic vasculitis in a routine clinical setting. From April 1996 to March 2000, 2734 samples from five hospital departments were tested for ANCA by indirect immunofluorescence (IIF) at a single laboratory. After April 1999, enzyme-linked immunosorbent assays (ELISAs) were performed on all IIF-positive samples. Clinical diagnosis was determined for all patients with a positive IIF ANCA, and a sample of the ANCA-negative patients. Some 2–18% of patients with suspected ANCA-associated systemic vasculitis (AASV) had positive IIF ANCA. The AASV diagnosis was confirmed in 0–56% of these cases. Analysis by department suggested that 88–100% of patients with a positive IIF ANCA did not have AASV, except in the Rheumatology department. The positive predictive value (PPV) of IIF ANCA for AASV was 59% and the negative predictive value (NPV) was 84%. Of the patients with proven AASV, 41% did not have ANCA on IIF. Combined ANCA testing by IIF/ELISA had a higher sensitivity and PPV but lower specificity than IIF alone for AASV. For the combined IIF/ELISA test, only the Rheumatology department had a sensitivity or PPV >0% for AASV. The PPV of ANCA by IIF/ELISA for AASV was 79% and the NPV was 63%. The ANCA test is being widely applied with very poor return. Guidelines for more effective usage are proposed.
|
Introduction |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
ANCA and diagnosis
There are no established diagnostic criteria for systemic vasculitis. Diagnosis of vasculitis is based on clinical assessment supported by investigation results, with tissue biopsy as the gold standard. However, much interest has focused on the role of anti-neutrophil cytoplasmic antibodies (ANCA) in the diagnosis of vasculitis since their discovery in 19821 and the report by van der Woude2 and colleagues in 1985 that they were useful in both the diagnosis and monitoring of disease activity in Wegener's granulomatosis. ANCA may have a role in the classification of systemic vasculitis,3 as certain types of vasculitis are strongly associated with their presence: Wegener's granulomatosis (WG), Microscopic polyangiitis (MPA), Churg-Strauss syndrome (CSS) and idiopathic rapidly progressive glomerulonephritis (iRPGN). The term ANCA-associated systemic vasculitis (AASV) has been used to group these disorders together.4 However, ANCA are not diagnostic for these conditions, and not all patients with AASV have a positive ANCA test.
ANCA testing methods
To understand the significance of ANCA, clinicians need to know about the methods used for their detection, and the associated problems and pitfalls. ANCAs are antibodies to human neutrophil cytoplasmic granule proteins. There are two main types of ANCA pattern that can be discerned on the basis of indirect immunofluorescence (IIF) on ethanol-fixed human neutrophils, namely cytoplasmic or cANCA staining, and perinuclear or pANCA staining. A minority of IIF patterns do not fit with either pANCA or cANCA and are termed atypical. A disadvantage of the IIF test for ANCA detection is that the assay is not antigen-specific.5 The main antigens are protease 3 (PR3), which closely correlates with the cANCA IIF pattern, and myeloperoxidase (MPO), which is responsible for the majority of pANCA patterns. However, there are several other antigens (e.g. lactoferrin, cathepsin-G, glucuronidase, lysozyme, bactericidal permeability inhibitor) which can give rise to a pANCA pattern and are found in a variety of non-vasculitic conditions, e.g. inflammatory bowel disease, infection, connective tissue disease and secondary to drugs.6 The development of enzyme-linked immunosorbent assay (ELISA) testing for circulating MPO and PR3 antibodies has been of value as an adjunct to the IIF test. International consensus guidelines recommend all sera are first tested by IIF, and subsequently by ELISA for PR3 and MPO if the IIF is positive.7
Role of ANCA in monitoring
There are conflicting data in the literature about the relationship between ANCA and disease activity. The first report of ANCA in systemic vasculitis indicated that patients with active WG tended to have higher titres than those in remission2 and the presence of ANCA has itself been associated with an increased incidence of relapse of systemic vasculitis.8 Rising ANCA titres have been used by some groups as an indicator of imminent disease relapse on which to base changes in treatment plans,9 but others have found changes in ANCA titre of limited prognostic value.10 By comparing patients with vasculitis and those without, Locke et al.11 found that a rising anti-MPO level by ELISA had a maximal PPV of only 50% in predicting increased disease activity.
Table 1 shows an overview of the percentage of positive ANCA reported with different types of vasculitis and different types of different controls (data from reference 5). It is important to bear in mind that while the majority of positive ANCA results were seen in patients with vasculitis, a significant proportion of disease controls (19%) had a positive pANCA, although further testing with ELISA increased the specificity of the ANCA test. Up to 35% of patients with WG and 42% of patients with MPA did not have ANCA. In the pulmonary renal syndrome, Westman et al.12 found that in a study of over 1000 patients, only 12% were either PR3- or MPO-positive by ELISA, using a rapid screening assay. Therefore, even in such a specific clinical setting where the diagnostic yield of vasculitis would be highest, many patients did not have ANCA.
|
We studied the use of ANCA testing by non-specialist physicians in a routine clinical setting in a single teaching hospital, and the value of ANCA testing in achieving a diagnosis of systemic vasculitis. We also investigated the effect of additional ELISA testing on the sensitivity and specificity of ANCA testing for AASV by IIF.
|
Methods |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
All samples submitted to a single teaching hospital laboratory for ANCA testing over a 4-year period (April 1996–March 2000) were identified by a computer search of the laboratory database. A total of 2734 samples from five hospital departments (Rheumatology, Neurology, General Medicine, Respiratory Medicine, other departments) were tested for both cytoplasmic (c) and perinuclear (p) ANCA by indirect immunofluorescence (IIF) (ANCA Test System, Immuno Concepts NA). During the course of the study, 5% of the patients had a subsequent ANCA test performed; these samples were included in the analysis. Because renal medicine is located at a different hospital, ANCA tests for renal patients were performed at a separate laboratory, and hence were not included in this study. Results were reported as: negative (titre
1/20), positive (titre 
1/40) or atypical positive for either cANCA or pANCA, or uninterpretable. From April 1999 to March 2000, enzyme-linked immunosorbent assays (ELISAs) for antibodies to PR3 and MPO (Bindazyme Anti-PR3 and Anti-MPO Enzyme Immunoassay Kit, The Binding Site) were performed on all IIF-positive samples and 13% of the IIF-negative samples.
Both our IIF and ELISA test systems are subject to rigorous external quality control measures as recommended by the International Consensus Statement on ANCA testing.7
We determined the clinical diagnosis by review of the medical records in all patients with a positive ANCA by IIF. In addition, a random sample of 12% of the patients with a negative ANCA from April 1999 to March 2000 also had their medical records reviewed to establish the clinical diagnosis. This method does have its limitations, and depended entirely on the clinician's pursuit of a diagnosis of systemic vasculitis. We sought supporting evidence within the case record both from the clinically documented features, as well as any histological, radiological or serological evidence to either confirm or refute a diagnosis of vasculitis. On this basis, patients were classified into the following groups: Wegener's granulomatosis (WG), microscopic polyangiitis (MPA), Churg-Strauss syndrome (CSS), cutaneous vasculitis (CV) and unclassified systemic vasculitis (USV), no AASV (if AASV could be definitely excluded) or unknown diagnosis (in cases where there was inadequate information available in the medical notes to ascertain the diagnosis). We defined the diagnostic yield as: (ANCA-positive patients with confirmed AASV/ANCA-positive patients)x100.
|
Results |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
ANCA testing by IIF
The departments of General Medicine, Neurology and Rheumatology were responsible for 92% of ANCA requests. The majority of IIF ANCA results throughout the hospital were negative, with the highest positive rate (18%) in the Rheumatology department (Figure 1). The rate of ANCA positivity by IIF varied considerably between departments (2–18%), as did the diagnostic yield for AASV (0–56%). In all departments apart from Rheumatology, the majority of patients (88–100%) with a positive ANCA by IIF did not have confirmed AASV (Figure 2).
|
|
ANCA testing by ELISA
The relationship between IIF and ELISA test results for 1999–2000 is shown in Table 2. For the majority of IIF-positive samples, cANCA was related to PR3 and pANCA was related to MPO. Negative, atypical and uninterpretable IIF ANCAs were usually ELISA-negative. From the point of view of further data analysis, atypical and uninterpretable IIF ANCAs were considered negative.
|
The relationship between results of ANCA testing by IIF and ELISA and clinical diagnosis 1999–2000 are shown in Table 3. The majority of patients with AASV who were ANCA-positive by IIF were also ANCA-positive by ELISA, whereas the majority of those without AASV or in whom the diagnosis was unknown were ANCA-negative by both IIF and ELISA. The diagnosis was determined for the 81 patients who were ANCA-negative by IIF. AASV was confirmed in nine patients with negative IIF ANCAs (11%), most of whom were Rheumatology patients. Of those patients confirmed to have AASV, 13/22 (59%) were ANCA-positive by IIF (sensitivity of IIF) and 12/13 (92%) were also ELISA-positive. Of the patients with confirmed AASV, 41% (9/22) were ANCA-negative by IIF, most of whom were Rheumatology patients. Of those patients in whom AASV was excluded, 46/55 (84%) were ANCA-negative by IIF (specificity of IIF). Of the remaining nine patients, all were ELISA-negative. The positive predictive value (PPV) of IIF ANCA for AASV was 59% and the negative predictive value (NPV) was 84%.
|
The sensitivity, specificity, PPV and NPV of ANCA by the combined IIF/ELISA test for AASV varied widely between departments in 1999–2000. Only the Rheumatology department had a significant sensitivity (68%) or PPV (79%) of ANCA by IIF/ELISA for AASV. The specificity and NPV of ANCA by IIF/ELISA for AASV in the Rheumatology department were 75% and 63%, respectively.
The cost of ANCA testing
We have estimated the cost of a combined ANCA test at £14.48 per test from our laboratory, and used this as the basis of the cost analysis for ANCA testing within our hospital. We calculated the estimated expenditure of each department on ANCA testing over the length of the study, on the assumption that the combined ANCA test was available from 1996 to 1999, rather than the IIF test alone for these years. This indicates the high cost of ANCA testing in a single teaching hospital (almost £40 000 over 4 years) with great variation between departments in cost per case of AASV. The cost per case for the Rheumatology Department was £144 per confirmed case of AASV. This contrasted starkly with the figures in General Medicine (>£3000 per case) and particularly in Neurology, where since there were no real cases of AASV detected despite spending over £12 000 during the period 1996–2000, we cannot place any value on ANCA testing.
|
Discussion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
In patients with suspected ANCA-associated systemic vasculitis attending a single teaching hospital (excluding renal medicine), our data suggest that the yield of a positive ANCA test by IIF is low (2–18%). Of these patients, 0–56% have confirmed AASV. This wide variation in diagnostic yield between departments was likely based on differences in clinical practice and experience. For example, the Neurology department routinely performed a ‘vasculitis screen’ that included an ANCA test in the majority of new patients presenting with both explained and unexplained neurological symptoms and signs. In clinical practice, the commonest neurological presentation likely in ANCA-associated disease would be peripheral neuropathy;13 less commonly cranial nerve palsy or cerebral involvement can occur.14 In the 4 years of the study, not a single case of systemic vasculitis was identified by the Neurology department based on detection of a positive ANCA test either by IIF alone or in combination with ELISA. Based on this example and as a general guideline, we suggest that ANCA testing be used in selected cases where there is a high suspicion of AASV with typical features as determined by initial clinical assessment.
There was a wide variation in PPV for the combined ANCA test between departments, with only Rheumatology having a PPV >0%. This indicates that in non-specialist hands, a positive test is highly likely to be a false positive, whereas in a specialist setting, a positive ANCA test has a high probability of reflecting true systemic vasculitis. The NPV in the Rheumatology department for the combined ANCA test was low at 63%, reflecting the high percentage of ANCA-negative vasculitis patients in our cohort. However, the NPV for the combined ANCA test was 92–100% in three-quarters of the remaining departments; this indicates that in a non-specialist setting, a negative test can be highly predictive of the absence of AASV, which may be very useful in certain circumstances, e.g. pulmonary renal syndrome. However, this use of the ANCA test to exclude vasculitis must be balanced against the cost of the test and the pre-test probability of a negative test. i.e. if the likelihood of the test being positive is very low, then a negative result does not provide much additional weight in clinical decision-making. In addition, the cost of a combined ANCA test needs to be borne in mind when using the test if the result is unlikely to affect clinical management. We have shown the high cost of ANCA testing in a single teaching hospital. The cost of ANCA testing can vary greatly between departments per case detected, and strongly supports the development of guidelines to avoid unnecessary expenditure.
Our results compare favourably with data published by others. The Johns Hopkins group reported on 856 consecutive patients with possible AASV (small- and medium-vessel vasculitis; CHCC guidelines) tested from 1995 to 1998.15 Only 8% (69/856) had confirmed AASV (WG, MPA, CSS, pauci-immune GN). For a positive IIF test alone (1/20), the sensitivity for AASV was 67%, the specificity was 93%, the PPV was 45% and the likelihood ratio for a positive test was 9.4 (95%CI 6.9–12.7). For a positive combined IIF/ELISA test, the sensitivity for AASV was 52%, the specificity was 99%, the PPV was 88%, and the likelihood ratio for a combined positive test was 88.1 (95%CI 33.3–202.5). Therefore, the addition of the ELISA test dramatically increased the specificity, PPV and likelihood ratio for a positive test for AASV. In their cohort, 12% of patients who met clinical definitions of AASV were ANCA-negative, which is less than in our study group.
In contrast to both our data (and that of Stone et al.15), in a study of patients with diseases that may mimic vasculitides (pulmonary diseases/pulmonary failure, renal failure or multi-organ dysfunction), none of 125 patients without WG or MPA had a positive ANCA by combined IIF/ELISA.16
In a one-year study of ANCA testing at a single regional laboratory in Southampton (1997–1998), 13% (157/1168) of ANCA tests were positive by IIF alone.17 Of the patients with a positive IIF ANCA, 58% had primary vasculitis. Of the patients with primary vasculitis, 78% were ANCA-positive by combined IIF/ELISA testing and 22% were ANCA-negative. The sensitivity of the combined ANCA test for primary vasculitis was 78%, the specificity was 59% and the PPV was 72%.
In a seven-year analysis (1992–1998) of ANCA requests in an academic centre in the USA performed at a single reference laboratory by indirect immunofluorescence (IIF), positive rates were as follows: Rheumatology 5/50 (10%), Nephrology 7/63 (11%), ENT 2/14 (14%), Respiratory Medicine 0/11 (0%).18 Their patient population had a high percentage of ANCA-negative WG at 71%. Specificity of cANCA for WG was 98%, sensitivity was 50% and PPV was 43%.
In a large European multicentre study of the diagnostic value of ANCA testing in systemic vasculitis, the sensitivity of IIF ANCA for WG, MPA and iRPGN was 81–85%.5 Of the disease controls, 19% were positive for pANCA and 4% were positive for cANCA. Among healthy controls, 4% had a positive pANCA and 2% a positive cANCA. The specificity of the IIF test for systemic vasculitis was 76% in disease controls and 94% in healthy controls.
With ELISA testing, the sensitivity of anti-PR3 was 65–67% for WG and the sensitivity of anti-MPO for MPA was 58%. The specificity of anti-PR3 in disease controls was 86–89% and that of anti-MPO was 91%. In healthy controls, the specificity of anti-PR3 was 98–99% and that of anti-MPO was 96%.
The sensitivities of the combined cANCA/anti-PR3 or pANCA/anti-MPO for WG, MPA or iRPGN were 73%, 67% and 82%, respectively, and the specificity was 98% in disease controls and 100% in healthy controls. A high specificity towards disease controls is of great importance, as the disease controls were selected on the basis of clinical signs and symptoms similar to primary vasculitis patients, and they posed a problem in terms of differential diagnosis. This combined testing method is absolutely necessary to improve the diagnostic significance of a positive ANCA.
A meta-analysis of published studies of combined ANCA testing in the diagnosis of vasculitis (based on CHCC definitions) calculated an overall sensitivity of 82% and an overall specificity of 99% for the use of the combined IIF and ELISA assays for cANCA/PR3 and pANCA/MPO.19
Based on our study results and those of others presented above, we suggest as a guideline for more appropriate ANCA testing in routine clinical practice that the emphasis should not be on ANCA testing but on clinical assessment of patients with multi-system disease. In selected cases where there is diagnostic difficulty, a combined ANCA test may be helpful in conjunction with standard investigation, e.g. tissue biopsy in establishing the diagnosis of AASV. We do not advocate the use of ANCA testing as a screening tool or diagnostic test in patients with suspected systemic vasculitis. Guidelines have been proposed20 to help determine clinical settings where the diagnostic value of the ANCA test is likely to be higher (Figure 3). We cannot understate the need for clinical assessment as the most important guide to selecting cases with a high index of suspicion of systemic vasculitis, and would strongly advocate this approach to a more rational use of the ANCA test in future.
|
|
Acknowledgments |
|---|
We thank Ms Diane M. Anderson, Biomedical Scientist, Clinical Microbiology, Western General Hospital and Mrs Anne Houghton for her secretarial support.
|
Notes |
|---|
Address correspondence to Dr R.A. Luqmani, Consultant Rheumatologist/Part Time Senior Lecturer, Rheumatic Diseases Unit, Western General Hospital, Crewe Road South, Edinburgh EH4 2XU, UK. e-mail: Raashid.Luqmani{at}ed.ac.uk
|
References |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
1. Davies DJ, Maran JE, Niall JF, Ryan GB. Segmental necrotising glomerulonephritis with antineutrophil antibody: possible arbovirus aetiology? Br Med J1982; 285:606.
2. van der Woude FJ, Rasmussen N, Lobatto S, Wiik A, Permin H, van Es LA, van der Giessen M, van der Hem GK, The TH. Autoantibodies against neutrophils and monocytes: tool for diagnosis and marker of disease activity in Wegener's granulomatosis. Lancet1985; 1: 425–9.[Medline]
3.
Scott DGI, Watts RA. Classification and epidemiology of systemic vasculitis. Br J Rheumatol1994; 33:897–900.
4. Rasmussen NJD, Abramowicz D, Andrassy K, Bacon PA, Cohen Tervaert JW et al. European therapeutic trials in AASV: disease scoring, consensus regimes and proposed clinical trials. Clin Exp Immunol1995; 101(Suppl. 1): 29–34.
5. Hagen EC, Daha MR, Hermans J, Andrassy K, Csernok E, Gaskin G, Lesavre P, Ludemann J, Rasmussen N, Sinico RA, Wiik A, van der Woude FJ for the EC/BCR Project for ANCA Assay Standardization. Diagnostic value of standardized assays for anti-neutrophil cytoplasmic antibodies in idiopathic systemic vasculitis. Kidney Int1998; 53:743–53.[Web of Science][Medline]
6. Bajema IM, Hagen EC. Evolving concepts about the role of antineutrophil cytoplasm autoantibodies in systemic vasculitides. Curr Opin Rheumatol1999; 11:34–40.[Medline]
7. Savige J, Gillis D, Benson E, Davies D, Esnault V, Falk RJ, Hagen EC, Jayne D, Jennette JC, Paspaliaris B, Pollock W, Pusey C, Savage CO, Silvestrini R, van der Woude F, Wieslander J, Wiik A. International Consensus Statement on Testing and Reporting of Antineutrophil Cytoplasmic Antibodies (ANCA). Am J Clin Pathol1999; 111:507–13.[Web of Science][Medline]
8.
Gordon M, Luqmani RA, Adu D, Greaves I, Richards N, Michael J, Emery P, Howie AJ, Bacon PA. Relapses in patients with a systemic vasculitis. Q J Med1993; 86:779–89.
9. Tervaert JWC, Huitema MG, Hene RJ, Sluiter WJ, The TH, van der Hem GK, Kallenberg CGM. Prevention of relapses in Wegener's granulomatosis by treatment based on antineutrophil cytoplasmic antibody titre. Lancet1990; 336:709–11.[Web of Science][Medline]
10. Kerr GS, Fleisher TA, Hallahan CW, Leavitt RY, Fauci AS, Hoffman GS. Limited prognostic value of changes in antineutrophil cytoplasmic antibody titer in patients with Wegener's granulomatosis. Arthritis Rheum1993; 36:365–71.[Web of Science][Medline]
11. Locke C, Leaker B, Cambridge G. A comparison of the characteristics of circulating anti-myeloperoxidase autoantibodies in vasculitis with those in non-vasculitic conditions. Clin Exp Immunol1999; 115:369–76.[Web of Science][Medline]
12.
Westman KW, Bygren PG, Eilert I, Wiik A, Wieslander J. Rapid screening assay for anti-GBM antibody and ANCAs: an important tool for the differential diagnosis of pulmonary renal syndromes. Nephrol Dial Transplant1997; 12:1863–8.
13. Kissel JT. Vasculitis of the peripheral nervous system. Semin Neurol1994; 14:361–9.[Web of Science][Medline]
14. Nishino H, Rubino FA, DeRemee RA, Swanson JW, Parisi JE. Neurological Involvement in Wegener's Granulomatosis: An Analysis of 324 Consecutive Patients at the Mayo Clinic. Ann Neurol1993; 33:4–9.[Web of Science][Medline]
15. Stone JH, Talor M, Stebbing J, Uhlfelder ML, Rose NR, Carson KA, Hellmann DB, Burek CL. Test characteristics of immunofluorescence and ELISA tests in 856 consecutive patients with possible ANCA-associated conditions. Arthritis Care Res2000; 13:424–34.[Web of Science][Medline]
16. Vassilopoulos D, Niles JL, Villa-Forte A, Arroliga AC, Sullivan EJ, Merkel PA, Hoffman GS. Prevalence of antineutrophil cytoplasmic antibodies (ANCA) in patients with various pulmonary diseases or multi-organ dysfunction. Arthritis Rheum1999; 42:S175.
17. Venkatachalam S, Major P, Morgan L, Smith J, Cawley MID. Clinical utility of antineutrophil cytoplasmic antibody (ANCA) test: an audit. Ann Rheum Dis1999; 58:25–6.
18. Suhail T, Wolf RE, Hearth-Holmes M. Seven year analysis of anti-neutrophil cytoplasmic antibody (ANCA) utilization in an academic hospital. Arthritis Rheum1999; 42:S316.
19. Choi HK, Liu S, Merkel PA, Colditz G, Niles JL. Diagnostic value of ANCA testing for idiopathic systemic vasculitis syndromes. A meta-analysis with focus on anti-myeloperoxidase, p-ANCA and combined testing systems. Arthritis Rheum1999; 42:S175.
20. Christiaan Hagen E. Antineutrophil Cytoplasmic antibodies and Systemic Vasculitis. PhD Thesis, 1997; Chapter 13, 169–79.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  E Suresh Diagnostic approach to patients with suspected vasculitis. Postgrad. Med. J., August 1, 2006; 82(970): 483 - 488. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. R. Friedman and S. D. Wolfsthal Cocaine-Induced Pseudovasculitis Mayo Clin. Proc., May 1, 2005; 80(5): 671 - 673. [Abstract] [PDF]
|
|
|
|
 |
|
 R J Lock Rational requesting or rationing testing? J. Clin. Pathol., February 1, 2004; 57(2): 121 - 122. [Full Text] [PDF]
|
|
|
|
|
|
D Sinclair, M Saas, and J M Stevens The effect of a symptom related "gating policy" on ANCA requests in routine clinical practice J. Clin. Pathol., February 1, 2004; 57(2): 131 - 134. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||