Rheumatic Diseases: Serological Aids to Early Diagnosis

Peter Maddison MD FRCP

Consultant Rheumatologist, North West Wales NHS Trust/Chair of Joint and Muscle Disorders, School of Sport, Health and Exercise Science, University of Wales, Bangor
Pearl Huey MSc

Head Biomedical Scientist, Immunology, North West Wales NHS Trust

Reports on the Rheumatic Diseases Series 5 : Topical Reviews

• Antineutrophil cytoplasmic antibodies (ANCA) of different specificity occur in a range of disorders and testing should only be performed when there is a likelihood of vasculitis
• Sera should be screened for ANCA by indirect immunofluorescence and positive tests confirmed by enzyme-linked immunoassays (ELISA) specific for anti-proteinase 3 (anti-PR3) or myeloperoxidase (MPO)
• Classical cytoplasmic ANCA (C-ANCA) and anti-PR3 have 99% specificity for Wegener's granulomatosis (WG)
• Anticyclic citrullinated peptide (anti-CCP) antibodies are highly specific for rheumatoid arthritis (RA) at any stage of the disease
• Anti-CCP is highly associated with the presence of shared epitope and is predictive of joint destruction even in seronegative arthritis

Introduction

Laboratory methods to detect particular autoantibodies have provided the clinician with valuable tools to assist in diagnosis and, to some extent, prognosis of patients with autoimmune rheumatic diseases. Serology is of particular value in the early stage of disease when clinical expression is often incomplete. Then the autoantibody profile can be diagnostic. In this context, serology has been particularly valuable in connective tissue diseases such as systemic lupus erythematosus. However, more recently, autoantibodies with high specificity have been found in some forms of primary vasculitis (antineutrophil cytoplasmic antibodies – ANCA) and in rheumatoid arthritis (anticitrulline antibodies). Their detection facilitates diagnosis and early institution of treatment and, increasingly, has become part of the diagnostic armamentarium of serology laboratories.

Antineutrophil cytoplasmic antibodies (ANCA)

Systemic necrotising vasculitis encompasses a heterogeneous group of conditions that often rapidly progress to irreversible tissue damage. Early diagnosis and institution of treatment is vital. Clinical diagnosis, however, is often difficult in the initial stages because early symptoms are frequently non-specific. An important development was the recognition that ANCA characterise a group of vasculitides, including Wegener's granulomatosis (WG), microscopic polyangiitis (MPA), and Churg–Strauss syndrome (CSS). These have in common the involvement of both small and medium-sized vessels, frequent development of glomerulonephritis, and a good response to immunosuppression with cyclophosphamide. Prior to serology, pathological examination of involved tissue was the only reliable diagnostic tool. This could introduce delay and sometimes the result was non-specific. Now, ANCA testing has become routine in the diagnostic serology laboratory.

However, because ANCA are found in a number of other settings including a range of other rheumatological diseases, inflammatory bowel disease, autoimmune liver diseases and infectious diseases as well as drug-induced syndromes, interpretation of serology results can be difficult, especially with indiscriminant ordering of ANCA testing.

ANCA patterns and antibody specificity

ANCA describes a number of circulating autoantibodies specifically directed against the cytoplasmic constituents of neutrophils and monocytes.

ANCA was first described over 40 years ago and associated with segmental necrosing glomerulonephritis and vasculitic like symptoms in 1982.1 Subsequently,2 it was demonstrated that ANCA had a strong association with WG and titres of antibody could be correlated with disease activity. In 1988, van de Woude et al3 reported that ANCA had a wide range of specificities and showed that the specificity of ANCA correlated with the pathological features of vasculitis.

Two ANCA patterns were originally identified by indirect immunofluorescence (IIF): the cytoplasmic (C-ANCA) and the perinuclear (P-ANCA) patterns. These apparent morphological differences are purely artefactual and based on the fixative used to preserve the neutrophil substrate.

What has come to be known as 'classical' C-ANCA is associated with antibodies reacting with the 29–30 kDa elastinolitic enzyme, serine proteinase 3 (PR3). This is composed of 229 amino acids and found in the azurophilic granules of neutrophils and monocytes. The 'classical' P-ANCA pattern is associated with antibodies to myeloperoxidase (MPO), a 140 kDa heterodymeric enzyme also associated with the antimicrobial properties of neutrophils.

However, additional IIF staining patterns have been described. These are termed 'atypical' ANCA and may be cytoplasmic or perinuclear. Atypical ANCA is generally anti-PR3- and anti-MPO-negative by enzyme-linked immunosorbent assays (ELISA). Many of these atypical patterns have now been attributed to other antigenic molecules in the neutrophil granules such as elastase, cathepsin G, cathelicidins, lactoferin, lysosyme, bactericidal/permeability increasing protein (BPI), calprotectin and defensins.4,5 In addition, there is a challenge in distinguishing ANCA from other autoantibody specificities, which frequently accompany ANCA in patients with vasculitis. For example, antimicrosomal, antiribosomal and smooth muscle antibodies can easily be confused with C-ANCA, whereas antinuclear antibodies such as anti-double stranded deoxyribonucleic acid (DNA) and anti-golgi body antibodies often resemble P-ANCA.

Methods of testing

The most commonly used assays for ANCA testing are IIF and ELISA specific for anti-PR3 or anti-MPO. Other methods occasionally used include whole neutrophil ELISA, flow cytometry and various immunoblotting methods. The International Consensus Statement for testing and reporting ANCA recommends that all sera are screened for ANCA by IIF and that positive results are confirmed by specific ELISA.6

IIF is still considered to be the gold standard despite interpretation being so subjective. The method is relatively simple to perform and utilises ethanol-fixed neutrophils and fluorescein-tagged anti-human immunoglobulin G (IgG). The test can be interpreted as positive or negative, or a titre of ANCA can be given. Because a very wide range of patterns can be observed, interpretation and consequently the specificity and sensitivity of the test depends on the skill and experience of the laboratory scientist. For example, restricting the term C-ANCA to the classical appearance of central/interlobular accentuation of immunofluorescence staining will strengthen the correlation with anti-PR3 and a diagnosis of WG.7

ANCA are mostly IgG antibodies, mainly belonging to the IgG1, G2 and G4 subgroups. IgM ANCA has also been identified in patients presenting with haemorrhagic renopulmonary syndrome and infections. IgA ANCA has been described in patients with Henoch–Schönlein purpura. However, routine testing of sera for IgA and IgM antibodies reduces the specificity of the test for vasculitis. Some laboratories routinely use formalin-fixed neutrophils and HEp-2 cells as well as ethanol-fixed neutrophils as part of the screening routine to help distinguish potentially interfering autoantibodies.

ANCA patterns determined by IIF do not always correctly predict the target antigen. Furthermore, about 5% of ANCA-positive sera are positive only by ELISA while approximately 10% of IIF-positive C-ANCA are negative by ELISA. Therefore, ideally, all sera should be tested by both IIF and by ELISA specific for antibodies to PR3 and to MPO. ELISA should also be carried out on all sera showing immunofluorescence staining associated with other autoantibodies that may interfere with the interpretation of ANCA by IIF.

Various immunoblotting techniques can be used as a quick and easy method to confirm the presence of PR3 or MPO antibodies. The Quikcard version of this method, when used in conjunction with IIF, will provide rapid confirmation of the presence of PR3 or MPO antibodies which later can be quantified by ELISA.

It should be mentioned that the National External Quality Assessment Scheme (NEQAS) has demonstrated considerable variation in results when a single serum is tested in different laboratories. For IIF this variation reflects not only observer error but also the quality of the neutrophil substrate and conjugate in different commercial preparations.8 Discrepancies are also seen in ELISA results.9 Antigen preparations are not standardised and loss of conformational epitopes on PR3 and MPO can lead to false-negative results. Conversely, the high sensitivity of the method can give false-positive results.

To help preserve the native PR3 molecule, an indirect ELISA (capture assay) has been developed which involves coating the ELISA plate with a monoclonal antibody that binds purified PR3 antigen. This method is more sensitive than the conventional direct ELISA. However, this has to be balanced against loss of specificity and an increased rate of persistently high ANCA titres during clinical remission. Some studies conclude that the capture ELISA is more reliable than the direct ELISA in both diagnosis and detecting clinical relapse in WG.10,11 However, this superiority has not been demonstrated in all studies.12

Disease associations

Studies have shown that when the IIF and ELISA results are combined, the presence of C-ANCA and anti-PR3 has 99% specificity for the diagnosis of primary systemic vasculitis, as does the combination of P-ANCA and anti-MPO.13

The classical C-ANCA pattern on IIF and/or anti-PR3 are predominantly found in patients with WG, whereas P-ANCA and anti-MPO are more often seen in MPA, CSS and idiopathic necrotising glomerulonephritis.14 In the context of vasculitis, C-ANCA in more than 90% of cases is directed against PR3, whereas in 80–90% of cases P-ANCA reacts with MPO. Although C-ANCA (anti-PR3) is predominately associated with WG and P-ANCA (MPO) with MPA and CSS, there is no absolute specificity. Between 10 and 20% of patients with classical WG demonstrate P-ANCA and anti-MPO and an even larger number of patients with MPA or CSS have C-ANCA and anti-PR3. Furthermore, it should be stressed that 10–20% of patients with WG or MPA and 45–50% of CSS have negative ANCA results. Nevertheless, when both IIF and ELISA are performed, the sensitivity and specificity of anti-PR3 are approximately 75% and 99% respectively. The sensitivity depends on the patient population selected. C-ANCA and/or anti-PR3 are found in at least 90% of patients with diffuse disease and renal involvement. They occur less commonly in patients with localised WG.

Atypical ANCA reflect different and often unknown antibody specificities and are found in a large variety of conditions (Table 1). The clinical significance of these systems is not yet known.

TABLE 1. The occurrence of ANCA in different clinical conditions. Disease category                 Systemic vasculitis C PR3 P MPO ATYP BPI Other Positive patients Wegener's granulomatosis +++ +++ + +       85–95% Idiopathic pauci-immune necrotising and crescentric glomerulonephritis + + ++++ +++       60–80% Microscopic polyarteritis nodosa + + ++ ++       50–80% Churg–Strauss syndrome   ++ ++ ++       30–70% Rheumatological diseases C PR3 P MPO ATYP BPI Other Positive patients Rheumatoid arthritis     + + ++   + 20–40% Systemic lupus erythematosus     + + ++   + 20–30% Sjögren's syndrome     + + ++   + 11–26% Polymyositis/dermatomyositis         ++     11–16% Reactive arthritis         +      56% Juvenile idiopathic arthritis         +     35% Antiphospholipid syndrome         +     27% Inflammatory bowel disease C PR3 P MPO ATYP BPI Other Positive patients Ulcerative colitis         +++ + + 50–80% Crohn's disease         + +   10–40% Autoimmune liver diseases C PR3 P MPO ATYP BPI Other Positive patients Primary sclerosing cholangitis         +   + 26–87% Autoimmune hepatitis Type 1         +   + 65–90% Other disorders C PR3 P MPO ATYP BPI Other Positive patients Diabetes mellitus Type 1         +   + <20 cases Thyroid disease       + +   + 30% Goodpasture's syndrome       + +   + 40% Infectious diseases C PR3 P MPO ATYP BPI Other Positive patients Subacute bacterial endocarditis + +     +     <20 cases Human immunodeficiency virus     ++ + + ++   17–40% Amoebiasis         + +   <20 cases Malaria         + +   <20 cases Entamoeba hystolytica +++ +++           75% Leptospirosis + +       +   <20 cases Drugs C PR3 P MPO ATYP BPI Other Positive patients Hydralazine       + +   + 50–60% Propyltiouracil       + +   + 20–25% Minocycline       + +   + 40–100% D-penicillamine       + +   + <20 cases Sulfasalazine       + +   + <20 cases ATYP atypical; BPI bactericidal/permeability increasing protein; C classical C-ANCA by IIF; MPO myeloperoxidase; P classical P-ANCA by IIF; PR3 proteinase 3

When to test for ANCA

Guidelines on testing for ANCA in order to maximise the positive predictive value of the test have been suggested15 (Table 2).

TABLE 2. Recommendations for ANCA testing (after Guillevin et al 199615): indications for ANCA.

Patients suspected of Wegener's granulomatosis, microscopic polyangiitis, Churg–Strauss syndrome or idiopathic necrotising glomerulonephritis Chronic destructive disease of the upper airways Pulmonary nodules (not obviously malignant) Subglottic stenosis of the trachea Pulmonary–renal syndrome Glomerulonephritis Vasculitis of the skin with evidence of systemic disease Mononeuritis multiplex Retro-orbital mass Any other condition resembling systemic vasculitis

Value of ANCA levels in monitoring treatment

There has been much debate about the relationship between ANCA titre and the risk of relapse. Many workers have noted that a rise in ANCA titres often precedes a clinical relapse.2,16-20 This is particularly true for anti-PR3. Han et al21 studied the significance of serial ANCA titres as a tool to predict relapses in ANCA-associated vasculitis. They found that in 74% of patients a rise in ANCA titre occurred prior to relapse and in 52% of cases a 4-fold rise in ANCA was detected.

There is a consensus that capture ELISA is a promising tool for monitoring patients and anticipating relapses based on increasing anti-PR3 titres. However, while ANCA levels are useful to monitor disease and serve to alert clinicians to the possibility of relapse, they should not be the sole measure to guide therapy.22

Anticitrulline antibodies

The consensus view, backed up by clinical evidence, is that rheumatoid arthritis (RA) should be treated as early as possible to improve long-term outcome. The challenge, however, is to identify the patients among those presenting with early synovitis who have a bad prognosis and need anti-rheumatic drug therapy. The American College of Rheumatology classification criteria for RA are more than 90% specific if a patient is followed for long enough but are of little use in early diagnosis. Rheumatoid factor (RF), which is neither specific nor completely sensitive for the diagnosis of RA, also has its limitations. Into this arena has come the detection of antibodies to citrullinated proteins that show great specificity for RA and the potential to indicate which patients are likely to have destructive joint disease.

Antibody specificity

Antibodies to citrullinated proteins comprise a family of autoantibodies which includes antiperinuclear factor (APF), antikeratin antibodies (AKA) and anti-Sa. They were identified over 40 years ago when Mandema et al23 described APF, which targeted an antigen present in the keratohyaline granules surrounding the nucleus of human buccal mucosal cells. The antigen responsible for APF and the subsequently described AKA24 was later shown to be citrulline residues in the epithelial intermediate filament-associated protein, filaggrin, which is involved in the aggregation of cytokeratin filaments during the cornification of the epidermis.25,26 Although APF, AKA and anti-filaggrin antibodies were found to be both specific and fairly sensitive for RA, these assays never really took off in routine clinical practice, largely for technical reasons. Subsequently, a reproducible and readily available assay was developed by Schellekens et al27 to detect antibodies (anticyclic citrullinated peptide – anti-CCP) to a single cyclic citrullinated peptide that has a 3-dimensional structure. This assay enables this peptide to be optimally recognised by corresponding antibodies in RA sera. A second-generation anti-CCP assay with enhanced sensitivity and specificity is now commercially available28 and third-generation assays are in development.

Citrulline is the product of post-translational modification of protein-bound arginine, catalysed by peptidylarginine deiminase (PAD).29 The high specificity of antibodies to citrullinated proteins for RA and their association with joint damage (see later) has excited interest in the potential involvement of this immune response in aetiopathogenesis of the disease. As well as occurring during the terminal differentiation of keratinocytes, citrullination occurs during cell death and injury.30 Anti-CCP antibodies are produced locally within the joint of RA patients.31 There are a number of intra- and extracellular autoantigen candidates including vimentin, histones and fibrin. Citrullinated proteins, such as fibrin, are present in RA synovium but the finding is not specific for RA and occurs in synovial tissue from patients with a variety of arthritides. However, there is a strong association between anti-CCP and major histocompatibility complex (MHC) class II molecules expressing the shared epitope (e.g. human leucocyte antigen (HLA)-DRB1 *0101, *0401 and *0404) with evidence that expression of shared epitope facilitates presentation of citrulline-containing peptides to T lymphocytes.32 This immune response may be important in sustaining the inflammatory response. Interestingly, a Japanese study33 showed an association between a particular haplotype of PADI4, one of the isotypes of PAD expressed in the joint. Since this haplotype encodes a messenger ribonucleic acid (mRNA) with greater stability, potentially increasing the citrullination of protein, this could be an additional explanation for anti-CCP in RA. However, this observation has not been reproduced in European populations.34

Clinical associations

A very high proportion of patients with established RA have (predominately IgG) anti-CCP antibodies35 (Table 3). Up to 80% with RA are positive. By contrast, only 2–5% disease controls and 0–1% healthy controls have been found to be positive (Table 4). These antibodies appear early in RA and can predate the clinical onset of disease by years.36 Most, but not all, patients are also positive for RF. Multiple studies have shown that anti-CCP is associated with joint damage,37-40 even in those who are RF-negative, but, unlike RF, not with extraarticular features.41 Also in contrast to RF, the titre of anti-CCP does not fall with 'standard' anti-rheumatic therapy.42 Significant falls have been reported in patients clinically responding to infliximab43 and to rituximab.44 Anti-CCP has been reported to help distinguish between polymyalgia rheumatica and a polymyalgic onset of RA in elderly people.45 They have been reported in 13% of children with polyarticular juvenile idiopathic arthritis (JIA),46 showing an association with DR4 positivity and erosive joint disease, but in only 2% of other types of JIA. Anti-CCP has also been reported in 8–10% of patients with psoriatic arthritis,47-49 again associated with shared epitope and severity of joint damage. By contrast, these antibodies have been found in approximately 7% of patients with primary Sjögren's, often at high titre, but not necessarily associated with erosive joint disease.50

TABLE 3. The sensitivity and specificity of anti-CCP in rheumatoid arthritis.
Condition	Sensitivity	Specificity
Established rheumatoid arthritis	70–80%	95–98%
Rheumatoid arthritis of less than 6 months' duration	45–65%	95–98%

TABLE 4. Occurrence of anti-CCP in conditions other than rheumatoid arthritis.
Condition	% frequency
Healthy controls	<1
All disease controls	2–5
Infections	1
Psoriatic arthritis	8–10
Juvenile idiopathic arthritis	2–13
Various connective tissue diseases	5–7

Early synovitis

Potentially, it is within the context of the early synovitis clinic that identification of anti-CCP can be of most clinical value. Several studies51-53 have demonstrated that anti-CCP predicts RA in patients presenting with an as yet undifferentiated arthritis. In this setting, the assay shows sensitivity ranging from 45–65% but uniformly high specificity and high predictive value.54

Conclusion

ANCA and anti-CCP meet the reqirements of clinically relevant biological markers. They have high specificity for diagnosis as well as being valuable for prognosis. However, as with all serological testing, the results need to be interpreted in the light of the clinical presentation. ANCA testing now has an established place in diagnostic serology laboratories. Many of the pitfalls in interpreting results can be overcome by combining an ANCA screen using IIF with a specific ELISA for antibodies to PR3 or MPO. This increases substantially the positive predictive value of serological testing for ANCA-associated vasculitis. However, it must be appreciated that not all patients with vasculitis are ANCA-positive, including a substantial proportion of patients with limited WG. Whether or not a positive test for antibodies to PR3 or MPO precludes the need for biopsy is controversial. In our practice, positive serology justifies immediate institution of treatment in somebody seriously ill. However, in other situations, evidence to date55 suggests that it is preferable to obtain histological confirmation of the diagnosis before committing someone to long-term immunosuppression. Similarly, although using ANCA titres in follow-up helps alert the clinician to relapse, serology should not be the sole measure on which to base management decisions.

Similarly, the presence of anti-CCP is the most specific marker to date of RA. It has been approved by the Food and Drug Administration (FDA) for diagnosis of RA and routine testing is becoming increasingly widespread in the UK. The combination of high specificity and the prognostic potential of anti-CCP for predicting joint damage means that testing is valuable in the context of early arthritis. We utilise anti-CCP testing in our 'early synovitis clinic' when a positive result influences the decision to introduce early anti-rheumatic drug therapy. Current evidence54,56 suggests there is value in combining testing for both anti-CCP and RF. Positive tests for both markers gives the highest positive predictive value in recent onset arthritis for the evolution of destructive RA. Negativity for both suggests an alternative diagnosis. The drawback of anti-CCP testing in early arthritis is its lowish sensitivity. This may be explained by the absence of relevant citrullinated epitopes on the test antigen. This is being addressed as future generations of tests are developed.

Other Titles in this series

• June 2008 No 14
  Sports-Related Musculoskeletal Disorders
• October 2007 No 13
  Glucocorticoids in Rheumatic Diseases
• June 2007 No 12
  Immunisation of Immunosuppressed Patients with Rheumatic Diseases
• February 2007 No 11
  Rheumatological Manifestations of Malignant Haematological Conditions
• October 2006 No 10
  Work-Related Disorders of the Upper Limb
• June 2006 No 9
  Qualitative Research: Its Relevance and Use in Musculoskeletal Medicine
• October 2005 No 7
  Malignancy and the Rheumatic Diseases
• June 2005 No 6
  Systemic Amyloidosis in the Rheumatic Diseases
• February 2005 No 5
  Heritable Collagen Disorders
• October 2004 No 4
  An Update on Gout
• June 2004 No 3
  Surgery for Arthritis: Total Hip and Knee Joint Replacement
• February 2004 No 2
  Idiopathic Inflammatory Myopathies
• September 2003 No 1
  Low Back Pain

download a pdf of this publication

6628.pdf
(272Kb)