What is the signifi cance in routine care of c-ANCA/PR3-ANCA in the absence of systemic vasculitis? A case series A. Knight1, A. Ekbom2, L. Brandt2, J. Askling2,3

1Department of Rheumatology, Uppsala ABSTRACT litides, and especially for Wegener’s University Hospital, Uppsala, Sweden; Objective. ANCA has come to play an granulomatosis (WG) (1), although not 2 Clinical Epidemiology Unit and important role in diagnosing vasculitis. pathognomonic as such ANCAs have 3Rheumatology Unit, Department of In selected populations c-ANCA/PR3- been reported in other conditions (2) Medicine, Karolinska Hospital and Institute, Stockholm Sweden. ANCA has a high specifi city and sensi- and in elderly healthy individuals (3). tivity for vasculitis. In clinical practice, Conversely, not all patients with sys- A. Knight, MD; A. Ekbom, MD, PhD; L. Brandt, BSc; J. Askling, MD, PhD. how individuals with c-ANCA/PR3- temic vasculitides, or WG, are ANCA ANCA but without suffi cient evidence Please address correspondence to: positive (4). Dr. Ann Knight, Department of of systemic vasculitis should be man- Although neither included in the Amer- Rheumatology, Uppsala University aged is unclear. We therefore retrospec- ican College of Rheumatology (ACR Hospital, S-751 85 Uppsala, Sweden. tively assessed the disease panorama 1990) – nor in the Chapel Hill Consen- E-mail: [email protected] and outcome in a consecutive series of sus Conference (CHCC 1994) – criteria Received on February 1, 2007; individuals with c-ANCA/PR3-ANCA, for the classifi cation of WG or any sys- accepted in revised form on May 16, 2007. and studied in detail those individuals temic vasculitis, testing for c-ANCA/ Clin Exp Rheumatol 2008; 26 (Suppl. 49): who turned out not to fulfi l criteria for PR3-ANCA is an important feature of S53-S56. vasculitic disease. the work-up for patients with suspected © CopyrightCopyright CLINICAL AND Methods. The study population con- systemic vasculitis. ANCAs should, EXPERIMENTAL RHEUMATOLOGY 2008.2008. sisted of 74 consecutive patients who however, always be used together with all had a positive test for C-ANCA and clinical assessment tools (5, 6). Key words: ANCA, non-vasculitic PR3-ANCA between 1992 and 2002 at The fi nding of a c-ANCA/PR3-ANCA disease, outcome, Sweden. the Immunology laboratory at Uppsala supportive of a clinical diagnosis of, University Hospital, Sweden. The pa- for example, WG poses little chal- tients’ medical fi les werewere reviewedreviewed and lenge. However, clinicians fi nding a their diagnosis re-evaluated through c-ANCA/PR3-ANCA in individuals June 2006. whose other investigations eventually Results. 18 of the 74 ANCA-positive provide insuffi cient evidence of WG individuals did not present clinical evi- or any other vasculitides are faced with dence supportive of, or insuffi cient to three important questions that each support, a diagnosis of systemic vas- prompts a medical decision: First, what culitis, but presented a range of other is the chance that such a patient actu- diseases. During a mean follow-up of ally suffers from a systemic vasculitis? 6.8 years, none of these 18 patients Second, what is the chance that such a developed vasculitis. patient will go on to develop a systemic Conclusions. Individuals with a posi- vasculitis? Third, will such patients de- tive c-ANCA and PR3-ANCA but no velop any other serious disease? vasculitis at the time of testing run an Just as it is important to be able to cor- unknown but likely small risk of later rectly diagnose a vasculitic disease in developing vasculitis. In this group, a an early stage, and institute adequate positive ANCA may represent back- treatment, it is vital to avoid erroneous ground noise (borderline titres) or be a immunosuppressive treatment in pa- marker of infl ammatory activity rather tients without vasculitides (7). than of vasculitic disease (high titres). To our knowledge, no previous study has assessed the disease spectrum in Introduction patients with c-ANCA/PR3-ANCA The presence of ANCA (anti-neu- but with insuffi cient signs and symp- trophil cytoplasmic antibodies) directed toms of WG or other systemic vascu- against proteinase 3 as tested with im- litis at the time of testing and ffollowedollowed munofl ourescence (IIF) (cANCA) and such patients over time. Other studies ELISA (PR3-ANCA) has been regarded have assessed the disease spectrum at Competing interests: none declared. as specifi c for certain systemic vascu- the time of testing in ANCA positive

S-53 ANCA but not vasculitis / A. Knight et al. patients without systemic vasculitis, litic disease (75%); 53 patients (72%) out a systemic vasculitides (1) and that and noted the low predictive value of met the ACR criteria for WG and 67% a false positive may lead to misdiagno- the test in unselected populations (8). of these cases also had a biopsy-verifi ed sis of vasculitis (7). diagnosis of WG. When we applied the In our study, we have approached the Patients and methods algorithm for the classifi cation of the clinically important issue of what the The department of clinical immunol- ANCA associated vasculitides, as sug- presence of c-ANCA and PR3-ANCA ogy at Uppsala University Hospital gested by Watts et al. (10), of the 74 in the absence of vasculitis represents. has served as a catchment area corre- patients, 53 patients were classifi ed as Our results suggest, with the reserva- sponding to Uppsala county and sev- WG and one as MPA. The two remain- tion of the limited number of subjects eral smaller primary referral hospitals ing cases had vasculitis associated with (n=74), that none of the ANCA-positive in central Sweden. Between 1992 and RA and ulcerative colitis. individuals who did not have vasculitis 2002, 4,997 tests were performed; of The remaining 18 (out of 74) patients but presented symptoms that might in- these, 444 tests (in 74 individuals) were suffered from a heterogeneous group dicate vasculitic disease (and which positive for IIF /c-ANCA and ELISA/ of medical conditions, signs and symp- prompted ANCA testing) developed PR3-ANCA, each individual having toms such as other rheumatic diseases, WG or any other vasculitic disease dur- one or more positive tests. The indirect bacterial and viral infections, and in- ing the mean follow-up of 6.8 years. immunofl ourescence (IIF-ANCA) was fl ammatory bowel disease (Table I). In WG, overall 75-90% of all patients performed with human granulocytes as The values of the ELISA tests among develop c-ANCA and PR3-ANCA, substrate, the result given as positive or the 18 patients without a confi rmed but whether this antibody may precede negative (Wieslab, Lund, Sweden) (9). systemic vasculitis at the time of test- clinical symptoms, as is the case with The enzyme linked immunosorbent as- ing varied from 10 to 140 U/ml (mean anti-CCP in RA (11) or anti-dsDNA in say (ELISA -ANCA) was performed 36, median 15). Although these values SLE (12) is still unknown. Accumulat- with a commercial kit (Wielisa R PR3- were lower than those of the 56 patients ing evidence supports the hypothesis 102X [9], with a cut-off value of 10 U/ with confi rmed systemic vasculitis (10 that c-ANCA is indeed involved in the ml, sensitivity and specifi city of the test to 910, mean 103, median 28), this dif- pathophysiology of WG, but the fi nding for WG, according to the manufacturer, ference did not reach statistical signifi - that none of the 18 c-ANCA positive in- being 92% and 99% respectively. Test- cance (Mann Whitney p=0.11). dividuals developed systemic vasculitis ing of blood donor sera revealed no Eleven of the 8 non-vasculitis patients may suggest that c-ANCA also refl ects positive [>20U] or borderline [10-20 had only one positive test, all with val- neutrophil-activating properties that are U] tests [9]). ues in the borderline region (10-20 U). not specifi c to systemic vasculitides. In The study population consisted of all in- By contrast, 7 out of the 18 non-vas- RA and SLE, ANCA has been found to dividuals (n=74), treated at any hospital culitic patients, all with more than one refl ected high infl ammatory activity in in the catchment area, whose blood was positive test, had PR3-ANCA values > the disease (13). The pro-infl ammatory sent to Uppsala for testing, and who 20 in their fi rst test. These 7 cases are properties of ANCA may enhance or tested positive for both ELISA and IIF more closely presented in Table II. maintain the infl ammatory process, per- ANCA at least once between 1992 and During the follow-up period from fi rst haps accounting for the greater relapse 2002. Through scrutiny of the medical positive ANCA until 30 June 2006 risk in WG-patients with a persisting fi les of these 74 individuals, one of us (range 3 years to 12 years, mean 6.8 ANCA. (AK) assessed the reason for testing, years), none of the 18 patients without Some of the positive tests reported the underlying medical condition at the a diagnosis of systemic vasculitis at the may have represented analytical false time of testing and the possible devel- time of testing developed such disease, values, especially those with only one opment of a systemic vasculitis or any nor any new signs or symptoms there- test and ELISA values in the borderline other serious disease during the remain- of. Four patients died, all above the age positive range. der of the study period, which ended in of 75, from causes unrelated to vascu- Our study was designed as a routine June 2006. The clinical diagnosis was litic disease. care assessment, and as such, a clinical re-evaluated, and patients with fi nd- counterpart to studies of ANCA per- ings compatible with vasculitis were Discussion formance in highly selected and clini- assessed and classifi ed according to Our population-based case-series indi- cally typical patients with vasculitides the ACR- criteria and when biopsy had cates that patients who test positive for from referral-centres (14). One limita- been performed, also by the CHCC-cri- c-ANCA/PR3-ANCA but do not have tion is our retrospective assessment of teria. Although these criteria primarily suffi cient clinical or other evidence of a the diagnoses and follow-up. We used are for classifi cation they have been systemic vasculitis display a wide array the ACR criteria to defi ne the systemic widely to defi ne vasculitides. of medical conditions but have a low vasculitides, but 67% of the patients probability of progressing to systemic also had their vasculitis-diagnosis con- Results vasculitides. Previous studies have fi rmed by biopsy. Since all individu- Of the 74 patients with c-ANCA and pointed out that c-ANCA/PR3-ANCA als were followed using their medical PR3-ANCA, 56 patients had a vascu- should not be used to rule in or to rule charts or with direct contact with their

S-54 ANCA but not vasculitis / A. Knight et al. treating physician, we feel confi dent The four patients that died were all of To conclude, our series suggests that that any serious disease during the fol- high age and none of the deaths were in clinical practice, the clinical sig- low-up period did not pass undetected. related to vasculitis. nifi cance of c/PR3-ANCA in patients

Table I. Characteristics of the 18 patients with a positive c-ANCA/PR3 but no concurrent evidence of Wegener’s granulomatosis or other systemic vasculitis.

Patient Sex Age PR3 - c-ANCA Medical condition at Signs and symptoms Medical diagnosis Observation Vital status at 1st ANCA 1st positive PR3 leading to PR3 testing after re-assessment time at end of positive (months) follow-up PR3 (yrs)

1 F 38 140 + HLA B27 related arthritis Fever, malaise, high ESR, Ankylosing spondylitis, 96 Alive, no anaemia, arthralgia possible pneumonia vasculitis

2 M 83 22 ++ Sarcoidosis + pneumonia Fever Sarcoidosis+ pneumonia 36 Deceased, +fi brosis 981027

3 F 66 20 + None Acute renal failure, Nephropathia epidemica 84 Alive, no hematuria vasculitis

4 F 17 127 +++ Ulcerative colitis Fever, back pain, hematuria, Ulcerative colitis +urinary 83 Alive, proteinuria, cough, pleuritis tract infection no vasculitis

5 M 31 110 ++ None Fever, exanthema, artritis, Streptococcal infection, 68 Healthy, no elevated infl ammatory rheumatic fever vasculitis parameters

6 M 75 44 +++ Systemic sclerosis Cutaneous vasculitis Systemic sclerosis, infection 5656 DeceasedDeceased 060103. No vasculitis

7 M 32 31 + None Blocked nose Observation for possible WG 58 Healthy, no vasculitis

8 M 22 10 ++ Ankylosing spondylitis Renal failure, proteinuria Ankylosing 146 Alive, no spondylitis,amyloidosis, vasculitis kidney transplant

9 F 35 14 ++ Diabetes mellitus Renal failure, proteinuria Diabetic nephropathy 147 Alive, no vasculitis

10 M 33 15 + None Renal failure,hematuria, Nephropathia epidemica 140 Alive, no proteinuria vasculitis

11 M 70 14 ++ None Upper respiratory tract Liver failure 60 Deceased symptoms, purpura , 2001. No liver failure vasculitis at post-mortem

12 M 66 10 + None Fever, ESR, malaise, Gr+ endocarditis 102 Alive, no back pain vasculitis

13 M 54 12 + Ulcerative colitis, Ulcerative colitis 100 Alive, no hypertension, diabetes vasculitis

14 F 56 15 + Primary Sjögren’s Upper respiratory tract Primary Sjögren’s 84 Alive, no syndrome symptoms, malaise syndrome vasculitis

15 M 66 12 ++ U.C., hypertension, ESR,Creatinine and Hepatitis C,U.C., 84 Alive, no diabetes liverenzymes elevated diabetes vasculitis

16 F 32 12 ++ None Artralgias, fatigue, upper Chronic fatigue syndrome 47* Alive, no respiratory tract symptoms vasculitis

17 F 84 11 +++ None Ear pain, epistaxis External otitis 39 Deceased 2005, no vasculitis

18 F 84 18 ++ None Fever, cutaneous vasculitis Infection NUD, erythema 49 Alive, no multiforme vasculitis

*Lost to follow-up after moving abroad after 47 months of observation.

S-55 ANCA but not vasculitis / A. Knight et al.

Table II. Description of the investigations performed and treatments given in the seven RA: Clinical and biological assessment in cases with repeatedly positive PR3-ANCAs >20, but without vasculitis. systemic necrotizing vasculitides. Clin Exp Rheumatol 22006;006; 2244 ((Suppl.Suppl. 441):1): SS92-9.92-9. No. Findings indicative Diagnostic work-up Treatments Final diagnosis 6. SANDERS JS, STASSEN PM, VAN ROSSUM AP, of vasculitis KALLENBERG CG, STEGEMAN CA: Risk fac- tors for relapse in anti-neutrophil cytoplas- 1 Pulmonary infi ltrates Repeated lung biopsies NSAID and Ankylosing mic antibody (ANCA)-associated vasculi- negative analgesics spondylitis tis: tools for treatment decisions? Clin Exp Rheumatol 22004;004; 2222 ((Suppl.Suppl. 336):6): SS94-101.94-101. 2 Progressive pulmonary ENT incl biopsies negative Bronchodilators Sarcoidosis and 7. DAVENPORT A: “False positive” perinuclear infi ltrates respiratory and cytoplasmic anti-neutrophil cytoplasmic insuffi ciency antibody results leading to misdiagnosis of Wegener’s granulomatosis and/or micro- 3 Acute renal failure ENT and x-ray lungs No specifi c Nefropathia scopic polyarteritis. Clin Nephrol 11992;992; 337:7: negative treatment epidemica 124-30. Hanta-virus serology 8. MANDL LA, SOLOMON DH, SMITH EL, LEW positive RA, KATZ JN, SHMERLING RH: Using an- 4 Pleurisy, fever, HRCT and angiography Antibiotics, Ulcerous colitis tineutrophil cytoplasmic antibody testing to hematuria of lungs neg.U-culture azathioprin diagnose vasculitis: can test-ordering guide- pos E.coli mesalazin lines improve diagnostic accuracy? Arch In- tern Med 22002;002; 1162:62: 11509-14.509-14. 5 Fever, exanthema, ENT, x-ray lungs neg. Penicillin β-Strep infection, 9. www.wieslab.se. arthritis Ekg T-wave inversion, rheumatic fever 10. WATTS R, LANE S, HANSLIK T et al.: Devel- throat swab pos β-strep opment and validation of a consensus meth- odology for the classifi cation of the ANCA- 6 Leucocytoclastic rash ENT and x-ray of lungs Cortisone short time Systemic sclerosis associated vasculitides and polyarteritis no- neg. dosa for epidemiological studies. Ann Rheum Dis 2006. 7 Blocked nose CT sinus, x-ray lungs neg. None Healthy 11. RÄNTAPÄÄ-DAHLQVIST S, DE JONG BA, Repeated blood BERGLIN E et al.: Antibodies against cyclic chemistry neg citrullinated peptide and IgA rheumatoid factor predict the development of rheuma- who test positive but do not present actual clinical testing conditions. Clin Immu- toid arthritis. Arthritis Rheum 2003; 48: 2741-9. suffi cient evidence for a systemic vas- nol 2002; 103: 196-203. 2. SCHMITT WH, VAN DER WOUDE FJ: Clini- 12. ARBUCKLE MR, MCCLAIN MT, RUBERTONE culitides at the time of testing, is lim- cal applications of antineutrophil cytoplas- MV et al.: Development of autoantibodies be- ited. The exact nature of these positive mic antibody testing. Curr Opin Rheumatol fore the clinical onset of systemic lupus ery- non-vasculitis associated ANCA tests 2004; 16: 9-17. thematosus. N Engl J Med 2003; 349: 1526- 3. ANDERSEN-RANBERG K, M HO-M, WIIK A, 33. (cross-reactivity, non-specifi c neutro- JEUNE B, HEGEDUS L: High prevalence of 13. MANOLOVA I, DANTCHEVA M: Antineu- phil-activating properties, marker of autoantibodies among Danish centenarians. trophil cytoplasmic antibodies in Bulgarian disease activity, analytical false values Clin Exp Immunol 22004;004; 1138:38: 1158-63.58-63. patients with rheumatoid arthritis: character- etc.) remains unknown. 4. BOOMSMA MM, STEGEMAN CA, VAN DER ization and clinical associations. Rheumatol LEIJ MJ et al.: Prediction of relapses in We- Int 22005;005; 226:6: 1107-14.07-14. gener’s granulomatosis by measurement of 14. NOLLE B, SPECKS U, LUDEMANN J, ROHR- References antineutrophil cytoplasmic antibody levels: a BACH MS, DEREMEE RA, GROSS WL: Anticy- 1. RUSSELL KA, WIEGERT E, SCHROEDER DR, prospective study. Arthritis Rheum 2000; 43: toplasmic autoantibodies: their immunodiag- HOMBURGER HA, SPECKS U: Detection of 2025-33. nostic value in Wegener granulomatosis. Ann anti-neutrophil cytoplasmic antibodies under 5. MUKHTYAR CB, FLOSSMANN O, LUQMANI Intern Med 11989;989; 1111:11: 28-40.28-40.

S-56