Classification, presentation, and initial treatment of Wegener's granulomatosis in childhood




To compare the criteria for Wegener's granulomatosis (WG) of the American College of Rheumatology (ACR) with those of the European League Against Rheumatism/Pediatric Rheumatology European Society (EULAR/PRES) in a cohort of children with WG and other antineutrophil cytoplasmic antibody (ANCA)–associated vasculitides (AAVs), and to describe the interval to diagnosis, presenting features, and initial treatment for WG.


Eligible patients had been diagnosed by site rheumatologists (termed the “MD diagnosis”) since 2004. This diagnosis was used as a reference standard for sensitivity and specificity testing of the 2 WG classification criteria. Descriptive analyses were confined to ACR-classified WG patients.


MD diagnoses of 117 patients (82 of whom were female) were WG (n = 76), microscopic polyangiitis (n = 17), ANCA-positive pauci-immune glomerulonephritis (n = 5), Churg-Strauss syndrome (n = 2), and unclassified vasculitis (n = 17). The sensitivities of the ACR and EULAR/PRES classification criteria for WG among the spectrum of AAVs were 68.4% and 73.6%, respectively, and the specificities were 68.3% and 73.2%, respectively. Two more children were identified as having WG by the EULAR/PRES criteria than by the ACR criteria. For the 65 ACR-classified WG patients, the median age at diagnosis was 14.2 years (range 4–17 years), and the median interval from symptom onset to diagnosis was 2.7 months (range 0–49 months). The most frequent presenting features by organ system were constitutional (89.2%), pulmonary (80.0%), ear, nose, and throat (80.0%), and renal (75.4%). Fifty-four patients (83.1%) commenced treatment with the combination of corticosteroids and cyclophosphamide, with widely varying regimens; the remainder received methotrexate alone (n = 1), corticosteroids alone (n = 4), or a combination (n = 6).


The EULAR/PRES criteria minimally improved diagnostic sensitivity and specificity for WG among a narrow spectrum of children with AAVs. Diagnostic delays may result from poor characterization of childhood WG. Initial therapy varied considerably among participating centers.

Primary systemic vasculitis (PSV) of childhood encompasses a group of rare conditions that have in common primary vascular inflammation, often in critical organs such as the heart, kidneys, lungs, and brain (1). They are acutely and chronically life-threatening and have high associated disease- and treatment-related morbidity. Childhood PSV affects fewer than 25 per 100,000 children, with the “acute” subtypes including Henoch-Schönlein purpura and Kawasaki disease accounting for the majority of cases (2). The frequencies of the distinct subtypes of “chronic” childhood PSV are difficult to ascertain. Among the chronic childhood primary systemic vasculitides, Wegener's granulomatosis (WG) is one of the most common, with incidence rates ranging between 0.03 and 3.2 per 100,000 children per year (1–3). Chronic childhood PSV is the diagnosis assigned to <3% of all patient referrals in some US pediatric rheumatology clinics (4), and individual US/Canadian pediatric rheumatologists will typically see 0–5 new patients per year (5). Consequently, most of our knowledge about the presentation, course, outcome, and optimal treatment of childhood PSV either comes from a few small case series or has been adapted from studies of adults, without validation of its true applicability to children.

Between one-third and two-thirds of children with chronic PSV are described as having unclassified vasculitis (3, 4), and one-third of children with WG diagnosed by expert opinion (6) never fulfill the American College of Rheumatology (ACR) 1990 classification criteria (7). It is not known whether these described “difficulties” in assigning a specific classification are related to the fact that the ACR 1990 criteria for the classification of vasculitis (8) and the subsequent Chapel Hill Consensus Conference (CHCC) disease definitions (9) were based largely on adult data, or whether they result from the fact that children with a developing immune system have a modified disease expression. The difficulties described for adult populations in distinguishing between the so-called antineutrophil cytoplasmic antibody (ANCA)–associated vasculitides (AAVs) (10), including WG, microscopic polyangiitis (MPA), and Churg-Strauss syndrome (CSS), have not been addressed in the limited pediatric reports of childhood PSV.

In 2005, under the auspices of the European League Against Rheumatism (EULAR) and the Pediatric Rheumatology European Society (PRES), existing ACR criteria were modified with a view to improving their applicability in children (11). The resulting proposed system of classification and criteria (EULAR/PRES criteria) have yet to be validated in a cohort of children with vasculitis, and specifically, it is not known whether the EULAR/PRES criteria for classifying WG in children will “capture” additional children diagnosed as having WG but not fulfilling ACR criteria, children with other diagnostic subsets of AAV, or children with previously unclassified chronic PSV.

In the context of establishing a network of pediatric investigators interested in advancing our knowledge about childhood PSV, we undertook a pilot, multicenter, contemporary inception cohort project called ARChiVe (A Registry for Childhood Vasculitis: e-entry) in collaboration with members of the Childhood Arthritis and Rheumatology Research Alliance (CARRA). This project initially focused on children with WG together with other AAVs including MPA, CSS, ANCA-positive pauci-immune glomerulonephritis (GN), and unclassified vasculitis. A list of collaborating centers and investigators in the ARChiVe network is shown in Appendix A.

The specific aims of this pilot study were to compare sensitivity and specificity of the adult-derived ACR criteria for WG with those of the proposed EULAR/PRES pediatric-specific criteria among a group of children with chronic vasculitis whose diagnosis fell within the spectrum of diseases related primarily by ANCA association and small-to-medium vessel size and whose clinical characteristics often overlapped. Additionally, we aimed to characterize the time from symptom onset to diagnosis, the presenting features, and the initial treatment of the patients in the cohort with ACR-defined WG.


Since the launch of the ARChiVe registry in March 2007, pediatric rheumatologists at 30 CARRA-associated geographically diverse institutions in the US (n = 26) and Canada (n = 4) have contributed patients. Eligible patients were all children followed up at participating centers who were diagnosed after January 1, 2004 as having WG, MPA, CSS, ANCA-positive pauci-immune GN, and unclassified vasculitis and who were age <18 years at the time of diagnosis. The diagnosis, established by the treating pediatric rheumatologist at each of the ARChiVe network sites, was entered in the database and considered to be the “reference standard” and described in this article as the “MD diagnosis.” Using this deliberately sensitive entry criterion, we anticipated that as many as one-third of all patients might be entered as unclassifiable (3, 4), one-third of the patients diagnosed as having WG might not be classifiable by ACR criteria (6, 7), and some patients with MPA might also fulfill criteria for WG (12). Specific patient data items for categorical capture included all criteria required for diagnosis using either the 1990 ACR (7) or EULAR/PRES (11) classification systems (Table 1); patients in the cohort fulfilling criteria for diagnosis of WG according to either of these criteria were identifiable by computation of data.

Table 1. Comparison of the ACR and EULAR/PRES classification criteria for WG*
  • *

    Differences between American College of Rheumatology (ACR) and European League Against Rheumatism/Pediatric Rheumatology European Society (EULAR/PRES) classification systems and for individual criteria are italicized. WG = Wegener's granulomatosis; CT = computed tomography; GN = glomerulonephritis; anti-PR3 = anti–proteinase 3; ANCA = antineutrophil cytoplasmic antibody; cANCA = cytoplasmic ANCA.

1990 ACR criteria
 A patient is said to have WG when 2 of the following 4 criteria are present:
  Nasal or oral inflammation
  Abnormal chest radiograph
  Abnormal urinary sediment (microhematuria [>5 red blood cells per high-power field] or red blood cell casts in urine sediment)
  Granulomatous inflammation on biopsy
EULAR/PRES criteria
 A patient is said to have WG when 3 of the following 6 criteria are present:
  Nasal or sinus inflammation
  Abnormal chest radiograph or chest CT scan
  Abnormal urinalysis (hematuria and/or significant proteinuria)
  Granulomatous inflammation on biopsy/necrotizing pauci-immune GN
  Subglottic, tracheal, or endobronchial stenosis
  Anti-PR3 ANCA or cANCA staining

For patients diagnosed between January 2004 and the launch of ARChiVe in March 2007, data were collected by review of available medical records. Data were collected prospectively for patients who were diagnosed between March 2007 and November 2008. In both instances data collected were for the period from clinical onset and presentation until 2 months after diagnosis. The ARChiVe uses a Web-based interface for data entry of predominantly categorical variables, which include the following: MD diagnosis; date of symptom onset; demographic data; family and past medical history; presenting/diagnostic features; physical examination findings (including height, weight, blood pressure, and pubertal staging); results or reports of laboratory testing including ANCA testing (cytoplasmic and/or perinuclear pattern of ANCA [cANCA and/or pANCA, respectively]), histopathology, and other procedures including diagnostic imaging, bronchoscopy, and pulmonary function tests; and initial therapy detailing the use of oral prednisone doses in mg/kg/day (low [<0.5], medium [0.5–1.5], high [>1.5]) and intravenous (IV) corticosteroids, oral or IV cyclophosphamide, other disease-modifying or “biologic” therapies, and other concurrently used drugs (e.g., anticoagulants, antihypertensive medications).

Other data set items captured included all items describing disease activity that are used in the Birmingham Vasculitis Activity Score (13, 14) in addition to common pediatric presenting features, unique diagnostic features for classification, or measurements of disease activity that were generated from a survey of CARRA members and incorporated into a standardized data collection form by a subsequent vasculitis consensus working group.

All data were reviewed at the main coordinating center for completeness and quality. Data were reviewed to ensure that there were no duplicate cases, since some of the institutions in the ARChiVe network reported patients who were either transferred from other medical institutions or referred to them for a “second opinion.”

The study protocol was approved by the local research ethics board at each participating center. Informed consent for participation was obtained from parents, and informed consent or assent was obtained from patients for both retrospective and prospective recruitment as applicable.

Because the data were of a cross-sectional nature, descriptive statistical analyses were performed. In order to make our data comparable with those of previously reported cohorts, we described presenting features and treatment of patients who fulfilled ≥2 of the ACR classification criteria; this does not mean to assume that the ACR criteria best define childhood WG. Frequency and distribution of the sociodemographic, clinical, and treatment characteristics of pediatric WG patients were provided using medians with ranges and percentages where applicable.

To compare the ACR and EULAR/PRES classification criteria for WG, all patients in the registry with any physician-assigned diagnosis (WG, MPA, CSS, ANCA-positive pauci-immune GN, or unclassified vasculitis) were analyzed. To assess the degree of agreement between the MD, ACR, and EULAR/PRES WG diagnoses, exploratory analyses to ascertain convergent validity (chance-corrected agreement) were performed using the kappa statistic (15). To determine discriminant validity of criteria in the absence of a gold standard is difficult; MD diagnosis of patients enrolled by >30 contributing physicians could not be considered a gold standard. However, for these exploratory analyses we used the MD diagnosis, not as a gold standard, but as the reference standard to calculate the ACR and EULAR/PRES classification criteria sensitivity and specificity. All analyses were performed using the Statistical package for the Social Sciences, version 16.0 (SPSS, Chicago, IL).


During the study period, 117 pediatric patients (82 female [70.1%]) from 30 centers were recruited into the ARChiVe cohort; 37 (31.6%) were collected prospectively. The MD diagnoses of the patients in the cohort were WG (n = 76), MPA (n = 17), ANCA-positive pauci-immune GN (n = 5), and CSS (n = 2). Seventeen patients were not considered classifiable under any of these categories (unclassified vasculitis).

Evaluating classification, and comparing existing ACR and proposed EULAR/PRES criteria for WG.

Agreement was moderate between the MD diagnosis and either the ACR criteria or the EULAR/PRES criteria (κ = 0.346 and κ = 0.476, respectively). However, there was substantial agreement between the ACR criteria and the EULAR/PRES criteria (κ = 0.686). The sensitivities of the ACR criteria and the EULAR/PRES classification criteria for WG among the spectrum of pediatric AAVs were 68.4% and 73.6%, respectively, and the specificities were 68.3% and 73.2%, respectively (Table 2).

Table 2. Discriminant validity of the ACR and EULAR/PRES classification criteria for WG*
 ACR criteriaEULAR/PRESS criteria
  • *

    Values are the percent. PPV = positive predictive value; NPV = negative predictive value (see Table 1 for other definitions).

Overall accuracy68.473.5

Figure 1 is a Venn diagram showing the number of patients in the cohort who were diagnosed as having WG by MD diagnosis (n = 76), by the ACR classification criteria (n = 65), or by the EULAR/PRES classification criteria (n = 67). Twenty-six patients were not considered to have WG by MD diagnosis, ACR criteria, or EULAR/PRES criteria.

Figure 1.

Venn diagram of assigned Wegener's granulomatosis (WG) diagnosis among the 117 patients in the ARChiVe (A Registry for Childhood Vasculitis: e-entry) cohort. The diagnosis of WG was established by the treating pediatric rheumatologist at each of the ARChiVe network sites (the MD diagnosis), by the American College of Rheumatology (ACR) classification criteria, and by the European League Against Rheumatism/Pediatric Rheumatology European Society (EULAR/PRES) classification criteria.

Among the 65 patients with WG according to the ACR criteria, the MD diagnosis differed in 13 (10 had MPA and 3 had unclassified vasculitis). Among the 52 patients who did not meet the ACR criteria for WG, 17 had an MD diagnosis of WG, 2 were classified as having WG by EULAR/PRES criteria, and 7 were classified as having WG by both the MD diagnosis and the EULAR/PRES criteria. The EULAR/PRES criteria that enabled WG classification in these 9 patients were sinus involvement and/or ANCA-positive serologic findings and/or significant proteinuria, and, less frequently, subglottic/tracheal/endobronchial stenosis and/or kidney biopsy showing necrotizing pauci-immune GN (Table 3).

Table 3. Data on the 16 patients from the ARChiVe study cohort who could fulfill a sufficient number of either EULAR/PRES classification criteria or ACR classification criteria (but not both) for diagnosis of WG*
 Nasal inflammationSinus inflammationAbnormal chest radiographAbnormal chest CT scanAbnormal urinalysis (hematuria)Abnormal urinalysis (significant proteinuria)Granulomatous inflammation on biopsyNecrotizing pauci-immune GN on biopsySubglottic, tracheal, or endobronchial stenosisAnti-PR3 ANCA or cANCA staining
  • *

    The presence of an individual component of the criteria in each patient is indicated by an X. Complete descriptions of the EULAR/PRES and ACR criteria are shown in Table 1. ARChiVe = A Registry for Childhood Vasculitis: e-entry (see Table 1 for other definitions).

  • Criterion shared by both EULAR/PRES and ACR.

  • EULAR/PRES criterion that is either a modification of an ACR criterion or a new criterion.

Patients fulfilling 3 of 6 EULAR/PRES criteria for WG, by age in years (n = 9)          
 3 XX  X   X
 4XX      XX
 6XX   X  XX
 7 X   X  X 
 10 X  XX X  
 12 X  XX X  
 14    XX X X
 14 XX      X
 14 XX      X
Patients fulfilling 2 of 4 ACR criteria for WG, by age in years (n = 7)          
 2  X XX    
 3XX    X   
 3XX    X   
 4  X XX    
 6    XXX   
 6XX  XX    
 15  X XX    

Among the 67 patients with WG according to the EULAR/PRES criteria, the MD diagnosis differed in 11 (7 had MPA, 3 had unclassified vasculitis, and 1 had ANCA-positive pauci-immune GN). Among the 50 patients who did not meet the EULAR/PRES criteria for WG, 17 had an MD diagnosis of WG, 4 had WG by ACR criteria, and 3 had WG by both MD diagnosis and ACR criteria. The 7 patients who met ACR criteria fulfilled only 2 of the 4 criteria that were also criteria of the EULAR/PRES classification system (Table 3).

WG cohort.

Sixty-five of the 117 patients met the ACR criteria for WG; 63.1% of these patients were female. The majority of patients were Caucasian (69.2%), 4 were of mixed ethnicity, 3 were of East Indian descent, 3 were African American, 2 were Hispanic, 1 was Asian, and 1 was of Middle Eastern origin; ethnicity data were not available for 6 patients. The median age at WG diagnosis was 14.2 years (range 4–17 years). The median interval from symptom onset to diagnosis was 2.7 months (range 0–49 months). Eight of these 65 WG patients (12.3%) had a symptom onset to diagnosis interval >12 months; patients with this longer interval to diagnosis had a higher frequency of ear, nose, and throat and cutaneous manifestations and less frequently had renal features (data not shown).

Frequencies of presenting features by organ system were as follows: constitutional/general (89.2%), pulmonary (80.0%), ear, nose, and throat (80.0%), and renal (75.4%). Additional categories and specific features are listed in Table 4. Significant renal involvement as manifested by serum creatinine elevated above the 95th percentile for age was found in 27 patients (41.5%); dialysis was necessary in 7 patients (10.8%), and end-stage renal disease was present in 1 patient. Of note, among those patients with pulmonary disease, severe involvement requiring continuous oxygen therapy or mechanical ventilation occurred in 19% and 11%, respectively. Of the 35 patients who had pulmonary function tests done, 27 had abnormal results; obstructive and restrictive abnormalities were found to be equally represented.

Table 4. Frequency of presenting clinical features in pediatric patients with WG in 3 different single-center cohorts and in our multicenter ARChiVe cohort*
Clinical featureNIH (n = 23)GOSH (n = 17)HSC (n = 25)§ARChiVe (n = 65)
  • *

    Values are the number (%) of affected patients. NR = frequency not reported (see Table 1 for other definitions).

  • National Institutes of Health (NIH) study, Bethesda, MD (Rottem et al, 1993 [16]). Absolute patient numbers were calculated from percentages provided in the article.

  • Great Ormond Street Hospital (GOSH) for Children, London, UK (Belostotsky et al, 2002 [6]). Some clinical features in this cohort correspond to the entire disease course since features at disease onset were not reported consistently.

  • §

    Hospital for Sick Children (HSC), Toronto, Ontario, Canada (Akikusa et al, 2007 [17]).

  • A Registry for Childhood Vasculitis: e-entry (ARChiVe), 2008 (present study). Patients met ≥2 of the ACR classification criteria for WG.

  • #

    From 62 children who underwent chest imaging.

  • **

    From 35 children who had pulmonary function tests done.

  • ††

    Nasal involvement features were reported separately, with epistaxis occurring in 40% and nasal ulcers in 24% of children at disease onset.

  • ‡‡

    Arthralgias and arthritis at disease onset were not reported separately.

Constitutional/generalNRNR24 (96.0)58 (89.2)
 Malaise, fatigueNRNRNR58 (89.2)
 Fever5 (22.0)NR18 (72.0)35 (53.8)
 Weight loss3 (13.0)5 (29.0)14 (56.0)28 (43.1)
Pulmonary5 (22.0)14 (82.4)20 (80.0)52 (80.0)
 Shortness of breathNRNRNR38 (58.5)
 Chronic coughNR9 (53.0)NR34 (52.3)
 Hemoptysis/alveolar hemorrhage2 (9.0)3 (18.0)11 (44.0)29 (44.6)
 Nodules3 (13.0)3 (17.6)11 (44.0)26 (41.9)#
 Abnormal pulmonary function test resultsNR3 (17.6)NR27 (42.5)**
 Fixed pulmonary infiltrates2 (9.0)1 (5.9)4 (16.0)14 (22.6)#
 Oxygen dependencyNRNRNR12 (18.5)
 Pleurisy1 (4.0)NR2 (8.0)15 (23.1)
RenalNR4 (23.5)22 (88.0)49 (75.4)
 Abnormal urinalysis resultsNR4 (23.5)22 (88.0)49 (75.4)
 Biopsy-proven GN2 (9.0)NR16 (64.0)34 (52.3)
 Elevated serum creatinineNRNR7 (28.0)27 (41.5)
Ear, nose, and throat20 (87.0)17 (100.0)21 (84.0)52 (80.0)
 Nasal involvement11 (48.0)7 (41.0)10 (40.0)††42 (64.6)
 Sinusitis14 (61.0)6 (35.0)11 (44.0)39 (60.0)
 Otitis/mastoiditis9 (39.0)NR6 (24.0)9 (13.8)
 Subglottic involvement1 (4.0)7 (41.0)1 (4.0)9 (13.8)
 Hearing loss6 (26.0)NR4 (16.0)7 (10.8)
 Oral ulcers1 (4.0)5 (29.0)7 (28.0)6 (9.2)
Eyes3 (13.0)9 (53.0)13 (52.0)24 (36.9)
 Nonspecific red eyeNRNRNR10 (15.4)
 Conjunctivitis0 (0.0)5 (29.0)11 (44.0)6 (9.2)
 Scleritis1 (4.0)3 (18.0)3 (12.0)3 (4.6)
CutaneousNR9 (53.0)8 (32.0)23 (35.4)
 Palpable purpura/petechiaeNRNR8 (32.0)15 (23.1)
GastrointestinalNRNR3 (12.0)27 (41.5)
 Nonspecific abdominal painNR5 (29.0)NR21 (32.3)
 Chronic nauseaNRNRNR11 (16.9)
MusculoskeletalNRNRNR37 (56.9)
 Arthralgia/myalgia7 (30.0)‡‡9 (53.0)16 (64.0)35 (53.8)
 ArthritisNRNR8 (32.0)13 (20.0)
Nervous systemNR2 (12.0)2 (8.0)16 (24.6)
 Severe headacheNRNRNR9 (13.8)
 DizzinessNRNRNR8 (12.3)
CardiovascularNRNRNR0 (0.0)
 Venous thrombosisNRNR3 (12.0)0 (0.0)

The numbers and frequencies of patients presenting with any of the criteria used in either the ACR or the EULAR/PRES classification were as follows: nasal bleed/ulcers/crusts, 64.6%; sinusitis, 60.0%; subglottic, tracheal, or endobronchial stenosis, 16.9%; oral ulcers, 9.2%; abnormal findings on chest imaging (multiple nodules, fixed infiltrates, and/or cavitation), 64.5% of 62 patients who underwent chest imaging; abnormal urinalysis results (>5 red blood cells [RBCs] per high-power field, RBC casts, or proteinuria), 70.8%; histologic evidence of necrotizing pauci-immune GN, 12.3%; and histologic evidence of granulomatous vasculitis at nonrenal sites, 24.6%. Serology tests for the presence of ANCAs by immunofluorescence yielded positive results in 58 children with WG (89.2%) (43 with cANCA [66.2%], 14 with pANCA [21.5%], and 1 with both cANCA and pANCA [1.5%]); 4 patients were negative for both cANCA and pANCA, and serology tests were not performed for 3 patients. Using enzyme-linked immunosorbent assay, anti–proteinase 3 (anti-PR3) positivity was reported in 44 patients (67.7%), antimyeloperoxidase (anti-MPO) positivity was reported in 8 patients (12.3%), and 1 patient was found to be both anti-PR3 and anti-MPO positive. Results of ANCA testing and other autoantibody tests are summarized in Table 5.

Table 5. Autoantibody testing results among children with and those without WG according to the ACR criteria*
Serologic testWith WG (n = 65)Without WG (n = 52)
  • *

    Values are the number (%) of patients. ELISA = enzyme-linked immunosorbent assay; anti-MPO = antimyeloperoxidase; pANCA = perinuclear ANCA (see Table 1 for other definitions).

  • Transiently positive for both patients.

Immunofluorescence for cANCA, positive43 (66.2)12 (23.1)
 Anti-PR3–positive on ELISA40 (93.0)11 (91.7)
 Anti-MPO–positive on ELISA0 (0.0)0 (0.0)
 Negative on ELISA2 (4.7)1 (8.3)
 ELISA not done0 (0.0)0 (0.0)
 ELISA result not available1 (2.3)0 (0.0)
Immunofluorescence for pANCA, positive14 (21.5)20 (38.5)
 Anti-PR3–positive on ELISA3 (21.4)1 (5.0)
 Anti-MPO–positive on ELISA8 (57.1)18 (90.0)
 Negative on ELISA2 (14.3)0 (0.0)
 ELISA not done1 (7.1)1 (5.0)
Immunofluorescence for pANCA and cANCA, positive1 (1.5)1 (2.0)
 Anti-MPO–positive on ELISA0 (0.0)1 (100.0)
 ELISA not done1 (100.0)0 (0.0)
Immunofluorescence not done3 (4.6)1 (2.0)
 Anti-PR3– and anti-MPO–positive on ELISA1 (33.3)0 (0.0)
 Anti-PR3–positive on ELISA1 (33.3)0 (0.0)
 ELISA not done0 (0.0)1 (100.0)
 ELISA result not available1 (33.3)0 (0.0)
Negative for ANCA (cANCA and pANCA)4 (6.2)17 (32.7)
 Anti-PR3–positive on ELISA0 (0.0)3 (17.6)
 Anti-MPO–positive on ELISA0 (0.0)1 (5.9)
 Negative on ELISA4 (100.0)12 (70.6)
 ELISA not done0 (0.0)1 (5.9)
Antinuclear antibodies, tested50 (76.9)44 (84.6)
 Positive17 (34.0)16 (36.4)
 Negative33 (66.0)28 (63.6)
Anticardiolipin antibodies and/or lupus anticoagulant, tested34 (52.3)19 (36.5)
 Positive6 (17.6)5 (26.3)
 Negative28 (82.4)14 (73.7)
Anti–glomerular basement membrane antibodies, tested9 (13.8)3 (5.8)
 Positive2 (22.2)0 (0.0)
 Negative7 (77.8)3 (100.0)


Fifty-four patients had ≥1 biopsy performed (36 renal biopsies and 38 nonrenal biopsies). Nonrenal histopathology showing granulomatous vasculitis included the following: lower airways/lungs (6 of 13 biopsy samples), paranasal sinuses (4 of 8 biopsy samples), upper airways (3 of 6 biopsy samples), skin (2 of 6 biopsy samples), and nasal septum (1 of 3 biopsy samples). No granulomatous vasculitis was found in 2 reported biopsy samples of the gastrointestinal tract. In all but 2 kidney biopsy samples, the histopathology showed GN, with 22.2% being described as necrotizing pauci-immune GN.


Fifty-four patients (83.1%) were treated with a combination of corticosteroids and cyclophosphamide; 27 of these received oral cyclophosphamide, 21 received IV cyclophosphamide, and 6 initially received IV cyclophosphamide and were switched to oral cyclophosphamide early in the disease course. Forty-five patients (69.2%) received high-dose IV pulse corticosteroids. Among 60 patients (92.3%) receiving oral corticosteroids, 2 were receiving a low-dose regimen, 31 were receiving a medium dose, and 27 were receiving a high dose. Four children were treated with corticosteroids only (2 received medium-dose corticosteroids and 2 received high-dose corticosteroids). Six patients (9.2%) received a combination of corticosteroids and methotrexate, and 1 patient (1.5%) was treated with methotrexate only. Thirty-four patients (52.3%) were started on trimethoprim/sulfamethoxazole (TMP/SMX), generally as Pneumocystis jiroveci pneumonia prophylaxis. Other therapeutic modalities used included plasmapheresis (9 patients), extracorporeal membrane oxygenation (2 patients), IV immunoglobulin (IVIG) (2 patients), and rituximab (1 patient).


This report of 65 patients with pediatric-onset WG (ACR defined) from 30 US and Canadian centers describes the largest cohort of such children to date. The 3 largest series previously reported from the US (n = 23) (16), the UK (n = 17) (6), and Canada (n = 25) (17) were single-center series that collected patients over 24, 17, and 21 years, respectively. Unlike the others, this current report does not describe any followup data. However, through the concurrent establishment of a network of interested investigators supported by a Web-based data entry interface system (ARChiVe), we have identified a relatively contemporary cohort of patients (diagnosed since 2004) for future prospective study. Because of the rarity of WG in children, it is not feasible for any single center to establish a similarly sized cohort (5). Furthermore, the nature of this multicenter network reduces the likelihood that patients are selected for greater disease severity, a potential bias that might be attributed to cohorts from single large referral centers.

In classifying WG, the limited convergence of MD diagnosis with either ACR criteria or EULAR/PRES criteria is not surprising, since the premise for diagnosis of patients as having any disease differs from that for classification. Because the EULAR/PRES criteria were built upon the existing ACR criteria, there was a reasonable convergence between them, and as such, the sensitivity and specificity were likely to be similar. Among the 76 patients referred to ARChiVe with the physician diagnosis of WG, 52 could be classified as having WG according to ACR criteria; in an additional 13 children with ACR-defined WG, the MD diagnosis differed. Use of the proposed pediatric EULAR/PRES criteria minimally improved diagnostic sensitivity and specificity over those of the ACR criteria; although the EULAR/PRES classification scheme identified 9 patients with WG in addition to those identified by the ACR criteria, 7 of the patients identified as having WG according to the ACR criteria could not meet >2 criteria required for EULAR/PRES classification (Table 3). Interestingly, by applying the modification proposed by the WG etanercept trial investigators (18) of adding a positive enzyme immunoassay finding for anti-PR3 as a fifth criterion, sensitivity of the ACR criteria increased substantially (to 86.8%) without a significant change in specificity.

The low sensitivity and specificity of both sets of criteria likely reflect the denominator. In our analyses, sensitivity and specificity for classification of WG were tested among a group of patients with vasculitides related by size of the vessel predominantly involved (small to medium) and the association with ANCAs. Patients in any diagnostic subset of AAV in this cohort were more likely to have WG than a vasculitis diagnosis other than AAV. Thus, we would anticipate that the sensitivity and specificity would improve significantly if the criteria were tested among a population of pediatric patients that included the wider spectrum of vasculitides of differing predominant vessel sizes, from Henoch-Schönlein purpura (small vessel) through polyarteritis nodosa (PAN) (medium vessel) to Takayasu arteritis (large vessel).

Among the patient population that we studied, the most difficult differential diagnosis was the distinguishing of WG from MPA. Neither the ACR nor the EULAR/PRES have defined categorical classification criteria for MPA. Using the CHCC characterization of MPA (9), 36 of 117 patients in our cohort could be defined as having MPA, and among these patients, 24 could concurrently be diagnosed as having WG according to ACR criteria. Since WG and MPA share many features, there are clearly difficulties in defining mutually exclusive classification criteria for these 2 disorders. Some authorities believe that these 2 entities should be considered together under the rubric of AAV (10). The algorithm by Watts et al (12) was developed to categorize adult patients among those with AAV or PAN into a single nonoverlapping classification; those investigators criticize the poor individual performance of ACR criteria, CHCC definitions, and Lanham criteria (19) with respect to defining mutually exclusive classification categories. This algorithm applies the different criteria in a stepwise approach, first defining WG patients where criteria are most specific, and subsequently and sequentially applying elements of different criteria to the remaining patients to classify each into a single category. The EULAR/PRES criteria, a “pediatric” modification of the ACR criteria, did not address this criticism.

Basic demographics, clinical features, and treatment modalities for ACR-defined WG patients in the ARChiVe cohort were comparable with those reported elsewhere in that the patients were predominantly female and Caucasian and the age at onset was mostly in the adolescent years. Our patients' median interval from symptom onset to diagnosis (2.7 months) was similar to those reported by Akikusa et al and Rottem et al (2 months and 8 months, respectively) (16, 17) in spite of the great diversity of geographic and economic backgrounds of our population and differences in health care policy between Canada and the US. The upper end of the range (49 months) remains disconcerting; the 8 ARChiVe patients (12.3%) with an interval to diagnosis of >12 months presented predominantly with ear, nose, and throat and cutaneous features and had fewer renal manifestations than those patients who were diagnosed within 12 months of symptom onset. The implication of this disparity is that diagnosis is delayed in the absence of initial critical organ involvement.

The frequencies of specific presenting clinical features of our cohort and those of the 3 previous largest cohorts are compared in Table 4. We corroborated the high incidence of renal involvement at disease onset described by Akikusa et al and by investigators in some smaller series (17, 20, 21) relative to the adult experience (22, 23), whereas there was a somewhat lower incidence in the other 2 large series. Pulmonary involvement was as common as renal disease among ARChiVe patients, with severe pulmonary involvement (oxygen dependency and/or requirement for mechanical ventilation) occurring in 15 patients (23.1%). Findings on imaging studies (pulmonary nodules and infiltrates) in our population were comparable with those reported in other pediatric WG series, with a lower proportion of abnormal imaging findings in patients without pulmonary symptoms in our cohort (data not shown) compared with that reported by Rottem et al (17% versus 41%) (16).

The frequency of ear, nose, and throat features among ARChiVe patients was similar to that reported in other pediatric WG series. Subglottic stenosis is noted in some reports to occur frequently in pediatric-onset WG compared with WG in adults, such that this feature was included in the EULAR/PRES criteria (derived by expert consensus) for classifying WG (11). In our cohort, after nasal and sinus involvement, subglottic stenosis was only slightly more frequent than otitis/mastoiditis or hearing loss. Nonetheless, it still may be valuable as a criterion because it may be more specific for WG, having a narrower differential diagnosis than other upper respiratory symptoms. Inclusion of subglottic stenosis as a criterion may enable classification of patients through EULAR/PRES who would not otherwise be described as having WG; one such patient was diagnosed in this cohort.

We found a lower frequency of ocular manifestations (conjunctivitis and scleritis in particular) than previously reported. In contrast, the frequency of gastrointestinal and neurologic symptoms was greater in our patients compared with their frequency in reported series.

The high frequency of thrombotic events described in 2 previous reports (17, 24) was not found in our series. Additionally, half of our patients were tested for antiphospholipid antibodies and lupus anticoagulant, with only 18% of those tested having a positive result at the time of diagnosis (the reasons for antibody testing were not documented in the registry). This proportion is similar to that reported in adult patients (25). Thrombotic events may potentially occur in our patient cohort later in the disease course, although in the existing reports, thrombotic events occurred within a few weeks of diagnosis, and such patients in our cohort would arguably have been identified within our data capture time frame.

In treating pediatric WG, the standard practice of initial treatment for 83% of ARChiVe patients was cyclophosphamide and corticosteroids, although there was diversity and several permutations of regimens and routes of administration. Early use of cyclophosphamide was similar to that reported by Rottem et al and Belostotsky et al (82.6% and 94%, respectively) (6, 16). Methotrexate was not frequently used in the early 1980s; however, in the ARChiVe cohort it was used in 10.7% of patients, a rate similar to that described in the Toronto cohort. Use of TMP/SMX in ARChiVe patients was higher than that reported by Belostotsky et al (6) (51% versus 29%); however, except for 1 patient, it was prescribed as P jiroveci pneumonia prophylaxis. Other therapies such as plasmapheresis, extracorporeal membrane oxygenation, IVIG, and rituximab were rarely used. No patients began therapy using azathioprine, mycophenolate mofetil, or biologic medications other than rituximab at the time of diagnosis. We recognize that there are recent reports of rituximab use as initial therapy in newly diagnosed adult and pediatric WG patients (26, 27), but at the time of this report those pediatric patients had not been entered into the registry.

Some considerations must be applied to the interpretation of our findings, particularly with respect to previous WG series. First, our data were collected cross-sectionally and predominantly in a retrospective way. However, data collection was done using a standardized Web-based data entry system that was regularly reviewed centrally for data consistency and quality to minimize incomplete information. Furthermore, the retrospective time frame was the shortest (5 years) and most contemporary. Patient recruitment to the ongoing ARChiVe registry cohort will be predominantly prospective, thereby improving the quality and completeness of data capture. Also, having data collected in many institutions will provide us with a broader overview of the different clinical approaches to pediatric vasculitis in practice across the US and Canada. Second, because ARChiVe data collection currently consists of a one-time data entry point at the time of diagnosis, we are not able to provide information on patient outcomes. Future longitudinal data collection from this patient cohort is planned.

In this cohort, the largest collection of pediatric patients with WG to date, some patients were found to have a considerable delay from symptom onset to time of diagnosis, particularly when there was initially noncritical organ disease. Lack of appreciation of the female preponderance in childhood disease and differences in presentation from adult patients may contribute to the diagnostic delay. The diversity of initial treatments used for patients in this cohort suggests that there may be no standard approach to therapy. This difference may predominate among patients with noncritical organ disease. Further analysis of treatment regimen stratified by disease severity at onset and subsequent outcome will be necessary to determine optimal outcomes while minimizing treatment toxicities. Estimations of disease severity by adapting severity classification systems used in adults will be integrated into future outcome studies of this cohort. We did not find that the newly proposed pediatric EULAR/PRES vasculitis classification criteria presented significant improvement over the ACR criteria for classifying WG; however, a second iteration of the criteria is being refined and will be testable in our growing cohort.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Cabral had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Cabral, Uribe, Benseler, O'Neil, Hashkes, Zeft, Lovell, Kingsbury, Stevens, Chira, Abramson, Arkachaisri, Eberhard, Hersh, Levy, Li, Morgan DeWitt, Reiff, Schikler, Singer.

Acquisition of data. Cabral, Uribe, Benseler, O'Neil, Hashkes, Higgins, Lovell, Kingsbury, Stevens, McCurdy, Chira, Abramson, Arkachaisri, Campillo, Eberhard, Hersh, Huber, Kim, Klein-Gitelman, Li, Mason, Morgan DeWitt, Muscal, Nassi, Schikler, Singer, Wahezi, Woodward.

Analysis and interpretation of data. Cabral, Uribe, Huber, Muscal, Schikler.

Study administration and coordination. Cabral.


We are indebted to all participating patients and their families, without whom this study would not be possible. The origins of this project were in the CARRA; although the ARChiVe network now extends beyond this, we gratefully acknowledge that it would not be sustainable without the endorsement and ongoing support of the CARRA and its membership.



Coordinating center.

British Columbia Children's Hospital, Vancouver, British Columbia, Canada: David A. Cabral (Study Principal Investigator), América G. Uribe (Study Coordinator), Victor Espinosa (Information Technology Manager, Statistician), Jaime Guzman, Kristin Houghton, Peter Malleson, Kimberly Morishita, Ross Petty, Lori Tucker, Stuart Turvey (Site Investigators). Participating centers. Case Medical Center, and Rainbow Babies and Children's Hospital University Hospitals, Cleveland, OH: Nora G. Singer (Site Principal Investigator), Elizabeth B. Brooks (Site Investigator), Mary Lesko (Site Coordinator). Children's Hospital of Boston, Boston, MA: Susan Kim (Site Principal Investigator), Fatma Dedeoglu, Robert Fuhlbrigge, Melissa Hazen, Mary Beth Son, Robert Sundel (Site Investigators). Childrens Hospital Los Angeles, Los Angeles, CA: Andreas Reiff (Site Principal Investigator), Diane Brown, Bracha Shaham (Site Investigators), Ana Cabrera (Site Coordinator). Children's Hospital at Montefiore, Bronx, NY: Norman T. Ilowite (Site Principal Investigator), Dawn Wahezi (Site Investigator). Children's Hospital of Pittsburgh, Pittsburgh, PA: Thaschawee Arkachaisri (Site Principal Investigator), Raphael Hirsh, Daniel Kietz, Paul Rosen, Margalit Rosenkrank, Kathryn Torok (Site Investigators). Children's Memorial Hospital, Chicago, IL: Marisa Klein-Gitelman (Site Principal Investigator), Lauren Pachman (Site Investigator), Aisha Ali (Site Coordinator). Cincinnati Children's Hospital Medical Center, Cincinnati, OH: Daniel J. Lovell (Site Principal Investigator), Hermine Brunner, Thomas Griffin, Alexi Grom (Site Investigators), Anne Johnson (Site Coordinator). Cleveland Clinic Foundation, Cleveland, OH: Philip J. Hashkes (Site Principal Investigator), Steven Spalding (Site Investigator), Deborah Bork (Site Coordinator). Duke Children's Hospital and Health Center, Duke University Medical Center, Durham, NC: Esi Morgan DeWitt (Site Principal Investigator), Stacy Ardoin, Egla Rabinovich, Laura Schanberg (Site Investigators), Rhonda Wilder (Site Coordinator). Hospital for Sick Children, Toronto, Ontario, Canada: Susanne Benseler (Site Principal Investigator). IWK Health Centre, and Dalhousie University, Halifax, Nova Scotia, Canada: Adam M. Huber (Site Principal Investigator), Bianca A. Lang, Suzanne Ramsey, Elizabeth Stringer (Site Investigators), Aleasha Warner (Site Coordinator). Joseph M. Sanzari Children's Hospital, Hackensack University Medical Center, Hackensack, NJ: Suzanne C. Li (Site Principal Investigator), Kathleen Haines, Yukiko Kimura, Jennifer Weiss (Site Investigators). Legacy Emanuel Children's Hospital, Portland, OR: Daniel J. Kingsbury (Site Principal Investigator). Lucile Packard Children's Hospital, Stanford University School of Medicine, Stanford, CA: Peter Chira (Site Principal Investigator), Imelda Balboni, Reuven Bromberg, Michal Cidon, Jennifer Frankovich, Dana Gerstbacher, Joyce J. Hsu, Tzielan Lee, Jane L. Park, Christy Sandborg, Steven Song (Site Investigators). Mayo Eugenio Litta Children's Hospital, Mayo Clinic, Rochester, MN: Thomas Mason (Site Principal Investigator), Ann Reed (Site Investigator). The Montreal Children's Hospital, McGill University Health Centre, Montreal, Quebec, Canada: Sarah Campillo (Site Principal Investigator), Gaëlle Chédeville, Ciarán Duffy, Karen Duffy, Rosie Scuccimarri (Site Investigators), Michele Gibbon (Site Coordinator). Nationwide Children's Hospital, Columbus, OH: Gloria Higgins (Site Principal Investigator). New York–Presbyterian Morgan Stanley Children's Hospital, New York, NY: Deborah M. Levy (Site Principal Investigator), Andrew Eichenfield, Lisa Imundo (Site Investigators), Candido Batres (Site Coordinator). Riley Children's Hospital, Indianapolis, IN: Suzanne Bowyer (Site Principal Investigator), Susan Ballinger, Thomas Klausmeier, Amy Woodward (Site Investigators), Andrea Hudgins (Site Coordinator). Schneider Children's Hospital, New York, NY: Anne Eberhard (Site Principal Investigator). Seattle Children's Hospital, Seattle, WA: Anne Stevens (Site Principal Investigator), Helen Emery, Kristin Hayward, Christi Inman, Sarah Ringold, Elizabeth Shaw, Troy Torgerson, Jennifer Turner, Carol Wallace, Jennifer Wargula (Site Investigators), Sarah Halford (Site Coordinator). Texas Children's Hospital, Baylor College of Medicine, Houston, TX: Eyal Muscal (Site Principal Investigator), Barry L. Myones (Site Investigator). Texas Scottish Rite Hospital, University of Texas Southwestern, Dallas: Marilynn Punaro (Site Principal Investigator), Lorien Nassi, Virginia Pascual (Site Investigators). University of California, Los Angeles: Deborah McCurdy (Site Principal Investigator). University of California, San Francisco: Aimee O. Hersh (Site Principal Investigator), Emily von Scheven (Site Investigator). University of Louisville Health Sciences Center, Louisville, KY: Kenneth Schikler (Site Principal Investigator), Adrienne Michels (Site Coordinator). University of Oklahoma Health Sciences Center, Oklahoma City: Kathleen M. O'Neil (Site Principal Investigator), Michael Henrickson, James Jarvis (Site Investigators). University of Utah Primary Children's Medical Center, Salt Lake City: Andrew S. Zeft (Site Principal Investigator), John Bonsack, Sampath Prahalad (Site Investigators). University of Vermont, Burlington: Leslie Abramson (Site Principal Investigator).