Limited versus severe Wegener's granulomatosis: Baseline data on patients in the Wegener's granulomatosis etanercept trial

Authors


Abstract

Objective

To report baseline data on 180 patients with Wegener's granulomatosis (WG) enrolled in the WG Etanercept Trial (WGET), and to examine demographic and clinical differences between patients with limited disease and those with severe disease.

Methods

Definitions of limited and severe disease were agreed upon by consensus of the investigators at a pretrial meeting and were incorporated into the protocol as a stratification criterion. These data were applied prospectively to the WGET patient cohort, based on clinical features and the intention to treat patients according to disease activity. Data related to disease onset, date of diagnosis, clinical features, antineutrophil cytoplasmic antibody assays, tissue biopsy findings, and other medical history were collected on a baseline medical history form. Physician-investigators from each center participated in the development of this form, and all were certified in its use prior to the start of the trial. Selected data on patients who were screened for the trial but were not enrolled were also collected.

Results

Several significant differences between the limited and severe disease subsets were observed. Patients with limited disease were nearly a decade younger at disease onset compared with patients with severe disease. Thirty-three percent of patients with severe disease were women, compared with 58% of those with limited disease. Despite their younger age at symptom onset, patients with limited disease tended to have longer disease duration, a greater likelihood of experiencing exacerbation of previous disease following a period of remission, and a higher prevalence of destructive upper respiratory tract disorders at the time of enrollment (e.g., saddle-nose deformity). Patients with limited WG were less likely than those with severe disease to have antibodies to either proteinase 3 or myeloperoxidase. Patients with severe disease had a higher likelihood of previous thyroid disease, particularly either Graves' disease or Hashimoto thyroiditis, suggesting the possibility of different pathogenetic factors within these disease subsets. Other observed differences between these subsets, such as the greater frequency of alveolar hemorrhage in the severe disease group, were related to the a priori definitions of limited and severe disease.

Conclusion

There are significant differences between patients with limited WG and those with severe WG with regard to sex, age, the likelihood of recurrent disease, the risk of damage in certain organ systems, and, possibly, etiologic factors. These differences (and perhaps other differences that are currently unrecognized) in patient subsets may have implications for mechanisms of pathogenesis, prognosis, response to treatment, and the design of future clinical investigations.

Wegener's granulomatosis (WG) is a multifocal inflammatory illness that most often affects the upper and lower respiratory tracts and the kidneys (1, 2). In its generalized form, untreated WG is associated with a 1-year mortality rate of >80% (3). In the early 1970s, the introduction of treatment regimens based on the combination of cyclophosphamide and glucocorticoids converted this once typically fatal illness into a disease in which most patients achieve remissions, some of which last for several years or longer (2, 4, 5). A significant proportion of patients (perhaps 20–40%) appear to be “cured” by conventional therapy (2, 6). However, disease flares following the tapering or discontinuation of treatment remain the rule, and treatment-induced side effects are a major source of morbidity and mortality (2, 6, 7).

In the mid-1960s, limited forms of WG were recognized (8–10). Specific definitions of limited WG have varied in the medical literature. For example, limited disease has also been referred to as “initial phase” (6, 11), “nonrenal” (12), “indolent” (13), “early-systemic” (14), “localized” (6), and “locoregional” (11, 15). Similarly, severe disease has also been termed “generalized” (6, 7), “classic” (12), or “classic generalized” (11). The current consensus among vasculitis experts is that limited disease, in contrast to severe disease, includes manifestations of WG that pose no immediate threat to either the patient's life or the function of a vital organ (2, 6, 14, 16, 17).

In terms of the current standard of care, the distinction between limited and severe disease subsets is important, because it has practical implications for treatment. Severe disease requires prompt institution of an aggressive therapeutic regimen that includes cyclophosphamide and glucocorticoids (18). In contrast, limited disease usually responds well to a less toxic alternative regimen consisting of methotrexate (MTX) and glucocorticoids (2, 19–21). Despite the important treatment implications of classifying patients into limited or severe disease subsets, few direct comparisons of the characteristics of these subsets have ever been performed. Furthermore, there have been no subset-specific analyses of patients with WG whose classifications were determined (and therapies allocated) by the same group of physicians using a uniform protocol.

Two large cohorts of patients with WG, both of which were recruited over periods of ∼25 years, have been previously described (2, 6). Hoffman et al (2) reported the NIH experience with 158 patients, 133 of whom received a standard regimen of daily cyclophosphamide and glucocorticoids (5). These patients were evaluated and followed up between 1968 and 1992. Reinhold-Keller et al (6) described 155 patients who were evaluated at one center (in Lübeck, Germany) between 1966 and 1997, 142 of whom were treated with daily cyclophosphamide and glucocorticoids. Because patients in both cohorts were enrolled before use of MTX became an accepted practice in the treatment of WG, differences between limited and severe disease subsets were not analyzed in detail in either study. Some of the patients in both studies, however, did receive MTX at some point in the course of treatment (18 patients in the NIH study and 45 patients in the German study).

The Wegener's Granulomatosis Etanercept Trial (WGET) is the first multicenter, randomized trial for patients with this disease in the US (members of the WGET Research Group are listed in Appendix A). Moreover, the WGET cohort is the largest reported group of patients studied in a prospective research protocol. Enrollment for this trial was completed in September 2002. We report the baseline data on all 180 patients with WG enrolled in the trial. When appropriate and possible, we confirmed our findings in the enrolled population (particularly those findings related to disease demographics) by reference to the group of patients who were screened for the trial but were not enrolled (n = 778). We compared the disease characteristics at the time of enrollment for the limited and severe disease subsets and reported the baseline clinical features for the entire cohort. We also compared features of the WGET patients with those of patients in the 2 other large cohorts (2, 6).

PATIENTS AND METHODS

Trial eligibility.

All patients with the diagnosis of WG and a Birmingham Vasculitis Activity Score for WG (BVAS/WG) (16, 22) of ≥3 were eligible for enrollment in the WGET, regardless of whether they had limited or severe disease. All patients enrolled fulfilled at least 2 of the 5 modified American College of Rheumatology (ACR) criteria for the classification of WG (in the modified ACR criteria, a positive serum enzyme immunoassay [EIA] for antibodies to proteinase 3 was added to the original 4 criteria) (16, 23). Issues related to the design and development of the WGET have been reported in detail previously (16).

Disease classification.

For the purpose of assigning conventional therapies and for stratification of the randomization, patients were classified as having either severe or limited disease (Table 1). These definitions were applied to the WGET cohort prospectively, based on each patient's clinical features and the intention to treat according to disease severity.

Table 1. Definitions of severe and limited Wegener's granulomatosis*
  • *

    From the Wegener's Granulomatosis Etanercept Trial Manual of Operations. See ref. 23 for a description of the American College of Rheumatology criteria.

Limited
 Fulfillment of the modified American College of Rheumatology criteria for the classification of Wegener's granulomatosis in the absence of disease features that pose immediate threats to either a critical individual organ or to the patient's life. Specifically, this means that:
 1. The patient has no red blood cell casts in the urine.
 2. If hematuria is present (but no red blood cell casts), the serum creatinine is ≤1.4 mg/dl, and there must be no evidence of a rise in creatinine >25% above the patient's baseline level.
 3. Pulmonary involvement must be circumscribed, such that the room air PO2 is >70 mm Hg or the room air O2 saturation by pulse oximetry is >92%. Pulmonary hemorrhage may be treated as limited disease provided there is no evidence of progression of the process. In the absence of data on progression, pulmonary hemorrhage may be treated as severe disease at the discretion of the physician.
 4. No disease may exist within any other critical organ (e.g., the gastrointestinal tract, eyes, central nervous system) that, without the immediate institution of maximal therapy (i.e., pulse methylprednisolone and daily oral cyclophosphamide), threatens the function of that organ and/or the patient's life.
Severe
 Any patient with Wegener's granulomatosis whose disease is not classifiable as limited has severe disease, by definition.

Screening logs.

Screening logs were maintained at all clinics, documenting the age, sex, race, disease severity (limited versus severe), enrollment status, and reason(s) for not enrolling. Any visit by a patient with the clinical diagnosis of WG (or possible WG) was considered a screening visit. For patients who underwent screening for the trial on more than 1 occasion, only data from the first screening visit were analyzed (i.e., screened patients were counted only once, even if they were evaluated for entry multiple times).

Baseline medical history form.

At enrollment, disease characteristics were captured on a baseline medical history form that addressed questions related to the date of symptom onset, the date of diagnosis, the type of organ involvement by WG, the results of biopsies, testing for antineutrophil cytoplasmic antibodies (ANCA), previous treatment, and other medical history. Disease duration was defined as the interval between the onset of unequivocal signs and symptoms of WG and the date of enrollment in the WGET. All principal physician-investigators at each center participated in the development of the baseline medical history form, and all physician-investigators were certified in its use prior to the start of the trial.

Disease assessment instruments.

Disease activity at baseline was measured by the BVAS/WG, an instrument adapted specifically for WG and validated by the investigators prior to the start of enrollment (22). BVAS/WG scores were weighted as described in the validation study, with 1 point for a minor item and 3 points for a major item. Disease activity was also assessed by both the physician-investigators and the patients, using 10-cm visual analog scales. Damage caused by WG was assessed using the Vasculitis Damage Index (VDI) (24).

Statistical analysis.

Data were maintained in an SAS (Cary, NC) database. We compared data between the limited and severe disease subsets using chi-square tests (or Fisher's exact tests, as appropriate) for categorical data, and the Wilcoxon's rank sum test for continuous data. We calculated either means or medians as appropriate to the clinical situation, and rounded percentages up to the nearest whole number. We calculated Spearman's rank correlation coefficients (25) to evaluate the relationships between the 3 different scores used in the assessment of disease activity: the physician's global assessment, the patient's global assessment, and the BVAS/WG. The physician's global assessment and patient's global assessment scores were compared with paired t-tests.

RESULTS

Enrollment.

Between June 9, 2000 and September 30, 2002, a total of 181 patients with the diagnosis of WG were enrolled in the trial. Several weeks after the enrollment of one patient, that patient's pulmonary infiltrates (which had been characterized by necrotizing granulomatous inflammation on lung biopsy) were determined to be secondary to the intramuscular injection of anabolic steroids rather than to WG. For the purpose of this study, data on that patient were excluded from the analysis.

Demographics.

The baseline characteristics of all patients enrolled are shown in Table 2, classified according to disease severity (limited versus severe). One hundred twenty-eight patients (71%) had severe disease at enrollment, and 52 (29%) had limited disease. One hundred eight patients were male (60%), and 72 (40%) were female. Ninety-two percent of the patients enrolled in the WGET were white, 4% were Hispanic, 2% were African American (non-Hispanic), and 2% were of other ethnicities.

Table 2. Baseline disease characteristics by disease severity*
CharacteristicLimited (n = 52)Severe (n = 128)All patients (n = 180)P
  • *

    IQR = interquartile range; BVAS/WG = Birmingham Vasculitis Activity Score for Wegener's granulomatosis; VDI = Vasculitis Damage Index.

  • Determined by the chi-square test or Fisher's exact test for categorical data and Wilcoxon's rank sum test for continuous data.

Age at symptom onset, mean ± SD years41 ± 1650 ± 1647 ± 160.005
Sex   0.002
 Male, %42.367.260.0 
 Female, %57.732.840.0 
Ethnicity   0.251
 White, non-Hispanic, %88.493.792.2 
 African American, non-Hispanic, %1.91.61.7 
 Hispanic, %3.93.93.9 
 Other, %5.80.82.2 
Newly diagnosed at enrollment, %32.749.244.40.043
Time since onset of symptoms, median (IQR) months25.0 (7.7, 63.9)14.2 (3.9, 46.7)16.7 (5.0, 50.4)0.057
 Total patient years, no.199.1368.4567.5 
Time since diagnosis, median (IQR) months13.3 (1.1, 44.6)2.7 (0.7, 32.4)4.7 (0.8, 35.8)0.097
 Total patient years, no.151.0230.3381.3 
BVAS/WG, mean ± SD4.5 ± 2.17.9 ± 3.46.9 ± 3.4<0.001
Physician's global assessment score, mean ± SD4.2 ± 2.06.2 ± 2.25.6 ± 2.3<0.001
Patient's global assessment score, mean ± SD5.7 ± 2.76.7 ± 2.76.4 ± 2.70.011
VDI score, mean ± SD1.5 ± 1.71.3 ± 1.71.3 ± 1.70.214

Within the limited and severe disease subsets, skewing between the sexes was observed. Among patients with severe disease, 67% were male and 33% were female. In contrast, 42% of the patients with limited disease were male, and 58% were female (P = 0.002). This skewing of the sex distribution across categories of disease severity was also observed among patients who were screened for the WGET but were not enrolled: 60.6% of the patients in this category who had limited disease were female, and 39.4% were male (P = 0.029). Although more men than women were enrolled in the trial, a slightly higher number of women were screened for enrollment (420 women versus 358 men). Among the most common reasons for not being enrolled in the trial were low BVAS/WG scores or remission (74.5% of those screened), patient choice (4.5%), travel distance to trial centers (3.0%), and recent malignancy (1.1%).

Compared with patients with severe disease, a higher percentage of patients with limited WG failed to meet the entry criteria for the WGET, usually because they were less likely to have the minimum BVAS/WG score of 3 required for enrollment. Among the patients enrolled, those with limited disease had lower baseline BVAS/WG scores (see below). Among the patients screened, only 15% of those with previous or current features of limited disease were enrolled. In contrast, 23% of patients with previous or current severe disease were enrolled (P < 0.001).

Age at onset and disease duration.

Eighty patients (44%) had newly diagnosed WG when they enrolled. The other 100 patients (56%) were experiencing flares of previously diagnosed disease. Compared with patients with limited disease, those with severe disease were more often newly diagnosed (49% versus 33%; P = 0.043). The mean ± SD age at the time of first symptom onset for the entire cohort of WGET patients was 47 ± 16 years (50 ± 16 years for those with severe disease and 41 ± 16 years for those with limited disease; P = 0.005). Despite their younger age, patients with limited disease had a longer mean time since diagnosis compared with those with severe disease (median 13.3 months versus 2.7 months; P = 0.097).

Baseline disease assessment.

The mean ± SD BVAS/WG score at trial entry was 6.9 ± 3.4. Patients with severe disease had higher BVAS/WG scores at enrollment than did those with limited disease (mean ± SD 7.9 ± 3.4 versus 4.5 ± 2.1; P < 0.001). This difference was reflected in the scores for physician's global assessment and patient's global assessment of disease activity, both of which correlated highly with the BVAS/WG (rs = 0.59 for physician's global assessment [P < 0.001] and rs = 0.28 for patient's global assessment [P < 0.001]). On a 10-cm horizontal scale, patients tended to rate their disease activity higher than did physicians (P < 0.001). The mean ± SD baseline damage scores, as indicated by the VDI (24) measured at enrollment, were equivalent (1.5 ± 1.7 in the limited disease subgroup and 1.3 ± 1.7 in the severe disease subgroup; P = 0.214).

Clinical organ involvement at baseline.

The patients' organ involvement at the time of enrollment is shown in Table 3. There were significant differences between the 2 disease subsets in the frequency of ear, nose, and throat manifestations (89% in the limited disease subset, compared with 72% in the severe disease subset; P = 0.017). Patients with limited disease were more likely than those with severe disease to have sinus involvement (60% versus 41%; P = 0.027), nasal septal perforations (17% versus 6%; P = 0.022), and nasal collapse (saddle-nose deformity [19% versus 6%; P = 0.009]). The frequency of subglottic stenosis was not significantly different between the 2 groups (19.2% in the group with limited disease and 9.4% in the group with severe disease; P = 0.067). In contrast, patients with severe disease were more likely to have fever at enrollment (21% versus 8%; P = 0.031).

Table 3. Organ involvement at enrollment, as recorded on the BVAS/WG form, by disease severity*
 Limited (n = 52)Severe (n = 128)All patients (n = 180)P
  • *

    Values are the percentage of patients. BVAS/WG = Birmingham Vasculitis Activity Score for Wegener's granulomatosis.

  • Determined by chi-square test or Fisher's exact test for categorical data.

  • See Table 5.

General76.970.372.20.370
 Arthralgia/arthritis69.264.866.1 
 Fever7.721.117.2 
Cutaneous15.421.920.00.324
 Purpura13.518.817.2 
 Ulcers3.93.13.3 
 Gangrene0.00.80.6 
Mucous membranes/eyes25.027.326.70.747
 Mouth ulcers7.711.710.6 
 Conjunctivitis/episcleritis13.514.814.4 
 Retroorbital mass/proptosis1.94.73.9 
 Uveitis0.00.00.0 
 Scleritis1.92.32.2 
Ear/nose/throat88.571.976.70.017
 Bloody nasal discharge/nasal  crusting/ulcer67.356.359.4 
 Sinus involvement59.641.446.7 
 Swollen salivary gland1.93.12.8 
 Subglottic inflammation11.57.08.3 
 Conductive deafness25.022.723.3 
 Sensorineural deafness1.93.12.8 
Cardiovascular1.91.61.70.864
 Pericarditis0.00.80.6 
Gastrointestinal1.92.31.70.862
 Mesenteric ischemia0.00.80.6 
Pulmonary51.963.360.00.159
 Pleurisy7.77.07.2 
 Nodules or cavities38.528.131.1 
 Other infiltrates23.132.029.4 
 Endobronchial involvement3.93.13.3 
 Alveolar hemorrhage0.025.017.8 
 Respiratory failure0.02.31.7 
Renal28.964.153.9<0.001
 Hematuria25.028.927.8 
 Red blood cell casts3.935.926.7 
 Elevated serum creatinine3.928.121.1 
Nervous system0.012.58.90.017
 Meningitis0.00.00.0 
 Cord lesion0.00.00.0 
 Stroke0.00.00.0 
 Cranial nerve palsy0.02.31.7 
 Sensory peripheral neuropathy0.010.27.2 
 Motor mononeuritis multiplex0.02.31.7 
Other involvement46.239.141.10.244
Table 5. Unusual disease features reported in the WGET cohort*
Disease manifestationNo. of patients
  • *

    WGET = Wegener's Granulomatosis Etanercept Trial.

  • Both serum aspartate aminotransferase and serum alanine aminotransferase levels were >1,000 mg/dl at enrollment. Other causes of aminotransferase elevations were excluded. Levels of these enzymes normalized with treatment.

Lacrimal gland involvement3
Supraglottic airway inflammation2
Aural chondritis2
Breast mass2
Peripheral 7th cranial nerve lesion2
Keratitis2
Optic neuropathy mimicking optic neuritis1
Splenic involvement; asymptomatic infarctions found on computed tomography1
Central nervous system involvement with parenchymal brain lesions and seizures1
Skull-based lesion in the infratemporal fossa, misdiagnosed as a tumor1
Diabetes insipidus1
Pituitary involvement1
Scalp tenderness1
Scalp lesion1
Myositis1
Myocarditis1
Hilar adenopathy1
Tongue ulcer1
Nodular vocal cord lesion1
Hepatic transaminase elevation1

Because significant alveolar hemorrhage, major renal involvement, and nervous system disease were designated as features of severe disease by definition, there were also significant differences in the numbers of patients in the 2 subsets with these complications at the time of enrollment. The overall proportion of patients in the 2 groups who had lung involvement was approximately equal (63% of patients with severe disease and 52% of those with limited disease; P = 0.159). However, 25% of patients in the severe disease subset had alveolar hemorrhage, compared with none in the limited disease group (P < 0.001). Among patients with limited WG who had lung disease, pulmonary involvement was characterized by lung nodules and cavities in 74% and by nonspecific pulmonary infiltrates in 45%.

Sixty-four percent of patients in the severe disease subset had glomerulonephritis at the time of enrollment, compared with only 29% of those with limited disease (P < 0.001). The median serum creatinine level at entry for those in the severe disease subset was 1.3 mg/dl (mean ± SD 2.06 ± 2.19 mg/dl), compared with 0.9 mg/dl (mean ± SD 0.97 ± 0.20 mg/dl) for patients in the limited disease subset. Ten patients with severe disease required dialysis at baseline or had been dialyzed because of WG prior to enrollment.

Neurologic complications of WG were considered, by definition, to be manifestations of severe disease. Thirteen percent of patients in the severe disease group had neurologic manifestations of active WG at the time of enrollment. The most common neurologic complication was a sensory peripheral neuropathy (10% of the severe disease group), followed by cranial nerve palsies (2%) and motor mononeuritis multiplex (2%).

History of organ involvement.

We also evaluated patients' histories of organ involvement before the period in which they became eligible for the WGET. These data are shown in Table 4, along with comparable data from the other 2 large WG patient cohorts (2, 6). Arthralgias, reported by more than three-fourths of all patients, were more common than was frank arthritis, a symptom in slightly fewer than half of all patients. Subcutaneous nodules (occurring in 10% of patients) were the second most common cutaneous manifestation of WG in the WGET cohort (purpura, occurring in 21% of patients, was the most common cutaneous manifestation). Subcutaneous nodules tended to occur over the extensor surfaces of joints (particularly the elbows) and were clinically consistent with cutaneous extravascular necrotizing granulomata (Churg-Strauss granulomas) (26). Patients were often unaware of the presence of such nodules. The number of patients with a history of hearing loss was approximately equal in the 2 disease subsets, but conductive hearing loss (33% of patients) was substantially more common than were sensorineural deficits (6% of patients).

Table 4. Features of major Wegener's granulomatosis cohorts*
Trial/period of collectionWGET/June 2000–September 2002 (n = 180)NIH/1966–1990 (n = 158)Lübeck/1966–1993 (n = 155)
  • *

    WGET = Wegener's Granulomatosis Etanercept Trial; cANCA = antineutrophil cytoplasmic antibody (cytoplasmic staining); NI = no information; NA = not applicable.

Mean age at diagnosis, years474148
Male/female, %58/4250/5049/51
Women with limited disease, %57.7NINI
Women with severe disease, %32.8
Mean followup, yearsNA87
Total patient years, no.3811,2292,144
cANCA positive, no.738884
Organ involvement ever, %   
 Upper respiratory tract909299
 Lung758566
 Kidney687770
 Eye315261
 Heart3.7825
 Skin38.54633
 Gastrointestinal tract6.806
 Peripheral nerves141540
 Central nervous system5811
 Arthralgias/arthritis816777

Because of the frequent difficulty in attributing a cause to the symptom of “fatigue” in patients with WG, this item was excluded intentionally from the baseline medical history form (which was derived from the BVAS/WG form) (22). Nevertheless, physician-investigators designated “fatigue” as an “other” symptom in 10% of patients at the time of enrollment. The fatigue was often noted to be overwhelming or profound. Finally, in addition to having the organ manifestations regarded as typical of WG, patients enrolled in the WGET also had a number of unusual disease complications (either at baseline or before their enrollment in the trial). These are shown in Table 5.

Laboratory features.

Selected laboratory features of the patients at baseline are shown in Table 6. By immunofluorescence testing, 87% of patients enrolled were known to have been ANCA positive at some time in the course of their disease. More than 90% of those in the severe disease group were ANCA positive, compared with 78% in the limited disease subset (P = 0.028). Eighty-seven percent of the patients had shown cytoplasmic staining upon immunofluorescence testing for ANCA, and 12% had shown perinuclear staining. The staining pattern was not specified in 2% of patients. Slightly fewer patients were positive for ANCA by EIA (86% overall; 73% anti–proteinase 3 and 12% antimyeloperoxidase). Again, those in the severe disease subset were more likely than those with limited disease to be ANCA positive by EIA (91% and 67%, respectively; P < 0.001).

Table 6. Selected laboratory measures by disease severity*
Laboratory measureLimited (n = 52)Severe (n = 128)All patients (n = 180)P
  • *

    ANCA = antineutrophil cytoplasmic antibody; IF = immunofluorescence; cANCA = ANCA (cytoplasmic staining); pANCA = ANCA (perinuclear staining); anti-PR3 = anti–proteinase 3; EIA = enzyme immunoassay; anti-MPO = antimyeloperoxidase.

  • Determined by chi-square test or Fisher's exact test for categorical data and by Wilcoxon's rank sum test for continuous data.

  • One hundred fifty-five patients had a positive immunofluorescence result for ANCA at any time; 152 of these patients had specific ANCA patterns on the baseline medical history form.

  • §

    Thirty-seven (71%) of 52 patients with limited disease had biopsy results showing either granulomatous inflammation or vasculitis. Ninety-seven (75%) of 128 patients with severe disease had biopsy results showing at least 1 of these features.

Creatinine at baseline, mean ± SD mg/dl0.97 ± 0.202.06 ± 2.191.74 ± 1.91<0.001
Dialysis ever, %0.07.85.60.065
Hematocrit, mean ± SD %37.3 ± 8.135.5 ± 5.236.0 ± 6.20.003
ANCA positive ever (by IF), %78.490.687.10.028
 cANCA, % of total ANCA positive by IF89.786.787.5 
 pANCA, % of total ANCA positive by IF10.313.312.5 
Anti-PR3 ANCA positive (by EIA), %58.878.172.60.028
Anti-MPO ANCA positive (by EIA), %8.013.411.90.134
Biopsy-proven diagnosis§    
 Granulomatous inflammation, %67.650.054.90.068
 Vasculitis/glomerulonephritis, %46.070.163.40.009

Pathology.

Subset patterns according to tissue biopsy also emerged. We categorized patients' biopsy results according to whether they demonstrated granulomatous inflammation, vasculitis, or both. The pathologic finding of glomerulonephritis on renal biopsy was considered to be an indication of small-vessel vasculitis in the kidney. One hundred thirty-four (74.4%) of the 180 patients had tissue biopsy results demonstrating either granulomatous inflammation or vasculitis. Seventy percent of patients with severe disease whose biopsy specimens showed at least 1 of these features had histopathologic features of vasculitis, compared with 46% of patients with limited disease (P = 0.009). Conversely, 68% of patients with limited disease had biopsy results indicating granulomatous inflammation, compared with 50% of patients with severe disease (P = 0.068).

Other medical history.

Because of the previously described associations between medications used to treat thyrotoxicosis and the occurrence of ANCA-associated vasculitis (27, 28), we collected information about patients' histories of thyroid disease. Nineteen patients with severe disease (15%) had a history of thyroid disease, compared with no patients with limited WG (P = 0.003). Among the patients with severe disease and a history of thyroid disorders, 6 had a history of Graves' disease, 10 had a history of hypothyroidism of unknown etiology (presumed Hashimoto thyroiditis), 1 had a history of multinodular goiter, and 2 had a history of asymptomatic thyroid nodules.

DISCUSSION

This study represents the first attempt to compare the disease features of patients with limited WG and those with severe WG in the setting of a clinical trial. Patients enrolled in the WGET represent the largest prospectively collected group of patients with WG reported to date. Moreover, the patients enrolled in the WGET were 1) assembled over a period of only 28 months, in contrast to >2 decades for the other 2 prospectively enrolled cohorts, 2) evaluated by a group of experts who were certified on the data collection procedures, and 3) prescribed treatments according to a uniform protocol based on disease severity. We observed important differences in the epidemiologic characteristics, disease features, and even risk factors between these 2 subsets. Patients in the limited disease subset were more likely to be women and were significantly younger at the time that symptoms of WG began, yet they also had a greater likelihood of having destructive upper airway disease and exacerbations of previous disease following periods of remission. Patients with severe disease were more likely to be men and to be ANCA positive (a finding reported previously by others) (11), and they had a strikingly greater likelihood of having thyroid disease prior to trial entry.

Because WG is an uncommon disorder, many of its basic epidemiologic features have not been described completely. Early reports on this disease, for example, indicated a male predominance (∼60%) (3). Subsequent studies suggested that the number of men and women with this disease was approximately equal (2). At least 1 epidemiologic study has even reported a slightly higher percentage of women with the disease (29). In the first description of patients with limited WG (8), women outnumbered men at a ratio of 10 to 6. Our data suggest a subtlety in sex distribution in WG: whereas severe disease appears to occur with a male predominance, limited disease appears to occur more commonly in women. These observations were made not only in patients enrolled in the WGET but also among the 778 subjects who were screened but were not enrolled. Our findings are consistent with those of a survey of 701 patients with WG in North America that found a higher percentage of women (56%), particularly among patients in whom the disease began between the ages of 9 and 40 years (30). Conversely, among the patients in that study in whom WG was diagnosed between the ages of 45 and 72 years, more patients were men, a finding that is also consistent with our data.

Earlier reports emphasized that limited WG may evolve into severe disease, implying that these 2 states are part of the same disease spectrum (9, 31). DeRemee et al (9) hypothesized that limited WG represents a different tempo of the same disease, and postulated that the lack of progression in some cases might be attributable to treatment. One might surmise, therefore, that the patients with limited WG in our cohort were younger than those in the severe disease subset merely because their disease was in an earlier stage of evolution (i.e., that the natural history of their disease would be to evolve into severe disease over time). Certain characteristics of patients in the limited disease subset, however, suggest that this is often not the case. First, the pattern of organ involvement over time differed in important ways between the 2 groups. Despite their younger age, those with limited disease were more likely to have WG-related upper respiratory tract damage (e.g., nasal septal perforation, saddle-nose deformity, sinus involvement) than were those with severe disease. Second, patients in the limited disease group were more likely to have disease flares following periods of remission (i.e., to have relapsing disease). Third, patients whose disease features were consistent with limited disease at the time of their original diagnosis tended to remain true to their initial presentations (although exceptions to this were noted). Finally, the disparity in histories of thyroid disease between the 2 subsets—19 patients (15%) in the severe disease group compared with none in the limited disease group—suggests the possibility of different risk factors (genetic or acquired) and/or triggers for disease subsets.

The percentage of patients with severe WG and a history of thyroid dysfunction was striking and significantly higher than the prevalence of hyperthyroidism and hypothyroidism in the general population. In the Whickham survey (32, 33), a population-based study from northeastern England, 2% of the population had a history of either hyperthyroidism or hypothyroidism, compared with 13% of the WGET patients with severe disease (P < 0.001). In addition, despite the fact that Graves' disease is an unusual disorder in men, 4 (5%) of the 86 men with severe disease in the WGET had a history of Graves' disease before the onset of WG. (One of the 2 women in the WGET was diagnosed as having Graves' disease several years after her WG diagnosis.)

Although the overall numbers of patients with thyroid disease in this cohort are relatively small, there are at least 2 mechanisms by which thyroid disease and severe WG could be linked. First, WG, Graves' disease, and Hashimoto thyroiditis all have major autoimmune features and may therefore share etiologic factors (e.g., susceptibility alleles). Second, it is known that >25% of patients with Graves' disease who are treated with propylthiouracil develop ANCA (usually directed against myeloperoxidase) (34), and that some of these patients develop ANCA-associated vasculitis, including WG (35).

The overall features of patients in the WGET cohort were similar to those of patients in the NIH and Lübeck cohorts (2, 6). The distinction between limited and severe disease had no implications for treatment until 1992, when the efficacy of MTX for limited WG was first reported (2). Consequently, in those cohorts, patients with limited WG and those with severe WG were treated largely in the same manner, and subset differences were not analyzed in detail. (It was noted in the Lübeck cohort, however, that 15% of the patients had “initial-phase” disease at the time of diagnosis. In addition, among the NIH cohort, although only 18% of the patients had renal involvement at the time of diagnosis, 77% developed this complication over time.) Perhaps because of greater recognition of the limited disease subset over the past decade, our cohort of screened and enrolled patients may have included a higher percentage of patients meeting this definition.

The ethnic distribution of our patient cohort confirms that WG is most common among white individuals, a finding reported by other investigators (2, 36). The clinical features of the WGET cohort were generally similar to those of the NIH and Lübeck cohorts. Among the patients in the Lübeck cohort, the higher percentages of patients with cardiac (25%) and peripheral nerve (40%) involvement stand out. The higher percentages of patients with these complications may be explained, at least in part, by the longer followup of that cohort (see Table 4).

In certain other systemic diseases associated with autoimmunity (e.g., systemic sclerosis), investigators distinguish between limited and diffuse disease subsets (37). In the case of systemic sclerosis, limited does not necessarily imply a lower degree of disease severity; rather, it refers to the extent of skin involvement, which is more circumscribed in that subset compared with the diffuse form. Patients with systemic sclerosis whose disease is limited may still have profound and often life-threatening complications, including digital loss (38), pulmonary hypertension, and gastrointestinal tract dysfunction (39). Under the definition of limited disease that we used, patients in this subset have, by definition, a lower risk of events that are immediately organ- or life-threatening. However, their risks for long-term complications from both the disease and its treatment (e.g., their requirement for ongoing immunosuppressive therapy and greater numbers of surgical interventions for upper respiratory tract complications) may be substantial. In part, this is reflected in the baseline damage scores in the 2 groups, which were equal (1.5 for the limited disease group and 1.3 for the severe disease group).

A number of studies in the literature suggest a possible pathophysiologic basis for the differences that we report here. In an examination of nasal biopsy specimens, Müller et al (40) noted differences in cytokine expression across the localized and generalized disease subsets. Balding et al (41) detected a Th2 response within nasal biopsy specimens obtained from patients with WG but noted that this bias appeared to be lost in both renal biopsy samples and among circulating T cells that responded to proteinase 3. Other investigators have observed differences in the expression of chemokines and chemokine receptors across disease subsets (42, 43). In addition, some studies are consistent with confinement of critical cytokine production to certain T cell subsets, and recruitment of such cells (particularly CD28-negative T cells) into granulomatous lesions of patients with WG (44, 45).

Our study has certain potential limitations. First, when enrollment centers are composed entirely of tertiary care medical centers, there is always the potential for referral bias because of the tendency for patients with severe disease to be referred to such institutions. There are, however, no large, population-based cohorts of WG patients against which the WGET cohort can be compared. The comparisons of our patients with those in the NIH and Lübeck cohorts indicate that our patients were quite similar. Furthermore, it might be argued that patients with major complications of limited disease (e.g., saddle-nose deformity, subglottic stenosis) are also quite likely to seek care at or be referred to tertiary medical centers. Second, because a certain degree of disease activity was required for trial entry, patients with severe disease were more likely to be enrolled. However, our data on patients who were screened for the trial but were not enrolled (although less comprehensive than that collected on the patients enrolled) suggest that those patients were similar to those included in the WGET.

Third, patients enrolled in the WGET do not represent an inception cohort consisting entirely of cases that were untreated prior to trial entry. The most ideal method of examining differences between patients with limited WG and those with severe WG would be to conduct long-term followup of a cohort of patients with newly diagnosed WG. It is possible, as suggested by other investigators (9, 46), that the likelihood of developing severe features of WG following presentation with limited disease is altered by therapy. The distinct features of the subsets reported herein, however, argue that the natural histories of patients with limited disease and those with severe disease often differ in important ways, despite the occurrence of overlap in many disease features.

In conclusion, this study represents the first attempt to delineate the characteristics of 2 subsets of patients with WG (those with limited disease and those with severe disease). It is possible that other WG subsets exist, and that our appreciation of these subsets will become more refined as our understanding of this disease advances. Further attempts to characterize and understand clusters of traits may be useful in studies of subset prognosis, treatment, and pathogenesis. At this point in time, the distinction between limited and severe disease is clearly important with regard to the choice of treatment. Is the distinction between limited and severe WG also important from the standpoint of response to therapy and that of prognosis? What might differences in the selection of target organs tell us about pathogenetic mechanisms? These are questions that can be resolved only by future investigations. Based on the data from our cohort, subset differences in epidemiologic characteristics, disease features, and risk factors suggest that there may also be differences in responses to treatment and in disease outcomes. Further analysis of these issues may impact the design of new studies and clinical trials, and may ultimately reshape our thinking about the nature of this disease.

APPENDIX A

THE WGET RESEARCH GROUP

Members of the WGET Research Group are as follows: for The Johns Hopkins Vasculitis Center, Baltimore, Maryland (chairman's office), John H. Stone, MD, MPH, (chairman), Misty L. Uhlfelder, MPH (assistant trial coordinator), Amanda M. Moore, BS (research coordinator); for The Cleveland Clinic Foundation Center for Vasculitis Research and Care, Cleveland, Ohio (co-chairman's office), Gary S. Hoffman, MD (co-chairman); for The Johns Hopkins University Center for Clinical Trials, Baltimore, Maryland (coordinating center), Janet T. Holbrook, PhD, MPH (director), Curtis L. Meinert, PhD (associate director), John Dodge (systems analyst), Jessica Donithan (research coordinator), Nancy Min, PhD (biostatistician), Laurel Murrow, MSc (trial coordinator, former), Jacki Smith (research data assistant), Andrea K. Tibbs, BS (trial coordinator), Mark Van Natta, MHS (biostatistician); for The Beth Israel Medical Center, New York, New York (clinical center), Robert Spiera, MD, Rosanne Berman, MPH, Sandy Enuha, MPH; for Boston University, Boston, Massachusetts (clinical center), Peter A. Merkel, MD, MPH, Rondi Gelbard, BS, Melynn Nuite, RN, Aileen Schiller, MS; for The Cleveland Clinic Foundation (clinical center), David Blumenthal, MD, Debora Bork, MFA, Tiffany Clark, CNP, Sonya L. Crook, RN, Leonard H. Calabrese, DO, Sharon Farkas, Sudhakar Sridharan, MD, Kimberly Strom, CNP, William Wilke, MD; for Duke University, Durham, North Carolina (clinical center), E. William St.Clair, MD, Nancy B. Allen, MD, Karen Rodin, RN, Edna Scarlett; for Johns Hopkins University, Baltimore, Maryland (clinical center), David B. Hellmann, MD, Lourdes Pinachos, RN, BSN, Michael J. Regan, MD, MRCP; for The Mayo Clinic, Rochester, Minnesota (clinical center), Ulrich Specks, MD, Kristin Bradt, Kimberly Carlson, Susan Fisher, RN, Boleyn Hammel, Kathy Mieras, Steven Ytterberg, MD; for the University of California, San Francisco (clinical center), John C. Davis, MD, MPH, Maureen Fitzpatrick, MPH, Ken Fye, MD, Steve Lund, MSN, NP; for the University of Michigan, Ann Arbor (clinical center), Joseph McCune, MD, Billie Jo Coomer, BS, Barbara Gilson, RN, Hilary Haftel, MD, Ana Morrel-Samuels, BA, Sandra Neckel, RN; and for The Johns Hopkins University Immune Diseases Laboratory (resource center), Noel R. Rose, MD, PhD, C. Lynne Burek, PhD, Jobert Barin, BS, Monica Talor, MS.

Members of the Data and Safety Monitoring Board of the WGET Research Group are as follows: Paul L. Canner, PhD (Maryland Medical Research Institute, Ft. Detrick), Doyt L. Conn, MD (Emory University, Atlanta Georgia), John H. Klippel, MD (Arthritis Foundation, Atlanta, Georgia), J. Richard Landis, PhD (University of Pennsylvania, Philadelphia), and Barbara White, MD (University of Maryland, College Park).

Ancillary