Article Text

Download PDFPDF

Extended report
Prospective long-term follow-up of patients with localised Wegener's granulomatosis: does it occur as persistent disease stage?
  1. Julia U Holle1,
  2. Wolfgang L Gross1,
  3. Konstanze Holl-Ulrich2,
  4. Petra Ambrosch3,
  5. Bernhard Noelle4,
  6. Marcus Both5,
  7. Elena Csernok1,
  8. Frank Moosig1,
  9. Susanne Schinke1,
  10. Eva Reinhold-Keller1
  1. 1Department of Rheumatology and Clinical Immunology, Vasculitis Center, University Hospital Schleswig-Holstein, Campus Luebeck and Klinikum Bad Bramstedt, Germany
  2. 2Institute for Pathology, University of Luebeck, Germany
  3. 3Department of Otorhinolaryngology, University of Kiel, Germany
  4. 4Department of Ophthalmology, University of Kiel, Germany
  5. 5Department of Diagnostic Radiology, University of Kiel, Germany
  1. Correspondence to Dr Julia U Holle, Department of Rheumatology and Clinical Immunology, Vasculitis Center of University Hospital Schleswig-Holstein/Klinikum Bad Bramstedt, Oskar-Alexander-Str. 26, 24576 Bad Bramstedt, Germany; holle{at}


Objective To identify patients with localised Wegener's granulomatosis (locWG) to assess whether it occurs as a long-term disease stage or phenotype and to characterise its outcome.

Methods Patients in a ‘localised stage’ with histological criteria compatible with WG and a follow-up period of ≥1 year were included. They were prospectively followed at the Vasculitis Center Schleswig-Holstein from 1989 to 2009 and the clinical manifestations, antineutrophil cytoplasmic autoantibodies (ANCA) status and damage were evaluated. Immunosuppression was adapted to disease activity and severity in a step-up regimen.

Results Of 1024 patients with suspected WG, 99 were clinically diagnosed with locWG and 50 fulfilled the inclusion criteria (72% women, median age 43 years, 46% ANCA-positive). The median follow-up was 48 months. All achieved a response to treatment, 34% achieved complete remission, 1–4 relapses occurred in 46%, 5 (10%) had generalised disease (median 6 years after onset). ANCA status was not associated with relapse (p=0.98), transition to generalised disease (p=0.51) or refractory manifestations (p=0.60). 47% required cyclophosphamide for localised manifestations, 36% of them for pulmonary masses and 24% for orbital masses. 66% developed organ damage, mostly due to bony destruction or space obturation (28% saddle nose, 24% septal perforation, 10% orbital wall destruction). There were two deaths that were not related to WG.

Conclusion There is evidence that locWG is a long-term disease stage or phenotype (5% of all patients with WG), 46% of whom are ANCA-positive. LocWG is characterised by destructive and/or space-consuming lesions associated with high relapse rates (46%) and local damage.

Statistics from


Wegener's granulomatosis (WG) is usually associated with fulminant life-threatening disease and unfavourable outcome if untreated, as described in many cohort studies.1,,6 However, milder disease variants restricted to the respiratory tract (localised WG, locWG) have been reported to appear sporadically.7,,12 In the American College of Rheumatology (ACR) classification criteria locWG is not considered,13 yet the Chapel Hill Consensus (CHC) definitions clearly state that ‘granulomatous variants’ can occur.14 Although there is no clear definition of ‘granulomatous disease’, it alludes to the formation of masses round in appearance and/or to the histopathological picture of granuloma. Mass formation is typically seen in upper and lower respiratory tract manifestations of WG (eg, orbital masses, pulmonary mass); granuloma is a characteristic histopathological finding in biopsy specimens of such lesions. This aspect of locWG was first proposed in 1990 by Boudes15 and has been discussed recently.16

In order to facilitate stage-adapted treatment, the European Vasculitis Study Group (EUVAS) defined disease stages in antineutrophil cytoplasmic autoantibody (ANCA)-associated vasculitides (AAV) that include localised disease defined as restricted to the respiratory tract with absence of clinical signs of systemic vasculitis.17 However, locWG is considered to be a short-term disease stage that occurs very early as most patients present to the clinician at the systemic stage of the disease.5 Furthermore, the definition of locWG is not accepted worldwide; in the USA, a systemic non-life-threatening stage (limited form) is differentiated from a life-threatening stage, but a strictly localised stage is not considered.18

A few studies have considered locWG and give an idea of the proportion of patients diagnosed with localised disease (ranging from 3.9% to 22%),2 3 5 but these data only refer to the proportion at diagnosis, not during follow-up. In an earlier study we found that more than half of patients with locWG were ANCA-positive and younger (median age 41 years) than those with generalised WG, but this study lacked a long-term follow-up.19 It is known that locWG manifestations may lead to local damage (eg, saddle nose deformity or bony destruction of sinuses) and to organ- or life-threatening complications (eg, acute respiratory insufficiency due to subglottic stenosis or visual loss due to orbital granuloma).8 10,,12 In single cases the transition to generalised disease was described after 20 years of localised disease,20 21 but this has never been evaluated systematically.

To the best of our knowledge, no systematic long-term study of patients with locWG has been performed to elucidate whether locWG occurs as a long-term disease stage or phenotype.

In this study we focus on patients with locWG for ≥1 year who have a biopsy specimen compatible with WG and have been prospectively followed at our centre from 1989 to 2009 to elucidate whether there is a localised long-term disease stage/phenotype of WG and to characterise these patients with regard to clinical manifestations, ANCA status, treatment and outcome.


Vasculitis centre and patients

All consecutive patients with WG or suspected WG were assessed at the Vasculitis Center from October 1989 to October 2009. For this cohort study only patients who had clinically localised disease restricted to the upper and/or lower respiratory tract as defined by the EUVAS22 and a biopsy specimen that was compatible with WG (defined below) irrespective of the ANCA status were selected. Patients with orbital masses were also included in the study as this manifestation does not correspond to systemic vasculitis and can therefore not be counted as early systemic/generalised disease,22 even if the orbit is not part of the respiratory tract. All other disease stages were defined according to the EULAR/EUVAS.22 The study was conducted according to the Declaration of Helsinki.


All patients underwent at least one biopsy procedure. Patients could be biopsied several times (at the same site) and biopsy criteria were added up until histopathological criteria were fulfilled. Biopsy specimens were formalin-fixed and paraffin-embedded and sections of 6 µm were cut and stained with H&E, PAS and for acid-fast bacteria. Infections by, for example, mycobacteria and fungi were excluded.

Biopsy specimens were reviewed at the German Reference Center for Vasculitis (Department of Pathology, University of Luebeck) by two independent pathologists and considered diagnostic if at least two of the three features of granulomas, small to medium size vessel vasculitis or geographic necrosis were present23 and reviewers' findings matched.

Assessment of clinical manifestations, activity and activity states

All patients underwent standardised interdisciplinary clinical, laboratory and technical examinations (such as ENT and ophthalmology assessment, MRI of the head24) to assess organ manifestations, disease stage, activity, damage and treatment-related side effects at regular 6–12-monthly intervals, as previously described.5 Complete remission, partial remission/response, relapse and refractory disease were defined according to the EULAR/EUVAS.22 Disease activity was measured according to the Birmingham Vasculitis Activity Score (BVAS) (retrospectively before 1999) and damage was defined as an irreversible manifestation present for >3 months and assessed by the Vasculitis Damage Index (VDI).25,,27 The VDI was applied retrospectively in patients followed before the publication of the VDI in 1997.

Treatment procedures

Treatment was adapted according to activity of disease and severity of the manifestation. Patients with endonasal activity (without destructive disease) received trimethoprim/sulfamethoxazole (T/S) and methotrexate (MTX) was used for refractory or ulcerative rhinitis/sinusitis with mass formation, subglottic stenosis, pulmonary masses. In case of side effects or contraindications, other immunosuppressants were used (azathioprine, leflunomide, mycophenolate), mostly in addition to low-dose glucocorticoids (eg, prednisone 5 mg/day). Orbital masses, all other localised manifestations resistant to medium potent immunosuppressants and generalised disease were treated with cyclophosphamide (Cyc) for remission induction. Additional rituximab, tumour necrosis factor (TNF) antagonists or intravenous immunoglobulins were used in Cyc-resistant/refractory cases. When remission induction with Cyc was necessary, glucocorticoids were given at 1 mg/kg prednisone/day (for dosages see online supplement). Some patients with subglottic stenosis or orbital mass underwent local treatment procedures or surgery.

ANCA testing

ANCA testing was performed by an indirect immuno- fluorescence test using ethanol- and formalin-fixed leucocytes as screening assay and direct ELISA for the detection of target antigen, as previously described according to the consensus guidelines.28

Statistical analysis

Categorical variables were reported in frequencies and percentages. χ2 and Mann–Whitney tests were applied where appropriate. Statistical analyses were performed using SPSS Version 12. A two-sided p value <0.05 was considered statistically significant. To assess whether there was a decline in interval from first manifestation to diagnosis, patients were split into two cohorts diagnosed before and after 2000.


Patient characteristics

A total of 1024 patients presented with WG or suspected WG in the period from 1989 to 2009. Ninety-nine patients with clinically suspected locWG were identified, 49 of whom had to be excluded (see online supplement), leaving 50 patients who fulfilled the entry criteria, 36 (72%) of whom were women. The interval from first symptoms to diagnosis was reduced in patients diagnosed before 2000 (n=28) and after 2000 (n=22) (from median 12 to 6.7 months), but this was not significant (p=0.17). The median follow-up period was 4 years (range 1–14.5 years) for all patients, 6 years (range 3.75–11 years) for patients who developed systemic disease (n=5) and 4 years (range 1–14.5 years) for patients remaining in the localised stage (n=45), 25 of whom had a follow-up of at least 4 years (range 4–14.5, median 6.17). Details of patient characteristics and outcome are summarised in table 1.

Table 1

Characteristics and outcome of patients with localised Wegener's granulomatosis

Disease manifestations and outcome

An overview of disease manifestations is given in figure 1; 46% of all patients presented with localised disease restricted to the upper respiratory tract (nose and sinuses only; orbits and lung excluded) and 56% had disease manifestations consistent with mass formation (eg, mass formation in the sinuses assessed by MRI, orbital or pulmonary masses).

Figure 1

Disease manifestations of patients with localised Wegener's granulomatosis at disease onset, beginning of the follow-up period and over the whole course of the follow-up period.

Five patients (10%) developed systemic disease (1 early systemic, 4 generalised) after a median disease course of 72 months (range 48–84). None developed a severe generalised disease stage (creatinine >500 µmol/l).

ANCA serology

Twenty-three patients (46%) were ANCA-positive (PR3-ANCA, n=20 or MPO-ANCA, n=3) over the whole course of the follow-up period (table 1). There was no significant difference between all ANCA-positive (PR3-ANCA plus MPO-ANCA) and ANCA-negative patients with regard to gender (p=0.73), age (p=0.98), relapses (p=0.982), number of relapses (p=0.51), VDI at the end of follow-up (p=0.75) and organ manifestations (data not shown). There were also no significant differences when MPO-ANCA- and PR3-ANCA-positive and -negative patients were analysed separately (data not shown). The presence of ANCA (over the whole course of follow-up) was not associated with the development of early systemic/generalised disease (p=0.51) or refractory disease (p=0.60).


All patients underwent at least one biopsy procedure (1–3 biopsies) and 12 patients had biopsies from ≥2 organs. In 30 patients >1 biopsy was required to match the histological criteria: 33 patients (66%) had 46 nasal mucosa biopsies and 18 patients (34%) had 19 paranasal sinus or mastoid biopsies (see also online supplement).

Geographical necrosis was the most frequent feature present in the biopsy specimens (92%) followed by granulomatous lesions (84%) and vasculitis (32%). Other frequent features were ulcers (26%), microabscesses (22%) and lymphofollicular structures (22%) (table 2). There was no significant difference in the presence of vasculitis (p=0.151) and granulomatous lesions (p=0.804) in the biopsy specimens of ANCA-positive and ANCA-negative patients.

Table 2

Frequency of histology features present in all biopsy specimens/patient

In three patients who developed early systemic/generalised disease, biopsies were also taken from sites other than the respiratory tract and all showed vasculitis (one focal necrotising intracapillary and extracapillary glomerulonephritis, one large vessel vasculitis of the temporal artery and one leucocytoclastic vasculitis). One patient who developed generalised disease after 4 years of localised disease initially had repeated nasal mucosa biopsies with non-specific results but developed biopsy-proven necrotising intracapillary and extracapillary glomerulonephritis after 4 years.

Initial treatment and treatment escalation and de-escalation

Initially, patients received either T/S (26 patients, 52%), MTX (8 patients, 16%) or Cyc (16 patients, 32%) (figure 2). Indications for Cyc as initial treatment were orbital masses (n=4), cavitating pulmonary masses (n=4), subglottic stenosis (n=1), progressive hearing loss (n=1) and sinusitis with mass formation and bony destruction (n=5).

Figure 2

Primary treatment of all patients and treatment escalation. T/S, trimethoprim/sulfamethoxasole; MTX, methotrexate; Cyc, cyclophosphamide.

In 19 of 26 patients (73%) initial T/S was not sufficient and they were switched to MTX (n=12) due to progression to early systemic disease (n=1), persistent rhinitis/sinusitis (n=4), development of sinusitis with mass formation (n=5) or subglottic stenosis (n=2) or to Cyc (n=7) for progression to generalised disease (n=4), cavitating pulmonary masses with a diameter >3 cm (n=2) or sinusitis with mass formation and bony destruction (n=1).

MTX was given to 8 patients (16%) initially and was sufficient in 4; the other 4 required escalation of treatment by adding leflunomide to MTX or switching to Cyc. Four patients (8%) could be successfully de-escalated from high- or medium-potent immunosuppression to T/S with or without glucocorticoids. Seven patients (14%) were successfully tapered and remained off treatment for 3–108 months (median 12 months).

Cumulative treatment

In four patients (8%) T/S was the only medication over the course of the follow-up period and in three patients (6%) only glucocorticoid plus T/S was given (figure 3). Half of the patients (n=25) required Cyc (pulse or oral) for remission induction at some point during the follow-up period (eight patients more than once) and 21 patients (47%) who did not progress to systemic disease also required Cyc therapy. Four of five patients who developed systemic disease received Cyc upon generalisation (figure 3). Most frequent indications for Cyc were pulmonary masses in 36% of patients and orbital masses in 24% (see also online supplement). The median lifetime cumulative dose of Cyc for all patients receiving Cyc and for patients not developing generalised disease was 23 g (range 6–300 g). Six patients were refractory (n=5) or intolerant to Cyc (n=1) and received TNF antagonists or rituximab (5 for localised disease, 1 for systemic disease). Six patients received TNF antagonists (1 for conductive hearing loss, 2 for pulmonary masses, 2 for orbital masses, 1 for vasculitis of the temporal artery with leucopenia under conventional immunosuppression), inducing a response in five of the six patients. One patient received rituximab for orbital masses refractory to Cyc and anti-TNF therapy which also led to a response. Three patients received intravenous immunoglobulin during concomitant disease activity and infection.

Figure 3

Cumulative treatment of all patients, patients remaining in the localised stage and patients with secondary generalisation. GC, glucocorticoid; T/S, trimethoprim/sulfamethoxasole; MTX, methotrexate; Cyc, cyclophosphamide; Aza, azathioprine; Lef, leflunomide; MMF, mycophenolate; IVIG, intravenous immunoglobulin; TNF-Ant, tumour necrosis factor antagonist; RTX, rituximab.

Surgical intervention and local treatments

Five of eight patients with subglottic stenosis required additional surgical intervention and/or local treatment (including three emergency tracheotomies). One patient became blind in spite of resection of orbital masses.

Remission, relapses and refractory disease

All patients achieved a response (formerly partial remission), 17 patients (34%) a complete remission. Nearly half of the patients (n=23, 46%) suffered 1–4 relapses (34 relapses in all). Most frequently, patients presented with relapses of inflammatory subglottic stenosis (32% of all relapses) and pulmonary masses (21% of all relapses) (for an overview of relapses see online supplement).

Eight of all patients (16%) and seven of those patients remaining localised (15.5%) presented with refractory disease, three with orbital masses, two with subglottic stenosis, and one patient each with local destructive mastoiditis, sinusitis with mass formation and bony destruction and glomerulonephritis (in one patient who developed generalised disease).


At the beginning of the follow-up period 34% had a damage score >0 (VDI=1 in 30% (n=15), VDI=2 in 4% (n=2)). At the end of the follow-up period 66% (n=33) had a damage score >0 (VDI=1 in 34% (n=17), VDI=2 in 20% (n=10), VDI=3 in 10% (n=5) and VDI=4 in 2% (n=1)). Most frequent manifestations of damage were related to destruction of the nasal skeleton (saddle nose deformity and septal perforation in 28% (n=14) and 24% (n=12), respectively, at the end of the follow-up period). Details are shown in figure 4.

Figure 4

Vasculitis damage index items scored as percentage of patients at the beginning and end of the follow-up period. Orbit wall destr, orbit wall destruction; radiol damage, radiological damage of sinuses.


Two patients died aged >80 years; in neither case was the reason for death associated with WG or immunosuppressive therapy.


To the best of our knowledge, this is the first systematic study to provide evidence for locWG as a long-term disease stage or disease phenotype.

Patients showed excellent long-term survival, rarely proceeded to generalised disease (10% of patients) and 53% required T/S and/or medium potent immunosuppression only. However, locWG is also characterised by the formation of masses with tumour-like invasion, local destruction of tissues, bones and cartilage or leading to space obturation.

As a consequence of an aggressive course, low or medium potent immunosuppression was not sufficient in a substantial proportion of patients; 32% of all patients required Cyc as initial treatment and 47% needed Cyc for localised disease manifestations over the whole course of follow-up. The most frequent indications for Cyc were cavitating pulmonary masses (36% of patients requiring Cyc) and orbital masses (24%). Fifty-two percent presented initially with mild disease and received T/S, which was not sufficient in most patients (19/26, 73%) and 7/26 (27%) even required highly potent immunosuppression with Cyc later. T/S should therefore be critically considered as a therapeutic option in locWG.

Moreover, nearly half of the patients (46%) suffered relapses within the study period, which is a similar rate to that reported in studies including mainly patients with generalised disease.5 Subglottic stenosis and pulmonary masses were the manifestations most frequently associated with relapse. Refractory disease also occurred with a similar frequency (16%) to that reported in studies of patients with generalised WG.5 Orbital masses were associated with the highest rate of refractory disease (in 3/6 patients with orbital masses).

According to the EULAR/EUVAS definition, localised disease (as disease stage) is expected to be ANCA-negative.22 In this study, we confirm former findings of a high frequency of ANCA-positive patients (46%) with localised disease,19 although approximately half of the patients were ANCA-negative. This observation probably relates to the fact that a positive ANCA status is more likely to be associated with clinical signs of systemic vasculitis (and therefore generalised disease) such as that seen in Churg-Strauss syndrome.29 Nevertheless, in our study a positive ANCA status was not associated with the development of systemic disease and can therefore not be considered as indicative of progression to systemic disease although, due to the small number of patients, this finding needs to be readdressed in a larger cohort.

With regard to the use of biopsy specimens to support the diagnosis of WG, 92% of specimens had necrosis, 84% granulomatous lesions and 32% vasculitis. Furthermore, we confirmed former findings of the frequent presence of microabscesses (22%) and lymphofollicular structures (22%).23 30 In clinical practice the formation of masses in WG is often called ‘granuloma’, although this term should be reserved for pathologists. Nevertheless, granulomatous lesions as defined by pathologists were found in 84% of our biopsy specimens, confirming that these lesions are indeed part of the mass. Which cells exactly may be responsible for the space-consuming and sometimes destructive mass formation needs to be clarified. Furthermore, the presence of vasculitis in 32% of the biopsy specimens despite the lack of clinical signs of vasculitis is an interesting finding and may further support the view that localised and generalised WG are indeed two phenotypes of the same disease.

There are several limitations to the study. First, the cohort is small and the follow-up period is limited (median 4 years). Although a significant proportion of patients were followed up for >4 years (range 4–14.5 years, median 6.17), a longer follow-up period is desirable to elucidate how many patients remain in the localised stage for the rest of their lives. Furthermore, a larger cohort of patients or a registry is required to further characterise this rare phenotype of patients. Second, this study was conducted at a tertiary referral centre for vasculitis; however, patients with locWG may be seen by an ENT specialist only and not be referred to us. Third, some of the data were collected retrospectively (BVAS, VDI before its introduction).

In conclusion, there is a localised long-term disease stage or phenotype of WG (around 5% of our cohort). LocWG is characterised by high rates of relapse and damage. Due to destructive and/or refractory disease, highly potent immunosuppression was required in nearly half of the patients. Although occurring in only 10% of patients, long-term localised disease still carries the danger of developing into a generalised stage. Larger cohorts of patients with locWG are needed to further characterise this disease phenotype.


View Abstract


  • Competing interests None.

  • Ethics approval All patients sign a form when entering our department in case they agree to have their data analysed for research purposes. This form is used for several of our observational studies and is approved by the ethics committee of the University of Luebeck. No specific approval for this study was therefore needed.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.