Article Text

Extended report
Rituximab for refractory granulomatosis with polyangiitis (Wegener's granulomatosis): comparison of efficacy in granulomatous versus vasculitic manifestations
  1. Julia U Holle1,
  2. Christin Dubrau1,
  3. Karen Herlyn1,
  4. Martin Heller2,
  5. Petra Ambrosch3,
  6. Bernhard Noelle4,
  7. Eva Reinhold-Keller1,
  8. Wolfgang L Gross1
  1. 1Department of Rheumatology and Clinical Immunology, Vasculitis Center, University Hospital Schleswig-Holstein, Campus Lübeck and Klinikum Bad Bramstedt, Bad Bramstedt, Germany
  2. 2Department of Diagnostic Radiology, University of Kiel, Kiel, Germany
  3. 3Department of Otorhinolaryngology, University of Kiel, Kiel, Germany
  4. 4Department of Ophthalmology, University of Kiel, Kiel, Germany
  1. Correspondence to Dr Julia U Holle, Department of Rheumatology and Clinical Immunology, Vasculitis Center, University Hospital Schleswig-Holstein/Klinikum Bad Bramstedt, Oskar-Alexander-Strasse 26, 24576 Bad Bramstedt, Germany; hollejulia{at}


Objective First, to investigate the overall efficacy and safety of rituximab (RTX) in refractory granulomatosis with polyangiitis (GPA) in a tertiary referral centre. Second, to compare the efficacy of RTX in granulomatous and vasculitic manifestations in GPA.

Patients and methods This study comprised a retrospective, standardised data collection from all patients who received RTX for refractory Wegener's granulomatosis from 2002 to 2010. Patients were assessed by a standardised interdisciplinary diagnostic procedure (including ear, nose and throat and ophthalmology assessment, MRI, immunodiagnostics, B-cell levels and Birmingham Vasculitis Activity Score) and were treated by standardised therapeutic regimens according to available evidence.

Results 59 patients received 75 cycles of RTX. 9.3% achieved complete remission. A response was documented in 61.3% (improvement in 52%, unchanged disease activity in 9.3%), 26.7% had refractory disease. Birmingham Vasculitis Activity Score, disease extent index, erythrocyte sedimentation rate, C-reactive protein and prednisolone demand decreased significantly. All patients achieved B-cell depletion. Granulomatous manifestations such as orbital granuloma and pachymeningitis were more frequently refractory to RTX than vasculitis or other granulomatous manifestations. Thus, for example, complete remission/improvement was found in 89.2% of patients with renal disease and in only 44.4% of those with orbital masses (p=0.003). The relapse rate was 44.4% after a median period of 13.5 months. Adverse events occurred in 29%, pneumonia in 15% and death in 3%.

Conclusion The overall response rate of refractory GPA to RTX was high (61.3% complete remission or improvement). Response rates of vasculitic manifestations were excellent; failure of response/progress was mostly due to granulomatous manifestations, especially orbital masses. Relapse rates were high (40%) despite maintenance treatment.

Statistics from

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.


The outcome for patients with granulomatosis with polyangiitis (GPA) has constantly improved1,,3 since the introduction of immunosuppressive treatment by Fauci et al4 and the implementation of treatment principles such as remission induction and maintenance on the basis of a number of controlled trials performed by the European Vasculitis Study Group (EUVAS).5 Yet, a considerable proportion of patients is resistant to standard treatment: refractory GPA was present in 15–20% of patients of large GPA cohorts.1,,3 Even in the localised stage of GPA, in which disease manifestations are restricted to the respiratory tract without any clinical signs of vasculitis (and granulomatous manifestations only), refractory disease is common and was reported in 16% of patients,6 7 showing that vasculitic manifestations and granulomatous inflammation follow a refractory course in a considerable proportion of patients.

There is almost no evidence from controlled trials to guide further treatment decisions when patients do not respond to standard remission induction in antineutrophil cytoplasmic antibody (ANCA)-associated vasculitides (AAV).5 Rituximab (RTX) has been shown equally efficient for standard remission induction in generalised AAV as with oral cyclophosphamide (Cyc).8 Observational studies investigating the efficacy of RTX in refractory AAV and/or GPA have been published in the past years,9,,23 most of which reported high rates of remission (>80%).23 Furthermore, it was found that relapses were common but that re-treatment was effective.23 However, there is continuing debate as to whether RTX may not be as effective in the treatment of granulomatous mass formation compared with vasculitic manifestations such as pulmonary capillaritis or glomerulonephritis. Additionally, whereas the mechanism of action of RTX seems obvious in vasculitis manifestations—namely, ANCA depletion, the mechanism remains unclear for granulomatous manifestations. An initial study reported a lack of efficacy in five of eight patients with granulomatous GPA manifestations,14 as did several other reports.9 10 16 In contrast to these reports, further studies on refractory ophthalmic manifestations and on refractory head and neck involvement showed a beneficial response in most patients.21 22 Moreover, some of the studies cannot be compared owing to different RTX dosage, treatment intervals and different regimens of concomitant immunosuppression. A large study is needed to examine the efficacy of RTX in granulomatous manifestations adequately. The largest report published is on 46 patients with GPA, which does not refer to the refractory disease manifestations in detail.23

The aims of this study were (1) to investigate the overall efficacy and safety of RTX in addition to standard treatment in refractory GPA and (2) to compare the efficacy of RTX for granulomatous and vasculitic manifestations in a large cohort at a tertiary referral centre.

Patients and methods

Patients and inclusion criteria

All patients who fulfilled the classification criteria of the American College of Rheumatology24 and the definition of the Chapel Hill Consensus Conference25 for Wegener's granulomatosis/GPA and who received RTX for refractory GPA in the period from the beginning of 2002 until 2010 were included in the study and followed up retrospectively. Refractory disease was defined according to the EULAR/EUVAS—namely, as unchanged or increased disease activity after 4 weeks of standard treatment in acute AAV or as lack of response after 6 weeks of treatment (<50% reduction of the Birmingham Vasculitis Activity Score (BVAS)) or chronic, persistent disease after ≥12 weeks of treatment.26

Clinical, laboratory assessment and treatment protocol

All patients were followed up by a strict interdisciplinary approach of our tertiary referral centre as previously described.3 5 6 All patients underwent a regular set of interdisciplinary clinical (including ophthalmology and ear, nose and throat assessment), technical (including MRI and high-resolution CT, if required), serological and immunological examinations before the introduction of RTX and at 4-monthly intervals after the first RTX infusion to document the effect of RTX on the disease activity as described earlier,3 5 unless complications required an earlier admission of the patient (eg, owing to infection).

RTX was given intravenously as four doses of 375 mg/m2 at weekly intervals (plus prednisolone 100 mg on the day of infusion) and conventional immunosuppression was continued after the start of RTX. When Cyc was given orally (2 mg/kg/day), this was continued until re-evaluation; patients receiving intravenous Cyc had three infusions (15–20 mg/kg body weight) at three weekly intervals in conjunction with RTX. Glucocorticoids (GCs) were tapered according to a standardised protocol (online supplementary material, table 1). Some patients received a second or third course or cycle of RTX if they relapsed and were refractory to standard treatment (figure 3, supplementary material).

Disease stages (localised, early systemic, generalised and severe) were defined according to EULAR/EUVAS.4 Orbital masses were counted as localised disease manifestation as they represent granulomatous disease manifestations that often originate from the ear, nose and throat tract.6 We differentiated between vasculitis manifestations and granulomatous manifestations (supplementary material, table 4). Disease activity was assessed by the BVAS,27 disease extent by the Disease Extent Index28 and chronic organ damage by the Vasculitis Damage Index.29 Remission (absence of disease activity for which stable maintenance immunosuppressive treatment may be continued, including prednisolone at ≤7.5 mg/day), response (≥50% reduction of disease activity assessed by the BVAS and absence of new manifestations) and relapse (reoccurrence of new onset of disease attributable to active vasculitis) were defined according to EULAR.26 For response, we further differentiated between improvement and stable disease in order to better evaluate treatment effects, especially with regard to granulomatous manifestations. We defined stable disease as an unchanged disease manifestation which equates to a 50% reduction of the BVAS upon re-evaluation (eg, in orbital granulomatous disease) and is therefore consistent with the current EULAR/EUVAS definition of response. A patient was considered refractory to RTX if neither remission nor a response had occurred 4 months after the introduction of RTX. Four months after the first RTX infusion, assessments of the overall efficacy of RTX were made. Furthermore, the organ involvements were assessed separately for their response to RTX (supplementary material, tables 4 and 5). Maintenance treatment after successful treatment with RTX is described in the supplementary material.

B-cell count was determined by flow cytometry: B-cell depletion was defined as a cell count <0.02×102/l and B-cell repopulation by a rise in cell count ≥0.02×102/l following depletion. ANCA were assessed by indirect immunofluorescence and direct enzyme-linked immunosorbent assay (ELISA) as described earlier.14 Patients received standardised education about RTX, including possible side effects and complications. Side effects and complications were assessed by a standardised interview. Serious adverse events (SAEs) were defined as events requiring admission to hospital, intravenous treatment, life-threatening situations or death.

Statistical analysis

Statistical analysis was performed using SPSS, version 18 (SPSS, Chicago). Changes in variables were compared by Wilcoxon and Mann–Whitney test; χ2 test was used to compare disease activity before and after RTX administration. Relapses were analysed by Kaplan–Meier survival analysis. p Values <0.05 were considered statistically significant.


Patient characteristics

Sixty patientswere included in the study, 59% were male (n=35) with a median age of 54 years (range 22–76) and a median disease duration of 37 months (range 3–211) at first RTX administration. Fifty (85%) had generalised disease. All patients had received Cyc previously (table 1, cumulative treatment). Directly before RTX treatment, 78.3%of patients (n=47) were receiving treatment with Cyc, eight of whom additionally received a tumour necrosis factor (TNF) α antagonist; the median cumulative Cyc dosage was 27.75 g (range 4.5–450). Further details are shown in table 1.

Table 1

Baseline characteristics of patients with refractory disease before the administration of RTX

Refractory disease manifestations at RTX start

Manifestations such as orbital masses (n=27), active renal disease (n=26, n=12 with non-impaired renal function, but erythrocyturia and proteinuria ≥1 g/day, n=14 with impaired renal function defined as creatinine clearance ≤60 ml/min + erythrocyturia and/or proteinuria), pachymeningitis (n=12), pulmonary masses (n=12) and alveolar haemorrhage (n=12) represented the most common indications for RTX. Thirty-seven patients had more than one refractory manifestation before the start of RTX (figure 1; supplementary material, table 6). Manifestations such as ground-glass infiltrates, neutrophil alveolitis and arthritis were accompanied by further refractory disease manifestations.

Figure 1

Refractory organ involvement/activity before rituximab (RTX) treatment. Thirty-seven patients had more than one refractory organ involvement before the start of RTX granulomatous sinusitis and bony destruction.

Treatment protocols and follow-up

One patient had an infusion reaction during the first administration of RTX and was therefore excluded from further analysis. Fifty-nine patients received one cycle of RTX, 12 patients received two and four patients three cycles of RTX (table 1). All patients were additionally treated with GCs (tables 2 and 3 supplementary material). No additional conventional immunosuppressant agent was administered in 5.3%; during all other cycles additional medium or highly potent immunosuppression was given. Co-medication with Cyc was given most frequently (54.7%) (figure 2). All patients were followed up for at least 4 months after the last administration of RTX, 36 patients were followed up for longer than 4 months and analysed for relapse frequency. The median follow-up period of the whole cohort was 7 months (4–58 months); the 36 patients who had entered remission/response were followed up for 13.5 months (3–54).

Figure 2

Co-medication during rituximab treatment in percentage per patient and cycle. All patients additionally received glucocorticoids. AZA, azathioprine; Cyc, cyclophosphamide; Lef, leflunomide; MMF, mycophenolate mofetil; MTX, methotrexate.

Table 2

Outcome after rituximab (RTX)

Figure 3

Efficacy of rituximab (RTX) by organ involvement given for the most frequent organ involvements (in at least four patients or more). For efficacy of RTX in organ involvements that were less common see online supplementary material.

Table 3

Effect of RTX on overall disease activity and disease extent assessed by BVAS and DEI, on laboratory parameters and prednisolone dose

Overall efficacy of RTX (clinical, laboratory)

Regarding overall efficacy of RTX, 9.3% of patients (n=7) achieved a complete remission of their disease manifestations; 52% (n=39) had an improvement/response and in 9.3% of patients (n=7), a stabilisation/response was reached. 26.7% (n=20) were refractory which was consistent with disease progression (table 2). Thirty-two patients received RTX for more than one refractory organ involvement. RTX usually had the same effect on organ involvements (table 6, supplementary material).

Efficacy of RTX on single organs

The effect of RTX on the most common organ manifestations is displayed in figure 3: 27 patients presented orbital masses, none of them achieved a complete remission (0%), 18 patients showed a response (66.7%) (12 improved (44.4%), six (22.2%) had stable disease) and nine patients (33.3%) were refractory (with progressive disease). Patients with pachymeningitis (n=12) had a similar rate of refractory disease (n=4, 33.3%) as patients with orbital disease, one (8.3%) reached a complete remission, six had a response (five (41.7%) with improvement, one (8.3%) with stable disease). Twelve patients were treated for refractory pulmonary masses with good response rates: two (16.7%) had a complete remission, eight (66.7%) improved and two (16.7%) had refractory disease (disease progression); similarly, refractory active renal disease and alveolar haemorrhage responded well: of 26 patients with glomerulonephritis, nine (34.6%) had a complete remission, 12 (46.2%) improved and four (15.4%) had refractory disease; of 12 patients with alveolar haemorrhage, eight (66.7%) had a complete remission, two (16.7%) improved and one patient (8.3%) had refractory disease. One patient presenting with refractory renal disease and one with refractory alveolar haemorrhage died. The efficacy of RTX on polyneuropathy/mononeuritis (PNP) and arthritis was also satisfactory: four patients with PNP achieved complete remission (n=1, 25%) or response/improvement (n=3, 75%); all patients with arthritis (n=6) had a complete remission (100%).

After RTX treatment, there was a significant decline of (refractory) disease activity in all organ systems/manifestations (refractory manifestation before RTX vs refractory/unchanged disease activity after RTX, p<0.05) except for granulomatous sinusitis (p>0.05), which was probably owing to the small number of patients with granulomatous sinusitis (n=3). When the response of all granulomatous disease manifestations (taken together) were compared with all vasculitic disease manifestations, the proportion of complete remissions/improvement versus unchanged activity/refractory disease was significantly higher for vasculitic manifestations (90.6% and 9.4% for vasculitic vs 58.2% and 41.8%, respectively, for granulomatous manifestations, p<0.05, table 4 supplementary material). Furthermore, the proportion of complete remissions/improvement versus unchanged/refractory disease was lower in some of the granulomatous manifestations than in some of the vasculitic manifestations, which reached statistical significance for the comparison of orbital masses with alveolar haemorrhage and glomerulonephritis (p=0.003 and p=0,003, respectively). The comparison of response rates of pachymeningitis and pulmonary masses versus alveolar haemorrhage and glomerulonephritis was not significantly different (p=0.15 and p=0.14 for pachymeningitis versus alveolar haemorrhage and glomerulonephritis and p=0.94 and p=0.67 for pulmonary masses, respectively). An overview of the efficacy of RTX on all disease manifestations is provided in the supplementary material (supplementary tables 1 and 2). BVAS, Disease Extent Index, laboratory parameters such as erythrocyte sedimentation rate, C-reactive protein, ANCA titre and required dose of prednisolone declined significantly after RTX treatment (table 3). B-cell count also declined, although this was not significant (see figure 1 supplementary material).

Relapses after successful induction of remission/response

For 36 patients, a follow-up after successful induction of remission/response was available (median follow-up period 13.5 months, range 3–54 months). Sixteen of 36 patients (44.4%) relapsed, 20 patients (56.6%) remained in remission until the end of follow-up. The median time to relapse was 13.5 months (3–54) in patients who had a relapse (see supplementary files for Kaplan–Meier relapse-free survival analysis, figure 2) and 12 months (3–27) in patients who did not have a relapse. There were no differences in disease stage (localised versus systemic), ANCA status or organ system involvement between patients who relapsed and patients who did not (data not shown). Furthermore, in patients who relapsed, there was no significant difference of ANCA titre at remission compared with the time point of relapse (p=0.673, data not shown). Twelve patients received a second cycle and four patients a third cycle of RTX with similar response rates as with the first cycle (supplementary material, figure 3).

Adverse events and complications of RTX

Twenty-four SAEs were documented during or after the administration of 75 cycles of RTX (table 2). Infection was the most common SAE and occurred in 28.9%. Pneumonia was most prevalent (and occurred in 11 cases, 14.5%). IgG/IgM levels declined significantly after RTX but there was no significant association with low IgG/IgM and infection. Two deaths occurred (supplementary material).


We confirm the finding of a good overall efficacy of RTX for the treatment of refractory GPA in this uncontrolled retrospective study (61.3% complete remission or improvement, 9.3% stable disease, 26.7% refractory disease). Refractory disease activity declined in all organ systems. However, the response rates in our study are lower than with other large uncontrolled trials: Jones et al23 reported complete remission rates of 75% and treatment failures of only 2%. This may be because of the high proportion of granulomatous manifestations in our study, some of which responded less well to RTX than the vasculitic manifestations: complete remission/improvement was documented in 89.2% of renal disease and in 80.8% of patients with alveolar haemorrhage, whereas in orbital masses and pachymeningitis, a complete remission or improvement was found in 44.4% and 49.9%, respectively. Interestingly, remission/response rates in pulmonary masses were higher than with the other granulomatous manifestations. The strength of the study lies in its strict interdisciplinary and standardised diagnostic approach; its weakness is that it is retrospective and that there was some variation in the GC dosing. Furthermore, we cannot definitely rule out the possibility that some of the responders with orbital masses might have been in complete remission as the orbital mass might have turned into a scar and, in turn, that a persistent mass might indeed be refractory with an increasing inflammatory activity but unchanged diameter. Yet, all patients with orbital masses termed refractory had further enlargement of the masses so that the percentage of refractory orbital masses is a correct observation. Nevertheless, it would be desirable to develop imaging strategies that differentiate between scar and active inflammation.

The reason for lower treatment response rates, especially of orbital masses, is not clear. Biopsy specimens demonstrated a classic cell composition of orbital lesions as found in other typical granulomatous lesions of GPA,30 but fibrosis and an onion-like arrangement of collagen are also well-documented features,31 32 which are usually not found in other granulomatous Wegener's granulomatosis lesions (such as pulmonary granuloma). It may be speculated that a different inflammatory environment of orbital lesions leading to sustained granuloma formation and induction of fibrous tissue may account for the relative resistance towards immunosuppressant agents, including RTX. Moreover, metabolism may be slower and/or drugs may not spread as easily to remote places such as the orbit surrounded by fibrous tissue. Debulking operations in conjunction with immunosuppression to treat orbital masses have been reported to be successful,32 33 yet this procedure bears the risk of structural damage to the eye. For the future, it will be crucial to understand the pathophysiology of granulomatous masses in order to develop targeted treatments to tackle the inflammatory process and prevent fibrosis.

Despite maintenance with conventional immunosuppression, relapse rates were high (44.1% after a median time of 13.5 months). Similar relapse rates have been documented by others with similar time intervals to relapse.23 Repeated courses of RTX were successful in four patients who did not respond to the first course,22 but this observation requires validation in a larger cohort. We found no differences in disease stage, ANCA status or organ system involvement (granulomatous vs vasculitic) between patients who relapsed and patients who did not. Similarly to other studies23 re-treatment with RTX upon relapse was successful in the majority of patients. Pre-emptive re-treatment may be of benefit23 and be a future option to reduce the burden of disease in patients resistant to standard treatment and prone to relapse, but long-term data on the safety of RTX are still lacking. Moreover, RTX plus GC should be assessed for refractory disease without additional conventional immunosuppression as this was successful in induction of standard remission.8

Infections were the most common serious adverse event which occurred in nearly 30%. One of the two deaths was due to infection. The infection rate was similar to that found by other studies of RTX in refractory GPA.23 Remarkably, the infection rate of our study and others,23 34 in which patients received RTX in conjunction with conventional medium to highly potent immunosuppression plus GC, was similar to that of patients receiving standard remission induction with Cyc plus GC,4 35 suggesting that concomitant RTX treatment is not associated with an additional increased risk of infections in the short term.

In summary, this study suggests a good overall efficacy or RTX in refractory GPA. Efficacy was less pronounced than with other uncontrolled studies of RTX in refractory AAV,23 which is probably owing to the high proportion of granulomatous manifestations in this study: treatment failures mostly occurred in granulomatous manifestations, in particular in orbital masses and pachymeningitis. A controlled trial is needed to further evaluate the principle of B-cell-depleting treatment with RTX for refractory granulomatous AAV.


Supplementary materials


  • Competing interests None.

  • Ethics approval Ethics approval was obtained from University of Luebeck.

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