Abstract
Objective. To evaluate the efficacy compared to the relapse risk and tolerance of systematic rituximab (RTX) infusions as maintenance therapy for patients with granulomatosis with polyangiitis (GPA) or microscopic polyangiitis (MPA), who entered remission taking conventional immunosuppressants or RTX.
Methods. A retrospective study of the main clinical characteristics, outcomes, and RTX tolerance of patients who had received ≥ 2 RTX maintenance infusions in our center, regardless of induction regimen, between 2003 and 2010.
Results. We identified 28 patients [4 MPA and 24 GPA; median age 55.5 yrs (range 18–78); 17 (60%) males] who received a median of 4 (range 2–10) RTX maintenance infusions, with median followup of 38 months (range 21–97) since diagnosis or last flare. None experienced a RTX infusion-related adverse event; 15 patients (among the 21 with available data) had hypogammaglobulinemia (predominantly IgM) prior to their last RTX maintenance infusion; 3 had infectious events (1 cutaneous abscess, 1 otitis, 1 fatal H1N1 flu). Two patients suffered pulmonary relapses shortly before a planned RTX maintenance infusion (both had increased antineutrophil cytoplasmic antibody levels and 1 had CD19+ lymphocyte reconstitution).
Conclusion. Rituximab maintenance therapy was well tolerated but did not completely prevent relapses and persistent “grumbling” disease. These preliminary results remain to be confirmed by a randomized controlled trial currently in progress.
- GRANULOMATOSIS WITH POLYANGIITIS
- MICROSCOPIC POLYANGIITIS
- RITUXIMAB
- MAINTENANCE THERAPY
- ANTINEUTROPHIL CYTOPLASM ANTIBODY-ASSOCIATED VASCULITIDES
Rituximab (RTX) has reportedly achieved successful remission of granulomatosis with polyangiitis (GPA) and microscopic polyangiitis (MPA) in 80%–100% of patients with severe and/or refractory diseases1,2,3. Recent results of the randomized RAVE4 and RITUXVAS5 trials demonstrated that an RTX-based regimen was as effective and safe as conventional cyclophosphamide (CYC)-based regimens to induce remission in these patients4,5. Notably, in both those trials, no maintenance therapy was prescribed after RTX-induced remission.
However, results of previous retrospective studies1,2, published after those randomized trials were started, indicated that, in populations of refractory/relapsing patients, relapses after RTX-induced remission could occur in more than half the patients2, especially within the 9–12 months after the last RTX infusion, and independently of antineutrophil cytoplasmic antibody (ANCA) status or CD19+ lymphocyte immune reconstitution. We hypothesized that maintenance therapy was required after RTX-based induction therapy and that regular RTX infusions might represent an effective maintenance regimen. No recommendations are available yet concerning the need for maintenance therapy for these patients and, if so, with which agent. We report our retrospective analysis of 28 patients with ANCA-associated vasculitis (AAV) who received preemptive RTX infusions as maintenance.
MATERIALS AND METHODS
Patient selection
For this study, patients had to have GPA or MPA satisfying American College of Rheumatology and/or Chapel Hill consensus conference criteria6,7; to have received ≥ 2 RTX maintenance infusions (minimum followup of 12 months), regardless of their induction regimen; and to be followed (or been referred, at least once), between 2003 and 2010, in the Internal Medicine Department of Cochin Hospital, Paris. Patients may have subsequently received some or all of their RTX infusions in another hospital (in Besançon, Caen, Colombes, Paris, or Tarbes).
Study measures
We retrospectively analyzed patients’ main characteristics (sex, diagnosis, age at diagnosis), clinical manifestations (duration of AAV since diagnosis, disease activity at the last flare), and treatments preceding and associated with RTX maintenance. We recorded why RTX maintenance was chosen (i.e., after RTX induction or because of contraindication and/or inefficacy of other conventional maintenance agents), the regimen (number, dose, timing), the efficacy of RTX at maintaining remission, tolerance for RTX, and patient outcomes. For the patients not followed entirely at our institution, information was updated by contacting the referral physicians (NMR, AG, AZ, AP, BB). Only the number of RTX maintenance infusions was counted. When available, we noted ANCA status and titer by immunofluorescence and enzyme-linked immunosorbent assay (ELISA; normal proteinase-3/myeloperoxidase values < 20 IU/ml), circulating CD19+ B lymphocyte counts by flow cytometry (normal range > 2%), and immunoglobulin levels (normal range 7.5–12 g/l) at diagnosis and during followup. Normal ranges of IgG and IgM were 6.8–15 g/l and 0.40–2.20 g/l, respectively. The normal CD19+ lymphocyte count exceeds 2% of the total lymphocyte population. CD19+ depletion is defined as below this normal value, and CD19+ reconstitution as the return of the CD19+ level towards normal.
Definitions
AAV activity was assessed using the Birmingham Vasculitis Activity Score (BVAS, version 3)8 with complete remission defined as BVAS = 0 for ≥ 3 consecutive months (i.e., no active or persistent disease). The p-BVAS (for persistent BVAS) rated disease status when patients had some BVAS items that were not new or worsening (BVAS = 0) but persistent (regressing or unchanged) for ≥ 3 months, corresponding to “partial remission” (p-BVAS > 0) and represented by mainly “grumbling” ear-nose-throat (ENT) disease and/or scarring lung nodule. However, irreversible damage, like renal insufficiency or sustained, minor, and/or nonworsening sensory neuropathy, was recorded, when present, but did not exclude complete remission. Relapse meant the recurrence, worsening, or first appearance of ≥ 1 BVAS items attributable to active AAV after entering remission.
Disease duration was defined as the time from diagnosis of AAV to the onset of the last flare (0 for newly diagnosed patients; > 0 for relapsers). Followup duration was calculated from the start of induction treatment for the last AAV flare (either diagnosis or relapse for the relapsers) and last patient assessment visit.
Statistical analyses
Quantitative variables (mean ± SD) were compared using Wilcoxon tests for immunoglobulin levels. ANCA levels at last relapse, first RTX maintenance infusion, and last consultation were compared using one-way analysis of variance and pair-wise 2-sample t tests. All analyses were carried out with SPSS 16 software (SPSS Inc., Chicago, IL, USA). Significance was defined as p < 0.05.
RESULTS
Patient characteristics
This analysis concerned 28 patients (Table 1 and Appendix): 4 with MPA and 24 with GPA; 17 (60%) males; median age 40.5 years (range 13–70 yrs) at diagnosis and 50.5 years (range 19–78 yrs) at induction treatment for last flare. AAV disease was newly diagnosed in 3 patients, whereas 25 had relapsed (10 had had ≥ 2 relapses). Median disease duration (from diagnosis to last flare) was 84 months (range 15–222 mo). Median BVAS was 15 (range 6–26) before induction [at diagnosis for the former (n = 3) and at last flare for the others], with 19 patients having ENT involvement, 18 lung, 11 arthralgias, 11 fever, 9 kidney, 6 eyes, and/or 3 peripheral or central neuropathy. Lung manifestations were characterized by nodules for 11 patients, interstitial syndrome for 4, or both for 3. Eye manifestations were scleritis (n = 3) and, in 1 patient each, retrobulbar optic neuritis, lacrimal gland inflammation and scleritis, and retroorbital mass exophthalmia.
Induction and RTX maintenance regimens
All the patients had received CYC at least once in the past, with a cumulative CYC dose of 48 g/patient (range 5–250 g) before RTX maintenance therapy. Induction therapy before RTX-based maintenance included corticosteroids combined with RTX for 21 patients, with CYC for 5, with intravenous immunoglobulins for 1, or with intravenous immunoglobulins and methotrexate (MTX) for the remainder (Table 1). RTX induction regimens consisted of 2 infusions of 1 g each, given 2 weeks apart (for 4 patients), or 4 infusions of 375 mg/m2 each, given once weekly (for 17 patients).
At the beginning of maintenance therapy, 6 patients were in complete remission (BVAS and p-BVAS = 0); 12 were in partial remission, with grumbling ENT disease (n = 12) and/or persistent lung nodules (n = 3); 10 were in complete remission with irreversible damage (including peripheral sensory neuropathy or renal insufficiency) from their last flare (n = 9) and/or an earlier flare (n = 3; Table 2).
RTX maintenance therapy was chosen for the 21 who had received RTX induction therapy. For the other patients, RTX maintenance was chosen for the following reasons: (1) side effects/intolerance of more conventional maintenance treatments for 2 patients [azathioprine (AZA)-related hepatitis and myelotoxicity; mycophenolate mofetil (MMF)-associated gastrointestinal intolerance]; (2) persistent and grumbling vasculitis manifestations after > 4 years of AZA or MMF in 1 patient each; or (3) previous relapse(s) under AZA therapy for 2 patients or (4) renal insufficiency for 1.
Patients had received 4 (range 2–10) RTX maintenance infusions over 38 months (range 21–97 mo) of followup since their diagnoses (in 3 patients) and since the last flare for the remaining 25. RTX doses were 375 mg/m2 biannually for 13 patients, 1 g biannually for 4, 1 g annually for 3, and in different regimens for 8 (Table 2). Combined treatments at the time of the first RTX maintenance infusion included corticosteroids [median dose 5 mg/day (range 2–20)] for 23 (82%) patients and/or other immunosuppressants for 14 (50%; AZA for 6, MMF for 5, leflunomide for 1, and/or MTX for 4). Immunosuppressants were subsequently stopped for 9 patients, 8 months (range 2–26 mo) after their first RTX maintenance infusion. At last visit, all patients were still receiving RTX maintenance, 16 (57%) patients still took corticosteroids [median dose 5 mg/day (range 4–10)], and 7 (25%) received other immunosuppressant(s) (AZA for 3, MMF for 2, MTX for 2, and/or leflunomide for 1 patient).
Outcomes of AAV disease
Two patients had pulmonary relapses. A 65-year-old woman with MPA developed alveolar hemorrhage 6 months after RTX maintenance infusion, given biannually. Intravenous CYC and plasma exchange generated no response and she again received RTX induction treatment, before switching to RTX maintenance again (she died of H1N1 flu infection 16 months later). A 62-year-old man with GPA developed new lung nodules 11 months after his sixth RTX maintenance infusion (375 mg/m2 every 6 months, then annually for the last 2). A new RTX induction regimen achieved remission, followed by RTX maintenance.
At last evaluation, 6 patients were in complete remission; 9 patients had partial remissions: 7 with persistent grumbling ENT manifestations, 2 with lung nodules; and 11 patients were in complete remission with irreversible damage: 6 with endstage renal disease, 1 hypoacousia, 1 subglottic stenosis, 1 pulmonary fibrosis, and 2 had residual peripheral neuropathy.
ANCA and CD19+ lymphocyte levels
CD19+ B lymphocytes were depleted before each of the RTX maintenance infusions in 12 of the 24 patients with available data (including the 2 who relapsed). CD19+ lymphocytes were never fully depleted before RTX infusions in 4 patients, and 8 had rising CD19+ lymphocyte levels at at least 1 determination before RTX infusion, but none of them subsequently relapsed. However, one relapser had CD19+ lymphocyte reconstitution at the time of relapse, 11 months after his last RTX infusion (this was not measured in the other patient).
Five patients were persistently ANCA-negative at time of diagnosis. At last flare, 5 of the remaining patients were ANCA-negative by ELISA (level < 20 IU/ml) and 18 were ANCA-positive. At the time of their first RTX maintenance infusion, 11 of these latter 18 were persistently ANCA-positive (including the 2 relapsers) and 6 remained so at their last evaluation. Mean titers for all 19 proteinase-3 ANCA-positive patients progressively declined, from the last flare to the first RTX infusion and last assessment (84.4 ± 93, 33.4 ± 58.6, and 14.3 ± 34.6 IU/ml, respectively; ANOVA, p < 0.0001), but they increased before at least 1 of the RTX maintenance infusions in 8 other patients. ANCA titers at relapse were higher than previous measurements in the 2 patients who relapsed (from 81 to 140 IU/ml and 157 to 240 IU/ml, by ELISA).
RTX tolerance/safety
No immediate RTX infusion-related reaction or serum disease was observed. No late-onset neutropenia was detected during followup. Among the 18 patients with available total gammaglobulin levels at 2 different times (i.e., just before starting RTX induction/maintenance and before the last RTX maintenance infusion), 11 were hypogammaglobulinemic before starting RTX induction/maintenance. Three patients developed hypogammaglobulinemia taking RTX, and total gammaglobulin levels decreased in 14 patients and normalized in 2 patients (Figure 1). Mean gammaglobulinemia was significantly decreased before starting and after RTX administration, i.e., at last evaluation (7.78 ± 2.5 g/l vs 6.97 ± 1.9 g/l, respectively; Wilcoxon test; p = 0.041). IgG and IgM levels were under the normal range at last RTX maintenance infusion in, respectively, 12 and 15 of the 21 patients with available data.
Three patients developed infections: 1 cutaneous abscess, 1 otitis media that resolved rapidly under antibiotics, and 1 fatal H1N1 flu infection. The latter patient had normal immunoglobulin levels at the time of her last RTX infusion 6 months before she died (not measured when she developed H1N1 flu), and the 2 others were hypogammaglobulinemic.
DISCUSSION
Our results suggest that RTX effectively maintained remissions of GPA and MPA, with a good global safety profile. The results confirm those reported by Jones, et al2 and Rhee, et al9, who showed that patients who entered remission with RTX may need maintenance therapy, possibly based on RTX itself.
Among the 65 patients with refractory vasculitis analyzed retrospectively by Jones, et al2, none of the 15 who had received preemptive 1-g RTX infusions biannually relapsed, in contrast to 57% of those not prescribed any maintenance therapy, with median followup at 11 months (range 5–23). Rhee, et al9 treated 39 patients with RTX maintenance (1 g every 4 months), and among 20 patients at 24 months of followup, 3 had relapsed. With our longer duration of followup, 2 (7%) relapses have occurred, at 12 and 48 months after induction treatment. For comparison, at similar duration of followup, the WEGENT trial relapse rate was 35% for patients who had received AZA or MTX maintenance, after having achieved remission with corticosteroids plus CYC10. However, half our study patients were receiving additional concomitant immunosuppressive therapies during followup, making the interpretation that the low relapse rate was due to RTX less certain.
Based on our results and to further examine the potential place of RTX maintenance therapy compared to more conventional agents (e.g., AZA), we have started a prospective open-label trial (MAINRITSAN; ClinicalTrials.gov identifier NCT00748644) for patients with AAV who achieved remission with conventional CYC-based induction therapy10,11. Patients are enrolled at remission and randomized to receive maintenance therapy with either RTX (1 systematic infusion every 6 months for 18 months) or AZA (for 21 months).
Early identification of potential relapsers could help adjust therapy individually, thereby also limiting the risk of prolonged treatment toxicity. Unfortunately, to date, no reliable marker exists to do so12,13,14. Although our 2 relapsers had persistent ANCA positivity and increases of ANCA level at relapse, ANCA status and level do not seem sufficiently predictive of relapse to guide retreatment. Similarly, the CD19+ lymphocyte count does not seem to be a reliable predictor of relapse. Jones, et al2 found no clear association between elevated ANCA levels and relapse. Rhee, et al9 reported complete CD19+ lymphocyte depletion in their 3 patients whose ANCA levels were positive but stable before they relapsed.
RTX, like corticosteroids and other immunosuppressants, carries a risk of infection, although the relationship between RTX maintenance and the 3 infectious events we observed remains hypothetical. Two of these 3 patients who developed infections were hypogammaglobulinemic. Hypogammaglobulinemia is common in patients treated with RTX, but most of our patients had also previously and/or concomitantly received other drugs such as CYC and corticosteroids, possibly further decreasing their gammaglobulin levels. Whether RTX associated hypogammaglobulinemia is associated with a higher risk of infections remains to be determined. Jones, et al2 found no clear association between infections and gammaglobulin levels. In patients with rheumatoid arthritis, gammaglobulin levels before RTX appeared to be more closely associated with the risk of infection than during or after RTX therapy15. Longterm RTX safety in patients with AAV requires further investigation, especially if regular maintenance infusions are given.
We acknowledge that our study had limitations, including its retrospective design, thus the heterogeneity of induction treatments and RTX maintenance regimens, for which protocols had not been fully developed until recently. This preliminary study does not allow conclusions about the precise role of RTX in maintenance regimens for AAV or an optimal RTX regimen. The heterogeneity of treatments given in this exploratory and retrospective analysis precludes definitive conclusions, but no “red flag” was raised concerning safety of RTX or relapse rate. Moreover, not all patients had regular or systematic monitoring of ANCA levels, white cell counts, and immunoglobulins, which hampers understanding of the safety and tolerance of RTX maintenance therapy. However, no recommendations have been made concerning the need for maintenance therapy (or not) for those patients with AAV achieving remission with RTX, and if so, with which drug and regimen. Longer followup of patients in the RAVE4 and RITUXVAS5 trials who entered RTX-induced remission should provide some insight into these issues.
Our results and other available data suggest that RTX could probably be used as maintenance treatment for AAV. We must now determine whether RTX is as effective and safe as other maintenance agents such as AZA10, and the optimal RTX maintenance regimen.
Acknowledgment
We thank Janet Jacobson for editorial assistance, and all French Vasculitis Study Group members who followed the patients and/or included them in the French Vasculitis Study Group database.
APPENDIX
- Accepted for publication September 7, 2011.