Objective To compare individually tailored, based on trimestrial biological parameter monitoring, to fixed-schedule rituximab reinfusion for remission maintenance of antineutrophil cytoplasm antibody (ANCA)-associated vasculitides (AAVs).
Methods Patients with newly diagnosed or relapsing granulomatosis with polyangiitis (GPA) or microscopic polyangiitis (MPA) in complete remission after induction therapy were included in an open-label, multicentre, randomised controlled trial. All tailored-arm patients received a 500 mg rituximab infusion at randomisation, with rituximab reinfusion only when CD19+B lymphocytes or ANCA had reappeared or ANCA titre rose markedly based on trimestrial testing until month 18. Controls received a fixed 500 mg rituximab infusion on days 0 and 14 postrandomisation, then 6, 12 and 18 months after the first infusion. The primary endpoint was the number of relapses (new or reappearing symptom(s) or worsening disease with Birmingham Vasculitis Activity Score (BVAS)>0) at month 28 evaluated by an independent Adjudication Committee blinded to treatment group.
Results Among the 162 patients (mean age: 60 years; 42% women) included, 117 (72.2%) had GPA and 45 (27.8%) had MPA. Preinclusion induction therapy included cyclophosphamide for 100 (61.7%), rituximab for 61 (37.6%) and methotrexate for 1 (0.6%). At month 28, 21 patients had suffered 22 relapses: 14/81 (17.3%) in 13 tailored-infusion recipients and 8/81 (9.9%) in 8 fixed-schedule patients (p=0.22). The tailored-infusion versus fixed-schedule group, respectively, received 248 vs 381 infusions, with medians (IQR) of 3 (2–4) vs 5 (5–5) administrations.
Conclusion AAV relapse rates did not differ significantly between individually tailored and fixed-schedule rituximab regimens. Individually tailored-arm patients received fewer rituximab infusions.
Trial registration number NCT01731561; Results.
- ANCA vasculitis
- CD19+ B lymphocytes
- granulomatosis with polyangiitis
- microscopic polyangiitis
Statistics from Altmetric.com
Standard induction-remission treatment of antineutrophil cytoplasm antibody (ANCA)-associated vasculitides (AAVs) combines glucocorticoids and cyclophosphamide or rituximab, a chimeric murine/human monoclonal IgG1 antibody directed against CD20, a specific B-cell antigen.1 Rituximab was shown to be non-inferior to cyclophosphamide.2 3 Remission maintenance with immunosuppressants, for example, azathioprine or methotrexate, is the conventional therapeutic approach.4 MAINRITSAN-trial results demonstrated rituximab superiority (500 mg on days 0 and 14, then at months 6, 12 and 18) to azathioprine to maintain remission.5 In that trial, at month 28, only 5% of rituximab recipients versus 29% taking azathioprine had experienced a major relapse.
Neither ANCA-positivity nor ANCA-titre change on conventional immunosuppressants is considered a reliable relapse predictor.6–8 Nevertheless, ANCA reappearance or titre increase, mainly anti-proteinase-3 (PR3), in patients in remission is frequently associated with relapses,7 especially those given rituximab.9 ANCA parameters are not recommended for monitoring treatment1 10 and circulating B-cell detection is not a good predictor of AAV relapse.11 However, when B cells are undetectable and ANCA remain negative, relapses are rare.11
The present trial, MAINRITSAN2, was undertaken to evaluate ANCA and circulating CD19+ B cells as indicators to reinfuse rituximab to maintain remission. To do so, an individually tailored rituximab regimen, adapted to ANCA-positivity or ANCA-titre change and/or circulating CD19+ B cell repopulation, was compared with fixed-schedule rituximab infusions, previously shown to maintain remission5 in patients with granulomatosis with polyangiitis (GPA) or microscopic polyangiitis (MPA), who were in complete remission at the time of inclusion.
Eligible patients >18 years old had newly diagnosed or relapsing GPA or MPA, as defined by the Chapel Hill Consensus nomenclature.12 They had to be in complete remission after induction therapy, combining glucocorticoids and cyclophosphamide, rituximab or methotrexate (as decided by each investigator), in accordance with French and international recommendations. Birmingham Vasculitis Activity Score V.3 (BVAS) of 0 (score range: 0–63, with higher scores indicating more active disease) defined complete remission.13 The following exclusion criteria were applied: another systemic vasculitis; induction with an agent not recommended; active disease; incapacity or refusal to understand or sign the informed consent form; non-compliance; allergy to the study medication; pregnancy; breastfeeding; human immunodeficiency, hepatitis B or C virus infection; severe infection declared during the 3 months before randomisation; cancer or malignant blood disease diagnosed during the 5 years preceding vasculitis diagnosis; participation in another clinical research protocol during the 4 weeks before inclusion; any clinical or psychiatric disorder that could expose the patient to a greater risk of an adverse event (AE) or could prevent treatment administration and patient follow-up according to the protocol; severe immunosuppression; administration of live vaccine during the 4 weeks before inclusion; severe chronic obstructive pulmonary diseases (maximum expiratory volume <50% or dyspnoea grade III); chronic heart failure (dyspnoea NYHA III or IV); history of recent acute coronary syndrome unrelated to vasculitis; patients not enrolled in the French national health insurance.
All patients provided written informed consent.
This trial, Maintenance of Remission using Rituximab in Systemic ANCA-associated Vasculitis-2 (MAINRITSAN2), is an open-label, pragmatic, multicentre, randomised controlled trial, with evaluation of the primary outcome by an independent Adjudication Committee, comprising three vasculitis experts (one each specialised in rheumatology (X Puéchal), internal medicine (O Lidove) or both (M Gayraud)), blinded to the treatment arm and circulating CD19+ B cell counts. The two coprincipal investigators (PC, LG) designed the trial and drafted the manuscript, with appropriate input from coauthors and other-site investigators. The trial was funded by the Programme Hospitalier de Recherche Clinique of the French Ministry of Health (PHRC National 2011 AOM11145). The Ethics Committee (Comité de Protection des Personnes Île-de-France 1 (Paris)) approved the study, which received legal, monitoring and administrative management support from the Assistance Publique–Hôpitaux de Paris. Hoffmann-La Roche provided rituximab for the study but was not involved in or consulted about the study design and did not have access to the data.
Patients were randomised at a 1:1 ratio to receive maintenance therapy with either an ‘individually tailored’ (according to laboratory findings every 3 months) or ‘fixed-schedule’ (control) rituximab regimen within 1 month after completing induction treatment, if they had received cyclophosphamide or methotrexate, or 4–6 months after the last rituximab infusion, if it had been used to obtain remission. An independent statistician provided the computer-generated randomisation sequence, stratified by newly diagnosed or relapsing AAV. Randomisation was centralised through electronic case-report forms (eCRF) to assure allocation concealment.
Treatment allocation was known by patients and clinicians. The Adjudication Committee that evaluated the primary endpoint was blinded to the treatment arm and circulating CD19+ B cell count.
Tailored-infusion-arm patients always received 500 mg of rituximab at randomisation; then ANCA and CD19+ B lymphocytes were assessed every 3 months. Another 500 mg were infused when ANCA status differed from the previous control (ie, reappearance after being negative, indirect immunofluorescence-determined ≥2-dilution–titre increase and/or at least doubled ELISA PR3 or myeloperoxidase (MPO) arbitrary units) or CD19+ B cell counts exceeded 0/mm3. That algorithm, implemented in the eCRF, specified rituximab reinfusion when the CD19+ B cell count and/or ANCA changes were documented. The last rituximab infusion could be given at month 18.
The control group received the MAINRITSAN trial regimen5: 500 mg rituximab infusion on days 0 and 14 postrandomisation and at months 6, 12, 18 after the first infusion.
Premedication before all rituximab infusions comprised intravenous methylprednisolone (100 mg), dexchlorpheniramine (5 mg) and acetaminophen (1000 mg).
Most patients were still taking low-dose prednisone at randomisation that was tapered and stopped or could be maintained at 5 mg/day at the discretion of each site investigator.
All patients received Pneumocystis jiroveci pneumonia prophylaxis (daily sulfamethoxazole (400 mg)–trimethoprim (80 mg) or pentamidine aerosolisations for patients allergic to sulfa drugs).
Study visits were scheduled at enrolment, then every 3 months until the endpoint, 28 months postrandomisation. At each visit, BVAS was calculated and blood samples were drawn from every patient. Patients were asked to record their study medication(s) weekly in a specifically designated diary. It was mandatory that all ANCA-testing and CD19+ B cell counts for a given patient be determined in the same laboratory.
The primary endpoint at month 28 was the number of relapses, defined as reappearance or worsening of AAV symptoms, that is, BVAS>0. Secondary endpoints included the number of major relapses, defined as life-threatening or involving at least one major organ; number of minor relapses; potential association of ANCA evolution and CD19+ B cell counts with relapses; glucocorticoid duration and cumulated dose; Vasculitis Damage Index (VDI) evaluated damage severity and number for each group and mortality. We also recorded all AEs, treatment-related or not and AAV evolution. The independent Adjudication Committee assessed all relapses.
The trial was designed to detect a 20% absolute between-arm difference for relapses, with a 5% alpha risk and 80% power in a two-sided test, with 35% relapses in the control group and 10% lost-to-follow-up for both groups.
The statistical analyses were conducted according to the intention-to-treat principle, including all randomised patients in their assigned group. For descriptive analyses, qualitative variables are expressed as numbers (%) or mean±SD, as appropriate and quantitative variables as median (IQR).
A generalised estimating equation model with Poisson distribution, adjusted for AAV type (newly diagnosed or relapsing) considered for randomisation and supposing a within-centre correlation (exchangeable correlation structure), compared the means of the numbers of relapses per patient per group at month 28. Kaplan-Meier estimations of the times to first relapse were compared between groups with a Cox regression model, also adjusted for AAV type and supposing within-centre correlation.
Statistical analyses were computed with R V.3.2.2 (R Core Team (2015). R: A language and environment for statistical computing (R Foundation for Statistical Computing, Vienna, Austria, https://www.R-project.org)).
Enrolment and baseline characteristics
Among 166 patients enrolled in 59 centres (figure 1) between November 2012 and November 2013, 3 who did not meet inclusion criteria were not randomised. Among the 163 patients randomised, one did not provide written consent and was excluded. Among the 162 randomised patients, 117 (72.2%) had GPA and 45 (28.8%) MPA; 104 (64.2%) and 58 (35.8%), respectively, were in remission after a first flare or at least one relapse. Preinclusion induction treatment included cyclophosphamide for 100 (61.7%), rituximab for 61 (37.6%) and methotrexate for 1 (0.6%). Eighty-one (50%) patients, whose characteristics were comparable (table 1), were randomised to each study arm.
At month 28, 21 patients had suffered 22 relapses: 14/81 (17.3%) in 13 tailored-infusion recipients and 8/81 (9.9%) in 8 fixed-schedule–infusion patients (p=0.22). Among the 21 patients who relapsed, 12 had newly diagnosed AAVs, 18 with GPA and 3 with MPA; 12 were anti-PR3 and 3 were anti-MPO ANCA-positive, 2 were ANCA-positive of unknown specificity and 4 were ANCA-negative. Two relapses (one in each arm) occurred after month 28 and were censured from the main analysis.
Comparing tailored versus fixed-schedule rituximab infusions, respectively: relapse-free survival rates were 83.8% (95% CI 76.1% to 92.3%) vs 86.4% (95% CI 79.2 to 94.2) (p=0.58) (figure 2) and major relapses occurred in 6 (7.4%) vs 3 (3.7%) patients (p=0.23). The six major relapses in the tailored-infusion arm manifested as two renal flares, two peripheral neuropathies, one pulmonary nodule and one pachymeningitis with orbital mass. To treat those relapses, two patients received cyclophosphamide and four rituximab. The three major relapses in the control arm (renal flare, myopericarditis with pulmonary infiltrates or subglottic stenosis) were treated with rituximab. All relapses are described in detail in online supplementary tables S1 and S2.
Supplementary file 1
VDIs (mean ±SD) for the tailored and fixed-schedule rituximab-infusion arms, respectively, were 1.64±1.41 and 1.86±1.70 at inclusion and 1.99±1.57 and 2.09±1.97 at 28 months.
The damage-accrual difference between the two arms was 0.14 (95% CI –0.07 to 0.35) (p=0.179), using a constrained longitudinal data-analysis model (unplanned analysis required by reviewers).
ANCA-cell and B-cell-repopulation-related AAV relapses
ANCA and circulating CD19+ B cell determinations and their evolutions were available for 161 patients (one relapsing patient’s data were missing). Five ANCA-evolution profiles were identified (table 2); none was associated with relapses. At month 28, 46/76 (60.5%) tailored-infusion recipients were ANCA-positive vs 26/71 (36.6%) fixed-schedule patients (p=0.06). In addition, circulating B cells were not detected in 10 (45.4%) patients who relapsed and 4 (18.2%) were ANCA-cell-negative and B-cell-negative. All relapses, including the two censured, were analysed.
Cumulated glucocorticoid dose and duration
Glucocorticoid doses and durations since inclusion did not differ significantly between the two arms. For tailored-infusion versus fixed-schedule regimen, respectively, cumulated glucocorticoid doses (mean ±SD) were 4915±2613 vs 4850±2444 mg (p=0.71) and treatment lasted 25.3 (95% CI 23.9 to 26.6) vs 24.5 (95% CI 22.8 to 36.3) months (p=0.52).
Patients’ diaries were collected at each visit (online supplementary table S3).
Twenty-five patients stopped glucocorticoids during the study; that discontinuation was not associated with relapses (online supplementary figure S1).
The tailored-infusion versus fixed-schedule groups, respectively, received 248 vs 381 infusions, with medians (IQR) of 3 (2–4) vs 5 (5–5) administrations. Four (4.9%), 23 (28.4%), 30 (37%), 15 (18.5%), 6 (7.4%), 2 (2.5%) or 1 (1.2%) tailored-regimen patients received, respectively, 1, 2, 3, 4, 5, 6 or 7 rituximab infusions, at a median between-infusion interval of 6.1 (IQR: 3.1–9.2) months. That group received 168 reinfusions because ANCA became positive or titres rose (22 (13.1%) infusions), circulating B cells reappeared (85 (50.6%) infusions) or both (61 (36.3%) infusions). Ten protocol deviations were observed in the tailored-regimen arm: 10 patients did not receive one reinfusion: eight because the investigator did not follow the reinfusion rules and two patients refused.
Sixty-nine (85.2%) tailored-infusion recipients vs 74 (91%) controls reported at least one AE (p=0.33), with, respectively, 26 (32.1%) vs 31 (38.3%) experiencing at least one severe AE (SAE) (p=0.51) and 37 vs 53 SAEs per group. Eighteen infectious complications occurred in each arm, with 9 (11.1%) tailored-regimen recipients vs 16 (19.8%) controls experiencing at least one infection. The only episode of neutropaenia occurred in the tailored-infusion arm. SAEs are reported in table 3 and online supplementary table S4.
Four patients died during the study: one tailored-infusion recipient (bronchospasm not related to rituximab infusion) versus three controls (nosocomial pneumonia, carcinomatous meningitis or cardiogenic shock unrelated to AAV or rituximab).
No significant between-group gammaglobulin-level differences or decreases were observed throughout the trial (online supplementary figure S2).
The usefulness of monitoring ANCA reappearance/titres and/or circulating CD19+ B cells for AAV treatment is controversial and not evidence-based.7 8 11 14–16 This prospective trial was designed to determine whether two of the most frequently prescribed laboratory tests during surveillance of AAV remission-maintenance therapy are reliable and could help decide whether or not to reinfuse rituximab during follow-up.
The results of this study confirmed rituximab efficacy for AAV-remission maintenance: relapse rates and major-relapse rates were low in both arms: 6/81 (7.4%) vs 3/81 (3.7%) for tailored and fixed-schedule regimens, respectively. Relapse rates were comparable to that of rituximab-treated patients in the MAINRITSAN trial (3/57 (5.2%) with a major relapse).5 They should be compared with the 32% observed at 18 months without maintenance therapy after remission induction with rituximab2 or the 29% major-relapse rate at month 28 under azathioprine after cyclophosphamide-induced remission.2 5 Moreover, because the high relapse rates observed without treatment2 or with azathioprine5 can impact survival and enhance AAV-induced damage,17 18 on top of iatrogenicity of a new induction treatment, our data further support that it is reasonable to maintain remission with rituximab and not wait to retreat.
This trial’s findings also demonstrated that it is indeed possible to maintain remission with fewer infusions, even though a non-significant trend towards more relapses was observed for patients receiving the individually tailored regimen. In that arm, no day-14 infusion was planned and patients received one less reinfusion than the fixed-schedule controls.
Herein, although ANCA evolution and/or circulating CD19+ B cells were not reliable predictors of AAV relapses, combining them achieved fewer infusions in the tailored-infusion arm without significantly more relapses. In the literature, the role of ANCA as a marker of relapse remains a source of debate.8 In an earlier randomised trial,19 more than half the patients not given any maintenance therapy and whose ANCA titres rose suffered a major relapse. In a previous retrospective study,7 we showed that relapse-free survival was longer when anti-PR3 ANCA remained negative, and clinical status and ANCA evolution were closely associated for only 60% of patients with GPA. More recently,9 16 anti-PR3 ANCA were associated with GPA relapses only in the subgroup of patients with a ‘vasculitis phenotype’ (eg, renal involvement or alveolar haemorrhage), particularly after receiving rituximab. Herein, five distinct ANCA-evolution profiles were identified, none of which was a good predictor of relapses. Patients receiving individually tailored infusions were more likely to be ANCA-positive at month 28. Among relapsing patients, 10/13 (76.9%) were ANCA-positive at month 28 vs only 3/8 (37.5%) fixed-schedule controls. The small number of events prevents us from drawing any definitive conclusions.
Pertinently, relapses could also occur in the absence of circulating B cells, perhaps because B-cell repopulation could be only at sites of active disease20 or because the CD19+ B lymphocyte count is not strictly associated with CD27 +memory B-cell reemergence.21 Four of our patients were ANCA-negative at relapse without any circulating B cells. This pattern, considered rare,11 represented 18.2% of the relapses in our study. Thus, we are unable to conclude as to the relevance of monitoring ANCA and/or CD19+ B cells. In the MAINRITSAN2 trial, patients assigned to the tailored-infusion strategy received fewer rituximab infusions and relapses in this group were not significantly different from those of the fixed-schedule controls. Although these laboratory findings seem to have no association with relapse, they are useful markers to guide infusion rhythm during follow-up.
Eighteen infections occurred in each arm, with lung infections being the most frequent. Even though rituximab was not associated with more infectious complications than azathioprine3 or methotrexate,4 safety remains an important issue. In large series,22–25 up to 29% of rituximab-treated patients experienced infectious complications. None of our patients developed Pneumocystis jiroveci pneumonia, probably because of mandatory trimethoprim–sulfamethoxazole prophylaxis. Also, progressive multifocal leukoencephalopathy was not seen within the 28 months of observation.
This trial’s findings have multiple repercussions for patients and physicians. The low relapse rates observed in the two arms after 28 months of follow-up confirmed that rituximab-maintenance therapy is justified. Moreover, patients were probably overtreated with previously administered regimens5 because remission could be sustained with a lower rituximab dose. Although ANCA and CD19+ B cell monitoring was not associated with relapse, those values contributed to deciding to reinfuse, with the clear benefit of fewer rituximab infusions in the tailored-infusion arm.
This study had some limitations. It was open-labelled but all relapses were assessed by an independent Adjudication Committee, unaware of treatment arm and the circulating CD19+ B cell count. Biological parameters were assayed in each centre, because it would have been impractical to centralise testing for the 59 participating centres throughout the country and impossible to decide rapidly to reinfuse patients. However, all ANCA-titering and CD19+ B cell counts for a given patient had to be done in the same laboratory. ANCA titres were measured with different techniques according to each centre’s practice. Interassay variability was not evaluated before the study. Fixed-schedule infusion-arm patients were more likely to be ANCA-positive at inclusion. It cannot be excluded that this difference might have impacted the relapse risks of patients in the two arms. By protocol design, two-thirds of the patients were included after their first flare; therefore, our results may not completely be fully generalisable to the subgroup of patients with relapsing disease.
Our study also has several strengths. It was a multicentre trial using a reinfusion-decision algorithm applied to the tailored-infusion arm, which received 248 infusions; only 10 protocol deviations were observed, probably because the algorithm included in the eCRF emitted an automatic directive indicating the need for reinfusion.
In conclusion, AAV relapse rates for patients treated according to individually tailored or fixed-schedule rituximab-infusion regimens did not differ significantly. However, those benefitting from personalised patient-centred care received fewer infusions and, hence, lower total rituximab doses.
The authors thank URC-CIC Paris Descartes Necker/Cochin (Séverine Aït El Ghaz-Poignant and Charly Larrieu) for implementation, monitoring and data management of the study, DEC-Agence Générale des Equipements et Produits de Santé for their logistics assistance with the study drugs and Janet Jacobson for editorial assistance.
Disclaimer : This is a summary of a scientific article written by a medical professional (“the Original Article”). The Summary is written to assist non medically trained readers to understand general points of the Original Article. It is supplied “as is” without any warranty. You should note that the Original Article (and Summary) may not be fully relevant nor accurate as medical science is constantly changing and errors can occur. It is therefore very important that readers not rely on the content in the Summary and consult their medical professionals for all aspects of their health care and only rely on the Summary if directed to do so by their medical professional. Please view our full Website Terms and Conditions.
Copyright © 2019 BMJ Publishing Group Ltd & European League Against Rheumatism. Medical professionals may print copies for their and their patients and students non commercial use. Other individuals may print a single copy for their personal, non commercial use. For other uses please contact our Rights and Licensing Team.
Handling editor Josef S Smolen
Contributors PCh and LG conceived and planned the study. ÉP and PR designed and carried out the statistical analyses. PCh wrote the first draft of the manuscript and all authors contributed to subsequent revisions. All other authors entered data, participated in devising the protocol and reviewed all versions of the manuscript.
Funding This study was funded by a research grant from the French Ministry of Health (PHRC National 2011 AOM11145) and sponsored by the Assistance Publique–Hôpitaux de Paris. Hoffmann-La Roche provided rituximab for the study.
Competing interests BT has received consulting and speaking fees (Roche, LFB, Grifols, GSK). MH has received personal fees from Roche. AK has received personal fees and non-financial support from Roche. XP has received speaking fees and honoraria (Pfizer, LFB, Roche) and a research grant (Pfizer).
Patient consent Obtained.
Ethics approval The Ethics Committee (Comité de Protection des Personnes Île-de-France 1 (Paris)) approved the study.
Provenance and peer review Not commissioned; externally peer reviewed.
Collaborators Other investigators and members of the French Vasculitis Study Group who participated in the study are listed in the online supplementary.
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.