Rituximab for remission induction and maintenance in refractory granulomatosis with polyangiitis (Wegener's): Ten-year experience at a single center


  • Rodrigo Cartin-Ceba,

    1. Mayo Clinic, Rochester, Minnesota
    Search for more papers by this author
  • Jason M. Golbin,

    1. Mayo Clinic, Rochester, Minnesota, and Long Island Lung Center, Bay Shore, New York
    Search for more papers by this author
  • Karina A. Keogh,

    1. Mayo Clinic, Rochester, Minnesota
    Search for more papers by this author
    • Drs. Keogh, Peikert, Ytterberg, and Fervenza were site coinvestigators on the Rituximab in ANCA-Associated Vasculitis (RAVE) trial, for which only Drs. Ytterberg and Fervenza received support.

  • Tobias Peikert,

    1. Mayo Clinic, Rochester, Minnesota
    Search for more papers by this author
    • Drs. Keogh, Peikert, Ytterberg, and Fervenza were site coinvestigators on the Rituximab in ANCA-Associated Vasculitis (RAVE) trial, for which only Drs. Ytterberg and Fervenza received support.

  • Marta Sánchez-Menéndez,

    1. Mayo Clinic, Rochester, Minnesota
    2. Centro Médico de Asturias, Oviedo, Spain
    Search for more papers by this author
  • Steven R. Ytterberg,

    1. Mayo Clinic, Rochester, Minnesota
    Search for more papers by this author
    • Drs. Keogh, Peikert, Ytterberg, and Fervenza were site coinvestigators on the Rituximab in ANCA-Associated Vasculitis (RAVE) trial, for which only Drs. Ytterberg and Fervenza received support.

  • Fernando C. Fervenza,

    1. Mayo Clinic, Rochester, Minnesota
    Search for more papers by this author
    • Drs. Keogh, Peikert, Ytterberg, and Fervenza were site coinvestigators on the Rituximab in ANCA-Associated Vasculitis (RAVE) trial, for which only Drs. Ytterberg and Fervenza received support.

    • Dr. Fervenza has received unrestricted research grants from Genentech and Biogen Idec.

  • Ulrich Specks

    Corresponding author
    1. Mayo Clinic, Rochester, Minnesota
    • Division of Pulmonary and Critical Care Medicine, Mayo Clinic, 200 First Street SW, Rochester, MN 55905
    Search for more papers by this author
    • Dr. Specks was coprincipal investigator for the RAVE trial, for which he received support. The RAVE trial was sponsored by the National Institute of Allergy and Infectious Diseases and was partially funded by Genentech and Biogen Idec. Participants in the RAVE trial were excluded from the present study.



This study was conducted to evaluate the efficacy and safety of repeated and prolonged B cell depletion with rituximab (RTX) for the maintenance of long-term remission in patients with chronic relapsing granulomatosis with polyangiitis (Wegener's) (GPA).


We conducted a single-center observational study of all patients with chronic relapsing GPA treated with at least 2 courses of RTX between January 1, 2000 and May 31, 2010. Participants in the Rituximab in ANCA-Associated Vasculitis (RAVE) trial were excluded from this analysis. Data were abstracted from electronic medical records.


Fifty-three patients with refractory GPA (median age 46 years [interquartile range (IQR) 30–61 years]; 53% women) received at least 2 courses of RTX to treat GPA relapses or to maintain remission. All but 1 patient had antineutrophil cytoplasmic antibodies (ANCA) against proteinase 3 (PR3). These patients received a median of 4 courses of RTX (IQR 3–5); all had depletion of B cells, and the median time to return of B cells was 8.5 months (IQR 6–11 months). All observed relapses occurred after reconstitution of B cells and were accompanied or preceded by an increase in ANCA levels, except for the 1 ANCA-negative patient. Infusion-related adverse events occurred in 16 patients. During the period of B cell depletion, 30 infections requiring antimicrobial therapy were recorded.


RTX appeared to be effective and safe for the induction and maintenance of remission in patients with chronic relapsing GPA. Repeated depletion of B lymphocytes seems to be associated with a low risk of infections. Preemptive re-treatment decisions can be individualized based on serial B lymphocyte and PR3 ANCA monitoring. The use of RTX for the maintenance of long-term remission merits further formal investigation.

Granulomatosis with polyangiitis (Wegener's) (GPA) is a primary systemic small-vessel vasculitis that often affects the respiratory tract and the kidneys (1). Most patients have circulating antineutrophil cytoplasmic antibodies (ANCA) against proteinase 3 (PR3) (2). The traditional treatment regimen for severe disease has been the combination of cyclophosphamide (CYC) and glucocorticoids. With this combination, initial remission rates of 70–90% have been reported (3–5). The remaining patients have persistent disease activity, and 50% of those who achieve initial remission will relapse within 2 years of diagnosis (3, 5). Patients with refractory disease, defined as having persistent disease activity while receiving the maximal tolerated dose of CYC or having contraindications for the use of CYC, have been extremely difficult to manage. Consequently, more effective and better tolerated treatment alternatives for such patients have been sought.

B lymphocytes were implicated in the pathogenesis of GPA 3 decades ago and seemed to be the primary target of the therapeutic effects of CYC in this disease (6, 7). Moreover, the frequency of activated B lymphocytes has been associated with both disease activity and disease severity (8). Rituximab (RTX), a chimeric anti-CD20 monoclonal antibody that specifically targets cells of B lymphocyte lineage, was first used successfully in 2000 to induce remission in a patient with refractory GPA (9). Since then, several studies have reported remission rates higher than 80% with RTX therapy in patients with ANCA-associated vasculitis (AAV) presenting with refractory disease (10–16). More recently, 2 randomized controlled trials (Rituximab in ANCA-Associated Vasculitis [RAVE] and Rituximab versus Cyclophosphamide in ANCA-Associated Vasculitis [RITUXVAS]) found RTX to be noninferior to CYC for the induction of remission in newly diagnosed severe AAV (17, 18) and superior in severe relapsing disease (17).

The need for repeated treatments with RTX as well as the efficacy and safety of prolonged B cell depletion for the long-term management of patients with chronic relapsing GPA is less clear. Moreover, the role of monitoring B lymphocyte counts and ANCA levels to guide re-treatment decisions in this group of patients has yet to be elucidated. Therefore, the present observational study was conducted to evaluate the efficacy and safety of RTX for remission induction and long-term maintenance therapy in patients with severe, chronic relapsing refractory GPA. To this end, we determined the effect of repeated courses of RTX on clinical disease activity, duration of B lymphocyte depletion, ANCA levels, severity of subsequent clinical relapses, adverse events, and immunoglobulin levels.


Patient identification.

After approval by the Institutional Review Board, we conducted a single-center historical cohort study of all patients with GPA treated with at least 2 courses of RTX between January 1, 2000 and May 31, 2010. Two complementary search strategies were applied. The first list of patients was derived from a prospectively collected clinical database of consecutive patients with AAV evaluated at our institution. The second list was established using the Mayo Clinic Life Sciences System. This collaborative effort between Mayo Clinic and IBM represents a clinical data warehouse providing approved users with a query tool called Data Discovery and Query Builder. With this tool, the entire electronic medical record, including all clinical notes, can be searched to extract clinical data about patients who received care at Mayo Clinic, based on user-defined search terms. This system was applied to identify patients whose records showed the following query terms anywhere in the electronic clinical notes: ANCA, vasculitis, Wegener's granulomatosis (WG), or microscopic polyangiitis (MPA). The system was also queried for International Classification of Diseases, Ninth Revision (ICD-9) codes 446.4 (for WG), 446.0 (for MPA), and 447.6 (for AAV). The query did not include patients with Churg-Strauss syndrome.

The generated lists were combined, and patients were eligible for the study if they had provided authorization for review of their medical records, were older than 18 years at the time of the search, had a biopsy-proven diagnosis of GPA or fulfilled the American College of Rheumatology criteria for GPA (19) and the Chapel Hill Consensus Conference definitions of GPA or MPA (1), and had received at least 2 courses of RTX for refractory disease.

Definitions and measurements.

Consistent with the definition used by the European Vasculitis Study Group (20), refractory disease was defined as disease that, at the time of the first RTX infusion, was unresponsive to maximum tolerated doses of CYC or disease for which CYC was deemed necessary to control its activity but could not be given because of previous toxicities associated with CYC use (profound leukopenia, hemorrhagic cystitis) or other contraindications, such as history of malignancy (21, 22). Participants in the RAVE trial were excluded from this analysis.

Disease activity was assigned according to the Birmingham Vasculitis Activity Score for WG (BVAS/WG) (23). All relapses were evaluated at our institution by at least 1 vasculitis expert. Remission was defined as a BVAS/WG score of 0, and complete remission as a BVAS/WG score of 0 and successful completion of the prednisone taper. Relapse was defined as an increase in the BVAS/WG score of ≥1.

B lymphocyte depletion and reconstitution were defined as <20 and ≥20 CD19+ cells/μl, respectively, as measured by fluorescence activated cell sorting. The ANCA staining pattern (cytoplasmic [cANCA] or perinuclear [pANCA]) was determined by indirect immunofluorescence, and the antigen specificity for myeloperoxidase (MPO) or PR3 was measured by enzyme-linked immunosorbent assay (until 2006) or by Multiplex Flow Immunoassay (after 2006), according to standard procedures (24, 25). Any conversion from negative to positive and any increase from the previously determined nadir were deemed clinically significant.

Our practice protocol aimed to measure B lymphocyte and PR3 ANCA levels prior to the first RTX infusion, about 1–2 weeks after the last RTX infusion of each course to ascertain successful B lymphocyte depletion, and then quarterly thereafter (12). In individual patients with established B lymphocyte reconstitution, PR3 ANCA, and related flare patterns, B lymphocyte or PR3 ANCA levels were measured monthly if the preceding flare occurred between 2 quarterly measurements.

Medical records were abstracted for demographic and clinical information, including the number of RTX courses, B cell counts, PR3 ANCA levels, duration of B cell depletion, reasons for RTX treatment, frequency and severity of relapses, immunoglobulin levels, and adverse events (infusion related, number and type of infections). Serious adverse events associated with infusion were defined as those resulting in hospitalization, life-threatening situations, or death.

Treatment protocol.

The remission induction regimen consisted of the combination of glucocorticoids and RTX. Patients received oral prednisone at a dosage of 1 mg/kg/day, not to exceed 80 mg/day. Intravenous methylprednisolone, 1 gm/day for up to 3 days preceding the oral prednisone, was used occasionally if deemed necessary by the treating physician. A previously described consistent prednisone dosage–tapering regimen was used (12, 17, 26). Briefly, the dosage of prednisone was tapered such that by month 5, all patients in whom remission was achieved without disease flare had completely discontinued glucocorticoids. The prednisone dosage was reduced to 40 mg/day no later than by the completion of week 4 and was maintained for 2 weeks. Stepwise dosage reductions then continued every 2 weeks: 30 mg, 20 mg, 15 mg, 10 mg, 7.5 mg, 5 mg, 2.5 mg, and 0 mg/day. Once RTX therapy was initiated, all of the other immunosuppressive agents were discontinued (except glucocorticoids) and were not restarted in any of the patients throughout the entire observation period.

The RTX regimen consisted of 4 weekly intravenous (IV) infusions, each at a dose of 375 mg/m2 of body surface area. More recently, RTX was given as 2 IV doses of 1 gm each, administered 2 weeks apart, to some patients for remission maintenance. To minimize potential infusion reactions associated with RTX, 100–125 mg of methylprednisolone was given once as premedication immediately preceding the first infusion of each series of infusions to most patients.

After successful depletion of B lymphocytes, remission induction, and glucocorticoid taper, patients were re-treated with RTX for 1 of the 2 following reasons. First, if a clinical relapse occurred, the entire remission induction regimen was repeated; depending on the severity of the disease flare, the starting dose of the prednisone tapering regimen varied between 30 and 60 mg/day. Second, for remission-maintenance therapy, RTX was administered without concomitant prednisone in the event of any of the following scenarios: 1) a rise in the PR3 ANCA levels following reconstitution of B lymphocytes; 2) reconstitution of B lymphocytes in patients with GPA who never had a positive ANCA test result; or 3) reconstitution of B lymphocytes in the absence of a concomitant rise in PR3 ANCA levels in patients with known PR3 ANCA positivity and a well-established history of relapses occurring promptly upon B lymphocyte reconstitution.

Statistical analysis.

All data are summarized as the median (interquartile range [IQR]) or as the percentage, except where indicated otherwise. Wilcoxon's rank sum test and Wilcoxon's signed rank test were used to compare unpaired and paired continuous variables, respectively. JMP statistical software (version 8.0; SAS Institute) was used for all data analyses. P values less than 0.05 were considered significant for all statistical tests.


Patient characteristics.

A total of 637 patients with GPA or MPA were evaluated at our institution between January 2000 and May 2010. During the study period, all patients (and only those patients) with refractory disease were treated with RTX, except for 54 participants in the RAVE trial who had enrolled at our institution. Fifty-three patients fulfilled the inclusion criteria (Figure 1). The median age of the cohort was 46 years (IQR 30–61 years); 28 patients were women, and all but 1 of the cohort were PR3 ANCA positive. Table 1 summarizes the baseline clinical characteristics, treatment history, and organ system involvement in these 53 patients at the time of their first RTX course. Twenty-one patients (40%) had granulomatous disease manifestations and 32 (60%) had only vasculitic disease manifestations prompting the first RTX treatment course.

Figure 1.

Study design showing the distribution of the patients identified with granulomatosis with polyangiitis (Wegener's) (GPA) or microscopic polyangiitis (MPA) during the study period January 2000 to May 2010. RAVE = Rituximab in ANCA-Associated Vasculitis; RTX = rituximab.

Table 1. Baseline characteristics of the 53 patients with relapsing granulomatosis with polyangiitis (Wegener's)*
  • *

    IQR = interquartile range; RTX = rituximab; PR3 = proteinase 3; ANCA = antineutrophil cytoplasmic antibody; BVAS/WG = Birmingham Vasculitis Activity Score for Wegener's Granulomatosis; ENT = ear, nose, and throat.

  • Includes serious side effects of cyclophosphamide therapy, such as hemorrhagic cystitis, recurrent severe leukopenia, and history of solid or hematologic malignancy.

Age, median (IQR) years46 (30–61)
No. (%) female28 (53)
Time from diagnosis, median (IQR) years10 (8–14)
Followup period since first RTX infusion, median (IQR) years4.4 (2.7–6.2)
PR3 ANCA, no. (%)52 (98)
Relapse with cyclophosphamide, no. (%)42 (79)
Cyclophosphamide contraindicated, no. (%)11 (21)
Previous methotrexate therapy, no. (%)30 (57)
Previous azathioprine therapy, no. (%)26 (49)
BVAS/WG score at first RTX infusion, median (IQR)5 (4–6)
Organ system involvement, no. (%) 
 Systemic (arthralgias, arthritis, fever)34 (64)
 Skin6 (11)
 Mucous membranes/eyes16 (20)
 ENT28 (53)
 Cardiovascular1 (2)
 Gastrointestinal0 (0)
 Pulmonary26 (49)
 Renal22 (42)
 Nervous system6 (11)

Treatment characteristics.

A total of 233 courses of RTX were given to these 53 patients over the study period, with a median number of courses per patient of 4 (IQR 3–5). Eighty-five (36%) of the courses were combined with the glucocorticoid regimen for induction of remission after a disease relapse. The majority of RTX courses (148 [64%]) were given preemptively.

The median time from the first RTX infusion to reconstitution of B lymphocytes was 8.5 months (IQR 6–11 [range 5–22]). Most of the courses (209 courses [90%]) consisted of 4 weekly doses of 375 mg/m2. Nine patients received a total of 24 courses (10%) consisting of 2 fixed doses of 1 gm each, based on the discretion of the treating physician and on patient preference. This 2-dose regimen was only given preemptively, whereas all relapses were treated with the body surface–based 4-dose regimen. A paired analysis comparing the last 4-dose regimen to the successive 2-dose regimen given to these 9 patients for remission maintenance showed no difference in the median time to reconstitution of B lymphocytes: 7.6 months (IQR 7–9 months) for the 4-dose regimen versus 8.6 months (IQR 7–9.5 months) for the 2-dose regimen (P = 0.51).

Figure 2 shows the distribution of the total number of RTX courses per patient and indicates that the majority of patients received between 3 and 5 courses. One patient received as many as 12 courses of RTX. In addition, Figure 2 shows that the time to repletion of B lymphocytes initially lengthened but tended to reach a plateau after more than 4 courses.

Figure 2.

Total number of rituximab courses (sets of 4 weekly or 2 fortnightly infusions) given during the study period and median time to reconstitution of B lymphocytes. Open bars represent the number of courses given for treatment of relapses; solid bars represent the number of courses given for relapse prevention. Numbers within the bars are the number of courses represented graphically by the bars. The plot above the bars is the median time (months) to reconstitution of B lymphocytes associated with each infusion category.

Clinical response.

All patients treated for relapses with the combination of RTX and glucocorticoids (85 courses) achieved complete remission. The only exception was 1 patient who achieved remission but required long-term hormone replacement therapy for panhypopituitarism, which was damage from a previous disease flare.

All 32 relapses observed among these 53 patients occurred after their first remission induction with RTX following reconstitution of B lymphocytes and were associated with, or preceded by, a rise in the PR3 ANCA levels from their nadir (except in 1 patient who was PR3 ANCA negative). The BVAS/WG score at the time of each clinical relapse was equal to or lower than that at the preceding relapse (Figure 3).

Figure 3.

Severity of disease flares, as determined by the Birmingham Vasculitis Activity Score for Wegener's Granulomatosis (BVAS/WG). The first flare shown represents the disease activity that prompted the first course of rituximab for remission induction. Subsequent disease flares occurring in individual patients were of similar or lower severity.

Remission was successfully maintained in all patients who had been re-treated preemptively (148 courses) with RTX. In 130 instances, these courses were given after both reconstitution of B lymphocytes and an increase in the PR3 ANCA levels. Fifteen courses were given after reconstitution of B lymphocytes in the absence of a concomitant rise in PR3 ANCA values (3 patients with known PR3 ANCA positivity and a well-documented history of relapse promptly upon B lymphocyte repletion). The only patient in this cohort without detectable ANCAs also received 2 preemptive courses of RTX following B lymphocyte reconstitution.

Adverse effects.

Adverse effects were recorded over a followup of 230.4 patient-years (Table 2). RTX was well tolerated, and there were no serious adverse events during the 233 courses of treatment given in this cohort of 53 patients. A total of 16 events were related to infusions; none of these events precluded the completion of treatment. Of these 16 events, 14 occurred during the first RTX infusion.

Table 2. Adverse events reported during the followup period (230.4 patient-years)
Adverse eventNo. of events
 Acute myelogenous leukemia1
 Pneumocystis jiroveci pneumonia1
 Upper respiratory tract9
 Urinary tract3
 Herpes zoster4
Infusion-related events16
 Throat tightness1
 Chest tightness1

Two patients died during the study period: one of respiratory failure caused by Pneumocystis jiroveci pneumonia (PCP), occurring 2 months following discontinuation of prophylaxis with trimethoprim/sulfamethoxazole; this patient had not received any glucocorticoid therapy. The other patient died of acute myelogenous leukemia; this patient had received a cumulative dose of 41 gm of CYC prior to receiving the first RTX infusion.

Overall, 30 infections were reported. Most of these were related to the respiratory tract (Table 2). No case of progressive multifocal leukoencephalopathy was identified.

Immunoglobulin levels.

Immunoglobulin levels were not measured routinely in all of the patients. IgM and IgA levels were available for 18 patients and IgG levels for 19 (Table 3). The time between baseline and followup measurements of immunoglobulin levels was also not standardized. The median number of RTX courses given between the measurements was 4 (IQR 3–6). There was a significant decline in the levels of all 3 immunoglobulins between measurements. However, IgA levels remained within the normal range, and IgG levels of <200 mg/dl were not observed. No link between lower IgG levels and the occurrence of infection could be detected.

Table 3. Immunoglobulin levels before and after RTX therapy*
 No. of patientsBaseline, median (IQR)Post-RTX therapy, median (IQR)P
  • *

    RTX = rituximab; IQR = interquartile range.

IgA, mg/dl (normal 61–356)18147 (69–202)86 (60–181)0.006
IgG, mg/dl (normal 767–1,590)19840 (678–949)574 (427–739)<0.001
IgM, mg/dl (normal 50–300)1851 (41–77)26 (13–48)<0.001


This 10-year followup study of a historical cohort at a single center demonstrates that repeated B lymphocyte depletion with RTX can effectively and safely be used to reestablish and maintain remission in patients with chronic relapsing refractory GPA. We found no evidence that patients develop resistance to RTX over time. However, even after repeated courses of RTX, PR3 ANCA–positive patients with relapse remained at risk for the recurrence of disease after reconstitution of B lymphocytes. Our study shows that relapses can be prevented by re-treatment with RTX timed individually based on serial B lymphocyte and PR3 ANCA determinations. Infusion-related adverse events were few, mild, and did not preclude re-treatment with RTX. The number of infections requiring antimicrobial therapy was very low, given the extended followup period and the large number of sequential courses of RTX received by these patients.

All patients in this cohort achieved complete remission with the RTX-based remission induction regimen, and disease flares were not observed in the absence of peripheral blood B lymphocytes. This finding supports the central pathogenic role of B lymphocytes in GPA. Moreover, each relapse was of equal or lesser severity than the preceding relapse. This is most likely the result of surveillance bias, since these patients were regularly and carefully monitored, allowing for earlier detection of disease flare. Not all disease flares occurring after the first remission induction with RTX were severe. However, since these patients had previously failed multiple remission induction and maintenance regimens other than RTX, even disease flares that were not severe were treated with the RTX-based regimen.

Remission was maintained in all patients treated preemptively with RTX. Patients who experience chronic relapse may benefit from re-treatment, as it spares them additional morbidity and damage from recurrent disease and exposure to glucocorticoids. The timing of B lymphocyte reconstitution, PR3 ANCA level increases, and disease relapses were highly variable between patients, but were very consistent in individual patients.

Our observations differ somewhat from other reports. In our cohort, relapses followed PR3 ANCA increases unless the patients were retreated preemptively. This contrasts with more general conventionally treated cohorts of GPA, in which serial PR3 ANCA determinations did not predict relapses (27). This may be because our cohort consisted entirely of patients with chronic relapses. Moreover, PR3 ANCA levels may be affected differently by RTX than by other immunosuppressive therapies (17).

We did not observe disease flares in the absence of peripheral blood B lymphocytes, but other investigators have (15, 16). Reasons for this may include differences in the methods of B lymphocyte and PR3 ANCA detection, differences in RTX dosing, concomitant use of other immunosuppressive agents at other centers, and the fact that peripheral blood B lymphocyte counts may not accurately reflect tissue B lymphocyte levels.

Relapses with documented B lymphocytes within tissue lesions despite undetectable peripheral blood B lymphocytes have been reported and have been attributed to impaired macrophage effector mechanisms (28, 29). In 2 cohorts that included patients with relapses despite peripheral blood B lymphocyte depletion, other immunosuppressive agents, most of which have documented effects on macrophage function, were used concomitantly with RTX (15, 16, 30). It is possible that by affecting macrophage function, the concomitant use of these agents impairs complete tissue B lymphocyte depletion, allowing for disease activity despite undetectable peripheral blood B lymphocytes. If the concomitant use of other immunosuppressants has an effect on tissue B lymphocyte depletion, it may also have an effect on ANCA production, as the short-lived plasma cells are dependent on antigen-specific B cell progenitors. This could explain why in the study by Rhee and colleagues (16), the 3 patients who experienced a flare in the absence of detectable peripheral blood B lymphocytes had remained ANCA-positive all along.

All clinical trials evaluating the efficacy of RTX for remission induction in AAV have used the 4-dose body surface–based dosing regimen that is now approved by the Food and Drug Administration for AAV (12, 17, 18). No controlled studies have been performed to determine optimum RTX dosing in AAV for either remission induction or maintenance. The 2-dose regimen of 1 gm per dose that was approved for rheumatoid arthritis may be of similar efficacy and may result in B lymphocyte depletion of similar duration (15, 31). One center reported successful remission maintenance using a single 1-gm dose given every 4 months (16). However, it is possible that the single-dose application of RTX rather than an infusion series is suboptimal for tissue B lymphocyte depletion and, thus, may have contributed to the 3 disease flares observed despite undetectable peripheral blood B lymphocytes (16). The optimal dosing and timing of RTX for remission maintenance in chronic relapsing GPA clearly deserves more investigation.

RTX was well tolerated in this cohort. As previously reported, the observed infusion reactions were relatively uncommon and mild, and most were related to the first infusion. They did not preclude the completion of the infusion or the subsequent use of RTX. We did not observe any increase in the severity of reactions with subsequent courses of RTX delivered to patients who had experienced infusion reactions with previous courses.

Overall, the rate of infections requiring antimicrobial therapy was low in this cohort. However, there was the notable exception of 1 instance of fatal PCP that occurred after discontinuation of prophylaxis and had a fatal outcome. All other patients had been maintained with PCP prophylaxis. We are aware of 1 other reported case of PCP following RTX therapy (32). In contrast to our patient, that patient was also receiving methotrexate and prednisone at the time PCP was diagnosed (32). The occurrence of PCP after RTX infusion serves as a reminder that RTX is an immunosuppressive agent, and its effects may not simply be limited to B lymphocytes (33). The function of T lymphocytes may also indirectly be affected by prolonged and repeated B cell depletion (33). Although PCP occurs primarily in patients with abnormal T lymphocyte function or low counts (34), the importance of B lymphocytes in host defense mechanisms against Pneumocystis carinii is well documented in mice (35–37). For these reasons, we recommend to maintain PCP prophylaxis for at least the duration of B lymphocyte depletion following RTX infusions in patients with GPA.

Our study has several limitations. First, this was a historical cohort study with inherent limitations, including open-label administration of RTX and the fact that not all laboratory data were collected uniformly or in a protocolized manner. Second, this was a single-center experience with a midwestern Caucasian population of predominantly Scandinavian and Northern European ethnic background. Third, only patients with chronic relapsing GPA whose disease could not satisfactorily be controlled with CYC and standard remission-maintenance agents were treated with RTX. Except for 1 patient, all patients were PR3 ANCA positive and had previously exhibited a correlation between PR3 ANCA levels and disease activity. Our results cannot therefore be extrapolated to patients with other ethnic backgrounds, patients with MPO ANCA, or patients in whom the condition is newly diagnosed. Fourth, the capture of adverse events may be incomplete. Except for the first RTX infusion, not all RTX infusions were administered at our institution. Minor infusion reactions not leading to discontinuation of therapy may thus have been underreported. Similarly, mild infections not requiring hospitalization may not always have been recorded by local health care providers and conveyed to us. It is extremely unlikely, however, that severe infusion reactions or severe infections were not brought to our attention by either the patients or their local providers, and thus recorded in the Mayo Clinic electronic medical record. Late-onset neutropenia is a well recognized phenomenon occurring after RTX therapy (38). It is usually self-limited and rarely of clinical significance (38). Its occurrence may have remained undetected in some of our patients, as white blood cell (WBC) counts were not measured routinely. Since WBC counts were obtained at each patient visit to our center and at times of recorded infections, it is unlikely that clinically relevant neutropenia remained undetected. Last, our long-term data on immunoglobulin levels are limited. Whether prolonged or repeated B cell depletion for autoimmune disease results in hypogammaglobulinemia-associated increased risk of infection remains unclear (39).

Our study has several important strengths. First, the homogeneity of our population allows us to draw firm conclusions about the use of RTX in patients with relapsing, PR3 ANCA–positive GPA. Second, all patients were evaluated by the same group of physicians with expertise in ANCA-associated vasculitis who followed consistent practice patterns. Furthermore, each disease flare was evaluated at our center by at least 1 of the authors. As a result, re-treatment decisions were standardized, and all flares were treated according to the same RTX dosing regimen and a standardized glucocorticoid taper. Third, B lymphocyte and PR3 ANCA levels were measured consistently according to the same methodology at least quarterly. Last, our results are not confounded by the effects of other concomitantly used immunosuppressant agents and thus reflect what can be achieved with RTX alone in chronic relapsing GPA.

In conclusion, RTX appears to be effective and safe for induction and maintenance of remission in patients with chronic relapsing refractory GPA. Repeated and prolonged B lymphocyte depletion seems to be associated with a low risk of infections, but PCP prophylaxis for the duration of B lymphocyte depletion is justified. Timing of re-treatment can be individualized based on serial B lymphocyte and PR3 ANCA determinations in these patients. Our results provide important information for the design of further trials conducted to determine the optimal timing and dosing of preemptive re-treatments with RTX that assure optimal efficacy and long-term safety.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Specks had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Cartin-Ceba, Golbin, Keogh, Peikert, Sánchez-Menéndez, Ytterberg, Fervenza, Specks.

Acquisition of data. Cartin-Ceba, Golbin, Keogh, Ytterberg, Specks.

Analysis and interpretation of data. Cartin-Ceba, Golbin, Peikert, Fervenza, Specks.