Optimizing Pemphigus Management With Rituximab and Short-Term Relapse Predictors

Key Points

Question

Is it possible to prevent early relapses after rituximab use for treatment of moderate to severe pemphigus?

Findings

This cohort study of 87 patients diagnosed with pemphigus who were treated with rituximab found that an additional rituximab infusion at month 6 in patients with predictive factors for early relapse was associated with reduced relapse rate, from 17.6% (reference study) to 2.6% in year 1.

Meaning

This finding suggests that factors predictive for relapse may be used to identify patients at risk and that an additional rituximab infusion at month 6 in high-risk patients may effectively prevent early relapse.

Abstract

Importance

Rituximab is approved for the treatment of moderate to severe pemphigus. However, 20% of patients in the RITUX 3 trial relapsed within the first year of treatment.

Objective

To assess the outcome of an additional rituximab infusion at month 6 in patients with pemphigus who were in complete remission (CR) after rituximab regimen but had 1 or more predictors of relapse at month 3.

Design, Settings, and Participants

This multicenter cohort study was conducted in France from September 2018 to June 2023 to assess patients with newly diagnosed pemphigus who were in CR after treatment with the RITUX 3 regimen but had predictors of relapse at month 3. Relapse factors were a Pemphigus Disease Area Index (PDAI) score of 45 or higher, desmoglein 1 (DSG1) antibodies greater than 20 IU/mL, and/or DSG3 antibodies greater than 130 IU/mL.

Exposure

Patients in CR at month 6 with at least 1 predictor of relapse were treated with an additional rituximab infusion at month 6.

Main Outcomes and Measures

Primary end point was the rate of CR without corticosteroid therapy for 2 months at month 12. Secondary end points were the rate of relapse, number of patients needing to be re-treated (NNT) with rituximab to avoid a relapse, and safety.

Results

The study population comprised 87 patients (44 females [50.6%] and 43 [49.4%] males), with a mean (SD [range]) age of 55.3 (15.2 [24-92]) years at pemphigus diagnosis. Of these, 64 patients (73.6%) had pemphigus vulgaris and 23 (26.4%) had pemphigus foliaceus. At month 6, CR had been achieved by 77 patients (88.5%), and 10 (11.5%) had persistent disease activity. Of the 77 patients in CR, 30 (39.0%) had at least 1 predictor of relapse and received an additional infusion of rituximab; 47 patients (61.0%) without a predictor did not. Two patients without a predictor and no patients with a predictor experienced relapse—an overall relapse rate of 2.6% and an NNT of 3.6 (95% CI, 1.6-46.5). The 10 patients (11.5%) with persistent disease activity at month 6 were re-treated with rituximab, 2000 mg. At month 12, the rate of CR without corticosteroid therapy for a minimum of 2 months was 72 of 77 (93.5%) among patients who had achieved CR at month 6, and 72 of 87 (82.7%) for the whole study population. Eight serious adverse effects were reported among 5 patients; there were no deaths.

Conclusion and Relevance

This multicenter cohort study indicates that using predictors such as baseline PDAI score, anti-DSG1 antibodies, and/or anti-DSG3 antibodies to initiate preemptive treatment with additional rituximab may reduce the rate of short-term relapse.

This multicenter cohort study assesses the association of an additional rituximab infusion at month 6 in patients with pemphigus who were in complete remission and had at least 1 predictor of relapse.

Introduction

Pemphigus is a rare and life-threatening autoimmune bullous disease that affects the skin and mucosa. It is characterized by pathogenic immunoglobulin G (IgG) antibodies directed against 2 desmosomal adhesion proteins, desmoglein (DSG) 1 and DSG3.^1,2,3,4 The Comparison Between Rituximab Treatment and Oral Corticotherapy Treatment in Patients With Pemphigus trial (RITUX 3; NCT00784589)⁵ demonstrated the higher efficacy of combining rituximab with a short-term oral corticosteroid (CS) regimen compared to the standard regimen of high doses of CS alone. However, at 1 year posttreatment, nearly 20.0% of the patients in the rituximab group had experienced a relapse within the first year of treatment in this latter study, while 67.4% were in complete remission (CR) off CS therapy after 1 year of treatment. Most relapses occurred between the sixth and the twelfth month after the first cycle of rituximab, whereas only 3 relapses were observed during the second year of follow-up after 2 maintenance infusions of rituximab performed at months 12 and 18, as scheduled by the trial. Consistent reports in the literature^6,7,8,9,10 have observed the occurrence of relapse after an initial cycle of rituximab in patients who were not receiving maintenance infusions of rituximab. Moreover, the PEMPHIX trial,¹¹ which compared the efficacy of rituximab vs mycophenolate mofetil, used a strategy in which all patients were systematically re-treated at month 6 with rituximab, 2000 mg, after the initial cycle.¹² This strategy may have led to overtreatment of patients.

To identify patients at a high risk of early relapse after rituximab treatment, we previously conducted a post hoc analysis¹³ of the RITUX 3 trial.⁵ This analysis identified 3 predictors of relapse evident at month 3: a baseline PDAI score of 45 or higher, anti-DSG1 antibodies (per enzyme-linked immunosorbent assay [ELISA] results), greater than 20 IU/mL, and/or anti-DSG3 antibodies greater than 130 IU/mL. These criteria yielded a positive predictive value of 50% and a negative predictive value of 94% for relapse occurrence during the first year of treatment.¹³ In particular, the 94% negative predictive value suggested that a maintenance infusion of rituximab could reasonably be avoided in patients without any of these predictors. These results led us to test a personalized treatment strategy based on these predictors, which consisted of performing an additional infusion of rituximab, 500 or 1000 mg, at month 6 in patients who had achieved CR and had at least 1 predictor of early relapse. No additional maintenance rituximab infusion at month 6 was performed in patients without any predictors. Additionally, patients who were not in CR and had persistent disease activity after the initial cycle or rituximab were treated with an additional cycle of rituximab, 2000 mg, at month 6.¹⁴

The primary objective of this study was to assess the rate of 1-year CR without CS for at least 2 months in patients with pemphigus who had achieved CR after the initial rituximab cycle, and who were subsequently treated according to whether they exhibited 1 or more predictors of early relapse.

Methods

This study was reviewed and approved by the Comité de Protection des Personnes Sud-Ouest et Outre-Mer 2 (2-20-039, No. 2020-A01312-37), and written informed consent was obtained. The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

Study Design and Patients

This multicenter cohort study was conducted in 12 dermatology departments in France from September 2018 to June 2023. Inclusion criteria were the same as in the RITUX 3 trial. We enrolled consecutive adult patients with newly diagnosed moderate to severe pemphigus who met the following criteria: clinical features suggestive of pemphigus vulgaris or pemphigus foliaceus; a histologic image of intraepidermal acantholysis; and deposition of immunoglobulin G, complement component 3, or both on keratinocyte membrane, detected by direct immunofluorescence (Figure).

Figure. Flow Diagram of Study Patients.

CR indicates complete remission; CS, corticosteroid; and PDAI, Pemphigus Disease Area Index score (≥45, severe; 15-44, moderate; and <15, mild).

Pemphigus severity was assessed using the Pemphigus Disease Area Index (PDAI).^15,16,17 A PDAI score less than 15 defined mild pemphigus; 15 to 44, moderate pemphigus; and 45 or greater, severe pemphigus. According to French and European guidelines,^1,14 we also included patients with an initially mild pemphigus (characterized by an initial PDAI score <15) who did not respond or worsened despite a 3-month treatment regimen with topical CS and dapsone.

All patients received 1 infusion of intravenous rituximab, 1000 mg, on days 1 and 14, combined with a short-term prednisone regimen: 0.5 mg/kg daily (mg/kg/d) tapered over 3 months for mild or 1.0 mg/kg/d over 6 months for moderate to severe pemphigus. Patients with at least 1 predictive factor of relapse (baseline PDAI score ≥45 and/or anti-DSG1 antibodies [per ELISA] >20 IU/mL and/or anti-DSG3 antibodies >130 IU/mL at month 3 posttreatment) received an additional infusion of rituximab, 500 or 1000 mg, at month 6. Considering that no study, to our knowledge, has compared the efficacy of a 500 vs 1000 mg dose of rituximab as maintenance infusion at month 6, and because the available guidelines¹⁴ do not recommend either dose over the other, investigators could choose the dosage (500 or 1000 mg). Patients without any predictors of relapse did not receive an additional infusion of rituximab at month 6; instead, they were monitored until either experiencing a relapse or reaching month 12, which was the primary end point. Additionally, patients with persistent clinical disease activity after the initial cycle received a second cycle of rituximab, 2000 mg, at month 6.

Clinical Assessments

Patients were followed up on days 1 and 14, and months 1 (evaluation of disease control), 3 (measurement of serum anti-DSG1 and 3), 6 (additional rituximab infusion in patients in CR with ≥ 1 predictor of relapse or with persistent clinical disease activity), and 12 (end of follow-up). Clinical data were recorded according to the consensus statement definitions, and included CR without CS for 2 months or more (primary end point); CR without CS for less than 2 months; CR receiving minimal CS (prednisone dose <10 mg/day); CR with nonminimal dosage of CS (prednisone dose ≥10 mg/day); and relapse.¹⁸ The doses of rituximab infused at M6, and the occurrence of non-severe and severe adverse events (SAEs) were also recorded.¹⁹

Serological Assessments

Serum samples were collected at baseline and at months 3 and 6. Detection of anti-DSG1 antibodies and anti-DSG3 IgG antibodies was performed centrally using ELISA, following the manufacturer’s instructions. Values equal to or greater than >20 IU/mL for anti-DSG1 antibodies and anti-DSG3 antibodies were deemed positive. However, thresholds of 20 IU/mL and 130 IU/mL, respectively, were considered relevant for predicting short-term relapse.

Outcomes

The primary end point was the rate of CR without CS therapy for 2 months or more at month 12 in patients who had achieved CR after the initial cycle of rituximab, and were treated according to the tested strategy using the predictors of early relapse. These predictors were identified previously through our ad hoc study¹³ of a randomized clinical trial.⁵ Secondary end points were (1) the relapse rate during the 12-month follow-up in patients who had achieved CR after the initial cycle of rituximab, (2) the number of patients needing to be re-treated with rituximab (NNT) to avoid a relapse, (3) the rate of CR without CS for 2 months or more at month 12 in the whole population; and (4) the rate of treatment-related adverse events.

Statistical Analysis

The calculation of the number of patients to be included was based on a 20% rate of relapse during the year after the initial cycle of rituximab, as observed in the RITUX 3 trial,⁵ and an expected 7% rate of relapse with the tested strategy. Seventy-four patients needed to be included for a statistical power of 90% and an α risk of 5%. The NNT was estimated by comparison with the RITUX 3 trial with corresponding 95% CIs.

Statistical tests were 2-tailed and P < .05 was considered statistically significant. The analyses were conducted using SAS, version 9.4 (SAS Institute).

Results

The study population comprised 87 patients (44 females [50.6%] and 43 [49.4%] males), with a mean (SD [range]) age of 55.3 (15.2 [24-92]) years at pemphigus diagnosis, who were initially treated with rituximab plus short-term oral prednisone (Table). All patients completed the 12-month follow-up, and none withdrew from the study. Among the 87 patients, 64 (73.6%) had pemphigus vulgaris, and 23 (26.4%) had pemphigus foliaceus. At baseline, the mean (SD) PDAI score was 29.7 (19.0). Regarding pemphigus severity, 21 patients (25.0%) had an initial PDAI score of 45 or higher (severe) and 63 patients (75.0%), had a score less than 45. The latter group included 41 patients with moderate (PDAI 15-44) and 22 with mild pemphigus (PDAI <15) whose disease was not controlled despite a 3-month treatment with topical applications of clobetasol propionate cream combined with dapsone. A baseline PDAI score was not available for 3 patients; given that their oral prednisone dosage was tapered over 6 months (same as for patients with moderate to severe pemphigus), they were analyzed accordingly.

Table. Baseline Characteristics of Patients With Pemphigus.

Characteristic	Patients, No. (%)
Total, No.	87
Age, mean (SD), y	55.3 (15.2)
Female	44 (50.6)
Male	43 (49.4)
Pemphigus type
Foliaceus	23 (26.4)
Vulgaris	64 (73.6)
Pemphigus severitya
Severe	21 (25.0)
Moderate	41 (48.8)
Mild	22 (26.2)
Karnofsky score,b mean (SD)	89 (10)

^a Determined from Pemphigus Disease Area Index score: ≥45, severe; 15-44, moderate; and <15, mild. Three were missing data.

^b Karnofsky performance status is a standardized measure of a patient’s ability to perform various ordinary tasks (from 0 to 100, with higher scores indicating greater function/less impairment).

Clinical Status at Month 6

Of the 87 patients (treated with rituximab plus short-term oral CS), CR was achieved by month 6 in 77 patients: 17 with mild pemphigus who had discontinued CS therapy for 2 months or longer; 15 who had discontinued CS for less than 2 months; 32 receiving minimal CS therapy; and 13 receiving prednisone dosage greater than 10 mg/d. Predictive factors of short-term relapse were tested in this subpopulation of 77 patients. Ten patients (11.5%) had persistent clinical disease activity at month 6 and were re-treated with a higher dose cycle of rituximab, 2000 mg.

Disease Course of Patients in CR at Month 6 Based on Relapse Predictors

Of the 77 patients who achieved CR at month 6, 30 patients (39.0%) had at least 1 predictor of relapse: 16 had a baseline PDAI score of 45 or higher; 10 had either anti-DSG1 antibody (ELISA) greater than 20 IU/mL or anti-DSG3 antibodies greater than 130 IU/mL at month 3; and 4 patients had both clinical and serological predictors. Therefore, 14 patients continued to have positive results for anti-DSG antibodies between months 3 and 6, and 3 patients had positive results for both anti-DSG1 and anti-DSG3 antibodies. The mean (SD) anti-DSG1 and anti-DSG3 antibody values were 51 (21) IU/mL and 186 (44) IU/mL, respectively. These 30 patients with at least 1 predictor of relapse were re-treated with a maintenance infusion of rituximab at month 6, including 14 patients who received 500 mg, and 16 patients who received 1000 mg. None of these 30 patients relapsed by month 12.

The 47 patients (61.0%) who were in CR at month 6 and had no predictors of relapse were not re-treated until the first maintenance infusion of rituximab, 500 mg, scheduled at month 12 (consistent with the RITUX 3 regimen⁶). Two of these patients relapsed between months 6 and 12, and were re-treated with a cycle of rituximab, 2000 mg. One of these 2 patients was in CR at month 12 and was still receiving prednisone, 20 mg/d. The other patient relapsed a few days before the month 12 evaluation, and consequently, was considered to have active disease at month 12.

Overall, 2 of the 77 patients (2.6%) who were treated according to the predictors relapsed between months 6 and 12, and 76 of 77 (98.7%) finally achieved CR at month 12. Among them, 72 patients (93.5%) were in CR and not receiving CS for at least 2 months; 3 patients were in CR and had discontinued CS for less than 2 months, and 1 patient was in CR and receiving a CS dose greater than 10 mg/d. When considering the 1-year relapse rate of 2.6% in the present study and the 17.6% rate of the RITUX 3 trial, the NNT to avoid 1 relapse using the predictors was 3.6 (95% CI, 1.6-46.5).

Disease Course of Patients With Active Disease at Month 6

Ten patients with persistent clinical pemphigus activity at month 6 received an additional cycle of rituximab, 2000 mg. Eight of them were in CR at month 12, including 6 patients in CR not receiving CS for less than 2 months, and 2 patients in CR on minimal CS therapy. The other 2 patients still had mild disease activity at month 12. Interestingly, these 2 patients both had a mild type of pemphigus initially, and therefore, were treated with rituximab after the failure of an initial treatment with dapsone and topical CS.

Overall Proportion of Patients in CR With Tested Treatment Strategy

When considering the entire population, including patients in CR and those with persistent clinical disease activity at month 6, 84 of 87 patients (96.5%) were in CR at month 12. Among them, 72 patients were in sustained CR without CS for at least 2 months; 9 in CR without CS for less than 2 months; 2 in CR on minimal CS therapy; and 1 in CR with a prednisone daily dosage of 20 mg. The median (IQR) time to achieve CR without CS therapy was 197 (168-252) days. Among the 3 patients who did not achieve CR at month 12, 1 patient relapsed between months 6 and 12 after having achieved CR. The other 2 patients initially had a mild type of pemphigus that did not respond to dapsone; they had persistent disease activity throughout the follow-up period despite receiving 2 infusions of rituximab, 2000 mg, at baseline and month 6.

Safety End Points

No patients died and none withdrew for safety reason. Eight nonsevere adverse effects were recorded in 6 patients during the first rituximab infusion: 2 experienced hyperthermia; 2, hypotension; 1, asthmatic dyspnea; and 1, skin rash. After the infusion of rituximab, there were 2 additional adverse effects: 1 patient experienced transient blood eosinophilia, and another, transient mild neutropenia.

There were 8 rituximab-related severe adverse effects (per the researchers’ opinion) reported: SARS-CoV-2 infection, 1 patient; obstruction of the retinal central vein, 1; papulopustular skin rash, 1; and psychiatric decompensation, 1. Additionally, 1 patient experienced 4 severe adverse effects: cytomegalovirus gut infection, campylobacter bacteriemia, Pneumocystis carinii pneumonia, and aspergillosis—all of which occurred during the first 6 months after the start of treatment (ie, before any further additional infusion of rituximab).

Discussion

This observational study validated the use of predictors of early relapse that we previously identified in a post hoc analysis¹³ of the RITUX 3 trial.⁵ Indeed, this series shows that use of these predictors was associated with reductions in early relapse, from 17.6% in the RITUX 3 trial (reference) to 2.6% in this study. Overall, this strategy allowed 93.5% of patients who were in CR at month 6 to achieve a CR without CS for at least 2 months at month 12, while re-treating with an additional infusion of rituximab for only 39% of patients, those with a high risk of relapse. According to the 50% positive predictive value of the predictors that we previously reported,¹³ 15 relapses were expected in the present study (ie, 50% of the 30 patients with the predictors), corresponding to an expected frequency of relapse of 15 of 77 (19.5%). This rate is very close to the rate reported in the RITUX 3 trial (8 of 46 patients [17.4%]). Lastly, in our present study, only 2 of 77 patients (2.6%) relapsed before month 12. This finding suggests the substantial usefulness of these predictors in clinical practice to avoid the occurrence of relapses.

Importantly, none of the 30 patients with 1 or more predictors relapsed after the maintenance rituximab dose received at month 6, regardless of the dosage (500 or 1000 mg). This finding validated our hypothesis that an additional infusion of rituximab at month 6 would be associated with prevention of early relapse in patients with the predictors.

The hypothesis of the effectiveness of a rituximab infusion at month 6 was based on the drug’s mechanism of action. Indeed, rituximab, an anti-CD20 monoclonal antibody, induces a rapid (within a few days), deep (less than 0.2% of B lymphocytes among total lymphocytes), and long-lasting (6 months) depletion.^5,6 At approximately month 6, B lymphocytes begin to reemerge from the bone marrow, so it seemed logical to us to resume infusions at this time point. Moreover, we believe this is why the majority of relapses in the RITUX 3 trial occurred between months 6 and concomitantly with the reappearance of anti-DSG antibodies. The role of B-lymphocyte resurgence was subsequently reduced by repeated rituximab infusions at months 12 and 18. Additionally, several pathophysiologic mechanisms may explain the prolonged remissions (or even cures) observed after rituximab use. These may involve various phenomena, such as the disappearance of certain pathogenic B-cell clones, a decrease in the number of anti-DSG IgG isotypes, a reduction in the expression of BAFF receptors on B lymphocytes, or the emergence of autoreactive follicular regulatory T cells.^{6,20,21,22,23} On the other hand, the mechanisms involved in explaining late relapses are more complex to elucidate, particularly due to the rarity of late relapses after a prolonged remission period.

Because no study, to our knowledge, has determined the optimal dosage of rituximab for patients with a high risk of relapse, investigators were free to choose between a 500 or 1000 mg dose of rituximab. Our results suggest that 500 mg may be sufficient to maintain CR in most patients with a high risk of relapse. Considering that 39% of patients in CR after the initial cycle of rituximab had to be re-treated at month 6, the efficacy of reduced maintenance doses reinforces the interest of this personalized treatment strategy compared to the higher doses used in the PEMPHIX trial (2000 mg at month 6).¹¹

When accounting for 10 patients who were not in CR at month 6 and were re-treated with rituximab, 2000 mg, a total of 40 of the 87 patients (46.0%) were re-treated at month 6. In total, at month 12, 72 of the 87 patients (82.7%) were in CR without CS for at least 2 months at month 12, a higher rate than in the RITUX 3 trial (67.4%).⁶ This personalized treatment strategy based on predictors of relapse could both decrease the cost of treatment—because 54.0% of patients were not re-treated with rituximab at month 6)—and limit the immunosuppression related to repeated infusions of rituximab. Additionally, the efficacy of an additional infusion of rituximab to maintain CR in patients with a high risk of relapse avoided the resumption of CS, which has been demonstrated to have deleterious effects.^5,24 Indeed, 8 nonsevere adverse effects (mostly infusion-related reactions) and 8 serious adverse effects were recorded during the present study—a low-frequency rate of 0.1 serious adverse effects per patient.

Furthermore, our study confirmed the effectiveness of re-treatment with 1 additional cycle of rituximab, 2000 mg, in patients with persistent disease activity at month 6. This is suggested by the finding that 8 of the 10 re-treated patients achieved CR without or with minimal CS therapy at month 12.

The use of 2000 mg of rituximab to treat patients with active pemphigus is recommended by the European guidelines on pemphigus.¹ The choice of this dosage, rather than a lower dose (500 or 1000 mg), is based on the experience of the French task force for autoimmune bullous diseases, as well as unpublished scientific data showing incomplete B-cell depletion in several patients who were found to be rituximab-recalcitrant after rituximab, 2000 mg.

Patients were followed up for 12 months in the present study, which may be considered a short follow-up period. However, in accordance with the RITUX 3 regimen, all patients were subsequently treated with 2 maintenance infusions of rituximab, 500 mg, at months 12 and 18—previously demonstrated to be effective in preventing the occurrence of delayed relapses.^5,25

Indeed, the recent results of the RITUX 3 extension study²⁵ showed that the strategy of using systematic rituximab infusions at baseline and months 12 and 18 for moderate to severe pemphigus achieved complete remission in 93% of patients. After a mean follow-up of more than 7 years, ie, more than 5 years after treatment discontinuation, relapses were observed in only 43% of patients. Thus, a last important point to address is the prevention of long-term relapses. In the extension study,²⁵ using the same methods as for early relapses, we observed that an increase in anti-DSG1 and anti-DSG3 antibodies yielded a very high positive predictive value (83.0%) for relapse, suggesting the possibility of re-treating patients with rituximab only when their serum anti-DSG ELISA values exceed the thresholds we determined: a combination of ELISA values of anti-DSG1 antibodies of 20 IU/mL or more, and/or anti-DSG3 antibodies of 48 IU/mL. Furthermore, the 94% negative predictive value provided by these thresholds offers confidence in not re-treating patients whose anti-DSG antibody levels are below the thresholds, which applies to most patients treated with the RITUX 3 regimen as a first-line therapy.

We are currently conducting a randomized clinical trial (NCT05898308) to test the strategy of re-treating patients at high risk of relapse based on anti-DSG antibody ELISA values to prevent clinical relapses and avoid the need for patients to restart CS therapy.

Limitations

The main limitation was the observational, noninterventional study design, which led to substantial differences in the management of the disease in these patients, particularly regarding the tapering speed of CS therapy used with rituximab. Although CS tapering was well described in the European guidelines,¹ some investigators modified them per the patient’s disease course or their clinical experience. Therefore, some patients were still on low-dose CS therapy (>10 mg daily) at month 6, whereas the dose should have been 5 mg or less by that time. However, we believe that this limitation did not undermine the value of using these predictive factors and the potential benefit of an additional rituximab infusion at month 6.

Conclusions

The findings of this cohort study validate the use of predictors of early relapse after an initial cycle of rituximab in patients with pemphigus. When baseline PDAI score is 45 or greater, anti-DSG1 antibodies are more than 20 IU/mL, and/or anti-DSG3 antibodies are more than 130 IU/mL in months 3 to 6, an additional maintenance infusion of rituximab should be proposed. The administration of this additional rituximab infusion at month 6 according to these predictors would allow a significant reduction in the rate of short-term relapse in patients with moderate to severe pemphigus.

Supplement.

Data Sharing Statement