Rituximab use as induction therapy for lupus nephritis: a systematic review

RCTs that compared RTX monotherapy or combination therapy to placebo or standard therapy

The renal clinical and serologic responses to RTX therapy are summarized in Table 1; the grouped results based on the categories of studies are presented in Table 3. Overall, the majority of the patients studied attained a complete renal response (CRR) or partial renal response (PRR) to RTX therapy; indexes for clinical renal response were determined by creatinine level, urine sediment, and proteinuria (see Table 2). However, due to the limited number of RCTs, some of the studies compared RTX monotherapy to the standard treatment or placebo, whereas in other studies, the experimental group utilized RTX as adjunctive induction therapy. Table 1 summarizes the doses of RTX employed and which therapies were used concurrently, as well as the renal response rates in patients who received RTX as either monotherapy or adjunctive induction therapy.

OPEN IN VIEWER

Table 3.

Summary of grouped results.

Types of studies	Number of studies/total patients	Comparisons	Ethnicities	Significant findings
RCTs comparing RTX monotherapy or combination therapy to placebo or standard therapy	3 studies,12,14,17 n = 247	RTX to placebo (with MMF and steroids), RTX vs. RTX+CYC, RTX+CYC vs. CYC	Caucasian, Hispanic, African American, and East Asian	Improvements in C3 and C4, anti-dsDNA levels, decreased anti-C1q and ANCANo difference in renal response with addition of CYC to RTX
Prospective studies that directly compared use of RTX monotherapy to MMF or CYC	2 studies,18,19 n = 276	RTX vs. MMF vs. CYC	Indian, Caucasian	RTX was as effective as CYC and MMF in achieving clinical renal responseComplete response averaged 72.3%, partial response averaged 20.4%Decrease in creatinine with RTX treatment was not statistically significant between the groupsDecrease in proteinuria was statistically significant in the RTX group itself when comparing pretreatment levels with post-treatment levels, but it was not statistically significant in comparisons with other groups
Prospective and retrospective studies that investigated the effectiveness of rituximab as monotherapy without comparisons with other immunosuppressive agents	6 studies,20,22–26 n = 227	RTX monotherapy in patients refractory to prior induction therapy with other agents (CYC, MMF, and AZA)	White Caucasian, East Asian, Hispanic, and patients of African descent	Clinical renal improvements with RTX, complete or partial renal response in 65.8% of patientsIn the Iwata et al. 23 and Chavarot et al. 24 studies, the improvement in creatinine was not clinically significant from 0.99 to 0.85 mg/dL in the Iwata et al. 23 study and remained stable at 0.7 mg/dL in Chavarot et al. 24 study; when taking into account weighted averages of four studies, the daily proteinuria (as designated by both urine protein-to-creatinine measure as a surrogate measurement and daily protein excretion) decreased from 3.43 to 0.45 g
Studies investigating how RTX induction with MMF maintenance impacted oral steroid therapy	2 studies,27,28 n = 68	RTX induction with MMF maintenance	White Caucasian, Indian, African descent, and East Asian patients	Complete renal response in 47.1%, partial response in 33.8%Statistically significantly reduced use of oral steroids during the maintenance period (the average doses of steroids were not specified) Overall, there were no significant changes in creatinine in both studies
Prospective and retrospective studies that examined the effect of RTX in combination with other immunosupp-ressives	11 studies,29–39 n = 330	RTX with CYC or MMF induction	African descent, Caucasian, East Asian, and Hispanic patients	In studies that evaluated clinical renal response rate, there was complete or partial renal response in 65.5% of patientsIn four studies that had specific outcomes of daily proteinuria, the average improvement was from 3.72 to 0.75 gIn four studies, the average serum creatinine changed from 1.34 to 1.29 mg/dL
The serologic, immunologic, and histologic benefit of RTX added to other immunosuppressives used concurrently	5 studies,40–44 n = 101	RTX added to concurrent immunosuppressive regimen, which included CYC in most patients	East Asian, Caucasian, and Hispanic patients	In studies that evaluated clinical renal response rate, 77.8% of patients had complete or partial responsesTwo studies40,43 showed significant improvements in histology resultsThere was a relation between clinical renal response and B-cell depletion
Immunologic responses to RTX monotherapy; relation to clinical response	2 studies,45,46 n = 10	RTX monotherapy	Greek	The degree of B-cell depletion obtained was related to the clinical renal response to RTX

When comparing RTX to placebo (with concomitant use of MMF and corticosteroids), the Lupus Nephritis Assessment with Rituximab (LUNAR) study ¹² found that there was not a superior renal response rate with RTX. Although the LUNAR trial did not meet its primary endpoint (a statistically significantly higher percentage of patients who obtained a CRR and PRR with RTX), there was a significant serologic benefit in lupus control, as those in the RTX group had statistically significant improvements in C3 and C4, and in anti-dsDNA levels, as well as a clinical benefit seen in patients of African descent. Since 46% of the patients in the LUNAR trial had a prior episode of LN, this confounds the applicability of the study in determining how RTX functions as an independent induction agent. In addition, the patients in the treatment group had a longer median duration of disease (11.1 months) compared to the control group (5.4 months), which further confounds the results.

As part of further data analysis of the LUNAR trial, Gomez Mendez et al. ¹³ investigated the results of 68 patients who were treated with RTX in order to determine if there was a relationship between B-cell depletion and renal response. They found that the achievement of complete peripheral depletion was statistically significantly associated with renal response, with an odds ratio of 5.8 (95% confidence interval 1.2–28; p = 0.03). These subgroup analyses of the data from the LUNAR trial may offer insight into an immunologic mechanism by which RTX could effect a clinical renal response.

Another RCT that included 84 patients ¹⁴ showed decreased levels of anti-C1q and ANCA in those who received RTX with CYC compared to those who received CYC alone. In their randomized study (in a Chinese population), Li et al. ¹⁵ compared the renal response rates of nine patients who received RTX monotherapy for their LN induction with 10 patients who received RTX with CYC. There were no statistically significant differences in renal response rates (as determined by clinical response as well as histological response on biopsy) or in B-cell depletion at 48 weeks.

Overall, these three RCTs, with Caucasian, Hispanic, African American, and East Asian participants, did not demonstrate a clinical renal benefit of RTX compared to placebo, but in two of those studies, there were significant improvements in serologic levels. Of the three RCTs, only one ¹⁴ demonstrated a clear clinical benefit of RTX therapy added to CYC when compared to the control group. However, in that trial, it is difficult to conclude how RTX would function as an independent induction therapy, since RTX was used in conjunction with CYC to provide its clinical benefit. There is a precedent for using RTX in combination with CYC based on protocols that were used in RA¹⁶ and later in SLE. ¹⁷ The rationale for this regimen was for the CYC to help achieve higher levels of B-lymphocyte depletion.

Prospective studies that directly compared the use of RTX monotherapy to other standard modalities

Moroni et al.¹⁸ assessed the renal response to induction therapy with RTX compared to MMF or CYC induction regimens. They found that RTX induction therapy was as effective as therapy with MMF or CYC in achieving renal remission and reducing levels of proteinuria among mainly Caucasian patients. However, in this study, the patients in the RTX group had a longer duration of SLE and LN, and more flares prior to the study when compared to both the MMF and CYC groups. A large prospective study in India¹⁹ that involved 222 patients compared low-dose CYC, high-dose CYC, MMF, and RTX in the induction phase of LN treatment. There was a statistically significantly higher renal response rate in the high-dose CYC and RTX groups compared to the low-dose CYC and MMF groups. However, the limitation with this study was that most patients in the RTX group received either MMF or CYC concurrently (see Table 1).

Although these two direct-comparison studies (where LN was WHO class III–V) did not show a clear benefit of RTX over traditional therapy for LN, they did demonstrate that RTX could be a viable alternative. However, with the inconsistencies in the baseline characteristics of the patient populations and the concurrent treatments offered to the patients while receiving RTX therapy, it is difficult to reach a definitive conclusion about the efficacy of RTX.

Prospective and retrospective studies that investigated the effectiveness of RTX as monotherapy without comparisons with other immunosuppressive agents

Much of the existing data about RTX use as monotherapy included patients who had been refractory to more common pulse induction therapy regimens, namely CYC. A large multicenter study²⁰ at 11 Italian centers found that among patients with LN who were refractory to standard therapy (which included prior exposure to immunosuppressive agents such as CYC, MMF, and azathioprine), at the 12-month follow-up, 30.9% of patients had a CRR and 63.2% had a PRR according to the European League Against Rheumatism (EULAR) criteria.²¹ The study was conducted in patients who were refractory to prior therapy, making it difficult to assess the true impact of RTX. Also, several patients in the study were re-treated with RTX due to relapse in renal disease.

In open-label prospective studies that specifically examined the use of RTX, there were also clinically significant renal disease activity response rates. For example, Tanaka et al. ²² found that there were response rates of 58.8% and 52.9% by American College of Rheumatology criteria and the LUNAR study criteria, respectively, in 17 Japanese LN patients receiving RTX induction therapy. In their study, Iwata et al. ²³ found that 83% of the 36 LN patients (who were refractory to prior immunosuppressive therapy) who obtained treatment with RTX showed a change in disease activity from British Isles Lupus Assessment Group (BILAG) A or B to C or D in one year. In a retrospective study by Chavarot et al., ²⁴ of 15 patients with membranous LN who received RTX as monotherapy for induction (who had also received prior regimens including CYC and MMF), there was disease remission in 13 with CRR in eight of those patients.

Contis et al. ²⁵ further expanded on this concept, specifically investigating the effectiveness of RTX in 17 patients who had LN that was refractory to prior treatment with CYC. They found that at one year, 53% of patients achieved a CRR or PRR. Interestingly, Arce-Salinas et al. ²⁶ found that among eight patients who were given RTX induction therapy (who were refractory to prior induction with CYC or MMF), although initially all of the patients had clinically inactive renal disease at six months, at two years, three of the patients had worsening or unimproved renal disease.

The studies that specifically investigated RTX as monotherapy demonstrated that in a shorter time frame (of about one year), at least half of the patients had CRR or PRR. However, as indicated by Arce-Salinas et al., ²⁶ at longer follow-up times, the benefits were not clear and therefore further investigation is warranted. The collective results from these six studies, all of which had patients who were refractory to standard induction immunosuppressive therapy, showed clinical renal improvements with RTX use in at least half of the patients. The patient populations were diverse and included predominantly Caucasian, East Asian, and Hispanic patients, with a few patients of African descent (in the study by Chavarot et al. ²⁴ ).

Studies that investigated how RTX induction therapy with MMF maintenance impacted the use of oral steroid therapy

Condon et al. ²⁷ sought to determine the effect of RTX induction therapy with MMF maintenance for LN treatment (they had a diverse population, with the predominant ethnic groups including mostly Caucasian, Indian, African descent, and East Asian patients). Their prospective study evaluated 50 patients who received what is known as the rituxilup protocol (two doses of RTX and methylprednisolone given two weeks apart) with subsequent maintenance with MMF. The results were significant for CRR in 52% of the patients in 52 weeks and PRR in 34% of the patients. Only two patients ended up requiring more than two weeks of oral steroids during their maintenance therapy.

In their prospective study of 18 patients with class III/IV/V LN who received RTX induction therapy with MMF maintenance, Pepper et al. ²⁸ similarly found that at one year, 67% had sustained CRR or PRR. Although both studies had patients with relapses of disease at later times, they suggested that RTX use can limit subsequent steroid use. In both studies, there was statistically significantly reduced use of oral steroids during the maintenance period. Although these studies provided insight into how RTX in combination with MMF may be beneficial in patients with severe disease, a RCT investigating this regimen compared to either MMF or CYC alone would better elucidate its effectiveness.

Prospective and retrospective studies that evaluated the effect of using RTX as therapy in combination with other major immunosuppressive agents

While there are still limited data on the use of RTX alone as induction therapy in LN, several studies have specifically investigated how RTX could be used with other agents, possibly to create a heightened effect. Table 1 summarizes the doses of immunosuppressants used, the comparison therapies, and adverse events to RTX treatment.

Davies et al. ²⁹ conducted a prospective study of 18 patients with refractory LN who received both RTX and CYC. Of these patients, 11 had a CRR and two had a PRR. Those that did not respond had more severe LN on biopsy at baseline. This analysis underscores the importance of stratifying patients based on the degree of LN. An ideal design for a RCT would be to either have all patients with the same class of nephritis or appropriately stratify patients in order to assess for response. Catapano et al. ³⁰ found similar effectiveness of adding RTX to CYC therapy, with 90.9% of patients achieving CRR or PRR at 30 months. All of the patients in the study had been previously treated with immunosuppressive therapy. Roccatello et al. ³¹ found that with a shorter but more intensive course of RTX and CYC, at 12 months, there were statistically significant improvements in proteinuria.

The participants in one prospective study ³² had previously received induction therapy with CYC for proliferative LN, and subsequently had biopsy-proven relapse of their disease. These patients, in addition to taking MMF and methylprednisolone, received a four-week course of RTX once a week. With the use of RTX, PRR was obtained in eight patients at a median of 3.5 months. Six of those patients had CRR at a median of 38 months, suggesting that the addition of RTX treatment allowed for greater disease response. Another prospective study ³³ found that a single dose of RTX (together with other forms of immunosuppressive therapy, namely CYC, MMF, and azathioprine) was similarly effective. Of the 14 patients in the study, at five months, two patients had a CRR and nine had a PRR. However, five patients relapsed at a median of 17 months. In this prospective study, prior therapies that had been used included CYC, azathioprine, and MMF.

When evaluating prospective data in 13 Hispanic patients with refractory LN, Garcia-Carrasco et al. ³⁴ also found a clinical response with the addition of RTX. There was a decrease in the mean prednisone dose during the subsequent follow-up from 25.7 to 11.4 mg/day. In their retrospective study, Melander et al. ³⁵ analyzed a variety of patients with LN: 12 who were refractory to prior induction therapy (mainly CYC) and two who had not received prior immunosuppressive induction therapy. Three patients received CYC in conjunction with RTX for their induction therapy instead of RTX alone. Of all of these patients, 60% achieved a CRR or PRR. Díaz-Lagares et al. ³⁶ also compared the effect of RTX therapy in refractory patients with different concurrent medications: one group received CYC with corticosteroids, one group received MMF with corticosteroids, and one group received corticosteroids. With the addition of RTX to these regimens, there was a CRR or PRR in 67% of patients at one year.

In their investigation, Jónsdóttir et al. ³⁷ found that among 25 patients who received RTX in addition to their concurrent regimens, which ranged from CYC to MMF to RTX alone, 64% of the patients maintained a CRR at two years. In this study, the patients also obtained renal biopsies to monitor response, and almost all of the patients had a significant reduction in the activity index according to WHO guidelines. These patients had also been exposed to prior immunosuppressive therapy that included CYC, MMF, azathioprine, and cyclosporine.

In a prior study, Jónsdóttir et al. ³⁸ compared the efficacy of RTX in patients with proliferative versus membranous LN, and found that both groups had statistically significant decreases in proteinuria and statistically significant increases in albumin and C3 levels. In the proliferative group, there was also a statistically significant decrease in anti-dsDNA titers. The creatinine level improved in the membranous nephritis group and worsened in the proliferative nephritis group.

Another study indicated that RTX may have had some detrimental effects. Manou-Stathopoulou and Robson ³⁹ retrospectively investigated the effect of RTX in seven patients (with prior therapy that included CYC, MMF, and azathioprine) with LN who had received RTX. As a result, four of those patients developed subsequent worsening in renal function, and one of those patients subsequently received rescue therapy with CYC and high-dose steroids. In these 11 studies (that included patients of African descent and Caucasian, East Asian, and Hispanic ethnicities), the addition of RTX to CYC or MMF was beneficial but resulted in more adverse renal effects in one study.

Studies that explored the serologic, immunologic, and histologic benefit of RTX added to other immunosuppressive regimens used concurrently

Gunnarsson et al. ⁴⁰ investigated the efficacy of a regimen that included both RTX and CYC. There were statistically significant improvements in SLE Disease Activity Index (SLEDAI) score, anti-dsDNA levels, and anti-C1q antibody levels. Repeat renal biopsies in all seven patients showed improvements in their histopathologic results. In their retrospective study, Lateef et al. ⁴¹ found that the addition of RTX to CYC therapy in seven patients with LN resulted in partial or complete renal response in all of the patients at 12 months. All of the patients had complete B-cell depletion within eight weeks of receiving their RTX dose, with a six-month median duration of B-cell depletion. The relation between the degree of B-cell depletion and renal response implied that the immunologic mechanism of the RTX itself likely contributed to this enhanced effect.

Lindholm et al. ⁴² found a similar effect in their study with 17 LN patients who had RTX added to their immunosuppression regimen. All of the patients experienced complete B-cell depletion, and 11 of those patients obtained a CRR or PRR in 6–12 months. Tsanyan et al. ⁴³ investigated the histopathologic response to RTX therapy in two different clinical scenarios: 29 patients who received RTX monotherapy and 16 who received RTX with CYC. Of both groups combined, 16 patients received a repeat renal biopsy after induction therapy. Subsequently, 10 of those patients had an improvement in the histological grade of their LN, with an overall statistically significant improvement in activity index but not chronicity index. A limitation, however, was that in their results, the authors did not delineate which patients received RTX monotherapy and which patients received RTX in combination with CYC.

Vigna-Perez et al. ⁴⁴ researched the clinical response to RTX in 22 patients who had LN that was refractory to other prior immunosuppressive therapy, including CYC and MMF. RTX was utilized as an adjunctive therapy to the current immunosuppression that patients received. In their immunologic analysis, they found that 20/22 patients in the study achieved B-cell depletion. The levels of regulatory T cells were significantly increased after RTX therapy, with a further increase in regulatory T-cell function, though there was not a significant correlation between levels of regulatory T cells and the level of renal clinical response. This finding implied that RTX impacted T-cell regulation and apoptosis, which is a mechanism through which RTX may be able to provide an additional therapeutic effect.

Immunologic responses to RTX monotherapy and their relation to renal response rate

The mechanism of action of RTX appears to correlate with clinical response. Sfikakis et al. ⁴⁵ analyzed both the immunologic and clinical responses to RTX therapy in their prospective study that included 10 patients of Greek ethnicity (six of whom had failed prior immunosuppressive therapy) who received RTX as monotherapy for induction. They found that at 12 months, 40% had achieved a CRR and 40% had achieved a PRR. It was interesting that immunologically, there was a direct correlation between the clinical renal response to RTX and the onset and duration of B-cell depletion. The participants had frequent evaluations of their laboratory and flow cytometry results at several times during their follow-up. At the time of their partial remission, which occurred at a mean of two months, all but one patient had a complete depletion of their B cells. Those who achieved a CRR at three months either had continued complete depletion of their B cells or were in the early phase of recovery of their B-cell levels. Patients who did not attain a clinical response or who began to relapse did not have significant decreases in their B-cell activity.

A follow-up analysis ⁴⁶ noted a significant increase in regulatory T-cell activity in the early stages of B-cell depletion. Furthermore, patients who experienced remission of their clinical LN had higher levels of FOXP3 protein activity. Conversely, patients who had active disease had decreased FOXP3 mRNA activity during their follow-up. These findings indicated that FOXP3 activity modulated the regulatory T-cell response in patients with B-cell depletion. In summary, the degree of B-cell depletion achieved with RTX treatment was related to the clinical renal response to therapy.

RTX in combination with belimumab in LN

A promising new prospect may be a combination of RTX with belimumab—an antibody that inhibits B-cell activating factor (BAFF). This agent is approved for treatment of SLE. A 2011 RCT showed efficacy in reducing general SLE disease activity compared to placebo, with comparable adverse effects. ⁴⁷ The group who received belimumab had statistically significant reductions in SELENA-SLEDAI score ⁴⁸ compared to the placebo group. A 2017 systematic review ⁴⁹ found that belimumab may have favorable effects on LN. Of 326 patients from 11 studies who had LN and who received belimumab, 55.1% had a renal response to treatment. The renal responses varied in the studies that were analyzed. Examples of renal response assessment included proteinuria, serum creatinine, and renal disease improvement as per the SELENA-SLEDAI, BILAG, and Systemic Lupus Activity Measure indexes. ⁵⁰

The 2019 BLISS-BELIEVE RCT, ⁵¹ which investigates the impact of belimumab with RTX on SLE disease activity in general, is still ongoing. In a Phase IIA, open-label, single-arm proof-of-concept study ⁵² of 13 patients with LN (of the 16 total SLE patients in the study), treatment with RTX followed by belimumab resulted in a CRR in five patients and a PRR in six patients, with statistically significant decreases in proteinuria. The combination of RTX followed by belimumab may be more effective than RTX alone, since RTX depletes B cells and belimumab inhibits BAFF, thus preventing the propagation of B cells, even after depletion.

The CALIBRATE study ⁵³ is examining whether adding belimumab to induction therapy with CYC and RTX may led to significant differences in renal response. The trial is in Phase II. At week 24, there were no significant changes in clinical renal response between the group that received belimumab as maintenance therapy and the control group. However, the trial is still ongoing. Should this trial demonstrate a clinical improvement with belimumab maintenance therapy, it would be of interest to study how RTX alone, with belimumab maintenance therapy, would impact clinical renal response. While an intriguing and novel approach for treating LN, the data regarding RTX treatment followed by belimumab in LN are still limited.

Summary of renal responses to RTX in various studies

Of the three RCTs that compared RTX monotherapy or combination therapy to placebo or standard therapy, only one trial showed a clinically significant improvement in renal response with RTX therapy added to CYC. The two prospective studies that compared RTX monotherapy to induction regimens with CYC and MMF found that with RTX therapy, the CRR improvement averaged 72.3% and the PRR averaged 20.4%. There were no clinically significant differences in clinical disease response between the different types of induction treatments.

The pooled results of six prospective and retrospective studies that investigated RTX as monotherapy for induction found CRRs or PRRs in 65.8% of patients. The changes in serum creatinine were not significant, while the changes in proteinuria were clinically significant. In the two studies that investigated how RTX induction therapy with MMF maintenance impacted the use of oral steroid therapy, there were significant reductions in the doses of maintenance steroids used. CRRs occurred in 47.1% of patients, and PRRs occurred in 33.8% of patients. Overall, there were no significant changes in serum creatinine in both studies.

In the 11 prospective and retrospective studies that investigated the effect of using RTX as therapy in combination with other major immunosuppressive agents, there were CRRs or PRRs in 65.5% of patients. There were also serologic, immunologic, and histologic benefits of RTX added to other immunosuppressive regimens used concurrently, with 77.8% of patients achieving a CRR or a PRR. Two of those studies showed significant improvements in histology results, and there was a relation between clinical renal response and B-cell depletion.

Overall, although the existing studies regarding RTX use in LN did not indicate a clear benefit compared to standard therapy, there were clinically significant renal improvements as determined by preset response criteria. The studies varied widely in terms of the ethnicities and nationalities represented. The nationalities and ethnic groups most represented among the different studies were East Asian (Chinese and Japanese), Hispanic, and Caucasian. Although all of the studies showed significant renal responses as determined by serologic values and urine sediment, the studies that involved predominantly Caucasian participants demonstrated a higher absolute percentage of patients who achieved a CRR. For the outcome of a CRR (see Table 3), in the studies that had predominantly Caucasian participants (14 studies), 77% of patients had a CRR. For the three studies with predominantly Hispanic patients, 27.9% had a CRR. For the three studies with predominantly East Asian patients, 37.5% had a CRR. In the LUNAR study, there was not an overall significant improvement in renal response rate with RTX compared to placebo, except in patients of African American descent. Further studies, particularly RCTs, are needed particularly among minority groups in order to determine the efficacy of RTX as induction therapy.