#NephJC Chat
Tuesday Jan 12 9 pm Eastern
Wednesday Jan 13 9 pm IST
Wednesday Jan 13 9 pm GMT
Kidney Int 2020 Nov 7;S0085-2538(20)31251-5. doi: 10.1016/j.kint.2020.10.014. Online ahead of print.
The STARMEN trial indicates that alternating treatment with corticosteroids and cyclophosphamide is superior to sequential treatment with tacrolimus and rituximab in primary membranous nephropathy
Gema Fernández-Juárez , Jorge Rojas-Rivera , Anne-Els van de Logt, Joana Justino, Angel Sevillano, Fernando Caravaca-Fontán, Ana Ávila, Cristina Rabasco, Virginia Cabello, Alfonso Varela, Montserrat Díez, Guillermo Martín-Reyes, Marian Goicoechea Diezhandino, Luis F Quintana, Irene Agraz, Juan Ramón Gómez-Martino, Mercedes Cao, Antolina Rodríguez-Moreno, Begoña Rivas, Cristina Galeano, Jose Bonet, Ana Romera, Amir Shabaka, Emmanuelle Plaisier, Mario Espinosa, Jesus Egido, Alfonso Segarra, Gérard Lambeau, Pierre Ronco, Jack Wetzels, Manuel Praga, STARMEN investigators
PMID: 33166580
Introduction
Membranous nephropathy (MN) is a common glomerular disease characterised histologically by subepithelial deposits of IgG and complement components and manifesting clinically as proteinuria, often of nephrotic range. In simple terms, MN is caused by autoantibodies directed against antigens (often phospholipase A2 Receptor - PLA2R) located on the podocytes (see detailed discussion of the topic from the NephJC MENTOR summary by Jennie Lin here). It is the most common cause of primary nephrotic syndrome in nondiabetic adults worldwide. Apart from PLA2R, there has been a veritable explosion of different antigens discovered on a regular basis, now dubbed as the ‘Moore’s Law’ of membranous nephropathy (Hayashi et al, Kidney Int 2020), which suggests that as scientific research advances, more and more antigens will be (and have been) implicated in the pathogenesis of MN. Similar to the diagnostic discoveries, another advance are the therapeutic trials. Priti Meena and Het Patel did a fantastic write up on this in association with the Landmark Nephrology and it can be accessed here. See table below for a summary.
Summary of trials, from the Landmark Nephrology RFN blog
The single most important trial that revolutionized treatment, has been the 1984 trial (Ponticelli et al, NEJM 1984)which established the role of alkylating agents (chlorambucil in this case) in the management of MN. The study protocol ( later on immortalized as the ‘original Ponticelli protocol’) consisted of a 6 month regimen in which patients received methylprednisolone (3 doses of 1 gram/day followed by oral prednisolone 0.4 mg/kg/day for 27 days) and chlorambucil (0.2 mg/kg/day) alternately for a month each for 6 months. This was an open label RCT in which the control group received only symptomatic treatment for 6 months. The researchers reported that 23 of 32 patients in the treatment arm achieved either complete or partial remission while only 9 of 32 patients did so in the control arm.
Prior to this landmark study, steroids alone were used in MN, which did show a good response, compared to placebo (Collaborative study, NEJM 1979). In 1992, Ponticelli again carried out an open label RCT, this time to compare the efficacy of his protocol compared to that of steroids alone (Ponticelli et al, NEJM 1992). At the end of one year of follow up, 58% patients in the steroid + chlorambucil arm attained either complete (CR) or partial remission (PR) as compared to 26% in the patients who received steroids alone (p = 0.002). This difference in remission rates was maintained on 2 year and 3 year follow up as well but became insignificant at 4 years. It was concluded that Ponticelli’s protocol brought about earlier remission but with diminishing response.
Cyclophosphamide (CYC) was another drug which was being increasingly used by nephrologists in glomerular diseases and in 1998, Dr Ponticelli designed an open label RCT in which his protocol was compared with steroids and CYC (2.5 mg/kg/day) given alternate monthly for 6 months (Ponticelli et al, JASN 1998). It was found that both protocols were equally effective in bringing about remission. This led to the emergence of the ‘modified’ Ponticelli protocol in which steroids were used with either chlorambucil or CYC. This particular protocol has been widely used and nephrologists worldwide vouch for its efficacy. It was also incorporated into the 2012 KDIGO guidelines as a first line treatment protocol for primary MN.
However, alkylating agents come with problems of their own and cause several adverse effects, the most important of which is an increased risk of malignancies. CYC particularly, owing to its acrolein metabolite, has been implicated in gonadal as well as bladder malignancies. This has led to the hunt for an alternative agent which can help avoid the risk associated with alkylating agents.
So far we have seen that:
Steroids were better than placebo
Ponticelli protocol (chlorambucil based) was better than conservative management
Ponticelli protocol (chlorambucil based) was better than steroids alone
Ponticelli protocol (chlorambucil based) seemed similar to modified Ponticelli protocol (cyclophosphamide based)
The ‘better effect’ we mention above is mostly based on remission of proteinuria - and when it comes to kidney function, this issue gets more tricky with calcineurin inhibitors (CNI). Along with CNI came MMF and rituximab (Rtx). In 2007, Branten et al obtained significant reduction in proteinuria with MMF as compared to cyclophosphamide (but historically matched controls). But these were not substantiated in subsequent RCTs (eg Dussol et al, AJKD 2008 and Hogg et al, AJKD 2015) so MMF monotherapy has fallen by the wayside. The same fate has not befallen the CNIs. In 2001, cyclosporine + steroid was compared to placebo + steroid and a significant reduction in proteinuria with cyclosporine was reported (Cattran et al, Kidney Int 2001). CNIs do have a problem though: they reduce proteinuria very well, but are nephrotoxic, and withdrawal often leads to relapse (unlike cyclophosphamide). Hence there has been much hope on Rtx, which brought the promise of efficacy and safety. But failed to live up to its promise and faltered at the very first trial. In 2017, the GEMRITUX trial (Dahan et al, JASN 2017) reported adding on and comparing Rtx (375 mg/m2 on days 1 and 8) on a background of ‘Non Immunosuppressive Antiproteinuric Treatment’ (NIAT). Alll patients received NIAT for 6 months and were then randomized to either Rtx (375 mg/m2 on days 1 and 8) and NIAT versus NIAT alone. Rituximab and NIAT at this dose and frequency failed to show any superiority at 6 months against NIAT alone. However, Rtx takes some time to reduce PLA2R - and presumably the subsequent reduction in proteinuria would also take some time - thus the argument that 6 months are not enough, and the GEMRITUX observational follow up did show a benefit at 17 months (Dahan et al, JASN 2017).
What really cemented the place for Rtx was the MENTOR trial (Fervenza et al, NEJM 2019, NephJC summary, podcast). An open label RCT (MENTOR), it compared Rtx (1000 mg on days 1 and 15) to cyclosporine (starting at 3.5 mg/kg/day). This was a noninferiority trial in which Rtx was noninferior to cyclosporine at 12 months and superior at 24 months. Note that cyclosporine was abruptly stopped at 12 months, and we know this inevitably leads to relapse. The effects of Rtx are indeed much more long lasting, hence its superiority, becoming clear at 24 months, is somewhat unsurprising. While the revised KDIGO guidelines for MN are yet to be published, the draft released for public review showed that Rtx now occupies a prominent position in treatment protocols for MN.
So to recap:
Rtx is superior to cyclosporine
Cyclosporine is superior to steroids alone
However, this therapeutic path seems to be in parallel with but ignoring the cyclophosphamide/Ponticelli regimen comparison. How would Rtx compare with cyclophosphamide?
STARMEN took one approach, of combining CNI (in this case tacrolimus) with one dose of Rtx. An observational study had reported that the high relapse rate with CNIs could be reduced with Rtx (375 mg/m2 per week for 4 weeks, with pulse steroids) at the time of tapering (Segarra et al, CJASN 2009). Another pilot trial (Waldman et al, KI Reports 2016) reported encouraging results of a combined therapy with cyclosporine plus rituximab in high-risk PMN patients. Rtx was dosed at 1 g, 2 weeks apart, with the dose repeated based on CD19 B cell counts. Which leads us to the STARMEN design.
The Study
Sequential Treatment with Tacrolimus and Rituximab versus Alternating Corticosteroids and Cyclophosphamide in PMN study (STARMEN)
Study Design
Multicenter (19 centers in Spain, 1 in Netherlands), prospective, open-label, randomized, controlled trial with a two-group parallel design
Inclusion Criteria
Adults with biopsy-proven primary MN
Estimated glomerular filtration rate (eGFR) ≥ 45mL/min/1.73m2
Nephrotic-range proteinuria (>4 g/24 h, without a decrease >50% during the observational period), and
Hypoalbuminemia (≤3.5 g/dL during the observational period
Exclusion Criteria
Secondary causes of MN
HIV infection
Liver disease
Treatment with another investigational drug
Suspected or known hypersensitivity and/or allergy reaction to drugs from study
Previous treatment with
corticosteroids (3 months before screening)
other immunosuppressive agent (6 months before screening)
rituximab or any other biological agent (2 years before screening)
Non-response to previous immunosuppressants
Any other severe condition or abnormal laboratory test with a potential risk for the patient outcome
Current drug or alcohol dependence
Interventions
All patients received a background of RAS blockade (ACEi/ARB for at least 2 months) and controlled BP (defined as BP < 150/90 for 3 months). Randomization: 1:1 randomization, via computer algorithm, to corticosteroid-cyclophosphamide (Group 1) or tacrolimus-rituximab (Group 2)
Figure 4 from STARMEN, KI 2020
Group 1 (modified Ponticelli protocol)
Patients received methylprednisolone (MPS) at months 1, 3 and 5 (1 g IV at days 1, 2, 3, then 0.5 mg/kg/day orally from day 4 to 30). At months 2, 4 and 6, patients received oral CYC adjusted for age and renal function (1–2 mg/kg/day for 30 days).
Group 2 (Rtx and Tacrolimus)
Patients received oral tacrolimus (0.05 mg/kg/day), to reach target blood levels of 5–7 ng/mL, for six months. At day 180, patients received one dose of Rtx (1 g) and tacrolimus dosage was then reduced by 25% per month, with complete withdrawal at the end of month 9. Tacrolimus dose was reduced in case of kidney function impairment. To minimize infusion reactions with Rtx, patients received methylprednisolone 100 mg as well with that dose.
Prophylaxis During Treatment
Oral trimethoprim/sulfamethoxazole 160/800 mg, 3 times/week
Outcomes and Analysis
Primary Outcome :
Complete or partial remission (composite endpoint) at 24 months
Complete remission (CR) was defined as reduction of proteinuria from baseline to a value ≤ 0.3 g/24 h + stable kidney function (eGFR ≥45mL/min/1.73m2)
Partial Remission (PR) was defined as reduction of proteinuria >50% from baseline and a value less than 3.5 g/24 h + stable kidney function (eGFR ≥45mL/min/1.73m2)
No remission was defined as reduction of proteinuria <50% from baseline values.
Relapse was defined as reappearance of proteinuria >3.5 g/24 h and at least ≥50% increase from the lowest value in three or more consecutive visits, in patients with previous CR/PR.
Secondary Endpoints:
Rate of complete and partial remission at 3, 6,12, 18 and 24 months
Relapse of nephrotic syndrome at 6, 12, 18 and 24 months
Immunological response at 3 to 24 months
Percentage of patients free of ≥50% increases of serum creatinine and with preserved kidney function (eGFR ≥45 mL/min/1.73 m2) at 24 months
Adverse events
Safety Endpoints
Proportion of patients with drug-related adverse events
Total cumulative dose received of each study treatment
PLA2R were measured as well. Baseline serum levels >14 RU/mL as measured with the standardized commercial ELISA (Euroimmun, Lubeck, Germany) were considered positive. Immunological response was defined by a level of anti-PLA2R ≤ 14 RU/mL
Statistical analysis : The investigators aimed to include 94 patients, assuming a statistical power of 80%, assuming a 60% response rate with the Ponticelli protocol and 85% with tacrolimus + Rtx intervention. Intention-to-treat and per-protocol analysis estimating the relative risk (RR) with 95% confidence interval (CI) and compared using Pearson chi-squared or Fisher exact test.
Funding source: The study was funded by ERA-EDTA, other granting agencies and local institutions (no industry).
Results
Of the 130 patients assessed for eligibility, 86 were randomised (figure 1). Their baseline characteristics are shown in Table 1.
Figure 1 from STARMEN, KI 2020
Table 1 from STARMEN, KI 2020
As can be seen above, they were fairly nephrotic with median 7 g proteinuria and hypoalbuminiemia. About 3 quarters were PLA2R positive.
Group 1
Total cumulative dose of oral MPS (methylprednisone) = 3.4 ± 0.9 g
Total cumulative dose of IV MPS = 8.2 ± 1.4 g
Total cumulative dose of CYC = 10 ± 3.5 g
Group 2
Doses and blood levels of tacrolimus were as shown in Table S3
Table S3 from STARMEN, KI 2020
Three patients received a second dose of rituximab at months 12, 12 and 18, respectively. Two patients received additional doses of tacrolimus beyond month 9
5 patients in each group were switched to a non-study intervention (crossed over) due to a lack of efficacy of the assigned treatment, as can be seen in figure 1. These were considered as 'no response.'
Primary Outcome
36 patients (83.7%) in Group 1 and 25 patients (58.1%) in Group 2 showed CR/PR at 24 months (RR 1.44; 95% CI 1.08 to 1.92); (Table 2 and Figure 2A).
Table 2 from STARMEN, KI 2020
Per-protocol analysis also confirmed significant differences in the primary outcome between the two groups. 26 patients (60%) in Group 1 achieved CR at 24 months as compared to 11 patients (26%) in Group 2 (RR 2.36; 95% CI 1.34 to 4.16).
1/36 patients (2.7%) in Group 1 and 3/25 (12%) patients in Group 2 who had achieved partial remission relapsed with nephrotic range proteinuria or >50% increase in urinary protein from baseline (Supplementary Table S8).
Figure 2A from STARMEN, KI 2020 (CR or PR)
Figure 2B from STARMEN, KI 2020 (CR only)
Subgroup analysis
Corticosteroid-cyclophosphamide treatment showed greater efficacy across different baseline values of proteinuria, serum albumin, serum creatinine, anti-PLA2R levels and age. The interaction for gender seems significant below, but the numbers were small.
Figure S1 from STARMEN, KI 2020
Secondary Outcomes
Proteinuria decreased from 7.4 g/24 hours (4.8-11.3) at baseline to 0.35 g/24 hours (0.2-9) at 24 months in the Group 1 and from 7.4 g/24 hours (6.7-11.6) at baseline to 1g/24 hours (0.3-3.3) at 24 months in Group 2 (between-group difference p=0.005) (Figure 3A, and Supplementary Table S5).
Figure 3A from STARMEN, KI 2020
Serum albumin increased from 2.6±0.1 g/dL at baseline to 4±0.1 g/dL at 24 months in Group 1 and from 2.6±0.1 g/dL at baseline to 3.9±0.1 g/dL at 24 months in Group 2 (between-group difference p=0.2) (Figure 3B, and Supplementary Table S5).
Figure 3B from STARMEN, KI 2020
With respect to kidney function, there was a non-significant trend for higher values of eGFR in Group 1 than in Group 2 throughout the follow-up (Figure 3C, and Supplementary Table S6). The number of patients with preserved renal function at 24 months was 40 (93%) in Group 1 and 37 (86%) in Group 2 (p=0.48). Only 1 patient developed CKD. Note that 2 patients discontinued CNI arm for development of AKI (from figure 1).
Figure 3C from STARMEN, KI 2020
The anti PLA2R levels decreased significantly in both groups (Table 3, Figure 3A). At 3 and 6 months, more patients in group 1 (77% and 92%, respectively), as compared to Group 2 (45% and 70%, respectively) achieved immunological response (PLA2R <14 RU/mL). Immunological response at 3 months (p=0.036) and 6 months (p=0.005) was associated with remission at 24 months.
Table 3 from STARMEN, KI 2020
Adverse Events
There were no statistically significant differences in the frequency of serious adverse events between groups. The only case of serious acute kidney injury occurred in the tacrolimus-rituximab group, while 4 of 5 serious infections occurred in the corticosteroids-cyclophosphamide group. Three cases of cancer were reported, although none of them was deemed related to the treatment received.
Table 4 from STARMEN, KI 2020
Discussion
This study was carried out on the premise that as tacrolimus is associated with increased rate of relapse on tapering and as rituximab is effective in preventing these relapses, a combination should probably have the advantage of sustained remission. Moreover, it was also thought that there would be fewer adverse effects as compared to the protocol with cyclophosphamide. However, the study surprisingly (or not!) showed that this protocol was actually inferior to the time-tested protocol developed by Ponticelli.
Steroid + CYC brought about faster, longer, and more complete remissions.
Although there were more adverse events with the later, there was no statistically significant difference between the protocols in terms of serious adverse events. This could be partly due to the lower cumulative dose of CYC used here as compared to previous trials. We should not give up on tacrolimus plus rituximab though, as STARMEN did not show a significant number of relapses after discontinuing tacrolimus. So, Rtx did achieve some of its role. Also, more number of patients did achieve complete remission with the addition of Rtx, which is in accordance with previous data.
It is to be noted here that the tacrolimus-rituximab group had more men and there were trends toward higher PLA2R-ab titers, a higher interquartile range of proteinuria and lower interquartile range of serum albumin in this group. Although not significant, these differences at baseline may have biased the result towards an inferior outcome in the tacrolimus-rituximab group.
Both protocols induced a significant reduction in anti-PLA2R levels, although immunological response occurred faster in Group 1. CYC induces a generalized leukocyte and mature plasma cell ablation, resulting in a more drastic reduction in antibody production than more specifically targeted drugs like rituximab or calcineurin inhibitors. Early immunological response was followed by clinical remission in the vast majority of patients, which confirms the usefulness of anti-PLA2R monitoring for a personalized treatment of the disease. It could also be useful in deciding when to stop CYC in these patients.
Let’s acknowledge though, that STARMEN was a head-to-head prospective controlled trial that compared the time honoured Ponticelli protocol with CNI and Rtx, something that has not been done before as we discussed above.
Limitations
This was an open label RCT with a lack of blinding.
The limited sample size prevented analysis by pre-specified subgroups or detailed analysis of anti-PLA2R kinetics
Anti-PLA2R were measured only in 69 of 86 patients
Since CD19+ B cells were not measured, no information about the adequacy of rituximab dose was available - and a single dose of Rtx at 6 months may have been insufficient.
An additional trial was the RI-CYCLO trial which was also y presented as a late breaking clinical trials poster at the same time as STARMEN, but has not been published yet. In this pilot RCT, 74 patients were randomized to receive Rtx (1g two weeks apart) or CYC as in Ponticelli's protocol. You can check out a thread here for more. Overall there was no difference, so the Rtx story in MN is still not yet done. Coming soon to a NephJC near you.
Conclusion
'Treatment with corticosteroid-cyclophosphamide induced remission (mostly complete) of nephrotic syndrome in a significantly greater number of patients than treatment with tacrolimus-rituximab, although side effects were more frequent in the former group.' Thus, old is definitely gold as far as membranous nephropathy is concerned. Almost 4 decades have passed since Ponticelli and colleagues described their original protocol and although a number of drugs have been introduced and tried since then, they are yet to show an advantage over this age-old formula. Perhaps future trials will enable us to achieve a fine balance between efficacy and safety as far as membranous nephropathy is concerned.
Summary by Namrata Parikh
Nephrologist, Divine Life Hospital, Gujarat, India
NSMC graduate, class of 2020