Comparative effectiveness and tolerance of immunosuppressive treatments for idiopathic membranous nephropathy: A network meta-analysis

PLOS ONE, Sep 2017

Background Immunosuppressive agents in general are shown to prevent renal progression and all-cause mortality in idiopathic membranous nephropathy (IMN) patients with nephrotic syndrome. However, the efficacy and safety of different immunosuppressive treatments have not been systematic assessed and compared. A network meta-analysis was performed to compare different immunosuppressive treatment in IMN. Methods Cochrane library, MEDLINE, EMBASE and trial register system were searched for randomized controlled trials reporting the treatments for IMN to May 3, 2016. Composite endpoint of mortality or end-stage kidney disease (ESKD), complete or partial proteinuria remission and withdrawal because of treatment adverse events were compared combing direct and indirect comparison using network meta-analysis. Ranking different immunosuppressive treatments in the outcomes were analyzed by using surface under the cumulative ranking curve (SUCRA). Results Total 36 randomized controlled trials (n = 2018) covering 11 kinds of treatments were included. Compared with non-immunosuppressive treatment, only cyclophosphamide (CTX) and chlorambucil significantly reduced the risk of composite outcome of mortality or ESKD while combining the direct and indirect comparison (OR = 0.31, 95%CI: 0.12–0.81 and OR = 0.33, 95%CI: 0.12–0.92). CTX increased the composite outcome of complete remission (CR) or partial remission (PR) (OR = 4.29, 95%CI: 2.30–8.00) but chlorambucil did not (OR = 1.58, 95%CI: 0.80–3.12) as compared with non-immunosuppressive treatment. Chlorambucil also significantly increased the withdrawal risk (OR = 3.34, 95%CI: 1.37–8.17) as compared to CTX. Both tacrolimus (OR = 3.10, 95%CI: 1.36–7.09) and cyclosporine (CsA) (OR = 2.81, 95%CI: 1.08–7.32) also significantly increased the rate of CR or PR as compared with non-immunosuppressive treatment (without significant difference as compared with CTX), while ranking results showed that cyclosporine or tacrolimus was with less possibility of drug withdrawal as compared to CTX. Conclusions Cyclophosphamide and chlorambucil reduce risk of ESKD or death in IMN with nephrotic range proteinuria, but carry substantial toxicity that may be lower for cyclophosphamide. Tacrolimus and cyclosporine increase the possibility of proteinuria remission with less drug withdrawal, but the effects on kidney failure remain uncertain.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

http://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0184398&type=printable

Comparative effectiveness and tolerance of immunosuppressive treatments for idiopathic membranous nephropathy: A network meta-analysis

et al. (2017) Comparative effectiveness and tolerance of immunosuppressive treatments for idiopathic membranous nephropathy: A network meta-analysis. PLoS ONE 12(9): e0184398. https://doi.org/10.1371/journal. pone.0184398 Comparative effectiveness and tolerance of immunosuppressive treatments for idiopathic membranous nephropathy: A network meta- analysis Song Ren 0 1 Ying Wang 1 Li Xian 0 1 Tadashi Toyama 1 Meg Jardine 1 Guisen Li 0 1 Vlado Perkovic 1 Daqing Hong 0 1 0 Renal Division and Institute of Nephrology, Sichuan Academy of Medical Science & Sichuan Provincial People's Hospital, School of Medicine, University of Electronic Science and Technology of China , Chengdu , China , 2 The George Institute for Global Health, University of Sydney , Sydney , Australia , 3 Division of Nephrology, Kanazawa University Hospital, Kanazawa city, Japan, 4 Concord Repatriation General Hospital , Concord , Australia 1 Editor: Giuseppe Remuzzi, Istituto Di Ricerche Farmacologiche Mario Negri , ITALY Immunosuppressive agents in general are shown to prevent renal progression and allcause mortality in idiopathic membranous nephropathy (IMN) patients with nephrotic syndrome. However, the efficacy and safety of different immunosuppressive treatments have not been systematic assessed and compared. A network meta-analysis was performed to compare different immunosuppressive treatment in IMN. - Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Background Methods Results Cochrane library, MEDLINE, EMBASE and trial register system were searched for randomized controlled trials reporting the treatments for IMN to May 3, 2016. Composite endpoint of mortality or end-stage kidney disease (ESKD), complete or partial proteinuria remission and withdrawal because of treatment adverse events were compared combing direct and indirect comparison using network meta-analysis. Ranking different immunosuppressive treatments in the outcomes were analyzed by using surface under the cumulative ranking curve (SUCRA). Total 36 randomized controlled trials (n = 2018) covering 11 kinds of treatments were included. Compared with non-immunosuppressive treatment, only cyclophosphamide (CTX) and chlorambucil significantly reduced the risk of composite outcome of mortality or ESKD while combining the direct and indirect comparison (OR = 0.31, 95%CI: 0.12±0.81 and OR = 0.33, 95%CI: 0.12±0.92). CTX increased the composite outcome of complete remission (CR) or partial remission (PR) (OR = 4.29, 95%CI: 2.30±8.00) but chlorambucil supported by John Chalmers Servier Post-doctoral fellowship. Competing interests: The authors have declared that no competing interests exist. did not (OR = 1.58, 95%CI: 0.80±3.12) as compared with non-immunosuppressive treatment. Chlorambucil also significantly increased the withdrawal risk (OR = 3.34, 95%CI: 1.37±8.17) as compared to CTX. Both tacrolimus (OR = 3.10, 95%CI: 1.36±7.09) and cyclosporine (CsA) (OR = 2.81, 95%CI: 1.08±7.32) also significantly increased the rate of CR or PR as compared with non-immunosuppressive treatment (without significant difference as compared with CTX), while ranking results showed that cyclosporine or tacrolimus was with less possibility of drug withdrawal as compared to CTX. Conclusions Cyclophosphamide and chlorambucil reduce risk of ESKD or death in IMN with nephrotic range proteinuria, but carry substantial toxicity that may be lower for cyclophosphamide. Tacrolimus and cyclosporine increase the possibility of proteinuria remission with less drug withdrawal, but the effects on kidney failure remain uncertain. Introduction Idiopathic membranous nephropathy (IMN) is one of the commonest causes of primary nephrotic syndrome in adults [ 1 ]. Approximately one-third of affected individuals will have a complete and spontaneous remission of proteinuria, while another third develop persistent proteinuria with long-term preservation of renal function. However, the remaining third of patients will progress to end-stage kidney disease (ESKD), and the risk is higher in people with nephrotic range proteinuria [ 2,3 ]. Overall, immunosuppressive therapy has been suggested to reduced proteinuria, all-cause mortality and progression to ESKD [ 4 ]. The Kidney Disease Improving Global Outcomes (KDIGO) guidelines recommend steroids plus alkylating agents in patients at high risk of developing ESKD based on severe proteinuria. People with contraindications for alkylating agents or who do not tolerant them, are recommended to receive steroids plus calcineurin inhibitors, (such as tacrolimus and cyclosporine) as second line therapy[ 5 ]. However, the quality of the evidence has been noted to be moderate leaving residual uncertainty about the optimal treatment approach, and a number of additional randomized-controlled trials (RCTs) evaluating novel immunosuppressive therapies for IMN have been published subsequent to these guidelines. In this study, we used network meta-analysis to directly and indirectly compare the effects of different immunosuppressive regimens on renal survival, proteinuria remission and tolerability in people with IMN and nephrotic range proteinuria. Materials and methods Data sources and searches The Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE and EMBASE databases were systematically searched for randomized controlled trials in people with IMN up to May 3, 2016. The search terms used consisted of IMN and randomized controlled trials (S1 Table). We searched two additional websites [www.controlled-trials.com, and www. clinicaltrials.gov] for ongoing clinical trials, and manually checked the reference lists of included studies to identify additional eligible studies. 2 / 15 Study selection Studies of immunosuppressive treatments in IMN were included. We included studies of peo ple with biopsy proven IMN and nephrotic range proteinuria (urinary protein excretion >3.5 g/24h). Membranous nephropathy secondary to autoimmune diseases, cancer, infections (including B and C hepatitis) and drugs, or atypical membranous nephropathy were excluded from our review. Type of interventions: We included any treatment expected to primarily act via an immu nosuppressive effect. Agents without clear mechanism of immunosuppressive effects (such as traditional Chinese medicine) were excluded. Comparators: Different immunosuppressive treatments were compared with each other or with non-immunosuppressive treatment (placebo, renin angiotensin system blockers or other supportive therapies). We excluded the trials with more than two immunosuppressive agents in one arm (except steroids) due to difficulty in differentiating the treatment effects as a result of the combined effect or the sum of the monotherapy effect. Studies that assessed the same immunosuppressive agents with different doses in different arms but no other comparator were excluded. Trials with a follow-up of less than 6 months were excluded, as were cross-over trials. The abstracts were initially screened by two researchers (SR and LX). Disagreements were resolved through discussions (SR, LX, DH and GL). Data extraction and quality assessment Data extraction was undertaken independently by 2 investigators (SR and LX) using a standardized electronic form. Disagreements in abstracted data were resolved by a third investigator (DH). The abstracted data included baseline characteristics of the participants such as age, sex, serum creatinine, serum albumin and proteinuria value. The main outcomes included 1) the composite of all-cause mortality or ESKD (defined as start of dialysis or renal transplantation); 2) complete or partial remission (defined by the authors in each study); 3) withdrawal due to the drug related toxicity. Other outcomes included adverse reactions such as infection, bone marrow suppression, abnormal liver function, incidence of diabetes mellitus or hypertension, or as defined by the study authors. Study quality was judged by the Jadad score[ 6 ] and Cochrane Collaboration tool[7] for assessing the risk of bias. We considered the study of low quality if the Jadad score was less than 3, or higher-quality if the Jadad score was more than or equal to 3. The risk of bias assessment included random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting and other bias[ 8 ]. We summarized both individual and aggregate risk of bias data for the included studies. Data synthesis and statistical analysis We use STATA (version 14.0) to perform network meta-analysis with a random-effects mixed-treatment comparisons model for multi-armed trials within the frequency probability method on the effects of main outcomes. Odds ratios (OR) and their 95% confidence intervals (CI) were used to compare treatment effects for each dichotomous outcome. To summarize the effectiveness and tolerance of all treatments, we also calculated the surface under the cumulative ranking curve (SUCRA). The heterogeneity of the included studies was large, and it is unavoidable. So we perform a consistency test to differences between direct and various indirect comparisons by the nodesplitting approach for the main outcomes. The p value was calculated by the node-splitting method, which separated evidence on a particular comparison into direct and indirect 3 / 15 evidence[ 9 ]. We used consistency model to do network meta-analysis when there was no statistically significant difference between direct and indirect comparison. For the main outcomes, we also performed sensitivity analysis according to study quality (excluding studies with Jadad scores less than 3). As the participants' follow-up time may influence effects on mortality or ESKD, we also performed a sensitivity analysis to exclude studies with follow-up time of less than two years. Results Description of the included studies The search retrieved 750 citations for screening, from which 36 studies [10±45] published from 1974 to 2016 involving 2018 participants and covering 11 kinds of treatments were eventually included (Fig 1). All studies were published in English except one in Japanese literature[ 42 ]. Four studies were three-armed trials [11,13,35,43] and the others had two arms. Five studies were only published in abstract form [24,34±36,45]. Eighteen studies used ACEI/ARB or placebo as the control group and immunosuppressive agents as the treatment group (S2 Table). Overall, 19 RCTs reported effects on all-cause mortality or ESKD, 31 RCTs reported effects on complete or partial remission, and 21 RCTs reported effects on treatment intolerance. The average follow-up time was 33months (9-120months). Trials generally had a small sample size with a median of 56 (9±158) patients (Fig 2). Seventeen of 36 studies had a Jadad score of less than 3. The risk of bias of each study was assessed using Review Manager 5.3 and summarized in the S1 and S2 Figs. Composite endpoint of mortality or ESKD Effects on all-cause mortality or ESKD were reported in 19 studies including 8 types of immunosuppressive therapies. Among these, 14 studies reported effects on all-cause mortality alone, Fig 1. Study selection flow chart. RCT, randomized controlled trial. 4 / 15 Fig 2. a. Network of eligible comparisons for total mortality or ESKD. The width of lines is proportional to the number of studies compared in every pair of treatments, and the size of nodes is proportional to the total sample size of each treatment. b. Network of eligible comparisons for the outcome of total remission. c. Network of eligible comparisons for the outcome of treatment withdrawal. including 7 types of immunosuppressive therapies, and 16 studies separately reported effects on ESKD, involving 6 types of immunosuppressive therapies. There was no endpoint events reported in trials involving mizoribine, ACTH or leflunomide. The results of the network analysis of eligible comparisons for total mortality or ESKD are shown in Fig 2A. There was no statistically significant difference between the direct and indirect comparisons using the node-splitting method (p > 0.05). We therefore used the consistency model to undertake network meta-analysis. The findings of the network meta-analysis for mortality or ESKD is shown in Fig 3 (direct and all comparison results). Ranking of treatments is presented in S3 Fig. Compared with nonimmunosuppressive treatment, only cyclophosphamide (CTX) and chlorambucil significantly reduced the risk of the composite outcome of mortality or ESKD when combining the direct and indirect comparison (OR = 0.31, 95%CI 0.12±0.81 and OR = 0.33, 95%CI 0.12±0.92, respectively). No clear effects were able to be identified with the other immunosuppressive agents assessed. Fifteen high-quality studies including 7 types of immunosuppressive therapies were included to perform further sensitivity analysis and found similar results (S6 Fig) Exclusion of studies with follow up period of <2 years also did not change the primary conclusions, although confidence intervals were widened. Complete or partial remission A total of 31 RCTs including 11 immunosuppressive therapies reported effects on complete or partial remission, including cyclophosphamide (15 studies), control (18 studies), chlorambucil (9 studies), steroids (9 studies), cyclosporine (6 studies), tacrolimus (5 studies), mycophenolate mofetil (4 studies), azathioprine (3 studies), mizoribine (2 studies), ACTH (1 study), and leflunomide (1 study). The network analysis of eligible comparisons for the total remission outcome of the network meta-analysis is shown in Fig 2. Direct and indirect estimates by the node-splitting method identified no statistically significant differences. Compared with nonimmunosuppressive therapies, four immunosuppressive agents showed significantly improve CR or PR. ACTH showed the best therapeutic effect among all the immunosuppressive thera pies (OR = 91.00, 95%CI 5.98±1384.15), followed by CTX (OR = 4.29, 95%CI 2.30±8.00), tacrolimus (OR = 3.10, 95%CI 1.36±7.09), CsA (OR = 2.81, 95%CI 1.08±7.32). Chlorambucil (OR = 1.58, 95%CI 0.8±3.12) and MMF (OR = 1.52, 95%CI 0.58±3.96) didn't significantly improve CR or PR compared with non-immunosuppressive therapies (Fig 4). The rank probabilities of all treatments calculated using the SUCRA is shown in S4 Fig. A sensitivity analysis that excluded lower quality studies showed that only two immunosuppressive agents (CTX and tacrolimus) significantly improve CR or PR (OR = 3.73, 95%CI 5 / 15 Fig 3. The results of network meta-analysis (total mortality or ESKD). All treatments were compared with each other. In the first group, 7 treatments were compared with the control group. In the second group, 6 treatments were compared with the azathioprine group. In the third group, 5 treatments were compared with the CTX group. In the fourth group, 4 treatments were compared with the chlorambucil group. In the fifth group, 3 treatments were compared with the CsA group. In the sixth group, 2 treatments were compared with the MMF group. In the last group, tacrolimus compared with the steroids group. 1.85±7.48 and OR = 3.69, 95%CI 1.49±9.14), respectively, compared with non-immunosuppressive therapies (S7 Fig). Withdrawal of treatments A total of 21 RCTs assessing 9 immunosuppressive agents reported effects on withdrawal from treatment. The network analysis of eligible comparisons for the total tolerance outcome of the network meta-analysis is shown in Fig 2. In the network meta-analysis comparison, the control group had the lowest withdrawal rate and best tolerance. Compared with the control group, chlorambucil was associated with the highest withdrawal rates (OR = 13.73, 95%CI 4.23±44.59), followed by MMF (OR = 6.17, 95%CI 1.10±34.58), mizoribine (OR = 4.46, 95%CI 1.21±91.66), CTX (OR = 4.11, 95%CI 1.52±11.10) and steroids (OR = 3.82, 95%CI 1.24±11.75). Compared with chlorambucil, the withdrawal rates of tacrolimus and CsA were lower (OR = 0.26, 95%CI: 0.06±1.18 and OR = 0.15, 95%CI 0.02±1.21). There were no statistically 6 / 15 Fig 4. The results of network meta-analysis (complete or partial remission). In the first group, 10 treatments were compared with the control group. In the second group, 9 treatments were compared with the ACTH group. In the third group, 8 treatments were compared with the azathioprine group. In the fourth group, 7 treatments were compared with the CTX group. In the fifth group, 6 treatments were compared with the chlorambucil group. In the sixth group, 5 treatments were compared with the CsA group. In the seventh group, 4 treatments were compared with the MMF group. In the eighth group, 3 treatments were compared with the LEF group. In the ninth group, 2 treatments were compared with the mizoribine group and in the last group, tacrolimus compared with the steroids group. significant differences between other immunosuppressive agents and the control group. Calcineurin inhibitors showed the best tolerance among all immunosuppressive agents. (Fig 5 and S5 Fig) 7 / 15 After excluding low-quality studies, the sensitivity analysis indicated that ranking of the drug tolerance remained unchanged, with alkylating agents having the highest withdrawal rates (S8 Fig). Other adverse reactions associated with use of immunosuppressive agents Overall, patients treated with immunosuppressive agents were reported to have an increased risk of infection (e.g. pneumonia, cystitis and skin infection), and alkylating agents have a higher risk of myelosuppression (e.g. leukopenia or thrombocytopenia). CTX, CsA and azathioprine was associated with hepatotoxicity, in particularly elevated serum transaminases. The following adverse events were analyzed: infection, bone marrow suppression, and abnormal liver function, incidence of diabetes mellitus (DM) or hypertension. The occurrence of adverse reactions according to treatments for IMN is listed in Table 1. Infection was the most common adverse events associated with various immunosuppressive therapies. The occurrence rate of infection was 7. 9% reported in 21 studies including 160 participants. Three immunosuppressive agents associated with the highest risk of infection were tacrolimus 21.5% (38/177), MMF 19.7% (14/71) and CTX 15.2% (68/448). There were no infections reported with three immunosuppressive agents: mizoribine, ACTH, leflunomide. Myelosuppression was seen with 4 immunosuppressive agents (CTX, chlorambucil, MMF and azathioprine). Among them, azathioprine (14.71%) was associated with highest incidence rates followed by MMF (8.45%), chlorambucil (7.79%) and CTX (2.23%). Deranged liver function tests were seen in patients treated with tacrolimus or CTX. Diabetes mellitus or glucose intolerance was reported in 30 patients (16.9%) treated with tacrolimus. 122 patients relapse after remission of proteinuria (6.0%), patients who are treated with CsA are more likely to develop relapse (21.2%), followed by chlorambucil (16.4%), azathioprine (14.7%). Overall, tacrolimus, azathioprine and MMF were more commonly associated with adverse reactions with the adverse events rates of 49%, 34%, and 32%, respectively. Discussion Membranous nephropathy is prevalent and could increase the risk of renal function lost and subsequent mortality. In 2009, Beck et al [ 46 ] first reported that M-type PLA2 receptor (PLA2R) was a major target antigen for IMN and about 70% of patients with IMN had autoantibodies to PLA2R in their serum. Thrombospondin type 1 domain±containing 7A (THSD7A) was described as a new autoantigen involved in adult idiopathic membranous nephropathy. Tomas et al [47] found 15 of 154 patients with idiopathic membranous nephropathy had circu lating autoantibodies to THSD7A but not to PLA2R receptor antibodies. Subsequently, more clinical studies have found THSD7A positivity by testing for circulating antibodies in IMN patients or THSD7A antigen deposition in renal tissue by immunohistochemistry[48±50]. We had evaluated the prevalence of THSD7A in the IMN patients (The article has not yet been published), and the result shows that the estimated prevalence of THSD7A in patients with IMN is 3% and higher prevalence in the PLA2R-negtive patients is 10%. Combined with the detection results of PLA2R and THSD7A can improve the diagnostic level of IMN. Immunosuppressive agents were used widely in patients with idiopathic membranous nephropathy. In this review, we focused on the outcomes most likely to be important to patients in making decisions, specifically ESKD or death, remission as well as tolerability and side effects. The results indicate that alkylating agents are the only immunosuppressive agents with proven benefits for the composite outcome of ESKD or mortality, and were the immunosuppressive agents most likely to induce remission of proteinuria. The likelihood of complete 8 / 15 Fig 5. The results of network meta-analysis (withdrawal of treatments). In the first group, 8 treatments were compared with the control group. In the second group, 7 treatments were compared with the azathioprine group. In the third group, 6 treatments were compared with the CTX group. In the fourth group, 5 treatments were compared with the chlorambucil group. In the fifth group, 4 treatments were compared with the CsA group. In the sixth group, 3 treatments were compared with the MMF group. In the seventh group, 2 treatments were compared with the mizoribine group and in the last group, tacrolimus compared with the steroids group. or partial remission was significantly increased with ACTH, mizoribine, CsA and tacrolimus, as compared with non-immunosuppressive treatment. Conversely, alkylating agents were also associated with a higher risk of withdrawal from the therapy, while calcineurin inhibitors (tacrolimus and CsA) appeared to be well tolerated despite high reported rates of adverse effects with tacrolimus in particular. Previous studies have demonstrated that glucocorticoids alone had little effective in increasing remission of proteinuria. Therefore, glucocorticoid plus immunosuppressive agents are often used for idiopathic membranous nephropathy. Our findings support the recommendations of the KDIGO guideline where alkylating agents are recommended as the preferred immunosuppressive agents for IMN[ 5 ]. In particular, they suggest that CTX should be used as the first line therapy for IMN. It not only reduced the long term events of ESKD or all-cause mortality, but also achieved high proteinuria remission rates that were superior to those achieved with chlorambucil. It was associated with high withdrawal rates, but these again 9 / 15 appeared lower than those seen with chlorambucil in direct and indirect comparisons. Chen et al [ 51 ] have performed a systematic review and meta-analysis which including RCT found that alkylating agents plus corticosteroids had long-term and short-term benefits for adult IMN, but resulted in more withdrawals or hospitalizations. This result is consistent with our analysis. In recent years, more and more patients have adopted intravenous CTX rather than oral to reduce the side effects of drug. Calcineurin inhibitors, including tacrolimus and CsA, are also widely used in IMN. Our results demonstrated no clear effect on mortality/ESKD events with these agents, although power was low due to a small number of events. As tacrolimus increases the risk of diabetes via reductions in insulin secretion and inhibition of insulin gene expression stimulus by hyperglycemia [ 52,53 ], long term monitoring blood glucose level is important for patients with IMN treated with this agent. A meta-analysis[54] including 21 RCTs was shown that calcineurin inhibitors had better short-term efficacy and safety than CTX, but the CsA has a higher relapse rate. Our analysis indicated that the patients who treated with CsA had hightest relapse rate. So, well-designed clinical trials are needed to further evaluate the long-term efficacy and safety of CNIs. The data regarding other agents was limited. We found that MMF was consistent with the recommendations from the KDIGO guidelines, which suggests that the monotherapy with MMF should not be used as a first line therapy for of IMN [5]. While the effects of ACTH appear promising, the data on this treatment was provided by only one study with 15 patients published in abstract form and of uncertain reliability. Mizoribine and leflunomide were also assessed in few patients. Further, appropriately powered, high-quality RCTs of these agents would appear worthwhile. We ranked different treatment strategies for each outcome by SUCRA probabilities and the posterior probabilities. The higher the SUCRA value, the higher the ranking. The results indicated that alkylating agent had best probability in reducing the risk of composite endpoint of mortality or ESKD. In terms of proteinuria remission, ACTH, CTX, tacrolimus and CsA had good probabilities of achieving proteinuria remission, however, only one small trial reported ACTH. Tacrolimus and CsA had best tolerance in all immunosuppressive agents. Overall, the total remission rate was 59.2% in the patients treated with immunosuppressive therapy, and 32.4% in patients treated with non-immunosuppressive agents. This difference highlights the impact of immunosuppressive therapy in these high risk individuals, but also 10 / 15 suggests that better treatments are still required to further increase this remission rate. It is hoped that novel immunsuppressive therapies in development may improve outcomes with less toxicity, and that our recent advances in understanding the role of anti PLA2 receptor antibodies will lead to better targeted treatments. In recent years, new immunosuppressive agents such as rituximab, Tripterygium wilfordii and others have been used in IMN. However, there are no RCTs to assess the efficacy and safety of used for IMN. Tripterygium wilfordii is a traditional Chinese medicine with immunosuppressive effects. It has been reported to be effective in Chinese patients, but the mechanism is not clear [ 55 ]. Advantages and innovations Previous studies had analyzed the effect of immunosuppressive therapy and (or) non-immunosuppressive therapy only by direct comparison. This study is the first time to comprehensive analysis the effectiveness and tolerance of different treatments for IMN by network meta-analysis, including direct and indirect comparison. The advantage of this new method is to compare a variety of treatment options which used for the same disease and rank the treatments according to the analysis results. Classic meta-analysis often focused on these treatments which common application in clinical, such as: CTX, tacrolimus and cyclosporine. But for the other treatments which RCTs were fewer or minimal application in clinical such as leflunomide, azathioprine, mizoribine, there is no evidence-based research. In this paper, all of the RCTs that met the inclusion criteria were included in the comparisons. For those critical positive results, we also confirmed by direct and indirect comparisons. Limitations Our study used a comprehensive network meta-analysis approach to compare the effectiveness and tolerability of immunosuppressive treatments for idiopathic membranous nephropathy. It has quantitatively assessed and compared different immunosuppressive agents used for IMN [ 56 ]. However our study has a number of limitations. First, the quality of the including studies varied, causing significant heterogeneity. Second, the potential for reporting bias exists, which may influence the results. Third, the sample sizes of these studies are relatively small, reducing statistical power. Conclusions Cyclophosphamide and chlorambucil reduce risk of ESKD or death in IMN with nephrotic range proteinuria, but carry substantial toxicity that may be lower for cyclophosphamide. Tacrolimus and cyclosporine increase the possibility of proteinuria remission with less drug withdrawal, but the effects on kidney failure remain uncertain. Supporting information S1 Fig. The individual risk of bias data for the included studies. (TIF) S2 Fig. The aggregate risk of bias data for the included studies. (TIF) S3 Fig. The cumulative probability graph for the total mortality or ESKD outcome. The solid line represents the estimated probabilities and the dotted line represents the predicted probabilities. The larger the area under the curve, the higher the ranking. (TIF) 11 / 15 S4 Fig. The cumulative probability graph for the total remission outcome. The solid line represents the estimated probabilities and the dotted line represents the predicted probabilities. The larger the area under the curve, the higher the ranking. (TIF) S5 Fig. The cumulative probability graph for the withdrawal outcome. The solid line represents the estimated probabilities and the dotted line represents the predicted probabilities. The larger the area under the curve, the higher the ranking. (TIF) S6 Fig. The result of sensitivity meta-analysis of total mortality or ESKD (excluding low quality studies). (TIF) S7 Fig. The result of sensitivity meta-analysis of total remission (excluding low quality studies). (TIF) S8 Fig. The result of sensitivity meta-analysis of withdrawal (excluding low quality studies). (TIF) S1 Table. Search strategy. (DOCX) S1 File. PRISMA checklist. (DOC) Author Contributions S2 Table. Characteristics of studies included in systematic review and meta-analysis. (DOCX) Conceptualization: Guisen Li, Vlado Perkovic, Daqing Hong. Data curation: Song Ren, Ying Wang, Li Xian, Tadashi Toyama, Daqing Hong. Formal analysis: Song Ren, Daqing Hong. Funding acquisition: Guisen Li. Investigation: Song Ren, Li Xian, Daqing Hong. Methodology: Guisen Li, Vlado Perkovic, Daqing Hong. Resources: Daqing Hong. Software: Song Ren, Daqing Hong. Visualization: Song Ren, Daqing Hong. Writing ± original draft: Song Ren, Ying Wang, Meg Jardine, Daqing Hong. Writing ± review & editing: Guisen Li, Vlado Perkovic. 12 / 15 13 / 15 14 / 15 1. Haas M , Meehan SM , Karrison TG , Spargo BH ( 1997 ) Changing etiologies of unexplained adult nephrotic syndrome: a comparison of renal biopsy findings from 1976±1979 and 1995±1997 . Am J Kidney Dis 30 : 621 ± 631 . PMID: 9370176 2. Hladunewich MA , Troyanov S , Calafati J , Cattran DC , Metropolitan Toronto Glomerulonephritis R ( 2009 ) The natural history of the non-nephrotic membranous nephropathy patient . Clin J Am Soc Nephrol 4 : 1417 ± 1422 . https://doi.org/10.2215/CJN.01330209 PMID: 19661220 3. Davison AM , Cameron JS , Kerr DN , Ogg CS , Wilkinson RW ( 1984 ) The natural history of renal function in untreated idiopathic membranous glomerulonephritis in adults . Clin Nephrol 22 : 61 ± 67 . PMID: 6478673 4. Hofstra JM , Wetzels JF ( 2012 ) Management of patients with membranous nephropathy . Nephrol Dial Transplant 27 : 6±9 . https://doi.org/10.1093/ndt/gfr371 PMID: 21737514 5. KDIGO Idiopathic membranous nephropathy . Kidney International Supplements 2 : 186 ± 197 . https:// doi.org/10.1038/kisup. 2012 .20 PMID: 25018932 6. Moher D , Cook DJ , Jadad AR , Tugwell P , Moher M , Jones A , et al. ( 1999 ) Assessing the quality of reports of randomised trials: implications for the conduct of meta-analyses . Health Technol Assess 3 : i±iv, 1 ± 98 . 7. Higgins J GS e Cochrane Handbook for Systematic Reviews of Interventions . Version 5. 1. 0 [updated March 2011 ]. The Cochrane Collaboration . 8. Higgins J GS, eds. Cochrane Handbook for Systematic Reviews of Interventions. Version 5. 1. 0 [updated March 2011 ]. The Cochrane Collaboration . 9. F S, T X , S P-B ( 2011 ) Inconsistency between direct and indirect comparisons of competing interventions: meta-epidemiological study . BMJ 343 : d4949 . 10. Xu J , Zhang W , Xu Y , Shen P , Ren H , Wang W , et al. ( 2013 ) Tacrolimus combined with corticosteroids in idiopathic membranous nephropathy: a randomized, prospective, controlled trial . Contrib Nephrol 181 : 152 ± 162 . https://doi.org/10.1159/000348475 PMID: 23689577 11. Howman A , Chapman TL , Langdon MM , Ferguson C , Adu D , Feehally J , et al. ( 2013 ) Immunosuppression for progressive membranous nephropathy: A UK randomised controlled trial . The Lancet 381 : 744 ± 751 . 12. Naumovic R , Jovanovic D , Pavlovic S , Stosovic M , Marinkovic J , Basta-Jovanovic G ( 2011 ) Cyclosporine versus azathioprine therapy in high-risk idiopathic membranous nephropathy patients: A 3-year prospective study . Biomed Pharmacother 65 : 105 ± 110 . https://doi.org/10.1016/j.biopha. 2010 . 10 .009 PMID: 21109389 13. Kosmadakis G , Filiopoulos V , Smirloglou D , Skarlas P , Georgoulias C , Michail S ( 2010 ) Comparison of immunosuppressive therapeutic regimens in patients with nephrotic syndrome due to idiopathic membranous nephropathy . Renal failure 32 : 566± 571 . https://doi.org/10.3109/08860221003728754 PMID: 20486839 14. Chen M , Li H , Li XY , Lu FM , Ni ZH , Xu FF , et al. ( 2010 ) Tacrolimus combined with corticosteroids in treatment of nephrotic idiopathic membranous nephropathy: A multicenter randomized controlled trial . American Journal of the Medical Sciences 339 : 233 ± 238 . https://doi.org/10.1097/MAJ. 0b013e3181ca3a7d PMID: 20220333 15. Senthil Nayagam L , Ganguli A , Rathi M , Kohli HS , Gupta KL , Joshi K , et al. ( 2008 ) Mycophenolate mofetil or standard therapy for membranous nephropathy and focal segmental glomerulosclerosis: a pilot study . Nephrol Dial Transplant 23 : 1926 ± 1930 . https://doi.org/10.1093/ndt/gfm538 PMID: 17989103 16. Dussol B , Morange S , Burtey S , Indreies M , Cassuto E , Mourad G , et al. ( 2008 ) Mycophenolate Mofetil Monotherapy in Membranous Nephropathy: A 1-Year Randomized Controlled Trial . American Journal of Kidney Diseases 52 : 699 ± 705 . https://doi.org/10.1053/j.ajkd. 2008 . 04 .013 PMID: 18585835 17. Praga M , Barrio V , Juarez GF , Luno J , Grupo Espanol de Estudio de la Nefropatia M ( 2007 ) Tacrolimus monotherapy in membranous nephropathy: a randomized controlled trial . Kidney Int 71 : 924 ± 930 . https://doi.org/10.1038/sj.ki. 5002215 PMID: 17377504 18. Jha V , Ganguli A , Saha TK , Kohli HS , Sud K , Gupta KL , et al. ( 2007 ) A randomized, controlled trial of steroids and cyclophosphamide in adults with nephrotic syndrome caused by idiopathic membranous nephropathy . J Am Soc Nephrol 18 : 1899 ± 1904 . https://doi.org/10.1681/ASN.2007020166 PMID: 17494881 19. Chan TM , Lin AW , Tang SC , Qian JQ , Lam MF , Ho YW , et al. ( 2007 ) Prospective controlled study on mycophenolate mofetil and prednisolone in the treatment of membranous nephropathy with nephrotic syndrome . 20. Shibasaki T , Koyama A , Hishida A , Muso E , Osawa G , Yamabe H , et al. ( 2004 ) A randomized openlabel comparative study of conventional therapy versus mizoribine onlay therapy in patients with steroid-resistant nephrotic syndrome (postmarketing survey) . Clin Exp Nephrol 8 : 117 ± 126 . https://doi. org/10.1007/s10157-004 -0276-0 PMID: 15235928 21. Cattran DC , Appel GB , Hebert LA , Hunsicker LG , Pohl MA , Hoy WE , et al. ( 2001 ) Cyclosporine in patients with steroid-resistant membranous nephropathy: A randomized trial . 22. Ponticelli C , Altieri P , Scolari F , Passerini P , Roccatello D , Cesana B , et al. ( 1998 ) A randomized study comparing methylprednisolone plus chlorambucil versus methylprednisolone plus cyclophosphamide in idiopathic membranous nephropathy . 23. Branten AJW , Reichert LJM , Koene RAP , Wetzels JFM ( 1998 ) Oral cyclophosphamide versus chlorambucil in the treatment of patients with membranous nephropathy and renal insufficiency . QJMÐMonthly Journal of the Association of Physicians 91 : 359 ± 366 . PMID: 9709470 24. Austin HA , Vaughan EM , Boumpas DT , Balow JE ( 1996 ) Randomized trial of pulse cyclophosphamide and prednisone vs. prednisone alone in idiopathic membranous nephropathy. [abstract no: A0410] . Journal of the American Society of Nephrology: JASN 7 : 1327 . 25. Ponticelli C , Zucchelli P , Passerini P , Cesana B , Locatelli F , Pasquali S , et al. ( 1995 ) A 10-year followup of a randomized study with methylprednisolone and chlorambucil in membranous nephropathy . Kidney International 48 : 1600± 1604 . PMID: 8544420 26. Cattran DC , Greenwood C , Ritchie S , Bernstein K , Churchill DN , Clark WF , et al. ( 1995 ) A controlled trial of cyclosporine in patients with progressive membranous nephropathy . Canadian Glomerulonephritis Study Group. 27. Reichert LJM , Huysmans FTM , Assmann K , Koene RAP , Wetzels JFM ( 1994 ) Preserving renal function in patients with membranous nephropathy: Daily oral chlorambucil compared with intermittent monthly pulses of cyclophosphamide . Annals of Internal Medicine 121 : 328 ± 333 . PMID: 8042821 28. Ponticelli C , Zucchelli P , Passerini P , Cesana B ( 1992 ) Methylprednisolone plus chlorambucil as compared with methylprednisolone alone for the treatment of idiopathic membranous nephropathy . 29. Falk RJ , Hogan SL , Muller KE , Jennette JC , Moser A , Crews D , et al. ( 1992 ) Treatment of progressive membranous glomerulopathy. A randomized trial comparing cyclophosphamide and corticosteroids with corticosteroids alone . Annals of Internal Medicine 116 : 438 ± 445 . 30. Cameron JS , Healy MJ , Adu D ( 1990 ) The Medical Research Council trial of short-term high-dose alternate day prednisolone in idiopathic membranous nephropathy with nephrotic syndrome in adults. The MRC Glomerulonephritis Working Party . Quarterly journal of medicine 74: 133±156. PMID: 2189149 31. Cattran DC , Delmore T , Roscoe J , Cole E , Cardella C , Charron R , et al. ( 1989 ) A randomized controlled trial of prednisone in patients with idiopathic membranous nephropathy . 32. Silverberg DS , Atkins EL , Ballon HC ( 1976 ) Controlled trial of azathioprine in the nephrotic syndrome secondary to idiopathic membranous glomerulonephritis . Canadian Medical Association Journal 115 : 1209 ± 1210 . PMID: 793706 33. Donadio JV , Holley KE , Anderson CF , Taylor WF ( 1974 ) Controlled trial of cyclophosphamide in idiopathic membranous nephropathy . Kidney international 6: 431±439. PMID: 4613934 34. Arnadottir M , Stefansson B , Berg A ( 2006 ) A randomized study on treatment with adrenocorticotrophic hormone in idiopathic membranous nephropathy [abstract no: F-PO1112] . Journal of the American Society of Nephrology: JASN 17 : 571A . 35. Braun N , Dietz K , Vonthein R , Erley C , Risler T ( 2004 ) Remission and relapses in idiopathic membranous nephropathy. Description by a longitudinal multi-state markov-model [abstract] . 41st Congress European Renal Association European Dialysis and Transplantation Association. 36. Dyadyk AI , Bagriy AE , Yarovaya NF , Dyadyk IA ( 2001 ) Results of long-term randomised study of immunosupressive treatment of patients with idiopathic membranous glomerulonephritis [abstract] . Nephrology Dialysis Transplantation 16 : A64 . 37. Murphy BF , McDonald I , Fairley KF , Kincaid-Smith PS ( 1992 ) Randomized controlled trial of cyclophosphamide, warfarin and dipyridamole in idiopathic membranous glomerulonephritis . Clinical nephrology 37: 229±234. PMID: 1606772 38. Coggins CH ( 1979 ) A controlled study of short-term prednisone treatment in adults with membranous nephropathy. Collaborative study of the adult idiopathic nephrotic syndrome . New England Journal of Medicine 301 : 1301 ± 1306 . https://doi.org/10.1056/NEJM197912133012401 PMID: 388220 39. Ahmed S , Rahman M , Alam MR , Islam S , Chowdhury MN , Chowdhury SMZ , et al. ( 1994 ) Methyl prednisolone plus chlorambucil as compared with prednisolone alone for the treatment of idiopathic membranous nephropathyÐA preliminary study . Bangladesh Renal Journal 13 : 51 ± 54 . 40. Pahari DK , Das S , Dutta BN , Banerjee D ( 1993 ) Prognosis and management of membraneous nephropathy . Journal of the Association of Physicians of India 41 : 350 ± 351 . PMID: 8005970 41. Tiller DJ , Clarkson AR , Mathew T , Thompson N , Row G , Lauer C , et al. ( 1981 ) A prospective randomised trial of cyclophosphamide, dipyridamole and warfarin in membranous and mesangiocapillary glomerulonephritis . Advances in Basic and Clinical Nephrology 8th International Congress of Nephrology. 42. Koshikawa S ( 1993 ) [Clinical evaluation of an immunosuppressive drug, mizoribine (HE-69) on steroid resistant nephrotic syndromeÐa multicenter double-blind comparison study with placebo] [Japanese] . Jin to Toseki [Kidney and Dialysis] 34 : 631 ± 650 . 43. Peng L , Wei SY , Li LT , He YX , Li B ( 2016 ) Comparison of different therapies in high-risk patients with idiopathic membranous nephropathy . J Formos Med Assoc 115 : 11 ± 18 . https://doi.org/10.1016/j.jfma. 2015 . 07 .021 PMID: 26315481 44. Ramachandran R , Hn HK , Kumar V , Nada R , Yadav AK , Goyal A , et al. ( 2016 ) Tacrolimus combined with corticosteroids versus Modified Ponticelli regimen in treatment of idiopathic membranous nephropathy: Randomized control trial . Nephrology (Carlton) 21 : 139 ± 146 . 45. Liu S ( 2015 ) Clinical Trial of Treatment for Idiopathic Membranous Nephropathy with Leflunomide Combined with Cyclophosphamide and Glucocorticoid . Hong Kong Journal of Nephrology 17 : S64 . 46. Beck LH Jr., Bonegio RG , Lambeau G , Beck DM , Powell DW , Cummins TD , et al. ( 2009 ) M-type phospholipase A2 receptor as target antigen in idiopathic membranous nephropathy . N Engl J Med 361 : 11 ± 21 . https://doi.org/10.1056/NEJMoa0810457 PMID: 19571279 47. Tomas NM , Beck LH , Jr., Meyer-Schwesinger C , Seitz-Polski B , Ma H , Zahner G , et al. ( 2014 ) Thrombospondin type-1 domain-containing 7A in idiopathic membranous nephropathy . N Engl J Med 371 : 2277 ± 2287 . https://doi.org/10.1056/NEJMoa1409354 PMID: 25394321 48. Larsen CP , Cossey LN , Beck LH ( 2016 ) THSD7A staining of membranous glomerulopathy in clinical practice reveals cases with dual autoantibody positivity . Mod Pathol 29 : 421 ± 426 . https://doi.org/10. 1038/modpathol. 2016 .32 PMID: 26847174 49. Iwakura T , Ohashi N , Kato A , Baba S , Yasuda H ( 2015 ) Prevalence of Enhanced Granular Expression of Thrombospondin Type-1 Domain-Containing 7A in the Glomeruli of Japanese Patients with Idiopathic Membranous Nephropathy . PLoS One 10 : e0138841. https://doi.org/10.1371/journal.pone. 0138841 PMID: 26393352 50. Hoxha E , Beck LH , Jr. , Wiech T , Tomas NM , Probst C , Mindorf S , et al. ( 2017 ) An Indirect Immunofluorescence Method Facilitates Detection of Thrombospondin Type 1 Domain-Containing 7A-Specific Antibodies in Membranous Nephropathy . J Am Soc Nephrol 28 : 520 ± 531 . https://doi.org/10.1681/ASN. 2016010050 PMID: 27436855 51. Chen Y , Schieppati A , Cai G , Chen X , Zamora J , Giuliano GA , et al. ( 2013 ) Immunosuppression for membranous nephropathy: a systematic review and meta-analysis of 36 clinical trials . Clinical Journal of The American Society of Nephrology: CJASN 8 : 787 ± 796 . https://doi.org/10.2215/CJN.07570712 PMID: 23449768 52. Drachenberg CB , Klassen DK , Weir MR , Wiland A , Fink JC , Bartlett ST , et al. ( 1999 ) Islet cell damage associated with tacrolimus and cyclosporine: morphological features in pancreas allograft biopsies and clinical correlation . Transplantation 68 : 396 ± 402 . PMID: 10459544 53. Burroughs TE , Lentine KL , Takemoto SK , Swindle J , Machnicki G , Hardinger K , et al. ( 2007 ) Influence of early posttransplantation prednisone and calcineurin inhibitor dosages on the incidence of new-onset diabetes . Clin J Am Soc Nephrol 2 : 517 ± 523 . https://doi.org/10.2215/CJN.00620206 PMID: 17699459 54. Qiu TT , Zhang C , Zhao HW , Zhou JW ( 2016 ) Calcineurin inhibitors versus cyclophosphamide for idiopathic membranous nephropathy: A systematic review and meta-analysis of 21 clinical trials . Autoimmun Rev. 55. Liu S , Li X , Li H , Liang Q , Chen J , Chen J ( 2015 ) Comparison of tripterygium wilfordii multiglycosides and tacrolimus in the treatment of idiopathic membranous nephropathy: a prospective cohort study . BMC Nephrol 16 : 200 . https://doi.org/10.1186/s12882-015-0199 -x PMID : 26637482 56. Lumley T ( 2002 ) Network meta-analysis for indirect treatment comparisons . Stat Med 21 : 2313 ± 2324 . https://doi.org/10.1002/sim.1201 PMID: 12210616


This is a preview of a remote PDF: http://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0184398&type=printable

Song Ren, Ying Wang, Li Xian, Tadashi Toyama, Meg Jardine, Guisen Li, Vlado Perkovic, Daqing Hong. Comparative effectiveness and tolerance of immunosuppressive treatments for idiopathic membranous nephropathy: A network meta-analysis, PLOS ONE, 2017, DOI: 10.1371/journal.pone.0184398