Non-immunosuppressive treatment for IgA nephropathy (Review)

Non-immunosuppressive treatment for IgA nephropathy (Review) Reid S, Cawthon PM, Craig JC, Samuels JA, Molony DA, Strippoli GFM This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library 2011, Issue 3 http://www.thecochranelibrary.com

T A B L E O F C O N T E N T S HEADER....................................... 1 ABSTRACT...................................... 1 PLAIN LANGUAGE SUMMARY.............................. 2 BACKGROUND.................................... 2 OBJECTIVES..................................... 3 METHODS...................................... 3 RESULTS....................................... 5 Figure 1...................................... 6 Figure 2...................................... 8 Figure 3...................................... 9 DISCUSSION..................................... 14 AUTHORS CONCLUSIONS............................... 16 ACKNOWLEDGEMENTS................................ 17 REFERENCES..................................... 18 CHARACTERISTICS OF STUDIES............................. 28 DATA AND ANALYSES.................................. 96 Analysis 1.1. Comparison 1 Fish oil versus placebo/no treatment, Outcome 1 ESKD............ 104 Analysis 1.2. Comparison 1 Fish oil versus placebo/no treatment, Outcome 2 > 50% increase in serum creatinine.. 104 Analysis 1.3. Comparison 1 Fish oil versus placebo/no treatment, Outcome 3 > 50% decrease in creatinine clearance. 105 Analysis 1.4. Comparison 1 Fish oil versus placebo/no treatment, Outcome 4 All-cause mortality........ 105 Analysis 1.5. Comparison 1 Fish oil versus placebo/no treatment, Outcome 5 Serum creatinine......... 106 Analysis 1.7. Comparison 1 Fish oil versus placebo/no treatment, Outcome 7 Change in serum creatinine..... 106 Analysis 1.9. Comparison 1 Fish oil versus placebo/no treatment, Outcome 9 Creatinine clearance....... 107 Analysis 1.11. Comparison 1 Fish oil versus placebo/no treatment, Outcome 11 Proteinuria.......... 108 Analysis 1.12. Comparison 1 Fish oil versus placebo/no treatment, Outcome 12 Change in proteinuria...... 108 Analysis 1.14. Comparison 1 Fish oil versus placebo/no treatment, Outcome 14 Change in haematuria (cells/ml x 105). 109 Analysis 2.1. Comparison 2 Fish oil versus symptomatic treatment, Outcome 1 ESKD............ 110 Analysis 2.2. Comparison 2 Fish oil versus symptomatic treatment, Outcome 2 > 50% increase in serum creatinine.. 110 Analysis 2.3. Comparison 2 Fish oil versus symptomatic treatment, Outcome 3 > 50% decrease in creatinine clearance. 111 Analysis 2.4. Comparison 2 Fish oil versus symptomatic treatment, Outcome 4 Serum creatinine........ 111 Analysis 2.6. Comparison 2 Fish oil versus symptomatic treatment, Outcome 6 Creatinine clearance....... 112 Analysis 2.8. Comparison 2 Fish oil versus symptomatic treatment, Outcome 8 Proteinuria.......... 112 Analysis 3.1. Comparison 3 Fish oil: high versus low dose, Outcome 1 ESKD............... 113 Analysis 4.1. Comparison 4 Fish oil+acei+arb versus ACEi+ARB, Outcome 1 Creatinine clearance...... 114 Analysis 4.2. Comparison 4 Fish oil+acei+arb versus ACEi+ARB, Outcome 2 Proteinuria......... 114 Analysis 5.1. Comparison 5 Antihypertensive agents versus placebo/no treatment, Outcome 1 ESKD or doubling serum creatinine.................................... 115 Analysis 5.2. Comparison 5 Antihypertensive agents versus placebo/no treatment, Outcome 2 Remission of proteinuria. 116 Analysis 5.4. Comparison 5 Antihypertensive agents versus placebo/no treatment, Outcome 4 Serum creatinine... 117 Analysis 5.5. Comparison 5 Antihypertensive agents versus placebo/no treatment, Outcome 5 Creatinine clearance.. 118 Analysis 5.6. Comparison 5 Antihypertensive agents versus placebo/no treatment, Outcome 6 Proteinuria..... 119 Analysis 6.1. Comparison 6 Antihypertensive agents versus other treatment, Outcome 1 ESKD........ 120 Analysis 6.2. Comparison 6 Antihypertensive agents versus other treatment, Outcome 2 > 50% increase in serum creatinine.................................... 121 Analysis 6.3. Comparison 6 Antihypertensive agents versus other treatment, Outcome 3 Doubling of serum creatinine. 121 Analysis 6.4. Comparison 6 Antihypertensive agents versus other treatment, Outcome 4 Serum creatinine..... 122 Analysis 6.5. Comparison 6 Antihypertensive agents versus other treatment, Outcome 5 Creatinine clearance.... 124 Analysis 6.7. Comparison 6 Antihypertensive agents versus other treatment, Outcome 7 Proteinuria....... 126 Analysis 6.8. Comparison 6 Antihypertensive agents versus other treatment, Outcome 8 Change in proteinuria... 128 Analysis 6.10. Comparison 6 Antihypertensive agents versus other treatment, Outcome 10 Haematuria...... 130 Analysis 7.1. Comparison 7 Antihypertensive agents+other treatment versus other treatment, Outcome 1 ESKD... 132 i

Analysis 7.2. Comparison 7 Antihypertensive agents+other treatment versus other treatment, Outcome 2 Serum creatinine.................................... 132 Analysis 7.3. Comparison 7 Antihypertensive agents+other treatment versus other treatment, Outcome 3 Change in serum creatinine.................................... 133 Analysis 7.4. Comparison 7 Antihypertensive agents+other treatment versus other treatment, Outcome 4 Creatinine clearance.................................... 134 Analysis 7.5. Comparison 7 Antihypertensive agents+other treatment versus other treatment, Outcome 5 Proteinuria. 135 Analysis 7.7. Comparison 7 Antihypertensive agents+other treatment versus other treatment, Outcome 7 Haematuria. 136 Analysis 8.1. Comparison 8 Anticoagulant versus placebo/no treatment, Outcome 1 ESKD.......... 137 Analysis 8.2. Comparison 8 Anticoagulant versus placebo/no treatment, Outcome 2 Serum creatinine...... 137 Analysis 8.3. Comparison 8 Anticoagulant versus placebo/no treatment, Outcome 3 Creatinine clearance..... 138 Analysis 8.5. Comparison 8 Anticoagulant versus placebo/no treatment, Outcome 5 Proteinuria........ 138 Analysis 8.7. Comparison 8 Anticoagulant versus placebo/no treatment, Outcome 7 Haematuria........ 139 Analysis 9.1. Comparison 9 Anticoagulant versus other treatment, Outcome 1 Serum creatinine........ 140 Analysis 9.2. Comparison 9 Anticoagulant versus other treatment, Outcome 2 Creatinine clearance....... 140 Analysis 9.3. Comparison 9 Anticoagulant versus other treatment, Outcome 3 Proteinuria.......... 141 Analysis 10.1. Comparison 10 Anticoagulant+other treatment versus other treatment, Outcome 1 ESKD..... 141 Analysis 10.2. Comparison 10 Anticoagulant+other treatment versus other treatment, Outcome 2 Serum creatinine. 142 Analysis 10.3. Comparison 10 Anticoagulant+other treatment versus other treatment, Outcome 3 Change in creatinine clearance.................................... 142 Analysis 10.5. Comparison 10 Anticoagulant+other treatment versus other treatment, Outcome 5 Proteinuria... 143 Analysis 11.1. Comparison 11 Tonsillectomy+other treatment versus other treatment, Outcome 1 Remission of proteinuria................................... 143 Analysis 11.2. Comparison 11 Tonsillectomy+other treatment versus other treatment, Outcome 2 Remission of haematuria................................... 144 Analysis 11.3. Comparison 11 Tonsillectomy+other treatment versus other treatment, Outcome 3 Serum creatinine. 145 Analysis 11.4. Comparison 11 Tonsillectomy+other treatment versus other treatment, Outcome 4 Creatinine clearance. 145 Analysis 11.5. Comparison 11 Tonsillectomy+other treatment versus other treatment, Outcome 5 Proteinuria... 146 Analysis 13.1. Comparison 13 Statins+other treatment versus other treatment, Outcome 1 Serum creatinine.... 147 Analysis 13.2. Comparison 13 Statins+other treatment versus other treatment, Outcome 2 Creatinine clearance... 148 Analysis 13.3. Comparison 13 Statins+other treatment versus other treatment, Outcome 3 Proteinuria...... 148 Analysis 13.4. Comparison 13 Statins+other treatment versus other treatment, Outcome 4 Haematuria...... 149 Analysis 14.1. Comparison 14 Phenytoin versus placebo/no treatment, Outcome 1 Remission of haematuria.... 149 Analysis 14.2. Comparison 14 Phenytoin versus placebo/no treatment, Outcome 2 Serum creatinine....... 150 Analysis 14.3. Comparison 14 Phenytoin versus placebo/no treatment, Outcome 3 Creatinine clearance..... 150 Analysis 14.4. Comparison 14 Phenytoin versus placebo/no treatment, Outcome 4 Proteinuria......... 151 Analysis 15.1. Comparison 15 Herbal medicine versus placebo/no treatment, Outcome 1 Serum creatinine.... 151 Analysis 15.2. Comparison 15 Herbal medicine versus placebo/no treatment, Outcome 2 Proteinuria...... 152 Analysis 15.3. Comparison 15 Herbal medicine versus placebo/no treatment, Outcome 3 Haematuria...... 153 Analysis 16.1. Comparison 16 Herbal medicine versus other treatment, Outcome 1 Serum creatinine...... 153 Analysis 16.2. Comparison 16 Herbal medicine versus other treatment, Outcome 2 Proteinuria........ 154 Analysis 16.3. Comparison 16 Herbal medicine versus other treatment, Outcome 3 Haematuria........ 154 Analysis 17.2. Comparison 17 Herbal medicine+other treatment versus other treatment, Outcome 2 Serum creatinine. 155 Analysis 17.3. Comparison 17 Herbal medicine+other treatment versus other treatment, Outcome 3 Proteinuria... 156 Analysis 18.1. Comparison 18 Urokinase+ACEi versus ACEi, Outcome 1 > 50% increase in serum creatinine.... 156 Analysis 18.2. Comparison 18 Urokinase+ACEi versus ACEi, Outcome 2 > 50% decrease in proteinuria..... 157 Analysis 18.3. Comparison 18 Urokinase+ACEi versus ACEi, Outcome 3 Serum creatinine.......... 157 Analysis 18.4. Comparison 18 Urokinase+ACEi versus ACEi, Outcome 4 Creatinine clearance......... 158 Analysis 18.5. Comparison 18 Urokinase+ACEi versus ACEi, Outcome 5 Proteinuria............ 158 Analysis 19.1. Comparison 19 Vitamin E versus placebo/no treatment, Outcome 1 ESKD.......... 159 Analysis 19.2. Comparison 19 Vitamin E versus placebo/no treatment, Outcome 2 Creatinine clearance..... 159 Analysis 19.3. Comparison 19 Vitamin E versus placebo/no treatment, Outcome 3 Proteinuria......... 160 Analysis 20.1. Comparison 20 Sodium cromoglycate versus placebo/no treatment, Outcome 1 Creatinine clearance. 161 ii

Analysis 20.2. Comparison 20 Sodium cromoglycate versus placebo/no treatment, Outcome 2 Proteinuria..... 161 APPENDICES..................................... 161 HISTORY....................................... 174 CONTRIBUTIONS OF AUTHORS............................. 175 DECLARATIONS OF INTEREST.............................. 175 DIFFERENCES BETWEEN PROTOCOL AND REVIEW..................... 175 INDEX TERMS.................................... 175 iii

[Intervention Review] Non-immunosuppressive treatment for IgA nephropathy Sharon Reid 1, Peggy M Cawthon 2, Jonathan C Craig 3, Joshua A Samuels 4, Donald A Molony 5, Giovanni FM Strippoli 6 1 Sydney School of Public Health, University of Sydney, Sydney, Australia. 2 Research Institute, California Pacific Medical Center, San Francisco, CA, USA. 3 a) Cochrane Renal Group, Centre for Kidney Research, The Children s Hospital at Westmead, b) Sydney School of Public Health, The University of Sydney, Sydney, Australia. 4 Nephrology / Pediatric Nephrology, UT-Houston Health Science Center, Houston, TX, USA. 5 Internal Medicine, UT-Houston Health Science Center, Houston, TX, USA. 6 a) Mario Negri Sud Consortium, Santa Maria Imbaro (Ch), Italy, b) Sydney School of Public Health, University of Sydney, Australia, c) Diaverum Medical- Scientific Office, Lund, Sweden Contact address: Giovanni FM Strippoli, Cochrane Renal Group, The Children s Hospital at Westmead, Locked Bag 4001, Westmead, NSW, 2145, Australia. Giovanni.Strippoli@diaverum.com. strippoli@negrisud.it. Editorial group: Cochrane Renal Group. Publication status and date: New, published in Issue 3, 2011. Review content assessed as up-to-date: 13 July 2010. Citation: Reid S, Cawthon PM, Craig JC, Samuels JA, Molony DA, Strippoli GFM. Non-immunosuppressive treatment for IgA nephropathy. Cochrane Database of Systematic Reviews 2011, Issue 3. Art. No.: CD003962. DOI: 10.1002/14651858.CD003962.pub2. Background A B S T R A C T IgA nephropathy (IgAN) is the most common primary glomerular disease with approximately 30% to 40% of patients progressing to end-stage kidney disease (ESKD) within 20 years. The most common regimens include immunosuppressive agents, however the risks of long-term treatment often outweigh the potential benefits. Non-immunosuppressive options, including fish oils, anticoagulants, antihypertensive agents and tonsillectomy have also been examined but not reviewed systematically. Objectives To assess the benefits and harms of non-immunosuppressive treatments for treating IgAN in adults and children. Search methods In July 2010 we searched the Cochrane Renal Group s specialised register, CENTRAL (in The Cochrane Library), MEDLINE (from 1966) and EMBASE (from 1980). We also searched reference lists of included studies, review articles and contacted local and international experts. Selection criteria Randomised controlled trials (RCTs) of non-immunosuppressive agents in adults and children with biopsy-proven IgAN were included. Data collection and analysis Two authors independently reviewed search results, extracted data and assessed study quality. Results were expressed as mean differences (MD) for continuous outcomes and risk ratios (RR) for dichotomous outcomes with 95% confidence intervals (CI) using a randomeffects model. 1

Main results We included 56 studies (2838 participants). Antihypertensive agents were the most beneficial non-immunosuppressive intervention for IgAN. The antihypertensives examined were predominantly angiotensin-converting enzyme inhibitors (ACEi), angiotensin receptor blockers (ARB) or combinations of both, versus other antihypertensives and other agents. The benefits of antihypertensive agents, particularly inhibitors of the renin angiotensin system, appear to potentially outweigh the harms in patients with IgAN. The benefits are largely manifest as a reduction in proteinuria, a surrogate outcome. There is no evidence that treatment with any of the antihypertensive agents evaluated affect major renal and/or cardiovascular endpoints or long-term mortality risk beyond the benefit that arises from controlling hypertension in patients with IgAN. The RCT evidence is insufficiently robust to demonstrate efficacy for any of the other non-immunosuppressive therapies evaluated here. Authors conclusions IgAN remains a disease in search of adequately powered RCTs to reliably inform clinical practice. More and better evidence is needed to understand the magnitude of benefit and the possible risks of anti-hypertensive or more specifically of ACEi/ARB therapy alone or in combination and which specific types of patients with the IgAN might have the greatest potential for benefit. For other nonimmunosuppressive therapies, where neither benefit nor significant harm has yet to be demonstrated, there remains some justification for further exploration of the potential benefits. P L A I N L A N G U A G E S U M M A R Y Non-immunosuppressive treatment for IgA nephropathy IgA nephropathy (IgAN) is the most common primary glomerular disease with approximately 30% to 40% of patients progressing to end-stage kidney disease (ESKD) within 20 years. The most common regimens include immunosuppressive agents, however the risks of long-term treatment often outweigh the potential benefits. Non-immunosuppressive options, including fish oils, anticoagulants, antihypertensive agents and tonsillectomy have also been examined but not reviewed systematically. We included 56 studies enrolling 2838 participants. Antihypertensive agents were the most beneficial non-immunosuppressive intervention for IgAN. The antihypertensives examined here were predominantly angiotensin-converting enzyme inhibitors (ACEi), angiotensin receptor blockers (ARB) or combinations of both, compared with other antihypertensives and other agents. The benefits of antihypertensive agents appear to potentially outweigh the harms in patients with IgAN. The benefits were mostly reported as a reduction in 24 hour proteinuria. There is no evidence that treatment with any of the antihypertensive agents evaluated to date affect major renal and/or cardiovascular endpoints or long-term mortality risk beyond the benefit that arises from controlling hypertension in patients with IgAN. The RCT evidence is insufficiently robust to demonstrate efficacy for any of the other non-immunosuppressive therapies evaluated here. IgAN remains a disease in search of adequately powered RCTs to reliably inform clinical practice. More and better evidence is needed to understand the magnitude of benefit and the possible risks of anti-hypertensive or more specifically of ACEi/ARB therapy alone or in combination and which specific types of patients with the IgAN might have the greatest potential for benefit. For other nonimmunosuppressive therapies, where benefit nor significant harm has yet to be demonstrated, there remains some justification for further exploration of the potential benefits. B A C K G R O U N D Description of the condition IgA nephropathy (IgAN) was first described in 1968 by Dr J Berger. Characterised by prominent diffuse mesangial IgA deposits at immunofluorescent microscopy, the condition was initially thought to be a rare and benign cause of recurrent haematuria (Berger 1968). It has since become apparent that IgAN is 2

neither rare nor entirely benign. Although biopsy practices differ from region to region, affecting the frequency of diagnosis of IgAN, it is now known that IgAN is the most common primary glomerular disease worldwide (D Amico 1987). While about 20% of patients affected by IgAN experience lasting remission (Nolin 1999), end-stage kidney disease (ESKD) will develop in 15% to 20% of patients within 10 years from the apparent onset of the disease and in up to 30% to 40% within 20 years (Schena 2001). Most individuals who develop IgAN have degalctosylated IgA involving 0-linked glycans in the hinge region of IgA1. This results in the potential for the generation of autoantibodies and the formation of circulating immune complexes which deposit in the glomerular mesangium and lead to the clinical manifestations of the disease (Glassock 2009; Julian 1999; Mestecky 1993). Despite the progression in understanding of the pathogenesis of IgAN, a consensus on optimal treatment for IgAN has yet to be established. Recent meta-analyses have examined the evidence for treatment of both adults and children (Nolin 1999) with IgAN to find optimal regimens. These analyses included studies of variable methodological quality, mostly case series and other forms of non-randomised evaluations. These data have resulted in conflicting information regarding the optimal therapy. The most commonly used regimens include immunosuppressive agents, however the risks of long-term treatment often outweigh the potential benefits (Samuels 2003a). How the intervention might work Non-immunosuppressive therapeutic options, including fish oils, anticoagulants, antihypertensive agents and tonsillectomy have also been examined. The rationale for administration of fish oils, anticoagulants and antihypertensive agents is an effect on the pathogenesis of renal damage with possible decrease of both renal inflammation and glomerulosclerosis. The potential benefits of tonsillectomy are proposed to be due to elimination of a source for the defective IgA1 (which is manufactured in the tonsils as well as in the bone marrow) (Glassock 2009). This review aims to assess the benefits and harms of non-immunosuppressive treatments (e.g. fish oils, antihypertensive agents, anticoagulants, tonsillectomy and other non-immunosuppressive agents) for the treatment of IgAN in adults and children. M E T H O D S Criteria for considering studies for this review Types of studies Any randomised controlled trials (RCTs) and quasi-rcts evaluating the benefits and harms of fish oil, antihypertensive agents, anticoagulant agents, tonsillectomy and other non-immunosuppressive agents versus placebo/no treatment or any other non-immunosuppressive treatment for IgAN patients were included. The first period of randomised, cross-over studies, where separate data was available, were included. Types of participants Inclusion criteria Adult and children with biopsy-proven IgAN (presence of IgA mesangial deposits evident as mesangial expansion by standard histopathology examination and IgA staining by immunofluorescence). Studies where chronic glomerulonephropathies were grouped together, but separate data for IgAN was not given, were included if > 75% of included participants had IgAN. Exclusion criteria Studies of non-immunosuppressive agents used in people with chronic nephropathies but not IgAN were excluded. Why it is important to do this review Given the burden of disease and the known risks of progression, a systematic review of the benefits and harms of fish oil treatments, antihypertensive agents and other non-immunosuppressive treatments for patients with IgAN will aid healthcare providers in their management of this condition. O B J E C T I V E S Types of interventions Fish oil supplements (any dose, type or duration) versus placebo/no treatment. Fish oil supplements (any dose, type or duration) versus any other non-immunosuppressive treatment (including head to head comparisons). Antihypertensive agents (any dose, type or duration) versus placebo/no treatment. Antihypertensive agents (any dose, type or duration) versus any other non-immunosuppressive treatment (including head to head comparisons). Antiplatelet agents/anticoagulants (any dose, type or duration), versus placebo/no treatment. 3

Antiplatelet agents/anticoagulants (any dose, type or duration) versus any other non-immunosuppressive treatment (including head to head comparisons). Tonsillectomy versus placebo/no treatment. Tonsillectomy versus any other non-immunosuppressive treatment (including head to head comparisons). Other non-immunosuppressive treatment for IgAN (e.g. statins, phenytoin, herbal medicine, urokinase, vitamin E, sodium cromoglycate) versus placebo/no treatment. Other non-immunosuppressive treatment for IgAN (e.g. statins, phenytoin, herbal medicine, urokinase, vitamin E, sodium cromoglycate) versus any other non-immunosuppressive treatment (including head to head comparisons). Types of outcome measures 1. Progression or improvement in kidney disease. Where possible, time to reach each end point was included in the analysis. End points include attainment of: ESKD requiring renal replacement therapy (RRT) Doubling of serum creatinine (SCr) concentration > 50% increase in SCr > 50% decrease in creatinine clearance (CrCl) Remission of haematuria Remission of proteinuria Absolute change in the SCr (mg/dl or µmol/l) Absolute change in CrCl (ml/min or ml/min/1.73 m²) Changes in daily proteinuria (g/24 h) 2. All-cause mortality 3. Any adverse event 4. Infection rate 5. Changes in any measure of bone density 6. Fractures 7. Changes in any measure of quality of life (QoL) Search methods for identification of studies Electronic searches 1. The Cochrane Renal Group s specialised register using keywords relevant to this review. The register is populated using the following strategies: i) Screening of weekly Ovid MEDLINE AutoAlerts using a search strategy covering the scope of the Group (see the Renal Group s module under Specialised Register at http:// www.mrw.interscience.wiley.com/cochrane/clabout/articles/ RENAL/frame.html) plus the optimally sensitive strategy developed for the Cochrane Collaboration for the identification of RCTs (Lefebvre 2008). ii) Quarterly searches of NEW records in the Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library using a search strategy covering the scope of the Group. iii) Records handsearched by the Renal Group. The Renal Group has undertaken extensive handsearching of renal-related journals and conference proceedings, therefore we did not specifically search conference proceedings for this review. For a full list of journals and conference proceedings handsearched by the Renal Group use the link to its module (Renal Group 2010). 2. MEDLINE (from 1966) using the optimally sensitive strategy developed for The Cochrane Collaboration to identify RCTs (Lefebvre 2008) with a specific search strategy developed with input from the Trial Search Coordinator. 3. EMBASE (from 1980) using a search strategy adapted from that developed for The Cochrane Collaboration to identify RCTs (Lefebvre 2008) together with a specific search strategy developed with input from the Trial Search Coordinator. See Appendix 1 for search terms used. Searching other resources Additional studies were located through reference lists of all included studies, review articles and through contact with local and international experts in the field. Information about unpublished or incomplete studies was asked by email correspondence of investigators known to be involved in previous studies. Data collection and analysis Selection of studies The search strategy described was used to obtain titles and abstracts of studies that appeared relevant to the review. The titles and abstracts were screened independently by the two authors, who discarded studies that were not applicable based on our inclusion criteria. Two authors independently assessed retrieved abstracts and, if necessary the full text, of these studies to determine which studies satisfied the inclusion criteria. Disagreements were resolved in consultation with other authors. Data extraction and management Data extraction was carried out by the same authors independently using standard data extraction forms. Studies reported in non- English language journals were translated, where possible, before assessment. Where more than one publication of one study existed, reports were grouped together and the publication with the most complete data was used in the analyses. Disagreements were resolved in consultation with other authors. 4

Assessment of risk of bias in included studies The quality of studies to be included was assessed independently by two authors without blinding to authorship or journal using the checklist developed for the Cochrane Renal Group (see Appendix 2). Discrepancies were resolved by discussion with other authors. The items assessed were; allocation concealment, blinding of investigators, participants, outcome assessors and data analysts, intention-to-treat analysis, and completeness to follow-up. Measures of treatment effect For dichotomous outcomes (e.g. ESKD, > 50% increase in SCr, > 50% decrease in CrCl, remission of proteinuria/haematuria, allcause mortality) results were expressed as risk ratios (RR) with 95% confidence intervals (CI) for individual studies. Where continuous scales of measurement were used to assess the effects of treatment (e.g. SCr, CrCl, daily proteinuria) the mean difference (MD) was used, or the standardised mean difference (SMD) if different scales were used. Data synthesis Data were pooled using a random effects model but the fixed effect model was also analysed to ensure robustness of the model chosen and susceptibility to outliers. Subgroup analysis and investigation of heterogeneity Subgroup analyses were performed wherever possible to explore how potential sources of heterogeneity (related to characteristics of the population: children versus adults, populations with different degrees of renal impairment, populations with different stage of IgAN based on renal biopsy; interventions: different types and doses of fish oil, different types and doses of antihypertensive agent, different types and doses of anticoagulant, different techniques of tonsillectomy) may influence treatment effect. R E S U L T S Unit of analysis issues For dichotomous outcomes, to avoid double counting where studies had multiple intervention groups, both the number of events and the total number of participants in the control group were divided up. For continuous outcomes, only the total numbers of participants in the control group were divided up (Higgins 2008). Dealing with missing data Any further information required from the original author was requested by written correspondence (email) and any relevant information obtained in this manner was included in the review. Assessment of heterogeneity Heterogeneity was analysed using a Chi² test on N-1 degrees of freedom, with a P of 0.05 used for statistical significance and the I² statistic (Higgins 2003). Description of studies See: Characteristics of included studies; Characteristics of excluded studies; Characteristics of ongoing studies. Results of the search The combined search of MEDLINE, EMBASE, CENTRAL, the Cochrane Renal Group specialised registry, handsearching of reference lists and contact with experts, identified a total of 2872 potentially relevant articles. Duplicate publications (433) were excluded. Of the remaining 2439 reports, 2305 were excluded after title and abstract review. This left 134 reports of potentially eligible studies that underwent full text review. Of these 22 were excluded (not RCTs, not relevant population or intervention) leaving 112 reports of studies eligible for inclusion. Of these there were 56 studies, with 56 secondary publications, enrolling a total of 2838 participants (Figure 1). 5

Figure 1. Flow chart showing number of reports of studies identified by searches, reports excluded and number of studies included 6

Included studies The 56 studies were grouped according to the category of intervention and the comparisons made (Appendix 3 Key characteristics of included studies). Fish oil (7 studies, 402 participants) Fish oil versus placebo/no treatment (4 studies, 265 participants: Bennett 1989; Donadio 1994; NA IgAN Study 2006; Pettersson 1994) Fish oil versus symptomatic treatment (1 study, 34 participants: Alexopoulos 2004) In Alexopoulos 2004 the symptomatic treatment included angiotensin converting enzyme inhibitors (ACEi) ± beta-blockers (BB) or calcium channel blockers (CCB) to treat hypertension and diuretics if there was evidence of severe oedema or congestive heart failure. High versus low dose fish oil(1 study, 73 participants: Donadio 2001) Fish oil + other treatment versus other treatment (1 study, 30 participants: Costanzi 2006) Other treatment was ACEi + angiotensin receptor blockers (ARB) Antihypertensive agents (30 studies, 1587 participants) Antihypertensive agents used were ACEi, ARB, BB, and CCB. The effect of the antihypertensive agents was examined as a group and where possible antihypertensive type/class effect was examined in subgroup analysis. Antihypertensive agents versus placebo/no treatment (4 studies, 251 participants: HKVIN Study 2006; IgACE Study 2001; Maschio 1994*; Nakamura 2000) Antihypertensive agents versus other treatment (22 studies, 1060 participants: Bannister 1995; Chen 2006; Cheng 1998; Hashizume 1998; Horita 2007; JLIGHT Study 2003; Kanno 2005; Liu 2007*; Nakamura 2000; Nakamura 2007; Nakao 2001*; Panichi 2000*; Park 2003; Perico 1998; Praga 2003; Roccatello 2000*; Russo 2001*; Shi 2002; Shimizu 2008; Tesar 1999*; Woo 1990; Woo 2008) The other treatment was either another antihypertensive (generally comparisons were made with ACEi versus other antihypertensive such as CCB or BB), symptomatic treatment (non ACEi antihypertensives such as CCB, BB or alpha blockers (AB)), herbal medicine, or an anticoagulant/ antiplatelet. Where possible antihypertensive type effect was examined in subgroup analyses. Antihypertensive agents + other treatment versus other treatment (8 studies, 408 participants: Bang 1994; Chen 2007; Cheng 1998; Guan 2005; Horita 2004; Horita 2007; Nakamura 2007; Song 2003*) Other treatments were used in both comparison arms and included another antihypertensive, anticoagulants ± steroids and herbal medicine. Horita 2004 and Nakamura 2007 compared ACEi + ARB versus ACEi or ARB Where possible antihypertensive type/class effect was examined in subgroup analysis. Antiplatelets and anticoagulants (5 studies, 152 participants) Antiplatelet and anticoagulant agents used were warfarin, dipyridamole, aspirin, defibrotide and ticoplidine. Antiplatelet/anticoagulant versus placebo/no treatment (2 studies, 33 participants: Camara 1991*; Lee 1997) Camara 1991 compared dipyridamole to placebo Lee 1997 compared dipyridamole + warfarin to no treatment Antiplatelet/anticoagulant versus other treatment (1 study, 52 participants: Chan 1987) Chan 1987 compared aspirin + dipyridamole to vitamin B complex. Antiplatelet/anticoagulant + other treatment versus other treatment (2 studies, 67 participants: Cheng 1998; Frasca 1997) Other treatments were used in both comparator arms. Cheng 1998 compared ticlopidine + ACEi to ACEi or BB. Frasca 1997 compared defibrotide + steroids to steroids. Tonsillectomy (2 studies, 79 participants) Tonsillectomy + other treatment versus other treatment (Hotta 1993; Kawasaki 2006). Hotta 1993 compared tonsillectomy + steroid pulse initially + cyclophosphamide + dipyridamole + warfarin during the steroid therapy versus steroid pulse initially + cyclophosphamide + dipyridamole + warfarin during the steroid therapy. Kawasaki 2006 compared tonsillectomy + methylprednisolone pulse + PWD (prednisolone, warfarin, dipyridamole) versus PWD + mizoribine. Statins (2 studies, 51 participants) 7

Statins (fluvastatin) versus placebo/no treatment (1 study, 21 participants: Buemi 2000*) Statins (fluvastatin) + other treatment versus other treatment (1 study, 30 participants: Kano 2003) Kano 2003 used dipyridamole both comparator arms. Li 2006 compared Astragalus IV injection + dipyridamole + benazepril to dipyridamole + benazepril. Wu 2003 compared Dan Shao Tang + western medicine (not specified) to western medicine. Phenytoin (3 studies, 100 participants) Phenytoin versus placebo/no treatment (Adamkova 1989*; Clarkson 1980; Houssin 1984) Herbal medicines (5 studies, 358 participants) Herbal medicine versus placebo/no treatment (2 studies, 131 participants: Qian 1987; Yoshikawa 1997) Qian 1987 compared Tripterygium Wilfordii Glycoside (TWG) to no treatment Yoshikawa 1997 compared Sairei-to with no treatment. Herbal medicine versus other treatment (1 study, 70 participants: Bo 2000) Bo 2000 compared Shenning Mixture to steroid + vitamin E + another drug (not translated). Herbal medicine + other treatment versus other treatment (2 studies, 157 participants: Li 2006; Wu 2003*) Others (3 studies, 156 participants) Urokinase + ACEi versus ACEi (1 study, 71 participants: Chen 2004) Vitamin E versus placebo/no treatment (1 study, 55 participants: Chan 2003) Sodium cromoglycate versus placebo/no treatment (1 study, 30 participants: Sato 1990) Some studies included more than 2 arms (e.g. ACEi versus CCB versus placebo) they were therefore considered in more than one comparator group. Also in some studies (*) the data could not be included in the meta-analysis for various reasons, which are described in Appendix 4 Studies not included in meta-analyses. Risk of bias in included studies The quality of most studies included in this review was suboptimal (see the Risk of bias tables under Characteristics of included studies and Appendix 4: Figure 2; Figure 3). Figure 2. Risk of bias graph: review authors judgements about each risk of bias item presented as percentages across all included studies. 8

Figure 3. Risk of bias summary: review authors judgements about each risk of bias item for each included study. 9

Allocation concealment Allocation concealment was unclear in most of the studies with only 3/56 (5%) of the studies (HKVIN Study 2006; IgACE Study 2001; Praga 2003) having adequate allocation concealment. One study concerning the effect of tonsillectomy (Hotta 1993) was assumed to be a quasi-rct (the authors stated that the participants were divided into two groups but did not clearly specify randomisation). Given the paucity of studies on tonsillectomy it was decided to include this study in the systematic review but the risk of bias must be considered in interpreting the results. Blinding The nature of blinding across participants, investigators, outcome assessors and data analysts was not clearly stated in 53/56 (94%) of the included studies. In only one study (2%) participants, investigators, outcome assessors and data analysts were blinded (NA IgAN Study 2006). In two studies (4%) participants and investigators were blinded (Donadio 1994; Maschio 1994). In one study (2%) participants and data analysts were blinded (Yoshikawa 1997). In five studies (10%) only participants were blinded (Buemi 2000; Camara 1991; HKVIN Study 2006; Nakamura 2007; Perico 1998). In five studies (9%) double blinding was stated but the authors did not clarify who and how (Chen 2006; Chen 2007; IgACE Study 2001; Nakao 2001; Song 2003). Intention-to-treat analysis Intention-to-treat analysis was reported and confirmed on review of the studies in 8/56 studies (14%) (Donadio 1994; Donadio 2001; HKVIN Study 2006; IgACE Study 2001; Lee 1997; NA IgAN Study 2006; Praga 2003; Song 2003). Completeness of follow-up The completeness of follow-up was not reported or assessable in 24/56 (43%) included studies. Eight studies (14%) reported total number of participants lost to follow-up but did not specify whether these participants were lost from the intervention or comparator arms. Overall incomplete follow-up ranged from 0% to 33% (Camara 1991; Chan 1987; Hashizume 1998; Horita 2004; Horita 2007; Maschio 1994; Song 2003). In the fish oil studies losses ranged from 0% to 33% with losses across arms slightly greater in the fish oil arm in two studies (Alexopoulos 2004; Bennett 1989). In the antihypertensive studies losses ranged from 0% to 33% with much greater losses in the nifedipine arm compared with an ACEi arm in one study (Bannister 1995). In the antiplatelet/anticoagulant studies losses ranged from 5% to 37%. In the tonsillectomy studies losses ranged from 0% to 6%. In the statins studies losses were nil in one study and not stated in the other. In the other non-immunosuppressive treatments studies losses were not stated. Effects of interventions The effects of interventions are presented in Data and analyses section. Some data not suitable for presentation in the forest plots are presented in other data tables. In the original protocol the outcomes: infection rate; decreased bone density or increased fractures or shorter stature; and QoL, were proposed but none of the included studies reported these outcomes. Fish oil Fish oil versus placebo/no treatment There was a significant reduction in the risk of a > 50% increase in SCr with fish oil compared with placebo/no treatment (Analysis 1.2 (1 study, 106 participants): RR 0.20, 95% CI 0.06 to 0.65). For the following outcomes there was no significant difference in the risk between fish oil and placebo/no treatment. > 50% decrease in CrCl (Analysis 1.3) all-cause mortality (Analysis 1.4) SCr (Analysis 1.5) change in SCr (Analysis 1.7) CrCl (Analysis 1.9) proteinuria (Analysis 1.11) change in proteinuria (Analysis 1.12) change in haematuria (Analysis 1.14) Fish oil versus symptomatic treatment There was a significant decrease in SCr with fish oil compared to symptomatic treatment (Analysis 2.4 (1 study, 28 participants): MD -318.24, 95% CI -525.58 to -110.90). For the following outcomes there was no significant difference in the risk between fish oil and symptomatic treatment. 10

> 50% increase in SCr (Analysis 2.2) > 50% decrease in CrCl (Analysis 2.3) CrCl (Analysis 2.6) proteinuria (Analysis 2.8) > 50% decrease in CrCl (Analysis 2.3) CrCl (Analysis 2.6) proteinuria (Analysis 2.8) High versus low dose fish oil There was no significant difference between high dose and low dose fish oil for ESKD (Analysis 3.1). Change in SCr and proteinuria could not be meta-analysed. Fish oil + other treatment versus other treatment When fish oil + ACEi + ARB was compared with ACEi + ARB: There was a significant increase in CrCl (Analysis 4.1 (1 study, 30 participants): MD 26.20, 95% CI 1.01 to 51.39) in the fish oil arm. There was a significant decrease in proteinuria (Analysis 4.2 (1 study, 30 participants): MD -0.99 g/24 h, 95% CI -1.70 to - 0.28) in the fish oil arm. Antihypertensive agents Antihypertensive agents versus placebo/no treatment The majority of antihypertensive agents used were either ACEi or ARB. ACEi versus placebo There was a significant increase in the remission of proteinuria with ACEi compared to placebo (Analysis 5.2 (1 study, 2 subgroups - complete and partial, 66 participants): RR 3.97, 95% CI 1.11 to 14.24; I² = 0%). Any antihypertensive agent versus placebo/no treatment Overall there was a significant decrease in proteinuria (Analysis 5.6 (3 studies, 207 participants): MD -0.69 g/24 h, 95% CI -1.00 to -0.37; I² = 0%). These antihypertensive agent studies were analysed in subgroups according to type of antihypertensive. When ACEi or ARB was compared with placebo/no treatment there was a significant decrease in proteinuria in the ACEi or ARB group (Analysis 5.6.1 (3 studies, 197 participants): MD -0.73 g/24 h, 95% CI -1.06 to -0.39; I² = 0%). When CCB was compared with placebo/no treatment there was no significant difference in proteinuria (Analysis 5.6.2 (1 study, 10 participants): MD -0.30 g/24 h, 95% CI -1.33 to 0.73) When CCB was compared with placebo/no treatment there was no significant difference in proteinuria (Analysis 5.6.2 (1 study, 10 participants): MD -0.30 g/24 h, 95% CI -1.33 to 0.73). No significant differences were detected for the following outcomes. ESKD (Analysis 5.1) SCr (Analysis 5.4) CrCl (Analysis 5.5) Antihypertensive agents versus other treatment Studies compared ACEi or ARB with other antihypertensive agents, and antihypertensive agents with other agents such as symptomatic treatment (which involved other antihypertensive agents from a different class to control blood pressure), anticoagulants or herbal (including Chinese medicine) medicine. ACEi versus ARB Woo 2008 had four intervention arms (ARB: high dose and normal dose ; and ACEi: normal dose and low dose ). When ACEi was compared with ARB there was a significant increase in the risk of ESKD (Analysis 6.1.4 (1 study, 226 participants): RR 1.84, 95% CI 1.06 to 3.18). There were no significant differences for the following outcomes. SCr (Analysis 6.4.1) CrCl (Analysis 6.5.1) Proteinuria (Analysis 6.7.1) Change in proteinuria (Analysis 6.8.1) Combined therapy (ACEi + ARB) versus monotherapy (ACEi or ARB) When combined therapy (ACEi + ARB) was compared to monotherapy (ACEi or ARB), there was a significant decrease in proteinuria with combination therapy (Analysis 6.7.2 (2 studies, 67 participants): MD -0.49 g/24 h, 95% CI -0.72, to -0.25; I² = 20%). There were no significant differences detected for these outcomes. SCr (Analysis 6.4.2) CrCl (Analysis 6.5.2) CrCl (Analysis 6.5.2) ACEi or ARB or ACEi + ARB versus symptomatic treatment There was a significantly lower risk of > 50% increase in SCr in the ACEi group compared to symptomatic treatment 11

(Analysis 6.2 (1 study, 44 participants): RR 0.23, 95% CI 0.07 to 0.70). There was a significant decrease in SCr in the ACEi or ARB or ACEi + ARB group compared to symptomatic treatment (Analysis 6.4.4 (3 studies, 168 participants): MD -39.37, 95 % CI -71.95 to -6.80; I² = 0%). There was a significant increase in CrCl in the ACEi or ARB or ACEi + ARB group compared to symptomatic treatment (Analysis 6.5.4 (2 studies, 127 participants): MD 23.26, 95% CI 10.40 to 36.12; I² = 20%). There was a significant decrease in proteinuria in the ACEi or ARB or ACEi + ARB group compared to symptomatic treatment (Analysis 6.7.5 (3 studies, 168 participants): MD - 1.16 g/24 h, 95% CI -1.52 to -0.81; I² = 0%) There was no significant difference in the change in proteinuria between the ACEi or ARB or ACEi + ARB group compared to symptomatic treatment (Analysis 6.8.2) ACEi or ARB versus CCB There was a significant decrease in proteinuria in the ACEi or ARB group compared to CCB (Analysis 6.7.4 (3 studies, 40 participants): MD -0.32, 95% CI -0.79 to 0.14) and a significant change in proteinuria in the ACEi or ARB group compared to CCB (Analysis 6.8.2 (1 study, 8 participants): MD 14.60, 95% CI 2.16 to 27.04). There were no significant differences for the following outcomes. ESKD (Analysis 6.1.1) Doubling of SCr (Analysis 6.3) SCr (Analysis 6.4.4) CrCl (Analysis 6.5.4) ACEi versus BB There was no significant difference in ESKD between the ACEi and BB groups (Analysis 6.1.3). ACEi + BB versus CCB + BB There was a significant decrease in proteinuria in the ACEi + BB group compared to the CCB + BB group (Analysis 6.7.5 (1 study, 49 participants): MD -0.54, 95% CI -0.83 to -0.25). There were no significant differences in the following outcomes. SCr (Analysis 6.4.5) CrCl (Analysis 6.5.5) SCr (Analysis 6.4.6) CrCl (Analysis 6.5.6) Proteinuria (Analysis 6.7.6) Haematuria (Analysis 6.10) ACEi versus herbal medicine There were no significant differences between ACEi and herbal medicine for the following outcomes. SCr (Analysis 6.4.7) CrCl (Analysis 6.5.7) Proteinuria (Analysis 6.7.7) Antihypertensive agents + other treatment versus other treatment Other treatments included anticoagulants, steroids and herbal or Chinese medicines. For the subgroup ARB + steroid + anticoagulant compared with steroid + anticoagulant there was significantly lower proteinuria (Analysis 7.5.1 (1 study, 38 participants): MD -0.20 g/24 h, 95% CI -0.26 to -0.14). When ARB + Huobahuagen (a Chinese medicine) was compared with Huobahuagen haematuria was significantly lower (Analysis 7.7 (1 study, 62 participants): MD -7.00 RBC/HPF, 95% CI -11.08 to -2.92). No other outcomes were significant. Antiplatelet and anticoagulants Anticoagulant agents used were warfarin and the antiplatelet agents used were aspirin, defibrotide, dipyridamole and ticlopidine. For the meta-analysis studies were grouped as anticoagulant versus placebo/no treatment, anticoagulant versus other and anticoagulant + other versus other. Anticoagulant versus placebo/no treatment One study (Lee 1997) considered dipyridamole + warfarin versus no treatment. No significant differences were found for the outcomes ESKD, SCr, CrCl, proteinuria and haematuria. Anticoagulant versus other treatment One study (Chan 1987) considered aspirin + dipyridamole versus vitamin B complex. No significant differences were found for the outcomes SCr, CrCl and proteinuria. ACEi or ARB versus anticoagulants There were no significant differences between ACEi or ARB and anticoagulants for the following outcomes. ESKD (Analysis 6.1.4) Anticoagulant + other treatment versus other treatment Two studies were included. Cheng 1998 compared anticoagulant + ACEi versus ACEi and Frasca 1997 compared anticoagulant + steroid versus steroid. 12

The only significant outcome was a decrease in proteinuria for anticoagulant + steroid compared to steroid (Analysis 10.5 (1 study, 20 participants): MD -0.50 g/24 h, 95% CI -0.89 to - 0.11). Other non-immunosuppressive treatments for IgAN Other treatments used for IgAN included tonsillectomy, statins, phenytoin, herbal medicines, urokinase, vitamin E and sodium cromoglycate. There was no significant difference in any of the outcomes for majority of the comparisons. The few exceptions are reported below. Tonsillectomy Two studies of tonsillectomy were included in this analysis. Hotta 1993 compared tonsillectomy + steroid pulse initially + cyclophosphamide + dipyridamole + warfarin during the steroid therapy versus as above without tonsillectomy Kawasaki 2006: tonsillectomy + methylprednisolone pulse + PWD (prednisolone, warfarin, dipyridamole) versus PWD + mizoribine. The outcomes assessed were proteinuria remission, haematuria remission, SCr, CrCl and proteinuria. The only significant finding was a significant increase in the risk of haematuria remission in the tonsillectomy arm compared with non-tonsillectomy arm (Analysis 11.2 (2 studies, 77 participants): RR 1.54, 95% CI 1.05 to 2.25; I² = 0%). This outcome had two subgroups. Microscopic haematuria was significant (Analysis 11.2.1 (1 study, 45 participants): RR 1.83, 95% CI 1.04 to 3.22) and macroscopic haematuria was not significant (Analysis 11.2.2 (1 study, 32 participants): RR 1.33, 95% CI 0.80 to 2.23). Statins There were two included studies. Buemi 2000 compared fluvastatin versus placebo Kano 2003 compared fluvastatin + dipyridamole versus dipyridamole Only Kano 2003 was suitable for inclusion in the meta-analysis. The outcomes assessed were SCr, CrCl, proteinuria and haematuria. The only significant outcome was an increase in CrCl in the statin group (Analysis 13.2 (1 study, 30 participants): MD 22.60, 95% CI 11.83 to 33.37). Herbal medicines The herbal medicines group included studies that looked at herbal versus placebo/no treatment, herbal versus other treatments (such as steroids, vitamins, and western medicine ), and herbal + other versus other (the other in this instance being anticoagulants and antihypertensives). Herbal medicine versus placebo/no treatment There were two included studies. Qian 1987 compared TWG versus no treatment Yoshikawa 1997 compared Sairei-to versus no treatment. The outcomes assessed for TWG versus no treatment were SCr and proteinuria. The only significant outcome was lower proteinuria with TWG compared with no treatment (Analysis 15.2.1 (1 study, 30 participants): MD -0.56 g/24 h, 95% CI -1.03 to -0.09). The outcomes assessed for Sairei-to versus no treatment were proteinuria and haematuria. The only significant outcome was a decrease in for haematuria for the Sairei-to group (Analysis 15.3 (1 study, 94 participants): MD -0.80, 95% CI -1.27 to -0.33). Herbal medicine versus other treatment One study (Bo 2000), comparing Shenning Mixture versus steroid + vitamin E + dipyridamole assessing the outcomes SCr, proteinuria and haematuria was included. Proteinuria was significantly lower for the Shenning Mixture group compared with the other treatment group (Analysis 16.2 (1 study, 70 participants): MD -0.90 g/24 h, 95% CI -1.23 to -0.57). Haematuria was significantly lower for the Shenning Mixture group compared with the other treatment group (Analysis 16.3 (1 study, 70 participants): MD -18.17, 95% CI - 22.39 to -13.95). SCr was not significant (Analysis 16.1). Herbal medicine + other treatment versus other treatment Two studies compared herbal medicine + other treatment versus other treatment (Li 2006; Wu 2003). The data from Wu 2003 could not be meta-analysed. There were no significant differences for SCr and proteinuria. Urokinase One study compared urokinase + ACEi versus ACEi (Chen 2004). There was a significant difference in the following outcomes > 50% decrease in proteinuria (Analysis 18.2 (1 study, 71 participants): RR 1.61, 95% CI 1.05 to 2.45) SCr (Analysis 18.3 (1 study, 71 participants): MD - 39.49, 95% CI -47.69 to -31.29) CrCl (Analysis 18.4 (1 study, 71 participants): MD 10.20, 95% CI 4.26 to 16.14) Proteinuria (Analysis 18.5 (1 study, 71 participants): MD -0.67 g/24 h, 95% CI -0.74 to -0.60). There was no significant difference in > 50% increase in SCr (Analysis 18.1) 13