Evaluating a New International Risk-Prediction Tool in IgA Nephropathy | Nephrology | JAMA Internal Medicine | JAMA Network
[Skip to Navigation]
Figure 1.  Enrollment Flowchart and Cumulative Incidence of the Primary Outcome in the Derivation and Validation Cohorts
Enrollment Flowchart and Cumulative Incidence of the Primary Outcome in the Derivation and Validation Cohorts

A, Derivation of the analytic cohorts used for the derivation and validation of the risk prediction models. B, The primary outcome was 50% decline in eGFR or ESRD. eGFR indicates estimated glomerular filtration rate; ESRD, end-stage renal disease.

Figure 2.  Calibration Curves Depicting the Predicted vs Observed 5-Year Risks of the Primary Outcome Using the Full Prediction Models With and Without Race/Ethnicity
Calibration Curves Depicting the Predicted vs Observed 5-Year Risks of the Primary Outcome Using the Full Prediction Models With and Without Race/Ethnicity

Results from the derivation cohort are on the left, and from the validation cohort on the right. Predicted 5-year risks are from the prediction model, and observed 5-year risks are from Kaplan-Meier estimates within deciles of predicted risk. The dotted line represents perfect calibration in which predicted and observed risks are identical.

Table 1.  Description of the Derivation and Validation Cohortsa
Description of the Derivation and Validation Cohortsa
Table 2.  Prediction Performance in the Derivation Cohort for the Clinical, Limited, and Full Modelsa
Prediction Performance in the Derivation Cohort for the Clinical, Limited, and Full Modelsa
Table 3.  Rate of Kidney Function Decline and the Mean Predicted 5-Year Risk of the Primary Outcome in Subgroups Based on the Linear Predictor
Rate of Kidney Function Decline and the Mean Predicted 5-Year Risk of the Primary Outcome in Subgroups Based on the Linear Predictor
1.
Donadio  JV, Grande  JP.  IgA nephropathy.   N Engl J Med. 2002;347(10):738-748. doi:10.1056/NEJMra020109PubMedGoogle ScholarCrossref
2.
Reich  HN, Troyanov  S, Scholey  JW, Cattran  DC; Toronto Glomerulonephritis Registry.  Remission of proteinuria improves prognosis in IgA nephropathy.   J Am Soc Nephrol. 2007;18(12):3177-3183. doi:10.1681/ASN.2007050526PubMedGoogle ScholarCrossref
3.
Cattran  DC, Feehally  J, Cook  H,  et al.  Kidney Disease: Improving Global Outcomes (KDIGO) Glomerulonephritis Work Group: KDIGO clinical practice guideline for glomerulonephritis.   Kidney Int Suppl. 2012;2(2):139-274. doi:10.1038/kisup.2012.9Google ScholarCrossref
4.
Barbour  SJ, Reich  HN.  Risk stratification of patients with IgA nephropathy.   Am J Kidney Dis. 2012;59(6):865-873. doi:10.1053/j.ajkd.2012.02.326PubMedGoogle ScholarCrossref
5.
Bartosik  LP, Lajoie  G, Sugar  L, Cattran  DC.  Predicting progression in IgA nephropathy.   Am J Kidney Dis. 2001;38(4):728-735. doi:10.1053/ajkd.2001.27689PubMedGoogle ScholarCrossref
6.
Pozzi  C, Bolasco  PG, Fogazzi  GB,  et al.  Corticosteroids in IgA nephropathy: a randomised controlled trial.   Lancet. 1999;353(9156):883-887. doi:10.1016/S0140-6736(98)03563-6PubMedGoogle ScholarCrossref
7.
Manno  C, Torres  DD, Rossini  M, Pesce  F, Schena  FP.  Randomized controlled clinical trial of corticosteroids plus ACE-inhibitors with long-term follow-up in proteinuric IgA nephropathy.   Nephrol Dial Transplant. 2009;24(12):3694-3701. doi:10.1093/ndt/gfp356PubMedGoogle ScholarCrossref
8.
Lv  J, Zhang  H, Wong  MG,  et al; TESTING Study Group.  Effect of oral methylprednisolone on clinical outcomes in patients with IgA nephropathy: the TESTING randomized clinical trial.   JAMA. 2017;318(5):432-442. doi:10.1001/jama.2017.9362PubMedGoogle ScholarCrossref
9.
Rauen  T, Eitner  F, Fitzner  C,  et al; STOP-IgAN Investigators.  Intensive supportive care plus immunosuppression in IgA nephropathy.   N Engl J Med. 2015;373(23):2225-2236. doi:10.1056/NEJMoa1415463PubMedGoogle ScholarCrossref
10.
Maes  BD, Oyen  R, Claes  K,  et al.  Mycophenolate mofetil in IgA nephropathy: results of a 3-year prospective placebo-controlled randomized study.   Kidney Int. 2004;65(5):1842-1849. doi:10.1111/j.1523-1755.2004.00588.xPubMedGoogle ScholarCrossref
11.
Pozzi  C, Andrulli  S, Pani  A,  et al.  Addition of azathioprine to corticosteroids does not benefit patients with IgA nephropathy.   J Am Soc Nephrol. 2010;21(10):1783-1790. doi:10.1681/ASN.2010010117PubMedGoogle ScholarCrossref
12.
Hogg  RJ, Lee  J, Nardelli  N,  et al; Southwest Pediatric Nephrology Study Group.  Clinical trial to evaluate omega-3 fatty acids and alternate day prednisone in patients with IgA nephropathy: report from the Southwest Pediatric Nephrology Study Group.   Clin J Am Soc Nephrol. 2006;1(3):467-474. doi:10.2215/CJN.01020905PubMedGoogle ScholarCrossref
13.
Li  PK, Leung  CB, Chow  KM,  et al; HKVIN Study Group.  Hong Kong study using valsartan in IgA nephropathy (HKVIN): a double-blind, randomized, placebo-controlled study.   Am J Kidney Dis. 2006;47(5):751-760. doi:10.1053/j.ajkd.2006.01.017PubMedGoogle ScholarCrossref
14.
Coppo  R, Peruzzi  L, Amore  A,  et al.  IgACE: a placebo-controlled, randomized trial of angiotensin-converting enzyme inhibitors in children and young people with IgA nephropathy and moderate proteinuria.   J Am Soc Nephrol. 2007;18(6):1880-1888. doi:10.1681/ASN.2006040347PubMedGoogle ScholarCrossref
15.
Cattran  DC, Coppo  R, Cook  HT,  et al; Working Group of the International IgA Nephropathy Network and the Renal Pathology Society.  The Oxford classification of IgA nephropathy: rationale, clinicopathological correlations, and classification.   Kidney Int. 2009;76(5):534-545. doi:10.1038/ki.2009.243PubMedGoogle ScholarCrossref
16.
Roberts  IS, Cook  HT, Troyanov  S,  et al; Working Group of the International IgA Nephropathy Network and the Renal Pathology Society.  The Oxford classification of IgA nephropathy: pathology definitions, correlations, and reproducibility.   Kidney Int. 2009;76(5):546-556. doi:10.1038/ki.2009.168PubMedGoogle ScholarCrossref
17.
Wakai  K, Kawamura  T, Endoh  M,  et al.  A scoring system to predict renal outcome in IgA nephropathy: from a nationwide prospective study.   Nephrol Dial Transplant. 2006;21(10):2800-2808. doi:10.1093/ndt/gfl342PubMedGoogle ScholarCrossref
18.
Goto  M, Wakai  K, Kawamura  T, Ando  M, Endoh  M, Tomino  Y.  A scoring system to predict renal outcome in IgA nephropathy: a nationwide 10-year prospective cohort study.   Nephrol Dial Transplant. 2009;24(10):3068-3074. doi:10.1093/ndt/gfp273PubMedGoogle ScholarCrossref
19.
Goto  M, Kawamura  T, Wakai  K, Ando  M, Endoh  M, Tomino  Y.  Risk stratification for progression of IgA nephropathy using a decision tree induction algorithm.   Nephrol Dial Transplant. 2009;24(4):1242-1247. doi:10.1093/ndt/gfn610PubMedGoogle ScholarCrossref
20.
Berthoux  F, Mohey  H, Laurent  B, Mariat  C, Afiani  A, Thibaudin  L.  Predicting the risk for dialysis or death in IgA nephropathy.   J Am Soc Nephrol. 2011;22(4):752-761. doi:10.1681/ASN.2010040355PubMedGoogle ScholarCrossref
21.
Bjørneklett  R, Vikse  BE, Bostad  L, Leivestad  T, Iversen  BM.  Long-term risk of ESRD in IgAN; validation of Japanese prognostic model in a Norwegian cohort.   Nephrol Dial Transplant. 2012;27(4):1485-1491. doi:10.1093/ndt/gfr446PubMedGoogle ScholarCrossref
22.
Pesce  F, Diciolla  M, Binetti  G,  et al.  Clinical decision support system for end-stage kidney disease risk estimation in IgA nephropathy patients.   Nephrol Dial Transplant. 2016;31(1):80-86. doi:10.1093/ndt/gfv232PubMedGoogle ScholarCrossref
23.
Xie  J, Kiryluk  K, Wang  W,  et al.  Predicting progression of IgA nephropathy: new clinical progression risk score.   PLoS One. 2012;7(6):e38904. doi:10.1371/journal.pone.0038904PubMedGoogle Scholar
24.
Barbour  SJ, Cattran  DC, Kim  SJ,  et al.  Individuals of Pacific Asian origin with IgA nephropathy have an increased risk of progression to end-stage renal disease.   Kidney Int. 2013;84(5):1017-1024. doi:10.1038/ki.2013.210PubMedGoogle ScholarCrossref
25.
Barbour  SJ, Espino-Hernandez  G, Reich  HN,  et al; Oxford Derivation, North American Validation and VALIGA Consortia; Oxford Derivation North American Validation and VALIGA Consortia.  The MEST score provides earlier risk prediction in lgA nephropathy.   Kidney Int. 2016;89(1):167-175. doi:10.1038/ki.2015.322PubMedGoogle ScholarCrossref
26.
Tanaka  S, Ninomiya  T, Katafuchi  R,  et al.  Development and validation of a prediction rule using the Oxford classification in IgA nephropathy.   Clin J Am Soc Nephrol. 2013;8(12):2082-2090. doi:10.2215/CJN.03480413PubMedGoogle ScholarCrossref
27.
Coppo  R, Troyanov  S, Bellur  S,  et al; VALIGA study of the ERA-EDTA Immunonephrology Working Group.  Validation of the Oxford classification of IgA nephropathy in cohorts with different presentations and treatments.   Kidney Int. 2014;86(4):828-836. doi:10.1038/ki.2014.63PubMedGoogle ScholarCrossref
28.
Zeng  CH, Le  W, Ni  Z,  et al.  A multicenter application and evaluation of the oxford classification of IgA nephropathy in adult Chinese patients.   Am J Kidney Dis. 2012;60(5):812-820. doi:10.1053/j.ajkd.2012.06.011PubMedGoogle ScholarCrossref
29.
Shi  SF, Wang  SX, Jiang  L,  et al.  Pathologic predictors of renal outcome and therapeutic efficacy in IgA nephropathy: validation of the oxford classification.   Clin J Am Soc Nephrol. 2011;6(9):2175-2184. doi:10.2215/CJN.11521210PubMedGoogle ScholarCrossref
30.
Herzenberg  AM, Fogo  AB, Reich  HN,  et al.  Validation of the Oxford classification of IgA nephropathy.   Kidney Int. 2011;80(3):310-317. doi:10.1038/ki.2011.126PubMedGoogle ScholarCrossref
31.
Katafuchi  R, Ninomiya  T, Nagata  M, Mitsuiki  K, Hirakata  H.  Validation study of oxford classification of IgA nephropathy: the significance of extracapillary proliferation.   Clin J Am Soc Nephrol. 2011;6(12):2806-2813. doi:10.2215/CJN.02890311PubMedGoogle ScholarCrossref
32.
Levey  AS, Stevens  LA, Schmid  CH,  et al; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration).  A new equation to estimate glomerular filtration rate.   Ann Intern Med. 2009;150(9):604-612. doi:10.7326/0003-4819-150-9-200905050-00006PubMedGoogle ScholarCrossref
33.
Trimarchi  H, Barratt  J, Cattran  DC,  et al; IgAN Classification Working Group of the International IgA Nephropathy Network and the Renal Pathology Society; Conference Participants.  Oxford classification of IgA nephropathy 2016: an update from the IgA Nephropathy Classification Working Group.   Kidney Int. 2017;91(5):1014-1021. doi:10.1016/j.kint.2017.02.003PubMedGoogle ScholarCrossref
34.
Inker  LA, Lambers Heerspink  HJ, Mondal  H,  et al.  GFR decline as an alternative end point to kidney failure in clinical trials: a meta-analysis of treatment effects from 37 randomized trials.   Am J Kidney Dis. 2014;64(6):848-859. doi:10.1053/j.ajkd.2014.08.017PubMedGoogle ScholarCrossref
35.
Levey  AS, Inker  LA, Matsushita  K,  et al.  GFR decline as an end point for clinical trials in CKD: a scientific workshop sponsored by the National Kidney Foundation and the US Food and Drug Administration.   Am J Kidney Dis. 2014;64(6):821-835. doi:10.1053/j.ajkd.2014.07.030PubMedGoogle ScholarCrossref
36.
Berthoux  F, Mariat  C, Maillard  N.  Overweight/obesity revisited as a predictive risk factor in primary IgA nephropathy.   Nephrol Dial Transplant. 2013;28(suppl 4):iv160-iv166. doi:10.1093/ndt/gft286PubMedGoogle ScholarCrossref
37.
Cattran  DC, Reich  HN, Beanlands  HJ, Miller  JA, Scholey  JW, Troyanov  S; Genes, Gender and Glomerulonephritis Group.  The impact of sex in primary glomerulonephritis.   Nephrol Dial Transplant. 2008;23(7):2247-2253. doi:10.1093/ndt/gfm919PubMedGoogle ScholarCrossref
38.
Royston  P, Altman  DG.  External validation of a Cox prognostic model: principles and methods.   BMC Med Res Methodol. 2013;13:33. doi:10.1186/1471-2288-13-33PubMedGoogle ScholarCrossref
39.
Chambless  LE, Cummiskey  CP, Cui  G.  Several methods to assess improvement in risk prediction models: extension to survival analysis.   Stat Med. 2011;30(1):22-38. doi:10.1002/sim.4026PubMedGoogle ScholarCrossref
40.
Moons  KG, Altman  DG, Reitsma  JB,  et al.  Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD): explanation and elaboration.   Ann Intern Med. 2015;162(1):W1-73. doi:10.7326/M14-0698PubMedGoogle ScholarCrossref
41.
Knoop  T, Vågane  AM, Vikse  BE,  et al.  Addition of eGFR and age improves the prognostic absolute renal risk-model in 1,134 Norwegian patients with IgA nephropathy.   Am J Nephrol. 2015;41(3):210-219. doi:10.1159/000381403PubMedGoogle ScholarCrossref
42.
Barratt  J, Feehally  J.  IgA nephropathy.   J Am Soc Nephrol. 2005;16(7):2088-2097. doi:10.1681/ASN.2005020134PubMedGoogle ScholarCrossref
43.
Debray  TP, Vergouwe  Y, Koffijberg  H, Nieboer  D, Steyerberg  EW, Moons  KG.  A new framework to enhance the interpretation of external validation studies of clinical prediction models.   J Clin Epidemiol. 2015;68(3):279-289. doi:10.1016/j.jclinepi.2014.06.018PubMedGoogle ScholarCrossref
44.
Haas  M, Verhave  JC, Liu  ZH,  et al.  A multicenter study of the predictive value of crescents in IgA nephropathy.   J Am Soc Nephrol. 2017;28(2):691-701. doi:10.1681/ASN.2016040433PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Original Investigation
    April 13, 2019

    Evaluating a New International Risk-Prediction Tool in IgA Nephropathy

    Author Affiliations
    • 1Division of Nephrology, University of British Columbia, Vancouver, British Columbia, Canada
    • 2BC Renal, Vancouver, British Columbia, Canada
    • 3Regina Margherita Children’s University Hospital, Torino, Italy
    • 4Peking University Institute of Nephrology, Beijing, China
    • 5Nanjing University School of Medicine, Nanjing, China
    • 6Faculty of Medicine, Juntendo University, Tokyo, Japan
    • 7National Fukuoka Higashi Medical Center, Fukuoka, Japan
    • 8Division of Nephrology, University of Toronto, Toronto, Ontario, Canada
    • 9The John Walls Renal Unit, Leicester General Hospital, Leicester, England
    JAMA Intern Med. 2019;179(7):942-952. doi:10.1001/jamainternmed.2019.0600
    Key Points

    Question  How can we better predict, at the time of kidney biopsy, the risk of a 50% decline in kidney function or end-stage renal disease in patients with IgA nephropathy?

    Findings  Large international multiethnic cohorts including 3927 patients were enrolled to both derive and externally validate 2 prediction models, one that included patient race/ethnicity, and one that did not. Both models outperformed clinical measures for prediction of kidney disease progression and patient risk stratification.

    Meaning  The 2 prediction models were shown to be accurate and validated methods to help clinicians improve management and treatment of IgA nephropathy in multi-ethnic cohorts and may aid international researchers in trial recruitment.

    Abstract

    Importance  Although IgA nephropathy (IgAN) is the most common glomerulonephritis in the world, there is no validated tool to predict disease progression. This limits patient-specific risk stratification and treatment decisions, clinical trial recruitment, and biomarker validation.

    Objective  To derive and externally validate a prediction model for disease progression in IgAN that can be applied at the time of kidney biopsy in multiple ethnic groups worldwide.

    Design, Setting, and Participants  We derived and externally validated a prediction model using clinical and histologic risk factors that are readily available in clinical practice. Large, multi-ethnic cohorts of adults with biopsy-proven IgAN were included from Europe, North America, China, and Japan.

    Main Outcomes and Measures  Cox proportional hazards models were used to analyze the risk of a 50% decline in estimated glomerular filtration rate (eGFR) or end-stage kidney disease, and were evaluated using the R2D measure, Akaike information criterion (AIC), C statistic, continuous net reclassification improvement (NRI), integrated discrimination improvement (IDI), and calibration plots.

    Results  The study included 3927 patients; mean age, 35.4 (interquartile range, 28.0-45.4) years; and 2173 (55.3%) were men. The following prediction models were created in a derivation cohort of 2781 patients: a clinical model that included eGFR, blood pressure, and proteinuria at biopsy; and 2 full models that also contained the MEST histologic score, age, medication use, and either racial/ethnic characteristics (white, Japanese, or Chinese) or no racial/ethnic characteristics, to allow application in other ethnic groups. Compared with the clinical model, the full models with and without race/ethnicity had better R2D (26.3% and 25.3%, respectively, vs 20.3%) and AIC (6338 and 6379, respectively, vs 6485), significant increases in C statistic from 0.78 to 0.82 and 0.81, respectively (ΔC, 0.04; 95% CI, 0.03-0.04 and ΔC, 0.03; 95% CI, 0.02-0.03, respectively), and significant improvement in reclassification as assessed by the NRI (0.18; 95% CI, 0.07-0.29 and 0.51; 95% CI, 0.39-0.62, respectively) and IDI (0.07; 95% CI, 0.06-0.08 and 0.06; 95% CI, 0.05-0.06, respectively). External validation was performed in a cohort of 1146 patients. For both full models, the C statistics (0.82; 95% CI, 0.81-0.83 with race/ethnicity; 0.81; 95% CI, 0.80-0.82 without race/ethnicity) and R2D (both 35.3%) were similar or better than in the validation cohort, with excellent calibration.

    Conclusions and Relevance  In this study, the 2 full prediction models were shown to be accurate and validated methods for predicting disease progression and patient risk stratification in IgAN in multi-ethnic cohorts, with additional applications to clinical trial design and biomarker research.

    Introduction

    The most common type of glomerulonephritis, IgA nephropathy (IgAN) has a worldwide incidence exceeding 1.5 per 100 000 persons/y and is a frequent causes of end-stage renal disease (ESRD) in Asian countries.1 A significant challenge in IgAN is the extremely heterogeneous risk of progressive kidney function decline, with a 10-year risk of ESRD varying between 5% and 60%.2 Although guidelines recommend risk stratifying patients with IgAN so that immunosuppressive treatment can be targeted to high-risk patients, there is currently no tool available to accurately predict kidney disease progression.3 Instead, risk stratification and treatment decisions rely on broad categories of clinical risk factors which can be highly inaccurate. A significant proportion of patients who qualify for immunosuppression therapy have nonprogressive disease, while many patients who do not qualify for treatment nonetheless experience kidney function decline.4-8 Several clinical trials in IgAN have failed efficacy end points partly owing to inadvertently recruiting low-risk patients, and future trials will continue to be limited by unreliable methods of risk stratification.9-14 This indicates a clear need for an accurate tool to predict disease progression in IgAN.

    Although there are well-established clinical and histologic risk factors for kidney function decline, when used individually they are unable to accurately identify high-risk patients.4,5,15,16 Previous efforts to integrate multiple risk factors into a prediction model have been limited by predictor variables that are not clinically meaningful; the use of histological scoring systems that have not been validated and are not routinely available in clinical practice; and lack of external validation, especially in different ethnic groups.5,17-23 This last limitation is particularly relevant given the highly variable incidence and severity of disease related to ethnicity.24 Although the Oxford MEST histologic score (mesangial [M] and endocapillary [E] hypercellularity, segmental sclerosis [S] and interstitial fibrosis/tubular atrophy [T]) in IgAN has been internationally validated, is widely available, and is independently associated with the risk of kidney progression, to our knowledge it has not been included in a risk prediction model validated in multiple ethnic groups.24-26 Owing to these substantial limitations, none of the existing prediction models in IgAN are sufficiently robust for use in clinical practice.

    We therefore used a large international collaboration of data sets with diverse ethnic representation to derive and externally validate a comprehensive risk prediction model in IgAN using readily available clinical and laboratory risk factors and the MEST histologic score.

    Methods
    Study Population

    The study population comprised cohorts that were collected independently for research purposes from Europe (VALIGA, n = 1406); Europe, Asia, and North and South America (Oxford derivation, n = 265); North America (Oxford validation, n = 187); China (Beijing, n = 410; Nanjing, n = 1026); and Japan (Fukuoka, n = 702) (Figure 1A). Details have been previously described.15,27-31 An additional Japanese cohort (Tokyo, n = 635) was collected from Juntendo University in Tokyo. All of these cohorts included only patients with biopsy-proven idiopathic IgAN, available MEST scores, and long-term follow-up after biopsy. Further details are provided in eMethods in the Supplement. The derivation cohort comprised the VALIGA, Nanjing, and Tokyo cohorts, and the validation cohort comprised the Beijing, Fukuoka, and both Oxford cohorts. We included patients in our analysis 18 years or older who did not have ESRD at the time of biopsy and who had available estimated glomerular filtration rate (eGFR) data. This project was approved by the University of British Columbia research ethics board, waiving patient written informed consent for de-identified data.

    Definitions

    Proteinuria, mean arterial blood pressure (MAP), eGFR (using the Chronic Kidney Disease Epidemiology Collaboration formula32), body mass index (BMI), age, and prior use of medications that block the renin-angiotensin system (RASBs, including angiotensin-converting enzyme inhibitors and angiotensin-receptor blockers) and the use of immunosuppression were determined at the time of biopsy. Race was self-reported by the patient as white, Chinese, Japanese, or other. Kidney biopsies were scored according to established criteria for the Oxford MEST scoring system, with the addition of crescents given recent evidence of their importance.15,16,33 The primary outcome was a composite of the first occurrence of either ESRD (eGFR <15 mL/min/1.73m2, dialysis, or transplantation) or a permanent reduction in eGFR to below 50% of the value at biopsy, which is an established kidney-related surrogate end point.34,35

    Statistical Analysis

    Time from kidney biopsy to the primary outcome (censored at death or end of follow-up) was analyzed using Cox proportional hazards models. In the derivation analysis, the clinical model contained only data on eGFR, proteinuria, and MAP at biopsy because these are the best recognized clinical predictors of outcome.4 The limited model included the following core predictor variables based on the existing literature: eGFR, MAP and proteinuria at biopsy, and the MEST score.4,25 The full model included the same core predictor variables, but also considered age, sex, race/ethnicity, crescents, BMI, RASB and immunosuppression at biopsy, and interaction terms. The additional predictors were determined based on the existing literature and selected for retention in the model using a backward elimination procedure with P = .20.4,24,25,31,36,37 A second full model was created in the same manner but without race/ethnicity as a potential predictor variable because the categories for race/ethnicity may not generalize to other populations. Prediction model performance was assessed using measures of model fit (R2D,38 Akaike Information Criterion [AIC]), discrimination (C statistic adapted for censoring39), reclassification (continuous net reclassification improvement [NRI] and integrated discrimination improvement [IDI] adapted for censoring39), and calibration plots.

    Results are presented according to the TRIPOD guidelines for risk-prediction models (see eTable 1 in the Supplement).40 Two-tailed P < .05 findings were considered statistically significant, except where otherwise indicated. Additional details regarding the study methods and statistical analyses are provided in eMethods in the Supplement.

    Results
    Derivation Analysis

    There were 3067 patients in the combined VALIGA, Nanjing, and Tokyo cohorts, of whom 2781 satisfied the inclusion criteria and formed the derivation cohort (Figure 1A). Characteristics of the derivation cohort are detailed in Table 1. The 5-year risk of the primary outcome (50% reduction in eGFR or ESRD) was 14.7% (95% CI, 13.1%-16.3%) (Figure 1B).

    Results of the clinical, limited, and full prediction models are detailed in eTable 2 in the Supplement. The clinical model contains data on eGFR, MAP, and proteinuria at biopsy. The limited model additionally contains the MEST score. The full model with race/ethnicity additionally contains the MEST score, age, race/ethnicity, RASB and immunosuppression at biopsy, and interactions between proteinuria and each of MAP and the T-score component of MEST. The full model without race/ethnicity contains the same predictors but with an interaction between RASB and proteinuria instead of race/ethnicity (which was selected for retention in the model only when race/ethnicity was removed). Race was categorized as Chinese, Japanese, white, or other. The distribution of predicted 5-year risk of the primary outcome is shown in eFigure 1 in the Supplement.

    The prediction performance details and all supporting data for the clinical, limited, and full models are reported in Table 2. Compared with the clinical model, the limited and full models all had better model fit, as demonstrated by higher R2D and lower AIC, better discrimination with significant increases in C statistics (ΔC), and significant improvement in reclassification with NRI and IDI 95% CIs above 0. Compared with the limited model, the full model with race/ethnicity had better model fit with higher R2D and lower AIC, better discrimination with a significant increase in C statistic to 0.82 (ΔC 0.02; 95% CI, 0.01-0.02), and significant improvement in reclassification as assessed by the IDI (0.03; 95% CI, 0.02- 0.04), but not the NRI (0.01; 95% CI, −0.08 to 0.16). The full model without race/ethnicity had a similar pattern, but with a significant NRI (0.19; 95% CI, 0.08-0.32). Both full models were well calibrated with very similar predicted and observed risks of the primary outcome 5 years after biopsy (Figure 2) and over the duration of follow-up (eFigure 2 in the Supplement). When the full models were compared with each other, there was no consistent trend in the C statistics, NRI, or IDI favoring one model over the other (data not shown). Based on these results, both full models were further assessed in the external validation analysis.

    Validation Analysis

    There were 1564 patients in the Oxford derivation, Oxford validation, Beijing, and Fukuoka cohorts, of whom 1146 satisfied the inclusion criteria and formed the validation cohort (Figure 1A). There were some differences, as expected, in patient characteristics compared with the derivation cohort (Table 1). Both the 5-year risk of the primary outcome (13.3%; 95% CI, 11.1%-15.5%) (Figure 1B) and the distribution of predicted 5-year risk (eFigure 1 in the Supplement) were similar to those in the derivation analysis.

    The R2D for the full models with and without race/ethnicity when applied to the validation cohort were both 35.3%, which were better than the R2D for the models applied to the derivation cohort (26.3% and 25.3%, respectively). The C statistics for both full models were 0.82 (95% CI, 0.81-0.83) and 0.81 (95% CI, 0.80-0.82), respectively, and were similar to the C statistics from the derivation analysis. The calibration slopes were 1.12 (95% CI, 0.98-1.25) and 1.19 (95% CI, 1.04-1.34) for the full models with and without race/ethnicity, respectively, indicating similar or better discrimination than was found in the derivation analysis.38 Both full models were well calibrated in the validation cohort, with good agreement between predicted and observed risk of the primary outcome at 5 years after biopsy (Figure 2) and over the duration of follow-up (eFigure 2 in the Supplement)

    As detailed in Table 3, for both full models, a higher predicted risk of the primary outcome was associated with a significantly faster rate of eGFR decline. The formulas for both full prediction models are listed in eTable 3 in the Supplement, and have been converted into mobile-app and web-based prediction tools available on Calculate by QxMD for iOS, Android and the web at https://qxmd.com/calculate-by-qxmd.

    The Role of Crescents

    Because crescents have been implicated as an important histologic risk factor, we explored reasons they were not selected in either full prediction model. Crescents were highly correlated with race/ethnicity (see eTable 4 in the Supplement) and with immunosuppression use after biopsy (56% vs 36% of those with and without crescents, P < .001). Even when crescents were added to the full model without race/ethnicity, they did not meet the P value threshold for selection in the prediction model (P < .20) without also including immunosuppression after biopsy, which was not a candidate predictor variable in our primary analysis (see eTable 5 in the Supplement).

    Discussion

    We derived and externally validated 2 risk-prediction models capable of accurately predicting a 50% decline in eGFR or ESRD in biopsy-proven IgAN using large international and ethnically diverse data sets of patients with a broad spectrum of disease and clinical predictor variables readily available at the time of biopsy, including the MEST score. We generated 2 prediction models, one that includes white, Chinese, or Japanese race/ethnicity, and one without race/ethnicity that can be used in other racial groups or if race/ethnicity is not available. An increase in predicted risk was associated with a more rapid rate of eGFR decline, demonstrating that the models robustly capture patients with more aggressive disease. Both prediction models have been converted into mobile-app and web-based calculators to facilitate clinical implementation.

    Our risk-prediction models are suitable for implementation worldwide and can potentially improve clinical treatment decisions and future research in IgAN. Risk stratification to determine immunosuppression treatment is currently based on simplistic categorization of clinical variables, which is highly inaccurate.3 Clinical trial data suggest that up to 75% of patients are unnecessarily treated because they have low-risk nonprogressive disease.6-8 Conversely, 33% of patients who do not meet clinical treatment criteria but have high-risk MEST scores eventually experience kidney function decline but are denied therapy.25 Our prediction models can now provide more accurate risk stratification early after a patient’s diagnosis, although further research is required to determine the optimal risk threshold for treatment that accounts for both the risks and benefits of immunosuppression. Previous clinical trials in IgAN, including the recent STOP-IgAN trial,9 have failed to achieve their primary end points partly due to inadvertently recruiting low-risk patients who did not experience kidney outcome events.9-14 The prediction models can overcome this limitation, improve study power, and facilitate future clinical trial design by allowing targeted recruitment of high-risk patients. The models can also be used in translational research to test the prediction benefit of adding biomarkers to the fully specified models detailed in eTable 3 in the Supplement. This will allow the use of smaller cohorts than would ordinarily be required for de novo model derivation, thus providing a critical tool for biomarker validation in the clinical domain.

    This study addresses the limitations of previous prediction models in IgAN related to (1) the use of smaller, single-ethnicity cohorts with comparatively few patients across the spectrum of disease severity, (2) the use of predictor variables requiring prolonged periods of follow-up thus limiting clinical utility, and (3) the use of histologic scoring systems that have not been validated.5,17-23,26,41 The large size of the present study cohorts from different international centers with few exclusion criteria ensured that the full spectrum of low- and high-risk disease was captured, with 36% to 44% of patients having 5-year predicted risks above 10%. This provides confidence that the prediction models can be applied to a diverse population of patients with IgAN. Because of known differences in the incidence and severity of IgAN across ethnic groups, we specifically assembled international cohorts to reflect this ethnic diversity and included race/ethnicity as a predictor variable.24,42 Since individual patients may not be adequately represented by the available race/ethnicity categories, we generated a second model without race/ethnicity but with similar overall prediction performance in multi-ethnic cohorts. We used the MEST histologic scoring system because it has been validated in multiple ethnic groups and is now a recommended component of kidney biopsy reports for IgAN.33 The other predictor variables are routinely used in clinical practice and are available at the time of kidney biopsy, making the present prediction models easy to implement.

    External validation using separate and autonomous data sets from those used to derive the prediction models is a strength of the present analysis and is missing from most previously developed prediction tools in IgAN.17-23,41 The derivation cohorts were specifically chosen because of their large size with sufficient outcome events, the availability of candidate predictors, and their multi-ethnic representation reflecting the diversity of IgAN. The validation cohort was from an older era, with less frequent use of RASB and immunosuppression, and different racial composition and frequency of histologic lesions. This strengthens the analysis because, in general, prediction models that perform well in a validation cohort that differs substantially from the derivation cohort provide greater generalizability to other populations with IgAN.43 Most of the patients (86.7%) in the derivation cohort were treated with RASB starting at or shortly after biopsy. The prediction models are thus likely best applied to similarly treated patients, which is consistent with KDIGO guideline recommendations for routine use of RASB in IgAN.3 The fully specified prediction models in eTable 3 in the Supplement can be used by other researchers for additional validation in more contemporary cohorts or different ethnic and age groups.

    The present results highlight the importance of different prognostic factors in IgAN. Despite varying statistical significance of the individual MEST components, the overall MEST score accounts for the improvement in prediction performance between the clinical and limited models. Crescents were not selected as a predictor variable for either full model likely because they were highly correlated with race/ethnicity, which was more strongly associated with the primary outcome, and because the association between crescents and the primary outcome was confounded by the subsequent use of immunosuppression consistent with the findings of several other studies.31,44 The addition of the MEST score, age, medication use at biopsy, and interaction terms with or without race/ethnicity account for the numerically small but statistically significant improvement in all the prediction performance metrics between the full and clinical models. Because these variables require no additional measurements beyond what is routinely available in clinical practice, the full models can be easily implemented to achieve this improvement in prediction performance. Chinese patients were at lower risk of the primary outcome during the first 3 years after biopsy, but at higher risk thereafter. This was not a finding unique to the derivation cohort; the same effect was observed when the full model with race/ethnicity was refit using the validation data. This highlights the importance of considering the time horizon when investigating ethnic differences in kidney outcome in IgAN.

    Limitations

    There are several limitations to our results. We included Chinese, Japanese, and white patients whose diagnosis and follow-up were within their countries of origin. Further research applying the models in other ethnic groups, or in countries with multi-ethnic populations or different biopsy practices, will be required. The prediction models apply only to biopsy-proven IgAN and are not applicable to other types of kidney disease. We included only adults in our analysis; therefore, the prediction models may not apply in children. Histologic data were only available for the presence or absence of crescents, and not the percentage of glomeruli involved, as proposed for the updated MEST-C score.33 Although the models can generate predicted risks at any time point after biopsy, we suggest using 5 years and no more than 7 years because these are the 50th and 75th percentiles of follow-up duration. This time horizon should be considered in the context of a lifelong, slowly progressive disease. Finally, the prediction models were designed to be applied near the time of biopsy, and additional research is needed to determine if they can be used at other time points in the trajectory of the disease.

    Conclusions

    This project was a large international research collaboration that derived and externally validated prediction models for kidney outcome in IgAN that use readily available clinical and histologic predictor variables and are suitable for clinical implementation in multiple ethnic groups worldwide.

    Back to top
    Article Information

    Accepted for Publication: February 15, 2019.

    Published Online: April 13, 2019. doi:10.1001/jamainternmed.2019.0600

    Correction: This article was corrected on January 25, 2021, to correct a number in eTable 3 of the Supplement. This article was also corrected on July 1, 2019, to add an omitted international group member’s name and affiliation.

    Corresponding Author: Sean J. Barbour, MD, MSc, Division of Nephrology, University of British Columbia, 2775 Laurel St, Fifth Floor, Vancouver, BC V5Z 1M9, Canada (sean.barbour@vch.ca).

    Author Contributions: Dr Barbour had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

    Study concept and design: Barbour, Katafuchi, Espino-Hernandez, Reich, Feehally, Cattran.

    Acquisition, analysis, or interpretation of data: Barbour, Coppo, Zhang, Liu, Suzuki, Matsuzaki, Katafuchi, Er, Espino-Hernandez, Kim, Reich, Cattran.

    Drafting of the manuscript: Barbour, Suzuki, Katafuchi, Er, Espino-Hernandez, Cattran.

    Critical revision of the manuscript for important intellectual content: Barbour, Coppo, Zhang, Liu, Matsuzaki, Katafuchi, Kim, Reich, Feehally, Cattran.

    Statistical analysis: Barbour, Katafuchi, Er, Espino-Hernandez, Kim, Reich, Cattran.

    Obtained funding: Barbour, Katafuchi.

    Administrative, technical, or material support: Barbour, Zhang, Liu, Matsuzaki, Katafuchi, Cattran.

    Study supervision: Barbour, Coppo, Suzuki, Reich, Cattran.

    Conflict of Interest Disclosures: Dr Barbour is a Scholar with the Michael Smith Foundation for Health Research. No other conflicts are reported.

    Funding/Support: Funding and support for this project was provided by grant funding from the Canadian Institutes of Health Research (PCG-155557). The VALIGA study was supported by a grant from the first research call and the Immunonephrology Working Group of the European Renal Association–European Dialysis and Transplant Association. The Oxford derivation and North American validation studies were supported by the International IgA Nephropathy Network, the Toronto GN Registry, and the Toronto General Hospital Foundation.

    Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    Group Information: The International IgA Nephropathy Network members are as follows: the VALIGA investigators: M.L. Russo (MA, PhD, Fondazione Ricerca Molinette, Torino, Italy); S. Troyanov (MD, Division of Nephrology, Department of Medicine, Hopital du Sacre-Coeur de Montreal, Montreal, Quebec, Canada); H.T. Cook (MD, Centre for Complement and Inflammation Research, Department of Medicine, Imperial College, London, England); I. Roberts (MD, Department of Cellular Pathology, Oxford University Hospitals NHS Foundation Trust, John Radcliffe Hospital, Oxford, United Kingdom); V. Tesar, (MD, Department of Nephrology, 1st Faculty of Medicine and General University Hospital, Charles University, Prague, Czech Republic); D. Maixnerova (MD, Department of Nephrology, 1st Faculty of Medicine and General University Hospital, Charles University, Prague, Czech Republic); S. Lundberg (MD, Nephrology Unit, Department of Clinical Sciences, Karolinska Institute, Stockholm, Sweden); L. Gesualdo (MD, Department of Nephrology, Emergency and Organ Transplantation, University of Bari “Aldo Moro,” Foggia-Bari, Italy); F. Emma (MD, Division of Nephrology, Department of Pediatric Subspecialties, Bambino Gesù Children’s Hospital IRCCS, Rome, Italy); L. Fuiano (MD, Division of Nephrology, Department of Pediatric Subspecialties, Bambino Gesù Children’s Hospital IRCCS, Rome, Italy); G. Beltrame (MD, Nephrology and Dialysis Unit, San Giovanni Bosco Hospital, and University of Turin,Turin, Italy); C. Rollino (MD, Nephrology and Dialysis Unit, San Giovanni Bosco Hospital, and University of Turin,Turin, Italy); A. Amore (MD, Nephrology Unit, Regina Margherita Children’s Hospital,Turin, Italy); R. Camilla (MD Nephrology Unit, Regina Margherita Children’s Hospital, Turin, Italy); L. Peruzzi (MD, Nephrology Unit, Regina Margherita Children’s Hospital, Turin, Italy); M. Praga (MD, Nephrology Unit, Hospital 12 de Octubre,Madrid, Spain); S. Feriozzi (MD, Nephrology Unit, Belcolle Hospital, Viterbo, Italy), R. Polci, (MD, Nephrology Unit, Belcolle Hospital,Viterbo, Italy); G. Segoloni, (MD, Division of Nephrology Dialysis and Transplantation, Department of Medical Sciences, Città della Salute e della Scienza Hospital and University of Turin,Turin, Italy); L.Colla (MD, Division of Nephrology Dialysis and Transplantation, Department of Medical Sciences, Città della Salute e della Scienza Hospital and University of Turin,Turin, Italy); A. Pani (MD, Nephrology Unit, G. Brotzu Hospital, Cagliari, Italy); D. Piras (MD, Nephrology Unit, G. Brotzu Hospital, Cagliari, Italy), A. Angioi (MD, Nephrology Unit, G. Brotzu Hospital, Cagliari, Italy); G. Cancarini, (MD, Nephrology Unit, Spedali Civili University Hospital, Brescia, Italy); S. Ravera (MD, Nephrology Unit, Spedali Civili University Hospital, Brescia, Italy); M. Durlik (MD, Department of Transplantation Medicine, Nephrology, and Internal Medicine, Medical University of Warsaw, Warsaw, Poland); E. Moggia (Nephrology Unit, Santa Croce Hospital, Cuneo, Italy); J. Ballarin (MD, Department of Nephrology, Fundacion Puigvert, Barcelona, Spain); S. Di Giulio (MD, Nephrology Unit, San Camillo Forlanini Hospital, Rome, Italy); F. Pugliese (MD, Department of Nephrology, Policlinico Umberto I University Hospital, Rome, Italy); I. Serriello (MD, Department of Nephrology, Policlinico Umberto I University Hospital, Rome, Italy); Y. Caliskan (MD, Division of Nephrology, Department of Internal Medicine, Istanbul Faculty of Medicine, Istanbul University, Istanbul, Turkey); M. Sever (MD, Division of Nephrology, Department of Internal Medicine, Istanbul Faculty of Medicine, Istanbul University, Istanbul, Turkey); I. Kilicaslan (MD, Department of Pathology, Istanbul Faculty of Medicine, Istanbul University, Istanbul, Turkey); F. Locatelli (MD, Department of Nephrology and Dialysis, Alessandro Manzoni Hospital, ASST Lecco, Italy); L. Del Vecchio (MD, Department of Nephrology and Dialysis, Alessandro Manzoni Hospital, ASST Lecco, Italy); J.F.M. Wetzels (MD, Departments of Nephrology, Radboud University Medical Center, Nijmegen, the Netherlands); H. Peters (MD, Departments of Nephrology, Radboud University Medical Center, Nijmegen, the Netherlands); U. Berg (MD, Division of Pediatrics, Department of Clinical Science, Intervention and Technology, Huddinge, Sweden); F. Carvalho (MD, Nephrology Unit, Hospital de Curry Cabral, Lisbon, Portugal); A.C. da Costa Ferreira (MD, Nephrology Unit, Hospital de Curry Cabral, Lisbon, Portugal); M. Maggio (MD, Nephrology Unit, Hospital Maggiore di Lodi, Lodi, Italy); A. Wiecek (MD, Department Nephrology, Endocrinology and Metabolic Diseases, Silesian University of Medicine, Katowice, Poland); M. Ots-Rosenberg(MD, Nephrology Unit, Tartu University Clinics, Tartu,Estonia); R. Magistroni (MD, Department of Nephrology, Policlinic of Modena and Reggio Emilia; Modena, Italy); R. Topaloglu (MD, Department of Pediatric Nephrology and Rheumatology, Hacettepe University, Ankara, Turkey); Y. Bilginer (MD, Department of Pediatric Nephrology and Rheumatology, Hacettepe University, Ankara, Turkey); M. D’Amico (MD, Nephrology Unit, S. Anna Hospital, Como, Italy); M. Stangou (MD, Department of Nephrology, Hippokration General Hospital, Aristotle University of Thessaloniki, Thessaloniki, Greece); F. Giacchino (MD, Nephrology Unit, Ivrea Hospital, Ivrea, Italy); D. Goumenos (MD Department of Nephrology, University Hospital of Patras, Patras, Greece); P. Kalliakmani (MD Department of Nephrology, University Hospital of Patras, Patras, Greece); M. Gerolymos (MD Department of Nephrology, University Hospital of Patras, Patras, Greece); K. Galesic (MD, Department of Nephrology, University Hospital Dubrava, Zagreb, Croatia); C. Geddes (MD, Renal Unit, Western Infirmary Glasgow, Glasgow, United Kingdom); K. Siamopoulos (MD, Nephrology Unit, Medical School University of Ioanina, Ioannina, Greece); O. Balafa (MD, Nephrology Unit,Medical School University of Ioanina, Ioannina, Greece); M. Galliani (MD, Nephrology Unit, S.Pertini Hospital, Rome, Italy); P. Stratta (MD, Department of Nephrology, Maggiore della Carità Hospital, Piemonte Orientale University, Novara, Italy); M. Quaglia (MD, Department of Nephrology, Maggiore della Carità Hospital, Piemonte Orientale University, Novara, Italy); R. Bergia (MD, Nephrology Unit, Degli Infermi Hospital, Biella,Italy); R. Cravero (MD, Nephrology Unit, Degli Infermi Hospital, Biella, Italy); M. Salvadori, (MD, Department of Nephrology, Careggi Hospital, Florence, Italy); L. Cirami (MD, Department of Nephrology, Careggi Hospital, Florence, Italy); B. Fellstrom (MD, Renal Department, University of Uppsala, Uppsala, Sweden); H. Kloster Smerud (MD, Renal Department, University of Uppsala, Uppsala, Sweden); F. Ferrario (MD, Nephropathology Unit, San Gerardo Hospital, Monza, Italy); T. Stellato (MD, Nephropathology Unit, San Gerardo Hospital, Monza, Italy); J. Egido (MD, Department of Nephrology, Fundacion Jimenez Diaz, Madrid, Spain); C. Martin (MD, Department of Nephrology, Fundacion Jimenez Diaz, Madrid, Spain); J. Floege (MD, Nephrology and Immunology, Medizinische Klinik II, University of Aachen, Aachen, Germany); F. Eitner (MD, Nephrology and Immunology, Medizinische Klinik II, University of Aachen, Aachen, Germany); A. Lupo (MD, Department of Nephrology, University of Verona, Verona, Italy); P. Bernich (MD, Department of Nephrology, University of Verona, Verona, Italy); P. Menè (Department of Nephrology,S. Andrea Hospital, Rome, Italy); M. Morosetti (Nephrology Unit, Grassi Hospital, Ostia, Italy); C. van Kooten, (MD, Department of Nephrology, Leiden University Medical Centre, Leiden, The Netherlands); T. Rabelink (MD, Department of Nephrology, Leiden University Medical Centre, Leiden, The Netherlands); M.E.J. Reinders (MD, Department of Nephrology, Leiden University Medical Centre, Leiden, The Netherlands); J.M. Boria Grinyo (Department of Nephrology, Hospital Bellvitge, Barcelona, Spain); S. Cusinato (MD, Nephrology Unit, Borgomanero Hospital, Borgomanero, Italy); L. Benozzi (MD, Nephrology Unit, Borgomanero Hospital, Borgomanero, Italy); S. Savoldi, (MD, Nephrology Unit, Civile Hospital, Ciriè, Italy); C. Licata (MD, Nephrology Unit, Civile Hospital, Ciriè, Italy); M. Mizerska-Wasiak (MD, Department of Pediatrics, Medical University of Warsaw, Warsaw, Poland); G. Martina (MD, Nephrology Unit, Chivasso Hospital, Chivasso, Italy); A. Messuerotti (MD, Nephrology Unit, Chivasso Hospital, Chivasso, Italy); A. Dal Canton (MD, Nephrology Unit, S. Matteo Hospital, Pavia, Italy); C. Esposito (MD, Nephrology Unit, Maugeri Foundation, Pavia, Italy); C. Migotto (MD, Nephrology Unit, Maugeri Foundation, Pavia, Italy); G. Triolo MD, Nephrology Unit CTO, Turin, Italy); F.Mariano (MD, Nephrology Unit CTO, Turin, Italy); C. Pozzi (MD, Nephrology Unit, Bassini Hospital, Cinisello Balsamo, Italy); R. Boero (MD, Nephrology Unit, Martini Hospital, Turin, Italy); the VALIGA pathology investigators: S. Bellur (MD, Department of Cellular Pathology, Oxford University Hospitals NHS Foundation Trust, John Radcliffe Hospital, Oxford, United Kingdom); G.Mazzucco (MD, Pathology Department, University of Turin, Turin, Italy); C. Giannakakis (MD, Pathology Department, La Sapienza University, Rome, Italy); E. Honsova (MD, Department of Clinical and Transplant Pathology, Institute for Clinical and Experimental Medicine, Prague, Czech Republic); B. Sundelin (MD Department of Pathology and Cytology, Karolinska University Hospital, Karolinska Institute, Stockholm, Sweden); A.M. Di Palma (Nephrology Unit, Aldo Moro University, Foggia-Bari, Italy); F. Ferrario (MD, Nephropathology Unit, San Gerardo Hospital, Monza, Italy); E. Gutiérrez (MD, Renal, Vascular and Diabetes Research Laboratory, Fundación Instituto de Investigaciones Sanitarias-Fundación Jiménez Díaz, Universidad Autónoma de Madrid, Madrid, Spain); A.M. Asunis (MD, Department of Pathology, Brotzu Hospital, Cagliari, Italy); J. Barratt (MD, The John Walls Renal Unit, Leicester General Hospital, Leicester, United Kingdom); R. Tardanico (MD, Department of Pathology, Spedali Civili Hospital, University of Brescia, Brescia, Italy); A. Perkowska-Ptasinska (MD, Department of Transplantation Medicine, Nephrology and Internal Medicine, Medical University of Warsaw, Warsaw, Poland); J. Arce Terroba (MD, Pathology Department, Fundació Puigvert, Barcelona, Spain); M. Fortunato (MD, Pathology Department, S. Croce Hospital, Cuneo, Italy); A. Pantzaki (MD, Department of Pathology, Hippokration Hospital, Thessaloniki, Greece); Y. Ozluk (MD, Department of Pathology, Istanbul University, Istanbul Faculty of Medicine, Istanbul, Turkey); E. Steenbergen (MD, Radboud University Medical Center, Department of Pathology, Nijmegen, The Netherlands); M. Soderberg (MD, Department of Pathology, Drug Safety and Metabolism, Huddinge, Sweden); Z. Riispere (MD, Department of Pathology, University of Tartu, Tartu, Estonia); L. Furci (MD, Pathology Department, University of Modena, Italy); D. Orhan (MD, Department of Pediatrics, Division of Rheumatology, Hacettepe University Faculty of Medicine, Ankara, Turkey); D. Kipgen (MD,Pathology Department, Queen Elizabeth University Hospital, Glasgow, United Kingdom); D. Casartelli (Pathology Department, Manzoni Hospital, Lecco, Italy); D. Galesic Ljubanovic (MD, Nephrology Department, University Hospital, Zagreb, Croatia; Zagreb, Croatia); H Gakiopoulou (MD, Department of Pathology, National and Kapodistrian University of Athens, Athens, Greece); E. Bertoni (MD, Nephrology Department, Careggi Hospital, Florence, Italy); P. Cannata Ortiz (MD, Pathology Department, IIS-Fundacion Jimenez Diaz UAM, Madrid, Spain); H. Karkoszka MD, (Nephrology, Endocrinology and Metabolic Diseases, Medical University of Silesia, Katowice, Katowice, Poland); H.J. Groene (MD, Cellular and Molecular Pathology, German Cancer Research Center, Heidelberg, Germany); A. Stoppacciaro (MD, Surgical Pathology Units, Department of Clinical and Molecular Medicine, Ospedale Sant’Andrea, Sapienza University of Rome, Rome, Italy); I. Bajema (MD, Department of Pathology, Leiden University Medical Center, Leiden, The Netherlands); J. Bruijn (MD, Department of Pathology, Leiden University Medical Center, Leiden, The Netherlands); X. Fulladosa Oliveras (MD, Nephrology Unit, Bellvitge University Hospital, Hospitalet de Llobregat, Barcelona, Spain); J. Maldyk (MD, Division of Pathomorphology, Children’s Clinical Hospital, Medical University of Warsaw, Warsaw, Poland); and E. Ioachim (MD, Department of Pathology, Medical School, University of Ioannina, Ioannina, Greece); the Oxford derivation and North American validation investigators: Bavbek N (MD, Department of Pathology, Vanderbilt University, Nashville, Tennessee); Cook T (MD, Imperial College, London, England), Troyanov S (MD, Division of Nephrology, Department of Medicine, Hopital du Sacre-Coeur de Montreal, Montreal, Quebec, Canada); Alpers C (MD, Department of Pathology, University of Washington Medical Center, Seattle, Washington), Amore A (MD, Nephrology, Dialysis and Transplantation Unit, Regina Margherita Children’s Hospital, University of Turin, Turin, Italy), Barratt J (MD, The John Walls Renal Unit, Leicester General Hospital, Leicester, England); Berthoux F (MD, Department of Nephrology, Dialysis, and Renal Transplantation, Hôpital Nord, CHU de Saint-Etienne, Saint-Etienne, France); Bonsib S (MD, Department of Pathology, LSU Health Sciences Center, Shreveport, Los Angeles); Bruijn J (MD, Department of Pathology, Leiden University Medical Center, Leiden, The Netherlands); D’Agati V (MD, Department of Pathology, Columbia University College of Physicians & Surgeons, New York, New York); D’Amico G (MD, Fondazione D’Amico per la Ricerca sulle Malattie Renali, Milan, Italy); Emancipator S (MD, Department of Pathology, Case Western Reserve University, Cleveland, Ohio); Emmal F (MD, Division of Nephrology and Dialysis, Department of Nephrology and Urology, Bambino Gesù Children’s Hospital and Research Institute, Piazza S Onofrio, Rome, Italy); Ferrario F (MD, Renal Immunopathology Center, San Carlo Borromeo Hospital, Milan, Italy); Fervenza F (MD PhD, Division of Nephrology and Hypertension, Mayo Clinic, Rochester); Florquin S (MD, Department of Pathology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands); Fogo A (MD, Department of Pathology, Vanderbilt University, Nashville, Tennessee); Geddes C (MD, The Renal Unit, Western Infirmary, Glasgow, Scotland); Groene H (MD, Department of Cellular and Molecular Pathology, German Cancer Research Center, Heidelberg, Germany); Haas M (MD, Department of Pathology and Laboratory Medicine, Cedars-Sinai Medical Center, Los Angeles, California); Hill P (MD, St Vincent’s Hospital, Melbourne, Australia); Hogg R (MD, Scott and White Medical Center, Temple, Texas (retired)); Hsu S (MD, Division of Nephrology, Hypertension and Renal Transplantation, College of Medicine, University of Florida, Gainesville, Florida); Hunley T (MD, Department of Pathology, Vanderbilt University, Nashville, Tennessee); Hladunewich (MD, Division of Nephrology, Sunnybrook Health Science Center, University of Toronto, Ontario, Canada M); Jennette C (MD, Department of Pathology and Laboratory Medicine, University of North Carolina, Chapel Hill, North Carolina); Joh K (MD, Division of Immunopathology, Clinical Research Center Chiba, East National Hospital, Chiba, Japan); Julian B (MD, Department of Medicine, University of Alabama at Birmingham, Birmingham, Alabama); Kawamura T (MD, Division of Nephrology and Hypertension, Jikei University School of Medicine, Tokyo, Japan); Lai F (MD, The Chinese University of Hong Kong, Hong Kong); Leung C (MD, Department of Medicine, Prince of Wales Hospital, Chinese University of Hong Kong, Hong Kong); Li L (MD, Research Institute of Nephrology, Jinling Hospital, Nanjing University School of Medicine, Nanjing, China); Li P (MD, Department of Medicine, Prince of Wales Hospital, Chinese University of Hong Kong, Hong Kong); Liu Z (MD, Research Institute of Nephrology, Jinling Hospital, Nanjing University School of Medicine, Nanjing, China); Massat A (MD, Division of Nephrology and Hypertension, Mayo Clinic, Rochester, Minnesota); Mackinnon B (MD, The Renal Unit, Western Infirmary, Glasgow, Scotland); Mezzano S (MD, Departamento de Nefrología, Escuela de Medicina, Universidad Austral, Valdivia, Chile); Schena F (MD, Renal, Dialysis and Transplant Unit, Policlinico, Bari, Italy); Tomino Y (MD, Division of Nephrology, Department of Internal Medicine, Juntendo University School of Medicine, Tokyo, Japan); Walker P (MD, Nephropathology Associates, Little Rock, Arkansas); Wang H (MD, Renal Division of Peking University First Hospital, Peking University Institute of Nephrology, Beijing, China(deceased)); Weening J (MD, Erasmus Medical Center, Rotterdam, The Netherlands); and Yoshikawa N (MD, Department of Pediatrics, Wakayama Medical University, Wakayama City, Japan); the International investigators: Cai-Hong Zeng (MD, Nanjing University School of Medicine, Nanjing, China); Sufang Shi (MD, Peking University Institute of Nephrology, Beijing, China); C.Nogi (MD, Juntendo University, Faculty of Medicine, Tokyo, Japan); H.Suzuki (MD, Juntendo University, Faculty of Medicine, Tokyo, Japan); K. Koike (MD, Division of Nephrology and Hypertension, Department of Internal Medicine, Jikei University School of Medicine,Tokyo, Japan); K. Hirano (MD, Division of Nephrology and Hypertension, Department of Internal Medicine, Jikei University School of Medicine, Tokyo, Japan); T. Kawamura (MD, Division of Nephrology and Hypertension, Department of Internal Medicine, Jikei University School of Medicine, Tokyo, Japan); T. Yokoo (MD, Division of Nephrology and Hypertension, Department of Internal Medicine, Jikei University School of Medicine, Tokyo, Japan); M. Hanai (MD, Division of Nephrology, Department of Medicine, Kurume University School of Medicine, Fukuoka, Japan); K. Fukami (MD, Division of Nephrology, Department of Medicine, Kurume University School of Medicine, Fukuoka,, Japan); K. Takahashi (MD, Department of Nephrology, Fujita Health University School of Medicne, Aichi, Japan); Y. Yuzawa (MD, Department of Nephrology, Fujita Health University School of Medicne, Aichi, Japan); M. Niwa (MD, Department of Nephrology, Nagoya University Graduate School of Medicine, Aichi, Japan); Y. Yasuda (MD, Department of Nephrology, Nagoya University Graduate School of Medicine, Aichi, Japan); S. Maruyama (MD, Department of Nephrology, Nagoya University Graduate School of Medicine, Aichi, Japan); D. Ichikawa (MD, Division of Nephrology and Hypertension, Department of Internal Medicine, St Marianna University School of Medicine, Kanagawa, Japan); T. Suzuki (MD, Division of Nephrology and Hypertension, Department of Internal Medicine, St Marianna University School of Medicine, Kanagawa, Japan); S. Shirai (MD, Division of Nephrology and Hypertension, Department of Internal Medicine, St Marianna University School of Medicine, Kanagawa, Japan); A. Fukuda (MD, First Department of Internal Medicine, Faculty of Medicine, University of Miyazaki, Miyazaki, Japan); S. Fujimoto (MD, Department of Hemovascular Medicine and Artificial Organs, Faculty of Medicine, University of Miyazaki, Miyazaki, Japan); H. Trimarchi (MD, Division of Nephrology, Hospital Britanico, Buenos Aires, Argentina).

    Meeting Presentation: This article was presented at the 2019 ISN World Congress of Nephrology; April 13, 2019; Melbourne, Australia.

    Additional Contributions: The authors would like to acknowledge all the investigators who contributed to the datasets used in this analysis as part of the International IgA Nephropathy Network. None of these investigators received financial compensation for their contributions.

    References
    1.
    Donadio  JV, Grande  JP.  IgA nephropathy.   N Engl J Med. 2002;347(10):738-748. doi:10.1056/NEJMra020109PubMedGoogle ScholarCrossref
    2.
    Reich  HN, Troyanov  S, Scholey  JW, Cattran  DC; Toronto Glomerulonephritis Registry.  Remission of proteinuria improves prognosis in IgA nephropathy.   J Am Soc Nephrol. 2007;18(12):3177-3183. doi:10.1681/ASN.2007050526PubMedGoogle ScholarCrossref
    3.
    Cattran  DC, Feehally  J, Cook  H,  et al.  Kidney Disease: Improving Global Outcomes (KDIGO) Glomerulonephritis Work Group: KDIGO clinical practice guideline for glomerulonephritis.   Kidney Int Suppl. 2012;2(2):139-274. doi:10.1038/kisup.2012.9Google ScholarCrossref
    4.
    Barbour  SJ, Reich  HN.  Risk stratification of patients with IgA nephropathy.   Am J Kidney Dis. 2012;59(6):865-873. doi:10.1053/j.ajkd.2012.02.326PubMedGoogle ScholarCrossref
    5.
    Bartosik  LP, Lajoie  G, Sugar  L, Cattran  DC.  Predicting progression in IgA nephropathy.   Am J Kidney Dis. 2001;38(4):728-735. doi:10.1053/ajkd.2001.27689PubMedGoogle ScholarCrossref
    6.
    Pozzi  C, Bolasco  PG, Fogazzi  GB,  et al.  Corticosteroids in IgA nephropathy: a randomised controlled trial.   Lancet. 1999;353(9156):883-887. doi:10.1016/S0140-6736(98)03563-6PubMedGoogle ScholarCrossref
    7.
    Manno  C, Torres  DD, Rossini  M, Pesce  F, Schena  FP.  Randomized controlled clinical trial of corticosteroids plus ACE-inhibitors with long-term follow-up in proteinuric IgA nephropathy.   Nephrol Dial Transplant. 2009;24(12):3694-3701. doi:10.1093/ndt/gfp356PubMedGoogle ScholarCrossref
    8.
    Lv  J, Zhang  H, Wong  MG,  et al; TESTING Study Group.  Effect of oral methylprednisolone on clinical outcomes in patients with IgA nephropathy: the TESTING randomized clinical trial.   JAMA. 2017;318(5):432-442. doi:10.1001/jama.2017.9362PubMedGoogle ScholarCrossref
    9.
    Rauen  T, Eitner  F, Fitzner  C,  et al; STOP-IgAN Investigators.  Intensive supportive care plus immunosuppression in IgA nephropathy.   N Engl J Med. 2015;373(23):2225-2236. doi:10.1056/NEJMoa1415463PubMedGoogle ScholarCrossref
    10.
    Maes  BD, Oyen  R, Claes  K,  et al.  Mycophenolate mofetil in IgA nephropathy: results of a 3-year prospective placebo-controlled randomized study.   Kidney Int. 2004;65(5):1842-1849. doi:10.1111/j.1523-1755.2004.00588.xPubMedGoogle ScholarCrossref
    11.
    Pozzi  C, Andrulli  S, Pani  A,  et al.  Addition of azathioprine to corticosteroids does not benefit patients with IgA nephropathy.   J Am Soc Nephrol. 2010;21(10):1783-1790. doi:10.1681/ASN.2010010117PubMedGoogle ScholarCrossref
    12.
    Hogg  RJ, Lee  J, Nardelli  N,  et al; Southwest Pediatric Nephrology Study Group.  Clinical trial to evaluate omega-3 fatty acids and alternate day prednisone in patients with IgA nephropathy: report from the Southwest Pediatric Nephrology Study Group.   Clin J Am Soc Nephrol. 2006;1(3):467-474. doi:10.2215/CJN.01020905PubMedGoogle ScholarCrossref
    13.
    Li  PK, Leung  CB, Chow  KM,  et al; HKVIN Study Group.  Hong Kong study using valsartan in IgA nephropathy (HKVIN): a double-blind, randomized, placebo-controlled study.   Am J Kidney Dis. 2006;47(5):751-760. doi:10.1053/j.ajkd.2006.01.017PubMedGoogle ScholarCrossref
    14.
    Coppo  R, Peruzzi  L, Amore  A,  et al.  IgACE: a placebo-controlled, randomized trial of angiotensin-converting enzyme inhibitors in children and young people with IgA nephropathy and moderate proteinuria.   J Am Soc Nephrol. 2007;18(6):1880-1888. doi:10.1681/ASN.2006040347PubMedGoogle ScholarCrossref
    15.
    Cattran  DC, Coppo  R, Cook  HT,  et al; Working Group of the International IgA Nephropathy Network and the Renal Pathology Society.  The Oxford classification of IgA nephropathy: rationale, clinicopathological correlations, and classification.   Kidney Int. 2009;76(5):534-545. doi:10.1038/ki.2009.243PubMedGoogle ScholarCrossref
    16.
    Roberts  IS, Cook  HT, Troyanov  S,  et al; Working Group of the International IgA Nephropathy Network and the Renal Pathology Society.  The Oxford classification of IgA nephropathy: pathology definitions, correlations, and reproducibility.   Kidney Int. 2009;76(5):546-556. doi:10.1038/ki.2009.168PubMedGoogle ScholarCrossref
    17.
    Wakai  K, Kawamura  T, Endoh  M,  et al.  A scoring system to predict renal outcome in IgA nephropathy: from a nationwide prospective study.   Nephrol Dial Transplant. 2006;21(10):2800-2808. doi:10.1093/ndt/gfl342PubMedGoogle ScholarCrossref
    18.
    Goto  M, Wakai  K, Kawamura  T, Ando  M, Endoh  M, Tomino  Y.  A scoring system to predict renal outcome in IgA nephropathy: a nationwide 10-year prospective cohort study.   Nephrol Dial Transplant. 2009;24(10):3068-3074. doi:10.1093/ndt/gfp273PubMedGoogle ScholarCrossref
    19.
    Goto  M, Kawamura  T, Wakai  K, Ando  M, Endoh  M, Tomino  Y.  Risk stratification for progression of IgA nephropathy using a decision tree induction algorithm.   Nephrol Dial Transplant. 2009;24(4):1242-1247. doi:10.1093/ndt/gfn610PubMedGoogle ScholarCrossref
    20.
    Berthoux  F, Mohey  H, Laurent  B, Mariat  C, Afiani  A, Thibaudin  L.  Predicting the risk for dialysis or death in IgA nephropathy.   J Am Soc Nephrol. 2011;22(4):752-761. doi:10.1681/ASN.2010040355PubMedGoogle ScholarCrossref
    21.
    Bjørneklett  R, Vikse  BE, Bostad  L, Leivestad  T, Iversen  BM.  Long-term risk of ESRD in IgAN; validation of Japanese prognostic model in a Norwegian cohort.   Nephrol Dial Transplant. 2012;27(4):1485-1491. doi:10.1093/ndt/gfr446PubMedGoogle ScholarCrossref
    22.
    Pesce  F, Diciolla  M, Binetti  G,  et al.  Clinical decision support system for end-stage kidney disease risk estimation in IgA nephropathy patients.   Nephrol Dial Transplant. 2016;31(1):80-86. doi:10.1093/ndt/gfv232PubMedGoogle ScholarCrossref
    23.
    Xie  J, Kiryluk  K, Wang  W,  et al.  Predicting progression of IgA nephropathy: new clinical progression risk score.   PLoS One. 2012;7(6):e38904. doi:10.1371/journal.pone.0038904PubMedGoogle Scholar
    24.
    Barbour  SJ, Cattran  DC, Kim  SJ,  et al.  Individuals of Pacific Asian origin with IgA nephropathy have an increased risk of progression to end-stage renal disease.   Kidney Int. 2013;84(5):1017-1024. doi:10.1038/ki.2013.210PubMedGoogle ScholarCrossref
    25.
    Barbour  SJ, Espino-Hernandez  G, Reich  HN,  et al; Oxford Derivation, North American Validation and VALIGA Consortia; Oxford Derivation North American Validation and VALIGA Consortia.  The MEST score provides earlier risk prediction in lgA nephropathy.   Kidney Int. 2016;89(1):167-175. doi:10.1038/ki.2015.322PubMedGoogle ScholarCrossref
    26.
    Tanaka  S, Ninomiya  T, Katafuchi  R,  et al.  Development and validation of a prediction rule using the Oxford classification in IgA nephropathy.   Clin J Am Soc Nephrol. 2013;8(12):2082-2090. doi:10.2215/CJN.03480413PubMedGoogle ScholarCrossref
    27.
    Coppo  R, Troyanov  S, Bellur  S,  et al; VALIGA study of the ERA-EDTA Immunonephrology Working Group.  Validation of the Oxford classification of IgA nephropathy in cohorts with different presentations and treatments.   Kidney Int. 2014;86(4):828-836. doi:10.1038/ki.2014.63PubMedGoogle ScholarCrossref
    28.
    Zeng  CH, Le  W, Ni  Z,  et al.  A multicenter application and evaluation of the oxford classification of IgA nephropathy in adult Chinese patients.   Am J Kidney Dis. 2012;60(5):812-820. doi:10.1053/j.ajkd.2012.06.011PubMedGoogle ScholarCrossref
    29.
    Shi  SF, Wang  SX, Jiang  L,  et al.  Pathologic predictors of renal outcome and therapeutic efficacy in IgA nephropathy: validation of the oxford classification.   Clin J Am Soc Nephrol. 2011;6(9):2175-2184. doi:10.2215/CJN.11521210PubMedGoogle ScholarCrossref
    30.
    Herzenberg  AM, Fogo  AB, Reich  HN,  et al.  Validation of the Oxford classification of IgA nephropathy.   Kidney Int. 2011;80(3):310-317. doi:10.1038/ki.2011.126PubMedGoogle ScholarCrossref
    31.
    Katafuchi  R, Ninomiya  T, Nagata  M, Mitsuiki  K, Hirakata  H.  Validation study of oxford classification of IgA nephropathy: the significance of extracapillary proliferation.   Clin J Am Soc Nephrol. 2011;6(12):2806-2813. doi:10.2215/CJN.02890311PubMedGoogle ScholarCrossref
    32.
    Levey  AS, Stevens  LA, Schmid  CH,  et al; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration).  A new equation to estimate glomerular filtration rate.   Ann Intern Med. 2009;150(9):604-612. doi:10.7326/0003-4819-150-9-200905050-00006PubMedGoogle ScholarCrossref
    33.
    Trimarchi  H, Barratt  J, Cattran  DC,  et al; IgAN Classification Working Group of the International IgA Nephropathy Network and the Renal Pathology Society; Conference Participants.  Oxford classification of IgA nephropathy 2016: an update from the IgA Nephropathy Classification Working Group.   Kidney Int. 2017;91(5):1014-1021. doi:10.1016/j.kint.2017.02.003PubMedGoogle ScholarCrossref
    34.
    Inker  LA, Lambers Heerspink  HJ, Mondal  H,  et al.  GFR decline as an alternative end point to kidney failure in clinical trials: a meta-analysis of treatment effects from 37 randomized trials.   Am J Kidney Dis. 2014;64(6):848-859. doi:10.1053/j.ajkd.2014.08.017PubMedGoogle ScholarCrossref
    35.
    Levey  AS, Inker  LA, Matsushita  K,  et al.  GFR decline as an end point for clinical trials in CKD: a scientific workshop sponsored by the National Kidney Foundation and the US Food and Drug Administration.   Am J Kidney Dis. 2014;64(6):821-835. doi:10.1053/j.ajkd.2014.07.030PubMedGoogle ScholarCrossref
    36.
    Berthoux  F, Mariat  C, Maillard  N.  Overweight/obesity revisited as a predictive risk factor in primary IgA nephropathy.   Nephrol Dial Transplant. 2013;28(suppl 4):iv160-iv166. doi:10.1093/ndt/gft286PubMedGoogle ScholarCrossref
    37.
    Cattran  DC, Reich  HN, Beanlands  HJ, Miller  JA, Scholey  JW, Troyanov  S; Genes, Gender and Glomerulonephritis Group.  The impact of sex in primary glomerulonephritis.   Nephrol Dial Transplant. 2008;23(7):2247-2253. doi:10.1093/ndt/gfm919PubMedGoogle ScholarCrossref
    38.
    Royston  P, Altman  DG.  External validation of a Cox prognostic model: principles and methods.   BMC Med Res Methodol. 2013;13:33. doi:10.1186/1471-2288-13-33PubMedGoogle ScholarCrossref
    39.
    Chambless  LE, Cummiskey  CP, Cui  G.  Several methods to assess improvement in risk prediction models: extension to survival analysis.   Stat Med. 2011;30(1):22-38. doi:10.1002/sim.4026PubMedGoogle ScholarCrossref
    40.
    Moons  KG, Altman  DG, Reitsma  JB,  et al.  Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD): explanation and elaboration.   Ann Intern Med. 2015;162(1):W1-73. doi:10.7326/M14-0698PubMedGoogle ScholarCrossref
    41.
    Knoop  T, Vågane  AM, Vikse  BE,  et al.  Addition of eGFR and age improves the prognostic absolute renal risk-model in 1,134 Norwegian patients with IgA nephropathy.   Am J Nephrol. 2015;41(3):210-219. doi:10.1159/000381403PubMedGoogle ScholarCrossref
    42.
    Barratt  J, Feehally  J.  IgA nephropathy.   J Am Soc Nephrol. 2005;16(7):2088-2097. doi:10.1681/ASN.2005020134PubMedGoogle ScholarCrossref
    43.
    Debray  TP, Vergouwe  Y, Koffijberg  H, Nieboer  D, Steyerberg  EW, Moons  KG.  A new framework to enhance the interpretation of external validation studies of clinical prediction models.   J Clin Epidemiol. 2015;68(3):279-289. doi:10.1016/j.jclinepi.2014.06.018PubMedGoogle ScholarCrossref
    44.
    Haas  M, Verhave  JC, Liu  ZH,  et al.  A multicenter study of the predictive value of crescents in IgA nephropathy.   J Am Soc Nephrol. 2017;28(2):691-701. doi:10.1681/ASN.2016040433PubMedGoogle ScholarCrossref
    ×