Evaluation of the Oxford Classification of IgA Nephropathy: A Systematic Review and Meta-analysis

Original Investigation Evaluation of the Oxford Classification of IgA Nephropathy: A Systematic Review and Meta-analysis Jicheng Lv, MD,1* Sufang Shi, MD,1* Damin Xu, MD,1 Hong Zhang, MD, PhD,1 Stéphan Troyanov, MD,2 Daniel C. Cattran, MD,3 and Haiyan Wang, MD1 Background: The Oxford Classification of the pathology of immunoglobulin A (IgA) nephropathy, developed in 2009, is highly predictive of renal prognosis. It has been validated in different populations, but the results remain inconsistent. Study Design: Systematic review and meta-analysis. Setting & Population: Patients with biopsy-proven primary IgA nephropathy. Selection Criteria for Studies: Studies assessing the Oxford Classification of IgA nephropathy published between January 2009 and December 2012 were included following systematic searching of the MEDLINE and EMBASE databases. Predictor: 4 pathologic lesions of the Oxford Classification: mesangial hypercellularity (M), endocapillary hypercellularity (E), segmental glomerulosclerosis (S), and tubular atrophy/interstitial fibrosis (T). Outcome: Kidney failure defined as doubled serum creatinine level, 50% decline in estimated glomerular filtration rate, or end-stage kidney disease. Results: 16 retrospective cohort studies with 3,893 patients and 570 kidney failure events were included. In a multivariate model, HRs for kidney failure were 0.6 (95% CI, 0.5-0.8; P ⬍ 0.001), 1.8 (95% CI, 1.4-2.4; P ⬍ 0.001), and 3.2 (95% CI, 1.8-5.6; P ⬍ 0.001) for scores of M0 (mesangial hypercellularity score ⱕ0.5), S1 (presence of segmental glomerulosclerosis), and T1/2 (⬎25% tubular atrophy/interstitial fibrosis), respectively, without evidence of heterogeneity. Pooled results showed that E lesions were not associated with kidney failure (HR, 1.4; 95% CI, 0.9-2.0; P ⫽ 0.1), with evidence of heterogeneity (I 2 ⫽ 54.1%; P ⫽ 0.01). Crescent (C) lesions were associated with kidney failure (HR, 2.3; 95% CI, 1.6-3.4; P ⬍ 0.001), with no evidence of heterogeneity (I 2 ⫽ 14.7%; P ⫽ 0.3). Limitations: All studies were retrospective. This was not an individual-patient-data meta-analysis. Conclusions: This study suggests that M, S, T, and C lesions, but not E lesions, are associated strongly with progression to kidney failure and thus should be included in the Oxford Classification system. Am J Kidney Dis. 62(5):891-899. © 2013 by the National Kidney Foundation, Inc. INDEX WORDS: Immunoglobulin A (IgA) nephropathy; Oxford Classification; meta-analysis.

I

mmunoglobulin A (IgA) nephropathy (IgAN), the most common glomerular disease worldwide, is characterized by IgA deposition in the glomerular mesangium with widely variable clinical and pathologic presentations.1-3 Histologic classification is essential for evaluating the severity of the lesions and guiding therapeutic strategies for IgAN in clinical practice,4,5 but most classifications are not widely accepted.6-8 The Oxford Classification of IgAN, developed by an international working group in 2009, was established by analyzing biopsy specimens from 265 patients and was intended to establish a consensus on specific pathologic features that reliably predict the risks of progression of IgAN. Four histopathologic features—mesangial hypercellularity (M), endocapillary hypercellularity (E), segmental glomerulosclerosis (S), and tubular atrophy and interstitial fibrosis (T)—were identified as histopathologic predictors of the renal prognosis of IgAN independent of clinical features.9,10 The original Oxford Classification study differed from previous classifications by using a stepwise development methodology that at first considered all plausible variables, eliminated those with poor reproAm J Kidney Dis. 2013;62(5):891-899

ducibility, avoided highly collinear biopsy findings, tested univariate predictive value, and finally, adjusted for clinical risk factors for progression. Since then, the Oxford Classification has been validated in many studies, as well as in different ethnic populations.11-26 However, the results are still debated, partly From the 1Renal Division, Department of Medicine, Peking University First Hospital; Peking University Institute of Nephrology; Key Laboratory of Renal Disease, Ministry of Health of China; Key Laboratory of Chronic Kidney Disease Prevention and Treatment (Peking University), Ministry of Education, Beijing, China; 2Department of Nephrology, Hôpital du Sacré-Coeur de Montréal, University of Montreal, Montreal, Quebec; and 3Division of Nephrology, Toronto General Hospital, University of Toronto, Ontario, Canada. * J.L. and S.S contributed equally to this work. Received October 10, 2012. Accepted in revised form April 25, 2013. Originally published online July 1, 2013. Address correspondence to Hong Zhang, Renal Division, Department of Medicine, Peking University First Hospital, No. 8 Xi Shi Ku Street, Xi Cheng District, Beijing 100034, PR China. E-mail: [email protected] © 2013 by the National Kidney Foundation, Inc. 0272-6386/$36.00 http://dx.doi.org/10.1053/j.ajkd.2013.04.021 891

Lv et al

due to the small sample size or low number of end points observed in most studies, which has limited the statistical power necessary to draw firm conclusions. In the present study, we aimed to perform a systematic review and pool the available validation data to evaluate the predictive value of the Oxford Classification of IgAN.

METHODS Data Sources, Search Strategy, and Selection Criteria We performed a systematic review of the published literature according to the approach recommended by the PRISMA (Preferred Reporting Items for Systematic Reviews and Metaanalyses) statement for the conduct of meta-analyses.27 Relevant studies were identified by searching MEDLINE by Ovid and EMBASE (from January 2009 to December 2012), with relevant text words and medical subject headings that included all spellings of “IgA nephropathy,” “IgA nephritis,” “IgAN,” “Berger’s disease,” “Berger’s syndrome,” “chronic glomerulonephritis,” “Oxford,” and “Oxford Classification” (Item S1, available as online supplementary material). The search was limited to studies validating the Oxford Classification in primary IgA, without language restriction. Studies that re-evaluated the Oxford cohort or subgroup analysis using the same population were excluded. To identify other relevant studies, we manually scanned the reference lists of the identified studies and review articles and searched conference proceedings on kidney disease from May 2009 to December 2012; one such study28 from an academic meeting was included in the present study (Fig 1). The literature search, data extraction, and quality assessment were undertaken independently by 2 authors (J.L. and S.S.) using a standardized approach. All completed studies that validated the Oxford Classification of primary IgAN were eligible for inclusion. Study end points had to include end-stage kidney disease, ⬎50% decrease in estimated glomerular filtration rate (eGFR), or doubling of serum creatinine concentration (approximately equivalent to halving the GFR). Studies that

159 articles found 27 PubMed 132 EMBASE 30 duplicate studies excluded 129 abstracts screened 1 article identified from meeting abstract

93 excluded 24: not original study (eg, review) 53: not IgAN (eg, lupus nephritis) 4: publication from Oxford cohort 12: basic research study

Full-text analysis (n=37) 21 articles that did not evaluate renal outcome 16 studies with 3,893 patients and 570 endpoint events Figure 1. Flow chart for study selection. 892

used eGFR decline as the outcome were limited and pooled analysis was not performed.

Data Extraction and Quality Assessment Published reports were obtained for each eligible study and standard information was extracted into a spreadsheet. The data sought included baseline patient characteristics (ethnicity, sex, and age), mean arterial blood pressure, eGFR, 24-hour urine protein excretion, follow-up duration, percentage of patients receiving hypertension/antihypertensive medication, percentage receiving renin-angiotensin-aldosterone system inhibition treatment, percentage receiving immunosuppressive therapy, number of end point events, definition of kidney disease outcome, pathologic methodology, statistical methodology, and characteristics of the Oxford Classification pathologic lesions (M, E, S, and T). Any disagreements about the abstracted data were adjudicated by a third reviewer (H.Z.).

Statistical Analysis Individual-study hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated from event numbers or extracted from each study before data pooling. Summary estimates of HRs were obtained from a random-effects model. The percentage of variability across studies attributable to heterogeneity beyond chance was estimated with the I 2 statistic. We explored potential heterogeneity in the predictive value of the 4 pathologic lesions by comparing the summary results obtained from subsets of studies grouped according to number of patients, number of end point events, child or adult, follow-up time, ethnicity, and immunosuppressive treatment. A 2-sided P ⬍ 0.05 was regarded as significant for all analyses. All statistical analyses were done using Stata, version 12.0 (StataCorp LP).

RESULTS Trial Flow and Study Characteristics The literature search yielded 160 articles. Sixteen studies11,13-26,28 including 3,893 patients with 570 end point events were eligible for inclusion. Reasons for exclusion are listed in Fig 1. All studies were retrospective, with a mean follow-up of 38-156 months. In 8 studies,11,14,16,19,20,23,25,28 more than 2 independent pathologists who were blinded to the clinical outcomes scored every feature according to the Oxford Classification. Characteristics of the included studies are summarized in Table 1. Three studies used inclusion criteria similar to those of the Oxford cohort16,23,26; 11 studies17-26,28 (2,996 patients) were of Asian populations including Chinese, Japanese, and Korean; and 5 studies11,13-16 were from North American or European countries (897 patients). Three studies13,20,24 with 478 participants included children and adolescents only. In most studies, the patients had received angiotensin-converting enzyme inhibitors or angiotensin receptor blockers; percentages ranged from 58%-99%. Three studies15,18,22 evaluated patients with minimal or no proteinuria, with relatively few patients (⬍43%) receiving reninangiotensin-aldosterone system inhibitors. All studies used multivariate regression models for statistical Am J Kidney Dis. 2013;62(5):891-899

Oxford Classification of IgAN

analysis; the adjusted confounders included at least blood pressure, proteinuria, kidney function, and the 4 pathologic lesions (M, E, S, and T). Applicability of the 4 Oxford Classification Pathologic Lesions Mesangial Hypercellularity Thirteen studies with 3,629 patients and 539 kidney failure events reported the predictive value of M lesions. As shown in Fig 2, a score of M0 (defined as mesangial hypercellularity score ⱕ0.5) was associated independently with lower risk of kidney failure in IgAN (M1 [mesangial hypercellularity score ⬎0.5] as the reference category; HR, 0.6; 95% CI, 0.5-0.8; P ⬍ 0.001), with no evidence of heterogeneity (I 2 ⫽ 18.1%; P ⫽ 0.3). This analysis was dominated by a large Chinese study (1,026 patients; accounting for 26.4% of the weight). Exclusion of this study did not affect the association of M0 lesions with lower risk of kidney failure (HR, 0.6; 95% CI, 0.4-0.8; P ⬍ 0.001). Endocapillary Hypercellularity

Thirteen studies (3,511 participants with 529 kidney failure events) reported the association of E lesions with kidney survival. Overall, a score of E1 (defined as any E lesion present) was not associated with poor kidney survival in the pooled analysis (E0 as reference; HR for kidney failure, 1.4; 95% CI, 0.9-2.0; P ⫽ 0.1). Evidence of heterogeneity lay in the magnitude of the effect across the included studies (I 2 ⫽ 54.1%; P ⫽ 0.01; Fig 3). Univariate metaregression analysis revealed the presence of heterogeneity in the effect of participant number, number of events recorded, age of study participants (children/adolescents or not), and race (Asian or not). E lesions were associated with greater likelihood of kidney survival in studies with a small sample size or less end point events (Fig 4; P ⬍ 0.05). Univariate regression found no interaction between immunosuppressive therapy and E lesions (P ⫽ 0.2; Table 2; Fig S1), which had been reported to be the case in the Oxford Classification. Segmental Glomerulosclerosis/Adhesion

Fourteen studies with 3,771 patients and 519 kidney failure events reported the association between S lesions and kidney failure. These studies showed that a score of S1 (defined as the presence of segmental glomerulosclerosis) was associated strongly with progression to kidney failure (S0 as reference; HR, 1.8; 95% CI, 1.4-2.4; P ⬍ 0.001), with no evidence of heterogeneity (I 2 ⫽ 34.5%; P ⫽ 0.1; Fig 5). This analysis was dominated by a large Japanese study (702 patients accounting for 30.5% of the weight). Am J Kidney Dis. 2013;62(5):891-899

Exclusion of this study did not affect the association of S1 lesions with kidney failure (HR, 1.7; 95% CI, 1.2-2.3; P ⫽ 0.002). Tubular Atrophy and Interstitial Fibrosis

Six studies with 2,719 patients and 364 kidney failure events reported the association of a score of T1 (defined as ⬎25% but ⬍50% tubular atrophy/interstitial fibrosis) with kidney failure. These studies showed that T1 was associated strongly with progression to kidney failure (T0 [⬍25% tubular atrophy/interstitial fibrosis] as reference; HR, 2.7; 95% CI, 1.6-4.6; P ⬍ 0.001), with moderate evidence of heterogeneity (I 2 ⫽ 59.1%; P ⫽ 0.03; Fig 6). Five studies with 2,558 patients and 357 kidney failure events described the association of T2 score (defined as ⬎50% tubular atrophy/interstitial fibrosis) with kidney failure (Fig 6). These studies showed that T2 was associated strongly with progression to kidney failure (T0 as reference; HR, 7.2; 95% CI, 4.9-10.6; P ⬍ 0.001), with no evidence of heterogeneity (I 2 ⫽ 0.0%; P ⫽ 0.8). Seven studies with 722 patients and 124 kidney failure events reported the association of the combined category of T1 or T2 scores with kidney failure events (Fig 6). These studies showed that T1/T2 scores were associated strongly with progression to kidney failure (T0 as reference; HR, 3.2; 95% CI, 1.8-5.6; P ⬍ 0.001), with no evidence of heterogeneity (I 2 ⫽ 37.7%; P ⫽ 0.1). Cellular/Fibrocellular Crescents Five studies (4 Asian and 1 European) with 1,487 patients and 206 end point events reported the association of crescents with kidney failure. These studies showed that a score of C1 (defined as any presence of crescents) was associated strongly with progression to kidney failure (C0 as reference; HR, 2.3; 95% CI, 1.6-3.4; P ⬍ 0.001), with no evidence of heterogeneity (I 2 ⫽ 14.7%; P ⫽ 0.3; Fig 7).

DISCUSSION In the new Oxford Classification, the International IgAN Network has named 4 pathologic lesions of singular prognostic importance: mesangial (M) and endocapillary (E) hypercellularity, segmental glomerulosclerosis (S), and tubular atrophy and interstitial fibrosis (T). Validation of the Oxford Classification is essential before it is recommended for implementation on a global scale. Although many studies have tried to evaluate the use of the Oxford Classification, validation is difficult. This is due mostly to the lack of hard end points observed in most studies. In multivariate models, there are at least 7 adjusted confounders 893

Lv et al Table 1. Clinical, Pathologic, and Statistical Characteristics of Included Studies (continued on following page) Validation Cohort Studies (continued on following page) Joh28

Oxford Cohort

El Karoui14

Alamartine11

Kang17

Herzenberg16

Yau25

Edstrom Halling13

Country

Multicountry

Japan

France

France

Korea

Canada/US

US

Sweden

Ethnicity White Asian African Other

66% 27% 3% 4%

— 100% — —

92.2% 7.8% — —

NA NA NA NA

— 100% — —

69% 20% 5% 6%

50% 26% 6% 19%

NA NA NA NA Single

Designa

Multi

Single

Multi

Single

Single

Multi

Single

No. pts

265

233

128

183

197

187

54

99

F/U (mo)

65

127

44

77

56.8

53

70

156

Age (y)

30 (4-73)

36 (18-70)

38.7 (18-78)

38

32.4

34

41

9.7

Age ⬍18 y

22.3%

0%

0%

NA

NA

24%

NA

100%

M:F

2.6:1

1:1

2.1:1

3:1

1.4:1

1.4:1

1.3:1

1.4:1

Proteinuria (g/d)

2.0

1.7

1.3

2.4

1.2

2.1

1.7

2.0

eGFR (mL/min/1.73m2) 83

78

52.1

72

87.1

82

61

100

CKD1/2/3/4/5c

27/50/23/0/0

NA

28/34/19/7/6

NA

NA

54 (1&2)/30/15/2

75/16/4/3/2

36/38/26/0/0

MAP (mm Hg)

98

94

NA

NA

91.4

96

NA

85.4

HTNd

31%

9%

NA

38%

24.4%

58%

48%

NA

RAAS blockade

74%

77%

99%

65%

83%

87%

78%

86%

Immunosuppressive therapy

29%

35%

0.8%

30%

38%

41%

35%

43%

End point definition

⬎100% increase ⬎100% increase 50% decline in 50% decline in 50% decline in 50% decline in eGFR or eGFR or of baseline of baseline SCr eGFR or eGFR or ESKD; slope ESKD SCr or ESKD or ESKD; slope ESKD; slope of ESKD; slope of of eGFR of eGFR eGFR eGFR

50% decline in eGFR or ESKD

50% decline in eGFR or ESKD

No. end point events

58

58

41

36

16

26

10

18

No. of ESKD events

34

40

35

30

NA

21

7

15

No. of pathologists

5, blinded

NA

2, blinded

1 (& 1 1 nephrologist)

2

2, blinded

1, blinded

Lesion M0/M1 E0/E1 S0/S1 T0/T1/T2 C0/C1

20/80 58/42 NA NA 55/45

NA NA NA NA NA

66/34 75/25 31/69 56/26/23 75.8/24.2

79/21 86/14 46/54 70/20/10 95/5

NA NA NA NA NA

28/72 80/20 19/81 65/13/22 81/19

69/31 90/10 77/23 84/12/3 82/18

26/74 89/11 44/56 66/26/8 NA

Statistical methodsf

Linear regression, Linear regression, Linear regression, Cox regression Cox regression Cox regression logistic regression, Cox regression

Cox regression Linear regression Logistic regression, Cox regression Cox regression

Adjusted factors

Initial GFR, MAP, Initial GFR, MAP, Initial GFR, MAP, Initial GFR, proteinuria MAP, proteinuria; proteinuria; proteinuria initial GFR & initial GFR & F/U MAP, F/U MAP, proteinuria proteinuria

None

Initial GFR & F/U Initial GFR, MAP, proteinuria MAP, proteinuria

Initial proteinuria & 1 y-F/U proteinuria

Note: Values for continuous variables are given as mean or mean (range). Abbreviations: CKD, chronic kidney disease; E, endocapillary hypercellularity; eGFR, estimated glomerular filtration rate; ESKD, end-stage kidney disease; F/U, follow-up; GFR, glomerular filtration rate; HTN, hypertension; M, mesangial hypercellularity; MAP, mean arterial pressure; NA, not available; pts, patients; RAAS, renin-angiotensin-aldosterone system; S, segmental glomerulosclerosis; SCr, serum creatinine; T, tubular atrophy/interstitial fibrosis; US, United States. (continued on following page)

(blood pressure, proteinuria, kidney function, and the 4 Oxford Classification pathologic lesions). This means that validation studies need more than 70 end points, but only 218,26 have achieved such study power to date. Systematic review and meta-analysis is a strategy to overcome these obstacles. In the present study, with more than 3,800 participants and more than 500 end point events, we have consistently confirmed that in patients with IgAN, the presence of the M, S, or T lesions in the Oxford Classification is associated strongly with progression to kidney failure independently of all clinical and laboratory parameters. In 894

particular, T1 and T2 were associated with increased risk of kidney failure (⬎2 and 7 times as likely, respectively, as T0), which indicates that tubular and interstitial lesions strongly predict kidney disease outcomes. E lesions were not an independent predictive factor. We also evaluated crescent lesions, which are not included in the Oxford Classification, and found that they were associated strongly with progression to kidney failure. Thus, our findings support the use of the M, S, T, and C pathology lesions for classification purposes. The Oxford Classification work group included E lesions mainly due to their interaction with the Am J Kidney Dis. 2013;62(5):891-899

Oxford Classification of IgAN Table 1 (Cont’d). Clinical, Pathologic, and Statistical Characteristics of Included Studies Validation Cohort Studies (continued from previous page) Shi23

Zeng26

Shima24

Katafuchi18

Gutierrez15

Kataoka19

Le20

Lee21

Moriyama22

China

China

Japan

Japan

Spain

Japan

China

Korea

Japan

— 100% — —

— 100% — —

— 100% — —

— 100% — —

100% — — —

— 100% — —

— 100% — —

— 100% — —

— 100% — —

Single

Multi

Multi

Single

Multi

Single

Multi

Single

Single

410

1,026

161

702

141

43

218

69

42

38

53

54

62

108

120

56

85

NA

31

34

11.7 (3.6-19.4)

30

23.7 (5-71)

40

14 (2-17.9)

34 (27-45)

34.2

0%

0%

100%

NA

NA

0%

100%

0%

0%

1:1

1:1

1.7:1

1:1.2

1.8:1

1.5:1

1.9:1

1.4:1

1:1

1.7

1.3

0.7

0.9b

0.2 (0.1-0.4)

1.8

1.5 (0.5-8.0)

1.2 (0.4-1.9)

5.7

85.8

85

103

82

111.7

78.3

134

90.8

51.1

43/38/19/0/0

41/35/24/0/0

NA

37/37/21/3/1

10 (1&2)/0/0/0/0

NA

NA

NA

NA

94

98

79

92

86.2

102.6

88

97

NA

NA

30.5%

NA

NA

16.3%

NA

NA

94%

NA

86.1%

89%

NA

37%

41.8%

58.1%

61.5%

90%

54.8%

42.7%

31%

16%

32%e

0

51.2%

56%

18.3%

64.2%

ESKD

50% decline in eGFR or ESKD; slope of eGFR

CKD stages 3-5

ESKD

⬎50% increase of baseline SCr

⬎150% increase of baseline SCr

50% decline in eGFR or ESKD

50% decline in eGFR or ESKD

ESKD

30

159

7

84

5

16

24

16

24

30

38

5

84

0

NA

NA

NA

24

2, blinded

2

NA

NA

1 at each center, blinded

2, blinded

2

NA

NA

44/56 43/57 25/75 78/14/8 40/60

57/43 89/11 17/83 72.7/24/3.3 52/48

36/64 42/58 92/8 99/1/0 48/52

88/12 58/42 21/79 71/18/12 37/63

67/33 91/9 84/16 95/5/0 NA

19/81 47/53 19/81 NA 47/53

55/45 77/23 38/62 93/7/1 44/56

39/61 68/32 20/80 64/25/12 NA

40/60 57/43 17/83 34/40/26 NA

Cox regression

Linear regression, Cox regression

Cox regression

Cox regression

Cox proportional hazard model

Cox proportional hazard model

Cox regression

Cox regression

Cox proportional hazard model

Initial GFR, MAP, proteinuria

Initial GFR, MAP, proteinuria; initial GFR & F/U MAP, proteinuria

Initial proteinuria

Initial GFR, MAP, proteinuria, therapy (steroid or not)

Time-averaged proteinuria

Age, sex, eGFR

Initial GFR, MAP, proteinuria

Initial GFR, MAP, proteinuria

Initial GFR, MAP, proteinuria, therapy (steroid or not), age, sex

a

d

b

e

Number of centers. Grams of protein per gram of creatinine. c Percentage number indicates CKD stage.

effect of immunosuppressive therapy in IgAN and not because of their predictive value.10 Studies from the North America Cohort16 and Peking University Chinese Cohort23 also found E lesions to be pathologic indicators for immunosuppressive therapy. However, overall, we did not confirm this in the present study. We found a strong interaction between immunosuppressive therapy and E lesions (an almost linear association) in small cohorts (Fig S1), but not in large studies. The European Validation Study of the Oxford Classification of IgAN (VALIGA) aims to determine the most useful treatAm J Kidney Dis. 2013;62(5):891-899

Either presence of hypertension or use of antihypertensive medications. Steroid. All analyses were multivariate.

f

ment for each lesion and the “point of no return” at which no therapy is effective. We anticipate that the role of E lesions will be clarified in VALIGA. The Oxford Classification cohort excluded patients with severely decreased kidney function or mild proteinuria.10 Also, it did not include patients with crescentic IgAN with rapidly progressive renal deterioration. In the present systematic review, we included a wide range of cases of IgAN, including patients who would have met the exclusion criteria of the Oxford cohort, and found the results to be consistent with the original Oxford study. This suggests 895

Lv et al

Figure 2. Hazard ratios of kidney failure for patients with M0 vs M1 (mesangial hypercellularity score ⱕ0.5 vs ⬎0.5).

Figure 3. Hazard ratios of kidney failure for patients with versus without endocapillary hypercellularity (E).

Figure 4. Subgroup analyses of the hazard ratios of kidney failure for patients with versus without endocapillary hypercellularity (E). 896

Am J Kidney Dis. 2013;62(5):891-899

Oxford Classification of IgAN Table 2. Univariate Metaregression Analysis of Possible Sources of Heterogeneity Across Studies Possible Source of Heterogeneity

No. of patients No. of events Duration of follow-up Ethnicity Age Immunosuppressive agents

Scale

Change in HRa

95% CI

P

Every 100-patient increase Every 10-event increase Every 1-y increase Asian (vs non-Asian) Child/adolescent (vs adult) Every 10% increase

0.92 0.94 0.98 0.38 1.55 0.82

0.80-1.05 0.85-1.04 0.80-1.19 0.10-1.39 0.38-6.35 0.56-1.19

0.2 0.2 0.8 0.1 0.5 0.3

Abbreviations: CI, confidence interval; HR, hazard ratio. Proportional change in HR. Results presented with exponentiated regression coefficients and their 95% CIs, showing the proportional change in risk ratio for the first listed factor compared with that of the second listed factor or for every scale increase in each factor. a

its utility for the whole IgAN population. However, we could not obtain individual patient data for the present study or formally validate the 4 Oxford Classification pathologic features in populations outside the Oxford cohort. Further studies are needed to evaluate the Oxford Classification in such populations. In the Oxford cohort, few patients with crescents were included the analysis. Therefore, the predictive value of C lesions for kidney failure was not determined. In the present analysis, we included more patients with crescentic IgAN. The pooled results show that compared with those without such lesions, patients with crescents had a 2.3-fold higher risk of kidney failure, with no evidence of heterogeneity across the included studies. Thus, our study supports the addition of crescent lesions to the Oxford Classification. This systematic review and meta-analysis has several strengths: the rigorous methodology, the importance of the clinical question, and the inclusion of a wide range of IgAN populations. The

number of patients was more than 10 times that of the original Oxford cohort and supports the use of M, S, T, and C lesions in the classification of IgAN. The major limitation is that all the included studies were retrospective; also, this was not an individualpatient-data meta-analysis. Thus, we could not evaluate the interaction between active lesions such as crescents or mesangial proliferative lesions and immunosuppressive therapy.29,30 In none of the studies was therapy included in the multivariate regression model. Another limitation is that kidney disease outcomes in the included studies were not consistent, and only 4 studies14,16,26,28 used creatinine slope and/or renal survival as in the Oxford Classification cohort; thus, we could not validate the Oxford Classification for eGFR decline rate. Finally, the quality of the included studies was variable. There currently is no generic robust tool for evaluating risk of bias in nonrandomized studies; thus, we could not evaluate the effect of study quality on the pooled results. There also is the possibility of publication bias because some studies with negative results may not be published

Figure 5. Hazard ratios of kidney failure for patients with versus without segmental glomerulosclerosis/adhesion (S). Am J Kidney Dis. 2013;62(5):891-899

897

Lv et al

Figure 6. Hazard ratios for kidney failure for patients with versus without various levels of tubular atrophy and interstitial fibrosis (T).

Figure 7. Hazard ratios for kidney failure for patients with versus without cellular/fibrocellular crescents (C).

or are reported only at meetings. For this reason, we searched conference proceedings and consulted experts in the field to try to identify unpublished studies. In conclusion, our study confirms that M, S, and T (but not E) of the Oxford Classification, together with crescent lesions, are associated strongly with progression to kidney failure. Our findings support the use of M, S, T, and C in classifying IgAN in clinical practice.

SUPPLEMENTARY MATERIAL Figure S1: Univariate meta-regression exploring interaction of E lesions and immunosuppressive therapy. Items S1: Search strategy. Note: The supplementary material accompanying this article (http://dx.doi.org/10.1053/j.ajkd.2013.04.021) is available at www.ajkd.org.

REFERENCES ACKNOWLEDGEMENTS Support: This study was supported by grants from the Major State Basic Research Development Program of China (973 program, 2012CB517700), the National Natural Science Foundation of China (81270795 and 30825021), the Natural Science Fund of China to the Innovation Research Group (81021004), and the Program for New Century Excellent Talents in University from the Ministry of Education of China (NCET-12-0011). Financial Disclosure: The authors declare that they have no other relevant financial interests. 898

1. Lee SM, Rao VM, Franklin WA, et al. IgA nephropathy: morphologic predictors of progressive renal disease. Hum Pathol. 1982;13:314-322. 2. Haas M. Histologic subclassification of IgAnephropathy: a clinicopathologic study of 244 cases. Am J Kidney Dis. 1997;29:829-842. 3. D’Amico G. Natural history of idiopathic IgA nephropathy and factors predictive of disease outcome. Semin Nephrol. 2004;24: 179-196. 4. Roufosse CA, Cook HT. Pathological predictors of prognosis in immunoglobulin A nephropathy: a review. Curr Opin Nephrol Hypertens. 2009;18:212-219. Am J Kidney Dis. 2013;62(5):891-899

Oxford Classification of IgAN 5. D’Amico G. Natural history of idiopathic IgA nephropathy: role of clinical and histological prognostic factors. Am J Kidney Dis. 2000;36:227-237. 6. Katafuchi R, Kiyoshi Y, Oh Y, et al. Glomerular score as a prognosticator in IgA nephropathy: its usefulness and limitation. Clin Nephrol. 1998;49:1-8. 7. Manno C, Strippoli GF, D’Altri C, Torres D, Rossini M, Schena FP. A novel simpler histological classification for renal survival in IgA nephropathy: a retrospective study. Am J Kidney Dis. 2007;49:763-775. 8. Lee HS, Lee MS, Lee SM, et al. Histological grading of IgA nephropathy predicting renal outcome: revisiting H.S. Lee’s glomerular grading system. Nephrol Dial Transplant. 2005;20:342348. 9. Roberts IS, Cook HT, Troyanov S, et al. The Oxford Classification of IgA Nephropathy: pathology definitions, correlations, and reproducibility. Kidney Int. 2009;76:546-556. 10. Cattran DC, Coppo R, Cook HT, et al. The Oxford Classification of IgA Nephropathy: rationale, clinicopathological correlations, and classification. Kidney Int. 2009;76:534-545. 11. Alamartine E, Sauron C, Laurent B, Sury A, Seffert A, Mariat C. The use of the Oxford Classification of IgA Nephropathy to predict renal survival. Clin J Am Soc Nephrol. 2011;6:23842388. 12. Coppo R, Troyanov S, Camilla R, et al. The Oxford IgA nephropathy clinicopathological classification is valid for children as well as adults. Kidney Int. 2010;77:921-927. 13. Edstrom Halling S, Soderberg MP, Berg UB. Predictors of outcome in paediatric IgA nephropathy with regard to clinical and histopathological variables (Oxford Classification). Nephrol Dial Transplant. 2012;27:715-722. 14. El Karoui K, Hill GS, Karras A, et al. Focal segmental glomerulosclerosis plays a major role in the progression of IgA nephropathy. II. Light microscopic and clinical studies. Kidney Int. 2011;79:643-654. 15. Gutierrez E, Zamora I, Ballarin JA, et al. Long-term outcomes of IgA nephropathy presenting with minimal or no proteinuria. J Am Soc Nephrol. 2012;23:1753-1760. 16. Herzenberg AM, Fogo AB, Reich HN, et al. Validation of the Oxford Classification of IgA Nephropathy. Kidney Int. 2011;80: 310-317. 17. Kang SH, Choi SR, Park HS, et al. The Oxford Classification as a predictor of prognosis in patients with IgA nephropathy. Nephrol Dial Transplant. 2012;27:252-258. 18. Katafuchi R, Ninomiya T, Nagata M, Mitsuiki K, Hirakata H. Validation study of Oxford Classification of IgA Nephropathy:

Am J Kidney Dis. 2013;62(5):891-899

the significance of extracapillary proliferation. Clin J Am Soc Nephrol. 2011;6:2806-2813. 19. Kataoka H, Ohara M, Shibui K, et al. Overweight and obesity accelerate the progression of IgA nephropathy: prognostic utility of a combination of BMI and histopathological parameters. Clin Exp Nephrol. 2012;16:706-712. 20. Le W, Zeng CH, Liu Z, et al. Validation of the Oxford Classification of IgA Nephropathy for pediatric patients from China. BMC Nephrol. 2012;13:158. 21. Lee H, Yi SH, Seo MS, et al. Validation of the Oxford Classification of IgA Nephropathy: a single-center study in Korean adults. Korean J Intern Med. 2012;27:293-300. 22. Moriyama T, Nakayama K, Iwasaki C, et al. Severity of nephrotic IgA nephropathy according to the Oxford Classification. Int Urol Nephrol. 2012;44:1177-1184. 23. 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: 2175-2184. 24. Shima Y, Nakanishi K, Hama T, et al. Validity of the Oxford Classification of IgA Nephropathy in children. Pediatr Nephrol. 2012;27:783-792. 25. Yau T, Korbet SM, Schwartz MM, Cimbaluk DJ. The Oxford Classification of IgA Nephropathy: a retrospective analysis. Am J Nephrol. 2011;34:435-444. 26. 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. 27. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. Ann Intern Med. 2009;151:W65-W94. 28. Joh K, Hashiguchi A, Okonogi H, Miyazaki Y, Utsunomiya Y, Kawamura T. Histological evaluation on IgA nephropathy in 233 Japanese adult patients: retrospective analysis [abstract; poster no. 12]. Oral and poster presentation in World Congress of Nephrology 2009 Satellite Symposium on IgA Nephropathy; May 22-26, 2009; Milan, Italy. 29. Tumlin JA, Lohavichan V, Hennigar R. Crescentic, proliferative IgA nephropathy: clinical and histological response to methylprednisolone and intravenous cyclophosphamide. Nephrol Dial Transplant. 2003;18:1321-1329. 30. Shoji T, Nakanishi I, Suzuki A, et al. Early treatment with corticosteroids ameliorates proteinuria, proliferative lesions, and mesangial phenotypic modulation in adult diffuse proliferative IgA nephropathy. Am J Kidney Dis. 2000;35:194-201.

899