Monday, December 11, 2017

Oxford Classification

Discussions which led to the development of the Oxford Classification of IgA Nephropathy began in 2004, eventually leading to the publication in Kidney International of two papers describing the Oxford Classification in September 2009[1].  This classification was the result of a collaboration between the International IgA Nephropathy Network and the Renal Pathology Society. 

Background

The need for a new classification for IgA nephropathy was discussed by John Feehally (nephrologist – Leicester, UK) and Terry Cook (renal pathologist – London, UK) in 2004.  Since the first description of IgA nephropathy in 1968, at least 15 different pathological classifications had been proposed in the literature, and several were still in use.  Furthermore there was no consensus about the preferred classification.  A variety of approaches had been used in developing these classifications and head to head comparisons between them were few.  There was enthusiasm among nephrologists for the development of a consensus classification with the goals of (a) improving accuracy of prognosis for individual patients; (b) allowing improved comparison of patterns and outcomes of IgA nephropathy in different populations; (c) refining recruitment into clinical trials of new therapies. 

In 2006 the Renal Pathology Society undertook a survey of its members.  Only 37% of renal pathologists who responded used a pathological classification in reporting renal biopsies for IgA nephropathy.  Five different classifications were being used, most commonly the Haas classification, but that was only used by 14% of all respondents. However 94% agreed that a more universal pathological classification system for IgA nephropathy would be worthwhile, as long as it was proven to have clinical utility.

Planning

A low cost approach was necessary if the classification was to be developed and published in a timely way.  Funding was obtained from the International Society of Nephrology, Kidney Research UK, & through an unrestricted educational grant from Vifor Aspreva Pharma.  Short planning meetings were held during ASN Kidney Week from 2005 to 2008.  Two three-day meetings were held at Magdalen College, Oxford, UK in 2006 and 2008.  A furhter face-to fce meeting for the pathologists on the working group was held in 2007.

A small planning committee co-chaired by John Feehally and Terry Cook identified a working group of 14 nephrologists and 8 renal pathologists with the necessary expertise and commitment (Table 1).

Table 1. Members of the Oxford Classification Working Group

Pathologists

Nephrologists

Charles Alpers (USA)

Stephen Bonsib (USA)

Jan Bruijn  (Netherlands)

Terry Cook (UK)

Vivette D’Agati  (USA)

Steven Emancipator (USA)

Franco Ferrario (Italy)

Sandrine Florquin (Netherlands)

Agnes Fogo (USA)              

Mark Haas (USA)

Andrew Hertzenberg (Canada)

Prue Hill (Australia)

Charles Jennette (USA)

Kensuke Joh  (Japan)

Fernand Lai (Hong Kong)

Ian Roberts (UK)

Patrick Walker (USA)

Jan Weening (Netherlands)

Jonathan Barratt (UK)

Francois Berthoux (France)

Dan Cattran  (Canada)

Rosanna Coppo (Italy)

Guiseppe D’Amico (Italy)

John Feehally (UK)

Ron Hogg (USA)

Stephen Hsu (USA)

Tetsuya Kawamura (Japan)

CB Leung (Hong Kong)

Philip Li (Hong Kong)

Stephan Troyanov (Canada)

Paolo Schena (Italy)

Yasuhiko Tomino (Japan)  

The working group was chosen to represent both adult and paediatric specialists, from the widest possible range of countries and continents. The principles agreed were that a new classification of IgA nephropathy must be evidence based, simple, reproducible, clinically relevant, and applicable to the widest possible range of subjects by age, gender, geography and race.

Work Programme

The development was divided into five phases

  1. Agreeing definitions of the histological features seen on light microscopy
  2. Testing among pathologists the reproducibility of each of those lesions
  3. Collecting a large cohort of patients with a dataset including clinical features at time of biopsy, treatments given, clinical outcome during follow up.
  4. Seeking clinicopathological correlations between pathological features which were proven to be reproducible and independent, and clinical features
  5. Proposing a classification.

It was agreed that the problem must be approached with an open mind – evaluating the full range of pathological features, and not restricting only to those expected to be markers of poor outcome.  The working group recognised the known usefulness of clinical parameters in predicting outcome, especially time-average proteinuria during follow up, and blood pressure.  It would therefore be necessary to demonstrate that any pathological features used would have added predictive value even when clinical features at presentation and during follow up were known. 

1. Agreeing definitions of the histological features seen on light microscopy

Definitions were agreed for IgA nephropathy (based on diffuse mesangial IgA deposition) and for 20 glomerular, tubulo-interstitial, and vascular features seen in IgA nephropathy biopsies. 

2. Testing among pathologists the reproducibility of each of those lesions

These definitions were tested by circulating 40 renal biopsies among the pathologists, each biopsy being scored independently by four pathologists, in order to resolve inconsistencies and confirm reproducibility.  This work confirmed that there were only six pathology features that were both reproducible and independent from each other. 

3. Collecting a large cohort of patients with a dataset including clinical features at time of biopsy, treatments given, clinical outcome during follow up.

A cohort of 265 cases was assembled from  Asia (62), Europe (94), and North America (109);  59 of the cases were children at the time of diagnosis.  A clinical cohort was deliberately selected which represented typical slowly progressive IgA nephropathy.  Those  with urine protein <1g/24 hours (or equivalent in children) at presentation were excluded to avoid the risk that inclusion of large numbers of low risk cases would make it harder to identify the predictive power of pathological features. Those with eGFR < 30ml/min were also excluded with the goal of preventing inclusion of those with very advanced slowly progressive disease; but with the disadvantage that many of the most rapidly progressive cases would also be excluded.

4. Seeking clinicopathological correlations between pathological features which were proven to be reproducible and independent, and clinical features

There were only four independent reproducible pathological features which predicted outcome in addition to available clinical information.  Furthermore, each of these four pathological features could be scored in a way which was straightforward and workable in routine clinical practice.  The four features were mesangial hypercellularity (M0 or M1 in > or < 50% of glomeruli); endocapillary hypercellularity (present or absent, E1 or E0); segmental sclerosis and adhesions (present or absent S1 or S0); tubular atrophy/interstitial fibrosis 0-25% (T0), 25-50% (T1), >50% (T2).

5. Proposing a classification

A descriptive classification was therefore proposed, recommending that the scores for all four features – M, E, S, T – should be included in routine reporting of renal biopsies in cases of IgA nephropathy in addition to the usual verbal report of the findings.

The Success of the Working Group

This project was successful because it brought together in collaborative partnership the key opinion leaders in IgA nephropathy and glomerular pathology.  The relationships and mutual respect which this work generated will give the opportunity for future project work.  The commitment of each individual member of the working group, and the low cost approach to the planning of each face-to-face meeting, resulted in a total  development and publication cost of only $75,000. 

Implications of the Classification

The approach taken in the development of the Oxford Classification of IgA Nephropathy represents a new standard for an evidence-based approach which has potential applicability for the classification of other glomerular diseases. 

The development of the Oxford Classification does not represent the completion of the task, but the beginning of a continuing process which will be needed for a truly universal clinico-pathological classification of IgA nephropathy being achieved.

  1. The published classification only describes pathological features which are predictive of outcome in addition to clinical information.  A next important stage is to develop a scoring system which combines pathological features and clinical features into a more precise risk score; perhaps with the addition of other key laboratory features.   
  1. The Oxford Classification was developed in a defined and restricted cohort of patients.  This was a deliberate restriction in order to maximise information about typical slowly progressive IgA nephropathy.  However, the classification needs further study in patients at lower risk of progression and also those at high risk of rapid progression.  It is this restriction which probably explains the unexpected finding that crescents were not significantly associated with outcome in the studied cohort.
  1. The Oxford Classification was based on a single cohort of patients.  It would therefore be very important to check the predictive value of the classification in other cohorts of patients retrospectively, and also to study other cohorts prospectively. 

Work since the Oxford Classification was published

In the three years up to July 2012, a further 12 studies have been published, or presented in abstract form, which have applied  the Oxford Classification in other cohorts of patients from the same countries as were represented in the original study, and also from other countries (Table 2).  Some of these cohorts have included broader inclusion criteria.  The largest of these studies (VALIGA) in Europe, which will report soon, has studied >1000 patients.

Table 2. Published studies assessing the predictive value of  the Oxford Classification of IgA nephropathy

Country of Origin

Reference

Cohort size

Multinational 2009

Working Group of the International IgA Nephropathy Network and the Renal Pathology Society.  

Kidney Int  2009; 76:534-45

265 adults & children

Multinational 2010

Working Group of the International IgA Nephropathy Network and the Renal Pathology Society.

Kidney Int 2010;77:921-7

59 children

Macedonia  2010

Grcevska L et al.

Prilozi   2010;31:7-16

40 adults

France  2011

Alamartine E et al.

Clin J Am Soc Nephrol 2011; 6:2384-8

183 adults

USA, Canada  2011

Herzenberg AM et al.

Kidney Int 2011; 80: 310-7

187 adults & children

China 2011

Shi SF et al.

Clin J Am Soc Nephrol  2011; 6:2175-84

410 adults

Japan 2011

Katafuchi R et al.

Clin J Am Soc Nephrol 2011; 6: 2806-13

702 adults

USA 2012

Yau T et al.

Am J Nephrol 2011;34(5):435-44

54 adults

Korea 2012

Kang SH et al. 

Nephrol Dial Transplant  2012; 27:252-8

197 adults

Sweden 2012

Edström Halling S et al.

Nephrol Dial Transplant 2012;27:715-22

99 children

Japan 2012

Shima Y et al.

Pediatr Nephrol 2012; 27:783-92

161 children

China 2012

Zeng CH et al.

Am J Kidney Dis 2012 Jul 19. [Epub ahead of print]

1,062 adults

Next Steps

The energy with which these studies have been undertaken by the IgA nephropathy community internationally, indicates that there is continuing enthusiasm for the further refinement and development of the Oxford Classification.  A meeting is being planned by IIgANN in Oxford, UK in June 2014 (more details will be posted on the IIgANN website in due course).  The goal of this meeting will be to present the full range of validation studies for IgA nephropathy, discuss further refinements of the classification which are then required, and to agree further studies which should be planned and initiated.


[1] 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 Sep;76(5):534-45

 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 Sep;76(5):546-56

Join Us

Become a member of the IgAN today. Expand your network and collaborate blank

Contact Us

For more information about our organization, please contact us here

IgA Nephropathy News

Stay abreast of news in the world of IgA Nephropathy research. Click here.