Classifying and predicting outcomes in ANCA-associated glomerulonephritis

Abstract

Histopathological classification systems have been applied to a number of renal conditions, to allow clinicians to make useful predictions regarding patient and kidney outcome and to aid in comparing similar lesions across clinical trials and cohorts. This has been successful for lupus nephritis, focal segmental glomerulosclerosis and more recently for both IgA disease (the Oxford classification) [1] and anti-neutrophil cytoplasm antibody (ANCA)-associated vasculitis (AAV) [the Berden or European vasculitis study group (EUVAS) classification] [2]. In IgA disease, numerous follow-up studies have now been published validating the classification, in different age groups and across different countries and ethnicities. The paper in Nephrology Dialysis Transplantation by Chang et al. [3] attempts to re-evaluate the Berden/EUVAS ANCA-associated glomerulonephritis classification in a Chinese population, and it therefore represents an important first step in assessing the general utility of the classification. AAV frequently causes a focal segmental necrotizing glomerulonephritis, and renal involvement exerts a profound influence on patient morbidity and mortality. Indeed, disease severity in AAV is mostly assessed by the severity of renal involvement. A significant proportion of patients present with advanced renal failure and require dialysis. With modern treatment protocols, ~50–60% of those dialysis-dependant patients who survive the first year recover independent renal function. Previous studies by EUVAS investigators had demonstrated the prognostic value of the renal biopsy by defining lesions that were associated with favourable renal outcome and treatment response. They demonstrated that those patients recovering independent renal function had less glomerulosclerosis, arteriosclerosis and tubular atrophy while they were more likely to have received plasmaphaeresis as adjunctive treatment [4]. Additionally, by combining presenting renal function measurement (estimated glomerular filtration rate; eGFR0) with histological parameters, they devised better predictors of renal function at 18 months than eGFR alone [5]. They found that the most inflammatory glomerular lesions, containing fibrinoid necrosis or cellular crescents, were the ones most likely to respond to treatment, being associated with the greatest improvement in renal function following 18 months of follow-up. By contrast, established sclerotic glomerular lesions were most predictive of final renal function at 18 months. These findings relied on assessment of cohorts who were treated in a relatively homogeneous manner (recruited from various EUVAS clinical trials), with pathologists who had demonstrated high degrees of concordance in defining the histopathological lesions [6]. However, despite the prognostic value of the biopsy, a formal histological classification system had been lacking, until recently. The Berden/EUVAS histopathological classification of AAV was proposed in 2010 [2] by an international group of renal pathologists and was validated in 100 patients from two large multicentre European vasculitis trials (CYCAZAREM and MEPEX, which recruited patients with ANCA-associated glomerulonephritis of differing severities). Renal histology from patients originating in nine European countries, with at least 1-year follow-up, and including both MPO-ANCA and PR3-ANCA were analysed. Importantly, patients with other co-morbid conditions, such as concurrent anti-glomerular basement membrane disease, were excluded. The classification system was based on glomerular pathology assessed by light microscopy. Interestingly, tubulointerstitial fibrosis and tubular atrophy, generally predictive of long-term renal outcomes for many diseases, did not add any prognostic value and were thus excluded. The key aspect of the classification was that it was simple and was defined by four separate glomerular categories: focal, crescentic, mixed and sclerotic. Biopsies in the focal category were defined by having 50% normal glomeruli. Crescentic biopsies were those with 50% glomeruli with cellular crescents. Mixed referred to biopsies having <50% normal, <50% crescentic and <50% globally sclerotic glomeruli. The last category was sclerotic, which included biopsies with 50% globally sclerotic glomeruli. The biopsies included in this classification were all pauci-immune and contained at least 10 glomeruli, which are clearly important standards for applying the classification. The validation study was able to demonstrate that the different classes correlated with the