IgA nephropathy is the most common glomerular disease worldwide, yet there is no international consensus for its pathological or clinical classification. Here a new classification for IgA nephropathy is presented by an international consensus working group. The goal of this new system was to identify specific pathological features that more accurately predict risk of progression of renal disease in IgA nephropathy, thus enabling both clinicians and pathologists to improve individual patient prognostication. In a retrospective analysis, sequential clinical data were obtained on 265 adults and children with IgA nephropathy who were followed for a median of 5 years. Renal biopsies from all patients were scored by pathologists blinded to the clinical data for pathological variables identified as reproducible by an iterative process. Four of these variables: (1) the mesangial hypercellularity score, (2) segmental glomerulosclerosis, (3) endocapillary hypercellularity, and (4) tubular atrophy/interstitial fibrosis were subsequently shown to have independent value in predicting renal outcome. These specific pathological features withstood rigorous statistical analysis even after taking into account all clinical indicators available at the time of biopsy as well as during follow-up. The features have prognostic significance and we recommended they be taken into account for predicting outcome independent of the clinical features both at the time of presentation and during follow-up. The value of crescents was not addressed due to their low prevalence in the enrolled cohort.

Pathological classifications in current use for the assessment of glomerular disease have been typically opinion-based and built on the expert assumptions of renal pathologists about lesions historically thought to be relevant to prognosis. Here we develop a unique approach for the pathological classification of a glomerular disease, IgA nephropathy, in which renal pathologists first undertook extensive iterative work to define pathologic variables with acceptable inter-observer reproducibility. Where groups of such features closely correlated, variables were further selected on the basis of least susceptibility to sampling error and ease of scoring in routine practice. This process identified six pathologic variables that could then be used to interrogate prognostic significance independent of the clinical data in IgA nephropathy (described in the accompanying article). These variables were (1) mesangial cellularity score; percentage of glomeruli showing (2) segmental sclerosis, (3) endocapillary hypercellularity, or (4) cellular/fibrocellular crescents; (5) percentage of interstitial fibrosis/tubular atrophy; and finally (6) arteriosclerosis score. Results for interobserver reproducibility of individual pathological features are likely applicable to other glomerulonephritides, but it is not known if the correlations between variables depend on the specific type of glomerular pathobiology. Variables identified in this study withstood rigorous pathology review and statistical testing and we recommend that they become a necessary part of pathology reports for IgA nephropathy. Our methodology, translating a strong evidence-based dataset into a working format, is a model for developing classifications of other types of renal disease.

The Oxford Classification of IgA nephropathy (IgAN) includes the following four histologic components: mesangial (M) and endocapillary (E) hypercellularity, segmental sclerosis (S) and interstitial fibrosis/tubular atrophy (T). These combine to form the MEST score and are independently associated with renal outcome. Current prediction and risk stratification in IgAN requires clinical data over 2 years of follow-up. Using modern prediction tools, we examined whether combining MEST with cross-sectional clinical data at biopsy provides earlier risk prediction in IgAN than current best methods that use 2 years of follow-up data. We used a cohort of 901 adults with IgAN from the Oxford derivation and North American validation studies and the VALIGA study followed for a median of 5.6 years to analyze the primary outcome (50% decrease in eGFR or ESRD) using Cox regression models. Covariates of clinical data at biopsy (eGFR, proteinuria, MAP) with or without MEST, and then 2-year clinical data alone (2-year average of proteinuria/MAP, eGFR at biopsy) were considered. There was significant improvement in prediction by adding MEST to clinical data at biopsy. The combination predicted the outcome as well as the 2-year clinical data alone, with comparable calibration curves. This effect did not change in subgroups treated or not with RAS blockade or immunosuppression. Thus, combining the MEST score with cross-sectional clinical data at biopsy provides earlier risk prediction in IgAN than our current best methods.

To study the predictive value of biopsy lesions in IgA nephropathy in a range of patient ages we retrospectively analyzed the cohort that was used to derive a new classification system for IgA nephropathy. A total of 206 adults and 59 children with proteinuria over 0.5 g/24 h/1.73 m(2) and an eGFR of stage-3 or better were followed for a median of 69 months. At the time of biopsy, compared with adults children had a more frequent history of macroscopic hematuria, lower adjusted blood pressure, and higher eGFR but similar proteinuria. Although their outcome was similar to that of adults, children had received more immunosuppressants and achieved a lower follow-up proteinuria. Renal biopsies were scored for variables identified by an iterative process as reproducible and independent of other lesions. Compared with adults, children had significantly more mesangial and endocapillary hypercellularity, and less segmental glomerulosclerosis and tubulointerstitial damage, the four variables previously identified to predict outcome independent of clinical assessment. Despite these differences, our study found that the cross-sectional correlation between pathology and proteinuria was similar in adults and children. The predictive value of each specific lesion on the rate of decline of renal function or renal survival in IgA nephropathy was not different between children and adults.

These findings support the hypothesis that hyperglycemia evokes an intrinsic pathway of proapoptotic signaling in mesangial cells. In addition, these results point to an important role for the intrinsic pathway in microvascular injury in the diabetic kidney in vivo.