The Oxford classification of IgAN provides a histopathological grading system for prediction of renal prognosis of IgAN independent of the clinical features
[8, 9]. However, the classification must be validated in the different cohorts of patients. This study was designed, using similar methods as in the Oxford study, to assess the validity of the new Oxford classification of IgAN in a multi-center cohort of pediatric patients from China. The clinical characteristics in our cohort were very similar to the pediatric patients in original Oxford cohort (Table
1). Our study shows that tubular interstitial fibrosis was the only pathological feature independently associated with renal outcomes in Chinese children with IgAN.
It is remarkable to notice that lesion M, E, C and N, which were all thought to be active glomerular lesions in patient with IgAN, were not independently associated with renal outcome in our study. The similar results were also showed in another validation study in 1026 Chinese adult patients
. Moreover, the prognostic values of M, E and C were also controversial in different validation studies
[8, 15–18, 20, 22–26]. The lesion E and C were also not statistically associated with renal outcome in the original Oxford cohorts
. The prognostic value of necrosis was not evaluated in the original Oxford study, as only six cases (2.3%) had this lesion in that cohort of patients. Compared with patients in the Oxford study, there were significantly more patients with necrosis (16%) in this study, however, an association between necrosis and renal outcome was not established, and similar results were also found in two other studies
[24, 27]. Taken together, those results indicate that there are only weak associations between present of these acute lesions (M, E, C, N) and renal outcome. Several possible explanations may account for these results. Firstly, those acute glomerular lesions only reflect the disease activity at the time of renal biopsy, and all of them are reversible after immunosuppressive treatment
. Secondly, the ration of glomeruli with these lesions is very important in patients with IgAN, as most of the patients have only small numbers of crescents in our study and similar finding were also showed in other studies
[8, 17, 21]. Shima et. al
 found that only those patients with C > 30% or E > 30% were associated with an unfavorable renal outcome in children with IgAN, indicating that the optimal cutoff ratios of these acute lesions for predicting a worse outcome should be determined in IgAN in the future. Thirdly, the inconsistent results among those validation studies may due to different inclusion criteria, as shown by Katafuchi et al.
 and Shima et al.
. Finally, the lack of predictive value of this lesion may reflect inadequate statistical power, as only a small subset of patients developed ESRD or 50% decline in GFR during the follow-up in most validation studies, including the current study.
Recently, two studies about validation of the Oxford classification for pediatric IgA nephropathy were published from Japan and Sweden respectively. The most obvious difference between our study and the two previous studies is lesion S. Both Shima et. al
 Edström et. al
 found that present of lesion didn’t indicate a poor prognosis in IgAN. In the present study, present of lesion S were showed to be significantly associated with renal outcome in univariate COX analysis, but it failed to attain independent significance in multivariate model. A similar predictive value was shown between S and S-alone in this cohort. This may due to the different health screening practice and inclusion criteria, various treatments during follow-up, and especially the poor reproducibility (ICC) of lesion S. Children in this study have more severe proteinuria at biopsy, and received more RASB and immunosuppressive therapy during follow-up than children from Japan
. Given that the ICC of adhesion was poor (0.2) in the original Oxford study, the frequency distribution of S was also remarkably different among the validation studies. Taken together, these findings indicate that lesion S seems to had a weak influence on renal survival.
One of the most exciting findings in the new Oxford IgAN classification, is the question of whether this classification can predict optimal treatment for patients with IgAN. The original Oxford study showed that, in patients who received no immunosuppression, the rate of renal function decline in those with E was faster than those without, while there was no such difference in patients treated with immunosuppression. Hence, the lesion E was finally involved in the Oxford classification, and this provided indirect evidence that lesion E is assumed responsive to immunosuppressive therapy. The similar indirect evidence was also shown in a validation study from four centers in North America
. However, in the current cohort of patients, we do not confirm these findings. Whether lesion C, E, and N can predict optimal treatment for patients with IgAN remains unclear, and prospective clinicopathological studies are needed to investigate this possibility.
IgAN is defined as dominant or codominant staining with IgA in glomeruli by immunofluorescence or immunoperoxidase
. It is important to note that it may in fact simply define a group of diseases sharing identical histopathologic sequelae
. If that is the case, a great limitation of this histopathological classification should be recognized, for an ideal classification system should be based on pathogenic mechanisms and should thus suggest an appropriate therapeutic strategy. The classification of IgAN should also be improved based on the biomarker of pathogenic mechanisms in the future.