Validation of the revised Oxford classification for IgA nephropathy considering treatment with corticosteroids/immunosuppressors

The Oxford classification for IgA nephropathy (IgAN) was updated in 2017. We have validated the revised Oxford classification considering treatment with corticosteroids/immunosuppressors. In this retrospective analysis, 871 IgAN patients were enrolled. Patients were divided into two groups, those treated with or without corticosteroids/immunosuppressors. The 20-year renal prognosis up to end-stage renal disease was assessed using the Oxford classification. In all patients, the renal survival rate was 87.5% at 10 years and 72.6% at 20 years. The T score alone was significantly related to renal prognosis in the Kaplan–Meier analysis and multivariate Cox regression analysis. In the non-treatment group (n = 445), E, S, T, and C scores were significantly related to renal survival rates, however, in the treatment group (n = 426), T score alone was significantly related to renal prognosis on Kaplan–Meier analysis, indicating that corticosteroids/immunosuppressors improved renal prognosis in E1, S1, and C1. In patients with E1, S1, or C1, the treatment group showed significantly better renal prognosis than the non-treatment group in univariate and multivariate analysis. The Oxford classification and T score were used to determine renal prognosis in IgAN patients. Corticosteroids/immunosuppressors improved renal prognosis, especially E1, S1, and C1 scores.


Results
Clinical and histological findings, initial treatment, and prognosis in all IgAN patients. The study group included 871 patients with IgAN, who had > 8 glomeruli and were observed over a period of ≥ 1 year. The renal prognosis was evaluated using the revised Oxford classification and compared between patients treated with and without corticosteroids/immunosuppressants. The baseline data of all patients are shown in Table 1a. The median age was 31.0 years, and there were 356 (40.9%) male and 515 (59.1%) female patients. The median systolic blood pressure (SBP) was 120.0 mmHg, and the median diastolic blood pressure (DBP) was 74.0 mmHg. The median duration of follow-up was 8.0 years. Regarding the laboratory findings, the median eGFR was 77.0 mL/min/1.73 m 2 , and the median urinary protein excretion (U-Prot) was 0.68 g/day. Notably, in our study group, 11 patients (1.2%) had an eGFR < 30 mL/min/1.73 m 2 , with 323 patients (37.0%) having a U-Prot level < 0.5 g/day. Histological findings were as follows: 49.4% had M1, 44.9% had E1, 72.0% had S1, 21.7/5.9% had T1/T2, and 45.3/5.3% had C1/C2. Several major treatments for IgAN were started within 1 year after renal biopsy as the initial treatment (Table 1b); 426 patients (48.9%) were treated with corticosteroids/ immunosuppressors. Among those 426 patients, 424 patients were treated with corticosteroids alone, and 13 patients were treated with corticosteroids and/or other immunosuppressive agents. One-hundred-and-ninetytwo (22.0%) patients underwent tonsillectomy, 293 (33.6%) were treated with renin-angiotensin system (RAS) inhibitors, and 177 (20.3%) were treated with fish oil (Table 1b). One-hundred-and-fifteen patients (13.2%) progressed to ESRD during the follow-up period, and five patients died before reaching ESRD.
The Cox regression multivariate analysis indicated that lower eGFR and higher U-Prot and mean arterial pressure (MAP) were independent risk factors for progression to ESRD in all patients (Table 3). According to the Oxford classification, only the T score was an independent risk factor for progression in all patients (HR, 1.48; 95% CI, 1.10-1.99; p = 0.0085) and the treatment group (HR, 1.76; 95% CI, 1.05-2.92; p = 0.0287).
Comparison of renal prognosis between the treatment and non-treatment group patients in the E1, S1, or S1 categories. Treatment with corticosteroids/immunosuppressors significantly improved the renal prognosis among IgAN patients in the E1, S1, and C1 categories compared to no treatment ( Fig. 4; E1, p = 0.008; S1, p = 0.0064; and C1, p = 0.0014). In the univariate analysis (

Discussion
This report is a validation study of the Oxford analysis of all 871 IgAN patients treated with or without corticosteroids/immunosuppressors. In our analysis, we included IgAN patients with eGFR < 30 mL/min/1.73 m 2 and progression to ESRD within 1 year who were excluded from the original Oxford classification because it seemed important to include rapidly progressive glomerulonephritis (RPGN) and deteriorated cases at the time of renal biopsy to analyse the risk of cellular and fibrocellular crescents, which were newly included in the Oxford classification as C scores. We also included IgAN patients with U-Prot < 0.5 g/day who were excluded from the original Oxford classification because, in Japan, many IgAN patients are found in the relatively early stages of the www.nature.com/scientificreports/ disease by health screening checks; therefore, those mild cases were important to the analysis of IgAN diagnosed in Japan. Moreover, IgAN is generally considered a slowly progressing disease, and those mild cases should be considered in the long-term renal prognosis as in our analysis. Salient features of our study cohort that are important to note include a relatively young age (median: 30 years old), a greater proportion of females (59.1%), well-controlled blood pressure (120/74 mmHg), relatively good renal function (eGFR 77.0 mL/min/1.73 m 2 ), and mild U-Prot (0.68 g/day), including 37.0% of patients with a U-Prot level < 0.5 g/day. The distribution of Oxford classification categories, based on histological findings, were as follows: M1, 49.4%; E1, 44.9%; S1, 72.0%; T1:T2, 21.7%:5.9%; and C1:C2, 45.4%:5.3%. Almost half of the patients (48.9%) were treated using corticosteroids/immunosuppressants, with corticosteroids being used in the majority of patients (424 of 426). In our institution, corticosteroids are generally used in the treatment of IgAN for patients with higher U-Prot and U-RBC levels, stable renal function, and presence of active histological findings; these criteria are reflected in Table 2. The indications for the use of immunosuppressants in the other 13 cases included rapid disease progression, presence of comorbidities, patient's request for a reduction of the dose and/or duration of corticosteroids or the detection of adverse effects of corticosteroid therapy, and the physician's decision. The RAS-I was used in 33.6% of patients. This rate of use of RAS-I does not reflect the current standard treatment for IgAN. This lower than expected percentage of RAS-I use does, however, reflect the extended relevant period for our study, which included patients from as far back as 1974. The first RAS-I (captopril) was used for the treatment of hypertension in the mid-1980s in Japan. We began using RAS-I for the treatment of IgAN at our institution in the 1990s, at a rate of 19.2% up to the year 2000, with this rate having since increased to 44.5%. The lower than expected rate of RAS-I use also reflects the characteristics of our study group, with a relatively young age (median age, 31 years), larger proportion of women than men, and the majority of patients being non-hypertensive (MAP ≤ 100 mmHg, 669 (76.8%) patients, and ≤ 90 mmHg, 461 (52.9%) patients). Therefore, there was no significant indication for the use of RAS-I in our study group. We used ESRD as the endpoint of our study, although 37% of our cohort was relatively early cases of IgAN, with a U-Prot level < 0.5 g/day and slow disease progression. As such, the longer period of observation to ESRD was deemed to be more appropriate for analysis than the change in eGFR that was used for the validation of the Oxford classification for IgAN (VALIGA) in the European multicentre cohort trial 28,29 . Use of the delta eGFR was appropriate in the VALIGA study as patients in that study cohort had more severe IgAN than our cohort and, as such, eGFR is a good predictor of short-term renal survival. www.nature.com/scientificreports/ In all patients, only the T score was an independent risk factor for progression to ESRD in the multivariate Cox regression analysis. In previous reports of validation studies of the Oxford classification 12-26 , the results varied because of differences in clinical backgrounds, treatments, and endpoints; however, the T score was shown to be the most important predictive factor for progression in almost of all those reports. These previous studies are in support of our results. For IgAN patients treated with corticosteroids/immunosuppressors, we found that the T score was an independent risk factor according to the multivariate Cox regression analysis. For IgAN patients without corticosteroids/immunosuppressors treatment, the E, S, T, and C scores were predictive factors for progression to ESRD according to the univariate analysis (Kaplan-Meier analysis and log-rank test). Moreover, among patients with E1, S1, or C1 lesions, renal prognosis was significantly better among those treated with than in those without corticosteroids/immunosuppressors on univariate analysis ( Fig. 4 and Model 1 in Table 4) as well as on multivariate analysis considering the clinical and histological background (Model 2 in Table 4) and treatment (Model 3 in Table 4). These results are indicative of the possibility that corticosteroids/ immunosuppressors can improve E1, S1, and C1 lesions. The renal survival rate of IgAN patients with a C1 score increased slightly more than that of IgAN patients with a C0 score, but it did not increase in IgAN patients with a C2 score. These results indicate that treatment with corticosteroids/immunosuppressors could improve the renal prognosis of IgAN patients with crescents in less than 25% of glomeruli, but it was difficult to improve the prognosis of IgAN patients with crescents in more than 25% of glomeruli. Recent validation studies of the Oxford classification, including the C score, showed varying results. In results from the VALIGA study 29 , for all patients (n = 1,130), the C score was not a predictive risk factor for a 50% decrease in eGFR or ESRD, and was not a predictive risk factor for the eGFR slope according to the multivariate analysis; however, for IgAN patients without corticosteroids/immunosuppressors treatment during follow-up (n = 582), the C score was an independent risk factor for the eGFR slope. These results indicated that corticosteroids/immunosuppressors improve the prognosis of IgAN patients with C scores. Furthermore, a multicentre validation study involving 3,380 IgAN patients performed in Korea indicated that C1 and C2 scores were valid predictive risk factors for progression to ESRD and decreased eGFR according to univariate and multivariate analyses 30 31 . A validation study performed in China involving two centres with 1,152 patients showed Table 2. Comparison of baseline characteristics between patients with or without corticosteroids/ immunosuppressors. BMI body mass index, SBP systolic blood pressure, DBP diastolic blood pressure, MAP mean arterial pressure, TP serum total protein, Cr serum creatinine, eGFR estimated glomerular filtration rate, UA serum uric acid, T-cho serum total cholesterol, TG triglyceride, U-Prot urinary protein excretion, U-RBC urinary red blood cells, HPF high power field, M mesangial hypercellularity, E endocapillary hypercellularity, S segmental sclerosis, T interstitial fibrosis/tubular atrophy, C crescents. www.nature.com/scientificreports/ that the 10-year renal survival rate was not significantly different among patients with C0, C1, and C2 scores, regardless of whether patients were treated with corticosteroids/immunosuppressors. However, in this study, for the patients with nephrotic syndrome, the C score was an independent factor for progression according to the multivariate analysis after adjusting for age, sex, eGFR, MAP, pathological findings, and immunosuppressors 32 . Considering all of these reports and our results, the C score is indicated to be a significant predictive factor for progression to ESRD, and the C1 score was reversible with corticosteroids/immunosuppressors, which is a good indication for this treatment. Interestingly, our results showed that treatment with corticosteroids/immunosuppressors also improved the prognosis of IgAN patients with E1 or S1 score. Endocapillary hypercellularity (E1) was considered as the active lesion leading to the inflammation of capillaries and crescent formation. The beneficial effects of corticosteroids/ immunosuppressors were seen as a reasonable result, like the previous VALIGA study 28 . When segmental sclerosis (S1) was considered as the chronic lesion, it was difficult to obtain a good response using corticosteroids/ immunosuppressors. However, chronic lesions can result from continuous inflammation in the glomeruli; therefore, some IgAN patients with chronic lesions also had active lesions, and immunosuppressors improved their renal outcomes. These results were also shown in a sub-analysis of the STOP-IgAN trial 31 and a sub-analysis 34 of the randomized controlled trial by the IgAN study group in Japan that compared tonsillectomy combined with steroid pulse therapy and steroid pulse monotherapy 33 31 . A randomized controlled trial performed by the IgAN study group indicated that only the S score was an independent factor for the disappearance of both proteinuria and haematuiria by steroid pulse therapy combined with tonsillectomy; however, the M, E, and T scores were not 33,34 . These results were observed during the short-term (only 1 to 3 years); therefore, we propose that our long-term observation study shows clearer results.
This study has some limitations. First, it was performed at a single centre in Japan. Therefore, almost all of the patients in our cohort were Japanese, which means that these results might not apply to other ethnicities than -year renal survival rates were 77.3% for T0, 60.0% for T1, and 29.0% for T2, which were significantly different among the three groups (p < 0.0001). The renal survival rate was significantly higher for T0 than for either T1 (p = 0.0009) or T2 (p < 0.0001). The renal survival rate was significantly higher for T1 than T2 (p = 0.0019). (f) The 20-year renal survival rate was significantly higher for C0 than for C1 + C2 (C0, 73.5%; C1 + C2, 60.3%; p = 0.0228). Note that as there were only 5 patients in the C2 category, this group was combined with the C1 group.  -year renal survival rates were 90.6% for T0, 55.4% for T1, and 51.7% for T2, which were significantly different among the three groups (p < 0.0001). The renal survival rate was significantly higher for T0 than T1 (p < 0.0001) or T2 (p < 0.0001). The renal survival rate for T1 was significantly higher than for T2 (p = 0.0165). (f) The 20-year renal survival rate was not significantly different among the three C-score groups (C0, 74.9%; C1, 82.6%; C2, 64.2%; p = 0.4954). The renal survival rate for the C0 category was similar to that for C1 (p = 0.6672) and for C2 (p = 0.4924). The renal survival rate for C1 was similar to that for C2 (p = 0.2219). Table 3. Independent risk factors for progression to ESRD in the multivariate Cox regression analysis. BMI body mass index, MAP mean arterial pressure, TP serum total protein, eGFR estimated glomerular filtration rate, U-Prot urinary protein excretion, U-RBC urinary red blood cells, M mesangial hypercellularity, E endocapillary hypercellularity, S segmental sclerosis, T interstitial fibrosis/tubular atrophy, C crescents. www.nature.com/scientificreports/ Asian. Second, we did not exclude cases of deteriorated renal function and/or mild proteinuria cases at diagnosis because our aim was also to evaluate the rapid progressive cases (cases of deteriorated renal function), which seem to have more crescent formation, and also evaluate early-stage cases (mild proteinuria cases) especially those diagnosed in Japan. We have shown the cohort results when applying the inclusion criteria stated in the original Oxford classification (Supplemental Materials). However, the results were clearer in the cohort where our criteria were applied than in the cohort with the criteria of the original Oxford classification. In Japan, IgAN was mainly found in health screening checks, and the criterion for biopsy was the early stage of the disease; therefore, the evaluation criteria of the Oxford classification might be different. Specifically, although delta eGFR might be the most appropriate marker to evaluate the short-term renal prognosis among patients with more than mild IgAN, the Oxford Classification was useful in our study, in which we included cases with a wide range of IgAN disease severity, from mild (slowly progressing) to deteriorating IgAN. Third, this study was a retrospective cohort analysis. To establish strong evidence, large multicentre prospective control trials, including patients of other races, should be performed.

conclusions
In this study, we report a validation of the Oxford analysis and found that the T score was the most important predictive factor of renal survival in all IgAN patients despite treatment with corticosteroids/immunosuppressors. However, corticosteroids/immunosuppressors improved the long-term renal prognosis for IgAN patients with E1, S1, and C1 scores.

Methods
Study population and study design. In    www.nature.com/scientificreports/ 1 year after renal biopsy as the initial treatment (treatment group), and 445 patients did not (non-treatment group). Validation of the Oxford classification was determined using all patients, and comparisons were made between the two groups. This retrospective cohort study was conducted in accordance with the guidelines of the Declaration of Helsinki and was approved by the Medical Ethics Committee of Tokyo Women's Medical University (reference #5104). Written informed consent to perform a renal biopsy was obtained from all patients; patients were able to opt-out of this study by visiting our institution's website.
Diagnosis of IgAN and histological evaluation of renal biopsy specimens. The indication for renal biopsy generally depended on a higher amount of proteinuria (> 1.0 g/day), U-RBC (> 50/HPF), stable renal function (chronic kidney disease (CKD) grade 1 or 2), and/or severe histological active lesions, such as endocapillary hypercellularity and cellular and fibro-cellular crescents. The criteria for renal biopsy included rapid progressive glomerulonephritis, with deteriorating renal function, and patients' background and assessment/treatment goals in cases with mild urinary and/or histological findings.
All renal biopsy specimens were obtained using a percutaneous needle biopsy. Specimens were fixed in 10% phosphate-buffered formalin (pH 7.2), embedded in paraffin, and cut into 4-μm-thick sections. The sections were stained with haematoxylin and eosin, periodic acid-Schiff, silver methenamine, and Masson trichrome; then, they were examined by light microscopy. For the immunofluorescence analysis, the specimens were fixed with cold acetone, and frozen sections were routinely subjected to fluorescence by IgG, IgA, IgM, C3, C4, C1q, fibrinogen, and fibronectin. IgAN was diagnosed based on mesangial proliferative changes in light microscopic findings, mesangial IgA and C3 deposition in immunofluorescence findings, and mesangial electron-dense deposits in electron microscopic findings.
Histological findings were graded according to the Oxford classification 10,11,27 . Clinical and laboratory data. Patient sex, age, body mass index (BMI), SBP, DBP, MAP, and duration of the observation period were recorded. Laboratory data included serum TP, creatinine (Cr), eGFR, uric acid (UA), total cholesterol (T-Cho), triglycerides, U-Prot, and urinary red blood cells (U-RBC) at the time of renal biopsy; these were evaluated as baseline data. The eGFR was calculated using the modified isotope dilution mass spectrometry-modification of diet in renal disease (IDMS-MDRD) study for Japanese individuals (eGFR = 194 × S-Cre -1.094 × age -0.287 × 0.739 [if female]) 35 . Time to progression to ESRD, defined as requiring dialysis or renal transplantation, was evaluated as the endpoint, and the risk factors associated with progression to ESRD were evaluated.
Statistical analysis. Data were expressed as mean ± standard deviation (SD) for normally distributed data and as the median and interquartile range (IQR) for skewed data. Cumulative renal survival rates until ESRD were calculated according to the Kaplan-Meier method and compared using the log-rank test. The unpaired Student's t-test for normally distributed data and Mann-Whitney's U test for skewed data were used to compare the clinical findings of patients treated with or without corticosteroids/immunosuppressors. The chi-squared test was used to compare the sex distribution, the number of patients with each grade of U-RBC at the time of renal biopsy, and the Oxford classification of groups treated with or without corticosteroids/immunosuppressors. Univariate and multivariate Cox regression analyses were performed to evaluate the risk of deterioration to ESRD. The univariate analyses indicated that sex (male/female) and Oxford classification were categorical variables, and that age, BMI, MAP, eGFR, UA, T-Cho, U-Prot, and U-RBC were quantitative variables. The results of these univariate and multivariate analyses are expressed as hazard ratios (HRs) with 95% confidence intervals (CIs). In all analyses, p < 0.05 was considered statistically significant. All analyses were performed using JMP Pro 13.0.0 (SAS Institute Inc., Cary, NC, USA).

Data availability
The datasets during and/or analysed during the current study available from the corresponding author on reasonable request.