Clinical impact of endocapillary proliferation according to the Oxford classification among adults with Henoch-Schönlein purpura nephritis: a multicenter retrospective cohort study

We performed a retrospective cohort study of adults (≥18 years) with HSPN who underwent a renal biopsy between 2004 and 2014 at fourteen different medical centers (Nagoya University Hospital, Japanese Red Cross Nagoya Daiichi Hospital, Kasugai Municipal Hospital, Anjo Kosei Hospital, Konan Kosei Hospital, Toyota Kosei Hospital, Ichinomiya Municipal Hospital, Handa City Hospital, Toyohashi Municipal Hospital, Yokkaichi Municipal Hospital, Ogaki Municipal Hospital, Chutoen General Medical Center, Gifu Prefectural Tajimi Hospital, and Nagoya Memorial Hospital). We included patients diagnosed with HSPN according to the EULAR/PRINTO/PRES HSP criteria [1] and the presence of IgA-dominant immune deposits on renal biopsy [9]. We excluded patients with less than 12 months of follow-up or with renal biopsy specimens with fewer than eight glomeruli. Patients were followed up until June 2018. During the study period, 89 patients with HSPN underwent a renal biopsy. We excluded ten patients followed up for less than 12 months and five patients with fewer than eight glomeruli in renal biopsy specimens. Seventy-four patients were analyzed.

We analyzed the clinical parameters from the medical records obtained when the patients underwent renal biopsy. A diagnosis of hypertension was defined as blood pressure (BP) values ≥140/90 mmHg or a prescription of antihypertensive drugs. A diagnosis of diabetes was defined as HbA1c levels ≥6.5% or treatment with antidiabetics medications. We calculated the estimated glomerular filtration rate (eGFR) using the modified equation for Japanese [eGFR = 194 × sCr^-1.094 × age^− 0.287 × 0.739 (if female)] [20].

We stained the renal biopsy samples by periodic acid-Schiff and Masson’s trichrome. Two nephrologists (K.I and AB.K) evaluated the renal biopsy specimens. Each histologic lesion was graded according to the ISKDC Classification, Oxford classification 2016, and JHC [16, 17].

The Oxford classification 2016 is based on the MEST-C score. M0 indicates the involvement of ≤50% of glomeruli with ≥4 cells/mesangial area and M1 indicates the involvement of ≥50% of glomeruli with ≥4 cells/mesangial area. E0 and E1 indicate the absence and presence of endocapillary proliferation, respectively, and S0 and S1 refer to the absence and presence of segmental glomerulosclerosis or tuft adhesion, respectively. T0, T1 and T2 indicate the degree of tubular atrophy or interstitial fibrosis in ≤25%, 26–50%, and > 50% of the cortical area, respectively [14]. C0, C1 and C2 refer to the amount of cellular or fibrocellular crescents of glomeruli as absent, ≤25%, and > 25%, respectively [21].

The ISKDC classification is used to analyze the degree of crescent formation. It includes six histological grades: grade I, minimal glomerular abnormalities without crescents; grade II, mesangial proliferation without crescents; grade III, mesangial proliferation with crescents in < 50% of glomeruli; grade IV, mesangial proliferation with crescents in 50–75% of glomeruli; grade V, mesangial proliferation with crescents in > 75% of glomeruli; and grade VI, membranoproliferative-like lesions [11].

The JHC is based on glomerular lesions in kidney biopsy samples. It includes four HGs, that reflect < 25%, 25–49%, 50–74%, and ≥ 75% of glomeruli. Histologically, the glomerular lesions are seen as follows; cellular crescent, fibrocellular crescent, fibrous crescent, segmental sclerosis, and global sclerosis [17].

The primary outcome of this study was a 30% decline in the eGFR from the baseline and/or ESKD.

Parametric and nonparametric variables were expressed as means ± standard deviations (SD) and medians and interquartile ranges (IQRs). They were compared using Student’s t-test and the Mann–Whitney U test. Categorical variables were expressed in percentages and were compared using the Fisher’s exact test. The Spearman correlation coefficient was used to determine the relationship between two histologic lesions. Kaplan–Meier survival curves were used to analyze freedom from the primary outcome in patients grouped according to the following classification: M0/1, E0/1, S0/1, T0/1 + 2, C0/1/2, ISKDC (grade I + II/III + IV), and JHC (HG I/II + III + IV). We used the univariate Cox regression to determine the factors predicting the primary outcome. Proteinuria was measured using time-collection urine, but we were not able to measure time-collection urine in four patients. Thus, they were assessed by calculating the spot urine protein-creatinine ratio. Variables with P values < 0.30 in the univariate analysis were examined in the multivariate analysis. These results were expressed as hazard ratios (HRs) with 95% confidence intervals (CIs). P values < 0.05 were considered statistically significant. All statistical analyses were performed with EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan). EZR is a graphical user interface for R statistics (The R Foundation for Statistical Computing, Vienna, Austria) [22].

The baseline clinical characteristics are shown in Table 1. The mean age was 47.8 ± 17.4 years, and 47.3% were male (the detailed data were shown in Additional file 1). The mean eGFR was 76.4 ± 25.8 mL/min/1.73 m², and the median proteinuria was 1.40 [IQR: 0.70–2.38] g/day. The median duration between lower limb purpura and the time of renal biopsy was 53 [IQR: 25–168] days. After a median kidney biopsy procedure period of 12 [IQR: 1–25] days, oral prednisolone (PSL) was prescribed to 61 (82.4%) patients.

Histologic characteristics are also shown in Table 1. Findings consistent with M1, E1, and S1 were observed in 5 (6.8%), 38 (51.4%), and 38 (51.4%) patients, respectively. In patients with E1, the median number of glomeruli with endocapillary proliferation was 14.3% [IQR: 8.03–22.4]. Findings consistent with T1 and T2 were observed in 14 (18.9%) and 4 (5.4%) patients, respectively. Crescents (ISKDC grade III + IV + V) were identified in 54 (73.0%) patients, including two (3.7%) with crescents in ≥50% of glomeruli (ISKDC grade IV + V). Findings consistent with C1 and C2 were found in 35 (47.3%) and 17 (23.0%) patients, respectively. HG I, II, and III in the JHC were diagnosed in 39 (52.7%), 27 (36.5%), and 8 (10.8%) patients, respectively.

Table 2 presents correlations between the different histologic variables. The strongest correlation was found between the Oxford C lesion and ISKDC classification (r = 0.81). Moreover, JHC strongly correlated with Oxford T (r = 0.56) and C (r = 0.62) and ISKDC classification (r = 0.55).

Table 3 shows a comparison of the clinical features by three pathological classifications. Oxford M is not shown because there were very few M1 lesions. There were significant age-related differences in Oxford T and C lesions, and JHC. eGFR varied significantly by T lesion, JHC, and ISKDC classification. Proteinuria varied significantly by Oxford E and C lesions, and ISKDC classification.

During a mean follow-up period of 68.0 ± 33.0 months, fourteen patients (18.9%) reached the renal outcome, defined by a 30% decline in eGFR from the baseline. Among them, only one patient developed ESKD after 11 years following the kidney biopsy. All patients who showed a ≥ 30% decline in eGFR had received oral PSL. Kaplan–Meier survival curves are shown in Fig. 1. The log-rank test revealed significant between-group differences in event-free renal survival in patients with E0 and E1 lesions (p = 0.0072). The M, S, T, and C lesions, ISKDC classification, and JHC did not show a significant association with renal outcome. In univariate Cox regression analysis, Oxford E lesion (HR: 6.13, 95% CI: 1.36–27.7) and patient age (HR: 1.64, 95% CI: 1.17–2.30) were significantly associated with renal outcome (Table 4). Male sex and C lesion showed tendencies of association with poor renal outcome, but these tendencies were not statistically significant. Table 4 presents multivariate models adjusted for clinical and pathological factors. In model 1, Oxford E lesion (HR: 8.94, 95% CI: 1.44–55.7) and patient age (HR: 1.58, 95% CI: 1.15–2.18) were significantly associated with an increased risk of poor renal outcome after adjustment for clinical factors (age, sex, and proteinuria). S lesion, T lesion, C lesion, ISKDC classification, and JHC were not associated with an increased risk of poor renal outcome after adjustment for clinical factors. In model 2, Oxford E lesion (HR: 6.71, 95% CI: 1.06–42.7) and patient age (HR: 1.57, 95% CI: 1.12–2.21) were significantly associated with an increased risk of poor renal outcome after adjusting for clinical factors, T lesion, and C lesion.