This study was conducted from January 2019 to September 2021 and involved children with idiopathic nephrotic syndrome who had not received prior steroid therapy. The study protocol involving human participants was reviewed and approved by the Ethics Committee.

The diagnosis of idiopathic nephrotic syndrome (INS) was based on the presence of edema, 24-h urinary protein excretion of ≥50 mg/kg, morning urinary protein/creatinine of >2 mg, hypoalbuminemia of <25 g/L and the disease of unknown causing. All children with INS received the standard steroid therapy and were classified into three categories, steroid-sensitive nephrotic syndrome (SSNS), frequent relapse nephrotic syndrome (FRNS) and steroid-resistant nephrotic syndrome (SRNS), on the basis of their clinical responses toward steroids. The SSNS group included patients with negative urinary protein for ≤4 weeks in those treated with sufficient prednisone [2 mg/(kg·d) or 60 mg/(m·d)]. The frequent relapse group included patients in whom INS recurred two times or more within half a year, or four times or more within 1 year in the course of the disease. The SRNS group included patients who failed to achieve remission after 4 weeks of daily sufficient prednisone. FRNS and SRNS were collectively referred to as refractory nephrotic syndrome (RNS). Further, the relapse group included patients in whom the quantity of urinary protein was ≥50 mg/kg, or the urinary protein/creatinine (mg/mg) of morning urine was ≥2.0, or the morning urinary protein changed from negative to positive for three consecutive days. The non-relapse group included patients in whom INS no recurred within 1 year after the first complete remission. The infrequent relapse group included patients in whom INS recurred once within 6 months or one to three times within 1 year after the first complete remission. Patients who were not finished four weeks of glucocorticoid therapy, or received other immunosuppressants, monoclonal antibodies, or cytotoxic drugs within four weeks, would be excluded from the study.

The clinical features of enrolled patients were recorded from the medical record, such as age, genders, creatinine, estimated glomerular filtration rate (eGFR), 24-h urine protein (Upro), occurred of complications and hypertension. eGFR was calculated according to Schwarz-formula (9).

The blood sample were collected into vacuum sampling vessel containing coagulant (Improve Medical, China) before steroid therapy from the enrolled patients. Serum were separated within 2 h after blood collection and then divided and stored in the refrigerator at 2–8°C immediately. All the serum samples were transferred to an ultra-low temperature refrigerator (−80°C) no more than 8 h. The samples were guaranteed not to thaw until the ELISA test.

Renal biopsies were recommended in patients with INS diagnosis. Pathology test should obtain the written informed consent of the guardian first. Renal biopsies were executed by skilled nephrologists, and all biopsies were assessed by pathologists through light microscopy and immunofluorescence. Each kidney sample was observed for the total number of glomeruli, glomerular sclerosis, mesangial proliferation, basement membrane thickening, tubular degeneration and atrophy, interstitial fibrosis, interstitial inflammation, etc.

Human Leptin ELISA KIT (4A biotech, China, Assay range: 31.25–2000 pg/ml), Human CCL22 ELISA KIT (4A biotech, China, Assay range: 15.625–1000 pg/ml) were used for patient serum test. Briefly, to assay each protein, serum samples at an optimal dilution were added to a microplate precoated with capture antibody, incubated, washed and followed by addition of capture antibody, horseradish peroxidase and substrate. The absolute levels of serum protein biomarkers were determined using standard curves run on each ELISA plate, and normalized by urine creatinine concentration.

The quantitative data with normal distribution was expressed as (Average ± SD), and the classified data indicated with the quantity of each component separately. Statistical analysis was performed using SPSS 22.0. The chi-square test was used to analyze gender differences among the groups, while the t-test was used to assess age differences. Outlier detected by histogram, and replaced with the average in statistical process. Pearson bivariate correlation analysis was used for correlation statistics. A P-value greater than 0.05 indicated no significant difference. After quantifying the ELISA results based on the standard curve, the t-test or Wilcoxon rank sum test was used for between-group comparisons, and analysis of variance or Kruskal-Wallis test was used for comparisons involving more than two groups. Statistical significance was set at P < 0.05. Receiver operating characteristic (ROC) curves were utilized to determine the effective area, sensitivity, and specificity of candidate indexes.

A total of 117 cases (86 males, 31 females) with childhood INS were included in the study, comprising the subgroups of SSNS (82 cases), FRNS (10 cases), and SRNS (25 cases). Additionally, 40 cases (28 males, 12 females) undergoing health check-ups and confirmed to be free of obvious diseases were selected as the healthy control group. Serum samples were collected before steroid therapy and stored according to the aforementioned protocol. The clinical features of children enrolled in the study were listed in Table 1. There were no significant differences in age, gender ratio, Creatinine, eGFR and Upro between the three groups (P > 0.05). There were 148 complications could be recorded in 117 INS patients, with 12 types. The first three kinds were inflammation (48 cases), hydrops (30 cases) and liver injury (23 cases). There was no difference in the incidence of complications among three subgroups (P > 0.05). The incidence of hypertension was different (P < 0.05), with the highest in SRNS.

A total of 41 (35.0%) patients underwent kidney biopsy, 11 cases (13.4%) in SSNS, 7 cases (70.0%) in FRNS and 24 cases in SRNS with an acceptance rate up to 96.0%. Minimal change disease (MCD) was the most common pathological type with 30 cases, mesangial proliferative glomerulonephritis (MsPGN) with 9 cases and focal segmental glomerulosclerosis (FSGS) with 3 cases. Pathological types in the three groups with serum concentration of CCL22 and Leptin were shown in Table 2.

There was no significant correlation between Leptin and Creatinine (P = 0.368), eGFR (P = 0.329), Upro (P = 0.241). At the same between CCL22 and Creatinine (P = 0.060), eGFR (P = 0.570), Upro (P = 0.963). These indicated that changes of Leptin and CCL22 in INS patients had a different significance compared to clinical features.

The concentration of serum Leptin was significantly increased in RNS, with even higher levels observed in the SRNS subgroup. No significant differences were found between the healthy group and SSNS. The results were depicted in Figure 1A. Patients in SSNS were distinguished as no relapse and infrequent relapse with fewer recurrences than FRNS. Leptin was associated with the frequency of relapse, and patients with FRNS had significantly elevated Leptin concentrations (Figure 1B). Regarding the different pathological types, serum Leptin concentration was increased compared to healthy children, with a significant increase observed in MCD and MsPGN (Figure 1C). In patients with steroid resistance, there were minimal differences in Leptin concentration among the three pathological types (Table 2). Compared with group SSNS, Leptin was significantly higher in RNS group (Figure 1D). All data were listed in Table 2 and Table 3.

The concentration of serum CCL22 was significantly decreased in SSNS compared with SRNS, and no significant differences with healthy group (Figure 1E). The recurrence of INS patients was not related to CCL22 concentration (Figure 1F). CCL22 differed significantly among the three pathological types, with elevated levels in MCD and FSGS, and reduced levels in MsPGN (Figure 1G). Compared between RNS and SSNS, there were no significant differences in CCL22 concentration (Figure 1H). All data were listed in Table 2 and Table 3.

Compared to SSNS group, the concentration of serum CCL22 and Leptin significantly increased in SRNS (P = 0.046 and P = 0.007). The diagnostic performance of serum CCL22 and Leptin in predicting RNS or SRNS was evaluated using the ROC curve shown in Figure 2. For SRNS prediction, the AUC of Leptin, CCL22, and the joint prediction index were 0.764, 0.640, and 0.806, respectively (Figure 2A). The cut-off of Leptin was 602.6 pg/ml, with positive predictive value (PPV) = 0.43, negative predictive value (NPV) = 0.93, The cut-off of CCL22 was 23.7 pg/ml, with PPV = 0.27, NPV = 0.87. For RNS prediction, the AUC of Leptin, CCL22, and the joint prediction index were 0.748, 0.578, and 0.756, respectively (Figure 2B). The cut-off of Leptin was 555.6 pg/ml, with PPV = 0.53, NPV = 0.88, The cut-off of CCL22 was 18.3 pg/ml, with PPV = 0.36, NPV = 0.80.