Baseline characteristics at time of kidney biopsy (Table 1) Table 1 Patient characteristics

Patients with IgAN were younger than controls and had higher eGFR and UACR compared to patients with non-IgAN CKD. Serum IgA and Gd-IgA1 were significantly elevated in patients with IgAN vs. patients with non-IgAN CKD and healthy controls (Fig. 1). Serum Gd-IgA1:IgA differed significantly between groups with highest values in patients with IgAN. Serum Gd-IgA1:IgA was significantly elevated in patients with IgAN vs. healthy controls (p = 0.001) but did not reach statistical significance vs. patients with non-IgAN CKD (p = 0.053) (Fig. 1). These results were not significantly altered after exclusion of the three patients on prednisolone treatment from the analysis.

Fig. 1

 S-IgA (A), Gd-IgA1 (B) and Gd-IgA1:IgA (C) at time of kidney biopsy

IgAN (n = 40), Non IgAN CKD patient (n = 21), Healthy controls (n = 19)

* denotes p < 0.05 and ** p < 0.001 between groups. Abbreviations: IgA: Immunoglobulin A, IgAN: IgA nephropathy, CKD: chronic kidney disease, Gd-IgA1: galactose deficient IgA1.

Serum levels of IgA, Gd-IgA1, and Gd-IgA1:IgA in patients with non-IgAN CKD were comparable to those in healthy controls. The urinary Gd-IgA1:creatinine was significantly increased in patients with IgAN vs. patients with non-IgAN CKD (p = 0.049). At the time of biopsy three patients with IgAN were on immunosuppressive therapy (all with prednisolone). A total of six IgAN patients (15%) were treated with immunosuppressive drugs at some point during the study period.

Associations between MEST-C scores and UACR, eGFR and Gd-IgA1 at time of kidney biopsy

There were no statistically significant differences in eGFR between patients with M1 vs. M0, E1 vs. E0 or S1 vs. S0 (data not shown). Patients with T1 had lower eGFR compared to patients with T0 (51 [IQR 49–65] vs. 86 [IQR 70–115] mL/min/1.73m2, p < 0.001 and patients with T2 had lower eGFR than patients with T1 (34 [23–49] vs. 51 [IQR 49–65] mL/min/1.73m2, p = 0.001).

Patients with M1 had elevated UACR compared to those with M0 (227 [IQR 107–297] vs. 54 [IQR 18–144] mg/mmol, p = 0.003) and patients with T2 had elevated UACR vs. those with T0 (226 [IQR 124–338] vs. 33 [IQR 14–83] mg/mmol, p = 0.003).

There were no statistically significant differences in serum Gd-IgA1 concentrations between patients with different MEST-C scores (Table 2). No patient showed C2 lesions. Similarly, there were no statistically significant differences in serum IgA, or Gd-IgA1:IgA, between patients with different MEST-C scores (data not shown).

Urinary Gd-IgA1:creatinine was elevated in patients with T2 scores vs. those with T0 (47.6 [IQR 26-71.4] vs. 6.5 [IQR 8.8–26.1] µg/mmol, p = 0.047). No additional, statistically significant, differences were found in urinary Gd-IgA1:creatinine between patients with different MEST-C scores (data not shown).

Table 2  S-Gd-IgA1 in patients with IgAN (n = 40) according to MEST-C subgroups

Values are median (IQR). IgA, Immunoglobulin A; IgAN, IgA nephropathy; Gd-IgA, galactose deficient IgA. For definitions of MEST-C scores see “Methods”. P values are for the Mann-Whitney U and Kruskal-Wallis tests.

Correlation analyses at time of kidney biopsy in patients with IgAN (n = 40)

There was a significant inverse correlation between UACR and eGFR (r=-0.39, p = 0.012). Neither serum Ig-A, Gd-IgA1, Gd-IgA1:IgA, IgA/C3, nor Gd-IgA1:C3 showed statistically significant correlations to either eGFR, UACR, systolic or diastolic BP, or age (values not shown). Urinary Gd-IgA1:creatinine showed a statistically significant positive correlation to UACR but not to serum Gd-IgA1 (Fig. 2). Both UACR (r = 0.47, p < 0.001) and urinary Gd-IgA1:creatinine (r = 0.47, p < 0.001) were significantly correlated to systolic blood pressure.

Fig. 2

Correlations between u-Gd-IgA1:creatinine UACR (A), and s-Gd-IgA1 (B) at time of kidney biopsy (n = 40). Abbreviations: IgA: Immunoglobulin A, Gd-IgA1: galactose deficient IgA1, UACR: urine albumin:creatinine ratio. r, Spearman’s correlation coefficient

Prognostic role of serum Gd-IgA1 at time of kidney biopsy in patients with IgAN (n = 40)

The median follow-up time after kidney biopsy was 78 months (IQR: 60–120 months). Median ΔGFR per year was − 2.1 (IQR − 3.8 - -0.6) mL/min/1.73m2. Median ΔUACR per year during follow up was − 5 (IQR − 19.3 − 0.25) mg/mmol. There was a statistically significant correlation between UACR at baseline and ΔUACR per year (r=-0.76, p < 0.001) but not with ΔGFR per year (r=-0.29, p = 0.072).

There was no statistically significant correlation between serum Gd-IgA1 and ΔGFR per year during follow up (r=-0.14, p = 0.39). In addition, neither serum IgA (r = 0.21, p = 0.19), Gd-IgA1:IgA (r=-0.04, p = 0.8), nor Gd-IgA1:C3 (r=-0.072, p = 0.74) correlated significantly with ΔGFR during follow up. These results were not significantly altered after exclusion of the three patients on prednisolone treatment from the analysis.

Similarly, there were no statistically significant correlations between serum Gd-IgA1, IgA or Gd-IgA1:IgA and ΔUACR per year during follow up (values not shown). Furthermore, there were no statistically significant correlations between serum Gd-IgA1, IgA or Gd-IgA1:IgA at baseline and changes in UACR or eGFR during the first 6 months after kidney biopsy (values not shown).

Prognostic role of urinary Gd-IgA1:creatinine at time of kidney biopsy in patients with IgAN (n = 40)

There was no statistically significant correlation between urinary Gd-IgA1:creatinine and ΔGFR per year during follow up (r=-0.22, p = 0.16). Urinary Gd-IgA1:creatinine showed a significant inverse correlation to ΔUACR during follow up (r=-0.67, p < 0.001).

Intra-individual changes in s erum and urinary Gd-IgA1 during follow-up in patients with IgAN (n = 19)

Nineteen patients had blood and urine samples taken after a median follow-up time of 116 months after the diagnostic kidney biopsy. Serum IgA concentrations did not change significantly during this interval (Fig. 3). However, both serum Gd-IgA1 and Gd-IgA1:IgA decreased significantly with time (Fig. 3). Similarly, urinary Gd-IgA1:creatinine was significantly reduced at follow-up (Fig. 4).

Fig. 3

Individual changes in serum IgA (A), GdIgA1 (B), and Gd-IgA1/IgA (C) between baseline and at follow up. Median follow up time was.116 months, (n = 19). Abbreviations: IgA: Immunoglobulin A, Gd-IgA1: galactose deficient IgA1.

Fig. 4

Individual changes in urinary Gd-IgA1:creatinine between baseline and at follow up. Median follow up time 116 mån, (n = 19). Abbreviations: Gd-IgA1: galactose deficient IgA1.

We assessed whether changes in serum Gd-IgA1 or Gd-IgA1:IgA were associated with changes in kidney function during the corresponding follow up period. However, changes in serum Gd-IgA1 and Gd-IgA1:IgA did not correlate significantly to ΔGFR or ΔUACR. Changes in urinary Gd-IgA1:creatinine during follow up correlated significantly to ΔUACR (r = 0.60, p = 0.008) but not to ΔeGFR.

Four out of 19 patients had received immunosuppressive therapy during the follow up period. After excluding these four individuals from the analyses reductions in serum Gd-IgA1 and Gd-IgA1:IgA over time were no longer statistically significant (p = 0.125 for both analyses).