Correlation of Serum Galactose-Deficient IgA1 and Oxford Class in Cases of IgA Nephropathy

This was a hospital-based prospective cohort study. The study was approved by the institutional ethics committee at Dr. Ram Manohar Lohia Institute of Medical Sciences, Lucknow, India (vide IEC No. 145/20). The sample size in the study was calculated by using G Power software (version 3.1). To analyze the difference in means of paired samples, using an effect size of 0.5, a total sample size of 35 was obtained, at an α of .05 and power of 0.8. An adjustment for loss to follow-up of 5% was added. Hence, a sample size of 40 was calculated for cases. To standardize the assay, a control size of 20 (ratio 2:1) was chosen. Patients attending the outpatient department or admitted in the Department of Nephrology were enrolled for the study. Forty renal biopsy–proven cases of IgAN with no history of immunosuppressive treatment were included in the study. A total of 20 healthy volunteers with no abnormalities on urine examination were taken as controls. Informed consent was obtained from the subjects after the nature of procedures was fully explained. Criteria for remission of proteinuria and hematuria at the end of 3 months were adopted from those used in the TESTING trial.¹²  Complete proteinuria remission was defined as 24-hour urine protein of less than 200 mg/d. Partial proteinuria remission was defined as proteinuria less than 50% of baseline by 24-hour urine protein and less than 1 g/d. No remission was taken as proteinuria persistence above 50% of baseline or more than 1 g/d. Those patients who did not have proteinuria at baseline were not graded. Hematuria remission was defined as less than 5 red blood cells (RBCs)/high-power field (HPF) in urine at the end of the third month. More than 5 RBCs in urine was considered as no hematuria remission. Those patients with fewer than 5 RBCs at baseline were not graded.

At least 2 renal cores were received, one in formalin (for light microscopy) and another in normal saline (for immunofluorescence studies). The renal core for light microscopy was processed routinely, paraffin embedded, and studied by using hematoxylin-eosin, periodic acid–Schiff, Masson trichrome, and periodic silver methenamine stains. The other core was processed for immunofluorescence by using fluorescein isothiocyanate–conjugated polyclonal rabbit anti-human antibodies to IgG, IgA, IgM, C3, and C1q (all from Dako, Denmark). The diagnosis of IgAN was performed with light and immunofluorescence staining patterns.¹³  Serum samples of these cases were collected on the day of biopsy and preserved at −70°C. A follow-up serum sample of the patients was collected after 3 months. Serum values of Gd-IgA1 were measured by human Gd-IgA1 autoantibody (anti–Gd-IgA1) using a sandwich enzyme-linked immunosorbent assay (ELISA) kit (Cat. No. MBS109532, MyBioSource, San Diego, California). Serum was diluted 800 times with EIA Buffer (MyBioSource) to obtain optical density within the measurement range of the kit (0.625–20 ng/mL).

Clinical data pertaining to age, sex, duration of illness, presence of edema and oliguria, and preceding history of hypertension and diabetes were recorded. Serology including serum creatinine, blood urea, serum albumin, and complement C3 and C4 levels were noted from patient records and hospital information system. Urinary albumin, 24-hour urine protein, and urine microscopy findings were similarly recorded. Data were collected on baseline (on day of biopsy) and after 3 months. Histopathologic grading of all renal biopsies was done by using the Oxford classification, and C-MEST scores were calculated as described earlier.¹⁴ 

The SPSS 24.0 statistical package was used for data management and analysis. The assessment of variables was done by analysis of variance (ANOVA) and χ² test. To estimate the prognostic value of markers, logistic regression analysis was applied. A P value of <.05 was considered significant, and the assessments were 2 tailed.

Forty consecutive renal biopsy–proven cases of IgAN were included. Serum samples from 40 cases were available at baseline and from 32 cases at the third month of follow-up (2 patients died, 1 underwent renal transplant, and 5 others were lost to follow-up). Among the 40 cases of IgAN included, the patient ages ranged from 15 to 72 years. Mean age was 34.65 years, with 30 males and 10 females. The maximum number of cases was noted in the 21- to 40-year age group, with a male predominance.

The duration of symptoms before biopsy ranged from 10 days to 24 months. Edema at presentation was noted in 28 patients. Diabetes was concomitantly present in 5 patients and hypertension in 24 patients. Oliguria was present in 2 cases only; both were male patients with ages of 21 and 36 years. Table 1 gives the baseline characteristics of the patients. The level of serum albumin recorded on the day of biopsy ranged from 1.15 to 7.45 g/dL, with a mean value of 3.516 g/dL. Mean baseline hemoglobin level was 11.23 g/dL. Hemoglobin concentration negatively correlated with baseline creatinine (Spearman rho [ρ] = −.33; P = .04) and urea level at 3 months (ρ = −.37; P = .02). The mean level of serum C3 was 115.58 mg/dL, and serum C4 was 34.98 mg/dL. Only 5 cases showed low serum complement C3 levels (<90 mg/dL reference); the rest showed normal C3 levels. Complement C3 and C4 levels positively correlated with each other (P = .39; P = .01). Serum creatinine level at baseline ranged from 0.55 to 7.47 mg/dL (mean, 2.27 mg/dL) (reference range for adults, 0.4–1.4 mg/dL). The mean blood urea level was 56.26 mg/dL (range, 16.8–133.0 mg/dL) on day of biopsy (reference range for adults, 10–45 mg/dL). The estimated glomerular filtration rate (eGFR; calculated by using the Modification of Diet in Renal Disease [MDRD] equation) had a mean of 48 mL/min/1.73 m² (range, 10–137 mL/min/1.73 m²) at baseline.

Thirty-nine cases showed presence of proteinuria at baseline. Twelve cases (30% of patients) showed nephrotic-range proteinuria (>3.5 g/d). Microscopic hematuria (>5 RBCs/HPF) was present in 37 cases at presentation. White blood cells (WBCs) (>5 WBCs/HPF) were present in urine sediment from 9 cases. Only 3 cases showed presence of more than 5 epithelial cells in urine sediment.

Serum creatinine and urea levels were recorded for each patient at a follow-up visit 3 months after biopsy. Serum creatinine levels on follow-up at 3 months ranged from 0.6 to 6.5 mg/dL (mean, 1.8 mg/dL). eGFR showed a mean of 60.65 mL/min/1.73 m² (range, 10–135 mL/min/1.73 m²) after 3 months of follow-up. Table 2 shows the values at baseline and on follow-up at 3 months.

The duration between the onset of symptoms and the acquisition of biopsy negatively correlated with baseline creatinine levels (ρ = −0.42; P = .007). Creatinine at baseline positively correlated with creatinine at 3 months (ρ = 0.874; P < .001) and urea at 3 months (ρ = 0.724; P < .001), besides its correlation with the duration, hemoglobin concentration, and urea levels at baseline. Creatinine concentration at baseline did not correlate with age, serum albumin, complement C3 or C4 levels, or 24-hour urine protein at baseline. Creatinine at 3 months positively correlated with urea at 3 months (ρ = 0.841; P < .001), besides its correlation with urea and creatinine at baseline. Creatinine at 3 months did not correlate with duration of illness, age, serum albumin, and complement C3 or C4 levels or 24-hour urine protein at baseline.

The mean blood urea level after 3 months of follow-up was 51.1 mg/dL (range, 12.6–136 mg/dL). Urea at baseline positively correlated with baseline creatinine (ρ = 0.652; P < .001), creatinine at 3 months (ρ = 0.592; P < .001), and urea at 3 months (ρ = 0.495; P = .001). No correlation was seen with the other parameters.

Of the 32 cases with follow-up samples, 12 (37.5%) showed complete proteinuria remission, 9 (28.13%) showed partial remission, and 11 (34.38%) did not show any remission for proteinuria. In 3 months of follow-up, among 29 cases (with follow-up samples available and excluding 3 with no hematuria at baseline) 9 (31.03%) presented with fewer than 5 RBCs/HPF. A significant difference was observed in hematuria before and after treatment for 3 months. Twenty of the 29 cases (69%) did not show any remission for hematuria. Five of the 9 cases (55.56%) with more than 5 WBCs in urine sediment at baseline had fewer than 5 WBCs in the urine sediment after 3 months. There was no significant difference in leukocytes in urine sediment in IgAN cases over this time.

All IgAN biopsies (n = 40) were scored as per the Oxford (C-MEST) scoring system. For C score, 30 cases showed C0 (75%), followed by C1 and C2, while for M score, 21 cases showed M1 (52.5%). For E and S scores, predominantly E0 was seen in 21 cases (52.5%) and S0 in 22 cases (55%). T1/T2 score in 27 cases (67.5%) was predominant as compared to T0 score. Photomicrographs of various histologic findings are given in Figure 1, A through H.

Gd-IgA1 levels were assessed at baseline in IgAN cases and in healthy controls. The mean levels and ranges in IgAN cases at baseline and in healthy controls were 4881.58 ng/mL (range, 1984.8–23 200 ng/mL) and 1912 ng/mL (range, 1076–2481.3 ng/mL), respectively. A significant difference was observed in the values in these 2 groups (P = .001). After a 3-month follow-up, the mean value was 4230.22 ng/mL. There was no significant correlation between serum Gd-IgA1 levels at baseline and after 3 months of follow-up (P = .31) (Figure 2).

No significant correlation was noted between Gd-IgA1 levels with different baseline characteristics, biochemical parameters, and immunofluorescence findings. However, the distribution of Gd-IgA1 levels at 3 months was found to differ significantly between different grades of Oxford S score (Kruskal-Wallis ANOVA, P = .008). The various correlations are detailed in Tables 3 and 4.

IgAN is the commonest form of glomerulonephritis worldwide in the adult population. Gd-IgA1 has been recently identified as a potential key factor in the pathogenesis of IgAN. Gd-IgA1 bound to IgG was identified in sera from patients with IgAN by using size exclusion chromatography, sucrose gradient ultracentrifugation, and ELISA techniques. It has been proposed that these heavy complexes cannot undergo hepatic degradation and their surplus amounts in serum are eventually deposited in renal mesangium, leading to IgAN.^15,16  Several studies have been done to find the utility of serum Gd-IgA1 levels as a diagnostic as well as prognostic marker. We studied the correlation of serum Gd-IgA1 with serologic, urinary, and histologic markers in IgAN cases reported at our center. Forty biopsy–proven IgA cases were analyzed in our study. Serum samples of these cases were collected on the day of biopsy and at follow-up after 3 months.

The age group of patients with IgAN ranged from 15 to 72 years, with a mean age at presentation of 34.65 years. Overall, among the 40 IgAN cases, 30 (75%) were male patients (male to female ratio, 3:1), which is similar to the higher prevalence in males reported by many other studies on IgAN.^17–20  The duration of symptoms before diagnosis of IgAN varied from 10 days to 4 years. Twenty-eight patients (70%) presented with edema. Hypertension was present in 24 IgAN cases (60%). A study by Zheng et al²¹  also showed prevalence of hypertension in 63.3% of IgAN cases. In 2 studies from India, hypertension was reported as the most common sign with a prevalence among 67.9% and 78.8% of IgAN cases.^22,23 

Renal biopsy samples were scored by using the Oxford classification.¹⁴  Ten cases (of 40, 25%) showed presence of crescents, for a C1/C2 score. This finding was quite similar to that reported by other authors.^24,25  Twenty-one cases (52.5%) showed mesangial hypercellularity with M1 score. M1 Oxford score was reported in various studies in 41.3%, 69.4%, and 86% of IgAN cases.^24–26  Presence of endocapillary hypercellularity (E1 score) was noted in 19 of 40 cases (47.5%), more frequently than reported in other studies.^24–26  Twenty-two cases (55%) presented with segmental glomerulosclerosis score S0 and 18 cases (45%), with S1. Segmental glomerulosclerosis of S1 grade was reported in 27.4%, 16.2%, and 68.4% of cases in different studies.^24–26  T1 and T2 grades were found in 27 cases (67.5%), as similarly reported by another Indian study.²⁴  However, lower rates (around 16%) have been reported in East Asian countries (South Korea and Japan).^25,26 

Results of Oxford scoring vary somewhat in our study and other reported series on IgAN. This may be a result of timing of biopsy and duration of disease prior to it.

Gd-IgA1 levels were assessed at baseline in IgAN cases and in healthy controls. The mean levels and ranges in IgAN cases at baseline and in healthy controls were 4881.58 ng/mL (range, 1984.8–23 432 ng/mL) and 1912 ng/mL (range, 1076–2481.3 ng/mL), respectively. A significant difference was observed in the values in these 2 groups (P = .001). Other studies also showed a significant difference in Gd-IgA1 levels between IgAN cases and healthy controls.^10,24,25,27  Other studies have also assessed Gd-IgA1 in non-IgAN glomerular diseases and found a significant difference from levels observed in IgAN cases.^24,25,28  This may be of utility in predicting IgAN and differentiating these cases from healthy controls and cases of non-IgAN glomerular diseases. Larger studies addressing this issue are warranted.

The mean value of Gd-IgA1 at baseline was 4881.58 ± 3417.81 ng/mL, with 3-month follow-up Gd-IgA1 value of 4230.22 ± 2000.87 ng/mL. No significant correlation between serum Gd-IgA1 levels at baseline and after 3 months of follow-up was found (P = .31).

In our study no correlation of age and Gd-IgA1 levels was noted (P = .13). Similar to ours, other studies in adults have also not shown any correlation between age and serum Gd-IgA1 levels.^24–26,29  In contrast to ours, a study by Irabu et al²⁸  in 2020 showed a significant positive correlation between serum Gd-IgA1 levels and age.

No significant correlation was noted in our study between eGFR and Gd-IgA1 (P = .18 at baseline and P = .83 at 3-month follow-up). Similar findings were reported by Bagchi et al²⁴  (P = .52) and Wada et al²⁶  (P = .12). Moldoveanu et al¹⁰  also did not find any significant correlation between serum Gd-IgA1 levels and eGFR. On the contrary, Kim et al²⁵  reported a significant correlation between eGFR and Gd-IgA1 levels (P = .03).

Proteinuria is also considered as a prognostic factor for IgAN in many studies. In our study no significant correlation was noted between baseline (P = .09) or 3-month (P = .14) Gd-IgA1 levels and proteinuria. A recent study of 33 patients with IgAN by Nguyen et al³⁰  from Switzerland found no significant association between serum Gd-IgA1 levels and proteinuria or eGFR at diagnosis. Similar lack of association between baseline Gd-IgA1 and proteinuria was reported by Bagchi et al²⁴  (P = .48) and Wada et al²⁶  (P = .06). Hematuria is one of the most important presenting features in IgAN. In ours as well as other notable studies there was no correlation found between Gd-IgA1 levels and hematuria.^10,25 

Apart from clinical parameters, which can predict prognosis in IgAN, Oxford classification is also an independent predictor of disease prognosis. Many studies done to validate the Oxford classification have shown that S and T scores better predict disease progression and worse renal outcome.^10,31,32  Our study found significant correlation between Gd-IgA1 at 3-month follow-up and segmental glomerulosclerosis (P = .008). Other parameters of Oxford classification did not show significant correlation with Gd-IgA1. A report from a Chinese population reported higher Gd-IgA1 levels in patients with more severe histologic class.³¹  Nguyen et al³⁰  showed that serum Gd-IgA1 level was significantly higher in patients with segmental glomerulosclerosis (S0 versus S1, P = .02) and tubular atrophy/interstitial fibrosis (T0 versus T1 and T2, P =.03; T0 versus T2, P = .02). Similarly, Kim et al²⁵  found Gd-IgA1 levels to correlate with tubular atrophy/interstitial fibrosis. However, some authors found no correlation between serum Gd-IgA1 level and histologic findings.²⁴  Irabu et al²⁸  also assessed serum Gd-IgA1 levels with mesangial IgA deposition, which they found significant. We did not find any such significance. Table 5 shows the comparison of mean values of serum Gd-IgA1 reported in various studies.

It has been established that λ light chains are identified more frequently in glomerular deposits of IgAN cases than in other glomerulonephritis cases.³³  Recent studies have reported that λ light chains are expressed more often in peripheral blood mononuclear cells from IgAN cases than in healthy controls and non-IgAN cases. Serum from IgAN cases has also shown more λ than κ light chains.³⁴  Lambda dominance of membrane-bound and secreted light chains in IgAN cases may relate to leukocyte aggregation and inflammatory damage of the glomeruli.³⁵  Some authors have found an association between λ light-chain deposition on immunofluorescence and endocapillary hypercellularity on renal biopsies, but no relation to clinical remission or ESRD development.³⁶  Others have found it to associate with crescents, tubular atrophy, interstitial fibrosis, and inflammation and with a poorer prognosis.³⁷  While we did not assess the light-chain subtype in our study, its significance may need to be evaluated in larger cohorts.

Our study showed that after 3 months of follow-up, serum Gd-IgA1 is significantly correlated with segmental sclerosis of Oxford classification. Other studies like those by Zhao et al²⁷  and Maixnerova et al³⁸  concluded that higher Gd-IgA1 levels correlate with the rate of decline of renal function. Berthoux et al³⁹  concluded that higher serum Gd-IgA1 level at time of biopsy correlated with need of dialysis or led to death.

Several studies have discussed the utility of Gd-IgA1 in prognosis of IgAN cases. Though our study was limited by sample size and short follow-up of 3 months’ duration only, we did not find any correlation between eGFR levels at baseline or 3-month follow-up, proteinuria remission status, hematuria remission status, and Gd-IgA1 levels (as predictors of prognosis). The only significant difference was found between S scores and Gd-IgA1 levels. However, regarding terminal events, which occurred during the course of the study, 2 patients died within 3 months. Of this poor prognostic group, 1 case had a markedly raised baseline Gd-IgA1 level of 23 432 ng/mL (and the other of 3096.8 ng/mL) as compared to an overall mean of 4881.58 ng/mL in the entire study group. This suggests that Gd-IgA1 may hold some relevance in predicting disease outcome. Nguyen et al³⁰  found similar correlation of Gd-IgA1 levels with Oxford S scores. They experimentally stimulated human mesangial cells in vitro with IgA1 from IgAN cases. Higher levels of monocyte chemoattractant protein 1 (a proinflammatory cytokine) in the cell culture were noted in cases with higher Gd-IgA1 levels and higher S score.³⁰  This indicates that higher mesangial inflammation may be the putative step through which higher Gd-IgA1 may induce segmental sclerosis and hence be a poor prognostic marker. This warrants longer follow-up studies on a larger number of cases to validate or refute the prognostic value of Gd-IgA1 levels in IgAN.