High serum IgA/C3 ratio better predicts a diagnosis of IgA nephropathy among primary glomerular nephropathy patients with proteinuria ≤ 1 g/d: an observational cross-sectional study

Immunoglobulin A nephropathy (IgAN), which was initially described by Berger and Hinglais in 1968 [1], is the most common type of primary glomerulonephritis worldwide [2‐4], especially in the Asia-Pacific region [2, 5, 6]. IgAN accounts for more than 20% of glomerular diseases [5, 6]. An epidemiological survey showed that approximately 45% of the cases of primary glomerulonephritis in China were IgAN [6]. IgAN is diagnosed by renal biopsy with immunofluorescence staining or immunohistochemistry, and is histologically characterized by the presence of polymeric IgA deposits in glomerular mesangial areas [1, 7]. The clinical presentation of IgAN is highly variable between individuals, and ranges from a subclinical disease with isolated microscopic hematuria to severe end-stage renal disease (ESRD). The stereotypical IgAN patient is a young adult with asymptomatic microscopic hematuria or episodic macroscopic hematuria occurring during an upper respiratory tract infection. IgAN was previously considered to be a benign condition, but recent studies have demonstrated that up to 25–40% of patients progress to uremia within 10–20 years after the diagnostic biopsy [6, 8‐10].

Until present, the diagnosis of IgAN relies on revealing IgA as the dominant or codominant immunoglobulin in the glomerular mesangium by kidney biopsy [1]. However, the pathogenesis of IgAN remains unclear. According to recent studies, IgAN is thought to be autoimmune in nature. Suzuki et al. put forward a four-hit hypothesis, in which the galactose-deficient IgA1 (GdIgA1) molecule plays a key pathogenic role [11, 12]. Although renal biopsy remains the gold standard for IgAN diagnosis, recent efforts have sought to find new biomarkers that can predict IgAN without the need for biopsy [13‐15]. Previous studies have shown that IgA deposition in the mesangial area is necessary for IgAN progression and accompanying C3 deposition. However, the correlation between the changes in serum complement levels and diagnosis remains unclear [12, 16]. Some researchers have reported that IgAN patients usually have higher serum IgA and GdIgA1 levels [17, 18], alongside normal-to-low serum C3 levels, compared to those of patients with other types of primary glomerulonephritis [19, 20]. In 2000, Yasuhiko T et al. reported that patients with higher serum IgA/C3 ratios were more likely to be diagnosed with IgAN [13]. Since then, several studies have sought to evaluate whether the IgA/C3 ratio can serve as a useful clinical indicator to predict the diagnosis and prognosis of IgAN [13‐15, 20‐22]. However, most of these studies focused on a Japanese patient population and had small sample sizes. Thus, these findings require replication in other patient populations and larger sample sizes to confirm the validity of the association, and to account for potential ethnic differences. Moreover, previous studies have not reported whether the IgA/C3 ratio is of similar value when diagnosing IgAN by proteinuria levels or renal function.

In non-randomized studies, good study methodology not only effectively evaluates the clinical program, but also minimizes confounding variables. In statistical analyses, propensity score matching (PSM) between treatment groups is a practical method to reduce confounding bias, particularly in evaluating comparative effectiveness of two or more variables [23, 24]. Unlike stratification or regression, the new PSM cohort is in equilibrium with the measured covariates during the PSM process, and is also in “clinical equipoise”. In other words, in this selected cohort, every patient should be a reasonable candidate for every treatment being studied [24‐26]. Increasing attention is being paid to the use of PSM as a statistical method in clinical studies.

The purpose of this study was to evaluate whether a high serum IgA/C3 ratio, the suggested novel biomarker, could be used as a diagnostic criterion for IgAN in a large Chinese patient population, using PSM methods.

In the present study, we assessed 1095 primary glomerular nephropathy patients, including 757 IgAN patients and 338 Non-IgAN patients, to investigate whether a high IgA/C3 ratio could be used as a diagnostic criterion for IgAN, using a PSM method. Our study revealed that both in the full cohort and PSM cohort, the serum IgA/C3 ratio of the IgAN was significantly higher than that of the non-IgAN. The same results were also obtained when comparisons were stratified by level of proteinuria or level of renal function. In the PSM cohort, there was no difference in the IgA/C3 ratio of IgAN patients between different proteinuria groups and different CKD stages. The AUROC for the IgA/C3 ratio was 0.767 in the full cohort, and 0.734 in the PSM cohort in distinguishing IgAN among primary glomerular disease. The highest AUROC of the IgA/C3 ratio for an IgAN diagnosis was in the proteinuria≤1 g/d group (0.801 in the full cohort, and 0.803 in the PSM cohort); however, there was no difference in the AUROC of the IgA/C3 ratio between all CKD groups. Meanwhile, the diagnostic accordance rate for a diagnosis of IgAN among all patients with an IgA/C3 ratio > 3.5304 was as high as 92.02% in the full cohort. A multivariate logistic regression analysis showed that IgAN was independently correlated with IgA/C3 ratio.

In recent years, studies have confirmed that IgAN is a complex immune disease that involves deposits of the circulating immune complexes (CIC) IgA, C3, and IgG in the mesangial region of the glomeruli [11]. Although the pathophysiology of IgAN is still unclear, researchers currently favor a four-hit hypothesis, wherein GdIgA1 plays a key pathogenic role [11, 12]. Firstly, increased levels of circulating GdIgA1 with aberrant O-glycosylation of its hinge region result in exposed N-acetylgalactosamine. And then, N-acetylgalactosamine results in the production of unique antiglycan antibodies Gd-IgA1-binding proteins, which are thought to target the terminal N-acetylgalactosamine in the hinge region of Gd-IgA1. The third hit is the formation of pathogenic Gd-IgA1–containing CICs. Finally, these complexes are deposited in the mesangial region, leading to glomerular injury [11, 12, 30]. Studies have also revealed that mesangial IgA can activate C3, which speeds the production of terminal complex c5b-9, which in-turn activates macrophage to produce inflammatory mediators and matrix proteins. Thus, both IgA and C3 play an important role in the development and occurrence of IgAN. Due to glomerular deposition of IgA1 or IgA1-IgG immune complexes, serum C3 levels are lower in IgAN, and together with elevated serum IgA, this result in a higher IgA/C3 ratio, which can be used as a biomarker for IgAN [31]. In addition to measurements of pure serum IgA or C3 levels, previous studies have also found that the serum IgA/C3 ratio can clearly distinguish between severe IgA histological lesions and mild IgA histological lesions [13, 22, 32]. Our study is in line with these previous results, as in our PSM cohort, the AUROC of the IgA/C3 ratio for the diagnosis of IgAN reached 0.734, while the AUROC of IgA and C3 alone were only 0.696 and 0.621, respectively. However, additional data from larger samples are still needed, especially as previous studies have all examined relatively small patient populations.

In the present study, IgAN patients had a significantly higher IgA/C3 ratio than that of non-IgAN primary glomerular disease patients, in both the full cohort and the PSM cohort. Our results are consistent with previous studies [13, 14, 21, 22], and are also the first to demonstrate the predictive value of IgA/C3 ratio among primary glomerular nephropathy patients between different levels of proteinuria and different renal functions. The cut-off value of the serum IgA/C3 ratio was 2.102 in the full cohort, with an AUROC as high as 0.767 and 2.410, and an AUROC as high as 0.734 in the PSM cohort. These results are close to the cut-off value of 2.14 with an AUROC of 0.71, which was reported by Yasuhiko T et al. in a comparison of IgAN and non-IgAN patients [13]. The Yasuhiko T team also reported that the serum IgA/C3 ratio might be used to diagnose and predict prognostic grading in patients with IgAN [33], and combined with microscopic hematuria and/or persistent proteinuria, high serum IgA levels, and the serum IgA/C3 ratio could be used to distinguish IgAN from other primary renal diseases [34]. A study that collected data from 44 children with IgAN treated with multi-drug combination therapy showed that combined serum IgA/C3 ratio and glomerular C3 staining could predict the prognosis of IgA nephropathy [35]. However, a study involving 1564 IgAN patients indicated that a decrease in serum C3 levels in IgAN was not associated with renal progression [36].

Persistent proteinuria has proven to be an adverse prognostic factor for IgAN, as its progression varies with different levels of proteinuria. A follow-up study by Heather N. R et al. showed that when proteinuria level was < 1 g/d, the renal function of IgAN patients decreased at a 90% slower rate than average. The rate of renal function decrease was accelerated with increased proteinuria, and in IgAN patients with continuous proteinuria > 3 g/d, the decrease in renal function was 25 times faster than that of patients with proteinuria < 1 g/d [29]. Studies have also shown that compared to other types of progressive glomerulonephritis, IgAN appears to be associated with a poorer prognosis when presenting with nephrotic proteinuria between 3 and 3.5 g/d. Therefore, it has been suggested that IgAN with proteinuria < 1 g/d often indicates a favorable prognosis [29, 37‐40]. Based on these observations, we divided proteinuria into three levels: ≤1 g/d, 1–3 g/d and > 3.5 g/d, and revealed that patients with proteinuria ≤1 g/d had the highest AUROC of the IgA/C3 ratio (which was higher than 0.80 in both the full cohort and the PSM cohort with cut-off points 2.229 and 2.352, respectively) in distinguishing IgAN among primary glomerular disease, while those in the proteinuria > 3.5 g/d group had the lowest. Meanwhile, renal function was also divided into three groups (eGFR ≥90, 90–60, and < 60 mL/min per 1.73 m²), but we found no difference in the AUROC of the IgA/C3 ratio for the diagnosis of IgAN between these groups.

In China, there are many nephropathy patients without access to renal biopsy. This is for a variety of reasons – such as the uneven distribution of medical resources, the prevalence of rural hospitals without access to biopsy technology, incidences of renal function impaired to a level unable to tolerate renal biopsy, and a prevalence of patients with a low level of proteinuria (i.e. < 1 g/d) with no obvious clinical symptoms, who elect not to undergo this invasive procedure. However, the prevalence of IgAN is very high in China, and therefore there is a strong need for noninvasive, readily available clinical indicators that better predict IgAN. Our study has shown that the IgA/C3 ratio, a non-invasive index, is of high predictive value for IgAN, especially in patients with primary glomerulonephritis with proteinuria < 1 g/d. Moreover, the predictive value of the IgA/C3 ratio is not affected by kidney function.

In analyzing non-experimental data without a random distribution, differences between groups must be considered, and the PSM approach simplifies this problem by considering only one dimension: the propensity score dimension. Rather than matching all variables, you may compare individual subjects based only on their propensity score [23‐26, 41]. As a novel and effective statistical tool that is increasingly being used in clinical studies, the PSM method is able to correct initial data mismatch between control groups. In the present study, there were significant differences between the two groups in baseline indicators such as age, gender, proteinuria, BMI, and eGFR prior to matching. In order to exclude the influence of confounders, we used the PSM method to balance these non-observational factors to reach more reliable conclusions. In the end, we achieved very similar results in both the full cohort and the PSM cohort, indicating that the clinical significance of these indicators is trustworthy. Using this improved statistical method, we calculated the 95% cut-off point for the IgA/C3 ratio in the non-IgAN population, which was 3.5304, and the diagnostic accordance rate of an IgAN diagnosis among all patients with an IgA/C3 ratio > 3.5304 was as high as 92.02% in the full cohort. Moreover, by multivariate logistic regression analysis of the full cohort, we demonstrated that IgAN was independently correlated with the IgA/C3 ratio. Together this analysis demonstrates that our measured indicators are consistent with observed clinical symptoms.

The present study has several advantages. Firstly, we used a large sample size to prove that the IgA/C3 ratio is of diagnostic value to IgAN. Secondly, a stratified analysis was carried out, controlling for different levels of renal function and levels of proteinuria. The results showed that the IgA/C3 ratio, as a non-invasive index, was of higher predictive value for IgAN in patients with primary glomerulonephritis, particularly those whose proteinuria was less than 1 g/d, and was not affected by renal function. Thirdly, we suggest that this index can be applicable to patients with glomerulonephropathy in remote areas without access to renal biopsy, and to patients with severe renal failure where there is an inability to perform renal biopsy. The limitations to our study include a relatively small number of non-IgAN patients, particular those with normal GFR and proteinuria≤1 g/d, the inclusion of only two units of data, lack of the correlation between of IgA/C3 ratio and histopatological MEST score and the lack of multicenter and follow-up data. Additionally, as a cross-sectional clinical study, we were unable to explain why the IgA/C3 ratio has higher diagnostic efficacy in patients with proteinuria < 1 g/d and the potential role of IgA/C3 ratio as prognostic marker of therapeutic response. Finally, as an observational study, we cannot firmly establish a cause and effect relationship between IgA/C3 ratio and IgAN, thus additional basic research is needed.

In conclusion, the results of the present study provide clear evidence that the IgA/C3 ratio is a valid predictor for diagnosis of IgAN, particularly in patients with proteinuria less than 1 g/d. Additional multicenter large-scale studies are necessary in order to investigate the validity of this biomarker, and determine a standardized cut-off value.