Diagnostic utility of serum and urine biomarkers in idiopathic membranous nephropathy: a systematic review and meta-analysis

Membranous nephropathy is an autoimmune nephropathy that is one of the most common pathological types of nephrotic syndrome. It is important to find and apply specific biomarkers for the noninvasive diagnosis of idiopathic membranous nephropathy (IMN). However, there are limited data about their diagnostic value. Therefore, an overall meta-analysis helps to identify effective biomarkers for the clinical diagnosis of IMN. A systematic literature search was carried out in PubMed, Embase, Cochrane and Web of Science from inception until December 31, 2020. Two researchers searched for studies that met the inclusion criteria. The results of the joint study were expressed in terms of sensitivity and specificity. The meta-analysis included 24 studies with biomarkers for the clinical diagnosis of IMN, including antibody against phospholipase A2 receptor (PLA2R-AB), antibody against thrombospondin type I domain-containing 7A (THSD7A-AB), lysosome membrane protein-2 (LIMP-2) and circular RNAs. The diagnostic efficiency of PLA2R-AB for IMN had a combined sensitivity of 60% and a combined specificity of 100%. The diagnostic efficiency of THSD7A-AB for IMN had a combined sensitivity of 3% and a combined specificity of 99%. The diagnostic efficiency of urinary LIMP-2 for IMN was 100%, and the specificity was 100%. The diagnostic efficiency of exosomal circRNAs for IMN was 100%, and the specificity was 100%. This meta-analysis shows that PLA2R-AB and THSD7A-AB are of important diagnostic value for IMN. More studies are needed in the future to reveal the diagnostic value of LIMP-2 and circRNAs for IMN.


Introduction
Membranous nephropathy (MN) is the most common cause of adult nephrotic syndrome [1]. Approximately 20% of MN patients will progress to end-stage renal disease, and approximately 10% of them will die within 5 to 10 years [2,3]. MN can be divided into idiopathic membranous nephropathy (IMN) and secondary membranous nephropathy (SMN). Approximately 75% of MN patients have idiopathic membranous nephopathy (IMN), while 20%-25% of patients are secondary to different diseases, such as autoimmune diseases, infection, drugs, and malignancy [4].
In the past 10 years, the incidence of IMN has increased significantly, and it has been the main pathological type of primary glomerular disease [5,6]. At present, the diagnosis of IMN mainly depends on kidney biopsy. Although kidney biopsy is the gold standard for diagnosing IMN, there are many potential complications in this method, such as perirenal hematoma, infection and other organ damage. Second, some patients cannot undergo renal biopsy, including isolated kidneys, abnormal coagulation function, hypertension dissatisfied with drug control and mental illness. Therefore, we have been committed to finding reliable biomarkers to guide clinical diagnosis through simple and noninvasive technology.
In recent years, several biomarkers in serum, PLA 2 R, THSD7A and IgG4 antibodies, have been assessed for their clinical significance in diagnosing idiopathic membranous 1 3 nephropathy [7][8][9]. However, the current research still has some limitations. PLA 2 R and THSD7A are a kind of macromolecular transmembrane glycoproteins, which are mainly expressed on the surface of podocytes and participate in inducing the humoral immune response dominated by IgG4. PLA 2 R and THSD7A are major autoantigens in IMN, but we can measure circulating autoantibodies to them. PLA 2 R-AB is the most commonly used method for the diagnosis of IMN, and the clinical value of other serum biomarkers still needs to be further explored. There are few studies on urine biomarkers, such as lysosome membrane protein-2 (LIMP-2) and circular RNAs [10,11], but they have broad prospects and need to be confirmed by large, multicenter studies.
In this article, we performed the first systematic review and meta-analysis of serum and urine biomarkers in IMN patients, with the hope of promoting clinical diagnosis through noninvasive techniques.

Data sources and search strategy
Two researchers, Gao and Zhao, conducted a systematic review of qualified articles on PubMed, Embase, Cochrane and Web of Science from the beginning until December 31, 2020. The search terms were "idiopathic membranous nephropathy and (phospholipase A2 receptor or PLA 2 R or the thrombospondin type I domain-containing 7A or THSD7A or IgG4 or lysosome membrane protein-2 or LIMP-2 or circular RNAs)". The literature search was limited to human studies and was published in English.

Study selection
We would include the original observational studies that reported the biomarkers of IMN and contained disease and healthy control groups. The disease control groups included SMN, other glomerular disease or other non-renal related diseases control groups. These were exclusion criteria: (1) IMN biomarkers measured in animal models; (2) cadaver specimens; (3) in vitro data; (4) no healthy controls; (5) complications with other serious diseases or complications.

Data extraction and quality assessment
Two researchers (D.G. and Z.Z.) extracted data independently from all eligible initial documents. Disagreements were discussed and resolved by a third person's point of view (L.L.). The extracted information included the year of article publication, author, country, sample type, type of markers, experimental method, numbers of case groups, control groups, true positive (TP), false positive (FP), false negative (FN), and true negative (TN) results in each included study. Two authors (Gao and Lu) assessed the quality of the included studies using the updated Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool [12].

Statistical analysis
All of the data were analyzed using Review Manager 5.3 and Stata MP 16.0 (Multiprocessor computers) software. TP, FP, FN and TN were used to describe various indicators in the studies. According to the Cochrane Handbook, I 2 is divided into 0.25, 0.50 and 0.75, representing mild, moderate and high heterogeneity, respectively [13]. When P < 0.05 and I 2 > 50%, we used the random effect model. When P > 0.05 and I 2 < 50%, we used the fixed effect model [14]. The results of the combination of the studies were expressed by sensitivity, specificity, PLR, NLR and DOR. Forest maps were used to describe the 95% CI (95% confidence interval) sensitivity and specificity in the study. Deeks' Funnel Plot Asymmetry Test was used to reflect literature publication bias.

Search results and study characteristics
We obtained 1003 records from the PubMed database and 872 records from the Web of Science, Embase and Cochrane databases. Excluding duplicated articles, 1103 articles remained. We browsed the titles and abstracts of the articles, because they were review articles, or irrelevant to the current analysis and 207 articles remained. 114 articles were excluded because of insufficient data. 76 articles were excluded for reasons (49 were mainly about IMN therapy, 8 did not provide enough data, 6 did not meet the accuracy of the test, 13 were without healthy controls). Finally, the metaanalysis included 17 articles, including 24 studies. Article selection flow chart is shown in Fig. 1.
The characteristics of the included studies are shown in Table 1. The total population of the studies was 7562. The studies included 20 serum samples and 4 urine samples. Biomarkers included PLA 2 R-AB, THSD7A-AB, LIMP-2 and circular RNA in exosomes. PLA 2 R-AB was detected by Western blotting (WB) in three studies, by enzymelinked immunosorbent assay (ELISA) in seven studies, by immunofluorescence test (IFT) in one study, by indirect immunofluorescence cell-based assay (IIF-CBA) in two studies, by indirect immunofluorescence (IIF) in two studies and by time-resolved fluoroimmunoassay (TRFIA) in one study. THSD7A-AB was detected by WB in two studies and by ELISA in three studies. LIMP-2 was detected by Proteomics. Circular RNAs in exosomes were detected by reverse transcription polymerase chain reaction (RT-PCR) followed by quantitative PCR (qPCR) in two studies. 15 studies used SMN patients as controls, and 14 studies used other glomerular disease patients as controls.

Quality evaluation
The quality evaluation of the selected studies was based on the QUADAS-2, which is shown in Fig. 2. Overall, the quality evaluation of the included studies was reliable, but 11 studies had unclear risks in terms of flow and timing, and 2 studies had higher risks in terms of flow and timing. At the same time, 6 studies were unclear on the risks of the index test.

Predicted posterior probability of PLA 2 R-AB and THSD7A-AB in IMN
As shown in Fig. 7, the pre-test probability of PLA 2 R-AB was 20%, and the post-test probability of PLA 2 R-AB was 97%. The pre-test probability of THSD7A-AB was 20%, and the post-test probability of THSD7A-AB was 50%.   This means that PLA 2 R-AB and THSD7A-AB can improve the diagnosis of IMN.

Subgroup and sensitivity analysis of PLA 2 R-AB
The causes of heterogeneity were analyzed by subgroup analysis. As shown in Table 2, the diagnostic accuracy rate of PLA 2 R-AB testing in Asia was higher than that in Europe. There were also other factors, such as method, sample, controls and sample size.     WB  3 -----IFT  1 -----IIF-CBA  2 -----TRFIA  1 -----IIF  2 -----ELISA  7 Table 3 shows that the diagnostic accuracy rate of THSD7A-AB in serum is higher than that in urine.

Publication bias evaluation
The publication bias of the included studies was evaluated by Deeks' funnel plot asymmetry test. As shown in Fig. 8, the results showed that the PLA 2 R-AB (P = 0.80) and THSD7A-AB (P = 0.61) studies had no publication bias. P < 0.05 indicates publication bias.

Discussion
This systematic review and meta-analysis focused on the diagnostic value of serum and urine biomarkers in IMN. At the same time, this is the first meta-analysis for the diagnostic value of different biomarkers of IMN. There was a meta-analysis of the diagnostic value of PLA 2 R-AB and THSD7A-AB separately. In this meta-analysis, the study group included healthy controls, and the criteria for inclusion in the literature were different from those of previous meta-analyses. The specimen type was obtained from serum and urine. There were several biomarkers (PLA 2 R-AB THSD7A-AB, LIMP-2 and circRNAs) that met the inclusion criteria of the study.
In 2009, Beck [15] found that PLA 2 R is specific to the antigen of adult MN, and its specific PLA 2 R antibody was a serum biomarker for detecting IMN, with high sensitivity and specificity. We included 16 studies about the diagnostic value of PLA 2 R-AB that met the inclusion criteria. The sensitivity was 60% (95% CI 53%-67%), and the specificity was 100% (95% CI 97%-100%). The AUC was 0.81 (95% CI 0.77-0.84). Serum PLA 2 R-AB testing is an important clinical diagnostic value of IMN. That is consistent with the research of Hu [30].
Therefore, there was a high level of heterogeneity in the sensitivity of our meta-analysis (I 2 = 88.47%), probably due to the region of studies, test method, specimen type, control group classification and sample size. Then, subgroup analysis further explored the source of heterogeneity. In our metaanalysis, studies were mainly distributed in Asia, followed by Europe and only America. More studies that meet the inclusion criteria are needed in the future. Then, detection methods and the grouping of studies can lead to sources of heterogeneity. Regarding the control group of the studies, we were included in the IMN and contained healthy controls, which was different from the control group of other studies. Studies in other meta-analyses may not have a healthy control group. However, we think it is necessary to include a healthy control group in the study and play the role of disease screening [31].  THSD7A is structurally similar to PLA 2 R, which has been determined to be the second autoantigen of IMN in adults [19]. We included 5 studies about the diagnostic value of THSD7A-AB that met the inclusion criteria. The sensitivity was 3% (95% CI 1%-5%), and the specificity was 99% (95% CI 97%-100%). The results are consistent with the research of Liu [7]. Although not sensitive enough, the diagnosis of IMN is very specific. The prevalence of THSD7A-AB in PLA 2 R-AB negative patients was higher than that in IMN patients [32]. THSD7A-AB testing is important for the clinical diagnostic value of IMN. In this meta-analysis, the small number of studies on THSD7A-AB explored the source of heterogeneity. We need more research on the diagnostic value of THSD7A-AB in IMN.
Noninvasive diagnosis of IMN was performed according to the actual clinical needs of patients, we first conducted a systematic meta-analysis, and reviewed the diagnostic efficiency of PLA 2 R-AB and THSD7A-AB for IMN patients without publication bias.
In our meta-analysis, LIMP-2 in urine and circular RNAs in exosomes had important clinical value in the diagnosis of IMN, although there were few articles included. They were highly specific and sensitive by proteomics.
In conclusion, this meta-analysis shows that PLA 2 R-AB and THSD7A-AB are of important diagnostic value for IMN. Future studies are needed to uncover the diagnostic value of LIMP-2 and circular RNAs for IMN.
Author contributions All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by MT, LL and ZZ. The first draft of the manuscript was written by DG, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Funding We received no financial support from any individual or organization.
Data availability All relevant data are within the paper.

Conflict of interest
The authors have no relevant financial or non-financial interests to disclose. The authors have no competing interests to declare that are relevant to the content of this article. All authors certify that they have no affiliations with or involvement in any organization or entity with any financial interest or non-financial interest in the subject matter or materials discussed in this manuscript. The authors have no financial or proprietary interests in any material discussed in this article.
Ethical approval For this type of study, ethical approval is not required.
Informed consent For this type of study, formal consent is not required.