Prognostic significance of albumin-to-globulin ratio in patients with renal cell carcinoma: a meta-analysis

Background Whether the albumin-to-globulin ratio (AGR) predicts the prognosis of renal cell carcinoma (RCC) remains controversial. Herein, we performed a meta-analysis to critically evaluate the relationship between the AGR and RCC prognosis, as well as the association between the AGR and the clinicopathological characteristics of RCC. Methods The PubMed, Web of Science, Embase, and Cochrane Library databases were thoroughly and comprehensively searched from their inception until 24 June 2023. To determine the predictive significance of the AGR, hazard ratios (HRs) and corresponding 95% confidence intervals (CIs) were calculated from the pooled data. The relationship between the AGR and the clinicopathological features of RCC was evaluated by estimating odds ratios (ORs) and 95% CIs in subgroup analyses. Results The meta-analysis included nine articles involving 5,671 RCC cases. A low AGR significantly correlated with worse overall survival (OS) (HR = 1.82, 95% CI = 1.37–2.41, p <0.001) and progression-free survival (PFS) (HR = 2.44, 95% CI = 1.61–3.70, p <0.001). Analysis of the pooled data also revealed significant associations between a low AGR and the following: female sex (OR = 1.48, 95% CI = 1.31–1.67, p <0.001), pT stage T3–T4 (OR = 4.12, 95% CI = 2.93–5.79, p <0.001), pN stage N1 (OR = 3.99, 95% CI = 2.40–6.64, p <0.001), tumor necrosis (OR = 3.83, 95% CI = 2.23–6.59, p <0.001), and Fuhrman grade 3–4 (OR = 1.82, 95% CI = 1.34–2.42, p <0.001). The AGR was not related to histology (OR = 0.83, 95% CI = 0.60–1.15, p = 0.267). Conclusion In patients with RCC, a low AGR strongly predicted poor OS and PFS and significantly correlated with clinicopathological features indicative of disease progression.


Introduction
Renal cell carcinoma (RCC) accounts for approximately 2.2% of all cancer cases, and the median age at diagnosis age is 64 years (1,2). Most (approximately 90%) malignant solid lesions in the kidney are RCCs (3). Globally, there were 431,288 new cases of RCC and 179,368 deaths from RCC in 2020 (1). Since 2012, the number of people developing RCC worldwide has increased by 22% according to the World Cancer Research Fund International (4).
RCC has three major subtypes: clear cell (approximately 80% of cases), papillary (approximately 15%), and chromophobe (approximately 5%) (5,6). Despite significant advances in therapeutic strategies, RCC still has a poor prognosis. The overall 5-year survival rate is 8-59% (7); for advanced disease, it is <20%, and for metastatic RCC (mRCC), the median overall survival (OS) time is 10 months (8,9). These poor outcomes may partly reflect the lack of powerful prognostic indicators (10). Consequently, identifying novel and effective biomarkers for prognosis prediction in patients with RCC in clinical settings is imperative.
Current evidence indicates that systemic chronic inflammation and malnutrition contribute to carcinogenesis and tumor progression (11,12). Albumin (ALB) and globulin (GLB) are two major serum proteins that reflect nutritional and inflammatory status. The ALB-to-GLB ratio (AGR) is an established marker in oncology; it is calculated as follows: AGR = serum ALB/(total serum protein -serum ALB) (13). A low AGR has been widely associated with poor outcomes in various cancers, such as non-small-cell lung cancer (14), esophageal cancer (15), cervical cancer (16), multiple myeloma (17), and pancreatic cancer (18).
Previous studies have explored the prognostic significance of the AGR in patients with RCC (19)(20)(21)(22)(23)(24)(25)(26)(27). However, inconsistent results were obtained: in some studies, a low AGR significantly predicted worse survival (23,26,27), whereas in others, the AGR was unrelated to prognosis (21,24,25). To provide resolution, we performed a meta-analysis that evaluated the relationship between the AGR and RCC prognosis, as well as between the AGR and the clinicopathological characteristics of RCC.

Study guideline
This meta-analysis was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines (28).

Ethics statement
This study used data from previous articles, and approval from an ethics committee or institutional review board was therefore waived.

Search strategy
We thoroughly and comprehensively searched the PubMed, Web of Science, Embase, and Cochrane Library databases from their inception until 24 June 2023. Only English publications were selected. The following combinations of Medical Subject Headings and other terms were used: (albumin to globulin ratio, albumin/ globulin ratio, AGR, or albumin-globulin ratio) and (renal cancer, kidney neoplasm, kidney cancer, renal cell carcinoma, renal carcinoma, or RCC). The references of the retrieved studies were manually examined to identify additional relevant studies.

Study eligibility criteria
We included English-language articles that reported the following: (1) the association between the AGR and survival outcome [e.g., OS, recurrence-free survival, cancer-specific survival (CSS), and progression-free survival (PFS)]; (2) hazard ratios (HRs) with 95% confidence intervals (CIs) or data that allowed their calculation; and (3) the threshold used to stratify a high/low AGR. Additional inclusion criteria were a pathological diagnosis of RCC and measurement of serum ALB and GLB levels before treatment. Meeting abstracts, reviews, case reports, letters, comments, studies with overlapping patients, and animal studies were excluded.

Data collection and quality evaluation
Two researchers (HM and FY) independently screened the retrieved articles and extracted and crosschecked the data. Disagreements between the two investigators were settled through negotiation until a consensus was reached. The data collected in this study included first author, country, publication year, sample size, study period, patient age and sex, number of study centers (single or multiple), metastatic status, treatment, follow-up period, AGR threshold, survival endpoints, type of survival analysis (univariate or multivariate), and HRs with 95% CIs. We used the Newcastle-Ottawa Scale (NOS) (29) to evaluate the quality of the study in three domains: selection (0-4 points), comparability (0-2 points), and outcome assessment (0-3 points), yielding a total score of 0-9. Articles with NOS scores ≥6 were considered high-quality.

Statistical analysis
HRs with 95% CIs were calculated to determine whether the AGR significantly predicted survival outcome. Inter-study heterogeneity was assessed using Higgin's I 2 statistic (30) and Cochran's Q test (31).  To further explore the ability of the AGR to predict RCC prognosis, subgroup analyses stratified by country, sample size, study center number, metastatic status, treatment, AGR threshold, and type of survival analysis were performed. Associations between the AGR and the clinicopathological features of RCC were evaluated by determining odds ratios (ORs) with 95% CIs. Begg's test (32) and Egger's test (33) were used to assess publication bias. Statistical analyses were performed using Stata software, version 12.0 (Stata Corporation, College Station, TX, USA). P values <0.05 indicated statistical significance.

AGR and PFS
Six articles with 4,511 patients examined the relationship between the pretreatment AGR and PFS (19, 21-23, 26, 27). Owing to obvious heterogeneity (I 2 = 75.9%, p = 0.001), we used a random-effects model to collectively analyze the data in these articles. We found that a low AGR strongly predicted worse PFS (HR = 2.44, 95% CI = 1.61-3.70, p <0.001; Figure 3 and Table 3). In subgroup analyses, a low AGR significantly predicted poor PFS regardless of country, sample size, study center number, metastatic status, treatment, cut-off value, or type of survival analysis (Table 3).

Relationship between the AGR and clinicopathological features
The relationship between the pretreatment AGR and the clinicopathological features of RCC was analyzed using data from six studies with 5,219 patients (19-21, 23, 26, 27). The clinicopathological characteristics examined were as follows: sex (female vs male), pT stage (T3-T4 vs T1-T2), pN stage (N1 vs N0), tumor necrosis (present vs absent), histology (clear cell RCC vs non-clear cell RCC), and Fuhrman grade (3-4 vs 1-2). As shown in Table 4 and Figure 4

Publication bias
Begg's and Egger's tests were used to assess publication bias. Funnel plots revealed rough symmetry regarding the distribution of many of the included articles, indicating the absence of obvious publication bias for OS (p = 0.266 and 0.236 for Begg's and Egger's tests, respectively) and PFS (p = 0.260 and 0.087, respectively; Figure 5).

Discussion
Whether the AGR predicts RCC prognosis is unclear, with previous reports providing conflicting results (19)(20)(21)(22)(23)(24)(25)(26)(27). The present meta-analysis, which synthesized data from nine studies with 5,671 RCC cases, revealed a robust association between a low pretreatment AGR and shorter OS and PFS times. As a reflection of the aggressive nature of RCC, a low AGR also correlated with pT stage T3-T4, pN stage N1, tumor necrosis, and Fuhrman grade 3-4. Hence, it might serve as a stable predictor of the short-and longterm prognosis of patients with RCC; notably, patients with low AGRs tended to experience tumor progression and metastasis. To our knowledge, this is the first reported meta-analysis of the significance of the AGR in predicting the prognosis and clinicopathological features of RCC.
The AGR compares ALB and GLB levels, with a low AGR indicating a low ALB and/or high GLB level. The potential FIGURE 2 Forest plots of the prognostic role of AGR for OS in patients with RCC. Forest plots of the prognostic role of AGR for PFS in patients with RCC. mechanisms underlying the correspondence between a low AGR and poor RCC prognosis reflect the anti-oncogenic and prooncogenic actions of ALB and GLB, respectively. Serum ALB, a liver-generated soluble protein, maintains capillary osmotic pressure, removes free radicals from the blood (13), inhibits systemic inflammatory reactions, and serves as a marker of nutritional status (34). Malnutrition and inflammation impede its synthesis, as does interleukin-6 during the acute phase of inflammation in hepatocytes (35). ALB levels have been useful for predicting the outcomes of various cancers (36). The GLB component of the AGR includes diverse proinflammatory proteins, such as immunoglobulins, C-reactive protein,  (37), a-2 macroglobulin, fibrinogen, prothrombin, and serum amyloid A (38,39). Because immunoglobulins are primarily metabolized in the liver, their clearance is impaired in patients with severe hepatic dysfunction, resulting in hyperglobulinemia (40, 41). In support of a potential role of GLB in apoptosis and carcinogenesis, a previous study associated increases in immunoglobulin levels with variations in the gene encoding tumor necrosis factor receptor 13B (42). Therefore, the AGR, which considers both ALB and GLB levels, is a reasonable and reliable prognostic marker. Notably, a recent large-scale multicenter real-world retrospective study used pretreatment clinical characteristics to identify patients with muscle-invasive bladder cancer (MIBC) who could benefit from neoadjuvant combination therapy (43). Based on the combined prognostic efficiency of four hematological indexes [platelet-to-lymphocyte ratio (PLR) and hemoglobin, globulin, and platelet levels], a new indicator, the PLR.GHR (PLR*Globulin/Hemoglobin), was developed for MIBC (43). This finding has important implications in terms of the findings of our meta-analysis on RCC Previous studies have evaluated various hematological parameters as predictors of RCC outcomes. In the report by De Giorgi et al., the systemic immune inflammation index and body mass index independently predicted OS in RCC patients treated with nivolumab (44). In studies of mRCC, a low pretreatment prognostic nutritional index was a potential risk factor after firstline therapy with tyrosine kinase inhibitors (45), and the PLR was an independent indicator of survival in a large cohort (n = 921) (46). Additionally, a recent meta-analysis associated a high pretreatment neutrophil-to-lymphocyte ratio and PLR with progression and mortality in mRCC patients treated with immune checkpoint inhibitors. Collectively, these studies implicate multiple hematological parameters in RCC prognosis (44)(45)(46)(47).
Both units of the AGR (ALB and GLB) are important components of the tumor microenvironment. Therefore, the prognostic capability of the AGR can be influenced by tumor immune markers, such as branched chain aminotransferase 2 (BCAT2) (48), 5 methylcytosine (5mC) (49), and Siglec15 (50), all of which have been shown to s hape the tumor microenvironment (48)(49)(50). The relationship between the AGR and these markers should be investigated in future studies.
Several meta-analyses suggest that a low pretreatment AGR can predict the prognosis of various cancers (51)(52)(53)(54)(55). In a meta-analysis of 3,211 patients with head and neck cancer, a low AGR significantly correlated with poor disease-free survival (DFS), distant metastasis-free survival, OS, T3-T4 status, lymph node metastasis, and stage III-IV disease (51). In a meta-analysis of 7,211 patients with metastatic prostate cancer, the AGR independently predicted PFS and CSS (52). Furthermore, in a meta-analysis of 8,397 patients with colorectal cancer, a low AGR robustly predicted poor OS and DFS/PFS (53). In additional metaanalyses, a low AGR remarkably predicted poor OS and DFS in patients (n = 3,496) with lung cancer (54) and was closely associated with poor OS and DFS/PFS in patients with gastric cancer (12 articles) (55). The present study had some limitations. First, because all included articles were retrospective, inherent heterogeneity existed. However, the appropriate model for analysis of heterogeneous data (the random-effects model) was used. Second, the sample size was relatively small. Although we searched several databases, only nine relevant studies were retrieved. Third, most of the eligible studies were conducted in Asia. Consequently, our findings might be more applicable to Asian vs non-Asian patients with RCC. Owing to these limitations, larger, multi-arm prospective studies are needed to validate the prognostic significance of the pretreatment AGR in patients with RCC.
In conclusion, a low AGR markedly predicted poor OS and PFS in patients with RCC and significantly correlated with clinicopathological features indicative of disease progression. Use of the AGR will aid the identification of high-risk individuals and expedite the development of effective therapeutic strategies for RCC.

Data availability statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.

Author contributions
HM and FY contributed to conception and design of the study. HM and FY were in charge of data collection. HM and FY performed the statistical analysis. HM wrote the first draft of the manuscript. FY edited the manuscript. All authors contributed to the article and approved the submitted version. Publication test by Begg

Publisher's note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.