Prognostic value of lymphocyte to monocyte ratio for metastatic renal cell carcinoma: a systematic review and meta-analysis

Abstract

Objective

To systematically evaluate the prognostic value of lymphocyte-to-monocyte ratio (LMR) for the prognosis of patients with metastatic renal cell carcinoma (mRCC).

Methods

This systematic review and meta-analysis followed the PRISMA 2020 guidelines and was registered in PROSPERO (CRD420251246337). Relevant literature up to November 2025 was systematically searched in PubMed, Embase, Web of Science, and Cochrane Library. Observational studies assessing the relationship between LMR and overall survival (OS) and progression-free survival (PFS) in mRCC patients were included. Two researchers independently screened the literature, extracted data, and assessed the quality using the Newcastle-Ottawa Health Scale (NOS). Pooled hazard ratios (HR) and their 95% confidence intervals (CIs) were calculated using a random-effects model. Heterogeneity was assessed using the I² statistic, and sensitivity analysis and Egger’s test were performed to assess publication bias.

Results

A total of 11 studies involving 1763 patients with mRCC were included. Meta-analysis results showed that lower LMR was significantly associated with shorter OS (HR = 1.64, 95% CI: 1.30–2.07, P < 0.0001) and shorter PFS (HR = 1.33, 95% CI: 1.14–1.56, P = 0.0003). OS analysis showed high heterogeneity (I² = 82%), with sensitivity analysis suggesting the Barua 2019 study was the primary source; while PFS heterogeneity was low (I² = 15%). Egger’s test indicated publication bias in OS (P = 0.04), but the association remained significant after trimming and complement correction (corrected HR = 1.44, 95% CI: 1.10–1.89), indicating robust results.

Conclusion

LMR is significantly associated with OS and PFS in patients with mRCC, and low LMR indicates poor prognosis. Although OS is heterogeneous, the results are robust. As an economical and readily available inflammatory marker, LMR can help with risk stratification and individualized treatment decisions in mRCC, but its standardization of threshold values and its role in assessing prognosis still require further validation through prospective studies.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12885-026-16689-6.

Introduction

Renal cell carcinoma (RCC) is one of the most common malignant tumors of the urinary system, accounting for more than 90% of all kidney tumors [1, 2]. Its incidence is rising globally, especially in developed countries, where the annual incidence is approximately 3–5 cases per 100,000 people [3, 4]. Although early-stage RCC can achieve a good prognosis through surgical resection, approximately 20–30% of patients have metastasized at diagnosis, a condition known as metastatic RCC (mRCC) [5]. mRCC has an extremely poor prognosis, with a 5-year survival rate of less than 10%, and is insensitive to conventional chemotherapy and radiotherapy [6]. While recent advancements in targeted therapies such as tyrosine kinase inhibitors and immune checkpoint inhibitors targeting vascular endothelial growth factor (VEGF) have led to significant differences in patient survival, highlighting the urgency of prognostic assessment [7, 8]. Identification of prognostic factors is crucial for individualized treatment strategies, risk stratification, and clinical decision-making. Traditional prognostic models, such as the International Metastatic Renal Cell Carcinoma Database Consortium (IMDC) score, incorporate clinical and laboratory parameters including anemia, neutrophilia, and thrombocytosis, which are themselves systemic inflammatory markers. However, these composite scores do not directly capture the lymphocyte-to-monocyte axis of immune regulation [6]. Therefore, exploring novel biomarkers, especially blood-based inflammatory markers, has become a research hotspot in recent years.

Inflammation plays a central role in tumorigenesis, development, and metastasis [9]. Cancer-associated inflammation can drive disease progression through mechanisms such as promoting angiogenesis and suppressing immune surveillance [10]. The lymphocyte-to-monocyte ratio (LMR), an easily accessible marker of systemic inflammation, is calculated by dividing the peripheral blood lymphocyte count by the monocyte count [11]. Lymphocytes represent the anti-tumor immune response, while monocytes can differentiate into tumor-associated macrophages, promoting immunosuppression [12]. A decreased LMR value may reflect immunosuppression and is associated with poor prognosis in various cancers, such as lung, gastric, and colorectal cancers [13–15]. However, the prognostic value of LMR in mRCC remains controversial. Existing studies have small sample sizes, high population heterogeneity, and inconsistent results. Some studies have reported that low LMR is associated with shorter overall survival (OS) and progression-free survival (PFS) [16, 17], while other studies have not found a significant association [18]. This inconsistency may stem from study design, population characteristics, or differences in LMR cutoff values, necessitating comprehensive quantification through systematic reviews and meta-analyses.

The purpose of this systematic review and meta-analysis is to comprehensively assess the prognostic value of LMR in mRCC. Specific objectives include: first, calculating the combined hazard ratio (HR) of LMR to OS and PFS; second, assessing and exploring the sources of heterogeneity among studies; and third, examining publication bias and outcome robustness. The results of this study are expected to provide evidence for optimizing prognostic models and promoting personalized treatment. By integrating global research, we hope to resolve existing controversies, promote the application of LMR as a routine prognostic tool, and ultimately improve the management of mRCC patients.

Methods

Literature search

This meta-analysis was conducted in accordance with the PRISMA 2020 statement [19] and was registered in PROSPERO (CRD420251246337). To identify relevant studies assessing the prognostic role of LMR in mRCC, a systematic literature search was performed in PubMed, Embase, Web of Science, and the Cochrane Library up to November 2025. The search query in PubMed is as follows: ((((“Lymphocytes“[Mesh]) OR (((Lymphocyte) OR (Lymphoid Cells)) OR (Lymphoid Cell))) AND ((“Monocytes“[Mesh]) OR (Monocyte))) AND (Ratio)) AND ((“Carcinoma, Renal Cell“[Mesh]) OR (((((((((Renal Carcinoma) OR (Nephroid Carcinoma)) OR (Renal Cell Adenocarcinoma)) OR (Renal Cell Carcinoma)) OR (Renal Cell Cancer)) OR (Renal Adenocarcinoma)) OR (Hypernephroma)) OR (Clear Cell Renal Carcinoma)) OR (Collecting Duct Carcinoma))). Detailed search strategies of all databases are presented in Supplementary Table 1. To ensure completeness, the bibliographies of all included articles were manually screened. The process of article selection and evaluation was conducted independently by two investigators, and any discrepancies were resolved through discussion.

Inclusion and exclusion criteria

Inclusion criteria:

• P: Patients diagnosed with mRCC.

• E: Low LMR.

• C: High LMR.

• O: Any survival outcome, including OS, and PFS, etc.

• S: Study design was cohort or case-control.

Exclusion criteria: Study protocols, unpublished studies, non-original publications such as letters, abstracts, comments, corrections, and replies. Additionally, single-arm studies, review articles, and studies lacking sufficient data—specifically those from which HRs and corresponding 95% confidence intervals (CIs) for survival outcomes could not be directly obtained or calculated—were also excluded.

Data abstraction

Two investigators independently extracted data. Discrepancies were resolved by the third author. The following data were extracted from each eligible study: first author’s name, year of publication, geographic location, study design, population characteristics, sample size, patient age and sex, cut-off values, multivariate adjusted HR and 95% CI of OS, PFS. When essential data were missing or incomplete, corresponding authors were contacted for clarification or additional information.

Quality evaluation

The methodological quality of the included cohort and case-control studies was evaluated using the Newcastle-Ottawa Scale (NOS) [20], with scores ranging from 7 to 9 indicating high quality [21]. Quality assessments were performed independently by two reviewers, and any differences in their judgment were resolved through discussion.

Statistical analysis

The analysis was performed using Review Manager software, version 5.4.1. HR were used for survival data synthesis. Each metric was presented with 95%CIs using a random-effects model. Chi-squared (χ²) test (Cochran’s Q) and inconsistency index (I²) were used to evaluate the heterogeneity of each outcome [22, 23]. Thresholds for high heterogeneity were defined as P < 0.1 and I² > 50%. For results with more than two included studies, sensitivity analysis was conducted to assess each study’s influence on the overall estimate. Potential publication bias was assessed using Egger’s regression test [24] using Stata 15.1 (Stata Corp, College Station, Texas, USA) for outcomes with 10 or more included studies, as the test has limited statistical power with fewer studies. A threshold of P < 0.05 was used to identify significant evidence of publication bias. For results with publication bias, the trim-and-fill method was used to assess its impact on the results. Additionally, according to GRADE, each outcome’s evidence was evaluated and graded as “high”, “moderate”, “low”, or “very low” quality to draw conclusions [25].

Results

Literature retrieval, study characteristics, and quality assessment

The process of literature identification and selection is illustrated in Fig. 1. A comprehensive search across PubMed (n = 79), Embase (n = 174), Web of Science (n = 90), and the Cochrane Library (n = 2) initially yielded a total of 345 records. After duplicate entries were removed, 213 unique titles and abstracts were screened for relevance. In the final analysis, 11 studies encompassing a total of 1763 patients were deemed eligible and included in the meta-analysis [16–18, 26–33]. Table 1 summarizes the key features and methodological quality of each eligible study. The included literature was published between 2017 and 2025. Across the 11 included studies, LMR was dichotomized using various methods, including receiver operating characteristic curve-derived optimal cutoffs, median split, and predefined thresholds from previous literature. The cut-off values ranged from 1.1 to 3.84, reflecting substantial variability in study populations, laboratory assays, and statistical approaches.

Fig. 1.

Flowchart of the systematic search and selection process

Table 1.

Characteristics and quality assessment of included studies

Study	Country	Study design	Population	No. of patients	Gender		Mean/median age	LMR cut-off	NOS score
					Male	Female
Barua 2019 [26]	India	Retrospective cohort	Metastatic Non-Clear Cell Renal Cell Carcinoma	31	21	10	62	2.61	7
Bolzacchini 2022 [27]	Italy	Retrospective cohort	Metastatic renal cell carcinoma treated with sunitinib	100	70	30	63.2	3	7
Fukuda 2018 [28]	Japan	Retrospective cohort	Metastatic renal cell carcinoma treated with cytoreductive nephrectomy	152	109	43	64	3.23	7
Gu 2017 [29]	China	Retrospective cohort	Metastatic Clear Cell Renal Cell Carcinoma	145	118	27	NA	3	7
Martini 2020 [30]	USA	Retrospective cohort	Metastatic Renal Cell Carcinoma Treated with Immune Checkpoint Inhibitors	100	66	34	NA	1.1	7
Numakura 2024 [31]	Japan	Retrospective cohort	Metastatic renal cell carcinoma treated using nivolumab plus ipilimumab therapy	112	93	19	68	3	7
Pacholczak-Madej 2025 [17]	Poland	Retrospective cohort	Metastatic renal cell carcinoma treated with nivolumab and ipilimumab	116	87	29	64	2.83	7
Wang Jichen 2023 [18]	China	Retrospective cohort	Metastatic renal cell carcinoma	198	154	44	57	3.84	7
Wang Zhaojuan 2023 [32]	China	Retrospective cohort	Patients Treated With Tyrosine Kinase Inhibitors for Metastatic Renal Cell Carcinoma	210	155	55	NA	3.33	8
Yildirim 2025 [16]	USA	Retrospective cohort	Patients with metastatic renal cell carcinoma treated with immune checkpoint inhibitors	401	283	118	NA	1.59	7
Yılmaz 2021 [33]	Turkey	Retrospective cohort	Patients with metastatic renal cancer	198	135	63	NA	2.43	7

NA indicates that the mean or median age was not reported in the original study by the primary authors

This study systematically reviewed 11 included cohort studies, all of which received high scores of 7 to 8, indicating good overall research quality. In terms of specific evaluation dimensions, all studies performed excellently in “Selection,” fully encompassing representative populations, clearly defined selection criteria for the non-exposed group, and the diagnostic process for exposure factors. Regarding “Outcome Assessment,” all studies demonstrated blinded outcome assessment and sufficient follow-up time (median ≥ 1 year), with good follow-up completeness. However, in terms of “Comparability,” the vast majority of studies failed to effectively control for significant confounding factors. In summary, while these studies demonstrated high reliability in their design and data collection, they generally exhibited limitations in controlling for confounding factors, suggesting that we should carefully consider the potential impact of bias on the conclusions when interpreting the results of this meta-analysis. Detailed NOS score results are shown in Supplementary Table 2.

Prognostic value of LMR for OS

OS results were synthesized from 10 studies. The meta-analysis indicated that a lower LMR was significantly associated with shorter OS (HR: 1.64; 95% CI: 1.30, 2.07; P <0.0001) (The prediction interval is 0.63, 42.6), with a substantial heterogeneity (I² = 82%, P <0.00001) (Fig. 2) (Table 2).

Fig. 2.

Forest plots of the prognostic value of LMR for OS

Table 2.

Summary of meta-analysis and evidence grading results

Outcomes	No. of studies	Pooled HRs	95%CI	P value of Z-test	I2	Prediction interval	Egger’s test P value	GRADE
OS	10	1.64	1.30, 2.07	< 0.0001	82%	0.63, 42.6	0.04	Very low
PFS	7	1.33	1.14, 1.56	0.0003	15%	1.04, 1.71	NA	Low

Prognostic value of LMR for PFS

PFS results were synthesized from 7 studies. The meta-analysis indicated that a lower LMR was significantly associated with shorter PFS (HR: 1.33; 95% CI: 1.14, 1.56; P = 0.0003) (The prediction interval is 1.04, 1.71), with no substantial heterogeneity (I² =15%, P = 0.31) (Fig. 3) (Table 2).

Fig. 3.

Forest plots of the prognostic value of LMR for PFS

Publication bias

Potential publication bias regarding the prognostic significance of LMR for OS was evaluated using Egger’s regression test and funnel plot analysis. The results indicated the presence of significant publication bias for OS (Egger’s test P = 0.04, Fig. 4A). The trim-and-fill analysis imputed one hypothetical missing study to correct for the asymmetry suggested by Egger’s test. After adjustment, the corrected pooled HR for OS remained statistically significant (HR 1.44; 95% CI: 1.10–1.89) (Fig. 4B). However, persistence of statistical significance after trim-and-fill adjustment does not demonstrate the absence of publication bias, nor does it establish that the pooled estimate was unaffected by such bias. The trim-and-fill method makes strong assumptions about the symmetry of the funnel plot and may not fully correct for the underlying bias. Therefore, the presence of publication bias should be considered when interpreting the OS results.

Fig. 4.

(A) Funnel plot of OS and (B) trimming and filling funnel plot of OS

Sensitivity analysis

A sensitivity analysis was performed to assess each study’s effect on the overall HR for the prognostic value of LMR for OS, and PFS using a leave-one-out approach. The overall HRs for OS (Fig. 5A) and PFS (Fig. 5B) remained consistent following the sequential exclusion of individual studies. Furthermore, it is worth noting that when the data from Barua 2019 were excluded, the heterogeneity of OS decreased from 82% to 45%, suggesting that this study was one of the main reasons for the significant heterogeneity of OS.

Fig. 5.

Sensitivity analysis of the prognostic value of LMR for (A) OS, (B) PFS

GRADE

Due to significant heterogeneity and publication bias, OS was rated as very low-quality evidence. PFS, on the other hand, was rated as low-quality evidence because no significant heterogeneity was detected (Table 2).

Discussion

This study, through systematic review and meta-analysis, synthesized data from 11 studies involving 1763 patients with mRCC, providing the first comprehensive assessment of the prognostic value of LMR for OS and PFS. Results showed that low LMR was significantly associated with shorter OS (HR: 1.64; 95% CI: 1.30–2.07; P < 0.0001) and shorter PFS (HR: 1.33; 95% CI: 1.14–1.56; P = 0.0003), suggesting that LMR is a potential prognostic biomarker for mRCC. High heterogeneity was observed in the OS analysis (I²=82%), while PFS heterogeneity was low (I²=15%), suggesting that the impact of LMR on different survival endpoints may be modulated by various factors. Sensitivity analysis confirmed the robustness of the results, and publication bias correction using the trimming and filling method did not alter the significance of OS, enhancing the reliability of the conclusions. These findings are consistent with the cancer inflammation theory, suggesting that reduced LMR reflects an immunosuppressive microenvironment, which may promote mRCC progression [34].

The strong association between LMR and OS (HR = 1.64) highlights the central role of inflammatory immune balance in the prognosis of mRCC. Lymphocytes represent adaptive immunity, and their reduced number may weaken the anti-tumor response [35], while an increase in monocytes often indicates the accumulation of myeloid-derived suppressor cells, promoting tumor tolerance [36]. In the context of mRCC, this imbalance may be amplified by treatment modality. In the studies included in this meta-analysis, patients received sunitinib, immune checkpoint inhibitors, or cytoreductive therapy, which may modulate inflammatory cell dynamics. OS showed high heterogeneity (I²=82%), with sensitivity analysis showing the Barua 2019 study as the main source of heterogeneity. This study focused on non-clear cell mRCC, a unique population that may explain some of the variation. Conversely, the association between PFS and LMR was more consistent (HR = 1.33, I²=15%), suggesting that the early impact of LMR on disease progression may be more stable and unaffected by long-term survival competition risks.

An intriguing finding of our study is the marked discrepancy in heterogeneity between OS (I²=82%) and PFS (I²=15%). We hypothesize that this difference may be explained by the differential susceptibility of these two endpoints to post-progression confounding. OS is influenced by subsequent lines of therapy, supportive care, and competing causes of mortality, all of which vary considerably across studies and clinical settings. In contrast, PFS is more directly tied to the efficacy of the initial treatment and is less affected by downstream clinical decisions [37]. Therefore, the high OS heterogeneity may partly reflect differences in post-progression management across the included cohorts rather than genuine variability in the LMR-prognosis association. Additionally, the longer follow-up periods required for OS assessment introduce greater opportunities for time-varying confounding, such as changes in LMR over the disease course, which may further attenuate the stability of the OS estimate.

Compared to other systemic inflammatory biomarkers evaluated in mRCC, LMR offers a distinct biological rationale by capturing the balance between adaptive anti-tumor immunity (lymphocytes) and innate immunosuppressive activity (monocytes). Previous meta-analyses have reported that elevated NLR and elevated PLR are also associated with poor prognosis in mRCC [38]. The prognostic magnitude of LMR observed in our study is comparable to these markers. However, LMR may provide complementary information, as it reflects a different immunological axis—specifically, the monocyte-macrophage lineage, which plays a central role in tumor immune evasion through mechanisms such as PD-L1 expression and immunosuppressive cytokine secretion. Future studies should evaluate whether a composite inflammatory score incorporating multiple biomarkers offers superior prognostic discrimination compared to any single marker alone.

Notably, this study presents a significant methodological and conclusive comparison with the meta-analysis published by Garcia-Rojo et al. in 2021 [39] for non-metastatic renal cell carcinoma. That study, focusing on seven observational studies of non-metastatic patients, found that low LMR was only significantly associated with poor CSS (HR = 3.04), but not with overall survival or disease-free survival, and its LMR cutoff value showed considerable variability (2.5-5.0). In contrast, this study, by including 11 studies, not only confirmed the independent prognostic value of LMR for overall survival (HR = 1.64) and progression-free survival (HR = 1.33) in mRCC, but also identified the Barua 2019 study as the primary source of heterogeneity through sensitivity analysis. This provides a new direction for the application of LMR in precision medicine for mRCC. Furthermore, this study employed a uniform random-effects model and pruning correction, effectively controlling for the effects of heterogeneity and publication bias, whereas the Garcia-Rojo study only used a fixed-effects model and did not adequately discuss the sources of heterogeneity. These methodological optimizations and in-depth analyses enhance the clinical applicability and generalizability of this study’s conclusions, providing a higher level of evidence supporting the application of LMR as a prognostic marker for metastatic renal cell carcinoma.

Although subgroup analyses by treatment type, histology, or LMR cut-off values were precluded by the limited number of included studies and incomplete reporting across studies, several plausible sources of the observed heterogeneity deserve consideration. First, treatment modality varied considerably across studies, with patients receiving sunitinib, immune checkpoint inhibitors, or cytoreductive therapy, each of which may differentially modulate inflammatory cell dynamics and thus affect the LMR-prognosis association. Second, the inclusion of both clear-cell and non-clear-cell histology, as exemplified by the Barua 2019 study, introduces histological heterogeneity that may contribute to the observed variability. Third, the wide range of LMR cut-off values across studies reflects differences in study populations and laboratory methodologies, representing an additional source of between-study variation. Future meta-analyses based on individual patient data are warranted to systematically investigate these potential effect modifiers.

While our findings support the prognostic value of LMR in mRCC, the question of whether LMR provides incremental prognostic value beyond established models such as the IMDC score remains unanswered by the current evidence. The included studies did not consistently report whether LMR remained an independent predictor after adjustment for IMDC components in multivariable analyses, and no study directly compared the discriminative ability of models with and without LMR. Therefore, rather than advocating for the replacement or immediate incorporation of LMR into clinical guidelines, we view LMR as a candidate biomarker that warrants further validation in prospective cohorts designed specifically to assess its added prognostic utility beyond existing risk stratification tools. Future studies should employ metrics such as the C-statistic or net reclassification improvement to quantify the incremental value of adding LMR to the IMDC score.

Clinically, this study demonstrates that LMR, as an inexpensive and easily reproducible blood routine indicator, may serve as a candidate marker for prognostic assessment. In resource-constrained healthcare institutions, LMR can supplement clinical scoring, helping to identify high-risk patients for intensified monitoring and risk-adapted follow-up strategies. For example, patients with low LMR may represent a higher-risk population warranting closer surveillance, although whether they derive differential benefit from specific therapies remains to be determined in prognostic biomarker studies. However, standardization of LMR cutoff values is a prerequisite for its application. This study shows a wide cutoff range, requiring future prospective studies to determine the optimal cutoff point or the development of dynamic monitoring models. Furthermore, the interaction between LMR and emerging biomarkers such as PD-L1 expression or mutation burden warrants further exploration to construct multidimensional prognostic tools. These results support the inclusion of LMR in clinical guidelines, but real-world validation is needed.

The limitations of this study include: First, all studies were retrospective, which, while of high quality, contained residual confounding; second, an important limitation that warrants emphasis is the lack of access to individual patient data (IPD), which precluded adjustment for key confounders such as prior nephrectomy, performance status, specific tyrosine kinase inhibitor or immune checkpoint inhibitor regimen, and number of prior lines of therapy. An IPD meta-analysis would represent the gold standard approach, as it would allow for standardized cut-off determination, time-to-event analysis with appropriate censoring, and multivariable adjustment for patient-level covariates. The aggregate data meta-analysis approach used in the present study, while informative, cannot overcome these inherent limitations. Future collaborative efforts to pool individual patient data from existing cohorts are strongly encouraged to validate and extend our findings.; third, some studies had incomplete data, potentially introducing reporting bias; and fourth, the literature search ended in November 2025, meaning the latest evidence may not have been covered. Finally, an important limitation that warrants explicit emphasis is the substantial treatment heterogeneity across the included studies. Patients in different studies received markedly different systemic therapies, including sunitinib monotherapy, nivolumab plus ipilimumab combination immunotherapy, and tyrosine kinase inhibitors, reflecting the evolving treatment landscape of mRCC over the study period. Since treatment modality may directly influence both peripheral blood inflammatory biomarker levels and patient prognosis, this heterogeneity introduces a potential confounder that cannot be fully accounted for in the present meta-analysis. Specifically, immunotherapies may differentially modulate lymphocyte and monocyte counts compared to targeted therapies, potentially altering the LMR-prognosis association in ways that are treatment-dependent. This consideration further tempers the generalizability of our findings and underscores the need for future studies to evaluate the prognostic value of LMR within homogeneous treatment cohorts. Future research should focus on prospective designs, standardizing LMR measurements, and integrating multi-omics data, such as genomic or inflammatory factor panels, to improve prognostic accuracy. Simultaneously, future studies should specifically evaluate whether LMR possesses prognostic value for treatment response to ICIs or TKIs, as this question is distinct from the prognostic association established in the present analysis.

Conclusion

This meta-analysis confirms that low LMR is a potential prognostic factor for OS and PFS in mRCC patients, with robust results and some degree of clinical applicability. Despite heterogeneity, LMR, as an inflammatory marker, holds promise for optimizing prognostic stratification. We recommend that future studies standardize LMR assessment and validate its added value in prospective studies to advance precision medicine for mRCC.

Abbreviations

CI

Confidence Interval

CSS

Cancer-Specific Survival

HR

Hazard Ratio

IMDC

International Metastatic Renal Cell Carcinoma Database Consortium

LMR

Lymphocyte-to-Monocyte Ratio

mRCC

metastatic Renal Cell Carcinoma

NLR

Neutrophil-to-Lymphocyte Ratio

NOS

Newcastle-Ottawa Scale

OS

Overall Survival

PFS

Progression-Free Survival

PLR

Platelet-to-Lymphocyte Ratio

RCC

Renal Cell Carcinoma

Authors’ contributions

X.Z. contributed to the study conception and design, performed the literature search, data extraction, and statistical analysis, and drafted the manuscript. Y.W. participated in literature screening, data extraction, risk of bias assessment, and helped in data interpretation and manuscript revision. Z.Z. supervised the entire project, contributed to the study design and data interpretation, and critically reviewed and edited the manuscript. All authors read and approved the final version of the manuscript.

Funding

This research was not supported by any funding.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.