Vempati R, Damarlapally N, Vasudevan SS, Patel V, Banda P, Mourad D, Polamarasetty H, Mathur G, Khan A, Desai R, Ratnani I, Surani S. Association of neutrophil-lymphocyte ratio with cardiovascular and all-cause mortality in patients receiving chronic hemodialysis: Systematic review and meta-analysis. World J Methodol 2025; 15(4): 107468 [DOI: 10.5662/wjm.v15.i4.107468]
Corresponding Author of This Article
Salim Surani, MD, Professor, Department of Medicine and Pharmacology, Texas A&M University, 40 Bizzell Street, College Station, TX 77843, United States. srsurani@hotmail.com
Research Domain of This Article
Medicine, General & Internal
Article-Type of This Article
Meta-Analysis
Open-Access Policy of This Article
This article is an open-access article which was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/
World J Methodol. Dec 20, 2025; 15(4): 107468 Published online Dec 20, 2025. doi: 10.5662/wjm.v15.i4.107468
Association of neutrophil-lymphocyte ratio with cardiovascular and all-cause mortality in patients receiving chronic hemodialysis: Systematic review and meta-analysis
Author contributions: Desai R conceptualized and designed the study, developed the methodology and software, performed formal analysis, provided resources, participated in writing and editing the manuscript, and managed project administration; Surani S and Ratnani I conducted formal analysis, provided resources, reviewed and edited the manuscript, and managed project administration; Vempati R and Patel V contributed to methodology, performed screening and feasibility assessments, conducted data curation and data extraction, drafted the original manuscript, reviewed and edited the manuscript, and supervised the project; Damarlapally N participated in screening and feasibility assessments, performed data extraction, and contributed to writing the original draft and reviewing and editing the manuscript; Vasudevan SS performed statistical analysis and contributed to writing the original draft and reviewing and editing the manuscript; Banda P contributed to data extraction, writing the original draft, and reviewing and editing the manuscript; Mourad D and Polamarasetty H participated in data extraction and manuscript review and editing; Mathur G and Khan A contributed to data extraction; All authors have read and approved the final manuscript.
Conflict-of-interest statement: None of the authors has any conflict of interest to disclose.
PRISMA 2009 Checklist statement: The authors have read the PRISMA 2009 Checklist, and the manuscript was prepared and revised according to the PRISMA 2009 Checklist.
Open Access: This article is an open-access article that was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution NonCommercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: https://creativecommons.org/Licenses/by-nc/4.0/
Corresponding author: Salim Surani, MD, Professor, Department of Medicine and Pharmacology, Texas A&M University, 40 Bizzell Street, College Station, TX 77843, United States. srsurani@hotmail.com
Received: March 25, 2025 Revised: April 28, 2025 Accepted: June 18, 2025 Published online: December 20, 2025 Processing time: 133 Days and 9.6 Hours
Abstract
BACKGROUND
The neutrophil-lymphocyte ratio (NLR) has been proposed as a potential prognostic marker for mortality outcomes in various conditions, yet its association with chronic hemodialysis (HD) remains underexplored. We aim to study its utility by conducting a meta-analysis of this specific population.
AIM
To determine whether elevated NLR is associated with all-cause mortality (ACM) and cardiovascular mortality (CVM) in patients undergoing chronic HD.
METHODS
A comprehensive search from PubMed, Google Scholar, and Scopus identified studies showing the association between NLR and mortality outcomes in patients with chronic HD. Random-effects models with 95%CIs were employed to pool adjusted hazard ratios (aHRs), odds ratios (ORs), and I² statistics for evaluating the heterogeneity of findings. Leave-one-out sensitivity and meta-regression analyses assessed changes in overall effects and identified confounders, respectively. The Joanna Briggs Institute (JBI) tool was used to assess the quality of studies.
RESULTS
19 studies comprising 9047 patients with a mean age of 59.5 ± 5.86 years and a mean follow-up duration of 46.7 months were included in our study. Our meta-analysis revealed a significant association between NLR > 2.5 and increased risks of ACM (aHR: 1.25, 95%CI: 1.14-1.37, P < 0.0001) and CVM (aHR: 1.24, 95%CI: 1.02-1.49, P = 0.03). Studies reporting outcomes in OR reported similar findings for ACM (OR: 4.59, 95%CI: 1.74-12.11, P = 0.002) and CVM (OR: 1.11, 95%CI: 1.01-1.23, P = 0.03). Sensitivity analysis revealed no variations. Meta-regression revealed increasing male proportion is positively associated with ACM. Pooled area under the curve (AUC) was 0.71 (95%CI: 0.63-0.80, P < 0.0001). The JBI tool revealed high-quality studies.
CONCLUSION
This meta-analysis suggests that elevated NLR may serve as a useful prognostic marker for ACM and CVM in patients on chronic HD and can be useful in planning for the prevention of mortality-related strategies.
Core Tip: This systematic review and meta-analysis evaluate the prognostic value of the neutrophil-to-lymphocyte ratio (NLR) in patients undergoing chronic hemodialysis (HD). Elevated NLR was significantly associated with increased all-cause and cardiovascular mortality in this population. By pooling adjusted hazard ratios and assessing heterogeneity and diagnostic performance via the area under the curve, the study highlights NLR as a potentially valuable and accessible biomarker for risk stratification in HD patients. Findings support its clinical use, while emphasizing the need for standardized cutoffs and prospective validation.
Citation: Vempati R, Damarlapally N, Vasudevan SS, Patel V, Banda P, Mourad D, Polamarasetty H, Mathur G, Khan A, Desai R, Ratnani I, Surani S. Association of neutrophil-lymphocyte ratio with cardiovascular and all-cause mortality in patients receiving chronic hemodialysis: Systematic review and meta-analysis. World J Methodol 2025; 15(4): 107468
Chronic kidney disease (CKD) is characterized by irreversible structural or functional abnormality and a decrease in the glomerular filtration rate (GFR) to less than 60 mL/minute per 1.73 m², along with or without biomarkers of renal damage lasting for more than or equal to three months[1]. The majority of them receive hemodialysis (HD). Still, it can progress to end-stage renal disease (ESRD) and necessitates renal replacement therapy, which includes HD, peritoneal dialysis (PD), and kidney transplantation[2]. Cardiovascular events and mortality are more likely to occur in these patients[2]. Based on data from the National Health and Nutrition Examination Survey from 2017 to March 2020, the United States Renal Data System's 2023 annual data report states that 14.0% of United States adults had an estimated GFR (eGFR) of less than 60 mL/minute/1.73 m², a urinary albumin-creatinine ratio (ACR) of at least 30 mg/g, or both. In the United States, the overall population's prevalence of CKD increased from 12.5% in 2009-2012 to 14% by March 2017-2020[3]. Between 2001 and 2019, the incidence of ESRD rose by 37.8%; in 2020, it decreased by 3.1%, and in 2021, it sharply increased. In 2021, home HD was initiated for 0.4% of incident ESRD patients[3].
Many conventional inflammatory markers, including tumor necrosis factor-α, interleukin-6, and C-reactive protein, exacerbate kidney function impairment in patients with CKD[4]. CKD's hyperinflammatory state encourages atherogenesis and advances through vascular endothelial infiltration and leukocyte adhesion[5]. Cardiovascular etiology is the primary cause of mortality among patients with ESRD. Accelerated atherosclerosis leads to peripheral vascular disease, ischemic heart disease, and stroke. It is also associated with an enhanced risk of thrombotic events and venous thromboembolism, and can also lead to heart failure[6-8]. Patients undergoing dialysis have always had a higher adjusted mortality rate than other ESRD patients, and HD patients have a higher adjusted mortality rate than PD patients. Arrhythmia or sudden cardiac arrest is the leading cause of death among HD patients, accounting for 44%[3].
A computed composite inflammatory marker, the NLR, is based on the innate immune represented by the neutrophil count, and the adaptive immune and nutrition represented by the lymphocyte count[9]. The normal NLR values for adults, non-geriatric, and the population in good health are between 0.78 and 3.53 (mean: 1.65)[10]. However, racial differences have been studied in adult, non-institutional patients, in which it was found that non-Hispanic Black (1.76) and Hispanic (2.08) participants have significantly lower mean NLR values when compared to non-Hispanic whites (2.24), while the overall mean NLR was 2.15[11]. Previous meta-analyses have already proved that NLR is associated with an increased risk of all-cause mortality (ACM) and cardiovascular mortality (CVM) among patients with CKD[9]. In our study, we have studied whether NLR is associated with ACM and CVM among patients with ESRD undergoing chronic HD, with more updated data, performing subgroup, meta-regression, and pooled area under the receiver operating characteristic (ROC) curve (AUCROC) analysis for a thorough understanding of the association.
MATERIALS AND METHODS
We have conducted this study by following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines[12].
Data sources and search strategy
We have used appropriate Medical Subject Headings (MeSH) terms, and a search was conducted in PubMed/MEDLINE, Google Scholar, and Scopus from 2013 to June 2024, which reported the association between the NLR and ACM or CVM among the patients receiving chronic HD. MeSH terms that we have utilized were “Hemodialysis,” “dialysis,” “Neutrophil-lymphocyte ratio,” “Mortality,” “survival,” “prognosis,” “Impact,” “risk,” “death,” “Died,” “cardiac arrest,” and relevant articles and their full texts were collected for further assessment. A full list of search terms and an example search strategy for PubMed has been included in the Supplementary material. We have created a PRISMA chart that depicts the flow of inclusion and exclusion of articles with reasoning (Figure 1).
Articles retrieved from the search strategy were imported into Rayyan, and duplicates were removed; title and abstract screening was done by two independent reviewers, Roopeessh Vempati and Nanush Damarlapally, and any disparity was ruled out by a third reviewer, SV. In our study, we have included those observational studies, both prospective and retrospective, where the study population included adults with ESRD and undergoing chronic HD, NLR, and its association with outcomes, including ACM and CVM documented in the form of odds ratio (OR) or hazard ratio (HR) along with their 95%CI. We have excluded those studies where the study population is of pediatric age group, undergoing PD, animal studies, review articles, case reports, conference articles, editorials, animal studies, non-English articles, and those that reported no outcomes for the association of NLR with ACM or CVM in the form of ORs or HRs. In addition, we have excluded studies with incomplete data, which refer to studies that lacked sufficient information on NLR thresholds, mortality outcomes, or did not provide effect estimates such as OR or HR with 95%CI, making them unsuitable for inclusion in meta-analysis.
Data extraction and quality assessment
A Microsoft Excel sheet was utilized to extract the data points from the studies, including demographics, study characteristics, the predictive validity of the NLR on outcomes through AUC under the ROC curve, ORs, and HRs on the primary outcomes, that is, ACM or CVM was included in the study. Overlapping datasets were checked by comparing study authors, locations, and timeframes. In case of overlap, the most comprehensive or recent dataset was included. These studies were assessed for quality through the Joanna Briggs Institute (JBI) quality appraisal tool for cohort studies and are represented as traffic light and summary plot utilizing a generic template from Robvis software (Figure 2 and Supplementary Figure 1). Findings from the data extraction were represented in Tables 1 and 2 and Supplementary Table 1. The P values in Tables 1 and 2 reflect statistical comparisons between high and low NLR groups within individual studies, based on either unpaired t-tests or Mann-Whitney U tests, depending on data normality as reported by the original authors.
All statistical analysis was conducted with the help of Comprehensive Meta-analysis software version 4, by Biostat, Inc. In our study, we pooled the ORs and adjusted HR (aHR) by employing a random effects model with a 95%CI. Forest plots were utilized to visualize the pooled ORs and HRs in Figure 3. In addition, I² statistics were applied to look for heterogeneity in the outcomes of interest, ACM and CVM. Mild heterogeneity was considered for I² values between 25%-50%, moderate heterogeneity for the I² values between 50%-75%, and severe heterogeneity for the values greater than 75%. Furthermore, a leave-one-out sensitivity analysis was performed by removing each study and re-running the meta-analysis to calculate the effect size, CI, and P value each time to identify the influence of each study on the overall effect size estimate (Supplementary Figure 2). We have also performed a subgroup analysis by dividing the data from the included studies into various categories and representing them in Table 3. On top of that, we have performed a meta-regression analysis to identify the confounding effect of other variables on ACM and CVM (Table 4 and Supplementary Figure 3). Pooled AUCROC analysis was performed to identify the discriminative or predictive ability and the diagnostic performance of NLR > 2.5 for the studied outcomes. An AUC value indicates the predictive performance of an intervention, it can be 1.0 (perfect), 0.9-1.0 (excellent), 0.9-0.9 (very good), 0.7-0.8 (good), 0.6-0.7 (fair), 0.5-0.6 (poor), while 0.5 (random) or less indicates no or worse than random discrimination (Supplementary Figure 4). The publication bias was assessed using Egger’s regression asymmetry test, and is represented as funnel plots. A P-value < 0.05 was considered indicative of small-study effects (Supplementary Figure 5).
Table 3 Subgroup analysis of the included studies on the association of neutrophil-lymphocyte ratio with all-cause mortality.
Variables
Number of studies
No. of patients
aHR (95%CI)
P value
I2
Study design
Prospective
5
786
1.63 (1.66-2.27)
0.004
67.1
Retrospective
6
7111
1.17 (1.07-1.29)
0.0008
87.7
Sample size
≥ 200
5
7131
1.11 (1.02-1.21)
0.016
81.6
< 200
6
766
1.35 (1.16-1.56)
< 0.0001
52
Age (years)
≥ 57
5
6722
1.07 (1.00-1.14)
0.048
72.2
< 57
6
1175
1.42 (1.21-1.67)
< 0.0001
57.4
Males%
≥ 60%
4
719
1.30 (1.16-1.46)
< 0.0001
45.5
< 60%
7
7178
1.14 (1.03-1.25)
0.008
79.2
Diabetes (%)
≥ 30
4
6730
1.14 (1.04-1.25)
0.008
91.2
< 30
5
1009
1.62 (1.20-2.19)
0.001
69.5
NLR
≥ 3.5
5
6806
1.08 (1.02-1.14)
0.01
71.8
2.5-3.5
6
1091
1.54 (1.33-1.78)
< 0.0001
14
Hemoglobin (g/dL)
≥ 10.5
5
1045
1.37 (1.06-1.76)
0.02
91
< 10.5
4
986
1.17 (1.00-1.36)
0.04
75.8
Creatinine (mg/dL)
≥ 7.7
5
1003
1.42 (1.16-1.73)
0.0006
60.7
< 7.7
3
718
1.43 (0.77-2.67)
0.26
64
Serum albumin (g/dL)
≥ 3.5
6
6685
1.53 (1.21-1.94)
0.0004
57.6
< 3.5
4
1026
1.13 (1.03-1.24)
0.007
90.6
Serum ferritin (ng/mL)
≥ 340
3
418
1.29 (1.11-1.49)
0.0008
56.4
< 340
2
381
1.34 (1.09-1.64)
0.004
< 0.0001
Table 4 Meta-regression analysis of variables associated with all-cause mortality and cardiovascular mortality.
Variables
Number of studies
Coefficient estimate
95%CI
P value
All-cause mortality
Age
11
-0.02
-0.03 to 0.01
< 0.0001
Males
11
0.005
-0.003 to 0.01
0.22
BMI
4
-0.05
-0.23 to 0.12
0.55
Hypertension
6
0.002
-0.005 to 0.01
0.5
Diabetes
9
-0.005
-0.01 to 0.0006
0.08
CRP
4
0.1
-0.03 to 0.23
0.13
Hemoglobin
9
0.08
-0.08 to 0.24
0.32
Serum creatinine
6
0.06
-0.02 to 0.13
0.17
Serum ferritin
5
0.0004
-0.0001 to 0.0009
0.12
Serum albumin
10
0.17
-0.13 to 0.47
0.26
Smoking
5
-0.02
-0.06 to 0.02
0.28
Cardiovascular mortality
Age
4
-0.02
-0.04 to 0.004
0.01
Males
4
0.005
0.003 to 0.008
< 0.0001
Diabetes
4
-0.004
-0.01 to 0.006
0.43
Hemoglobin
4
0.26
-0.04 to 0.56
0.09
Serum albumin
4
0.54
-0.75 to 1.83
0.41
Ethical considerations
As this study was done by pooling individual published studies for a systematic review and meta-analysis of previously published studies, no ethical approval was required.
RESULTS
Literature search
A thorough search in Google Scholar, PubMed, and Scopus databases matching our selection criteria yielded 334 articles. Two independent reviewers (Roopeessh Vempati, Nanush Damarlapally), after removing 50 duplicates and 159 articles by applying filters through automation tools, screened the remaining 125 articles through their titles and abstracts, out of which 64 articles were removed which included those studies that were not on HD or NLR, those studied outcomes other than ACM and CVM, and those meeting the exclusion criteria. 61 articles were screened for full-text, out of which 42 articles were excluded. A third reviewer, Srivatsa Surya Vasudevan, resolved all the disparities. Finally, 19 studies[6,13-30] with a collective sample of 9047 patients were included in this meta-analysis. The PRISMA diagram shows the detailed selection process (Figure 1).
Baseline characteristics
The total sample size of the study population was 9047 individuals with a mean age of 59.5 ± 5.86 years and an average follow-up duration of 46.7 months. These studies included both prospective and retrospective designs. Several studies reported a high prevalence of comorbidities among the participants, including hypertension (ranging from 74% to 98.7% in some studies), diabetes mellitus (up to 69%), and cerebrovascular diseases. The mean hemoglobin levels were generally around 10-11 g/dL, indicating the presence of anemia. Inflammatory markers such as CRP and ferritin were elevated, and serum albumin levels were low in many studies. Overall, these baseline characteristics highlight the diverse nature of the study populations and the significant burden of comorbidities in patients undergoing chronic HD. Further details about demographic and clinical characteristics of the population from the included studies, including age, sex distribution, dialysis therapy duration, and comorbid conditions like hypertension, diabetes mellitus, and cerebrovascular diseases, were reported in the baseline characteristics in Tables 1 and 2.
Quality assessment and publication bias
The JBI’s quality appraisal tool for cohort studies was utilized and is represented as a traffic light plot by using the generic template in Robvis software[31,32] as in Figure 2, and a representation of a summary plot as in Supplementary Figure 1. The JBI tool questionnaire revealed high scores for almost all the studies, indicating a low risk of bias among the selected studies. Egger’s regression asymmetry test showed significant publication bias in the association of NLR with ACM (P = 0.0007). However, no publication bias was found in the association of NLR with CVM (P = 0.60). The asymmetric shape of the funnel plot for the association of NLR with ACM and the symmetric shape for the association of NLR with CVM are both discernible in Supplementary Figure 5, which further demonstrates this point.
Clinical outcomes
In Supplementary Table 1, we have presented the various NLR cutoffs that were studied in the included studies, effect sizes of the outcomes, along with the various confounding variables that were adjusted for by the included studies. Figure 3 illustrates a forest plot showing the pooled HRs and ORs for ACM and CVM that conveys a significant association between a higher NLR (> 2.5) with an increased risk of both ACM and CVM in patients undergoing chronic HD. Most studies demonstrate a rightward shift from the null value of 1, which indicates that there is an increased risk of the outcomes studied associated with an elevated NLR. 95%CI not including 1 and P value < 0.05 indicates a statistically significant relationship.
ACM: The pooled analysis demonstrated that an NLR greater than 2.5 is significantly associated with an increased risk of ACM, with a pooled aHR (11 studies) of 1.25 (95%CI: 1.14-1.37; P < 0.0001; I2= 85.2%) and a pooled OR (4 studies) of 4.59 (95%CI: 1.74-12.11; P = 0.002; I2 = 91.5%). The analysis revealed severe heterogeneity (I² > 75%) (Figure 3). In addition, subgroup analysis of the various variables associated with ACM revealed that most of the variables’ categories correspond to increased risk of ACM, including the NLR ranges of 2.5-3.5 and ≥ 3.5 (Table 3).
CVM: The association between NLR and CVM was similarly significant, with a pooled aHR (6 studies) of 1.24 (95%CI: 1.02-1.49; P = 0.03; I2 = 88.4%). The pooled OR for CVM (2 studies) was 1.11 (95%CI: 1.01-1.23; P = 0.03; I2 < 1%). The analysis for ACM displayed severe heterogeneity (I² > 75%) (Figure 3).
Sensitivity and meta-regression analysis
A leave-one-out sensitivity analysis, in which each study was sequentially excluded and the meta-analysis was repeated, demonstrated that no single study altered the overall effect size, confirming the robustness of the results (Supplementary Figure 2). Furthermore, to detect the presence of any confounding variables such as age, sex, or comorbid conditions, a meta-regression analysis was done. Meta-regression analysis showed that males proportion in the studies was positively associated with worsening CVM (coefficient: +0.005, 95%CI: 0.003-0.008, P < 0.0001) (Supplementary Figure 3). Additionally, increasing patients' age was inversely associated with both ACM (coefficient: -0.02, 95%CI: -0.03 to -0.01, P < 0.0001) and CVM (coefficient: -0.02, 95%CI: -0.04 to -0.004, P = 0.01), however strength of the co-efficient was very weak. Meta-regression of other variables for ACM or CVM did not show a statistically significant association, as presented in Table 4.
Pooled analysis of AUCROC
We have pooled the AUC values on the association of NLR and ACM to validate our findings and also to determine the discriminative or predictive ability of NLR for the studied association. Pooled AUC-ROC of 0.711 (95%CI: 0.63-0.80, P < 0.0001) indicates a good (AUC range: 0.7-0.8) discriminative ability of NLR > 2.5 in determining the risk of ACM in patients undergoing chronic HD (Supplementary Figure 4). Pooled AUC-ROC analysis for CVM was not feasible as only very few studies reported the values.
DISCUSSION
Principal findings
Our study demonstrates that elevated NLR > 2.5 is significantly associated with both ACM (aHR = 1.25, OR = 4.59) and CVM (aHR = 1.24, OR = 1.11) in patients undergoing chronic HD (Figure 3). These findings underscore the potential of NLR as a non-invasive, cost-effective prognostic biomarker in a high-risk ESRD population. Our findings align with and expand upon the previous meta-analysis by Ao et al[9], which reported a significant association between elevated NLR and risk of mortality in patients with CKD. However, Ao et al[9] focused on patients at various stages of CKD, whereas our analysis was exclusive to patients with ESRD receiving chronic HD.
Pathophysiological rationale
CKD results in a sustained systemic inflammatory response that impacts vascular and myocardial structures. This inflammation facilitates the development of atherosclerotic plaques in the vessels, contributes to vascular calcification and senescence, and leads to myocardial fibrosis as well as calcification of heart valves[33]. When a systemic inflammatory immune response brought on by tissue damage results in a decrease in the number of lymphocytes and the suppression of neutrophil apoptosis and neutrophil-mediated killing as part of the innate immune response, the NLR can rise[34]. The immune system's natural physiological response to supraphysiologic insults controlled by the nervous system, including the sympathetic and parasympathetic systems as well as the endocrine system, is a change from the adaptive to the innate immune system[35]. This is caused by stress hormones like catecholamines, cortisol, and prolactin, causing neutrophilia and lymphocytopenia, respectively, by demarginating the lymphocytes and marginating the neutrophils[35]. Neutrophils play a pivotal role in the innate immune system, including phagocytosis, release of cytokines, and inflammatory mediators[35]. High NLR was found to be an independent predictor of long-term death from coronary heart disease[36]. The pathophysiology of atherosclerosis and endothelial dysfunction is specifically influenced by inflammation and oxidative stress, which are typified by the activation of the NLRP3 inflammasome[36].
Clinical implications of NLR in CKD and ESRD
The non-microbial inflammatory state indicated by neutrophils, and the overall stress and nutritional status, which is indicated by lymphocytes, causes the progression of CKD[37]. In a way, NLR generally reflects the degree of inflammation in CKD patients[37]. Tumor necrosis factor-a, C-reactive protein, and interleukin-6 are among the inflammatory markers that are elevated in patients with ESRD[38]. Furthermore, NLR can indicate the progression of stage IV CKD to the dialysis stage[39]. The conventional CKD risk factors, such as ACR and eGFR, have no bearing on this correlation of progression among CKD patients[37]. When GFR declines, CKD progresses to ESRD, and the risk of cardiovascular disease and death increases[40]. Among ESRD patients receiving HD, we found a significant correlation between NLR, a measure of inflammatory status, with ACM and CVM.
Dialysis modality and comparative mortality
HD, PD, or continuous renal replacement therapy are options for patients with CKD[41]. Whether patients receiving HD have a higher mortality rate than those receiving PD is not clear[41,42]. In a recent meta-analysis by Chander et al[41], PD was associated with lower mortality risk than HD, but the significance was lost in the sensitivity analysis. A large retrospective study of 449652 patients using the United States renal data system and Medicare found similar mortality for HD and PD[42].
NLR across various diseases
In addition to atherosclerosis[43], NLR was also studied in various pathologies including myocardial infarction[44], acute stroke[45], bacterial or fungal infection[34], severe trauma[46], cancer[47], complications from post-op surgery[48], type 2 diabetes[49], pulmonary arterial hypertension[50], obstructive sleep apnea syndrome[51], chronic obstructive pulmonary disease[52], spontaneous intracerebral hemorrhage[53], peripheral arterial occlusive disease[54], sepsis[55], and coronavirus disease 2019[56].
Comparison with other biomarkers
Studies had reported various similar biomarkers, and their independent predictability was debated. In a study by Yaprak et al[17], platelet lymphocyte ratio (PLR) but not NLR, Xiang et al[20] found monocyte lymphocyte ratio (MLR), but not NLR, was found to predict ACM independently. In addition, Zhang et al[29] found that high PLR predicts both ACM and CVM, while high NLR could predict only ACM. On the other hand, Chen et al[57] found that NLR, but not CRP or PLR, is an important prognostic predictor of all major clinical outcomes in patients with advanced CKD. In addition to NLR, serum ferritin could also predict the ACM among patients undergoing HD.
Insights from meta-regression analysis
Meta-regression analysis revealed increasing male proportion across the studies was positively associated with cardiovascular mortality, which is in line with the current evidence[58,59]. CKD/ESRD was found to be associated with disruption of sex hormone physiology[59]. A decrease in testosterone due to CKD or ESRD is found to be associated with CVM, but no association between a decrease in estradiol level to CVM was found in females[59]. On the other hand, in our meta-regression analysis, we found a weak inverse yet a debatable relationship between increasing age and ACM or CVM, which raised concerns that NLR may be a less useful predictor of mortality in the older population. Furthermore, with increasing age, there exists a state of chronic, low-grade inflammation, raising baseline inflammatory markers, including NLR, which may diminish their discriminative ability or effectiveness in risk assessment[60,61].
Limitations
Severe heterogeneity was found in the pooled estimates (> 75%) (Figure 3). It's important to acknowledge this as a limitation, and it can be due to various factors inherent in the included studies, as they were observational by study design, and are susceptible to certain biases, such as selection bias and confounding. There was also variability in the definition of NLR (Supplementary Table 1), the patient populations, and the endpoints reported, which might have contributed to heterogeneity. Despite having a large number of patients included in our study, the population of each study may not represent the real-world population. On top of that, the included studies varied considerably in sample size, ranging from 59 to 5782 participants. Studies with larger sample sizes provide more precise estimates of the association between NLR and mortality. Differences in sample size, along with variations in study populations and methodologies, and NLR thresholds, likely contributed to the severe heterogeneity (I2 > 75%) (Figure 3). Additionally, the effect size may vary depending on various factors like comorbidities and dialysis duration. While most studies scored well on the JBI assessment, some had limitations in terms of clearly reporting follow-up procedures or addressing potential confounders, which could introduce some degree of bias (Figure 2). Egger's regression test revealed significant publication bias in studies examining the relationship between the NLR and ACM (P = 0.0007). This indicates that smaller or non-significant studies may be underrepresented, which could inflate the pooled effect estimate. The funnel plot for ACM displayed clear asymmetry, reinforcing this concern. In contrast, no publication bias was found for cardiovascular mortality (P = 0.60). These findings emphasize the need for cautious interpretation of the results related to ACM and highlight the importance of balanced publication practices (Supplementary Figure 5). Nevertheless, we reported sensitivity, subgroup, meta-regression analyses, and pooled AUC-ROC analysis, making our study reliable (Table 3, Supplementary Figures 2-4). Additionally, we were unable to study the potential confounding effect of infections, medication effects (e.g., anti-inflammatory agents or phosphate binders), and dialysis adequacy (e.g., Kt/V) on the studied outcomes as they were inconsistently or not uniformly reported in the included studies, limiting our ability to perform a subgroup or meta-regression analysis with these potential confounders.
Strengths
Nevertheless, our study had various strengths, including the large sample size, and analysis based on aHRs allowed us to account for potential confounding variables. Many studies had a prospective design (11/19), which strengthens the ability to assess the association between NLR and future mortality outcomes (Table 1). Our assessment using the JBI’s quality appraisal tool revealed higher scores, as indicated through the traffic light plot that the included studies had a low risk of bias (Figure 2 and Supplementary Figure 1). In addition, we have identified no variations in the leave-one-out sensitivity analysis, where no single study disproportionately influenced the overall effect size for either ACM or CVM, and the association between NLR and mortality outcomes remained consistent across various study designs and populations, conveying the robustness of our study findings (Supplementary Figure 2). It implies that the significance of the association between NLR and mortality was not dependent on any individual study. While no single study altered the significance of our results, studies with larger sample sizes, such as Ouellet et al[16], naturally had a greater influence on the pooled estimates.
NLR is an easily available and cost-effective index for clinical practice, which makes it an attractive prognostic index for patients with CKD[9]. The pooled area under ROC curve analysis indicated that NLR could correctly classify the ACM in 71.1% (P < 0.0001) of the cases, and also indicates moderate effectiveness as a predictive biomarker (Supplementary Figure 4 ).
Furthermore, we performed a meta-regression analysis to study the influence of various confounders on the ACM and CVM in patients undergoing HD (Supplementary Figure 3), and we also assessed the publication bias using Egger's test, represented by funnel plots enhancing the reliability of our study (Supplementary Figure 5) which indicates the credibility of our meta-analysis.
Future directives
Future studies should investigate finding the most appropriate NLR threshold for a specific HD population by considering age, sex, and other comorbidities, as NLR is a non-specific marker. Large, multicenter prospective studies with a standardized population and subpopulation analysis are needed to generalize the findings. In addition, interventional studies should investigate the modifications in dialysis, like increased dose, altered composition, and the role of anti-inflammatory agents, nutritional interventions, and lifestyle modifications, to see if there is a reduction in the high mortality burden in this population.
CONCLUSION
This meta-analysis suggests that an elevated NLR is associated with an increased risk of all-cause and cardiovascular mortality among patients undergoing chronic HD. The pooled aHR and area under the curve analyses support the potential of NLR as a readily available and cost-effective prognostic biomarker. However, due to heterogeneity in study designs and thresholds used, these findings should be interpreted with caution. Future prospective studies with standardized methodologies are warranted to better establish the clinical utility of NLR in this high-risk population.
Footnotes
Provenance and peer review: Invited article; Externally peer reviewed.
Peer-review model: Single blind
Corresponding Author's Membership in Professional Societies: American College of Physician.
Specialty type: Medical laboratory technology
Country of origin: United States
Peer-review report’s classification
Scientific Quality: Grade A, Grade B, Grade B, Grade C
Novelty: Grade B, Grade B, Grade B, Grade C
Creativity or Innovation: Grade B, Grade B, Grade B, Grade C
Scientific Significance: Grade A, Grade B, Grade B, Grade B
P-Reviewer: Itoh K; Jiao Y; Li H S-Editor: Liu H L-Editor: A P-Editor: Zhang YL
Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, Shamseer L, Tetzlaff JM, Akl EA, Brennan SE, Chou R, Glanville J, Grimshaw JM, Hróbjartsson A, Lalu MM, Li T, Loder EW, Mayo-Wilson E, McDonald S, McGuinness LA, Stewart LA, Thomas J, Tricco AC, Welch VA, Whiting P, Moher D. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews.BMJ. 2021;372:n71.
[RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)][Cited by in Crossref: 44932][Cited by in RCA: 39806][Article Influence: 9951.5][Reference Citation Analysis (2)]
Woziwodzka K, Dziewierz A, Pawica M, Panek A, Krzanowski M, Gołasa P, Latacz P, Burkat M, Kuźniewski M, Krzanowska K. Neutrophil-to-lymphocyte ratio predicts long-term all-cause mortality in patients with chronic kidney disease stage 5.Folia Med Cracov. 2019;59:55-70.
[RCA] [PubMed] [DOI] [Full Text][Cited by in RCA: 10][Reference Citation Analysis (0)]
Kular D, Chis Ster I, Sarnowski A, Lioudaki E, Braide-Azikiwe DCB, Ford ML, Makanjuola D, Rankin A, Cairns H, Popoola J, Cole N, Phanish M, Hull R, Swift PA, Banerjee D. The Characteristics, Dynamics, and the Risk of Death in COVID-19 Positive Dialysis Patients in London, UK.Kidney360. 2020;1:1226-1243.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 6][Cited by in RCA: 14][Article Influence: 2.8][Reference Citation Analysis (0)]
Parmelia M, Sutarka IN. Neutrophil-Lymphocyte Ratio as Predictor of Mortality in Regular Hemodialysis Patients at Tabanan General Hospital.J Kedokteran. 2023;12:150-156.
[PubMed] [DOI] [Full Text]
Zhang Y, Zhang A, Wei L, Ren K, Wang Q, Shao B, Zhao C, Ren Z, Bai J, Cao N. A high platelet-to-lymphocyte ratio predicts all-cause mortality and cardiovascular mortality in maintenance hemodialysis patients.Ren Fail. 2023;45:2258228.
[RCA] [PubMed] [DOI] [Full Text][Cited by in RCA: 6][Reference Citation Analysis (0)]
Yuan Q, Wang J, Peng Z, Zhou Q, Xiao X, Xie Y, Wang W, Huang L, Tang W, Sun D, Zhang L, Wang F, Zhao MH, Tao L, He K, Xu H; C-STRIDE study group. Neutrophil-to-lymphocyte ratio and incident end-stage renal disease in Chinese patients with chronic kidney disease: results from the Chinese Cohort Study of Chronic Kidney Disease (C-STRIDE).J Transl Med. 2019;17:86.
[RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)][Cited by in Crossref: 38][Cited by in RCA: 75][Article Influence: 12.5][Reference Citation Analysis (0)]
Lee MJ, Park SD, Kwon SW, Woo SI, Lee MD, Shin SH, Kim DH, Kwan J, Park KS. Relation Between Neutrophil-to-Lymphocyte Ratio and Index of Microcirculatory Resistance in Patients With ST-Segment Elevation Myocardial Infarction Undergoing Primary Percutaneous Coronary Intervention.Am J Cardiol. 2016;118:1323-1328.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 13][Cited by in RCA: 32][Article Influence: 3.6][Reference Citation Analysis (0)]
Lee PY, Oen KQX, Lim GRS, Hartono JL, Muthiah M, Huang DQ, Teo FSW, Li AY, Mak A, Chandran NS, Tan CL, Yang P, Tai ES, Ng KWP, Vijayan J, Chan YC, Tan LL, Lee MB, Chua HR, Hong WZ, Yap ES, Lim DK, Yuen YS, Chan YH, Aminkeng F, Wong ASC, Huang Y, Tay SH. Neutrophil-to-Lymphocyte Ratio Predicts Development of Immune-Related Adverse Events and Outcomes from Immune Checkpoint Blockade: A Case-Control Study.Cancers (Basel). 2021;13:1308.
[RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)][Cited by in Crossref: 10][Cited by in RCA: 49][Article Influence: 12.3][Reference Citation Analysis (0)]
Erturk M, Cakmak HA, Surgit O, Celik O, Aksu HU, Akgul O, Gurdogan M, Bulut U, Ozalp B, Akbay E, Yildirim A. Predictive value of elevated neutrophil to lymphocyte ratio for long-term cardiovascular mortality in peripheral arterial occlusive disease.J Cardiol. 2014;64:371-376.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 67][Cited by in RCA: 75][Article Influence: 6.8][Reference Citation Analysis (0)]
