Role of neutrophil to lymphocyte ratio in osteoarthritis: A systematic review and meta-analysis

Abstract

BACKGROUND

Osteoarthritis (OA) involves low-grade inflammation. The neutrophil-to-lymphocyte ratio (NLR) may serve as a simple biomarker, but its role in OA remains unclear.

AIM

To review the existing scientific literature on the role of NLR in OA, a classic age-related disorder, to perform a meta-analysis of the available data.

METHODS

The electronic databases PubMed, ProQuest, and Scopus were systematically searched from inception to March 1, 2024. The inclusion criteria were retrospective and prospective case-control studies involving human subjects with OA and healthy controls. The included studies needed to provide NLR levels for both OA patients and healthy controls and perform a comparative analysis of NLR levels between these groups.

RESULTS

According to the PRISMA guidelines, fifteen articles were included in the meta-analysis after multiple screenings. The pooled results demonstrated a significant overall elevation of NLR in OA patients compared to healthy controls. (standardized mean difference = 0.39, 95% confidence interval: 0.03-0.75, P = 0.03). However, the subgroup analysis shows no significant differences in NLR levels when considering study design (retrospective vs prospective) and OA severity (severe vs mild-moderate). This suggests variability and potential limitations in using NLR as a consistent marker across different study types and OA severity.

CONCLUSION

Our study found that OA patients have higher NLR than healthy individuals. However, NLR did not significantly differ by study type or disease severity, suggesting its limited use in indicating OA severity.

Key Words: Neutrophil to lymphocyte ratio; Osteoarthritis; Meta-analysis; Inflammatory biomarker; Systemic inflammation

Core Tip: This systematic review and meta-analysis evaluate the role of the neutrophil-to-lymphocyte ratio (NLR) in osteoarthritis (OA). Our findings indicate that OA patients have significantly higher NLR levels than healthy controls, suggesting its potential as an inflammatory biomarker. However, NLR does not significantly correlate with OA severity or study design, limiting its utility in assessing disease progression. Further research is needed to clarify NLR’s clinical relevance in OA diagnosis and management.

INTRODUCTION

Osteoarthritis (OA) is characterized by low-grade inflammation and destruction of articular joint structures such as cartilage, bones, and synovium[1]. The most recent criteria for clinical diagnosis of OA are based on symptoms of reduced joint function and pain that worsens throughout the day with signs of crepitus, restricted motion, and bony remodeling[2]. Although the etiology of OA is commonly associated with factors like prior injury, obesity, aging, and heredity, the disease process remains poorly understood[3]. In recent literature, OA has been linked to increased systemic inflammatory response through mediators in bodily fluids, including blood, synovial fluid, and urine[4-7]. In a radiologic study, Attur et al[5] demonstrated that inflammatory mediators in OA were associated with the narrowing of joint space and disease progression.

The neutrophil-to-lymphocyte ratio (NLR) is a sensitive indicator of systemic infection, inflammation, and inflammatory-related prognosis[8-12]. It is obtained by dividing the absolute neutrophil count by the absolute lymphocyte count collected from a complete blood count, making it a readily available, cost-effective measure[9]. The NLR is dependent on the degree of immunological stress experienced in an infectious or non-infectious disease state[9]. Previous studies have demonstrated that elevated NLR is a sensitive biomarker for OA, serving as a possible prognostic factor of disease progression[8,10].

The current diagnostic criteria may lead to delayed diagnosis and treatment of patients with OA. Adding further measures of systemic inflammation, like the NLR, to these criteria would help accelerate diagnosis and improve patient outcomes. A link between the presence of heightened systemic inflammation and OA is present in current literature[4-7]. Recent studies indicate that delays in OA diagnosis are common, often due to a reliance on symptomatic and radiographic criteria that may appear late in the disease process. Since NLR is a cost-effective and routinely available blood test, its inclusion in early diagnostic workups could improve timeliness. This systematic review and meta-analysis aim to clarify the diagnostic and prognostic role in OA and assess its relationship with disease severity.

MATERIALS AND METHODS

Eligibility criteria

The PRISMA standards were applied to prepare this meta-analysis (Figure 1). Studies comparing the NLR levels of OA patients to healthy controls in retrospective and prospective case-control articles were deemed eligible.

Figure 1 PRISMA Flow diagram for new systematic reviews, which includes searches of databases, registers, and other sources. NLR: Neutrophil-to-lymphocyte ratio.

Literature search

Two authors independently searched electronic sources, including Web of Science, PubMed, and Scopus, to find relevant research that included only human participants. On March 1, 2024, the last systematic search was carried out using the following search criteria: “Osteoarthritis” AND (“Neutrophil to lymphocyte ratio” OR NLR). There were no constraints on language or publication year. The following items were excluded throughout various databases: Publications with non-human subjects, case reports or case series, reviews, editorials, clinical trials, comments, letters, and duplicate articles. EndNote software (version X8) was used to delete and monitor any duplicate studies that could have emerged from the electronic database search. Included studies were cross-references for completeness.

Data extraction

The following items were extracted from the included articles: Publication year, name of the first author, study design, study location, the number of healthy controls and their NLR level, the number of OA patients and their NLR level, sex, and the mean age of participants. Three writers had several discussions to solve any discrepancies missing or unclear data presentation. If unsolved, the first or the corresponding author resolved the disagreements between other writers.

Heterogeneity and quality assessments

The heterogeneity (I²) of the included articles was evaluated using the Tau-squared test, where I² measured the level of consistency between the studies (P < 0.10 was considered significant). A random effect model was used for the meta-analysis of heterogeneous results. The Newcastle-Ottawa scale (NOS) for case-control studies was used to assess the quality of each of the included studies. High-quality studies were those with NOS scores above or equal to the median. A contour-enhanced funnel plot was created to inspect any asymmetry brought on by publication bias visually. Additionally, the quantitative analysis of publication bias was conducted using Egger’s test, where P < 0.05 was considered statistically significant.

Statistical analysis

Standardized mean difference (SMD) with a 95% confidence interval (CI) was applied to assess the NLR level of patients with OA compared to healthy controls. A P value under 0.05 was considered significant. STATA 12.0 software was employed to conduct statistical analysis.

RESULTS

Search and selection of literature

A total of 956 records were obtained from the database search and manual review of the citation list of the included articles. After excluding non-relevant records and duplicates, 16 studies were included in this review[13-21]. The exclusion and inclusion process are explained in detail in the PRISMA flow diagram in Figure 1.

Characteristics of the included studies

This meta-analysis included 15 studies, of whom 12 were conducted in Turkey[14-16,19-25], two in China[8,18], and one in Iraq[13]. Regarding article language, 14 documents were written in English[13,15-24] and one in Turkish[14]. In terms of study design, 11 of them were retrospective[14-16,18,20-24], and 4 of them were prospective[13,17,19]. Table 1 shows the general characteristics and the quality scores of the studies. In total, nine studies compared NLR levels of patients with OA and healthy controls[13,14,17-21,26,27], including 879 patients with OA and 701 healthy controls, and 12 studies compared NLR levels of patients with severe OA and those with mild-moderate OA, including 959 patients with mild-moderate OA, 603 with severe OA[14-16,18,21-26]. NOS scores of included studies ranged between 6-8.

Table 1 General characteristics of included studies.

Ref.	Country	Design	Age	Male, %	BMI	Healthy control, n, NLR	OA, total, n, NLR	Mild-moderate OA, n, NLR	Severe OA, n, NLR	NOS score
Taşoğlu et al[10], 2016	Turkey	Retrospective	61.75	78.4	33.07	NR	NR	146, 1.79 ± 0.80	30, 2.18 ± 1.04	7
Atar et al[14], 2017	Turkey	Retrospective	56.21	88	28.36	57, 1.76 ± 0.66	92, 2.02 ± 0.75	44, 2.11 ± 0.79	48, 1.93 ± 0.70	6
Hira and Tamam[18], 2017	Turkey	Retrospective	50	64	NR	145, 1.79 ± 0.97	118, 2.15 ± 1.12	NR	NR	8
Büyükavcı et al[16], 2018	Turkey	Retrospective	63.5	70.5	27.04	NR	NR	165, 1.91 ± 0.82	69, 2.16 ± 0.84	8
Gundogdu et al[7], 2018	Turkey	Prospective	61.75	83	32.78	30, 2.25 ± 0.91	60, 2.37 ± 1.67	NR	NR	6
Özler et al[21], 2018	Turkey	Retrospective	59.55	NR	NR	NR	NR	112, 2.04 ± 0.08	106, 2.40 ± 0.14	7
Gao et al[17], 2019	China	Retrospective	55.47	51.66	NR	120, 1.49 ± 0.40	119, 2.39 ± 1.73	86, 2.22 ± 1.59	33, 2.85 ± 2.00	6
Koca et al[20], 2019	Turkey	Retrospective	59.06	82.29	NR	NR	NR	105, 2.10 ± 1.50	104, 2.90 ± 2.40	7
Cai et al[8], 2021	China	Prospective	59.4	75.35	23.15	150, 1.42 ± 0.62	142, 2.81 ± 1.14	NR	NR	8
Karataş et al[19], 2021	Turkey	Retrospective	52.67	84.7	27.74	40, 2.45 ± 2.60	122, 2.33 ± 0.74	70, 2.08 ± 1.92	52, 2.40 ± 1.09	7
Al-Janaby[13], 2021	Iraq	Prospective	60.4	65	29.1	50, 1.97 ± 0.88	100, 2.19 ± 0.72	NR	NR	7
Bakılan et al[15], 2022	Turkey	Retrospective	60.85	92	NR	NR	NR	81, 1.81 ± 0.67	55, 1.94 ± 0.80	6
Korkmaz et al[24], 2022	Turkey	Prospective	57.6	23.7	NR	59, 1.7 ± 0.7	76, 2.1 ± 1.3	37, 2.1 ± 1.5	39, 2.1 ± 0.9	7
Yasar et al[22], 2022	Turkey	Retrospective	65.7	36	NR	NR	NR	35, 2.11 ± 0.9	30, 1.82 ± 0.7	7
Ünal Enginar et al[23], 2023	Turkey	Retrospective	65	100	NR	50, 1.7 ± 0.45	50, 1.60 ± 0.53	43, 1.60 ± 0.53	7, 1.66 ± 1.10	8

Some studies did not report all variables, such as body mass index or age. The neutrophil to lymphocyte ratio value is expressed as mean ± SD. BMI: Body mass index; NR: Not reported; NLR: Neutrophil to lymphocyte ratio; NOS: Newcastle-Ottawa scale; OA: Osteoarthritis.

Difference in NLR level between patients with OA and healthy controls

There was statistically significant heterogeneity among included studies (I² = 91.3%, P < 0.001); thus, the random-effect model was applied in the analysis (Figure 2A). The pooled results from 16 studies showed that OA patients had higher NLR levels than healthy controls, with a SMD of 0.39 (95%CI: 0.03 to 0.75, P = 0.03). This indicates a statistically significant overall elevation of NLR in OA patients, highlighting its potential as an inflammatory marker. Subgroup analysis divides the data based on study design (retrospective vs prospective) and OA severity (severe vs mild-moderate), to explore whether the effect size varies across these subgroups.

Figure 2 Meta-analysis of neutrophil-to-lymphocyte ratio levels. A: Meta-analysis of neutrophil-to-lymphocyte ratio (NLR) levels in patients with osteoarthritis (OA) compared to healthy controls; B: Subgroup meta-analysis of NLR levels in patients with OA compared to healthy controls, according to study design; C: Meta-analysis of NLR levels in patients with severe OA compared to those with mild-moderate OA; D: Funnel plot showing publication bias. Studies on NLR levels in patients with OA compared to healthy controls; E: Studies on NLR levels in patients with severe OA compared to those with mild-moderate OA. SMD: Standardized mean difference; CI: Confidence interval.

Subgroup analysis

Subgroup analysis by study design were as followed (Figure 2B).

Retrospective studies: The analysis of retrospective studies showed no significant difference in NLR levels between OA patients and healthy controls (SMD = 0.25, 95%CI: -0.07 to 0.58, P = 0.12).

Prospective studies: Similarly, the analysis of prospective studies also showed no significant difference in NLR levels (SMD = 0.58, 95%CI: -0.15 to 1.30, P = 0.13).

Subgroup analysis by OA severity: We used the random-effect model to pool the data because of high heterogeneity between included studies (I² = 94.8%, P < 0.001) (Figure 2C). The pooled results showed that there were not any differences in NLR levels of patients with severe OA compared to those with mild-moderate OA (SMD = -0.42, 95%CI: -0.91-0.07, P = 0.09).

Subgroup analysis by geographic location: To further explore the sources of heterogeneity, we performed additional subgroup analyses based on geographic location (Turkey vs other countries) and joint type (knee vs hip). The subgroup analysis by country indicated that studies conducted in Turkey showed a moderate pooled effect size (SMD = 0.35, 95%CI: 0.01 to 0.69, P = 0.04), while studies from other countries yielded a slightly higher SMD but with wider confidence intervals (SMD = 0.47, 95%CI: -0.22 to 1.16, P = 0.18), indicating more variability.

Subgroup analysis by joint type: Regarding joint type, the data allowed comparison in a subset of studies that reported knee vs hip OA. The subgroup for knee OA showed a significant elevation in NLR (SMD = 0.40, 95%CI: 0.04 to 0.77), while hip OA studies were limited in number and showed no significant association (SMD = 0.20, 95%CI: -0.32 to 0.71). It is important to note that the subgroup analysis of prospective studies was based on only four studies (n = 4), which limits the statistical power to detect meaningful differences. Similarly, subgroup analyses by severity included variable sample sizes.

Publication bias

Figure 2D and E illustrates that studies comparing the NLR levels of healthy controls and OA patients showed no evidence of publication bias (P Egger’s test = 0.07) and among studies comparing patients with severe OA and mild-moderate OA (P Egger’s test = 0.73)[14-16,18,21-27].

DISCUSSION

In this study, we pull available data to examine the NLR in OA; three key findings emerged. Patients with OA had higher NLR levels compared to healthy controls overall. The subgroup analysis reveals that while the overall data suggest elevated NLR in OA patients, this effect is not consistently observed when studies are divided by design (retrospective or prospective). This indicates potential variability in study designs and the importance of considering these differences in future research. The lack of significant differences in NLR levels between severe and mild-moderate OA patients suggests that NLR may not be a useful marker for distinguishing OA severity. The importance of NLR in diagnosing and treating OA lies in its potential role in the disease’s pathogenesis. While OA’s exact causes and processes are not well understood, hypotheses suggest that mechanical stress and immune responses play crucial roles. Evidence indicates that inflammatory molecules, such as proinflammatory cytokines, are critical mediators in these processes. Symptoms like joint pain, swelling, stiffness, and the association with low-grade synovitis support the inflammatory nature of OA[10,28,29].

One hypothesis is that degraded cartilage fragments trigger a sterile inflammatory response, releasing catabolic mediators (like chemokines and proinflammatory cytokines) and suppressing anabolic mediators (like growth factors and anti-inflammatory cytokines). These inflammatory cytokines, such as interleukin-1β (IL-1β), IL-6, IL-15, IL-17, IL-18, and tumor necrosis factor α, result from aging, overuse, and trauma. Consequently, cartilage degradation products and inflammatory markers can indicate OA progression, but no low-cost laboratory test efficiently detects OA severity in clinical settings[30,31]. Unfortunately, these markers can’t be used efficiently in a clinical setting to detect OA. Still, some inflammatory-induced factors, such as cartilage oligomeric matrix protein (COMP), C-terminal telopeptide of type II collagen (CTX-II), and NLR, could be potentially related to OA. COMP is a non-collagenous extracellular matrix protein that aids in collagen secretion and aggregation, contributing to the extracellular matrix’s stability. The literature on COMP is contradictory, and it is currently used only for scientific purposes, with its significance not yet established. Fernandes et al[28] suggested that serum COMP could be a diagnostic marker for early asymptomatic OA (OA) that shows no radiological abnormalities.

Similarly, Verma and Dalal[29] concluded that serum COMP is linked to pain but not to radiological changes. However, Riegger et al[30] found some cofounders for this marker, such as renal function, exist. Moreover, the association between disease severity and COMP level is highly controversial[32,33]. The urinary level of CTX-II is the other marker for OA. Although some studies have shown a relationship between the level of CTX-II and pain, there are many controversies about the aforementioned level and the radiographic severity of OA[34,35]. Neutrophils and lymphocytes are central players in the inflammation process, and their numbers are found to be significantly altered during systemic inflammation[27]. Neutrophils participate in OA pathogenesis through the degradation of cartilage itself and do so by the release of matrix metalloproteinase-8, as well as numerous other cytokines[25]. In both early and advanced OA, infiltrating perivascular T and B lymphocytes can be found at the cellular level[36]. A study by Taşoğlu et al[10] revealed a high NLR in peripheral blood and supports the idea that neutrophils play a part in OA pathogenesis. Animal models have also revealed cellular infiltrates and synovial fluid with a high NLR[10]. Neutrophil counts are elevated in contrast to decreased lymphocyte counts. The accumulation of lymphocytes at the inflammation site can be the reason for the reduction in lymphocyte counts[27].

NLR is a promising marker because it reflects the immune response balance and can be measured from routine blood tests[27]. Elevated NLR levels are linked to poorer outcomes in various systemic diseases. However, the relationship between NLR and OA, particularly in differentiating unilateral from bilateral OA, requires further study[25]. Considering these studies’ findings and the currently proposed pathophysiology of OA, we can say that the NLR function in the clinical management of OA is still controversial but offers clinical utility combined with other measures. Our study showed a strong relationship between OA and NLR, but limitations include the small number of analyzed articles and high heterogeneity among studies. Despite using a random-effects model, heterogeneity remains an issue. Nonetheless, our systematic and thorough literature search strengthens this meta-analysis. Further studies are needed to reinforce our findings and clarify NLR’s clinical utility in OA management. We performed a systematic search and manually reviewed the reference list of the included studies. This ensured a reliable and thorough literature search, one of the present meta-analysis’s remarkable strengths.

Clinical implications and future directions

Although our findings confirm a statistically significant elevation in NLR levels among OA patients, its limited association with disease severity and study design reduces its immediate clinical utility. NLR may be most effective when combined with other markers of inflammation, such as C-reactive protein or erythrocyte sedimentation rate, or as part of a multimodal diagnostic panel including imaging data. We recommend future research focus on: (1) Longitudinal cohort studies to assess whether changes in NLR predict OA progression; (2) Comparative biomarker analysis between NLR, C-reactive protein, erythrocyte sedimentation rate, COMP, and CTX-II; and (3) Integration of NLR into clinical algorithms to assist with early detection or patient stratification.

CONCLUSION

This study found that patients with OA have higher NLR than healthy individuals. Therefore, it could be used as an indicator for OA. However, when we looked at different types of studies and compared patients with varying disease severity, the differences in NLR levels were insignificant. This suggests that while NLR might be a useful marker for detecting OA, it may not reliably indicate the severity of the condition. Further research is needed to confirm these findings and better understand NLR’s role in OA.