Altered blood-based biomarkers in carpal tunnel syndrome. A systematic review and meta-analysis

Abstract

Introduction

Carpal tunnel syndrome (CTS) is the most common entrapment neuropathy, yet its pathophysiology remains unclear. Inflammation, fibrosis, and metabolic factors have been implicated in animal models, but evidence from human studies is inconsistent.

Objectives

To comprehensively characterize blood-based biomarker alterations in CTS compared with healthy controls. A secondary aim was to summarize reported associations between biomarkers and clinical outcomes. Methods: Seven databases were searched from inception to February 2026. Two reviewers independently screened the included studies, extracted data and assessed methodological quality using the Newcastle–Ottawa Scale. Meta-analyses were performed when at least two comparable studies were available with a multilevel random-effects approach for the main meta-analyses, and random-effects or fixed-effects for subgroups metaanalyses. Sensitivity analyses were performed through “leave-one-out” method and certainty of evidence was evaluated through GRADE approach.

Results

Meta-analyses showed significantly higher blood levels of fibrosis-related markers [SMD (95% CI) = 1.64 (0.96/2.33)], anti-inflammatory markers [SMD (95% CI) = 1 (0.1/1.89)], proinflammatory markers [SMD (95% CI) = 0.63 (0.13/1.12)], and lower levels of vitamins and minerals [SMD (95% CI) = −0.75 (−1.49/-0.005)]. At individual level, increased concentrations of CCL4, CXCL10, CCL2, CXCL8, CRP, IL4, neutrophils, VEGF, and MDA were observed, along with decreased vitamin D levels.

Conclusion

CTS might be associated with systemic alterations in inflammatory and fibrotic pathways, and reduced vitamin D levels. Despite low certainty of evidence, high heterogeneity and the inability to confer causality from observational data, these biomarkers may inform future prospective research.

1. Introduction

Carpal tunnel syndrome (CTS) is the most prevalent entrapment neuropathy (EN), affecting approximately 4% of the general population (Atroshi, 2011; Padua et al., 2016). CTS patients initially present intermittent paresthesia and later develop sensory loss, followed by weakness and atrophy of the thenar muscles in more advanced stages, leading to difficulties in performing activities of daily living (Padua et al., 2016, 2023). Beyond its clinical and socioeconomic relevance (Atroshi et al., 2015; Brodeur et al., 2022), CTS provides a well-characterized, accessible, and common human model to study mechanisms of nerve compression (Baskozos et al., 2020; Schmid et al., 2014).

While numerous risk factors for CTS have been identified, like female sex, diabetes, obesity and even genetic factors (Wiberg et al., 2019), the majority of cases are considered “idiopathic”, as the etiology of CTS remains largely unknown (Erickson et al., 2019; Pourmemari and Shiri, 2016; Wiberg et al., 2019, 2022). However, several local pathophysiological alterations have been observed in affected individuals. Carpal tunnel pressure has been shown to be increased (Ahn et al., 2009; Chen et al., 2013) and, when maintained over time, it can be associated with functional and structural alterations of the median nerve as well as fibrotic changes in the subsynovial connective tissue and transverse carpal ligament (Schmid et al., 2013b; Schmid et al., 2020). Compressed median nerves have shown intraneural oedema (Schmid et al., 2012), altered microcirculation (Coppieters et al., 2012; Seradge et al., 1995), degeneration of small and large nerve fibers, signs of demyelination and remyelination, and altered architecture of nodes of Ranvier (Schmid et al., 2014).

These responses may be preceded by or associated with a low grade local and systemic proinflammatory process, which has been extensively described in animal models (Gupta et al., 2005; Otoshi et al., 2010; Pham and Gupta, 2009; Schmid et al., 2013). Moreover, animal models have not only demonstrated an inflammatory response and activation of the immune system after neural injury, but also their involvement in the onset and maintenance of neuropathic pain (Austin and Moalem-Taylor, 2010; Calvo et al., 2012). However, the translation of these findings to human populations has been limited, primarily due to the challenges of obtaining tissue samples from patients (Villa et al., 2025).

Recent studies in humans have begun to reveal similar local immune-mediated processes. Increased intraneural macrophage density has been reported in Morton’s neuroma (another EN) compared with healthy controls (O. P. Sandy-Hindmarch et al., 2025), and in CTS higher density of CD3⁺ T cells were found within the subsynovial connective tissue (O. Sandy-Hindmarch et al., 2024). Since nerve sampling from CTS patients is unfeasible, blood analyses have been carried out as an indirect measure of systemic inflammation. These analyses show higher blood concentrations of central and effector memory T cells (Moalem-Taylor et al., 2017) compared to healthy controls, as well as higher concentrations of TGF-β, CCL5, VEGF, CXCL8 and CXCL10 (Moalem-Taylor et al., 2017; Sandy-Hindmarch et al., 2022), thus confirming the presence of a proinflammatory response.

Beyond inflammation, alterations in other biological systems have also been suggested, such as cardiovascular markers (Shiri et al., 2011), sex hormones (Mohammadi et al., 2016; Toesca et al., 2008) or vitamins (Tanik et al., 2016). Nevertheless, despite the growing body of observational studies, our understanding of the pathophysiology of EN remains incomplete, as these approaches provide only partial insights and do not allow a comprehensive view of which biological systems are altered or involved in the process.

To address this, we conducted a systematic review and meta-analysis with the aim of understanding the type and concentration of blood-based biomarkers altered in CTS patients compared with healthy controls. By doing so, we sought to clarify the potential physiological processes involved, including whether a clear inflammatory component which can be detected systemically is present. Furthermore, this review aims to explore potential associations between these biomarkers and clinical outcomes, including nerve conduction studies, disease duration and patient-reported symptoms.

2. Methods

This systematic review was conducted and reported following the guidelines of the Cochrane Handbook for Systematic Review of Intervention (Higgins and Green) and the most recent update from 2021 from the original guide Preferred Reporting Items for Systematic Reviews “PRISMA” (Page et al., 2021). The protocol has been prospectively registered in the International Prospective Register of Systematic Reviews (PROSPERO, CRD42024604729). Deviations for the protocol were made for unplanned multilevel and fixed effects analyses, conducted to better align with the statistical assumptions of the included studies and the implementation of the GRADE approach to assess the certainty of the evidence (see more details in sections 2.5 and 2.6).

2.1. Literature search

The following databases were searched from inception to 18th of February 2026: MEDLINE (via EBSCO), PubMed, EMBASE, CINAHL, Scopus, WOS and Cochrane database. The search strategy for each database is described in the supplementary material (SUPPLEMENTARY MATERIAL S1) and is based on the combination of population terms (Carpal tunnel syndrome) with search words representing biomarkers.

2.2. Selection criteria

2.2.1. Types of studies

We included observational studies (cross-sectional, cohort, or case-control) and baseline measurements of randomized clinical trials (RCT). Studies need to include a quantitative blood marker analysis in patients with CTS compared to a control group of healthy participants or comparison to the original study authors’ own normative data. The potential bias associated with sample selection in this type of study was taken into account in the assessment of methodological quality.

Case series, conference abstracts, observational studies or RCTs in which patients have been treated previously with surgery or corticosteroids, RCT follow up, studies without a comparator (either control group or own normative data), non-human studies and studies that use non-quantitative methods to assess biomarker concentrations (e.g., qualitative immunohistochemistry) were excluded.

2.2.2. Patients/controls

We included studies with patients diagnosed with idiopathic CTS by clinical evaluation, nerve conduction studies or both. CTS caused by systemic diseases such as diabetes, dialysis, rheumatoid arthritis, CTS related to pregnancy or traumatic CTS were excluded.

The inclusion criteria for the control group were healthy participants aged greater than or equal to 18 years old. The exclusion criteria were participants with a concomitant diagnosis of peripheral neuropathy or central nervous system disorder, or any other exclusion criteria mentioned for patients.

2.2.3. Outcome measures

Our main outcome measure was the type and concentration of blood-based biomarkers (e.g., cytokines, chemokines, immune cells) in patients with CTS compared to healthy controls. These biomarkers were categorized into several functional groups: proinflammatory and anti-inflammatory markers, fibrosis markers, white blood cell counts and ratios, growth and stimulation factors, oxidative stress markers, cardiovascular health-related markers, and vitamins and minerals.

2.2.4. Study selection

Identified articles were imported into Rayyan application to facilitate screening (Ouzzani et al., 2016). First, duplicates were identified and removed. In a second stage, two reviewers (J.M.-C. and E.F.-C.) independently assessed the eligibility of the identified studies based on information from title, abstract and keywords. During the third stage, the remaining full text articles were again independently reviewed for eligibility by both reviewers (J.M.-C. and C.R.-R.). A third reviewer (M. M.-A.) acted as a mediator if consensus was not reached at both title/abstract and full text screening stages.

2.3. Data extraction and management

Data of eligible studies was independently extracted into an excel file by two different reviewers (J.M.-C. and C.R.-R.).

Extracted data include: year of publication; authors; study design; number of patients and their characteristics (age, sex, diagnostic criteria and duration of symptoms); number of controls and their characteristics (age and sex); scales used for the measurement of CTS symptoms severity (e.g., visual analogical scale [VAS]) and disability (e.g. Boston carpal tunnel syndrome questionnaire [BCTQ]); electrodiagnostic test (EDT) severity; reported correlations between biomarkers and symptoms/EDT/duration (e.g. Pearson/Spearman correlation coefficients); medications consumed; biomarker characteristics (type of biomarker, summary statistics of concentration e.g. mean and 95% CI e.g., picograms/mL or nanograms per liter) and type of biomarker analysis (e.g., Simoa®, electrochemiluminescence or ELISA; sample time points and collection times).

Means, standard deviations (SD), sample sizes and p-values for biomarker concentrations were extracted by two independent investigators. If data were only available in graphs, we extracted data using the Web Plot Digitizer online version (apps.automeris.io/wpd/). Accuracy of the double extracted data was checked, and consensus reached between investigators. In case of disagreement, a third investigator (A.A.-R.) made the final decision.

2.4. Methodological quality assessment

No clinical trials made it past the study screening, so the studies were assessed using the Newcastle-Ottawa Scale (NOS) adapted for cross-sectional studies or cohort studies if required. This scale assesses selection bias, comparability bias and outcome bias, with a score from 0 to 10 in cross-sectional studies. Established cutoffs are high risk (0-3 points), moderate risk (4-7 points) and low risk (8-10 points). In the case of cohort studies, the scale ranges from 0 to 9 points with no established cutoff (Ottawa Hospital Research Institute).

Two independent reviewers (J.M.-C. and J. Z.-Z.) assessed each study for methodological quality. Disagreements between reviewers were resolved by discussion and a third reviewer (L.M.-G.) mediated if required.

2.5. Data analysis and synthesis

Statistical calculations were performed using R version 4.4.1, packages meta and metafor (Viechtbauer, 2010). Separate multilevel meta-analyses were performed to analyze the effect estimates of each category of blood-based biomarkers in patients with CTS compared to controls. Given that some studies reported multiple biomarkers for a single category, the multilevel approach was used to appropriately account for dependency among effect sizes within studies (Fernández-Castilla et al., 2020). Suitability for meta-analysis had to include at least two studies reporting biomarkers of the same category (e.g. proinflammatory or anti-inflammatory biomarkers) with comparable laboratory analysis.

Multilevel meta-analyses used a random-effects model fitted using restricted maximum likelihood (REML). For meta-analyses with fewer than five studies, the Hartung-Knapp adjustment (IntHout et al., 2014) was applied to improve the robustness of confidence interval estimates in the context of small-study meta-analyses, where conventional methods like DerSimonian-Laird may underestimate uncertainty. Effect estimates (standardized mean differences [SMD]) and measures of precision (95% confidence intervals and prediction intervals [PI]) were reported.

Effect sizes were measured by Hedge’s method considering effects sizes as small if they fall between 0.2 ≤ 0.49, medium between 0.5 ≤ 0.79 and large if they are >0.8 (Tagliaferri et al., 2024). Between and within clusters heterogeneity was calculated and reported using the I² statistic considering low heterogeneity 0-40%, moderate 30-60%, substantial 50-90% and considerable >75% (J. P. T. Higgins and Thompson, 2002).

Sensitivity analyses were performed using the “leave-one-out” method, in which each study is sequentially excluded to evaluate its individual impact on the overall pooled effect estimate. Publication bias was assessed for each meta-analysis through funnel plots and trim-and-fill method.

Additionally, subgroup analyses were performed for individual biomarkers if two or more studies reported values of a single biomarker with comparable analysis methodology. For these subgroup analyses, a random effects model with the Hartung-Knapp adjustment when needed was performed. When only two studies were available for meta-analysis, a fixed-effect model was applied (unplanned analysis), as estimation of between-study variance is imprecise and unstable with very few studies (Borenstein et al., 2010). The fixed-effect model assumes a single true underlying effect shared by both studies. This assumption was considered plausible given the homogeneous nature of the sample resulting from strict inclusion criteria focused on idiopathic CTS. Accordingly, results from these analyses should be interpreted as the common effect among the included studies and should not be generalized beyond this evidence base (Borenstein et al., 2010).

For secondary objectives, the association between biomarkers and patients’ symptoms and electrodiagnostic test severity was reported qualitatively in the text as the low number of studies evaluating this association prevented a meta-regression.

Results from single studies which could not be meta-analyzed where narratively synthesized using the “Guidance on the Conduct of Narrative Synthesis in Systematic Reviews: A Product from the ESRC Methods Programme” (Popay et al., 2006) as guidance to report these findings.

2.6. Certainty of evidence

The Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) approach (Guyatt et al., 2011) was used to assess the certainty of evidence for each meta-analysis by two independent reviewers (J.M.-C. and A. A.-R). Briefly, downgrading by one or two levels was applied according to the following criteria: studies designs (cross-sectional or cohort studies), risk of bias (NOS scale ratings and sensitivity analyses), inconsistency (comparison of confidence versus prediction intervals following CINeMA guidance), indirectness (related to population representativeness), imprecision (confidence intervals crossing minimal clinically important improvement thresholds), and publication bias (visual inspection of funnel plots or Egger and Begg’s test). An upgrade of one level was considered when confounding is expected to increase the effect but no effect was observed, as reported in GRADE Handbook (GRADE handbook).

2.7. Data statement

The dataset and R scripts used for all analyses in this study are publicly available at [https://github.com/CTS-Biomarkers].

3. Results

3.1. Identification and selection

The study selection process is illustrated in Fig. 1. The bibliographic search yielded 7424 results, from which 3289 duplicates were removed. A total of 4135 articles were screened by title, abstract and keywords, resulting in the exclusion of 4088 articles. The remaining 47 articles were assessed for eligibility via full text review, leading to 28 studies being included in this review. Excluded articles after title and abstract screen are listed in SUPPLEMENTARY MATERIAL S2.

Fig. 1. PRISMA 2020 flow diagram.

3.2. Main characteristics

The characteristics of each study are described in detail in Table 1.

Table 1. Included articles - population characteristics and biomarkers analyzed.

	AUTHOR YEAR	CTS (N, SEX, AGE [Mean ± SD] years)	CONTROL (N, SEX, AGE [Mean ± SD] years)	DIAGNOSTIC CRITERIA	DISEASE DURATION (Months)	BIOMARKERS	LABORATORY TECHNIQUE	QUESTIONNAIRES	CONCLUSIONS
1.	Karacan G ölen and Yilmaz Okuyan, 2023	N: 300 (F: 226; M: 74) Age: 53.9 ± 13.03	N: 100 (F: 78; M: 22) Age: 42 ± 15	Clinical and electrophysiological	-	Neutrophils Lymphocytes Platelets NLR SII	-	BCTQ	A significant relationship was found between age, increased BMI, and the SII index and CTS suggesting that inflammation may play a role in the pathophysiology of CTS
2.	Freeland et al., 2002	N: 41 (F: 31, M: 10) Age: 46.7	N: 21(–) Age: -	Clinical and electrophysiological	-	IL-1 IL-6 PGE2 MDA	ELISA	-	Biochemical mediators, PGE2, and IL-6, may be involved in the pathophysiology of idiopathic CTS
3.	Nageeb et al., 2018	N: 50 (F: 35, M: 15) Age: 42.3 ± 11.8	N: 50 (F:30, M: 20) Age: 40.4 ± 12.9	Clinical and electrophysiological (Wang criteria)	-	Vitamin D 25-hydroxyvitamin D (25-OHD)	Radioimmunoassay method using 25-OH Vitamin D EIA Kit	BCTQ	CTS patients were found to have lower levels of vitamin D and higher BMI compared to controls. Additionally, pain severity was highly related to low vitamin D levels
4.	Bergsten et al., 2022	N: 189 (F: 128, M: 61) Age: 57 ± 5.7	N: 4568 (F: 2,731, M: 1.837) Age: 57.5 ± 6.0	Clinical and in some cases combined with electroneurography	-	Caspase-3 Caspase-8 HSP27	OLINK Proseek R© Multiplex	-	No association between CTS and plasma levels of peripheral nervous tissue biomarkers (i.e., caspase-3, caspase-8 and HSP27) were found
5.	a Utrobi či ć et al., 2014	N: 71 (F: 56, M: 15) Age: 59 ± 11.2	N: 68 (F: 54, M: 14) Age: 55 ± 11.3	Clinical and electrophysiological	-	Fibrinogen ESR	Fibrinogen concentrations: Multifibren U test. Rates of erythrocyte sedimentation: Westergren method.	-	An association was identified between idiopathic CTS and increased fibrinogen levels in plasma and transversal carpal ligament deposits
6.	Abdul-razzak and Kofahi, 2020	N: 48 (F: 40, M: 8) Age: 48.55 ± 8.3	N: 48 (F: 40, M: 8) Age: 42.33 ± 8.3	Clinical and electrophysiological	-	Vitamin D 25-hydroxyvitamin D (25-OHD)	Chemiluminescent assay	-	Low vitamin D levels and HADS Anxiety were significant independent predictors of CTS
7.	Yeo et al. 2010	N: 114 (F: 91, M: 23) Age: 56.04 ± 9.61	N: 74 (F: 56, M: 18) Age: 51.72 ± 10.88	Electrophysiological	-	TC TG HDL LDL	-	-	High serum TG may act as an aggravating factor of CTS
8.	Güneş andBüyükgöl, 2020	N: 407 (F: 310, M:97) Age 54.6 ± 12.7	N: 206 (F: 146, M:60) Age: 46.6 ± 12.8	Clinical and electrophysiological	-	WBC CRP Neutrophils Lymphocytes NLR Platelets PLR	Autoanalyzer (Sysmex XN-1000 hematology analyzer, Kobe, Japan)	-	Neurophysiologically more severe CTS was associated with higher NLR levels
9.	Ajeena et al., 2021	N: 64 (F:57, M: 7) Age: -	N: 15 Age: -	Clinical and electrophysiological	-	CCL5 (RANTES)	ELISA	-	CCL5 levels were found to be increased in patients with CTS, indicating its potential use as a predictor for diagnosis
10.	Deveci and Matur, 2023	N: 166 (F: 127, M: 39) Age: 46.19 ± 11.93	N: 80 (F: 54, M:26) Age: 43 ± 15.1	Clinical and electrophysiological	-	Hb Ferritin Ca Mg Vitamin D 25-hydroxyvitamin D (25-OHD) Vitamin B12 Folic acid	-	BCTQ	Anemia and lower ferritin levels were common among patients with CTS; however, these factors did not directly contribute to symptom severity, and the significance of this finding remains unclear
11.	Arshad et al., 2024	N: 32 (F: 21, M:11) Age: 43.8 ± 7.2	N: 32 (F: 22, M: 10) Age: 39 ± 5.2	Electrophysiological	1.35 ± 1.12	IL-6 TNF-α NO MDA SOD	IL-6, TNF-a and NO ELISA SOD Absorbance MDA Spectrophotometer	BCTQ	CTS might be a disease of sterile inflammation and oxidative stress imbalance, with higher inflammatory and oxidative stress markers indicating more severe disease
12.	Nakamichi and Tachibana, 2005	N: 185 (F:153, M.32) Age: 54 ± 4.75	N: 219 (F: 138, M: 81) Age: 52 ± 5.5	Clinical and electrophysiological	-	TG HDL LDL	-	-	High LDL levels in middle age were identified as a risk factor for idiopathic CTS, suggesting that LDL-correlated median nerve enlargement may increase the volume of carpal tunnel contents
13.	Moalem-Taylor e al. 2017	N: 26 (F: 21, M: 5) Age: 56.4 ± 10.1	N: 26 (N: 18, M: 8) Age: 54 ± 10.5	Clinical and electrophysiological	-	CCL4 (MIP-1 β) CCL5 (RANTES) Eotaxin CXCL10 (IP-10) CCL2 (MCP-1) CXCL8 (IL-8) IL - [1RA, 2, 4, 5, 6, 7, 9, 10, 12, 13, 17A] IFNγ TNFα VEGF PDGF GM-CSF Neutrophils Monocytes Lymphocytes Leukocytes	Multiplex Bio-Plex Pro Human Cytokine 27 Assays kit	DN4 NPSI	CTS was found to be associated with adaptive changes in memory T cell homeostasis and increased systemic inflammatory modulating cytokines and chemokines, potentially regulating neuropathic symptoms
14.	Sağir et al., 2021	N: 100 (F:84, M:16) Age: 52,65 ± 11,74	N: 40 (F:35, M:5) Age: 50,08 ± 11,75	Clinical and electrophysiological	-	Vitamin D 25-hydroxyvitamin D (25-OHD) Vitamin B12 Folic acid	-	BCTQ SLANSS	No statistically significant correlation was observed between serum levels of vitamin D, vitamin B12 and folic acid and clinical scales in CTS patients
15.	Yusifov et al., 2022	N: 41 (F: 33, M:8) Age: 46.07 ± 8.31	N: 43 (F: 32, M: 11) Age 42.35 ± 10.34	Electrophysiological	-	TG HDL LDL TC HbA1c	-	-	Metabolic syndrome was more prevalent in CTS patients. Some clinical and electrophysiological features, mainly sensory thresholds, worsened in the presence of metabolic syndrome
16.	Sariçam, 2018	N: 130 (F: 113, H: 17) Age: 47.51 ± 8.14	N: 130 (F: 101, M: 29) Age: 45.65 ± 13	Electrophysiological	-	NLR PLR NMR LMR CRP Leukocytes	Complete blood count Automatic blood count (ADVIA 2120 I) CRP Turbidimeter (BECKMAN COULTER AU680).	-	High levels of NMR, platelets, and CRP were found; however, no relationship was detected between disease severity and these parameters
17.	Oh el al., 2013	N: 6 (F: 5, M: 1) Age: 50.7 ± 7.4	N: 6 (F: 5, M: 1) Age: 51.8 ± 7.4	Clinical and electrophysiological	2.08 ± 0.73	Vitamin D-binding protein Chain A, heat shock 70- kDa protein, 42-kDa ATPase N-terminal domain Fibrinogen gamma chain Apolipoprotein A-IV Clusterin Heterogeneous nuclear ribonucleoprotein H1 Glutathione-insulin transhydrogenase (216 AA) cAMP-dependent protein kinase inhibitor Alpha Mutant beta-globin	Two Dimensional Electrophoresis, Silver Staining, Image analysis, destaining and Trypsin Digestion.	-	Ten proteins were identified as significantly and consistently altered in the serum of CTS patients, with four upregulated and six downregulated. Further studies will be required to determine the pathophysiologic role of these proteins in CTS
18.	Baričiće al., 2023	N: 17 (F:8, M: 9) Age: 59.5 (48-81)	N: 15 (F: 3, M:12) Age: 60 (60-64)	Clinical and electrophysiological	-	BMP-7 IL-1β TGF-β1 TNFα CRP	ELISA	-	In end-stage CTS, serum levels of IL1-β and TNFα were not altered, while TGF-β1 and BMP-7 levels were significantly higher, especially in patients with coexisting OA and CTS
19.	Karim et al., 2021	N: 40 (F: 40) Age: 45.07 ± 8.52	N: 40 (F: 40) Age: 45.32 ± 8.42	Clinical and electrophysiological	-	TNF-α IL1 β IL-6 IL-10	ELISA	BCTQ	Serum levels of inflammatory cytokines (IL1, IL6, IL10, and TNFα) did not show meaningful changes in CTS patients
20.	Fattah et al., 2023	N: 100 (F:70; M:30) Age: 47.5 ± 14	N: 100 (F: 60, M:40) Age: 46.4 ± 11.8	Clinical and electrophysiological	-	TGF-β1 MIP-1β	ELISA	-	Increased serum TGF-β1 and MIP-1β levels were observed in CTS patients.
21.	Sandy-Hindmarch et al., 2022	N: 55 (F: 37, M: 18) Age: 64 (16)	N: 21 (F: 14, M: 7) Age: 63 (21)	Clinical and electrophysiological	36 (42)	TGF-β1 CCL5 IL - [1β, 2, 4, 6, 9, 10, 12, 17] CXCL10 CCL2 VEGF CXCL8 IFN-γ CRP GM-CSF TNF-α CXCL5 Fractalkine	Cytokines: Multiplex U-PLEX Plate Custom Biomarker Multiplex Assay Kit CCL5: R-PLEX plates CRP: ELISA RNA: Rt-PCR	BCTQ NPSI	Increased TGF-β and CCL5 protein levels were observed in the active stage of CTS. Specific dysregulation of systemic cytokine expression was demonstrated in both the active and resolution phases of nerve injury and neuropathic pain
22.	Tekcan et al., 2016	N: 155 (F:135, M: 20) Age: 46.85 ± 11.45	N: 140 (F: 106, M: 34) Age: 46.59 ± 11.03	Clinical and electrophysiological	3.02 ± 1.85	IL4	PCR	-	The IL-4 70 bp VNTR polymorphism was not identified as a relevant CTS marker, although the P1 allele may be related to CTS
23.	Gürsoy et al, 2016	N: 108 (F: 94, M: 14) Age: 44.7 ± 11.6	N: 52 (F: 46, M: 6) Age: 41.3 ± 11.7	Clinical and electrophysiological	-	Vitamin D 25-hydroxyvitamin D (25-OHD)	Radioinmuno assay	BCTQ	Vitamin D deficiency was found to be common among patients with CTS symptoms, but not correlated with clinical symptoms
24.	Cevik et al., 2017	N: 158 (F: 136, M: 22) Age: 48.21 ± 11.22	N: 151 (F: 116, M: 35) Age: 45.79 ± 11.02	Clinical and electrophysiological	3.02 ± 1.85	ACE polymorphisms IL1RA	PCR	-	No associations were found between IL- 1Ra and ACE I/D polymorphisms and susceptibility to CTS.
25.	Tutoglu et al., 2014	N: 48 (F:40, M:8) Age: 70.64 ± 5.13	N: 72 (F: 35; M: 37) Age: 68.87 ± 4.20	Electrophysiological	-	CRP ESR Leukocytes	-	-	Although CRP levels were significantly elevated in patients, they remained within normal ranges. Increased mean platelet volume levels in older patients may be associated with CTS etiology
26.	Demirkol et al., 2012	N: 43 (F: 38, M: 5) Age: 43.30 ± 10.48	N: 43 (F: 36, M: 7) Age: 41.76 ± 9.92	Clinical and electrophysiological	30.9	TAS TOS OSI	Novel automated method	-	Changes in oxidative stress and antioxidant defenses were observed in CTS patients. Increased TOS and OSI, along with decreased TAS, might stimulate fibrosis in the tenosynovium and median nerve, potentially playing a role in CTS occurrence and progression.
27.	Şanli and Çetin, 2024	N: 99 (F: 74, M:25) Age: 50 (42-61)	N: 43 (F: 33, M:10) Age: 38 (31-44)	Electrophysiological	-	Platelets Neutrophils Lymphocytes CRP NLR PLR 25-hydroxyvitamin D (25-OHD)	WBC: Beckman Coulter UniCel DxH 800 hematology analyzer	-	NLR, PLR, and 25(OH)D levels might be useful biomarkers for the diagnosis and evaluation of CTS severity
28.	Aykurt Karlibel et al. 2025	N: 47 (F: 41, M:6) Age: 45.4 ± 9	N: 34 (F: 30, M:4) Age: 42.5 ± 10.8	Clinical and electrophysiological	5 (3–15)	CRP Albumin HNA% 25-hydroxyvitamin D (25-OHD)	Vitamin D Architect i2000 analyzer CRP BN II system nephelometer HNA Colorimetry with bromocresol purple	BCTQ	HNA% could serve as a novel biomarker for CTS, with a positive correlation with subjective symptom severity and functional status. Additionally, low vitamin D levels may contribute to oxidative stress by increasing HNA% levels.

ACE: angiotensin-converting enzyme; ATPase: adenosine triphosphatase; BCTQ: Boston Carpal Tunnel Questionnaire; BMP-7: bone morphogenetic protein 7; Ca: calcium; cAMP: cyclic adenosine monophosphate; CCL2: chemokine (C-C motif) ligand 2; CCL5: chemokine (C-C motif) ligand 5 (RANTES); CRP: C-reactive protein; CXCL5: chemokine (C-X-C motif) ligand 5; CXCL8: chemokine (C-X-C motif) ligand 8 (IL-8); CXCL10: chemokine (C-X-C motif) ligand 10; DN4: Douleur Neuropathique 4; ELISA: enzyme-linked immunosorbent assay; ESR: erythrocyte sedimentation rate; GM-CSF: granulocyte-macrophage colony-stimulating factor; Hb: hemoglobin; HbA1c: glycated hemoglobin; HDL: high-density lipoprotein; HDL-C: high-density lipoprotein cholesterol; HNA: Human non-mercaptoalbumin; HSP27: heat shock protein 27; IFN-γ: interferon gamma; IL: interleukin; IL1Ra: interleukin 1 receptor antagonist; LDL: low-density lipoprotein; LDL-C: low-density lipoprotein cholesterol; LMR: lymphocyte-to-monocyte ratio; MDA: malondialdehyde; Mg: magnesium; MIP-1β: macrophage inflammatory protein 1 beta; N: sample size; NLR: neutrophil-to-lymphocyte ratio; NMR: neutrophil-to-monocyte ratio; NO: nitric oxide; NPSI: Neuropathic Pain Symptom Inventory; OSI: oxidative stress index; PCR: polymerase chain reaction; PDGF: platelet-derived growth factor; PGE2: prostaglandin E2; PLR: platelet-to-lymphocyte ratio; Rt-PCR: reverse transcription polymerase chain reaction; SII: systemic immune-inflammation index; SLANSS: Self-report Leeds Assessment of Neuropathic Symptoms and Signs; SOD: superoxide dismutase; TAS: total antioxidant status; TC: total cholesterol; TGF-β1: transforming growth factor beta 1; TG: triglycerides; TNF-α: tumor necrosis factor alpha; TOS: Total oxidative stress; VEGF: Vascular endothelial growth factor.

^aResults from (Utrobičić et al., 2014) reported in supplementary material were not included, as measures of central tendency and deviation are not reported.

3.2.1. Population characteristics

A total of 2840 CTS patients and 6437 healthy controls were evaluated across the included studies. The CTS group had a mean age of 52 years old (SD ± 13.3) and was predominantly female (79.0%), whereas the control group age was 52.9 years (SD ± 11.7) and included 63.05% females. Only seven studies reported disease duration, spanning from a mean of 1.35 months (SD ± 1.12) in the shortest disease duration study (Arshad et al., 2024) to a median of 36 months (IQR = 42) in the longest one (Sandy-Hindmarch et al., 2022).

3.2.2. Carpal tunnel syndrome diagnosis

Six studies report a diagnosis of CTS made exclusively through nerve conduction studies (Arshad et al., 2024; Şanlı and Çetin, 2024; Sariçam, 2018; Tutoglu et al., 2014; Yeo et al., 2010; Yusifov et al., 2022). The other 22 articles report a diagnosis made with a combination of both clinical and electrophysiological measures.

3.2.3. Biomarkers

Among the 28 included studies, a total of 77 distinct biomarkers were identified and classified into 10 functional categories: proinflammatory cytokines/chemokines (27.6%), anti-inflammatory cytokines/chemokines (4 %), white blood cell counts (6.6%), white blood cell ratios (9.2%), fibrosis-related markers (2.6%), growth/stimulating factors (1.3%), oxidative stress markers (7.9%), vitamins and minerals (6.6%), cardiovascular health related markers (7.9%) and others (26.3%) (Fig. 2A).

Fig. 2.

(A) Distribution of biomarker categories. Percentages are adjusted by splitting the weight of biomarkers included in two categories. (B) Summary of effect sizes from each biomarker category meta-analysis. Dot sizes are proportional to the frequency (percentage) of each category.

Proinflammatory markers including cytokines/chemokines were evaluated in 13 studies (Ajeena et al., 2021; Arshad et al., 2024; Aykurt Karlıbel et al., 2025; Baričić et al., 2023; Fattah et al., 2023; Freeland et al., 2002; Güneş and Büyükgöl, 2020; Karimi et al., 2021; Moalem-Taylor et al., 2017; Sandy-Hindmarch et al., 2022; Şanlı and Çetin, 2024; Sariçam, 2018; Tutoglu et al., 2014), covering 23 markers. The most commonly evaluated were CRP (C-reactive protein) (seven studies) (Aykurt Karlıbel et al., 2025; Baričić et al., 2023; Güneş and Büyükgöl, 2020; Sandy-Hindmarch et al., 2022; Şanlı and Çetin, 2024; Sariçam, 2018; Tutoglu et al., 2014) followed by IL-6 (interleukin 6) (five studies) (Arshad et al., 2024; Freeland et al., 2002; Karimi et al., 2021; Moalem-Taylor et al., 2017; Sandy-Hindmarch et al., 2022), and by TNF-α (tumor necrosis factor alpha) (five studies) (Arshad et al., 2024; Baričić et al., 2023; Karimi et al., 2021; Moalem-Taylor et al., 2017; Sandy-Hindmarch et al., 2022).

Anti-inflammatory cytokines/chemokines were analyzed in seven studies (Baričić et al., 2023; Fattah et al., 2023; Güneş and Büyükgöl, 2020; Karimi et al., 2021; Moalem-Taylor et al., 2017; Sandy-Hindmarch et al., 2022; Utrobičić et al., 2014), accounting for four biomarkers. The most commonly evaluated biomarkers were TGF-β1 (transforming growth factor beta-1) (Baričić et al., 2023; Fattah et al., 2023; Sandy-Hindmarch et al., 2022) and IL-10 (Karimi et al., 2021; Moalem-Taylor et al., 2017; Sandy-Hindmarch et al., 2022), each reported in three studies. Two studies were found for IL-4 (Moalem-Taylor et al., 2017; Sandy-Hindmarch et al., 2022) and one for IL-1RA (Moalem-Taylor et al., 2017).

Six studies analyzed white blood cell-related markers (Moalem-Taylor et al., 2017; Güneş and Büyükgöl, 2020; Karacan Gölen and Yilmaz Okuyan, 2023; Şanlı and Çetin, 2024; Sariçam, 2018; Tutoglu et al., 2014). Total leukocyte count (Güneş and Büyükgöl, 2020; Moalem-Taylor et al., 2017; Sariçam, 2018; Tutoglu et al., 2014) (as a measure of the overall number of white blood cells) was evaluated in four studies, and neutrophils (Güneş and Büyükgöl, 2020; Karacan Gölen and Yilmaz Okuyan, 2023; Moalem-Taylor et al., 2017; Şanlı and Çetin, 2024; Sariçam, 2018) and lymphocytes (Moalem-Taylor et al., 2017; Güneş and Büyükgöl, 2020; Karacan Gölen and Yilmaz Okuyan, 2023; Şanlı and Çetin, 2024; Sariçam, 2018) were examined in five studies each. Monocytes (Karacan Gölen and Yilmaz Okuyan, 2023; Moalem-Taylor et al., 2017; Sariçam, 2018) were assessed in three studies.

Some of these studies reported data regarding white blood cells ratios. PLR (Platelet/lymphocyte ratio) and NLR (neutrophil/lymphocyte ratio) were assessed in three studies (Güneş and Büyükgöl, 2020; Şanlı and Çetin, 2024; Sariçam, 2018). Also, NMR (neutrophil/monocyte ratio) (Sariçam, 2018), LMR (leukocyte/monocyte ratio) (Sariçam, 2018) and composite measure of SII (systemic immune inflammation index) (Karacan Gölen and Yilmaz Okuyan, 2023) were analyzed in one study each. SII was described as neutrophil count × platelet count/lymphocyte count (Karacan Gölen and Yilmaz Okuyan, 2023).

Five studies included fibrosis associated markers (Baričić et al., 2023; Fattah et al., 2023; Moalem-Taylor et al., 2017; Sandy-Hindmarch et al., 2022; Utrobičić et al., 2014), analyzing four different biomarkers. TGF-β1 was assessed in three of them (Baričić et al., 2023; Fattah et al., 2023; Sandy-Hindmarch et al., 2022), followed by VEGF (vascular endothelial growth factor) assessed in two (Moalem-Taylor et al., 2017; Sandy-Hindmarch et al., 2022). Fibrinogen (Utrobičić et al., 2014) and PDGF (Moalem-Taylor et al., 2017) (platelet-derived growth factor) were evaluated in one study each.

Cardiovascular health-related markers were evaluated in three studies (Nakamichi and Tachibana, 2005; Yeo et al., 2010; Yusifov et al., 2022), analyzing five different markers. Triglycerides, HDL (high--density lipoprotein) and LDL (low-density lipoprotein) were analyzed in all three studies (Nakamichi and Tachibana, 2005; Yeo et al., 2010; Yusifov et al., 2022), followed by cholesterol (Yeo et al., 2010; Yusifov et al., 2022), assessed in two studies. Additionally, HbA1c (glycated hemoglobin) (Yusifov et al., 2022) was evaluated in one study.

As for vitamins and minerals, seven studies were included (Abdul-razzak & Kofahi, 2020; Aykurt Karlıbel et al., 2025; Deveci and Matur, 2023; Gürsoy et al., 2016; Nageeb et al., 2018; Sağir et al., 2021; Şanlı and Çetin, 2024). Vitamin D was the most common (25-hydroxyvitamin D (25-OHD)), assessed in seven studies (Abdul-razzak & Kofahi, 2020; Aykurt Karlıbel et al., 2025; Deveci and Matur, 2023; Gürsoy et al., 2016; Nageeb et al., 2018; Sağir et al., 2021; Şanlı and Çetin, 2024), followed by vitamin B12 and folic acid, both assessed in two studies (Deveci and Matur, 2023; Sağir et al., 2021). Calcium and magnesium were evaluated in one study (Deveci and Matur, 2023).

Oxidative stress-related markers were reported in three studies (Arshad et al., 2024; Demirkol et al., 2012; Freeland et al., 2002). MDA (malondialdehyde) was the only one analyzed in two studies (Arshad et al., 2024; Freeland et al., 2002), while SOD (Arshad et al., 2024) (superoxide dismutase), NO (Arshad et al., 2024) (nitric oxide), TAS (Demirkol et al., 2012) (total antioxidant status), TOS (Demirkol et al., 2012) (total oxidative stress) and OSI (Demirkol et al., 2012) (oxidative stress index) were evaluated in one study.

Growth and stimulation factors were assessed in three studies (Baričić et al., 2023; Moalem-Taylor et al., 2017; Sandy-Hindmarch et al., 2022), being VEGF and GM-CSF (granulocyte-macrophage colony-stimulating factor) the most common, evaluated in two (Moalem-Taylor et al., 2017; Sandy-Hindmarch et al., 2022). BMP-7 (Baričić et al., 2023) (bone morphogenetic protein 7), PDGF (Moalem-Taylor et al., 2017) were only assessed in one study.

Lastly, other biomarkers not included in the previous categories were evaluated. This accounts for platelets (Güneş and Büyükgöl, 2020; Karacan Gölen and Yilmaz Okuyan, 2023; Şanlı and Çetin, 2024; Sariçam, 2018; Tutoglu et al., 2014), ESR (erythrocyte sedimentation rates) (Tutoglu et al., 2014; Utrobičić et al., 2014), ferritin (Deveci and Matur, 2023), hemoglobin (Deveci and Matur, 2023), albumin and human non-mercaptoalbumin (HNA) (Aykurt Karlıbel et al., 2025), fasting glucose (Yusifov et al., 2022), ACE (I/D) polymorphisms (angiotensin-converting enzyme) (Cevik et al., 2017), markers related to cellular stress Caspase-3, Caspase-8 and HSP27 (Heat shock protein 27) (Bergsten et al., 2022), as well as other proteins involved in stress response, coagulation, lipid metabolism, RNA processing, redox regulation and hemoglobin function (e.g., vitamin D-binding protein, HSP70, fibrinogen gamma chain, apolipoprotein A-IV, clusterin, hnRNP H1, glutathione-insulin transhydrogenase, PKI-α, and mutant beta-globin) (Oh et al., 2013).

3.2.4. Laboratory techniques

As for the laboratory techniques involved, ELISA technique was the most commonly used, reported in seven studies (Ajeena et al., 2021; Arshad et al., 2024; Baričić et al., 2023; Fattah et al., 2023; Freeland et al., 2002; Karimi et al., 2021; Sandy-Hindmarch et al., 2022). Multiplex assays where reported in three studies, two based on electrochemiluminescence (Moalem-Taylor et al., 2017; Sandy-Hindmarch et al., 2022) and one on quantitative polymerase chain reaction (qPCR) (Bergsten et al., 2022). Seven studies did not report the techniques used (Deveci and Matur, 2023; Karacan Gölen and Yilmaz Okuyan, 2023; Nakamichi and Tachibana, 2005; Sağir et al., 2021; Tutoglu et al., 2014; Yeo et al., 2010; Yusifov et al., 2022). Additional methods are summarized in Table 1.

3.3. Risk of bias

Out of the 28 included studies, 27 were evaluated with the NOS for cross-sectional studies and one with the cohort studies adaptation. Among the cross-sectional studies, 20 were graded as moderate risk of bias (Abdul-razzak & Kofahi, 2020; Arshad et al., 2024; Aykurt Karlıbel et al., 2025; Baričić et al., 2023; Cevik et al., 2017; Demirkol et al., 2012; Deveci and Matur, 2023; Fattah et al., 2023; Freeland et al., 2002; Karacan Gölen and Yilmaz Okuyan, 2023; Nageeb et al., 2018; Nakamichi and Tachibana, 2005; Oh et al., 2013; Sağir et al., 2021; Şanlı and Çetin, 2024; Tekcan et al., 2016; Tutoglu et al., 2014; Utrobičić et al., 2014; Yeo et al., 2010; Yusifov et al., 2022) and seven as low risk of bias (Ajeena et al., 2021; Güneş and Büyükgöl, 2020; Gürsoy et al., 2016; Karimi et al., 2021; Moalem-Taylor et al., 2017; Sandy-Hindmarch et al., 2022; Sariçam, 2018). The cohort study was graded with an 8 out of 9 points (Bergsten et al., 2022).

Detailed scores are provided in SUPPLEMENTARY MATERIAL S3.

3.4. GRADE approach

Three of the included meta-analyses were rated as moderate quality of evidence (white blood cell count and ratios, and fibrosis-related markers), five as low (proinflammatory and anti-inflammatory makers, cardiovascular health-related markers, oxidative stress markers and growth/stimulation factors) and one as very low quality (vitamins and minerals). Results of GRADE approach are described in detail in SUPPLEMENTARY MATERIAL S3.

3.5. Meta-analyses

A summary of the results of each overall meta-analysis can be found in Fig. 2B. A more detailed description of the multilevel meta-analyses, which complements the forest plots, can be found in SUPPLEMENTARY MATERIAL S4.

Given the use of random-effects models and the considerable heterogeneity observed across analyses (reflected in high I² values), the summary effects presented below should be interpreted as the average alteration across a distribution of true effects. Prediction intervals further describe this dispersion.

3.5.1. Proinflammatory cytokines/chemokines

The overall meta-analysis for proinflammatory markers at protein level shows, with low certainty, a significant increase in CTS patients compared to controls with moderate effect size [SMD (95% CI) = 0.63 (0.13/1.12), p < 0.013, I²-total = 90.21%, PI= (-1.10/2.35), GRADE = Low]. (Fig. 3A).

Fig. 3.

(A) Meta-analysis of proinflammatory markers (multilevel random-effects model). Zero readings in Moalem-Taylor et al., 2017 for IL-2 and IL-3 and in Sandy-Hindmarch et al., 2022 for IL-2 and IL-17 represent flooring effects. (B) Meta-analysis of anti-inflammatory markers (multilevel random-effects model).

3.5.2. Proinflammatory cytokines/chemokines biomarker-specific meta-analyses

Subgroup meta-analyses were performed for CCL5, IL-6, IL1β, TNFα, CCL4, CXCL10, CCL2, CXCL8, INF-γ, IL-12, IL9 and CRP. Significantly elevated concentrations with large and medium effect sizes were found in CTS patients compared to controls for CCL4 [SMD (95% CI) = 2.14 (1.82/2.46), p < 0.0001, I² = 97.4%], CXCL10 [SMD (95% CI) = 0.80 (0.41/1.19), p = 0.0005, I² = 73.4%], CCL2 [SMD (95% CI) = 0.47 (0.1/ 0.85), p = 0.013, I² = 0%], CXCL8 [SMD (95% CI) = 0.57 (0.19/0.95), p = 0.0031, I² = 31.2%] and CRP [SMD (95% CI) = 0.32 (0.12/0.53), p = 0.0199, I² = 63.4%]. (SUPPLEMENTARY MATERIAL S5.A).

No significant alterations were found for INF-γ, IL2, IL12, IL9, CCL5, IL-1β, TNF-α and IL6.

Several markers were only reported in one study, so no subgroup meta-analyses could be performed. Among them, CCL11, IL-7 and IL-17A were found to be increased (Moalem-Taylor et al., 2017) in CTS patients compared to healthy controls, whereas no significant differences were found for IL-1 (Freeland et al., 2002), PGE2 (Freeland et al., 2002), IL-5 (Moalem-Taylor et al., 2017), IL-2 (Moalem-Taylor et al., 2017; Sandy-Hindmarch et al., 2022), IL-13 (Moalem-Taylor et al., 2017), IL-17 (Sandy-Hindmarch et al., 2022), CXCL5 (Sandy-Hindmarch et al., 2022) and Fractalkine (Sandy-Hindmarch et al., 2022).

In addition to protein analysis, Sandy-Hindmarch et al., 2022 analyzed gene expression through real-time qPCR. Only PTGES2, a gene encoding a membrane bound enzyme which catalyzes the conversion of Prostaglandin H2 to Prostaglandin E2, was found to be significantly decreased in patients compared to healthy controls (adjusted p = 0.013).

3.5.3. Anti-inflammatory cytokines/chemokines

The overall meta-analysis for anti-inflammatory markers shows, with low certainty, a large and significant increase in CTS patients compared to controls [SMD (95% CI) = 1 (0.1/1.89), p = 0.029, I²-total = 92.52%, PI= (-1.21/3.21), GRADE = Low]. (Fig. 3B).

3.5.4. Anti-inflammatory cytokines/chemokines biomarker-specific metaanalyses

Subgroup meta-analyses were performed for TGF-β1, IL-10 and IL-4. Significant differences with medium effect size were found between CTS patients and healthy controls for IL-4 [SMD (95% CI) = 0.74 (0.36/1.13), p = 0.0001, I² = 35.8%], but no significant differences were found for TGF-β1 [SMD (95% CI) = 1.75 (−0.54/4.05), p = 0.081, I² = 82.7%] and IL-10 [SMD (95% CI) = 0.09 (−1.07/1.24), p = 0.778, I² = 67.1%] (SUPPLEMENTARY MATERIAL S5.B).

3.5.5. White blood cell count

The overall meta-analysis for white blood cell count shows, with moderate certainty, no significant alteration between CTS patients and controls [SMD (95% CI) = −0.004 (−0.2/0.19), p = 0.96, I²-total = 78.67%, prediction interval (PI)= (−0.69/0.69), GRADE = Moderate]. (Fig. 4A).

Fig. 4.

(A) Meta-analysis of white blood cell count markers (multilevel random-effects model). (B) Meta-analysis of white blood cell ratio markers (multilevel random-effects model).

3.5.6. White blood cell count biomarker-specific meta-analyses

Subgroup meta-analysis revealed with small effect size an increase in neutrophils counts between CTS patients and controls [SMD (95% CI) = 0.2 (0.09/0.31), p = 0.0008, I² = 0%]. No significant differences were found for lymphocytes [SMD (95% CI) = −0.21 (−0.67/0.25), p = 0.37, I² = 93.6%], monocytes [SMD (95% CI) = −0.20 (−0.67/ 0.28), p = 0.22, I² = 46.7%] or leukocytes [SMD (95% CI) = 0.13 (0.03/0.28), p = 0.083, I² = 0%] (SUPPLEMENTARY MATERIAL S5.C).

3.5.7. White blood cells ratios

The overall meta-analysis for white blood cells ratios shows, with moderate certainty, no significant increase in CTS patients compared to controls [SMD (95% CI) = 0.32 (−0.008/0.65), p = 0.05, I²-total = 86.31%, PI= (−0.32/0.97), GRADE = Moderate] (Fig. 4B).

3.5.8. White blood cells ratios biomarker-specific meta-analyses (PLR and NLR)

Our subgroup meta-analysis on PLR and NLR shows no significant changes between patients and controls [PLR - SMD (95% CI) = 0.07 (−0.41/0.55), p = 0.59, I² = 65.2%] [NLR - SMD (95% CI) = 0.16 (−0.39/0.71), p = 0.34, I² = 85.9%] respectively (SUPPLEMENTARY MATERIAL S5.D).

Other ratios like NMR, LMR and SII could not be meta-analyzed due to a lack of studies. Among them, only SII showed a significant increase in CTS patients in a single study (Karacan Gölen and Yilmaz Okuyan, 2023).

3.5.9. Fibrosis-related markers

The overall meta-analysis for fibrosis-related markers shows, with moderate certainty, a significant increase in CTS patients compared to controls with a large effect size [SMD (95% CI) = 1.64 (0.96/2.33), p < 0.0001, I²-total = 85.55%, PI = (0.036/3.25), GRADE = Moderate]. (Fig. 5A).

Fig. 5.

(A) Meta-analysis of fibrosis-related markers (multilevel random-effects model). (B) Meta-analysis of cardiovascular health–related markers (multilevel random-effects model).

3.5.10. Fibrosis-related markers biomarker-specific meta-analyses (VEGF and TGF-β1)

The VEGF subgroup meta-analysis showed significant changes with large effect size in CTS patients compared to controls [SMD (95% CI) = 0.87 (0.48/1.26), p = 0.0001, I² = 84.4%] (SUPPLEMENTARY MATERIAL S5.E). The TGF-β1 results are detailed above.

Both PDGF (Moalem-Taylor et al., 2017) and Fibrinogen (Utrobičić et al., 2014) were analyzed only in one study each. Both fibrotic markers showed a significant increase in patients compared to controls (PDGF: p = 0.00383) (Fibrinogen: p < 10^-7).

3.5.11. Cardiovascular health-related markers

The overall meta-analysis for cardiovascular risk-related markers shows, with low certainty, no significant differences between CTS patients and controls [SMD (95% CI) = 0.26 (−0.1/0.63), p = 0.154, I²-total = 88.79%, PI= (−0.76/1.29), GRADE = Low]. (Fig. 5B).

3.5.12. Cardiovascular health-related markers biomarker-specific meta-analyses

No significant differences were identified in subgroups meta-analyses; triglycerides [SMD (95% CI) = 0.36 (−0.19/0.92), p = 0.105, I² = 62.9%], HDL [SMD (95% CI) = −0.01 (−0.92/0.9), p = 0.962, I² = 84.9%], LDL [SMD (95% CI) = 0.47 (−1.83/2.76), p = 0.474, I² = 97.9%], and cholesterol [SMD (95% CI) 0.02 (−0.23/ 0.26), p = 0.89, I² = 0%] (SUPPLEMENTARY MATERIAL S5.F).

Biomarker Hba1c was reported in only one study (Yusifov et al., 2022), showing no significant differences between CTS patients and controls.

3.5.13. Vitamins and minerals

Significant differences were found, with very low certainty, in blood concentrations of vitamins and minerals between CTS patients and controls, with a moderate effect size [SMD (95% CI) = −0.75 (−1.49/−0.005), p = 0.0048, I² = 95.74%, PI= (−2.53/1.04), GRADE Very Low]. (Fig. 6A).

Fig. 6.

(A) Meta-analysis of vitamin and mineral markers (multilevel random-effects model). (B) Meta-analysis of oxidative stress–related markers (multilevel random-effects model). (C) Meta-analysis of growth and stimulating factors (multilevel random-effects model).

3.5.14. Vitamins and minerals biomarker-specific meta-analyses (vitamin D, B12 and folic acid)

Our subgroup meta-analysis revealed a significant decrease with large effect size in vitamin D in CTS patients compared to controls [SMD (95% CI) = −0.73 (−1.25/−0.21), p = 0.005, I² = 90.7%]. However, no significant differences were found for vitamin B12 [SMD (95% CI) = 0.14 (−0.08/0.35), p = 0.22, I² = 0%] and folic acid [SMD (95% CI) = 0.02 (−0.20/0.23), p = 0.88, I² = 0%] (SUPPLEMENTARY MATERIAL S5.G).

Magnesium and calcium could not be meta-analyzed, but no significant differences were found between patients and controls in one study (Deveci and Matur, 2023).

3.5.15. Oxidative stress markers

With low certainty, no significant changes were found for markers related to oxidative stress [SMD (95% CI) = 1.56 (−0.15/3.27), p = 0.074, I²-total = 96.52%, PI = (−2.06/5.18), GRADE Low]. (Fig. 6B).

3.5.16. Oxidative stress markers biomarker-specific meta-analysis (MDA)

A subgroup meta-analysis could only be performed for MDA, showing significant changes with large effect size in patients compared to controls [SMD (95% CI) = 2.00 (−1.53/2.56), p < 0.0001, I² = 95.5%] (SUPPLEMENTARY MATERIAL S5.H).

Several biomarkers were analyzed in only one study; Demirkol et al. (2012) (Demirkol et al., 2012) showed that TAS blood concentration was higher in healthy controls compared to CTS patients (p = 0.008), but TOS and OSI concentrations were higher in CTS (p = 0.002 and p < 0.001 respectively). Arshad et al. (2024) (Arshad et al., 2024) found that SOD and NO levels were increased in CTS compared to healthy controls (p < 0.0001).

3.5.17. Growth/stimulation factors

The overall meta-analysis for growth/stimulation factors showed, with low certainty, no significant increase in blood-based concentrations in CTS patients compared to controls [SMD (95% CI) = 0.49 (-0.08/1.06), p = 0.09, I²-total = 72.51%, PI= (-0.61/1.59), GRADE = Low]. (Fig. 6C).

3.5.18. Growth/stimulation factors biomarker-specific meta-analyses (VEGF and GM-CSF)

Subgroup meta-analyses were performed for VEGF as described above and for GM-CSF. which showed no significant alteration [SMD (95% CI) = 0.36 (-0.01/0.74), p = 0.056, I² = 0%] (SUPPLEMENTARY MATERIAL S5.I).

Both BMP7 and PDGF were analyzed in only one study and showed no significant differences between CTS and controls (BMP7: p = 0.28) (Baričić et al., 2023) (PDGF: p = 0.00383) (Moalem-Taylor et al., 2017).

3.6. Sensitivity and publication bias analyses

Sensitivity and publication bias analyses for each overall meta-analysis are described in SUPPLEMENTARY MATERIAL S6. In general, all overall meta-analyses remained stable throughout the “leave-one-out” sensitivity analyses, indicating that no study had a decisive influence on the meta-analysis.

Regarding “Trim & fill”, analyses revealed possible publication bias in the meta-analyses of white blood cell count, cardiovascular healthrelated markers and oxidative stress markers. Potential publication bias may be driven mainly by individual studies; however, these studies did not substantially alter the overall effect estimates, as confirmed by the leave-one-out analysis.

3.7. Others

Due to their heterogeneity, several biomarkers could not be grouped into broader categories and are therefore reported individually in Table 2.

Table 2. Changes in biomarkers not included in previous categories.

INCREASED IN CTS	DECREASED IN CTS	NO CHANGE
Fasting glucose1	Hemoglobin2	IL-1RA polymorphisms3
HSP704	Vitamin D-binding protein (VDBP) 4	ACE (I/D) polymorphisms3
Glutathione-insulin transhydrogenase (216 A A)4	Fibrinogen gamma chain4	Caspase-3 and -85
cAMP-dependent protein kinase inhibitor alpha4	Clusterin4	HSP275
Mutant β-globin4	Heterogeneous nuclear ribonucleoprotein H1 (hnRNP H1)4	Platelets6–10
	Apolipoprotein A-IV (ApoA-IV)4	Ferritin2
HNA%11		Albumin11
aErythrocyte sedimentation rate (ESR)10		aErythrocyte sedimentation rate (ESR)6

ACE (I/D): angiotensin-converting enzyme (insertion/deletion); cAMP: cyclic adenosine monophosphate; CTS: carpal tunnel syndrome; HNA %: Human non-mercaptoalbumin HSP27: heat shock protein 27; HSP70: heat shock protein 70; IL-1RA: interleukin-1 receptor antagonist.

^aErythrocyte sedimentation rate showed increased values in CTS in Utrobičić et al. (2014)¹² and no change in Tutoglu A. et al. (2014)⁶.

3.8. Biomarkers correlation with patients’ symptoms

Biomarker correlations with patients’ symptoms were reported in eleven studies (Abdul-razzak & Kofahi, 2020; Aykurt Karlıbel et al., 2025; Güneş and Büyükgöl, 2020; Gürsoy et al., 2016; Moalem-Taylor et al., 2017; Nageeb et al., 2018; Nakamichi and Tachibana, 2005; Sağir et al., 2021; Sandy-Hindmarch et al., 2022; Şanlı and Çetin, 2024; Yeo et al., 2010). Meta-regression analysis could not be performed due to a lack of studies hence results are reported visually in Fig. 7, with a more detailed description in SUPPLEMENTARY MATERIAL S7.

Fig. 7. Correlation of biomarkers with patients’ symptoms.

4. Discussion

This meta-analysis included a total of 2840 CTS patients and 6437 healthy from 28 included studies, with a total of 77 biomarkers analyzed.

Meta-analyses revealed increased blood-based biomarker concentrations with large effect sizes for fibrosis-related markers and anti-inflammatory markers, and medium effects size for proinflammatory markers and vitamins and minerals. No significant changes were found for oxidative stress markers, growth factors, white blood cells and cardiovascular health-related markers, however, subgroup analyses found neutrophils, MDA and VEGF to be increased, and vitamin D decreased in CTS compared to controls. While these results highlight that CTS involves a complex pathophysiology involving several systems, our findings have to be interpreted in the context of low to very low grade of certainty in six out of nine overall meta-analyses.

Our synthesis of correlation analyses revealed several statistically significant associations between biomarkers and clinical outcomes. However, most of these findings derived from single studies and lacked replication, limiting the robustness of these conclusions. Vitamin D levels showed a moderate negative correlation with electrodiagnostic grade, whereas PLR, NLR, and CRP were positively correlated, suggesting that higher inflammatory burden may correlate with more severe nerve conduction studies. Additionally, LDL was positively correlated with distal motor latency and negatively correlated with nerve conduction velocity.

4.1. Inflammation and immune response

Half of the biomarkers analyzed (47.6%) are involved in immune and inflammatory responses and our findings highlight, albeit indirectly in blood, the involvement of inflammation in this focal nerve injury.

Overall, both proinflammatory and anti-inflammatory markers may be increased in CTS patients. In subgroup analyses an increase in the pro-inflammatory biomarkers CCL4, CXCL10, CCL2, CXCL8, CRP and anti-inflammatory marker IL4 was observed. The simultaneous presence of both pro- and anti-inflammatory mechanisms may reflect concomitant axonal or myelin degeneration and regeneration (Kuffler, 2026; Uçeyler et al., 2009), and/or a failure in resolving inflammation. The resolution of inflammation is a complex process involving multiple immune pathways (Fiore et al., 2023) and previous studies have shown that distinct trajectories of inflammation resolution are associated with pain chronicity or symptom remission in other disorders, including CTS (Parisien et al., 2022; Sandy-Hindmarch et al., 2022). In addition to ongoing compression, dysfunction in these resolution mechanisms could also explain the only short-lasting rather than curative effects of corticosteroid injections in CTS (Ashworth et al., 2024).

However, despite significant changes in inflammatory markers at protein/gene level, systemic alterations in immune cell populations could not be found in CTS patients at an overall level. In subgroup analyses only neutrophils showed a small increase while monocytes and lymphocytes counts did not differ significantly from controls. Taking into account that neutrophil infiltration peaks at early stages after nerve injury (Austin and Moalem-Taylor, 2010), this result supports the hypothesis of a sustained injury likely mediated by ongoing compression. However, this requires further investigation. The lack of detectable differences in monocytes and lymphocytes in blood likely reflects the indirect nature of blood measurements, since both cell types are recruited to the lesion site following peripheral nerve injury (Kim and Moalem-Taylor, 2011). Supporting this, a recent gene expression and histological study in human nerves demonstrate increased densities of intraneural T cells as well as macrophage populations in Morton's neuroma compared to controls (O. P. Sandy-Hindmarch et al., 2024).

The consistent indication of an altered systemic immune response identified here in CTS differs from other systematic reviews that could not find a clear dysregulation of inflammatory mediators in lumbar radiculopathy (Jungen et al., 2019), another entrapment neuropathy. The authors attributed a lack of clear blood biomarker changes in lumbar radiculopathy to heterogeneity in populations, biomarkers, and laboratory methods. Given that lumbar radiculopathy often presents more acutely and with a more severe symptom profile than CTS, it cannot be excluded that the presence of comorbidities in CTS patients may contribute to changes in blood-based biomarkers. Indeed, patients with CTS often present with comorbid conditions such as elevated BMI, osteoarthritis, or rheumatoid arthritis, which are also associated with systemic inflammation. Future research will have to carefully adjust for potential confounders such as medication use, immune diseases, age, BMI or physical activity (Koop et al., 2021).

As for the relationship between altered inflammatory and immune markers with neuropathic pain, only Il-4 showed a weak positive correlation with neuropathic pain (Sandy-Hindmarch et al., 2022), and Il-9 and CCL5 showed inverse correlations (Moalem-Taylor et al., 2017; Sandy-Hindmarch et al., 2022). The weak nature of these correlations could be explained by the fact that pain is not the main symptom in CTS patients (Sobeeh et al., 2022), but also highlight the complexity of neuropathic pain, making it unlikely that a single biomarker could fully capture this multidimensional experience. Additionally, weak positive correlations have been reported between NLR and PLR with CTS severity assessed by electrodiagnostic grade in two studies (Güneş and Büyükgöl, 2020; Şanlı and Çetin, 2024), suggesting a possible correlation between the degree of nerve injury and immune response that needs to be further explored.

4.2. Fibrotic and oxidative stress biomarkers

Altered concentrations of fibrosis-related markers were found, but not of those associated with oxidative stress or tissue remodeling (10.5% of the biomarkers analyzed). In subgroup analyses only VEGF and MDA showed statistically significant increases.

Increased blood concentrations of fibrotic markers are consistent with histological evidence from both animal and human studies, which describe fibrosis as one of the major pathomechanisms associated with nerve compression syndromes (Ettema et al., 2004; Jinrok et al., 2004; Mackinnon, 2002; Pham and Gupta, 2009). In CTS, histological analyses have shown that the subsynovial connective tissue undergoes fibrotic thickening (Ettema et al., 2004; Gingery et al., 2014; Jinrok et al., 2004), which can affect the gliding of the median nerve, thus promoting entrapment (Ettema et al., 2007).

Although blood levels of TGF-β1 did not reach statistical significance, probably due to the indirect nature of blood analyses, these markers mediate fibrotic thickening (Chikenji et al., 2014; Ettema et al., 2004; Gingery et al., 2014; Saito et al., 2017) and might have potential as therapeutic targets in CTS since local inhibition of TGF-β1 and VEGF receptors in SSCT fibroblasts has shown promise in attenuating profibrotic signaling (Yamanaka et al., 2018).

Oxidative stress has also been implicated with nerve injury mainly in animal models (Bittar et al., 2017, 2017, 2017; Shim et al., 2019; Teixeira-Santos et al., 2020). Free radical production has been attributed to inflammation or neurogenic inflammation, and ischemia; these reactive species can exacerbate tissue damage, nociception and sustain inflammation (Grace et al., 2016; Lee et al., 2010; Niella et al., 2024). Although no significant changes were found in markers related to oxidative stress, maybe due to the indirect nature of blood analyses, elevated levels of MDA were found, one of the most widely recognized oxidative markers (Del Rio et al., 2005). Additionally, two other individual markers – NO and SOD – were found to be increased in one study (Arshad et al., 2024). While NO, released by different cell types such as neutrophils (Witko-Sarsat et al., 2000) and macrophages (Boscá et al., 2005), contributes to sensitization of primary afferents and peripheral hyperalgesia, SOD plays an opposing analgesic role by reducing pro-algesic mediators like reactive oxygen species and NO (Twining et al., 2004). This dual elevation suggests that SOD upregulation could represent an adaptive although insufficient counter-regulatory mechanism.

Given the limited number of studies included for subgroup analysis, additional research is needed to clarify the contribution of oxidative stress to CTS pathophysiology and the interplay between NO and SOD.

4.3. Cardiovascular health and vitamins/minerals

Lastly, cardiovascular health-related markers and vitamins/minerals account for 14.5% of the biomarkers analyzed. Obesity as well as hypercholesterolemia or high triglycerides have been proposed as risk factors for the development of CTS (Erickson et al., 2019). Our analysis shows no significant systemic alteration for triglycerides, LDL, HDL or total cholesterol. Given the cross-sectional design of the included studies and methodological limitations, such as variability in adjustment for confounders or potential medication consumption, it is not possible to confirm or refute the role of these cardiovascular factors as risk factors for CTS.

Our subgroup meta-analysis revealed a decrease in vitamin D levels in CTS patients, with all the included studies reporting mean levels indicative of hypovitaminosis (vitamin D < 20 ng/dL) (Holick et al., 2011). However, in three of the five studies, the healthy control groups also fall under this category (Deveci and Matur, 2023; Gürsoy et al., 2016; Sağir et al., 2021).

Lower levels of Vitamin D has been linked to pain in older adults (Hirani, 2012), but its association with neuropathies remains controversial (Alkhatatbeh and Abdul-Razzak, 2019; Jennaro et al., 2020; Morrison et al., 2024; Tobias et al., 2025; Yammine et al., 2020). Extensive evidence associates vitamin D with critical processes in the development and maintenance of the central nervous system (Groves et al., 2014) and in the peripheral nervous system preclinical studies have reported a role of vitamin D in axon regeneration (Chabas et al., 2008), reducing oxidative stress markers (Moslemi et al., 2022) and exerting anti-inflammatory effects in diabetic patients (Karonova et al., 2020; Yammine et al., 2020), highlighting a complex interaction of biomarkers identified here.

In our review, inverse correlations between vitamin D levels and both anxiety and depression (Abdul-razzak & Kofahi, 2020), pain intensity, distal motor latency, motor conduction velocity and electrodiagnostic grade were found (Nageeb et al., 2018; Şanlı and Çetin, 2024). Although Aykurt Karlıbel et al. (2025) found no associations with other electrodiagnostic variables such as motor amplitude and latency or sensory amplitude, latency, and conduction velocity (Aykurt Karlıbel et al., 2025). From a clinical perspective, a recent systematic review suggest that vitamin D supplementation could help manage the symptoms of patients with CTS (Anusitviwat et al., 2021), although evidence is still scarce. Further research is needed in this area to elucidate the role of vitamin D and its supplementation on patients’ symptoms and their improvement.

Our primary approach involved pooling biomarkers into functional categories to provide a broad overview of the systemic pathways involved in CTS pathophysiology. This prespecified strategy helps to identify which biological systems are most consistently altered. However, this approach carries inherent assumptions that warrant careful consideration. Pooling different molecules under a single construct assumes a degree of interchangeability. However, they may exhibit different kinetics, be influenced by different confounders, or even change in opposite directions for instance depending on the disease stage. Consequently, a statistically significant summary effect for a broad category could be driven by a strong signal from individual biomarkers, while others within the same category show no or even opposite effects (e.g. vitamins and minerals meta-analysis). Although we have attempted to mitigate this risk by also providing the individual biomarker meta-analyses as subgroup analyses, the category-level estimates should be interpreted as a global, high-level indicator of pathway involvement rather than a precise or homogeneous effect.

5. Strengths and limitations

This review has some limitations that need to be considered. Regarding our methodology, we did not search grey literature, which implies a potential inclusion bias, although this remains a subject of debate (Schmucker et al., 2017). A wide array of different methodologies were used in the included studies for the biomarkers analyses, which could impact subgroup meta-analyses of individual biomarkers. Reassuringly, the “leave-one-out” sensitivity analysis confirmed stability of our results.

An additional consideration relates to our analytical approach, while we allocated biomarkers to their primary roles (maximum two categories), various biomarkers may have pleiotropic functions and might interact with each other, thus complicating the interpretation of results.

With respect to the primary studies themselves, several inherent limitations constrain the inferences that can be drawn. Due to the observational design of the included studies, causality cannot be inferred, and it is uncertain whether these systemic changes precede or result from nerve compression. The lack of standardized adjustment for key confounders like physical activity levels, medication use, or comorbidities represents a source of potential bias. Furthermore, the duration of CTS symptoms was inconsistently reported or not reported at all, precluding any meaningful subgroup analysis based on disease chronicity. This is particularly relevant as biomarker profiles might evolve over time. Finally, a fundamental consideration when interpreting the results is that findings from systemic blood-based measurements represent indirect surrogates and do not necessarily reflect the magnitude or exact nature of the local pathological processes occurring within the median nerve.

Unlike previous reviews in neuropathic conditions that focus on specific biomarker categories (Jungen et al., 2019), this review’s strength is the inclusion of all blood biomarkers reported in CTS. We thus compiled a comprehensive overview of the physiological processes potentially involved in CTS pathophysiology.

6. Clinical implications

The alterations identified in this review, particularly those related to inflammatory and fibrotic responses, could have significant clinical relevance. Since pharmacological treatments like corticosteroids injections are routinely used to treat CTS in clinical practice (Ashworth et al., 2024; Lusa et al., 2024), phenotyping based on blood-based biomarkers could help clinicians and researchers identify potential treatment responders. In addition, the identification of altered fibrosis-related biomarkers highlights potential new therapeutic avenues for antifibrotic agents, which, although not yet standard in CTS care, show promise in preliminary studies (Yamanaka et al., 2018).

7. Conclusion

Our systematic review provides a comprehensive overview of blood biomarker alterations present in patients with CTS compared to healthy controls. Increased levels of fibrosis-related markers, proinflammatory and anti-inflammatory markers, CRP, neutrophils, as well as a decrease in vitamin D levels have been found. Together, they reveal a complex pathophysiology and offer hypothesis-generating signals to inform the design of future longitudinal and mechanistic studies. Further research is required to help us understand the relationship of these markers to patients’ symptomatology, symptom progression or treatment stratification and response. These findings must be interpreted with caution due to low to very low grade of certainty and high heterogeneity in most of the meta-analyses.

Supplementary Material

Appendix A. Supplementary data