Inflammatory biomarker alterations in patients with fibromyalgia: a systematic review and meta-analysis

Abstract

Objective

To analyze the available scientific evidence on differences in inflammatory biomarker levels between patients with fibromyalgia and healthy controls.

Methods

A systematic review and meta-analysis were conducted in accordance with PRISMA guidelines. Methodological quality was assessed using the Newcastle–Ottawa Scale. Inflammatory, neuroendocrine, and oxidative stress biomarkers were analyzed, along with their association with clinical symptoms.

Results

Eighteen observational studies (case–control and cross-sectional) involving a total of 1,605 participants were included. Of these 18 studies, quantitative data suitable for meta-analysis could be extracted from only 10 studies. The fibromyalgia group showed elevated levels of p =high sensitive C-reactive protein (hsCRP) levels confirmed by meta-analysis too (SMD = 0.34; 95% CI: 0.16 to 0.52; p = 0.0002). In the quantitative synthesis, only hsCRP reached statistical significance, whereas the meta-analyses of hair cortisol, IL-6 and IFN-γ showed no significant between-group differences. In the qualitative synthesis were observed alterations in oxidative stress markers (including MDA and glutathione) and dysregulation of the hypothalamic–pituitary–adrenal axis (cortisol, adrenaline, and serotonin), with lower serotonin levels compared to healthy subjects. These biomarkers were associated with greater severity of clinical symptoms, including pain, fatigue, sleep disturbances, anxiety, and depression. Overall methodological quality was moderate, and heterogeneity between studies was high.

Conclusions

Significant differences in hsCRP levels were observed between patients with fibromyalgia and healthy controls, supporting a role of low-grade inflammatory mechanism in the pathophysiology of the disease. However, methodological heterogeneity limits direct clinical applicability. Further well-designed studies with better control of confounding factors are needed to advance the diagnostic and therapeutic use of these biomarkers.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s41927-026-00677-1.

Introduction

Fibromyalgia (FM) is a syndrome characterized by chronic musculoskeletal pain of unknown origin. In addition to widespread pain, patients often experience other symptoms such as muscle and joint stiffness, fatigue, sleep disturbances, mood alterations, cognitive dysfunction, anxiety, depression, generalized hypersensitivity, and difficulty carrying out daily activities [1, 2].

Although no single event has been identified as the cause of FM, its onset and exacerbation appear to be associated with multiple physical and emotional stressors, including infections, physical or psychological trauma, and other environmental triggers [3]. Epidemiologically, FM affects women more frequently than men (between 80 and 96% of female predominance), with an estimated prevalence of approximately 2–5% of the world population [4], 6.4% in the United States population (7.7% in women and 4.9% in men). Studies conducted in Eurozone report prevalence rates of approximately 2.64% [5], with prevalence increasing with age [3].

Gender differences in FM prevalence may be explained by several factors, including greater susceptibility to anxiety and depression, differences in pain processing, alterations in the central nervous system, and hormonal influences related to the menstrual cycle [6, 7]. From a neurobiological perspective, FM is associated with alterations in central pain modulation. Hypersensitivity to pain has been observed in FM patients, along with imbalances in key neurotransmitters: increased levels of glutamate and substance P, decreased serotonin and noradrenaline in the descending inhibitory pathways of the spinal cord, dopaminergic dysregulation, and altered activity of endogenous opioids [8].

In addition, other mechanisms appear to be involved in the pathophysiology of FM, including neuroendocrine dysfunction, genetic predisposition, oxidative stress, and environmental and psychosocial factors [3]. It is important to note that FM rarely presents as an isolated syndrome; it frequently coexists with other chronic pain disorders or psychiatric conditions, which may complicate diagnosis and increase the likelihood of misdiagnosis [8].

One of the main clinical challenges in addressing FM is the absence of specific biomarkers, which limits diagnostic objectivity. Currently, diagnosis is based exclusively on clinical criteria. The American College of Rheumatology (ACR) published the first classification criteria in 1990, which required widespread pain and tenderness in at least 11 of 18 specific tender points. In 2010, new criteria were introduced based on the Widespread Pain Index (WPI) and the Symptom Severity Scale (SSS), which were later revised in 2016 to improve clinical applicability and to acknowledge the coexistence of FM with other conditions [9, 10].

It has been hypothesized that fibromyalgia may have a peripheral as well as a central inflammatory origin. Chronic low-grade systemic inflammation drives the elevation of pro-inflammatory cytokines (e.g., IL-6, TNF-α, IL-8) and chemokines, which induces peripheral nociceptor sensitization and triggers central neuroinflammation through microglial and astrocytic activation. This reciprocal neuroimmune crosstalk promotes the central sensitization and pain amplification that characterize the pathology [11]. In this context, there is a need to explore potential inflammatory biomarkers to improve diagnostic accuracy and deepen understanding of the pathophysiological mechanisms of FM. Research on biomarkers in FM has gained increasing relevance, as several studies have reported alterations in circulating cytokines, neuroendocrine markers, and oxidative stress indicators in affected patients, suggesting that biological profiling may support clinical assessment even in the absence of a validated diagnostic biomarker [12].Therefore, the main objective of this systematic review was to examine the available scientific evidence regarding differences in the levels of various biomarkers between patients with FM and healthy controls.

Methodology

Study design

This systematic review and meta-analysis was conducted according to a predefined protocol and structured into four phases following the standards of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [13]. The review was registered in PROSPERO on September 25, 2025, under registration number CRD420251111687.

Eligibility criteria

Studies included in this systematic review were required to meet the following criteria:

1. full-text availability in a peer-reviewed journal (no restrictions were applied regarding publication date);

2. written in English or Spanish, based on feasibility considerations and available resources for accurate translation and data extraction;

3. observational study designs, including case-control, cohort, and cross-sectional studies;

4. studies including adult patients diagnosed with FM according to established classification criteria (e.g., the ACR 1990, 2010, or 2016 criteria) and healthy controls; (e) studies evaluating pro-inflammatory biomarkers (such as IL-6, IL-1β, TNF-α, and C-reactive protein), anti-inflammatory biomarkers (such as IL-10), biomarkers related to the hypothalamic–pituitary–adrenal (HPA) axis (such as cortisol, adrenaline, noradrenaline, serotonin, and substance P), and oxidative stress markers (such as malondialdehyde [MDA] and glutathione), as well as their association with clinical manifestations (pain intensity, sleep disturbances, anxiety, and depression), compared with healthy controls This review was restricted to humoral (circulating) biomarkers. Non-humoral measures, including conditioned pain modulation (CPM), gait and balance parameters, and digital biomarkers, were not considered primary outcomes, although some included studies additionally reported functional and psychophysical endpoints. Exclusion criteria were: (a) absence of a healthy control group; (b) studies not reporting circulating (humoral) biomarkers; and (c) reviews, case reports, conference abstracts, editorials, and animal studies.

Search strategy

The systematic search was conducted by two independent reviewers (CS and EQ). The following electronic databases were searched: MEDLINE (via PubMed), Scopus, ScienceDirect, Web of Science, CINAHL, and Cochrane Library. No restrictions were applied regarding publication date, and the last search was performed on June 30, 2026. The complete search strategy, reported separately for each database, is provided as Supplementary Material to ensure reproducibility.

Descriptors based on DeCS and MeSH were used. Search terms were derived from combinations of the following MeSH terms: “fibromyalgia,” “biomarker,” “interleukin-6,” “interleukin-1beta,” “interleukin-10,” “tumor necrosis factor-alpha,” “cortisol,” “C-reactive protein,” “substance P,” “pain,” “fatigue,” “sleep disturbance,” “hyperalgesia,” “quality of life,” “depression,” and “anxiety,” as well as the following free-text terms: “IL-6,” “IL-1β,” “IL-10,” “TNF-alpha,” and “CRP.”

The following Spanish-language terms were also used: “fibromialgia,” “biomarcadores,” “interleuquina-6,” “interleuquina-1beta,” “interleuquina-10,” “factor de necrosis tumoral alfa,” “cortisol,” “proteína C reactiva,” “sustancia P,” “dolor,” “fatiga,” “alteración del sueño,” “hiperalgesia,” “calidad de vida,” “depresión,” and “ansiedad.” These terms were combined using the Boolean operators “AND” and “OR” to define the search strategy.

Data extraction process

Two reviewers (CSO and EQR) independently performed study selection and data extraction using a specifically designed data collection form. The following information was extracted: study characteristics (authors, year of publication, title, objectives, study design, and language), participant characteristics (study population and sample size), variables analyzed, and main conclusions.

After the initial selection process, duplicate studies were removed. Studies that were not relevant to the research objective were excluded after title and abstract screening. The remaining articles underwent full-text review to determine eligibility according to the inclusion criteria.

Discrepancies regarding study inclusion were resolved through discussion and consensus. When consensus could not be reached, a third investigator (FJFR) made the final decision.

Methodological quality assessment

The methodological quality of the included observational studies was independently assessed by two reviewers (CSO and EQR) using the Newcastle–Ottawa Scale (NOS) [14], a recommended tool for evaluating non-randomized studies in systematic reviews.

This scale evaluates three main domains: selection of participants (up to 4 points), comparability of study groups (up to 2 points), and assessment of outcomes or exposure (up to 3 points). Stars are assigned for each domain according to predefined criteria, with a maximum possible score of 9 points. Higher scores indicate better methodological quality and lower risk of bias. Studies were classified as low quality (0–3 stars), moderate quality (4–6 stars), or high quality (7–9 stars) [15].

Discrepancies in quality assessment scores were resolved through discussion. When agreement could not be reached, a third investigator (FJFR) was consulted to resolve disagreements.

Level of quality of evidence

The quality of the evidence was assessed using the GRADE approach.

Statistical analysis

Statistical analyses were performed using Review Manager version 5.3 (RevMan 5.3). For quantitative synthesis, mean differences (MD) were calculated when studies used the same measurement scale for the variable of interest, whereas standardized mean differences (SMD) were calculated when different measurement methods or scales were used across studies.

A random-effects model was applied to pool effect sizes, as this approach is more conservative and accounts for both within-study variability and between-study heterogeneity.

Statistical heterogeneity was assessed using Cochran’s Q (Chi-square test) and quantified with the I² statistic, which represents the percentage of total variability attributable to heterogeneity rather than chance. According to Cochrane guidelines, I² values of 0–40% may not be important, 30–60% may represent moderate heterogeneity, 50–90% substantial heterogeneity, and 75–100% considerable heterogeneity. In addition, p-values associated with the Chi-square test were reported to evaluate the statistical significance of the observed heterogeneity.

Effect sizes were presented with their corresponding 95% confidence intervals (95% CI). Overall statistical significance was assessed using the Z-test, with a p-value < 0.05 considered statistically significant. Leave-one-out sensitivity analyses were performed to assess the robustness of the pooled estimates and to identify individual studies that contributed substantially to heterogeneity. Assessment of publication bias through funnel-plot inspection was not undertaken, as fewer than ten studies were available for each outcome, a number below the threshold at which such methods are adequately powered.

All analyses were conducted in accordance with methodological recommendations for meta-analyses of biomarker studies, ensuring robustness, transparency, and reproducibility in the interpretation of the results.

Results

Study selection

A total of 18 articles met the inclusion criteria and were included in this systematic review (Fig. 1). In accordance with PRISMA recommendations, the literature search was conducted independently by two reviewers and initially identified 1082 records.

Fig. 1.

PRISMA flowchart of studies selection

After removing duplicates (n = 267),815 records remained for title and abstract screening. Of these, 453 were excluded based on title/abstract relevance or unavailability of full text, leaving 362 for full-text assessment.

Of these, 48 were selected for full-text review., of which, 30 were excluded due to lack of relevance to the research objective or publication in a language other than English. Therefore, a total of 18 observational studies (case-control design) were included in the qualitative and 10 studies were included in the quantitative synthesis.

Characteristics of included studies

Of the 18 studies included in this systematic review, all involved participants diagnosed with FM; case–control studies additionally included healthy control groups. Overall, a total of 1,712 participants were assessed, comprising 188 men and 874 women, with a mean age of 44 years (range: 18–70 years). Notably, most studies included exclusively female samples, while nine studies incorporated both male and female participants [16–24], the ratio was approximately 1:9.

The included studies were conducted across a geographically diverse set of countries, reflecting the international scope of biomarker research in fibromyalgia. Several studies originated from Turkey, including those by Baris et al. [8], Cure et al. [17], Ellergezen et al. [25], Karabacak et al. [22], and Barut et al. [19]. Spain contributed multiple studies, including two from the same research group at the University of Jaén (Reyes del Paso et al. [9, 26] and García-Hernández et al. [27]), as well as independent studies from Toledo (Arroyo-Fernández et al. [24]), the University of Extremadura (Hinchado et al. [28]), and the University of Santiago de Compostela (Gil-Ugidos et al. [29]). Germany was represented by two studies from the same research group in Heidelberg (Beiner et al. [20, 21]), while Belgium (Polli et al. [10]) and Sweden (Gerdle et al. [30]) each contributed one study. Additional studies were conducted in Taiwan (Hung et al. [18]) and Japan (Izuno et al. [16]), reflecting the involvement of Asian research centers. Norway (Jacobsen et al. [23]) and Brazil (Loçasso et al. [31]) were each represented by a single study. This geographic diversity strengthens the generalizability of the findings, although it also introduces potential sources of heterogeneity related to differences in population characteristics, healthcare systems, and laboratory methodologies across countries.

With regard to biomarkers and assessment instruments, the included studies investigated a wide range of biological, physiological, and psychometric variables. These included serum levels of interleukin-6 (IL-6) and serotonin [31]; circulating concentrations of choline, leptin, and IL-6 [14]; and serum levels of high-sensitivity C-reactive protein (hsCRP) [20]. Several studies evaluated perceived and endocrine stress markers, including diurnal cortisol concentrations or its presence in hair [9, 21, 23, 26]. Genetic and epigenetic mechanisms regulating catechol-O-methyltransferase (COMT) activity and cytokine expression (TNF-α, IFN-γ, IL-6, and TGF-β) were also examined [10, 25]. Additional research explored biochemical, metabolic, and inflammatory profiles, identifying associations between IL-6, IL-8, and symptom severity [17], as well as broader cytokine and chemokine patterns [30]. Furthermore, studies assessed subjective sleep quality [28], oxidative stress markers in relation to muscle pain and fatigue [18], pressure-induced pain sensitivity [24], and conditioned pain modulation (CPM) [29] (Table 1).

Table 1.

Characteristics of included studies

Author (year)	Design	Sample	Variables studied / measurement instrument	Main findings
Arroyo Fernández et al. (2020) [24]	Cross-sectional	n = 26 patients with FM (Female 84,6%); n = 26 healthy controls (Female 84,6%)	Pressure-induced pain: pressure algometer; pain sensitivity: questionnaires	Greater pressure-induced pain in FM patients.
Ellergezen et al. (2021) [25]	Cross-sectional	n = 29 patients with FM; n = 25 controls	IL-2, IL-6, IL-12, IL-17, IFN-γ, TNF-α: ELISA; sleep/fatigue/symptoms: WPI, FIQ and Likert scales	IL-17 was associated with non-restorative sleep.
Jacobesen et al. 2023 [23]	Cross-sectional	n = 64 chronic pain (24 PNP, 40 FM) (Female FM 75%) n = 53 Healthy controls (Not reported)	Hair cortisol; executive functions (EF)	Higher hair cortisol in pain patients vs. healthy controls; no group difference in EF; higher cortisol predicted worse spatial working memory.
Beiner et al. (2023) [20]	Case–control	n = 80 patiens with FM (Female 88,75%); n = 30 healthy controls (Female 40%)	hsCRP: ELISA; clinical symptoms: questionnaires; covariates: age, weight, smoking	Higher hsCRP levels associated with greater symptom burden.
Hinchado et al. (2023) [28]	Cohort	n = 68 patients with FM; n = 68 healthy controls	Sleep: questionnaires; melatonin and cortisol: ELISA	Hormonal and sleep alterations were observed in FM patients.
Hung et al. (2023) [18]	Cross-sectional	n = 51 patients with FM (Female 88,2%); n = 41 healthy controls (Female 83,8%)	Oxidative stress: blood analysis; pain/fatigue intensity: VAS	Increased oxidative stress and oxidative metabolism dysfunction.
Beiner et al. (2024) [21]	Case–control	n = 99 patients with FM (Female 87,9%); n = 50 healthy controls (Female 64%)	Perceived stress: Perceived Stress Scale; salivary cortisol: ELISA; hair cortisol: ELISA	Higher perceived stress in FM, without clear cortisol alterations.
Polli et al. (2024) [10]	Case–control	n = 28 patients with CFS/FM; n = 26 healthy controls	Cytokines (TNF-α, IFN-γ, IL-6, TGF-β): ELISA	Altered cytokine expression without clear diagnostic specificity.
Reyes del Paso et al. (2024) [9]	Case–control	n = 47 patients with FM; n = 36 healthy controls	Cardiovascular regulation: postural recordings; healthy hair cortisol: ELISA	Cardiovascular alterations associated with chronic stress and hair cortisol levels.
Reyes del Paso et al. (2024) [26]	Case–control	n = 47 patients with FM; n = 36 healthy controls	Chronicity/clinical severity: questionnaire; hair cortisol: ELISA	Higher cortisol associated with greater clinical severity, but lower levels with longer chronicity.
Cure et al. (2024) [17]	Retrospective cross-sectional	n = 176 patients with FM (Female 95,5%); n = 184 healthy controls (Female 64,7%)	Biochemical, metabolic and inflammatory profiles: blood analysis and metabolic tests	IL-6 and IL-8 were associated with greater symptom severity.
Gerdle et al. (2024) [30]	Cross-sectional	n = 32 patients with FM; n = 30 healthy controls	Cytokines and chemokines: ELISA; pain intensity: VAS	Inflammatory patterns differentiated patients from controls.
Gil-Ugidos et al. (2024) [29]	Cross-sectional	n = 23 patients with FM; n = 23 healthy controls	CPM: thermal pain threshold test; FM severity: FIQ; anxiety and depression: HADS	CPM differentiated FM patients, although results depended on the protocol used.
Izuno et al. (2025) [16]	Case-control	n = 36 patients with FM (Female 78%); n = 30 healthy controls (Female 90%)	CH50, C3, C4, IgG, IgM, hs	CRPCH50 and hsCRP were elevated in FM; IgG and IgM were reduced
Baris et al. (2025) [8]	Case-control	n = 38 women with FM, n = 38 healthy controls	Choline and leptin: ELISA; IL6; ELISA; pain: numeric rating scale; dietary choline: questionnaire.	Lower choline and leptin levels, and higher IL-6 levels in FM patients.
Barut el al. (2025) [19]	Case-control (Pilot)	N = 18 patients with FM (Female 77,8%); n = 18 controls (Female 61,1%)	miRNA-233-3p, serum IL-1β	miRNA-223-3p reduced; IL-1β not significant
Karabacak et al. (2026) [22]	Case-control	n = 46 patients with FM (Female 95,7%); n = 35 controls (Female 91,4%)	Prealbumin, CRP, complete blood count	CRP and prealbumin were altered in FM
García-Hernández et al. (2026) [27]	Case-control	n = 43 patients with FM; n = 36 controls	Hair cortisol concentration (HCC)	Hair cortisol was higher in women with FM than in healthy women, without reaching statistical significance, consistent with the non-significant meta-analytic finding of this review.

CPM: Conditioned Pain Modulation; ELISA: Enzyme-Linked Immunosorbent Assay; FIQ: Fibromyalgia Impact Questionnaire; FM: Fibromyalgia; HADS: Hospital Anxiety and Depression Scale; HPLC: High-Performance Liquid Chromatography; hsCRP: high-sensitivity C-reactive protein; IFN-γ: Interferon gamma; IL: Interleukin [2, 6, 12, 17]; n: sample size; CFS: Chronic Fatigue Syndrome; TGF-β: Transforming Growth Factor beta; TNF-α: Tumor Necrosis Factor alpha; VAS: Visual Analogue Scale; WPI: Widespread Pain Index; CH50, total complement hemolytic activity; C3, complement component 3; C4, complement component 4; IgG, immunoglobulin G; IgM, immunoglobulin M

Methodological quality assessment

The methodological quality of the included observational studies was moderate to high, with a mean score of 6.79 out of 9 points. According to the Newcastle-Ottawa Scale (NOS) criteria, one study achieved the highest score [28], while ten showed high quality [8, 16, 18–20, 23, 24, 29–31]. On the other hand, six studies obtained a rating of six stars, which indicates moderate quality [9, 10, 17, 21, 25, 26]. Finally, 2 studies showed a low quality [22, 27] In the latter, the main limitation was identified in the exposure domain, specifically due to the omission of the non-response rate; this factor introduces a potential risk of bias that affects the external validity of their results. The details of this assessment are presented in Table 2. Six studies were rated as moderate quality (6/9 stars), their main limitation lying in the outcome/exposure domain (omission of the non-response rate). Because moderate-quality studies were included in the quantitative synthesis, the pooled estimates should be interpreted with corresponding caution.

Table 2.

Methodological quality results

Study (Year)	NOS Score (Quality)	Selection (4)	Comparability (2)	Outcome / Exposure (3)	Methodological quality
Hinchado et al. (2023)	9/9	4	2	3	High
Gerdle et al. (2024)	8/9	4	2	2	High
Gil-Ugidos et al. (2024)	8/9	4	2	2	High
Baris et al. (2025)	8/9	4	2	2	High
Barut et al. (2025)	8/9	4	2	2	High
Hsien-Hung et al. (2023)	7/9	3	2	2	High
Locasso et al. (2024)	7/9	4	1	2	High
Beiner et al. (2023)	7/9	4	2	1	High
Arroyo Fernandez R et al. (2020)	7/9	3	1	3	High
Izuno et al. (2025)	7/9	3	2	2	High
Jacobsen et al. 2023	7/9	2	2	3	High
Cüre et al. (2024)	6/9	3	1	2	Moderate
Ellergezen et al. (2021)	6/9	3	1	2	Moderate
Beiner et al. (2024)	6/9	3	2	1	Moderate
Polli et al. (2022)	6/9	3	2	1	Moderate
Reyes del Paso et al. (2024)	6/9	3	2	1	Moderate
Reyes del Paso et al. (2023)	6/9	3	2	1	Moderate
Karabacak et al. (2026)	5/9	3	0	2	Low
García-Hernández et al. (2026)	5/9	3	0	2	Low

Synthesis of results

Qualitative synthesis

Overall, the included studies consistently highlight the involvement of inflammatory processes, oxidative stress, and dysregulation of stress-response systems in the pathophysiology of FM, as well as their association with the core clinical manifestations of the disorder, including pain, fatigue, and functional impairment.

Pro-inflammatory cytokines

Across the reviewed studies, pro-inflammatory cytokines—particularly interleukin-6 (IL-6) and tumor necrosis factor alpha (TNF-α) tended to be elevated in individuals with FM compared with healthy controls. Although some studies did not observe statistically significant between-group differences, higher IL-6 concentrations were consistently reported among patients presenting greater symptom severity, as assessed using the Fibromyalgia Impact Questionnaire (FIQ). For example, Loçasso et al. (2024) reported a trend toward increased IL-6 levels in patients with more severe clinical presentations, suggesting that low-grade systemic inflammation may be associated with symptom burden in FM.

Similarly, several studies identified elevated concentrations of IL-6 and TNF-α in patients with FM relative to healthy individuals [17, 26, 30]. In addition, positive correlations between IL-6 and IL-8 levels and clinical severity indicators—particularly pain intensity and functional disability—were reported in multiple studies [18, 24, 28]. These findings support the hypothesis that pro-inflammatory cytokines may contribute to symptom amplification and maintenance through mechanisms related to central sensitization.

In contrast, evidence regarding interleukin-1 beta (IL-1β) was less consistent, with no clear or reproducible differences observed between patients and controls across studies [19].

Beyond classical cytokines, alterations in acute-phase reactants have also been described. Karabacak et al. [22] reported significantly elevated C-reactive protein (CRP) and reduced prealbumin levels in patients with FM compared with healthy controls, along with a higher CRP/prealbumin ratio, proposing this composite index as a potentially useful and low-cost inflammatory marker [22]. These findings reinforce the notion that systemic low-grade inflammation in FM may be reflected not only in cytokine profiles but also in routinely available acute-phase markers. Complementing this evidence, Izuno et al. (2025) reported significantly higher complement activity (CH50) and hsCRP, together with lower immunoglobulin G and M levels, in patients with FM relative to sex- and age-matched healthy controls, pointing toward broader immune dysregulation involving both the complement system and humoral immunity, beyond cytokines and acute-phase proteins alone [16].

Oxidative stress

Biomarkers of oxidative stress also demonstrated alterations in individuals with FM compared with healthy controls. In particular, increased levels of malondialdehyde (MDA), a marker of lipid peroxidation, together with alterations in antioxidant systems such as glutathione metabolism, were reported [18]. Oxidative imbalance may contribute to neuronal sensitization, mitochondrial dysfunction, and tissue damage, thereby facilitating the persistence of symptoms such as widespread pain and fatigue. Furthermore, alterations in oxidative metabolism have been associated with phenotypic manifestations commonly observed in FM, including persistent musculoskeletal pain and increased fatigue severity.

Hypothalamic–pituitary–adrenal axis markers

Regarding biomarkers related to hypothalamic–pituitary–adrenal (HPA) axis activity, cortisol and catecholamine levels (adrenaline and noradrenaline) showed variable patterns across studies. Reyes del Paso et al. (2024) observed a trend toward higher cortisol concentrations in patients with FM, particularly among those with greater clinical severity, suggesting potential chronic activation or dysregulation of the stress-response system [26]. Beiner et al. (2024) similarly reported elevated hair and salivary cortisol concentrations in FM, without a clear discriminative pattern relative to healthy controls [21].

Conversely, other studies did not identify consistent differences in cortisol levels between patients and controls, highlighting substantial heterogeneity in findings.

Jacobsen et al. (2023) also reported higher hair cortisol concentrations in the FM group relative to healthy controls, without reaching statistical significance, and found an association between cortisol and impaired working memory performance [23].

In line with this heterogeneity, García-Hernández et al. (2026) reported higher hair cortisol concentrations in women with FM compared with healthy women, although the difference did not reach statistical significance; within the FM group, cortisol levels correlated positively with central sensitization severity, suggesting a possible link between HPA axis activity and symptom severity rather than a robust discriminative biomarker [27].

Baris et al. (2025) also reported alterations in circulating choline and leptin concentrations in individuals with FM [8]. Although these biomarkers are not direct indicators of HPA axis activity, both molecules have been implicated in neuroendocrine regulation and pain modulation pathways, suggesting potential indirect interactions with stress-response mechanisms [8].

Overall, variability in sampling methods (e.g., salivary, serum, or hair cortisol), differences in circadian measurement protocols, and interindividual variability complicate the interpretation of HPA axis biomarkers across studies. Nevertheless, the available evidence suggests a possible role of stress-system dysregulation in the clinical expression and severity of FM.

Clinical symptoms associated

Across all included studies, patients with FM exhibited significantly higher scores for pain (VAS), anxiety, and depression, as well as poorer quality of life compared to healthy controls. Beiner et al. reported a correlation between elevated hsCRP levels and greater symptom burden, including the number of tender points [20]. Similarly, Loçasso et al. identified an association between reduced serotonin levels and poorer quality of life [31].

Additionally, Hinchado et al. incorporated variables such as sleep quality and hormone levels (melatonin and serotonin), identifying both subjective and objective sleep disturbances in patients with FM [28]. These findings further support the concept of FM as a multisystem disorder.

Finally, Gil-Ugidos et al. evaluated the Conditioned Pain Modulation (CPM) paradigm as an objective biomarker of pain in FM [29]. Their results showed that the parallel protocol correctly classified 70% of patients, supporting its potential diagnostic utility, although standardization of measurement methods remains necessary.

Quantitative synthesis of results

IL-6 results

The meta-analysis did not show a statistically significant difference in IL-6 levels between patients with FM and healthy controls. The pooled effect showed a positive but non-statistically significant standardized mean difference (SMD = 2.07; 95% CI: -0.37 to 4.52; p = 0.10). Heterogeneity among the studies was considerable (I² = 97%), indicating substantial variability in the primary results (Fig. 2).

Fig. 2.

Forest plot of IL-6

Hair cortisol results

Regarding hair cortisol, the results showed a positive but non-statistically significant standardized mean difference between patients with FM and healthy controls (SMD = 0.27; 95% CI: -0.07 to 0.60; p = 0.12). Heterogeneity among the studies was moderate (I² = 63%), suggesting some variability in the direction and magnitude of the effect (Fig. 3).

Fig. 3.

Forest plot of hair cortisol

IFN-γ results

The meta-analysis of IFN-γ levels did not show significant differences between groups (SMD = -2.02; 95% CI: 6.00 to 1.97; p = 0.32). Heterogeneity was very high (I² = 98%), reflecting great dispersion in results and limiting the overall interpretation of this biomarker (Fig. 4).

Fig. 4.

Forest plot of IFN-γ

hsCRP results

Regarding hsCRP, the meta-analysis showed a statistically significant increase in patients with FM compared to healthy controls, with a positive standardized mean difference (SMD = 0.34; 95% CI: 0.16 to 0.52; p = 0.0002). Heterogeneity among the studies was low and non-significant (I² = 0%), indicating a consistent effect across studies (Fig. 5).

Fig. 5.

Forest plot of hsCRP

Sensitivity analysis

Sensitivity analyses were conducted to assess the robustness of the findings and explore sources of heterogeneity. For interleukin-6 (IL-6), the sequential omission of individual studies did not materially alter the magnitude, direction, or statistical significance of the primary effect size, despite the persistence of residual heterogeneity. For C-reactive protein (CRP), the pooled effect remained highly stable across all iterations, with heterogeneity remaining consistently null (I² = 0%). These findings demonstrate that the overall conclusions are robust and not driven by any single dominant study or outlier.

For hair cortisol, exclusion of Beiner et al. also reduced heterogeneity to null (I² = 0%) and revealed a statistically significant difference between groups (SMD = 0.43; 95% CI: 0.18 to 0.73; p = 0.0007).

Quality of the evidence

The overall quality of the evidence for the assessed variables was evaluated using the GRADE approach. Following this evaluation, the quality of the evidence was rated as very low (Table 3).

Table 3.

Quality of the evidence

Certainty assessment							№ of patients		Effect		Certainty
№ of studies	Study design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Fibro	healthy controls	Relative (95% CI)	Absolute (95% CI)
IL-6
3	non-randomised studies	not serious	not serious	not serious	not serious	publication bias strongly suspecteda	95	89	-	SMD 2.07 higher (0.37 lower to 4.52 higher)	⨁◯◯◯ Very lowa
Hair cortisol
4	non-randomised studies	not serious	not serious	not serious	not serious	publication bias strongly suspecteda	229	175	-	SMD 0.27 higher (0.07 lower to 0.6 higher)	⨁◯◯◯ Very lowa
hsCRP
3	non-randomised studies	not serious	not serious	not serious	not serious	publication bias strongly suspecteda	282	244	-	SMD 0.34 higher (0.16 higher to 0.52 higher)	⨁◯◯◯ Very lowa
IFN- γ
2	non-randomised studies	not serious	not serious	not serious	not serious	publication bias strongly suspecteda	57	51	-	SMD 2.02 lower (6 lower to 1.97 higher)	⨁◯◯◯ Very lowa

CI: confidence interval; SMD: standardised mean difference. Explanations a. Publication bias was rated as suspected due to the small number of eligible studies

Discussion

Based on the hypothesis that FM is associated with alterations in inflammatory and neuroendocrine biomarkers consistent with a state of low-grade inflammation and physiological dysregulation, the findings of this review provide partial support for this premise.

Specifically, hsCRP levels were significantly elevated in patients with FM compared to healthy controls (SMD = 0.34; 95% CI: 0.16 to 0.52; p = 0.0002), with a consistent direction of effect despite and low heterogeneity (I² = 0%). This finding supports the role of hsCRP as a possible marker of low-grade inflammation in FM [16, 17, 20]; reinforces its interpretability relative to other biomarkers assessed in this review.

In contrast, IL-6 did not show a significant difference between groups (SMD = 2.07; 95% CI: -0.37 to 4.52; p = 0.10), and the very high heterogeneity observed (I² = 95%) further constrains the applicability of IL-6 as a standalone biomarker, and its elevation most likely reflects a broader low-grade inflammatory profile rather than a disease-specific signature. Consistent with this, the pooled estimate was sensitive to individual studies: exclusion of Polli et al. or Baris et al. markedly altered both the effect size and the heterogeneity, indicating that the result is driven in part by single studies and should be interpreted with caution.

In contrast, hair cortisol and IFN-γ did not show significant differences in the primary analysis, suggesting that neuroendocrine and immunological alterations are not consistently observed across all biomarkers.

These findings are consistent with previous literature. Studies included in this review evaluated classical inflammatory biomarkers such as IL-6, IL-8, and hsCRP [8, 17, 20, 31], generally reporting higher levels in patients with FM. This aligns with the systematic review by Kumbhare et al. [12], which included 54 studies and observed significantly higher levels of IL-6, IL-8, TNF-α, IFN-γ, CRP, and BDNF, along with lower levels of IL-1β in patients with FM. However, that review also cautioned that while these biomarkers reflect a pro-inflammatory low-grade state, they are not specific to FM, limiting their diagnostic utility. Similarly, the review by Sedda et al. [32] highlighted elevated levels of inflammatory biomarkers such as IL-6, TNF-α, and CRP in patients with FM, as well as their potential shared associations with depression and autoimmune diseases. The authors also identified serotonergic dysfunction as a potential mediator linking inflammation with clinical manifestations, including pain and mood disturbances. However, given that the underlying mechanisms of depression and autoimmune conditions and their influence on inflammatory processes and pain regulation in FM remain poorly understood, further research is warranted to elucidate the behaviour of these inflammatory pathways independently of such comorbid processes. This approach may provide more precise insights into their specific role in the pathophysiology of FM. Altogether, these findings reinforce the need to standardize methodologies and validate results in larger, multicenter studies.

In comparison to the systematic review by Favretti et al. [7], the obtained results confirm that there is still no single validated biomarker for the diagnosis of FM, in part due to the clinical and methodological heterogeneity of available studies. Nonetheless, there is value in exploring biomarkers to better understand the pathophysiological mechanisms of the disease. The accumulated evidence suggests that a panel of biomarkers, possibly adapted to patient subtypes, could be more useful than any single marker in isolation. Unlike the previous review, which questioned the direct relationship between biomarkers and clinical symptoms, several individual studies in this review reported associations between IL-6, cortisol, hsCRP, and serotonin and pain, fatigue, and quality of life—reported by primary studies rather than confirmed in the pooled quantitative synthesis, in which only IL-6 reached statistical significance—in patients with FM. This discrepancy could be explained by differences in methodological approach, study selection, and the sensitivity of the measurement tools used.

This perspective is relevant because it suggests the possible existence of subgroups of patients with different combinations of implicated mechanisms. Additionally, the review pointed to the usefulness of indirect biomarkers, such as deficiencies in vitamin D, vitamin B12, and zinc, and highlighted the potential of large databases and machine learning tools to identify diagnostic and prognostic biomarkers with greater accuracy.

In relation to the findings of this review, although IL-6 could appear to be associated with FM, neither hair cortisol nor interferon-gamma seem to provide meaningful or consistent information. In this context, future research should focus on specific components of the inflammatory cascade involving IL-6, independently of cortisol and interferon pathways [33]. Our findings cannot highlight the role of IL-6 as a consistent biomarker in patients with FM, contrary to data reported by Kumbhare et al. (2022) [12]. Although IL-6 is involved in other inflammatory and metabolic conditions, its significant elevation in this cohort underscores the existence of a neuroimmune profile that may be common in FM and its comorbidities, but it is still unclear. It is noteworthy that frequent conditions such as depression and obesity are also independently associated with this cytokine; far from constituting a mere confounding factor, this suggests that IL-6 may act as a biological link integrating the clinical and metabolic domains of these patients. Therefore, despite the limitations inherent to data scarcity and heterogeneity in its representation across the available studies, the current findings support the need to further investigate the role of IL-6 as a biological indicator of the inflammatory and neuroendocrine dysregulation characteristic of FM, opening the door to its potential utility in the clinical characterization of these patients. Rather than serving as standalone diagnostic tests, inflammatory biomarkers such as IL-6 may prove more useful for patient stratification, prognosis, or treatment monitoring, in line with current trends in personalised medicine. Furthermore, FM is increasingly recognised as a biologically heterogeneous condition, and interindividual differences in genetic and psychological susceptibility may contribute to the variability in symptom expression and biomarker profiles observed across studies [34]. Future research integrating inflammatory biomarkers with other domains—including pain modulation, psychological variables, sleep quality, and genetic factors—within multidimensional models may provide a more comprehensive understanding of the disorder than isolated biomarker approaches.

Importantly, the observational and predominantly cross-sectional nature of the included studies precludes causal inference; it therefore remains unclear whether the observed biomarker alterations contribute to the pathophysiology of FM or represent secondary consequences of the disease and its comorbidities, and all reported associations should be interpreted accordingly. Overall, although the evidence gathered is useful for understanding biomarkers associated with FM, methodological heterogeneity and risk of bias support the need for more homogeneous, multicentre, longitudinal studies with larger and more diverse populations in terms of sex and age.

Limitations

Several limitations should be acknowledged. The number of studies investigating biomarkers in FM remains limited (n = 18), which constrains the robustness and generalisability of the available evidence. In addition, existing studies show substantial methodological heterogeneity, particularly regarding data collection procedures, biomarker selection, and analytical approaches, which complicates direct comparison and synthesis of findings.

The marked sex imbalance in FM populations (approximately 7:1 female-to-male ratio) introduces further heterogeneity that may influence biomarker expression and reduces the interpretability of pooled estimates, particularly in meta-analytical frameworks.

Furthermore, two of the included studies were rated at high risk of bias using the Newcastle-Ottawa Scale, mainly due to the absence of matching or adjustment for confounding variables in their design, which may limit confidence in their individual contributions to the pooled estimates.

Collectively, these sources of variability hinder adequate control of potential confounding factors and limit the systematic assessment of associated comorbid conditions, as previously discussed, making it difficult to disentangle FM-specific alterations from those related to overlapping clinical entities such as chronic fatigue syndrome and other co-occurring pain or fatigue-related disorders. In addition, the meta-analyses pooled data obtained from different biological matrices (serum, plasma, saliva, and hair) and different assay methods, which probably contributed to the high heterogeneity observed. The small number of studies available for each quantitative synthesis precluded a formal assessment of publication bias and the calculation of prediction intervals.

Conclusion

The findings of this review suggest that immunological, neuroendocrine, and metabolic biomarkers could provide relevant information about the pathophysiology of FM, particularly in relation to low-grade inflammation and clinical severity with hsCRP emerging as a candidate biomarker of interest in this context, pending confirmation in larger, more homogeneous cohorts. In the quantitative synthesis, hsCRP was the only biomarker showing a statistically significant elevation in patients with FM, whereas the meta-analyses of IL-6, hair cortisol and IFN-γ did not reveal significant between-group differences. Associations of IL-6, cortisol, and serotonin with symptom burden were reported by individual studies but were not confirmed in the pooled analysis and should therefore be regarded as preliminary. However, the methodological heterogeneity of the included studies, along with limitations in controlling confounding variables and the moderate-to-high risk of bias identified across studies, currently prevents their use as diagnostic or therapeutic tools in clinical practice. Therefore, more robust studies with larger samples, longitudinal designs, and standardized measurement criteria are needed to validate these biomarkers and explore their usefulness in patient stratification and personalized treatment. Integrating these markers with clinical assessment could, in the future, contribute to a more accurate and earlier approach to FM.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization: FJFR; Methodology: CSSO, EAQR; Formal analysis and investigation: MDSR; Writing - original draft preparation: FJFR; Writing - review and editing: EDP, MDSR; Resources: CSSO, EAQR; Supervision: MDSR.

Funding

This article did not receive any form of funding.

Data availability

All data supporting the findings of this study are available within the paper and its Supplementary Information. If additional data are required, the authors may be contacted.

Declarations

Ethics approval and consent to participate

Not aplicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.