Abstract
Biomarkers are becoming pivotal in understanding the complex pathophysiology of disc herniation and guiding novel therapeutic strategies. Recent research highlights the value of molecular and cellular biomarkers in delineating disease progression, treatment monitoring and patient stratification. This review summarizes current advances in the identification and validation of emerging biomarkers across genomic, transcriptomic, proteomic, and metabolomic domains, emphasizing their potential to bridge basic mechanistic insights with clinical translation. Particular attention is given to the interplay between inflammatory mediators, extracellular matrix turnover, and immune cell activity as indicators of lumbar disc herniation diagnosis and prognosis. Despite encouraging progress, standardization of biomarker validation protocols, inter-study comparability, and large-scale clinical implementation remain major challenges. Future directions include the integration of multi-omics technologies and bioinformatic tools to identify predictive biomarker panels with diagnostic and prognostic utility towards personalized medicine.
-
Keywords: Disc herniation, Proteomics, miRNA, Personalized diagnostics
INTRODUCTION
Radiculopathy resulting from lumbar disc herniation (LDH) is an important and identifiable cause of low back pain (LBP). LDH can be described as a focal displacement of disc material that extends beyond the physiological margins of the intervertebral disc (IVD) space, giving rise to a series of symptoms including pain, motor weakness, or sensory disturbances such as numbness, typically with myotomic or dermatomic distribution. Indeed, the distinctive feature of LDH lies not simply in the structural alteration of the disc, but in its ability to produce a recognizable neuroanatomical pattern of symptoms, which provides a key diagnostic element and differentiates it from more common or nonspecific sources of LBP [
1,
2]. Diagnostic and prognostic biomarkers can provide important information about the link between structural herniation and symptoms.
Diagnosis and prognosis of LDH have traditionally relied on clinical imaging methodologies such as x-ray and magnetic resonance imaging (MRI). However, noninvasive biomarkers are emerging that may be used to provide mechanistic insight into the pathophysiology of the disease [
3], to predict the disease course and to aid in patient stratification towards individualized therapies [
4].
The importance of each category of circulating biomarkers is still unclear. However, as more information is being gathered and the molecular networks are building up, consistent diagnostic and therapeutic targets are emerging. The role of noninvasive biomarkers has been demonstrated in other diseases [
5], and their potential value for LDH patients is substantial, especially considering that LDH disease progression and response to therapy is highly unpredictable.
According to the Food and Drug Administration-National Institutes of Health biomarker working group, a biomarker is defined as ‘an indicator of normal biological processes, pathogenic processes, or biological responses to an exposure or intervention, including therapeutic interventions’ [
6]. In the context of LDH, biomarkers can be investigated to confirm the presence of LDH or sciatica (diagnostic biomarker), to monitor the response to treatment (monitoring biomarker), or to predict the likelihood of recurrence and progression (prognostic biomarker). The biomarkers of symptomatic LDH reflect its biological and pathological mechanisms. These biomarkers are usually associated with IVD inflammation, angiogenesis [
7], IVD matrix breakdown [
8], and lesions in the nervous system [
9]. Broadly, noninvasive circulating molecular biomarkers can be classified into protein, nucleic acids (mainly noncoding RNAs), lipid biomarkers, extracellular vesicles (EVs) and immune cells (
Table 1,
Fig. 1).
In this review, a comprehensive literature search was conducted across multiple scientific databases, including PubMed, Scopus, and Google Scholar, to ensure broad coverage of peer-reviewed biomedical and translational research. Articles published between 1994 and 2025 were considered, encompassing both fundamental studies and recent advances in the field. Key search terms included combinations of ‘lumbar disc herniation,’ ‘circulating biomarkers,’ ‘proteomics,’ ‘non-coding RNAs,’ ‘extracellular vesicles,’ ‘lipid biomarkers,’ and ‘immune cells.’
PROTEIN BIOMARKERS
1. Protein Biomarkers for LDH Diagnosis and Differentiation From Other LBP Sources
The purpose of a diagnostic biomarker for LDH is to provide a clear indication of the occurrence of a disc herniation as compared to other conditions leading to LBP. In several studies, protein biomarker panels were measured in the peripheral blood of LBP patients undergoing spine surgery. One study that included 77 patients found a correlation between the presence of disc herniation and elevated plasma levels of fibroblast growth factor-2. This may be related to an angiogenic or tissue repair response. On the other hand, negative correlations were observed between disc herniation and inflammatory markers such as C-reactive protein (CRP), serum amyloid A or soluble intercellular adhesion molecule 1 [
10]. This may indicate a negative feedback response or potential treatment with anti-inflammatory medication. Since that study analyzed data from a heterogeneous group of patients undergoing spine surgery, confounding factors such as multiple disorders may also play a role.
With emphasis on serum cytokine levels, symptomatic LDH and MRI based herniation severity were analyzed in a group of 78 patients, requiring surgery for treatment of LDH, compared to matched control subjects [
7]. Multiple cytokines, chemokines and angiogenic factors were elevated in LDH compared to the control group. Furthermore, age, body mass index (BMI) and sex were found to be significant predictors of various markers, highlighting the importance of confounding parameters. Herniation severity, but not Pfirrmann grade, was associated with altered levels of insulin-like growth factor (IGF), platelet derived growth factor BB and interleukin (IL)-9. Interestingly, these associations were observed in patients with chronic symptoms but not after acute episodes. These findings provide valuable insight into the regulation of circulating inflammatory markers after LDH, although their potential use as diagnostic markers was not addressed. C-C motif chemokine ligand 5 (CCL5), one of the markers found up-regulated in LDH patients was also evaluated in another study comparing subjects with disc degeneration on MRI with a matched control group without degeneration [
11]. Here the suitability of CCL5 as a diagnostic marker was evaluated based on receiver operating characteristic analysis, adding more information about its diagnostic potential.
A prospective study including 50 patients with radiological and clinical manifestation of LDH found significantly elevated levels of endothelin-1 (ET1) in the serum of LDH patients compared to a control group of outpatients with no history of lumbar pain [
12]. Serum ET1 levels correlated with Pfirrmann grades but not with visual analogue scale (VAS) or Oswestry Disability Index scores in the patient group. This indicates that the severity of IVD degeneration may be determined by measuring the ET1 levels, although the validity of this analyte as an effective biomarker needs verification.
Brisby and colleagues [
13] measured the levels of inflammatory cytokines in the serum and cerebrospinal fluid (CSF) of 39 LDH patients and compared them to normal values used for clinical routine investigations. Interestingly, the serum cytokine levels of the LDH patients were within the normal values. There was however an increase in IL-8 concentrations in the CSF of patients with short duration of sciatic pain symptoms, indicating an early inflammatory response, and after disc extrusion or sequestration. This hints towards a mechanical effect of the nerve root compression or a biochemical response of the surrounding to the herniated tissue. The absence of increased cytokine levels in the circulation of patients with a more chronic condition or at later stages of the disease progression is frequently encountered and can be related to the restriction of cytokine release to peak levels in the acute inflammatory state at onset of sciatica in the case of LDH. Hence, circulating levels of standard inflammatory cytokines may not be suitable biomarkers for nonacute cases.
A similar outcome was reported from a study aiming to differentiate sciatica with and without MRI findings of nerve compression [
14]. An array of cytokines, chemokines and matrix metalloproteinases (MMPs) was analysed in the serum of patients with sciatica. No difference was found in any marker tested in patients with (n=93) or without (n=26) clinically confirmed sciatica, nor between those with (n=44) or without (n=49) sciatica and MRI confirmed nerve root compression. It was concluded that larger longitudinal studies would be required to determine whether serum biomarker levels are associated with persistence of sciatica symptoms.
A different, nontargeted approach, namely proteome profiling, was applied to identify markers that may distinguish discogenic pain and painful LDH from unknown origin pain sources [
15]. Different sets of serum samples, 30 per group, were used for training and verification, respectively, aiming to generate an accurate model. Two proteins, complement C3 and fibrinogen, were identified as potential biomarkers to distinguish between different causes of chronic LBP, although further validation is needed. The advantage of proteome profiling compared to single marker or marker panel analysis is the unbiased approach while screening the entire proteome. On the other hand, several validation steps are needed for effective translation into clinical practice.
In another unbiased approach, transcriptomic and metabolomic profiles of peripheral blood mononuclear cells (PBMCs) were compared among LDH patients with (1) mild MRI findings and severe symptoms, (2) severe MRI findings and mild symptoms, and (3) a control group [
16]. Strikingly, 38 metabolites were different between the 2 LDH patient groups. Moreover, 2 proteins detected in the serum were increased in patients with severe symptoms that had mild findings on MRI. The model created based on the combined profiles demonstrated excellent predictability with a mean area under the curve (AUC) of 0.875. This model has the potential to facilitate early diagnosis and personalized treatment strategies of LDH, filling the gap between MRI findings and clinical symptoms.
Levels of cytokines and chemokines were measured in CSF and serum in 40 patients with disc degeneration and 40 LDH patients, and outcomes associated with neuroinflammation and pain [
9]. Serum monocyte chemoattractant protein (MCP)-1 concentrations correlated with higher global pain ratings and increased spinal pressure pain sensitivity. Also, IL-6 serum levels correlated with the intensity of the neuropathic pain component with leg pain in LDH patients. The findings of that study are exploratory but warrant further investigation of circulating biomarker regulation in neuroinflammation and pain episodes in LDH.
Plasma biomarkers indicating recovery 2 months after surgery were analyzed in patients with different spinal pathologies [
10]. The investigators found that patient age combined with plasma levels of CRP and CCL22 had a strong predictive value for efficient recovery, with an AUC of 0.883 based on the ROC curve. Accordingly, younger patients with high CCL22 and low CRP plasma levels would have the highest probability of total recovery. Nonetheless, patients with different spinal disorders were assessed and the performance of the biomarker combinations for patients with LDH was not reported separately.
An inflammatory protein profile was generated from the serum of patients with sciatica to identify proteins characteristic of patients with persistent lumbar radicular pain after LDH [
17]. Patients were classified into low and high (VAS>6) pain groups at 12 months after LDH. The investigators identified 41 proteins with false discovery rate (FDR) <0.10 and 13 proteins with FDR <0.05, which were up-regulated in the patients with severe pain one year after LDH. There was increased expression of many inflammation-related proteins in the high pain group compared to the low pain group. On the top of the list, ranked by estimated increase, C-X-C motif chemokine ligand 5, epidermal growth factor and MCP-4 were found. For each patient, an inflammation score, i.e., weighted average of 41 protein levels, was also constructed that clearly separated the 2 patient groups. This study demonstrates the feasibility of detecting serum protein differences between patients with persistent versus nonpersistent pain.
A similar study [
18] found significantly higher IL-6 and IL-8 concentrations in the serum of patients with VAS>3 compared to patients with VAS<3 at 12-month follow-up. These observations suggest that chronic lumbar radicular pain may be associated with a persistent increase of proinflammatory factors in serum after LDH. Interestingly, high baseline levels of serum IL-6 were also associated with less favorable recovery in patients with lumbar radicular pain [
19]. The prognostic value of these inflammatory proteins, or of the proposed inflammation score, remains to be investigated.
Pre- and postoperative cutoff values of IL-6 serum levels (4.36 and 1.16 pg/mL, respectively) were determined to predict postoperative pain relieve and disability improvement [
20]. The cutoff values showed about 66% sensitivity and high specificity of IL-6 before and after surgery, indicating its value as a predictor for prognosis of outcome, though the study included a limited number of 32 patients and further studies are required for validation.
Serum levels of S100 calcium binding protein B (S100B) and brain-derived neurotrophic factor (BDNF) were measured as outcome predictors for the effect of radiofrequency treatment in LDH patients [
21]. The aim was to correlate the serum levels with outcomes at 6 months after radiofrequency. The results indicate that serum levels of S100B could predict functional outcomes after pulsed radiofrequency in patients with lumbar disc prolapse. A low baseline serum level of S100B could predict postinterventional pain and functional improvement, ultimately assisting in identifying suitable candidates for pulsed radiofrequency.
The turnover of extracellular matrix (ECM) molecules can be monitored as indication of tissue regeneration. Procollagen type I C-peptide (PICP) and C-terminal telopeptide of type I collagen (CTx) indicate the synthesis and degradation of collagen type I, the major component of the annulus fibrosus (AF). These markers were monitored in the serum of 67 patients undergoing discectomy due to LDH [
22]. While the ratio of PICP to CTx increased 6 weeks following surgery, indicating enhanced synthesis and decreased degradation of collagen, there was no significant correlation with the outcome in terms of pain and disability improvement. This indicates that these matrix molecules can reliably be measured to monitor the metabolism of collagenous tissue but may not be suitable predictors of a favorable outcome in patients undergoing surgery for LDH.
Levels and activities of MMP2, MMP9 and tissue inhibitors of metalloproteinases (TIMP1, TIMP2) were measured in the serum of 70 LDH patients, before surgery and at 1 and 3 months after surgery [
8]. Correlations between levels of TIMP2 and leg pain or pain intensity were found at 3-month postsurgery. Nevertheless, the validity of these markers as monitoring tools or therapeutic targets was not confirmed.
Proinflammatory cytokine levels in patients’ plasma were measured to compare the influence of open discectomy compared to computerized tomography navigation percutaneous spinal endoscopy [
23]. Not surprisingly, plasma concentrations of IL-6, tumor necrosis factor (TNF), CRP and creatine phosphokinase were higher in the open discectomy group, confirming the minimal invasiveness of the percutaneous approach. Cytokine quantification can therefore be a valuable method to assess and monitor the invasiveness of a procedure in terms of inflammatory response.
A study with 262 patients with lumbar radiculopathy aimed to evaluate the plasma concentrations of IL4 and TNF at 1 month and 12 months after microdiscectomy [
24]. Cytokine levels were compared between patients with VAS<3 and patients with VAS≥3. In both groups, TNF decreased over time after surgery, whereby the higher VAS group had increased levels at baseline. In contrast, levels of the anti-inflammatory IL-4 were lower in the higher VAS group at baseline and were slightly increasing over time. This indicates an association between pain, pro- and anti-inflammatory markers, though their validity as monitoring markers needs verification.
NONCODING RNA BIOMARKERS
A variety of noncoding RNAs (ncRNAs) are differentially expressed in degenerative disc disease (DDD), including miRNAs (reviewed in [
25]), lncRNAs (reviewed in [
26]) and circRNAs (reviewed in [
27]). They have been shown to be involved in multiple pathological processes during DDD, including apoptosis, ECM turnover, cell proliferation and inflammatory response. The identification of circulating ncRNAs remains limited, with only a few studies reporting correlations with LDH.
Hasvik et al. [
28] studied a cohort of 97 patients with LDH and leg pain on MRI and the association between circulating miR- 17 and leg pain intensity. They found an association between the level of miR-17 in serum and the intensity of lumbar radicular pain. Furthermore, a rat model was used to examine possible changes in miR-17 expression in the nucleus pulposus (NP) associated with a leak of NP tissue out of the herniated disc, and the functional role of miR-17 was addressed by transfection of miR-17 into the human monocytic THP-1 (THP-1) cell line. They found up-regulation of miR-17 in the rat NP tissue when applied onto spinal nerve roots and increased TNF release following transfection of miR-17 into THP-1 cells. This study suggests that miR-17 may be involved in the pathophysiology underlying lumbar radicular pain after LDH.
Zhang et al. [
29] analyzed miR-29a expression in the plasma of 3 different study groups: healthy subjects, patients affected by LDH and lumbar spinal stenosis (LSS). Plasma expression levels of MMP9 and a disintegrin and metalloproteinase with thrombospondin motifs 5 (ADAMTS5) were also measured by enzyme-linked immunosorbent assay (ELISA). The expression levels of miR-29a in plasma and IVD tissue were significantly lower in patients with LSS compared to patients with LDH, as well as healthy controls. Conversely, the protein expression levels of MMP9 and ADAMTS5 were significantly higher in patients with LSS compared to patients with LDH, as well as healthy controls. miR-29a has therefore potential to be studied as a potential biomarker of LSS.
Cui et al. [
30] performed RNA sequencing on peripheral blood serum from patients with LDH, identifying 73 differentially expressed miRNAs between LDH patients and patients that had other spinal diseases. Many of these miRNAs were linked to ECM regulation, inflammation, and apoptosis and involved in diverse signaling pathways. The profile of miR-766-3p, miR-6749-3p, and miR-4632-5p was significantly enriched in multiple pathways associated with IVD degeneration. Although further studies are needed to confirm the functional role of the identified miRNAs, this panel has potential as non invasive biomarkers of LDH.
Recently, in a rat model of IVD herniation, we have identified plasma levels of miR-143-3p, miR-10b-5p, miR-27a-3p, miR-140-5p, miR-155-5p, miR-146a-5p, and miR-21-5p as positively correlated with hernia size. Moreover, the protein tenascin C-miR-155-5p protein-miRNA pair was highlighted as promising candidate to be part of a putative regulatory module worth investigating as a prognostic tool of IVD hernia regression [
31]. Indeed, the combined analyses of RNAs and proteins will certainly be necessary in such a complex and multifactorial disease as LDH.
Other studies have specifically analyzed circulating RNAs and demonstrated that they were not correlated with LDH. Zou et al. [
32] analyzed the association of the local and plasma expression of miR-21 with disease severity of LDH patients with sciatic pain. 92 LDH patients with sciatic pain and 25 scoliosis patients as painless controls were enrolled in the study. Local miR- 21 expression was found to be linked with clinical severity of LDH. However, no significant difference in plasma miR-21 expression was found. Wei et al. [
33] studied the expression of circRNA GRB10 in the plasma from patients with lumbar DDD, sacroiliac joint pain (SJP), LDH, piriformis syndrome (PS), entrapment neuropathy (EN), as well as from healthy controls (n=60 per group). Decreased expression of GRB10 RNA was only observed in the DDD group, but not in SJP, LDH, PS, and EN groups, compared to controls.
LIPID BIOMARKERS
Besides proteins and nucleic acids, metabolites are gaining increasing attention as circulating biomarkers. For example, Deng et al. [
16] studied metabolomic profiles in the serum of patients with LDH. They found 38 metabolites to differ between patients with mild MRI but severe symptoms and those with severe MRI but mild symptoms, whereas only 13 metabolites were different between the severe MRI group and healthy controls. Hider et al. [
14] did not find biomarker levels to be different when comparing sciatica with and without MRI findings of nerve compression. These studies imply a decoupling between MRI signs of herniation and symptoms. This agrees with an earlier study showing that a significant proportion of asymptomatic subjects have IVD bulges and protrusions [
34]. These findings highlight that structural displacement of the IVD is not the only factor that causes back and leg pain.
One hypothesis is that LDH is caused by an insufficient blood supply, like ischemic heart disease. Blood lipid biomarkers, usually used to predict ischemic heart disease, have also been evaluated in LDH patients [
35-
37]. Nonetheless, the correlation between LDH and lipid biomarker level is rather weak. In the study by Tan et al. [
36], triglyceride (TG) level was positively correlated with the risk of LDH, although the TG level in the LDH group was only marginally higher compared to the control group (1.78±1.17 mmol/L and 1.38±0.88 mmol/L, respectively). In the study by Zhang et al. [
37], the ratio of total cholesterol/high-density lipoprotein cholesterol (TC/HDL-C) and total cholesterol/high-density lipoprotein cholesterol (LDL-C)/HDL-C were higher in the LDH group than the control in terms of statistical significance, but the level difference was small (3.71 vs. 3.43 for TC/HDL-C, 2.3 vs. 2.06 for LDL-C/HDL-C). Hence, there is still a lack of evidence demonstrating that lipid biomarkers can be used to predict LDH similarly to ischemic heart disease prediction.
EXTRACELLULAR VESICLES
EVs are nanoscale, membrane-bound structures secreted by all cell types, including exosomes, microvesicles, and apoptotic bodies, each defined by distinct biogenesis pathways [
38]. Their cargo (proteins, lipids, DNA, mRNAs, and microRNAs) reflects the physiological or pathological state of the parent cell and mediates intercellular communication [
39]. EVs are abundant in blood, urine, saliva, CSF, breast milk, and synovial fluid, and their stability and ability to cross biological barriers make them ideal “liquid biopsy” tools [
40]. While widely studied in oncology, cardiovascular, neurological, and autoimmune diseases, their application in musculoskeletal disorders remains limited, despite the importance of early diagnosis to enable regenerative interventions. Blood-derived EVs offer minimally invasive biomarkers for spinal conditions such as LDH, DDD, and chronic LBP, with miRNAs like miR-223, miR-146a and proteomic signatures correlating with disease state and therapeutic response [
41-
44]. These findings highlight the translational potential of circulating EVs as accessible biomarkers for diagnosis, monitoring, and treatment evaluation in musculoskeletal disease, though challenges in isolation and characterization persist. Advances in proteomics and immunoaffinity enrichment now allow more precise profiling, supporting the role of EVs as both biomarkers and active effectors in precision musculoskeletal medicine [
45-
48].
IMMUNE CELLS
To distinguish between LDH patients with and without IVD rupture, percentages of Th17 lymphocytes were analyzed in patients’ PBMCs, whereas concentrations of IL17 and prostaglandin E2 (PGE2) were measured in their peripheral blood plasma [
49]. Both the percentage of Th17 cells and the concentration of IL17 in the blood were highest in the LDH patients with IVD rupture, followed by the LDH patients without IVD rupture, and the healthy control group. This outcome corroborates that rupture of the AF and herniation of the NP initiate an autoimmune response. The Th17 cell proportion and serum levels also correlated with the VAS score and with PGE2 levels, indicating their association with inflammation and pain. However, the suitability of IL17 as biomarker to identify ruptured IVD requires further investigation.
The autoimmune response after LDH was also targeted in another study [
50], where investigators found that LDH patients had higher proportions of cells with autoimmune markers within PBMCs than the control group of patients without LDH. Interestingly, LDH patients had lower levels of “rescue” CD4+ CD25+ T cells than the control group. Finally, serum concentrations of activation-inducible tumor necrosis factor receptor (AITR) and AITR ligand, important costimulatory molecules in the pathogenesis of autoimmune diseases, along with interleukins (IL-2, IL-6, IL-8, IL-1b) and TNF, were all enhanced in the patients compared to the control group. While these data provide important insight in the pathological mechanisms of LDH, their suitability as diagnostic biomarkers was not assessed.
GENERAL CONSIDERATIONS
1. Sample Processing and Biomarker Quantification
Existing studies often discover different sets of biomarkers with limited consistency between different reports. Besides the different patient cohorts, the disparity in sample harvesting and processing protocols is another potential reason for the inconsistency. Part of the studies have used peripheral blood serum for biomarker discovery [
7,
9,
12,
15,
22,
51,
52], Serum is obtained by triggering the coagulation of blood in commercially available tubes, followed by centrifugation to remove blood cells and fibrin clots. Other studies have used plasma which is obtained by adding anticoagulant to the blood and a following centrifugation step to remove blood cells [
10,
11,
23,
53]. Several studies have placed the blood on ice for some time and collected the supernatant after centrifugation, where the coagulation may or may not be completed [
17-
19,
51]. Due to the different processing method, it is evident that serum and plasma have different metabolomic and lipoprotein profiles [
54]. The choice of anticoagulant also influences plasma measurements [
55]. It is also frequently overlooked whether blood cells are completely removed from the serum or plasma. Peripheral blood cells express and secrete proteins that are associated with symptomatic LDH [
49,
50]. It is therefore important to standardize the blood processing protocols to improve the consistency of biomarker studies.
In many investigations, specific multiplex panels or ELISA kits are chosen for protein measurement based on the hypothesis of the respective study. However, the set of proteins evaluated using these kits may not be comprehensive, whereby important biomarkers associated with unknown LDH mechanisms may be missed. Improved understanding about the onset of LDH may help to generate new hypotheses for clinical biomarker discovery. Alternatively, more comprehensive protein measurements such as proteomic profiling are emerging in this field. For example, based on proteomic profiling, plasma alpha-2-macroglobulin, coagulation factor XIII B chain (F13B), MMP2 and IGF1 were found to be correlated with histological grades of IVD degeneration [
53]. In proteomics, the protein quantification is usually inferred from peptide peaks measured by mass spectrometry. Instead of proteins, Zhang et al. [
15] used peptide quantification in serum to differentiate between discogenic LBP and LDH. They identified 23 differential peptide peaks and a classification model based on these peaks showed a sensitivity of 93.3% and a specificity of 80% in differentiating DDD and LDH.
Another interesting aspect is the modulation of biomarker expression by circadian rhythms. Numerous cytokines, chemokines, and ncRNAs exhibit circadian oscillations, with peaks in inflammatory mediators such as IL-6 and TNF coinciding with the worsening of pain and stiffness in the early morning, often reported by patients. Circadian disruption, whether due to sleep disturbances, shift work, or systemic inflammation, can exacerbate degenerative cascades by altering the release of MMPs, TIMPs, or stress-related metabolites into the circulation. Consequently, accounting for time of day in sample collection and adopting chronobiology-based study designs is critical to improving the reproducibility and predictive value of biomarker research in LDH [
56].
To discover diagnostic biomarkers of LDH, diverse control groups have been defined. They include healthy volunteers [
7,
49], spine fracture patients [
50], outpatient clinical cases with symptoms but without LDH [
12], DDD cases without LDH [
9,
11] or subjects from other irrelevant studies [
13]. Depending on the group setting, the discovered biomarker may be used to diagnose the herniation itself [
49,
53] or herniation associated sciatica [
9]. The main goal of discovering novel biomarkers is to improve diagnostic accuracy compared to an existing “gold standard.” [
57] Currently, MRI is the mostly used diagnostic tool for LDH. There is a lack of studies that compare the accuracy of blood biomarkers with this “gold standard.” thus to date there is insufficient rationale to introduce these discovered biomarkers in clinical practice.
For prognostic biomarkers, an ideal high-quality study is a prospective follow-up, and the patients of the entire cohort need to be enrolled at the same stage of their disease. The serum obtained at the start of the follow-up is measured and is then correlated with the outcome of the follow-up, e.g. chronic pain [
17,
18], recurrent sciatica [
58] or functional recovery [
19-
21]. Prognostic studies investigating the ‘recurrence of herniation postsurgery’ [
58] become challenging because the incidence of herniation recurrence is rare [
59]. Even if a large surgery cohort is followed-up, only few cases will have the outcome of herniation recurrence. In this case, the retrospective case control design [
60] is a compromising solution. A cohort is chosen based on the outcome of postsurgery recurrence and the matched control group can be constructed based on the patients’ characteristics. Case control design is ranked as level III evidence because of the potential recall bias [
57]. Nevertheless, while these studies are labelled as ‘low level of evidence,’ their value in addressing unmet needs in predicting postsurgery recurrence is still important.
Heterogeneity across studies regarding patient demographics, symptom duration, disease stage and outcome measures represents a major challenge in biomarker research, limiting both generalizability and reproducibility. In particular, differences in age, sex, BMI, comorbidities, and treatment history may confound biomarker levels and contribute to inconsistent results across cohorts. Similarly, variability in symptom duration and disease chronicity may influence inflammatory and metabolic markers, making direct comparison between acute and chronic LDH populations difficult. Outcome measures also differ widely across studies, ranging from pain intensity scores and disability indices to imaging-based assessments, further complicating interpretation across studies. These sources of heterogeneity are particularly relevant when comparing small-scale exploratory studies, often designed to identify candidate biomarkers, with larger clinical cohorts aimed at validation or prognostic assessment. Although exploratory studies are valuable for generating hypotheses, their results must be interpreted with caution and require confirmation in well-characterized, adequately powered cohorts with standardized protocols.
Finally, it is fundamental to validate the accuracy of potential biomarkers. The ROC curve is a useful tool to quantify biomarker accuracy. The AUC takes consideration of the trade-off between sensitivity and specificity [
10,
20]. The accuracy can be improved by integrating multiple biomarkers and other clinical profiles into a prediction model [
10]. It is interesting to note that when combined with age, the prognostic biomarkers CRP and CCL22 give the highest prediction of pain score recovery after surgery [
61]. Additionally, serum biomarker level can be confounded by age, sex, and BMI. These factors must be corrected for when evaluating the correlation between biomarker level and clinical outcome [
7].
Usually, the learning of parameters in these multivariant models and the validation are performed with the same dataset. These models may have a good ROC score in the learning data, but can still fail in predicting future data, namely in the situation of overfitting [
62]. Ideally the data used for biomarker discovery should be different from the data used for testing biomarker accuracy. Wang et al. [
61] discovered NAD kinase 2 (NADK2) in blood as a potential diagnostic biomarker based on existing transcriptomic microarray data. They used an external cohort of 200 clinical cases which was independent from the dataset used for biomarker discovery to validate the accuracy of NADK2’s diagnostic potential.
The effective integration of biomarkers into clinical practice will be dictated by their feasibility, cost-effectiveness, and their ability to significantly influence clinical decision-making. From a practical standpoint, several circulating biomarkers can be measured using standardized and relatively cost-effective platforms, such as ELISA or multiplex immunoassays, which are already widely available in clinical laboratories. However, more advanced multiomic approaches currently remain resource-intensive and require specialized infrastructure and expertise. It is important to emphasize that the clinical value of these biomarkers will ultimately depend on their ability to modify patient management beyond existing standards of care. Potential applications include identifying patients at risk of persistent pain, guiding treatment selection, monitoring therapeutic response, or supporting decisions regarding conservative versus surgical intervention.
In this context, blood biomarkers and multiomic approaches offer complementary information to the spatially informative imaging tools, by capturing dynamic biological processes, such as inflammation, immune activation, ECM turnover, and metabolic stress, which are not directly accessible through imaging. While MRI remains indispensable for anatomical assessment and localization of disc pathology, numerous studies have demonstrated a limited correlation between structural imaging findings and symptom severity, disease progression, or treatment response. Indeed, disc herniations and degenerative changes are often observed in asymptomatic individuals, underscoring the limited specificity of imaging alone. Recent studies have shown that molecular signatures derived from proteomic, metabolomic, and transcriptomic analyses can stratify patients, predict symptom persistence, and identify endotypes that cannot be detected by imaging alone. The strength of multi-omic approaches lies not in replacing MRI [
16], but in complementing informative imaging with biologically meaningful data. Integrating molecular biomarkers with imaging and clinical parameters may enable more refined patient stratification, improved prognostic accuracy, and better identification of patients likely to benefit from specific interventions. This integrative perspective aligns with emerging concepts of molecular endotypes [
4] and personalized medicine in musculoskeletal disorders.
CONCLUSIONS
The landscape of the current biomarker discovery-phase evidence in LDH and DDD is rapidly evolving, with circulating markers offering promising possibilities for improving diagnosis, prognosis, and therapeutic monitoring. ncRNAs, including miRNAs, lncRNAs, and circRNAs, are emerging as regulators of ECM remodeling, inflammation, apoptosis, and immune activation. Although their specific contribution to LDH remains incompletely understood, accumulated data suggest that coherent ncRNA-protein interaction networks are being established, paving the way for targeted interventions. Beyond their mechanistic role, circulating ncRNA signatures show remarkable prognostic potential in a disease characterized by unpredictable progression and variable therapeutic response. In addition to these regulatory molecules, protein serum and plasma biomarkers provide a complementary dimension. Proinflammatory cytokines and chemokines such as IL-6, IL-8, TNF, and MCP-1, as well as vascular and neuroinflammatory markers such as ET-1, S100B, and BDNF, have been associated with pain severity, disc degeneration, and postoperative recovery. Similarly, proteins reflecting matrix turnover, including MMP2, MMP9, and their inhibitors TIMP1, TIMP2, along with systemic markers such as CRP, highlight the inflammatory and catabolic processes occurring within the disc environment. Lipid markers and metabolomic signatures add further insight into systemic metabolic contributions to the pathophysiology of LDH. Although many of these serum and plasma biomarkers are promising, inconsistencies between studies underscore the need for methodological standardization and validation in large prospective cohorts. While diagnostic biomarkers require robust discrimination between disease and control states, prognostic and monitoring biomarkers rely on longitudinal associations with clinical outcomes and treatment response over time. Biomarker data must be interpreted within a rigorous translational framework that prioritizes robustness, reproducibility, and clinical relevance. A more robust and personalized biomarker toolkit will be achieved by the integration of serum and plasma signatures of ncRNA, protein and different metabolites. In this scenario, stratifying patients based on molecular, clinical, and imaging biomarkers will be essential for personalizing interventions and improving therapeutic outcomes. Identifying specific subgroups of patients based on their biomarker profiles will enable more precise targeting of therapies, ensuring that interventions are not only aligned with patients’ clinical needs but also adaptable to their changing conditions over time, allowing for dynamic adjustments as the disease progresses or responds to treatment. Incorporation of imaging and clinical biomarkers in combined multidimensional models [
58] have the potential to develop risk prediction tools that should facilitate the development of more effective and personalized treatments for LDH.
NOTES
-
Conflict of Interest
The authors have nothing to disclose.
-
Funding/Support
CC acknowledges funding from FCT -
Fundação para a Ciência e a Tecnologia (10.54499/CEECIND/00184/2017/CP1392/CT0001) and FEDER - European Regional Development Fund (COMPETE2030-FEDER-00691600-15806, NORTE2030-FEDER-01801000). SG and JM acknowledge funding from AO Foundation and AO Spine. GV and VT acknowledge funding from NextGenerationEU (NRRP M6C2-PNRR-MAD-2022-12376692_VADALA) and PNRR-MCNT2-2023-12378359.
-
Author Contribution
Writing – original draft: CC, JM, VT, IHH, GV, SG; Writing – review & editing: CC, JM, VT, IHH, GV, SG.
Fig. 1.Circulating biomarkers of lumbar disc herniation can be identified in the serum and plasma collected from the patients’ peripheral blood and classified as proteins (e.g., cytokines, chemokines), noncoding RNAs, metabolites and other small molecules, extracellular vesicles (EVs) and immune cells.
Table 1.Circulating molecular biomarkers in lumbar disc herniation, describing the aim of the biomarker study, the detection method and key findings
Table 1.
|
Study |
Aim of biomarker |
Detection method |
Key findings |
|
Coquelet et al. [10] 2025 |
Diagnostic biomarker of LDH, myelopathy and sciatica |
Multiplex assays and single molecule array assays |
Herniated discs correlated with elevated levels of FGF-2 and decreased levels of NfL, CRP, SAA, and s-ICAM-1. Myelopathy (spinal cord damage) was positively associated with the plasma levels of NfL, CRP, and TNFα. Nerve involvement showed a positive correlation with IL-18 and CXCL10. |
|
Multicomponent and prognostic biomarker for recovery |
Higher presurgery CRP, SAA or NfL were less likely to have a good recovery postsurgery. |
|
The best-performing model was obtained when age, CRP, and CCL22 were combined; the AUC reached 0.883. |
|
Jacobsen et al. [7] 2020 |
Diagnostic biomarker for LDH |
Multiplex immune-bead assay |
CCL2/MCP-1, CCL4, CCL5/RANTES, FGF basic, G-CSF, GM-CSF, HMGB1, IFN-g, IL-1b, IL-1ra, IL-2, IL-5, IL-6, IL-7, IL-9, IL-12(p70), IL-15, IL-17, MMP-9, PDGF-BB, and VEGF were higher in LDH compared to control. CCL3, CCL11, CXCL10, IL-4, IL-8, IL-10, IL-13, MMP-1, MMP-3, and TNF-α did not exhibit a significant contribution of LDH on levels in the multivariate regression. Herniation severity had a significant effect on serum levels of HMGB1, IL-9, and PDGF-BB. MMP-1 levels were inversely correlated with VAS score. |
|
Monitoring biomarker postsurgery |
MIF significantly decreased at 3-mo postsurgery compared to presurgery. CCL11, CCL3, CXCL1, and CXCL10 all significantly increased. |
|
Grad et al. [11] 2016 |
Diagnostic biomarker for disc displacement |
ELISA |
A suggestive trend towards significance was noted with elevated levels of CCL5 in subjects with disc displacement (mean: 23.4 vs. 14.0 ng/mL; p = 0.073). |
|
No significant difference was found in CXCL6 between with or without disc displacement. |
|
Yıldırım Uslu et al. [12] 2024 |
Diagnostic biomarker for LDH |
ELISA |
ET-1 was 123.829 ± 48.909 ng/L in the LDH group and 73.761 ± 47.554 ng/L in the healthy control group (p < 0.01). There was a positive correlation between Pfirrmann grades and ET-1 (p < 0.01), while no correlations were determined between ET-1 and VAS, ODI, and MacNab grades (p = 0.137, p = 0.218, and p = 0.397, respectively). |
|
Dube et al. [53] 2025 |
Diagnostic biomarker for severity of IVD degeneration |
Proteomic profiling |
Plasma levels of A2M, F13B, MMP2, and IGF1 were correlated with histological grades of IVD degeneration. ROC curve analyses revealed that A2M has the highest AUC score of 0.79, demonstrating its high accuracy in distinguishing between mild and severe IVD degeneration. |
|
Zhang et al. [15] 2014 |
Diagnostic biomarker to differentiate discogenic low back pain (DLBP), chronic low back pain of unknown origin (CLBP) and LDH and compare them with normal control (N) |
Matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF-MS) |
Paired comparison analyses showed that the classification model for DLBP vs. CLBP had the lowest accuracy, with a recognition capability of 82.66%, a forecasting ability of 67.24%, and a sensitivity of 70% in the blind test. The classification models for DLBP vs. LDH and LDH vs. N were more accurate, with forecasting abilities of nearly 90% and sensitivities greater than 90% in the blind test. |
|
They next combined the DLBP and LDH groups as a disc degeneration group (DLBP + LDH) and compared it with group N using the above-mentioned algorithms. This resulted in a recognition capability of 95.82%, a forecasting ability of 73.28%, and a sensitivity as high as 95% in the blind test. |
|
Cheng et al. [49] 2013 |
Diagnostic biomarker for IVD rupture |
ELISA |
The level of IL-17 in group P (non-ruptured disc) (5.06 ± 3.78 pg/mL, p = 0.003), and in group E (ruptured disc) (12.13 ± 3.62 pg/mL, p < 0.001) were larger than healthy controls (1.31 ± 0.77 pg/mL), VAS score was positively correlated with IL-17 concentration (r = 0.458, p = 0.007). PGE2 was positively correlated with IL-17 concentration (r = 0.500, p = 0.003). |
|
Park et al. [50] 2007 |
Diagnostic marker for symptomatic LDH |
ELISA |
Symptomatic LDH PBMC showed increased expression ratio of AITR (resting state 10.36 ± 5.51% vs. 2.18 ± 1.05%, p < 0.05, stimulated state 27.79 ± 8.54% vs. 10.30 ± 3.56%, p < 0.05) and AITRL (resting state 23.15 ± 3.78% vs. 17.02 ± 4.57%, p < 0.05, stimulated state 45.00 ± 10.75% vs. 24.96 ± 4.75%, p < 0.05) than spinal fracture sample control. Mean serum concentrations of soluble AITR, AITRL, IL-2, IL-6, IL-8, and TNF-α were significantly higher in the herniation patients. |
|
Palada et al. [9] 2019 |
Diagnostic biomarker for neuropathic pain |
Tenplex immunoassay kits and a uniplex |
The MCP-1 concentration in serum was associated with higher global pain ratings and increased spinal pressure pain sensitivity. IL-6 serum concentration correlated with the intensity of the neuropathic pain component (leg pain) in LDH patients. |
|
Wang et al. [61] 2024 |
Diagnostic biomarker for sciatica |
Quantitative real-time PCR (RT-PCR) |
NADK2 emerged as the sole significant marker for sciatica. |
|
Moen et al. [17] 2016 |
Prognostic biomarker for chronic pain |
Multiplex proximity extension assay technology |
CCL7, CSF-1, VEGF-A, CXCL10, MCP-2, CXCL5, CCL-4, IL-15-R-alpha, MCP-4, TGF-beta-1, CASP-8, EGF, STAMPB were significantly increased. The highest increase: CXCL5; 217% increase, EGF; 142% increase, and MCP-4; 70% increase. |
|
Pedersen et al. [18] 2015 |
Prognostic biomarker for chronic pain |
ELISA |
Significantly higher levels of IL-6 and IL-8 in serum were found in patients with VAS ≥ 3 at 12-mo follow-up, than in patients with VAS < 3. Covariates for IL-6: age, smoking; covariates for IL-8: smoking, treatment. |
|
Schistad et al. [19] 2014 |
Prognostic biomarker for chronic pain |
ELISA |
High serum IL-6 levels, but not disc degeneration or Modic changes, were associated with less favorable recovery in patients with lumbar radicular pain. |
|
Haddadi et al. [20] 2020 |
Prognostic biomarker for postsurgery pain recovery |
ELISA |
Presurgery IL-6 had inverse correlation with recovery in VAS score (correlation coefficient = -0.443 and p = 0.011). IL-6 concentration at 4.36 and 1.16 pg/mL before and after the surgery, respectively were determined as cutoff to predict severe disability after surgery. Their AUC of ROC curves were 0.598 and 0.667, respectively. |
|
Fathy et al. [21] 2024 |
Prognostic biomarker for postradiofrequency treatment recovery in pain and function |
ELISA |
S100B level before radiofrequency was positively correlated with NRS (p < 0.001) and FRI (p = 0.001) 6 months after radiofrequency. BDNF serum level before radiofrequency was negatively correlated with both NRS and FRI 6 mo following radiofrequency (p = 0.022, p = 0.041, respectively). |
|
Kamieniak et al. [8] 2019 |
Monitoring biomarkers for pre- and postsurgery pain score |
MMP-2 and MMP-9 activities were determined through gelatin zymography; TIMPs levels were assessed using ELISA |
A higher level of MMP-2 (p < 0.001) and TIMP-2 (p < 00.1), but not MMP-2 activity was correlated with higher Numeric Rating Scale for leg pain (NRS-L). |
|
Lower level and activity of MMP-2 was weakly correlated with longer signs of LDH (rs = −0.23, p = 0.053 and rs = −0.26, p = 0.044). The level of MMP-9 (p < 0.001) and MMP-9 activity were correlated with a higher level of NRS-L (rs = 0.27, p < 0.05 and rs = -0.31, p < 0.05, respectively), and a higher level of TIMP-2 was correlated with a higher Pain Rating Index (PRI, rs = 0.27, p < 0.005) and present pain intensity (rs = 0.35, p < 0.01). |
|
Kamieniak et al. [63] 2020 |
Monitoring biomarkers for pre- and postsurgery pain score |
ELISA |
IFN-γ in patients increased after surgery and was higher in ruptured disc (9.54 ± 7.41 pg/mL and 10.70 ± 11.88 pg/mL for pre-and postsurgery, respectively) than unruptured disc (6.36 ± 1.67 pg/mL and 7.99 ± 2.30 pg/mL for pre- and postsurgery, respectively). IFN-γ before surgery was correlated with 3-mo postsurgery pain scores (PRI score, rs = 0.462; p = 0.023; NRS back score, rs = 0.528; p = 0.008). |
|
Ran et al. [23] 2021 |
Monitoring biomarkers for postsurgery evaluation |
ELISA |
The serum concentrations of IL-6, TNF-α, CPK, and CRP in the open discectomy group were higher than those in the computerized tomography navigation percutaneous spinal endoscopy group postoperatively. |
|
Zu et al. [24] 2016 |
Monitoring biomarkers of inflammatory cytokines after microdiscectomy |
ELISA |
TNF-α decreased over time in the VAS ≥ 3 and VAS < 3 groups, while IL-4 increased in both groups at 1 month and then gradually decreased until month 12. The changes in serum levels of TNF-α and IL-4 over time between the VAS ≥ 3 and VAS < 3 groups were significantly different. |
|
Fathy et al. [51] 2022 |
Monitoring biomarkers to evaluate and compare outcomes of treatments |
ELISA |
SOD and GSH serum levels were increased 2 wk after magnesium sulphate+ steroids injection (p = 0.002, p = 0.005, respectively) and ozone+ steroids injection groups (p < 0.001), but no significant change in SOD and GSH over 2 wk was detected after transforaminal steroids injection alone treatment (p = 0.059, p = 0.494, respectively). |
|
Tan et al. [36] 2025 |
Lipid biomarker for LDH mechanism |
CIBA Corning 550 Express Auto |
High-ApoB (0.93 ± 0.21 g/L vs. 0.82 ± 0.21 g/L), high A1 (1.33 ± 0.24 g/L vs. 1.25 ± 0.23 g/L), and high-ApoB/ApoA1 ratio (0.72 ± 0.22 vs. 0.67 ± 0.19) were significantly higher in the LDH group compared to the spine fracture group. |
|
Zhang et al. [37] 2016 |
Lipid biomarker for LDH mechanism |
Cobas 8000 system |
Patients with LDH had significantly higher serum TC (4.75 mmol/L; range, 2.03–10.27 mmol/L) (p < 0.001) and LDL-C (2.92 mmol/L; range, 0.68–8.33 mmol/L) (p < 0.001) compared with patients with wounded lower limbs group, TC (4.41 mmol/L; range, 1.69–7.35 mmol/L) and LDL-C (2.62 mmol/L; range, 0.53–4.99 mmol/L). |
|
Longo et al. [35] 2011 |
Lipid biomarker for LDH mechanism |
CIBA system |
Patients with symptomatic LDH showed significantly higher TC (82–321 mg/dL vs. 44~813 mg/dL) and TG (137–320 mg/dL vs. 82–321 mg/dL) concentration than arthroscopic meniscectomy patients. |
REFERENCES
- 1. Deyo RA, Mirza SK. Clinical practice. Herniated lumbar intervertebral disk. N Engl J Med 2016;374:1763-72.
- 2. Kreiner DS, Hwang SW, Easa JE, et al. An evidence-based clinical guideline for the diagnosis and treatment of lumbar disc herniation with radiculopathy. Spine J 2014;14:180-91.
- 3. Chen X, Wang W, Cui P, et al. Evidence of MRI image features and inflammatory biomarkers association with low back pain in patients with lumbar disc herniation. Spine J 2024;24:1192-201.
- 4. Mobasheri A, Loeser R. Clinical phenotypes, molecular endotypes and theratypes in OA therapeutic development. Nat Rev Rheumatol 2024;20:525-6.
- 5. de Gramont A, Watson S, Ellis LM, et al. Pragmatic issues in biomarker evaluation for targeted therapies in cancer. Nat Rev Clin Oncol 2015;12:197-212.
- 6. FDA-NIH Biomarker Working Group. BEST (Biomarkers, EndpointS, and other Tools) resource. Silver Spring (MD): Food and Drug Administration (US); Bethesda (MD): National Institutes of Health (US); 2016-.
- 7. Jacobsen HE, Khan AN, Levine ME, et al. Severity of intervertebral disc herniation regulates cytokine and chemokine levels in patients with chronic radicular back pain. Osteoarthritis Cartilage 2020;28:1341-50.
- 8. Kamieniak P, Bielewicz J, Kurzepa J, et al. The impact of changes in serum levels of metalloproteinase-2 and metalloproteinase-9 on pain perception in patients with disc herniation before and after surgery. J Pain Res 2019;12:1457-64.
- 9. Palada V, Ahmed AS, Finn A, et al. Characterization of neuroinflammation and periphery-to-CNS inflammatory crosstalk in patients with disc herniation and degenerative disc disease. Brain Behav Immun 2019;75:60-71.
- 10. Coquelet P, Da Cal S, El Hage G, et al. Specific plasma biomarker signatures associated with patients undergoing surgery for back pain. Spine J 2025;25:32-44.
- 11. Grad S, Bow C, Karppinen J, et al. Systemic blood plasma CCL5 and CXCL6: potential biomarkers for human lumbar disc degeneration. Eur Cell Mater 2016;31:1-10.
- 12. Yıldırım Uslu E, Gülkesen A, Akgol G, et al. Serum endothelin-1 level can reflect the degree of lumbar degeneration: a cross-sectional study. Cureus 2024;16:e59966.
- 13. Brisby H, Olmarker K, Larsson K, et al. Proinflammatory cytokines in cerebrospinal fluid and serum in patients with disc herniation and sciatica. Eur Spine J 2002;11:62-6.
- 14. Hider SL, Konstantinou K, Hay EM, et al. Inflammatory biomarkers do not distinguish between patients with sciatica and referred leg pain within a primary care population: results from a nested study within the ATLAS cohort. BMC Musculoskelet Disord 2019;20:202.
- 15. Zhang YG, Jiang RQ, Guo TM, et al. MALDI-TOF-MS serum protein profiling for developing diagnostic models and identifying serum markers for discogenic low back pain. BMC Musculoskelet Disord 2014;15:193.
- 16. Deng Q, Ren S, Zhang N, et al. Unravelling lumbar disc herniation severity beyond MRI: integrated transcriptomic and metabolomic analyses highlight glycerophospholipid metabolism and inform a machine-learning diagnostic model: a pilot study. Bone Joint Res 2025;14:434-47.
- 17. Moen A, Lind AL, Thulin M, et al. Inflammatory serum protein profiling of patients with lumbar radicular pain one year after disc herniation. Int J Inflam 2016;2016:3874964.
- 18. Pedersen LM, Schistad E, Jacobsen LM, et al. Serum levels of the pro-inflammatory interleukins 6 (IL-6) and -8 (IL-8) in patients with lumbar radicular pain due to disc herniation: a 12-month prospective study. Brain Behav Immun 2015;46:132-6.
- 19. Schistad EI, Espeland A, Pedersen LM, et al. Association between baseline IL-6 and 1-year recovery in lumbar radicular pain. Eur J Pain 2014;18:1394-401.
- 20. Haddadi K, Abediankenari S, Alipour A, et al. Association between serum levels of interleukin-6 on pain and disability in lumbar disc herniation surgery. Asian J Neurosurg 2020;15:494-8.
- 21. Fathy W, Hussein M, Magdy R, et al. Predictive value of S100B and brain derived neurotrophic factor for radiofrequency treatment of lumbar disc prolapse. BMC Anesthesiol 2024;24:161.
- 22. Kuiper JI, Verbeek JH, Frings-Dresen MH, et al. Exploration of the use of biomarkers to monitor recovery after surgery for lumbar disc herniation: a prospective cohort study. Clinical Spine Surgery 2002;15:398-403.
- 23. Ran B, Wei J, Yang J, et al. Quantitative evaluation of the trauma of CT navigation PELD and OD in the treatment of HLDH: a randomized, controlled study. Pain Physician 2021;24:E433-41.
- 24. Zu B, Pan H, Zhang XJ, et al. Serum levels of the inflammatory cytokines in patients with lumbar radicular pain due to disc herniation. Asian Spine J 2016;10:843-9.
- 25. Cazzanelli P, Wuertz-Kozak K. MicroRNAs in intervertebral disc degeneration, apoptosis, inflammation, and mechanobiology. Int J Mol Sci 2020;21:3601.
- 26. Statello L, Guo CJ, Chen LL, et al. Gene regulation by long non-coding RNAs and its biological functions. Nat Rev Mol Cell Biol 2021;22:96-118.
- 27. Li Z, Chen X, Xu D, et al. Circular RNAs in nucleus pulposus cell function and intervertebral disc degeneration. Cell Prolif 2019;52:e12704.
- 28. Hasvik E, Schjolberg T, Jacobsen DP, et al. Up-regulation of circulating microRNA-17 is associated with lumbar radicular pain following disc herniation. Arthritis Res Ther 2019;21:186.
- 29. Zhang G, Zhang W, Hou Y, et al. Detection of miR‑29a in plasma of patients with lumbar spinal stenosis and the clinical significance. Mol Med Rep 2018;18:223-9.
- 30. Cui S, Zhou Z, Liu X, et al. Identification and characterization of serum microRNAs as biomarkers for human disc degeneration: an RNA sequencing analysis. Diagnostics (Basel) 2020;10:1063.
- 31. Correia C, Ribeiro-Machado C, Caldeira J, et al. Circulating tenascin-C/-miR-155-5p identified as promising prognostic candidates of intervertebral disc herniation. Bioengineering 2026;13:74.
- 32. Zou ZF, He JP, Chen YL, et al. Increased local miR-21 expressions are linked with clinical severity in lumbar disc herniation patients with sciatic pain. Adv Clin Exp Med 2022;31:723-30.
- 33. Wei L, Guo J, Zhai W, et al. CircRNA GRB10 is a novel biomarker for the accurate diagnosis of lumbar degenerative disc disease. Mol Biotechnol 2023;65:816-21.
- 34. Jensen MC, Brant-Zawadzki MN, Obuchowski N, et al. Magnetic resonance imaging of the lumbar spine in people without back pain. New Eng J Med 1994;331:69-73.
- 35. Longo UG, Denaro L, Spiezia F, et al. Symptomatic disc herniation and serum lipid levels. Eur Spine J 2011;20:1658-62.
- 36. Tan B, Xiang S, Zheng Y, et al. Association of dyslipidemia with intervertebral disc degeneration: a case-control study. Eur J Med Res 2025;30:194.
- 37. Zhang Y, Zhao Y, Wang M, et al. Serum lipid levels are positively correlated with lumbar disc herniation--a retrospective study of 790 Chinese patients. Lipids Health Dis 2016;15:80.
- 38. Welsh JA, Goberdhan DCI, O'Driscoll L, et al. Minimal information for studies of extracellular vesicles (MISEV2023): from basic to advanced approaches. J Extracell Vesicles 2024;13:e12404.
- 39. Irmer B, Chandrabalan S, Maas L, et al. Extracellular vesicles in liquid biopsies as biomarkers for solid tumors. Cancers (Basel) 2023;15:1307.
- 40. Ragni E. Extracellular vesicles: recent advances and perspectives. Front Biosci (Landmark Ed) 2025;30:36405.
- 41. Jin Y, Wu O, Chen Z, et al. Exploring extracellular vesicles as novel therapeutic agents for intervertebral disc degeneration: delivery, applications, and mechanisms. Stem Cell Res Ther 2025;16:221.
- 42. Moen A, Jacobsen D, Phuyal S, et al. MicroRNA-223 demonstrated experimentally in exosome-like vesicles is associated with decreased risk of persistent pain after lumbar disc herniation. J Transl Med 2017;15:89.
- 43. Sung SE, Seo MS, Park WT, et al. Extracellular vesicles: their challenges and benefits as potential biomarkers for musculoskeletal disorders. J Int Med Res 2025;53:3000605251317476.
- 44. Tilotta V, Vadala G, Ambrosio L, et al. Mesenchymal stem cell-derived exosomes: the new frontier for the treatment of intervertebral disc degeneration. Appl Sci-Basel 2021;11:11222.
- 45. Bernardi S, Balbi C. Extracellular vesicles: from biomarkers to therapeutic tools. Biology (Basel) 2020;9:258.
- 46. De Sousa KP, Rossi I, Abdullahi M, et al. Isolation and characterization of extracellular vesicles and future directions in diagnosis and therapy. Wiley Interdiscip Rev Nanomed Nanobiotechnol 2023;15:e1835.
- 47. DiStefano TJ, Vaso K, Danias G, et al. Extracellular vesicles as an emerging treatment option for intervertebral disc degeneration: therapeutic potential, translational pathways, and regulatory considerations. Adv Healthc Mater 2022;11:e2100596.
- 48. Piazza N, Dehghani M, Gaborski TR, et al. Therapeutic potential of extracellular vesicles in degenerative diseases of the intervertebral disc. Front Bioeng Biotechnol 2020;8:311.
- 49. Cheng L, Fan W, Liu B, et al. Th17 lymphocyte levels are higher in patients with ruptured than non-ruptured lumbar discs, and are correlated with pain intensity. Injury 2013;44:1805-10.
- 50. Park MS, Lee HM, Hahn SB, et al. The association of the activation-inducible tumor necrosis factor receptor and ligand with lumbar disc herniation. Yonsei Med J 2007;48:839-46.
- 51. Fathy W, Hussein M, Ibrahim RE, et al. Comparative effect of transforaminal injection of Magnesium sulphate versus Ozone on oxidative stress biomarkers in lumbar disc related radicular pain. BMC Anesthesiol 2022;22:254.
- 52. Morimoto T, Kobayashi T, Ito H, et al. Serum periostin levels correlate with severity of intervertebral disc degeneration. Eur Spine J 2024;33:2007-13.
- 53. Dube CT, Gilbert HT, Rabbitte N, et al. Proteomic profiling of human plasma and intervertebral disc tissue reveals matrisomal, but not plasma, biomarkers of disc degeneration. Arthritis Res Ther 2025;27:28.
- 54. Vignoli A, Tenori L, Morsiani C, et al. Serum or plasma (and which plasma), that is the question. J Proteom Res 2022;21:1061-72.
- 55. World Health Organization. Use of anticoagulants in diagnostic laboratory investigations. Geneva (Switzerland): World Health Organization; 2002.
- 56. Song Z, Yan M, Zhang S, et al. Implications of circadian disruption on intervertebral disc degeneration: the mediating role of sympathetic nervous system. Ageing Res Rev 2025;104:102633.
- 57. Kreiner DS, Matz P, Bono CM, et al. Guideline summary review: an evidence-based clinical guideline for the diagnosis and treatment of low back pain. Spine J 2020;20:998-1024.
- 58. Cai M, Yin J, Jin Y, et al. A nomogram model integrating inflammation markers for predicting the risk of recurrent sciatica after selective nerve root blocks. Risk Manag Healthc Policy 2025;18:1279-89.
- 59. Mariscal G, Torres E, Barrios C. Incidence of recurrent lumbar disc herniation: a narrative review. J Craniovertebr Junction Spine 2022;13:110-3.
- 60. Tenny S, Kerndt CC, Hoffman MR. Case control studies. StatPearls. Treasure Island (FL): StatPearls Publishing; 2025.
- 61. Wang X, Ren Z, Wang B, et al. Blood expression of NADK2 as a diagnostic biomarker for sciatica. iScience 2024;27:111196.
- 62. Ojala M, Garriga GC. Permutation tests for studying classifier performance. Paper presented at: 2009 Ninth IEEE International Conference on Data Mining; 2009 Dec 6-9; Miami (FL), USA.
- 63. Kamieniak P, Bielewicz JM, Grochowski C, et al. IFN-γ correlations with pain assessment, radiological findings, and clinical intercourse in patient after lumbar microdiscectomy: preliminary study. Dis Markers 2020;2020:1318930.
, Junxuan Ma2
