Introduction

Low back pain (LBP) is now the number one cause of disability globally [1], posing a major health problem that causes an enormous social and economic burden on the community and health systems [2, 3]. It has a calculated global prevalence of 9442.5 per 100,000 adults (9%) [4]. Chronic low back pain (CLBP) is defined as pain, muscle tension or stiffness localized below the costal margin and above the inferior gluteal folds, with or without neurological symptoms in the lower limb, and is defined as chronic when it persists for 12 weeks or more [5]. CLBP affects approximately 20–25% of the elderly population (older than 65 years) [6,7,8]. At any given time, 12–33% of the adult population has low back pain [9], being more common in women than in men [10,11,12]. Causes of CLBP can be distinguished into specific (e.g. degenerative process to the spinal segments of the lumbar spine such as lumbar spinal stenosis, spondylolisthesis, disc degeneration or herniation) or non-specific, when there is no identifiable cause of pain [13,14,15,16].

Cytokines are regulatory proteins (pro-inflammatory biomarkers) which, in the case of inflammation, modulate the inflammatory response of all cells of the immune system. Pro-inflammatory cytokines, such as IL-1B, IL-6 and TNF-α, can be objectively measured in the CNS and circulation and have been implicated in the processes of central sensitization and chronic LBP [16,17,18]. On the other hand, anti-inflammatory cytokines, such as IL-4 and IL-10, inhibit pro-inflammatory cytokine response. Elevated IL-6 and reduced IL-10 levels were described in peripheral blood of non-specific chronic low back pain (NsCLBP) patients, thereby suggesting that an imbalance between pro-inflammatory and anti-inflammatory mediators contributes to the pathophysiology of LBP [19, 20]. Other studies reported an association between increased pro-inflammatory cytokines and pain intensity levels in a population with NsCLBP [21,22,23,24,25,26].

The purpose of this study was to systematically review the evidence on inflammatory biomarkers as analytic predictors of NsLBP, assess whether patients with NsLBP present changes in several inflammatory biomarkers and analyse whether LBP severity is associated with the magnitude of changes in inflammatory biomarkers. A deeper understanding of the aetiology and pathophysiology associated with non-specific LBP may lead to a better stratification of these patients and to the development of better targeted interventions.

Materials and methods

Search strategy and selection criteria

The recommendations from the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [27] were followed to perform this systematic review, and its protocol was registered on the International prospective register of systematic reviews (PROSPERO—ID CRD42022375174).

A systematic search was conducted on July 2022 in Cochrane Library, MEDLINE and Web of Science for all the available literature, based on the following query: (“low back pain” OR “lower back pain” OR “back pain” OR “backache” OR "back ache" or “lumbago” OR “lumbar pain” OR “sciatica” OR “back disorder”) AND (“inflammation mediators” OR “inflammatory markers” OR “cytokines” OR “interleukins” OR “monokines” OR “chemokines” OR “tumor necrosis factor” OR “C-Reactive Protein” OR "CRP" OR “C-reactive protein” OR “hs-CRP” OR "hsCRP" OR “biological markers” OR “biomarkers”). Studies were retrieved and duplicates removed electronically and manually. References were transferred to the reference management software Rayyan (https://rayyan.qcri.org). This software was developed specifically to expedite the initial screening of abstracts and titles for systematic reviews and to allow for blinded screening, in this case, between two authors.

Cross-sectional, longitudinal cohort or case–control studies that evaluated the relationship between inflammatory biomarkers and low back pain in humans were considered eligible for inclusion, as well as prospective and retrospective studies.

Literature screening

Each full-text article was searched for reports studying a correlation between inflammatory biomarkers and the prevalence and intensity of chronic LBP. Observational cohort studies (with and without control group), cross-sectional studies and randomized clinical trials (RCT) were included. Studies that comprised patients older than 18 years old, in which one or more inflammatory biomarkers were measured in blood plasma, were included. We excluded publications in a population with previous lumbar pathology, with underlying systemic pathology (e.g. autoimmune diseases or osteoarthritis) that could influence the systemic inflammatory biomarkers measurement and studies with fewer than 20 subjects. The duration of LBP was retrieved to distinguish between acute (< 6 weeks) and chronic (≥ 6 weeks) NsLBP, and studies in a population with different sources of pain besides LBP were excluded, as well as studies with evaluation of areas other than the lumbar spine (e.g. cervical and thoracic spine) or evaluating inflammatory biomarkers in the intervertebral discs. Studies specifically evaluating the effectiveness of pharmaceutical intervention on pain or anti-cytokine therapy, biomechanical and cadaver studies, duplicate publications and studies that did not meet the inclusion criteria were also excluded (Fig. 1).

Fig. 1
figure 1

Strategy used to identify published reports

Full size image

Data extraction

The following items were recorded from all eligible studies: study design and study purpose, patient demographics and related characteristics (number, age, gender, body mass index), inclusion and exclusion criteria, patient outcomes associated with biomarker concentrations and included biomarkers. Data concerning the measuring instrument or scale used to evaluate the severity of LBP, quality of life and return to work were also extracted. The type, number and concentration of the assessed inflammatory biomarkers were extracted, and no particular biomarker associated with changes in inflammatory processes was excluded. Data of titles and abstracts were analysed independently by two reviewers, and any discrepancies were resolved by consensus. After the primary selection, full-text eligibility was done by two reviewers using screening tools developed a priori. The most common limitations amongst the studies were a low number of patients included, lack of ascertainment of exposure or definition of controls, lack of blinding for outcome assessments and adequacy of follow-up of cohorts. We present the results of the cross-sectional studies and the longitudinal studies separately. When odds ratios (ORs) were presented, the p-value and the 95% confidence interval (CI) were extracted. For other measures of association, the p-value was used to assess whether the association was statistically significant.

Assessment of study quality

The methodological quality of included studies was evaluated using two versions (for case–control and cohort studies) of the Newcastle–Ottawa Assessment scale (NOS) [28] (Table 2). As described in Table 2, amongst all studies, the minimum score was 4 out of 9 and maximum score was 7 out of 9, with a median score of 6.

Results

Study selection

The systematic database search resulted in a total of 4016 records (Web of Science (n = 1962), MEDLINE (n = 2051), Cochrane Library (n = 3)). Following deduplication, a total of 3022 potentially relevant studies were identified and screened for retrieval, of which 2831 articles were excluded as they did not fulfil the eligibility criteria. Regarding the selection of articles for full-text screening, there was 95% consensus between the two reviewers, and 191 studies were included, of which 176 were excluded. The reasons for exclusion were as follows: review articles (n = 10), articles with non-serum biomarkers evaluation (i.e. inflammatory biomarkers evaluated by intervertebral disc, ligamentum flavum or muscle biopsy) (n = 33), evaluation of patients with specific spinal disorders (n = 28) and articles that did not meet the inclusion criteria (n = 105). Hence, 15 articles were included for synthesis in this systematic review. The entire selection process is presented in the flowchart of Fig. 1.

Study characteristics

Studies characteristics and demographics are presented in Table 1. All the papers were published between 2006 and 2022, with 12 of the 15 included studies published from 2016 to 2022. Eleven studies [19, 29,30,31,32,33,34,35,36,37,38,39] were cross-sectional and four were longitudinal [40,41,42,43]. Regarding the characteristics of the population included, seven studies [34, 35, 37, 39,40,41,42,43] assessed non-specific acute low back pain (NsALBP) and ten studies [19, 29, 32,33,34, 36,37,38,39,40,41] included persons with NsCLBP. The differentiation cut-off between acute low back pain (ALBP) and CLBP was 6 weeks. Ten studies [19, 29, 32,33,34, 37,38,39,40,41] presented a control group. Regarding the restrictions in patient inclusion, two studies included only female patients [32, 44] and one studied a multi-generation cohort with evaluations at 14, 17, 20 and 22 years [42]. Sample size comprised a total of 14,555 patients with LBP (ALBP (n = 2073); CLBP (n = 12,482)) and 494 controls. Excluding the multi-generation cohort study [42], the mean age ranged from 29 to 71 years. Regarding the most common inflammatory biomarkers evaluated (Table 3), nine studies included C-reactive protein (CRP) or Hs-CRP (high-sensitive C-reactive protein) [30, 34, 36, 38, 40,41,42,43], eight studies included IL-1 and IL-1β [32,33,34, 37,38,39, 41, 43], ten studies included IL-6 [19, 29, 32, 34, 35, 37,38,39, 41, 43], two studies included IL-10 [19, 34] and eight studies included TNF-α [32, 34, 35, 37,38,39, 41, 43]. Other inflammatory biomarkers that were not analysed in more than one study are described in Table 1, but an exhaustive analysis of these was not carried out. Pain severity was evaluated most commonly using the visual analogue scale (VAS) in five studies [32, 34, 37, 39, 40], Roland–Morris Disability Questionnaire (RMDQ) in four studies [32, 34, 41, 43], numerical rating scale (NRS) in three studies [35, 41, 43], McGill Pain Questionnaire (MPQ) in three studies [29, 35, 38] and Oswestry Disability Index (ODI) in two studies [37, 38].

Table 1 Studies characteristics and demographics
Full size table

Clinical features in relationship to biomarkers

LBP assessment, duration of symptoms, biomarkers studied, source and technique of collection, analysis of inflammatory biomarkers and the type of association (positive, negative, or non-existent) between clinical parameters and inflammatory biomarkers are presented in Tables 2 and 3. A positive association means that the presence of LBP is associated with higher levels of the pro-inflammatory biomarkers measured in the blood serum. Regarding Hs-CRP, only one study [40] demonstrated a positive association with the presence of LBP and two studies [40, 42] did not find any relation between these variables. In studies that evaluated only C-reactive protein (CRP) [30, 34, 36, 39, 41, 43], all showed a positive association with the presence of LBP. Within the eight studies evaluating interleukin 1 (IL-1) and IL-1β, in seven studies [33,34,35, 37, 39, 41, 43] no association was found with the presence of LBP. On the contrary, in a cross-sectional study including 148 patients with CLBP and 150 controls, Dadkhah et al. [38] found a positive correlation between these variables. The pro-inflammatory biomarker interleukin 6 (IL-6) was assessed in ten studies, and a positive association was found between IL-6 and CLBP in six studies [19, 34, 35, 38, 39, 43]. Klyne et al. [34], in a study comparing a group of 99 patients with ALBP (acute low back pain) (divided into a high and low pain intensity subgroups) with 55 controls, found a positive correlation between IL-6 and patients with high low back pain intensity (p = 0.045), but there were no significant differences between low pain intensity subgroup and control group (p = 0.141). The rest of the studies found no association between these variables [29, 32, 41]. The evaluation of the interleukin 10 (IL-10) biomarker was only assessed in two studies, with a negative association with the presence of LBP. These studies presented a low [19] and moderate [37] risk of bias. The association between changes in serum TNF-α and presence of LBP was achieved in eight studies [32, 34, 35, 37,38,39, 41, 43], with five studies [32, 35, 37, 38, 43] demonstrating a positive association between these variables. Four studies [37, 39,40,41] have made direct comparisons between ALBP and CLBP groups regarding their inflammatory biomarkers profile. Gebhardt et al. [40] found no association between Hs-CRP and acute or chronic LBP. On the other hand, Injeyan et al. [37] demonstrated that IL-1, IL-1β and IL-6 values were significantly higher in the ALBP group (p = 0.0001), but not in the CLBP group, and that TNF-α values were significantly higher in the ALBP group (p = 0.0001) and CLBP group (p = 0.003).

Table 2 Quality assessment and risk of bias using the modified Newcastle–Ottawa scale
Full size table
Table 3 Studies results according to the type of inflammatory biomarker
Full size table

Discussion

There is a lack of evidence regarding the association between inflammatory biomarkers and LBP. In this systematic review, most studies found a positive correlation between classic pro-inflammatory biomarkers and NsLBP, namely CRP (6 out of 6), IL-1 and IL-1β (7 out of 8), IL-6 (6 out of 10) and TNF-α (5 out of 8). On the other hand, Hs-CRP did not present a significant correlation with LBP (1 out of 3) and anti-inflammatory biomarker IL-10 demonstrated a negative association with NsLBP (2 out of 2).

Previous research has demonstrated a link between chronic inflammation and central sensitization development [15, 45]. Khan et al. [46] have demonstrated an association between increasing levels of circulating pro-inflammatory biomarkers, such as CRP and IL-6, and an increase in pain and disc degeneration in a population with specific and NsLBP. In previous reviews, Berg et al. [47] and Lim et al. [48] found consistent evidence for an association between elevated levels of serum or plasma CRP, TNF and IL-6 and the presence of NsLBP. However, the previous two review studies [47, 48] included both acute and chronic non-specific LBP, and the results were not reported separately for each population, which reveals a limitation in the interpretation of results, as these cytokines are reported to play different roles in the acute and chronic phase of pain [43, 49]. Of the included studies that directly compared groups with acute and chronic LBP, Gebhardt et al. [40] have found no differences in Hs-CRP plasma levels between patients with acute and chronic LBP. On the other hand, Injeyan et al. [37] demonstrated increased levels of IL-1, IL-6 and TNF-α in the acute LBP group, but only increased levels of TNF-α in the chronic LBP group.

The characterization of LBP, particularly its intensity and duration, presents important heterogeneity between studies, given the multiplicity of pain assessment questionnaires. However, the assessment through similar evaluation scales (NRS and VAS) is verified in eight studies [32, 34, 35, 37, 39,40,41, 43].

Successful validation of candidate biomarkers is a demanding and time-consuming process that requires multiple rigorous and carefully designed clinical studies. Indeed, there is moderate evidence for the existence of an increase in the pro-inflammatory systemic profile in association with NsLBP. This relationship and the scarce evidence of the association between the severity of NsLBP and the magnitude of changes in the systemic inflammatory profile are matters of extreme importance to be studied in the future. This could provide a set of potential biomarkers that can more objectively monitor the degree of disease activity and become targets for new treatments.

Strengths and limitations

This study presents several strengths, including the use of PRISMA guidelines for reporting systematic reviews and the pre-registration in the PROSPERO database, reducing unplanned duplication and potential publication bias. Contrary to previous reviews, a survey was carried out excluding patients with some type of underlying pathology of the spine or systemic pathology that could influence inflammatory biomarkers measurement. On the other hand, the differentiation between ALBP and CLBP regarding changes in the inflammatory profile was performed and is detailed in Tables 1 and 3. The differentiation between studies evaluating CRP and Hs-CRP was also performed.

This study has some limitations. Studies on this topic are mostly case–control studies, which entail certain intrinsic limitations such as sampling bias and recall bias, in addition to the fact that this type of studies could only reach a level B of evidence at most. In this type of research, a publication bias is also a possibility, given that funnel plot is not possible due to the heterogeneity of markers and publications. On the other hand, screening references may result in an over representation of positive studies, as trials with a negative result are less likely to be published and hence referred. There was a cross-sectional limitation to the studies presented regarding the presentation of mean differences and confidence intervals or biomarker concentration levels. One article [38] did not accurately describe the method for biomarker analysis, which can compromise the quality of the analysis. On the other hand, although the included studies present gender percentages of the included patients and no significant differences were found between control and patient groups for age and gender, the comparative analysis of the differences in systemic cytokine level between males and females was not investigated. On the other hand, the included studies do not differentiate external factors, such as the use of corticosteroids or NSAIDs, which may influence the inflammatory biomarkers value [50].

Finally, previous research [51,52,53] has demonstrated that a significant percentage of patients with ALBP still have some degree of symptomatology after one year of follow-up. Thus, a longer follow-up is advisable in the future studies.

Conclusion

In this systematic review, considering the overall risk of bias of the included studies, individuals with NsLBP were found to have an associated increase of systemic inflammation, demonstrated by the increased levels of pro-inflammatory biomarkers CRP, IL-6 and TNF-α and decreased anti-inflammatory biomarker IL-10. IL-1 and Hs-CRP were not correlated with NsLBP. There is insufficient evidence to associate these findings with the degree of pain severity or the activity status of the lumbar pain over time. The utilization of blood biobanks and longitudinal evaluation of prospective cohorts would be pertinent in examining the relationship between biomarkers and low back pain, as they can provide a valuable resource for obtaining samples and exploring potential mechanisms underlying the association.