Adipokines and Inflammatory Biomarkers Linking Obesity and Periodontitis: A Critical Review

The main purpose was to critically examine the associations between adiposity, periodontal status, and the four principal biomarkers. The secondary objective was to assess changes after periodontal treatment and determine what they support regarding diagnosis, treatment monitoring, and clinical decision-making. We addressed the observational associations and treatment effects separately because they answer different questions. Molecular studies are considered as supporting biological evidence, not as substitutes for clinical validation.

Although observational research has linked adiposity and periodontal status with altered biomarker levels, group-level associations do not determine individual diagnostic cut-offs (Table 1).

2. MATERIALS AND METHODS

This critical narrative review covered selected clinical and molecular studies related to adiposity, periodontitis, and host inflammatory biomarkers. The literature pertaining to clinical studies was limited to papers available up to July 31, 2025. Primary research articles were collected using focused search strategies in PubMed (last updated on July 31, 2025) and on publisher’s article webpages, utilising different sets of keywords (obesity, periodontitis, adipokines, IL-6, TNF-alpha, resistin, adiponectin, biomarkers) and author/DOI searches. Study inclusion was determined by relevance and not exhaustivity, focusing on those studies that (1) tested one or more of the four main biomarkers; (2) studied both adiposity and periodontitis conditions; and (3) presented primary data.

Clinical evidence included patients with documented obesity or metabolic syndrome and periodontal disease. Evidence for biomarkers also includes host inflammatory mediators or adipokines. Experimental studies involving treatment included longitudinal measurements pre- and post-periodontal treatment. Microbial and patient data were not considered as evidence for predicting host protein response. Molecular work was excluded from our analysis; however, experimental and review work provided background information.

We collected information on the study design, population, sample size, biomarkers, specimens, assay methods, treatment components, comparators, follow-up, and outcomes. Full texts were examined when they were accessible. The evidence was organised into observational studies, non-randomised periodontal treatment studies, randomised controlled trials, and molecular investigations.

We organised the findings according to study design, biomarker, specimen type, comparison group, and follow-up duration. Baseline associations, within-group changes, and between-group treatment effects were analysed separately. Circulating, gingival crevicular fluid, and salivary measurements were distinguished, as were the biomarker concentrations and total amounts. Critical appraisal addressed comparator suitability, confounding, baseline balance, missing data, specimen collection, assay reporting, treatment content, and clinical applicability.

Statistically significant and non-significant findings were retained. A significant within-group change was not considered to indicate a between-group treatment effect, and nonsignificance was not considered proof of equivalence. No meta-analysis was conducted, and neither a formal assessment of risk of bias nor certainty grading was performed.

3. CLINICAL EVIDENCE

Observational studies have linked adiposity and periodontal severity to changes in host biomarkers. However, the association varies by comparison. In 112 adults, Jia et al. [6] found more severe periodontal disease and higher IL-6 and TNF-α levels in gingival crevicular fluid (GCF) among obese participants. Compared with the normal-weight participants, serum resistin and leptin levels were higher, and adiponectin levels were lower [6]. These findings link clinical periodontal measurements with both a local inflammatory profile and a circulating adipokine profile. Because not all participants had periodontitis and disease severity differed across weight groups, the results do not isolate an obesity effect within otherwise identical periodontal disease.

Not all metabolic comparisons produced the same associations. For example, Nilsen et al. [7] examined 52 adults with obesity and observed that metabolic syndrome was not significantly associated with periodontitis [7], in contrast to the obesity-periodontitis association reported by Jia et al. [6]. In contrast, Rangé et al. [8] found that circulating orosomucoid levels were associated with periodontal disease severity in 32 adults with severe obesity and periodontitis [8]. This result concerns an acute-phase marker within a defined clinical population; it does not establish that every adipokine reflects the disease severity.

Alkan and Guzeldemir-Akcakanat [9] compared 71 women across obesity–periodontal phenotypes and reported different adipokine patterns in serum, saliva, and GCF [9]. Their findings reinforce the importance of specifying the specimen rather than assuming a common systemic and oral fluid signature. Table 1 summarizes the populations, measured outcomes, and clinically relevant findings.

The practical message is that clinicians should document adiposity and periodontal status when interpreting a biomarker. These group-level associations do not define individual diagnostic thresholds.

Non-randomized reports demonstrate that patients with obesity can show periodontal improvement after nonsurgical periodontal therapy (NSPT). Altay et al. [10] treated 46 patients with generalized chronic periodontitis, including 22 with obesity and dyslipidemia. Both weight groups improved clinically, and at three months the obese group showed reductions in serum IL-6 and TNF-α [10]. Al-Zahrani and Alghamdi [11] studied 40 women and reported reduced gingival inflammation and serum CRP at two months. CRP reduction did not differ significantly between the normal-weight and obese groups [11]. These are favorable clinical observations, but both studies compared treated weight groups rather than treatment against no treatment.

Gonçalves et al. [12] reported periodontal improvement in 48 patients (24 with obesity and 24 without) after SRP, without significant changes in serum leptin or adiponectin levels. At six months, probing depth decreased more in patients without obesity [12]. A separate one-year report in 40 patients showed that serum mediators did not change significantly, although selected GCF measurements changed according to weight group and sampling site [13]. The clinical and biochemical outcomes therefore provide different information: a measurable periodontal benefit does not necessarily produce a corresponding change in circulating adipokine levels.

Tahir et al. [14] analysed 48 of the 54 recruited participants: 18 with obesity and 30 with normal weight. Plaque and gingival bleeding measures improved after treatment, but serum resistin did not significantly decrease. Six participants were excluded from the obesity group, which constrains the interpretation of the complete-case comparison [14]. The actual instrumentation and chlorhexidine regimen are listed in Table 2.

Clinicians should directly assess periodontal outcomes rather than infer them from serum biomarker changes. Their before-and-after design also leaves time, behavioural change, regression to the mean, and concurrent systemic management as possible contributors to laboratory changes. A treated non-obese group informs response by adiposity but does not replace an untreated or randomised alternative-treatment comparator.

Akram et al. [15] randomized 66 adults with obesity and periodontitis; 62 were analyzed. The intervention combined SRP, oral-hygiene instruction, and chlorhexidine, whereas the control arm received no treatment during follow-up. Plaque outcomes favored treatment, while change in salivary resistin showed no significant association with the improvement of periodontal variables. Although resistin decreased within the treated group, the change did not differ significantly between groups (p = 0.418), and baseline concentrations differed [15]. This trial provides a clinical example of why a favorable paired comparison does not establish a biomarker treatment effect.

Montero et al. [16] randomised 63 patients with MetS and severe periodontitis to subgingival instrumentation plus azithromycin or supragingival plaque removal plus placebo. At six months, the adjusted serum high-sensitivity C-reactive protein (hsCRP) level was 1.2 mg/L lower after intensive treatment (95% confidence interval, 0.4–2.0 mg/L; p = 0.004) [16]. This is evidence of a controlled biomarker benefit for the complete treatment package. It does not separate the contributions of instrumentation and antibiotic treatment, nor does it establish a reduction in future cardiovascular or diabetes complications.

Table 2 distinguishes treatment components, comparators, specimens, and findings. Chlorhexidine-containing protocols and a regimen containing systemic azithromycin should not be presented as interchangeable with mechanical instrumentation alone. Similarly, the outcome of a MetS trial cannot be automatically generalised to every patient with obesity classified according to body mass index (BMI).

Resistin illustrates a distinction between association and responsiveness. Higher circulating resistin is associated with obesity, but neither the serum treatment study nor the randomized salivary comparison established a consistent treatment-related benefit [6, 14, 15]. Adiponectin provides a different pattern: lower circulating concentrations are associated with obesity, yet serum concentrations did not consistently increase after SRP [6, 12, 13]. Leptin similarly appeared more closely related to adiposity than to short-term treatment response in the selected studies [6, 12].

CRP is relevant as a circulating inflammatory outcome rather than an adipokine or a periodontal-specific diagnostic marker. The controlled hs-CRP finding supports the possibility that a defined periodontal treatment package can reduce systemic laboratory measures [16]. This finding should not be generalised to a uniform response across all biomarkers. These distinctions are consistent with the mediator- and specimen-dependent findings of an earlier systematic synthesis [4].

4. MOLECULAR EVIDENCE: FROM DISCOVERY TO CANDIDATE BIOMARKERS

Molecular studies provide complementary evidence on shared pathways but require distinct interpretation. Cai et al. [17] integrated separate public obesity and periodontitis expression datasets [17]. The findings generate hypotheses about convergent pathways but do not represent a newly recruited combined-disease cohort or a measured periodontal treatment effect on circulating proteins. Expression results from different datasets are not equivalent to an integrated clinical test measured in the same participants.

Foratori-Junior et al. [18] investigated label-free salivary proteomics in 40 women classified by obesity and periodontal status during pregnancy and after delivery [18]. Their findings add a protein discovery layer to the defined physiological setting. Follow-up across pregnancy and the postpartum period cannot be interpreted as follow-up after periodontal therapy, and the results require validation before being extrapolated to other adult populations.

In 292 men, Wilkins et al. [19] found that the association between central adiposity and periodontal progression depended on interleukin-1 inflammatory genetic variation [19]. This longitudinal finding supports investigating host susceptibility but does not establish a biomarker-guided treatment strategy. Chen et al. [20] reported no significant global genetic correlation between obesity and periodontitis, and their nominal BMI–periodontitis causal estimate was not robust across additional analyses [20].

Molecular studies in Table 3 identify candidates and hypotheses. Stronger clinical integration would require reproducible links between molecular measurements and a defined periodontal phenotype or outcome in independently characterised participants. Combining separate genetic, expression, and protein studies does not establish a chain or validate a multi-omics panel.

5. BIOLOGICAL AND ANALYTICAL INTERPRETATION

Adipose tissue and periodontal inflammation may share host response pathways; however, circulating biomarkers do not distinguish their separate contributions [2]. Human resistin illustrates the importance of cellular context because inflammatory activation of monocytes and macrophages contributes to its expression [5]. Adiponectin has demonstrated anti-inflammatory activity in experimental endothelial cells [21]. These findings explain the biological interest in the markers but do not establish that treatment-related changes mediate systemic clinical benefits.

The combination of altered circulating adipokines and local cytokines is compatible with interacting metabolic and periodontal inflammatory environments. This is not specific to a single causal pathway. To establish mediation, a study would need a treatment-to-biomarker effect, an appropriate temporal relationship with a subsequent clinical outcome, and an analysis addressing alternative explanations. A correlation with probing depth or an isolated laboratory change does not meet these requirements.

Blood-based measurements, site-specific GCF, and saliva represent distinct biological environments. Their responses can diverge, as shown in clinical reports [6, 9, 13]. Serum and plasma should be explicitly identified rather than described only as “blood” in an assay report. For GCF, the concentration and total recovered amount are different outcomes; a change in fluid volume may change the concentration without a corresponding change in the amount of mediator.

Relevant sampling details included the selected periodontal sites, duration of collection, recovered volume, contamination criteria, and whether the sites were pooled. A sample from a deep inflamed pocket is not interchangeable with a mixture of healthy and diseased sites. For blood and saliva, collection time, fasting status where relevant, stimulated or unstimulated saliva, processing delay, storage, and freeze–thaw exposure should be documented consistently across visits. These are reporting requirements for interpretable measurements, not assumptions about procedures used in every reviewed study.

Several studies have used enzyme-linked immunosorbent assays (ELISA) [12–15]. Comparability across studies requires additional details, including the platform, calibration range, quantification limits, units, dilution procedures, handling of out-of-range values, and within- and between-run precision. However, not all reviewed abstracts provided this level of detail. Platform, calibration range, quantification limits, units, dilution procedures, handling of out-of-range values, and within- and between-run precision are all relevant. Repeated or blinded laboratory measurements help assess analytical reliability but do not correct an unsuitable clinical comparator.

Transcriptomics measures ribonucleic acid (RNA) expression, discovery proteomics may report normalised relative abundance, and microbial quantitative polymerase chain reaction (qPCR) measures specified microbial targets [14, 17, 18]. These outputs should not be considered equivalent to the targeted host protein concentrations. A reliable assay must validate its clinical use; utility requires showing that the results improve decisions or outcomes. Table 4 summarises the current clinical interpretation of the principal markers.

6. DISCUSSION

Clinical associations involving IL-6, TNF-α, resistin, and adiponectin are marker- and compartment-dependent. The reviewed studies support a pattern in which greater adiposity and worse periodontal findings coexist with higher local inflammatory cytokine levels and an altered circulating adipokine profile. In particular, the reported pattern of higher GCF IL-6 and TNF-α levels, higher serum resistin levels, and lower serum adiponectin levels is consistent with a more inflammatory phenotype in the obese group [6]. This is a group-level association, not a universal four-marker signature or proof that periodontal inflammation causes systemic disease.

Additional observations refine this answer. Compartment-dependent findings in women with different obesity–periodontal phenotypes suggest that an oral fluid result cannot replace a circulating result [9]. The association of orosomucoid with periodontal severity, alongside the non-significant ETS–periodontitis comparison in another cohort, shows that the answer also depends on the clinical population and the marker selected [7, 8]. Collectively, the evidence supports biological and clinical heterogeneity rather than a single concentration pattern applicable to all patients with obesity.

Periodontal treatment was followed by improvements in clinical measures in the selected reports, including settings in which serum adipokine levels did not change [10–15]. This is clinically important because the failure of a serum marker to normalise should not be interpreted as the failure of otherwise successful periodontal treatment. Conversely, lower levels of circulating inflammatory markers do not demonstrate attachment recovery or resolution of residual periodontal pockets. Report clinical and laboratory outcomes alongside each other, without using one to substitute the other.

The two randomised trials reviewed provided selective rather than uniformly positive or negative biomarker findings. The MetS trial supports a six-month hsCRP benefit of the tested treatment package [16]. This study on saliva resistin did not show a biomarker effect between groups [15]. These findings supplement the available systematic reviews [3, 4] by determining the exact intervention, sample, and comparison used for each finding. These findings do not prove that all periodontal treatments cause the same systemic reaction.

In diagnosis, group-level associations warrant research into biomarkers but do not set criteria that can replace physical examination. For prognosis, the selected papers do not demonstrate that;

The four main proteins can predict future periodontal disease progression beyond clinical information already available. For monitoring, an interpretable serial change would need to exceed analytical and biological variability and correspond to a clinically meaningful outcome. The reviewed reports did not establish these requirements as a complete testing strategy [6–16].

Therefore, management must still be guided by periodontal examination and radiography, risk factors, oral hygiene, subgingival instrumentation when indicated, and supportive periodontal care [1]. In a patient whose plaque, bleeding, and pocket measurements improve while the serum adipokine remains unchanged, the clinical response should guide the periodontal reassessment; an isolated biomarker result does not establish a need to intensify treatment. In the reverse situation, a favourable laboratory result should not override persistent clinical signs of disease.

Obesity and metabolic disturbances have separate evaluations and treatments. The evidence from the available biomarkers can help in discussing the inflammatory burden among experts but should not be regarded as proof of the ability of periodontal therapy to prevent cardiovascular disease or diabetic complications. Likewise, hsCRP reduction in a regimen containing azithromycin is not an indication to prescribe antibiotics solely to modify biomarkers [1, 16, 22–24]. These conclusions concern the markers and populations reviewed here, not every possible periodontal biomarker application.

FUTURE DIRECTIONS

Future studies should specify the intended use: explaining the biology of association, diagnosis, prognosis, monitoring of treatment, or choice of treatment. Association studies need to have clear adiposity–periodontal phenotypes with proper adjustment for confounders. In the treatment studies, the analyte, sample type, primary time point, and effect need to be specified, with emphasis on between-group effect estimation as against paired analyses. The study reporting needs to include estimates of effect and uncertainty, and not just statistical significance.

Quantitative synthesis needs to discriminate between associations, uncontrolled differences, and randomisation effects. Combining data requires comparable samples, analytes, units, times, and treatment comparisons. Normalising effect sizes cannot remedy a difference in the biological measure itself. Molecular targets demand independent validation, ideally with coordinated measures at the appropriate molecular levels within the same subjects. Fig. (2) presents a graphic overview of the distinctions in study designs and the proposed path from discovery to proving clinical utility.