Association of Salivary α₂-Macroglobulin Levels with Glycaemic Status and Diabetic Retinopathy in Patients with Type 2 Diabetes Mellitus

2.1. Study Design and Ethical Approval

This analytical cross-sectional study was conducted over ten months from January 2025 to October 2025. The study followed methodological standards recommended by the STROBE guidelines [11]. Ethical approval was obtained from the Institutional Ethics Committee, and written informed consent was obtained from all participants prior to enrolment; study procedures adhered to the principles of the Declaration of Helsinki [12].

2.2. Participants and Grouping

A total of 165 individuals were screened for eligibility. Of these, three did not meet the inclusion criteria, 17 met one or more exclusion criteria, 10 declined to participate, three provided inadequate saliva samples for analysis, and two had incomplete clinical or laboratory data. Consequently, 130 participants were enrolled in the study, comprising 52 participants with uncontrolled T2DM, 52 with controlled T2DM, and 26 healthy controls. Diabetes was defined and stratified according to the 2024 American Diabetes Association Standards of Medical Care [13]. Healthy controls had no history of diabetes and an HbA1c < 5.7%. Controlled diabetes was defined as HbA1c ≤ 7%, and uncontrolled diabetes as HbA1c > 7%.

Inclusion criteria comprised adults aged ≥18 years who provided informed consent and met predefined study-group criteria, including non-diabetic controls or individuals with established type 2 diabetes classified according to HbA1c status. Exclusion criteria included acute infections, chronic inflammatory or autoimmune diseases, hepatic or renal disease (including clinically diagnosed diabetic nephropathy based on medical history and serum creatinine), pregnancy, smoking, alcohol consumption, salivary gland disorders, xerostomia-inducing medications, recent dental procedures, and active oral infectious conditions. Participants with healthy gingiva or clinically stable periodontal conditions without evidence of acute oral infection were eligible for inclusion. Periodontal status was subsequently assessed using a CPITN probe and recorded for all dentulous participants in accordance with established recommendations for salivary biomarker research [14]. Healthy controls were age-matched individuals without a history of diabetes or other systemic diseases. Eligibility was confirmed through medical history, clinical assessment, HbA1c measurement, and application of the same exclusion criteria used for the diabetic groups, including the absence of acute infections, chronic inflammatory or autoimmune diseases, hepatic or renal disease, pregnancy, smoking, alcohol consumption, salivary gland disorders, xerostomia-inducing medications, recent dental procedures, or other conditions known to influence salivary biomarkers.

Previous studies have reported moderate-to-strong correlations between salivary α₂-macroglobulin and glycaemic markers, supporting the assumption of a moderate effect size for sample size estimation. Owing to heterogeneity in study design and reporting, including the absence of consistently reported means and standard deviations, a conventional effect size could not be calculated directly from the published literature. Therefore, the sample size estimation was based on an assumed moderate effect size derived from the available evidence. Using a two-sided significance level of 5% and a statistical power of 80%, the required sample size was calculated using IBM SPSS Statistics. Accordingly, a total of 130 participants were included in the study.

2.3. Clinical and Biochemical Measurements

Demographic and clinical information, including age, sex, Body Mass Index (BMI), duration of diabetes, medication use, and comorbidities, were recorded. After an overnight fast of at least ten hours, venous blood samples were collected to measure fasting and postprandial glucose, HbA1c, total cholesterol, triglycerides, High-Density Lipoprotein Cholesterol (HDL-C), Low-Density Lipoprotein Cholesterol (LDL-C), fasting insulin, and serum creatinine. Glucose, lipid parameters, hs-CRP, and fasting insulin were measured using an automated clinical chemistry analyser (cobas c 111; Roche Diagnostics GmbH, Mannheim, Germany). HbA1c was measured using the Merilyzer GluQuant A1c™ HPLC analyser (Meril Diagnostics Pvt. Ltd., Vapi, Gujarat, India), which is based on ion-exchange chromatography. Serum creatinine was measured using the modified Jaffe kinetic method on an automated biochemical analyser (Beacon Diagnostics Pvt. Ltd., Navsari, Gujarat, India).

Indices of insulin resistance and β-cell function were calculated using the Homeostatic Model Assessment (HOMA) described by Wallace et al. [15]. Homeostatic Model Assessment of Insulin Resistance (HOMA-IR) and Homeostatic Model Assessment of β-cell function (HOMA-β) (%) were calculated using the HOMA2 calculator (Version 2.2.3; Diabetes Trials Unit, University of Oxford), which remains a reliable surrogate method for assessing insulin dynamics in metabolic research.

2.4. Saliva Collection and Processing

Unstimulated whole saliva was collected between 8:00 AM and 10:00 AM to minimise diurnal variation. Participants refrained from food, beverages, caffeine, smoking, and oral hygiene procedures for at least 90 minutes before sampling. While seated upright, saliva was allowed to accumulate passively and was collected by drooling into sterile polypropylene tubes, following the protocol of Navazesh and Kumar [16].

Samples were immediately placed on ice, centrifuged at 3000 rpm for 10 minutes, aliquoted to prevent repeated freeze–thaw cycles, and stored at –80°C until analysis. Sample handling followed current recommendations for salivary proteomic research [14].

2.5. Estimation of Salivary α₂MG

Salivary α₂MG concentrations were measured using a Quantitative Enzyme-Linked Immunosorbent Assay (ELISA) kit (EM1115-1; AssayMax™ Human Alpha-2-Macroglobulin ELISA Kit, Cat. No. EM1115-1, Assaypro LLC, St. Charles, MO, USA) according to the manufacturer's instructions [17]. Standards and samples were analysed in duplicate. Optical density was measured at 450 nm, and concentrations were calculated using a standard curve ranging from 0 to 2000 ng/mL. ELISA-based measurement of α₂MG is a validated method with high analytical sensitivity in biomarker studies [7].

2.6. Assessment of Diabetic Retinopathy

Participants underwent ophthalmic evaluation by a qualified ophthalmologist. Slit-lamp biomicroscopy and indirect ophthalmoscopy were performed, and DR was classified according to the International Clinical Diabetic Retinopathy Severity Scale [18].

2.7. Statistical Analysis

Statistical analyses were performed using IBM SPSS Statistics Version 26.0. Normality was assessed using the Shapiro–Wilk test. Continuous variables were expressed as mean ± standard deviation, and categorical variables as proportions. Group comparisons were conducted using one-way Analysis of Variance (ANOVA) with Tukey post hoc testing.

Associations between salivary α₂MG and metabolic variables were evaluated using Spearman’s rank correlation. Multivariate linear regression identified independent factors associated with salivary α₂MG levels. Diagnostic performance for distinguishing uncontrolled from controlled diabetes was assessed using ROC curve analysis, with Area Under the Curve (AUC), optimal cut-off value, sensitivity, specificity, positive predictive value, and negative predictive value calculated.

Univariate logistic regression evaluated associations between clinical and biochemical variables and DR. Significant variables were entered into a multivariate logistic regression model to determine the independent association of salivary α₂MG with DR after adjustment for potential confounders. A two-sided P-value < 0.05 was considered statistically significant.

4.1. Comparison with Prior Salivary α₂MG Studies in T2DM

Early investigations first established salivary α₂MG as a potential surrogate of glycaemic exposure, with Aitken et al. reporting clear correlations with HbA1c and suggesting an 840 ng/mL diagnostic threshold [7]. Subsequent work progressively strengthened this association: Chung et al. demonstrated that α₂MG not only correlated with fasting glucose and HbA1c but also mirrored short-term glycaemic shifts over a three-month period [20], while Feng et al. confirmed higher concentrations in uncontrolled diabetes, although their weaker correlations were likely influenced by early-morning sampling, highlighting the importance of methodological standardisation [21].

As analytic techniques improved and sample sizes increased, more recent studies began to delineate the biomarker gradient across glycaemic states with greater precision. Rastogi et al. reaffirmed strong HbA1c correlations [22], whereas more recent evidence from Goel et al. further demonstrated a stepwise rise in salivary α₂MG across glycaemic states, with strong correlation to HbA1c and promising diagnostic accuracy for identifying inadequate glycaemic control [23]. Sarkar et al. reached similar conclusions, reporting strong correlations and an AUC of 0.915 despite lower absolute concentrations attributable to inter-assay variability [24]. Likewise, Nsr-Allah et al. refined diagnostic thresholds by identifying a 645 ng/mL cut-off for predicting HbA1c ≥7% [25], while El-Assal et al. and El-Alfy et al. demonstrated comparable three-tier separation of glycaemic states using a different ELISA platform [26, 27]. Caixeta et al.’s meta-analysis demonstrated a robust pooled correlation between salivary α₂MG and HbA1c [10], and Le et al.’s work emphasised that α₂MG reliably complements multi-marker panels integrating inflammatory and metabolic indicators [28]. Such integrated analyses support the individual observations.

Our diagnostic performance (AUC 0.692; cut-off ≥764 ng/mL) falls within the range of published estimates and aligns well with patterns reported in earlier studies [7, 23, 25-27]. Differences in absolute thresholds appear to reflect variation in population characteristics, sampling protocols, and assay methodologies. Within-sample ROC analysis demonstrated modest discriminatory performance of salivary α₂MG for poor glycaemic control; however, these estimates should be interpreted as exploratory due to the modest sample size and lack of external validation.

4.2. Insulin Resistance and β-cell Function

To the best of our knowledge, this is the first study to relate salivary α₂MG to both HOMA-IR and HOMA-β. Its positive association with HOMA-IR and inverse association with HOMA-β indicate that α₂MG reflects broader metabolic dysfunction, not just hyperglycaemia. These observations align with the established role of adipokines, including tumour necrosis factor-α, interleukin-6, interleukin-18, leptin, and resistin, in impairing insulin signalling and promoting β-cell stress [3], implying that α₂MG up-regulation may integrate both inflammatory and metabolic pathways. Larger longitudinal studies are needed to validate this mechanistic link.

4.3. Salivary α₂MG and DR

DR represents a key microvascular marker of chronic glycaemic injury, and its greater prevalence among those with uncontrolled diabetes in our cohort is consistent with established epidemiology [2]. In this context, the marked elevation of salivary α₂MG in individuals with DR and its persistence as an independent association after multivariable adjustment provides clinically meaningful support for its microvascular relevance. Given that α₂MG is a high–molecular-weight protein whose circulating levels may be influenced by renal dysfunction, individuals with clinically diagnosed renal disease, including diabetic nephropathy, were excluded to minimise renal-related confounding.

Previous evidence in this domain has been limited, with most studies focusing on serum or type 1 diabetes: Calixto et al. showed that salivary α₂MG reflects innate immune activation in T2DM with complications [19], and Chee et al. identified α₂MG among 119 salivary proteins differentially expressed across DR stages [29]. The findings extend this body of work by demonstrating a clear and independent association between salivary α₂MG and DR in type 2 diabetes, suggesting that the marker may capture inflammatory microvascular stress beyond glycaemic burden alone.

4.4. Associations with Cand Inflammatory Markers

In the uncontrolled diabetes group, salivary α₂MG correlated negatively with systolic and diastolic blood pressures, differing from Chung et al.’s findings [20]. The reason for this discrepancy remains unclear and may reflect differences in participant characteristics or study methodology.

Patients with uncontrolled diabetes had higher triglycerides and lower HDL-C, and salivary α₂MG correlated positively with total cholesterol and triglycerides and negatively with HDL-C. This observation may reflect shared inflammatory pathways.

Hs-CRP did not differ significantly between groups, likely due to inflammatory comorbidities in some controls. The absence of a significant difference in hs-CRP between groups may reflect the influence of comorbid inflammatory conditions in some participants or the known variability of hs-CRP as a systemic inflammatory marker. This finding supports salivary α₂MG as a more diabetes-specific inflammatory marker rather than a broad systemic signal.

4.5. Oral Environment and Methodological Considerations

Diabetes alters salivary composition and flow [28]. Both diabetic groups in our study exhibited reduced salivary flow, consistent with earlier reports. Mid-morning unstimulated samples were used to minimise the reported diurnal variation. Sandwich ELISA was chosen due to superior sensitivity over older precipitation assays [30].

Although gingivitis and periodontitis were common among participants with diabetes, no significant association was observed between salivary α₂-macroglobulin levels and periodontal status in the study cohort. Together with the absence of significant differences in periodontal status across the study groups, these findings suggest that the observed associations between salivary α₂-macroglobulin and glycaemic status or diabetic retinopathy are unlikely to be explained solely by periodontal inflammation. Nevertheless, residual confounding cannot be completely excluded.

4.6. Rationale for Selecting α₂-Macroglobulin over other Salivary Biomarkers

Several salivary proteins have been investigated as non-invasive biomarkers in diabetes, including albumin, immunoglobulins, amylase, cytokines, and salivary HbA1c. However, many of these markers primarily reflect passive diffusion, transient inflammatory changes, or oral health status rather than systemic metabolic injury [24]. Salivary albumin is particularly sensitive to gingival inflammation and mucosal microbleeding, while salivary HbA1c reflects erythrocyte glycation and exists at very low concentrations in saliva, limiting its biological and analytical robustness. Conversely, α₂MG represents an acute-phase protein integrally involved in inflammatory regulation and protease inhibition, and its salivary elevation plausibly reflects diabetes-related systemic inflammation and microvascular barrier dysfunction, mechanisms that are central to the pathogenesis of DR [10].

4.7. Clinical Implications and Future Directions

Salivary α₂MG may represent a promising, non-invasive biomarker integrating glycaemic exposure, insulin resistance, β-cell dysfunction, and microvascular injury. Saliva-based assessment may be particularly valuable in resource-limited settings. Its moderately strong correlations with glycaemic markers, within-sample ROC performance, and independent association with retinopathy suggest that α₂MG may support non-invasive assessment of microvascular involvement in patients with T2DM.

The observed AUC of 0.692 indicates only modest discriminatory performance, suggesting that salivary α₂-macroglobulin should be interpreted as a complementary biomarker rather than a standalone diagnostic test. Although promising, salivary α₂MG should currently be considered an exploratory biomarker pending validation in larger multicentre cohorts. Future studies should assess longitudinal changes in α₂MG, validate predictive ability for DR progression, and evaluate integration into multi-marker algorithms. Standardised assays and cross-population validation are essential, as emphasised by the relevant systematic reviews [10, 28].