Salivary Glucose Testing for Diabetes Mellitus: A Systematic Review and Meta-Analysis of Current Evidence and Methodological Heterogeneity

TL;DR

This study reviews current research on using saliva to measure glucose levels as a non-invasive alternative to blood tests for diabetes, finding weak and inconsistent results.

Contribution

The study provides a systematic review and meta-analysis of salivary glucose testing, highlighting methodological heterogeneity and weak correlations.

Findings

Whole-mouth saliva shows a weak correlation with blood glucose (r2 = 0.05).

Parotid saliva has a slightly stronger correlation (r2 = 0.11) but still weak.

High heterogeneity (I2 near 100%) limits the reliability of pooled estimates.

Abstract

Background/Objectives: As current methods of measuring blood glucose levels are inconvenient and painful for patients, using salivary glucose as a non-invasive biomarker to estimate glucose levels may improve patient compliance. This study aims to quantify the association between salivary glucose levels and blood glucose levels and to assess how heterogeneity between studies and the methodological differences affect the potential clinical use of salivary glucose. Methods: The PRISMA guidelines were used for this review, and the protocol was registered on PROSPERO (CRD42023471213). Four databases were searched: PubMed, EMBASE, Cochrane’s Library Trials, and Web of Science. The search was conducted on 22 November 2023 and updated on 18 August 2025. No filters were applied for the search. Human studies, where paired salivary and blood samples were taken both from patients with diabetes and healthy individuals after at least 8 h of fasting, were included in the analysis. We extracted correlation coefficients and group means differences. Risk of bias was assessed with QUADAS-2, and the between-study heterogeneity was examined using random-effects models. Results: Through the systematic search, 15,162 articles were found, 25 of which were included in our meta-analysis. The analysis showed a weak correlation between whole-mouth saliva samples and blood samples (r2 value: 0.05) and a slightly stronger correlation between parotid saliva samples and blood samples (r2 value: 0.11). These low r2 values reflect weak associations and are reported descriptively. The difference between the mean salivary glucose level of patients with diabetes and controls was 4.43 mg/dL (95% CI: 2.05; 6.80). The high heterogeneity (I2 values approaching 100% for mean difference analyses) limits the interpretability of pooled estimates. Conclusions: Current evidence indicates that salivary and blood glucose levels associate weakly, and the study results are highly heterogeneous. Given the weak and highly heterogeneous associations observed across studies, current evidence does not support the use of salivary glucose measurement as an alternative to blood glucose measurement for the time being. Further standardized research is required before any conclusion about clinical applicability can be drawn.

Figures