The clinical logic behind the search is compelling. Gestational diabetes mellitus, or GDM, is one of the most common metabolic complications of pregnancy, and current screening depends on venous blood sampling, most notably the oral glucose tolerance test, which requires fasting, repeated blood draws over several hours, and a glucose load that many women find unpleasant. Timely detection matters because untreated hyperglycemia in pregnancy raises risks of macrosomia, birth injury, and later metabolic disease for both mother and child. Saliva offers an attractive alternative sampling matrix: it can be collected non-invasively, repeatedly, and even at home, which would allow far more frequent monitoring than blood-based testing. The catch, as the review makes clear, is that saliva is a complex fluid whose composition reflects not only systemic metabolism but also oral health, hydration, circadian rhythm, and collection technique, all of which can confound comparisons between patient groups.

The most consistent findings emerged for salivary glucose itself. Both studies included in the review that measured glucose in saliva found higher postprandial or post-load values in women with GDM compared with normoglycemic controls, and both reported positive correlations between blood and saliva glucose concentrations, supporting the physiological premise that salivary glucose tracks circulating glucose. One study reported an area under the receiver operating characteristic curve, or AUC, of 0.865 (95% confidence interval 0.765 to 0.966) for two-hour postprandial unstimulated saliva, a discriminatory performance that, if replicated, would approach the range considered useful for clinical screening. The technical appeal is obvious: a marker that mirrors blood glucose but can be sampled without needles would transform the logistics of glucose monitoring during pregnancy, particularly for repeated postprandial assessment.

Oxidative stress markers formed the second major domain of evidence. Hyperglycemia generates reactive oxygen species, and malondialdehyde, or MDA, a lipid peroxidation product that serves as a readout of oxidative damage, emerged as the leading candidate in this category. One study reported an AUC of 0.812 (95% confidence interval 0.694 to 0.929) for salivary MDA, though the review notes that oxidative stress findings were not uniform across studies, with other measures of antioxidant capacity and oxidant status showing more variable results. The biological rationale is sound, since pregnancy already imposes oxidative load and hyperglycemia amplifies it, but the heterogeneity of assays, from ferric reducing ability of plasma to oxygen radical absorbance capacity, makes cross-study comparison difficult and underscores the need for standardized methodology.

Adipokines, the hormone-like signaling molecules secreted by fat tissue, provided a third and more modest signal. Selected adipokines showed study-specific discriminatory performance, with reported AUCs of 0.705 (95% confidence interval 0.574 to 0.836) for visfatin and 0.641 (95% confidence interval 0.520 to 0.761) for resistin. These values fall into the range that might support a screening aid but would not suffice as stand-alone diagnostic tests. Adipokines are biologically plausible candidates because insulin resistance is central to GDM pathophysiology and adipose-derived signaling is altered in metabolic disease, yet the review found that individual studies often examined different panels of molecules, limiting the ability to build a coherent picture across the literature.

Inflammatory markers told a more complicated story, one that illustrates why the oral environment cannot be ignored. Inflammatory findings varied with periodontal context, meaning that gum disease, which is itself common in pregnancy and independently alters salivary inflammatory profiles, appears to act as a confounder for markers such as interleukins, matrix metalloproteinases, and related mediators. A woman with periodontitis may show salivary inflammation patterns that mimic or mask those attributable to hyperglycemia. The review’s authors highlight this as a critical methodological gap: few studies adequately assessed or controlled for periodontal status, which means that some reported differences between GDM and control groups could reflect oral health rather than glucose metabolism. Future studies, they argue, must document and adjust for periodontal conditions as a matter of course.

Perhaps the most forward-looking portion of the review concerns the oral microbiome. Rather than measuring single molecules, some studies examined the composition of the salivary microbial community, identifying taxon-specific differences between hyperglycemic and normoglycemic pregnancies. One study went further and internally validated saliva-only classification models, machine-learning style classifiers built from microbial profiles, while another examined early-pregnancy microbial patterns in a separate cohort, probing whether the microbiome might flag risk before overt hyperglycemia develops. The idea that a bacterial fingerprint in spit could predict metabolic disease is undeniably viral in its appeal, but the review tempers the enthusiasm: internal validation is not external validation, and models trained on one cohort frequently lose performance when applied to another population with different demographics, diets, and oral hygiene practices.

Methodological caveats run throughout the review and shape its conclusions. Most of the thirty-one reports used cross-sectional or case-control designs rather than prospective cohorts, and several arose from the same or potentially overlapping study populations, which raises the possibility of double-counting evidence. Critical appraisal using design-specific JBI tools revealed limited control of confounding in many studies, and the collection and processing of saliva varied widely, from stimulated to unstimulated samples and differing timing relative to meals or glucose loads. The absence of external validation for reported diagnostic-performance estimates was identified as the central constraint on interpretation. In plain terms, an AUC reported in a single study tells you how well a marker performed in that dataset, not how it will perform in your clinic tomorrow.

What the review does establish is a roadmap. The findings for salivary glucose, MDA, and selected adipokines support further investigation in hyperglycemia in pregnancy, particularly GDM, but the authors are careful to state that their diagnostic utility remains unestablished. They call for prospective studies with standardized collection and assay methods, appropriate consideration of periodontal status, and validation in independent cohorts, the sequence of steps that any biomarker must clear before regulatory approval and clinical adoption. Multi-center prospective designs would also allow researchers to test whether combining markers, for example glucose with MDA or microbial features, improves discrimination beyond any single analyte, an approach that has proven fruitful in other diagnostic fields.

For pregnant women and their clinicians, the practical takeaway is patience. The needle-based oral glucose tolerance test remains the standard of care, and no salivary test currently exists that can replace or supplement it. But the direction of travel is clear: the review demonstrates a growing, if uneven, evidence base and identifies exactly which molecules and study designs deserve investment. If salivary glucose or MDA can replicate their reported performance in large, well-controlled prospective cohorts, the familiar scenario of a pregnant woman sitting in a clinic for hours after a sugary drink, having blood drawn repeatedly, could eventually give way to something as simple as spitting into a tube. That future is not here yet, but the scientific groundwork for it is now, for the first time, systematically mapped.

Subject of Research: Salivary biomarkers for detecting hyperglycemia in pregnancy and gestational diabetes mellitus

Article Title: Salivary biomarkers in hyperglycemia in pregnancy: a scoping review

Article References: Wu, Z., Huang, Y., Lyu, L., Huang, Y., Chen, Y., Hang, Q., Zhang, Q., & Ping, F. (2026). Salivary biomarkers in hyperglycemia in pregnancy: a scoping review. BMC Endocrine Disorders. https://doi.org/10.1186/s12902-026-02647-7

Image Credits: AI Generated

DOI: 10.1186/s12902-026-02647-7

Keywords: salivary biomarkers, hyperglycemia in pregnancy, gestational diabetes mellitus, salivary glucose, malondialdehyde, oxidative stress, adipokines, visfatin, resistin, oral microbiome, scoping review, non-invasive diagnostics