New Blood Test Index Flags Prediabetes Risk in Adults With Normal Lab Results

Millions of adults walk out of their annual health checks each year with a clean bill of metabolic health. Their fasting glucose sits comfortably below the diabetic threshold, their cholesterol panel reads normal, and their doctor sends them home with routine reassurance. But a new study published in BMC Endocrine Disorders suggests that some of these apparently healthy people may already be on a trajectory toward prediabetes, and a simple composite number calculated from four routine measurements could reveal who is most at risk.

The research, led by Huibao Lyu, Yiting Zhang, and colleagues at Guangdong Medical University and Jieyang People’s Hospital in China, focused on the triglyceride-high-density cholesterol-glucose-body index, abbreviated as TyHGB. This index weaves together four parameters that virtually every clinic already measures: triglycerides, high-density lipoprotein cholesterol, fasting plasma glucose, and body mass index. Rather than viewing each value in isolation, the index integrates them into a single score that captures the intertwined disturbances in blood fat handling, sugar metabolism, and body composition that precede overt metabolic disease.

The study team set out to answer a question that had never been formally addressed: does a higher TyHGB score predict the future development of impaired fasting glucose, the critical prediabetic stage defined by fasting glucose between 5.6 and 6.9 millimoles per liter, in adults whose baseline glucose and lipid profiles are entirely normal? To find out, the researchers conducted a secondary retrospective analysis of a multicenter cohort, drawing on a fully anonymized, open-access dataset from the Dryad repository originally collected under ethical approval from the Rich Healthcare Group Ethics Review Board.

The scale of the analysis is striking. The investigators included 57,097 Chinese adults who were initially free of impaired fasting glucose, all of whom had baseline fasting plasma glucose below 5.6 millimoles per liter and lipid profiles below dyslipidemia thresholds. In other words, every participant started the study in the metabolic safe zone by conventional standards. The researchers calculated each person’s baseline TyHGB index from triglycerides, high-density lipoprotein cholesterol, fasting glucose, and body mass index, then tracked who developed impaired fasting glucose over time.

During a median follow-up of 2.97 years, 5,491 participants, or 9.62 percent of the cohort, developed incident impaired fasting glucose. That means nearly one in ten adults with normal baseline labs crossed into the prediabetic range within roughly three years. The central finding was unambiguous: a higher TyHGB index was significantly associated with an elevated risk of this transition. In the fully adjusted Cox proportional hazards model, each 1-unit increase in the TyHGB index carried a hazard ratio of 2.41, with a 95 percent confidence interval of 2.30 to 2.52, meaning the risk more than doubled with every unit of the index.

The quartile analysis made the gradient even more vivid. Compared with participants in the lowest quartile of the TyHGB index, those in the highest quartile faced a hazard ratio of 5.85, with a 95 percent confidence interval of 5.26 to 6.51. Adults in the top quarter of the index were nearly six times more likely to develop impaired fasting glucose than those in the bottom quarter, despite all of them beginning the study with glucose and lipid values that no clinician would flag as abnormal. The association held across prespecified subgroups, although the effect sizes varied across several clinical strata, and the results remained robust in sensitivity analyses designed to test the stability of the primary finding.

One of the most technically interesting aspects of the study involved the shape of the relationship. The researchers employed restricted cubic splines, a flexible modeling technique that allows the association between an exposure and an outcome to bend rather than forcing it into a straight line. The spline analysis revealed a significant nonlinear, piecewise relationship, with the test for nonlinearity yielding a p value below 0.001. The curve showed a threshold inflection point at a TyHGB index of 6.97. Below this value, the association between the index and impaired fasting glucose risk was comparatively gentle; above it, the risk climbed steeply. Piecewise models confirmed this threshold effect, suggesting that the index may behave less like a linear dial and more like a tripwire that marks a point of accelerating metabolic deterioration.

The biological logic behind the TyHGB index rests on the concept of insulin resistance, the metabolic state in which tissues respond sluggishly to insulin and the body compensates by pumping out more of the hormone. Elevated triglycerides combined with reduced high-density lipoprotein cholesterol form a classic lipid signature of insulin resistance, even when both values remain within the normal laboratory range. Fasting glucose reflects how well the insulin-glucose feedback loop is holding, and body mass index captures the adiposity that drives much of the metabolic disturbance. By multiplying these components together, the index amplifies the signal of subclinical metabolic dysfunction that any single parameter would miss. The study team also used the triglyceride-glucose index and its body mass index extension, known as TyG-BMI, as conceptual building blocks, and the TyHGB index represents a refinement that incorporates the protective high-density lipoprotein fraction as well.

The methodological rigor of the analysis deserves attention. The investigators adhered to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines, used multivariate imputation by Chained Equations to handle missing data, checked multicollinearity with variance inflation factors, and validated the proportional hazards assumption underlying the Cox models. They also ran logistic regression models alongside the survival analyses to quantify the association in odds terms, and conducted subgroup and sensitivity analyses to evaluate the robustness of the primary associations. The study was exempt from further institutional review board approval because it used a fully anonymized public dataset, and the authors declared no competing interests and no use of generative artificial intelligence in preparing the manuscript.

The authors are careful about what their findings do and do not establish. A higher TyHGB index was associated with incident impaired fasting glucose among Chinese adults with normal baseline fasting glucose and lipid levels, but whether the index provides additional clinical value beyond conventional risk factors requires prospective validation and formal predictive-performance assessment, including receiver operating characteristic analysis of the area under the curve. The study population was Chinese, the follow-up was under three years, and the design was observational, so causation cannot be inferred. Still, the implications are compelling for a health system that routinely measures all four components of the index at negligible cost. If future prospective studies confirm the threshold behavior observed here, the TyHGB index could become a routine early-warning score, identifying adults whose normal lab reports conceal an accelerating metabolic risk long before their fasting glucose crosses the prediabetic line.

Subject of Research: Metabolic Disease

Article Title: Triglyceride-high-density cholesterol-glucose-body index and risk of impaired fasting glucose in normolipidemic adults: a secondary analysis of a multicenter cohort

Article References: Lyu, H., Zhang, Y., Lu, Y., Lin, R., Luo, Z., Li, S., Zhang, H., Yang, D., He, Q., & Zhang, L. (2026). Triglyceride-high-density cholesterol-glucose-body index and risk of impaired fasting glucose in normolipidemic adults: a secondary analysis of a multicenter cohort. BMC Endocrine Disorders. https://doi.org/10.1186/s12902-026-02667-3

Image Credits: AI Generated

DOI: 10.1186/s12902-026-02667-3

Keywords: TyHGB index, impaired fasting glucose, prediabetes, insulin resistance, triglycerides, HDL cholesterol, fasting plasma glucose, body mass index, cohort study, metabolic syndrome, normolipidemic adults, risk prediction