Repeated Triglyceride-Glucose Scores Predict Diabetes Risk in Prediabetic Adults, Study Finds

A simple calculation based on two routine blood tests may reveal which people with prediabetes are most likely to slide into full-blown diabetes, according to a new prospective cohort study drawing on data from one of China’s largest longitudinal health surveys. The research, published in BMC Endocrine Disorders, suggests that tracking a metabolic marker over time—rather than measuring it once—offers a far clearer picture of who is genuinely at risk.

The marker in question is the triglyceride-glucose index, commonly abbreviated as TyG. It is derived from fasting plasma glucose and fasting triglyceride levels, both of which are inexpensive and widely available in standard blood panels. The index serves as a practical surrogate for insulin resistance, the underlying metabolic dysfunction in which the body’s tissues respond poorly to insulin and glucose regulation gradually deteriorates. Because insulin resistance precedes type 2 diabetes by years or even decades, researchers have long sought accessible ways to quantify it without resorting to costly clamp studies or specialized assays.

What makes the new study distinctive is its treatment of time. A single TyG measurement captures insulin resistance at one moment, but metabolic health is dynamic: people’s glucose and lipid levels fluctuate with diet, weight, illness, and aging. To account for this, the research team, led by Fang Duan and Yujian Fu of the People’s Hospital of Anji in Zhejiang Province, calculated a cumulative average TyG index, or CumAvgTyG. This was defined as the arithmetic mean of fasting TyG values measured at two separate waves of the survey, in 2011–2012 and again in 2015. The approach essentially averages a person’s metabolic burden across a multi-year window, smoothing out short-term noise and capturing sustained insulin resistance rather than a transient spike.

The data came from the China Health and Retirement Longitudinal Study, known as CHARLS, a nationally representative survey of middle-aged and older Chinese adults. The investigators identified participants with prediabetes at the 2011–2012 baseline wave. Prediabetes—blood glucose levels elevated above normal but below the diabetic threshold—is a critical clinical juncture: many people with the condition progress to diabetes, but others remain stable or even revert to normal glucose metabolism. Distinguishing between these trajectories has been a persistent challenge, because conventional single-point measures offer limited predictive power.

From the original prediabetic cohort, the researchers excluded anyone who already had diabetes by 2015, as well as participants whose diabetes-free status in 2015 could not be reliably ascertained. The final analytic sample comprised 1,958 adults with a mean age of 58.5 years, just over half of whom were women. The outcome of interest was clinically recognized diabetes by 2018, defined as a participant-reported physician diagnosis of diabetes and/or the use of glucose-lowering medication or insulin. This definition captures diabetes that has been detected and treated in real-world clinical practice, rather than diabetes identified solely through research screening.

The results were striking. Over the follow-up period, 88 participants—4.5 percent of the cohort—developed clinically recognized diabetes. In a logistic regression model fully adjusted for age, sex, body mass index, smoking status, drinking status, and baseline glycated hemoglobin (HbA1c), each one-standard-deviation increase in CumAvgTyG was associated with 50 percent higher odds of developing clinically recognized diabetes (odds ratio 1.50, 95 percent confidence interval 1.21 to 1.86, P less than 0.001). The adjustment for baseline HbA1c is particularly important, because it means the association held even after accounting for how elevated participants’ blood sugar already was at the start of the study.

When the researchers divided participants into tertiles—three groups based on their CumAvgTyG values—the contrast between the extremes was even more pronounced. Those in the highest tertile had nearly three times the odds of clinically recognized diabetes compared with those in the lowest tertile (odds ratio 2.89, 95 percent confidence interval 1.56 to 5.35, P less than 0.001), and the trend across tertiles was statistically significant. Restricted cubic spline analysis, a flexible statistical technique for examining dose-response relationships, confirmed an overall association (P for overall association 0.001) while finding no evidence of nonlinearity (P for nonlinearity 0.850). In plain terms, the relationship between cumulative TyG and diabetes risk appears to rise steadily across the range of values, without a threshold effect or a plateau—every increment in sustained insulin resistance carries additional risk.

The findings carry practical implications for clinical practice and public health. TyG requires no equipment beyond a standard fasting blood draw and a calculator, making it feasible even in resource-limited settings where sophisticated insulin assays are unavailable. The study’s message is that serial measurement matters: a person whose TyG remains persistently high across several years is in a fundamentally different risk category than someone whose value is high at one visit and normal at the next. For the vast population of people living with prediabetes—estimated in the hundreds of millions worldwide—repeated TyG tracking could help clinicians decide who needs the most intensive lifestyle intervention, closer monitoring, or earlier pharmacological consideration.

Several caveats deserve attention. The outcome was clinically recognized diabetes, meaning the study captures diagnosed and treated disease; some participants may have developed undiagnosed diabetes that would not have been counted. The outcome relied on self-reported physician diagnosis or medication use, which introduces the possibility of recall or reporting error. Residual confounding cannot be excluded in any observational study, even one with careful statistical adjustment, and the findings derive from a Chinese cohort of middle-aged and older adults, so generalizability to other populations and age groups requires further study. The study was supported by the Medical and Health Science Program of Zhejiang Province and the Zhejiang Province Traditional Chinese Medicine Science and Technology Project, with the funding bodies having no role in the design, analysis, or publication decisions.

Even with those limitations, the study adds to a growing body of evidence that cumulative exposure measures outperform single snapshots in metabolic epidemiology. The same cumulative-average logic has been applied to blood pressure, cholesterol, and body mass index, and the present findings extend it to insulin resistance surrogates in a prediabetic population. For researchers, the work underscores the value of longitudinal cohorts like CHARLS, which make such analyses possible. For clinicians and patients alike, the takeaway is deceptively simple: in prediabetes, what matters most may not be any single blood test result, but the sustained metabolic trajectory it reflects. Two cheap measurements taken years apart, averaged together, may flag the people who need help before diabetes takes hold.

Subject of Research: Cumulative triglyceride-glucose index and diabetes risk in adults with prediabetes

Article Title: Association between CumAvgTyG and clinically recognized diabetes among middle-aged and older adults with prediabetes: a prospective cohort study from CHARLS

Article References: Duan, F., Fu, Y., Pan, S., Gao, F., Zheng, X., & Zhao, W. (2026). Association between CumAvgTyG and clinically recognized diabetes among middle-aged and older adults with prediabetes: a prospective cohort study from CHARLS. BMC Endocrine Disorders. https://doi.org/10.1186/s12902-026-02629-9

Image Credits: AI Generated

DOI: 10.1186/s12902-026-02629-9

Keywords: triglyceride-glucose index, prediabetes, type 2 diabetes, insulin resistance, CHARLS, prospective cohort, HbA1c, metabolic syndrome, epidemiology, endocrinology, blood glucose, China