Inflammation Adds Little to Diabetes-Linked Heart Risk Score, Chinese Cohort Study Finds

For millions of people living with prediabetes or diabetes, the question of which laboratory numbers truly predict a future heart attack or stroke has become one of the most consequential debates in preventive cardiology. A new analysis of nearly 3,000 middle-aged and elderly Chinese adults with abnormal glucose metabolism now offers a strikingly nuanced answer: a much-discussed composite index that combines inflammation, blood fats, and blood sugar does indeed track cardiovascular risk, but a far simpler measure built from just two routine blood tests performs just as well. The finding, published in BMC Endocrine Disorders, challenges the assumption that layering inflammatory markers onto metabolic scores automatically buys better prediction, and it points clinicians back toward triglycerides and glucose as the core signals worth watching.

The study, led by Yan Yang and Danping Zhu of Chongqing Traditional Chinese Medicine Hospital together with Chunguang Xie of the Hospital of Chengdu University of Traditional Chinese Medicine, drew its participants from the China Health and Retirement Longitudinal Study, better known as CHARLS. This nationally representative survey follows community-dwelling Chinese adults over the age of 45 with detailed questionnaires, physical examinations, and blood biomarkers. From the CHARLS cohort, the researchers identified 2,980 individuals with impaired glucose metabolism, a category that encompasses prediabetes and diabetes, and tracked them for a median of three years. During that window, 440 participants experienced a first cardiovascular event, and 182 of those events were strokes, giving the team a meaningful number of outcomes with which to test their statistical models.

At the heart of the paper is a quantity called the cumulative C-reactive protein–triglyceride–glucose index, abbreviated cumCTI. To understand it, one must first unpack its parent measure, the triglyceride–glucose index, or TyG. The TyG index is calculated as the natural logarithm of fasting triglycerides multiplied by fasting glucose, divided by two. Because elevated triglycerides and elevated glucose are the biochemical hallmarks of insulin resistance, TyG has emerged over the past decade as a cheap, widely available proxy for how resistant a person’s tissues have become to insulin, without requiring the cumbersome insulin assays or glucose clamps that direct measurement demands. C-reactive protein, meanwhile, is a liver-produced protein that surges during systemic inflammation and has long been studied as a harbinger of cardiovascular disease.

The CTI fuses these concepts by multiplying TyG by the natural logarithm of C-reactive protein, on the theory that insulin resistance and inflammation act in concert to damage blood vessels. The cumulative version goes a step further: rather than taking a single snapshot, the researchers averaged CTI values measured at Wave 1 and Wave 3 of CHARLS, capturing repeated exposure over time. This cumulative approach matters biologically, because atherosclerosis is not driven by a single bad lab result but by years of sustained metabolic insult. A single measurement can be distorted by acute illness, recent meals, or laboratory variation, whereas a cumulative index smooths those fluctuations and better reflects a person’s long-running internal environment.

The results were consistent in direction but revealing in their details. In Cox proportional hazards models that sequentially adjusted for age, sex, and a battery of cardiovascular risk factors, each one-unit increase in cumCTI was associated with a 4.3 percent higher risk of overall cardiovascular disease and, more strikingly, a 9.0 percent higher risk of stroke. Both relationships followed a linear dose–response pattern, confirmed with restricted cubic splines, meaning risk rose steadily with the index rather than jumping only beyond some threshold. The association with stroke was stronger than the association with cardiovascular disease as a whole, a hint that the metabolic and inflammatory processes captured by the index may be particularly pertinent to cerebrovascular disease.

Yet the study’s most provocative findings came from its forensic dissection of what the index actually contains. Using weighted quantile sum regression, a technique designed to apportion the contribution of each component within a mixture, the researchers found that triglycerides carried 72.70 percent of the predictive weight for stroke and a full 80.78 percent for cardiovascular disease. C-reactive protein, the very ingredient added to create the fancier index, contributed a mere 0.20 percent to stroke prediction and only 7.81 percent to cardiovascular disease prediction. In other words, the incremental value of adding inflammation to the metabolic mix was nearly negligible, and the entire signal of the composite index was essentially riding on triglycerides.

The researchers also probed mechanism. Causal mediation analysis suggested that mean arterial pressure accounted for 15.6 percent of the effect of cumCTI on stroke and 12.5 percent of its effect on cardiovascular disease. This implies that part of the reason a high metabolic–inflammatory score precedes vascular events is that the same underlying dysfunction pushes blood pressure upward, which in turn damages arteries. For stroke specifically, blood pressure mediation was strongest, reinforcing the longstanding clinical axiom that hypertension is the dominant modifiable driver of cerebrovascular events and that glucose, lipids, and blood pressure should be managed as an integrated package rather than in isolation.

The statistical verdict on predictive performance, however, was the study’s most sobering conclusion. When the team compared models using Harrell’s C-index, a measure of how well a risk score discriminates between people who will and will not experience an event, the cumulative composite index scored 0.5639 while its simpler parent, the TyG index alone, scored 0.5744. A C-index of 0.5 represents chance-level discrimination, so both scores are modest predictors in absolute terms, but the crucial point is that the elaborate version did not beat the simple one. The extra blood draw, the added cost, and the added complexity of measuring C-reactive protein bought nothing. For population screening in resource-constrained settings, the authors conclude that TyG remains the more rational choice: cheaper, simpler, and no less informative.

Why does this matter beyond the statistics? Cardiovascular disease is the leading cause of death in China and much of the world, and people with impaired glucose metabolism face an elevated baseline risk that makes early identification especially valuable. If a two-component index computed from tests that cost pennies can flag higher-risk individuals, clinics can intervene earlier with lifestyle counseling, lipid-lowering therapy, and blood pressure control. Conversely, mandating inflammatory markers for everyone would inflate healthcare costs without improving triage. The study also delivers a broader methodological lesson for the biomarker literature: composite indices can sound impressive and publish well, but component-weighting analyses and head-to-head discrimination testing are essential to show whether a new score genuinely adds information or merely repackages the same signal under a longer name.

The findings come with caveats that the researchers themselves acknowledge implicitly. The follow-up period was short, with a median of three years, and both discrimination statistics were weak in absolute terms, so neither index is ready to serve as a standalone decision tool. The cohort comprised Chinese adults aged 45 and older with abnormal glucose metabolism, and whether the same component weights hold in younger populations, in other ethnic groups, or among people with normal glucose tolerance remains to be tested. Still, the paper’s central message is clear and clinically actionable. Among people with dysregulated blood sugar, sustained elevations in the triglyceride–glucose index track future cardiovascular and cerebrovascular events, partly through blood pressure, while C-reactive protein adds almost nothing to stroke prediction. Until more powerful risk models emerge, the humble TyG index, coupled with vigilant blood pressure management, appears to be the best low-cost screening strategy this line of research has produced.

Subject of Research: Prediction of cardiovascular disease and stroke risk using the cumulative C-reactive protein–triglyceride–glucose index in Chinese adults with impaired glucose metabolism

Article Title: The association between the cumulative C-reactive protein-triglyceride-glucose index and the incidence of cardiovascular disease in middle-aged and elderly Chinese individuals of abnormal glucose metabolism

Article References: Yang, Y., Xie, C., & Zhu, D. (2026). The association between the cumulative C-reactive protein-triglyceride-glucose index and the incidence of cardiovascular disease in middle-aged and elderly Chinese individuals of abnormal glucose metabolism. BMC Endocrine Disorders. https://doi.org/10.1186/s12902-026-02536-z

Image Credits: AI Generated

DOI: 10.1186/s12902-026-02536-z

Keywords: cumCTI, TyG index, C-reactive protein, triglycerides, insulin resistance, cardiovascular disease, stroke, impaired glucose metabolism, CHARLS, cohort study, risk prediction, mean arterial pressure