📄 medicine

Dyslipidemia and risk of incident stroke: the mediating role of insulin resistance assessed by the triglyceride–glucose index in the CHARLS cohort

In a cohort of Chinese adults aged 45 and older, insulin resistance assessed by the triglyceride–glucose (TyG) index was found to partially mediate the association between dyslipidemia (specifically total cholesterol and HDL-C) and incident stroke risk, suggesting TyG serves as a useful adjunct marker for metabolic risk assessment beyond traditional LDL-C measures.

Original authors: Yiming Huang, Xi Chen, Yujia Yan, Yang Zhang, Ming Li, Xunming Ji

Published 2026-07-14

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Yiming Huang, Xi Chen, Yujia Yan, Yang Zhang, Ming Li, Xunming Ji

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine your body is a bustling city, and your blood vessels are the main highways keeping everything running smoothly. For a long time, doctors thought the biggest traffic jam on these highways was caused by one specific villain: a sticky, cholesterol-heavy truck called LDL-C. The standard advice was simple: "Keep the LDL trucks small, and the roads will stay clear."

But in this new study, researchers from China decided to look closer at the city's traffic patterns. They used data from over 9,000 adults aged 45 and older, tracking them for nearly a decade to see who ended up having a stroke (a major crash on the highway). They wanted to know: Is LDL-C the only bad guy, or are there other sneaky culprits hiding in plain sight?

The "Sugar-Fat" Detective: The TyG Index
The researchers introduced a new detective tool called the TyG index. Think of this as a "metabolic smoke detector." It doesn't just look at fat; it measures how well your body handles both sugar (glucose) and fat (triglycerides) together. A high TyG score suggests your body's engine is struggling to process fuel efficiently—a state known as insulin resistance. It's like a city where the power grid is flickering, causing traffic lights to malfunction and fuel to pile up in the streets.

The Big Discovery: Two Different Types of Trouble
The study found that the "traffic jams" leading to strokes actually come from two very different sources, and they need different solutions.

The LDL-C Truck: The study confirmed that high levels of LDL-C (specifically ≥ 130 mg/dL) are indeed a risk factor. However, the researchers found that the TyG "smoke detector" didn't explain why this truck causes crashes. It seems the LDL-C truck causes problems on its own, independent of the body's sugar-fat processing issues.
The Sugar-Fat Pile-Up: Here is where it gets interesting. The study found that high Triglycerides (≥ 200 mg/dL) and high Total Cholesterol (≥ 240 mg/dL) are strongly linked to strokes, but only because they are often part of that "sugar-fat processing" failure. The TyG index acted as a bridge, showing that when your body struggles with insulin resistance, it creates a messy mix of high sugar and high fat that clogs the arteries. In fact, the TyG index was a slightly better predictor of a future crash than looking at any single lipid number alone.

The "150 mg/dL" Speed Bump
One of the most vivid findings was about Triglycerides. The researchers discovered that the risk doesn't rise in a straight, boring line. Instead, it's like a hill with a sudden, steep drop-off.

Below 150 mg/dL: The risk stays relatively flat. Your body can handle these levels without much trouble.
Above 150 mg/dL: The risk of a stroke shoots up sharply. It's as if the city's drainage system can only handle so much water; once the level passes 150 mg/dL, the streets start flooding.

What This Means for the City
The study suggests that for many people, especially in middle and older age, the real danger isn't just the LDL truck. It's the underlying "engine trouble" (insulin resistance) that causes the sugar and fat to pile up.

The researchers found that about 824 people in their study had a stroke during the follow-up. Those who did had higher TyG scores and higher triglycerides. The study suggests that if we want to keep the highways clear, we can't just focus on the LDL truck. We also need to fix the engine's fuel processing. If someone has a TyG index above 8.605 or triglycerides above 150 mg/dL, they might be at high risk even if their LDL looks "okay."

The Bottom Line
This research doesn't say LDL is safe; it just says it's not the whole story. The study suggests that the TyG index is a useful, simple tool to spot people whose bodies are struggling to manage sugar and fat, a struggle that significantly increases the chance of a stroke. While the TyG index isn't a magic bullet (the prediction accuracy was modest, with a score of 0.567 compared to 0.531–0.563 for other lipids), it offers a new way to look at the problem. It hints that to truly prevent these crashes, we might need to manage the "sugar-fat" engine trouble just as carefully as we manage the cholesterol trucks.

Technical Summary: Dyslipidemia, Insulin Resistance, and Incident Stroke in the CHARLS Cohort

Problem Statement
Despite established prevention strategies targeting low-density lipoprotein cholesterol (LDL-C), a significant residual risk for stroke persists. This residual risk is hypothesized to be linked to insulin resistance and metabolic dysregulation, which are not fully captured by traditional lipid metrics. While the triglyceride–glucose (TyG) index has emerged as a simple, cost-effective surrogate for insulin resistance, its specific role in mediating the relationship between various dyslipidemias (triglycerides, total cholesterol, HDL-C, LDL-C) and incident stroke remains to be fully elucidated, particularly in middle-aged and older Asian populations where "metabolically obese normal weight" phenotypes are common.

Methodology
This study conducted a secondary analysis of data from the China Health and Retirement Longitudinal Study (CHARLS), a nationally representative cohort of Chinese adults aged ≥45 years.

Population: The final analytic sample included 9,273 participants who had baseline fasting lipid and glucose measurements in 2011 and no history of stroke at baseline. Participants were followed through 2020.
Exposure: Dyslipidemia was defined using guideline-based cutoffs for Total Cholesterol (TC), Triglycerides (TG), LDL-C, and HDL-C. The TyG index was calculated as $ln[\text{fasting glucose (mg/dL)} \times \text{triglycerides (mg/dL)} / 2]$ .
Outcome: Incident stroke was defined as a new self-reported, physician-diagnosed stroke occurring during follow-up waves (2013, 2015, 2018, 2020).
Statistical Analysis:
- Association: Multivariable Cox proportional hazards models were used to estimate Hazard Ratios (HRs) with three levels of adjustment (demographics; lifestyle factors; and comorbidities).
- Non-linearity: Restricted cubic splines (RCS) were employed to examine dose–response relationships.
- Mediation: Structural Equation Modeling (SEM) with 5,000 bootstrap resamples was used to test whether the TyG index statistically mediated the associations between lipid parameters and stroke.
- Prediction: Time-dependent Receiver Operating Characteristic (ROC) curves and Area Under the Curve (AUC) values were calculated to assess discriminative performance.

Key Results

Incidence: During the follow-up period, 824 participants (8.9%) experienced an incident stroke.
Lipid Associations: In fully adjusted models, elevated Triglycerides (≥200 mg/dL) were the most robust predictor of stroke (HR 1.40, 95% CI 1.18–1.67). High LDL-C (≥130 mg/dL) also predicted stroke (HR 1.23), while low HDL-C showed a weaker association (HR 1.17). The association for high Total Cholesterol (≥240 mg/dL) was attenuated after full adjustment (HR 1.20, 95% CI 0.98–1.46).
Non-Linear Thresholds: A significant non-linear relationship was observed for Triglycerides, where stroke risk remained stable below approximately 150 mg/dL but escalated steeply above this threshold. TC and LDL-C showed predominantly linear associations, while HDL-C demonstrated an inverse linear association.
Mediation Analysis: The TyG index statistically accounted for a portion of the association between Total Cholesterol and stroke, and partially mediated the association between HDL-C and stroke. In contrast, the association between LDL-C and stroke was largely independent of the TyG index, suggesting distinct pathogenic pathways.
Predictive Value: The TyG index demonstrated slightly better discriminative ability for incident stroke (AUC 0.567) compared to individual lipid markers (LDL-C AUC 0.531; TC AUC 0.537; HDL-C AUC 0.541; TG AUC 0.563). The combined model of all lipids plus TyG achieved an AUC of 0.577.

Significance and Claims
The authors conclude that in middle-aged and older Chinese adults, insulin resistance (proxied by the TyG index) represents a distinct pathway that statistically mediates the stroke risk associated with elevated Total Cholesterol and Triglycerides, but not LDL-C. This supports the "common soil" hypothesis, suggesting that hypertriglyceridemia and elevated TC in this population are often downstream consequences of systemic metabolic dysfunction rather than isolated lipid abnormalities.

The study highlights a specific, actionable threshold for Triglycerides (150 mg/dL), below which stroke risk does not significantly increase, and above which risk rises sharply. Furthermore, the findings suggest that the TyG index may serve as a useful adjunct marker for metabolic risk assessment, particularly for identifying high-risk individuals who might be missed by traditional BMI or LDL-C-centric screening, such as those with the "metabolically obese normal weight" phenotype.

The authors remain modest regarding the clinical utility, noting that while the TyG index offers better discrimination than individual lipids, the overall predictive accuracy (AUC) for stroke remains modest. They emphasize that these findings support a dual-target approach to stroke prevention: managing LDL-C while simultaneously addressing insulin resistance and metabolic dysregulation. The study acknowledges limitations, including the self-reported nature of stroke events, the lack of stroke subtype data, and the single-time-point measurement of exposures, which precludes definitive causal inference regarding mediation.

Technical Summary: Dyslipidemia, Insulin Resistance, and Incident Stroke in the CHARLS Cohort

More like this