Relationship between triglyceride-glucose index and stroke risk in middle-aged and older Chinese population: a national longitudinal study based on explainable machine learning
Authors' conclusion
This research employs explainable machine learning to delineate factors that contribute to stroke risk, highlighting how the TyG index's impact on stroke risk varies by age and gender. As an established surrogate marker for insulin resistance, the TyG index monitoring may play a crucial role in stroke prevention and management strategies.
Study context
- Population
- Middle-aged and older adults without history of stroke at baseline, China Health and Retirement Longitudinal Study (CHARLS)
- Participants
- 6,863 participants
- Exposure
- Quartiles of baseline triglyceride-glucose (TyG) index
- Outcome
- Incident first stroke
- Follow-up
- 9-year follow-up
Methodological detailsnational longitudinal, explainable machine learning, Least Absolute Shrinkage and Selection Operator (Lasso) Cox regression, multivariate Cox proportional hazards model
- National longitudinal
- Explainable machine learning
- Lasso Cox regression
- Multivariate Cox proportional hazards model
Main finding
As reported in the abstract
Stroke risk
Adjusted hazard ratio (95 % CI)P value
Q2vs Q1
1.45(1.16–1.82)
P < 0.01Q3vs Q1
1.64(1.31–2.04)
P < 0.01Q4vs Q1
1.36(1.08–1.72)
P < 0.01Observational data: associations, not effects.
Further findings
Incidence of first stroke (all participants) · 9-year follow-up
- Participants with a first stroke
- 787 (11.5 %)
Subgroups by age and gender
- Associations consistent across all subgroups except age <60 years
Predictor importance (explainable machine learning)
- Age the most significant predictor of stroke risk; TyG index a relatively important factor
Original publication
Relationship between triglyceride-glucose index and stroke risk in middle-aged and older Chinese population: a national longitudinal study based on explainable machine learning. Diabetes Res Clin Pract. Published 7 August 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →