The trendy TyG index adds almost nothing to standard heart risk scores

The triglyceride-glucose index, or TyG, has become a fashionable marker of insulin resistance because it can be calculated from routine blood tests. A stream of studies has linked higher TyG values to cardiovascular disease, prompting suggestions that it should join, or even improve on, established risk tools such as the Framingham Risk Score. The practical question is not whether TyG is associated with risk, but whether adding it to a standard risk score actually sharpens prediction, which is what this study set out to quantify.

The researchers analysed adults aged 45 to 80 without cardiovascular disease at baseline from the China Health and Nutrition Survey, counting new diagnoses of myocardial infarction or stroke as incident disease. They tested the association between TyG and cardiovascular events, then measured whether adding TyG to the Framingham Risk Score improved its ability to pick out future cases, using several complementary metrics. Finally, they pooled effect estimates from five cohort studies, with cardiovascular disease incidence ranging from 3.3% to 14.4%, to quantify the incremental predictive value overall.

People in the highest TyG quartile had roughly twice the risk of cardiovascular disease and of stroke compared with the lowest, with a non-linear dose-response pattern. Yet adding TyG to the Framingham score nudged the discrimination statistic from 0.731 to just 0.732, a change that was not significant, although adding TyG as quartiles did improve some reclassification metrics. The pooled analysis of five cohorts found an overall gain of 0.0004, essentially zero; excluding one influential study raised it to 0.0051, still tiny. The authors rated the certainty of this evidence as very low.

This work draws a line that many risk-marker studies blur: a variable can be strongly associated with disease and still add almost nothing to prediction, because standard risk factors already carry most of the same information. It therefore complicates enthusiasm for bolting TyG onto existing risk scores, while leaving intact its role as a marker of metabolic risk. Whether the index performs differently in other populations or at particular risk levels remains open, and the very low certainty rating means the question is not fully closed.

This is a reason to treat the TyG index as an interesting metabolic signal rather than a superior crystal ball for heart risk. If cardiovascular risk has already been estimated with an established tool, this evidence suggests a TyG value would be unlikely to change the picture meaningfully. Higher TyG was still linked with roughly double the risk in this cohort, so it is not meaningless; it simply may not add new information beyond what standard risk factors capture.