In modern society, diabetes is like a “time bomb” threatening people’s health, severely affecting their quality of life and well – being. As the third – largest non – communicable disease globally, in 2021, the number of adult diabetic patients worldwide reached 537 million, and 6.7 million people died from diabetes. China, which is entering an aging society, is also facing a rapidly increasing burden of diabetes. In 2019, there were about 111.6 million adult diabetic patients, and it is predicted that this number will increase to 639 million by 2045.
Currently, most diabetes risk prediction models mainly focus on insulin resistance but overlook the crucial role of inflammation in the development of diabetes. It’s like building a bridge while only considering some of the supporting structures and ignoring other equally important parts, which makes the stability of the bridge difficult to guarantee, and the reliability of such prediction models is also in question.
To fill this research gap, researchers from multiple institutions carried out an important study. They focused on the C – reactive protein – triglyceride – glucose index (CTI index), an emerging indicator that can comprehensively reflect the body’s metabolic state and the impact of inflammation. The researchers used data from the China Health and Retirement Longitudinal Study (CHARLS) to deeply explore the relationship between the CTI index and the risk of diabetes.
The research results are of far – reaching significance. During the nine – year follow – up period, the incidence of diabetes was 15.9%. Through COX regression analysis, it was found that the higher the CTI index, the gradually increasing risk ratio of diabetes. The restricted cubic spline (RCS) curve further confirmed the linear relationship between them. Decision tree analysis showed that the CTI index is a key indicator for predicting the risk of diabetes. Moreover, compared with the TyG index, the CTI index is more accurate in predicting the risk of diabetes onset, just like a more precise “ruler” that helps people more accurately assess their risk of developing diabetes.
During the research process, the researchers adopted a variety of technical methods. They conducted a prospective study based on the CHARLS cohort and selected 6,728 participants without a history of diabetes at the baseline. For missing data, they used the multiple imputation method to handle it. They screened key predictors through the least absolute shrinkage and selection operator (LASSO) technique, used multiple COX regression to evaluate the relationship between the CTI index and the risk of new – onset diabetes, identified high – risk groups with decision tree analysis, and calculated the time – dependent Harrell’s C index and other methods to assess the predictive ability.
