Related Experiment Video
Updated: Jan 8, 2026
Implementation of a Real-Time Psychosis Risk Detection and Alerting System Based on Electronic Health Records using CogStack
Published on: May 15, 2020
Prognostic value of C-reactive protein triglyceride glucose index versus triglyceride glucose index in critically ill
Wen-Liang Shuai1, Jie Liu1, Tao Zhou2
1Department of Cardiovascular Medicine, The Second Affiliated Hospital, Jiangxi Medical College, Nanchang University, NanChang, Jiangxi, China.
Background:
Critically ill patients have high mortality influenced by insulin resistance (IR)and inflammation.The C-reactive protein-triglyceride-glucose index (CTI), integrating IR and inflammation, has unclear prognostic value in this population compared to the traditional triglyceride-glucose index (TyG).
Methods:
This cohort study analyzed 1963 critically ill patients from MIMIC-IV. Outcomes included in-hospital mortality, acute kidney injury (AKI), and length of stay. Analyses employed multivariable regression, mediation, and machine learning.
Results:
Elevated CTI showed a graded association with mortality. Per unit increase conferred 22% higher risk (OR 1.22, 95% CI 1.08-1.38); the top quartile had double the risk (OR 2.07, 1.36-3.13) and a mortality rate of 26.48% vs. 9.98% in the lowest quartile. CTI also independently predicted higher AKI risk (OR 1.28, 1.13-1.45; incidence 91.24% vs. 73.32% in Q4 vs. Q1) and prolonged ICU/hospital stay (β 1.99 and 3.71 days, p < 0.05). AKI mediated 5.53% of the CTI-mortality relationship. CTI's discriminative power surpassed TyG (DeLong p = 0.001). In prognostic modeling, CTI was a key feature, with the Gradient Boosting Machine model achieving optimal performance (AUC 0.789).
Conclusion:
CTI is strongly associated with adverse outcomes and may outperform TyG in critically ill patients, supporting its utility for risk stratification.
