The Triglyceride-glucose Index Predicts Short-Term Mortality in Critically Ill Patients With Cancer-Associated Thrombosis: A Retrospective Cohort Study.

BackgroundThe triglyceride-glucose (TyG) index as an indicator of insulin resistance is seldom studied in severely ill patients with cancer-associated thrombosis (CAT), but we sought to evaluate its performance on all-cause mortality and if it adds any added value over current risk scores.MethodsData for CAT patients were extracted from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database. Patients were stratified by TyG tertiles. Multivariable Cox regression, restricted cubic splines, and subgroup analyses were performed. Incremental predictive value was evaluated using area under the curve (AUC), integrated discrimination improvement (IDI) and net reclassification improvement (NRI).ResultsAmong 959 patients, 28-day and 90-day mortality rates were 16.27% and 20.02%. The TyG index was independently associated with increased mortality (28-day: HR 1.35, 95% CI: 1.13-1.61; 90-day: HR 1.32, 95% CI: 1.13-1.56), with linear dose-response relationships (P for nonlinear = 0.418 and 0.520, respectively). The association was stronger in non-diabetic patients. Adding TyG modestly improved risk discrimination and reclassification.ConclusionThe TyG index is associated with short-term mortality in critically ill CAT patients. Adding it to conventional risk scores yielded only modest improvements in discrimination metrics, with uncertain clinical significance. Routine clinical use is not supported by the current data, and residual confounding from unmeasured cancer-related variables cannot be excluded. Prospective validation is warranted.