Cardiometabolic indices as predictors of clinical outcomes in palliative care patients.

Abstract

Multiple cardiometabolic indices have been proposed for prognostic assessment, yet their comparative performance in palliative care remains unclear. The triglyceride-glucose body mass index (TyG-BMI) integrates metabolic dysfunction with adiposity, but whether it outperforms traditional lipid-based, inflammatory, and nutritional indices requires systematic evaluation. To comprehensively compare TyG-BMI against eleven established cardiometabolic indices for predicting sepsis, mechanical ventilation requirement, and 30-day mortality in palliative care patients, with specific focus on performance in diabetic subpopulations. This retrospective cohort included 318 palliative care patients. Twelve indices were calculated: TyG-BMI (primary); lipid-based (AIP, CRI-I, CRI-II, Non-HDL, TG/HDL); inflammatory (NLR, PLR, SII, MHR); and nutritional (PNI, CAR). ROC analysis compared discriminative ability for sepsis, mechanical ventilation, and 30-day mortality. Subgroup analyses stratified by diabetes mellitus status were performed with interaction testing. Of 318 patients (mean age 67.4 ± 14.8 years, 55% male), 121 (38.1%) had diabetes, 58 (18.2%) developed sepsis, 42 (13.2%) required mechanical ventilation, and 30 (9.4%) died within 30 days. TyG-BMI achieved the highest AUCs: 0.84 (95% CI 0.78-0.90) for sepsis, 0.82 (0.75-0.89) for ventilation, and 0.87 (0.82-0.92) for 30-day mortality-significantly superior to all comparator indices (p < 0.001). In multivariate analysis, TyG-BMI independently predicted mortality (OR 2.38 per SD, 95% CI 1.78-3.18, p < 0.001). In diabetic patients, TyG-BMI's discriminative ability was markedly enhanced (mortality AUC 0.92, 95% CI 0.87-0.97; OR 2.65, 95% CI 1.88-3.74, p < 0.001), while other indices showed minimal performance improvement (interaction p < 0.001). TyG-BMI demonstrates superior prognostic performance compared to traditional cardiometabolic indices for predicting sepsis and 30-day mortality in palliative care, with exceptional discriminative ability in diabetic patients.