Objective The aim of this study is to evaluate the prognostic value of CT-based body composition parameters (skeletal muscle quantity and muscle quality), fat distribution, Global Leadership Initiative on Malnutrition (GLIM) malnutrition criteria and inflammatory/metabolic markers (C-reactive protein/albumin ratio (CAR) and triglyceride-glucose (TyG) index) on 90-day mortality. Methods A total of 118 patients admitted to the palliative care unit between January 2020 and December 2021 were included in the study. The Psoas Area Index (PAI), Psoas Density Index (PDI) and visceral/subcutaneous fat area ratio were calculated from CT images at the L3 vertebra level. Malnutrition diagnosis was established according to GLIM criteria. Independent risk factors associated with 90-day mortality were evaluated using multivariate logistic regression analysis, and model performance was validated using the 1000-repetition bootstrap method. Results Out of 118 patients, the 90-day mortality rate was 39.8% (n=47). In deceased patients, PAI (5.4±1.2 vs 6.9±1.4 cm²/m²) and PDI (33.8±7.5 vs 39.5±7.0 HU (Hounsfield unit)) values were significantly lower (p<0.001 for both). In multivariate logistic regression analysis, low PAI (OR 0.65), low PDI (OR 0.89), high CAR (OR 3.78), high TyG index (OR 2.11), GLIM-severe malnutrition (OR 2.94) and presence of dementia (OR 2.74) were identified as independent predictors of 90-day mortality. The model’s discriminatory power was excellent (area under the curve: 0.91). Conclusions CT-based skeletal muscle quantity and muscle quality measurements, when evaluated together with inflammatory and metabolic indices, offer a robust and objective approach to predicting short-term mortality risk in palliative care patients. These parameters may contribute to the early identification of high-risk patients and the planning of individualised care strategies.
Kaya et al. (2026) studied this question.