Chest CT-derived body composition parameters for outcome prediction in sepsis patients with pneumonia.

OBJECTIVE: We aim to evaluate the potential contribution of chest computed tomography (CT)-derived body composition parameters in predicting adverse events in sepsis patients with pneumonia. METHODS: A retrospective study was conducted on sepsis with pneumonia cases who visited Shengjing Hospital of China Medical University from January 2023 to September 2024. We used chest CT scans to quantify skeletal muscle area (SMA) at the fourth thoracic vertebra (T4) and the first lumbar vertebra (L1) levels, as well as abdominal circumference (AC), subcutaneous adipose tissue (SAT), and intramuscular adipose tissue (IMAT) at the L1 level. RESULTS: A total of 303 patients (203 men; median age 70 years, interquartile range 63-79) were included in the study. Fully adjusted models identified low SMAT4, low SATL1, and high ACL1 as independent risk factors for medical intensive care unit (MICU) admission, with odds ratios (ORs) of 0.795, 0.897, and 2.095, respectively. Low SMAT4 (OR 0.880, 95% confidence interval [CI] 0.800-0.967, p = 0.008) and high ACL1 (OR 1.527, 95% CI 1.122-2.079, p = 0.007) were both independently associated with in-hospital mortality. High IMATL1 (β: -2.360, p = 0.003) was associated with a greater decrease in the PaO2/FiO2 ratio during hospitalization. Models using chest CT-derived body composition parameters to predict MICU admission, septic shock, and in-hospital mortality in patients with sepsis were as effective as sequential organ failure assessment scores. CONCLUSIONS: The assessment of frailty status and visceral obesity, determined by chest CT measurements of low thoracic muscle mass and elevated AC, is independently correlated with an increased risk of admission to the MICU and mortality among sepsis patients with pneumonia. This underscores the significance of CT-derived body composition as a critical imaging biomarker that reflects the physiological reserve of sepsis patients and their associated risk of adverse events.