Abstract Introduction Acute respiratory failure (ARF) is a major cause of morbidity and mortality in intensive care units (ICUs), often requiring invasive mechanical ventilation (IMV). Early liberation from IMV is associated with improved outcomes including earlier mobility, preserved neurocognition, avoidance of critical illness complications, and enhanced survival. However, it is unclear whether non-pulmonary characteristics such as pre-existing body composition, especially thoracic chest wall features, impact early ventilator liberation. The respiratory muscles are key to breathing efforts and airway clearance, while excess thoracic obesity may add resistive load. Therefore, we examined the association between thoracic body composition metrics and clinical outcomes in IMV patients with ARF. We hypothesized that higher thoracic skeletal muscle mass, lower thoracic subcutaneous fat mass, and a higher muscle-to-fat ratio would predict reduced 30-day mortality and faster successful ventilator liberation within 60 days among patients with ARF from any cause and in a subgroup with acute respiratory distress syndrome (ARDS). Methods We analyzed 238 mechanically ventilated ARF patients from the Pittsburgh Acute Lung Injury Registry (2012-2024), who had chest CT scans between 7 days prior to and 3 days post intubation. Thoracic skeletal muscle mass and thoracic subcutaneous fat mass were quantified from chest CTs using validated 3D convolutional neural networks spanning the lung apex to base. Multivariable logistic regression models predicted 30-day mortality and Cox proportional hazards models analyzed time to ventilator liberation within 60 days. Muscle mass models were adjusted for age, sex, and BMI. Fat mass and ratio models excluded BMI to avoid collinearity. A subgroup analysis included 81 patients meeting ARDS criteria. Results The ARF cohort had a mean age of 57.3 (SD 16.3) years and mean BMI of 30.34 (SD 8.7). The ARDS subgroup was younger with mean age of 51.3 (SD 16.8) and similar BMI of 30.1 (SD 7.6). Body composition metrics were not significantly associated with 30-day mortality in either cohort, though higher thoracic subcutaneous fat mass trended toward increase mortality risk (OR 1.08 1.00-1.17, p = 0.055). Importantly, a higher muscle-to-fat ratio was significantly associated with faster ventilator liberation (HR 1.16 1.01-1.31, p = 0.018) in the ARF cohort but not in the ARDS subgroup. Conclusion In mechanically ventilated patients with ARF, thoracic body composition measured by CT was not significantly associated with 30-day mortality. However, higher muscle-to-fat ratio independently predicted more rapid liberation from ventilation. Imaging-derived body composition may help prognosticate outcomes and warrants further research in larger cohorts. This abstract is funded by: R01HL174570 (JP), R01CA237277 (JP), Breathe PA Foundation – AirWISE Award (GDK; JP)
Jain et al. (Fri,) studied this question.