Development and validation of a CT-based body composition model for predicting adverse outcomes in small bowel...

Postoperative outcomes of small bowel obstruction (SBO) are unpredictable. Current risk models, based on basic clinical parameters, remain imprecise. While body composition abnormalities such as sarcopenia and visceral obesity are known to affect surgical outcomes, a comprehensive model that integrates muscle quality, systemic inflammation, nutrition, and intraoperative factors is lacking.

We conducted a retrospective cohort study of 270 patients diagnosed with SBO from the First Hospital of Shanxi Medical University between January 2018 and December 2024. Body composition parameters were assessed using computed tomography (CT) at the level of the third lumbar vertebra. Univariate and multivariate logistic regression analyses were performed to identify risk factors for adverse outcomes, including postoperative sepsis, ICU admission, and postoperative complications.

Of the 270 patients included in the final analysis, 68 required ICU admission, 105 experienced clinically significant postoperative complications (Clavien-Dindo grade ≥ II), and 53 developed postoperative sepsis. Multivariable analysis identified low skeletal muscle density (SMD) as the most consistent and potent independent predictor across all adverse outcomes: postoperative sepsis (aOR 4.58, 95% CI 1.61-13.04, p = 0.004), ICU admission (aOR 3.76, 95% CI 1.55-9.13, p = 0.003), and postoperative complications (aOR 3.73, 95% CI 1.59-8.71, p = 0.002). Other significant predictors included a prolonged time from symptom onset to surgery, longer operative duration, lower serum albumin levels, elevated D-dimer, greater resected bowel length, and prolonged prothrombin time. The predictive models demonstrated strong discriminative ability, with area under the curve (AUC) values ranging from 0.78 to 0.84 in the training sets and 0.72 to 0.80 in the validation sets. Calibration and decision curve analyses confirmed excellent model fit and clinical utility.

We developed and validated a novel multidimensional nomogram that accurately predicts the risk of major postoperative adverse events in SBO patients. By leveraging routinely available end-of-surgery data, this tool facilitates early postoperative risk stratification, which is crucial for intensifying postoperative monitoring and optimizing ICU resource allocation.