Development and Validation of a Novel Hollow Viscus Injury Prediction Score for Abdominal Seatbelt Sign: A Pacific Coast Surgical Association Multicenter Study

Abstract

BACKGROUND: High-quality CT can exclude hollow viscus injury (HVI) in patients with abdominal seatbelt sign (SBS) but performs poorly at identifying HVI. Delay in diagnosis of HVI has significant consequences necessitating timely identification. STUDY DESIGN: This multicenter, prospective observational study conducted at 9 trauma centers between August 2020 and October 2021 included adult trauma patients with abdominal SBS who underwent abdominal CT before surgery. HVI was determined intraoperatively and physiologic, examination, laboratory, and imaging findings were collected. Least absolute shrinkage and selection operator– and probit regression–selected predictor variables and coefficients were used to assign integer points for the HVI score. Validation was performed by comparing the area under receiver operating curves (AUROC). RESULTS: Analysis included 473 in the development set and 203 in the validation set. The HVI score includes initial systolic blood pressure <110 mmHg, abdominal tenderness, guarding, and select abdominal CT findings. The derivation set has an AUROC of 0.96, and the validation set has an AUROC of 0.91. The HVI score ranges from 0 to 17 with score 0 to 5 having an HVI risk of 0.03% to 5.36%, 6 to 9 having a risk of 10.65% to 44.1%, and 10 to 17 having a risk of 58.59% to 99.72%. CONCLUSIONS: This multicenter study developed and validated a novel HVI score incorporating readily available physiologic, examination, and CT findings to risk stratify patients with an abdominal SBS. The HVI score can be used to guide decisions regarding management of a patient with an abdominal SBS and suspected HVI.