137 Multi-Center Analysis of Point-of-Care Ultrasound in the Diagnosis of Small Bowel Obstruction and Prediction of Surgical Intervention

Abstract

The management and diagnosis of small bowel obstruction (SBO) relies on diagnostic imaging. While multiple studies have evaluated the performance of POCUS in diagnosis of SBO, further studies are needed to help establish if and how the test can be used in clinical practice. We sought to analyze the prognostic ability of POCUS to diagnosis SBO and identify patients requiring surgical intervention. 401 patients with suspected SBO who were prospectively enrolled at five separate emergency departments (ED) of tertiary care hospitals were included in this study. Clinician pretest probability, patient symptoms, physical examination and POCUS findings including bowel dilation, peristalsis, wall edema, abdominal free fluid, and transition points were used to determine the most reliable findings in diagnosis and severity grading of SBO. All POCUS images were reviewed and interpreted by an independent reviewer, who made a prediction of whether or not a SBO was present. Radiologist CT interpretations and applicable surgeon intra-operative findings were used as a reference standard. All findings were placed into a nomogram, to calculate an appropriate predictive model for POCUS as an imaging modality for SBO. The discriminative ability of the nomogram was tested using C statistics, calibration plots, and Kaplan-Meier curves. The derivation cohort included 401 patients who underwent POCUS for suspected SBO (mean age 55 years, 41% female). All patients underwent abdominal CT and 31% of the patients had a SBO by CT. Of the 124 patients diagnosed with SBO, 35% required surgical intervention and 91% of these patients were positive for SBO by POCUS. There was an 84% agreement between POCUS reviewer and radiologist/surgeon findings. A stepwise selection of clinically important variables showed identified small bowel diameter, positive free fluid in the abdomen, clinician pretest probability, and patient age as relevant factors in the identification of SBO. The nomogram revealed good predictive ability as indicated by a high C-statistic for SBO diagnosis. When combined with clinical factors including physician gestalt and patient demographics, POCUS is a reliable method of accurately diagnosing SBO and identifying patients who may need a surgical intervention. This holds true across multiple emergency departments and ultrasound users.