Abstract
BACKGROUND: Sand enteropathy is a common disease in horses worldwide. The presence of sand does not always cause disease. The amount of sand required to cause clinical disease is not well established. OBJECTIVES: To establish a weight-indexed cut-off for the diagnosis of clinically relevant sand enteropathy based on radiographic area. STUDY DESIGN: Cross-sectional. METHODS: Abdominal radiographs were acquired from clinically normal horses and compared with horses with clinical sand enteropathy. Sand area and sand area per kilogram of body weight were calculated. A receiver operating characteristic (ROC) curve was used to differentiate groups. RESULTS: Seventy clinically normal horses and 57 sand enteropathy cases were included. Median sand area in control horses was 1 cm(2) per 100 kg (interquartile range 0-6 cm(2) per 100 kg), and in sand enteropathy horses was 118 cm(2) per 100 kg (interquartile range 83-180 cm(2) per 100 kg). Receiver operator characteristic (ROC) analysis revealed that using a cut-off of 21 cm(2) per 100 kg had a sensitivity of 98.25% (CI 90.71%-99.1%) and a specificity of 92.54% (CI 83.69%-96.77%) for a diagnosis of clinical sand enteropathy. CONCLUSIONS: A weight-based cut-off of >21 cm(2) per 100 kg for the diagnosis of sand enteropathy was determined with excellent sensitivity and good specificity. This may aid in determining if the amount of sand accumulation in a horse is of clinical consequence.