Radiographic Predictors of Disease Severity in Neonates and Infants With Necrotizing Enterocolitis

The objective of our study was to validate a radiographic scale, the Duke abdominal assessment scale (DAAS), as a tool for predicting the severity of disease in neonates and infants with suspected necrotizing enterocolitis (NEC). Study group patients (n = 43) underwent at least two two-view abdominal radiographic series within 48 hours of surgical intervention for suspected NEC complications. Control group patients (n = 86) were patients with suspected NEC who did not undergo surgery for suspected NEC complications. DAAS scores were assigned by two pediatric radiologists with 20 and 6 years' experience. The initial radiographs of 26 of 43 (60.5%) patients in the study group showed fixed bowel loops (10/43, 23.3%), highly probable or definite pneumatosis (9/43, 20.9%), or portal venous gas (7/43, 16.3%). These findings had progressed to pneumoperitoneum on the follow-up series in 20 (46.5%) study group patients. Among the control group, three patients (3.5%) had highly probable or definite pneumatosis, and none had fixed bowel loops, portal venous gas, or pneumoperitoneum. Patients with higher DAAS scores were more likely to undergo surgical intervention than patients with lower scores (odds ratio, 1.69; 95% CI, 1.40-2.03). A receiver operating characteristic curve analysis showed good overall performance (c statistic = 0.83) for predicting eventual surgical intervention in the study group with higher DAAS scores. The DAAS provides a standardized 10-point radiographic scale that increases with disease severity when using need for surgical intervention as a surrogate for severe NEC. For every 1-point increase in the DAAS score, patients were statistically significantly more likely to have severe disease as measured by need for surgical intervention.