Distinguishing pathologic from benign pneumatosis intestinalis and validating a five-factor score: a retrospective cohort study

Purpose: Pathologic pneumatosis intestinalis (PI) generally requires surgery, but diagnostic uncertainty risks negative laparotomy. This study aimed to determine its incidence, identify distinguishing variables, assess radiographic accuracy, and validate a five-factor algorithm.Methods: This multihospital, single health system retrospective cohort study analyzed electronic health record encounters with radiographic pneumatosis to assess incidence, imaging concordance, and five-factor score performance. Secondary outcomes included differences at presentation, operative management, time to surgery, and mortality. Pearson chi-square, Fisher exact, and Wilcoxon rank sum tests compared categorical and continuous variables; confusion matrices assessed imaging concordance. Statistical significance was defined as P<0.05.Results: Of 252 cases, 90 (35.7%) involved pathologic PI. Compared with benign PI, pathologic PI was associated with abdominal pain (76.7% vs. 58.0%, P=0.003), tenderness (70.0% vs. 52.5%, P=0.008), vasopressor use (21.1% vs. 5.6%, P<0.001), lactate ≥2.0 mmol/L (P=0.011), portal venous gas (48.9% vs. 32.1%, P=0.009), in-hospital mortality among operative cases (42.2% vs. 17.4%, P=0.004), and higher median aspartate aminotransferase level (36 U/L vs. 29 U/L, P=0.044), white blood cell count (13.8×103/µL vs. 11.7×103/µL, P=0.024), and neutrophil to lymphocyte ratio (13.6 vs. 8.5, P<0.001). Negative laparotomy occurred in 46 benign cases (28.4%). Computed tomography (CT) correctly localized 97.9% of cases with and 83.7% without small bowel involvement, and 76.4% of cases with and 97.1% without colonic involvement. Corresponding positive and negative predictive values were 87% and 97% for the small bowel and 98% and 72% for the colon, respectively. Among 198 complete cases, 89 (44.9%), 49 (24.7%), 33 (16.7%), and 27 (13.6%) were classified as low, intermediate, high, and very high risk, with observed pathologic PI rates of 26%, 47%, 39%, and 56%, respectively.Conclusion: Pathologic PI and negative laparotomy were common. Several clinical, laboratory, and CT findings distinguished pathologic from benign PI. CT localization was generally more concordant with operative findings in the small bowel than in the colon. The five-factor score poorly predicted pathologic PI.