Computed tomography findings of pneumatosis and portomesenteric venous gas in acute bowel ischemia

Aim: To use more representative sample size to evaluate whether computed tomography (CT) scan evidence of the concomitant presence of pneumatosis and portomesenteric venous gas is a predictor of transmural bowel necrosis. Methods: Data from 208 patients who were referred for a diagnosis of bowel ischemia were retrospectively reviewed. Only patients who underwent a surgical intervention following a diagnosis of bowel ischemia who also had a post-operative histological confirmation of such a diagnosis were included. Patients were split into two groups according to the presence of histological evidence of transmural bowel ischemia (case group) or partial bowel ischemia (control group). CT images were reviewed for findings of ischemia, including mural thickening, pneumatosis, bowel distension, portomesenteric venous gas and arterial or venous thrombi. Results: A total of 248 subjects who underwent surgery for bowel ischemia were identified. Among the 208 subjects enrolled in our study, transmural bowel necrosis was identified in 121 subjects (case group), and partial bowel necrosis was identified in 87 subjects (control group). Based on CT findings, including mural thickening, bowel distension, pneumatosis, pneumatosis plus portomesenteric venous gas and presence of thrombi or emboli, there were no significant differences between the case and control groups. The concomitant presence of pneumatosis and porto-mesenteric venous gas showed an odds ratio of 1.95 (95%CI: 0.491-7.775, P = 0.342) for the presence of transmural necrosis. The presence of pneumatosis plus porto-mesenteric venous gas exhibited good specificity (83%) but low sensitivity (17%) in the identification of transmural bowel infarction. Accordingly, the positive and negative predictive values were 60% and 17%, respectively. Conclusion: Although pneumatosis plus porto-mesenteric venous gas is associated with bowel ischemia, we have demonstrated that their co-occurrence cannot be used as diagnostic signs of transmural necrosis.