Background
Distinguishing strangulated bowel obstruction (StBO) from simple bowel obstruction (SiBO) still poses a challenge for emergency surgeons. We aimed to construct a predictive model that could distinctly discriminate StBO from SiBO based on the degree of bowel ischemia.
Methods
The patients diagnosed with intestinal obstruction were enrolled and divided into SiBO group and StBO group. Binary logistic regression was applied to identify independent risk factors, and then predictive models based on radiological and multi-dimensional models were constructed. Receiver operating characteristic (ROC) curves and the area under the curve (AUC) were calculated to assess the accuracy of the predicted models. Via stratification analysis, we validated the multi-dimensional model in the prediction of transmural necrosis both in the training set and validation set.
Results
Of the 281 patients with SBO, 45 (16.0%) were found to have StBO, while 236(84.0%) with SiBO. The AUC of the radiological model was 0.706 (95%CI, 0.617–0.795). In the multivariate analysis, seven risk factors including pain duration ≤ 3 days (OR = 3.775), rebound tenderness (OR = 5.201), low-to-absent bowel sounds (OR = 5.006), low levels of potassium (OR = 3.696) and sodium (OR = 3.753), high levels of BUN (OR = 4.349), high radiological score (OR = 11.264) were identified. The AUC of the multi-dimensional model was 0.857(95%CI, 0.793–0.920). In the stratification analysis, the proportion of patients with transmural necrosis was significantly greater in the high-risk group (24%) than in the medium-risk group (3%). No transmural necrosis was found in the low-risk group. The AUC of the validation set was 0.910 (95%CI, 0.843–0.976). None of patients in the low-risk and medium-risk score group suffered with StBO. However, all patients with bowel ischemia (12%) and necrosis (24%) were resorted into high-risk score group.
Conclusion
The novel multi-dimensional model offers a useful tool for predicting StBO. Clinical management could be performed according to the multivariate score.