Multivariate Logistic Regression Analysis And Model Establishment Of Adult Small Bowel Infarction

Objective To investigate the main risk factors of small bowel infarction, to establish the Logistic regression model and to evaluate the accuracy,sensitivity and specificity for predicting small intestine infarction in adult patients, to provide the basis for clinical treatment.Methods Case-control study was used to retrospectively investigate the clinical data of392cases who was admitted in the Affiliated Hospital of Qingdao University between January2007and January2014,of which186cases of small bowel infarction were selected as study group, other206patients with adhesive small bowel obstruction(who were excluded bowel infarction by surgery)as the control group were matched. Firstly,single factor analysis was used. Secondly, multivariate Logistic regression analysis was used in the selected variables. Then, Logistic regression equation was established,which can be used to evaluate the accuracy,sensitivity and specificity of small bowel infarction in adult.Results14clinical variables,including high fever,history of abdominal operation, peritonitis, recurrent obstruction,melena, bowel wall thickening, mesenteric edema, preoperative count of leukocyte, pH value, Lactate, preoperative creatinine, preoperative urea nitrogen, creatine kinase level and CRP level were selected through single factor analysis. By binary Logistic analysis found that melena, peritonitis, creatine kinase level, CRP level, preoperative creatinine, mesenteric edema, high fever and preoperative count of leukocyte were the independent risk factors associated with small bowel infarction in adult. Logistic regression equation was P=1/(1+e (-2.365+2.366X1+2.144X2+1.795X3+1.762X4+1.541X5+0.583X6+0.540X7+0.450X8)). A test sample showed the accuracy,sensitivity and specificity of the Logistic model were81.12%,75.27%and86.41%.Conclusion melena, peritonitis, creatine kinase level, CRP level, preoperative creatinine, mesenteric edema, high fever and preoperative count of leukocyte were the main risk factors to predict the small bowel infarction in adult. The Logistic regressive model based on these factors is reliable in predicting the small intestine infarction in adult.