To explore the risk factors of bile leakage in patients with cystic echinococcosis undergoing surgery, and to establish a regression prediction model.

Clinical data of 183 patients with cystic echinococcosis who underwent surgical treatment in Qinghai Welfare Charity Hospital from June 2014 to December 2018 were retrospectively analyzed. The informed consents of all patients were obtained and the local ethical committee approval was received. Among them, 67 cases were male and 116 female, aged 12-78 years old with a median age of 41 years old. The risk factors of bile leakage were selected by Logistic regression analysis. The effect of every index in Logistic regression analysis was illustrated in the forest plot. The Logistic regression prediction model for bile leakage of cystic echinococcosis patients undergoing surgery was established. The fitting degree of the prediction model was evaluated by Hosmer-Lemeshow test. The prediction value of this model was assessed by receiver operating characteristic (ROC) curve.

The incidence of bile leakage was 22.4% (41/183). Multivariate Logistic regression analysis demonstrated that ALP>135 U/L (OR=10.797, 95%CI: 4.860-23.927, P<0.05) and cyst diameter>10 cm (OR=3.576, 95%CI: 1.911-6.268, P<0.05) were the independent risk factors for bile leakage in cystic echinococcosis patients undergoing surgery. These two factors were included in the Logistic regression equation: logit (P)=-2.580+2.412X_ALP+1.215X_{cyst diameter}. The probability model was: P=1/[1+Exp(-2.580+2.412X_ALP+1.215X_{cyst diameter})]. The area under ROC curve was 0.792 (95%CI: 0.719-0.864, P<0.05), the sensitivity was 0.74 and the specificity was 0.70.

ALP and cyst diameter are the risk factors for bile leakage in cystic echinococcosis patients undergoing surgery. The prediction model established by regression analysis can effectively predict the incidence of bile leakage in patients with cystic echinococcosis undergoing surgery.