Abstract:

Objective To establish a prediction model for failure of intussusception in children under fluoroscopy. Methods In this study， 439 children with intussusception were retrospectively included in our hospital， with 362 successful and 77 unsuccessful air enema. The total samples were randomly divided into 308 training sets and 131 verification sets according to a ratio of 7∶3. The parameters and clinical information of children with air enema were collected， the variables were screened by single factor Logistic regression analysis， and then the prediction model was constructed by multi?factor Logistic regression analysis and the Nomo diagram was drawn. The differentiation， calibration and decision curve of the model were evaluated. Results Statistical differences were found between the successful and unsuccessful groups except gender and intestinal obstruction （all P < 0.05）. The independent risk factors screened by multivariate Logistic regression analysis were age， course of disease， intestinal obstruction， head position and empty irrigation pressure for modeling. The model had good differentiation in predicting the failure of recirculation， with AUC = 0.854 （95%CI： 0.799 ~ 0.909）， truncation value of 0.24， sensitivity and specificity of 79% and 80%， respectively. The calibration curve showed that the risk predicted by the model was highly consistent with the actual situation. Hosmer Lemeshow test χ² = 8.159， P = 0.518； The decision curve shows that the model has clinical applicability. Conclusion Air enema pressure combined with clinical information has a better ability to predict the risk of failure of enema reduction， which can provide some references for clinical decision?making in children with emergency intussusception.

Key words: intussusception, children, air enema, prediction model