Risk factors and predictive models for early recurrent intussusception in children: a retrospective cohort study.

Abstract

Early recurrent intussusception (ERI) in children is common and seriously affects the physical and mental health of the children. There are few reports discussing risk factors for ERI in children, and this study aims to identify risk factors for ERI in children and build predictive models. We conducted a retrospective study of 787 children with no relapse intussusception (NRI) and 82 children with ERI between January 2011 and December 2021. Univariate and multifactorial stepwise logistic regression analysis was used to analyze the correlation between 11 factors and ERI, to determine the independent risk factors for ERI in children. The prediction model was established by independent risk factors and then verified. Age, vomiting, bloody stools, and monocyte ratios were independently correlated with the composite endpoint (P<0.05). A nomogram was constructed and a calibration curve was plotted, using independent risk factors. Based on the disease's diagnostic score, the predictive model's performance was validated by using logistic regression receiver operating characteristic (ROC) curve detection, with area under the curve (AUC) value of 0.883 [95% confidence interval (CI): 0.846-0.920], and the calibration curve was close to the ideal diagonal line. In addition, the decision curve analysis (DCA) showed that the model had significant net benefits. Independent risk factors for ERI in children are age, vomiting, bloody stool, and monocyte ratio. Children older than 1 year in age, who lacked vomiting and bloody stool symptoms, and who exhibited an elevated ratio of monocytes were more likely to relapse early. The predictive model constructed herein can predict the early recurrence of children with ERI, providing a reference for clinicians' individualized judgments.