关键词: 先天性肛门直肠畸形, 排便功能障碍, 预测模型, 列线图, 儿童, 验证

Abstract: Objective To explore the influencing factors of postoperative defecation dysfunction in children with anorectal malformations(ARM),and to construct a risk prediction model of postoperative defecation dysfunction after ARM.Methods 150 cases ARM children who underwent surgery in our department of general surgery between March 2021 and April 2024 were enrolled.The children were divided into the training set (120 cases) and the internal validation set (30 cases) at a ratio of 8∶2 by the retention method..Children in the training set were collected and divided into obstacle groups (n=36,children with postoperative defecation dysfunction) and barrier-free group (n=84) according to whether postoperative defecation dysfunction occurred.The demographic data,clinical characteristics,surgical approaches,pelvic-floor muscle strength,and other variables were collected.influencingfactors were identified by univariate and multivariate logistic regression analyses.A nomogram was developed with R software.The discrimination (C-index),calibration curves,and decision-curve analysis were assessed in both the internal and external validation sets.Results The incidence of postoperative defecation dysfunction was 30.00 % (36/120) in the training set,33.33% (10/30) in the internal validation set,and 32.00 % (16/50) in the external validation set.There was no significant difference in 3 sets (χ2=0.261,P=0.877).In the training set,variables showing significant differences between the dysfunction and non-dysfunction groups were preterm birth,hypoglycaemia,irregular diet,artificial feeding,associated congenital anomalies,cardiac malformations,poor defecation habits,lower-extremity manual muscle testing(MMT) grade 3,lower-extremity MMT grade 5,Oxford pelvic-floor muscle grade 4,Oxford pelvic-floor muscle grade 5 and bowel function score(BFS)＜17.Stepwise Logistic regression revealed that BFS score < 17,preterm birth,Oxford pelvic-floor muscle grade,hypoglycaemia,cardiac malformations and poor defecation habits were independent predictors of postoperative defecation dysfunction (P＜0.05).The Oxford pelvic-floor muscle grade was a protective factor.In the training set,the area under the curve (AUC) was 0.69(0.60~0.77),accuracy 0.62(0.52~0.72),sensitivity 0.86(0.75~0.97),specificity 0.91(0.84~0.99) and a cutoff value of 0.223.AUC of the validation set 0.70(0.61~0.78),accuracy 0.62(0.52~0.72),sensitivity 0.89(0.79~0.99),specificity 0.93(0.86~0.99).Both the internal and external validation sets confirmed good predictive performance and stability of the model.Conclusion The risk prediction model,incorporating BFS＜17points,preterm birth,Oxford pelvic-floor muscle grade,hypoglycaemia,cardiac malformations,and poor defecation habits,accurately predicts the risk of postoperative defecation dysfunction in children with ARM and has been internally and externally validated.

Key words: anorectal malformations, defecation dysfunction, prediction model, nomogram, children, validation