先天性巨结肠症患者术后短期并发症的危险因素及列线图模型

Risk factors and nomogram model for short-term postoperative complications in patients with hirschsprung disease.

作者信息

Song Aohua, Zhang Bobin, Feng Wei, Hou Jinping, Die Xiaohong, Wang Yi, Guo Zhenhua

机构信息

Department of General & Neonatal Surgery, Children's Hospital of Chongqing Medical University, National Clinical Research Center for Child Health and Disorders, Ministry of Education Key Laboratory of Child Development and Disorders, Chongqing Key Laboratory of Structural Birth Defect and Reconstruction, Chongqing, China.

出版信息

BMC Med Inform Decis Mak. 2025 Jul 1;25(1):214. doi: 10.1186/s12911-025-03053-0.

DOI:10.1186/s12911-025-03053-0

PMID:40598014

原文链接:https://pmc.ncbi.nlm.nih.gov/articles/PMC12211763/

Abstract

OBJECTIVE

Short-term postoperative complications (SPCs) in patients with Hirschsprung disease (HD) are a topic of concern. The aims of this study were to identify the risk factors for different severities of SPCs based on the Clavien–Dindo classification system and to develop a nomogram model for clinical prediction.

METHODS

The medical records of 304 eligible patients who were diagnosed with HD were retrospectively reviewed. Univariate/multivariate logistic regression analysis was used to identify the independent risk factors for SPCs of varying severity. Then, the enrolled patients were randomly divided into a training set ( = 212) and a testing set ( = 92) at a 7:3 ratio. Logistic regression analysis was conducted using R software, incorporating selected features from the LASSO regression model to construct a predictive model.

RESULTS

The SPCs of 304 patients was 21.7% (66/304). LASSO regression identified six candidate predictors: type of HD, feeding methods, preoperative nutritional status, preoperative hypoproteinemia, preoperative HAEC, and length of postoperative hospital stay. Multivariate analysis further confirmed formula feeding (OR: 5.150, 95%CI: 1.912–14.29, = 0.001), malnutrition (OR: 3.696, 95%CI: 0.987–13.37, = 0.047), and L-HD (OR: 14.45, 95%CI: 4.690–51.44, < 0.001) were independent factors for SPCs, and clinical nomogram model was developed based on these determinants. The area under of the receiver operating characteristics curve of the training set was 0.841 and that of the testing set was 0.834. Calibration curve analysis, decision curve analysis, and clinical impact curve analysis verified the accuracy and practicability of the predictive model.

CONCLUSION

The feeding method, preoperative nutritional status, and type of HD were found to be independent risk factors for SPCs. Formula feeding lacks immune protection, malnutrition delays recovery, and L-HD requires extensive surgery. A nomogram model based on the above factors could be utilized to predict whether HD patients are at risk of developing SPCs, thereby guiding timely nutritional interventions, personalized postoperative care, and close monitoring to reduce complication rates.

SUPPLEMENTARY INFORMATION

The online version contains supplementary material available at 10.1186/s12911-025-03053-0.

摘要

目的

先天性巨结肠（HD）患者术后短期并发症（SPCs）是一个备受关注的话题。本研究旨在基于Clavien-Dindo分类系统确定不同严重程度SPCs的危险因素，并建立一个列线图模型用于临床预测。

方法

回顾性分析304例确诊为HD的符合条件患者的病历。采用单因素/多因素逻辑回归分析确定不同严重程度SPCs的独立危险因素。然后，将纳入的患者按7:3的比例随机分为训练集（n = 212）和测试集（n = 92）。使用R软件进行逻辑回归分析，纳入LASSO回归模型中选定的特征来构建预测模型。

结果

304例患者的SPCs发生率为21.7%（66/304）。LASSO回归确定了6个候选预测因素：HD类型、喂养方式、术前营养状况、术前低蛋白血症、术前HAEC和术后住院时间。多因素分析进一步证实配方奶喂养（OR：5.150，95%CI：1.912 - 14.29，P = 0.001）、营养不良（OR：3.696，95%CI：0.987 - 13.37，P = 0.047）和长段型HD（OR：14.45，95%CI：4.690 - 51.44，P < 0.001）是SPCs的独立因素，并基于这些决定因素建立了临床列线图模型。训练集的受试者操作特征曲线下面积为0.841，测试集为0.834。校准曲线分析、决策曲线分析和临床影响曲线分析验证了预测模型的准确性和实用性。