Risk of complications in vascular surgery: development of a clinical predictive model

血管手术并发症风险:临床预测模型的建立

阅读:2

Abstract

BACKGROUND: Postoperative complications in vascular surgery are associated with high morbidity, mortality, and hospital costs, highlighting the need for reliable predictive tools for risk stratification. OBJECTIVES: To develop and validate a clinical model to estimate the risk of postoperative complications in vascular surgery. METHODS: This retrospective study included 510 patients who underwent vascular surgeries between 2021 and 2024, divided into arterial, venous, and vascular access subgroups. Clinical and surgical variables were analyzed using multivariate logistic regression, and model performance was evaluated using the receiver operating characteristic curve. RESULTS: The overall complication rate was 17.6%, being higher in arterial procedures (35.6%) than venous procedures (11.3%) or vascular access surgeries (6.9%). In the total sample, age (odds ratio [OR] 1.03; p = 0.006), chronic kidney disease (OR 9.94; p < 0.001), smoking (OR 3.29; p = 0.001), and procedure time (p = 0.038) were independent predictors, while chronic anticoagulant use had a protective effect (OR 0.39; p = 0.036). In the specific subgroup models, type 2 diabetes mellitus (OR 13.54; p < 0.001) and chronic kidney disease (OR 15.30; p = 0.007) were significant predictors in the venous access group, smoking was associated with risk in the vascular access group (OR 9.57; p = 0.081), and chronic kidney disease was significant in the arterial group (OR 6.50; p < 0.001). The model showed good discriminatory performance (overall area under the curve [AUC] = 0.806). CONCLUSIONS: The proposed model demonstrated good accuracy and clinical applicability, allowing individualized risk stratification across different vascular surgery contexts. External validation is needed to confirm its usefulness.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。