糖尿病患者足溃疡复发风险预测模型的系统评价

Risk prediction models for foot ulcer recurrence in diabetic patients:a systematic review

目的系统评价糖尿病足溃疡(DFU)复发风险预测模型,以期为预防DFU复发提供理论指导。方法系统检索Web of Science、Embase、PubMed、CINAHL、Cochrane Library、中国生物文献数据库、中国知网、维普网、万方数据库中关于DFU风险预测模型的相关文献,并追溯纳入文献的引证文献及参考文献。检索时限均为建库始至2024年3月1日,检索语种不限。由2名研究者独立进行文献筛选、资料提取,并根据预测模型偏倚风险评估工具对纳入研究进行质量评价。结果共纳入12篇文献,涉及18个预测模型,样本总量为101~1333例,结局事件数为29~209个。模型重复报告次数较多的预测因子为溃疡位置(n=9)、吸烟史(n=8)、截肢史(n=7)、性别(n=6)、糖化血红蛋白水平(n=5)以及体重指数(n=5),受试者工作特征曲线下面积为0.660~0.943,9篇文献进行内部验证、5篇文献内外部验证相结合。18个预测模型适用性较好,但存在较高的偏倚风险,主要集中在数据分析领域。结论对于罹患足底溃疡且具备吸烟史的男性DFU患者,应予以特别关注。纳入的模型具有良好的预测效能和适用性,有助于医护人员早期识别DFU复发的高危患者。未来可在已有模型的基础上进行广泛而深入的验证,或开展多中心、大样本的前瞻性研究,以减少偏倚并持续优化模型,从而更好地服务临床实践。

Objective To systematically evaluate the risk prediction model of diabetic foot ulcer(DFU)recurrence,so as to provide theoretical guidance for the prevention of DFU recurrence.Methods The literature on DFU risk prediction models was systematically searched on the Web of Science,Embase,PubMed,CINAHL,Cochrane Library,China Biology Medicine,CNKI,VIP,WanFang Data,and the citations and references of the included literature were traced back.The search period was from the establishment of the database to March 1,2024,and the search language was not limited.Two researchers independently conducted literature screening and data extraction,and evaluated the quality of the included studies using the predictive model bias risk assessment tool.Results A total of 12 articles were included,involving 18 prediction models,with a sample size ranging from 101 to 1333 cases and a number of outcome events ranging from 29 to 209.The predictive factors with the high number of repeated model reports were ulcer location(n=9),smoking history(n=8),gender(n=6),amputation history(n=6),glycated hemoglobin(n=5),and body mass index(n=5).The area under the receiver operating characteristic curve reported by 18 models ranged from 0.660 to 0.943.Nine articles underwent internal validation,and five articles underwent a combination of internal and external validation.These 18 prediction models had good applicability,but there was a high risk of bias,mainly concentrated in data analysis.Conclusions Special attention should be paid to male DFU patients with foot ulcers and a history of smoking.The included model has good predictive performance and applicability,which helps medical and nursing staff identify high-risk patients for DFU recurrence early.Extensive and in-depth validation can be conducted on existing models,or prospective studies with multiple centers and large samples can be conducted to reduce bias and continuously optimize the model,thereby serving clinical practice well.