糖尿病足溃疡截肢患者预后预测模型的构建与评估

Construction and evaluation of prognosis prediction model for amputation patients with diabetes foot ulcer

目的 建立糖尿病足溃疡截肢患者预后不良影响因素的列线图预测模型并进行评估。方法 选取2018年8月—2021年6月上海交通大学医学院附属新华医院伤口中心和内分泌科收治的糖尿病足溃疡截肢患者122例,术后随访1年,根据随访期间是否出现不良事件划分为预后良好组76例和预后不良组46例。收集患者一般资料及术前实验室指标;采用单因素分析与多因素Cox风险回归分析筛选糖尿病足溃疡患者截肢术后预后不良的影响因素,R软件纳入影响因素建立列线图预测模型,并通过受试者工作特征曲线(ROC)及校准曲线评价列线图对糖尿病足溃疡患者截肢预后不良的预测能力。结果 预后不良组有糖尿病家族史比例、中性粒细胞/淋巴细胞比值(NLR)、C反应蛋白(CRP)、D-二聚体(D-D)水平高于预后良好组[χ^(2)(t)/P=5.025/0.025、6.803/<0.001、3.229/0.002、6.752/<0.001],白蛋白(Alb)水平低于预后良好组(t/P=2.564/0.012)。多因素Cox风险回归分析结果显示,NLR高、CRP高、D-D高为糖尿病足溃疡截肢患者术后预后不良的危险因素[HR(95%CI)=1.520(1.193~1.935)、1.041(1.018~1.066)、3.771(2.059~6.908)],Alb高为其保护因素[HR(95%CI)=0.934(0.888~0.982)]。基于影响因素建立的列线图预测模型,验证结果显示,ROC曲线下面积为0.861(95%CI 0.786~0.936),校准曲线为斜率近似于1的直线。结论 基于糖尿病足溃疡患者截肢术后预后不良的影响因素建立的列线图预测模型具有良好的区分度与一致性,对于糖尿病足溃疡截肢患者术后治疗对策的制定具有一定的指导价值。

Objective To establish and evaluate the nomogram prediction model of the factors influencing the poor prognosis of amputation patients with diabetes foot ulcer. Methods From August 2018 to June 2021, 122 amputees with diabetes foot ulcer were selected from the Wound Center and Endocrine Department of Xinhua Hospital Affiliated to Shanghai Jiaotong University School of Medicine. They were followed up for one year after surgery. According to whether there were adverse events during the follow-up period, they were divided into a good prognosis group of 76 cases and a bad prognosis group of 46 cases. Collect general data of patients and preoperative laboratory indicators;Single factor analysis and multivariate Cox risk regression analysis were used to screen the influencing factors of poor prognosis of diabetes foot ulcer patients after amputation, R software was included in the influencing factors to establish a nomogram prediction model, and the ability of the nomogram to predict the poor prognosis of diabetes foot ulcer patients after amputation was evaluated through the subject’s work characteristic curve(ROC) and calibration curve. Results The proportion of family history of diabetes, neutrophil/lymphocyte ratio(NLR), C-reactive protein(CRP), D-dimer(D-D) levels in poor prognosis group were higher than those in good prognosis group [χ^(2)(t)/P=5.025/0.025, 6.803/<0.001, 3.229/0.002, 6.752/<0.001], the level of albumin(Alb) was lower than that of the group with good prognosis(t/P=2.564/0.012). Multivariate Cox risk regression analysis showed that high NLR, CRP and D-D were risk factors for poor prognosis of amputees with diabetes foot ulcer [HR(95% CI)=1.520(1.193-1.935), 1.041(1.018-1.066), 3.771(2.059-6.908)], and high Alb was a protective factor [HR(95% CI)=0.934(0.888-0.982)]. The validation results of the nomograph prediction model based on the influence factors show that the area under the ROC curve is 0.861(95% CI 0.786-0.936), and the calibration curve is a straight line with a slope of approximately 1. Conclusion The nomogram prediction model based on the influencing factors of poor prognosis of amputation patients with diabetes foot ulcer has good discrimination and consistency, and has certain guiding value for the formulation of countermeasures for amputation patients with diabetes foot ulcer.