胫骨远端骨折患者经微创接骨板内固定术后出现切口并发症的列线图预测模型构建与评价

Construction and evaluation of nomogram prediction model for incision complications in patients with distal tibial fractures after minimally invasive bone plate fixation

目的分析胫骨远端骨折患者经微创接骨板内固定(MIPO)术后出现切口并发症的相关因素,构建列线图预测模型并评价其性能。方法选取2019年10月至2023年3月新余北湖医院收治的300例胫骨远端骨折患者作为研究对象,所有患者均行MIPO术,根据患者术后是否出现切口并发症分为观察组(n=28)和对照组(n=272)。对两组的临床资料进行统计分析,采用logistic多因素回归分析患者术后出现切口并发症的独立危险因素;R软件构建预测患者术后出现切口并发症的列线图模型;校正曲线对预测模型进行内部验证;决策曲线评估预测模型的临床净收益。结果两组年龄、糖尿病、吸烟、术前使用抗生素比较,差异有统计学意义(P<0.05);多因素分析结果提示,术前没有使用抗生素(β=0.611,OR=1.843,95%CI:1.242~2.013)、年龄(β=0.718,OR:2.051,95%CI:1.574~2.689)、糖尿病(β=0.578,OR=1.872,95%CI:1.513~2.529)、吸烟(β=1.140,OR=3.128,95%CI:2.573~3.975)是胫骨远端骨折患者经微创接骨板内固定术后出现切口并发症的独立影响因素,差异有统计学意义(P<0.05)。将差异有统计学意义的连续变量纳入ROC曲线:年龄、BMI的曲线下面积(AUC)是0.786、0.602,最佳截断值分别是52.65岁、27.23 kg/m2。内部验证结果显示,列线图预测模型的C-index为0.823(95%CI:0.452~0.896),观测值与预测值之间保持较好一致性,差异有统计学意义(P<0.05)。决策曲线显示当风险阈值>0.05时,此预测模型可以提供额外的临床净收益。结论胫骨远端骨折患者MIPO术后出现切口并发症的列线图预测模型,具有较好的预测价值,有助于医务人员认识胫骨远端骨折患者MIPO术后出现切口并发症的相关因素并早期进行针对性干预,可为临床控制患者术后切口并发症发生率提供一些参考。

Objective To analyze the factors related to incision complications in patients with distal tibial fractures after minimally invasive percutaneous osteosynthesis(MIPO),and construct a nomogram prediction model and evaluate its perfor-mance.Methods A total of 300 patients with distal tibial fracture admitted to Beihu Hospital of Xinyu from October 2019 to March 2023 were selected as the study objects.All patients underwent MIPO surgery,and were divided into observation group(n=28)and control group(n=272)according to whether postoperative incision complications occurred.The clinical data of the two groups were statistically analyzed,and the independent risk factors of postoperative incision complications were analyzed by logistic multivariate regression.R software built a nomogram model to predict postoperative incision com-plications.The correction curve verifies the prediction model internally.The decision curve evaluates the clinical net benefit of the predictive model.Results There were significant differences in age,diabetes,smoking and preoperative antibiotic use between the two groups(P<0.05).The results of multivariate analysis suggested that no antibiotics were used before surgery(β=0.611,OR=1.843,95%CI:1.242-2.013),age(β=0.718,OR=2.051,95%CI:1.574-2.689),diabetes mellitus(β=0.578,OR=1.872,95%CI:1.513-2.529),smoking(β=1.140,OR=3.128,95%CI:2.573-3.975)were the independent in-fluencing factors for incision complications in patients with distal tibial fractures after minimally invasive bone plate fixation,and independent difference was statistically significant(P<0.05).Continuous variables with statistically significant differ-ences were included in the ROC curve:the area under the curve(AUC)of age and BMI were 0.786 and 0.602,and the best cutoff values were 52.65 years and 27.23 kg/m2,respectively.Internal verification results showed that the C-index of the nomogram prediction model was 0.823(95%CI:0.452-0.896),and there was a good agreement between the observed and predicted values,and the difference was statistically significant(P<0.05).The decision curve showed that this predic-tive model could provide additional clinical net benefit when the risk threshold was>0.05.Conclusion The nomogram pre-diction model for postoperative incision complications in patients with distal tibial fractures has good predictive value,which helps medical staff to understand the related factors of postoperative incision complications in patients with distal tibial frac-tures and conduct targeted intervention in the early stage,and can provide some references for clinical control of postopera-tive incision complications in patients.