全膝关节置换术后早期膝关节功能改善情况预测列线图模型的建立与验证

Development and validation of a nomogram model for predicting knee function improvement in early postoperative period after total knee arthroplasty

目的建立全膝关节置换术(total knee arthroplasty,TKA)后早期膝关节功能改善情况的列线图预测模型并进行验证。方法前瞻性选择2018年1月—2021年2月在四川省骨科医院行TKA的168例患者,收集可能影响TKA术后早期膝关节功能改善情况的因素,在术后6个月时采用美国特种外科医院(Hospital for Special Surgery,HSS)膝关节评分量表评估膝关节功能改善情况。根据术后膝关节功能改善情况是否理想将患者分为两组。比较两组术前、术中及术后因素,以LASSO回归筛选潜在因素后行多因素logistic回归,以R 4.1.3语言建立列线图预测模型并进行内部验证。结果术后6个月所有患者均获随访,患者平均HSS评分由术前的(55.19±8.92)分升高至术后6个月的(89.27±6.18)分(t=−40.706,P<0.001)。LASSO回归筛选出8个潜在影响因素,以此行多因素logistic回归分析结果显示,术前体质量指数、病因、术前关节活动度、术前HSS评分、术后下肢力线、术后镇痛情况为TKA术后早期膝关节功能改善情况的独立影响因素(P<0.05)。根据多因素结果建立列线图模型,校准曲线结果显示,预测曲线与标准曲线基本拟合;受试者操作特征曲线分析结果显示,列线图模型预测TKA术后早期膝关节功能不理想的曲线下面积为0.894[95%置信区间(0.825,0.963)]。结论TKA术后患者膝关节功能有明显改善,TKA术后早期膝关节功能改善情况受术前体质量指数、病因、术前关节活动度等因素的影响,依此建立的列线图模型用于预测TKA术后早期膝关节功能改善情况具有较高的区分度与准确度。

Objective To develop and validate a nomogram prediction model of early knee function improvement after total knee arthroplasty(TKA).Methods One hundred and sixty-eight patients who underwent TKA at Sichuan Province Orthopedic Hospital between January 2018 and February 2021 were prospectively selected to collect factors that might influence the improvement of knee function in the early postoperative period after TKA,and the improvement of knee function was assessed using the Knee Score Scale of the Hospital for Special Surgery(HSS)at 6 months postoperatively.The patients were divided into two groups according to the postoperative knee function improvement.The preoperative,intraoperative and postoperative factors were compared between the two groups;multiple logistic regression was performed after the potential factors screened by LASSO regression;then,a nomogram predictive model was established by R 4.1.3 language and was validated internally.Results All patients were followed up at 6 months postoperatively,and the mean HSS score of the patients increased from 55.19±8.92 preoperatively to 89.27±6.18 at 6 months postoperatively(t=−40.706,P<0.001).LASSO regression screened eight influencing factors as potential factors,with which the results of multiple logistic regression analysis showed that preoperative body mass index,etiology,preoperative joint mobility,preoperative HSS scores,postoperative lower limb force line,and postoperative analgesia were independent influencing factors for the improvement of knee function in the early postoperative period after TKA(P<0.05).A nomogram model was established based on the multiple logistic regression results,and the calibration curve showed that the prediction curve basically fitted the standard curve;the receiver operating characteristic curve showed that the area under the curve of the nomogram model for the prediction of suboptimal knee function in the early postoperative period after TKA was 0.894[95%confidence interval(0.825,0.963)].Conclusions There is a significant improvement in knee function in patients after TKA,and the improvement of knee function in the early postoperative period after TKA is influenced by preoperative body mass index,etiology,and preoperative joint mobility,etc.The nomogram model established accordingly can be used to predict the improvement of knee function in the early postoperative period after TKA with a high degree of differentiation and accuracy.