老年女性全髋关节置换术后低蛋白血症的相关因素△（开放获取）

Factors related to hypoproteinemia after total hip arthroplasty in elderly women

摘要:

[目的]探讨老年女性全髋关节置换患者围手术期低蛋白血症的危险因素。[方法]本院 2022 年 1 月—2023 年 6 月收治的 476 例全髋关节置换术的老年女性患者纳入本研究，采用单因素比较和多元逻辑回归分析探索低蛋白血症发生的相关因素。通过 R 软件构建列线图预测模型，并绘制出 ROC 曲线、校准曲线以及决策曲线分析评价模型的效能。[结果]476 例女性患者中 109 例围手术期发生低蛋白血症，占比 22.9%。单因素分析显示，低蛋白组年龄 [(75.7±4.8) 岁 vs (69.7±5.0) 岁, P＜0.001]、 病因中股骨颈骨折占比 [例, 股骨颈骨折/股骨头坏死, (73/36) vs (191/176), P=0.006]、C 反应蛋白 （C- reactive protein, CRP） [(22.1±25.9) mg/L vs (14.8±21.5) mg/L, P=0.008]、 血红细胞沉降率（erythrocyte sedimentation rate, ESR）[(29.6±22.1) mm/h vs (22.0± 16.7) mm/h, P＜0.001]、D-二聚体（D-dimer, D-D）[(5.2±5.7) μg/mL vs (3.1±1.1) μg/mL, P＜0.001]、住院时间 [(10.9±4.4) d vs (9.7± 3.2) d, P=0.002] 显著高于正常组，但糖尿病占比 [例, 有/无, (9/100) vs (77/290), P=0.002]、BMI [(23.3±3.8) kg/m2 vs (24.5±3.6) kg/m2 , P=0.002]、RBC [(4.0±0.4)×10¹²/L vs (4.2±0.5)×10¹²/L, P＜0.001]、PLT [(184.9±64.8)×10⁹ /L vs (200.9±66.2)×10⁹ /L, P=0.027]、Ca [(2.26± 0.11) mmol/L vs (2.32±0.11) mmol/L, P＜0.001] 显著低于正常组。多因素分析显示，年龄 （OR=1.266, P＜0.05）、术前 ESR （OR= 1.029, P＜0.001） 是术后发生低白蛋白血症的危险因素；糖尿病 （OR=0.336, P=0.013）、BMI （OR=0.924, P=0.040）、术前 PLT （OR=0.995, P=0.015）、术前 Ca 水平（OR=0.050, P=0.016）是保护因素。列线图模型的 ROC 曲线下面积（AUC）为 0.857，表明该预测模型具有良好的区分度。[结论] 根据年龄、BMI、糖尿病、术前血小板、术前血钙以及术前血沉这 6 个独立危险因素， 列线图可以有效预测老年女性全髋关节置换患者术后发生低蛋白血症的风险。该列线图预测模型具有良好的区分度和准确性， 并可能为患者带来更好的临床净收益。

Abstract:

[Objective] To explore the factors related to perioperative hypoproteinemia of total hip arthroplasty (THA) in elderly women. [Methods] A total of 476 elderly female who had THA performed in our hospital from January 2022 to June 2023 were included in this study. Univariate comparison and multiple logistic regression analysis were used to explore the related factors of hypoproteinemia. R software was used to construct the nomogram prediction model, and ROC curve, calibration curve and decision curve were drawn to analyze and evaluate the efficiency of the model. [Results] Among 476 female patients, 109 suffered from perioperative hypoproteinemia, accounting for 22.9%. As consequence of univariate comparison, the hypoproteinemia group proved significantly higher than the normal group in terms of age [(75.7±4.8) years vs (69.7±5.0) years, P＜0.001], the proportion of femoral neck fracture [femoral neck fracture/femoral head necrosis, (73/ 36) vs (191/176), P=0.006], Creactive protein (CRP) [(22.1±25.9) mg/L vs (14.8±21.5) mg/L, P=0.008], erythrocyte sedimentation rate (ESR) [(29.6±22.1) mm/h vs (22.0±16.7) mm/h, P＜0.001], D-dimer (DD) [(5.2±5.7) μg/mL vs (3.1±1.1) μg/mL, P＜0.001], hospital stay [(10.9±4.4) days vs (9.7±3.2) days, P=0.002], whereas the former was significantly lower than the latter in terms of diabetes [yes/no, (9/100) vs (77/290), P=0.002], BMI [ (23.3±3.8) kg/m2 vs (24.5±3.6) kg/m2 , P=0.002], RBC [(4.0±0.4) 10¹²/L vs (4.2±0.5) 10¹²/L, P＜0.001], PLT [(184.9±64.8) 10⁹ /L vs (200.9±66.2) 10⁹ /L, P=0.027], Ca [(2.26±0.11) mmol/L vs (2.32±0.11) mmol/L, P＜0.001]. Regarding to multivariate logic regression, age (OR=1.266, P＜0.05), preoperative ESR (OR=1.029, P＜0.001) were risk factors for postoperative hypoalbuminemia, while the diabetes (OR=0.336, P=0.013), BMI (OR=0.924, P=0.040), preoperative PLT (OR=0.995, P=0.015) and preoperative serum Ca(OR=0.050, P=0.016) were protective factors. The area under curve (AUC) of the nomogram model was of 0.857, indicating that the prediction model had a good differentiation capacity. [Conclusion] Based on six independent risk factors: age, BMI, diabetes mellitus, preoperative platelet, preoperative calcium and preoperative sedimentation rate, the histogram can effectively predict the risk of postoperative hypoproteinemia of THA in elderly women. The nomogram model has good differentiation ability and accuracy, and might lead better clinical net benefits for patients.