全髋关节置换后的低蛋白血症:危险因素及列线图预测模型建立

Hypoproteinemia after total hip arthroplasty:risk factors and nomogram prediction model establishment

背景:患者接受全髋关节置换后多发低蛋白血症,影响术后愈合及康复。目的:探讨及筛选全髋关节置换后发生低蛋白血症的危险因素,建立列线图为临床中判断全髋关节置换后是否发生低蛋白血症提供指导。方法:共纳入行全髋关节置换患者355例,根据术后第1天是否出现低蛋白血症分为低蛋白血症组238例、正常组117例,低蛋白血症发生率为67%。收集患者年龄、性别、糖尿病、高血压、高尿酸血症、高血脂、麻醉方式、置换前白细胞、置换前红细胞、置换前血红蛋白、置换前血小板、置换前血浆凝血酶原时间、置换前活化部分凝血酶原时间、置换前国际标准化比值、置换前凝血酶时间、置换前纤维蛋白原、置换前血沉、置换前C-反应蛋白、置换前D-二聚体、置换前平均红细胞血红蛋白含量、置换前平均红细胞体积、手术时间、体质量指数、置换前降钙素原、置换前红细胞压积等资料。利用SPSS 27.0软件进行单因素分析,随后利用R语言(4.3.1)再次对观察指标进行最小绝对收缩和选择算子回归分析及10倍交叉验证,取二者危险因素交集,使用SPSS 27.0进行多因素二元Logistic回归分析得到最终危险因素,通过R语言构建全髋关节置换后低蛋白血症的预测模型,并构建受试者特征曲线曲线、校准曲线、临床决策曲线评估预测模型预测能力。结果与结论:①经过单因素分析、最小绝对收缩和选择算子回归、多因素二元Logistic回归,筛选出年龄(OR=1.024,P=0.023)、置换前血小板(OR=0.995,P=0.028)、置换前血沉(OR=1.031,P=0.045)在判断置换后是否会发生低蛋白血症方面具有统计学意义(P<0.05);②基于多因素Logistic回归所筛选出的最终危险因素构建列线图预测模型,并通过构建受试者特征曲线评估模型预测能力,经计算受试者工作特征曲线的曲线下面积达到0.835(95%CI=0.779-0.891),C-index=0.835,经临床决策曲线计算,模型在阈值为0-0.83可带来更好的临床效能,该预测模型具有良好的灵敏度与准确性,可为医护人员及患者在全髋关节置换前更好地识别置换后发生低蛋白血症的风险。

BACKGROUND:The patient underwent multiple hypoproteinemia after total hip arthroplasty,which affected postoperative healing and rehabilitation.OBJECTIVE:To investigate and screen the risk factors for hypoproteinemia after total hip arthroplasty,and to establish a nomogram prediction model so as to provide guidance for judging whether hypoproteinemia occurs after total hip arthroplasty.METHODS:A total of 355 patients who underwent total hip arthroplasty were included,and according to whether hypoproteinemia occurred on the first day after surgery,they were divided into 238 cases in the hypoproteinemia group and 117 cases in the normal group,with a hypoproteinemia rate of 67%.Data were collected,including age,gender,diabetes mellitus,hypertension,hyperuricemia,hyperlipidemia,anesthesia method,preoperative leukocytes,preoperative erythrocytes,preoperative hemoglobin,preoperative platelets,preoperative plasma prothrombin time,preoperative activated partial prothrombin time,preoperative international normalized ratio,preoperative thrombin time,preoperative fibrinogen,preoperative erythrocyte sedimentation rate,preoperative C-reactive protein,preoperative D-dimer,preoperative mean corpuscular hemoglobin content,preoperative mean corpuscular volume,operation time,body mass index,preoperative procalcitonin,and preoperative hematocrit.SPSS 27.0 software was used for univariate analysis,followed by R language(4.3.1)to perform least absolute shrinkage and selection operator regression and 10-fold cross-validation of the observation indicators to obtain the intersection of the two risk factors.SPSS 27.0 software was used to perform multivariate binary logistic regression to obtain the final risk factors.The prediction model of hypoproteinemia after total hip arthroplasty was constructed by R language.The receiver operating characteristic curve,calibration curve,and clinical decision curve were constructed to assess the predictive model predictive ability.RESULTS AND CONCLUSION:(1)Univariate analysis,least absolute shrinkage and selection operator regression,and multivariate logistic regression were used to screen out significant differences in age(OR=1.024,P=0.023),preoperative platelets(OR=0.995,P=0.028),and preoperative erythrocyte sedimentation rate(OR=1.031,P=0.045)in judging whether hypoproteinemia would occur after surgery(P<0.05).(2)The nomogram prediction model was constructed based on the final risk factors screened by multivariate Logistic regression,and the prediction ability of the model was evaluated by constructing the receiver operating characteristic curve,and the area under the calculated receiver operating characteristic curve reached 0.835(95%CI=0.779-0.891),C-index=0.835.A threshold of 0-0.83 could bring better clinical efficacy calculated by the decision curve analysis.The model has good sensitivity and accuracy,which can better identify the risk of postoperative hypoproteinemia for medical staff and patients before total hip arthroplasty.