摘要
目的基于临床实验室指标构建平坦型突发性耳聋预后不良的数学预测模型,旨在为临床工作提供参考。方法前瞻性选取2020年8月至2024年8月郑州大学第一附属医院225例平坦型突发性耳聋患者作为研究对象,根据预后情况分为预后不良(n=150)与预后良好(n=75)。研究对象按照16∶4∶5的比例随机分为训练集、验证集和测试集。采用多因素Logistic回归分析筛选独立影响因素,构建数学预测模型,并通过R软件进行模型内部验证及评价。结果预后不良患者的总胆固醇(TC)、纤维蛋白原(FIB)、低密度脂蛋白胆固醇(LDL-C)、D-二聚体(DD)、同型半胱氨酸(HCY)、凝血酶原时间、活化部分凝血活酶时间(APTT)、血小板计数(PLT)、红细胞压积(Hct)分别为(4.79±1.02)mmol/L、(2.97±0.59)g/L、(3.17±0.90)mmol/L、(0.62±0.18)mg/L、(16.94±3.65)mmol/L、(18.25±1.90)s、(28.84±4.17)s、(219.54±18.27)×10^(9)/L、0.23±0.06,明显高于预后良好组患者的(4.30±0.84)mmol/L、(2.73±0.45)g/L、(2.59±0.76)mmol/L、(0.42±0.11)mg/L、(14.37±2.86)mmol/L、(15.17±1.42)s、(26.02±3.19)s、(212.37±12.83)×10^(9)/L、0.21±0.05,高密度脂蛋白胆固醇(HDL-C)为(1.17±0.24)mmol/L,明显低于预后良好组患者的(1.38±0.25)mmol/L,差异均有统计学意义(P<0.05);LDL-C、HCY、APTT、PLT、Hct均为平坦型突发性耳聋预后不良的独立危险因素,HDL-C为其独立保护因素(P<0.05);平坦型突发性耳聋预后不良的数学预测模型为ln(P/1-P)=-6.637-0.167×HDL-C+0.411×LDL-C+0.551×HCY+0.418×APTT+0.323×PLT+0.394×Hct;受试者工作特征(ROC)曲线、校准曲线结果显示,该数学模型在训练集、验证集及测试集中均具有较强的预测能力及较高的校准度。结论LDL-C、HCY、APTT、PLT、Hct均为平坦型突发性耳聋预后不良的独立危险因素,HDL-C为其独立保护因素,基于此构建的数学预测模型预测准确性较高且预测能力较强,对临床潜在平坦型突发性耳聋预后不良的高风险人群筛查具有一定指导价值,可为临床工作提供参考。
Objective To develop a mathematical prediction model for poor prognosis in flat-type sudden deafness based on clinical laboratory indicators,aiming to provide a reference for clinical practice.Methods A prospective study was conducted on 225 patients with flat-type sudden deafness admitted to the First Affiliated Hospital of Zhengzhou University from August 2020 to August 2024.Based on prognosis,patients were divided into a poor prognosis group(n=150)and a good prognosis group(n=75).The subjects were randomly allocated into training,validation,and test sets in a 16∶4∶5 ratio.Multivariate logistic regression analysis was used to identify independent influencing factors and construct the mathematical prediction model.Internal validation and evaluation of the model were performed using R software.Results Total cholesterol(TC),fibrinogen(FIB),low-density lipoprotein cholesterol(LDL-C),D-dimer(DD),homocysteine(HCY),prothrombin time(PT),activated partial thromboplastin time(APTT),platelet count(PLT),and hematocrit(Hct)in the poor prognosis group were(4.79±1.02)mmol/L,(2.97±0.59)g/L,(3.17±0.90)mmol/L,(0.62±0.18)mg/L,(16.94±3.65)μmol/L,(18.25±1.90)s,(28.84±4.17)s,(219.54±18.27)×10^(9)/L,and 0.23±0.06,respectively,which were significantly higher than(4.30±0.84)mmol/L,(2.73±0.45)g/L,(2.59±0.76)mmol/L,(0.42±0.11)mg/L,(14.37±2.86)μmol/L,(15.17±1.42)s,(26.02±3.19)s,(212.37±12.83)×10^(9)/L,and 0.21±0.05 in the good prognosis group(P<0.05).High-density lipoprotein cholesterol(HDL-C)was(1.17±0.24)mmol/L in the poor prognosis group,significantly lower than(1.38±0.25)mmol/L in the good prognosis group(P<0.05).LDL-C,HCY,APTT,PLT,and Hct were identified as independent risk factors for poor prognosis in flat-type sudden deafness,while HDL-C was an independent protective factor(P<0.05).The mathematical prediction model for poor prognosis in flat-type sudden deafness was ln(P/1-P)=-6.637-0.167×HDL-C+0.411×LDL-C+0.551×HCY+0.418×APTT+0.323×PLT+0.394×Hct.Receiver operating ibration across the training,validation,and test sets.Conclusion LDL-C,HCY,APTT,PLT,and Hct are independent risk factors for poor prognosis in flat-type sudden deafness,while HDL-C is an independent protective factor.The mathematical prediction model constructed based on these factors shows high predictive accuracy and strong predictive ability,providing valuable guidance for screening high-risk populations and offering a reference for clinical practice.
作者
郝少娟
赵堃
朱晓丹
王乐
HAO Shao-juan;ZHAO Kun;ZHU Xiao-dan;WANG Le(Department of Otology,the First Affiliated Hospital of Zhengzhou University,Zhengzhou 450000,Henan,CHINA)
出处
《海南医学》
2025年第24期3583-3589,共7页
Hainan Medical Journal
基金
河南省科技攻关项目(编号:192102310115)。
关键词
平坦型突发性耳聋
预测
数学模型
实验室指标
影响因素
预测价值
Flat-type sudden hearing loss
Predictive
Mathematical model
Laboratory indicators
Influencing factors
Predictive value
