利奈唑胺诱导血小板减少症的危险因素及其预警价值分析

Risk Factors and Warning Value Analysis of Linezolid Induced Thrombocytopenia

目的 采用Logistic模型和ROC曲线探讨利奈唑胺诱导血小板减少症的危险因素及其预警价值分析。方法 回顾性分析连云港市东方医院2020年1月-2021年9月接受利奈唑胺注射液治疗患者的临床资料,收集患者的基本信息、实验室检查结果、利奈唑胺应用情况、联合用药和不良反应等数据。根据有无发生血小板减少症将患者分为两组,比较两组的临床特征。采用二元Logistic回归法分析住院患者使用利奈唑胺发生血小板减少症的危险因素。通过ROC曲线分析危险因素对血小板减少症的预测价值。结果 纳入分析的患者共118例(男性83例、女性35例),治疗期间15例患者(12.71%)发生血小板减少症。两组临床数据比较,发现LI-TP组用药前血小板计数(PLT)、肾小球滤过率(eGFR)低于未发生血小板减少症组(P<0.05)。二元Logistic回归分析显示用药前PLT、eGFR是发生LI-TP的独立危险因素。分别用上述独立危险因素建立Logistic回归方程,经变换后得到联合预测因子的计算公式,Y联合=XeGFR+0.8XPLT,联合预测因子的ROC曲线下面积(0.867,95%CI:0.793~0.942,P<0.001)优于其他指标,具有一定的预测价值,Youden指数最大时(0.653)的切点为ROC曲线上的最佳界值(205.49)。结论 用药前PLT、eGFR是利奈唑胺注射液致住院患者血小板减少的独立危险因素,其临界值分别为193.5×109/L和58.06mL/min。临床中,可将用药前PLT、eGFR带入联合预测因子的计算公式,推测发生血小板减少症的概率,加强用药监护,确保患者用药安全。

Objective To explore risk factors and early warning value of linezolid induced thrombocytopenia with logistic models and ROC curves.Methods The paper reviewed and analyzed clinical data of patients treated with linezolid injection in Lianyungang City Dongfang Hospital from January 2020 to September 2021,collected basic information,laboratory test results,application of linezolid,combined medication and adverse reactions,and divided them into two groups based on thrombocytopenia situation,and compared clinical characteristics between two groups.Risk factors for thrombocytopenia in hospitalized patients treated with linezolid were analyzed with binary logistic regression method,predictive value of risk factors for thrombocytopenia was analyzed based on ROC curve.Results The paper chose 118 patients(83 males and 35 females)included in analysis,and 15 patients(12.71%)showed thrombocytopenia during treatment period.Comparison of clinical data between two groups showed,platelet count(PLT)and glomerular filtration rate(eGFR)before medication in LI-TP group was lower than non thrombocytopenic group(P<0.05).Binary logistic regression analysis showed,pre medication PLT and eGFR were independent risk factors for occurrence of LI-TP.Logistic regression equations were established with independent risk factors mentioned above,after transformation,calculation formula for the joint predictive factor was obtained.Y joint=XeGFR+0.8XPLT,area under ROC curve of joint predictive factor was(0.867,95%CI:0.793~0.942,P<0.001),better than other indicators,which had some certain predictive value.Tangent point at the maximum Youden index(0.653)was the best boundary value on the ROC curve(205.49).Conclusion Pre medication PLT and eGFR are independent risk factors for thrombocytopenia in hospitalized patients caused by linezolid injection,with critical values of 193.5 respectively×109/L and 58.06mL/min.In clinical practice,pre medication PLT and eGFR can be incorporated into calculation formula of joint predictive factors to speculate probability of thrombocytopenia,strengthen medication monitoring and ensure patient medication safety.