TDM联合Logistic模型对利奈唑胺致血小板减少症危险因素的研究

Study on the risk factors of linezolid induced thrombocytopenia by TDM combined with Logistic model

目的:研究利奈唑胺治疗重症感染患者发生血小板减少的危险因素。方法:采用逻辑回归分析方法,结合TDM技术,针对患者用药治疗前生理特点、治疗期间临床指标特征情况、治疗药物监测结果、感染类型与合并用药情况这五个方面,通过构建Logistic回归模型,并联合ROC曲线对血小板减少症的发生情况进行预测分析。结果:108例患者(男性65例,女性43例),35例患者出现了血小板减少症状。采用逐步回归的Logistic分析法显示,患者年龄、血肌酐浓度、血小板基线值、谷浓度是发生血小板减少症的独立危险因素。综合上述独立危险因素,建立Logistic回归方程:Logit (P)=0.056×XC-0.094×XPLT+0.086×XScr-0.082×XAge+1.927,并通过构建联合预测因子的ROC曲线,对Logistic回归方程进行变换得到联合预测因子的计算公式,Y联合=XC-1.678×XPLT+1.535×XScr-1.464×XAge。结论:患者应用利奈唑胺治疗时,可将该患者的年龄、血肌酐浓度、血小板基线值、谷浓度代入方程中(Y联合=XC-1.678×XPLT +1.535×XScr-1.464×XAge),若计算出的结果小于520.62(ROC最佳临界值),则该患者在治疗期间出现血小板减少症的风险较高,此时需要加强血常规监测,并根据结果及时调整给药方案。

OBJECTIVE To study the risk factors of thrombocytopenia in severe infection patients treated with linezolid? METHODS Logistic regression analysis and TDM technique were used to predict the occurrence of thrombocytopenia according to the physiological characteristics of patients before treatment, the characteristics of clinical indicators during treatment, the monitoring results of therapeutic drugs, the types of infection and the combination of drugs. RESULTS 108 patients (65 males, 43 females) were included in the study, among which thrombocytopenia occurred in 30 patients. Logistic analysis using stepwise regression showed that age, serum creatinine, baseline platelet counts and trough concentration were significant risks for linezolid-associated thrombocytopenia. Based on the above independent risk factors, Logistic regression equation was established: Logit(P)=0.056×XC-0.094×XPLT+0.086×XScr-0.082×XAge+1.927. By constructing the ROC curve of the joint predictor, the Logistic regression equation was transformed to obtain the calculation formula of the joint predictor : Y联合=XC-1.678×XPLT+1.535×XScr-1.464×XAge. CONCLUSION In clinical practice, we could calculate the joint predictor by the age, serum creatinine, baseline platelet counts and trough concentration of the patients by using calculation formula(Y=XC-1.678×XPLT +1.535×XScr-1.464×XAge). If the calculated result was less than 520. 62 (the optimal cut-off value of ROC), the patient had a higher risk of thrombocytopenia during treatment. At this time, it is necessary to strengthen blood routine monitoring and timely adjust the dosing regimen according to the results.