基于PLS-DA算法的中药注射剂不良反应个体差异性分析
|
摘要
目的采用偏最小二乘-判别分析(PLS-DA)算法分析中药注射剂不良反应的个体差异性。方法根据患者使用中药注射剂后的临床表现将患者分为药物不良反应(ADR)组和正常组。提取2组患者包括基本信息、血常规检查、尿液流式分析、生化系列检查、凝血功能检查结果在内的94项变量,构建多变量数据矩阵。建立PLS-DA模型,根据主成分得分对比2组患者个体因素的整体差异,根据主成分载荷分析患者个体因素与中药注射剂ADR的相关性。结果同时具有外伤手术史、药物过敏史和单核细胞比例偏高特征的患者ADR发生的风险较高。在临床使用中药注射剂时,对具有多项危险因素的患者应加强监护。结论 PLS-DA算法可适用于中药注射剂ADR的个体差异性分析。
AIM To investigate the individual differences for adverse drug reaction(ADR) induced by traditional Chinese medicine injections by using partial least squares discriminant analysis(PLS-DA) algorithm.METHODS The patients were divided into 2 groups according to their clinical manifestation after using traditional Chinese medicine injections. The patients who had ADR were included in the ADR group, while the patients who had no ADR were included in the normal group. The 94 variables, including the basic information of patients, the results of blood routine examination, urine flow analysis, biochemical series examination and coagulation function examination were extracted and the multivariate data matrix was built. The PLS-DA model was then established. The principal component scores were applied to investigate the overall differences between the ADR group and normal group. The correlations between individual factors of patients and ADR were analyzed according to the principal component loading. RESULTS There was higher risk of ADR for patients who had history of trauma surgery, history of drug allergy and high rates of monocyte. Intensive care should be paid to the patients who had multiple risk factors when using traditional Chinese medicine injections. CONCLUSION It is effective to employ PLS-DA algorithm for analyzing the individual differences for ADR induced by traditional Chinese medicine injections.
AIM To investigate the individual differences for adverse drug reaction(ADR) induced by traditional Chinese medicine injections by using partial least squares discriminant analysis(PLS-DA) algorithm.METHODS The patients were divided into 2 groups according to their clinical manifestation after using traditional Chinese medicine injections. The patients who had ADR were included in the ADR group, while the patients who had no ADR were included in the normal group. The 94 variables, including the basic information of patients, the results of blood routine examination, urine flow analysis, biochemical series examination and coagulation function examination were extracted and the multivariate data matrix was built. The PLS-DA model was then established. The principal component scores were applied to investigate the overall differences between the ADR group and normal group. The correlations between individual factors of patients and ADR were analyzed according to the principal component loading. RESULTS There was higher risk of ADR for patients who had history of trauma surgery, history of drug allergy and high rates of monocyte. Intensive care should be paid to the patients who had multiple risk factors when using traditional Chinese medicine injections. CONCLUSION It is effective to employ PLS-DA algorithm for analyzing the individual differences for ADR induced by traditional Chinese medicine injections.
引文
[1] 国家食品药品监督管理总局. 国家药品不良反应监测年度报告(2016年). [2017-02-20]. http://samr.cfda.gov.cn/WS01/CL0844/172167.html
[2] 刘福梅, 王连心, 谢雁鸣. 基于文献报道的1998-2015年间24例疑似清开灵注射液死亡不良事件病例分析 [J]. 中南药学,2017,15(7):993.
[3] 尹文杰. 参麦注射液致过敏性休克死亡[C]. 2013年中国药学大会暨第十三届中国药师周.2013.
[4] 沈斌, 周月红. 127例中药注射剂不良反应分析及原因探讨 [J]. 中华中医药杂志,2015,30(2):428.
[5] 祁麟, 文彦丽. 参麦注射液不良反应流行病学特点分析 [J]. 天津药学,2014,26(5):31.
[6] 唐军, 廖享, 童红, 等. 基于气质与PLS-DA对新疆薰衣草精油判别分析 [J]. 计算机与应用化学,2014,31(6):701.
[7] ANZANELLO M J, ORTIZ R S, LIMBERGER R, et al. PLS-DA wavenumber selection for the categorization of medicine samples based on multiple criteria [J]. Forensic Sci Int,2014,242:111.
[8] WONG K H, RAZMOVSKI-NAUMOVSKI V, LI K M, et al. Differentiation of Pueraria lobata and Pueraria thomsonii using partial least square discriminant analysis (PLS-DA) [J]. J Pharm Biomed Anal,2013,84:5.
[9] 田刚, 周明术, 杭永伦, 等. 肿瘤标志对肺癌诊断价值PLS-DA和ANN-MPL模型评估 [J]. 中华肿瘤防治杂志,2014,21(5):380.
[10] 蒋程, 韩冰, 吕健刚, 等. 基于偏最小二乘-判别分析算法的癌痛规范化治疗评价 [J]. 中国临床药理学与治疗学,2015,20(12):1378.
[11] 阚全程. 医院药学高级教程 [M]. 北京: 人民军医出版社,2014.
[12] 蒋程, 丁静, 吕健刚, 等. 基于偏最小二乘-判别分析算法的参麦注射剂不良反应关键影响因素分析 [J]. 中国临床药学杂志,2017,26(4):221.
[2] 刘福梅, 王连心, 谢雁鸣. 基于文献报道的1998-2015年间24例疑似清开灵注射液死亡不良事件病例分析 [J]. 中南药学,2017,15(7):993.
[3] 尹文杰. 参麦注射液致过敏性休克死亡[C]. 2013年中国药学大会暨第十三届中国药师周.2013.
[4] 沈斌, 周月红. 127例中药注射剂不良反应分析及原因探讨 [J]. 中华中医药杂志,2015,30(2):428.
[5] 祁麟, 文彦丽. 参麦注射液不良反应流行病学特点分析 [J]. 天津药学,2014,26(5):31.
[6] 唐军, 廖享, 童红, 等. 基于气质与PLS-DA对新疆薰衣草精油判别分析 [J]. 计算机与应用化学,2014,31(6):701.
[7] ANZANELLO M J, ORTIZ R S, LIMBERGER R, et al. PLS-DA wavenumber selection for the categorization of medicine samples based on multiple criteria [J]. Forensic Sci Int,2014,242:111.
[8] WONG K H, RAZMOVSKI-NAUMOVSKI V, LI K M, et al. Differentiation of Pueraria lobata and Pueraria thomsonii using partial least square discriminant analysis (PLS-DA) [J]. J Pharm Biomed Anal,2013,84:5.
[9] 田刚, 周明术, 杭永伦, 等. 肿瘤标志对肺癌诊断价值PLS-DA和ANN-MPL模型评估 [J]. 中华肿瘤防治杂志,2014,21(5):380.
[10] 蒋程, 韩冰, 吕健刚, 等. 基于偏最小二乘-判别分析算法的癌痛规范化治疗评价 [J]. 中国临床药理学与治疗学,2015,20(12):1378.
[11] 阚全程. 医院药学高级教程 [M]. 北京: 人民军医出版社,2014.
[12] 蒋程, 丁静, 吕健刚, 等. 基于偏最小二乘-判别分析算法的参麦注射剂不良反应关键影响因素分析 [J]. 中国临床药学杂志,2017,26(4):221.