目的 评价国内劳拉西泮片生产过程的控制能力。方法 利用近红外光谱法结合聚类分析和主成分分析法表征国内生产的劳拉西泮片不同生产工艺和工艺控制空间;建立近红外光谱通用定量模型,获取样品每个独立单位(片)的含量预测值,通过单变量统计分析获得描述批内、批间工艺均值和变异以及分布情况的统计量,用于相同产品不同工艺的评价。结果 聚类分析和主成分分析法对劳拉西泮片的3个不同生产工艺进行分类表征,主成分2、3重构的工艺控制空间显示B厂家的工艺比A厂家的工艺具有较小的变异范围;劳拉西泮通用定量模型主成分数5,r²值93.89%,偏差-0.008 56;劳拉西泮含量统计分布结果显示,27个批次中有9个批次具有较大的批内差异,B厂家比A厂家具有较小的批间差异。结论 本实验所建方法能反应劳拉西泮片的不同工艺控制水平,为仿制药一致性评价提供一种有效的药品质量一致性评价方法。

OBJECTIVE To evaluate the process control capability of lorazepam tablets produced in China. METHODS Near-infrared spectroscopy combined with cluster analysis(CLA) and principal component analysis(PCA) were used to characterize the different processes and process control space of lorazepam tablets produced in China. Universal quantitative model was built to obtain the content predictions of individual units(tablets), on base of which process mean value,intra-batch and inter-batch differences and distribution status were calculated by univariate statistics analysis methods. RESULTS Three different manufacturing processes of lorazepam tablets were characterized by both CLA and PCA. The process control spaces reconstructed by the second and third principal components indicated that the process of manufacturer B had smaller variation than that of manufacturer A. The universal quantitative model had a principal component number of 5, r² square value of 93.89% and bias of -0.008 56. The statistic distribution of API contents showed that 9 batches out of the total 27 batches had relative lager intra-batch differences and manufacturer B had better inter-batch differences than manufacturer A. CONCLUSION The method this study established can reveal the control levels of different processes of lorazepam tablets, which provide an efficient quality consistency evaluation means for generic drug consistency assessment.