Vrije Universiteit Brussel, Department of Analytical Chemistry and Pharmaceutical Technology, Laarbeeklaan 103, 1090 Brussels, Belgium.
Anal Chim Acta. 2009 Nov 23;655(1-2):43-51. doi: 10.1016/j.aca.2009.09.050. Epub 2009 Oct 8.
The development of a new drug substance is an expensive and time-consuming process. Therefore, the developers want to maximize the profit from the drug by patenting the concerned molecule as well as its synthesis pathway. In a later stage a faster or cheaper manufacturing process can be developed and patented. The aim of this study is to recognize paracetamol-containing drug formulations in relation to their synthesis pathways, in order to demonstrate the possibility to reveal fraudulently synthesized paracetamol. Since different synthesis pathways require different starting materials, solvents, reagents and catalysts and since they can produce different intermediates, it is expected that drug products originating from a different synthesis pathway will exhibit a different impurity profile. Therefore, in this study several paracetamol samples, synthesized in four different ways, are analyzed using trace-enrichment high-performance liquid chromatography (HPLC). The resulting chromatographic data were chemometrically treated in order to reveal clustering tendencies in the hope of distinguishing the different pathways. When performing principal component analysis (PCA) only 3 vaguely outlined clusters appeared. Projection pursuit (PP) was able to reveal 4 clusters and the samples with known synthesis pathway, except one, were classified in the different clusters. When hierarchical clustering and auto-associative multivariate regression trees (AAMRT) were applied, the samples of the four synthesis pathways could also be distinguished. AAMRT has an added value, since it can indicate the variables (peaks and thus also the impurities) that are responsible for the differences between the samples synthesized differently.
新药物的研发是一个昂贵且耗时的过程。因此，研发者希望通过对相关分子及其合成途径进行专利保护，从药物中获得最大的利润。在后期，可以开发并专利保护更快或更便宜的制造工艺。本研究旨在识别含有对乙酰氨基酚的药物制剂与其合成途径之间的关系，以证明揭示欺诈性合成对乙酰氨基酚的可能性。由于不同的合成途径需要不同的起始材料、溶剂、试剂和催化剂，并且可以产生不同的中间体，因此预计来自不同合成途径的药物产品将表现出不同的杂质特征。因此，在这项研究中，使用痕量富集高效液相色谱（HPLC）分析了以四种不同方式合成的几种对乙酰氨基酚样品。对所得的色谱数据进行了化学计量学处理，以揭示聚类趋势，希望能够区分不同的途径。当进行主成分分析（PCA）时，仅出现了 3 个模糊的聚类。投影寻踪（PP）能够揭示 4 个聚类，除了一个之外，具有已知合成途径的样品被分类到不同的聚类中。当应用层次聚类和自联想多元回归树（AAMRT）时，也可以区分四种合成途径的样品。AAMRT 具有附加值，因为它可以指示导致不同样品之间差异的变量（峰，也就是杂质）。
