Hofer Jeffrey D, Olsen Bernard A, Rickard Eugene C
Analytical Sciences Research & Development, Eli Lilly & Company, Lilly Corporate Center, Indianapolis, IN 46285, USA.
J Pharm Biomed Anal. 2007 Aug 15;44(4):906-13. doi: 10.1016/j.jpba.2007.04.023. Epub 2007 Apr 24.
HPLC is a generally accepted method for assay of drug substances. However, recent claims cast doubts on the utility of HPLC assay methods for characterizing quality [S. Görög, J. Pharm. Biomed. Anal. 36 (2005) 931-937]. This study examines the utility of the traditional drug substance HPLC assay as a quality control parameter. HPLC assay data from more than 100 batches for each of eight drug substances were compared to results from a mass balance approach (100-impurities%). Estimates of the variability of HPLC assays from our data and from the literature ranged from 0.6 to 1.1% R.S.D. This variability is an appreciable portion of a typical acceptance range (e.g., 98.0-102.0%) and frequently exceeds the variability of the manufacturing process. Therefore, the results of the HPLC assay are questionable at best to determine the acceptability of the drug substance batch. The high variability also can generate a significant percentage of false out-of-specification (OOS) results, even when the "true" purity is 99.0-100.0%. Each false OOS leads to inefficiencies because of unwarranted investigations for a root cause and/or implementation of countermeasures for a problem that does not exist. Lastly, low precision makes it nearly impossible to detect significant changes in the process mean and/or degradation during a stability study. The use of a mass balance approach for assay retains essentially the same average results as the HPLC assay but gives standard deviations that are up to 10 times less. Monitoring the assay by mass balance allows for more precise process and stability monitoring and facilitates more rapid and accurate identification of process changes.
高效液相色谱法（HPLC）是一种普遍认可的原料药分析方法。然而，最近有观点对HPLC分析方法在表征质量方面的实用性提出了质疑[S. 戈罗格，《药物分析杂志》36 (2005) 931 - 937]。本研究考察了传统原料药HPLC分析作为质量控制参数的实用性。将八种原料药中每种超过100批的HPLC分析数据与质量平衡法（100% - 杂质%）的结果进行了比较。根据我们的数据以及文献中HPLC分析的变异性估计，相对标准偏差（R.S.D.）范围为0.6%至1.1%。这种变异性在典型的合格范围（例如98.0 - 102.0%）中占相当大的比例，并且经常超过生产过程的变异性。因此，HPLC分析结果充其量只能勉强用于确定原料药批次的可接受性。即使“真实”纯度为99.0 - 100.0%，高变异性也可能导致相当比例的假不合格（OOS）结果。每个假OOS结果都会导致效率低下，因为会对不存在的根本原因进行不必要的调查和/或针对不存在的问题采取对策。最后，低精密度使得在稳定性研究期间几乎不可能检测到过程均值和/或降解的显著变化。采用质量平衡法进行分析，其平均结果与HPLC分析基本相同，但标准偏差可降低至十分之一。通过质量平衡监测分析可以实现更精确的过程和稳定性监测，并有助于更快速准确地识别过程变化。
