数据完整性是FDA新的检查重点

近些年，FDA警告信中关于数据造假和欺骗的内容越来越多，尤其是使用了色谱数据系统CDS的GMP分析实验室。由2007年Able 实验室造假事件引发，在FDA批准前检查过程中需要对数据完整性进行审计的要求就被写进了CPG7346.832 。而且，FDA的检查官都被培训过计算机系统和它们包含的数据。

2014年，三项新的内容加入到了FDA的关于CGMP的Q&A网页中，它们反映了发展现状：

问题5：为什么FDA关心电脑系统的共用账户?

问题6：主生产和控制记录中，电子签名可以取代手写签名么?

问题7：在对公司的警告信中，为什么FDA反对使用实际的样品来进行系统适应性测试(有时被称作“试验”、“测试”、“prep”运行)?

以前-早在2010年-FDA已通过问题3说明FDA希望能够看到HPLC色谱的电子原始数据。仅仅保留打印的色谱图已经无法满足CGMP的要求了，因为太多的重要信息可能丢失。

这些背景情况揭示了FDA检查过程中发现越来越多的数据完整性问题的原因。目前，很多警告信中可以看到这种情况，尤其是2015年1月。现在，典型的缺陷就是所谓的“试验样品的进样操作”，通过这个，公司可以掩盖一些OOS结果。例如，可以看看Apotex公司2015年1月30日收到的警告信和Micro Labs Limited2015年1月9日收到的警告信。

更多信息可以登录FDA's Q&A webpage on Current Good Manufacturing Practices。

In recent years, the number of complaints in FDA Warning Lett ers with regard to data falsification and fraud - especially in analytical GMP laboratories regarding the use of chromatography data systems (CDS) - has been increasing. Triggered by the cases of fraud at Able Laboratories in 2007, the requirements for audits of data integrity during FDA Pre-Approval inspections have been set in the Compliance Programme Guide (CPG) 7346.832. Moreover, FDA's inspectors have been explicitly trained on computer systems and the data they contain.

In 2014, three new contributions were added to FDA's Q&A webpage on Current Good Manufacturing Practices which reflect those current developments:

Question 5: Why is FDA concerned with the use of shared login accounts for computer systems?

Question 6: Can electronic signatures be used instead of handwritten signatures for master production and control records?

Question 7: In warning letters to firms, why has FDA objected to the practice of using actual samples to perform system suitability testing (sometimes also referred to as "trial," "test," or "prep" runs)?

Previously - already in 2010 - clarification had been provided through question 3 about the fact that the FDA always expects the electronic raw data for HPLC chromatograms. Keeping the printed chromatogram only is no longer sufficient to meet the CGMP expectations, since too many significant pieces of information would get lost.

This background explains the increased number of complaints about data integrity observed during FDA inspections. At present, this can be seen in many Warning Letters, especially in January 2015. Currently, typical findings are so-called "trial sample injection practices" through which companies try to disguise unfavourable OOS results. For example, please see the Apotex Warning Letter from 30 January 2015 and the Warning Letter for Micro Labs Limited from 9 January 2015.

Further, please also access FDA's Q&A webpage on Current Good Manufacturing Practices.