PDA TR88 制药行业微生物数据偏差调查

或许可以从这个角度去理解微生物实验室的数据可靠性

根据多次经历FDA检查经验，审计官是直接冲到微生物检查实验室，核对培养基数量是否与台账一致，然后核对最近的计数数据，实际观察培养基长菌情况，检查仪器使用记录。

由于目前微生物实验室相关操作主要由人工进行操作，其中部分仪器涉及数据完整性部分包括：

1.粒子计数器部分是有完整审计追踪功能，人员权限等功能，在使用和验证过程中需注意完整性验证及定期数据更新备份。

2.培养箱，灭菌锅等设备有温湿度的记录，偏离报警等功能，需注意此类设备的备份，核对其开关门与仪器使用记录填写时间是否对应。

3.部分微生物实验室会配置有视频和拍照记录功能的仪器设备，作为实验结果的辅助记录，也许考察其备份。但注意视频记录会有人员隐私问题，防止引发此类争端。

USP 1117 DATA INTEGRITY OF MICROBIOLOGICAL DATA

Microbiological methods are classically performed manually and on the basis of visual evaluation by an analyst performing
the test. Therefore, the interpretation of test results or the number of colonies tested may be prone to a certain subjectivity and
variability. To further improve data integrity and reduce subjectivity, alternative methods for the reading of plates, such as the
use of automated plate readers or high-resolution photographs of the plate, may be used. However, these systems can have
inherent challenges such as difficulty with the following: counting colonies embedded in the agar gel from pour-plated dishes,
counting satellite colonies, differentiating overlapping colonies, differentiating particles from colonies, and interpreting the
photo consistently from one individual to another. Automated enumeration methods that stack images to capture colonies
growing in time may overcome some of these challenges.
It should be noted that, in order to avoid jeopardizing the asepsis status required during testing, contemporaneous recording
of actions during execution of the microbiological testing cannot be followed in all cases. It is acceptable for recording of actions
to take place immediately after the working session upon exiting the UDAF hood instead of constantly having to go back and
forth from aseptic to non-aseptic areas to write down test actions executed.
Counts from Petri plates are considered original data on the day that the method requires the plates to be read and recorded.
After reading, if these same plates are subsequently stored at room temperature or under refrigeration, it is not possible to
confirm the original results because the microbial counts may increase during storage. Many microbial colonies continue to
grow during refrigeration but at a slower rate. The colony counts derived from conducting a microbe test using a compendial
method depend on adherence to the incubation time and temperature stated in the prescribed method.

总结概述如下：

1. 减少主观性和变异性

• 传统微生物检测方法依赖人工操作和视觉评估，容易产生主观性和结果差异

• 可采用自动化平板读数仪或高分辨率拍照等替代方法来提高客观性

• 自动化计数系统通过时间序列图像堆叠技术，有助于克服菌落计数的技术挑战

2. 技术挑战与应对

• 自动化系统需要解决以下技术难题：

◦ 倾注平板中嵌入琼脂凝胶的菌落计数

◦ 卫星菌落的识别和计数

◦ 重叠菌落的区分

◦ 颗粒与真实菌落的鉴别

◦ 确保不同人员对图像解释的一致性

3. 记录时间的灵活性

• 考虑到无菌操作要求，并非所有操作都能实时记录

• 允许在完成工作离开超净工作台后立即记录，避免在无菌区和非无菌区间频繁往返

• 这种灵活性不影响数据完整性，但需确保记录的及时性

4. 原始数据的确认与保存

• 平板计数在规定的读取日被视为原始数据

• 读取后如继续室温或冷藏保存，无法确认原始结果的准确性

• 微生物在冷藏条件下仍会缓慢生长，导致菌落计数发生变化

5. 严格遵守培养条件

• 药典方法的微生物计数结果依赖于严格遵守规定的培养时间和温度

• 任何偏离都会影响结果的准确性和可比性

这些管理要点强调了在保证无菌操作的前提下，通过技术改进、合理的记录程序和严格的实验条件控制来确保微生物数据的完整性和可靠性。

微生物检测目前大多都是手工操作的，所以微生物的数据完整性更多靠制度和复核去解决，只对辅助设备职环境设备、培养设备等及纸质记录、打印记录去做重点管理。