1.法规、指南和标准

1.1.中国GMP附录 无菌药品

第九条 无菌药品生产所需的洁净区可分为以下4个级别：A级：高风险操作区，如灌装区、放置胶塞桶和与无菌制剂直接接触的敞口包装容器的区域及无菌装配或连接操作的区域，应当用单向流操作台（罩）维持该区的环境状态。单向流系统在其工作区域必须均匀送风，风速为0.36-0.54m/s（指导值）。应当有数据证明单向流的状态并经过验证。

第三十二条：在任何运行状态下，洁净区通过适当的送风应当能够确保对周围低级别区域的正压，维持良好的气流方向，保证有效的净化能力。

第三十三条：应当能够证明所用气流方式不会导致污染风险并有记录（如烟雾试验的录像）。

1.2.欧盟GMP附录 无菌药品

4.15 Airflow patterns within cleanrooms and zones should be visualised to demonstrate that there is no ingress from lower grade to higher grade areas and that air does not travel from less clean areas (such as the floor) or over operators or equipment that may transfer contamination to the higher grade areas. Where unidirectional airflow is required, visualisation studies should be performed to determine compliance, (see paragraphs 4.4 & 4.19). When filled, closed products are transferred to an adjacent cleanroom of a lower grade via a small egress point, airflow visualization studies should demonstrate that air does not ingress from the lower grade cleanrooms to the grade B area. Where air movement is shown to be a contamination risk to the clean area or critical zone, corrective actions, such as design improvement, should be implemented. Airflow pattern studies should be performed both at rest and in operation (e.g. simulating operator interventions). Video recordings of the airflow patterns should be retained. The outcome of the air visualisation studies should be documented and considered when establishing the facility's environmental monitoring programme.

洁净室及洁净区内的气流模式应当进行可视化呈现，以证明不存在从较低级别区域到较高级别区域的空气流入情况，且空气不会从较不洁净的区域（比如地面）流动过来，也不会流经操作人员或设备（这些可能会将污染物传播到较高级别区域）。若要求单向气流，应当开展可视化研究以确定是否符合要求（见第 4.4 和 4.19 款）。当已灌装、密封的产品通过一个小出口转移至相邻的较低级别洁净室时，气流可视化研究应当证明空气不会从较低级别洁净室流入 B 级区域。若气流流动被证实对洁净区或关键区域存在污染风险，则应当采取诸如改进设计之类的纠正措施。气流模式研究应当在静态及运行状态下（例如模拟操作人员干预的情况）均开展。气流模式的视频记录应当予以留存。气流可视化研究的结果应当记录在案，并在制定设施的环境监测方案时加以考虑。

4.20 The background environment for isolators or RABS should ensure the risk of transfer of contamination is minimized.

隔离器或RABS的背景环境应确保将污染转移的风险降至最低。

i. Isolators:

隔离器：

c. Airflow pattern studies should be performed at the interfaces of open isolators to demonstrate the absence of air ingress.

应在开放式隔离器的接口处进行气流流型研究，以证明没有空气进入。

ii. RABS:

The background environment for RABS used for aseptic processing should correspond to a minimum of grade B and airflow pattern studies should be performed to demonstrate the absence of air ingress during interventions, including door openings if applicable.

用于无菌工艺的RABS的环境应至少为B级，并且应进行气流流型研究以证明在干预过程中没有空气进入，包括开门（如适用）。

7.18 Activities in clean areas that are not critical to the production processes should be kept to a minimum, especially when aseptic operations are in progress. Movement of personnel should be slow, controlled and methodical to avoid excessive shedding of particles and organisms due to over-vigorous activity. Operators performing aseptic operations should adhere to aseptic technique at all times to prevent changes in air currents that may introduce air of lower quality into the critical zone. Movement adjacent to the critical zone should be restricted and the obstruction of the path of the unidirectional (first air) airflow should be avoided. A review of airflow visualisation studies should be considered as part of the training programme.

洁净区中对生产过程不重要的活动应保持在最低限度，特别是在进行无菌操作时。人员移动应缓慢、受控并有条不紊，以避免由于过度活动导致的微粒和微生物的过量脱落。执行无菌操作的操作人员应始终遵循无菌技术，以防止气流的变化可能将较低质量的空气引入关键区域。应限制关键区附近的移动，应避免单向流（初始气流）通路的阻塞。应考虑将气流可视化研究的回顾作为培训计划的一部分。

8.126 Points to consider for the design of loading (and unloading, where the lyophilised material is still unsealed and exposed), include but are not limited to:

装载（如果冻干后物料仍未密封并暴露，还包括卸载）的设计要点，包括但不限于：

iii. Airflow patterns should not be adversely affected by transport devices and venting of the loading zone.

转移装置和装载区的通风不应对气流模式造成不良影响。

9.22 Where aseptic operations are performed, microbial monitoring should be frequent using a combination of methods such as settle plates, volumetric air sampling, glove, gown and surface sampling (e.g. swabs and contact plates). The method of sampling used should be justified within the CCS and should be demonstrated not to have a detrimental impact on grade A and B airflow patterns. Cleanroom and equipment surfaces should be monitored at the end of an operation.

当进行无菌操作时，应采用多种方法经常进行微生物监测，例如沉降碟、定量空气釆样、手套、洁净服和表面采样（例如棉签擦拭和接触碟）。所使用的采样方法的合理性应在CCS中进行论证，并证明不会对A级和B级的气流模式造成不利影响。在操作结束后应对洁净室和设备表面进行监测。

1.3.无菌工艺模拟试验指南（无菌制剂）

6.1.4 无菌生产区域的气流及环境达到了设计要求，并能稳定运行。但不得采用对环境或者器具进行过度灭菌或消毒的方式提高无菌保证水平。

1.4.无菌工艺模拟试验指南（无菌原料药）

6.1.4 无菌生产区域的气流及环境达到了设计要求，并能稳定运行。依据规定的消毒方法和频次进行环境消毒，应避免消毒剂的过度使用。

1.5.FDA ：无菌加工生产的无菌药品 GMP 指南

Proper design and control prevents turbulence and stagnant air in the critical area . Once relevant parameters are established, it is crucial that airflow patterns be evaluated for turbulence or eddy currents that can act as a channel or reservoir for air contaminants （e.g., from an adjoining lower classified area）. In situ air pattern analysis should be conducted at the critical area to demonstrate unidirectional airflow and sweeping action over and away from the product under dynamic conditions. The studies should be well documented with written conclusions, and include evaluation of the impact of aseptic manipulations （e.g., interventions） and equipment design. Videotape or other recording mechanisms have been found to be useful aides in assessing airflow initially as well as facilitating evaluation of subsequent equipment configuration changes. It is important to note that even successfully qualified systems can be compromised by poor operational, maintenance, or personnel practices.

合理的设计与管控可防止关键区域出现湍流和空气滞留现象。一旦确定了相关参数，对气流形态进行评估就至关重要，要查看是否存在可能成为空气污染物（例如来自相邻较低洁净级别区域的污染物）传播通道或聚集之处的湍流或涡流。应当在关键区域进行现场气流形态分析，以证实在动态条件下气流是单向的，并且能对产品上方进行吹扫并使其远离产品。这些研究应当做好详细记录并附上书面结论，还要包含对无菌操作（例如人工干预情况）以及设备设计所产生影响的评估。实践发现，录像或其他记录手段有助于初步评估气流情况，也便于对后续设备配置变更进行评价。需要着重指出的是，即便那些已成功通过验证的系统，也可能会因操作、维护不善或人员操作不当等情况而受到影响。

1.6.ISO 14644-4:2022 洁净室及相关受控环境：设计、建造、启动

单向气流通过供应清洁空气来排出污染空气。气流最常见的是垂直(向下)或水平，但也可以是对角线或向上。重要的设计特点是在工艺核心处保持气流的均匀性。

1.7.GB：GB 50457-2019 医药工业洁净厂房设计规范

气流流型的设计应符合下列要求：

高级别区域A级通过单向流来实现，B级的洁净级别可采用非单向流的紊流空气形式。在这种混合流的洁净室内，气流的形态应从该空间洁净度高的一端流向洁净度低的一端，以避免污染风险。

2.检查缺陷

气流流型测试目前是被 NMPA、WHO 和 EU等各级监管机构提出缺陷项较多的一个项目，GMP现场检查中气流流型常见的问题有以下几方面：

2.1.测试项目不充分或测试项目缺失

Our inspection found that you lacked smoke studies to evaluate whether unidirectional airflow exists on your (b)(4) ointment aseptic processing line.我们的检查人员发现你们的烟雾研究不足以评估 b4 软膏无菌生产线上是否存在单向流。

Your response states that you completed dynamic airflow studies, and you provided three brief smoke study videos. While you state that these studies were conducted under “dynamic” conditions, we note that they still lack an evaluation of operational conditions and aseptic interventions (e.g., reloading tubes and caps; filling the (b)(4)). In addition, the view of the aseptic processing zone was obstructed, and the smoke manifold was not stationary for sufficient time.

你们的回答表明你们完成了动态气流研究，并提供了三个简短的烟雾研究视频。虽然贵公司说这些研究是在“动态”条件下进行的，但我们注意到，它们仍然缺乏对操作条件和无菌干预措施的评估（例如，重新安装胶管和轧盖；灌装(b) (4)）。此外，无菌处理区的视野被遮挡，烟雾在足够长的时间内是不稳定的。

You did not perform smoke studies under “at rest” and “dynamic” conditions to evaluate air flow characteristics of your open Restricted Access Barrier System (RABS). You subsequently released the sterile (b)(4) products manufactured on this aseptic processing line without studies to demonstrate unidirectionalairflow over the exposed sterile product during processing.

贵公司没有在“静态和“动态”条件下进行烟雾研究，以评估贵公司的 ORABS 的气流特性。随后，贵公司放行了在无菌生产线上生产的无菌(b)(4)产品，而没有对在生产过程中对暴露的无菌产品进行单向气流的研究。

You produced and distributed drug product purporting to be sterile without first conducting adequate process simulations. Specifically, your airflow simulations (smoke studies) for filling line(b)(4) did not adequately demonstrate unidirectional airflow during (b)(4) interventions within the (b)(4)-RABS ISO 5 area. For example, airflow was observed flowing outward and upward during (b)(4) interventions. However, the simulation did not adequately depict airflow (the smoke was not visible) to determine whether HEPA filtered air was reaching the critical (b)(4).

你在没有首先进行充分的工艺模拟的情况下生产和销售声称无菌的药品。具体来说，您对灌装线(b)(4)的气流模拟(烟雾试验)没有充分证明在(b)(4)-RABS ISO 5区域内(b)(4)干预期间的单向气流。例如，在(b)(4)干预期间，观察到气流向外和向上流动。然而，模拟没有充分描绘气流(烟雾不可见)，以确定HEPA过滤的空气是否达到关键(b)(4)。

2.2.设计缺陷导致气流不合格

During the airflow analysis (smoke study) of aseptic connections on your (b)(4) equipment inside the laminar air flow (LAF) ISO-5 area, our investigator identified air flow disturbances and turbulence. Under dynamic conditions, air did not sufficiently sweep across and away from sterile con- nections, so the sterility of any product processed under these conditions could be compromised.

在 ISO-5 区域内(b)(4)设备无菌连接的气流分析（烟雾研究）中，我们的研究人员发现了气流扰动和湍流。在动态条件下，空气不能充分地从无菌区扫过和离开，因此在这些条件下生产的任何产品的无菌性都可能受到影响。

Furthermore, in our review of the smoke study, we identified multiple aseptic technique breaches during aseptic connection of the (b)(4) equipment. Your equipment design and aseptic processing operator competencies appear to contribute to the lack of unidirectionality. Aseptic processing equipment should provide for appropriate ergonomics that enable operators to reproducibly conduct aseptic manipulations. In addition, it is critical that your aseptic processing operators have the knowledge and skills to practice strict aseptic techniques. Even operations that have been successfully qualified can be compromised by poor operational, maintenance, or personnel practices.

此外，在我们对烟雾研究的审核中，我们发现了(b)(4)设备在无菌连接期间多次出现无菌操作不符合无菌规范。贵公司的设备设计和无菌操作人员的能力似乎是导致气流缺乏单向性的原因。无菌工艺设备应符合人体工程学，使操作人员能够重现无菌操作。此外，具备无菌知识和技能的工艺操作人员执行严格的无菌技术至关重要。即使是已经成功通过资格确认的操作，也可能会受到操作、维护或人员实践不佳的影响。

During the inspection, our investigator reviewed and noted turbulent airflow in the September 2015 smoke studies (airflow visualization studies) conducted on your aseptic processing line in room (b)(4) where you manufacture (b)(4) and (b)(4) for the U.S. market. This turbulent airflow poses a significant contamination hazard to your product.

检查员在审查你们 2015 年 9 月在用于无菌灌装线生产(b)(4)和(b)(4)的(b)(4)房间中做的烟雾研究（气流流型试验）时发现存在乱流。这种乱流会对你们的产品造成重大污染危害。

In your response, you submitted additional smoke studies conducted in December 2016. Like yourSeptember 2015 studies, the December 2016 smoke studies show turbulent airflow in multiple locations on the aseptic filling line. (FDA, 2017ah)

在你们的回复中，你们提交了 2016 年 12 月所做的其他气流流型研究。同你们 2015 年 9 月的试验一样，2016 年 12 月的烟雾流型研究也显示，在无菌灌装线的多个位置存在乱流。

You have not established that unidirectional airflow exists at the station where the cap is applied to the container. Additionally，your dynamic smoke study videos show turbulent airflow when operators manually (b)(4) the sterile container-closure components into (b)(4) bowls, which are located outside of the filling and sealing enclosure. Operators reach over the (b)(4) bowls while loading sterile container-closure components to overcome a limitation in your current equipment and process design. The ergonomics of these manual manipulations pose a significant hazard in your aseptic processing operation.

你们没有为轧盖操作工位提供单向流。此外，你们的动态烟雾流型研究视频显示，当操作员将无菌密封件手工加料至料斗时出现乱流，这些乱流位于灌装和密封区域外部。操作员将无菌密封件加入料斗时需把手伸过(b)(4)料斗，以克服现有设备和工艺设计的缺陷。这些手工操作的人体工程学问题会给无菌操作造成重大风险。

2.3.测试方法存在问题

The most recent smoke study conducted by your firm to validate the xxx installed in Building xxx room xxx which is used to fill xxx finished drug product vials following xxx was not conducted in a manner that accurately simulated actual processing conditions. Multiple video recordings made during the smoke study conducted on this xxx in September 2021 show the operator responsible for holding and positioning the smoke wand during intervention simulations standing inside of the xxx. These video recordings also show a second operator using the xxx to perform interventions, such as the changing of environmental monitoring settle plates, while the body of the operator holding the smoke wand was standing within approximately 50 cm or less of the location the intervening operator is working. Per your records, xxx lots of xxx finished drug product, totalling approximately xxx vials, have been filled in this xxx and released for commercial distribution in the US. since this smoke study was conducted.

贵公司最近进行的烟雾研究用以验证安装在xxx大楼xxx室的xxx(用于在xxx之后灌装xxx成品药瓶)，并不是以准确模拟实际生产条件的方式进行的。在2021年09月对该xxx进行的烟雾研究过程中制作的多个视频记录显示，进行干预模拟的过程中，负责握住和定位烟雾棒的操作员站在xxx内部。这些视频记录还显示，第二名操作员使用xxx进行干预，例如更换环境监测沉降皿，而手持烟棒的操作员的身体站在距离干预操作员工作位置大约50 cm或更近的地方。根据您的记录，自从这项烟雾研究开展以来，xxx批次的xxx成品药品，总计约xxx瓶，已在此xxx灌装并在美国进行商业销售。

2.4.无菌操作有问题

For example, during the aseptic filling of vials, an operator used restricted access barrier system (RABS) (b)(4) to remove a jammed stopper by reaching over exposed sterile stoppers in the stopper bowl. The RABS (b)(4) disrupted the unidirectional airflow over the stopper bowl, creating a risk for microbial contamination. After the operator removed the jammed stopper, the filling line was restarted, but the affected stoppers were not cleared.

例如，在西林瓶无菌灌装过程中，操作员使用限制进入屏障系统(RABS) (b)(4)通过在胶塞震荡锅中的暴露的无菌胶塞上面移除一个堵塞。RABS (b)(4)破坏了在胶塞震荡锅上方的单向气流，造成了微生物污染的风险。在操作员移除堵塞后，灌装线重新启动，但受影响的塞子没有清除。

2.5.测试未达到标准/法规要求

During the airflow analysis (smoke study) of aseptic connections on your (b)(4) equipment inside the laminar air flow (LAF) ISO-5 area, our investigator identified air flow disturbances and turbulence. Under dynamic conditions, air did not sufficiently sweep across and away from sterile con- nections, so the sterility of any product processed under these conditions could be compromised.

在 ISO-5 区域内(b)(4)设备无菌连接的气流分析（烟雾研究）中，我们的研究人员发现了气流扰动和湍流。在动态条件下，空气不能充分地从无菌区扫过和离开，因此在这些条件下生产的任何产品的无菌性都可能受到影响。

Air flow visualization studies for the (b)(4) line used to aseptically fill (b)(4) for the US market did not meet the acceptance criteria of airflow that is unidirectional and free from turbulence or follow the established execution instructions in the study protocol.

用于美国市场的 (b)(4) 无菌灌装线 (b)(4) 的气流可视化试验不符合单向且无湍流的气流的验收标准，或不符合试验方案中已建立的执行说明。

a. In the area where empty(b)(4) are opened and exposed to the environment there is a gap between the overhead HEPA filters of approximately (b)(4) Raw video footage obtained during the smoke studies of this area show air turbulence and upward flowing air, The raw video footage showing this deficient air flow pattern was not included in the final edited versions of the videos discussed in the validation report

在空的 (b)(4) 打开并暴露于环境的区域，高处的HEPA过滤器之间存在大约 (b)(4) 的间隙，该区域烟雾试验期间获得的原始视频显示空气湍流和向上流动的空气，显示该缺陷气流的视频在并没有包含在验证报告中最终编辑的视频版本中。

b. The videos show upward flowing smokealong the RABS barrier near (b)(4) inside the filling and stoppering RABS. This area is below an approximately (b)(4) gap between the edge of the HEPA filter and the RABS barrier. The validation report did not identify any deficiencies in this area. A similar gap between the RABS barrier and the HEPA filters exists on all (b)(4) sides of the RABS filling barrier. The air flow visualization studies have not thoroughly evaluated this gap.

视频显示，在灌装和加塞RABS内部 (b)(4) 附近，烟雾是沿着RABS屏障向上流动的。该区域在HEPA过滤器的边缘与RABS屏障之间的间隙大约 (b)(4) 下方。验证报告没有发现这方面的任何缺陷。RABS屏障和HEPA过滤器之间的类似间隙存在于RABS灌装屏障的所有侧面 (b)(4) 上。气流可视化试验尚未彻底评估这一缝隙。

c. The videos show upward flowing and turbulent air flow near a gap between the HEPA filter edge and the barrier(b)(4) outside of the filling barrier, near (b)(4) There is an approximately (b)(4) gap between the edge of the HEPA filter and the RABS barrier, This Grade A classified area is used during (b)(4) assembly of the machine, and interventions, The raw video footage showing this deficient air flow pattern was not included in the final versions of the videos discussed in the validation report.

视频显示了HEPA过滤器边缘和灌装屏障外部的屏障 (b)(4) 之间的间隙附近的向上的气流和湍流，在 (b)(4) 附近。在HEPA过滤器的边缘和RABS屏障之间存在大约 (b)(4) 的间隙，该等级A分类区域在机器的组装和干预期间使用 (b)(4) ，验证报告中讨论的视频的最终版本中不包括显示该缺陷气流模式的原始视频片段。

d. The(b)(4) barrier used to open and load empty (b)(4) has a support for the (b)(4) positioned about (b)(4) below the HEPA filter, The smoke studies did not thoroughly evaluate the impact of this support on the air flow in this area.

d. 用于打开和装载空的(b)(4)的(b)(4) 屏障具有用于 (b)(4) 的支撑件，其位于HEPA过滤器下方大约 (b)(4)，烟雾试验没有彻底评估这种支撑对该区域气流的影响。

e. Protocol(b)(4) OA/AFVP/017 for the (b)(4) line states the smoke needs to be introduced by placing the nozzle with the smoke upwards and the nozzle should be moved to cover the entire area of the filter. Raw video files show the smoke nozzle pointed in downward direction and in fixed locations. The final edited videos did not show the smoke from where it was introduced near the filter to the working location.

e.(b)(4)线的方案(b)(4)OA/AFVP/017 指出，需要通过将带有烟雾的喷嘴向上放置来引入烟雾，并且应该移动喷嘴以覆盖过滤器的整个区域。原始视频文件显示烟雾喷嘴指向向下的方向和固定的位置。最终编辑的视频没有显示烟雾从过滤器附近的哪里引入到工作位置的地方。

2.6.气流流行测试结果未能指导环境监测

Your firm failed to establish an adequate system for monitoring environmental conditions in aseptic processing areas (21 CFR 211.42(c)(10)(iv)).

贵公司未能建立一套完备的无菌加工区域环境状况监测系统。

You do not have a scientific rationale for the environmental monitoring sampling locations in aseptic filling Suites (b)(4). You did not include factors such as smoke study findings, number and location of operators, and historical microbial data in your assessment of hazardous points.

贵公司对于无菌灌装室环境监测采样点缺乏科学依据（此处信息暂缺）。在对危险点进行评估时，并未将烟雾研究结果、操作人员数量及位置以及历史微生物数据等因素纳入考量范围。

For example, we found that settling plates are not appropriately placed in critical areas. Your smoke study showed that during set-up and filling, air flows toward the front (when (b)(4) open) or back of the RABS. However, two relevant sampling points were recently eliminated. As a result, these points of increased risk are not monitored.

例如，我们发现沉降平板在关键区域的放置并不恰当。贵公司的烟雾研究显示，在设备安装及灌装过程中，气流会朝着隔离操作器（RABS）的前部（当此处信息暂缺打开时）或后部流动。然而，近期却取消了两个相关采样点。结果就是，这些风险增加的点位未得到监测。