FDA 检查德国第一三共欧洲制剂厂并出具 483,含手动积分等 3 项观察项
FDA 483|德国:第一三共欧洲(Daiichi Sankyo Europe GmbH):一条手动积分,暴露了整个数据完整性体系
FDA 对德国巴伐利亚 Pfaffenhofen 的第一三共欧洲制剂厂 2025 年 7 月 14 日至 18 日检查后出具 483,共列 3 项观察项。首要观察项指 Empower 中手动积分无程序、审计追踪无理由、无授权记录,且流动相与样品配制记录不完整;另两项涉偏差与 OOS 调查不彻底、清洁与压片模具维护失控。
材料译出德国制剂厂 3 项 483 观察项原文并附作者解读,可对照手动积分、偏差调查与清洁状态的检查逻辑。
在 Empower 里,把基线拉开一点、重新积分,只要点几下鼠标。难点从来不是操作,而是点完之后有没有人写下'为什么'。2025 年 7 月 14 日到 18 日,FDA 检查员 Joseph A. Piechocki 在德国巴伐利亚的这家工厂,就顺着这个'为什么',拿到了 3 项观察。
◆ 检查档案
检查时间 | 2025-07-14 至 2025-07-18(5 天) |
设施类型 | Drug Product Manufacturer(制剂生产) |
地点 | 德国巴伐利亚州 Pfaffenhofen a.d. Ilm,Luitpoldstr. 1 |
检查员 | Joseph A. Piechocki, Investigator |
观察项 | 3 项(AMENDMENT 1 版,已删除观察项 2A.2) |
公开时间 | 2026-11-19(FOIA 电子阅览室) |
观察项 1 | 手动积分无理由无授权
Laboratory records do not include complete data derived from all tests, examinations and assay necessary to assure compliance with established specifications and standards.
Specifically,
A. There is no procedure that describes the appropriate use of manual integration for acquired data within the Empower chromatography system nor do manual integration events require explanation to ensure complete rationale for use. For example, during the review of the release data supporting (b)(4) mg Batch (b)(4), specific test samples chromatograms (and (b)(4) test sample overall within the sample set) contained manually integrated peaks with no reason provided within the audit trail, nor was there documented authorization to perform this manual integration.
In addition, during the review of chromatographic results obtained within Empower within the project folder "uplc 0251 a.04", including results supporting the release testing of (b)(4) drug products, approximately (b)(4) % of the individual injections made had been manually integrated.
B. Records related to analytical testing do not include all the information related to the execution of the test. For example, test methods K-PA 0251_I HPLC1 and K-PA 0251 I HPLC2 are used for assay, content uniformity, and related substance determination of (b)(4) drug products for the US market. The executed laboratory records do not include complete information related to the execution of the test, which includes the following:
1.1 Mobile phases and solutions used for testing do not adequately document the preparations. For example, method K-PA 0251I-HPLC2 outlines preparations of (b)(4) solutions and mobile phases with specific volumes. However, the only documentation within the laboratory journal includes the reagent lot numbers.
Test sample and standard solutions used for assay, content uniformity, and degradation products do not document the steps used to prepare the solutions, including sonication, volumetric flasks, solutions used for dilution, and any subsequent dilutions. In addition, the content uniformity sample preparation requires (b)(4). The steps to this preparation are not documented.
中译实验室记录未包含为确保符合既定质量标准与规格而必需的全部检测、检查与含量测定所产生的完整数据。 具体而言: A. 没有程序规定 Empower 色谱系统中已采集数据的手动积分(manual integration)应如何正确使用,手动积分事件也不要求给出说明以确保完整的使用理由。例如,在审核支持 (b)(4) mg 批号 (b)(4) 的放行数据时,部分供试品色谱图(以及该样品集中 (b)(4) 个供试品的整体)存在手动积分的峰,审计追踪中未提供任何理由,也没有文件记录授权进行该手动积分。 此外,在审核 Empower 项目文件夹 "uplc 0251 a.04" 中获得的色谱结果(包括支持 (b)(4) 制剂放行检测的结果)时,约 (b)(4)% 的单次进样曾被手动积分。 B. 与含量测定相关的记录未包含检测执行的全部信息。例如,检测方法 K-PA 0251_I HPLC1 与 K-PA 0251 I HPLC2 用于 (b)(4) 美国市场制剂的含量、含量均匀度与有关物质测定,但已执行的实验室记录未包含检测执行的全部信息,具体包括: 1.1 用于检测的流动相与溶液未充分记录配制过程。例如,方法 K-PA 0251I-HPLC2 规定了 (b)(4) 溶液与流动相的具体体积配比,但实验室日志中唯一的记录只是试剂批号。 用于含量、含量均匀度与降解产物测定的供试品溶液与对照品溶液,未记录配制步骤,包括超声、容量瓶、稀释用溶液及任何后续稀释。此外,含量均匀度样品制备需要进行 (b)(4),该制备步骤未予记录。
解读21 CFR 211.194(a) 要求实验室记录包含'全部检测数据'(complete data)。FDA 与 MHRA 的数据完整性指南都把手动积分列为高风险操作:必须有 SOP 规定何时允许、必须记录理由、必须经授权、必须被审计追踪复核。本例三个要素全缺——无程序、审计追踪里无理由、无授权记录。 更关键的是'量':在某项目文件夹里,约 (b)(4)% 的进样曾被手动积分。这说明它不是偶发个案,而是日常习惯。误区是以为'审计追踪自动记录=合规';其实审计追踪只证明'谁在何时改了什么',不证明'为什么改、谁批准'。 B 部分则是另一层问题:流动相只记试剂批号、样品制备不记超声与稀释步骤,检测过程无法重现(reproducibility)。
风险手动积分可改变积分基线、峰面积与含量结果,且无理由即无法排除'倒推合格'的可能;一旦某批数据被质疑,整段区间的放行数据都可能被判为不可靠。配制步骤缺失,则方法转移(method transfer)与 OOS 调查都失去可复现的原始依据。 整改方向:建立手动积分 SOP:定义允许情形、强制填写理由、双人授权、审计追踪定期复核(建议每批/每月);对 uplc 0251 a.04 文件夹内全部放行数据做回顾性评估;修订记录模板,把流动相与样品制备的每一步(超声时间、容量瓶编号、稀释体积)设为必填项。
观察项 2 | 偏差与OOS调查不彻底
There is a failure to thoroughly review any unexplained discrepancy and the failure of a batch or any of its components to meet any of its specifications whether or not the batch has been already distributed.
Specifically,
A. Deviation 200007918 was initiated during the compression of (b)(4) mg Batch (b)(4) due to tablets having a tendency to stick on the (b)(4) punch while filling container (b)(4). The following deficiencies were noted:
1. The deviation stated that the error may occur sporadically from time to time, however no definitive root cause was identified which could have caused the issue, including but not limited to a dirty punch or a material issue.
2. The corrective action associated was to visually inspect containers (b)(4), which yielded an additional (b)(4) of rejected material. However, there was no heightened scrutiny of the batch to ensure that the batch did not result in any additional defective tablets after the cleaning of the tooling.
3. The deviation stated that this had occurred within the last 12 months, however there was no further assessment performed to determine the number of deviations where this had occurred and if any additional corrective actions needed to be implemented to prevent the repetition of the deviation.
B. OOS investigations for (b)(4) mg Batch (b)(4) (Investigation 200008934) and (b)(4) mg Batch (b)(4) (Investigation 200010548) were initiated due to one sample being outside the specification for content uniformity. The OOS investigations concluded that two tablets must have been used instead of one since the values were doubled for that specific sample. However, there was no documented interview with the analyst to corroborate this root cause determination, nor does the test sample preparation documentation have adequate documentation to support this determination. In addition, the corrective action for both OOS investigations was to retrain the specific analyst which caused the issue. However, the corrective action did not extend to the remainder of the Quality Control personnel or address the specific root cause, nor was a preventive action adequately established to address the root cause identified for having two tablets during the content uniformity to prevent reoccurrence. In addition, based on the results of the OOS, the invalid result was isolated to the individual sample. The original results from the content uniformity test were invalidated without scientific justification.
中译未能对任何不可解释的差异、以及批次或其任何组分不符合质量标准的情形进行彻底审核,无论该批次是否已经销售。 具体而言: A. 在压片过程中启动偏差 200007918,原因是 (b)(4) mg 批号 (b)(4) 在灌装容器 (b)(4) 时片剂易粘附于 (b)(4) 冲头。发现以下缺陷: 1. 该偏差称该差错可能'偶发',但未确定任何可导致该问题的确定根因,包括但不限于冲头不洁或物料问题。 2. 相关纠正措施为对容器 (b)(4) 进行目检,结果又挑出 (b)(4) 的额外不合格物料。但在清洁模具之后,并未对该批次实施加严检查,以确保该批次不再产生其他缺陷片。 3. 该偏差称此事在过去 12 个月内发生过,但未进一步评估此类偏差发生的次数,也未评估是否需要采取额外的纠正措施以防止偏差重复。 B. (b)(4) mg 批号 (b)(4)(调查 200008934)与 (b)(4) mg 批号 (b)(4)(调查 200010548)因一个样品的含量均匀度超出规格而启动 OOS 调查。该 OOS 调查结论认为,由于该特定样品的数值恰好翻倍,应为使用了两片而非一片。但没有对分析员进行书面记录的访谈以支持该根因判定,供试品制备记录也不足以支持该判定。此外,两项 OOS 调查的纠正措施均为对造成该问题的特定分析员重新培训;该纠正措施未扩展至其余 QC 人员,也未针对具体根因,且未充分建立预防措施以解决含量均匀度检测中使用两片的根因、防止再次发生。此外,基于 OOS 结果,无效结果被限定在单个样品;含量均匀度检测的原始结果在缺乏科学依据的情况下被无效化。
解读21 CFR 211.192 要求调查彻底并追溯根本原因。2A 的偏差暴露三个典型缺口:根因未定('偶发'不是根因)、CAPA 只是目检且未对清洁后批次加严检查(等于默认风险已消除)、未统计 12 个月内同类偏差次数(没有趋势视角)。 2B 更敏感:OOS 结论是'数值翻倍,所以用了两片'。这个推理在技术上可能成立,但没有对分析员的访谈记录、没有可支撑的样品制备记录,就只是推测。CAPA 只培训当事人、不扩展到全体 QC,等于把系统问题当成个人问题。 最后一句最重:原始结果在缺乏科学依据的情况下被无效化。FDA 会把这类情形读成'反复检测直至合格'(testing into compliance)。
风险粘冲与模具缺陷未彻底处理,可能持续产生外观/重量差异片,影响含量均匀度与溶出;OOS 数据被无依据作废,则同一产品的放行结论建立在不可信的数据上,一旦被认定,可能触发批次召回与稳定性数据重评。 整改方向:对该偏差补做根因分析(冲头表面状态、物料黏性、压片参数),对清洁后批次实施加严抽检并记录结果;建立偏差趋势看板(同设备/同缺陷/12 个月窗口);修订 OOS 调查模板,强制包含分析员访谈、制备记录复核、复算与假设验证;CAPA 从'培训个人'升级为'系统防错'(如 CU 制备改用单片工装或双人核对)。
观察项 3 | 清洁与模具维护失控
Written procedures are not established and followed for the cleaning and maintenance of equipment, including utensils, used in the manufacture, processing, packing or holding of a drug product.
Specifically,
Equipment, tools, and ancillary containers used in the manufacture of drug products are not adequately labeled or maintained during manufacturing or cleaning. For example, during the walkthrough of the facility on 14 Jul 2025, I observed the following:
1. Inspection of the (b)(4) (Room (b)(4)) equipment noted an apparent (b)(4) residue observed on the exterior of the (b)(4) (confirmed during the inspection as from the (b)(4)), apparent (b)(4) residue remaining on the (b)(4) used for (b)(4) in exterior space outside of the (b)(4) but within the (b)(4), and an apparent (b)(4) residue on the (b)(4) gasket. This equipment was documented as clean following (b)(4) mg Batch (b)(4).
2. During the inspection of the tooling used for compression operations, I observed apparent residue remaining within the debossing of tablet punch set #(b)(4) (for (b)(4) mg tablets), after the documentation showed that the punches were cleaned. In addition, during the inspection of tablet punch set #(b)(4) (used for (b)(4) mg tablets), I observed at least one punch with an apparent dent and one punch with an apparent nick.
Furthermore, the procedure related to the maintenance, cleaning, and inspection of compression tooling does not outline the instances when an investigation would need to be opened for defects observed at different inspection points throughout tooling inspection processes nor is it required to define the type of defects if found prior to when tooling is removed from use.
3. During (b)(4) of (b)(4) on (b)(4) (Room (b)(4)), the (b)(4) are contained within (b)(4). During the walkthrough, the (b)(4) used was not adequately labeled with its contents or identification, was not documented as clean prior to use, and was found open to the environment and not actively being used during the walkthrough.
中译未建立并遵循用于药品生产、加工、包装或贮存的设备(包括器具)清洁与维护的书面程序。 具体而言:药品生产中所用的设备、工具与辅助容器在生产或清洁过程中未被充分标识或维护。例如,在 2025 年 7 月 14 日的厂区巡查中,我观察到: 1. 检查 (b)(4)((b)(4) 房间)设备时,发现 (b)(4) 外表面有明显 (b)(4) 残留(检查期间确认来自 (b)(4))。(b)(4) 外部、但处于 (b)(4) 内部的、用于 (b)(4) 的 (b)(4) 上仍有明显 (b)(4) 残留,且 (b)(4) 垫圈上有明显 (b)(4) 残留。该设备在 (b)(4) mg 批号 (b)(4) 之后被记录为'已清洁'。 2. 在检查压片用模具时,发现 1 号冲头组(用于 (b)(4) mg 片)的字模(debossing)内仍有明显残留,而记录显示冲头已清洁。此外,在检查 2 号冲头组(用于 (b)(4) mg 片)时,发现至少一个冲头有明显凹痕、一个冲头有明显缺口。 而且,与压片模具维护、清洁和检查相关的程序,未规定在模具检查各环节发现缺陷时应在何种情形下启动调查,也未要求在模具停用前对已发现缺陷的类型作出界定。 3. 在 (b)(4)((b)(4) 房间)进行 (b)(4) 期间,(b)(4) 存放于 (b)(4) 内。巡查时发现所用 (b)(4) 未充分标识其内容物或身份,使用前未被记录为已清洁,且处于敞口对环境开放状态,巡查时并未在使用中。
解读21 CFR 211.67 要求设备清洁与维护有书面程序并被遵循,核心逻辑是'防止交叉污染与携带'。这条观察最致命的地方不在残留本身,而在记录与现场不一致:设备被记录为 clean,现场却有残留;冲头被记录为 cleaned,字模里还卡着东西。 FDA 对'清洁状态'的判断从不看记录,只看现场。第 2 点还叠加了模具完整性:冲头有凹痕与缺口,会直接影响片重、外观与崩解,而程序竟未规定'什么情况下必须开偏差'——等于把判断权交给了个人。第 3 点是一个敞口、无标识、无清洁记录的容器,这是交叉污染最经典的入口。
风险残留物可能带入下一批,造成交叉污染与有关物质异常;字模残留会形成片剂表面缺陷,缺口冲头可产生金属异物或片重超差;敞口无标识容器则使清洁状态与内容物均不可追溯,患者端风险为异物与效价偏差。 整改方向:对观察涉及的设备与模具重新清洁并做目视+擦拭(TOC/微生物)确认;修订模具管理程序,规定凹痕/缺口的判定标准、强制偏差触发点与停用流程;对清洁后设备实行标签(清洁状态+有效期+执行人)与上锁管理;增加'清洁后现场抽查'作为批放行的前置条件。
整体诊断
这 3 项观察表面分属实验室、调查与设备,实质是同一条链:记录无法回答'为什么'。手动积分有操作无理由(数据完整性),OOS 有结论无证据(调查完整性),清洁有签名无状态(现场真实性)。三者都指向'纸面合规、实质缺失'的模式化执行。
值得注意的是,这是一家位于德国的制剂厂,处于欧盟 GMP 体系之下,且该 483 为 AMENDMENT 1 版本——观察项 2A.2 因偏差文件中的证据被删除,说明沟通仍有空间;但 3 项核心未被撤销。对集团层面而言,Empower 的全局治理(是否所有 site 都缺手动积分 SOP)与偏差/OOS 模板化答复,是更值得自查的系统性弱点。
给中国药企的启示
01手动积分必须'三件套':SOP 允许情形、逐次理由、授权记录。审计追踪只解决'看得见',不解决'说得通'。
02OOS 的根因不能靠推测:访谈记录、制备记录、复算验证,缺一即不成立;只培训一个人,不叫 CAPA。
03清洁状态是现场事实,不是纸面结论——记录写 clean,现场就要经得起翻垫圈、翻字模。
本文基于 FDA 公开的 Form 483 原文整理,译文与解读为作者观点,不构成合规结论;如需原文请以 FDA 官方 PDF(fda.gov)为准。企业名与 FEI 均来自 FDA 公开数据。
来源:PharmaBuddy · mp.weixin.qq.com