FDA 对 Tailstorm Health 钱德勒厂签发 5 条 483 观察项，文章逐条解读

警告信、召回之后再看这台灭菌柜：Tailstorm 的 5 条 483，把"氮气、水、灭菌"三条命脉全点了名

2026 年 8 月 28 日，FDA 两名检查员在亚利桑那州 Chandler 结束为期两周的复查，给 Tailstorm Health（dba Medivant Healthcare）签下5 条观察项。这家企业上一次进入公众视野，是 2024 年 11 月因"无菌保证不足"自愿召回 6 批贝伐珠单抗（Avastin）眼内注射针；2025 年 4 月 8 日，FDA 发出警告信 WL #703497。而这次签字的，正是 2025 年 5 月在同一场地签过 483 的同一对检查员。

互动 同一对检查员、同一个场地、相隔 15 个月——如果第二次来还是这些问题，说明什么？

一、背景：从 Avastin 召回，到 15 个月后的第二次复查

这家公司是做什么的。Tailstorm Health, Inc.，商业名Medivant Healthcare / Medivant Health，在亚利桑那有两个场地：凤凰城（24416 N. 19th Avenue Suite 200）与本次受检的钱德勒厂（158 S Kyrene Rd, Chandler, AZ）。2019 年 10 月 23 日首次注册为 503B 外包设施。它的业务定位是"平价无菌注射剂"——预充式注射器、单剂量西林瓶，产品包括氯胺酮 10%、利多卡因/肾上腺素、布比卡因/肾上腺素、硫酸沙丁胺醇、劳拉西泮注射液等麻醉、急救与常用注射剂。这类产品走的是医院与诊所的常规用药通道，批量大、批次多。

签字的是谁。483 发给执行副总裁 Mitesh M. Gandhi（2025 年 5 月那份 483 上他的职务是运营与财务副总裁）。签署人两名：Nicholas L. Hunt（高级消费者安全官员，Senior CSO）与Ronda R. Loyd Jones（调查员），来自丹佛地区办公室（6th & Kipling St.）。这两人正是 2025 年 5 月 23 日钱德勒厂那一份 483 的签署人——15 个月后重回同一场地，对"上次说的改了没有"最有发言权。

监管史：警告信 + 召回。时间线：2024 年 9 月 24 日至 10 月 10 日，FDA 检查凤凰城厂，10 月 10 日签发 483 及修订版 483；2024 年 11 月 6 日，企业自愿召回 6 批贝伐珠单抗 1.25mg/0.05mL 单剂量注射器（批号 D24005、D24006、D24007、D24008、D24009、D24012），原因是缺乏无菌保证；2025 年 4 月 8 日，FDA 发出警告信 WL #703497（收件人 CEO Viraj Gandhi），认定产品不符合 503B 豁免条件（标签缺项）并存在 cGMP 缺陷与不卫生条件。2025 年 5 月 14–23 日钱德勒厂检查，5 月 23 日签 483（悬浮粒子超标、OOS 调查不充分、QC 验收标准不足，后两项为重复项）。2025 年 11 月 3–14 日凤凰城厂再检，签 483（实验室系统适用性、审计追踪审核缺失；灌装贝伐珠单抗时 ISO 灌装间天花板出现不明白色线头）。

这次检查怎么跑的。2026 年 8 月 17 日至 28 日，10 个工作日。检查员看的是三条"命脉"：灭菌柜（PE001）、制药用水与氮气系统、实验室控制。现场目击包括：灌装利多卡因 2% 注射液（批号 2608002）时 ISO 7 缓冲间内的人数、用水系统的排水软管与地漏、brown residue、氮气连接点。也就是说，这次的问题不是从文件里读出来的，很大一部分是站在设备前看到的。

二、逐条拆解：5 条，英文原文 + 中文还原

（以下每条中文表述均以 483 原文为据，并按上下文补全原文未言明之处以便阅读；(b)(4) 为 FDA 官方脱敏处，照原文保留；依上下文推测的内容以「（推断）」标出，非原文内容。）

观察项 1｜洁净室"动态确认"是在比日常轻松得多的状态下做的

Separate or defined areas to prevent contamination or mix-ups are deficient regarding operations related to aseptic processing of drug products. Specifically, (b)(4) operational/dynamic certification of the (b)(4) Vial Filling, Stoppering and Sealing Machine, ISO 5 (b)(4) Filling Line (b)(4) cleanroom suite does not represent routine production operations. For example, ISO 5 Filling Line (b)(4) was not fully assembled or in-operation, and there was one cleanroom technician outside the (b)(4) during dynamic nonviable particulate (NVP) monitoring. We observed up to (b)(4) technicians inside the ISO 7 buffer room for the ISO 5 (b)(4) during filling of Lidocaine 2% Injection, 10 mL vial, batch 2608002. Records reviewed included: Project AZ-MVDT261801C-1, (b)(4) Cleanroom Testing and Certification, dated April 2026; and Project AZ-MVDT251301C-1, Cleanroom Suite Certification, dated October 2025. All sterile drugs are manufactured on ISO 5 Filling Line (b)(4). The personnel, line setup, and production operations inside the ISO 5 (b)(4) are the same for aseptically filled and (b)(4) sterilized drugs. Examples of aseptically filled drugs include Lorazepam Injection USP, 2 mg/mL, lot 2510002; and Lorazepam Injection USP, 2 mg/mL, lot 2511001.

对灌装—压塞—封口机与 ISO 5 灌装线所在洁净套间所做的"运行/动态"确认，不代表常规生产状态：做确认时，ISO 5 灌装线既没有完全组装、也没有开机运行，动态悬浮粒子采样期间只有一名洁净技术员在 (b)(4) 外面。而检查员现场看到的是另一番景象——灌装利多卡因 2% 注射液（10mL 西林瓶，批号 2608002）时，ISO 7 缓冲间里最多有 (b)(4) 名技术员。

被审核的记录包括 2026 年 4 月的 Project AZ-MVDT261801C-1 与 2025 年 10 月的 Project AZ-MVDT251301C-1 两份洁净室确认报告。需要强调的是：企业所有无菌药品都在 ISO 5 灌装线 1 上生产，而且无论是无菌灌装的产品还是 (b)(4) 灭菌的产品，人员、装线和生产操作都是同一套（无菌灌装的例子包括劳拉西泮注射液 USP 2mg/mL，批号 2510002 与 2511001）。

FDA 的真实关切：洁净室分级确认必须在"最差的实际状态"下做——设备全装好、在跑、人最多的时候。用"半装配 + 一个人"的状态去代表"满负荷 + 多人"的日常，等于把最容易出问题的场景排除在验证之外。法规：21 CFR 211.42、211.113(b)；ISO 14644 / USP <1116> 理念。

观察项 2｜灭菌柜、培养基灌装、人员监测：三处都少了"证明力"

Procedures designed to prevent microbiological contamination of drug products purporting to be sterile are not established, written and followed. Specifically, Autoclave Qualification and routine monitoring of the (b)(4) equipment ID #PE001, is deficient. A. There is inadequate written justification to place (b)(4) in the (b)(4) of the (b)(4) as the only process verification for routine sterilization loads. The rationale for the location of the (b)(4) within the (b)(4) is inadequate. B. Operators do not use Biological Indicators (BIs) (b)(4) sterilization (b)(4). The approved procedures only require BIs during (b)(4) qualification of (b)(4) loads. The most recent (b)(4) qualification is documented in )4.PQ.0004, Performance Qualification-(b)(4) Report, (b)(4) PE001, report approved 06/19/26. C. Your procedures only require a (b)(4) air removal test during (b)(4) qualification of the (b)(4). Technicians do not perform (b)(4) tests before any routine (b)(4) cycles. There is inadequate written justification for the frequency of air removal tests. Technicians use (b)(4) PE001 to (b)(4) sterilize finished drug products and sterilize reusable product-contact parts, (b)(4) tubing, (b)(4), utensils and other items used during production of all commercial sterile drugs on ISO 5 Filling Line (b)(4). Examples of drugs manufactured with product-contact equipment sterilized in PE001 and then (b)(4) sterilized in PE001 include: Ketamine 10% USP, 100 mg/mL, 5 mL, SDV, lot 2602003; Lidocaine HCl 1% / Epinephrine Bitartrate 10 mcg/mL Injection USP, 10 mL, SDV, lot 2603001; Albuterol Sulfate 0.5%, 10 mL, lot 2512002; and Bupivacaine Hydrochloride 0.5% / Epinephrine Bitartrate 5 mcg/mL Injection, 10 mL vial, lot 2608001. Media Fill D. Operators do not verify the (b)(4) of prepared (b)(4) culture medium used in aseptic process simulation media fills on ISO 5 Filling Line (b)(4). There is inadequate assurance that the (b)(4) is suitable to validate the aseptic filling process for Lorazepam Injection. We reviewed the batch record for Media Fill, batch CHMF/26/001, completed (b)(4), approved 06/11/26. E. Technicians do not use BIs for routine or periodic monitoring of the (b)(4) cycle used for (b)(4) of ISO 5 Filling Line (b)(4) and the ISO 7 buffer room before filling. F. The (b)(4) surface remains folded inside the open (b)(4) pouch during the (b)(4) cycle for ISO 5 Filling Line (b)(4) and the ISO 7 buffer room. There are no BIs or (b)(4) placed inside the partially open pouches to demonstrate the (b)(4) adequately (b)(4) the (b)(4) that are exposed to the environment during installation. Personnel monitoring G. Personnel monitoring specifications for (b)(4) hands, forearms, chest, and forehead are not scientifically justified. Your firm uses the ISO 7 microbial monitoring limits ((b)(4) CFU/plate alert and (b)(4) CFU/plate action) to evaluate samples from operators that use (b)(4) to perform closed-door interventions in the ISO 5 Filling Line (b)(4). Samples collected from operators after open-door interventions are the only personnel samples held to the ISO 5 microbial limits ((b)(4) CFU/plate).

这条分三块，每一块都指向同一个问题：做了动作，但没有"证明力"。

灭菌柜 PE001（A–C）。A：把 (b)(4)（推断：生物指示剂或温度探头）放在 (b)(4)（推断：装载中的某个固定位置）作为常规灭菌装载的唯一过程验证手段，缺乏充分的书面论证——"为什么放在这里"的理由不成立。B：操作员在 (b)(4) 灭菌（推断：常规灭菌）时不使用生物指示剂（BI）；批准的规程只要求在 (b)(4) 装载的确认阶段用 BI；最近一次确认报告是 )4.PQ.0004（PE001 性能确认，2026 年 6 月 19 日批准）。C：规程只要求在确认阶段做一次 (b)(4) 空气排除测试（推断：Bowie-Dick 类的空气排除/真空泄漏测试），技术员在任何常规灭菌周期之前都不做这类测试，而且连"多久做一次"都没有书面的频次依据。这台 PE001 的分量很重：技术员用它对成品药做 (b)(4) 灭菌，也用它灭菌可复用的产品接触部件、(b)(4) 软管、(b)(4)、器具，以及 ISO 5 灌装线 1 上所有商业无菌药品生产用的其他物品——氯胺酮 10%（批号 2602003）、利多卡因/肾上腺素（2603001）、硫酸沙丁胺醇（2512002）、布比卡因/肾上腺素（2608001）都在其列。

培养基灌装（D–F）。D：操作员不验证 ISO 5 灌装线 1 上无菌工艺模拟所用 (b)(4) 培养基的 (b)(4)（推断：促生长能力 / growth promotion），因此无法充分保证培养基适用于验证劳拉西泮注射液的无菌灌装工艺；检查员审核了批号 CHMF/26/001 的培养基灌装批记录（完成于 (b)(4)，2026 年 6 月 11 日批准）。E：灌装前对 ISO 5 灌装线与 ISO 7 缓冲间进行的 (b)(4) 循环，技术员既不做常规也不做定期的 BI 监测。F：在 (b)(4) 循环期间，(b)(4) 表面始终折叠在敞口的 (b)(4) 袋内，而半开的袋子里既没有 BI 也没有 (b)(4)（推断：化学指示剂）——也就是说，那些在安装过程中会暴露在环境里的 (b)(4)（推断：部件表面）到底有没有被充分 (b)(4)，没有任何证据。

人员监测（G）。对 (b)(4) 手部、前臂、胸部、前额的人员监测标准没有科学依据：企业用ISO 7 的微生物限度（警戒 (b)(4) CFU/皿、行动 (b)(4) CFU/皿）去评判那些使用 (b)(4)（推断：袖套式手套端口/隔离装置）在 ISO 5 灌装线内做关门干预的操作员样品；只有开门干预之后采集的人员样品，才按 ISO 5 的限度（(b)(4) CFU/皿）来判。

FDA 的真实关切：BI 是灭菌有效性的生物学证据，Bowie-Dick 类测试是蒸汽穿透与空气排除的证据，培养基促生长是无菌模拟"考题"有效的证据——这三样在常规运行中全部缺席，等于每一批的"无菌"都只建立在物理参数上。人员监测则把 ISO 5 里最频繁的关门干预放宽到了 ISO 7 的标准。法规：21 CFR 211.113(b)、211.111、211.67。

互动 你们的常规灭菌每柜都放 BI 吗？还是只有做确认那年才放？

观察项 3｜用水系统：漏阀、软管插进地漏、褐色残留、20 英尺死角

Equipment used in the manufacture, processing, packing or holding of drug products is not of appropriate design to facilitate operations for its intended use and cleaning and maintenance. Specifically, System (b)(4): A. We observed a leaking diaphragm valve after the final (b)(4) skid used to supply (b)(4) for all sterile injectable drugs that contain (b)(4). Examples of drugs that contain (b)(4) from this system include: Ketamine 10% USP, 100 mg/mL, 5 mL, SDV, lot 2602003; Lidocaine HCl 1% / Epinephrine Bitartrate 10 mcg/mL Injection USP, 10 mL, SDV, lot 2603001; Albuterol Sulfate 0.5%, 10 mL, lot 2512002; and Bupivacaine Hydrochloride 0.5% / Epinephrine Bitartrate 5 mcg/mL Injection, 10 mL vial, lot 2608001. B. There is an approximately one-foot-long (b)(4) connected to the (b)(4) for the (b)(4) on the (b)(4) System. (b)(4) System: We observed numerous deficiencies with the design, use, testing, and maintenance of the ambient temperature, (b)(4) construction (b)(4) System dedicated to supply (b)(4) to the (b)(4) equipment ID #PE001, inside the (b)(4) cleanroom suite. Operators use (b)(4) PE001 to sterilize reusable product-contact parts and (b)(4) sterilize injectable drugs on ISO 5 Filling Line (b)(4). C. The System is not sampled or tested for conformance with any written specifications for chemical, microbiological (b)(4). D. (b)(4) after the final (b)(4) step which was the beginning of the distribution line is closed (b)(4). (b)(4) System (b)(4) distribution line, flexible hose connection (b)(4) when there is no production scheduled. There was no written run-up procedure (b)(4) for the System (b)(4) PE001. E. The System includes numerous (b)(4) hose clamp connections, and (b)(4). This includes connections after the final (b)(4) step. F. The System is not sanitized periodically and there is no written requirement. G. (b)(4) brown residue throughout the inside of the (b)(4) housing for the (b)(4). H. (b)(4) closed on the approximately 8-inch section of pipe that supplies approximately 20 feet of pipe for the backup filling (b)(4) or the break tank. The 8-inch section of pipe leading from the flow line to the closed (b)(4) (b)(4). I. The (b)(4) for (b)(4) was open. The flexible (b)(4) line was (b)(4) approximately six inches (b)(4) a hole in the floor drain cover plate. Draining (b)(4) was emptying into the drain and contacting the open drain hose for the (b)(4) System. This potentially allows (b)(4) into the hose under specific conditions. J. There is no validation data to support (b)(4) submersion (b)(4) that production operators use to test (b)(4) after they remove them from ISO 5 Filling Line (b)(4). K. There are inadequate written controls for replacement frequency (b)(4). There is no validation data to support unlimited (b)(4) sterilization cycles of (b)(4) before replacement. (b)(4) sterilization cycles are not tracked.

这一条几乎可以当作"制药用水系统缺陷清单"来读：

A. 漏阀。在为所有含水无菌注射剂供 (b)(4)（推断：注射用水/WFI）的系统上，检查员在最终 (b)(4) 之后的隔膜阀处发现渗漏——这个位置已经过了最后一道屏障。涉及氯胺酮 10%（2602003）、利多卡因/肾上腺素（2603001）、硫酸沙丁胺醇（2512002）、布比卡因/肾上腺素（2608001）。

B. 一英尺长的 (b)(4)。系统上接了一段约一英尺长的 (b)(4)（推断：软管或短管），用于 (b)(4)。

C–K（常温 (b)(4) 分配系统）。这套为 PE001 供水（推断：灭菌柜用水）的常温系统在设计与维护上问题成串：C——系统未按任何书面标准做化学、微生物 (b)(4)（推断：内毒素等）取样和检测；D——最终步骤之后的分配管路与柔性软管连接，(b)(4) 在没有生产计划时 (b)(4)，而系统与 PE001没有任何书面的启动（run-up）程序；E——系统里有大量 (b)(4) 软管夹连接和 (b)(4)，包括最终步骤之后的连接；F——系统不做周期性消毒，也没有书面要求；G——(b)(4) 外壳内壁遍布褐色残留；H——从主管接出、通往备用灌装/缓冲罐的约 20 英尺管段，其前端约 8 英寸处的 (b)(4) 是关闭的，等于形成了一段长期不流动的死角（推断：dead leg）；I——(b)(4) 的排放口是敞开的，柔性排放软管 (b)(4) 地漏盖板破洞处约六英寸，排放的 (b)(4) 正流进地漏并接触到这根敞开的排放软管——在特定条件下可能让 (b)(4)（推断：地漏中的污染物）进入软管；J——操作员把 (b)(4)（推断：滤芯/部件）从灌装线取下后做的 (b)(4) 浸没 (b)(4)（推断：完整性测试，如起泡点），没有任何验证数据支持；K——(b)(4) 的更换频次没有充分的书面控制，也没有数据支持"可以无限次灭菌后再更换"的做法，而且灭菌次数根本没有被记录。

FDA 的真实关切：注射用水系统是无菌注射剂的"血液"——褐色残留、不消毒、死管段、软管插在地漏里，任何一条单拎出来都足以让整批产品失去无菌/内毒素保证，何况同时存在。法规：21 CFR 211.63（设备设计适用于预期用途）、211.65、211.67（清洁与维护）、211.48。

互动 你们水系统的排放软管，末端是悬空排入地漏，还是插进了地漏里？

观察项 4｜微粒、天平、库房温度：三类偏差都没有书面调查

Written records are not always made of investigations into unexplained discrepancies. Specifically, Particulate Matter A. Production processes, procedures, and equipment do not adequately exclude particulates and other potential contaminants from sterile injectable and oral inhalation drug product vials. I reviewed the two most recent (b)(4) reports for Trending of Visible Particle at Visual Inspection. The trend reports include results from 100% visual inspection and AQL inspection of finished sterile drug vials. a. PROT.GSP.0025, Visible Particle Trending August-Dec 2025, approved 02/03/26, documents personnel found visible particulates in multiple vials from (b)(4) batches manufactured and inspected during the review period. The number of vials rejected for particulate matter ranged from (b)(4) (0.03%) to (b)(4) (0.46%) vials in batches that were released. b. PROT.GSP.0035, Visible Particle Trending January 2026-June 2026, approved 08/18/26, documents personnel found visible particulates in at least one vial from all (b)(4) batches manufactured and inspected during the review period. The number of vials rejected for particulate matter ranged from (b)(4) (<0.01%) to (b)(4) (0.16%) vials. B. There is no written investigation of how extrinsic controlled particulates such as non-shedding cleanroom wipes, cleanroom (b)(4), and goggles material routinely enter vials during filling. C. There is inadequate scientific justification for the (b)(4) Reject Acceptance Rate Limit for critical defects per batch. Balances D. Dispensing room Balance PE026 was found out-of-tolerance at the (b)(4) and/or (b)(4) weights during the 10/08/25 calibration. Balance PE026 was found out-of-tolerance at the (b)(4) weight during the 04/15/25 calibration. There was no investigation documented. E. Dispensing room Balance PE027 was found out-of-tolerance at the (b)(4) weights during the 10/08/25 calibration. Balance PE027 was found out-of-tolerance at the (b)(4) weights during the 04/15/25 calibration. There was no investigation documented. Warehousing F. The (b)(4) Environmental Monitoring Trend Report for the Raw Material (b)(4) storage area (May 1-31, 2026) documents the temperature readings from sensors (b)(4) and (b)(4) exceeded the High Warning Limit of (b)(4) on multiple occasions throughout the month of May 2026. Maximum temperature readings on both sensors approached or reached the High Alarm Limit of (b)(4) during the same period. Quality management stated they notified the engineering department but did not document the excursion or initiate an investigation. Your firm does not have an SOP that addresses maintaining adequate storage conditions, and temperature controls, for raw materials and finished goods stored in the warehouse with instructions on documenting and investigating high warning or high alarm temperature limits excursions.

微粒（A–C）。生产工艺、程序和设备都不能充分把微粒和其他潜在污染物挡在无菌注射剂与吸入用溶液的西林瓶之外。检查员审了最近两份可见微粒趋势报告（含 100% 灯检与 AQL 抽检结果）：PROT.GSP.0025（2025 年 8–12 月，2026-02-03 批准）显示，审核期内生产的 (b)(4) 个批次中，多支西林瓶发现可见微粒，因微粒剔除的瓶数占比在已放行批次中介于 (b)(4)（0.03%）到 (b)(4)（0.46%）；PROT.GSP.0035（2026 年 1–6 月，2026-08-18 批准）显示，审核期内所有 (b)(4) 个批次都至少有 1 支瓶发现可见微粒，剔除比例介于 (b)(4)（<0.01%）到 (b)(4)（0.16%）。但 B：没有任何书面调查去解释这些"外来的、本应受控的"微粒——不掉屑的洁净室抹布、洁净室 (b)(4)、护目镜材质——是怎么在灌装过程中反复进到瓶里的；C：对"每批关键缺陷的拒收接受率限度"也缺乏科学依据。

天平（D–E）。配料间天平 PE026 在 2025 年 10 月 8 日校准中于 (b)(4) 和/或 (b)(4) 砝码点超出允差，在 2025 年 4 月 15 日校准中于 (b)(4) 砝码点也超差——两次都没有调查记录。天平 PE027 同样在 10/08/25 与 04/15/25 两次校准中多个砝码点超差，同样没有调查。

库房温度（F）。原料 (b)(4) 储存区 2026 年 5 月 1–31 日的环境监测趋势报告显示，两个传感器的温度读数在 5 月里多次超过高限警戒限，多个最高读数已接近或达到高限报警限。质量部门说他们通知了工程部，但既没有记录这次超标，也没有启动调查。更根本的是：企业根本没有一份 SOP来规定原料和成品在仓库中的储存条件与温度控制，也没有规定如何记录和调查高警戒/高报警温度超标。

FDA 的真实关切：微粒、称量、仓储温度，三条链路都有数据、有告警，却没有一份书面调查——"发现了但不记录"比"没发现"更严重，因为它意味着调查机制本身不运转。法规：21 CFR 211.192、211.68(a)、211.142。

观察项 5｜培养基选错、氮气不测、丙二醇/PEG400 的 DEG/EG 还要混样

Laboratory controls do not include the establishment of scientifically sound and appropriate specifications, standards, sampling plans and test procedures designed to assure that components and drug products conform to appropriate standards of identity, strength, quality and purity. Specifically, Growth Media A. QC microbiology personnel use Standard Methods (b)(4) Plate Count (b)(4) for all (b)(4) bioburden samples. There is inadequate justification to use this high (b)(4) instead of (b)(4) designed to recover microorganisms from low (b)(4) systems. Growth promotion on the manufacturer COA for Plate Count (b)(4) only lists Staphylococcus aureus, Escherichia coli, and Enterococcus faecalis. It does not include waterborne pathogenic microorganisms of interest such as B. cepacia or P. aeruginosa that are known to colonize high purity (b)(4) systems. B. The firm accepts Plate Count (b)(4) plates based on manufacturer COA and does not perform growth promotion before use. Nitrogen Your firm uses nitrogen for pre-gassing vials and blanketing drug product solution in all lots of Lidocaine/Epinephrine Injection, Bupivacaine/Epinephrine Injection, and Lorazepam Injection. We observed multiple apparent deficiencies related to nitrogen. C. There was inadequate written justification to confirm that (b)(4) testing of nitrogen generated in-house ensured the nitrogen met all applicable standards each time it was used as a product-contact gas in sterile injectable and oral inhalation drugs. Drugs that require a nitrogen overlay include all lots of Lidocaine/Epinephrine Injection, Bupivacaine/Epinephrine Injection, and Lorazepam Injection. Examples of drugs manufactured with nitrogen generated in-house are Lidocaine HCl 1% / Epinephrine Bitartrate 10 mcg/mL Injection USP, 10 mL, SDV, lot 2603001; and Lorazepam Injection USP, 2 mg/mL, lot 2510002. D. Manufacturing personnel do not (b)(4) test the (b)(4) attached to ISO 5 Filling Line (b)(4). There is no scientific justification for the (b)(4) replacement interval for this point-of-use nitrogen (b)(4). E. The firm did not conduct ID tests or any other analyses on Ultra High Purity (UHP) Nitrogen, lot (b)(4), that was accepted based on supplier COA. This is a product-contact gas used during production of Bupivacaine Hydrochloride 0.5% / Epinephrine Bitartrate 5 mcg/mL Injection, 10 mL vials, lot 2608001. F. Operators confirmed the (b)(4) of the (b)(4) they attach to the connection point for UHP Nitrogen cylinders in February 2026. Engineering technicians remove the (b)(4) when not in-use and install it when requested by production personnel. Technicians used the (b)(4) in August 2026 without performing a (b)(4) test. There is no written justification for the (b)(4) test interval. Ethylene Glycol and Diethylene Glycol Testing G. Your firm did not perform all specific identity tests for Propylene Glycol, lot (b)(4) and Polyethylene Glycol 400, lot (b)(4). Also, your SOP LAB.0025-Sample Pooling Strategy for Diethylene Glycol and Ethylene Glycol (DEG/EG) effective 21 Oct 2025, instructs personnel to pool (b)(4) for testing of high-risk drug components. These high-risk components were used in the manufacturing of Lorazepam Injection USP, 2 mg/mL, lot 2510001 (first distribution 12/02/25), lot 2510002 (first distribution 12/01/25), and 2511001 (first distribution 05/11/26).

培养基（A–B）。QC 微生物人员对所有 (b)(4) 生物负荷样品使用的是 Standard Methods (b)(4) 平板计数 (b)(4)（推断：高营养型培养基），却拿不出充分理由说明为什么不用专为低营养 (b)(4) 系统设计的 (b)(4)（推断：R2A 一类的贫营养培养基）去回收微生物。厂家 COA 上的促生长试验只列了金黄色葡萄球菌、大肠杆菌和粪肠球菌，不包括洋葱伯克霍尔德菌（B. cepacia）或铜绿假单胞菌（P. aeruginosa）这类已知会在高纯水系统中定植的水生致病菌。此外，企业凭厂家 COA 接收平板，使用前不做促生长试验。

氮气（C–F）。企业用氮气对西林瓶做前充气、并对药液做氮封，涉及利多卡因/肾上腺素、布比卡因/肾上腺素、劳拉西泮注射液的所有批次。检查员看到的缺陷包括：C——对于自产氮气，没有充分的书面论证能确认"每次作为产品接触气体使用前"都符合所有适用标准；D——生产人员不对 ISO 5 灌装线上那个使用点氮气 (b)(4)（推断：终端过滤器）做 (b)(4) 测试，也没有科学依据支撑它的更换周期；E——对某批超高纯（UHP）氮气既没做鉴别也没做任何分析，仅凭供应商 COA 接收，而这批氮气用在了布比卡因/肾上腺素（批号 2608001）的生产中；F——操作员 2026 年 2 月确认过 UHP 氮气钢瓶连接点上那个 (b)(4) 的 (b)(4)（推断：安装/完整性确认），工程人员在不用时拆下、需要时装上，而 2026 年 8 月技术员在没有做 (b)(4) 测试的情况下就使用了它，且测试间隔同样没有书面依据。

DEG/EG（G）。企业对丙二醇（批号 (b)(4)）和聚乙二醇 400（批号 (b)(4)）没有做完所有专属鉴别项。此外，SOP LAB.0025《二甘醇与乙二醇（DEG/EG）样品合并策略》（2025 年 10 月 21 日生效）要求人员对高风险组分进行混样（pooling）后检测——而这些高风险组分，正是用于劳拉西泮注射液 USP 2mg/mL 批号 2510001（首次分销 2025-12-02）、2510002（2025-12-01）和 2511001（2026-05-11）的生产。

FDA 的真实关切：这一条把三件"实验室该守住"的事摆在一起——培养基选得对不对决定了你能不能检出污染菌；氮气是产品接触气体，不测就等于不设防；而 DEG/EG 用混样检测，会让某一桶被污染的丙二醇/PEG 在合并后被稀释到检出限以下。丙二醇与 PEG 正是 DEG/EG 污染的历史高发载体（1937 年美国磺胺酏剂事件直接催生了现代药品法；近年在多国糖浆事件中又反复出现）。法规：21 CFR 211.160(b)、211.84(d)、211.94；USP <71>/<788> 理念。

互动 高风险辅料做 DEG/EG 检测时，你们是逐桶取样，还是合并成一个样？

三、主线收敛

5 条的共同点是：凡是"该有证据的地方"，都只剩下了动作本身。洁净室在最轻松的状态下做了动态确认（1）；灭菌柜常规周期既不放 BI 也不做空气排除测试，培养基灌装的培养基没验证促生长能力（2）；水系统有褐色残留、不消毒、有死管段、排放软管插在地漏里（3）；微粒、天平超差、库房温度超标三类异常都没有书面调查（4）；培养基选得不对、氮气不测、DEG/EG 还要混样（5）。

把它们收成两句话：一是"验证与实际两张皮"——确认做得比实际轻松，监测做得比实际宽松；二是"异常与调查之间断路"——数据出来了，调查没启动。再叠加 2024 年的 Avastin 召回与 2025 年 4 月的警告信，这套体系最需要的不是再多一份整改计划，而是让"发现异常"这一步能自动接通"必须调查"。

四、严重度一览

五、把这份 483 放回执法链条上看

Tailstorm / Medivant 的监管轨迹密度在本批里最高：2024-10 凤凰城 483（含修订版）→ 2024-11 自愿召回 6 批贝伐珠单抗（无菌保证不足）→2025-04-08 警告信 WL #703497→ 2025-05 钱德勒厂 483 → 2025-11 凤凰城厂 483（实验室系统适用性、审计追踪、天花板异物）→ 2026-08 钱德勒厂本次 483。

两个细节值得单独拎出来。其一，检查员的连续性：2025 年 5 月与 2026 年 8 月两次钱德勒厂检查，签署人都是 Nicholas L. Hunt 与 Ronda R. Loyd Jones。同一双眼睛看第二次，"上次说的改了没有"这个问题会被逐条对表——本次观察项 1（动态确认不代表实际）、4（调查无记录）与 2025 年那份里的"OOS 调查不充分（重复项）""QC 验收标准不足（重复项）"是同一条脉络。

其二，风险性质的升级：上一轮的核心问题是"无菌保证"（Avastin 召回即因此而起）；这一轮除了无菌，还多了一层化学污染——氮气作为产品接触气体未做确认、丙二醇与 PEG 400 的 DEG/EG 检测还要混样。混样会把单桶污染稀释到检出线以下，这与 2022–2023 年多国糖浆 DEG 事件的失守方式如出一辙。放在本批序列里看，它与同批 RC Outsourcing（同为 Avastin 分装、同样有警告信与召回）构成了一组"503B 二次执法样本"：都在警告信之后被复查，都出现了"旧账 + 新账"的叠加。

需要强调的是，Form 483 不构成 FDA 的最终判定。企业仍有回复与申诉的程序空间，FDA 也尚未公布本次检查的分类结论——以上是风险方向，不是判决书。

互动 如果让你挑一条最该立刻动手的，是氮气、水系统，还是 DEG/EG 的混样？

声明：本文基于美国 FDA 公开的执法文书（Form 483 / Warning Letter / Import Alert 等）撰写，仅供医药行业质量管理与合规学习交流之用，不构成对任何企业的法律或商业评价。原始文书以 FDA 官网公布版本为准；文书内容不构成 FDA 的最终判定。