Q.01

What does EU GMP Annex One require for sterile manufacturing?

EU GMP Annex 1 (EudraLex Volume 4, "Manufacture of Sterile Medicinal Products") requires every sterile manufacturer to build a facility-wide Contamination Control Strategy (CCS). Controls must be designed in first, procedures come second, and monitoring only confirms they work, because "Monitoring or testing alone does not give assurance of sterility" [1]. The revised Annex had to be applied from 25 August 2023. The lyophilizer sterilisation clause, point 8.123, had until 25 August 2024 [2].

Nearly all requirements are worded "should". The Annex says any limits, frequencies or ranges it gives "should be considered as a minimum requirement" [3].

Approach: I read the Annex 1 text (the August 2022 revision) held in the EU GMP guidelines collection, EudraLex Volume 4. This is a summary of the main obligations, not every one of its roughly 290 numbered clauses.

Scope and principles (Sections 1–3)

  • What it covers: all sterile products, meaning active substances, excipients, primary packaging and finished dosage forms. It applies Quality Risk Management (QRM) to prevent microbial, particulate and endotoxin/pyrogen contamination. Firms that choose to apply it to non-sterile products should document which principles they used [3].
  • Key areas (2.1):
    • Design, qualify and validate the facility, equipment and process.
    • Consider barrier and other technologies: RABS, isolators, robotics, rapid/alternative methods and continuous monitoring.
    • Use qualified and trained personnel.
    • Control the bioburden and endotoxin of raw and packaging materials [4].
  • CCS (2.3–2.6):
    • It should define all critical control points and assess how well the controls and monitoring work.
    • It should be actively reviewed and drive continual improvement.
    • Its effectiveness should form part of the periodic management review [5].
    • It should cover 16 listed elements, including utilities, vendor approval, outsourced activities, validation of sterilisation processes, cleaning and disinfection, monitoring, and CAPA [6].
    • Changes should be assessed for impact on the CCS before and after implementation [7].
  • Pharmaceutical quality system (3.1): it should integrate risk management across the product lifecycle and require root-cause analysis and CAPA. Senior management should oversee the state of control. Those certifying batches should have adequate sterile-manufacturing knowledge [8].

Premises and barrier technology (Section 4)

  • Four cleanroom grades (4.4):
    • Grade A is the critical zone (filling, stopper bowl, aseptic connections). Unidirectional airflow should be demonstrated across the whole zone, and direct operator intervention minimised.
    • Grade B is the background for Grade A when an isolator is not used, with pressure differences continuously monitored.
    • Grades C and D are for less critical steps and isolator backgrounds [9].
  • RABS and isolators should be considered in the CCS, and any alternative to them should be justified [10].
  • Bio-decontamination (4.22):
    • For isolators, it should be automated and validated, with a sporicidal agent (e.g. gaseous or vaporised).
    • For RABS, routine sporicidal disinfection should use a validated method [11].
  • Airlocks: doors should not open at the same time. Airlocks into Grade A/B should be interlocked; those into Grade C/D need at least a visual and/or audible warning [12].
  • Pressure: adjacent rooms of different grades should differ by at least 10 Pa (guidance value) [13]. Critical pressure differences should be continuously monitored, with an alarm that is not overridden without assessment [14].
  • Classification (4.27, 4.29): particles ≥0.5 µm and ≥5 µm are measured "at rest" and "in operation" [15]. At-rest limits should be restored after a clean-up period, with a guidance value of less than 20 minutes [16].

Table 1: maximum total particles per m³ [15]

Grade≥0.5 µm at rest≥0.5 µm in operation≥5 µm at rest≥5 µm in operation
A3 5203 520Not specifiedNot specified
B3 520352 000Not specified2 930
C352 0003 520 0002 93029 300
D3 520 000Not predetermined29 300Not predetermined

For Grade D, the manufacturer sets its own in-operation limits based on risk assessment.

Utilities and personnel (Sections 6–7)

  • Water for injections (6.10): WFI should be made by distillation or an equivalent purification process. That may include reverse osmosis combined with EDI, ultrafiltration or nanofiltration. It should be stored and distributed so microbial growth is minimised, e.g. by constant circulation above 70°C [17].
  • Water system sanitisation (6.12): it should follow a set schedule. Microbiological and endotoxin results should be approved before batches made with that water are released [18].
  • Gowning qualification (7.4):
    • Staff entering Grade A/B should be trained in aseptic gowning.
    • Gowning should be reassessed at least annually, both visually and microbiologically.
    • Unsupervised access is limited to people who passed the gowning assessment and took part in a successful aseptic process simulation (APS) [19].
  • Unqualified people should enter Grade B, or Grade A in operation, only in exceptional cases, under a written procedure and supervision [20].

Production and sterilisation (Section 8)

  • Preparation of components and materials: at least Grade D. Use at least Grade C for high-risk products and for ointments, creams, suspensions and emulsions before terminal sterilisation [21]. If filling is unusually risky (e.g. slow filling or wide necks), fill in Grade A with at least a Grade C background [22].
  • Fusion-sealed containers (8.22):
    • Glass ampoules, BFS units and containers of ≤100 ml closed by fusion should get 100% integrity testing.
    • For containers over 100 ml, reduced sampling may be acceptable if justified.
    • Visual inspection is not an acceptable integrity test [23].
    • Container closure integrity validation should consider transport stresses [24].
  • Visual inspection (8.30, 8.33): every filled parenteral container should be inspected individually, against a maintained defect library. Defects should be trended, and a critical defect found later should trigger an investigation [25][26].
  • Sterilising filters (8.87):
    • Integrity should be tested before use, after sterilisation (PUPSIT).
    • A non-destructive test should follow use, before the filter is removed from its housing.
    • If PUPSIT is not possible, an alternative is allowed only with a thorough risk assessment and mitigating controls [27].
  • Lyophilizers (8.123, the clause with the 2024 deadline): those loaded or unloaded manually without barrier separation should be sterilised before each load. For automated or closed-barrier systems, the frequency should be justified in the CCS [28].

Monitoring and aseptic process simulation (Section 9)

Table 6: viable action limits (9.30) [29]

GradeAir (CFU/m³)Settle plate, 90 mm (CFU/4 h)Contact plate, 55 mm (CFU/plate)Glove print, 5 fingers (CFU/glove)
ANo growthNo growthNo growthNo growth
B10555
C1005025–
D20010050–

Any growth in Grade A should lead to an investigation. In Grades A and B, settle plates should be exposed for the whole operation and changed after at most 4 hours.

  • APS design (9.33): it should imitate the routine process from sterilisation to sealing. Inert gas should be replaced by air (unless an anaerobic simulation is intended), and lyophilization should be simulated under worst-case conditions [30].
  • APS frequency (9.38):
    • Initial validation needs at least three consecutive successful runs covering all shifts.
    • Normally, repeat twice a year per process, filling line and shift.
    • Each operator should take part in at least one successful APS a year [31].
    • For manual aseptic operations, each operator needs 3 initial runs, then one about every 6 months [32].
  • APS read-out (9.45): inspection by trained staff, plus positive controls using reference organisms and local isolates [33].

Quality control (Section 10)

  • Staff with microbiology and sterility-assurance expertise should be available to support process design, environmental monitoring and investigations [34].
  • Sterility test samples (10.6): the test should be done under aseptic conditions, with samples taken from the most at-risk parts of the batch:
    • For aseptic fills: the start and end of filling.
    • For heat-sterilised products: worst-case load locations.
    • For lyophilized products: different lyophilization loads.
    • For sub-batches: a separate test for each [35].

I can go further into any one section, for example the terminal sterilisation methods or the BFS requirements.

Q.02

What belongs in a contamination control strategy?

EU GMP Annex 1 (point 2.5) says a contamination control strategy (CCS) should cover at least 16 elements, "but are not limited to" them. They run from plant and process design through vendor approval, cleaning, monitoring and CAPA to continuous improvement [6]. The Annex defines a CCS as "a planned set of controls for microorganisms, endotoxin/pyrogen and particles, derived from current product and process understanding that assures process performance and product quality" [36]. Beyond the list, about 30 other clauses say a particular decision should be defined, justified or documented in the CCS.

Approach: I read Section 2 of Annex 1 (August 2022 revision) in EudraLex Volume 4, then every other Annex 1 clause that mentions the CCS.

The 16 elements (point 2.5)

The Annex says that building a CCS "requires detailed technical and process knowledge". Contamination sources include microbial and cellular debris (pyrogen, endotoxin) and particles, such as glass and other visible and sub-visible particles [6].

  1. Design of the plant and processes, including the documentation
  2. Premises and equipment
  3. Personnel
  4. Utilities
  5. Raw material controls, including in-process controls
  6. Product containers and closures
  7. Vendor approval: key component suppliers, sterilisation of components and single-use systems (SUS), and critical service providers
  8. Management of outsourced activities and the availability and transfer of critical information between parties (e.g. contract sterilisation)
  9. Process risk management
  10. Process validation
  11. Validation of sterilisation processes
  12. Preventative maintenance, planned and unplanned, so that equipment, utilities and premises add no extra contamination risk
  13. Cleaning and disinfection
  14. Monitoring systems, including whether scientifically sound alternative methods could detect environmental contamination better
  15. Prevention mechanisms: trend analysis, detailed investigation, root cause, CAPA and investigational tools
  16. Continuous improvement based on all of the above

How the CCS should work (points 2.3, 2.4, 2.6, 3.1)

  • Scope: it should be implemented across the whole facility. It should define all critical control points and assess how well all design, procedural, technical and organisational controls and monitoring measures work [5].
  • Existing systems: they "may not require replacement". They should be referenced in the CCS, and the interactions between systems should be understood [5].
  • Looking at controls together: each measure is usually assessed on its own, but their "collective effectiveness should be considered together" [37].
  • Review:
    • The CCS should be actively reviewed and updated, and should drive continual improvement.
    • Its effectiveness should form part of the periodic management review [5].
    • Changes should be assessed for impact on the CCS before and after they are implemented [7].
  • Risk management: QRM is used to build and maintain the CCS, to identify, assess, reduce or eliminate, and control contamination risks [8].

Decisions other clauses place in the CCS

Premises and barriers

  • Whether RABS or isolators are used [10].
  • Moving material into higher grades, and the risk assessment and mitigation for that [38].
  • Pressure-differential set points and any alarm delays [14].
  • The isolator background grade [39].
  • How often gloves are replaced [40].
  • Any departure from the 0.36 to 0.54 m/s unidirectional airflow guidance value [41].
  • Air velocity testing in Grades B, C and D [42].
  • Whether disinfectants used in Grades C and D must be sterile [43].

Utilities and personnel

  • Risk assessment of how utilities affect the product [44].
  • Water sampling programmes [45].
  • Use and disinfection of the equipment that clause 7.9 refers to [46].
  • Extra gowning in Grades C and D [47].

Production

  • Acceptance criteria, monitoring and effectiveness review for the controls that clause 8.7 covers [48].
  • How often other sterilisation load patterns are validated [49].
  • A second sterile filtration as close to the point of fill as possible [50].
  • Justifying campaign manufacture for filters [51].
  • Bioburden and endotoxin of FFS films and of BFS raw polymer [52][53].
  • How often automated or closed-barrier lyophilizers are sterilised [28].
  • Specific risks of single-use systems [54].

Monitoring, APS and QC

  • The environmental and process monitoring programme "forms part of the overall CCS" [55].
  • Any action limits stricter than the standard ones [56].
  • The air sampling method [57].
  • Extra gown monitoring for higher-risk operations [58].
  • Use of the methods that clause 9.36 says "should not replace routine APS" [59].
  • How many units are filled in an APS [60].
  • Microbial, particulate and endotoxin limits in specifications where the CCS shows they are needed [61].

These clauses mostly use "should". Point 2.7 is clear that sterility should not rest only on a terminal process or a finished-product test [62].

I can turn this into a CCS gap-assessment checklist mapped to the clauses, if that would help.

Q.03

How do cleanroom grades differ?

Under EU GMP Annex 1 the four grades differ by how critical the work in them is. Grade A is the critical zone where product, containers or sterilised surfaces are open to the air. Grade B is the cleanroom around Grade A when no isolator is used. Grades C and D are for less critical steps, as backgrounds for isolators, or for preparing and filling terminally sterilised products [9]. Each step up in grade brings tighter particle and microbial limits, more frequent requalification and monitoring, and stricter gowning.

Approach: I read the grade-related clauses of Annex 1 (August 2022 revision) in EudraLex Volume 4: Sections 4, 7, 8 and 9 and Tables 1 to 6.

What each grade is for (4.4, Tables 3 and 4)

GradeRoleTypical operations: aseptic processing (Table 4)Typical operations: terminal sterilisation (Table 3)
ACritical zone, normally unidirectional airflow within RABS or isolators. Unidirectional airflow should be qualified across the whole zone, and operators reaching in directly should be kept to a minimum.Aseptic assembly of filling equipment, aseptic connections after the final sterilising filter, aseptic compounding, aseptic filling and sealing, lyophilizer loadingFilling when the product is unusually at risk
BBackground for Grade A when no isolator is used. Pressure differences should be continuously monitored.Background support for A; staging protected items before they go into A–
CLess critical stagesPreparing solutions to be filtered, including sampling and dispensingFilling of products; preparing solutions when unusually at risk
DLess critical stagesCleaning equipment, assembly under HEPA airflow before sterilisation, assembling closed sterilised single-use systemsPreparing solutions and components for later filling

Sources: [9][63][64]

Fixed pairings:

  • Aseptic work: handling sterilised equipment and components, setting up the filling line and aseptic filling should be done in Grade A with a Grade B background [65].
  • Terminally sterilised products: filling should be done in at least Grade C [66].
  • Barrier backgrounds:
    • Open isolators: generally at least Grade C.
    • Closed isolators: at least Grade D, with the choice justified in the CCS.
    • RABS: at least Grade B [39].

Total particle limits per m³

Classification limits (Table 1) vs monitoring limits (Table 5): at rest / in operation

Grade≥0.5 µm (same in both tables)≥5 µm, classification (Table 1)≥5 µm, monitoring (Table 5)
A3 520 / 3 520Not specified / Not specified29 / 29
B3 520 / 352 000Not specified / 2 93029 / 2 930
C352 000 / 3 520 0002 930 / 29 3002 930 / 29 300
D3 520 000 / not predetermined29 300 / not predetermined29 300 / not predetermined

Sources: [15][67]

  • In Grades A and B, classification "may" include 5 µm particles where the CCS or historical trends point to it [15].
  • For Grade D, the manufacturer sets its own in-operation limits based on risk assessment and routine data [67].
  • At-rest limits should be met again after a clean-up period, guidance value under 20 minutes [16].
  • Unidirectional airflow should give a uniform 0.36–0.54 m/s (guidance value) at the working position [41].

Microbial limits (Table 2 for qualification, Table 6 action limits)

GradeAir (CFU/m³)Settle plate 90 mm (CFU/4 h)Contact plate 55 mm (CFU/plate)Glove print, 5 fingers (CFU/glove; Table 6 only)
ANo growthNo growthNo growthNo growth
B10555
C1005025–
D20010050–

Sources: [68][29]

Settle plates and gown monitoring:

  • Grades A and B: settle plates should be exposed for the whole operation and changed after at most 4 hours.
  • Grades C and D: exposure time (maximum 4 hours) and frequency should be based on QRM.
  • Grade A: any growth should lead to an investigation.
  • Grades C and D: routine gown monitoring is "not normally required", depending on the room's function [29].
  • Organism identification: organisms found in Grades A and B should be identified to species level. For Grades C and D, identification should be "considered", for example after excursions [69].

Requalification and routine monitoring

  • Requalification: at most every 6 months for Grades A and B, and every 12 months for Grades C and D. Air velocity testing in Grades B to D follows a risk assessment, but it is required for filling zones with unidirectional airflow [42].
  • Grade A: particles should be monitored continuously at ≥0.5 and ≥5 µm, at a flow of at least 28 L/min, for the whole of critical processing [70][71]. Continuous viable air monitoring should also run throughout [72].
  • Grade B: a similar system is "recommended", but sampling can be less frequent. Alarms should trigger when alert levels are exceeded [73].
  • Grades C and D: monitoring is based on qualification and routine data. Action limits may be tighter than Tables 5 and 6 [74].
  • Disinfectants: those used in Grades A and B should be sterile. For Grades C and D, they may need to be sterile where the CCS says so [43].

Gowning (7.13–7.15)

  • Grade B (including access to Grade A):
    • Sterilised suit worn over dedicated undergarments, with sterile headgear enclosing all hair.
    • Sterile facemask and eye coverings, sterilised over-boots, and two pairs of sterile gloves.
    • A fresh sterilised gown at every entry [47][75].
  • Grade C:
    • Hair and beards covered.
    • Single or two-piece trouser suit, gathered at the wrists, with a high neck.
    • Disinfected shoes or overshoes.
  • Grade D:
    • Hair and beards covered.
    • General protective suit and disinfected shoes or overshoes.
  • Extras in C and D: gloves and a facemask may be needed in Grades C and D where the CCS says so [47].
  • Clothing rules: outdoor clothing (other than underwear) should not go into change rooms leading directly to Grades B and C [76].

I can set these out as a single grade-by-grade comparison sheet for qualification or audit use.

Q.04