Regulatory Focus™ > News Articles > 2019 > 7 > FDA and EU GMP Annex 1 Differences in Cleanroom Specifications

FDA and EU GMP Annex 1 Differences in Cleanroom Specifications

Posted 22 July 2019 | By Aleksandr Fedotov, PhD 

FDA and EU GMP Annex 1 Differences in Cleanroom Specifications

Based on a presentation at Interphex in April 2019,1 this article discusses the contradictions between the US and EU requirements for cleanroom Good Manufacturing Practices (GMPs). The author discusses particle concentration for cleanrooms “at rest,” particle contamination in the air, start-up testing vs. routine monitoring, risk analysis and “interlocking” doors. The author concludes that there should be globally harmonized GMP norms for cleanroom operation based on “hands on” experience and scientific evidence.
 
Introduction
 
Although there is no clear explanation, critical cleanroom norms differ in the US and the European Union (EU). Requirements for cleanrooms are among the most critical issues of Good Manufacturing Practices (GMPs), but differences between requirements in the US and EU are responsible for confusion and misunderstandings and may even pose safety risks. To make matters worse, in some cases, the new draft of EU GMP Annex 1 adds even more confusion. This article focuses attention on the following “bullet points” from the EU draft and should be taken into consideration when updating GMP regulations for cleanrooms:
 
  • airborne particles with sizes ≥ 5.0 μm have different EU and FDA requirements
  • the criticism of EU GMP Annex 1 and other proposals
  • terminal sterilization and aseptic processes: EMA and FDA have different positions
  • confusion in Annex 1 Draft between testing at start-up versus monitoring
  • dangers posed by doors and interlocking airlocks
  • multi-cascade entering procedure in Annex 1 Draft over Grade types Clean Not Classified (CNC) – D – C – B
  • biocontamination in air – what do we really measure?
Cleanroom Requirements
 
Globally, there are two main GMP systems that regulate requirements for cleanrooms and clean zones. First, the EU GMP Guidance Annex 1: Manufacturing of Sterile Medicinal Products is now under revision.2 It specifies norms for both aseptic and terminal sterilization processes. The Draft of Annex 1 has important changes. The second is the US Food and Drug Administration’s (FDA’s) Guidance for Industry: Sterile Drug Products Produced by Aseptic Processing Current Good Manufacturing Practice.3 It concerns aseptic processes only. Both regulatory systems are in use worldwide. And, therein lies the problem, because the two systems are at time contradictory in terms of requirement. EU GMP and FDA requirements for cleanrooms were approved more than 30 years ago and Table 1 provides a sample of how they might differ:
 
Table 1. Differences Between FDA and EU Cleanroom Requirements
  FDA EU
Particle sizes ≥ 0.5 µm only both ≥ 0.5 µm and ≥ 5.0 µm
Types of sterile process aseptic only both aseptic and terminal sterilization
Occupancy state rooms/zones “in operation” only both “at rest’ and “in operation”
Zone D no analogue exists
 
 
This raises the following two questions:
 
  1. Why do FDA’s Aseptic Guidance and the EU’s GMP Guidance, Annex 1, specify mainly the same information, but indicate it differently?
  2. Is harmonization and mutual recognition possible?
These fundamental issues have not changed for decades. However, knowledge and practical experience gained over the last several decades has provided better understandings that necessitate change. EU GMP Guidance has been approved for application not only in EU, but also in Russia, China and other countries. It has become a nearly global guide. But, are its requirements for air cleanliness and for sterile production (Annex 1) based on science and practical analysis? Or, are they arbitrary opinions without a proper scientific grounding?
 
Because mutual recognition of FDA-EU inspections expects equivalency of norms, it is time to review these important requirements on basis of scientific knowledge and evidence rather than on opinions that may be outdated and nonscientific in nature.
 
Particles ≥ 5.0 µm
 
Sampling of particles ≥ 5.0 µm requires big volumes of air and a rather long period of time. These factors have entered into the invention of particle counters with big air flow rates of 50 l/min, and even 100 l/min, rather than to common 28 l/min (1 ft3/min).
 
Deposition of particles on tubing walls depends on particle size. For particles ≥ 5.0 µm this becomes critical that restricts length of tubes. There are also contradiction between normative limits and real statistics for particles of different sizes. However, is control of particles ≥ 5.0 µm necessary? Some experts say that microorganisms can deposit on particles 5.0 – 10.0 µm, so control of large particles helps to assure sterility.
 
But what does real experience and statistics say? Existing norms are rather old and there was insufficient experience 30-40 years ago and more likely, no one cared to investigate the issue in depth and provide opportunities for discussion regarding FDA/EU differences. However, there are now scientific arguments regarding the differences – Figure 1.
 
Figure 1. Arguments vs. Opinions in GMP History
Fedotov-1.png


 
FDA’s Experience
 
FDA’s Aseptic Guidance says to control particles ≥ 0.5 µm only, not ≥ 5.0 µm. But, is there any evidence that US drugs are worse than European drugs where particles of both ≥ 0.5 µm and ≥ 5.0 µm are to be controlled? There is no evidence to that effect and since FDA controls almost half the manufacturing of drugs globally and has decades of experience, their experience and opinion should be taken into account. It is quite possible to rely on particles ≥ 0.5 µm only because US FDA practice confirms it. There are also statistics, not only “practical” knowledge (Table 2), that give realistic ratios between numbers of:
 
  • particles ≥ 0.5 and ≥ 5.0 µm in air
  • particles ≥ 0.5 μm and microorganisms (CFU)
 
Table 2. Limits for Airborne Particles Concentration in Existing GMP EC Annex 1
Grade Maximum Permitted Number of Particles/m3
At Rest In Operation
≥ 0.5 µm ≥ 5.0 µm ≥ 0.5 µm ≥ 5.0 µm
A 3 520 20 3 520 20
B 3 520 29 352 000 2 900
C 352 000 2 900 3 520 000 29 000
D 3 520 000 29 000 Not defined Not defined
 
Limit 29 particles ≥ 5.0 µm/m3 for Grade B was taken from ISO 14644-1:1999 standard.4
Limit 20 particles ≥ 5.0 µm/m3 for Grade A was taken arbitrary by experts for it “should be tighter than for Grade B.”
 
The ratio between particle concentrations for zones B “at rest” is: Cn≥0.5 µm/Cn≥5.0 µm = 3520/29 ≈ 121.
But, does it correspond to reality? It does not because there is a difference in order of magnitude.
The real ratio is approximately between 5:1 and 50:1 for both “at rest” and “in operation” occupancy states, not 100:1 or 121.But Annex 1 limits are based on ratio 121.
 
Testing of cleanrooms by the author provided experimental data for ratio between particle concentrations Cn≥0.5 µm and Cn≥5.0 µm (Figure 2). Grades A/B are of the key interest.
 
Figure 2. Example of Particle Concentration for Cleanrooms “at Rest”

Fedotov-2.png

 
It was shown for zones A (clean and correctly designed unit with unidirectional airflow) that at rest, there were zero counts for particles both ≥ 0.5 µm and ≥ 5.0, i.e., particles were absent at all; and in operation, there were zero counts, even when an operator in a cleanroom was shaking a hand slightly on proximity of 10 cm aside the sampling probe at the same level (height). Particles were counted only when the hand moved was very close to the probe or above it. The ratio Cn≥0.5 µm/Cn≥5.0 µm  was  ≈ 5–10 for zones B in most cases.
 
However, existing Annex 1 and the new draft set limits: 3520 particles/m3 for particles ≥ 0.5 µm both for Grades A and B; 20 and 29 particles/m3 for particles ≥ 5.0 µm for Grades A and B, respectively. So, the ratio Cn≥0.5 µm/Cn≥5.0 µm = 121 and EU GMP noncompliance with reality exceed the order of magnitude.
 
The levels of ISO 4 or even ISO 3 for zones A can be achieved easily, a fact shown for many facilities. This data can be regarded as representative. Similar results were presented by a well-known British GMP expert John Sharp.5 Counts for particle sizes ≥ 5.0 µm for zones B “in operation” were 20 particles/m3 in his investigation that was much below limit 2900. Control of particles ≥ 5.0 µm gives no added value to particles ≥ 0.5 µm.
 
Contamination in the air
 
NASA NHB 5340 handbook says that approximately 1 particle from 1000 particles ≥ 0.5 µm carries 1 Colony Forming Unit (CFU).6 More realistic picture considers some deviations, but the sense remains (Figure  3).
 
Figure 3. Approximate Correlation Between Number of Particles and Microorganisms (CFU) in air

Fedotov-3.png

 
Swedish scientists Bengt Lungquist and Berit Reinmüller showed that ratio of particles with sizes ≥ 0.5 µm to CFU concentration is about 1500 ± 500.7 This is similar to NASA numbers. But, let’s compare it with EU GMP norms. The Limit for Grade A (ISO 5) is 3520 particles/m3 ≥ 0.5 µm. So, concentration of microorganisms in air can be expected as 3520/1000 = 3,5 CFU/m3. ISO  Class 4 limit for particles is equal to 352 particles/m3 ≥ 0.5 µm and an expected number of CFUs is 352/1000 = 0.35 CFU/m3. ISO Class 5 is weak for aseptic processes (filling) and ISO Class 4 can be recommended for these processes.
 
Start-up Testing vs. Routine Monitoring
 
Annex 1 Draft separates requirements for start-up and routine monitoring, although it is done in a very interesting way. The draft takes a step forward regarding ≥ 5.0 µm particles. It cancelled them from classification (start-up) testing. But, at the same time, it retains monitoring of particles ≥ 5.0 µm for routine control. Thus, the draft specifies a more detailed and rigorous routine for control than for start-up testing (Table 3 and Table 4).
 
Table 3. Maximum Permitted Airborne Particle Concentration During Classification in the Annex 1 Draft
Grade Maximum permitted number of particles ≥ 0.5 μm/m3
At Rest In Operation ISO Class
At Rest/In Operation
А 3 520 3 520 5/5
В 3 520 352 000 5/7
С 352 000 3 520 000 7/8
D 3 520 000 Not defined 8/-
 
 
Table 4. Recommended Limits for Airborne Particle Concentration for the Monitoring of non-Viable Contamination in the Annex 1 Draft
Grade Recommended maximum limits
for particles ≥ 0.5 μm/m3
Recommended maximum limits
for particles ≥ 5 μm/m3
In Operation At Rest In Operation At Rest
А 3 520 3 520 20 20
В 352 000 3 520 2 900 29
С 3 520 000 352 000 29 000 2 900
D Set a limit based on the risk assessment 3 520 000 Set a limit based on the risk assessment 29 000
 
 
The draft breaches a fundamental rule:
 
“A scope of start-up testing shall be always bigger (wider) than for routine monitoring” (Table 5). Otherwise, a fault can be overlooked at start-up and discovered at routing control only. These are two steps reversed.
 
Table 5. Volumes of Testing for Start-up and Monitoring
  Normal Practice Annex 1 Draft
Start-up Bigger Smaller
Monitoring Smaller Bigger
 
Figure 4. Relations Between Testing Efforts at Start-up and Routine Control

Fedotov-4.png

 
Aseptic Process and Terminal Sterilization
 
FDA sets requirements for aseptic conditions only. EU GMP Annex 1 sets requirements for both and for Grade A they are equivalent (Table 6).
 
Table 6. FDA and EU GMP Differences for two Kinds of Sterile Processes
Process FDA EU GMP
Aseptic Yes Yes
Terminal Sterilization No
Why?
Yes
 
Is this correct? To answer this question requires a short risk analysis (Figure 5 and Figure 6).
The key difference between two processes is that terminal sterilization provides reliable protection against microorganisms and air cleanliness is not so critical. For aseptic processes, such sterilizing barrier does not exist and air cleanliness in aseptic core is the last line of defense. Requirements for such last line shall be more severe than for terminal sterilization. So equivalent requirements of EU GMP for zones A in aseptic and terminal sterilization processes are in obvious contradiction. Aseptic processes are riskier, yet ISO Class 4 with filters U15 will protect the critical core more reliably.
 
Figure 5. Aseptic Process (Open): Risk Analysis
Fedotov-5.png


 
Figure 6. Terminal Sterilization: Risk Analysis

Fedotov-6.png

 
FDA does not require cleanrooms for processes with terminal sterilization. Does that mean that air cleanliness is not necessary for them? The answer is “no” because the sterilization process is not absolute. Rather, it is relative. It reduces bioburden on certain orders of magnitude by, perhaps, 10-6 or 10-12 for “overkill” process. Achieving of minimal Sterility Assurance Level (SAL = 10-6) depends greatly on biodurden load on/in product before sterilization (Table 7).
 
Particles can deposit on surfaces. If a spore of microorganism is covered by a particle or another contaminant, then a spore is protected from sterilizing agent and risk of its survival is greater (Figure 7). So contamination protects microorganisms against heat and reduces effect of sterilization, but cleanliness of air and surfaces is a priority demand for sterile manufacturing.
 
Figure 7. Microorganisms Under Protection of Particle or Other Kind of Contamination
Fedotov-7.png


 
Table 7. Existing Grades and ISO Classes for air Cleanliness “in Operation”
  EU GMP FDA
Grade ISO Class
Filling – Aseptic A ISO 4,8 ISO 5
Filling – Terminal Sterilization A ISO 4,8 -
Background for Grade A – Aseptic B ISO 7 ISO 7
Background for Grade A – Terminal Sterilization C ISO 8 -
Supporting Clean Areas D No control,
risk analysis
ISO 8
 
There are two clear weaknesses in both GMPs. The EU GMP is weaker for aseptic processes – Grade D is very weak and corresponds almost to air cleanliness in offices. Concentration of particles ≥ 0.5 μm/min offices “at rest” is around 5 – 7 Mio or even less. FDA GMP is weaker for terminal sterilization where no limits are specified.
 
What is a Grade D for?
 
Grade D, or its equivalent, cannot be found in the FDA Aseptic Guide (Table 3). Interestingly, FDA specifies ISO Class 8 “in operation” for supporting areas equivalent to EU GMP Grade C. But the EU GMP Guide specifies the same areas as “nonclassified.” This is an order of magnitude weaker than in the US.
 
Aseptic and Terminal Sterilization Processes are Very Different
 
EU GMP Annex 1 specifies the same Grade A for both (filling). Sterilization is not an absolute process. It is relative to initial bioburden. Contamination reduces effectiveness of sterilization and contaminants from the air can deposit on surfaces and reduce the effects of sterilization. Thus, air cleanliness for processes with terminal sterilization is necessary.
 
There should be different norms for filling (aseptic and terminal sterilization) and replace Grade D in Annex 1 with Grade C; it is easy, and is achieved according to FDA Aseptic Guidance, to start working on the Global GMP specification for cleanrooms (Table 8).
 
Table 8. Recommendations for “in Operation” State for Global Norms Harmonized
  Global Norms Offered
Grade ISO Class
Filling – Aseptic A0 ISO 4
Filling – Terminal Sterilization A ISO 5
Background for Grade A – Aseptic B ISO 7
Background for Grade A – Terminal Sterilization C ISO 8
Supporting Clean Areas C ISO 8
 
Doors Interlocking
 
The existing Annex 1 allows both an “interlocking system or a visual and/or audible warning system to prevent the opening of more than one door at a time.” The draft changes it only to “doors interlocking.” But this is a huge safety violation for personnel where safety should be a priority. This may lead to some dangerous situations provided below.
 
Situation 1: a person cannot leave a cleanroom in emergency if interlocking system is faulty. A person will essentially be in a “mouse trap.” An alarm situation is unusual for people who can easily fall into a panic and escape actions must be accomplished quickly and encounter no barriers, such as may be posed by doors interlocking.
 
Situation 2: it is opposite to situation 1. There is no emergency case, but only a single person is present in the room. He or she may feel sick, open the inner door of an airlock and fall faint. Because the outer door is interlocked, a sufficient amount of time is needed to resolve the situation. This necessary time can exceed what is needed to save the person. This scenario reflects a real case at one factory and suggests that there should be no mandatory door interlocking.
 
Should the existing item 52 in Annex 1 be retained as it is or is it even better to delete the word “interlocking?”
 
A Cascade Enter Procedure
 
Annex 1 Draft specifies a cascade concept for personnel airlocks and changing rooms:
 
1) outdoor → 2) CNC → 3) D → 4) C → 5) B/A. This is an example of unnecessary over-specification. This suggestion also contradicts the well-established safety practice of having reliable operation of very many cleanrooms around the globe. This is also in contradiction to FDA Aseptic Guidance requirements with no analogue to zone D. The weakest cleanliness class, according to FDA, is ISO Class 8 in operation, which is equivalent to EU grade C so that one changing room can be arranged for CNC room → grade C room without a problem (Table 9).
 
Table 9. Comparison of EU GMP Annex 1 and US FDA Requirements, Class in Operation
EU GMP Annex 1 US FDA
Grade ISO Class
А ISO 4,8 ISO 5
В ISO 7 ISO 7
С ISO 8 ISO 8
D No control
 
 
See below cascades using ISO classes.
 
The “at rest” state cascade of airlocks according to the draft is:
 
1) Non-classified area
1-st airlock
2) CNC
2-d airlock
3) ISO 8
3-d airlock
4) ISO 7
4-th airlock
5) ISO 5 (ISO 4,8)
 
However, there are no personnel, material or other sources of contamination in cleanrooms that are “at rest.” There is also no movement through airlocks “at rest” and no contamination transfer. So, this case can be omitted. The “in operation” status is the only occupancy state of interest.
 
An “in operation” state for a cascade of airlocks according to the draft looks like the below diagram:
 
Non-classified area
1-st airlock
CNC
2-d airlock
CNC
3-d airlock
ISO 8
4-th airlock
ISO 7
 
Grade D means CNC “in operation.” So, the draft specifies a separate airlock (2-nd airlock) between two CNC rooms “in operation.” This is obviously confusing and a case of over-specification. A practical schema can look like the chart below. A realistic cascade of airlocks “in operation” state for transfer from outdoor area to grade B:
 
Non-classified area
1-st airlock
CNC
2-d airlock
ISO 8
or
zone C

3-th airlock
ISO 7
or
zone B
 
This requirement is not dogmatic and other flexible solutions can be applied. However, there may be two airlocks between CNC and B zones, with only one changing room. Today’s microelectronics facilities specify much more rigid requirements for cleanliness than pharmaceutical production, but they use only one changing room.
 
Biocontamination Limits
 
FDA and EU GMP Annex 1 specify limits for biocontamination in air (Table 10).
 
Table 10. EU GMP and US FDA Requirements for Microbial Contamination in Operation
ISO Class/
Annex 1 Grade
FDA EU GMP Annex 1
Active air sampling, CFU/m3 Settle plates
90 mm/4 hours
Active air sampling, CFU/m3 Settle plates
90 mm/4 hours
5/A 1 1 ˂1 ˂1
6 7 3 - -
7/B 10 5 10 5
8/C 100 50 100 50
-/D - - 200 100
 
 
Limits for active sampling and settle plates for ISO Class 5/zone A are the same; but for others the ratio is 2, so the limit for active sampling is two times bigger than for settle plates. But what is happening in real air? Different sources give different data for this ratio (Table 11). The second line of Table 11 shows ratios of CFU numbers for active sampling to settle plates with diameter 90 mm and sedimentation time 5 min.8 The third line presents the same date but recalculates to four hours for comparison with Table 10. Numbers in line 2 are divided on 48 (60 min in hour /5 min =12; 12x4=48).
 
Table 11. Summary Researches on Biocontamination in air
Ratio of CFU numbers for:
active sampling/settle plates
Country
USSR (Russia) US (NASA) Japan China
5 min sampling time 157 86 30-600 290
4 hours sampling time 3,3 1,8 0,63 – 12,6 6
 
Different sources provide different numbers for this ratio for four hours: from 1,8 to even 12,6. The mean number is 6, i.e., in real clean zones active sampling shows approximately six times more CFUs than settle plates. So, it would be logical to reduce limits for settle plates appropriately.
 
Any sampling method reflects real biocontamination in air only partly, depending on its sensitivity. How small is this part and what is the real biocontamination level in air? Swedish researchers have shown that a “real time” microbiological air monitoring based on fluorescence of bacteria (rapid method) determines 100 more germs that an active sampling.9 This method does not disturb aseptic zone and sets no risk for the process. Limits for rapid methods should be different from existing ones and special researches are needed to offer them.
 
Conclusion
 
It is necessary to establish global harmonized GMP norms based on cleanroom operations experience and withdraw testing and monitoring of cleanrooms for particles 5.0 µm from EU GMP Annex 1.
It is also important to change the limits for Grade A for aseptic and terminal sterilization processes.
Tighter limits for particles ≥ 0.5 µm (ISO 4) for aseptic filling should be changed. Grade D should be deleted from Annex 1 and harmonization with FDA Guidance is applicable. In addition, for safety reasons, interlocking doors in airlocks should not be mandatory. Finally, existing limits for biocontamination in cleanrooms are not informative enough and research on rapid microbiological (fluorescent) test methods should be conducted with special attention to establishing relevant limits.
 
References
 
  1. Fedotov A. “FDA and EU GMP Annex 1 Differences of Cleanroom Specification. Is it Not Time to Eliminate Them?” Presentation at Interphex 2019, April 2019.
  2. EU GMP Annex 1: Manufacture of Sterile Medicinal Products draft for comments, December 2017.
  3. US FDA Guidance for Industry. Sterile Drug Products Produced by Aseptic Processing Current Good Manufacturing Practice. September 2004.
  4. ISO 14644-1:1999. Cleanrooms and Associated Controlled Environments–Part 1: Classification of air Cleanliness.
  5. Sharp J, Bird A, Brzozowski S and O’Hagan K. “Contamination of Cleanrooms by People.” European Journal of Parenteral and Pharmaceutical Sciences (EJPPS). 2010:15(3).
  6. NASA “Biological Handbook for Engineers.” Marshall Space Flight Center. October 1976.
  7. Reinmüller B and Ljungqvist B. “Evaluation of Cleanroom Garments in a Dispersal Chamber: Some Observations.” European Journal of Parenteral Sciences. 2000:5(3);55-58.
  8. Wang Lai, Tu Guang-bei. “The Relation Between the Density of Airborne Bacteria and the Numbers of Sedimented Bacteria.Swiss Contamination Control. 3 (1990) Nr. 4a, p. 317‒319.
  9. Andersson K, Bjerner G, Kene V and Åkerlund E. “Real Time Microbiological air Monitoring: RenhetsTeknik.” R3 Nordic Magazine. Sweden 2:2011;7‒10.
About the Author
 
Aleksandr Fedotov, PhD, is director at Clean Technologies LLC, Moscow, Russian Federation.
He has 30 years of cleanroom and GMP standardization experience. Fedotov is the Russia Delegate to ISO/TC 209 “Cleanrooms and Associated Controlled Environments” and serves as chair of Russian Mirror Committee, which has developed more than 70 Russian standards on cleanrooms and GMP rules. He may be contacted at fedotov@invar-project.ru.
 
Cite as: Fedotov A. “FDA and EU GMP Annex 1 Differences in Cleanroom Specifications. Is it time to eliminate them?” Regulatory Focus. July 2019. Regulatory Affairs Professionals Society.

Regulatory Focus newsletters

All the biggest regulatory news and happenings.

Subscribe