Guest Column | February 12, 2021

Cleaning Verification & Validation Of Multipurpose API Plants: 9 Rules To Follow

By Grant Mordue

Follow The Rules

Cleaning validation continues to be a hot topic during regulatory inspections and industry discussions. The cleaning required in an active pharmaceutical ingredient (API) plant between one manufacturing process and the next can present a huge challenge. This challenge is greatest when multipurpose plants are used, which are configured, cleaned, and reconfigured to manufacture a wide variety of intermediates and APIs. I’ve experienced many examples of poor practice and best practice during my involvement in the design, execution, and assessment of API plant cleaning methods, analytics, and validation principles and protocols. This article provides a sample of important rules to follow when verifying and validating cleaning the cleaning process for multipurpose API plants.

1. Know your limits.

First, it’s important to understand and determine the scientifically derived acceptance criteria you will need to demonstrate you have achieved when cleaning between one manufacturing process and the next. The acceptance criteria have evolved over the years, from the initial 1,000th of the dose (worst case is the minimum dose) of the contaminating molecule, per dose (worst case is the maximum dose) of the receiving product. This was based on the principle at the time that a patient receiving 0.1x the dose of the contaminating molecule will not experience an adverse reaction; however, a safety factor of 10 was then included, together with an additional robustness factor of 10, hence the 0.001 or 1,000th criteria. It is no longer acceptable to calculate a maximum allowable carryover (MACO) based on the adjustment of an LD50. The calculated criterion was often translated into a default limit of 10 ppm of the contaminant per minimum batch size of the next production (unless the calculated dosage criteria were lower). In the past, this 10-ppm limit was found to be acceptable in the majority of situations.

Exceptions could be made for manufacturing highly potent APIs, where the calculation of acceptance criteria recently evolved further, to require the use of health-based exposure limits (HBELs). HBEL calculation requires input from qualified toxicologists and usually results in limits that are higher (less stringent) than the limits applied in the past, unless the API is highly potent, which would probably require the use of dedicated facilities to avoid cross-contamination.

Once a MACO has been calculated for a specific changeover, it needs to be translated into the analytical limits that can be applied to rinse and swab samples taken from the cleaned equipment. This requires calculating the surface area inside each piece of equipment that becomes contaminated and is required for reuse in a different manufacturing process. Typically, qualified engineers provide internal surface area calculations for each piece of equipment, resulting in inspectable documentation showing the m2 total for each item and for the total equipment train required for reuse.

The analytical limits are calculated by dividing the MACO (per minimum batch size of the next production in grams [g]) by the combined surface area of all the equipment being reused. This is relatively easy for direct surface (swab) analysis; however, rinse analysis is not a direct measure of the contamination remaining on a cleaned surface and requires care. Swab test limits are typically calculated as g or mg of contaminant per 100 cm2 of swabbed surface area (10 cm x 10 cm). A rinse sample provides an indication of the contaminant removed (dissolved) by the rinsing conditions and not specifically the amount of contaminant remaining on the surfaces. It is therefore preferable to combine the rinse analysis with the results of a rinse efficiency study performed in the laboratory (as described below).

2. Don’t double up.

In some API manufacturing sites, the equipment used to manufacture the crystallized API (usually isolated as a wet cake) is in a different building than the equipment used to dry and offload the finished API. A common mistake is to apply the entire MACO to the equipment in the first facility and again in the second facility, resulting in the output potentially containing two times the MACO. The MACO to surface area calculation must be applied to the complete equipment train from input to dried output.

3. Maximize the recovery.

If rinse samples are tested as an indirect measure of cleaning effectiveness, the rinsing conditions, e.g., the solvent mix, temperature, agitation, and time, should be optimized and the recovery efficiency determined by performing laboratory studies. A known amount of representative residue should be spiked onto the surfaces of an equivalent test item, e.g., a stainless steel or glass beaker or flask, dried (as the worst-case challenge to the rinse), and subjected to the rinsing conditions. Samples of the rinse should be taken at intervals and tested to quantify the concentration of the residue rinsed from the surface. The results can be used to determine the optimized rinsing conditions and also the recovery efficiency when using the rinse samples as an indication of the cleanliness of the surfaces.

The analysis of direct surface (swab) samples is accepted as the most reliable indication of surface cleanliness. However, like rinse analysis above, the efficiency of the swab test conditions also needs to be optimized and determined. The choice and use of a swab media that is absorbent and does not generate background interference during analysis is preferred. Similar to the rinse study, a known amount of representative residue is spiked onto a representative surface area, e.g., stainless steel, glass, plastic, etc. A useful approach is to spike the acceptance criteria amount (R) onto a marked 10 cm x 10 cm area and also spike 2xR and R/2 onto additional marked 10 cm x 10 cm areas of the same material. In this way, the swab test and analytical conditions are demonstrated for efficiency and linearity across the range of results, R/2 to 2xR, which should be sufficient.

The moistened swab dissolves the residue from the test surface and retains some of this solution on the swab for analysis. However, a quantity of the residue solution is left behind on the surface. This can be recovered by using a dry swab immediately after the moistened swab, followed by analysis of both swabs combined. In this approach, the use of a non-volatile solvent for the swab testing is preferred. If a volatile solvent must be used, using a second swab immediately after the first can improve the recovery in a similar way. The recovery study described above provides three results for the efficiency of the residue removal, which should be comparable if the conditions are suitable. The results are then used to determine a correction factor, which is applied when calculating the amount of residue remaining on a surface after cleaning. A recovery of less than 50% should not be accepted, and the conditions should be further optimized. The optimized “wet then dry” principle described above typically provides recoveries of 80% or more.

4. Follow the map.

A “residue map” should be prepared that lists the composition of the residue remaining inside each piece of equipment after use. Note: Only the equipment used for the later stages of synthesis will contain the API molecule; therefore, the analysis used to determine the cleanliness of earlier equipment may need to look for residual raw materials, reactants, and intermediates instead.

Where a multistage synthetic route is used to manufacture an API, the carryover of residue to the next stage might represent the presence of raw materials, reactants, or contaminants with respect to that next stage; therefore, the residue map should be used to determine the appropriate cleaning limits for changeover from one stage to the next. Typically, the changeover to non-consecutive stages requires cleaning to a higher standard.

5. Validate the dirty hold time.

After the offload of the last batch, the equipment should be left “dirty” for the worst-case time period (delay) before the start of the cleaning. This should be stated in the cleaning validation protocol.

6. Confirm the appropriate method of cleaning.

When mapping the composition of the residue remaining in the equipment after use, it is useful to examine the design and use of the equipment to determine where residue might be retained and present a challenge to the cleaning conditions available for use. This will determine which equipment can be cleaned in-situ (while connected) and which equipment will require dismantling to expose the surfaces for cleaning.

The initial cleaning typically involves contacting the contaminated surfaces with a cleaning solution, which is then passed through the equipment chain from top to bottom. The cleaning solution composition should be chosen by examining the solubility of the residue composition shown on the residue map under trial conditions of temperature, time, and agitation. If organic solvents cannot be used, aqueous reagents may be trialled and applied, although the output residue composition might be different due to reaction or derivatization that needs to be incorporated into the analytical strategy used to examine surface cleanliness.

The use of (food-grade approved) detergents is seen as introducing contamination and therefore the removal of the detergent after cleaning is an additional requirement that needs to be included. The use of rinsing studies and rinse analysis using total organic carbon (TOC) methodology is an option. In this case, the optimized rinse should remove the detergent until the rinse and the input water have an equivalent TOC output.

Typically, ancillary equipment used directly during manufacturing, such as condensers, pumps, samplers, etc., should be cleaned using conditions that are optimized to ensure maximum surface contact and residue removal. Flexible hoses should initially be cleaned while connected to the equipment. Each flexible hose should be uniquely identified by a code that is clearly shown on a label and/or stamped onto the hose. A register should be maintained that shows the duty and usage history of each hose. If flexible hoses are used for (non-waste) slurry or suspended solids transfer, these should be dedicated for use due to the challenges associated with cleaning and the risk of cross-contamination during reuse in a different duty. A best practice is to clean all flexible hoses separately using a recirculated cleaning solution that is pumped through the hose offline from the plant, using a dedicated cleaning assembly. The interior cleanliness after cleaning should be demonstrated by swab testing each end and inspection of the internal surfaces using an endoscope. If an endoscope is not available, a worst-case study can be performed, which involves cutting the hose into sections after cleaning, followed by the swab testing of each section. Of course, this destroys the hose; however, if the worst case is chosen correctly, this needs to be done only once.

The (manual) cleaning of disassembled equipment is difficult to replicate and therefore typically cannot be validated, although I visited one firm that had installed large automatic washing machines, which are used following a validated loading pattern, cleaning agents, and fixed programming of conditions to avoid the need for manual cleaning and verification of cleanliness after cleaning. If manual cleaning cannot be avoided, the method of cleaning should be carefully described to be as reproducible as possible, with scripted training for the operators involved. After cleaning, the manually cleaned equipment should be verified as clean when dry by visual inspection followed by swab testing. If only a random selection of the cleaned equipment is tested, consideration must be given to re-cleaning all the equipment if a single item does not pass the acceptance criteria.

7. Ensure the standard of cleanliness is acceptable.

The initial criterion for all cleaned equipment is “visually clean when dry.” The reliability of visual inspection, however, is influenced by factors that require careful control:

  • The eyesight of the inspectors, who must pass eyesight proficiency checks at specific intervals.
  • The clarity of the view of the surface to be inspected. Vessels and other large items of equipment must be opened for inspection. The internal surface areas requiring visual inspection must be described and indicated, e.g., by using diagrams or photographs. A best practice is to illuminate the surface for inspection using a beam of light (torch/flashlight) reflected from a mirror (preferably not made from glass), carefully angled and positioned at the end of a telescopic pole. In this way, it is possible to inspect inlets and outlets, together with some otherwise difficult to see areas of the equipment, e.g., the dome and venting of vessels, seals, and the outer surfaces of dip-pipes, etc.
  • The personnel performing the visual inspection must be trained and certified as competent (preferably by QA) to perform the inspection reliably and reproducibly, using the carefully derived and described conditions, equipment, and methods.

If the equipment is not visually clean, it must be re-cleaned and re-inspected (with the generation of the appropriate documentation to track and record all of this additional work). If, after validation, the equipment is not visually clean after cleaning has been completed, a deviation must be initiated to document and investigate the exception and take action accordingly, which might trigger the need for further cleaning optimization and re-validation.

If the equipment is “visually clean when dry,” analysis can then be performed to determine if the standard of cleanliness is acceptable for reuse or not. This will typically involve rinse and direct surface swab analysis. Rinse analysis might also be used during optimization to determine if the equipment is sufficiently clean to move to examination by “visual inspection when dry.” The location of swab testing positions should be determined after a study of the design and use of the equipment, to identify the locations that represent the “most difficult to clean” areas. The locations should be described using diagrams or photographs and training should be completed and recorded for the personnel who obtain the swab samples.

8. Determine an acceptable clean hold time.

When the equipment has been demonstrated as cleaned to the required standard, the length of time during which it can remain sealed and empty before reuse should be determined. This typically involves determining how the equipment surfaces become microbiologically contaminated over a worst-case period of time and, hence, whether a sanitizing rinse should be applied before reuse. The cleaned and held equipment must be confirmed as still visually clean when dry before reuse.

9. Continue verification post-validation.

The above requirements are described in a validation protocol and practical compliance is described and demonstrated in a validation report. Thereafter, the use of identical, reproducible cleaning methods should not require further assessment, although it is typical that partial (or even full) analytical verification continues post-validation.

The minimum criteria post-validation should be “visually clean when dry” as described above. The validation approach can use a “worst-case” demonstration of cleaning effectiveness, based on a combination of factors such as the lowest MACO, the residue potency and solubility, and the difficulty in cleaning specific pieces of equipment. Alternatively, each specific cleaning can be validated after the equipment has been configured and used for a specific manufacturing process. I have also seen the use of a default worst-case standard for all cleaning, followed by a check, based on the reconfiguration of the equipment for reuse, that the residue remaining in the reconfigured full equipment train (based on the analytical data obtained) does not exceed the relevant MACO for the specific changeover before the plant is released for use by QA.

About The Author:

GrantGrant Mordue is the director of Pro-Active GMP Consulting Ltd., a U.K.-based consultancy founded in April 2020 to help companies to successfully implement a proactive level of quality management and cGMP compliance. Mordue has more than 30 years of management experience across the cGMP compliance of manufacturing and supply operations at local (national) and global levels, including the management of regulatory inspections. He has a BSc (Hons) degree in applied chemistry and is a Chartered Chemist and Member of the Royal Society of Chemistry in the U.K. You can connect with him on LinkedIn.