Allan Ader

	The revised Chapter 5 of Part 1 of the European Union’s GMP guidelines and the European Medicines Agency’s (EMA) 

Guideline on Setting Health Based Exposure Limits for Use in Risk Identification in the Manufacture of Different Medicinal Products in Shared Facilities

 came into effect in 2015 (1, 2). Both publications expect the industry to use a “quality risk management process, which includes a potency and toxicological evaluation” when deciding whether to use dedicated or shared facilities to produce pharmaceuticals. Consequently, the same comprehensive approach based on health risks should be followed to determine the acceptable limits set for cleaning shared equipment. The authors believe it is the right time to abandon the 10-ppm rule in favor of more scientific and data-driven health based limits.


	Health authorities do not require shared equipment to be cleaned to 10 ppm of the next product. Careful reading of GMP regulatory texts (3-11), FDA’s 

Guide on Inspections--Cleaning Validation 

(12), and the Pharmaceutical Inspection Convention and Pharmaceutical Inspection Co-operation Scheme’s (PIC/S) Recommendations on Cleaning Validation (13) yields no indication that any of these health authorities or PIC/S ever explicitly prescribed numerical cleaning limits. When it comes to cleaning limits, these guidelines all follow a similar discourse--they expect companies to abide by a general duty of care, sometimes adding that the limits should be “based on the knowledge of the materials involved.” Some of the texts mention the three traditional rules originally proposed by Fourman and Mullen (14) (i.e., visually clean equipment, residue levels ≤ 1/1000th daily dose, and ≤10 ppm of next product) as examples but never as binding rules. In this case, consistent with other performance-based regulations, the responsibility to establish and justify cleaning limits for shared facilities remains with pharmaceutical manufacturers.

	The 10-ppm limit goes against the basic principle of toxicology that it is the dose that makes the poison. Parts per million (ppm) is a concentration unit (i.e., a relative number and not an absolute one). In chemistry parlance, it is an intensive property (such as density or melting point) that is independent of the amount of the material, and as such, it is not suitable to describe a dose, which is an extensive property (such as mass or volume) that is dependent on the amount of material. Defined as a concentration, the limit dose will depend on the maximum amount of something else, in this case, on the amount of the following product (or API, the approaches vary) administered to the patient. Assuming that therapeutic doses range from 0.1 mg to 5000 mg per day, the amount of a contaminating substance corresponding to 10 ppm will span the same range of 4 logs. Using a limit of 10 ppm without risk characterization of exposure is equivalent to pretending that it makes no difference if a drug is contaminated by 1 ng (the equivalent of 10 ppm of 0.1 mg) or by 50 mcg (the equivalent of 10 ppm of 5000 mg) of the same substance.

	Whether the contaminant is Botox or a mild antacid, the limit amount computed by 10 ppm will be the same. Therefore, involvement of a risk assessment scientist (e.g., toxicologist or physician) is crucial for the characterization and differentiation of health risks and the overall risk management process. 

 compares the weight of scientific justification for three different approaches to compute the maximum safe surface residue (MSSR). In contrast to approaches targeting 10 ppm and 1/1000th of the daily dose, the permitted daily exposure (PDE) approach ensures that available preclinical and clinical data are considered in computation of the MSSR.

Does the 10-ppm limit ensure good cleaning?


	No; 10 ppm is a concentration unit that gives no indication on cleanliness. Operationally, a sensible physical unit for cleanliness must express the amount of residue per unit surface area (e.g., mg/m

). By switching from carryover (amount of residue [g]) to cleanliness [mg/m

], the surface is normalized, and the cleanliness of all pieces of equipment of a facility can be compared; cleaning results between facilities and even between companies can also be compared. When speaking of cleanliness, the size of the equipment does not matter; “cleanliness” is an intensive property of the equipment. Therefore, until the 10 ppm has been converted into a cleanliness value in [mg/m

], it is not possible to say what level of cleanliness the 10 ppm eventually represent. Depending on the particular situation, 10 ppm may mean 1 mg/m

Can 10 ppm be used to identify risks according to the EMA guideline on shared facilities?


	The EMA guideline (2) describes primarily a method for computing health-based limits based on all available toxicological and pharmacological information, which are termed permitted daily exposure (PDE). The 

guideline’s focus is not on how to set cleaning limits. The PDEs are only one element needed to set cleaning limits, which are calculated as MSSR. The other element is the knowledge of the efficiency of a cleaning process, which is independent from the next product. Therefore, while there is a link between PDEs and cleaning limits, they are absolutely not the same. This is expressed in the EMA guideline, which says, “these levels can be used as a risk identification tool and can also be used to justify carryover limits used in cleaning validation.”

	An evaluation of the effect of using 10 ppm in addition to PDE to identify risks was performed with the example of an actual product portfolio of a multiproduct API manufacturing facility making 11 different APIs. The basic data of the APIs are given in 

	All changeovers from a preceding product (α) to a following product (β) were considered. The matrix of maximum safe carryovers (MSC) was first calculated according to 

. In a second step, the matrix of MSSR was calculated by dividing the MSC by the equipment surface area used to make each following API (see 

	Each cell of the MSSR matrix shown in Table II reflects an individual product change from an API α to an API β and contains the MSSR for the particular change (i.e., the cleanliness required to make the corresponding API β when it is made after API α).

 shows a vivid landscape with strong contrasts, with values from 5 to 146,970 mg/m

. It is easy to identify the changeovers that pose the greatest risks. Comparing the calculated MSSRs with 

the results of approximately 250 swab results that were gathered over two years, 99% of all swabs gave results <25 mg/m

; high-risk changeovers are identified as the ones that require a MSSR of < 25 mg/m

	For example, the changeover from product D to product E was calculated as follows:
	

	Five products C, G, I, J, and K are potentially at risk, but only when they follow certain products, like A, G, I, and K. Similarly A, G, I, and K represent a risk for the other products, but not for all of them.

	The comparison of the matrix of MSSR based on PDE with the known efficiency of the cleaning processes allows for a sound identification of risks in line with the spirit of the EMA guideline. The comparison may give the concept of “worst-case approach to cleaning validation” a new orientation by shifting the focus from the hazards of worst-case products (with regard to solubility, cleanability, toxicity, etc.) to the worst-case risks, (i.e., to the “risky” changeovers between two products). Such risks will have to be mitigated, for example, by development of better cleaning procedures, avoiding high-risk changeovers by careful campaign scheduling and dedication of certain process lines or parts of the equipment to a group of products. It is important, however, to accept that cleaning will have its limits: There will be cases where the risk cannot be mitigated by cleaning alone.

	How does the risk identification work with the 10-ppm rule? Because the calculation is independent of the preceding products, the matrix collapses to a list of batch sizes divided by 105 (see 

based MSSRs all come out in a narrow range, between 9 and 59 mg/m2 and they do not really allow clear differentiation between the high- and low-risk changeovers. Compared to PDE, the 10-ppm calculation does not improve the detection of high risks.

	In practice, the 10-ppm methodology gives the impression that there are no low-risk changeovers. If the 10-ppm row in Table III is followed, whenever C, D, I, or K is made, the manufacturer has to meet the MSSR dictated by 10 ppm. The assumed cleaning limit of 25 mg/m

 will not meet the requirement to make C, D, I, or K in any case, independently of what the preceding product was.

 merely reflect the low risk associated with the corresponding changeovers. They should not be read as cleaning limits. The suspicion that production will set the cleaning limit at the level of the PDE-derived MSSR is not justified. Even if the MSSR values for particular changeovers are high, this does not justify reducing the cleaning effort, because cleaning will still have to fulfill the requirement of visually clean equipment. This will be sufficient to ensure a suitable minimal hygiene standard.

	The cleaning limit should also be based on the knowledge of the capability and the reliability of the cleaning processes, and it will ideally be set as far below the MSSR as reasonably possible.


	There is no scientific justification for a limit concentration that ignores the hazards presented by potential contaminants. Implementation of the 10-ppm criterion in cleaning validation does not ensure safe cleaning. and it may distort the identification and analysis of cross contamination risks. Moreover, it may impose more stringent limits for cleanliness than are required by health authorities without adding to the patient safety. Eliminating an unnecessary requirement serves to focus attention and resources on high-risk situations. The 10-ppm criterion offers no scientific or practical benefit and can be discontinued in cleaning validation practice.


	The authors would like to thank Ed Sargent and Courtney M. Callis for their views during compilation of this manuscript.


	1. European Commission, The Rules Governing Medicinal Products in the European Union; Volume 4 EU Guidelines for Good Manufacturing Practice for Medicinal Products for Human and Veterinary Use (Brussels, August 2014).
	2. EMA, Guideline on Setting Health Based Exposure Limits for Use in Risk Identification in the Manufacture of Different Medicinal Products in Shared Facilities (London, November 2014).
	3. European Commission, Volume 4 Good Manufacturing Practice (GMP) Guidelines, Parts I & II, http://ec.europa.eu/health/documents/eudralex/vol-4/index_en.htm, accessed Dec. 1, 2015.
	4. European Commission, Volume 4: EU Guidelines for Good Manufacturing Practice for Medicinal Products for Human and Veterinary Use, Annex 15: Qualification and Validation (Brussels, February 2014).
	5. FDA, Code of Federal Regulations, Part 211 Current Good Manufacturing Practice for Finished Pharmaceuticals, www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?CFRPart=211, accessed Dec. 1, 2015.
	6. Ministério Da Saúde Agência Nacional De Vigilância Sanitária, Dispõe sobre as Boas Práticas de Fabricação de Insumos Farmacêuticos Ativos (Brazil, December 2014).
	7. Agência Nacional de Vigilância Sanitária--Anvisa, Validação de limpeza para farmoquímicas, http://portal.anvisa.gov.br/wps/wcm/connect/2443e9804f1af51ea138bdc88f4b6a31/Guia+de+valida%C3%A7%C3%A3o+de+limpeza+para+Farmoqu%C3%ADmicas.pdf?MOD=AJPERES, accessed Dec. 1, 2015.
	8. Health Canada, Cleaning Validation Guidelines, Guide-0028 (Ontario, January 2008).
	9. South Korea, Good Manufacturing Practice for Medicinal Products for Human Use, Enforcement Regulation of Pharmaceutical Affairs Act [Annex 2], http://wenku.baidu.com/view/42de6028ed630b1c59eeb53e.html, accessed Dec. 1, 2015.
	10. South Korea, Provisions for Conducting Validation for Drugs, etc, The KFDA Notification No. 2009-173 (December 2009)
	11. China FDA, Good Manufacturing Practice for Drugs (2010 Revision), http://eng.sfda.gov.cn/WS03/CL0768/65113.html, accessed Dec. 1, 2015.
	12. FDA, Validation of Cleaning Processes (7/93), Guide to Inspections: Inspections of Cleaning Processes (7/93), www.fda.gov/ICECI/Inspections/InspectionGuides/ucm074922.htm, accessed Dec. 1, 2015.
	13. PIC/S, Recommendations on Validation Master Plan, Installation and Operational Qualification, Non-Sterile Process Validation, Cleaning Validation; PI 006-03 (September 2007).
	14. G.L. Fourman and M.V. Mullen, Pharmaceutical Technology, April issue, 54-60 (1993).

:
	When referring to this article, please cite it as M. Crevosier, "Cleaning Limits-Why the 10-ppm Criterion should be Abandoned," 

Articles

The 10-ppm criterion for the acceptable concentration of potential API in cleaning validation to minimize cross contamination into next product has been employed for many years. This article describes why the 10-ppm criterion, which was established based on analytical limitations and estimates of acceptability, is no longer necessary and why a risk-based approach should be universally adopted.