Do Visible Residue Limits Make the 10-ppm Carryover Limit Obsolete?

Cleaning validation is documented evidence that provides a high degree of assurance that a cleaning procedure consistently removes residues to predetermined acceptable levels. These acceptable residue limits (ARL) for drug products are often based on health and adulteration criteria (1–3). The limit used to determine the appropriate cleaning level is the lower of the two criteria. A health-based limit is generated from toxicity data, which can be expressed as acceptable daily intake (ADI) (4, 5). The health limit is calculated using the ADI, or an alternative toxicity factor, and the parameters of the equipment used to manufacture the formulation (2, 6). For example, a health-based limit can be calculated as follows:

(ADI ÷ MDD) × (DUB ÷ SSA) = (μg/cm²)

where the ADI units are μg/day of the residue in question, which would have no pharmacological effect; MDD is the maximum daily dose (units per day) of the product manufactured in the equipment; DUB is the dose units per batch (units) of the subsequent product; and SSA is the shared surface area (cm²) for the product contact surface area of the manufacturing equipment.

For the adulteration-based limit, a carryover limit of 10 ppm or a baseline limit of 100 μg/swab is often used in industry. The following equation provides an example:

10 μg/g × (MBS ÷ SSA) = ARL (μg/cm²)

where 10 μg/g (10 ppm) is the adulteration-based limit; MBS is the minimum batch size (g) of the product; and SSA is the shared surface area (cm²) for the product contact surface area of the manufacturing equipment. Alternatively, an adulteration limit of 4 μg/cm² or 100 μg/swab can be used as an adulteration-based cleaning limit (7).

A third level of acceptance criteria of any cleaning evaluation is that all equipment surfaces must be visibly clean. The visual cleanliness of the equipment must be established before any swabbing can take place to confirm compliance with a health-based or adulteration-based cleaning limit.

The 10-ppm limit

The concept of the 10-ppm carryover limit has been used since the beginning of cleaning validation. The idea of a carryover adulteration limit goes back much further. The latter is used in conjunction with a health-based residue limit; together, the two limits provide a well-defined residue limit to prevent carryover that causes adulteration in pharmaceutical formulation manufacturing. A defined limit of carryover adulteration makes sense for compounds, which have a relatively high health-based residue limit. Not only does this limit prevent unacceptable carryover, but it also ensures a pharmacologically safe formulation.

Fourman and Mullin popularized the 10-ppm criterion for cleaning validation (6). Their program was the first to couple the health-based and carryover criteria in a logical, straightforward manner. They took the 10-ppm limit, calculated it with the equipment-product contact-surface area and the subsequent batch size, and arrived at a swab limit for carryover. It has since been demonstrated that a flat, unadjusted 10-ppm or 100 μg/swab limit can also be appropriate for some applications without affecting compliance, which accompanies a constantly changing cleaning limit in a clinical-manufacturing facility (7). The US Food and Drug Administration cited the 10-ppm limit in its guide to Inspection of Validation of Cleaning Processes as one of several appropriate options for cleaning limits (8). Health Canada and the Pharmaceutical Inspection Cooperation Scheme also use the 10 ppm limit in their cleaning validation guidelines (9, 10). The European Medicines Agency uses similar wording regarding cleaning-validation limits, although the agency does not specifically use the 10 ppm limit (11). Several literature references (12–14) refer to the use of the 10-ppm limit and numerous commercial and research facilities (2, 15–17) have adopted the 10-ppm carryover limit as part of their internal cleaning validation programs.

The 10-ppm limit is used when it is lower than the corresponding health-based limit for the residue of interest. Although the 10-ppm limit has some historic precedent, there is a lack of scientific justification and validation as a limit for cleaning validation. However, the idea of a carryover limit was logical to industry and the 10-ppm limit not only filled this need but also made sense.

Visible residue limits

Fourman and Mullen described a visible residue limit (VRL) at approximately 100 μg per 2 × 2 in. swab area (6) or about 4 μg/cm². Jenkins and Vanderwielen observed various residues down to 1.0 μg/cm² with the aid of a light source (1). Neither offered details or data to substantiate the numbers and neither speculated about the use of visual limits in relation to the 10 ppm limit. LeBlanc questioned whether a VRL as the sole acceptance criterion could be justified (15).

Work at Merck's West Point, Pennsylvania, facility quantitated the use of VRLs for both pharmaceutical pilot plants and commercial manufacturing facilities (18, 19). The experiments included a series of active pharmaceutical ingredients (APIs), excipients, formulations, and detergents spotted onto stainless- steel coupons at decreasing concentrations until a group of observers were unable to detect the residues. The experimentally determined VRLs compared favorably to the health-based and carryover cleaning limits. For those compounds where the VRL was lower than the health-based and carryover limits, the VRL became the primary measure of equipment cleanliness.

Parameters explored included distance, viewing angle, ambient-light level, and residue composition. Established, acceptable viewing parameters for the vast majority of product residue on pharmaceutical manufacturing equipment were: < 10 ft, > 30°, and > 200 lux respectively. The viewing angle proved to be the most critical factor when viewing cleaned equipment, particularly when viewing corners and other non-flat surfaces. VRL training continues to emphasize the importance of the viewing parameters.

Additional studies demonstrated the effectiveness of VRL use. Swab data compared favorably with VRL data as part of a validation study in a clinical packaging facility (20). A VRL study involving five APIs at three different sites demonstrated the ability to transfer VRL data between sites (21). A statistical analysis of all available VRL data (21) demonstrated that as technique improved, VRL data variability decreased. The analysis also concluded that the VRL of an API was representative of a given formulation.

VRLs have replaced swab analysis for several applications, including: the introduction of new APIs or equipment into a facility, routine use inspection after cleaning, periodic assessment of program effectiveness, technology transfer of cleaning methodology, campaign-length extension, cleaning-procedure optimization, and reduced cleaning documentation during routine cleaning.

Comparison of VRL to 10 ppm

The 10-ppm limit in the author's facility equated to 100 μg/25 cm² swab or 4 μg/cm² based on the swab area and the solvent volume used to extract the residue form the swab. As VRL limits were established, they were compared to the 10-ppm swab limit for compounds for which the 10-ppm limit was lower than the health-based cleaning limit. The majority of the VRL data generated were well below 4 μg/cm² (see Table I). Once the sample- preparation parameters and spotting technique were refined, 89% of experimentally determined VRLs were less than 2 μg/cm² and 98% of the VRLs were below 4 μg/cm² (see Table I).

Table I : Comparison of visible residue limit (VRL) data.

The VRL data for several commercial formulations are shown in Table II. All of the VRLs are well below the constant 4μg/cm² limit. The data range from < 1.88 μg/cm² and 1.45 μg/cm² for Demser (metyrosine) and Emend (aprepitant) respectively, to <0.07 μg/cm² for Sinemet (carbidopa-levodopa) and < 0.06 μg/cm² for Aldomet (methyldopa). Differences between the 10-ppm adulteration cleaning limit and the VRL cleaning limit range from a factor of about 2 for Demser to a factor of almost 70 for Aldomet (see Figure 1).

Table II: Visible residue limit (VRL) data.

Comparison of VRL with the health-based limit

The health-based limit is directly related to the ADI as well as to manufacturing equipment and batch size parameters. Health-based swab limits were calculated for several commercial formulations (see Table II). Other than ADI, a representative set of large batch parameters was used: MDD = 1 tablet, DUB = 1,000,000 tablets, and SSA = 25,000 cm². The health-based limits range from 2000 μg/cm² for Aldomet, Isentress and Sinemet, to 10 μg/cm² for Zocor (simvastatin). Using the same VRL data, the margin between the VRL data and the health-based limits is shown in Table II and Figure 2. Differences between the health-based cleaning limit and the VRL cleaning limit range from a factor of < 18 for Zocor to well over 30,000 for Aldomet.

Figure 1: Visible residue limit versus 10 ppm (4 μg/cm2) limit. ADI is acceptable daily intake. Refer to Table II for letter key. (ALL FIGURES ARE COURTSEY OF THE AUTHOR)

Comparison of 10 ppm VRL with health-based limits

The cleaning limit for a pharmaceutical residue will initially be determined by the health-based limit calculation. The cleaning limit for residues that are not highly potent (i.e., that have a health-based limit greater than 10 ppm) or that defaulted to the 10-ppm or adulteration cleaning limit. However, the visual cleanliness of equipment is a regulatory requirement. For the category of pharmaceutical compounds that are not highly potent, the VRL is almost always lower than both the health-based limit and the 10 ppm adulteration limit (see Table I). The average VRL for all APIs, excipients ,and formulations tested to date is 1.1 μg/cm² (21). For these compounds, the visual inspection is the most stringent assessment of equipment cleanliness.

Figure 2: Visible residue limit versus health-Based and adulteration-based limits. ADI is acceptable daily intake. Refer to Table II for letter key.

With the health-based limit driving patient safety and the visual limit as the lowest cleaning limit, the adulteration-based limit ceases to be of importance. The 10 ppm adulteration limit is bracketed by the visual limit on the low end and the health-based limit on the high end. As a result, the VRL is the determining factor for equipment cleanliness.

It is also misleading to compare the visual limit with the adulteration limit as a margin of safety against cleaning failure. Two examples illustrate this point. The adulteration limit of 10 ppm or 4 μg/cm² for Demser is two times greater than the VRL of about 2μg/cm². The health-based limit for Demser is 400μg/cm², which is about 200 times greater than the VRL, which is a much greater margin of safety. The adulteration limit of 10 ppm or 4 μg/cm² for Zocor is about seven times greater than the VRL of < 0.57 μg/cm². The health-based limit for Zocor is 10 μg/cm², which is about 18 times greater than the VRL. Relating the VRL to the 10-ppm limit gives an artificially greater risk of failure compared with an evaluation of the VRL to the health-based limit. Data has consistently shown (18, 21) that when the ADI for a compound is > 100μg/day, the health-based cleaning limit is higher than the 10-ppm adulteration-based cleaning limit and the margin is greater compared to the VRL of the compound. It is scientifically justifiable to compare the VRL with the health-based limit in these cases.

Conclusion

A scientifically determined VRLprogram makes a 10 ppm adulteration cleaning limit obsolete for a large number of pharmaceutical compounds. When the VRL is lower than 4 μg/cm², the visual limit satisfies the dual regulatory requirements of providing visually clean manufacturing equipment and eliminating residual carryover that can lead to toxicological or adulteration concerns of the subsequent formulation.

Richard Forsyth is a private consultant for cleaning validation and good manufacturing practice issues, tel. 610.948.2970, mr.rich.forsyth@gmail.com

Submitted: Mar. 23, 2009. Accepted: Apr. 10, 2009.

References

1. K.M. Jenkins and A.J. Vanderwielen, Pharm. Technol., 18 (4), 60–73 (1994).

2. R. J. Forsyth and D. Haynes, Pharm. Technol., 22 (9), 104– 112 (1998).

3. D. A. LeBlanc, D. D. Danforth, and J. M. Smith, Pharm. Technol., 17 (10), 118–1124 (1993).

4. B. D. Naumann and P. A. Weideman, Human Ecol. Risk Assess. 1 (5), 590–613 (1995).

5. B. D. Naumann and E. V. Sargent, Occup. Med. 12 (1), 33–42 (1996).

6. G. L. Fourman and M. V. Mullen, Pharm. Technol., 17 (4), 54–60 (1993).

7. R. J. Forsyth, A. LeBlanc, and M. Voaden, Pharm. Technol., 31 (1), 74- 83 (2007).

8. FDA, "Guide to Inspection of Validation of Cleaning Processes" (Rockville, MD, July 1993).

9. Health Canada, "Cleaning Validation Guidelines" (June 2002).

10. PIC/S, Recommendation on Validation Master Plan, Installation and Operational Qualification, Non-sterile Process Validation, Cleaning Validation (PI 006-2, July 2004).

11. EudraLex, "Qualification and Validation" (Annex 15, July 2001).

12. M. Lazar, Pharm. Technol., 21 (9), 56–73 (1997).

13. D. A. LeBlanc, PDA J. Pharm. Sci. and Technol., 56 (1), 31–36 (2002).

14. J.M. Cardot and E. Beyssac, Cont. Environ., 11 (6), 11–14 (2008).

15. R. J. Romanach et. al., Pharm. Technol., 23 (1), 46–58 (1999).

16. M. C. Oliver, Eur. J. Parent. Sci., 5 (4), 97–100 (2000).

17. J. E. Martinez, Pharm. Technol., 26 (11), 62–74 (2002).

18. R. J. Forsyth and V. Van Nostrand, Pharm. Technol., 28 (10), 58–72 (2004).

19. R. J. Forsyth and V. Van Nostrand, Pharm. Technol., 29 (10), 152– 161 (2005).

20. R.J. Forsyth et al., Pharm. Technol., 30 (11), 90–100 (2006).

21. R.J. Forsyth, Pharm. Technol., 33 (3), 102–111 (2009).