[
New Perspectives on Cleaning.
Rizwan Sharnez and Angela To
Strategies for Setting
Rational MAC-Based Limits
Part III–Leveraging Characterization
and Toxicological Data
Rizwan Sharnez and Angela To
“New Perspectives on Cleaning” is an ongoing
series of articles dedicated to cleaning-process
development, validation, and monitoring. This column addresses scientific principles, strategies, and
approaches associated with cleaning that are relevant in everyday work situations.
Reader questions, comments, and suggestions
are requested for future discussion topics. These
can be submitted to the column coordinator Rizwan
Sharnez at [email protected].
SUMMARY
Conventional approaches for setting cleaning validation acceptance limits for multiproduct (i.e., shared)
equipment are often more stringent than necessary.
Furthermore, the acceptance limits derived using these
approaches can be impractical to achieve or below the
limit of quantitation (LOQ) of the analytical method.
Practical solutions to these issues are discussed in
this series.
The first two parts of this series described strategies for setting rational acceptance limits based on
the conventional maximum allowable carryover
(MAC) approach (1, 2). This article describes strategies for setting rational acceptance limits for process
For more Author
information,
go to
gxpandjvt.com/bios
24
Journal
residues based on characterization and toxicological
data. These strategies are based on clearance, acceptable daily exposure (ADE), and degradation of the
previously manufactured or contaminating active
pharmaceutical ingredient (API) (i.e., the API in the
product that is manufactured before the execution
of cleaning).
INTRODUCTION
The conventional MAC-approach for setting cleaning validation acceptance limits for APIs has several
drawbacks. One drawback is that it is based on overly
restrictive assumptions. For instance, the surface concentration of the contaminating API is assumed to be
equal to that measured at the worst-case swab location. Furthermore, the contaminating API is assumed
to be uniformely distributed across the entire surface
of the equipment. It is also assumed that all of the
residual contaminating API will be transferred into
the subsequently manufactured batch. An important
implication of these restrictive assumptions is that the
estimated carryover of the previously manufactured
API into the subsequently manufactured product can
greatly exceed the maximum carryover that is physically possible. As a result, the calculated acceptance
[
ABOUT THE AUTHORS
Rizwan Sharnez, Ph.D., is principal engineer at Amgen, Colorado. He has more than 15 years of
experience in the pharmaceutical industry. He may be reached by e-mail at [email protected].
Angela To has a Bachelors degree in chemical engineering from Rice University.
of
Validation T echnology [Summer 2011]
iv thome.com
Rizwan Sharnez and Angela To.
limits can be much more stringent than necessary,
and in some instances, well below the LOQ of the
analytical method. Strategies for addressing these
issues were discussed in the first two parts of this
series (1, 2). The strategies were used to justify higher
acceptance limits for residues associated with APIs.
Higher acceptance limits can reduce cycle times and
enhance success rates for cleaning operations. They
also facilitate the elimination of complex productspecific assays (PSA) in favor of simpler and more
cost-effective nonspecific assays, such as total organic
carbon (TOC).
Other drawbacks of the conventional MAC approach
are that it does not account for degradation and clearance of the contaminating API during the cleaning and
subsequent purification operations. Furthermore, it is
based on default safety limits such as 1/1000 of the lowest
clinical dose. As a result, limits based on the conventional
MAC approach can be overly stringent or not sufficiently
restrictive. This article discusses the drawbacks of the
conventional MAC approach. Strategies for addressing
these issues and setting rational acceptance limits for
APIs on the basis of characterization and toxicological
data are also discussed.
CLEARANCE OF CONTAMINATING API
The conventional MAC approach does not account for
clearance of the contaminating API by the purification process of the subsequently manufactured API.
Clearance of the contaminating API can justify higher
acceptance limits for multiproduct cleaning. This strategy can be effective if the purification process for the
subsequently manufactured product can selectively
remove the contaminating API. For instance, if the
contaminating API is a monoclonal antibody (mAb)
and the subsequently manufactured API is a microbial
product, it is very likely that the mAb will be selectively
removed by the purification process of the microbial
product, and vice versa. In principle, for every log reduction in the concentration of the contaminating API, a
tenfold higher acceptance limit could be justified for
the shared equipment train upstream of the relevant
purification step.
To implement the above approach, additional characterization studies would have to be performed to evaluate
the clearance (i.e., log reduction) of the contaminating API by the purification process of the subsequently
manufactured product. These studies can be laborious
and time-consuming. Depending on the scope and rigor
of the characterization studies, there may also need to be
additional validation work to demonstrate the effectiveness of the approach at full scale.
gxpandjv t.com
DEGRADATION OF
API DURING CLEANING
Another drawback of the conventional MAC approach is
that it does not account for degradation of the contaminating API during cleaning. This factor is important to
consider when using aggressive cleaning conditions. For
example, biopharmaceutical cleaning operations are typically designed to expose process soils to extremes of pH
(<2 and >13) and high temperatures (60–80∘C) for several
minutes. The equipment may also be steam sterilized or
sanitized after cleaning. Under these conditions, monoclonal antibodies and other biologicals are known to degrade
and denature rapidly, thereby becoming inactivated (3–7).
Degradation of the API during cleaning can be characterized by exposing the process soil to simulated cleaning conditions at bench scale (3, 4). The bench-scale
studies are designed to simulate the mildest conditions
for degradation. For example, for alkaline washes, the
cycle that exposes the process soil to the lowest pH,
temperature, duration, and ratio of cleaning solution to
process soil is simulated. The sample is then neutralized
and cooled to minimize further degradation. The degree
of degradation is evaluated by subjecting the sample
and an untreated control to the appropriate assay (e.g.,
SDS PAGE for biological products). If the results indicate that the API degrades, developing product-specific
assays (PSAs) for cleaning may be of little value. PSAs
are difficult and laborious to qualify and can result in
false negatives if the epitopes that they are designed to
detect degrade during cleaning (8). Instead, it may be
more appropriate to develop simpler nonspecific assays
such as TOC for detecting degradants. Methodologies for
setting acceptance limits for degradants will be discussed
in a forthcoming article.
MAC BASED ON ACCEPTABLE
DAILY EXPOSURE OF API
The conventional MAC approach is based on default exposure limits such as 1/1000 of the lowest clinical dose. A
drawback of this approach is that the resulting acceptance
limits can be overly stringent or not sufficiently restrictive.
A more science-based approach is to use toxicological
data to estimate an acceptable daily exposure (ADE) for
the contaminating API (9–21). The ADE for a substance
represents a dose that is unlikely to cause an adverse effect
if an individual is exposed to that substance at or below
this dose every day for a lifetime. It is derived from no
observed adverse effect level (NOAEL) or lowest observed
adverse effect level (LOAEL)—the dose at which a significant adverse effect is first observed based on animal and
human toxicological data according to the following
equation:
Journal
of
Validation T echnology [Summer 2011]
25
New Perspectives on Cleaning.
Table I: Uncertainty factors for estimating ADEs.
Uncertainty Factor
Accounts for:
Published values
References
F1 -Intraspecies Differences
Variability between individuals
Default value of 10
(24)
F2 – LOAEL-to-NOAEL
Extrapolation
Extrapolation of LOAEL data to
estimate NOAEL
3
(28-30)
F3 – Subchronic to chronic
extrapolation
Duration of study being shorter
than that of actual exposure
2-3
(24-27)
F4 - Interspecies differences
Variability between species
2-12
(2, 22-24)
F5 - Database completeness
Incomplete overall toxicity
database
3
(24)
noael or Loael × BW
ADE (mg/day) =
[Equation 1]
Fc × MF × PK
in which, the NOAEL or LOAEL is in mg/kg, BW is
body weight (kg), FC is the composite uncertainty factor,
MF is the modifying factor, and PK is the pharmacokinetic adjustment. FC, MF, and PK are dimensionless.
The composite uncertainty factor, FC, is the product
of the individual uncertainty factors described in Table
I (22-30), as follows:
Fc = F1 × F2 × F3 × F4 × F5
[Equation 2]
The scientific basis for the uncertainty factors has been
described by Naumann et al. (19).
The modifying factor (MF) is a number between 0
and 10 that may be used to address uncertainties not
accounted for by the composite uncertainty factors (e.g.,
the number of animals tested) (24). The pharmacokinetic
adjustment (PK) accounts for route-to-route extrapolation. For example, data obtained from a study conducted
by one route (e.g., oral) can be used to estimate an ADE
for a different route of exposure (e.g., parenteral).
The above methodology for estimating ADEs is an
accepted approach for noncarcinogens. The uncertainty
factors are typically based on scientific data and determined by toxicologists. In the absence of adequate scientific data, default uncertainty factors of 10 each are
considered adequately protective for noncarcinogens (31,
32). However, this practice can result in overly stringent
acceptance criteria for cleaning validation, which in turn
can lead to cleaning procedures that are overdesigned
and to higher operating costs.
For small molecules with limited or no toxicity data,
the threshold of toxicological concern (TTC) can be used
to derive ADEs (33). TTCs for three broad categories of
compounds are given in Table II.
For the conventional MAC approach, the maximum
allowable carryover of the contaminating API (A) into
26
Journal
of
Validation T echnology [Summer 2011]
Table II: Threshold of toxicological
concern (TTC) (33).
Compounds that are:
TTC (µg/day)
Likely to be carcinogenic
1
Likely to be potent or toxic
10
Not likely to be potent, toxic,
or carcinogenic
100
a batch of the subsequently manufactured product (B)
(abbreviated MACAB) is given by Equation 3 (1):
MACAB ≤
TDmin, a × Bb, min
SF × Db, max
[Equation 3]
in which, TDMIN, A is the minimum therapeutic dose of
A, BB, MIN is the smallest integral batch of B (worst-case)
that is manufactured in the facility, DB, MAX is the largest
(worst-case) dose of B, and SF is the safety factor. The
above MAC limit applies to the equipment train that is
shared between the two products A and B.
For a toxicity based assessment, TDMIN, A /SF in Equation 3 is replaced with the ADE, as seen in Equation 4,
MACAB ≤
ADE × Bb, min
Db, max
[Equation 4]
where, ADE is given by Equation 1.
Note that if toxicity studies for the contaminating API (A)
are robust and comprehensive, MF and PK can be dropped
from Equation 1. Furthermore, if the toxicity studies are
appropriately designed and executed, F4 and F5 can also
be dropped from Equation 1. The composite uncertainty
factor, FC, would then simplify to the product of the remaining three uncertainty factors, F1 x F2 x F3. Based on the
iv thome.com
Rizwan Sharnez and Angela To.
default value of 1/10 for each of these uncertainty factors,
FC would be 1/1000. This is equal to the safety factor (SF)
that is typically used with the conventional MAC approach.
most contaminants, and the approach described here can
be extended to other applications, such as the carryover
of cleaning agents.
CONCLUSION
REFERENCES
The conventional approach for setting MAC-based acceptance limits for multiproduct cleaning validation is based
on some overly restrictive assumptions. An important
implication of these assumptions is that the calculated acceptance limits based on the convntional MAC
approach can be much more stringent than necessary.
Further, the acceptance limits are often impractical to
achieve or significantly below the LOQ of the analytical method. Strategies for addressing these issues were
discussed in this series.
Other drawbacks of the conventional MAC approach
are that it does not account for clearance and degradation
of the contaminating API. Further, it is based on default
safety limits such as 1/1,000th of the lowest clinical dose.
As a result, acceptance limits based on the conventional
MAC approach can be overly stringent or not sufficiently
restrictive. A more science-based approach is to use toxicological data to estimate an ADE for the contaminating API.
The strategies described in this series can be used to
derive rational MAC-based higher acceptance limits for
cleaning validation. Rational acceptance limits can lead
to more efficient cleaning cycles, fewer unwarranted
cleaning validation failures, and lower operating costs.
They can also eliminate complex PSAs in favor of simpler and more cost-effective non-specific assays, such as
TOC. This strategy is particularly useful for biological
APIs for which product-specific immunoassays (PSIA)
can give false negatives, and are difficult to develop for
cleaning applications (8). The false negatives stem from
the fact that the epitopes that these assays are designed
to recognize can degrade when exposed to extremes
of pH and temperature during cleaning. If an epitope
degrades during cleaning, the PSIA may not provide
an accurate estimate of the residual API; therefore, the
assay may not be appropriate for verifying MAC-based
acceptance limits.
Degradation of the API during cleaning and steaming
operations has important implications for cleaning validation of multiproduct equipment. Demonstrating that the
API degrades during these operations obviates the need to
perform MAC-based assessments for APIs. It also eliminates
the need to develop PSAs for cleaning validation.
The analyses presented in this series are based on the carryover of one product into another (i.e., cross-contamination of multiproduct equipment). However, the underlying
principles for converting MAC-based limits to acceptance
criteria for cleaning validation samples are the same for
1. Sharnez, R., “Strategies for Setting Rational MAC-based Limits: Part I - Reassessing the Carryover Criterion,” Journal of
Validation Technology, p. 71-74, Winter 2010.
2.Sharnez, R. To, A., Klewer, L., “Strategies for Setting Rational
MAC-based Limits: Part II–Application to Rinse Samples,”
Journal of Validation Technology, Spring 2011.
3.Kendrick, K., Canhoto, A., Kreuze, M., “Analysis of Degradation Properties of Biopharmaceutical Active Ingredients as Caused by Various Process Cleaning Agents and
Temperature,” Journal of Validation Technology, Vol 15, No.
3, p. 69, 2009.
4.Rathore, N., Qi, W., Chen, C., Ji, W., “Bench-scale Characterization Of Cleaning Process Design Space for Biopharmaceuticals,” Biopharm Int, Vol 22, No. 3, 2009.
5.Nottingham, P.M., “The Alkaline Hydrolysis of Haddock
Actomyosin II- Peptide Bond Hydrolysis,” J Sci Food Agric,
Vol 6, p. 86, 1955.
6.Warner, R.C., “The Alkaline Hydrolysis of Egg Albumin,” J
Biol Chem, Vol 142, p. 741, 1942.
7. Sheen, S.J., Sheen, V.L., “Characteristics of fraction-1-protein
degradation by chemical and enzymatic treatments,” J Agric
Food Chem, Vol 35, p. 948, 1987.
8.Health Canada, Cleaning Validation Guideline (GUIDE0028); Section 8.3, Health Products and Food Branch Inspectorate, 2008. http://www.hc-sc.gc.ca/dhp-mps/compliconform/gmp-bpf/validation/index-eng.php
9.ISPE, Risk-Based Manufacture of Pharmaceutical Products: A
Guide to Managing Risks Associated with Cross-Contamination
(First ed. Vol. 7), Tampa, FL: ISPE., 2010.
10.EPA, A Review of the Reference Dose and Reference Concentration Processes; EPA/630/P-02/002F; Risk Assessment
Forum, U.S. Environmental Protection Agency: Washington,
DC, December 2002, p. 192.
11.EPA, Guidelines for Carcinogen Risk Assessment (EPA/630/
P-03/001F). Washington, DC: Risk Assessment Forum,
NCEA, 2005.
12.FDA, Guidance for Industry: Estimating the maximum
safe starting dose in initial clinical trials for therapeutics
in adult healthy volunteers, Center for Drug Evaluation
and Research (CDER), 2005. http://www.fda.gov/cder/
guidance/index.htm.
13.ICH, Impurities In New Drug Substances: Q3A(R): International
Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use, 2002.
14.JECFA, “Evaluation of certain food additives and contaminants. Safety evaluation of flavouring agents,” Forty-ninth Report of the Joint FAO/WHO Expert Committee on Food Additives,
gxpandjv t.com
Journal
of
Validation T echnology [Summer 2011]
27
New Perspectives on Cleaning.
WHO Technical Report Series 884; World Health Organization: Geneva, 1999.
15.IPCS, Assessing human health risk of chemicals: The derivation of guidance values for health-based exposure limits,
Environmental Health Criteria 170, Geneva, World Health
Organization, 1994.
16.IPCS, Guidance document for the use of data in development of Chemical Specific Adjustment Factors (CSAFs) for
interspecies differences and human variability in dose/concentration response assessment; World Health Organization,
Geneva, 2001. http://www.who.int/ipcs/methods/harmonization/areas/uncertainty/en/
17. Conine, D.L., Naumann, B.D., and Hecker, L.H., “Setting
Health-Based Residue Limits for Contaminants in Pharmaceuticals And Medical Devices,” Qual Assur, 1, 171-180, 1992.
18.Galer, D.L., Leung, H.L., Sussman, R.G., and Trzos, R.J., “Scientific And Practical Considerations For The Development
Of Occupational Exposure Limits (OELs) for Chemical Substances,” Regul Toxicol Pharmacol, 19, 317-337, 1992.
19.Naumann, B.D. and Weideman, P.A., Scientific basis for
uncertainty factors used to establish occupational exposure
limits for pharmaceutical active ingredients; Human Ecol. Risk
Assess, 1: 590-613, 1995.
20.Schwartz, C.S., “A Semiquantitative Method for Selection
of Safety Factors in Establishing OELs for Pharmaceutical
Compounds,” Human Ecol Risk Assess, 1:527-543, 1995.
21.Silverman, K.C., Naumann, B.D., Holder, D.J., Dixit, R., Faria,
E.C., Sargent, E.V., and Gallo, M., “Establishing Data-Derived
Adjustment Factors From Published Pharmaceutical Clinical
Trial Data,” Human Ecol Risk Assess, 5(5), 1059-1089, 1999.
22.Davidson, I.W.F., Parker, J.C., and Beliles, R.P., “Biological
Basis For Extrapolation Across Mammalian Species,” Regul
Toxicol Pharmacol, 6, 211-237, 1986.
23.Mordenti, J. and Chappell, W., “The Use Of Interspecies Scaling In Toxicokinetics,” Toxicokinetics and New Drug Development, pp. 42-96, New York, Pergamon Press, 1989.
24.Dourson, M.L., Felter, S.P., and Robinson, D., “Evaluation Of
Science-Based Uncertainty Factors In Noncancer Risk Assessment,” Regul Toxicol Pharmacol, 24, 108-120, 1996.
25.Weil, C.S. and McCollister, D.D., “Relationship Between
Short- And Long-Term Feeding Studies In Designing An Effective Toxicity Test,” Agr Food Chem, 11, 486-491, 1963.
26.McNamara, B.P., “Concepts In Health Evaluation Of Commercial And Industrial Chemicals,” New Concepts in Safety
Evaluation, pp. 61-140; Washington, DC, Hemisphere, 1976.
27. Lewis, S.C. and Nessel, C.S., “Extrapolating Subchronic Test
Results To Estimate Chronic No-Observed-Adverse-Effect
Levels: Factors Of 10 Are Larger Than Necessary,” The Toxicologist, 14(1), 401, 1994.
28.Weil, C.S., “Statistics vs. Safety Factors And Scientific Judgment In The Evaluation Of Safety For Man,” Toxicol Appl Pharmacol, 21, 454-463, 1972.
28
Journal
of
Validation T echnology [Summer 2011]
29.Lewis, S.C., “Reducing uncertainty with adjustment factors; Improvements in quantitative noncancer risk assessment (A symposium overview),” Fund Appl Toxicol, 20,
1-14, 1993.
30.Abdel-Rahman, M.S. and Kadry, A.M., Studies On The Use Of
Uncertainty Factors In Deriving RfDs; Human Ecol Risk Assess, 1:
614-624, 1995.
31.Lehman, A.J. and Fitzhugh, O.G., “100-Fold Margin of
Safety,” Assoc Food Drug Off. U.S.Q. Bull, 18, 33-35, 1954.
32.Dourson, M.L. and Stara, J.F., “Regulatory History And Experimental Support Of Uncertainty (Safety) Factors,” Regul
Toxicol Pharmacol, 3, 224-238, 1983.
33. Dolan DG, Naumann BD, Sargent EV, Maier A, Dourson,
M., “Application Of The Threshold Of Toxicological Concern
Concept To Pharmaceutical Manufacturing Operations,”
Regul Tox Pharm, 43:1-9, 2005. JVT
ACKNOWLEDGEMENTS
The authors would like to thank Joel Bercu, David Dolan,
and Arun Tholudur for their helpful suggestions.
ARTICLE ACRONYM LISTING
A
Product A (previously manufactured
product)
AB Product A to product B
ADE Acceptable Daily Exposure
API Active Pharmaceutical Ingredient
B Product B (subsequently manufactured
product)
BB,MIN Smallest Batch of Product B
BW Body Weight
CComposite
DB,MAX Largest (worst-case) Dose of Product B
FC Composite Uncertainty Factor
LOAEL Lowest Observed Adverse Effect Level
LOQ Limit of Quantitation
mAb Monoclonal Antibody
MACAB Maximum Acceptable Carryover of Product A
into a Manufacturing Batch of
Product B
MF Modifying Factor
NOAEL No Observed Adverse Effect Level
PK Pharmacokinetic Adjustment
PSA Product-Specific Assay
PSIA Product-Specific Immunoassay
SDS PAGE Sodium Dodecyl Sulfate Polyacrylamide Gel
Electrophoresis
SF Safety Factor
TDMIN,A Minimum Therapeutic Dose of Product A
TOC Total Organic Carbon
TTC Threshold of Toxicological Concern
iv thome.com