Using Health-Based Exposure

An Executive
Limits to Assess Risk in
Summary Cleaning Validation
Setting data-driven, scientific cleaning validation
limits reduces inconsistencies and risk with
pharmaceutical dosage form product carryover.

Ester Lovsin Barle, PhD, Background

MScTox, ERT In December 2016, the European Medicines Agency (EMA) released for public review a
Corporate Occupational
Toxicologist questions-and-answers document focused on the risk-based prevention of cross contamination
Lonza in pharmaceutical manufacturing and setting health-based exposure limits (HBELs) (1). EMA
made clear that HBELs should be established for all products. EMA also noted that limits for
cleaning purposes should continue to be based on risk assessments including the concept of
safety margins to help account for variability in cleaning processes and analytical methods (2).
Since HBELs were introduced in the European Union in 2015, inspections have shown that
manufacturers have been slow to adopt EMA guidelines and many do not have the expertise
within their organizations to create good-quality HBELs. A few companies that have developed
HBELs have not been using them other than to determine cleaning limits. This has revealed
limited use in cross-contamination and risk assessment.

The History of HBELs

The term HBEL was first introduced in EMEA guidelines in 2014 with the intention of establishing
health-based exposure limits when identifying cross-contamination risks while manufacturing
different medicinal products in shared facilities. Occupational exposure limits (OELs) are based
on the same data and the same scientific principles as the HBEL limits used in carryover limits.
In 2010, an International Society for Pharmaceutical Engineering (ISPE) guidance was pub-
lished describing the use of acceptable daily exposure (ADE) limits in cleaning validation and
risk assessment. The EMA followed with a regulatory document in 2015 that detailed permitted
daily exposure (PDE) limits.
HBELs, PDEs, and ADEs are defined as doses that are unlikely to pose an adverse effect if
an individual is exposed by any route of administration at or below the dose every day for their
entire lives. A lifetime for a patient is considered to be approximately 80 years; a working lifetime,
considered in the OEL, is approximately 40 years.

Reliability of Limits Used in Product Carryover

The risk of product carryover in pharmaceutical manufacturing came to light in 1988 with the recall
of cholestyramine USP. The API used to make this drug had become contaminated with low levels
of intermediates and degradants from the production of agricultural pesticides. The contamination

SPONSORED BY
 USING HEALTH-BASED EXPOSURE LIMITS TO ASSESS RISK IN CLEANING VALIDATION

was traced to inadequate cleaning and cleaning validation of • The routes of patient exposure that were not intended
drums that were previously used to recover solvents during originally.
pesticide manufacturing at another facility. In addition, HBEL monographs must contain an expert
Five years later, Gary Fourman and Michael Mullen of Eli assessment to identify:
Lilly suggested a combination of limits for product carryover, • Any critical effects on the patient and the worker. Critical
including 10 ppm as the maximum amount of drug that could effect is defined as the most sensitive adverse effect
appear in another product (3). This amount was considered to that is considered relevant to the target population.
be the threshold beyond which dose could potentially create • The assignment of adjustment factors. These values
an effect in individuals. provide a margin of safety between an unwanted effect
Another criteria for limiting product carryover is LD50, which and the HBEL.
is a dose that is used in animal studies via oral administration • Several calculations of HBEL, if data allows. However,
and is lethal in 50% of the population. This endpoint is unreliable the selected HBEL must be clearly explained and
for human toxicity and is considered less accurate in assessing justified.
patient risk. Thus, the use of LD50 for determining cleaning limit A good HBEL monograph can be lengthy and may take up to
is not appropriate. 40 hours of expert work. However, for the purposes of cleaning
Third is the widely used 1/1,000 minimum daily dose (MinDD) validation experts, a summary page should be designed to
criteria. Also proposed by Fourman and Mullen (3), this exposure quickly identify the parameters for setting cleaning validation
limit suggests that no more than 1/1000 of a drug’s therapeutic limits, as well as for industrial hygiene exposure assessments.
dose can be present in another product or as residual product An important aspect of HBELs is that all its parameters are
on a piece of equipment. looked at for each drug individually.
And last are HBELs, which, in contrast to the aforementioned
criteria, address who can be exposed and how they can be Regulatory Requirements
exposed. HBELs take into account all the relevant data that With regard to exposure limits and associated data, the
are important for determining a scientifically justified limit for the Rules Governing Medicinal Products in the European Union
target population, including sensitive patients such as children, chapter on good manufacturing practices (GMPs) state that
pregnant women, and the elderly. In addition, they address the manufacturing authorization holders must ensure that active
route by which the exposure may occur. substances are produced in accordance with GMPs (4). Audits
In summary, many criteria have been used in the past to should be carried out at the manufacturer and distributor of
determine a safe cleaning limit. Limits have been interpreted active substances to confirm that they comply with the relevant
individually, perhaps even subjectively. Due to difficulties and GMP practice. Finally, consideration should be given to potential
inconsistencies in the application of these approaches to the cross-contamination from other materials on site.
wide variety of pharmaceutical dosage forms across entire The owner of the data used for HBELs must clearly com-
industry, no single, consistent approach exists for establishing municate it to the manufacturer, whether it is internal or external.
limits for cleaning validation. The manufacturer should ensure that the quality of the incoming
data is sufficient to provide safe manufacturing of all products
The Beauty of HBELs produced in the shared facility for patients and workers alike.
So, how do we identify a reliable, scientific limit? HBELs are
an important part of the solution because there are consistent, Good Practices in the Use of HBELs
harmonized, and set by experts. HBELs are relevant for the OELs, as well as PDEs and ADEs, are mathematical represen-
manufacture of specific products or routes of administration, tations (as ranges) of potential biologic effects. This is similar
and they are adequate for risk assessment. They also eliminate to the number of calories that individuals are recommended
the need to use undefined terms such as “certain cytotoxic” or to eat in a given day, which is based on age, height, weight,
“certain hormone.” and activity level, but it is not an absolute number. OELs are
Creating a good HBEL and monograph is based on evalu- conservatively set for a daily lifetime exposure via inhalation.
ating all relevant available pharmacological and toxicological OELs are often the basis for banding in the pharmaceutical
data including both non-clinical and clinical data. Data collection industry. In a recent survey of pharmaceutical companies, most
inspection must identify: firms applied some sort of hazard or exposure banding with
• The chemical and its mechanism or mode of action; various numbers of bands and categories in their systems.
• Any health hazards found in the preclinical studies; These banding concepts are largely linked to historical pro-
• Any health hazards and effects found in clinical cesses within companies and are unlikely to be harmonized
studies, unless a drug is not yet in the clinic; across the industry. OELs—like PDs and ADEs—do not require
• The pharmacokinetic and pharmacodynamic additional banding by a toxicologist. Banding or defaulting
aspects of the drug; and OELs to a safe limit is appropriate for drugs in early-stage
development.
 USING HEALTH-BASED EXPOSURE LIMITS TO ASSESS RISK IN CLEANING VALIDATION

After an OEL is calculated, it can be directly compared to a potential risk for patients if 1/1000 MinDD criteria was used
industrial hygiene in monitoring and the risk for workers can be previously (5).
identified and mitigated. Many times, the attractiveness of easy PDEs can be less than 10 ug per day for several reasons.
banding and categorization concepts is so appealing to end For example, it may stem from low therapeutic dose, pharma-
users that they disregard the need for a proper risk assessment. cokinetic factors such as accumulation in the tissues, adverse
Banding based on hazards or OEL should never be a substitute effects in low doses, or a combination of those factors. PDE
for industrial hygiene expertise or define engineering solutions. calculation is, however, only the first step in assessing the
A high-quality HBEL can be compared to the previously risks for cross-contamination in a multi-purpose manufac-
used method for deriving maximum safe carry-over based turing facility (see Figure 2). Other criteria that are included
in 1/1000 minimum daily dose (MinDD). If the comparison in determining risk is batch size, maximum daily dose of the
shows that the PDE is higher than 1/1000 MinDD, cleaning next product, as well as criteria associated with cleaning that
is sufficient and no action is required. However, if the HBEL are not related to toxicology.
or PDE is lower than 1/1000 MinDD, a retrospective check of PDE of the previous product, batch size, and the maximum
previous cleaning is in order (see Figure 1). daily dose of the next product are used to calculate the max-
A recently published study of 140 reported drug substances imum safe carryover (MSC) from the previous product. The
from a typical pharma portfolio was compared to the minimum MSC value for each possible changeover in the facility is then
daily dose criteria with calculated PDE. Approximately 10% of calculated with the shared surface of the equipment between
the substances had PDE lower than 1/1000, which could be the previous and the next product, and plotted in a table

Permitted daily exposure (PDE) by the route/ 0.001 MinDD (R ratio)

Figure 1: Permitted daily exposure (PDE) by the route/0.001 MinDD (R ratio).

10000.00

1000.00

100.00
R ratio

10%
10.00 PDE > 1/1000 MinDD

1.00

0.10
Individual drug substances (140 DS) PDE < 1/1000 MinDD
0.01

Reasons
Figure for for
2: Reasons “low” permitted
“low” permitted dailydaily exposure
exposure (PDE)
(PDE) (under (under
10 ug/day). 10 ug/day)
10
9
8
PDE ug/day

7
6
5
4
3
2
1
0

Pharma & Biotech | Ester Lovsin Barle, PhD, MscTox, ERT | September 2017

Low therapeutic dose

Pharmacokinetics

Adverse effects in low doses

 USING HEALTH-BASED EXPOSURE LIMITS TO ASSESS RISK IN CLEANING VALIDATION

(see Figure 3). The lower table

Maximum safe carryover (MSC) for each
Figure 3: The MSC value for each possible changeover in the facility is calculated.
identifies products at potential possible changeover in the facility.
risk for contamination because Example of a shared facility with 11 PDE
Maximum
Batch size Equipment
API daily dose
their surface residue is below the APIs by using cleaning capability of
[µg/day]
[mg/day]
[kg] surface [m2]

cleaning capability in this facility, 20 mg/m2 (6) 1 7 40 580 100

2 65 200 690 125
which was determined to be PDEα [µg] × Batch sizeβ [kg]
3 200 2000 180 100
MSCα [g] =
20 mg/m2. Such risks must be Maximum Daily Doseβ [mg]
4 85 300 330 180

Margin of safety for health-based ex

mitigated by developing cleaning MSCα [g] × 1000
5 85 100 470 188
M S S R 6 30 300 1075 188
procedures, avoiding difficult
α
=
[mg/m2] Surface for β [m2] 7 10 1440 216 60
changeovers, and careful cam- Risks are identified when surface 8 330 100 310 78

(HBELs)
paign scheduling (6). residue is under the cleaning
9 15 750 207 220
10 1000 3000 335 100
capability of 20 mg/m2 11 5 450 120 100
Margin of Safety Following Product (β)
MSSR
After an HBEL is established, 1 2 3 4 5 6 7 8 9 10 11
1 195 6 43 175 133 18 278 9 8 19
should the10 ppm and 1/1,000 2 9850 60 410 1675 1276 170 2665 84 76 180

Preceding Product (α)

3 29400 5530 1222 5000 3809 510 7949 250 226 540
MinDD criteria still be applied? 4 12200 2295 75 2075 1581 211 3300 104 94 225
5 12200 2295 75 507 1581 211 3300 104 94 225
The margin of safety is the 6 4850 910 30 202 825 85 1310 41 35 90
7 1470 276 10 61 250 190 397 13 11 27
distance between the analytical 8 48941 9204 300 2035 8325 6341 847 418 376 897
9 2205 415 15 90 375 286 38 596 15 40
data and the HBEL base limit, 10 146970 27640 904 6110 25000 19043 2542 39744 1255 2695
which is indicated by a green 11 735 140 5 30 125 95 13 199 6 6

arrow on Figure 4. This margin Pharma & Biotech | Ester Lovsin Barle, PhD, MscTox, ERT | September 2017

of safety indicates the probability

Figure 4: Improved process capability moving from 0.001 dose to acceptable daily
of patient exposure to the API exposure (ADE) limits.
residues resulting from ineffec-
tive cleaning (7).
If there is a low margin of safety,
the process cleaning capability is
poor because of the risk that the
drug will contain a contaminant
above the HBEL may be high. As
seen in Figure 4, the true margin
of safety cannot be measured
when imposing the 10 ppm or
1/1000 MinDD criteria in addition
to the HBEL.

Summary
HBELs are scientifically justified
limits, and other criteria such as
10 ppm and 1/1000 MinDD are
not needed in addition to show Dec. 15, 2016, http://www.ema.europa.eu/docs/en_GB/
the level of the cleaning process capability. As a consequence, document_library/Other/2017/01/WC500219500.pdf
cleaning validation efforts should be focused on where the risks (2) Z. Brennan, “EMA Releases Q&A on Exposure Limits and
are high with the science behind the HBEL pointing companies (11)
Hazardous Compounds,” Jan. 9, 2017, http://raps.org/
Regulatory-Focus/News/2017/01/09/26523/EMA-Releases-
toward products that have the greatest risk.
QA-on-Exposure-Limits-and-Hazardous-Compounds
(3) G.L. Fourman and M.V. Mullen, Pharm. Technol. 17(4), 54–60
References (1993).
(1) European Medicines Agency, Questions and Answers (4) EU Guidelines for Good Manufacturing Practice for Medicinal
on Implementation of Risk Based Prevention of Cross Products for Human and Veterinary Use, https://ec.europa.eu/
Contamination in Production and “Guideline on Setting health//sites/health/files/files/eudralex/vol-4/chapter_5.pdf
Health Based Exposure Limits for Use in Risk Identification (5) E. Lovsin Barle et al., Pharm. Technol. 41(5), 42–53 (2017)
in the Manufacture of Different Medicinal Products in (6) M Crevoisier et al., Pharm. Technol. 40(8), 52–56 (2016).
Shared Facilities” (EMA/CHMP/CVMP/SWP/169430/2012), (7) A. Walsh et al., Pharm. Technol. 40(8), 45–55 (2016).
The content of this slide deck is accurate to the best of the presenter’s knowledge at the time of production. The views and opinions expressed in this presentation
are those of the author and do not necessarily reflect the official policy or position of Lonza or any of its officers.