Features

Setting Health-Based Exposure Limits to Support Pharmaceutical Development and Manufacturing

By B.D. Naumann, P.J. Nigro, U. Bruen, L. Parola, D. Cragin and A. Schatz, Merck & Co., Inc. and Ashland, Inc. | May 5, 2016

Health-based exposure limits are needed to support risk assessments

Setting health-based exposure limits is an important initial step in managing risks associated with the development and manufacture of active pharmaceutical ingredients (APIs) and drug products. Acceptable daily exposure (ADE) values are used within quality risk management programs to protect patients from product cross-contamination. The term ADE is generally synonymous with Permitted Daily Exposure (PDE) as used by the Food & Drug Administration (FDA), European Medicines Agency (EMA), the International Conference on Harmonization (ICH) and others. Occupational Exposure Bands (OEBs) and Occupational Exposure Limits (OELs) support industrial hygiene programs to protect employees handling the same APIs and associated process intermediates.
 
These values are established for small molecules and large molecules (i.e., biological materials) that are handled in laboratory, pilot plants, and manufacturing operations. The goal is to protect patients and employees from possible undesirable health effects from an unintended exposure. 

The Industrial Toxicology Advisory Committee (ITAC) is the formal committee at Merck, established in 1977, that sets health-based exposure limits for Merck compounds and, when appropriate, for non-Merck compounds. This committee includes toxicologists, occupational physicians, industrial hygienists, and engineers. This paper describes the procedures for establishing ADEs, OEBs and OELs at Merck. 

Overview of the limit setting process
Toxicologists in the Merck Global Safety and the Environment (GSE), Product Stewardship and Science group identify new development compounds by following internal communications of key Merck Research Laboratory committees responsible for approving new preclinical candidates (PCCs). When a PCC is approved for development, a GSE toxicologist will assign it to an OEB. Formal numerical ADE values and OELs are typically established later in development when sufficient repeat-dose human data and animal are available. Figure 1 shows when the assessments are completed during the typical drug development timeline.

At the time of PCC approval, GSE toxicologists perform an initial health hazard evaluation (typically found in the PCC Approval Document), and will establish a preliminary occupational exposure band (OEB). Merck compounds are assigned to one of several OEBs that reflect increasing levels of hazard potential. The goal is to set an OEB that is sufficiently protective, such that the later ADE is unlikely to be lower than the assigned OEB.

GSE toxicologists typically derive an ADE value and OEL at the decision to GO/TO Phase IIb, when initial multiple-dose data are available in human subjects; however, numerical ADE values and OELs or changes in OEB assignments may be recommended earlier in the drug development timeline if warranted by the available data. The criteria used to establish numerical OELs and control bands are summarized below under “Derivation of ADEs, OELs and OEBs.”

ADE values, OELs, OEBs and recommendations for medical surveillance and/or medical response established by ITAC represent formal Merck standards and expectations. ITAC decisions are documented and communicated through formal meeting minutes. The rationale and supporting data for these limits/bands is provided in the exposure limit monograph prepared for each compound. Additional special communications may also be made if the warranted by the specific recommendations.
These formal limits and bands are included in safety data sheets and can also be accessed in a corporate database. Established limits are reviewed as new information becomes available, and modified, as necessary, to properly reflect the available data.

Use of ADE values, OELs and OEBs
The ADE value is used by manufacturing sites to support cleaning validation. It is a key parameter in the standard equations used, in combination with batch sizes, shared surface area, maximum daily doses, etc., to demonstrate a sufficient margin of safety relative to the trace residues that may be left inside equipment, and available for carry-over to another product, following robust cleaning and current process capability. ADEs are also used by sites in risk assessments to determine if existing processes have sufficient engineering and administrative controls to ensure patient safety from cross-contamination from other processes. 

Numerical OELs are used to assist in facilities design and evaluation of control strategies for worker protection. OEB assignments are also used by engineers and safety and industrial hygiene professionals to design facilities and to develop safe handling practices. An engineering design standard for new facilities or major renovations provides details on specifications recommended to meet the facility control requirements defined by each OEB. A separate matrix with handling guidance based on OEBs is available for use in laboratories. OEBs can also be used to estimate ADE values, when needed, early in development.

Methods for assigning OEBs and deriving ADE values and OELs
The first step GSE toxicologists take when assigning an OEB for a material or establishing/updating an OEL is to gather all available information for the compound and related materials. This information is obtained by reviewing internal studies and conducting an extensive literature search on potential adverse effects, including pharmacological effects, and the dosages required to produce these effects. Information may come from internal sources or external sources (e.g., outsourcing or licensing partners). For new compounds, the key documents that summarize all of the available information are the PCC background approval package and, later, the Investigator’s Brochure. For existing marketed compounds, the current version of the product labeling published in the Physician’s Desk Reference is consulted. Other sources, such as standard texts, on-line proprietary databases are also consulted and supplemented with an online literature search on TOXNET/PubMed. The depth of the literature search is up to the judgment of the reviewing toxicologist in an effort to identify the information of greatest importance for establishing a safe exposure limit, including: the intended indication for the drug; the mechanism of action; the clinical dose(s); results of preclinical, clinical, and epidemiological studies; the probable route(s) of entry; occupational experience; and identification of any susceptible subpopulations.

Assigning OEBs
OEBs, which were previously termed performance-based exposure control limits (PB-ECLs), represent a control banding concept and classification system that utilizes a set of criteria applied to chemical compounds and biological materials to rate them according to their inherent hazardous properties and characterized them with respect to the severity of each (Naumann et al., 1996). In general–for both OELs and OEBs–the evaluation will often focus on one or more primary health concern(s) or lead endpoint(s) potentially associated with acute or chronic exposure, as well as on the reversibility and severity of these effects. For most materials, the intended pharmacologic effect is also the effect that occurs at the lowest dose. When the ADE is set to protect against this effect, it also protects against other adverse effects that occur at higher doses (visit the online version of this article for a complete OEB assignment criteria chart).

Types of OEBs
OEBs are typically established by GSE toxicologists early in the drug development process when a compound is approved for development. These initial assignments are intended primarily for use in laboratories and pilot plants. As a default, these relatively unstudied compounds, including biological materials, are automatically assigned to OEB 3 unless there are mechanistic data or data from compounds with a similar pharmacology that would support a higher or lower assignment. This also applies to compounds in drug discovery.

In special circumstances compounds in an entire therapeutic class may be assigned to a different default OEB (typically OEB 4 or OEB 5 depending on the underlying data). An example is the default OEB 4 assignment used for oncology drugs, which corresponds to airborne concentrations in the range of 1-10 μg/m3. OEB 4 is typically used for compounds with novel mechanisms involving cell-cycling pathways that are still not well understood. Exceptions can be considered when the mechanism of action clearly supports use of the standard default (OEB 3) (e.g., compounds that modulate growth hormone receptors with no concerns for genetic damage and monoclonal antibodies). Oncology compounds will be classified as OEB 5 materials if available data show, or strongly suggest, the possibility of genetic damage or serious target organ effects at low dosages. Reclassification from OEB 4 to OEB 3 or OEB 5 to OEB 4 will only occur when there is convincing evidence, most likely from results of clinical trials, that initial concerns have not been confirmed. Supportive data may come from mechanistic or animal toxicology studies.

Establishment of ADE values and OELs
The methods used by ITAC to establish OELs at Merck have been published in the technical literature (Sargent and Kirk, 1988; Naumann and Weideman, 1995; Naumann and Sargent, 1997). The same basic limit setting methods are used to derive ADE values, with minor differences. These methods for setting health-based exposure limits are based on sound science and are consistent, and essentially the same as, described in several more recent publications and guidance documents (ISPE 2010; EMA 2014; Sargent, et al. 2013; Dankovic et al. 2015) and a number of papers documenting the deliberations of an ADE Workshop held in 2014 (Weideman et al. 2015).

In general, based upon an evaluation of the available, pertinent information, ADE values and OELs are derived by identifying a no-effect level for the most sensitive clinically relevant endpoint, applying an appropriate composite adjustment factor, with possible further adjusts for differences in bioavailability, and calculating an ADE value. If the critical effect and susceptible subpopulations are similar, the OEL is calculated by dividing ADE (mg/day) by the volume of air a worker breathes during an eight-hour workday (10 m3) when engaged in light work. Wipe limits are typically calculated for OEB 3, 4 and 5 materials and compounds that have the potential to produce systemic effects following dermal exposure by dividing the mass represented by the ADE value by 100 cm2, the approximate surface area of the palm of one hand (EPA 2011). Sensitizer notations (e.g., DSEN and RSEN), skin notations and/or eye notations are applied, when supported by data, to alert employees of the potential for skin or respiratory sensitization or potential toxicity following dermal or ocular contact, respectively.

Data collection and analysis
Hazard evaluation requires a formal review of all available data for each compound. For the API, the data used in the analysis shall at a minimum include data from the sources listed at the beginning of this section. Consequently, the reviewing toxicologist considers the full range of preclinical and clinical data required for approval of the drug and from post-marketing clinical experience that have been published. Once all of the information is compiled the toxicological assessment including identifying the critical effect and dose-response assessment can begin.

Identification of the critical effect
The initial hazard evaluation is intended to identify all possible health hazards associated with a compound and to assess their level of severity. When combined with information on the dose-response relationship, as described below, the critical effect can be defined. The critical effect is typically the first adverse effect that is observed with increasing doses. By using this effect as the basis for the derivation of an ADE, it is protective of all other effects, as these will only occur at higher doses. The toxicologist uses professional judgment to decide which of the effects reported at the low end of the dose-response curve is the critical effect. It is important to recognize the distinction between biological effects and clinically significant effects. Biological effects may include minor physiological or adaptive changes, which would be considered subclinical pharmacological effects of the drug. These effects, while also effects that should be avoided if possible, are less serious than clinically significant pharmacologic effects, which would clearly be considered “adverse” effects for anyone.

Dose-response assessment
During the toxicological assessment, all key health effects associated with the compound are reviewed and documented in the exposure limit monograph. For each effect, as the dose increases, the incidence and severity of the effect increases. For systemic toxicity endpoints, there is a threshold for toxicity, which is the dose below which no effects are expected. For some endpoints, a threshold cannot be identified. An example includes carcinogenicity that results from direct damage to DNA, where it is often assumed that some effects (e.g., mutations) could be expected at low doses. The mechanism-of-action determines the approach used to set safe exposure limits. For compounds with thresholds, a “safety factor” approach can be used. For compounds without a threshold, it is necessary to determine the “acceptable” level of response for the effect (e.g., 1/100,000 excess cancer risk).

Identification of the Point-of-departure
Most compounds have toxicity endpoints where there is a clear dose-response relationship and threshold for the critical effect. For derivation of the ADE value, the objective is to define the no-observed-adverse-effect level (NOAEL) for the clinically significant critical effect. This is the dose used as the point-of-departure (PoD) to derive the ADE value.  Alternatively, in the absence of a NOAEL, it may be necessary to use the dose at which a significant adverse effect is first observed, or the lowest-observed-adverse-effect level (LOAEL).

When human data are available, they are preferentially used. The exception is if animal studies suggest a potential systemic or reproductive risk. In this case, an ADE is developed for both the critical effect in humans and the adverse effect(s) in laboratory animals. Whichever value is lower then becomes the formal ADE. Generally, the critical effect in human results in an ADE lower than that calculated from animal repeat-dose and reproductive toxicity studies, but this in particular is an important exercise to ensure the safety of potentially pregnant workers or patients.

For some toxicology data sets only a LOAEL is available for the critical effect. Rather than using this as the PoD, a benchmark dose (BMD) can be used, which is a value that is derived mathematically by fitting a curve to the dose-response data for the critical effect, and is equivalent to a NOAEL. The BMD software is available on the US EPA web site at www.epa.gov/ncea/bmds

Application of adjustment factors
In order to derive a health-based exposure limit such as ADE values, the PoD is adjusted lower using various adjustment factors to extrapolate to an anticipated no-effect level in the most sensitive sub-population. Depending on the target subpopulations, different adjustment factors may be used. Adjustment factors, also called uncertainty or safety factors, have been defined for each of the main sources of uncertainty and variability as described below. The product of these factors forms a composite adjustment factor (AFC) that is used in the denominator of the general equation used to set health-based limits (Figure 1).  The ADE value is used to derive swab or rinse limits for cleaning validation purposes and is intended to protect sensitive subpopulations. OELs are established to protect workers. Additional adjustments may be necessary in either case to address differences in bioavailability when extrapolating between different routes of exposure (see below).

Calculation of the Acceptable Daily Exposure (ADE) Value, OEL and Wipe Limit

ADE (mg/day) =      PoD x BW
                          AFc x MF x a x S

OEL (mg/m3) =       PoD x BW
                       AFc x MF x a x S x V

Wipe Limit (mg/100 cm2) =    OEL(mg/m3) x 10m3
                                                          100 cm2

where:
ADE = Acceptable daily exposure (mg/day)
OEL = Occupational exposure limit (mg/m3)
PoD = Point-of-departure (NOAEL or LOAEL, mg/kg/day)
BW = Body weight (50 kg)
AFC = Composite adjustment factor
MF = Modifying factor
a = Bioavailability correction factor
S = Steady state factor
V = Volume of air breathed in 8 hours (10 m3)
Wipe Limit = Surface limit (mg/100 cm2)

The safe level of exposure for patients may be different than that established for workers. The critical effect used to derive the limits may be different due to differences in the anticipated route of exposure and, even if the same, the choice of adjustment factors may differ. Workers are generally healthy individuals, whereas patients may be children or elderly and, by definition, have at least one active medical condition. Adjustments may also be necessary to address route-to-route extrapolation, even if the same critical effect is used. For example, an OEL derived from the ADE may need to be adjusted for differences in bioavailability between the route used to identify the critical effect (e.g., oral) and the anticipated route of exposure (i.e., inhalation).

If the same assumptions apply regarding the critical effect and susceptible subpopulations, an occupational exposure limit (OEL) can be calculated directly from the ADE by dividing the dose allowed by the volume of air breathed by a worker, typically 10 m3, engaged in light work for 8 hours. Likewise, wipe limits can also be derived by dividing the dose corresponding to the OEL by a standard surface area (e.g., 100 cm2).

Table 1 summarizes the various adjustment factors used to derive health-based exposure limits recommended for each source of uncertainty and variability. Detailed information on the scientific basis for the use of these adjustment factors is available for review (Naumann and Weideman (1995); Dankovic et al. 2015; Sussman et al. 2016). Alternative factors may be used if supported by available data for a specific compound and a sound scientific rationale. Toxicological expertise is required to determine the appropriate adjustment factor(s) to apply. The rationale for the choice of adjustment factors is documented in the compound monograph.  Reference is made to the corresponding adjustment factors described in the EMA guide on setting health-based exposure limits (EMA 2014).

Use of CSAFs
Default adjustment factors, by definition, should only be used in the absence of relevant data. During new product development, a significant amount of data on pharmacokinetics (PK) and pharmacodynamics (PD) is generated that can be used to define interindividual and interspecies differences in kinetics and dynamics (Silverman et al. 1999; Naumann et al., 2001; Sargent et al., 2002). Chemical-specific adjustment factors (CSAFs) are intended to replace the default factors where PK and PD data are available. Guidelines are available for proper application of CSAFs in risk assessment (IPCS 2005; Meek et al., 2001).  Figure 3 illustrates how PK data can be used to calculate a CSAF for known susceptible subpopulations (e.g., bimodal distribution).

Derivation of ADEs for compounds with limited data
Dolan et al. (2005) published a paper with recommendations on how the “Threshold of Toxicological Concern” (TTC) concept could be applied to relatively unstudied compounds encountered in the pharmaceutical industry. The TTC concept is used to recommend ADE values for three categories of compounds with limited or no toxicity data: 1) compounds that are likely to be carcinogenic, 2) compounds that are likely to be potent or toxic, and 3) compounds that are not likely to be potent, toxic or carcinogenic.  The corresponding ADEs recommended for these three categories are 1, 10 and 100 µg/day, respectively. Subsequent to the publishing of Dolan et al. (2005), the ICH M7 guidance on “Assessment and Control of DNA Reactive (Mutagenic) Impurities in Pharmaceuticals to Limit Potential Carcinogenic Risk” (ICH, 2014) was issued and it offers appropriate mutagenic impurity limits based on the duration of treatment (e.g., 1.5 µg/day for a chronic use scenario).

Setting ADEs for animal health
The methods described above are also appropriate for setting ADEs for use in Animal Health. For many animal health APIs, there are no human data so the ADE must be based on available animal data. In this case, the choice of the critical endpoint, point-to-departure and adjustment factors follows the methodology described above in this guideline. Alternatively, human health ADEs can be extrapolated for use with animal health medicines based using the following approach:

Human health ADEs are derived using a 50 kg bodyweight assumption. ADEs for animal health can be derived using this value and a default 1 kg animal body weight assumption, as recommended by EMA guideline on setting health-based exposure limits (EMA, 2014). As an example, the human health ADE for a veterinary drug is 20 mg/day. Expressed on a mg/kg basis, this is equivalent to 0.4 mg/kg/day for a 50 kg adult. For a 1 kg target species, the animal health ADE would be 0.4 mg/day.  If a higher body weight is appropriate to use based on the site-specific product mix and target animals, the alternate ADE, recommended on a body weight basis using 0.4 mg/kg/day, can be used to derive appropriate cleaning limits. If this product-specific ADE is calculated, the site must document the basis for this adjustment and provide the appropriate body weight.

Use of human health ADEs is appropriate if there are no indications that a specific relevant animal species shows a special sensitivity. If there are special species sensitivities, the ADE will be derived appropriately.

Assessment of trace contaminants in veterinary medicines given to food producing animals must take into account both target animal safety and consumer safety, i.e., ADEs for both animal health and human health should be applied.

Medical surveillance and medical response
A medical surveillance program may be recommended for compounds with unique risks to workers. Medical surveillance is the periodic medical evaluation for early detection of adverse effect(s) from workplace exposures that are below an OEL or the upper boundary of an OEB. This may include workplace medical surveillance questionnaires, medical testing/examination or any other appropriate medical screening protocol.

A compound-specific medical response may be recommended in the event of an over-exposure in workers. This may include medical assessments and actions taken in order to evaluate, monitor, mitigate and/or treat anticipated health effects of exposure above an OEL or OEB. It also includes development of communications for use at the affected site and for affected employees.

Significant figures, rounding and use of units
When setting ADE values and OELs it is important not to overstate the precision associated with the recommended values. Considering the significant figures in the dose used as the point of departure (e.g., NOAEL) and the various uncertainty and adjustment factors used, the resulting ADE value or OEL rarely exceeds one significant figure.

Significant figures
The number of significant figures in the ADE value and OEL should not exceed the least number of significant figures in the various adjustment factors used to derive these values. Typically, the ADE value and OEL will be expressed as one significant figure.

Rounding
When a number is obtained by calculation, its accuracy depends on the accuracy of the numbers used in the calculation. When calculating exposure limits, an extra significant figure is retained, and rounding should only occur at the end of the calculation for the final answer and not within the calculation (Moore et al., 2008).

Use of units
Presentation of ADE and OELs should also conform to certain conventions that are commonly used. As a rule, the choice of unit should minimize the number of zeros in the recommended value. For example, an OEL should be expressed as 5 mg/m3 rather than 5,000 ug/m3.

Documentation and other considerations
The rationale for the establishment of health-based limits should be documented very clearly in a written document that can be used by internal and external stakeholders, including regulators. The presence of this document demonstrates that the product owner has completed an appropriate hazard assessment and provides a scientific rationale for the recommended health-based limit to ensure patient and worker protection. In addition, it informs and facilitates communication between different operational groups, e.g., engineering and manufacturing groups charged with implementing quality risk management program elements, e.g., cleaning validation. It also facilitates communication with external partners and regulators. The reader should be able to readily determine: 1) the health endpoint (critical effect) on which the ADE value was established, 2) the values chosen for adjustment factors, and which sources of variability and uncertainty they address, and 3) any further adjustments used (e.g., a bioavailability correction factor).

There are several other considerations when evaluating an internal program for establishing health-based exposure limits for patient and worker safety. It is important to establish and maintain a strong underlying governance/peer review process, including upper management support, with an attendant formality assigned to the limit setting process. The technical qualifications of authors/reviewers must be sound and consistent with the importance of the task. There is also a need to establish a robust process to review new information to determine if updates to the limits are needed, along with formal periodic reviews on an appropriate revision cycle say every 5-years. 

Special thanks are given to the following individuals that contributed to the development of the exposure limit setting methods described in this paper: Drs. Dave Dolan, Ellen Faria, Ed Sargent, Keith Silverman, Jeffrey (Fei) Wang and Trish Weideman.

References:
For a full list of references please visit the online version of this story at ContractPharma.com.

Related Contract Manufacturing:

Related Compliance: