Case Studies of Categorical Data-Derived Adjustment Factors

Investigations were performed on representative compounds from five different therapeutic classes to evaluate the use of categorical data-derived adjustment factors to account for interindividual variability. The five classes included antidepressants, angiotensin converting enzyme (ACE) inhibitors, nonsteroidal anti-inflammatory drugs (NSAIDS), cholesterol lowering agents, and antibiotics. Each of the case studies summarized the mode of action of the class responsible for both the therapeutic and adverse effects and the key pharmacodynamic (PD) and pharmacokinetic (PK) parameters that determine the likelihood of these responses for individual compounds in the class. For each class, an attempt was made to identify the key factors that determine interindividual variability and whether there was a common basis to establish a categorical default adjustment factor that could be applied across the class (or at least across specific subclasses within the class). Linking the PK and PD parameters to the critical endpoint used to establish a safe level of exposure was an important underlying theme throughout the investigations. Despite the wealth of PK and PD information in the published literature on the surrogate compounds representing these classes, it was difficult to derive a categorical adjustment factor that could be applied broadly within each class. The amount of information available may have hindered rather than helped the evaluations. Derivation of categorical defaults for different classes of “common” chemicals may be more straightforward if sufficient data are available. In a few cases (e.g., tricyclic antibiotics, ACE inhibitors and selected antiinflammatory agents) categorical defaults could be proposed, although it is unclear whether the reduction in uncertainty resulting from their application would be offset by the additional uncertainties that may have resulted from their application. Residual uncertainties may remain depending on the level of confidence in the underlying assumptions used to support the categorical defaults. Regardless of the conclusions on the utility of categorical defaults, these investigations provided further support for the use of data-derived adjustment factors on a compound-specific basis.     

Establishing Data-Derived Adjustment Factors from Published Pharmaceutical Clinical Trial Data

In non-cancer risk assessment the goal traditionally has been to protect the majority of people by setting limits that account for interindividual variability in the human population. The Environmental Protection Agency (EPA) has assigned a default uncertainty factor (?UF) of 10 to account for interindividual variability in response to toxic agents in the general population. Previous studies have suggested that it is appropriate to equally divide this factor into sub-factors of 3.2 each for variability in human pharmacokinetics (PK) and pharmacodynamics (PD). As an extension of this model, one can envision using scientific data from the literature to modify the default sub-factors with compound-specific adjustment factors (AFs) and to create new and more scientifically based defaults. In this paper, data from published clinical trials on six pharmaceutical compounds were used to further illustrate how to calculate and interpret data-derived AFs. The clinical trial data were analyzed for content and the reported mean and standard deviation values for two key PK parameters, area under the curve of blood concentration by time (AUC) and peak plasma concentration (C_max), were evaluated. The mean PK values for each study were subsequently analyzed for variability within the population (unimodal distributions) and for the presence of potentially susceptible sub-populations (bimodal distributions). A method based on the proportion of the population covered was applied and data-derived AFs were calculated for these six compounds. Our results showed that, of the 15 possible data-derived AFs calculated using unimodal and bimodal distributions, only three exceeded a value of 3.2. This study further illustrates the value of calculating data-derived values when sufficient PK data are available.     

Rationale for the Chemical-Specific Adjustment Factors Used to Derive an Occupational Exposure Limit for Timolol Maleate

Timolol maleate is a non-selective beta-adrenergic blocking agent currently used primarily to reduce intraocular pressure in the treatment of glaucoma. It also produces effects on the heart and bronchial smooth muscle and all of these effects are of potential concern in workers handling this active pharmaceutical ingredient. The disposition of timolol maleate is influenced by a polymorphism in oxidative metabolism by CYP2D6 and two distinct phenotypes have been identified (i.e., poor and extensive metabolizers). These properties of timolol maleate provided an opportunity to use the compound as a case study to demonstrate the derivation of chemical-specific adjustment factors for pharmacokinetics and pharmacodynamics to replace the default uncertainty factor for interindividual variability. Overall, the available data on the pharmacodynamic endpoints showed very little variability and most pharmacokinetic studies failed to discern significant differences in relatively small groups of healthy volunteers or patients. Reports of bradycardia and bronchoconstriction in patients receiving therapeutic doses are relatively rare. In one study, there was a significant reduction in heart rate 24 hours post-dose that was associated with elevated area under the curve (AUC) values. A chemical-specific adjustment factor (CSAF) for kinetics of 9.8 based on these AUC data was combined with a CSAF for dynamics of 1.2 and applied to the extrapolated no-effect level for clinically significant cardiovascular effects (with correction for oral bioavailability) to establish an occupational exposure limit (OEL) for timolol maleate which is expected to be protective of workers that may be poor metabolizers or asthmatics.     

Application of Uncertainty Factors in the Priority Substances Program and International Harmonization

The Canadian Environmental Protection Act (CEPA) authorizes the Ministers of the Environment and of Health in Canada to investigate a wide variety of substances that may contaminate the environment and cause adverse effects on the environment and/or on human health. Under the Act, assessments have been completed for 44 environmental contaminants on the first Priority Substances List (PSL) and are relatively advanced for 25 compounds on the second PSL. The principles developed for the application of uncertainty factors in assessment of risks to human health for Priority Substances under CEPA are outlined, with emphasis on those aspects which are somewhat unique and/or evolving. The interface of developments in the Priority Substances program with an initiative of the International Programme on Chemical Safety in this area to effect greater harmonization of approaches is also described.     

Accounting for Missing Data in Noncancer Risk Assessment

Noncancer risk assessments are generally forced to rely on animal bioassay data to estimate a Tolerable Daily Intake or Reference Dose, as a proxy for the threshold of human response. In cases where animal bioassays are missing from a complete data base, the critical NOAEL (no-observed-adverse-effect level) needs to be adjusted to account for the impact of the missing bioassay(s). This paper presents two approaches for making such adjustments. One is based on regression analysis and seeks to provide a point estimate of the adjustment needed. The other relies on non-parametric analysis and is intended to provide a distributional estimate of the needed adjustment. The adjustment needed is dependent on the definition of a complete data base, the number of bioassays missing, the specific bioassays which are missing, and the method used for interspecies scaling. The results from either approach can be used in conjunction with current practices for computing the TDI or RfD, or as an element of distributional approaches for estimating the human population threshold.     

Some Comments on the Selection of Human Intraspecies Uncertainty Factors

The Reference Dose (RfD) is used in the risk assessment of non-carcinogenic chemicals. It is derived by dividing a point of departure by the product of the uncertainty (UFs) and modifying factors (MFs). Separate UFs are used for different variables, e.g., intraspecies variation and, in general, each UF is an order of magnitude (10-fold). On the other hand, the MF is usually based on some known variable such as differences in absorption of a chemical from food and water and its default value is one. The USEPA's Integrated Risk Information System (IRIS) has 14 chemicals that have RfDs based on human studies. We examined those IRIS files to determine the rationale for setting human intraspecies uncertainty factors (UF_H). The first consideration was that the chemical had an adequate peer-reviewed human database. Without such, it would not be possible to derive an RfD based on human data. Ten of the 14 chemicals had an UF_H of 1 or 3; four of these were essential trace elements (ETEs). The rationales for using less than a 10-fold UF_H for the ETEs included; 1) nutritional data, 2) large human exposure groups, 3) minimal effect levels and/or 4) several studies with similar effect levels. For the other compounds, reasons included; 1) large human exposure groups, 2) a critical effect that was not adverse (cosmetic), 3) the most sensitive population was exposed, 4) the compound was on the FDA's “generally regarded as safe” (GRAS) list, 5) database uncertainties and 6) less-than-lifetime exposure adjusted for 70 years exposure. It is important to understand the reasons for selecting a UF_H of 1, or 3 as they will apply to future chemicals considered by the USEPA and other agencies.     

Differential Sensitivity of Children and Adults to Chemical Toxicity

Children, particularly neonates, can be biologically more sensitive to the same toxicant on a body weight basis than adults. Current understanding of the rates of maturation of metabolism and evidence from case studies indicate that human infants up to 6 months of age typically lack the capacity to detoxify and eliminate substances as readily as adults. For most chemicals, the infant physiologic systems usually produce higher blood levels for longer periods. The newborn's metabolic capacity rapidly matures and, by 6 months of age, children are usually not more sensitive than adults based on their pharmacokinetic competence. Whether children are at greater risk from chemical exposures is another question. Drawing conclusions about the ability of the U.S. Environmental Protection Agency's intraspecies (UF_H) and database (UF_D) uncertainty factors to protect children on the basis of the modest data available is challenging. However, virtually all studies available suggest that a high percentage of the population, including children, is protected by using a 10-fold UF_H or by using a 3.16-fold factor each for toxicokinetic and toxicodynamic variability. Based on specific comparisons for newborns, infants, children, adults and those with severe disease, the population protected is between 60% and 100%, with the studies in larger populations that include sensitive individuals suggesting that the value is closer to 100%. UF_D is likewise protective when used with databases that are missing substantive studies.     

Categorical Default Uncertainty Factors-Interspecies Variation and Adequacy of Database

The potential application of categorical (i.e., species, pathway, or group specific) defaults for several components of uncertainty relevant to development of tolerable or reference concentrations/doses is considered-namely, interspecies variation and adequacy of database. For the former, the adequacy of allometric scaling by body surface area as a species-specific default for oral tolerable or reference doses is considered. For the latter, the extent to which data from analyses of subchronic:chronic effect levels, LOAELs/NOAELs, and critical effect levels for complete versus incomplete datasets informs selection of defaults is examined. The relative role of categorical defaults for these aspects is considered in the context of the continuum of increasingly data-informed approaches to characterization of uncertainty and variability that range from default (“presumed protective”) to “biologically based predictive”.     

Pathway-Related Factors: The Potential for Human Data to Improve the Scientific Basis of Risk Assessment

The default uncertainty factors used for risk assessment are applied either to allow for different aspects of extrapolation of the dose-response curve or to allow for database deficiencies. Replacement of toxicokinetic or toxicodynamics defaults by chemical-specific data allows the calculation of a chemical-specific “data-derived factor”, which is the product of chemical-specific values and default uncertainty factors. Such chemical-specific composite values will improve the scientific basis of the risk assessment of that chemical, but the necessary chemical-specific data are rarely available. Categorical defaults related to pathways of elimination and mechanisms of toxicity could be used when the overall fate or mechanism is known, but there are no chemical-specific data sufficient to allow replacement of the default, and the development of an overall data-derived factor. The development of pathway-related categorical defaults is being undertaken using data on selected probe substrates for which adequate data are available. The concept and difficulties of this approach are illustrated using data for CYP1A2.     

Noncancer Risk Assessment: A Probabilistic Alternative to Current Practice

Based on imperfect data and theory, agencies such as the United States Environmental Protection Agency (USEPA) currently derive “reference doses” (RfDs) to guide risk managers charged with ensuring that human exposures to chemicals are below population thresholds. The RfD for a chemical is typically reported as a single number, even though it is widely acknowledged that there are significant uncertainties inherent in the derivation of this number.

In this article, the authors propose a probabilistic alternative to the EPA's method that expresses the human population threshold as a probability distribution of values (rather than a single RfD value), taking into account the major sources of scientific uncertainty in such estimates. The approach is illustrated using much of the same data that USEPA uses to justify their current RfD procedure.

Like the EPA's approach, our approach recognizes the four key extrapolations that are necessary to define the human population threshold based on animal data: animal to human, human heterogeneity, LOAEL to NOAEL, and subchronic to chronic. Rather than using available data to define point estimates of “uncertainty factors” for these extrapolations, the proposed approach uses available data to define a probability distribution of adjustment factors. These initial characterizations of uncertainty can then be refined when more robust or specific data become available for a particular chemical or class of chemicals.

Quantitative characterization of uncertainty in noncancer risk assessment will be useful to risk managers who face complex trade-offs between control costs and protection of public health. The new approach can help decision-makers understand how much extra control cost must be expended to achieve a specified increase in confidence that the human population threshold is not being exceeded.     

Harmonizing the Incorporation of Uncertainty and Variability Into Risk Assessment: An International Approach

International harmonization of risk assessment approaches affords a number of opportunities and advantages. Overall, harmonization will lead to more efficient use of resources, but also will lead to better understanding amongst scientists and regulators worldwide. It is with these goals in mind that in 1994 the International Programme on Chemical Safety (IPCS) initiated its Project on the Harmonization of Approaches to the Assessment of Risk from Exposure to Chemicals (Harmonization Project). An ongoing activity under this project addresses uncertainty and variability in risk assessment. The goal of the overall activity is to promote harmonization of risk assessment methodologies for noncancer endpoints. However, given the common links in uncertainty and variability that apply across a range of end-point-specific activities, these links are identified wherever possible. This paper provides an overview of the IPCS Harmonization Project and reviews the activity and future plans related to uncertainty and variability.     

Effect of Variability in Pharmacokinetic Parameters on the Derivation of Exposure Limits for Pharmaceutical Compounds

The purpose of this study was to investigate how variability in pharmacokinetic parameters influences the determination of occupational exposure limits (OEL) for pharmaceutical compounds in potentially susceptible subpopulations. A compartmental pharmacokinetic model for quinidine was applied to derive OELs based on target blood concentrations in humans but relied on pharmacokinetic parameters in subjects with cirrhosis rather than normal subjects. Quinidine was used as the sample compound as this was used in the development of the methodology. The intent was not to set an OEL for quinidine for a particular population but rather to use the methodology to investigate how factors, which may influence susceptibility, could be incorporated into the analysis. The model was used to simulate exposure concentrations that would result in levels below those that cause undesirable pharmacological effects taking into account variability in parameters through incorporation of Monte Carlo sampling. Results indicate that cirrhotic patients did not require additional protection from occupational exposure to quinidine. These results cannot be extrapolated to other compounds, as the effects of variability in pharmacokinetics on systemic exposure are compound specific. However, this methodology does provide a framework for addressing issues related to the contribution of pharmacokinetics to susceptibility from occupational exposure to pharmaceutical compounds.     

Evaluating the 10X Uncertainty Factor for Children and Elderly with Data-Derived Values for Neuromuscular Blocking Agents

Conventional risk assessment process utilizes a 10-fold uncertainty factor (UF) to extrapolate from the general human population to sensitive subgroups, such as children and elderly. The purpose of this investigation was to evaluate whether the magnitude of the 10X-UF can be reduced when pharmacokinetic and pharmacody-namic data are incorporated to characterize human sensitivity. An extensive literature search was conducted on seven neuromuscular blocking agents (mivacurium, atracurium, rocuronium, vecuronium, doxacurium, pancuronium, pipecuronium). Composite factors (kinetics × dynamics) were calculated using the highest data-derived kinetic and dynamic values. For the drugs examined, all of the composite factors for the sensitivity of children were lower than 5. In the elderly, all of the composite factors were lower than 10, and five of seven composite factors were less than 5. From this study, it was concluded that relevant compound-specific kinetic and dynamic data can reduce the uncertainties associated with sensitive subgroups.     

Risk Analysis,Uncertainty Factors,and the Susceptibilities of Children

This article examines the use of uncertainty factors (UFs) by the Agency for Toxic Substances and Disease Registry (ATSDR) when developing health guidance values known as minimal risk levels (MRLs) in environmental risk analysis as it applies to children. Improvements in the chemical-specific databases often reveal new information and thereby reduce uncertainty or alternatively raise new concerns. As a result, MRLs can and will change. Children, in particular, are not “small adults” and in some instances demonstrate greater risks of exposure to environmental toxicants and greater susceptibility for adverse health effects once exposed. Recent experience with risk analysis for three toxicants (organic mercury, dioxin, and manganese) is recounted to demonstrate how ATSDR has revised MRLs as the emerging science generates greater knowledge and awareness of children's special vulnerabilities to toxic substances in the environment.     

The Precautionary Principle and Statistical Approaches to Uncertainty

The central challenge from the Precautionary Principle to statistical methodology is to help delineate (preferably quantitatively) the possibility that some exposure is hazardous, even in cases where this is not established beyond reasonable doubt. The classical approach to hypothesis testing is unhelpful, because lack of significance can be due either to uninformative data or to genuine lack of effect (the Type II error problem). Its inversion, bioequivalence testing, might sometimes be a model for the Precautionary Principle in its ability to ‘prove the null hypothesis.’ Current procedures for setting safe exposure levels are essentially derived from these classical statistical ideas, and we outline how uncertainties in the exposure and response measurements affect the No Observed Adverse Effect Level (NOAEL), the Benchmark approach and the “Hockey Stick” model. A particular problem concerns model uncertainty: usually these procedures assume that the class of models describing dose/response is known with certainty; this assumption is however often violated, perhaps particularly often when epidemiological data form the source of the risk assessment, and regulatory authorities have occasionally resorted to some average based on competing models. The recent methodology of Bayesian model averaging might be a systematic version of this, but is this an arena for the Precautionary Principle to come into play?     

Relevance of the 10X Uncertainty Factor to the Risk Assessment of Drugs Used by Children and Geriatrics

Conventional risk assessment practices utilize a tenfold uncertainty factor (UF) to extrapolate from the general human population to sensitive subgroups, such as children and geriatrics. This study evaluated whether the tenfold UF can be reduced when pharmacokinetic and pharmacodynamic data for pharmaceuticals used by children and geriatrics are incorporated into the risk assessment for human sensitivity. Composite factors (kinetics X dynamics) were calculated from data-derived values for bumetanide, furosemide, metoprolol, atenolol, naproxen, and ibuprofen. For the compounds examined, all of the composite factors were lower than 10. Furthermore, 8 of the 12 composite factors were less than 5.5. Incorporation of human kinetic and dynamic data into risk assessment can aid in reducing the uncertainties associated with sensitive subgroups and further study is encouraged.