5
Physiologically Based Pharmacokinetic Modeling

The discussion of cardiac sensitization in Chapter 4 highlights the need to predict the blood concentration of an agent and the duration of exposure needed to achieve a critical blood concentration. A tool that allows the investigation of those pharmacokinetic factors is the physiologically based pharmacokinetic (PBPK) model. Such models allow one to examine the relationship between external exposure scenarios and internal concentrations in a target tissue, for example, blood. PBPK models incorporate a mathematical description of the uptake, distribution, metabolism, and elimination of chemicals by the body. PBPK models have provided toxicologists with an advantage with respect to understanding the modes of action of chemicals. A number of investigators have used them for various cancer and noncancer end points (Andersen 1981; Conolly and Andersen 1991) as the regulatory agencies, such as the U.S. Environmental Protection Agency (EPA), pursue quantitative risk assessment in public policy. This chapter discusses the application of a PBPK model to estimate blood concentrations of halocarbons after exposure to them at various concentrations.

PBPK MODELING OF EXPOSURE TO FIRE SUPPRESSANTS

Modeling of airborne exposure to cardiac sensitizing agents requires that accounting for short-term (up to 5 min) events. Such a model has been described by Vinegar and colleagues (1998). It includes a respiratory-tract compartment containing a dead space and a pulmonary-exchange volume. The pulmonary-exchange volume contains air space, tissue, and capillary



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 49
Iodotrifluoromethane: Toxicity Review 5 Physiologically Based Pharmacokinetic Modeling The discussion of cardiac sensitization in Chapter 4 highlights the need to predict the blood concentration of an agent and the duration of exposure needed to achieve a critical blood concentration. A tool that allows the investigation of those pharmacokinetic factors is the physiologically based pharmacokinetic (PBPK) model. Such models allow one to examine the relationship between external exposure scenarios and internal concentrations in a target tissue, for example, blood. PBPK models incorporate a mathematical description of the uptake, distribution, metabolism, and elimination of chemicals by the body. PBPK models have provided toxicologists with an advantage with respect to understanding the modes of action of chemicals. A number of investigators have used them for various cancer and noncancer end points (Andersen 1981; Conolly and Andersen 1991) as the regulatory agencies, such as the U.S. Environmental Protection Agency (EPA), pursue quantitative risk assessment in public policy. This chapter discusses the application of a PBPK model to estimate blood concentrations of halocarbons after exposure to them at various concentrations. PBPK MODELING OF EXPOSURE TO FIRE SUPPRESSANTS Modeling of airborne exposure to cardiac sensitizing agents requires that accounting for short-term (up to 5 min) events. Such a model has been described by Vinegar and colleagues (1998). It includes a respiratory-tract compartment containing a dead space and a pulmonary-exchange volume. The pulmonary-exchange volume contains air space, tissue, and capillary

OCR for page 49
Iodotrifluoromethane: Toxicity Review subregions. Respiratory-tract uptake is described on a breath-by-breath basis that allows successful simulation of exhaled-breath concentration of agent during the first minute of exposure. The model was used to simulate exposure of two persons who were trapped in an armored personnel carrier in Israel after the release of a fire extinguishing agent, Halon 1211; one of them died (Vinegar et al. 1998). The investigators re-enacted the release in an identical vehicle. They measured the Halon 1211 concentrations in various portions of the vehicle and found very high concentrations—exceeding 50,000 ppm—within 1 min of the release. Later, they used the PBPK model to simulate the arterial blood concentration at the lowest-observed-adverse-effect level (LOAEL) of Halon 1211 (1.0% or 10,000 ppm). The cardiac-sensitization LOAEL was determined with the Reinhardt et al. (1971) protocol described in Chapter 4. With this PBPK simulation, the authors reported that within 5 min, the arterial blood concentration of Halon 1211 would be about 22 mg/L, which is the critical blood concentration for inducing cardiac sensitization as determined in the dog. At 1 min, the blood concentration would be about 15 mg/L. The investigators then simulated the blood concentrations at the airborne concentrations encountered in the vehicle. In this simulation, the survivor’s arterial concentration at 1 min approached 80 mg/L, at about 20 sec it was closer to 20 mg/L. Hence, this person was able to survive the incident because escape from the vehicle was presumably very quick. For the other person, however, the arterial blood concentration rose very rapidly in the first few seconds to about 30 mg/L, which exceeded the critical blood concentration; at 1 min, the arterial blood concentration was about 170 mg/L. The person died from the exposure because his escape was impaired either because of the physical environment or because of nervous system effects (central nervous system depression) of Halon 1211. The cause of death was judged to be due to cardiac sensitization. The authors suggested that the simulation under actual exposure conditions was consistent with the model predictions when compared with the simulation conducted at the cardiac-sensitization LOAEL. The validated PBPK model can be used to assess exposure to cardiac-sensitizing agents in a number of ways. Each method, however, depends on the determination of the critical blood concentration, typically the peak (steady-state) blood concentration resulting from exposure to the LOAEL. Such data are not often available from dog studies and less often available on humans, the target population. The most direct way to obtain them is to perform a pharmacokinetic study with arterial blood samples taken at various times during exposure. Because the exposure of interest is the threshold concentration or LOAEL, the pharmacokinetic study in dogs

OCR for page 49
Iodotrifluoromethane: Toxicity Review should be performed at this concentration but without epinephrine challenge. An alternative is to use a PBPK model to simulate the blood concentrations in humans subjected to the critical airborne exposure level, that is, the cardiac-sensitization LOAEL. Data on the arterial and venous blood concentrations in dogs during the first 5 min of exposure (and beyond) to iodotrifluoromethane (CF3I) are available (Huntingdon Life Sciences 2000) and have been used in the PBPK model to simulate blood concentrations in humans during exposure to CF3I (Vinegar et al. 2000). This approach is detailed later in this chapter. A limitation to this approach is that the lack of human data makes it difficult to validate the model for predicting human blood concentrations. For ethical reasons it is unlikely that human data will ever be available. Use of a PBPK model to simulate human blood concentrations based on dog and rodent data is a scientifically based approach to assess human health risk from exposure to CF3I and other compounds for which human data are unavailable. With the target arterial concentration determined either experimentally or by simulation, various exposure scenarios can be examined. Vinegar and Jepson (1996) and Vinegar et al. (2000) have proposed using PBPK models for estimating egress times after the release of fire-suppression agents. This model has been extensively reviewed as an approach to developing guidelines for safe exposure to halocarbon fire-extinguishing agents (ISO 2004). Vinegar et al. (2000) have suggested that if one can determine the critical arterial blood concentration at the 5-min cardiac-sensitization LOAEL, egress times can be estimated from the shape of the blood concentration-time curve and thus from the time before cardiotoxic blood concentrations are reached. For this PBPK model, the investigators used experimentally determined human blood:air partition coefficients, which lent confidence to uptake values of the compound in humans. They also used human anatomic and physiologic parameters. However, other tissue:air partition coefficients included in the model were usually determined in rats, not humans. Biochemical parameters included in the model were scaled1 from rodent data to humans because rodent data were usually the only data available. With those parameters, the investigators constructed a model of time versus arterial blood concentration from various airborne concentrations, most notably the cardiac-sensitization no-observed-adverse-effect level (NOAEL) or LOAEL. Monte Carlo methods 1   In the paper by Vinegar et al. (2000), the authors used a 0.75 exponential scaling factor to convert rodent metabolic parameters to humans. That is common practice in risk assessment, but it represents an added variable and assumption to the model.

OCR for page 49
Iodotrifluoromethane: Toxicity Review were used to account for exposure population variability in physiological and biochemical parameters, which control blood concentration. Vinegar et al. (2000) published results of PBPK modeling that simulated human arterial blood concentrations during the first 5 min of exposure to fire-suppression agents. Halon 1301, CF3I, and the four halofluorocarbons (HFCs) HFC-125, HFC 134a, HFC-227ea, and HFC-236fa were examined. The authors used target arterial concentrations based on the lowest measured 5-min value observed in dogs exposed to the agent of interest at the LOAEL and cited data published in a Huntingdon Life Sciences study (2000). Critical arterial blood concentrations are shown in Table 5-1 for Halon 1301, CF3I, HFC-125, HFC-227ea, and HFC-236fa (Vinegar et al. 2000). They used the PBPK model to simulate arterial blood concentrations for various exposure scenarios. A central objective of the modeling was to determine, at efficacious fire-suppressant concentrations, how long a person could safely be exposed, that is, for what duration the blood concentration would remain below the critical blood concentration as determined in dog cardiac-sensitization studies. For each agent, simulations were run at increasing exposure concentrations to determine the concentrations that would be considered safe for 5 min and at higher concentrations to determine the duration of safe exposure before the critical blood concentration was reached. The results of the studies are summarized in Table 5-1. On the basis of measured dog arterial blood concentrations, results of model simulations for Halon 1301 indicated that at the LOAEL (7.5%), humans could be safely exposed for only 0.42 min (25.2 sec) before their blood concentrations reached the critical point measured in dogs. Similarly, for HFC-236fa, humans could be safely exposed at the LOAEL (15.0%) for 0.49 min (29.4 sec). For HFC-125, humans could be safely exposed at the LOAEL (10.0%) indefinitely, at up to 11.5% for 5 min, and at higher concentrations for shorter periods. For HFC-227ea, humans could be safely exposed at the LOAEL (10.5%) for 5 min and at higher concentrations for shorter periods. For CF3I, humans could be safely exposed at the LOAEL (0.4%) for 0.85 min (51 sec), at 0.35% for 4.30 min, and up to 0.3% for 5 min or more; the NOAEL of CF3I is 0.2% (Vinegar et al. 2000). Confidence in this PBPK model may increase if experimental data were available for human tissue partition coefficients of the halocarbons. However, Monte Carlo simulations included lognormal distributions of values for each tissue partition coefficient with upper and lower limits of two geometric standard deviations. Therefore, experimental determination of human tissue partition coefficients would probably not change the model predictions as determined with rodent tissue partition coefficients.

OCR for page 49
Iodotrifluoromethane: Toxicity Review TABLE 5-1 Simulated Egress Times for Exposure to Fire-Suppression Agents Fire-Suppression Agent Critical Blood Concentration,a mg/L Cardiac-Sensitization Exposure LOAEL,b % Simulated Egress Time at Exposure LOAEL,c min Highest Simulated “Safe” 30-sec Exposure Concentrationc Halon 1301 25.7 7.5 0.42 7.0% (0.59 min) CF3I 12.9 0.40 0.85 0.40% (0.85 min) HFC-125 47.8 10 >5 13.5% (0.50 min) HFC-227ea 26.3 10.5 5 11.5% (0.60 min) HFC-236fa 90.37 15 0.49 14.5% (0.55 min) aBased on arterial blood concentrations in dogs exposed to the cardiac-sensitization LOAEL. bBased on cardiac-sensitization studies in dogs. c“Safe” human egress times, based on lowest measured 5-min arterial blood concentration in exposed dogs. Source: Vinegar et al. 2000. MODEL SIMULATIONS TO DETERMINE SAFE EXPOSURES The Army reviewed the toxicity of CF3I in 1999 (McCain and Macko 1999) and updated its review in 2002 (Chaney 2002) (see Appendix B). At the time of the 1999 review, CF3I was accepted as a substitute for Halon 1301 in normally unoccupied areas under the EPA Significant New Alternatives Policy (SNAP). On the basis of SNAP, any employee that could possibly be in the area must be able to escape within 30 sec, and employers were required to ensure that no unprotected employees entered the area during agent discharge. The Army concluded in 1999 that a potential hazard for cardiac sensitization was associated with acute exposure to CF3I at over 0.2% (2,000 ppm), the reported NOAEL based on cardiac sensitization in the dog model. It further concluded that the available data indicated that toxicity of CF3I precludes its use in many Army systems without further evaluation. The Army noted in its 2002 update that EPA had rescinded use conditions imposed under the SNAP program that limited human exposure to halocarbon and inert-gas agents used in the fire-suppression and explosion-protection industry, including CF3I. In April 2002, the EPA SNAP program recommended that use of CF3I and several other halocarbons be in accordance with the safety guidelines in the latest edition of the National Fire Protection Association (NFPA) 2001 Standard on Clean Agent Fire-Extinguishing Systems (NFPA 2000).

OCR for page 49
Iodotrifluoromethane: Toxicity Review According to NFPA 2001 Standard, on the basis of PBPK modeling it would be considered safe for a human to be exposed to CF3I above the NOAEL (0.2%) and up to 0.3% for as long as 5 min. At concentrations above 0.3%, “safe” exposure time decreases, but exposure is still allowed. The EPA-approved PBPK model simulates how long it will take the human arterial concentration to reach the critical point (as determined in the dog cardiac-sensitization test) during human inhalation of any particular concentration of a halocarbon. As long as the simulated human arterial concentration remains below the critical point, the exposure is considered safe. Inhaled halocarbon concentrations that produce human arterial concentrations equal to or greater than the critical concentration are considered unsafe because they represent inhaled concentrations that potentially yield arterial concentrations at which cardiac sensitization occur in dogs. The PBPK model predicts that at concentrations of up to 0.4% a human could be exposed for up to 30 sec without exceeding the critical arterial concentration. The Army concluded that because of the acute toxicity of CF3I at concentrations over 0.2%, it could not endorse the NFPA 2001 Standard recommendations for “safe” exposure to CF3I in as much as the recommendations were determined by using PBPK modeling based on a LOAEL (0.4%) for cardiac sensitization in the dog that resulted in death of the animal. In the NFPA standard (NFPA 2000), egress times have historically reflected knowledge of the NOAEL and LOAEL with recognition that cardiac sensitization to fluorocarbons will occur within 5 min. Establishing egress times by using PBPK modeling predictions based on the human blood:air partition coefficients in combination with Monte Carlo simulations to account for sensitive individuals in the population adds a level of quantification to the risk assessment for the safe use of fire-suppression agents. In the design of cardiac-sensitization studies, an airborne concentration is selected for administration on the basis of the test compound’s structural relationship with other compounds, the known acute toxicity of the compound in question, and the physical and chemical properties of the material. Furthermore, because this protocol was established to rank compounds according to cardiac-sensitization potency, there was no attempt to study incremental increases in airborne concentrations. For example, CFC-12 is known to induce cardiac sensitization at 5.0% in five of 12 dogs, and this concentration has been recognized as the LOAEL of CFC-12. No responses were observed in 12 dogs at 2.5%, this is the NOAEL of CFC-12. On the basis of the incidence at 5.0%, one would reasonably estimate that a true LOAEL could be lower. Therefore, assigning a LOAEL to such a

OCR for page 49
Iodotrifluoromethane: Toxicity Review compound and using it as a point of departure in a risk assessment may not be appropriate. Using the NOAEL, although possibly conservative, offers a more reliable starting point. In spite of that limitation, the use of PBPK models does have merit for quantifying acceptable exposure magnitudes relative to egress times. However, it is imperative that the input parameters in the model be known with confidence, for example, LOAEL and partition coefficients. Most important, the model should be accurately validated for each compound under investigation before it is used to recommend exposure-egress time relationships. During cardiac-sensitization testing, collection of blood to determine concentrations of the compound over time, particularly at shorter intervals, such as, less than 5 min, will yield the most robust data. Alternatively, if the model has been validated with data for similar compounds, that information and the blood:air partition coefficient may be used. In the absence of such data, the NOAEL would be the conservative determinant for establishing egress time for fire-suppression. A validated PBPK model does exist for determining arterial blood concentrations of CF3I and other halon replacements during short-term exposure, and arterial concentrations of CF3I in dogs during the first 5 min of exposure in the absence of an epinephrine challenge are also available. It is unlikely that blood concentrations of CF3I would be substantially different in the presence of an exogenous epinephrine challenge of 8 μg/kg using the cardiac-sensitization protocol. Because the circulating blood concentrations of epinephrine are low under normal physiologic conditions, intravenous administration of epinephrine at 8 μg/kg would not markedly elevate the existing amounts of endogenous epinephrine. Use of arterial CF3I concentrations measured in dogs in the absence of exogenous epinephrine is a reasonable approach to estimate the critical arterial blood concentration that would result in a cardiac event in epinephrine-challenged dogs. The dog cardiac-sensitization model was developed to rank potency of halocarbons, not as a risk-assessment tool; however, the resulting data are available and are useful in a PBPK model. PBPK models have been evaluated for many chemicals starting with publications by Andersen (1981). What is unique about this application is that the dose is calculated as a function of time over fairly short periods—1-5 min. This approach has also been studied extensively (Vinegar and Jepson 1996; Vinegar et al. 1998, 1999, 2000). Thus, the subcommittee finds that the use of a validated PBPK model is a reasonable scientifically based approach to determining safe egress times for exposure to CF3I.

OCR for page 49
Iodotrifluoromethane: Toxicity Review ARMY CONCERNS Although they were not part of the statement of work of the subcommittee, the Army posed several questions with regard to the cardiac-sensitization potential of CF3I. The subcommittee addressed the first in Chapter 4. The others and the subcommittee’s responses follow here. Historically, the cardiac-sensitization dog studies were designed to identify cardiac-sensitization potential, not to quantify risks. Please comment on the use of these data in PBPK modeling to estimate blood concentrations, which pose a threat to human health. The subcommittee finds that although the dog cardiac-sensitization model was developed to rank the potency of halons and not as a risk assessment tool; nevertheless, these data are available and cannot be ignored and are appropriate for use in a PBPK model. PBPK models have been evaluated for many chemicals. What is unique about this PBPK application is that the dose is calculated as a function of time over fairly short time periods, that is, 1-5 min. This approach has also been studied extensively for a variety of halons and halon substitutes, including CF3I. According to the NFPA 2001 Standard, based on PBPK modeling, it would be considered safe for a human to be exposed to levels of CF3I above the NOAEL and up to 0.3% (3,000 ppm) for as long as 5 minutes. At concentrations above 0.3% (3,000 ppm), the time for “safe” exposure decreases, but exposure is still allowed. [The Army] feels that, given the severe toxic effect (death) that was observed in the dog model at the LOAEL of 0.4% (4,000 ppm), and considering that in many Army applications there is still significant potential for human exposure in unoccupied areas, a conservative approach is justified in defining “safe” exposure levels for military applications of CF3I. [The Army’s] current recommendations are in agreement with previous EPA SNAP guidelines for CF3I, but [the Army is] not comfortable adopting the NFPA 2001 recommendations, based on PBPK modeling data for this particular agent because of the severe toxic effect (death) that was observed in the dog model at the LOAEL of 0.4%. Based on the available information, is this position reasonable?

OCR for page 49
Iodotrifluoromethane: Toxicity Review The subcommittee finds that the use of a validated, EPA-approved PBPK model is a reasonable scientifically based approach to determining safe egress times for exposure to CF3I. The NOAEL and LOAEL for CF3I as determined with the dog cardiac-sensitization model are 0.2% and 0.4%, respectively. According to the PBPK model, people could be safely exposed at 0.4% for about 51 sec before the critical blood CF3I concentration for cardiac sensitization is reached. Furthermore, people could be exposed to concentrations as high as 0.3% for more than 5 min without reaching the critical blood concentration. The Army’s decision to use an exposure limit of 0.2% (2,000 ppm) in normally unoccupied areas is a conservative policy decision to protect military personnel from health effects of CF3I exposure in (undefined) Army applications.