Approaches to Characterizing HumanHealth Risks of Exposure to FibersVanessa T. Vu and David Y LaiOffice of Pollution Prevention and Toxics, U.S. EnvironmentalProtection Agency, Washington, D.C.

Naturally occurring and man-made (synthetic) fibers of respirable sizes are substances thathave been identified by the U.S. Environmental Protection Agency (U.S. EPA) as prioritysubstances for risk reduction and pollution prevention under the Toxic Substances ControlAct (TSCA). The health concern for respirable fibers is based on the link of occupationalasbestos exposure and environmental erionite fiber exposure to the development of chronicrespiratory diseases, including interstitial lung fibrosis, lung cancer, and mesothelioma inhumans. There is also considerable laboratory evidence indicating that a variety of fibers ofvarying physical and chemical characteristics can elicit fibrogenic and carcinogenic effects inanimals under certain exposure conditions. This paper discusses key scientific issues andmajor default assumptions and uncertainties pertaining to the risk assessment of inhaledfibers. This is followed by a description of the types of assessment performed by the U.S.EPA to support risk management actions of new fibers and existing fibers under TSCA. Thescope and depth of these risk assessments, however, vary greatly depending on whetherthe substance under review is an existing or a new fiber, the purpose of the assessment, theavailability of data, time, and resources, and the intended nature of regulatory action. Ingeneral, these risk assessments are of considerable uncertainty because health hazard andhuman exposure information is often incomplete for most fibers. Furthermore, how fiberscause diseases and what specific determinants are critical to fiber-induced toxicity andcarcinogenicity are still not completely understood. Further research to improve ourknowledge base in fiber toxicology and additional toxicity and exposure data gathering areneeded to more accurately characterize the health risks of inhaled fibers. Environ HealthPerspect 105(Suppl 5):1329-1336 (1997)

Key words: respirable fibers, mineral fibers, man-made vitreous fibers, synthetic fibers,asbestos, risk assessment

IntroductionA major goal of the U.S. EnvironmentalProtection Agency (U.S. EPA) is the pre-vention, reduction, or elimination ofharmful pollutant releases into the generalenvironment. Naturally occurring andman-made (synthetic) fibers of respirablesizes are substances that have been identi-fied as priority substances for risk reduc-tion and pollution prevention. The healthconcern for respirable fibers is based on

the link of occupational asbestos exposureand environmental erionite fiber expo-sure to the development of chronic respi-ratory diseases, including interstitial lungfibrosis, lung cancer, and mesotheliomain humans (1). Moreover, there is exten-sive experimental evidence indicatingthat a variety of fibers of varying physicaland chemical characteristics can elicitfibrogenic and carcinogenic effects in

This paper is based on a presentation at The Sixth International Meeting on the Toxicology of Natural and Man-Made Fibrous and Non-Fibrous Particles held 15-18 September 1996 in Lake Placid, New York. Manuscriptreceived at EHP26 March 1997; accepted 15 April 1997.

The views and opinions expressed by the authors do not necessarily reflect the official position or policies ofthe U.S. Environmental Protection Agency. Mention of trade names or commercial products does not consti-tute endorsement or recommendation for use.

Address correspondence to Dr. V.T. Vu, U.S. Environmental Protection Agency (7403), 401 M Street, SW,Washington D.C. 20460. Telephone: (202) 260-1243. Fax: (202) 260-1283. E-mail: vu.vanessa@epamail.epa.gov

Abbreviations used: IT, intratracheal instillation; MF, modifying factor; MOE, margin of exposure; NOAEL, noobserved adverse effect level; OPPT, Office of Pollution Prevention and Toxics; PMN, premanufactured notice;RCF, refractory ceramic fiber; ROS, reactive oxygen species; SNUR, significant new use rule; TSCA, ToxicSubstances Control Act; UFs, uncertainty factors; U.S. EPA, U.S. Environmental Protection Agency; WOE,weight of evidence.

laboratory animals under certain exposureconditions (2,3).

The U.S. EPA's Office of PollutionPrevention and Toxics (U.S. EPA/OPPT)is responsible for the evaluation of humanhealth risk from occupational, consumer,and environmental exposure to respirablefibrous particles, and the development ofregulation under the Toxic SubstancesControl Act (TSCA) to prevent, reduce, oreliminate these potential risks. TSCAauthorizes the U.S. EPA to restrict or pro-hibit the manufacture, processing, dis-tribution, use, or disposal of chemicalsubstances already in commerce, and ofnew substances when there is a reasonablebasis to conclude that any such activityposes an unreasonable risk to humanhealth or the environment. TSCA alsoempowers the agency to require the manu-facturers and processors of a chemical todevelop the necessary toxicity and exposuredata, if the agency determines that there issignificant human or environmental expo-sure to such chemical, or that such chemi-cal may pose an unreasonable risk but lackssufficient data to take action. TSCA, how-ever, requires the U.S. EPA to consider andweigh health and environmental risks,potential costs and benefits, and availabilityof alternative materials or technologiesbefore taking any risk reduction actions.

This article describes the risk assess-ment approaches utilized by the U.S.EPA/OPPT in characterizing humanhealth risks of exposure to new and exist-ing fibers i.e., fibers that are listed on theTSCA inventory other than asbestos. Thispaper begins with a brief overview of thegeneral principles of human health riskassessment and a discussion of the majorscientific issues, uncertainties, and defaultassumptions pertaining to the risk assess-ment of inhaled fibers, and of researchneeds to improve the scientific basis forfuture risk assessments. This is followedby a description of the types of assess-ment performed by the U.S. EPA/OPPTto support risk management actions ofnew fibers and existing fibers underTSCA. Attention is focused on the kindsof scientific information and key factorsthat are considered in the hazard anddose-response assessments of the riskassessment process. Exposure assessment isonly briefly addressed. The risk assessmentof asbestos conducted by the U.S. EPAwill not be discussed here, as it can befound elsewhere (4).

Environmental Health Perspectives * Vol 105, Supplement 5 * September 1997 l1329

VUAND IAI

General Principlesof Risk AssessmentRisk assessment involves the analysis andsynthesis of the entire knowledge base onan environmental agent to characterize theanticipated risk from human exposure tothe agent. The U.S. EPA has followed thebasic National Research Council risk assess-ment paradigm (5,6) as a foundation for itshuman health risk assessment guidance(7-9). Risk assessment consists of fourcomponents: hazard identification, dose-response relationship assessment, exposureassessment, and risk characterization.

The U.S. EPA's risk assessmentguidelines contain detailed guidance on theapplication of default assumptions, i.e., sci-ence policy choices, and the associateduncertainties. It is generally accepted thatdefault assumptions are necessary tools inperforming risk assessment to bridge thegaps in general scientific knowledge and indata for a particular agent. Default assump-tions are developed generically, not on anagent-by-agent basis. The intent is thatdefault positions are taken for all agentsunless case-specific data clearly indicatethat the default assumption is no longerplausible and another inference positionmay be more appropriate.

Hazard identification is a qualitativeassessment of available scientific data tomake an informed judgment about thepotential adverse human health effectsposed by an agent. The principal questionis what is known about the capacity of anenvironmental agent for causing adverseeffects in humans, and under what expo-sure conditions an identified adverse effectmay be expressed. The U.S. EPA uses aweight of evidence (WOE) approach toevaluate the potential human health effectsof an agent. Information for this WOEevaluation is derived from available studiesin humans and laboratory animals on theagent and on structurally related sub-stances, together with other relevant infor-mation such as chemical and physicalproperties, toxicokinetic data, results ofshort-term assays for biochemical, molecu-lar, genetic, and cellular effects, and anyother mechanistic data.

Because human data are not alwaysavailable, most hazard evaluations are, ofnecessity, based on animal data. Con-sequently, a major default assumption inhazard assessment is that an agent causingadverse effects in laboratory animals will alsohave the potential to cause adverse effects inhumans unless the available data prove oth-erwise. In addition, in the absence of

evidence to the contrary, data from animalstudies in the most sensitive animal speciesare used for dose-response assessment pur-poses. The underlying scientific basis forthis default assumption is the possibility thathuman sensitivity is as high as the mostsensitive responding animal species.

The dose-response assessment evalu-ates the quantitative relationship of expo-sure and dose to the degree of response inexisting studies in which adverse healtheffects have been observed. When environ-mental exposures of interest are outsidethe range of observation, extrapolationsare necessary to estimate the likelihood ofadverse effects in populations potentiallyat risk, if it is deemed appropriate andscientifically supportable.

Because cancer and noncancer effectshistorically are believed to occur by differ-ent modes of action, different approachesto dose-response relationship assessmenthave been developed. It is believed thatmost cancers, if not all, develop as a conse-quence of mutations of critical genes.Because a single chemical-DNA interac-tion may lead to a mutation, and becausecancer is thought to arise from single cells,it follows that any dose of an agent may beassociated with some finite risk. This hasled the agency to employ a default proce-dure that cancer risk should be estimatedby a linear, nonthreshold dose-responseapproach when the mode of action is notknown, supportive of linearity, or is insuf-ficient to support a nonlinear or thresholdmode of action (8). It is recognized thatsuch extrapolation does not necessarily givea realistic prediction of the risk but is con-sistent with the agency's goal to provide anestimate of the upper limit to risk.

In contrast, based on our understandingof homeostatic and adaptive mechanisms,the default approach for dose-responseassessment of noncancer toxicity, i.e.,toxicity other than cancer and gene muta-tions, assumes an identifiable thresholdbelow which effects are not observable(10,11). It should be noted, however, thatfuture case-specific knowledge of themechanisms of chemically induced toxicityand carcinogenicity may blur this distinc-tion between approaches for noncancertoxicity and carcinogenicity.

The exposure assessment identifies thelikely sources of human exposures to theenvironmental agent, environmental path-ways for exposure (e.g., air, water, soil,food), potential routes of exposure (e.g.,oral, dermal, inhalation), populations atrisk, including those of highly exposed

groups and highly susceptible groups, andestimates exposure and dose levels thatimpact the exposed individuals or popula-tions. The exposure assessment relies onmany kinds of information, some based onactual measurements and some developedusing predictive models and surrogate data.

Risk characterization, the last step in riskassessment, is the integrated analysis of thepreceding three steps in risk assessment toreach a conclusion about the nature andmagnitude of expected risk. The predictedrisk can be qualitative (e.g., high or lowprobability) or quantitative (e.g., one in amillion probability of occurrence). Riskcharacterization also includes a discussion ofthe confidence, limitations, and uncertain-ties of the risk assessment, given the con-straints of available data, and the state ofscientific knowledge, significant issues, andscientific assumptions and policy choices.

It should be noted that not every U.S.EPA risk assessment contains all four com-ponents. The scope and depth of the riskassessment depend on the availability ofscientific data, resources, and time, thepurpose of the assessment, any legislativemandates, and the nature of the intendedregulatory action.

Issues in Risk Assessmentof Inhaled FibersThe following section discusses some keyscientific issues and uncertainties to be keptin mind when assessing the potentialhealth risk from exposure to fibers. Theseinclude a) the incomplete knowledge ofhow certain mineral fibers cause disease inhumans and laboratory animals and whichspecific fiber properties are important ininfluencing their biologic and toxicologiceffects; b) the limitations of availableexperimental models to accurately predictadverse outcomes in humans; and c) thelack of reliable data on human exposures toa wide variety types of fibers of concern.All of these factors contribute in part to theuncertainties of the risk assessment process.

Mechanisms ofFiber-inducedToxicity and CarcinogenicityAlthough there have been significantadvances in our knowledge about thehealth effects of asbestos and other fibers inrecent years, the mechanisms by whichmineral fibers cause fibrogenic and car-cinogenic effects in humans and laboratoryanimals are not clearly understood. Fiber-induced fibrogenesis appears to arise from achronic inflammation process that involvesthe release of cellular mediators (e.g.,

Environmental Health Perspectives * Vol 105, Supplement 5 * September 19971 330

HUMAN HEALTH RISK ASSESSMENT OF INHALED FIBERS

lysosomal enzymes, arachidonic acidmetabolites, neutral proteases, cytokines,growth factors, chemotactic factors, orreactive oxygen species [ROS]) by activatedalveolar macrophages and other inflamma-tory cells. These events are believed to befollowed by the proliferation of fibrolastsand deposition of collagen within thealveolar spaces and the interstitium.

With regard to fiber-induced carcino-genicity, several hypotheses have been pro-posed. These include: a) DNA damage byROS induced by fibers; b) direct DNAdamage by physical interactions betweenfibers and target cells; c) enhancement ofcell proliferation by fibers; d) fibers thatelicit a chronic inflammatory reaction,which leads to prolonged release of ROS,cytokines, and growth factors; and e) fibersthat act as cocarcinogens or carriers ofchemical carcinogens to the target tissue. Itis likely that all of these mechanisms con-tribute to the carcinogenicity of fibersbecause all these effects have been observedin various in vitro systems of human andmammalian cells using different types offibers (12).

The role of fiber-induced fibrosis incarcinogenesis is not clear. Macrophagelysis and chronic inflammation inducedby fiber exposure may not only lead tolung fibrosis but also to the developmentof some lung tumors in rats. Other ratlung tumors, however, appear to have norelationship to fibrotic scar.

Critical Determinants ofFiberToxicity and CacinogenicityCurrent scientific knowledge indicates thata major determinant of fiber toxicity andcarcinogenicity is fiber dimension. There isextensive evidence relating to the impor-tance of fiber size in lung deposition andclearance of fibers, which in turn governthe bioavailability of fibers at target tissues(13). Fiber diameter is the major determi-nant for the deposition of fibers. Onlyfibers less than about 3.5 pm in diametercan reach the alveolar spaces. Fiber lengthalso influences the deposition and clearanceof fibers. Fibers longer than 20 pm aremore readily deposited by interception atairway bifurcations. In general, short fibers(< 5 pm) are cleared more rapidly thanlong fibers (> 5 pm). Fiber size also playsan important role in relation to cellularmechanisms of toxicity and carcinogenic-ity. Long fibers of a given fiber type aregenerally more biologically active thanshorter fibers, regardless of the measuredbiological end points such as cytoxicity,

cell transformation, or aneuploidy.Furthermore, long fibers (> 5 pm) are morecarcinogenic and fibrogenic than shortfibers of asbestos and other fibers (< 5 pum)in chronic studies in rats by inhalation(14,15) or intracavitary injection (2,3).

Fiber biopersistence may also be animportant determinant of fiber toxicity andcarcinogenicity. The concept of biopersis-tence arose from the observation that differ-ent natural and synthetic fibers havedifferent lung retention characteristics; somepersisting over long periods, others beingless persistent. It has been hypothesized thata fiber with critical dimensions will be car-cinogenic if it is sufficiently durable toremain chemically and physically intact inlung tissue in close contact with the targetcells (3). Although more durable fibersappear to be more carcinogenic than moresoluble fibers e.g., refractory ceramic fiber(RCF) versus glass fiber, neither the influ-ence of biopersistence on fibrogenesis andcarcinogenesis nor the length of timerequired for a fiber to remain in the lung toexert a pathogenic effect has been ade-quately defined. For instance, chrysotileasbestos fiber is significantly less biopersis-tent than amphibole asbestos, yet the fiberis clearly carcinogenic and fibrogenic inhumans. In view of the many differentmechanisms by which fibers might influencethe carcinogenic process, it is difficult todetermine the degree of biopersistence nec-essary for fiber carcinogenicity. In addition,available models of fiber solubility or dura-bility in physiological systems, as well asfiber biopersistence in the lung, remain tobe validated. Nevertheless, it is generallybelieved that fibers capable of biopersistencein the lung are of greater concern.

Other fiber characteristics such assurface area, chemistry, and chemicalleaching are also likely to play a role infiber-induced toxicity and carcinogenicity,although these are much less well under-stood. For example, acid-leached chrysotilefibers are less carcinogenic than nativefibers (16).

Definition ofFibers ofConcernA major issue for the U.S. EPA is todefine the category of fiber of concern. Asdiscussed above, it is recognized that thereare a number of fiber properties e.g., fibersize, chemical composition, biopersis-tence, and surface chemistry that are likelyto exert an influence on the pulmonarytoxicity and carcinogenicity of inhaledfibers. However, it is the presence of along thin fibrous shape that appears to be

the most important determinant of fiber-induced pathogenicity. Still, it is notknown whether there is any single cutoffof fiber length or diameter that implieshuman safety. Moreover, the conventionaldefinition of a fiber used for industrialhygiene purposes continues to be apractical index for risk assessment.Consequently, fibrous particles of res-pirable sizes are considered potentially haz-ardous unless there are available data todemonstrate otherwise.

For regulatory purposes, the U.S. EPAdefines fiber as a particle of length > 5 pmwith an aspect ratio (ratio of fiber length tofiber diameter) of at least 3:1 (4). This def-inition has been widely adopted by otherorganizations (17,18). Fibers with aero-dynamic diameters of 10 to 12 pm or less,or actual fiber diameter of about 3.0 to 3.5pm or less, are generally considered res-pirable for humans (19). The upper limitof fiber length for human respirability isabout 200 pm (20). This definition appliesto any particle that fulfills these criteriaregardless of chemical composition ormineralogic characteristics.

Exposure CharterizationGiven that fiber exposure characteristicsgreatly influence the nature of the hazard, itis essential that human exposures to fibersare measured to the extent feasible and fullycharacterized (e.g., airborne fiber dimen-sions and physicochemical characteristics,fiber concentrations per unit air volume,frequency and duration of exposure) to per-mit a more reliable characterization ofpotential human risk. Available data onhuman exposures to fibers, however, areoften limited. Furthermore, standard ana-lytical methods to characterize and monitorfiber exposures have been confined to onlya few fiber types such as asbestos and man-made vitreous fibers. Improved techniquesfor characterizing human and animalexposure to other types of fibers are needed.

It is also important to note that fibersspecified by the same name are used innumerous applications or products thatmay have varying physicochemical charac-teristics. Each fiber type, therefore, maypose different degrees of health hazard andshould not be treated as a single entity.Moreover, fibers released into ambient airthroughout the product cycle of the fibercan also vary greatly in size and characteris-tics. For example, a fiber may undergobreakage across the fiber axis, as in glassfibers, or may generate fibrils due to split-ting longitudinally e.g., chrysotile. Fibrils

Environmental Health Perspectives * Vol 105, Supplement 5 - September 1997 1331

VU AND LAI

may also be peeled off from a core fibere.g., p-aramid. A fiber may undergostructural changes during its use. Forexample, RCF may partially transformfrom an amorphous form into crystallinesilica after use under high-temperatureconditions. Thus, each fiber type may posea different nature and magnitude of healthrisk under different exposure conditionse.g., workplace, user environments.

Hazard EvaluationThe question of fiber fibrogenicity andcarcinogenicity in humans should beanswered within the framework of all avail-able evidence. Judgment about the WOEinvolves consideration of the quality, ade-quacy, and consistency of responsesinduced by the fiber in question. The ques-tions to be asked may indude: a) Can thefiber be inhaled and deposited in thehuman lung? b) What is known about itsdeposition pattern, clearance rate, reten-tion, and translocation pathways? c) Doesthe deposited fiber have certain physicaland chemical characteristics that are criticaldeterminants of toxicity and carcinogenic-ity? d) What is the degree of evidence for acausal relationship between chronic respi-ratory effects in humans and fiber expo-sure? e) Is the fiber toxic to target cells?J) Does the fiber induce chronic inflam-mation and/or pathological lesions in theexposed animals after prolonged exposureto the fiber? g) What do we know aboutthe toxicity and carcinogenicity potentialof closely related fibrous particles?

Human data, when available, aregiven first priority in establishing thepresence of an adverse effect in exposedhuman populations and a quantitativerelationship between environmentalexposure and adverse effects. For mostfibers, there is insufficient information oneffects in humans. In such cases, hazardevaluation relies primarily on animalstudies, most often in the rat, on the fiberitself or on closely related fibrous partides.Information on fiber disposition (deposi-tion, translocation, dearance), dissolution,biopersistence, in vitro biological activity,and mechanistic data can provide addi-tional insights into possible toxicity andcarcinogenicity of a fiber.

It should be pointed out that consider-able uncertainties exist when extrapolatingexperimental findings to hazard assessmentin humans. When making such extra-polations, one must recognize that thequalitative and quantitative aspects of fiberdeposition, retention, and clearance in

rodents are considerably different fromthose in humans due to differences in theanatomy and physiology of the respiratorytract (21,22). First, the respirabiity of rats isdifferent from respirability in humans.Fibers with aerodynamic diameters > 3 pimare not respirable by the rat but more than20% of these fibers can be deposited in thehuman lung. Second, fibrous particles arepreferentially deposited at the alveolar ductbifurcations in rats, whereas they aredeposited mainly at the bronchiolar bifurca-tions in humans. These differences maycontribute to the different pattern of lesionsamong species. For example, asbestos causesbronchogenic carcinomas in humans;asbestos and other mineral fibers induceperipheral lung tumors in rats. Hamsters, onthe other hand, seem to be more susceptibleto the development of fiber-inducedmesothelioma than lung tumors. The over-all consequence is that inhalation studies inrodents may not necessarily be predictive ofhuman toxicity and carcinogenicity.Interpretations of animal study resultsshould be put into perspective with regardto potential human hazard and risk.

There has been considerable debateconcerning the relevance of various routesof exposure to cancer hazard assessment offibers in humans. The advantages and limi-tations of each route of administration arewell recognized (23,24). Because inhala-tion is the major route of human exposure,positive results of an inhalation study inanimals might have significant implicationsfor hazard and dose-response assessment inhumans. This finding would be bolsteredby positive results from studies usingnonphysiological routes of administrationsuch as intratracheal instillation (IT),intrapleural inoculation or implantation,and ip injection.

On the other hand, lack of tumorigenicresponses in an inhalation study does notnecessarily mean that the fiber is non-hazardous to humans, because of speciesdifferences in the respirability and suscepti-bility between humans and rodents as dis-cussed above. Such a finding, however,would strongly indicate that the fiber doesnot exhibit carcinogenic potential inhumans if it was demonstrated that the tar-get tissues (i.e., the lung) were exposed tosufficient quantities of critical-size fiberscompared with a positive control (e.g.,asbestos). Evidence for the absence of car-cinogenic potential of the fiber in humansmust be corroborated by consistent lack ofbiological and toxicological effects fromother studies.

Studies using instillation or injectionmethods of administration are of consider-able value for the evaluation of the poten-tial human hazard of fibers. Positive resultsin instillation or injection studies wouldsuggest a potential hazard to humans, butfurther investigation would be needed for afirm evaluation ofthe inhalation hazard forhumans. However, a negative result in suchstudies would suggest that the fiber proba-bly is of low human hazard potential andthe need for additional inhalation testingwould be mitigated. An exception to thiswould be for fibers that have the ability toagglomerate (certain organic fibers such asp-aramid) and tend to reduce the actualnumber of single fibers at target tissues.

Dose-Responsew rapolatonTo more accurately predict and characterizehuman health risks from inhaled fibers, it isnecessary to understand the mechanisticlinkage between fiber exposure and bio-logically effective dose and between thebiologically effective dose and response.Currently, validated biologically basedmodels and mathematical models describingfiber exposure/dose-response relationshipsfor fiber-induced toxicity and carcinogenic-ity in quantitative terms are not yet avail-able. There is still a lack of information onthe disposition of inhaled fibers both inlaboratory animals and humans. Moreover,the knowledge of how fibers cause biologi-cal and pathological effects is still incom-plete, and the question of how long a fiberhas to persist in the target tissue to inducea biologic and pathologic response has yetto be fully investigated. Given the incom-plete knowledge of fiber toxicology, theU.S. EPA often uses fiber exposure con-centration as a surrogate for dose for riskassessment purposes.

There are other uncertainties associatedwith performing fiber exposure/responseassessments when animal data are used. Anumber of scientific judgment and sciencepolicy choices must be made concerning therelevance of the animal model to humansand the appropriateness of extrapolatingresults from high experimental exposure torelatively low occupational and environmen-tal exposures. The rat inhalation model isgenerally accepted as suitable for establish-ing the exposure/response relationship, inspite of known limitations. An exception tothis would be cases where available dataindicate that other animal species may bemore sensitive (e.g., the hamster is moresusceptible to RCF-induced carcinogenicityof the pleura than the rat). This policy

Environmental Health Perspectives * Vol 105, Supplement 5 * September 19971 332

HUMAN HEALTH RISK ASSESSMENT OF INHALED FIBERS

position is considered reasonable because itis assumed that humans are at least asinnately sensitive as the most sensitivespecies tested. Furthermore, there isevidence that the rat inhalation model mayunderestimate the risk in humans, especiallyfor mesotheliomas (3).

Dose-response assessment for fiber-induced respiratory noncancer toxicity uti-lizes the uncertainty factor approach. Thisapproach assumes that a safe exposure levelexists, i.e., at exposures below the thresh-old, clinical manifestations of pulmonaryfibrosis or pleural changes are unlikely. Itinvolves the identification of the highestfiber exposure concentration at which thereare no statistically or biologically signifi-cant increases in the frequency or severityof adverse effects between the exposed pop-ulation and its appropriate control, and isknown as the no observed adverse effectlevel (NOAEL). To conclude that a partic-ular effect is adverse requires sound profes-sional judgment and a clear articulation ofthe scientific rationale. Because a fiber mayelicit more than one end point, the criticalend point used in the dose-response assess-ment is the effect with the lowest NOAEL.

With regard to dose-response assess-ment for carcinogenic effects of inhaledfibers, a linear, nonthreshold dose-responseapproach is generally used. This assump-tion is based on the current knowledgethat asbestos fibers can influence the car-cinogenic process at either early or latestages by both genetic and epigeneticmechanisms (25). However, a nonlinearthreshold approach may be appropriate ifthere is sufficient evidence to support themode of action of certain fibrous particleshaving a threshold, e.g., if the fiber-induced carcinogenesis is a secondaryeffect of fibrosis/scarring, which itself is athreshold phenomenon.

Risk Assessment of NewFibrous Substancesunder TSCASection 5 of TSCA requires anyone whointends to manufacture or import into theUnited States a chemical substance not onthe TSCA inventory and not exemptedfrom TSCA, to submit a formal noticeknown as the premanufactured notice(PMN) to the U.S. EPA. The U.S. EPA isrequired to review each PMN and make adecision within 90 days about potentialrisk to human health and environment ofallowing the manufacture, import, process-ing, distribution, use, and disposal ofthat chemical.

Risk assessment of new fibers is generallyhampered by lack of information submittedin the PMN, as the submitter is notrequired to generate toxicologic informa-tion on the new fibrous substance. As aresult, the U.S. EPA/OPPT relies on thehazard and exposure information submitted(if any) and on readily available toxicologicand epidemiologic data on related fibroussubstances to make a scientific judgment ofwhether the new fiber may present sig-nificant risk to human health or the envi-ronment. Thus, risk assessments of newfibers may be based almost entirely oncurrent scientific knowledge about criticaldeterminants of fiber toxicity and car-cinogenicity, surrogate dose response andexposure data, and policy assumptions.A regulatory decision is subsequently

made either to drop the case from furtherconsideration or take control action to pro-tect against that risk. If it is determinedthat the fiber is of concern but there areinsufficient data to assess the potentialhealth effects of the fiber, control actionmay be imposed until additional informa-tion is obtained. As more informationbecomes available, the risk assessment isupdated and risk management action isadjusted if deemed necessary. Screening-level testing may be imposed on a newfiber of low health concern if it is to beproduced in large quantities and widelyused in commercial products, which couldresult in substantial human exposure.Testing requirements are considered neces-sary to ensure that the new fiber does notpose significant risk to exposed humans.For fibers determined to be potentially haz-ardous, the agency may issue a significantnew use rule (SNUR) to prevent the fibersfrom reentering commerce without U.S.EPA notification.

Hazard IdenficationKey factors that are generally considered inpredicting the health hazard potential ofinhaled fibers are: a) the ability of the fiberto generate airborne respirable partides; b)the fiber size distribution of the airbornefibers; c) the morphologic and chemicalcharacteristics of the fiber; d) the in vitrosolubiity of the fiber; e) the biopersistenceof the fiber in the lung; f) the ability of thefiber to cause cytoxicity, cell proliferation,chromosomal damage, and other biologicalendpoints in in vitro and in vivo assays;g) the ability of the fiber to cause patholog-ical changes in short term or lifetimestudies in laboratory animals by inhalation,IT, and/or ip injection; and h) the toxicity

and carcinogenicity profiles of chemicallyand structurally related fibrous particles.

In the absence of any toxicologicinformation, a new fibrous material is pre-sumed to pose a fibrogenic hazard if it isrespirable. If the new fiber also contains asignificant proportion of particles with along, thin, fibrous shape (fiber diameter< 1 pm and fiber length 2 10 pm), it isconsidered potentially fibrogenic and car-cinogenic. This science policy position isjustified, as fiber dimension is an impor-tant determinant of fiber toxicity and car-cinogenicity. Health concerns aboutpotential carcinogenic and fibrogeniceffects of a fiber increase if available infor-mation indicates that it is relatively insol-uble in physiological systems andbiopersists in the lung. The concern isheightened if chemically and structurallyrelated fibrous substances are known tocause in vitro and in vivo toxicity in short-term studies and/or carcinogenicity inlong-term studies.A fiber is not considered a significant

hazard to human health if available infor-mation indicates that the fiber is nonres-pirable or has a low degree of respirability.A respirable fiber may be of low healthconcern if the fiber is relatively short, doesnot biopersist in the lung, is relatively solu-ble, exhibits low biological and toxicologiceffects in in vitro and in vivo short-termstudies, and/or demonstrates a lack offibrogenic and carcinogenic effect in long-term animal studies. Some short fibers maybe fibrogenic at high exposure levels butmay not be carcinogenic. On the otherhand, a carcinogenic fiber is also likely tobe fibrogenic.

The minimum data set required by theagency for a screening-level determinationof potential health hazards of a new fibershould include a complete physical andchemical characterization of the fiber andthe results from a well-designed and wellconducted 90-day subchronic inhalationstudy in the rat. Other toxicologic infor-mation such as in vitro solubility, in vitrocellular assays (e.g., cytotoxicity, genotoxi-city, cell proliferation, or generation ofROS), and lifetime IT or ip carcinogenic-ity studies, are considered highly relevantand desirable in the evaluation of thepotential hazard of the fiber. However, atthe present time, these studies are not partof the regulatory testing requirement fornew fibers. The U.S. EPA is in the processof developing fiber toxicity and carcino-genicity testing guidelines for new andexisting fibers (24).

Environmental Health Perspectives * Vol 105, Supplement 5 * September 1997 1 333

VUAND LAI

Dose-Response AssessmentFor new fibers, it is preferable to conductdose-response assessments using data fromwell-conducted lifetime inhalation studiesin laboratory rodents. However, toxico-logic information on new fibers is oftenlacking. Dose-response risk assessments forchronic respiratory noncancer toxicity ofnew fibers are generally based on animaldata from the most closely related fibrousparticles. Cancer risk assessment of newfibers, on the other hand, is based primar-ily on unit risk estimates derived from epi-demiologic studies of asbestos workers.This approach is considered valid only ifthe fibers under evaluation do not differgreatly from those of asbestos in terms offiber morphology, dimensions, and bioper-sistence. However, because most fibers dif-fer widely in their physical and chemicalcharacteristics, this approach may overesti-mate or underestimate the cancer risk ofthe fibers under evaluation. This uncer-tainty is taken into consideration in therisk characterization step of the risk assess-ment process. Therefore, these types ofassessments are considered screening leveland often only qualitative in nature.

Exposure AssessmentThe first step in exposure assessment is theevaluation of the life cycle of the fiber (howthe fiber is manufactured, processed, used,removed, and disposed of) and the identi-fication of any points in the product lifecycle that may result in human exposureby inhalation in the occupational setting,as an end user, or from environmentalreleases. This is followed by the estima-tion of the population, frequency, andmagnitude of exposure for each potentialexposure scenario, and an evaluation ofany alterations in the physical and chemi-cal characteristics of the fiber throughoutits life cycle. This is important becausethe physicochemical properties of thefiber are major determinants of toxicityand carcinogenicity.A major source of uncertainty in the

assessment of anticipated human exposureto new fibers is the absence of monitoringdata on fiber exposure in the workplace,during use, and from environmentalrelease. As a result, compliance with thepermissible exposure limit for respirablenuisance dust (5 mg/m3 as an 8-hr time-weighted average, set by the OccupationalSafety and Health Administration) is gen-erally assumed. Considerable uncertaintiesalso exist for the conversion of gravimetricconcentration into fiber concentration.

Risk CharacterizationRisk characterization includes an integrativeanalysis of hazard, dose response, andexposure assessments to characterize therisk posed by a fiber throughout itsexpected product life cycle. Major resultsof the risk assessment are presented in arisk characterization summary. Thesummary generally includes a) the qualita-tive WOE conclusions as to the likelihoodthat the new fiber may pose a hazard tohuman health; b) a discussion of thedose-response information considered inthe assessment of risk from chronic respira-tory toxicity and carcinogenicity; c) esti-mates of the nature and extent of theexposure, and the number and types ofpeople exposed; d) a conclusion withregard to the nature and extent of the risk;and e) a discussion of the overall confi-dence and uncertainty in the analysis,including the major assumptions made, thescientific judgments employed, and thedegree of conservatism involved.

In the integrated analysis for noncancerrespiratory effects, a margin of exposure(MOE) analysis is derived to determinethe likelihood of a health risk. MOE isdefined as the ratio of the NOAEL dividedby the estimated human exposure of inter-est. When MOE is equal to or greater thanthe product of uncertainty factors (UFs)and a modifying factor (MF), the need forregulatory concern is likely to be small.UFs are used to account for interspeciesvariation in sensitivity and intraspeciesextrapolation, and an MF is used toaccount for the completeness of the database. With regard to cancer risk assessmentof inhaled fibers, an excess lifetime cancerrisk of > 1 in 10,000 is generally consideredof low regulatory concern.

Risk Assessmentfor Existing FibrousSubstances under TSCAFew risk assessments have been conductedto date on existing fibrous substances listedon the TSCA inventory. The scope anddepth of the risk assessments vary depend-ing on the regulatory purposes. Testingaction, significant new use rulemaking,identification of potential candidates forrisk reduction action, or priority settingmay be included.

Assessment in SupportofTesting ActionSection 4 ofTSCA gives the U.S. EPA theauthority to gather information about toxi-city of existing chemicals and the extent to

which humans are exposed to them if theagency determines that insufficient dataexist to evaluate risks to human health orthe environment. Testing action can betriggered if the fiber is produced insubstantial quantities and if one of the fol-lowing applies: a) the fiber enters the envi-ronment in substantial quantities, b) thereis substantial human exposure (number ofpeople exposed), or c) there is significanthuman exposure (magnitude of exposure).Testing action can also be supported ifthere is evidence to indicate that the chem-ical may present an unreasonable risk tohuman health or the environment.

U.S. EPA investigators believe thatthere are sufficient reasons, based on haz-ard information, to suspect possible healtheffects from long-term inhalation exposureto respirable fibers. Therefore, the agencyhas added a respirable fibers category as apriority substance for hazard and exposuretesting on the U.S. EPA's Master TestingList (26).

To date, the agency has taken testingaction on only one class of fibers. Theagency has concluded that RCF is likely tobe carcinogenic and fibrogenic, but expo-sure data are inadequate to determine ifRCF poses an unreasonable risk to workers(27). Consequently, the U.S. EPA and themanufacturers of RCF developed an expo-sure monitoring program pursuant to anenforceable consent order to obtain addi-tional worker data (28). The agency ispresently developing guidelines for chronictoxicity and carcinogenicity testing forfibrous particles (24).

The objective of risk assessment insupport of testing action is to determine ifthe available data indicate a potentialhealth hazard and the extent of humanexposure, and if the information issufficient to perform a quantitative riskassessment to justify risk reduction actions.Thus, risk assessment in support of testingaction generally includes only twocomponents: a hazard assessment and anexposure assessment.

The hazard assessment involves anin-depth review and evaluation of healtheffects data available in the open literatureand/or directed to the U.S. EPA via TSCAsubmissions. The questions to be asked are:a) On the basis of available information,what can be concluded about the carcino-genic and fibrogenic potential of the fiberunder review? b) Are the available dataadequate for hazard characterization anddose-response assessment? The types ofscientific information used in the hazard

Environmental Health Perspectives * Vol 105, Supplement 5 * September 19971 334

HUMAN HEALTH RISK ASSESSMENT OF INHALED FIBERS

evaluation are similar to those discussed inthe section on new fibers. The evaluationincludes a qualitative WOE conclusion asto the likelihood that the fiber may pose ahazard to human health, identification ofdata gap, if any, and recommendations fortests to fill the data gap.

The purpose of the exposure assessmentis to estimate the nature and extent ofhuman exposure to the fiber of interest andto determine whether the available expo-sure information is adequate for assessingpotential health risk. The exposure assess-ment indudes consideration of the productcycle for the fiber, estimates of the size andnature of the populations exposed to thefiber, and the source, magnitude, frequency,and duration of inhalation exposure to thefiber, based on available monitoring ormodeling results. A discussion of the confi-dence and uncertainties of the analysisis included.

Assessment in Supportofa Sigificant New Use RuleThe technical support for a SNUR involvesan in-depth hazard assessment similar tothat used in support of testing actions. TheU.S. EPA has promulgated a SNUR undersection 5(e) ofTSCA for erionite (29) andissued a proposed SNUR for RCF (28) onthe basis that these fibers are hazardous tohuman health, and any use of these fibersmay result in significant human exposure.The SNUR would allow the U.S. EPA to

evaluate the intended new use and, ifnecessary, to prohibit or restrict that activ-ity if such use would pose an unreasonablerisk to human health.

Assesment for Priority SettingThe U.S. EPA is developing a screeningassessment to identify fibers of high con-cern for control action. The assessmententails a review of readily available hazardand exposure data from reliable sources tomake a qualitative judgment regarding thenature and magnitude of possible healthrisk. The scope and depth of the assess-ment is comparable to the assessment fornew fibers. The assessment may be basedon surrogate dose response and exposuredata and on policy assumptions. Outcomesof the review may include withdrawal fromthe review process because of low hazardconcern and/or limited exposure potential,recommendation for additional informa-tion, and a need for more in-depth reviewfor risk reduction action.

Assessment for RiskReduction ActionA comprehensive risk assessment isconsidered necessary to support an unrea-sonable risk finding before any risk reduc-tion action is enacted under section 6 ofTSCA. To date, the U.S. EPA/OPPT hasconducted no comprehensive risk assess-ments on any existing fibers exceptasbestos. A comprehensive assessment of a

fiber such as asbestos entails a criticalanalysis of all relevant hazard, doseresponse, and environmental exposuredata, and an assessment of quantitativeexpression of cancer risk (4).

ConclusionsRisk assessment is the starting point forrisk management consideration and thefoundation for regulatory decisionmaking.Risk assessment approaches used by theU.S. EPA to support regulatory decisionsunder TSCA for naturally occurring andsynthetic fibers differ, depending onwhether the fiber in question is an existingor a new fiber. This is due to differences inthe purpose of the assessment, the avail-ability of data, time, and resources, and theintended nature of regulatory action.

Risk assessments of new and existingfibers are prone to uncertainties becausehealth hazard and human exposureinformation is incomplete for most fibers.Furthermore, how fibers cause diseases andwhat specific determinants are critical tofiber-induced toxicity and carcinogenicityare still not completely understood. Ofnecessity, many default assumptionsbridge both data and knowledge gaps.Further research to improve our know-ledge base in fiber toxicology andadditional toxicity and exposure datagathering are needed to more accuratelycharacterize the health risks from exposureto inhaled fibers.

REFERENCES

1. IPCS. IPCS Environmental Health Criteria, EHC 53: Asbestosand Other Natural Mineral Fibres. Geneva:InternationalProgramme on Chemical Safety, World Health Organization,1986.

2. Stanton MF, Layard M, Tegeris A, Miller E, May M, MorganE, Smith A. Relation of particle dimension to carcinogenicityin amphibole asbestos and other fibrous minerals. J NatiCancer Inst 67:965-975 (1981).

3. Pott F, Roller M, Kamino K, Bellmann B. Significance of dura-bility of mineral fibers for their toxicity and carcinogenicpotency in the abdominal cavity of rats in comparison with thelow sensitivity of inhalation studies. Environ Health Perspect102(Suppl 5):145-150 (1994).

4. U.S. EPA. Airborne Asbestos Health Assessment Update.EPA/600/8-84/003F. Washington:U.S. EnvironmentalProtection Agency, 1986.

5. NRC Committee on the Institutional Means for Assessment ofRisks to Public Health, Commission on Life Sciences. RiskAssessment in the Federal Government: Managing the Process.Washington:National Academy Press, 1983.

6. NRC Committee on Risk Assessment of Hazardous AirPollutants, Commission on Life Sciences. Science andJudgment in Risk Assessment. Washington:National AcademyPress, 1994.

7. U.S. EPA. The Risk Assessment Guidelines of 1986.

EPA/600/8-87/045. Washington:U.S. EnvironmentalProtection Agency, 1986.

8. U.S. Environmental Protection Agency. Proposed guidelinesfor carcinogen risk assessment. Fed Reg 61(79):17960-18011(1996).

9. U.S. Environmental Protection Agency. Guidelines for expo-sure assessment. Fed Reg 57(104):22888-22938 (1992).

10. U.S. EPA. Methods for derivation of inhalation reference concen-trations and application of inhalation dosimetry. EPA/600/8-90/066F. Research Triangle Park, NC:U.S. EnvironmentalProtection Agency, 1994.

11. Barnes DG, Dourson M. Reference dose (RfD): descriptionand use in health risk assessments. Regul Toxicol Pharmacol8:471-486 (1988).

12. Kane, AB. Particle- and fiber-induced lesions: an overview. In:Toxic and Carcinogenic Effects of Solid Particles in RespiratoryTract (Dungorth DL, Mauderly JL, Oberdorster G, eds).International Life Sciences Institute Monographs. Washington:ILSI Press, 1994;3-16.

13. Lippmann M. Biophysical factors affecting fiber toxicity. In:Fiber Toxicology (Warheit DB, ed). San Diego, CA:AcademicPress, 1993;259-303.

14. Davis JMG, Addison J, Bolton RE, Donaldson K, Jones AD,Smith T. The pathogenicity of long versus short fibre samples ofamosite administered to rats by inhalation and intraperitoneal

Environmental Health Perspectives * Vol 105, Supplement 5 * September 1997 1335

VU AND LAI

injection. Br J Exp Pathol 67:473-491 (1986).15. Davis JMG, Jones AD. Comparisons of the pathogenicity of

long and short fibre of chrysotile asbestos in rats. Br J ExpPathol 69:717-737 (1988).

16. Monchaux G, Bignon J, Jaurand MC, LaFuma J, Sebastien P,Masse R, Hirsh A, Goni T. Mesotheliomas in rats followinginoculation and acid-leached chrysotile asbestos and other min-eral fibers. Carcinogenesis 2:229-236 (1981).

17. IARC. General remarks on man-made mineral fibres. In: IARCMonographs on the Evaluation of Carcinogenic Risks toHumans. Vol 43: Man-made Mineral Fibres and Radon.Lyon:International Agency for Research on Cancer,1988;33-35.

18. NIOSH. Criteria for a Recommended Standard. OccupationalExposure to Fibrous Glass. DHEW/NIOSH Publ No 77-152.Cincinnati, OH:National Institute for Occupational Safety andHealth, 1977.

19. Schlesinger RB. Comparative deposition of inhaled aerosols inexperimental animals and humans: a review. J Toxicol EnvironHealth 15:197 (1985).

20. Timbrell V. Deposition and retention of fibers in the humanlung. In: Inhaled Particles (Walton H, ed). Oxford, U.K.:Permagon, 1982; 347-369.

21. Warheit DB. Interspecies comparisons of lung responses toinhaled particles and gases. Crit Rev Toxicol 20: 1-29 (1989).

22. Yu CP, Zhang L, Oberdorster G, Mast RW, Glass LR, Utell

MJ. Clearance of refractory ceramic fibers (RCF) from the ratlung: development of a model. Environ Res 65:243-253(1994).

23. McClellan RO, Miller FJ, Hesterberg TW, Warheit DB, BunnWB, Kane AB, Lippmann M, Mast RW, McConnell EE,Reinhardt CF. Approaches to evaluating the toxicity and car-cinogenicity of man-made mineral fibers: summary of a work-shop held November 11-13, 1991, Durham, North Carolina.Regul Toxicol Pharmacol 16:321-364 (1992).

24. U.S. EPA. Workshop Report on Chronic Inhalation Toxicityand Carcinogenicity Testing of Respirable Fibrous Particles.EPA-748-R-96-001. Washington:U.S. EnvironmentalProtection Agency, 1996.

25. Barrett CJ. Cellular and molecular mechanisms of asbestos car-cinogenicity: implications for biopersistence. Environ HealthPerspect 102(Suppl 5):19-23 (1994).

26. U.S. Environmental Protection Agency. Master testing list;notice of availability. Fed Reg 57:61240 (1992).

27. U.S. Environmental Protection Agency. Testing consent orderfor refractory ceramic fibers. Fed Reg 58:28517-28520 (1993).

28. U.S. Environmental Protection Agency. Refractory ceramicfiber; significant use of a chemical substance. Fed Reg59:13294 (1994).

29. U.S. Environmental Protection Agency. Erionite fiber; signifi-cant new use of a chemical substance. Fed Reg 56:56470-56472(1991).

1336 Environmental Health Perspectives a Vol 105, Supplement 5 - September 1997