103x Filetype PDF File size 0.14 MB Source: ftp.cdc.gov
FEATURE Anoverview of setting occupational exposure limits (OELs) for pharmaceuticals Setting appropriate occupational exposure limits is an integral component in assuring the health and safety of workers ByRobert H. Ku Since adhering to OELs is consid- testing for the toxicity of chemicals ered an effective and proven way to became more common and more epi- INTRODUCTION protect workers from the deleterious demiologic studies were done in ccupational exposure limits health effects caused by chemicals, workplaces, the approach used to set 1 many pharmaceutical companies have OELswasbasedonthe“no-observed- (OELs) for the protection of opted to determine OELs for their drug effect-level/safety factor” (NOEL/SF) Oworkershavebeenaroundat substances for internal use. approach.1 In this approach, all of the least since 1939 when the National OELs, if appropriately determined pertinent animal and humanstudiesare (later changed to American) Confer- and periodically monitored in the reviewed and the highest dose that did ence of Governmental Industrial Hy- workplace air, would offer a first line not cause an effect in the most sensitive gienists (ACGIH) published its inau- indication of whether exposures are health end point (the NOEL) is identi- gural acceptable workplace exposure acceptable or not. The challenge to fied.2 Once a NOEL has been identi- limits (now known as threshold limit toxicologists and other health profes- fied, a set of uncertainty (or safety) fac- values, or TLVs). In 1970, the U. S. sionals involved in setting OELs is to tors are applied to this value to Occupational Safety and Health Act determine a value that has an ade- accommodate limitations in the data incorporated by reference the 1968 quatemarginbetweenalevelthatpro- and to try to assume that workers are ACGIH TLVs as enforceable limits. duces undesired health effects and protected. The Occupational Safety and Health one that does not. This is the safety Thenumberandmagnitudeofthese Administration (OSHA) refers to margin. If an OEL is set with an inor- safety factors depend on the quality of these enforceable limits as permissible dinatelylargesafetymargin,thenvalu- the data. In general, some of these exposure limits (PELs). able resources may be expended un- safety factors may include: (1) a factor Most TLVs and PELs are for com- necessarily in the form of extensive from 1 to 10 for animal-to-human (in- monly used industrial chemicals engineering containment equipment terspecies) extrapolation (if the NOEL where a large number of workers po- or overly protective personal protec- is based on animal data), (2) a factor tentially may be exposed. Very few tive equipment. If an OEL is set too from 1 to 10 for human-to-human in- pharmaceuticals fit this description, high, then employee health may be traspecies variability in response, (3) a hence, very few pharmaceuticals are compromised. factor from 1 to 10 to consider study on the ACGIH list of TLVs or the This article provides an overview of duration (a long-term study being OSHAlist of PELs. howOELshavebeendeterminedhis- morehelpfulthanashort-termstudy), torically, new approaches that are on (4) a factor to consider the persistence Robert H. Ku, Ph.D., D.A.B.T., the horizon, and approaches that are of the drug in the body (or elimination C.I.H., is a Principal Toxicologist unique to pharmaceuticals. half-life), and (5) a factor to accom- with SafeBridge Consultants, Inc., modate for absorption efficiency by Mountain View, California (e-mail: different routes of exposure.2–4 robert.ku@safebridge.com). APPROACHESTOSETTINGOELS If a NOEL is not available, then a SafeBridge has expertise in The historical approach used to set lowest-observed-effect-level (LOEL) occupational and environmental OELs was based on human experi- can be used The LOEL is the lowest toxicology, industrial hygiene, and enceintheworkplace.Ifairbornelev- level that causes an effect in the most 3 analytical chemistry, and els of a chemical were causing adverse sensitive end point. A safety factor specializes in providing these health effects, then these levels were from 1 to 10 may be considered for services to the pharmaceutical/ reducedtoalevelthatdidnotproduce extrapolating a LOEL to a NOEL. biotech industries (Web site: adverse health effects. In the latter Whetherthesafetyfactoriscloser to 1 safebridge.com). half of the 1900s, as laboratory animal or closer to 10 depends in part on the 34 ©Division of Chemical Health and Safety of the American Chemical Society 1074-9098/00/$20.00 Published by Elsevier Science Inc. PII S1074-9098(99)00070-2 severity of the end point. For example, LINEARIZED MULTISTAGE (LMS) LMS or any other mathematical if the adverse health effect is consid- MODELFORCARCINOGENS model for carcinogenic drugs when ered serious (e.g., severe birth defect) WhentheNOEL/SFapproachwasin- setting OELs. It has preferred to con- then a factor closer to 10 may be more troduced, it was applied to all chemi- tinue to use the NOEL/SF approach appropriate; if the adverse health ef- cals. This approach assumed that if for all chemicals and simply consid- fect is considered minor (e.g., de- exposure to a chemical was kept be- ered cancer as a severe adverse health creased body weight) then a factor lowsomelevel(e.g.,athresholddose), effect. Alternatively, some companies closer to 1 may be more appropriate. noadverse effects would occur. How- have adopted approaches to manage The NOEL or LOEL is dependent ever, as more information became worker exposure by recommending on the number of test subjects used available on the biologic mechanism specific personal protective equip- and the doses selected in the studies. of action of carcinogens (vs noncar- ment and requiring handling proce- Thegreaterthenumberoftestsubjects cinogens), there was a concern that a dures be followed. This approach is pergroup,thegreaterthelikelihoodof cancer risk may exist at any level of based on the job assignment and po- observing an effect. The number of exposure. In other words, contrary to tential for exposure (e.g., perfor- doses used per study is frequently lim- noncarcinogens, where a threshold mance-based exposure control levels) ited, to 3 or 4 dose levels. If the doses dose was believed to exist, there was in lieu of setting OELs. are properly chosen, one of the doses nodoseofacarcinogenthat was con- will turn out to represent the NOEL, sidered without risk (i.e., no threshold andanother,theLOEL.However,the dose). Because testing at very low BENCHMARKDOSEAPPROACH true NOEL or LOEL will probably be doses and with a large number of an- Recently, a new approach has been somewhere in between these two imals was not possible or practicable, proposed for setting acceptable expo- tested dose levels. If the gap between mathematical models were developed sure limits called the benchmark dose these two tested dose levels is large, to try to predict the hypothetical in- 8 cremental cancer rate at these very approach. For noncarcinogens, this then there could be greater uncer- low doses. The model that was ulti- approach attempts to identify a dose tainty as to what the true NOEL or mately adopted by government agen- that corresponds to a measurable re- LOELmaybe. cies was the linearized multistage sponse rate (e.g., dose that results in a An equation that is used to deter- 10%responserate,orED - (LMS) model. This model essentially 10). This des mine an OEL for a chemical can be resulted in a linear dose-response re- ignated response rate also has been 5 represented as follows : lationship at very low doses while ac- termed the “point of departure.” In es- OEL5@~NOEL! comodating for the observable dose- sence, the point of departure replaces response relationship at the high the NOEL or LOEL. The point of de- 3~humanbodyweight!#/ doses tested. All that was needed now parture is preferred over the NOEL or @~safety factor!n wasforgovernmentagenciestodecide LOEL in that it represents a dose that 3~humanbreathingrate!# what the acceptable cancer risk corresponds to a specific and measur- NOELusually is in units of milligram shouldbe.Fortheprotectionofpublic able response rate. of chemical administered/kilogram of health, the U.S. Environmental Pro- Thebenchmarkdoseapproachthen animal body weight/day. tection Agency has set the acceptable requires an estimation of exposure in 25 the most heavily exposed individuals. Humanbody weight typically is as- cancer risk at around 1 3 10 to 1 3 1026 (one in one hundred thousand Theratioofthedosecorrespondingto sumed to be 70 kiligrams for an adult to one in one million) for potential the point of departure and exposure male. exposure of individuals to chemical estimate in the most heavily exposed Safety factors are numeric values for carcinogens in the environment. The individuals becomes an indicator of accomodating limitations in the data, dose that corresponded to the ac- acceptability and has been termed the as described above. ceptable cancer risk based on the marginofexposure(MOE).Thelarger Breathing rate in workers typically LMS model became the acceptable the MOE, the less likely that an ad- is assumed to be 10 m3 - /8-hour work exposure limit. verse effect will occur. There is still day. Some researchers found that debate over the upper limit of the Hereisanexampleofthisapproach ACGIHsTLVsforcarcinogens(deter- MOE(e.g., 10, 100, 1000) to be con- for the synthetic estrogen, ethinyl es- mined by approaches other than the sidered acceptable. tradiol. The NOEL in humans has LMS approach) corresponded to ap- One criticism of the LMS model is been estimated to be around 3.5 mg/ 23 that all carcinogens are treated the proximately a 1 3 10 cancer risk day. (Because it is reported in these 6 sameway.Butasmorebecameknown using the LMS model. Others have units, there is no need to multiply by the inferred that this cancer risk repre- about the mechanism of action of car- human body weight). If one assumes a sented OSHA’sacceptablecancerrisk cinogens, it appears that there may be safety factor of 10 for the human vari- policy and believe this is the appropri- at least two general groups, threshold ability in response and a breathing rate 7 dose and nonthreshold-dose carcino- ate cancer risk level to use for workers. 10m3 /8-hourworkday,thentheOELis The pharmaceutical industry with gens. Under the benchmark dose ap- 3 estimated to be 0.035 mg/m . some exceptions has not adopted the proach, threshold-dose carcinogens Chemical Health & Safety, January/February 2000 35 are treated similarly to non-carcino- tionally and for which an OEL is in- the drugs are related to its pharmaco- gens. tended? If not, then one must con- logical mechanism of action. As the For nonthreshold-dose carcino- sider the possible difference in saying goes, “The dose makes the poi- gens, a straight line is drawn from the bioavailability. For example, alendro- son.” What may be therapeutic at a point of departure to the origin. The nate is a bisphosphonate drug ap- low dose may cause serious toxici- acceptable exposure limit is deter- proved for osteoporosis and for other ty at a high dose. For drugs where mined for any acceptable response bonediseases.Itisadministeredorally the mechanism of toxicity is related 26 and has a bioavailability of less than to exaggerated pharmacology, then rate (e.g., 1 3 10 ) belowthepointof departure. Thus, the benchmark ap- 1% orally. If this drug is readily ab- by protecting a worker against the proach attempts to address a major sorbed through the respiratory tract, “normal” pharmacologic effects of shortcoming of the LMS model, by thenthe100-foldsafetyfactormaynot the drug also will also protect treating threshold- and non-thresh- be adequately protective. Unfortu- against the “exaggerated” or adverse old-carcinogens differently. nately, it is often the case that for effects of the drug. drugs that are not intended for admin- There are other drugs where toxic istration by the inhalation route, no effects are unrelated to the mechanism APPROACHESSPECIFICTO inhalation bioavailability data may be of pharmacologic action. For in- PHARMACEUTICALS available. To be conservative (i.e., stance, a drug may cause developmen- Pharmaceuticals are chemicals. How- most health protective), one may as- tal toxicity (i.e., birth defects) or can- ever, they differ from most industrial sume that the inhalation bioavailabil- cer by a mechanism unrelated to its chemicals because there often is a ity is complete (i.e., 100%). Then, if pharmacologic effects. In these cases, wealth of human data available on the oral bioavailability of the drug is using the 100-fold safety factor on the pharmaceuticals. In setting OELs for 1%, an additional safety factor of 100 therapeuticdosemaystillbeappropri- certain pharmaceuticals, some have may be needed to accommodate for ate, but would need to be used with proposed simply to divide the thera- this difference in bioavailability by greater caution. If a promising drug peutic dose by a safety factor. For cer- these two routes. If the oral bioavail- causesbirthdefectsnearthetherapeu- tain pharmaceuticals that mimic some ability is closer to 100% (or closer to tic dose range, then its likelihood of biologic activity of a chemical pro- the inhalation bioavailability), then a ever reaching the marketplace (and duced by the body, some have pro- smaller bioavailability adjustment fac- thus manufactured in large quantities) posed limiting exposure to a fraction tor may be considered. is small. However, teratogenic drugs of this endogenous biologic activity. A second issue to note when con- couldmakeitintothemarketplaceif Both of these approaches actually are sidering applying the 100-fold safety thedrugisintendedforolderwomen similar to the NOEL/SF approach de- factor to the lowest recommended (i.e., those past menopause) or for scribed above. therapeutic dose is the intended use of men. the drug. If a drug is intended to be In summary, applying the 100-fold usedinlife-threateningdiseaseswhere safety factor to the lowest therapeutic THERAPEUTICDOSE/SAFETY significant toxicities may be consid- dose is simple to do. However, there FACTORAPPROACH ered acceptable, the 100-fold safety are a number of issues that must be In the therapeutic dose/safety factor factor may not be adequately protec- considered to determine whether approach, the lowest recommended tive. Less than life-threatening side ef- this method is appropriate. These in- therapeutic dose of the drug is identi- fects may be acceptable to the patient clude: fied. This therapeutic dose then is di- when trying to save the patient’s life, videdbyasafetyfactor.Asafetyfactor but these same side effects are not ac- 1. the difference in bioavailability be- of 100 usually is suggested because it ceptable to healthy workers. Clearly, tweentherouteofdrugadministra- can be thought of as a factor of 10 for the 100-fold safety factor approach for tion and the inhalation route, the adjusting a therapeutic effective dose ananticancerdrugwhereheroicdoses route by which workers are ex- to a therapeutically noneffective dose may be needed to save the patient of- posed; (somewhat similar to adjusting a fers less protection than for an anti- 2. the intended use of the drug or LOELtoaNOEL)andafactor of 10 cough drug where the effects of the risk-benefit considerations (i.e., to accommodate for individual vari- disease are not life-threatening and whether it is for serious or life- ability in response. may merely be a temporary nuisance. threateningsituations or for less se- For many drugs, this approach may Whatever side effects are produced rious diseases); and be reasonable and has produced from an anticough medication ought 3. possible differences in the mecha- OELs similar to the traditional not to cause any side effects more un- nism of action for the therapeutic NOEL/SFapproachdescribed above. comfortable than the cough itself. In effects versus the toxic effects. This approach cannot be used indis- the case of anticough medication, the criminately, however, for several rea- 100-fold safety factor approach may Using ethinyl estradiol as an example sons. First, is the drug administered be overly conservative. here, the lowest therapeutic dose is throughtherespiratorytract,theroute Another issue to consider is to de- around 20 mg/day. Dividing by the towhichworkersareexposedoccupa- termine whether the toxic effects of 100-fold safety factor and the breath- 36 Chemical Health & Safety, January/February 2000 3 timated by several approaches, one ing rate of 10 m /8-hour workday, the production rate in women is higher 3 may need to decide the most appro- estimated OEL is 0.02 mg/m . than the production rate in men.) The relative estrogenic potency of ethinyl priate approach (i.e., the approach on estradiol to 17-beta estradiol is around which the underlying data are most INCREMENTALINCREASEOF 2. Assuming a 1% incremental in- solid) especially if the approaches re- ENDOGENOUSBIOLOGICAL crease in biologic activity from ethinyl sult in OELs that vary significantly ACTIVITY estradiol is deemed acceptable and (e.g., by orders of magnitude). For the More and more frequently nowadays, 3 ethinyl estradiol example, the OEL es- drugs are being developed to replace using a breathing rate of 10 m /8- or mimic the function of a substance hour workday, the estimated OEL timated by the three approaches pre- for ethinyl estradiol is 0.035 mg/m3. sented above varied by less than a fac- naturally produced by the body. If the tor of two. When they vary by such a levels of these naturally occurring sub- small amount, the values essentially stances are decreased, the risk of cer- SUMMARY can be considered to be equivalent. tain diseases may increase. By supple- Setting appropriate OELs is an inte- It should be kept in mind that the menting the levels of these naturally gral component in assuring the health determination of an OEL for a phar- occurring substances with an exoge- and safety of workers. The approach maceutical agent or for any other nous source, these diseases may be proposed historically and still most chemical entity is an inexact science. ameliorated or eliminated. This is commonly used is the NOEL/SF ap- Each approach has its own set of as- the basis for hormone replacement proach. In this approach, a NOEL for sumptions and limitations. It often therapy (HRT) in postmenopausal the most sensitive effect is identified comesdowntoprofessionaljudgment women.Inthese women, the levels of and modified by safety factors to ac- as to which approach is preferred and naturally occurring estrogens are re- commodateforuncertainties and data how large the uncertainty factor duced subjecting them to an elevated gaps. This approach presumes that should be. It is important for anyone risk of certain types of diseases, e.g., both carcinogenic and noncarcino- whorelies on OELstounderstandthe osteoporosis. For individuals who do genic effects do not occur if exposure assumptions used in deriving the not have a reduced level of estrogens, is kept below the NOEL. As some ev- OEL. The real test for finding out additional exposure may be undesired idence suggested that a finite risk for whether an OEL has been appropri- and need to be controlled. cancer may exist at any level of expo- ately set comes from actual workplace An approach that can be used for sure to a carcinogen, the LMS model experience. The responsibility of toxi- determining an OEL for these types of was developed and adopted for car- cologists and other health professionals drugs is to limit exposure to a level cinogens. This mathematical model tasked with setting OELs is to estimate where exposure contributes a small predicted the level of carcinogenic a value that is protective of workers yet fraction, like 1%, of one’s endogenous risk for any level of exposure. Once an without being so overly protective that biological activity of that substance. acceptable risk level is specified, the resources are unnecessarily spent. For instance, if the daily production acceptable exposure level could be rate of 17-beta estradiol in the body is determined. As further information References X mg/day, then the acceptable expo- about the mechanism of action of car- 1. Lehman, A. J.; Fitzhugh, O. G. Assoc. sure could be set at 1% of that rate, or cinogens became available, it became Food Drug Off. U. S. Q. Bull. 1954, 0.01X mg/day. If the drug is not iden- apparent that NOELs may indeed ex- 18(1), 33. tical to the endogenous substance, ist for some carcinogens. This resulted 2. Dourson, M. L.; Stara, J. F. Regulat. then the relative potencies of the drug in the development of the benchmark Toxicol. Pharmacol. 1983, 3, 224, and the endogenous substance also dose approach. One of the major ad- 3. Baird, S. J. S.; Cohen, J. T.; Graham, must be determined to be able to limit vantages of the benchmark dose ap- J. D.; Shlyakhter, A. I.; Evans, J. S. exposure to a 1% increase in biologic proachistoallowforgreaterflexibility Hum.Ecol.RiskAssess.1996,2(1),79. activity. That is, if the limit is set at 1% in the determination of acceptable ex- 4. Dourson, M. L.; Felter, S. P.; Robinson, D. Regulat. Toxicol. Pharmacol. 1996, of the endogenous biologic activity and posure limits of carcinogens depending 24, 108–120. the drug is 10 times more potent than onwhatisknownaboutthemechanism 5. Sargent, E. V.;Kirk, G. D. Amer. Ind. the endogenous substance, then expo- of action. For pharmaceuticals, in addi- Hyg. Assoc. J. 1988, 49(6), 309. suretothedrugmustbelimitedto0.1% tion to the approaches described above, 6. Rodricks, J. V.; Brett, S. M.; Wrenn, of that of the endogenous substance. several other approaches have been G. C. Regulat. Toxicol. Pharmacol. Again, using ethinyl estradiol as an used to set OELs. For instance, the 1987, 7, 307. illustration, the production rate of en- use of the therapeutic dose and the 7. Herr, M. L. Amer. Ind. Hyg. Assoc. J. dogenous estrogen, 17-beta estradiol, use of an incremental increase in 1998, 59, 83. has been reported in men to be around some level of endogenous biologic 8. Benchmark Dose Technical guidance 70 mg/day. (Using the production rate activity have been considered as ap- Document: 1996; U.S. Environmental ProtectionAgency.RiskAssessmentFo- inmenismoreprotectivethanusingthe proaches for OEL determination. rum. Washington, D.C.; 1996; EPA/ production rate in women because the In the case where OELs can be es- 600/P-96/002A. Chemical Health & Safety, January/February 2000 37
no reviews yet
Please Login to review.