Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
CCase Studies on Strategies for Testing the Toxicity of Complex Mixtures We present here several case studies that involve testing of the toxicity of complex mixtures. A data base on each of these mixtures existed before any formalized system of testing or strategies, and it is instructive to look at the cases with the aid of the strategies presented in Chapter 3. How was the prob- lem defined? Were the data developed along predetermined lines? Was a strat- egy evident? Was it substantially different from that proposed in this report? Was it successful? Might another strategy have been of more value? The cases chosen represent a range of mixtures and a range of toxic end points. The end points were carcinogenesis and mutagenesis caused by ciga- rette smoke and fractions, systemic toxicity of fire atmospheres, and neuropa- thy associated with hydrocarbon and oxygenated solvents. The reader is encouraged to review these cases in the light of the strategies used that is, with respect to their success and to whether any surprises re- sulted from asking the wrong questions (as in the case of species specificity of sedation by thalidomide) or from failure to ask any questions at all (as in the case of unplanned release of methyl isocyanate from bulk storage). One further purpose of these case studies is to show that complex mixtures can and should be addressed by looking at basic and simple elements. Rather than being concerned with finding the perfect approach to a problem, we should look for feasible approaches. Complex mixtures can be thought of as representing laboratory questions in terms of composition and exposure. CIGARETTE-SMOKE TOXICITY Careful observation of the effects of cigarette-smoke inhalation in humans teaches many lessons. Among the most striking observations are the following: 168
APPENDIX C 169 · The large number of agents in cigarette smoke (more than 3,6001. · The development of both chronic and acute effects most unusual is the unpredictable and diverse character of the chronic lesions, which cannot be foretold from the nature of the acute injury. · The variety of organ systems that appear to be affected. · The synergistic interactions that can occur between cigarette smoke and environmental or occupational agents and accelerate or alter pathologic change. More than 3,600 components are generated in the distillation zone distal to the combustion zone. Mainstream smoke is drawn into the smoker's mouth from the burning cigarette; sidestream smoke arises from the burning end and is released into the environment. The composition of mainstream and sides- tream smoke varies with the smoking conditions. Standard cigarette-smoking conditions are created to produce an average for a male smoker of a nonfiltered cigarette. For filtered cigarettes, the average pufflasts 1.94-2.06 seconds and is repeated every 26.9-30.0 seconds, and the puff volume is 35.9-47.8 ml. For cigar-smoking, the standard conditions used are a 1 .5-second puff every 40 seconds with a volume of 20 ml and a butt length of 33 mm. For pipe-smokers, the standardization is less rigorous, but 2-second puffs every 18 seconds with a volume of 50 ml have often been used. Only 30% of the mainstream-smoke weight originates in the tobacco; the remaining weight is from the air drawn in with the smoke. Of the effluent of a nonfiltered cigarette, 5-9 % by weight is moist particulate matter, 55-65 % is nitrogen gas, 8-14% is oxygen, and the remainder is other gas-phase material. Undiluted cigarette smoke at the mouth contains 5 x 109 particles/ml in a size range of 0.2-1.0 mm (median particle size, 0.4 ,um). For filtered cigarettes, the particle size is 0.35-0.4 ,um; for cigarettes with perforated filter sips, the particle num- ber is substantially lower. The smoke particles are charged with 10~2 electrons per gram of smoke and have reducing activity when freshly inhaled. The gaseous phase of tobacco smoke does not induce malignant tumors of the respiratory tract in laboratory animals. That observation suggests that the carcinogenic activity of smoke requires the particulate phase. Benign and ma- lignant tumors have been induced with tobacco tar in skin and ear of rabbits, by intratracheal instillation in rats, and by topical application on mouse skin. Ex- perimental studies with classically defined initiators and promoters have dem- onstrated that the effect of a tumor initiator is irreversible, but promoters act only during the treatment period. The mouse-skin assay has been used to estab- lish a dose-response relationship for the induction of tumors by smoke parti- cles. Smoke particles have been fractionated into neutral, acidic, basic, and insol- uble fractions. Fractionation has established that the tar subfractions contain the bulk of the polynuclear aromatic hydrocarbons (PAHs) and are the only
170 APPENDIX C substances that produce carcinomas. The subfractions of PAH contain neutral cocarcinogens that potentiate the carcinogenic PAHs. The weakly acidic frac- tion has been shown to contain promoters, as well as cocarcinogens; these include phenolic compounds and catechols. Recent studies (reviewed in IARC, 1986) have demonstrated that tobacco smoke has transplacental carcinogenic effects in hamsters. Compounds that require metabolic activation to an active carcinogenic form such as N-nitro- samine, benzoLa~pyrene, o-toluidine, ethylcarbamate, and vinyl chloride act as transplacental carcinogens. Nitrosamine can also be formed in the fetus from unmetabolized nicotine. More than 90% of the total weight of mainstream smoke is made up of vapor-phase components (i.e., material of which more than 50% passes through a Cambridge glass filter). Nonfiltered and conventional filtered ciga- rettes contain CO at 3-5 vol % per puff and 13-26,ug per cigarette. Cigar smoke contains CO at up to 11 vol %. Concentrations of many vapor-phase components in cigarette smoke vale directly with concentrations of tar and nicotine; that is not the case with nitrogen oxides (NOX), most of which (95 %) is nitric oxide. The NOX content of an average American cigarette is 270-280 ,ug per cigarette. Ciliotoxic agents inhibit lung clearance. Many of those agents are vapor- phase components of cigarette smoke; they include hydrocarbons (280-550 fig per cigarette), NO2 (0-30 fig per cigarette), and formaldehyde (20-90,`4g per cigarette). Hydrazines are effective carcinogens. Vinyl chloride is acted on by the cytochrome P~50 enzyme system to form a halogenated epoxide that can yield halogenated aldehydes or alcohols. The activated forms of vinyl chloride can react with adenosines or cytidines to form new rings that interfere with base . . palnngs. Cigarette-smokers have an increased risk of cancer in organs other than the lungs, such as the esophagus, the pancreas, and the urinary bladder. Cigarette smoke does not contact organs directly; therefore, mechanisms other than di- rect contact must be involved in carcinogenesis. The increased risk might be due to the formation of organ-specific carcinogens or to a shift in metabolism of common components toward localized formation of carcinogenic metabolites. Examples of organ-specific carcinogens are 2-naphthylamine, 4-amino- biphenyl, and benzidine. Volatile nitrosamines are present in tobacco and to- bacco smoke; nitrosamines are also derived from residues of agricultural chemicals on tobacco. The Ames bacterial assay has been used to identify and characterize muta- gens in tobacco smoke. In that assay, mutagen content of smoke can be ob- tained from the equivalent of 0.01 cigarette, whereas only 0.0025 cigarette is needed for the sister-chromatic-exchange study. Substances with mutagenic potential are generally carcinogens as well. Activated carcinogens or mutagens
APPENDIX C 171 form covalent adducts with cellular macromolecules. Those binding to DNA form bulky adducts that produce errors in DNA replication and promote muta- tion. The urine of cigarette-smokers contains 5-10 times the quantity of muta- gens found in nonsmokers. Mutagen concentrations are highest in the evening and lowest in the morning. There appears to be no threshold in the dose- response relationship of smoking and lung-cancer incidence. That observation makes the incidental passive exposure of nonsmokers a potential mechanism of importance in the development of lung cancer in nonsmokers (IARC, 1986; NRC, 19861. Sidestream smoke contains larger quantities of N-nitrosamines and small particles. Cotinine in urine is a marker of cigarette-smoke exposure, and infants with care givers who smoke have been shown to excrete this mate- rial, which is evidence that the infants have been exposed to cigarette smoke. The toxic components in cigarettes include simple diatomic gases that can have direct and long-term toxic effects, as well as compounds that are incom- plete carcinogens and cocarcinogens and can produce their effects only after long periods of exposure (several decades). If we were faced with the question of finding the toxic component in a sample of cigarette smoke, without knowl- edge of how it had been generated or of the composition of the original mate- rial, a study of its "toxic component" might be extremely complex. Even after decades of serious investigation, we do not understand the role of tobacco- smoke components in producing chronic diseases, such as arteriosclerosis, emphysema, and malignant neoplasms. The task of identifying the toxic com- ponents is overwhelming and must be considered currently impossible. We can, however, identify groups of agents from a knowledge of their chemical similarity to agents generated in a standard control substance. We can learn something about the potency of tobacco smoke by comparing the composition of tobaccos grown in different soils, after different periods of aging, after the use of various fertilizer additives, and after treatment with different methods, such as nicotine and tar extraction. The potency of each class of agents must be compared separately. The acute effects, chronic ef- fects, and carcinogenic effects must be separately analyzed and compared with those of standard or control compounds in each group. The toxicologic evalua- tion of tobacco-smoke components might be both useful and confusinguseful as a means of understanding the effects of individual components, and confus- ing because the end product of analysis might not represent the active transient agent that causes an effect. To be hazardous, cigarette smoke must be inhaled over long periods. How- ever, acute effects can be produced in persons with hyperactive airways, such as persons with allergic asthma or bronchitis, or in persons with lung or heart diseases. Thus, the host conditions are important in determining risk. Recent work has demonstrated that not only the smoker, but also persons close by, can be affected by the smoke.
1;'2 APPENDIX C FIRE ATMOSPHERES The thermal-degradation products that evolve when materials pyrolyze, smolder, or bum provide a prime example of complex mixtures known to inca- pacitate, injure, and cause death. The assessment of the toxicity of these com- plex mixtures has to begin with the determination of what is known about them. For example: · All fire atmospheres are toxic. · The main route of entry into the body is inhalation, although dermal ab- sorption can play a role. · Acute exposures (i.e., single exposures of less than 30 minutes) are the principal concerns (except for fire fighters, who receive chronic exposures). · In most situations, the exposure concentrations are high; that is, one does not need to extrapolate results from high experimental doses to low doses that are considered to represent normal population exposures. What is unknown stems from the fact that no two fires are alike. They differ in type and size of burning compartment, burning materials, and ventilation conditions. Each fire is a dynamic phenomenon that affects and is affected by the rapidly changing environmental conditions. Most fires involve multiple materials, each of which might produce hundreds of combustion products (Levin, 19861. The magnitude of the problem becomes apparent with the rec- ognition that the toxicities of most of the products identified are unknown and that the number of toxic product combinations that could lead to possible inter- . . . actions Is Immense. Thus, posing the correct question to determine the best testing strategy is important. The question that initially concerned fire investigators was: Do any materials produce unusually toxic or extremely toxic thermal degradation products? ("Unusually toxic" implies that unexpected adverse effects are pro- duced; "extremely toxic" implies that relatively small amounts of material produce measurable toxicity.) The testing strategy to answer these questions is obviously in the category of comparative toxic potency and is "effect-driven." The concern is not why a material produces an effect, but rather whether it produces the effect, how it ranks toxicologically in comparison with other materials that are suitable for the uses in question, and whether it produces an expected or unexpected effect. In response to those questions, several test methods were designed (Alarie and Anderson, 1979; Crane et al., 1977; Hilado, 1976; Klimisch et al., 1980; Levin et al., 1982; Yusa, 19851. Although different, these methods have some similarities. For example, all include the testing of complete samples and mea- surement of the acute inhalation toxicity in experimental animals. Complete chemical analysis of combustion products was not considered the proper ap- proach, because toxic components might be missed, either through lack of
APPENDIX C 173 detection or through lack of knowledge of their toxicity. Animals, however, would breathe a combination of the gases and respond to its toxicity, regardless of whether it was due to one component or the interaction of several compo- nents. To avoid false-negative results, the tests must be designed to simulate real fire conditions or worst-case conditions. That approach provides a mea- sure of the relative intrinsic toxicity of the atmospheres produced by materials that are thermally decomposed under the specified conditions of each test method. It can be used by material manufacturers or product developers who need to choose materials to market. The comparative toxic-potency approach, however, is scientifically unsatis- factory, because it fails to provide any clues as to which components are re- sponsible for the toxicity and makes it nearly impossible to predict toxicity of future samples. The testing strategy to determine the causative agents in such mixtures therefore involves extensive chemical analysis. The analysis must be preceded by toxicity testing of the complete mixture to indicate whether there is a cause for concern. The effect-driven chemical analysis was used to determine that a fire-retarded laboratory formulation of rigid polyurethane foam pro- duced a highly toxic bicyclic phosphate ester when thermally decomposed (Petajan et al., 19751. Thus, the testing strategy used to determine the toxicity of the thermal- degradation products of materials has historically consisted of tests of compar- ative toxic potency combined with effect-driven chemical analysis. That ap- proach, however, has raised even more questions. For example, do materials with minor differences between formulations or dye lots, or with minor changes in additives or fire retardants, and so on, all need to be tested? If the components of materials have been tested, do composites made of the compo- nents need to be tested? What additional hazard is generated when a new mate- rial is added to a room full of furniture? Toxicity is only one of many kinds of information needed for a hazard assessment, and it is the hazard that must be considered when one is making decisions on the suitability of material for a particular use. Other information needed for a hazard assessment includes the quantity of material present, its configuration, the proximity of other combusti- ble materials, the volume of the compartments to which combustion products can spread, ventilation conditions, ignition and combustion properties of mate- rials present, the presence of ignition sources, the presence of fire protection systems, and the building occupancy. Determination of toxicity for a hazard assessment requires a testing strategy different from that needed for the relative ranking of materials. One needs to develop a model to predict toxicity while avoiding expensive and time- consuming multiple tests. An approach that examines the toxicity of both the primary component (individually and combined) and the total mixture has been used (Levin et al., in press). To use that single- and mixed-component ap- proach, toxicity data on the primary toxic gases present in most fires were
1 ~ j 1 6 14 - 12 x 1 o Z 8 o m 6 4 2 o . ~ No Deaths 1 ~ I ~ 1 1 000 174 APPENDIX C needed. Research began on a few of these gases (CO, CO2, and HCN) and on the effect of oxygen concentration to determine their individual toxicities and the occurrence of interactions. It was demonstrated in rats exposed to CO and HCN for 30 minutes at various concentrations that these two gases act in an additive fashion, such that some animals will die if the sum of [CO]/LC50CO and tHCN]/LCsoHCN is approximately equal to 1 + 15 % . If the sum is greater or less than that value, all the animals should die or live, respectively. It has also been shown in 30-minute exposures that CO2 acts synergistically with CO; in the presence of approximately 5 % CO2, the toxicity of CO (as indicated by lethality) doubles (Figure C-1) (Levin et al., 1987~. Relationships among CO, CO2, and HCN were also determined, as was the effect of O2 deprivation. (02 deprivation is probably more of a problem in the room of fire origin than at any distance away from the fire.) The experimental design was based on a matrix that indicated what concentrations and combinations of gases were lethal (Fig- ure C-21. With that matrix, a testing strategy that uses single and combined primary gases can be applied to predict toxicity as follows. First, a material is 1° i ' I CO Deaths / Co & CO2 I Deaths I l /o ~~ 0 ~ 0 0 \ o 1 ^\ 1 /\ ,0^ 51 o ~~ odd 2000 3000 4000 CARBON MONOXIDE (ppm) 5000 6000 FIGURE C-1 Deaths resulting from exposure to CO or CO2 alone and from CO plus CO2. No deaths (/\); deaths during exposure (O); deaths during and after exposure (A). Solid line separates experiments in which no deaths occurred from those in which one or more animals died.
APPENDIX C 20 - 16 - ~ 12 by Al X 8 - o 4 o 175 it\ DEATH FROM MEL DEATH / / / / FROM CO / / / POISONING // \ DEATH FROM \ CO + HCN NON -LETHAL _ LATH FROM CO, LOW O2 + 5% CO2 CO2 DEATHS \ // \ \ \ \ \ \ ' ~ \ \ \ \ ~ \ \ DEATH FROM O2 DEPRIVATION \ \ \\\\\\\\\\\\ , I r T o 2000 4000 6000 CARBON MONOXIDE (ppm) 160 ~ Q Q - 120 ~ z IS 80 c, 40 O c: O I FIGURE C-2 Lethal concentrations of CO, 02, and HCN individually and in combination with 5 % co2. thermally decomposed, and the concentrations of the primary gases are mea- sured. An LCso of each is then predicted on the basis of the previously deter- mined relationships between the gases. To test the validity of the prediction, animals are exposed at the predicted LCso. If some fraction of the exposed animals dies, it can be assumed that the LCso has been estimated fairly closely. (That is possible in combustion toxicology, because concentration-response curves are usually very steep.) If none of or all the animals die, more experi- ments are conducted to determine the LCso more precisely. If the material produces combustion products that are significantly more toxic than was pre- dicted, according to the pure-gas relationships, further studies are necessary to determine what other toxic products or interactions have occurred. As more data are collected on materials, more gases will be added to the model, and the predictions will become more accurate.
176 APPENDIX C HEXACARBON NEUROPATHY Hexacarbon neuropathy is a relatively newly discovered clinical disorder. The traditional belief was that aliphatic hydrocarbons have no organ-specific activity and are inert metabolically. The emergence of hexacarbon neuropathy established that the belief had no foundation. The investigation of the disease and the unraveling of the causative agentts) provide an illustration of strategies for testing the toxicity of complex mixtures. The following description is based on three published reviews of the topic (Allen, 1980; Couri and Milks, 1985; Scala, 19761. The initial reports of hexacarbon neuropathy were associated with exposure to "n-hexane" (composition unknown). They arose in Japan in 1964, and the first American cases were reported in 1971. A large outbreak of peripheral neuropathy took place in 1973 in an Ohio coating plant. By 1976, the American episodes had been linked to a possible common neurotoxic metabolite formed from the suspect agents. In each case, despite some initial "identification" of an active agent, there was considerable uncertainty. There was, for example, an unwillingness to believe that otherwise inert agents or small changes in formulation could produce the reported effects in humans. The index cases of hexane neuropathy in the United States were ascribed to n-hexane, but the disbelief of some investigators led to chemical analyses. A sample of solvent from the workplace was found to contain 68 % acetone, 15 % isomeric C6 and C7 hydrocarbons, and only 17 % n-hexane. Other case reports from outside the United States used solvents that contained a variety of paraff~nic, cycloparaf- finic, and aromatic hydrocarbons and occasionally some esters. The causative agent was determined through a form of bioassay-directed fractionation. Vari- ous hydrocarbon solvents were tested in animals to determine whether n-hex- ane was the active agent, whether any other hexane or heptane isomers were active, and whether some oxygenates could increase the effect. The bioassay system was validated with respect to producing clinical or histologic evidence of neuropathy. The potential of n-hexane alone to produce neuropathy was well established. The final link was to create a synthetic blend that simulated com- mercial hexanes, but contained no n-hexane. The final experiment (Egan et al., 1980), not cited in the major reviews referred to earlier, showed that both "n- hexane-free hexane mixture" and the possibly interacting agent methyl ethyl ketone fail to produce the characteristic neuropathy. The second mystery involving hexacarbons yielded in part to a form of screening study. The inquiry was guided to some degree by the circumstances surrounding uncovers of the disease. The peripheral-neuropathy "outbreak" in Ohio in 1973 arose from an index case in a person who complained of progressive weakness. That case generated the names of five co-workers with similar complaints who had been seen individually elsewhere. Before the in- vestigation was completed, over 1,000 workers had been studied; 86 had some manifestation of a peripheral neuropathy ultimately related to workplace expo-
APPENDIX C 177
178 APPENDIX C evaluating the complex mixtures were largely effect-dr~ven. Agents and work- ers were screened for a definite effect. The success of the screening was evi- dent, but was only a portion of the larger success in solving two related indus- tnal-health problems. REFERENCES Alarie, Y. C., and R. C. Anderson. 1979. Toxicologic and acute lethal hazard evaluation of thermal decomposition products of synthetic and natural polymers. Toxicol. Appl. Pharmacol. 51 :341-362. Allen, N. 1980. Identification of methyl n-butyl ketone as the causative agent, pp. 834-845. In P. S. Spencer and H. H. Schaumburg (eds.). Experimental and Clinical Neurotoxicology. Williams & Wilkins, Baltimore. Couri, D., and M. M. Milks. 1985. Hexacarbon neuropathy: Tracking a toxin. Neurotoxicology 6(4):65-71. Crane, C. R., D. C. Sanders, B. R. Endecott, J. K. Abbott, and P. W. Smith. 1977. Inhalation Toxicol- ogy. I. Design of a Small Animal Test System. II. Determination of the Relative Toxic Hazards of 75 Aircraft Cabin Materials. FAA-AM-77-9. U.S. Federal Aviation Administration, Office of Aviation Medicine, Washington, D.C. (49 pp.) Egan, G., P. Spencer, H. Schaumburg, K. J. Murray, M. Bischoff, and R. Scala. 1980. n-Hexane- "free" hexane mixture fails to produce nervous system damage. Neurotoxicology 1 :515-524. Hilado, C. J. 1976. Relative toxicity of pyrolysis products of some foams and fabrics. J. Combust. Toxicol. 3:32-60. IARC (International Agency for Research on Cancer). 1986. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans. Vol. 38. Tobacco Smoking. International Agency for Research on Cancer, Lyon, France. (421 pp.) Klimisch, H.-J., H. W. M. Hollander, and J. Thyssen. 1980. Comparative measurements of the toxic- ity to laboratory animals of products of thermal decomposition generated by the method of DIN 53 436. J. Combust. Toxicol. 7:209-230. Levin, B. C. 1986. A Summary of the NBS Literature Reviews on the Chemical Nature and Toxicity of the Pyrolysis and Combustion Products from Seven Plastics: Acrylonitrile-Butadiene-Styrenes (ABS), Nylons, Polyesters, Polyethylenes, Polystyrenes, Poly(Vinyl Chlorides) and Rigid Polyure- thane Foams. NBSIR 85-3267. National Bureau of Standards, Gaithersburg, Md. Levin, B. C., A. J. Powell, M. M. Birky, M. Paabo, A. Stolte, and D. Malek. 1982. Further Develop- ment of a Test Method for the Assessment of the Acute Inhalation Toxicity of Combustion Products. NBSIR 82-2532. National Bureau of Standards, Gaithersburg, Md. Levin, B., M. Paabo, J. L. Gurman, and S. E. Harris. In press. Effects of exposure to single or multiple combinations of the predominant toxic gases and low oxygen atmospheres produced in fires. Fun- dam. Appl. Toxicol. NRC (National Research Council). 1986. Environmental Tobacco Smoke: Measuring Exposures and Assessing Health Effects. National Academy Press, Washington, D.C. (337 pp.) Petajan, J. H., K. J. Voorhees, S. C. Packham, R. G. Baldwin, I. N. Einhorn, M. L. Grunnet, B. G. Dinger, and M. M. Birky. 1975. Extreme toxicity from combustion products of a fire-retarded polyurethane foam. Science 187:742-744. Scala, R. A. 1976. Hydrocarbon neuropathy. Ann. Occup. Hyg. 19:293-299. Yusa, S. 1985. Development of laboratory test apparatus for evaluation of toxicity of combustion products of materials in fire, pp. 471-487. In 7th Joint Meeting of the U.S.-Japan Panel on Fire Research and Safety, Washington, D.C., October 24-28, 1983. Proceedings. NBSIR-85/3118. Na- tional Bureau of Standards, Gaithersburg, Md. (Available from NTIS as PB85-199545/XAB.)
