VI. Health Effects Discussion and Determination of Final PEL 

A. General Principles of Toxicology and Dose Response

Introduction

As long ago as the 16th century, people recognized that there is no such thing as an absolutely safe chemical. The Swiss physician Paracelsus, who lived from 1493 to 1541, said:

All substances are poisons; there is none which is not a poison. The right dose differentiates a poison and a remedy.

On the other hand, methods have been devised to permit any chemical, no matter how poisonous, to be handled safely; this is done either by limiting the dose or controlling the exposure. However, before the necessary degree of control can be determined for a particular exposure or situation, the toxicity of the substance in question must be known. The paragraphs that follow describe the methods used by scientists to measure the relative toxicity of substances and to select exposure limits that will prevent exposed individuals from suffering adverse effects from such exposures. As this discussion demonstrates, methods of choosing exposure limits must, because of the lack or inadequacy of dose-response information for many chemicals, rely on experience in the use of these substances and on scientific and professional judgment(1).

__________ Footnote(1) The material in this section derives principally from the following sources: Klaasen, Amdur, and Doull 1986; National Research Council 1986; Cohen 1986a, b; and Tardiff and Rodricks 1987.

Chemicals range in inherent toxicity from those that are relatively harmless even after large doses have been administered to others that cause death if encountered even in small quantities. Toxicologists rank chemicals by categories that range from practically nontoxic (an adult human would have to consume a quart) to supertoxic (fewer than 7 drops would be lethal for most people).

In the occupational setting, it is the risk associated with a particular use of a chemical rather than its inherent toxicity that is important. Risk can be defined as the probability that a substance will produce harm under certain conditions of use. The converse of risk is safety, which is the probability that no harm will occur under specific circumstances.

The degree of hazard associated with exposure to a specific substance depends on the manner in which it is handled in a particular situation: a supertoxic chemical that is processed in a closed, isolated system may be less hazardous in actual use than a low-toxicity compound handled in an open batch process. Another factor affecting the ability of a chemical to elicit a toxic response is the susceptibility of the biological system or individual. For the relative degree of hazard to be known in a particular instance, this requires knowledge about the chemical agent, the exposure situation, and the exposed subject. In addition, the route of administration and the duration and frequency of exposure must be known.

Route of Exposure

There are four principal routes of exposure by which toxic substances can invade humans or animals. These are inhalation, ingestion, dermal absorption, and parenteral administration (i.e., administration through routes other than the intestinal canal). The route of administration of a toxin also affects the relative toxicity of the agent. For example, a chemical that can be detoxified in the liver will be less toxic if it is administered orally than if it is given systemically (i.e., inhaled). Studies that provide information about the relative toxicity of an agent via different routes of exposure can provide a considerable amount of information about the absorbability of the agent. For example, if exposure to a certain dose of a chemical via all routes of administration causes death within the same time period, it can be assumed that the substance in question is easily and rapidly absorbed. On the other hand, if the dermal dose of a chemical that is required to kill a subject is much higher than the dose required to produce the same effect when the chemical is ingested, one can deduce that the skin provides, to some degree, a barrier against that agent's toxicity. Other, less important, elements affecting the response to a toxic substance include the relative concentration of the substance, the volume of the vehicle used to administer the chemical, the chemical and physical properties of the vehicle, and the dose rate (i.e, the period of time over which the dose is administered).

Duration and Frequency of Exposure

Scientists conduct animal experiments that involve four different types of exposure: acute, subacute, chronic, and subchronic. Acute exposures are limited to periods of less than 24 hours and can involve either single or repeated exposures within that period. Subacute exposures are repeated exposures that last for one month or less, while subchronic exposures have a duration of one to three months. When a research project having a chronic regimen is conducted, the test animals are dosed repeatedly for a period lasting more than three months. Animals exposed acutely can have both immediate and delayed-onset responses. Similarly, chronic exposures can cause immediate reactions as well as long-term effects.

The frequency of dosing also has an important influence on the magnitude of the toxic effect: a large single dose of an acute toxin will usually have more than three times the effect of one-third the dose given at three different times, and the same dose administered in 10 or 15 applications might have no effect whatsoever. The pattern of dosing is important because it is possible for some of the substance to be excreted between successive administrations or because the lesion caused by the toxin has a chance to be partially or completely repaired between applications. Thus a chronic effect is said to occur: (1) if a toxic substance accumulates in the system of an exposed person or animal because the dose absorbed is greater than the body's ability to transform or eliminate the substance; (2) if it produces adverse effects that are not reversible; or (3) if it is administered in a manner that permits inadequate time for repair or recovery.

Variation in Response

Responses to toxic insults vary in a number of ways. For example, some toxicants have immediate effects, while others are associated with delayed symptom onset. The latency period for carcinogenic agents may be as long as 40 years for some types of cancer, and even some acute agents, such as some chemicals that have adverse ocular effects, may not cause overt symptoms until hours after exposure.

Another difference in type of response concerns the reversibility or irreversibility of the effect. Reversibility depends on the site of action as well as the magnitude of the insult. That is, some tissues of the body, such as the liver, have considerable ability to regenerate; others, like the kidney or central nervous system, do not.

The site of action associated with toxic substances also varies widely. Local effects are those lesions caused at the site of first contact between the agent and the organisms. Examples of localized effects are skin burns caused by contact with a caustic substance and site-of-contact tumors that develop at the locus of the injection of the carcinogen.

In contrast to localized effects, systemic effects involve the absorption and distribution of the toxic agent from the point of entry to a distant site; the toxic response is manifested at this distant point. An example of a systemic poison is mercury, which produces its toxic effect on the central nervous system. Often, the site of deposition for a chemical is not the organ system most affected by the toxin. For example, although lead is deposited and concentrated in the bone, it affects the central nervous system. Any sites that are adversely affected by the toxic effects of exposure to a substance, whether they are sites of contact or distal sites, are called the target organs of toxicity.

In cases of systemic poisoning, the system most often affected is the central nervous system (CNS); it is common for the CNS to be involved even when another target, such as the liver, is the primary target organ of toxicity. In descending order of frequency, the systems or organs most often involved in cases of systemic poisoning are the central nervous system, the circulatory system, the blood and hematopoietic system, the visceral organs (liver, kidneys, lungs), and the skin.

Dose-Response

The relationship that associates the dose of a chemical with the effects it causes is called the dose-response relationship. A single data point relating a dose to a response is sufficient to establish a dose-response relationship. As additional data become available, it is possible to expand our understanding of the dose-response relationship to cover a range of doses or exposures. Dose-response is an important principle in toxicology, and an understanding of dose-response is important in establishing occupational or other exposure limits. Knowing how toxic substances act makes it easier to predict the potential effects of exposure. (It is, of course, generally true that lowering dose reduces response, and data are often available to demonstrate that lower doses reduce responses, at least on the grossly observable level. However, data showing that more subtle responses (e.g., those at the subcellular level) have been reduced are rarely available.)

To apply dose-response relationships, it is helpful if several types of data are available. First, it must be possible to relate a response to a particular chemical. Although basic data pointing toward causality may be available, it is often difficult to refine the dose-response relationship further. For example, epidemiological studies often identify an association between a disease and one or more causative agents. However, since information on the precise identity of the etiologic agent, the actual dose received, and the true site of the response is usually not available, it is often impossible to use data from epidemiological studies to establish a precise dose-response relation between a specific dose of a toxin and an effect.

The second condition to be met before dose-response can be established is that it must be possible to relate the response to the dose. It is relatively easy to determine that a large dose causes an obvious response. Refining the relationship, however, involves three other requirements: (1) that there be a receptor site; (2) that the response and the intensity of the response be related to the concentration of the toxin at the receptor site; and (3) that the concentration of the toxin at the site be related to the dose given.

The third principle underlying the concept of dose-response is that there must be a quantifiable means of measuring the toxicity of a substance and a method of expressing this measured toxicity. Although lethality in test animals is often used to measure toxicity, the best form of measurement would involve quantification of the sequence of molecular events occurring during the toxic response. In the absence of such endpoints, other good methods are available. For example, it is common to measure an effect believed to be related to the substance in question. The level of activity of an enzyme in the blood is often used as a measure of effect, e.g., serum glutamic-oxaloacetic transaminase (SGOT) levels are used to measure liver damage. Many different endpoints can be used to measure toxic effects, such as changes in muscle tone, heart rate, blood pressure, electrical activity of the brain, motor functioning, and behavior.

The most widely used endpoint, especially when a new substance is involved, is lethality in an animal test system. Lethality studies allow scientists to make comparative assessments of a chemical's toxicity as it relates to that of many other substances. Research of this type also permits the gathering of essential information on dose, duration, route of administration, site of action, and the target organ of toxicity.

Form of the Response

The classic form of dose-response is sigmoidal (Figure 1). This form characterizes the relationship between the amount of a toxin administered and the degree of response to that dose. The response is measured on the ordinate, and the dose is represented on the abscissa.

Dose-response can be thought of in two ways:

As exposure increases, the proportion of the population that manifests the response increases (quantal response); and

As exposure increases, the intensity of an individual's response increases (graded response).

A relatively flat dose-response curve means that a large change in dose is required before there is a significant change in response. A steep curve, on the other hand, means that a small change in dose will elicit a large increase in response. Although it is sometimes possible to generate a curve of the type shown in Figure 1, it is not necessary to do so to demonstrate that exposure at a given level is associated with a particular response. That is, it is not necessary to have sufficient data to define, in mathematical terms, the dose-response relationship to know that exposure at a given level is associated with adverse consequences.

Figure 1 - Diagram of Dose-Response Relationship (For Figure 1, Click Here)

In the regulatory context, it is most common to express dose-response relations in terms of the percentage of the population responding. However, before this information can be evaluated, the endpoint being considered must be known. For every substance, there are several dose-response relationships, depending on endpoint: a substance that produces irritation at low doses may cause more severe symptoms or even death at high doses and in other conditions. For example, many substances that are mucosal irritants at low doses will produce pulmonary edema and nervous system effects at high doses.

Plotting the cumulative percentage of individuals responding against dose produces the typical sigmoid curve. Such a curve reflects the fact that at the lowest dose, zero percent of the population responds, while 100 percent of the population will respond at the highest dose. However, if the percentage responding is plotted against incremental rather than total dose, the curve produced is a normal distribution (Figure 2). This curve says that a relatively small percentage of the population will manifest the response at the lowest dose and that a similarly small percentage of the population will exhibit the effect only at the highest dose. What this normal distribution of response reflects is individual and species variation in exposed populations. A wide degree of variation occurs even in inbred, homogeneous laboratory animals, and such variability increases dramatically when a heterogeneous population, such as workers, is involved. Individuals responding at the left end of the curve shown in Figure 2 are hypersusceptible, while those at the right end could be termed resistant.

Figure 2 - Diagram of Quantal Dose-Response Relationship (For Figure 2, Click Here)

Because the relationship between dose and response is sigmoidal, response approaches zero as dose approaches zero. However, because of the mathematical form used to express this relationship, a true zero response can never be achieved. In the strictest sense, therefore, a true threshold dose level (i.e., the dose with which a zero response is associated) can never be established on the basis of experimental research. Instead, scientists attempt to define the minimum dose associated with a specific endpoint, which is customarily termed the "threshold" dose for that particular endpoint. However, unless a specific endpoint (such as respiratory irritation, cholinesterase inhibition, the development of a tumor, or death) is specified, the concept of a threshold is essentially meaningless. In fact, a separate threshold could be said to exist for each of these endpoints.

The extent to which an experimentally derived "threshold" actually reflects the true threshold for a substance (i.e., the level above which a response will occur and below which no response will occur) depends on several factors, such as the number of animals used to determine the experimental threshold, the number of dose levels tested, and the degree of variation represented in the test subjects. For example, to determine an LD(50) (the lethal dose that will kill 50 percent of the animals tested) with a high degree of precision requires the use of a minimum of 50 test animals and five dose groups (10 animals in each group). Other factors that can influence the magnitude of the median lethal dose include the sources involved, the sex and age of the animals, the environmental conditions prevailing during the test conditions, diet, the health status of the subjects being tested, and the subjects' past exposure to other toxic substances.

In toxicological research, the experimentally observed threshold dose is called the low-observed-effect level (LOEL) or the low-observed-adverse-effect level (LOAEL). Alternatively, the threshold may be expressed as the highest no-observed-effect level (NOEL), i.e., the highest dose administered and found not to produce a given response. Determination of an accurate NOEL requires both a careful interpretation of the toxicological data and the use of an adequate number of test animals. The National Academy of Sciences (1985) has concluded that the chance of finding a no-adverse-effect level (that is, of missing an adverse effect) at a given dose is statistically greater in experiments having a small number of animals than in studies involving a large number of animals. Thus, the degree of confidence one has that a NOEL actually represents a "safe" dose, rather than a research design artifact, increases with the number of animals tested. The greatest degree of confidence is associated with studies involving a large number of animals that were tested at several doses that were administered at close intervals.

In a recent publication (Tardiff and Rodricks 1987), David W. Gaylor of the National Center for Toxicological Research explained that experimentally derived thresholds represent statistical limitations in study design rather than biological characteristics:

The existence of dose-response relationships might lead one to assume incorrectly the existence of threshold doses below which no toxic effects could occur. As dosage is decreased, the prevalence of an observable toxic effect...diminishes to zero. Eventually, a dosage is reached below which the experiment has essentially no resolving power to distinguish between the spontaneous background rate and small induced toxic effects....

If no toxic effects are detected at a specified dosage, this dosage is called the no-effect, or more correctly the no-observed-effect dosage. Because of the limitations of any given experiment, the no-observed-effect dosage is not a precise estimate of a true no-effect level. Lack of statistical significance is not equivalent to no toxic effect. It may or may not be, and further experimentation would be required to resolve this equivocal issue.... The no-observed-effect level is not a biological property, but, rather, a statistical property or operational threshold that is highly dependent on sample size.

The scientific issues surrounding the concept of no-observed-effect levels or experimentally derived thresholds have important implications for their use in establishing protective occupational exposure limits. Because the no-observed-effect level cannot represent the "true" threshold for an adverse effect, given the design of most toxicologic studies, regulators and others have used the concept of safety factors (also known as uncertainty factors) to aid them in setting permissible exposure limits; that is, the exposure limit is established at some interval below the no-observed-effect level to provide additional assurance that exposed populations are not likely to suffer harm.

The size of the interval between the permissible exposure limit and the no-observed-effect level depends on a professional judgment as to whether the no-observed-effect level is likely to represent a level that is not harmful to humans. Thus, if the available data include a NOEL derived from a well-conducted human study, a smaller safety factor might be used to establish an exposure limit than would be used if the data to be used to establish the limit consisted of a NOEL from an animal study; in the latter case, there is greater uncertainty regarding the relationship between the animal NOEL and human NOEL. Safety factors have also been used to recognize the fact that the human population is heterogeneous and that there may be a wide variation in individual responses to toxic substances (the wide range in the odor thresholds reported for some substances is a good illustration of individual variability in response).

The use of NOELs, LOAELs, and safety factors to develop permissible exposure limits is not a recent development:

For more than half a century, evaluation of the safe use of chemicals has been focused mainly on the development of toxicity data and on the application of professional judgment to the ad hoc interpretation of such data to derive acceptable levels of exposure for humans. Generally, this practice has taken the form of identifying from studies in laboratory animals the no-observed-effect level and dividing it by a safety factor (usually 100 for NOELs derived from chronic studies) reflecting the uncertainties of relating data to humans under their conditions of exposure and the quality and appropriateness of the data base....

Safety factors are usually chosen prospectively to address the uncertainties of interspecies extrapolation. Although safety factors as small as 2 and as large as 2000 have been used...the safety factor of 100 is used most commonly, at least for NOELs derived from chronic toxicity studies, and incorporates adjustments for interspecies variability (usually 10) and intrahuman variability (usually 10).... The resulting value is equivalent to a NOEL in humans (Tardiff and Rodricks 1987, pp. 391, 421).

Tardiff and Rodricks caution, however, that the use of safety factors has been questioned because these factors "often create the impression that human population thresholds have been identified and that there is virtually no risk below that level of exposure" (Tardiff and Rodricks 1987, p. 421).

Although safety factors have traditionally been used to establish exposure limits for chronic or lifetime exposure situations, they have also been applied to establish limits for acute effects resulting from short-term exposure. The National Academy of Sciences' Committee on Toxicology has been using a safety-factor approach to establish emergency exposure guidance levels (EEGLs), which are exposure levels judged to be acceptable for military personnel performing tasks during emergency situations. An EEGL is not considered to be a safe exposure level for routine or normal operations, but these levels are considered acceptable when tasks must be performed to prevent greater risks (e.g., death or injury caused by fires or explosions). In developing EEGLs, safety factors are generally applied to account for uncertainties in the use of animal data and when extrapolating between different dose routes. The NAS also develops short-term public emergency exposure guidance levels (SPEGLs) to apply to the exposures of the general public to contaminants during airborne chemical releases; SPEGLs are generally set at a level of 0.1 to 0.5 times the EEGL (i.e., an additional safety factor of from 2 to 10) (Criteria and Methods for Preparing Emergency Exposure Guidance Level (EEGL), Short-Term Public Emergency Guidance Level (SPEGL), and Continuous Exposure Guidance Level (CEGL) Documents. Washington, D.C.: National Academy Press, National Academy of Sciences 1986).

The use of the safety factor approach in establishing occupational exposure limits was addressed by many rulemaking participants (Exs. 3-744, 3-1095, 8-16, 8-47, 116, and 144; Tr. 1-221, Tr. 2-163 to 2-164). NIOSH (Ex. 8-47) stated that safety factors cannot be used to estimate human risk and are therefore not related to the magnitude or significance of a risk; instead, NIOSH believes that safety factors are intended to reflect uncertainty in the available data. This comment echoes the observation made by Tardiff and Rodricks, i.e., that safety factors do not necessarily identify a human population threshold. NIOSH (Ex. 8-47) also endorsed the use of safety factors as a "pragmatic method" of developing standards (except when a nonthreshold process, such as the induction of cancer, is the outcome of concern). NIOSH also believes that "standards based on a margin of safety...as well as standards derived from a case-by-case evaluation, [should] be periodically reviewed to determine what new information is available" (Ex. 8-47).

Dr. Marcus Key, Professor of Occupational Medicine at the University of Texas School of Public Health, also testified on the appropriateness of using safety factors to establish occupational exposure limits:

We seldom, if ever, know with any precision where a significant risk level begins or ends; hence, the need for safety factors. Safety factors depend on several considerations,...mainly on toxicity and the nature of the health effects, but also on the availability of scientific evidence of effects at lower levels.

Professional judgment must be relied on in selecting safety factors, with one to three orders of magnitude being commonly used for serious effects, and 50 percent, or [a] safety factor of 2, [being used] for acute, less harmful effects (Tr. 1-221).

Both Dr. Key (Tr. 1-221) and Dr. Ernest Mastromatteo, Chairman of the ACGIH TLV Committee (Tr. 2-163 to 2-164) testified that safety factors are frequently used by the ACGIH to develop recommended exposure limits.

Some commenters (Exs. 8-16, 116, and 144; Tr. 7-121) were of the opinion that OSHA should adopt a uniform system of assigning safety factors to establish permissible exposure limits. For example, the Workers' Institute for Safety and Health (WISH) (Ex. 116, p. 13) commented that OSHA should review the toxicology profiles prepared by the Agency for Toxic Substances and Disease Registry (ATSDR), in which Reference Doses (RfD) are computed. The RfD, as described by WISH, is "an estimate (with uncertainty spanning perhaps an order of magnitude) of the daily exposure of the human population to a potential hazard that is likely to be without risk of deleterious effects during a lifetime" (Ex. 116, p. 13). The RfD is derived by applying uncertainty factors to experimentally derived NOAELs in a consistent manner. The uncertainty factors used by ATSDR include factors of 10 to account for each of the following:

* Human variation in response; * Extrapolation from animals to humans; * Extrapolation of effects associated with lifetime exposure from less-than-lifetime studies; and * Additional uncertainty in relying on a LOAEL rather than a NOAEL.

In addition, ATSDR applies a factor of from 1 to 10 to account for the overall quality of the scientific evidence.

EPA uses the same approach to develop RfDs for noncarcinogens; EPA's application of this approach is described in a concept paper presented by the EPA Reference Dose Work Group (Ex. 144, Appendix A). As explained by the Work Group:

The RfD is useful as a reference point for gauging the potential effects of other doses. Usually, doses that are less than the RfD are not likely to be associated with any health risks, and are therefore less likely to be of regulatory concern ....Nonetheless, a clear conclusion cannot be categorically drawn that all doses below the RfD are "acceptable" and that all doses in excess of the RfD are "unacceptable" (Ex. 144, Appendix A, p. A-10).

The EPA has been compiling dose-response data and information on RfDs for almost 2,000 chemicals in a database called the Integrated Risk Information System (IRIS). The system is described by Dr. Rebecca T. Zagraniski, Assistant Commissioner of the Division of Occupational and Environmental Health, New Jersey Department of Health (Exs. 144 and 144A). In her posthearing submission, Dr. Zagraniski presents an analysis in which EPA RfDs from the IRIS system are converted to Workday Ambient Air Concentrations (WACs) for 43 of the substances included in this rulemaking. These WACs were then compared to OSHA's proposed PELs for the same substances. After converting the RfDs to equivalent lifetime occupational exposure levels, Dr. Zagraniski found that all but three of the resulting WACs were lower than 1 mg/m(3) and that the WACs for noncarcinogens were generally 100 to 1,000 times lower than the PELs being proposed by OSHA in this rulemaking. Dr. Zagraniski commented on these findings as they relate to OSHA's proposal:

The WACs are not recommended exposure limits because they do not take into account numerous significant considerations including feasibility, anecdotal reports of effects following human exposure, routes of exposure other than inhalation, and other critical information. Also, the WACs for non-carcinogens are based primarily on oral exposure studies. In some cases, there may be inhalation studies which are more appropriate for use in setting an occupational exposure guideline, but which were not discussed in IRIS due to their focus on the oral exposure route. In spite of these constraints, the WACs may be considered preliminary health-based guidelines which are useful as indicators that current PELs and TLVs may need reevaluation (Ex. 144A, p. 4).

In response to Dr. Zagraniski's comments, OSHA notes that the approach suggested by this commenter is new and was not supported by other participants. It is also inconsistent with the recommendations of most expert organizations in this field and would require extensive analysis by OSHA before its merit could be ascertained. Accordingly, OSHA finds this approach inappropriate for use in the present rulemaking.

In this rulemaking, OSHA has evaluated the efficacy of the final rule's limits on a case-by-case basis; although the initial evaluation presented in the NPRM relied heavily on analyses conducted by the ACGIH and NIOSH, the limits promulgated in the final rule are based on an expanded toxicologic assessment using information contained in the rulemaking record. OSHA believes that, at this time, this case-by-case assessment is the best way to establish new and revised limits for the numerous substances addressed in this rulemaking.

Types of Toxicological Evidence

The evidence available to scientists wishing to evaluate the toxicity of a substance can be derived from studies in laboratory animals, in vitro studies in cell or tissue systems, reports of clinical observations, studies of exposed human populations, or from intervention studies conducted with human volunteers. The preceding paragraphs have described animal studies (or "bioassays"). The following section discusses the two most common types of human evidence: data derived from clinical observations and information from epidemiological studies.

Clinical observations. Much of the data on the toxic effects associated with human exposures have come from industrial accidents, fatal poisonings, or other such tragedies. This information is generally more useful in delineating broad categories of pathological effects than in refining a specific dose-response relationship, because the exposure levels causing the accident are known to be high but cannot be quantified with precision.

Epidemiological studies. Studies conducted by epidemiologists are designed to reveal the patterns of disease or mortality prevailing in certain groups of people (usually workers) exposed to a single toxin or to a group of substances. One of the advantages of epidemiological studies is that they involve humans and their responses to actual situations. The interpretation of the results of epidemiological studies is complicated by the inevitable presence of confounding variables that occur whenever human populations are involved. Ideally, the populations being studied (i.e., the study population and the control population) should be fully comparable with regard to every variable except the single characteristic under study. Because it is rarely possible to achieve this degree of comparability, statistical techniques are often used to attempt to adjust for this lack of comparability. In addition, if the measured effect is relatively large, it is unlikely that confounding factors will obscure the true picture.

Broadly speaking, epidemiological studies can have two possible outcomes: they can report an effect or they can report no effect; in the former case, the study is termed a positive study, and in the latter, a negative one. Within each of these categories, it is possible for the study to be correct (that is, to give a true-positive or true-negative result) or to be incorrect (that is, to give a false-positive or a false-negative result). A false-positive result reports that there is an increased risk when in fact there is not, and a false-negative study reports that there is no increased risk when in fact there is.

The probability that a study will detect a statistically significant effect if that effect is actually present is called the power of the study. As the power of a study increases, the likelihood of producing a false-negative error decreases. Power is dependent on two factors: the level of relative risk being evaluated and the number of cases of the effect (i.e., disease) that are expected in the population being studied. The number of expected cases depends both on the sample size and the expected disease frequency in the comparison population. For example, a study involving a small population and a common disease can have the same power as a study of a rare disease in a large population. Consequently, studies of larger samples have sufficient power to detect smaller increases in risk, and studies of smaller samples will be able only to detect large increases in relative risk.

Because epidemiological studies have limitations, it is essential that the power of such studies, particularly of negative studies, be examined to ensure that their sample sizes are adequate to detect the absence of increased risk with validity. When the power of a study is not adequate, negative studies cannot be said either to contradict or to support the conclusion that increased risk exists. Essentially, a negative epidemiologic study identifies a NOAEL, which, as discussed above, reflects the statistical limitations of a study more than the "true" population threshold for an effect. However, a study with a positive result may indicate a relationship if the excess risk is high, even if the study's sample size is small and the effects of some factors are not controlled for.

Quality of Evidence

Dose-response models have often been used in the quantitative assessment of the risks associated with exposures to carcinogenic substances. However, less scientific effort has been devoted to models to be used with non-carcinogenic substances. Mathematically precise methods to establish the true no-effect level or to define the dose-response curves have not been developed for most of the more than 400 substances involved in this rulemaking.

Most of the scientific work that has been done was designed to identify lowest observed effect or no-effect levels for a variety of acute effects. As described above, experts in industrial hygiene and occupational health have developed factors to be used to offset, at least to some extent, the insensitivity of NOELs and LOELs to such factors as subcellular effects, sensitive individuals, and chronic effects. It is possible to use these data, combined with professional judgment and OSHA's expertise and experience, to determine that significant risk exists at current levels of exposure and that a reduction in these levels will substantially reduce this risk of material impairment of health. OSHA is also confident that it is not attempting in this rulemaking to reduce exposures to insignificant levels. However, additional analysis may well reveal that the levels being established in the final rule can be refined further in the future.

B. Historical Development of Occupational Exposure Limits Early Limits

Until the development of occupational health standards, the occurrence of adverse health effects resulting from exposures to hazardous substances or conditions in the workplace could only be determined ex post facto - after impairment had already occurred to the health and welfare of exposed employees. In her 1910 studies of lead poisoning, Dr. Alice Hamilton was forced to rely on "personal observations of working conditions and the illness and deaths of workers to demonstrate the existence of harmful exposures" (Paull 1984/Ex. 1-255). The concept of occupational exposure limits thus represents a dramatic breakthrough in the battle against occupational disease and remains "one of the most useful and indispensable tools yet devised for safeguarding the health and well-being of industrial workers" (Thomas 1979/ Ex. 1-96).

Occupational exposure limits are air quality values that apply in workplaces, and they are derived by studying the correlation between the amount of a toxic substance absorbed by the body and its effects on health. Within the context of occupational exposure, knowledge of this relationship permits quantification of the etiology "of a large number of occupational health impairments, [evaluation of] the risk of such impairments and, if necessary, [consideration of] the effectiveness of preventive measures" (Parmeggiani 1973/Ex. 1-229). More specifically, an understanding of the levels at which disease or other health effects occur can be used to establish limits of occupational exposure below which health hazards are unlikely to occur in most workers.

The historical development of occupational exposure limits began with the published reports of a German scientist whose investigations in 1883 into the effects on experimental animals (and on himself) of carbon monoxide in known air concentrations caused him to conclude that "the boundary of injurious action of carbon monoxide lies at a concentration in all probability of 500 parts per million, but certainly [not less than] 200 parts per million" (Cook 1987/Ex. 1-187). Shortly after the appearance of this first documented dose-response value, another German researcher, K. N. Lehmann, published a series of reports on a number of chemical substances under the title "Experimental Studies on the Effect of Technically and Hygienically Important Gases and Vapors on the Organism." This series culminated in 1936 with a comprehensive paper on chlorinated hydrocarbons, published as Volume 116 of Archiv fur Hygiene.

In 1912, Rudolf Kobert published a table of exposure limits, based on animal studies, for 20 compounds. One of the first tables of hazardous air concentrations to originate in the United States was a technical paper published in 1921 by the U.S. Bureau of Mines. The 33 substances included in this table were those frequently encountered in the workplace. In addition to limits based on acute toxic effects, this table provided some information on the least detectable odor concentration and the lowest airborne concentration required to cause irritation (Paull 1984/Ex. 1-255; Cook 1987/Ex. 1-187).

Throughout the 1920s and 1930s, data became available that correlated concentrations of harmful substances with observed effects on worker health for such materials as lead and mercury compounds, benzene, and granite dusts. These early occupational health studies, which were based on animal experiments and on findings in exposed workers, provided the kind of data needed to link human exposures "to concentrations that were capable of producing not only acute, but chronic health effects" (Paull 1984/Ex. 1-255).

After 1935, the emphasis of researchers had shifted, for the most part, from the reporting of a series of values for a range of acute effects to results that yielded a single limit based on studies of repeated exposures. Over the years, a sizable amount of data about the levels of exposure that would not produce injurious effects had been amassed for a considerable number of substances. "By the early 1940s, control of the occupational environment to prevent the harmful absorption of toxic materials was becoming an accepted principle, and the practical problem of defining what was `harmful' was beginning to be met by employing maximum allowable concentrations" (Paull 1984/Ex. 1-255). In 1943, Sterner (Ex. 1-806) explained the meaning of the term maximum allowable concentrations as "the upper limit of concentration of an atmospheric contaminant which will not cause injury to an individual exposed continuously during his working day and for indefinite periods of time" (Paull 1984/Ex. 1-255).

The first lists of maximum allowable concentrations of airborne toxic substances were issued between 1933 and 1938. The Union of Soviet Socialist Republics (U.S.S.R.) was the first country to make occupational exposure limits a statutory obligation; in 1933 it published a list that included 14 substances (although health standards for some air pollutants apparently were used in the Soviet Union during the 1920s). The first American list was published four years later by the State of Massachusetts, and in 1938 Germany issued occupational health standards for a number of organic solvents (Holmberg and Winell 1977/Ex. 1-141). Additionally, the United States "imposed limited occupational safety and health requirements on certain contractors with the Federal government" when the Walsh-Healey Act was passed in 1936 (Mintz 1984/Ex. 1-840).

Standards Developed by Professional Organizations

During the 1940s, American organizations led in the development of occupational health standards, beginning with the American Standards Association (now the American National Standards Institute, or ANSI) list of "maximum acceptable concentrations" (MACs), which appeared in 1941. This list represented a consensus of opinion by the ASA and a number of industrial hygienists who had formed the American Conference of Governmental Industrial Hygienists (ACGIH) in 1938 (Baetjer 1980/Ex. 1-223). Originally conceived of as a time-weighted concentration to be maintained as an average over the working shift, the MAC was redefined in 1957 to mean an upper level (ceiling level) that should never be exceeded (Turner 1976/Ex. 1-79).

An important contribution to occupational health standard-setting was made in 1945 by Warren Cook (Ex. 1-726), who published a list of maximum allowable concentrations for 132 industrial atmospheric contaminants. These limits had been developed by six states, the U.S. Public Health Service, and the American Standards Association, and included Cook's own list of "accepted or tentative values" based on industrial experience, animal experimentation, human sensory response, or a combination of these factors. This table was followed by:

Documentation supported by 187 specific references, indicating the basis and reliability of each value. Cook was the first investigator to codify all of the available data on MAC's and present it in one publication. His list of recommended values was incorporated, practically without changes, by the ACGIH in establishing the TLVs. In support of Cook's inferences, it should be noted that 50 of the...values that he recommended in 1945 were subsequently adopted as federal standards, and are still in use today (Paull 1984/Ex. 1-255).

The American Conference of Governmental Industrial Hygienists Subcommittee on Threshold Limits presented its second report at the Eighth Annual Meeting of the ACGIH in 1946. The report included values for 131 gases, vapors, dusts, fumes, mists, and 13 mineral dusts "compiled from the list reported by this subcommittee...in 1942, from the list published by Warren Cook in...1945, and from published values of the Z-37 Committee of the American Standards Association" (Cook 1987/Ex. 1-87). The Committee's report noted that:

Considerable difficulty attends the fixing of satisfactory values for maximal allowable concentrations of chemicals in respirable atmospheres because of the lack of a uniform definition of the maximum allowable concentration concept. One concept is that the M.A.C. value should represent as accurately as possible that concentration at which a worker exposed for a sufficient period of time will just escape physiological or organic injury and occupational disease.

A second concept is that the M.A.C. should represent some fraction of that concentration which will injure the worker in order to allow a margin of safety in the design of protective equipment and guard against possible synergistic effects in the case of multiple exposures. A third concept is that the M.A.C. should perform the functions of the former concepts and in addition provide a work environment free of objectionable but non-injurious concentrations of smokes, dusts, irritants and odors. Obviously all of these concepts cannot be fulfilled with the establishment of a single value. M.A.C. values in use at the present time represent examples of all of these concepts. The committee feels that the establishment of dual lists or a single definition is not possible at the present time (ACGIH 1946).

The report concluded by stressing that the 1946 list of M.A.C. values was presented "with the definite understanding that it be subject to annual revision" (ACGIH 1946).

Papers presented at both the Ninth International Congress on Industrial Medicine in London (1948) and at the Fifteenth International Congress of Occupational Health in Vienna (1966) also dealt with maximum acceptable concentrations. The first of these proposed that zones of toxicity be set up to facilitate an understanding of the relative hazards of substances, "since the boundaries of MAC values were not sharp lines of demarcation" (Cook 1987/Ex. 1-87). At the 1966 meeting, discussion took place on the advantages of the concept of a "peak level" of exposure - an extension of the "ceiling level" notion inherent in the definition of a MAC since 1957. A "peak level" was defined as one "that can be applied to certain substances for brief designated periods and for a strictly limited number of times during the work shift, with a designated time interval between peaks. The `peak' concept places a limit on the intermittent higher exposures that occur in many industrial operations. The time-weighted average exposure limit is of course to be observed [even when a peak has also been assigned to a substance]" (Cook 1987/Ex. 1-87).

Terminology and definitions throughout this early period were ambiguous and imprecise, reflecting uncertainty as to exactly what needed to be and could be done in the realm of occupational health standard setting. Initially, the ACGIH designated its recommended limits as "maximum allowable concentrations," although this term was often used interchangeably with "threshold limit values." Confusion about the meaning, interpretation, and relative significance of the terms being employed during this embryonic period was common. After 1953, the ACGIH defined the concept of threshold limit values in the preface to its annual published list of occupational health standards as "maximum average atmospheric concentrations...for an eight-hour day." This definition of the TLVs as average concentrations differed from the general understanding of the original term "maximum allowable concentrations," which were essentially ceiling values (Stokinger 1962/Ex. 1-998).

Documentation for the 238 substances included in the TLV list for 1956 was provided by Smyth (Ex. 1-759) in a separate paper in which the author:

Recommended that the TLV's include references to the underlying data, and that the concepts represented by the values be restated in more realistic toxicological terms. In his analysis of the TLVs, he [Smyth] concluded that nine categories of objectionable action were guarded against: chronic toxicity, acute toxicity, narcosis, irritation, asphyxiation, fume fever, eye pigmentation, allergic response, and cancer (Paull 1984/Ex. 1-255).

At about the same time, Stokinger stated that, in his opinion, the Threshold Limits Committee had avoided grappling with the issue of developing a method for establishing limits for industrial carcinogens and noted that, with the exception of nickel carbonyl, limits had not been assigned for potential carcinogens (Paull 1984/Ex. 1-255). In 1962, however, the TLV Committee included three carcinogens as additions to the TLV list, although these were listed separately in an appendix and did not have assigned TLVs. Despite the fact that the ACGIH had stressed early on that TLVs were intended as guides and not as rigidly enforceable limits, the American Standards Association's MAC values (or, where none was available, the TLV) were included as mandatory limits in the Safety and Health Standards for Federal Supply Contracts, which were published in 1960 under the Walsh-Healey Act. Following this action, the ACGIH issued a statement on the definitions and interpretations of TLVs and MACs (Stokinger 1962/Ex. 1-998). At the same time, the ACGIH announced the production of the first edition of the Documentation for Threshold Limit Values (ACGIH 1962); this was followed by another paper in which the work and intentions of the Threshold Limits Committee were reviewed. Turner states that:

[a]t this time the concept of ceiling values and excursion factors around the timeweighted average values was introduced in order to reduce conflict or confusion with the "maximal" values in the American [ANSI] Standards. A "C" (ceiling value) listing was to be given to those fast-acting substances thought likely to be injurious if the concentration exceeded the limit value by more than a designated factor for a relatively short period (about 15 min.). The factor varied between 3 and 1.25, depending inversely upon the magnitude of the TLV. A corollary was that the factor would also indicate the limit of permissive excursion of the concentration above the TLV for a substance not given a "C" listing, always provided that the time-weighted average concentration did not exceed the TLV. This rule of thumb approach to limiting exposure is no doubt appropriate to certain substances when they are used routinely throughout the working day. It seems to have little relevance in other instances where exposure is irregular or where the basis for fixing the TLV is on grounds other than toxicity (Turner 1976/Ex. 1-79).

Several commenters (Tr. pp. 6-30 to 6-31, 7-119, 8-139 to 8-141, and 8-167) were of the opinion that the ACGIH's procedures for establishing TLVs were not open to comment and that its reasons for selecting certain TLVs were not clear. Dr. Ernest Mastromatteo, Chairman of the ACGIH's TLV Committee, explained the organization's limit-setting process at the hearing (Tr. pp. 2-113 to 2-128). He stated that the Committee's minutes have recently been made public and explained that the committee often invited industry or union consultants to help the committee in its work on the TLVs (however, these consultants do not vote on the recommended limits). In addition, Dr. Mastromatteo described the ACGIH's process of placing new or revised limits on an "Intended" list for a period of two years, during which time comments on the proposed limits are invited, and considered.

Permissible Exposure Limits in the Era of OSHA

The enactment of the Occupational Safety and Health Act of 1970 marked the first "comprehensive and serious attempt...to protect the health and safety of American workers" (Mintz 1984/Ex. 1-840); it also greatly extended the use of MACs and TLVs by authorizing the newly established Occupational Safety and Health Administration (OSHA) to adopt as its own standards "national consensus standards" and established federal standards (29 USC 655(a)). Mintz notes that "in addition to the safety standards adopted under Section 6(a), OSHA also adopted permissible exposure limits for approximately 400 toxic substances. These [start-up] health standards, now appearing in 29 CFR 1910.1000,...were derived from both national consensus and established federal standards. The national consensus standards had been issued by ANSI, while the established federal standards had been adopted under the Walsh-Healey Act from the TLVs... recommended by the...ACGIH" (Mintz 1984/Ex. 1-840).

Since OSHA's large-scale adoption of the ANSI consensus standards and the 1968 ACGIH TLVs, the Agency has promulgated standards under Section 6(b) of the OSH Act to regulate the industrial use of 24 substances, most of which have been identified as occupational carcinogens, but the ANSI and ACGIH start-up standards continue to comprise the major part of the Agency's occupational health and safety program.

In the interval since the establishment of OSHA and the adoption of the ACGIH and ANSI limits by the Agency, the ACGIH has continued to revise, update, and document the recommended limits that appear in its annual list of TLVs. Since 1968, annual revisions have been made to these limits by the ACGIH. During this time, the TLVs have been "accepted on an international basis as the best available guides for providing healthful occupational environments, and at least 18 countries, including the United States, have either adopted them as legal standards or as guides to legal action, thus verifying their efficacy in accomplishing this purpose" (Paull 1984/Ex. 1-255).

The action OSHA takes today initiates the process of updating the Agency's Z-table permissible exposure limits. That these limits were seriously out of date is attested to by the fact that the ACGIH has found it necessary to revise or add nearly 400 limits to its list in the 20 years since the limits that were later adopted by OSHA were initially published. Recognition that OSHA's Z-table limits need updating to reflect recent developments in toxicology and new data on the health effects associated with exposure to these substances is widespread throughout industry: for example, OSHA's Hazard Communication standard (29 CFR 1910.1200) requires organizations that develop Material Safety Data Sheets (MSDSs) to include on these MSDSs the ACGIH's current TLV values as well as OSHA's limits.

The following section describes the methodology used by OSHA in selecting the limits it is promulgating today. The Agency believes that promulgation of these limits will address a broad range of significant risks now prevalent in industry. As many industrial hygienists and occupational safety and health professionals have noted, the use of permissible exposure limits continues to be the single most efficacious way of protecting the health, functional capacity, and well-being of the American worker.

C. Description of the Substances For Which Limits Are Being Established

In this rulemaking, OSHA considered revising 428 substances, and the final rule is revising existing or adding new limits for several hundred toxic substances currently being manufactured, used, or handled in workplaces throughout general industry. This section of the preamble identifies the PELs being established, describes the available toxicological data, and explains the Agency's rationale for selecting the final permisible exposure limits for these substances.

The universe of substances included in this rulemaking is bounded by the substances for which the American Conference of Governmental Industrial Hygienists (ACGIH) has established a Threshold Limit Value (TLV) for exposures in the work environment. That is, OSHA is not at this time establishing exposure limits for any hazardous substance that is not included in the ACGIH's 1987-88 List of TLVs. In addition, where the limit included in the current ACGIH list was identical to OSHA's existing Z-table limit for the same substance, OSHA did not consider revising its existing limit.

Although new limits are not being established for chemicals excluded from the ACGIH's 1987-88 list, OSHA has not limited its initial consideration of appropriate limits to those levels established by the ACGIH. The Agency has also carefully evaluated the exposure limits recommended by the National Institute for Occupational Safety and Health (NIOSH), OSHA's sister agency. In instances where both NIOSH and the ACGIH have recommended substantially different limits for the same substance, OSHA has thoroughly analyzed the evidence presented by each organization and has made its own judgment of the appropriate level at which to establish the PEL. In addition OSHA has fully considered levels recommended by commenters and levels supported by evidence. For all substances addressed in this rulemaking, OSHA has also evaluated the extensive record evidence. The limits being established today thus represent, in the Agency's professional judgment, those levels found to be most consistent with the best available toxicological data, OSHA's mandate, and the case law that has subsequently developed to interpret that mandate. (For a discussion of the relevant legislative and judicial principles, see the sections of this preamble entitled Pertinent Legal Authority, History and Need for Revision of the PELs, and Approach).

For ease of analysis and presentation, the substances included in the scope of this rulemaking have been grouped into 18 separate sub-sections. In general, these groupings reflect the primary basis underlying the ACGIH or NIOSH recommended limits for these substances. In addition, three additional sections cover substances for which the ACGIH has increased its limits, substances for which OSHA is adding short-term limits, and those for which the Agency is adding skin notations.

The following sections are included:

1. Substances for which Limits Are Based on Avoidance of Neuropathic Effects

2. Substances for which Limits Are Based on Avoidance of Narcotic Effects

3. Substances for which Limits Are Based on Avoidance of Sensory Irritation

4. Substances for which Limits Are Based on Avoidance of Liver or Kidney Effects

5. Substances for which Limits Are Based on Avoidance of Ocular Effect

6. Substances for which Limits Are Based on Avoidance of Respiratory Effects

7. Substances for which Limits Are Based on Avoidance of Cardiovascular Effects

8. Substances for which Limits Are Based on Avoidance of Systemic Toxicity

9. Substances for which Limits Are Based on Observed-No-Adverse-Effect Levels

10. Substances for which Limits Are Based on Avoidance of Physical Irritation and Other Effects

11. Substances for which Limits Are Based on Avoidance of Odor Effect

12. Substances for which Limits Are Based on Analogy to Related Substances

13. Substances for which Limits Are Based on Avoidance of Biochemical/Metabolic Effects

14. Substances for which Limits Are Based on Avoidance of Sensitization Effects

15. Substances for which Limits Are Based on Avoidance of Cancer

16. Substances for Which Current ACGIH TLVs Are Less Stringent than Former OSHA PELs

17. Substances for Which OSHA Is Establishing Short-Term Exposure Limits

18. Substances for Which OSHA Is Adding Skin Notations.

A list of the references that OSHA relied on in evaluating the toxicological evidence pertaining to these chemicals appears in Section VI-D.

1. Substances for Which Limits Are Based on Avoidance of Neuropathic Effects

Introduction

Many industrial chemicals have been shown to cause severe neurological effects in exposed workers, and in many cases these effects are irreversible. Limits have been set on the basis of avoidance of neuropathic effects for 20 substances. Table C1-1 lists the former, proposed, and final rule limits, CAS number, and OSHA HS number for each of these substances. The table shows time-weighted averages (TWAs), ceiling limits, and short-term exposure limits (STELs). For this group of 20 substances, OSHA is lowering its former TWA - PEL for three substances; adding a STEL to a former or a revised TWA for four substances; changing a ceiling to a TWA or a TWA to a ceiling for four substances; establishing permissible exposure limits for seven substances not formerly regulated by OSHA; retaining an existing TWA but changing its accompanying ceiling to a STEL for one substance; and lowering the former TWA and changing its accompanying ceiling to a STEL for one substance.

Description of the Health Effects

The human nervous system comprises the central nervous system (CNS) and peripheral nervous system (PNS). The CNS is made up of the brain and spinal cord, while the PNS consists of a network throughout the body of nerves that communicate with the CNS via connections to the spinal cord. The brain and spinal cord are bathed in cerebrospinal fluid, which supplies nutrients to the CNS and also acts as a barrier against some foreign substances. This barrier protects the central nervous system. In general, fat-soluble substances readily diffuse across this barrier and water soluble substances do not.

TABLE C1-1. Substances for Which Limits Are Based on Avoidance of Neuropathic Effects NOTE: Because of its width, this table has been divided; see continuation for additional columns. _____________________________________________________________________ H.S. Number/ Chemical Name CAS No. Former PEL _____________________________________________________________________ 1051 n-Butyl alcohol 71-36-3 100 ppm TWA 1078 Chlorinated camphene 8001-35-2 0.5 mg/m(3) TWA, Skin 1114 Decaborane 17702-41-9 0.05 ppm TWA, Skin 1116 Di-sec-octyl-phthalate 117-81-7 5 mg/m(3) TWA 1123 Dichloroacetylene 7572-29-4 -- 1149 Dipropylene glycol 34590-94-8 100 ppm TWA, methyl ether Skin 1200 n-Hexane 110-54-3 500 ppm TWA 1202 2-Hexanone 591-78-6 100 ppm TWA 1216 Iron pentacarbonyl 13463-40-6 -- (as Fe) 1236A Manganese, fume 7439-96-5 5 mg/m(3) (as Mn) Ceiling 1237 Manganese 12079-65-1 -- cyclopentadienyl tricarbonyl (as Mn) 1238 Manganese tetroxide 1317-35-7 -- (as Mn) 1240 Mercury (aryl and 7439-97-6 0.1 mg/m(3) TWA inorganic compounds) (as Hg) 1241 Mercury, vapor 7439-97-6 0.1 mg/m(3) TWA (as Hg) 1242 Mercury, (organo) 7439-97-6 0.01 mg/m(3) TWA alkyl compounds 0.04 mg/m(3) (as Hg) Ceiling 1251 Methylacrylonitrile 126-98-7 -- 1253 Methyl bromide 74-83-9 20 ppm Ceiling, Skin 1304 Pentaborane 19624-22-7 0.005 ppm TWA 1316 Phenyl mercaptan 108-98-5 -- 1342 1,2-Propylene glycol 6423-43-4 -- dinitrate _______________________________________________________________________

TABLE C1-1. Substances for Which Limits Are Based on Avoidance of Neuropathic Effects (Continuation) _______________________________________________________________________ H.S. Number/ Chemical Name Proposed PEL Final Rule PEL(1) _______________________________________________________________________ 1051 n-Butyl alcohol 50 ppm Ceiling 50 ppm Ceiling Skin 1078 Chlorinated camphene 0.5 mg/m(3) TWA, 0.5 mg/m(3) TWA, 1 mg/m(3) STEL, 1 mg/m(3) STEL, Skin Skin 1114 Decaborane 0.05 ppm TWA, Skin 0.05 ppm TWA, Skin 0.15 ppm STEL 0.15 ppm STEL 1116 Di-sec-octyl-phthalate 0.5 mg/m(3) TWA 0.5 mg/m(3) TWA, 10 mg/m(3) STEL 10 mg/m(3) STEL 1123 Dichloroacetylene 0.1 ppm Ceiling 0.1 ppm Ceiling 1149 Dipropylene glycol 100 ppm TWA, Skin 100 ppm TWA, Skin methyl ether 150 ppm STEL 150 ppm STEL 1200 n-Hexane 50 ppm TWA 50 ppm TWA 1202 2-Hexanone 5 ppm TWA 5 ppm TWA 1216 Iron pentacarbonyl 0.1 ppm TWA 0.1 ppm TWA (as Fe) 0.2 ppm STEL 0.2 ppm STEL 1236A Manganese, fume 1 mg/m(3) TWA 1 mg/m(3) TWA (as Mn) 3 mg/m(3) 3 mg/m(3) STEL 1237 Manganese 0.1 mg/m(3) TWA, 0.1 mg/m(3) TWA, cyclopentadienyl Skin Skin tricarbonyl (as Mn) 1238 Manganese tetroxide 1 mg/m(3) TWA 1 mg/m(3) TWA (as Mn) 1240 Mercury (aryl and 0.1 mg/m(3) Ceil. 0.1 mg/m(3) Ceil. inorganic compounds) Skin Skin (as Hg) 1241 Mercury, vapor 0.05 mg/m(3) TWA 0.05 mg/m(3) TWA (as Hg) Skin Skin 1242 Mercury, (organo) 0.01 mg/m(3) TWA 0.01 mg/m(3) TWA alkyl compounds 0.03 mg/m(3) STEL 0.03 mg/m(3) STEL (as Hg) Skin Skin 1251 Methylacrylonitrile 1 ppm TWA, Skin 1 ppm TWA, Skin 1253 Methyl bromide 5 ppm TWA, Skin 5 ppm TWA, Skin 1304 Pentaborane 0.005 ppm TWA 0.005 ppm TWA 1316 Phenyl mercaptan 0.5 ppm TWA 0.5 ppm TWA 1342 1,2-Propylene glycol 0.05 ppm TWA 0.05 ppm TWA dinitrate Skin _______________________________________________________________________ Footnote(1) OSHA's TWA limits are for 8-hour exposures; its STELs are for 15 minutes unles otherwise specified; and its ceilings are peaks not to be exceeded for any period of time.

Chemicals that affect the central nervous system may manifest their toxic effects peripherally. An example of this is the tremor associated with elemental and organic mercury poisoning. Exposure to some chemicals (for example, n-hexane) is associated with axonal degeneration of the nerves in both the central and peripheral nervous systems. Baker (1983/ Ex. 1-230) refers to this dual-system effect as centralperipheral distal axonopathy.

Nervous system toxicants can affect motor function, sensory function, or integrative processes, and they can also cause changes in the behavior of exposed persons. Substances that cause demyelination or neuronal damage can produce motor dysfunction that is expressed as muscular weakness or unsteadi-ness of gait, while exposures to chemicals that are associated with loss of sensory function may result in alterations in touch, pain, or temperature sensation or damage to sight or hearing. Other neuropathic chemicals affect the way in which information is processed in the brain and can interfere with learning and memory. All of the health effects described above constitute material impairments of health within the meaning of the Act.

Although mature neurons cannot divide and be replaced, the nervous system has considerable ability to restore function lost as a result of exposure to toxic chemicals. This capability to restore function even after neurons have been killed is achieved by two mechanisms: plasticity of organization and redundancy of function. That is, when some neurons die, other cells that perform the same function may be able to maintain an adequate level of functioning, or other neurons may be able to "learn" how to perform the lost function. However, even when one of these mechanisms comes into play to compensate for neuronal damage, the overall reserve capacity of the nervous system will have been diminished. The loss of this reserve could be critical in a situation in which additional demands are placed on the nervous system. Thus, even so-called reversible neuropathic effects should be seen as toxic effects causing alterations in and material impairment of the normal functioning of the nervous system.

The neurological effects potentially associated with chemical exposures are numerous, and it is not always easy to identify the precise target site. However, recent medical advances have made tests available that can detect neurological damage that was not detectable several years ago. For example, electrophysiological methods have been developed to measure damage to the visual pathway caused by such exposures. Because of the variation in individual responses to chemical exposures, exposure limits should be set with a view toward this range of susceptibility and the avoidance of any neuropathic effects.

Peripheral Nervous System Effects

The pathological mechanisms associated with peripheral neuropathies result from segmental demyelination or axonal degeneration. Segmental demyelination destroys the myelin sheath but leaves the axon intact; this causes a slowing in nerve conduction velocity. Muscle weakness is often the first sign of such segmental demyelination, and this effect can progress to a decline in motor function or paralysis. Although remyelination may occur within weeks after injury, even a temporary loss in motor or sensory function places the affected worker or others at risk of injury.

Axonal degeneration is a more serious effect in that recovery is often slow or incomplete. It causes demyelination secondary to the degeneration of the distal portion of the nerve. This effect occurs when a chemical interferes with the physiologic dynamics of the nerve, e.g., when it decreases the transport of nutrients to the nerve. The axon will degenerate (die back) sufficiently to accommodate the cell's capacity to supply it with nutrients. Axonal degeneration can also occur as a result of biochemical or metabolic derangement of the central nervous system. Alkyl mercury and elemental mercury are examples of chemicals causing this type of effect (Cavanaugh 1977/Ex. 1-202).

Central Nervous System Effects

The mechanism of action of central nervous system toxins is not well understood but is believed to be associated with neurochemical alteration in the brain. Seizures, Parkinsonism, intellectual impairment, narcosis, dementia, cranial neuropathy, and visual disturbances are all examples of effects that can occur after overexposures to neuropathic chemicals. The more serious CNS effects, such as Parkinsonism, dementia, intellectual impairment, and cranial neuropathy, are generally irreversible (Baker 1983/Ex. 1-230). Before these effects are manifested, subtle changes in behavior may occur; if these subtle signs are interpreted correctly, exposure can be stopped before irreversible damage occurs.

Dose-Response Relationships and Neuropathic Effects

The development of chemically induced neurological effects is believed to follow a dose-response pattern. At an exposure intensity or duration below the no-effect level, detectable effects are unlikely to be evident. As exposure intensity/ duration increases to and beyond this level, the toxin begins to interfere with the normal cellular processes of the neurological system. At this early stage, transient signs and symptoms may appear. Overt effects become more severe as exposure continues and finally progress to serious loss of neurological function and possible permanent damage to neural tissue. Increases in our ability to detect neurological changes at lower levels of exposure have shown that neurobehavioral changes or impairment may occur at levels previously thought to be innocuous. These early effects can be important indicators of potential functional impairment at exposure levels below those that produce either transient or permanent damage. Heavy metals, solvents, and pesticides are examples of chemicals that can cause symptoms that include nausea, sensory and motor function impairments, depression, sleep disturbances, cognitive impairment, and sexual dysfunction. Limits for substances in this group are generally designed to maintain worker exposures below the level associated with such symptoms. This approach ensures that employees will not be likely to suffer these material impairments of health and provides a margin of safety against the risk of more severe or permanent neurological impairment.

The following discussions describe the record evidence and OSHA's findings for all of the substances in this group and illustrate the material impairments of health faced by workers exposed to these toxicants.

n-BUTYL ALCOHOL CAS: 71-36-3; Chemical Formula: CH(3)CH(2)CH(2)CH(2)OH H.S. No. 1051

OSHA's former PEL for n-butyl-alcohol was a 100-ppm 8-hour TWA; the ACGIH limit is a 50-ppm ceiling, with a skin notation. The proposed and final rule PEL is a 50-ppm ceiling, with a skin notation. NIOSH (Ex. 8-47, Table N1) concurs that these limits are appropriate. n-Butyl alcohol is a colorless, highly refractive liquid with a mild vinous odor that has long been known to cause irritation of the eyes and headaches in occupational settings.

Systemic effects in the form of vestibular and auditory nerve injuries have been reported in workers in France and Mexico (Seitz 1972 and Velasquez 1964, both as cited in ACGIH 1986/Ex. 1-3, p. 76; Velasquez, Escobar, and Almaraz 1969/Ex. 1-1174). Contact dermatitis of the hands may occur due to the defatting action of liquid n-butyl alcohol, and toxic amounts can be absorbed through the skin. Based on data describing the rate of n-butyl alcohol uptake through the skin of dogs, DiVincenzo and Hamilton (1979, as cited in Patty's Industrial Hygiene and Toxicology, 3rd rev. ed., Vol. 2C, pp. 4571-78, Clayton and Clayton 1982) suggested that direct contact of human hands with n-butyl alcohol for one hour results in an absorbed dose that is four times that resulting from inhalation of 50 ppm for one hour.

The former OSHA limit of 100 ppm (TWA) was based on the studies of Tabershaw, Fahy, and Skinner (1944, as cited in ACGIH 1986/Ex. 1-3, p. 76) and of Smyth (1956/Ex. 1-759). These studies indicated that workers experienced no narcotic or systemic effects at levels lower than 100 ppm. However, irritation has been reported in humans exposed to 24 ppm; this irritation became uncomfortable and was followed by headaches at 50 ppm (Nelson, Enge, Ross et al. 1943/Ex. 1-66).

More recent data reported by Seitz (1972, as cited in ACGIH 1986/Ex. 1-3, p. 76), Velasquez (1964, as cited in ACGIH 1986/Ex. 1-3, p. 76), and Velasquez, Escobar, and Almaraz (1969/Ex. 1-1174) indicate serious exposure-related long-term systemic effects on the auditory nerve and hearing loss (hypoacusia); the magnitude of the hearing loss was related to length of exposure. Nine of 11 workers exposed without hearing protection to 80 ppm for periods of from 3 to 11 years displayed impaired hearing. This phenomenon was particularly evident in younger workers (Velasquez 1964, as cited in ACGIH 1986/Ex. 1-3, p. 76; Velasquez, Escobar, and Almaraz 1969/Ex. 1-1174).

Three commenters, ConAgra (Ex. 3-635), the Motor Vehicle Manufacturers Association (MVMA) (Ex. 3-902), and ARCO (Tr. p. 3-237) submitted comments on n-butyl alcohol. ConAgra (Ex. 3-635) misinterpreted OSHA's discussion of a 1964 study (Velasquez, as cited in ACGIH 1986/Ex. 1-3, p. 76) to mean that OSHA attributed all hearing loss found in the workers in this study to n-butyl alcohol exposure. ARCO (Tr. p. 3-237) also questioned n-butyl alcohol's effect on hearing. In response to these commenters, OSHA notes that n-butyl alcohol has been shown in many studies to damage the auditory nerve and further, that workplace noise may also have contributed to the hearing loss observed in these studies. The MVMA comment (Ex. 3-902) lists n-butyl alcohol as a substance for which rulemaking should be delayed, but provides no other details.

OSHA finds that the former PEL of 100 ppm is not sufficiently protective against the acute effects associated with exposure to n-butyl alcohol; in addition, the possibility of auditory nerve damage from exposures below the 100-ppm level makes the former PEL inadequate. A skin notation is necessary because data in beagle dogs suggest that dermal contact with n-butyl alcohol can result in a systemic dose greater than that obtained by inhalation (DiVincenzo and Hamilton 1979). The Agency is establishing a permissible exposure limit of 50 ppm as a ceiling, with a skin notation, for n-butyl alcohol. OSHA concludes that this limit will protect workers against the significant risks of possible vestibular and auditory nerve injury as well as of headaches and irritation, which constitute material impairments of health and are associated with exposure to this substance at levels above the new limit.

CHLORINATED CAMPHENE (60 Percent) CAS: 8001-35-2; Chemical Formula: C(10)H(10)Cl(8) H.S. No. 1078

Previously, OSHA had a limit of 0.5 mg/m(3), with a skin notation, for chlorinated camphene. The ACGIH has a TLV-TWA limit of 0.5 mg/m(3) and a TLV-STEL of 1 mg/m(3) for chlorinated camphene (60 percent), with a skin notation, and these were the limits proposed. The final rule retains the 0.5-mg/m(3) 8-hour TWA and the skin notation, and adds a 1-mg/m(3) STEL for chlorinated camphene, an amber waxy solid with a pleasant, pine-like odor.

Chlorinated camphene has demonstrated a moderately high acute toxicity in animal studies (ACGIH 1986/Ex. 1-3, p. 115). Toxic doses cause varied central nervous system effects, including nausea, muscle spasms, confusion, and convulsions (Hayes 1963/Ex. 1-982). Data indicate that rats and guinea pigs show no significant effects at dietary levels of 800 ppm daily for a six-month period (Alderson Reporting Co., as cited in ACGIH 1986/Ex. 1-3, p. 115). Monkeys tolerate daily feeding at 10 ppm but show toxic symptoms after two weeks' feeding at the 60-ppm level (Sosnierz, Szczurek, Knapek, and Kolodziejczyk 1972/Ex. 1-760). Although chlorinated camphene may accumulate in fatty tissues, it clears quickly when ingestion is terminated (Sosnierz, Szczurek, Knapek, and Kolodziejczyk 1972/Ex. 1-760).

In humans, the acute lethal dose of chlorinated camphene is between 2 and 7 grams, and a dose of 10 mg/kg causes nonfatal convulsions in some exposed individuals. The ACGIH (1986/Ex. 1-3, p. 115) concludes that the acute toxicity of chlorinated camphene is equivalent to that of chlordane, for which the fatal human dose is estimated to be around 6 grams; the ACGIH TLV-TWA for chlordane is 0.5 mg/m(3). One study of 25 human volunteers failed to reveal toxic responses to daily 30-minute exposures to 500 mg/m(3) for 10 consecutive days, followed by similar exposures for three consecutive days three weeks later (Shelansky 1947, as cited in ACGIH 1986/Ex. 1-3, p. 115). There are no reports of occupational poisonings, and a review of the medical records of employees engaged in the manufacture and handling of chlorinated camphene showed no ill effects in workers exposed for an average of 3.7 years (Frawley 1972, as cited in ACGIH 1986/Ex. 1-3, p. 115).

NIOSH does not concur with OSHA's PELs for this substance; NIOSH believes that chlorinated camphene is a potential occupational carcinogen and should have lower exposure limits (Ex. 8-47, Table N6B; Tr. pp. 3-97, 3-98). No other comments on the health effects of this substance were submitted to the record.

OSHA is retaining the 8-hour TWA PEL of 0.5 mg/m(3) TWA and adding a 15-minute STEL of 1.0 mg/m(3) for this insecticide. The Agency's skin notation is retained. OSHA concludes that both a TWA and a STEL are required to protect exposed workers against the significant risks of bioaccumulation and neuropathic and systemic effects; the Agency finds that these effects constitute material impairments of health. The STEL ensures that TWA exposures will be maintained under good industrial hygiene control.

DECABORANE CAS: 17702-41-9; Chemical Formula: B(10)H(14) H.S. No. 1114

OSHA's former limit for decaborane was 0.05 ppm TWA, with a skin notation. The ACGIH has a TLV-TWA of 0.05 ppm and a TLV-STEL of 0.15 ppm, also with a skin notation. The proposal retained the 8-hour TWA of 0.05 ppm and added a 0.15-ppm STEL, with a skin notation, and the final rule establishes these limits. NIOSH (Ex. 8-47, Table N1) concurs that these limits are appropriate. Decaborane forms colorless crystals that are stable at ordinary temperatures and have a pungent odor.

The acute toxicity of decaborane is extremely high for small laboratory animals. The 40-hour LC(50)s for rats and mice are 46 and 12 ppm, respectively (Schechter 1958/ Ex. 1-363). Dermal LD(50)s for rabbits and rats are 71 and 740 mg/kg, respectively (Svirbely 1954a/Ex. 1-385). Acute exposures to decaborane cause loss of coordination, convulsions, weakness, tremors, and hyperexcitability. Decaborane's primary effects are on the kidneys and liver. Studies of repeated exposures to this substance suggest that the toxicity of decaborane is intermediate between that of pentaborane and diborane. The ability of decaborane to penetrate the skin is particularly notable, as is its toxicity to the central nervous system in some species, e.g., rats and rabbits (Svirbely 1954a/Ex. 1-385, 1954b/Ex. 1-530, and 1955/Ex. 1-386). Monkeys showed decreased ability for certain operant behaviors when injected with doses of 3 to 6 mg/kg decaborane (Reynolds et al. 1964, as cited in ACGIH 1986/Ex. 1-3, p. 169). Central nervous system toxicity has been observed in humans exposed occupationally (Krackow 1953/Ex. 1-344). No comments other than NIOSH's were received on the health effects of decaborane.

OSHA is retaining its 8-hour TWA PEL of 0.05 ppm TWA and skin notation, and adding a 15-minute STEL of 0.15 ppm for decaborane. The Agency concludes that these limits will provide protection against the significant risks of material health impairment in the form of neuropathy and kidney and liver damage possible in the absence of a short-term limit for decaborane.

Di-sec-OCTYL PHTHALATE CAS: 117-81-7; Chemical Formula: C(24)H(38)O(4) H.S. No. 1116

OSHA formerly had a limit of 5 mg/m(3) TWA for di-sec-octyl phthalate. The ACGIH has a TLV-TWA of 5 mg/m(3) and a TLV-STEL of 10 mg/m(3), and these are the limits that were proposed. In the final rule, OSHA is retaining the 8-hour TWA limit of 5 mg/m(3) and adding a 15-minute STEL of 10 mg/m(3) for this light-colored, viscous, odorless, combustible liquid.

Di-sec-octyl phthalate (DEHP) is not acutely toxic in small laboratory animals via the oral route. The oral LD(5)0 reported for mice is 26.3 g/kg; for rats, it is 33.8 g/kg (Krauskopf et al. 1973/Ex. 1-495). No skin irritation or sensitization potential has been demonstrated in either animals or humans, and the lethal dermal dose in rabbits is about 25 ml/kg (Singh, Lawrence, and Autian 1972/Ex. 1-436). Shaffer, Carpenter, and Smyth (1945/Ex. 1-369) and Lawrence (unpublished data, as cited in ACGIH 1986/Ex. 1-3, p. 223) have reported deaths in rats and chronic diffuse inflammation of the lung in mice exposed to DEHP at unspecified levels.

Long-term dietary toxicity studies in rats, guinea pigs, and dogs have established a no-effect dose level of about 60 mg/kg/day, and no carcinogenic or histologic abnormalities were observed at this level (Gesler 1973/Ex. 1-481). Higher doses were associated with growth retardation and increased liver and kidney weights but not histologic abnormalities. Metabolic studies have demonstrated that laboratory animals do not appreciably metabolize DEHP (Dillingham and Autian 1973/ Ex. 1-477). Teratogenicity studies in pregnant rats indicated that fertility is unaffected at doses of 0.1, 0.2, or 0.33 percent of the acute intraperitoneal LD(50) dose for rats, although slight effects on embryonic and fetal development were observed in these animals; skeletal deformities were the most common teratogenic effects observed (Dillingham and Autian 1973/Ex. 1-477). Mutagenic effects were observed at intravenous doses of one-third, one-half, and two-thirds of the acute LD(50); these effects are consistent with DEHP's ability to produce dominant lethal mutations (Dillingham and Autian 1973/Ex. 1-477).

A study of workers exposed to a mixture of the vapors of diethyl phthalate, dibutyl phthalate, and di-2-ethylhexyl phthalate reported that exposures to 1 to 6 ppm caused no peripheral polyneuritis (Raleigh, as cited in ACGIH 1986/Ex. 1-3, p. 223). However, Russian investigators examined male and female workers exposed to between 1.7 and 66 mg/m(3) of various combinations of airborne phthalates (including butyl phthalate, higher aryl phthalates, dioctyl phthalate and others) and noted complaints of pain, numbness, and spasms in the upper and lower extremities after six to seven years of exposure. Polyneuritis was observed in 32 percent of the workers studied, and 78 percent of these workers showed depression of vestibular receptors (Milkov, Aldyreva, Popova et al. 1973/Ex. 1-646).

OSHA received a comment from the Chemical Manufacturers Association Phthalate Esters Program Panel (Ex. 3-900). Although the Panel did not oppose the proposed PEL for di-sec-octyl phthalate, it objected to this substance's categorization as a neuropathic agent on the grounds that (1) confounding exposures to tricresyl phosphate and vinyl chloride, which are known neurotoxicants, occurred in the study referenced in the NPRM; and (2) other studies (in humans or animals) have not substantiated that this substance is neuropathic:

Including [di-sec-octyl phthalate] in this category of compounds [i.e., neuropathic agents] is not justified and could lead to improper labeling of the material or unwarranted regulations, and restrictions on the use of the material based on unfounded conclusions (Ex. 3-900, p. 1).

In response to this comment, OSHA notes that the classification scheme used in the preamble to the proposed and final rules is not intended to have regulatory implications. As explained earlier in the preamble, OSHA is using this scheme simply to facilitate generic rulemaking; the various categories reflect the health endpoint used by the ACGIH or NIOSH as the point of reference in setting a limit. Most of the substances included in this rulemaking produce multiple health effects and could be classified in more than a single health effects category. Di-sec-octyl phthalate is no exception, and exposure to this substance has been associated with liver damage, testicular injury, and teratogenic and carcinogenic effects in experimental animals, as well as with possible neuropathic effects.

Another commenter, Lawrence H. Hecker of Abbott Laboratories, feels that the STEL for di-sec-octyl phthalate is unwarranted (Ex. 3-678, p. 8). OSHA disagrees with Dr. Hecker and finds that, for substances posing serious health hazards, such as those associated with di-sec-octyl phthalate exposure, the STEL further protects workers from the significant adverse effects that could occur in the short-term excursions above the TWA limit permitted in the absence of a STEL.

NIOSH concurs in OSHA's selection of limits for di-sec-octyl phthalate but believes it should be designated as a potential occupational carcinogen (Ex. 8-47, Table N6A). On the other hand, the Chemical Manufacturers Association's (Ex. 140) analysis of the evidence for DEHP's carcinogenicity led the CMA to conclude that this substance is not a carcinogen. OSHA is aware of di-sec-octyl phthalate's carcinogenic effects in experimental animals and notes that IARC has determined that sufficient evidence exists to designate it as an animal-positive carcinogen. However, adequate data are not available to evaluate the risk of cancer to humans. The Agency will continue to monitor the scientific evidence for di-sec-octyl phthalate and will re-evaluate this substance in the future if such evidence suggests that this is appropriate.

In the final rule, OSHA is retaining the 8-hour PEL of 5 mg/m(3) and adding a 15-minute STEL of 10 mg/m(3) for di-sec-octyl phthalate. The Agency concludes that these limits together will protect workers from the significant risks of neuropathic, hepatic, and other systemic injuries, which constitute material health impairments and are associated with exposure to this substance.

DICHLOROACETYLENE CAS: 7572-29-4; Chemical Formula: ClC - CCl H.S. No. 1123

OSHA previously had no limit for dichloroacetylene. The ACGIH has a TLV-ceiling of 0.1 ppm for this liquid, which explodes upon boiling. OSHA proposed a ceiling limit of 0.1 ppm, and this is the limit established by the final rule.

In preliminary inhalation exposure studies, guinea pigs demonstrated a 4-hour LC(50) of 20 ppm; death occurred two or three days after exposure and was caused by pulmonary edema. In rats, similar exposures to dichloroacetylene in the presence of 330 ppm of trichloroethylene indicated an LC(50) of 55 ppm (Siegel 1967, as cited in ACGIH 1986/Ex. 1-3, p. 177). When dichloroacetylene was mixed with 9 parts of ether, the 4-hour LC(50) in rats was 219 ppm; in combination with 7 parts of trichloroethylene, the 4-hour LC(50) in rats was 55 ppm; and exposure to dichloroacetylene with 10 parts of trichloroethylene caused a 4-hour LC(50) in guinea pigs of 15 ppm (Siegel, Jones, Coon, and Lyon 1971/Ex. 1-371).

In humans, dichloroacetylene exposure causes headache, loss of appetite, extreme nausea, and vomiting; it affects the trigeminal nerve and facial muscles and exacerbates facial herpes. Disabling nausea was experienced by approximately 85 percent of individuals exposed for prolonged periods of time (not further specified) at concentrations from 0.5 to 1 ppm (Saunders 1967/Ex. 1-361). A number of occupational fatalities have been attributed to exposure to dichloroacetylene (Humphrey and McClelland 1944/Ex. 1-491; Firth and Stuckey 1945, as cited in ACGIH 1986/Ex. 1-3, p. 177). Humphrey and McClelland (1944/Ex. 1-491) reported 13 cases of cranial nerve palsy, nine of which had labial herpes, following exposure to dichloroacetylene. These patients also had symptoms of nausea, headache, jaw pain, and vomiting. Autopsies of two of these fatalities revealed edema at the base of the brain (Humphrey and McClelland 1944/Ex. 1-491).

NIOSH concurs with OSHA's limit for dichloroacetylene but believes that this substance should be designated as a potential occupational carcinogen (Ex. 8-47, Table N6A). However, as explained elsewhere in the preamble, OSHA has decided not to designate substances specifically as carcinogens since so many other organizations already do so. OSHA received no other comments regarding the health effects of dichloroacetylene.

In the final rule, OSHA is establishing a ceiling limit of 0.1 ppm for dichloroacetylene. The Agency concludes that this limit will substantially reduce the significant risks of disabling nausea and serious systemic effects posed to workers exposed to dichloroacetylene at the levels formerly permitted by the absence of any OSHA limit. OSHA finds that these health effects constitute material impairments of health.

DIPROPYLENE GLYCOL METHYL ETHER CAS: 34590-94-8; Chemical Formula: CH(3)OC(3)H(6)OC(3)H(6)OH H.S. No. 1149

OSHA formerly had an 8-hour TWA limit of 100 ppm for dipropylene glycol methyl ether (DPGME), with a skin notation. The ACGIH recommends a TLV-TWA of 100 ppm and a TLV-STEL of 150 ppm, with a skin notation, for this colorless liquid with a mild, pleasant, ethereal odor and a bitter taste. OSHA proposed to retain the 8-hour permissible exposure limit of 100 ppm TWA, to add a 150-ppm STEL, and to retain the skin notation for dipropylene glycol methyl ether. NIOSH (Ex. 8-47, Table N1) concurs that these limits are appropriate, and the final rule establishes these limits.

Intact dogs receiving intravenous injections of DPGME exhibited central nervous system depression and died as a result of respiratory failure (Shideman and Procita 1951/Ex. 1-667). Rowe and associates (1954/Ex. 1-435) reported a single acute oral LD(50) for rats of 5.4 ml/kg. Even at the highest levels tested (not further specified), no single application of DPGME to the skin of rabbits was lethal, although some narcosis and transient weight loss did occur. However, a significant number of deaths occurred in a group of rabbits treated with 65 repeated dermal applications containing DPGME concentrations of 3 ml/kg or higher during a 90-day period. Four animal species, including the monkey, were exposed repeatedly to seven-hour daily inhalation exposures of between 300 and 400 ppm DPGME; the animals exhibited narcosis and changes in the lung and liver (Rowe, McCollister, Spencer et al. 1954/Ex. 1-435).

Humans inhaling DPGME concentrations of 300 to 400 ppm judged this level to be very disagreeable, but 100 ppm was tolerable and, in the opinion of the authors, was unlikely to produce organic injury (Rowe, McCollister, Spencer et al. 1954/Ex. 1-435). Patch tests on the skin of 250 human subjects produced neither irritation nor sensitization (ACGIH 1986/Ex. 1-3, p. 221). Humans exposed to DPGME vapor concentrations at levels between 50 to 2000 ppm experienced eye, nose, and throat irritation before the onset of CNS impairment, which occurred at 1000 ppm in one of two subjects (Stewart, Baretta, Dodd, and Torkelson 1970/Ex. 1-379).

NIOSH (Ex. 150, Comments on Dipropylene Glycol Monomethyl Ether) reported that it is developing a criteria document on the glycol ethers; NIOSH submitted recent toxicity data on DPGME, including the following: rats and mice inhaling concentrations of 50, 140, or 330 ppm DPGME six hours/day for nine days showed increased liver weights (at 50 and 140 ppm for the rat and at 330 ppm for the mouse), but no effects were observed when rats inhaled 15, 50, or 200 ppm DPGME six hours/day, five days/week for 13 weeks (Landry and Yano 1984, as cited in Ex. 150). NIOSH also reported results of a 1985 study by Miller et al. indicating that DPGME is metabolized via the same routes to the same types of metabolites - propylene glycol, and sulfate and glucuronide conjugates of DPGME - as previously identified for PGME (1-methoxy-2-propanol) (Miller, Hermann, Calhoun et al. 1985, as cited in Ex. 150). The Landry and Yano study (1984, as cited in Ex. 150) further indicated that at the concentrations tested, DPGME exerted no teratogenic or reproductive effects (NIOSH/Ex. 150, Comments on Dipropylene Glycol Monomethyl Ether).

The ARCO Chemical Company (Ex. 3-740) questioned the appropriateness of a skin notation for this substance. In response to ARCO, the Agency notes that DPGME, applied essentially according to the Draize method, is absorbed in sufficient quantities through rabbit skin to cause transient narcosis, although the absorption rate was not considered acutely dangerous (Patty's Industrial Hygiene and Toxicology, 3rd rev. ed., Vol. 2C, p. 3990, Clayton and Clayton 1982). Topical administration of 10 mg/kg DPGME five times per week for 13 weeks to shaved rabbit skin caused six deaths among seven animals (Chemical Hazards of the Workplace, 2nd ed., p. 221, Proctor, Hughes, and Fischman, 1988). To date, there are no human data demonstrating that dermal contact with DPGME is without a significant adverse health risk; therefore, in accordance with the policy described in Section VI.C.18, OSHA finds that the available evidence does not meet the criterion for deleting an existing skin notation.

In the final rule, OSHA is retaining a PEL of 100 ppm TWA and adding a STEL of 150 ppm for dipropylene glycol methyl ether; the skin notation is retained. The Agency concludes that this combined limit will substantially reduce the significant risks of central nervous system effects and irritation, which constitute material health impairments, that exist when workers are exposed to DPGME for short periods above the 100-ppm PEL.

n-HEXANE CAS: 110-54-3; Chemical Formula: CH(3)(CH(2))(4)CH(3) H.S. No. 1200

OSHA's former PEL for n-hexane was 500 ppm. The ACGIH has a 50-ppm TWA limit for this substance, and the NIOSH REL is 100 ppm as a 10-hour TWA. OSHA proposed a limit of 50 ppm TWA for n-hexane, and the final rule establishes this limit. NIOSH (Ex. 8-47, Table N1) concurs that a PEL of 50 ppm is appropriate for n-hexane. Normal hexane is a clear, volatile liquid.

n-Hexane has been shown to produce distal axonopathy in both experimental animals and humans; it is metabolized to 2,5-hexanedione (2,5-HD), which is thought to be the causative agent of most of the adverse neurological effects observed after exposure to hexane (Schaumburg, Spencer, and Thomas 1983/ Ex. 1-228).

In the preamble to the proposed rule, OSHA asked:

Does the most current scientific information generally support acceptance of the hypothesis that the C(5-8) alkanes are not equally toxic because a metabolite of n-hexane exhibits unique neurotoxic properties?

Several commenters (Exs. 3-593, 3-1246, and 124; Tr. III, pp. 109-110) responded to this question, and their detailed responses are discussed in Section V of this preamble, Summary of Commenters' Responses to NPRM Questions.

The C(5)-(8) alkanes include pentane, n-hexane, the hexane isomers, n-heptane, octane, and the refined petroleum solvents. Whether all of these alkanes exhibit the same degree of toxicity or whether one (or more) is uniquely toxic has a direct bearing on the appropriate exposure limits for these substances. Based on a thorough review of the chemical and toxicological literature and the responses of these commenters, OSHA has determined that n-hexane is uniquely toxic to the peripheral nervous system. The Agency finds that 2,5-hexanedione (2,5-HD), a metabolite of n-hexane, is likely to be responsible for this unique toxicity, and the American Petroleum Insitute (Ex. 124) agrees with this finding. NIOSH (Tr. III, pp. 109-110), on the other hand, is of the opinion that any ketone or related chemical that can be metabolized to a gamma diketone has the potential to cause peripheral neuropathy. However, representatives of the Texaco Company (Ex. 3-1246) agree with OSHA that n-hexane is uniquely toxic because its toxicity is mediated by 2,5-HD.

The ACGIH established a TLV of 50 ppm for this substance, based primarily on studies (Miyagaki 1967/Ex. 1-198; Inoue, Takeuchi, Takeuchi et al. 1970/Ex. 1-75) showing peripheral neuropathies at exposure levels as low as 210 ppm. NIOSH based its 100-ppm REL on the same studies as those cited by the ACGIH (Miyagaki 1967/Ex. 1-198; Inoue, Takeuchi, Takeuchi et al. 1970/Ex. 1-75). NIOSH reasoned as follows:

The absence of definitive epidemiologic or toxicologic evidence makes it difficult to determine how much lower the environmental limit should be. Professional judgment suggests [that] a TWA concentration of 350 mg/m(3) (100 ppm) offers a sufficient margin of safety to protect against the development of chronic nerve disorders in workers (NIOSH 1977a/ Ex. 1-233, p. 74).

The adverse neurological effects of hexane exposure are manifested as both sensory and motor dysfunctions. Initially, there is a symmetric sensory numbness of the hands and feet, with loss of pain, touch, and heat sensation. Motor weakness of the toes and fingers is often present; as the neuropathy becomes more severe, weakness of the muscles of the arms and legs may also be observed (Schaumburg, Spencer, and Thomas 1983/Ex. 1-228). There are no known conditions that predispose an individual to hexane neurotoxicity (Schaumburg, Spencer, and Thomas 1983/Ex. 1-228). The onset of neurological symptoms may not be evident for several months to a year after the beginning of exposure. Recovery may be complete, but severely exposed individuals often retain some degree of sensorimotor deficit.

OSHA received comments on n-hexane from several participants, including NIOSH, the National Cotton Council, the American Petroleum Institute, the Corn Refiners Association, the AFL-CIO, and the United Auto Workers. Two commenters, the National Cotton Council (Tr. pp. 9-45 to 9-47) and the Corn Refiners Association (Ex. 177), stated that the revised PEL for n-hexane would impact their members, but did not provide further detail.

Some commenters (Exs. 194 and 197: Tr. pp. 3-290 to 3-293) urged OSHA to regulate all of the refined petroleum solvents on the basis of neurotoxicity. For example, the AFL-CIO recommended a 10-ppm PEL for all such solvents, and Dr. Franklin Mirer of the United Auto Workers described feasible controls that could be used, in his opinion, to achieve this level. Dr. Philip Landrigan (Tr. pp. 3-290 to 3-293) described the neurotoxic effects of exposure to any of the refined petroleum solvents. In response to these commenters, OSHA notes that it is reducing the limits for a number of these solvents in this rulemaking; however, the scale of this undertaking is such that OSHA was unable to perform the detailed analysis necessary to evaluate the health effects, risks, and feasibility for all of the solvents in this large group of substances.

The dose-response relationship for n-hexane exposure in humans is not well defined, although it is clear that the severity of the resulting neuropathy increases as the exposure level of n-hexane increases. A number of studies have shown a consistent relationship between exposure levels of 500 ppm (OSHA's former exposure limit) to 2000 ppm and the development of characteristic peripheral neuropathies (Yamamura 1969, as cited in ACGIH 1986/Ex. 1-3, p. 305; Yamada 1967/Ex. 1-192). Neuropathic effects have also been shown to occur at levels between 210 and 500 ppm (Takeuchi, Maluchi, and Takagi 1975/Ex. 1-217).

Reports of effects occurring at levels of 210 to 500 ppm indicate that the former OSHA PEL of 500 ppm was not adequate to protect exposed workers from adverse sensorimotor neuropathic effects, and exposure at this level thus represents a significant risk to workers. The decreased sensitivity to pain, touch, and temperature associated with n-hexane exposure can also make a worker more susceptible to injuries and accidents. Further, the delayed onset of a clinical response, which is typical of hexane exposure, increases the probability that exposure will continue until irreversible effects occur.

Both the presence of peripheral neuropathies at 210 ppm and the delay in onset of neurological symptoms indicate that workers exposed at levels above the new limit are at significant risk of developing these symptoms. OSHA therefore establishes a PEL of 50 ppm TWA for n-hexane. The Agency concludes that this PEL will substantially reduce the significant risk of peripheral neurophathies and other adverse neuropathic effects, which constitute material impairments of health and are associated with the exposures permitted at levels above the new limit.

2-HEXANONE (METHYL n-BUTYL KETONE) CAS: 591-78-6; Chemical Formula: CH(3)CO - CH(2)CH(2)CH(2)CH(3) H.S. No. 1202

OSHA's former PEL for 2-hexanone was 100 ppm TWA; the NIOSH REL is a 1 ppm (10-hour) TWA; and the ACGIH recommends a TLV-TWA of 5 ppm. The Agency proposed, and the final rule establishes, a permissible exposure limit of 5 ppm as an 8-hour TWA for 2-hexanone. 2-Hexanone is a colorless, volatile liquid with a characteristic acetone-like odor that is more pungent than that of acetone.

Industrial exposure to 2-hexanone causes distal neuropathy manifesting as interference with motor and sensory function; even in cases characterized by minimal intensity, electrodiagnostic abnormalities were seen (ACGIH 1987/Ex. 1-16). In animals, exposure to 2-hexanone causes axonal swelling and thinning of the myelin sheath. A metabolite of 2-hexanone, 2,5-hexanedione, appears to be responsible for the neural damage; this same metabolite is formed when n-hexane (discussed above) is metabolized. Exposures of rats, cats, dogs, monkeys, hens, and guinea pigs to 2-hexanone have resulted in peripheral neuropathies (O'Donoghue 1985). Krasavage, O'Donoghue, and Terhaar (1978) reported that 2,5-hexanedione is 3.3 times more neurotoxic than 2-hexanone and 38 times more neurotoxic than n-hexane in rats. Thus, 2-hexanone would be approximately eleven times more neurotoxic than n-hexane in rats.

The limit of 5 ppm TWA for 2-hexanone recommended by the ACGIH is based on the results of several studies. These include studies showing evidence of peripheral neuropathy at levels of 50 ppm and above after exposures lasting six months or more (Johnson, Anger, Setzer et al. 1979/Ex. 1-984; Streletz, Duckett, and Chambers 1976/Ex. 1-1067). Another study identified 2,5-hexanedione (the metabolite believed responsible for neurotoxic effects) in the serum of humans after a one-day exposure to 50 ppm (DiVincenzo, Kaplan, and Dedinas 1976/Ex. 1-1049).

The NIOSH REL for 2-hexanone of 1 ppm (10-hour TWA) is based on an epidemiologic study describing an outbreak of neurologic disease among workers in a plant that manufactures printed fabrics (Allen, Mendall, Billmaier et al. 1975/ Ex. 1-80). This study reported that a screening of 1,157 exposed workers revealed 86 verified cases of distal neuropathy. 2-Hexanone was suspected of being the neurotoxicant because it had only recently been introduced into the process (Allen, Mendall, Billmaier et al. 1975/Ex. 1-80). When recommending its limit, NIOSH relied on an industrial hygiene survey of the plant conducted by Billmaier, Yee, Allen et al. (Ex. 1-76) in 1974, which showed that 2-hexanone concentrations near the textile printing machines ranged from 1 to 156 ppm (10-minute area samples). After reviewing this evidence, NIOSH concluded that 1 ppm could not be considered a no-effect level for 2-hexanone-induced neuropathy, and NIOSH (Ex. 8-47, Table N2; Tr. p. 3-86) continues to recommend a limit of 1 ppm TWA for 2-hexanone. The AFL-CIO (Ex. 194) also supports the adoption of the lower NIOSH REL. Dr. Franklin Mirer of the AFL-CIO (Ex. 197) described controls for use in workplaces where solvents present exposure problems.

The ACGIH (1987/Ex. 1-16) stated that interpretation of the results of the Billmaier, Yee, Allen et al. (1974/Ex, 1-76) study was complicated because the exposure measurements reported in the study had been taken after the outbreak of neuropathic effects had occurred. In addition, the ACGIH pointed out that Billmaier and colleagues (1974/Ex. 1-76) found poor work practices at the plant (gloves were rarely used, employees washed their hands with solvent, etc.); thus, dermal exposure may have contributed substantially to the outbreak.

Both human and animal studies show the development of disease at exposure levels well below the former 100-ppm PEL, clearly indicating the need to reduce this significant risk. In the final rule, OSHA is establishing a 5-ppm (8-hour TWA) PEL for 2-hexanone. The Agency concludes that this limit will substantially reduce the significant risk of distal neuropathy, which constitutes a material impairment of health and has been demonstrated to occur at concentrations above the new limit.

IRON PENTACARBONYL CAS: 13463-40-6; Chemical Formula: Fe(CO)(5) H.S. No. 1216

OSHA previously had no exposure limit for iron pentacarbonyl. The ACGIH has a TLV-TWA of 0.1 ppm with a TLV-STEL of 0.2 ppm, measured as iron, for this highly flammable, oily, colorless to yellow liquid. The Agency proposed, and the final rule establishes, permissible exposure limits of 0.1 ppm TWA and 0.2 ppm STEL for iron pentacarbonyl, measured as Fe. NIOSH (Ex. 8-47, Table N1) concurs that these limits are appropriate.

In studies of rats, iron pentacarbonyl has been reported to have approximately one-third the acute toxicity of nickel carbonyl (for which the ACGIH has recommended a TLV of 0.05 ppm TWA) (Sunderman, West, and Kincaid 1959/Ex. 1-384). In 1970, Gage found that a 5.5-hour exposure at 33 ppm caused fatalities in three of eight rats; four of eight animals died after two 5.5-hour exposures at 18 ppm. At 7 ppm, no ill effects were observed in rats exposed 18 times in 5.5 hours (Gage 1970/ Ex. 1-318). There are no reports of long-term dose-response exposure studies in laboratory animals, and no evidence exists that iron pentacarbonyl is carcinogenic in either humans or animals (ACGIH 1986/Ex. 1-3, p. 327).

Immediate symptoms of acute exposure to high concentrations of iron pentacarbonyl include headache and dizziness, followed in 12 to 36 hours by fever, cyanosis, cough, and shortness of breath. Another clinical effect of overexposure to this substance is lung injury, and degenerative changes in the central nervous system have also been reported (ACGIH 1986/Ex. 1-3, p. 327). No comments (other than NIOSH's) on the health effects of iron pentacarbonyl were submitted to the rulemaking record.

In the final rule, OSHA establishes a permissible exposure limit of 0.1 ppm TWA and a STEL of 0.2 ppm for iron pentacarbonyl. The Agency concludes that these limits will protect workers from the significant risks of material health impairment in the form of headache, dizziness, fever, dyspnea, cyanosis, pulmonary injury, and central nervous system effects, which are potentially associated with exposures at levels above the new limits.

MANGANESE FUME CAS: 7439-96-5; Chemical Formula: MnO H.S. No. 1236a

OSHA previously had a ceiling limit of 5 mg/m(3) for manganese fume, measured as manganese. Because of this substance's potential for damage to the lungs and central nervous system, the ACGIH recommends an 8-hour TWA of 1 mg/m(3) and a 3-mg/m(3) STEL for manganese fume. These limits were proposed and are now established by the final rule. NIOSH (Ex. 8-47, Table N1) concurs that these limits are appropriate.

Symptoms of manganese poisoning range from sleepiness and weakness in the legs (Fairhall 1957a, as cited in ACGIH 1986/ Ex. 1-3, p. 354) to difficulty in walking and uncontrolled laughter (Fairhall and Neal 1943, as cited in ACGIH 1986/ Ex. 1-3, p. 354). Health surveys of employees exposed to manganese fume have demonstrated a high incidence of pneumonia in these workers (Davies 1946, as cited in ACGIH 1986/Ex. 1-3, p. 354). Tanaka and Lieben (1969/Ex. 1-388) found seven cases of pneumonia and 15 borderline cases of pneumonia among 144 workers exposed to manganese dust or fume concentrations greater than 5 mg/m(3); three of these cases were associated with fume rather than dust exposure. Those workers exposed to fume levels below 5 mg/m(3) exhibited no signs of pneumonia. In a separate study by Smyth, Ruhf, Whitman, and Dugan (1973/Ex. 1-990), three cases of manganese poisoning were detected among 71 employees exposed to levels of 13.3 mg/m(3) manganese fume.

OSHA received several comments on manganese fume and dust (Exs. 3-189, 3-673, 3-675, 3-829, 8-22, and 129). Some commenters stated that reducing the PEL for manganese fume would have a large impact on their industries but did not provide any details (Exs. 3-673, 3-675, and 8-22). Chemetals, Inc., a manufacturer of manganese products, supports the reduction in the PEL for manganese fume from a ceiling of 5 mg/m(3) to an 8-hour TWA of 1 mg/m(3) and a STEL of 3 mg/m(3). According to Chemetals:

[We] agree that the fumes of metals and their compounds have higher toxicities than the dusts....and that a time-weighted average is more appropriate for manganese (Ex. 3-189, p. 2).

However, Chemetals urged OSHA to also revise the Agency's limit for manganese dust from a ceiling to an 8-hour TWA (Ex. 3-189). OSHA did not propose a change to its existing 5-mg/m(3) ceiling limit for manganese dust. In response to this comment, OSHA notes that manganese dust is not a substance that is included in this rulemaking; the Agency did not propose to regulate manganese dust and is not revising its limits for this substance in the final rule (see the preamble section entitled "Boundaries to the Regulation").

One other commenter, the Specialty Steel Industry of the United States (Ex. 3-829), stated that, in its opinion, there was no basis for reducing OSHA's former PEL for manganese fumes or for supplementing this limit with a STEL. OSHA does not agree with the views of this commenter, because exposures to these fumes have been demonstrated to cause toxic effects in both humans and animals. Workers exposed to manganese fumes developed pneumonia (Tanaka and Lieben 1969/Ex. 1-388), and Stokinger (1981f, in Patty's Industrial Hygiene and Toxicology, 3rd rev. ed., Vol. 2A, p. 1767) reports that the 1-mg/m(3) limit "is supported by the finding in animals that the higher oxides are more toxic, and the report of an occasional case of Mn poisoning in susceptible workers exposed to ferro Mn fumes around the 1-mg/m(3) limit."

Based on a review of all of the record evidence, the final rule establishes a 1-mg/m(3) TWA and a 3-mg/m(3) STEL for manganese fume. The Agency concludes that both a TWA limit and a STEL are required to protect employees from the significant risks of manganese poisoning, lung damage, and pneumonia, all of which constitute material health impairments, associated with exposure to these fumes.

MANGANESE CYCLOPENTADIENYL TRICARBONYL CAS: 12079-65-1; Chemical Formula: C(5)H(5) - Mn(CO)(3) H.S. No. 1237

OSHA formerly had no limit for exposure to manganese cyclopentadienyl tricarbonyl (MCT). The ACGIH has a TLV-TWA of 0.1 mg/m(3) (measured as manganese), with a skin notation. The Agency proposed, and the final rule establishes, a permissible exposure limit of 0.1 mg/m(3) TWA (measured as manganese), with a skin notation, for this substance. NIOSH (Ex. 8-47, Table N1) concurs that these limits are appropriate.

A Russian study reported that a single two-hour exposure to MCT at 120 mg/m(3) was fatal to 80 percent of albino rats, although rabbits, guinea pigs, and rats survived a single two-hour exposure at 20 to 40 mg/m(3). Chronic exposure of rats for 11 months at levels averaging 1 mg/m(3) for four hours daily showed delayed effects (seven months from onset of exposure) of neuromuscular excitability, evidence of kidney damage, and decreased resistance to infection (Arkhipova, Tolgskaya, and Kochetkova 1965/Ex. 1-1046). The tails of 10 white mice were dipped in a gasoline mixture containing 1 gram MCT per 100 ml; a second group of mice had their tails immersed in gasoline without MCT. An equal number of fatalities were observed in the gasoline plus MCT and gasoline only groups after four or five two-hour applications, and all tails exhibited necrosis. The authors concluded that these effects were caused by the gasoline and not by the MCT (Arkhipova, Tolgskaya, and Kochetkova 1965/Ex. 1-1046). Further studies in rabbits showed that MCT applied dermally as an oil emulsion caused irritation of the skin. These authors also investigated the dermal toxicity of MCT solutions in tetrahydrofuran versus solutions of tetrahydrofuran in oil. All animals whose tails had been dipped in the hydrofuran solution of MCT died within an hour, while animals whose tails had been dipped in pure tetrahydrofuran did not (Arkhipova, Tolgskaya, and Kochetkova 1965/Ex. 1-1046). The same authors concluded that MCT is toxic at low concentrations, has cumulative properties, affects the nervous system, is irritating to the skin, and causes early histological changes in the respiratory tract.

More recent reports describe MCT-induced pulmonary edema and convulsions in the rat (Penney, Hogberg, Traiger, and Hanzlik 1985/Ex. 1-431). The ED(50s) for convulsions were 32 mg/kg orally and 20 mg/kg intraperitoneally; LD(50s) were 24 mg/kg orally and 14 mg/kg intraperitoneally. Necrosis of the bronchiolar tissue and pulmonary parenchymal damage were seen in mice and rats given intraperitoneal doses (Haschek, Hakkinen, Witschi et al. 1982/Ex. 1-1083). No comments other than NIOSH's were received on MCT.

OSHA has concluded that occupational exposure to MCT poses a risk of neuropathic effects, kidney damage, skin irritation, pulmonary edema, and tissue damage, which together constitute material health impairments. The Agency is therefore establishing an 8-hour TWA PEL of 0.1 mg/m(3) for manganese cyclopentadienyl tricarbonyl, with a skin notation, to protect workers against the significant risk of these effects, which have been shown to occur at levels above the new standard.

MANGANESE TETROXIDE CAS: 1317-35-7; Chemical Formula: Mn(3)O(4) H.S. No. 1238

OSHA previously had no exposure limit for manganese tetroxide (compound and fume). The ACGIH recommends a TLV-TWA of 1 mg/m(3), measured as manganese, for this brownish-black powder and its dust and fume. The Agency proposed a PEL of 1 mg/m(3) TWA for manganese tetroxide, measured as Mn, and the final rule establishes this limit. Ferromanganese fume has been determined by X-ray diffraction analysis to consist primarily of manganese tetroxide.

Findings from a Russian study indicated that intratracheal suspensions of manganese oxide, manganese dioxide, and manganese tetroxide particles (particle size less than 3 um) produced pneumonitis and other similar pulmonary effects in rats (Levina and Robachevskiau 1955/Ex. 1-1041). These investigators also determined that manganese tetroxide has a greater toxicity than do the lower oxides of manganese and that freshly prepared oxides were more potent than those stored for six months to one year.

Two cases of manganese fume poisoning were reported in a plant where concentrations were between 2.7 and 4.7 mg/m(3) (Whitlock, Amuso, and Bittenbender 1966/Ex. 1-455), but other investigators have questioned these air sampling results and believe that exposures to manganese tetroxide concentrations of 5 mg/m(3) or less cause no harmful effects (Whitman and Brandt 1966/Ex. 1-1103). In a seven-year study, Smyth and co-workers (1973/Ex. 1-990) investigated chronic manganese poisoning in workers exposed to both ferromanganese fumes and dust. Five of 71 employees suffered from chronic manganism; of these five cases, three resulted from fume exposure and two from dust exposure. Two of the three fume-exposure victims were exposed over a five-year period to an estimated average ferromanganese concentration of 13.3 mg/m(3); however, the third victim worked in an operation where air concentrations of manganese were less than 1 mg/m(3), which suggests that certain individuals may be hypersusceptible to manganese poisoning. The dust-exposed victims worked in areas where air concentrations were in the range of 30 to 50 mg/m(3) throughout the study period (Smyth, Ruhf, Whitman, and Dugan 1973/Ex. 1-990).

Martonik (1976, as cited in ACGIH 1986/Ex. 1-13, p. 357) reported that the fume of manganese has greater toxicity than does the dust. During a two-year period, at least one case of acute manganese poisoning was documented at a fume concentration level of 7.5 mg/m(3), and another case at the same welding operation may also have involved manganism.

OSHA received two comments on this substance, one from NIOSH (Ex. 8-47; Tr. p. 3-86), and one from Chemetals, a manganese manufacturer (Ex. 3-189). NIOSH (Ex. 8-47, Table N2) does not concur with the limits being established by OSHA. NIOSH (Ex. 8-47, Table N2) notes that, based on the results of the Smyth and co-workers study (1973/Ex. 1-990), the 1-mg/m(3) PEL being established by OSHA "may not be protective, especially to the potentially sensitive individual." In response to this NIOSH comment, OSHA states that the Agency intends to monitor the literature on manganese tetroxide closely in the future to determine whether the new limit for this substance is adequately protective.

Chemetals (Ex. 3-189) asked OSHA to promulgate separate limits for the dust and fume of manganese tetroxide based on the relative toxicities of these two particulate forms. OSHA recognizes that some information in the literature (including some discussed above) points to the greater toxicity of the fume and that fumes are generally the more toxic form of particulate. However, the Agency notes that intratracheal suspensions of manganese tetroxide dust caused pneumonitis and other pulmonary effects in Russian workers (Levina and Robachevskiau 1955/Ex. 1-1041) and that several cases of manganism have been caused by dust exposure (Smyth, Ruhf, Whitman, and Anger 1973/Ex. 1-990). The Agency believes it prudent not to distinguish at this time between the dust and the fume but to set the TWA PEL at a level that will protect against the effects of exposure to both forms of particulate.

OSHA is establishing a 1-mg/m(3) 8-hour TWA for manganese tetroxide (compound and fume). The Agency concludes that this limit will provide protection against the significant risks of material health impairment in the form of chronic manganese poisoning, pneumonitis, and other respiratory effects that are associated with exposure to manganese tetroxide at levels above 1 mg/m(3).

MERCURY (ARYL AND INORGANIC COMPOUNDS) CAS: 7439-97-6; Chemical Formula: Hg H.S. No. 1240

The former OSHA limit for all inorganic forms of mercury (Hg) was 0.1 mg/m(3) as a ceiling limit, as indicated on Table Z-2; this limit was adopted from ANSI standard Z37.8 (1943). In a compliance directive issued in 1978 (OSHA Instruction CPL 2-2.6), however, the Agency stated that the PEL for inorganic mercury should be expressed as an 8-hour TWA of 1 mg/10 m(3) (0.1 mg/m(3)) rather than as a ceiling. The ACGIH has a 0.1-mg/m(3) TLV-TWA for aryl and inorganic mercury compounds. NIOSH (1973b, as cited in ACGIH 1986/Ex. 1-3, p. 358) has recommended a 0.05-mg/m(3) limit as an 8-hour TWA. OSHA proposed to return to its 0.1-mg/m(3) ceiling limit (measured as mercury) and this limit is being established, together with a skin notation, in the final rule. This action cancels the 1978 compliance directive.

In 1971, shortly after OSHA had adopted the 0.1-mg/m(3) ceiling, the ACGIH reduced the TLV-TWA for all forms of mercury, including the inorganic compounds, to 0.05 mg/m(3). ANSI also reduced its standard to 0.05 mg/m(3) in 1972, and NIOSH recommended the same limit in 1973. The 0.05-mg/m(3) limit was based largely on the study of Smith, Vorwald, Patil, and Mooney (1970/ Ex. 1-373) of workers exposed to mercury levels ranging from less than 0.1 to 0.27 mg/m(3) in chlor-alkali plants. The authors reported a significant dose-related increase in the incidence of weight loss, tremors, abnormal reflexes, nervousness, and insomnia among workers exposed to concentrations of 0.1 mg/m(3) or more. There were slight increases in incidences of insomnia and loss of appetite among workers exposed to 0.1 mg/m(3) or less. Smith, Vorwald, Patil, and Mooney (1970/Ex. 1-373) concluded that a limit of 0.1 mg/m(3) contained little or no margin of safety. Other studies (Bidstrup, Bonnell, Harvey, and Locket 1951/Ex. 1-1014; Turrian, Grandjean, and Turrian 1956, as cited in ACGIH 1986/ Ex.1-3, p. 358) have also reported symptoms of mercury poisoning among workers exposed below 0.1 mg/m(3). The 0.05-mg/m(3) limit established by the ACGIH, ANSI, and NIOSH also follows the 1968 recommendation of an international committee (Permanent Commission & International Association on Occupational Health 1968, as cited in ACGIH 1986/Ex. 1-3, p. 358).

In 1980, the ACGIH revised its recommended TLV for aryl and inorganic mercury compounds to 0.1 mg/m(3). In revising this limit, the ACGIH cited discrepancies in the literature regarding the ratio of blood and urinary mercury levels to airborne concentrations of mercury (Bell, Lovejoy, and Vizena 1973/Ex. 1-1078; Stopford et al. 1978/Ex. 1-1100). These studies reported lower ratios of mercury body burden to airborne concentration when personal sampling is used rather than area sampling. According to Bell, Lovejoy, and Vizena (1973/Ex. 1-1078), the lower ratio results because mercury exposure measurements are generally found to be higher when personal sampling is conducted, presumably as a consequence of contamination of clothing. The ACGIH argued that the 0.05-mg/m(3) limit may be too stringent to apply when personal sampling is conducted. The ACGIH also stated that, in cont