- Email: [email protected]
Risk Analysis
Chapter 2
Risk Analysis In a nutshell, the job of the environmental engineer is to reduce the risks of environmental hazards, both long- and short-term. To accomplish this objective in an orderly fashion, the risks associated with various hazards must be quantified and evaluated. In this chapter, risk analysis is introduced as a necessary tool of environmental engineering, crossing all media boundaries.
RISK
A number of laws regulate release of pollutants into the environment. However, major issues must be addressed. For instance, how is a pollutant identified as such, so that steps may be taken to control it? A substance is considered a pollutant because it is perceived to have an adverse effect on the environment and, in particular, an adverse effect on human health. However, it is sometimes difficult to determine whether there is an effect, or whether the effect has been deleterious or detrimental to the organism that is affected. For example, we are now certain that cigarette smoke is unhealthy. Specifically, we have identified inhaled cigarette smoke as contributing significantly to lung cancer, chronic obstructive pulmonary disease, and heart disease. Notice that we do not say that cigarette smoking causes these health problems, because we have not identified the causes—the etiology—of any of them, at least in the sense that we have identified the poliomyelitis virus as the cause of polio. How, then, has cigarette smoking been identified as a contributing factor, if it cannot be identified as the cause? Cigarette smoking serves as a good example of how the health effects of pollutants are determined, although cigarette smoke is not regulated the way other pollutants are regulated. In the latter half of the twentieth century, with the discovery of antibiotics and sulfa drugs, infectious diseases ceased to be a primary cause of death. The life span of people in the developed countries of the world lengthened consid11
12
ENVIRONMENTAL ENGINEERING
erably, and heart disease and cancer became leading causes of death. It was thus observed, in the early 1960s, that lifelong heavy cigarette smokers often died from lung cancer.
EXPRESSION OF RISK
Let us investigate the details of such an observation: how do you identify the cause of death, and how is it related to a habit like heavy smoking? The cause of death is listed on death certificates, which are the source of much epidemiological information. To relate death from lung cancer to smoking, one must show that more lung cancer deaths occur in smokers than in nonsmokers. Such a showing can be made by determining the standard mortality ratio (SMR), which is defined as SMR = observed deaths/expected deaths The number of "expected deaths" in the above equation is the number of deaths from the particular disease, lung cancer, in this case, that happen without any identifiable cause. In the general population of smokers and nonsmokers, there are a certain number of lung cancer deaths. Even in nonsmokers, there are a certain, albeit small, number of lung cancer deaths. In this instance, then, the SMR may be defined as SMR = D S /D n s where
D s = lung cancer deaths in a population of smokers D ns = lung cancer deaths in a nonsmoking population of the same size
In this instance, the SMR is approximately 1 1 / 1 . Since the SMR is greater than 1.0, then we can say that cigarette smoke contributes to lung cancer or, as it is usually phrased, smoking cigarettes increases the risk of death from lung cancer. To be precise, the risk of death from lung cancer is 11 times as high for a heavy smoker as for a nonsmoker. Determination of the SMR tells something about the epidemiology of smoking, but not about its etiology. There are three characteristics of epidemiological reasoning in this example that are important: • • •
Everyone who smokes heavily will not die of lung cancer. Some nonsmokers die of lung cancer. Therefore, one cannot unequivocally relate any given individual lung cancer death to cigarette smoking.
Risk may be expressed in other ways as well. For example, in the United States there are 350,000 deaths each year from lung cancer and heart disease
Risk Analysis
13
Table 2-1. Adult Deaths/105 for 1985 Deaths/10s Population
Cause of Death Cardiovascular disease Cancer Chronic obstructive pulmonary disease Motor vehicle accidents Alcoholism
408 193 31 18.6 11.3
All causes
870
that are attributable to smoking. The United States has a population of 226 million. The risk of death associated with the effects of cigarette smoking may thus be expressed as deaths per 100,000 population, or 350,000/226 x 10 6 = 155/100,000 In other words, a heavy smoker in the United States has an annual probability (or risk) of 155 in 100,000, or 1.55 in 1,000, of dying of lung cancer or heart disease. Table 2 - 1 presents some typical statistics for the United States.1 A third way to present risk is as deaths per 1,000 deaths. For example, in 1985, there were 2.084 x 10 6 deaths in the United States. Of these, 350,000 or 168 deaths per 1,000 deaths, were related to heavy smoking. These figures show that meaningful risk analyses can be conducted only with very large populations. Moreover, the health risk posed by most pollutants is observed to be considerably lower than the risks cited in Table 2—1 and may simply not be observed in a small population. Chapter 18 cites several examples of statistically valid risks from air pollutants. The Environmental Protection Agency (EPA) has adopted the concept of unit risk in discussions of potential risk. Unit risk is defined as the risk to an individual from exposure to a concentration of 1 jug/m3 of an airborne pollutant, o r l O ~ 9 g / L o f a water borne pollutant. Unit lifetime risk is the risk to an individual from exposure to the above concentration for 70 years, or a lifetime, while unit occupational risk implies exposure for 40 hours per week for 50 years, or a working lifetime. Example 2.1 The EPA has calculated that the unit lifetime risk from exposure to ethylene dibromide (EDB) in drinking water is 0.85/10 5 population. What is the risk to an individual who, for 5 years, drinks water with an average EDB concentration of 5pg/L? Risk = (concentration)(unit risk)(exposure time)/70 yr
14
ENVIRONMENTAL ENGINEERING Risk = (5 x 10" 12 g/L)(0.85 x 10" 5 L/10" 9 g)(5 yr/70 yr) = 3.0xl0"8
or approximately one chance of death in 33 million.
ASSESSMENT OF RISK
Risk of death (mortality risk) is easier to determine for populations in the developed countries than risk of illness (morbidity), because all deaths and their apparent causes are reported. Death certificate data may still be misleading: an individual who suffers from high blood pressure but is killed in an automobile accident becomes an accident statistic rather than a cardiovascular disease statistic. In the United States until very recently, statistics for deaths caused by occupational exposure could be determined only for males, because the majority of women did not work outside the home for much of their adult lives. These particular uncertainties may be overcome in assessing risk from a particular cause by isolating the influence of that particular cause. This requires studying two populations whose environment is virtually identical, except that the risk to be studied is present in one population and not in the other. Such a study is called a cohort study and may be used to determine morbidity as well as mortality risk. One cohort study1, for example, showed that residents of copper smelter communities, who were exposed to airborne arsenic, had a higher incidence of a certain type of lung cancer than residents of similar industrial communities where there was no airborne arsenic. Retrospective cohort studies are almost impossible to perform because of uncertainties in data on habits, other exposures, etc. Cohorts must be well matched in cohort size, age distribution, lifestyle, and other environmental exposures and must be large enough for an effect to be distinguishable from the deaths or illnesses that occur anyway.
EXPOSURE A N D LATENCY
Most malignant neoplasms (cancers) grow very slowly and are found (expressed) many years after the exposure to the responsible carcinogen. The length of time between exposure to a pollutant and expression of the adverse effect is called the latency period. Malignant neoplasms occurring in adults have apparent latency periods ranging in length from about 10 years to about 40 years. Relating a cancer to a particular exposure is fraught with inherent inaccuracy: it is exceedingly difficult to isolate the effect of a single carcinogen when examining 30 or 40 years of a person's life. Thus, many carcinogenic effects are simply not identifiable over the lifetime of an individual.
Risk Analysis
15
There are a few instances in which a particular neoplasm is found only on exposure to a given agent (e.g., a certain type of hemangioma is found only on exposure to vinyl chloride monomer), but for most cases, the connection between exposure and effect is far from clear. Many carcinogens are identified through animal studies, but one cannot always extrapolate from animal results to human results. Finally, chronic low-level exposure may (or may not) have different effects than acute high-level exposure, even when the total dose is the same. The cumulative uncertainty surrounding the epidemiology of pollutants has resulted in a conservative posture towards regulation and control. That is, if there is any evidence, even inconclusive evidence, that exposure to a substance results in adverse health effects, release of that substance into the environment is regulated and controlled.
DOSE-RESPONSE EVALUATION The effect of a pollutant (an organism's response to the pollutant) always depends in some way on the amount or dose of the pollutant to the organism. The magnitude of the dose, in turn, depends on the pathway into the organism. Pollutants have different effects depending on whether they are inhaled, ingested, or absorbed through the skin or whether exposure is external. Ingestion or inhalation determines the biochemical pathway of the pollutant in the organism. In general the human body detoxifies an ingested pollutant more efficiently than an inhaled pollutant. The relationship between the dose of a pollutant and the organism's response can be expressed in a dose-response curve, as shown in Figure 2 - 1 . Some characteristic features of the dose-response relationship are: 1.
Threshold. The existence of a threshold in health effects of pollutants has been debated for many years. With reference to Figure 2 - 1 , there are four basic dose-response curves possible for a dose of a specific pollutant (e.g., carbon monoxide) and the response (e.g., reduction in the blood's oxygen-
threshold ®
I
100 response detectable response 0
TLV
dose
Figure 2—1. Possible dose-response curves.
16
2.
3.
4.
ENVIRONMENTAL ENGINEERING carrying capacity). Curve A shows that if this dose-response relationship holds, there is no effect on human metabolism until a critical concentration is reached. This critical concentration is called the threshold and is indicated in the figure. Curve A is a linear dose-response curve. Curve B suggests that there is a detectable response for any finite concentration of the pollutant; this curve is also linear but shows no threshold. Curve C is a sigmoidal dose-response curve and is characteristic of many pollutant dose-response relationships. Although curve C has no clearly defined threshold, the point at which a response can be detected (shown in the figure) is called the threshold limit value (TLV). Occupational exposure guidelines are frequently set at the TLV. Curve C is sometimes called a "sublinear" dose-response relationship. Curve D represents a "supralinear" relationship, which is found when low doses appear to provoke a disproportionately large response. The sum of the doses given to the organism from all pathways is the total body burden and includes concentrations of pollutants remaining in the body from previous exposures. The rate at which pollutants are eliminated from the body is measured by the physiological half-life. Time Versus Dosage. Most pollutants require time to react, and thus the time of contact is as important as the level. The best example of this is the effect of carbon monoxide. CO reduces the oxygen-carrying capacity of the blood by combining with the hemoglobin and forming carboxyhemoglobin. At about 60 percent carboxyhemoglobin concentration, death results from lack of oxygen. The effects of CO at sublethal concentrations are usually reversible. Because of the time-response problem, ambient air quality standards are set at maximum allowable concentrations for a given time (see Chapter 20). Synergism. Synergism is defined as an effect that is greater than the sum of the parts. For example, black lung disease in coal miners occurs only when the miner is also a cigarette smoker. Coal mining by itself or cigarette smoking by itself will not cause black lung, but the synergistic action of the two puts miners who smoke at high risk. The opposite of synergism is antagonism: a. phenomenon that occurs when two pollutants counteract each other's effects. LCS0 and LDS0. Dose-response relationships for humans are generally determined from health data or epidemiological studies. Human volunteers obviously cannot be subjected to doses of pollutants that produce major or lasting health effects. Toxicity may, however, be determined by subjecting nonhuman organisms to increasing doses of pollutants until the organism dies. The term LDS0 refers to the dose that is lethal for 50 percent of the experimental animals used (LC 50 refers to lethal concentration instead of lethal dose). LD 50 values are of most use in comparing toxicities, as for pesticides and agricultural chemicals; no direct extrapolation is possible,
Risk Analysis
17
either to humans or to any species other than the one used for the LD 50 determination. 5. Bioaccumulation and Bioconcentration. The term bioaccumulation is used when a substance is concentrated in one organ or type of tissue. Iodine, for example, bioaccumulates in the thyroid gland. The dose to an organ may thus be considerably greater than the body burden or whole-body dose. Bioconcentration occurs with movement up the food chain. A study2 of a Lake Michigan ecosystem found the following bioconcentration of DDT: 0.014 ppm (wet weight) in bottom sediments 0.41 ppm in bottom-feeding Crustacea 3 to 6 ppm in fish 2,400 ppm in fish-eating birds POPULATION RESPONSES Individual responses to pollutants are not identical. Dose-response curves will differ from one person to another; in particular, thresholds will differ. In general, threshold values in a population follow a Gaussian distribution. Figure 2 - 2 shows the distribution of odor thresholds for hydrogen sulfide in a typical population. Individual responses and thresholds also depend on age, sex, and general state of physical and emotional health. As might be anticipated, healthy young adults are less sensitive to pollutants than are elderly people, chronically or acutely ill people, and children. Allowable releases of pollutants are, in theory, restricted to amounts that assure protection of the health of the entire population, including its most sensitive members. In many cases, however, such restriction would mean zero release.
100 r ■σ Φ Φ
50 h ω o
u. Φ Û.
Odor Threshold for H 2 S (ppb)
Figure 2-2. Distribution of odor thresholds in a population.
18
ENVIRONMENTAL ENGINEERING
The levels actually chosen take technical and economic control feasibility into account and usually are set below threshold level for 95 percent or more of the entire population. For nonthreshold pollutants, however, no such determination can be made. In these instances, there is no release level for which protection can be assured for anyone, so that comparative risk analysis is necessary. Many pollutants, including carcinogens, are considered to have no measurable threshold. The best available control for such pollutants still entails a residual risk. There is a continuing need in our industrial society, therefore, for accurate quantitative risk assessment. We must also remain aware that we have not precisely identified carcinogens or mutagens, nor do we understand their mechanism of action. We can only identify apparent associations between most pollutants and a given health effect. In almost all cases, doses to the general public are so small that excess mortality and morbidity are not identifiable. In fact, almost all of our knowledge of adverse health effects of pollutants comes from occupational exposure, in which doses are orders of magnitude higher.
PROBLEMS
2.1 Using the data given in the chapter, calculate the SMR (from all diseases) for heavy smoking in the United States. 2.2 The EPA has determined the lifetime unit risk for cancer for lowenergy ionizing radiation to be 2.8 x 10~ 4 per rem of radiation. The allowed level of ionizing radiation (EPA standard) above background is 25 mrem per year. Average nonanthropogenic background is about 100 mrem per year. How many excess cancers would result in the United States each year if the entire population were exposed at the level of the EPA standard? How many cancers may be attributed to background? If 10 percent of these are fatal each year, what percentage of the annual cancer deaths in the United States may be attributed to exposure to background radioactivity? 2.3 The allowed occupational dose for ionizing radiation is 5 rem per year. By what factor does a worker exposed to this dose over a working lifetime increase his or her risk of cancer? 2.4 By what factor was DDT bioconcentrated at each level of the food chain in Lake Michigan in the example given in the text? 2.5 The EPA has set standards for the emission of arsenic from nonferrous smelters that project a maximum unit cancer incidence risk of 150 in 10,000 (per μ^/τη3 of airborne arsenic) to a population of 190,000 persons living within 20 km of a smelter. Assume that the population exposed is exposed throughout life and that death from this type of cancer occurs 25 years after incidence. What would you consider an acceptable level of ambient airborne arsenic from such a smelter, in /Ag/m , based on the risk? Suppose a resident
Risk Analysis
19
lived within this radius of the smelter for 10 years, on the average. Would your answer change? By how much? 2.6 Using the data of the preceding problem, estimate an acceptable workplace standard for ambient arsenic. 2.7 It is widely assumed that the Love Canal and Times Beach incidents, in which toxic chemicals were found to have leaked into the environment over a period of years, resulted in adverse health effects to the residents of these areas. Design an epidemiological study that would determine whether there actually were adverse health effects resulting from these incidents. Assume that the leaks occurred over 5 years and that there were leaks into the air, soil, and drinking water. 2.8 How does synergism make it difficult to establish cause/effect relationship between air pollutants and disease? Give at least one example.
REFERENCES
1. National Center for Health Statistics, U.S. Department of Health and Human Services, 1985. 2. Hickey, J.J., et al. "Concentration of DDT in Lake Michigan," Journal of Applied Ecology 3:141(1966).
Risk Analysis In a nutshell, the job of the environmental engineer is to reduce the risks of environmental hazards, both long- and short-term. To accomplish this objective in an orderly fashion, the risks associated with various hazards must be quantified and evaluated. In this chapter, risk analysis is introduced as a necessary tool of environmental engineering, crossing all media boundaries.
RISK
A number of laws regulate release of pollutants into the environment. However, major issues must be addressed. For instance, how is a pollutant identified as such, so that steps may be taken to control it? A substance is considered a pollutant because it is perceived to have an adverse effect on the environment and, in particular, an adverse effect on human health. However, it is sometimes difficult to determine whether there is an effect, or whether the effect has been deleterious or detrimental to the organism that is affected. For example, we are now certain that cigarette smoke is unhealthy. Specifically, we have identified inhaled cigarette smoke as contributing significantly to lung cancer, chronic obstructive pulmonary disease, and heart disease. Notice that we do not say that cigarette smoking causes these health problems, because we have not identified the causes—the etiology—of any of them, at least in the sense that we have identified the poliomyelitis virus as the cause of polio. How, then, has cigarette smoking been identified as a contributing factor, if it cannot be identified as the cause? Cigarette smoking serves as a good example of how the health effects of pollutants are determined, although cigarette smoke is not regulated the way other pollutants are regulated. In the latter half of the twentieth century, with the discovery of antibiotics and sulfa drugs, infectious diseases ceased to be a primary cause of death. The life span of people in the developed countries of the world lengthened consid11
12
ENVIRONMENTAL ENGINEERING
erably, and heart disease and cancer became leading causes of death. It was thus observed, in the early 1960s, that lifelong heavy cigarette smokers often died from lung cancer.
EXPRESSION OF RISK
Let us investigate the details of such an observation: how do you identify the cause of death, and how is it related to a habit like heavy smoking? The cause of death is listed on death certificates, which are the source of much epidemiological information. To relate death from lung cancer to smoking, one must show that more lung cancer deaths occur in smokers than in nonsmokers. Such a showing can be made by determining the standard mortality ratio (SMR), which is defined as SMR = observed deaths/expected deaths The number of "expected deaths" in the above equation is the number of deaths from the particular disease, lung cancer, in this case, that happen without any identifiable cause. In the general population of smokers and nonsmokers, there are a certain number of lung cancer deaths. Even in nonsmokers, there are a certain, albeit small, number of lung cancer deaths. In this instance, then, the SMR may be defined as SMR = D S /D n s where
D s = lung cancer deaths in a population of smokers D ns = lung cancer deaths in a nonsmoking population of the same size
In this instance, the SMR is approximately 1 1 / 1 . Since the SMR is greater than 1.0, then we can say that cigarette smoke contributes to lung cancer or, as it is usually phrased, smoking cigarettes increases the risk of death from lung cancer. To be precise, the risk of death from lung cancer is 11 times as high for a heavy smoker as for a nonsmoker. Determination of the SMR tells something about the epidemiology of smoking, but not about its etiology. There are three characteristics of epidemiological reasoning in this example that are important: • • •
Everyone who smokes heavily will not die of lung cancer. Some nonsmokers die of lung cancer. Therefore, one cannot unequivocally relate any given individual lung cancer death to cigarette smoking.
Risk may be expressed in other ways as well. For example, in the United States there are 350,000 deaths each year from lung cancer and heart disease
Risk Analysis
13
Table 2-1. Adult Deaths/105 for 1985 Deaths/10s Population
Cause of Death Cardiovascular disease Cancer Chronic obstructive pulmonary disease Motor vehicle accidents Alcoholism
408 193 31 18.6 11.3
All causes
870
that are attributable to smoking. The United States has a population of 226 million. The risk of death associated with the effects of cigarette smoking may thus be expressed as deaths per 100,000 population, or 350,000/226 x 10 6 = 155/100,000 In other words, a heavy smoker in the United States has an annual probability (or risk) of 155 in 100,000, or 1.55 in 1,000, of dying of lung cancer or heart disease. Table 2 - 1 presents some typical statistics for the United States.1 A third way to present risk is as deaths per 1,000 deaths. For example, in 1985, there were 2.084 x 10 6 deaths in the United States. Of these, 350,000 or 168 deaths per 1,000 deaths, were related to heavy smoking. These figures show that meaningful risk analyses can be conducted only with very large populations. Moreover, the health risk posed by most pollutants is observed to be considerably lower than the risks cited in Table 2—1 and may simply not be observed in a small population. Chapter 18 cites several examples of statistically valid risks from air pollutants. The Environmental Protection Agency (EPA) has adopted the concept of unit risk in discussions of potential risk. Unit risk is defined as the risk to an individual from exposure to a concentration of 1 jug/m3 of an airborne pollutant, o r l O ~ 9 g / L o f a water borne pollutant. Unit lifetime risk is the risk to an individual from exposure to the above concentration for 70 years, or a lifetime, while unit occupational risk implies exposure for 40 hours per week for 50 years, or a working lifetime. Example 2.1 The EPA has calculated that the unit lifetime risk from exposure to ethylene dibromide (EDB) in drinking water is 0.85/10 5 population. What is the risk to an individual who, for 5 years, drinks water with an average EDB concentration of 5pg/L? Risk = (concentration)(unit risk)(exposure time)/70 yr
14
ENVIRONMENTAL ENGINEERING Risk = (5 x 10" 12 g/L)(0.85 x 10" 5 L/10" 9 g)(5 yr/70 yr) = 3.0xl0"8
or approximately one chance of death in 33 million.
ASSESSMENT OF RISK
Risk of death (mortality risk) is easier to determine for populations in the developed countries than risk of illness (morbidity), because all deaths and their apparent causes are reported. Death certificate data may still be misleading: an individual who suffers from high blood pressure but is killed in an automobile accident becomes an accident statistic rather than a cardiovascular disease statistic. In the United States until very recently, statistics for deaths caused by occupational exposure could be determined only for males, because the majority of women did not work outside the home for much of their adult lives. These particular uncertainties may be overcome in assessing risk from a particular cause by isolating the influence of that particular cause. This requires studying two populations whose environment is virtually identical, except that the risk to be studied is present in one population and not in the other. Such a study is called a cohort study and may be used to determine morbidity as well as mortality risk. One cohort study1, for example, showed that residents of copper smelter communities, who were exposed to airborne arsenic, had a higher incidence of a certain type of lung cancer than residents of similar industrial communities where there was no airborne arsenic. Retrospective cohort studies are almost impossible to perform because of uncertainties in data on habits, other exposures, etc. Cohorts must be well matched in cohort size, age distribution, lifestyle, and other environmental exposures and must be large enough for an effect to be distinguishable from the deaths or illnesses that occur anyway.
EXPOSURE A N D LATENCY
Most malignant neoplasms (cancers) grow very slowly and are found (expressed) many years after the exposure to the responsible carcinogen. The length of time between exposure to a pollutant and expression of the adverse effect is called the latency period. Malignant neoplasms occurring in adults have apparent latency periods ranging in length from about 10 years to about 40 years. Relating a cancer to a particular exposure is fraught with inherent inaccuracy: it is exceedingly difficult to isolate the effect of a single carcinogen when examining 30 or 40 years of a person's life. Thus, many carcinogenic effects are simply not identifiable over the lifetime of an individual.
Risk Analysis
15
There are a few instances in which a particular neoplasm is found only on exposure to a given agent (e.g., a certain type of hemangioma is found only on exposure to vinyl chloride monomer), but for most cases, the connection between exposure and effect is far from clear. Many carcinogens are identified through animal studies, but one cannot always extrapolate from animal results to human results. Finally, chronic low-level exposure may (or may not) have different effects than acute high-level exposure, even when the total dose is the same. The cumulative uncertainty surrounding the epidemiology of pollutants has resulted in a conservative posture towards regulation and control. That is, if there is any evidence, even inconclusive evidence, that exposure to a substance results in adverse health effects, release of that substance into the environment is regulated and controlled.
DOSE-RESPONSE EVALUATION The effect of a pollutant (an organism's response to the pollutant) always depends in some way on the amount or dose of the pollutant to the organism. The magnitude of the dose, in turn, depends on the pathway into the organism. Pollutants have different effects depending on whether they are inhaled, ingested, or absorbed through the skin or whether exposure is external. Ingestion or inhalation determines the biochemical pathway of the pollutant in the organism. In general the human body detoxifies an ingested pollutant more efficiently than an inhaled pollutant. The relationship between the dose of a pollutant and the organism's response can be expressed in a dose-response curve, as shown in Figure 2 - 1 . Some characteristic features of the dose-response relationship are: 1.
Threshold. The existence of a threshold in health effects of pollutants has been debated for many years. With reference to Figure 2 - 1 , there are four basic dose-response curves possible for a dose of a specific pollutant (e.g., carbon monoxide) and the response (e.g., reduction in the blood's oxygen-
threshold ®
I
100 response detectable response 0
TLV
dose
Figure 2—1. Possible dose-response curves.
16
2.
3.
4.
ENVIRONMENTAL ENGINEERING carrying capacity). Curve A shows that if this dose-response relationship holds, there is no effect on human metabolism until a critical concentration is reached. This critical concentration is called the threshold and is indicated in the figure. Curve A is a linear dose-response curve. Curve B suggests that there is a detectable response for any finite concentration of the pollutant; this curve is also linear but shows no threshold. Curve C is a sigmoidal dose-response curve and is characteristic of many pollutant dose-response relationships. Although curve C has no clearly defined threshold, the point at which a response can be detected (shown in the figure) is called the threshold limit value (TLV). Occupational exposure guidelines are frequently set at the TLV. Curve C is sometimes called a "sublinear" dose-response relationship. Curve D represents a "supralinear" relationship, which is found when low doses appear to provoke a disproportionately large response. The sum of the doses given to the organism from all pathways is the total body burden and includes concentrations of pollutants remaining in the body from previous exposures. The rate at which pollutants are eliminated from the body is measured by the physiological half-life. Time Versus Dosage. Most pollutants require time to react, and thus the time of contact is as important as the level. The best example of this is the effect of carbon monoxide. CO reduces the oxygen-carrying capacity of the blood by combining with the hemoglobin and forming carboxyhemoglobin. At about 60 percent carboxyhemoglobin concentration, death results from lack of oxygen. The effects of CO at sublethal concentrations are usually reversible. Because of the time-response problem, ambient air quality standards are set at maximum allowable concentrations for a given time (see Chapter 20). Synergism. Synergism is defined as an effect that is greater than the sum of the parts. For example, black lung disease in coal miners occurs only when the miner is also a cigarette smoker. Coal mining by itself or cigarette smoking by itself will not cause black lung, but the synergistic action of the two puts miners who smoke at high risk. The opposite of synergism is antagonism: a. phenomenon that occurs when two pollutants counteract each other's effects. LCS0 and LDS0. Dose-response relationships for humans are generally determined from health data or epidemiological studies. Human volunteers obviously cannot be subjected to doses of pollutants that produce major or lasting health effects. Toxicity may, however, be determined by subjecting nonhuman organisms to increasing doses of pollutants until the organism dies. The term LDS0 refers to the dose that is lethal for 50 percent of the experimental animals used (LC 50 refers to lethal concentration instead of lethal dose). LD 50 values are of most use in comparing toxicities, as for pesticides and agricultural chemicals; no direct extrapolation is possible,
Risk Analysis
17
either to humans or to any species other than the one used for the LD 50 determination. 5. Bioaccumulation and Bioconcentration. The term bioaccumulation is used when a substance is concentrated in one organ or type of tissue. Iodine, for example, bioaccumulates in the thyroid gland. The dose to an organ may thus be considerably greater than the body burden or whole-body dose. Bioconcentration occurs with movement up the food chain. A study2 of a Lake Michigan ecosystem found the following bioconcentration of DDT: 0.014 ppm (wet weight) in bottom sediments 0.41 ppm in bottom-feeding Crustacea 3 to 6 ppm in fish 2,400 ppm in fish-eating birds POPULATION RESPONSES Individual responses to pollutants are not identical. Dose-response curves will differ from one person to another; in particular, thresholds will differ. In general, threshold values in a population follow a Gaussian distribution. Figure 2 - 2 shows the distribution of odor thresholds for hydrogen sulfide in a typical population. Individual responses and thresholds also depend on age, sex, and general state of physical and emotional health. As might be anticipated, healthy young adults are less sensitive to pollutants than are elderly people, chronically or acutely ill people, and children. Allowable releases of pollutants are, in theory, restricted to amounts that assure protection of the health of the entire population, including its most sensitive members. In many cases, however, such restriction would mean zero release.
100 r ■σ Φ Φ
50 h ω o
u. Φ Û.
Odor Threshold for H 2 S (ppb)
Figure 2-2. Distribution of odor thresholds in a population.
18
ENVIRONMENTAL ENGINEERING
The levels actually chosen take technical and economic control feasibility into account and usually are set below threshold level for 95 percent or more of the entire population. For nonthreshold pollutants, however, no such determination can be made. In these instances, there is no release level for which protection can be assured for anyone, so that comparative risk analysis is necessary. Many pollutants, including carcinogens, are considered to have no measurable threshold. The best available control for such pollutants still entails a residual risk. There is a continuing need in our industrial society, therefore, for accurate quantitative risk assessment. We must also remain aware that we have not precisely identified carcinogens or mutagens, nor do we understand their mechanism of action. We can only identify apparent associations between most pollutants and a given health effect. In almost all cases, doses to the general public are so small that excess mortality and morbidity are not identifiable. In fact, almost all of our knowledge of adverse health effects of pollutants comes from occupational exposure, in which doses are orders of magnitude higher.
PROBLEMS
2.1 Using the data given in the chapter, calculate the SMR (from all diseases) for heavy smoking in the United States. 2.2 The EPA has determined the lifetime unit risk for cancer for lowenergy ionizing radiation to be 2.8 x 10~ 4 per rem of radiation. The allowed level of ionizing radiation (EPA standard) above background is 25 mrem per year. Average nonanthropogenic background is about 100 mrem per year. How many excess cancers would result in the United States each year if the entire population were exposed at the level of the EPA standard? How many cancers may be attributed to background? If 10 percent of these are fatal each year, what percentage of the annual cancer deaths in the United States may be attributed to exposure to background radioactivity? 2.3 The allowed occupational dose for ionizing radiation is 5 rem per year. By what factor does a worker exposed to this dose over a working lifetime increase his or her risk of cancer? 2.4 By what factor was DDT bioconcentrated at each level of the food chain in Lake Michigan in the example given in the text? 2.5 The EPA has set standards for the emission of arsenic from nonferrous smelters that project a maximum unit cancer incidence risk of 150 in 10,000 (per μ^/τη3 of airborne arsenic) to a population of 190,000 persons living within 20 km of a smelter. Assume that the population exposed is exposed throughout life and that death from this type of cancer occurs 25 years after incidence. What would you consider an acceptable level of ambient airborne arsenic from such a smelter, in /Ag/m , based on the risk? Suppose a resident
Risk Analysis
19
lived within this radius of the smelter for 10 years, on the average. Would your answer change? By how much? 2.6 Using the data of the preceding problem, estimate an acceptable workplace standard for ambient arsenic. 2.7 It is widely assumed that the Love Canal and Times Beach incidents, in which toxic chemicals were found to have leaked into the environment over a period of years, resulted in adverse health effects to the residents of these areas. Design an epidemiological study that would determine whether there actually were adverse health effects resulting from these incidents. Assume that the leaks occurred over 5 years and that there were leaks into the air, soil, and drinking water. 2.8 How does synergism make it difficult to establish cause/effect relationship between air pollutants and disease? Give at least one example.
REFERENCES
1. National Center for Health Statistics, U.S. Department of Health and Human Services, 1985. 2. Hickey, J.J., et al. "Concentration of DDT in Lake Michigan," Journal of Applied Ecology 3:141(1966).