Document Nzmk7J2pZakrrZvJzx3OvyQw

ABDOO134440 POLLUTION ENGINEERING A Primer on Risk Calculations Assessing the chronic toxic effects ofa chemical can be difficult and expensive. But here is a simple model for assessing cancer risks due to exposure to an environmental pollutant. by Gerald Rich, PE, REP In recent years, federal statutes such as the Toxic Substances Control Act, the Resource Conservation and Recov ery Act, the Clean Air Act, and the Clean Water Act have given the U.S. Environmental Protection Agency (USEPA) the responsibility for regulat ing the release of toxic chemicals into the environment. In order to fulfill this function effec tively, the EPA must first determine which of the thousands of chemicals currently in use or proposed for use are toxic. Detecting a chemical's ability to cause immediate or acute toxic effects is a relatively straightforward task. As sessing the long-term or chronic toxic effects is much more difficult. Chronic effects such as cancer, birth defects and genetic disease characteristically appear several years or decades after the initial chemical exposure has occurred, and long-term studies using live animals must be conducted in order to detect these latent effects. Such studies are expensive and timeconsuming, requiring the use of highly specialized facilities and personnel. A single test for a chemical's carcinogeni city may take as long as three years and cost $250,000 or more. The number of compounds whose chronic toxicity has not been deter mined is overwhelming. Over 50,000 chemicals are currently in commercial production, and most of them have never been examined for chronic ef fects. The world laboratory capacity for long-term studies has been estimated at only 500 compounds per year, not enough to keep up with the 700 to 1000 new chemicals introduced each year. Short-term tests have been developed to serve as rapid and relatively inexpen sive predictors of a chemical's potential to cause chronic effects. These tests em ploy bacteria, yeast, plants, insects, iso lated mammalian cells, whole animals and computer modeling. Short-term tests can detect a chemi cal's genotoxicity, or its ability to alter a cell's genetic material -- the DNA. An increasing amount of evidence indi cates that latent diseases such as cancer, birth defects and genetic disease may be initiated by alterations in the DNA. Short-term tests enable a large num ber of chemicals to be screened for their genotoxic potential at a fraction of the time and cost required for long-term tests. Results from short-term tests can be used to make more informed deci sions as to which chemicals should be examined in the limited number of long-term testing facilities available. Other uses for short-term tests in clude determining which of several al ternative chemicals under development will be the least hazardous to human health; identifying the toxic compo nents of complex environmental pollut ants; and providing interim guidance for using a chemical when no other data is available. A distinction is made between car cinogenic and non-carcinogenic effects. Two general criteria are used to de scribe risk: Excess lifetime cancer risk for constituents which are thought to be potential human carcinogens, and the hazard index (HI) for all constitu ents. All chemicals have non-carcino genic effects, but only a few are known to cause cancer. For potential carcinogens, the current regulatory approach uses an extremely conservative approach in which it is as sumed that even a trace level of expo sure to a carcinogen can cause cancer. This is contrary to the traditional ap proach to toxic chemicals, in which fi nite thresholds are said to exist, below which the toxic effect will not occur. This traditional approach is still ap plied to non-carcinogenic chemicals. Because of these differing approaches to calculating potential risks, the risks associated with carcinogenic effects are generally much higher than those asso ciated with non-carcinogenic effects. Cancer is the end result of a multi stage process in which a large number of biological and environmental factors interact, simultaneously or in sequence, to disrupt normal cell growth and divi- 94 Pollution Engineering June 1990 ABDOO134441 TURNING WASTE INTO ENERGY? TURN TO THE FAN EXPERTS: ROBINSON. We'll meet your needs for high-efficiency/low-cost operation based on nearly a century of fan engineering, application and fabrication experience. Robinson fans perform dependably in applications such as in fluidized bed systems, boilers, heat recovery operations, incineration and pollution control. Robinson air-moving equipment actually saves money over the years thanks to our design experience and a quality assurance pro gram that meets or exceeds stringent industry requirements. INDUSTRIES, INC. P.O. Box 100 Zelienople, PA 16063 Phone: 1-412-452-6121 Telex: 866-311 Fax:1-412-452-0388 ROBINSON FANS FLORIDA, INC P.O. Box 6260 Lakeland, FL 33807^6260* Phone: 1-813-646-5270 Telex: 866-31T:*^ Let us improve the performance of your existing fans. Robinson can redesign and upgrade, rebuild or repair your existing equipment whether we built it in the first place or not. And Robinson repairs last because we begin by identifying and correcting the causes of the problem! Send for FREE literature. For more information on airy / moving equipment for any,/^' waste-to-energy application, send for BuJ 505. For repair and rebuild informatics Bulletin 551. Hotlines are answered 24 locations. For information circle 83 ABDOO134442 sion. The first stage, called initiation, involves the creation of errors in ge netic coding. Because the effects of in itiation are thought to occur at the mo lecular level, current regulatory policy assumes there is no finite dose below which the initiation effect cannot occur. In other words, there may be no abso lutely safe dose, because at any dose there is some finite probability associ ated with the occurrence of the initia tion event. Once a cell is initiated, many other processes are involved in either pro moting or blocking the development of cancer. Therefore, the frequency of can cer occurrence is very low in compari son to the hypothetical frequency of in itiation events, especially at the low chemical concentrations which occur in the environment. Further, many chemicals which are classified as carcinogens may not actu ally initiate cancer, but rather act at a RISK ASSESSMENT TERMS Synergism -- On a scale of toxicity of 0 to 25 with 25 being the most toxic, synergism is the term used to describe the effect when the outcome of combining two chemicals is greater than the sum of the inputs. An example would be carbon tetrachloride and ethanol. Antagonism -- Similarly, antagonism is the term used to describe the effect when the outcome is less than the sum of the two inputs. An example would be chelators and metals. Potentiation -- When one chemical has no toxic effect but combined with another chemical that is toxic produces a much more toxic effect, the result is called potentiation. An example would be carbon tetrachloride and isopropanol. f 4 9> ,. June 1990 Pollution Engineering 95 ABDOO134443 POLLUTION ENGINEERING The carcinogen potency factor is used to calculate the probability of cancer later stage on cells which were already For chemicals which seem to cause exposure dose and the cancer potency initiated. Although these chemicals are or increase cancer, the results of these factor for each constituent. Excess life treated as if one molecule could cause high-dose animal studies are extrapo time cancer risks can be summed across significant damage, this has not been lated to the low-dose human exposure routes of exposure and constituents. proven. situation using mathematical models An excess lifetime cancer risk of less As an example, in the case of chro which do not depict biological reality. than 10'7 is generally considered accept mium, only the hexavalent form has A variety of mathematical models are able. An excess lifetime cancer risk of been proven to be carcinogenic, al available for extrapolating from high to greater than 10'7 is generally considered though it is a potent carcinogen. There low doses. The USEPA currently favors unacceptable. Excess lifetime cancer is now insufficient evidence to deter a linearized multistage model which risks in the range of 10"* to 10'7 are mine that the trivalent form also is car provides a 95 percent upper-bound esti potentially acceptable. A range of ICri* cinogenic. A NESHAP analysis, how mate of cancer incidence at a given to 1 O'7 is the common target for reme ever, assumes that total chromium re dose. The slope of the extrapolated diation activities. leases are carcinogenic and that curve, called the carcinogen potency A finite dose or threshold, below trivalent chromium is as po which adverse effects will not tent as hexavalent chromium. occur, is believed to exist for There is no information non-cancer effects. Non-can now available on the. ratio of cer effects include birth de trivalent to hexavalent for fects, organ damage, death emissions or ambient concen trations, but some occupa The excess lifetime cancer risk and many others. A single compound might elicit sev tional exposure studies sug eral adverse effects depend gest the trivalent form may is an estimate ofthe increased risk of dominate in some source ing on the dose, the route and the duration of exposure. categories. On the other For a given chemical, the hand, several source catego ries are known to emit at least cancer which resultsfrom exposure to dose which elicits no effect when evaluating the most some hexavalent chromium, sensitive response (the ad and there is some evidence that changes in the valence constituents at a given concentration. verse effect which occurs at the lowest dose) in the most state can occur in the sensitive species is used to es atmosphere. Thus, in the case of chro One infourAmericans will have cancer. tablish an acceptable dose for non-carcinogenic effects. Ac mium, the problem of specia- ceptable doses which are tion adds one more layer of sanctioned by the USEPA are uncertainty to the risk esti called reference doses (RfDs). mate. factor, is used to calculate the probabil The hazard index is a representation Identification of constituents which ity of cancer associated with the expo which addresses non-cancer health ef are included in the carcinogen category sure dose. fects of compounds. The hazard index is based on a USEPA classification The probability of developing cancer is the ratio of the exposure dose to the scheme in which chemicals are system as a result of exposure to constituents acceptable dose or RfD. A hazard index atically evaluated for their ability to associated with the facility is called the greater than one is an indication that cause cancer in mammalian species and excess lifetime cancer risk. exposure is too high; it means that, for conclusions are reached about the po The excess lifetime cancer risk is an a given constituent, exposure exceeds tential to cause cancer in humans. estimate of the increased risk of cancer acceptable levels for protection against The USEPA classification scheme which results from exposure to con non-cancer effects. Hazard indexes can contains six categories: stituents at a given concentration. Ap be summed across exposure routes and A -- Human carcinogen; proximately one out of four Americans constituents. A hazard index of less B1 -- probable human carcinogen, will have cancer during their lifetimes, than one suggests non-cancer health ef limited evidence in humans; and about half of those will die from fects would not occur, less than one is B2 -- probable human carcinogen, the cancer. a common USEPA remediation goal. inadequate evidence in humans; The risk values provided in a risk However, many state agencies prefer C -- possible human carcinogen; assessment are an indication of the in the hazard index be in the range of 0.2. D -- inadequate evidence to classify; creased risk, above that applying to the This preference is based on the realiza E -- no evidence of carcinogenicity. general population, which results from tion that people may be exposed to Categories A, Bl, B2 and C are in the exposure scenarios described in an these same constituents from sources cluded in assessments as potential hu exposure characterization. The actual unrelated to a specific site. man carcinogens. Much of these evalu cancer risk is unknown and could be In general, cancer potency factors, ations are based on laboratory animal anywhere between zero -- no risk of carcinogen classifications and RfDs studies. In order to limit the number cancer --and the value provided, but (separated by exposure route) are taken of animals required for testing, very is not likely to exceed the estimate de from either USEPA's Superfund Public high doses of chemicals are used in rived in the risk assessment. The excess Health Evaluation Manual (1986) or laboratory studies. lifetime cancer risk is the product of IRIS (Integrated Risk Information 96 Pollution Engineering June 1990 ABDOO134444 System)--an EPA-approved computer data base. Three routes of exposure are consid ered: Dermal (skin), inhalation, and in gestion. Whenever possible, routespecific values are used. Because ac ceptable levels for dermal exposure are virtually never available, oral values are used in their place. When inhalation values are not available, oral values are substituted in their place unless com pound-specific information clearly in dicates this would be inappropriate. In some cases, route of exposure deter mines the type of toxicity that will be expected. Considerable uncertainty is inherent in the overall risk assessment process. The basic building blocks for risk as sessment include source assessment, monitoring/sampling and analysis, ex posure assessment, health effects meas urement and risk management, all of which are combined into the regulatory decision-making process. Each of these contributes uncertainties in its own fashion. Refer to Dr. Stephen Hall's December, 1988, article in POLLU TION ENGINEERING, "Health Risk Assessment of Chemical Exposure" for the health risk assessment aspects. Risk assessment is based, at least par tially, on the assumption that the moni toring data adequately describes the situation. Environmental sampling it self introduces uncertainty because of the potential for uneven distribution of constituents in the environmental media. Two sampling methods, grab and composite, are typically used. Grab samples determine the distribution of constituents and the variation in con centrations. Composite samples indicate average constituent concentrations but can ob scure information about hot spots. Composites of adjacent samples mini mize this problem. Uncertainty associated with sample analysis can be minimized by using ap propriate analytical methods and equipment, documenting the chain of custody of samples, and implementing strict laboratory data validation and quality assurance procedures. It is also critical that sample detection limits be lower than both the applicable stan dards or criteria and the concentration which may present a health risk. Exposure scenarios and constituent transport models also contribute uncer tainty to the risk assessment. Transport models typically over-simplify reality, contributing uncertainty. Exposure routes include: Inhalation, ingestion, dermal absorption, intravenous, subcutaneous (injected below the skin), intraperitoneal (subject to metabo lism in the liver) The toxicity values and other toxi cologic (health effects) information are associated with significant uncertainty. Most health effects information has TABLE 1 AGGREGATE RISK CALCULATIONS Ambient Number of Concentration People City G*g/m3) Exposed A6 3 0.5 5000 20,000 50,000 B3 10,000 1 50,000 0.1 200,000 C7 5 1 1000 5000 10,000 been developed using laboratory ani mals exposed to high doses. Although species differences in absorption, distri bution, metabolism, excretion and tar get organ sensitivity are well docu mented, available data are not suffi cient to allow compensation for these differences. Most laboratory studies strictly con trol as many factors as possible, yet the human population is genetically diverse and affected by a variation in diets, oc cupations, genetics, pharmaceuticals, pre-existing illnesses and other factors. There also is considerable uncer tainty associated with the toxicity of mixtures. For the most part, data about the toxicity of chemical mixtures are unavailable even though the effect of combining two chemicals may be known to be synergistic, antagonistic, or under potentiation. Rather, toxicity studies are generally performed using a single chemical. Chemicals present in a mixture can interact chemically to yield a new Aggregate Risk 5.75 5 2.1 June 1990 Pollution Engineering 97 ABDOO134445 POLLUTION ENGINEERING Toxicity data on chemical mixtures is for the most part, unavailable chemical or one can interfere with the absorption, distribution, metabolism, or excretion of another. Chemicals may also act by the same mechanism at the same target organ or can act completely independently. In general, risk assess ments assume toxicity is additive. Unit risk Assessing the cancer risks of exposure to an environmental pollutant requires three pieces of information: An esti mate of the carcinogenic potency, or the unit risk factor, of the pollutant be ing considered; an estimate of the ambi ent concentration that an individual or group of people may breathe; and an estimate of the number of people ex posed to those concentrations. Unit risk is a quantitative estimate of the carcinogenic potency, or the unit risk value or factor, and often is ex- pressed as the chance of contracting cancer (though not necessarily lifethreatening) from a 70-year lifetime continuous exposure to a concentration of one microgram per cubic meter in air (1 p,g/m3) of a given substance. The unit risk can be expressed in terms of parts per million (ppm) in diet, milligrams per kilogram of body weight per day (mg/kg/day) or, most frequently for air pollutant risk assessment, (ig/m3 in the air. Unit risk is usually expressed as a two-digit number times a power of 10 per microgram per cubic meter, such as 5.5 x 10'6 (|ig/m3)-l. A unit risk value of 5.5 x lO"* (pg/m3)-l would indicate an estimated 5.5 cancer cases per 1,000,000 people exposed to an average of 1 p.g/nr of a specific carcinogen for 70 years. For chemicals with good quantitative TABLE 2 CALCULATING INDIVIDUAL RISK* Pollutant Factor Acetaidehyde Acrylonitrile Arsenic Benzene BaP Beryllium 1,3-butadiene Cadmium Carbon tetrachloride Chloroform Chromium VI 1,2-dichloroethane 1,1-dichioroethylene Epichlorohydrin Ethylene dibromide Ethylene oxide Formaldehyde Gasoline (marketing) Hexachlorobenzene Methylene chloride Nickel-refinery dust Nickel-subsulfide Propylene oxide Styrene 2,3,7,8 - tetrachloridibenzo-p-dioxin Tetrachloroethylene Trichloroethylene Vinyl chloride 3.96 x 10-a 1.48 x 10-4 5.27 x 10-2 2.65 x 10-5 1.75 x 10-2 8.85 x 10-4 6.19 x 10-4 8.28 x 10-3 9.44 x 10-5 1.12 x 10-4 2.55 x 10-2 1.05 x 10-4 1.98 x 10-* 4.54 x 10- 1.69 x 10-3 1.80 x 10-4 1.60 x 10-5 1.69 x 10-e 5.71 x 10-3 1.63 x 10-5 2.36 x 10-3 1.18 x 10-3 8.79 x 10-5 2.43 x 10- 4.35 x 10 -4 3.93 x 10-5 9.28 x 10-s 1.05 x IQ-5 Factor = (unit risk factor) x (molecular weight)/(24.45 liters/mole @ 25C) `These are factors in common use by EPA They are subject to and frequently do change as additional information is evaluated. For further up-to-date information, contact Pollutant Assessment Branch (MD-12), Research Triangle Park, NC 27711. human data available, maximum likeli hood or best estimate procedures are used to estimate unit risk factors. In contrast to the 95 percent upper confidence limit procedures used for animal data, maximum likelihood pro cedures will generally overestimate risk 50 percent of the time and will underes timate risk 50 percent of the time. This less conservative approach is generally justified by citing lack of species to spe cies extrapolation and less extensive low-dose extrapolation. The unit risk estimates for a number of air pollutants have been derived from human data using maximum like lihood or best estimate procedures, as for asbestos, arsenic, cadmium, chro mium and nickel. For animal or human data with sig nificant quantitative uncertainty, EPA most frequently uses a linearized multi stage procedure to estimate unit risk. This procedure leads to an upper bound estimate of the risk -- the 95 percent confidence limit -- that fits the data and also may lead to a most likely estimate (MLE) of risk. The unit risk factor is a plausible upper bound esti mate of the risk associated with a pol lutant. With such an estimate, the true risk is not likely to be higher than the estimate, but could be lower. Individual lifetime risk vs. aggregate risk Cancer risks may be reported in terms of risk estimates for an individual (lifetime individual risk) or risk esti mates for an exposed population (aggre gate risk). Lifetime individual risk is a measure of the probability of an individual's de veloping cancer as a result of exposure to an ambient concentration of an air pollutant over a 70-year period. It is calculated by multiplying the unit risk, factor by a long term average exposure estimate (jig/ra3). The maximum life time individual risk, which usually ap plies to people living in the area of high est average ambient air concentrations, usually nearest a source, is determined by multiplying the unit risk factor for the specific chemical of concern by the highest average exposure estimate in the area under examination. If the unit risk factor for a chemical is 4.0 x lO"5 (|xg/m3)-l and the highest predicted life time exposure concentration is 3 p.g/m3, then the maximum individual lifetime risk estimate would be 4.0 x 10"3 (|xg/ m3)-l) x (3p. g/m3) = 1.2 x 10"4. 98 Pollution Engineering June 1990 ABDOO134446 This means there is one chance in 10.000 that a person residing in the area will contract cancer as a result of the exposure to the pollutant. Aggregate risk applies to all people within the given area of analysis. Aggre gate risk is the summation of the num ber of people multiplied by the esti mated concentration to which they are exposed multiplied by the unit risk fac tor of the pollutant or pollutants of concern. The resulting aggregate risk estimate is expressed as the expected incidence or rate of occurrence of cancer among all people in the analysis after 70 years of exposure. The incidence is often di vided by 70 to obtain an estimated an nual incidence. Two examples follow. The first exam ple illustrates the calculation of aggre gate risk in a single geographical area. The second example illustrates how ag gregate risk can vary among geographi cal areas. The aggregate risk calculation for City B is [(3 x 10,000) + (1 x 50,000) + (0.1 x 200,000)] x 5.0 x 10'5 = 5. The aggregate risk calculation for City C is [(7 x 1,000) + (5 x 5,000) + (1 x 10,000)] x 5.0 x 10'5 = 2.1. Individual risks calculated from concentrations Where the concentration is given in terms of micrograms per cubic meter (M-g/m3), the individual risk is calcu lated as follows: Individual Risk = M-g/ ent concentration of 2.52 ppbv has been measured, what is the individual risk? The individual risk = (2.52) x (2.2 x 10-6) x (44.05) divided by 24.45 = 1.0 x 10'5. The equation for calculating individ ual risk from a given concentration in terms of ppbv can be simplified for each pollutant as: The individual risk = ppbv x factor. The factor in this equation is equal to the unit risk factor of the pollutant multiplied by its molecular weight di Calculating aggregate risk Example 1 -- Within a geographical area, three levels of ambient concentra tions of chemical X have been identi fied: 2 p-g/m3, 1 n-g/m3, and 0.5 jtg/m3. Exposed populations to these three am bient concentrations are 1000, 10,000 and 100,000 people, respectively. The unit risk factor for chemical X is 4.0 x 10'5 (jig/m3)'1). Aggregate risk is then calculated as follows: (2 [lmglg/m3 x 1000 people x 4.0 x 10'5) + (1 n-g/m3 x 10,000 people x 4.0 x 10'5) + (0.5 M-g/m3 x 100,000 people x 4.0 x 10-*) = 2.48. In Example 1, the aggregate risk is estimated to be 2.48 lifetime cancer cases in the exposed population. An nual incidence, which expresses the number of cancer cases per year, would be equal to 0.004 (or 2.48/70). Example 2 -- Three cities have meas ured ambient concentrations for chemi cal Y. The different ambient concentra tions and the number of people exposed to each are summarized below. The unit risk factor of chemical Y is 5.0 x 10'5 (M-g/m3)1. In this example, aggregate risk ranges from two to six cancer cases. The small est city. City C, has the highest maxi mum ambient concentration, but the lowest aggregate risk. The aggregate risk calculation for City A is [(6 x 5,000) + (3 x 20,000) + (0.5 x 50,000)] x 5.0 x 10'5 = 5.75. m3 x unit risk factor. For example, an ambient concentra tion of coke oven emissions has been measured at 1.613 M-g/m3. Coke oven emissions have a unit risk factor of 6.2 x 10"4 (M-g/m3)'1. What is the individual risk associated with this concentration? The individual risk = (1.613 M-g/m3) x 6.2 x 1 O'4 (M-g/m3)'1 = 1.0 x 10'3. Where the concentration of a pollut ant is given in terms of parts per billion by volume (ppbv), the individual risk can be calculated. The individual risk = ppbv x unit risk factor x molecular weight divided by 24.45 L/mole at 25C. As another example, acetaldehyde has a unit risk factor of 2.2 x 10-6 and a molecular weight of 44.05. If an ambi vided by 24.45 L per mole at 25 C. Table 1 presents these factors for 28 pollutants. Gerald Rich is senior field editor for Pollution Engineering. He is a registered professional engineer and a registered environmental professional. He has extensive experience with governmental regulations and environmental consulting. Reader Interest Review Please circle the appropriate number on the Reader Service Card to indicate your level of interest in this article. High 459 Medium 460 Low 461 June 1990 Pollution Engineering 99 ABDOO134447 RESPONSIBILITY FOR A CLEAN ENVIRONMENT IS YOUR BUSINESS. YOUR OPPORTUNITY TO FULFILL IT IS It's finally happened. One event, one investment in the Environmental Technology Expo (ETE) is all you need to serve many industrial markets for your pollution abate ment and control technology. Developed in response to the emerging pollution abate THE ENVIRONMENTAL TECHNOLOGY EXPO ment and control industry, it is a broad- CONFERENCE AND EXPOSITION based marketplace for suppliers who offer solutions for air pollution, water pollution, solid waste management, hazardous waste April 8-11, 1991 McCormick Place East C h i c a go. Illinois management, and recycling. Exposition highlights include: A Qualified attendance by engineers and managers responsible for industrial pollution control and abatement--the people who must make critical investment decisions to com ply with government regulations and public demand. A A hands-on, up-to-the-minute Conference focused on regulations/legislation, new tech nology, and applications. A Sponsorship by nationaland international publications and associations, including Pollution Engineering magazine. A Co-location with the National Plant Engineering & Maintenance Show and the National Design Engineering.Show, providing an extended attendee base of 50,000 engineers and managers. A Chicago location near major Midwest industries and providing a convenient transpor tation hub for all national and international participants. RESERVE YOUR EXHIBIT SPACE TODAY. As a supplier of pollution abatement technology, you cannot ignore industry's demand for your solutions--and you can't ignore your best opportunity to respond to this demand. Plan to exhibit at ETE, the one event that puts you at the heart of the pollution abatement and control marketplace. For more information on exhibiting or attending ETE, contact: Jill Vanderlin, Show Manager Cahners Exposition Group Cahners Ploza 1350 EastTouhyAvenue P.O.Box5060 Des Plaines, II60017-5060 Telephone: (708) 299-9311 FAX: (708) 635-1571 Part of REED INTERNATIONAL P.L.C.