Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
6 Hazard Assessment After completing a hazard identification and dose-response assessment, it is important to place the conclusions of the assessment in the context of similar analyses and in the context of a real-worId situation to ensure that the risk estimates are reasonable in view of available information. Therefore, this chapter will summarize the findings ofthe subcommittee, compare the results of the subcommittee's dose-response assessment with those of the previous NRC ( 1999) subcommittee and EPA (2001), and, finally, examine whether the estimated risks are plausible when considered in the context of the U.S. population. FINDINGS OF THE SUBCOMMITTEE There is increasing evidence that chronic exposure to arsenic in drinking water may be associated with an increased risk of hypertension and diabetes. The existing studies from Taiwan and Bangladesh, discussed in Chapter 2, have observed substantial increases in the risk of these medical conditions at levels of arsenic exposure that are within one to two orders of magnitude of the Tower levels of current regulatory concern in the United States. Pending 214
HAZ4RD ASSESSMENT 215 further research that characterizes the dose-response relationship for these end points, the magnitude of possible risk that exists at low levels is nonquantifi- able. Nevertheless, because these endpoints are common causes of morbidity and mortality, even small increases in relative risk at Tow dose could be of considerable public-health importance. This potential impact should tee quaTi- tatively considered in the risk-assessment process. A sound and sufficient database exists on the carcinogenic effects of arsenic in humans. The main end points for a quantitative risk assessment following exposure to arsenic in drinking water are lung and bladder cancers. The human data from southwestern Taiwan used by EPA in its risk assessment remain the most appropriate for determining quantitative risk estimates. Hu- man data from more recent studies provide additional information for use in the risk assessment. Based on some ofthese studies, the subcommittee recom- mends using an external comparison population when analyzing the earlier studies from southwestern Taiwan, rather than comparing high- and low-expo- sure groups within the exposed population, because of concerns regarding probable exposure misclassification in the low-exposure villages within the data set and because of new data from southwestern Taiwan that suggest that confounding is unlikely. The data on the mode of action of arsenic do not indicate what form of-extrapolation should be used below the exposure range of human data. The observed data should be modeled using a biologically plausible model form that best fits the data to determine a ~ % effective dose (EDo~. The subcommittee used an additive Poisson model with a linear term in dose for the southwestern Taiwan cancer data. The dose-response relation- ship should be extrapolated linearly from the EDGE to zero. Because the hu- man data include exposures to arsenic concentrations relatively close to some U.S. exposures, the distance of extrapolation is very small less than 1 order of magnitude. The subcommittee calculated EDo,s based on the southwestern Taiwanese data (Chen et al. 1985, 1992; Wu et al. 1989), the Chilean data (Ferreccio et al. 2000), and the northeastern Taiwanese data (Chiou et al. 2001). It caTcu- lated cancer risk estimates for the southwestern Taiwanese data (Chen et al. 1985, 1992; Wu et al. 1989), and the Chilean data (Ferreccio et al. 2000) discussed below. Cancer risks were not estimated for northeastern Taiwan (Chiou et al. 2001) because of instability of the model calculated with the small number of cases in that study. . .
21 6 ARSENICINDR]NKING WATER: 2001 UPDA TE COMPARISONS OF RESULTS OF DOSE-RESPONSE ASSESSMENTS Estimates of Effective Dose for a I% Response: EDGE Doses of an agent associated with the onset of a defined rate of observable response in a study population (often termed EDGE when referring to a re- sponse rate of 1°/O) can be useful in several key respects in risk assessment. The EDGE can be used as a point of departure for extrapolation to lower doses (typically to the origin) when insufficient data exist to characterize the shape of the dose-response curve in a region. They can also be used to assess a margin of exposure (MOE) between a dose with observed adverse effects and the level of exposure that exists in the general population. A MOE is calcu- lated by dividing the dose associated with a defined level of response, such as a 1% (EDo~) or 10°/O (EDIT) response in animal or epidemiological studies, by the actual or projected human exposures (EPA ~ 996~eciding which level of response to use (e.g., 1% or 10°/O) is a policy choice that depends, in part, on the size and quality of the epidemiological or animal data sets available. Therefore, the smaller the MOE is for a given population, the closer the popu- lation exposures are to exposures shown to have an adverse effect. The MOE can provide risk managers with information about the extent of apparent pro- tection for the population. The MOE approach is complementary to more traditional approaches for determining a safe level of exposure, each approach providing different information to the riskmanagers (Presidential/ Congressio- nal Commission on Risk Assessment and Risk Management 1997a,b). Be- cause the human epidemiological data set for arsenic encompasses exposure levels close to those for which the subcommittee calculated Edgers, the sub- committee elected to present its EDo~s rather than Eddies used in EPA's margin-of-exposure analyses. Table 5-3 presents the EDo~s estimated by the subcommittee (based on mortality or incidence data, depending on the study) for the Chilean data (Ferreccio et al. 2000), the northeastern Taiwanese data (Chiou et al. 2001), and the southwestern Taiwanese data (Chen et al. ~ 985, ~ 992; Wu et al. ~ 989~. Those values were estimated using a number of statistical models to fit the data, including additive and multiplicative models using linear or logarithmic terms in dose. The EDo~s were estimated using the published or calculated relative risk values and a modification of the BEIR IV (NRC 1988) formula, as described in Chapter 5. Despite the variability, it is evident that most ofthe
HAZARD ASSESSMENT 21 7 EDo~s are less than a factor of 10 higher than the current U.S. maximum con- taminant level (MCL`) of 50 ,ug/~. The subcommittee determined EDo~s (i.e., the dose at which there is a I% response in the study population) for various studies using a number of statis- tical models. For example, the estimated EDo~s from the Chilean study on lung cancer ranged from 5 to 27 ~g/L, depending on the exposure data used. The previous Subcommittee on Arsenic in Drinking Water estimated EDo~s of 404 to 450 vigil, depending on the model used, for arsenic and male bladder cancer mortality. Those values are approximately within the range of EDo~s estimated by this subcommittee. However, because the EDGY values reported by the current and prior subcommittees were derived using different biostatistical approaches, they are not directly comparable. The EDGY values in NRC (1999) reflect a I% increase relative to background cancer mortality in Taiwan, whereas the current subcommittee's approach, using a modifica- tion of the BE1:R IV analysis (NRC 1988), reports EDo~s based on a 1°/O in- crease relative to the background cancer mortality in the United States. This is an important difference because the background rates for lung and bladder cancer are substantially different between Taiwan and the United States. Background rates for lung cancer in the United States are approximately 3- and 2.3-fold higher than in Taiwan for females and males, respectively; and bladder cancer risks are approximately 1.4- and 3-fold higher in females and males, respectively, in the United States when compared to Taiwan. Cancer Risk Estimates The subcommittee presents the theoretical lifetime excess cancer risks for lung and bladder cancer incidence for the U.S. population (females and males calculated separately) at fixed arsenic concentrations in drinking- water of 3, 5, 10, and 20 go/. Table 6-1 presents the maximum-likelihood estimates (MLEs) of the risk of bladder and lung cancer combined based on the data from southwestern Taiwan. Estimates calculated using the U.S. background cancer incidence data or Taiwanese background cancer incidence data are presented. The U.S. background cancer incidence data is taken from SEER (2001~. The Taiwanese background cancer incidence data were estimated by multiplying the subcommittee's corresponding U.S. lifetime incidence rate (Tables 5-7 and 5-~) by the ratio of the Taiwanese annualized rate (You et al. 2001) to the U.S. annualized rate (Ferlay et al. 2001~. The relatively small confidence limits around the MLE (+/- less than 12% ofthe MLE) reflect the
21~ ARSENIC IN DRINKING WA TER: 2001 UPDA TE TABLE 6-1 Theoretical Maxi Likelihood Estimates of Excess Lifetime Risk (Incidence per 10,000 people) of Lung Cancer and Bladder Cancer for U.S. Popula- tions Exposed at Vanous Concentrations of Arsenic in Drinking Watera Bladder Cancer Arsenic Concentration (fig/ L) Lung Cancer U.S. Back- ground Rateb T. alwanese Background Ratec Taiwanese U.S. Background Background Rateb Rater - Fernales Males Females Males Females Males Females Males 3 3.6 6.8 2.3 2.0 5.4 4.0 1.8 1.7 5 6.0 1 1 3.8 3.2 8.9 6.8 3.0 3.0 10 12 23 7.5 6.8 18 14 6.2 6.1 20 24 45 15 13 36 27 12 1 1 a Estimates were calculated using data from individuals in the arsenic-endemic region of south- western Taiwan and data from an external comparison group from the overall (mostly unex- posed) southwestern Taiwan area. The risks are estimated using what the subcommittee consid- ered reasonable assumptions. (A U.S. resident weighs 70 kg compared with 50 kg for the typical Taiwanese, and the typical Taiwanese drinks just over 2 L of water per day compared with 1 L per day in the United States. Therefore, it assumes that the Taiwanese exposure per kilogram of body weight is 3 times that of the U.S. population.) It is possible to get higher and lower estimates using other assumptions. Risk estimates are rounded to two significant figures. All 95% confidence limits are less than +/- 12% ofthe maximum-likelihood estimate and are not presented. It should be noted that those confidence limits are a function of sample size and are not indicative of the true uncertainty associated with the risk estimates. The individual risk estimates for bladder and lung cancer were added together to estimate combined risks. b Risks are estimated using the U.S. background cancer rate (SEER 2001~. c Risks are estimated using the Taiwanese background cancer rate. The Taiwanese background cancer incidence data were estimated by multiplying the subcommittee's corresponding U.S. lifetime incidence rate (Tables 5-7 and 5-8) by the ratio of the Taiwanese annualized rate (You et al. 2001) to the U.S. annualized rate (Ferlay et al. 2001~. relatively large sample size, end they are not indicative ofthe true uncertainty associated with the risk assessment discussed in Chapters 4 and 5. The MLEs of the lifetime excess risks for combined lung and bladder cancer incidence for females range from 9 per 10,000 from exposure to drinking water with arsenic at 3 ,ug/L to 60 per 10,000 from exposure to drinking water with ar- senic at 20 ,ug/~. The corresponding risk estimates for males are ~ ~ to 72 per 10,000. Those values are estimates of the combined lifetime excess risk of lung and bladder cancer (incidence) in a given population following lifetime exposure to arsenic in drinking water at the given concentration.
HAZARD ASSESSMENT 2~9 As presented in Chapter 5, the subcommittee used data from a study per- formed in northern Chile (Ferreccio et al. 2000) to estimate the theoretical lifetime excess risk of incident lung cancer in U.S. males and females at ar- senic concentrations in drinldug water of 3, 5, 10, and 20 ,ug/~. Using the peak period of arsenic exposure in Chile from 1958 to 1970 as a dose metric, the resulting estimates for excess lung cancer incidence in the United States were 3 to 4 times higher than the risks derived from the Taiwanese data. in contrast, when the dose metric used in the Chilean data was the average ar- senic concentration in drinking water from 1930 to 1994, the corresponding risk estimates were an order of magnitude higher. The previous Subcommittee on Arsenic in Drinking Water presented lifetime excess cancer risk estimates for bladder cancer mortality in males based on its analyses of the southwestern Taiwanese data (Chen et al. 1985, 1992; Wu et al. 19894. Some ofthose risk estimates are presented in Table 5- l. Those risks were estimated using an external comparison population and a multiplicative linear model. At an arsenic concentration of 50 ~g/L of drink- ing water, the excess risk of bladder cancer mortality for males was estimated to be 10 tol5 per 10,000 (NRC 1999~. Assuming linearity and dividing by 5, that corresponds to a mortality risk estimate of 2 to 3 per 10,000 at 10 ,ug/~. If the U.S. mortality rate for bladder cancer is 20% (SEER 2001), that corre- sponds to an estimated risk of bladder cancer incidence of 10-15 per 10,000. Using the southwestern Taiwanese data, this subcommittee's estimate for lifetime excess bladder cancer incidence in mates in the United States at an arsenic concentration of 10 ,ug/L is 23 per 10,000 (see Table 6-~. Therefore, although some analytical approaches were different, the estimates for bladder cancer riskin males for arsenic at 10 ,ug/L, of drinking water determined by the subcommittee in this report are generally consistent with those presented in the previous NRC report. As discussed in Chapter 5, EPA did not present theoretical lifetime excess cancer risk estimates for arsenic in drinking water in its notices in the Federal Register (2000, 2001~. The risk estimates it presents (EPA 2001) are adjusted for the occurrence of arsenic in U.S. drinking water; consideration of such an adjustment is beyond the charge to this subcommittee. It is not possible to directly compare the theoretical lifetime cancer risks estimated by this sub- committee with those presented by EPA. The different assumptions used by EPA (2001) and this subcommittee are presented in Table 6-2. The subcommittee did, however, use a linear extrapolation from the EDo~s estimated in the analysis on which EPA based its risk estimates (Morales et al. 2000) to estimate the theoretical lifetime excess bladder and lung cancer risks at 3, 5, 10, and 20 ,ug/L, presented in Table 5-2. Thus, the subcommittee
220 ARSENIC IN DRINKING WA TER: 2001 UPDA TE TABLE 6-2 Summary of Assumptions Used by EPA and the Subcommittee for Dose-Response Analyses and Their Impact on the EPA's Risk Estimates Relative to the Subcomm~ttee's Risk Estimatesa Study Parameter EPA (2001) Subcommittee Impact Choice of End Lung and bladder cancer Lung and bladder cancer No Point difference Choice of Study Southwestern Taiwanesecan- Southwestern Taiwanese No cer mortality data from Chen et cancer mortality data from difference al. (1985, 1988, 1992) Chen et al. (1985, 1988, 1 992) Model Choice Multiplicative Poisson regres- sion model with linear extrapo- lation Additive Poisson regression Decrease with linear extrapolation; BEIR IV (NRC 1988) Selection of No external comparison group External comparison group Decrease Comparison used used Group Adjustments for U.S. population: Monte Carlo Water Intake analysis of CSFII (EPA 2000) water intakes Taiwan population: water con- sumption is 3.5 L for males and 2.0 L for females U.S. population: Mean daily average from CSFII of 1 today for males and fe- males Taiwan population: expo- sures equal to 3 times U.S. default value, i.e., 3 IJday for males and females Decrease Adjustmentsfor Taiwan: Adjustedlower Taiwan: added a constant Decrease Dietary Intake bound estimates to account for concentration of arsenic (30 of Arsenic arsenic from cooking water by payday) to exposure rates adding 1 L of water; therefore, for all individuals in study total water intake for males villages was 4.5 LJday and for females was 3.0 IJday. Also to account for intake from food directly, multiplied lower bound estimate by fraction of arsenic consumed per kilogram contributed by drinking water Adjustments for Used Taiwanese mortality data Used U.S. background inci- Decrease Mortality ver- for bladder and lung cancers; dence data for bladder and sus Incidence adjusted upper bound by 1.25 lung cancers from SEER for bladder cancer to reflect (2001 ) database mortality, assumed all lung cancer is fatal in Taiwan a More detailed information about these assumptions can be found in Chapter 5. Abbreviations: CSFII, Continuing Survey of Food Intakes by Individuals; SEER, Surveillance, Epidemiology, and End Results.
HAZARD ASSESSMENT 221 compared its risk estimates with those estimates calculated from the published analyses (Morales et al. 2000) on which EPA based its risk estimates (Table 5-2~. The subcommittee notes, however, that the estimates in Table 5-2 are not adjusted for water consumption or arsenic in food in the same manner as used by EPA, nor by this subcommittee in its analysis in Chapter 5. (The adjustments used by EPA for food and water consumption would decrease the risk estimates.) However, even without those adjustments, the risk estimates on which EPA based its analyses are Tower than this subcommittee's esti- mates, regardless of whether the U.S. or Taiwanese background cancer rates are used to estimate the risks. Several factors contribute to that difference. Unlike the subcommittee ' s estimates, EPA' s analyses were based on estimates that were calculated without using an external comparison population. The subcommittee also used a different statistical method than EPA to estimate lifetime cancer risks. The subcommittee has presented lifetime excess cancer risk estimates calculated using either the U.S. or the Taiwanese background rates; Morales et al. (2000) estimated the EDo~s using Taiwanese background rates. The magnitude of the difference between the estimates can be seen in Table 6-1 . In addition, the method the subcommittee used to adjust for arsenic in food and its assumptions regarding water intake in the U.S. and Taiwanese populations were different from those used by EPA in its analyses. It should be noted that the subcommittee was split on whether using the U. S. background rates was preferable to using the Taiwanese background rates for estimating arsenic risks in the United States. Some members of the sub- committee felt strongly that using U.S. background rates was the preferred approach, while others felt that there was not sufficient justification to select one set of background rates over the other, and that both should be presented. Thus, the results from both approaches are presented in Table 6-l . The sub- committee agreed, however, that if there was a multiplicative interaction between a complex array of risk factors, including smoking, that establish the background rates, then using the U.S. background cancer incidence rates would be preferred over the Taiwanese background rates for estimating ar- senic cancer risks in the U.S. population. PLAUSIBILITY OF CANCER RISK ESTIMATES Upon completion of an assessment of the potential health effects of an environmental contaminant, it is wise to compare the results ofthe assessment with a real-world situation that is, the adverse health effects observed among the people most exposed to the contaminant. The key factors triggering public-
222 ARSENIC IN DRINKING WA TER. 2001 UPDA TE health concern regarding arsenic in drinking water have been the high inci- dences of different types of cancer in populations exposed to increased con- centrations of arsenic in drinking water (greater than 100 ,ug/~) in Taiwan, Chile, and Argentina. The cancer with the highest increases in relative risk in these countries is cancer of the bladder. It has been suggested that, if the risks of bladder cancer from arsenic in drinking water were indeed as high as estimated in this report (see Table 6-~), high cancer rates would have been anticipated in areas of the United States with increased concentrations of arsenic in groundwater, and these high rates would have readily attracted public-health attention. Some simple calcula- tions demonstrate how risk estimates for low-level arsenic exposure in this report might be difficult to detect by observing geographical differences in cancer incidence or mortality. To illustrate that point, the subcommittee used its risk estimate of 45 per 10,000 for bladder cancer incidence in U.S. males (based on the Taiwanese data, U.S. cancer incidence data, and a ratio of 3 for water ingestion on a per-body-weight basis for the Taiwanese population compared with the U.S. population) exposed to arsenic at a concentration of 20 ,ug/L (Table 6-~. The lifetime risk of being diagnosed with bladder cancer in U.S. mates is 3.42% for the period of 1 996-1998 (or 342 per 1 0,000) (SEER 2001~. An increased risk of 45 per 10,000 over a background risk of 342 cases in 10,000 mates would be difficult to detect. In terms of bladder cancer mor- taTity, if it is assumed that only about one in five bladder cancer cases in the United States results in death (the ratio of mortality to incidence is approxi- mately 20% for U.S. males, SEER 2001), a lifetime excess risk for mortality from bladder cancer in U.S. mates is about 9 in 10,000 following lifetime exposure to arsenic in drinking water at 20 ,ug/~. The subcommittee further explored how that risk contributes to overall U.S. mortality for bladder cancer. Lifetime mortality for bladder cancer in the United States for mates is 0.72% (72 per 10,000) for the period of 1996-1998 (SEER 2001~. That increase in mortality risk of 9 per 10,000 would be difficult to detect against that back- ground rate of 72 per 10,000. Indeed, it would represent only about 13% of the total risk ofbladder cancermortality. Furthermore, the denominator ofthe risk estimate for arsenic assumes that all 10,000 individuals are at risk (e.g., all consume arsenic at 20 ,ug/L of their drinking water for a lifetime). Detec- tion is further complicated by the variability in the actual exposure to arsenic in drinking water (not considered by this subcommittee), the unknown distri- bution of other risk factors (especially smoking), and the mobility of the U.S. population. However, if the risks for arsenic-related bladder cancer were
HAZARD ASSESSMENT 223 higher than the estimate used in this example, then bladder cancer incidence and mortality at exposures of 20 ~g/L would be proportionately higher and thus might be easier to detect in a population. Because background lung cancer mortality is almost 10-fold greater than bladder cancer, it would be even more difficult to demonstrate an association between Tow concentrations of arsenic in drinking water and lung cancer risk. Therefore, although the subcommittee's risk estimates are of public-health concern, they are not high enough to be easily detected in U.S. populations by comparing geographical differences in the rates of specific cancers with geographical differences in the concentrations of arsenic in drinking water. In accordance with its charge, the subcommittee has not conducted an exposure assessment and subsequent risk characterization and risk assessment. The theoretical lifetime excess cancer risks estimated by the subcommittee and presented in this report, however, should be interpreted in a public-health context using an appropriate risk-management framework, such as that pro- posed by the Presidential/Congressional Commission on Risk Assessment and Risk Management (1997a,b). SUMMARY AND CONCLUSIONS · The subcommittee's evaluation and analyses of the data from south- western Taiwan indicate that the lifetime excess cancer risks in the United States for bladder and lung cancers combined at arsenic concentrations in Winking wafer between 3 and 20 ~g/L (ppb) are estimated to be between 9 and 72 per 10,000 people based on U.S. background cancer incidence data. (The corresponding range based on Taiwanese background cancer incidence data is 4 to 24 per 10,000.) These estimates can be interpreted in light of EPA's stated goals for public-health protection (EPA ~ 992~. · Depending on the dose metric used in the study, excess risk estimates for cancer in the United States derived from a recent investigation in Chile are either similar to or higher than risk estimates derived from the Taiwanese data. · Although these risk estimates are high, they would not be detected in U.S. populations by comparing geographical differences in the rates of spe- cif~c cancers with geographical differences in the concentration of arsenic in drinking water.
224 ARSENIC IN DRINKING WA TER: 2001 UPDA TE REFERENCES Chen, C.J., Y.C. Chuang, T.M. Lin, and H.Y. Wu. 1985. Malignant neoplasms among residents of a blackfoot disease-endemic area in Taiwan: high-arsenic artesian well water and cancers. Cancer Res. 45~11 Pt 2~:5895-5899. Chen, C.J., M. Wu, S.S. Lee, J.D. Wang, S.H. Cheng, and H.Y. Wu. 1988. Atherogenicity and carcinogenicityofhigh-arsenic artesian well water. Multiple risk factors and related malignant neoplasms of blackfoot disease. Arteriosclero- sis 8~5~:452-460. Chen, C.J., C.W. Chen, M.M. Wu, and T.L. Kuo. 1992. Cancer potential in liver, lung, bladder and kidney due to ingested inorganic arsenic in Mung water. Br. J. Cancer 66~5~:888-892. Chiou, H.Y., S.T. Chiou, Y.H. Hsu, Y.L. Chou, C.H. Tseng, M.L. Wei, and C.J. Chen. 2001. Incidence of transitional cell carcinoma and arsenic in ~king water: a follow-up study of 8,102 residents in an arseniasis-endemic area in Northeastern Taiwan. Am J. Epidemiol. 153~5~:411-418. EPA (U.S. Environmental Protection Agency). 1992. Drinking water; national pri- mary drinking water regulations-synthetic organic chemicals and inorgaruc chem- icals; national prunary ~ng water regulations implementation. Fed. Regist. 57~138~:31797. EPA (U.S. Environmental Protection Agency). 1996. Proposed Guidelines for Car- cinogen Risk Assessment. Fed. Regist. 61 (799: 17959- 18011. EPA (U.S. Environmental Protection Agency). 2000. Estimated Per Capita Water Ingestion in the United States: Based on Data Collected by the United States Department of Agriculture's (USDA) 1994-1996 Continuing Survey of Food Intakes by Individuals. EPA-822-00-008. Office of Water, Off~ce of Standards and Technology, U.S. Environmental Protection Agency. April 2000. EPA (U.S. Environmental Protection Agency). 2001. 40 CFR Parts 9,141 and 142. National Pruna~y Drin~g Water Regulations. Arsenic and Clarifications to Compliance end New Source Contaminants Monitoring. Final Rule. Fed. Regist. 66~14~:6975-7066. (January22~. Ferlay, J., F. Bray, P. Pisani, and D. M. Parkin. 2001. GLOBOCAN 2000: Cancer Incidence, Mortality and Prevalence Worldwide, Version 1.0. IARC CancerBase, No.5. IARC Press, Lyons, France. "Online]. Available: http://www.dep.iarc.fr/ globocan/globocan.hun. [Sept. 20, 2001]. Ferreccio, C., C. Gonzalez, V. Milosavljevic, G. Marshall, A.M. Sancha, and A.H. Smith. 2000. Lung cancer and arsenic concentrations in ~inking water in Chile. Epidemiology 11(6):673-679. Morales, K.H., L. Ryan, T.L. Kuo, M.M. Wu, and C.J. Chen. 2000. Risk of internal cancers from arsenic in ~nking water. Environ. Health Perspect. 108~7~:655- 661. NRC (National Research Council). 1988. Health Risks of Radon and Other Inter- nally Deposited Alpha-Emitters: BEIR IV. Washington, DC: National Academy Press.
HAZARD ASSESSMENT 225 NRC (National Research Council). 1999. Arsenic in Drinking Water. Washington, DC: National Academy Press. PresidentiaVCongressional Commission on Risk Assessment and Risk Management. 1997a. Framework for Environmental Health Risk Management. Final Report. Vol.1. Washington, DC: GPO. Presidential/Congressional Commission on Risk Assessment and Risk Management. 1997b. Risk Assessment and Risk Management in Regulatory Decision-Making. Final Report. Vol.2. Washington, DC: GPO. SEER (Surveillance, Epidemiology, and End Results). 2001. Surveillance, Epidemi- ology, and End Results Program Public-Use Data (1973-1998), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch. "Online]. Available: ht~://www-seer.ims.nci.nih.gov/Publications/. [August20, 2001]. Wu, M.M., T.L. Kuo, Y.H. Hwang, and C.J. Chen. 1989. Dose-response relation between arsenic concentration in well water and mortality from cancers and vascular diseases. Am.J.Epidemiol.130~6~:1123-1132. You, S.L., T.W. Huang, and C.J. Chen. 2001. Cancer registration in Taiwan. Asian Pac. J. Cancer Prev. 2( IACR suppl.~:75-78.