Document xjDYvgojyv3dab57YQj6MQq2Q

.T "v Br. W si fl environmental research 34, 250-261 (1984) Lung Cancer in Relation to Environmental Pollutants Emitted from Industrial Sources Linda Morris Brown, Linda M. Pottern, and William J. Blot Environmental Epidemiology Branch, National Cancer Institute, Landow Building, Room 3CIS, Bethesda, Maryland 20205 Received January 23, 1983 | A case-control study of residents who lived in the vicinity of a primary zinc smeller and a large steel manufacturing plant in eastern Pennsylvania was undertaken to investigate the role of environmental pollutants in the etiology of lung cancer. Lifetime residential, occupational, and smoking histories were obtained from the next of kin of 335 white male lung cancer cases and 332 white male controls. Soil samples were collected and analyzed for content in ppm of arsenic, copper, lead, manganese, zinc, and cadmium. Relative risks were determined according to distance of residence from the zinc smelter and the steel plant, and according to residence in areas with heavy and light levels of various pollutants. Twofold risks for lung cancer were associated with residence near the zinc smelter and with residence in areas with heavy levels of arsenic and cadmium, although the number of individuals living in these higher risk areas was small. These increases were not explained by the effects of cigarette smoking or by employment in the zinc or steel industry. No excess risk was associated with living near the steel plant. The limited size of the study precludes causal interpretation, but the findings suggest the need for further investigation of metallic air pollution and lung cancer. j 1 I ) f I I I 1 INTRODUCTION The finding of increased rates of lung cancer among both males and females residing in counties with copper, lead, or zinc smelting and refining operations (Blot and Fraumeni, 1975) raised the possibility that air pollution from the nonferrous smelters may present a health hazard. Each of these ores contains inor ganic arsenic, a known lung carcinogen (Landrigan, 1981). A case-control study of lung cancer in a tricounty area of Pennsylvania with primary zinc smelting and refining offered the opportunity to investigate this association further. Since the study area also contained a large steel manufacturing and processing plant, we assessed the risk of lung cancer associated with living in the vicinity of either the zinc or the steel plant. I > Study Area The area studied involved three contiguous counties in eastern Pennsylvania (Fig. 1). Two industries of interest in the area were primary zinc smelting and refining and steel manufacturing and processing, each major employers (approx imately 1500 and 15,000 employees, respectively) with large physical plants con taining over 100 stacks. The zinc facility included two smelters, the west plant and the east plant, which began operations in 1898 and 1910, respectively. They are located approximately 4 km apart in a small town in a narrow valley bordered ? Fig. and ma. by !o\ (Fig.: steel i 1904. along and is the w< Cor identil white 1976were respir; with r certifi. next c 1980 \ structi and sr All jo categc cation one in worke LUNG CANCER IN RELATION TO POLLUTANTS 251 its Emitted i J. Blot ng, Room 3CIS, c smelter and ivestigate the iential, occu- lite male lung analyzed for ive risks were c steel plant, lutants. Two- Fig. I. Diagram of the tricounty area showing locations of the zinc plant, steel plant, soil samples, and major rivers and mountains. Itcr and with te number of not explained industry. No of the study investigation by low mountains. The prevailing winds are from the northwest and southwest (Fig. 2). The integrated steel facility consisting of a coke oven, blast furnaces, i steel making furnaces, finishing mill, and foundry has been in operation since 1904. Its major product line is structural steel. It extends for nearly five miles ) along the south bank of the Lehigh River in close proximity to an urbanized area l and is bounded on. the south by small mountains. The prevailing winds are from i the west and northeast. ^^and females operations MATERIALS AND METHODS from the noncontains inor-control study tc smelting and ther. Since the tsing plant, we ty of either the I Computerized mortality tapes from the state of Pennsylvania were used to identify ail 360 deaths attributed to primary lung cancer (ICDA 162.1) among white male residents aged 30-79 of Lehigh and Northampton counties during 1976-1977 and Carbon county during 1974-1977. An equal number of controls were randomly selected from among other causes of death, excluding chronic respiratory diseases, respiratory cancer, and suicide, to be similar to the cases with respect to sex, year and age at death, and county of usual residence. Death certificates were obtained for the cases and controls and were used to locate the next of kin. Personal interviews of the next of kin were conducted during 1979-- i Pennsylvania c smelting and oyers (approxcal plants conthe west plant rcctiveiy. They alley bordered i 1980 by local interviewers unaware of the case-control status of decedents. A structured questionnaire was used to elicit complete residential, occupational, and smoking histories in addition to demographic and other variables of interest. All jobs held for 6 months or longer were classified into one of 15 industrial categories using a coding scheme adapted from the Standard Industrial Classifi cation Manual (OMB, 1972). An individual's usual industry was defined as the one in which he had spent the longest number of years, provided that he had worked in this industry for at least 15 years. The usual number of cigarettes 252 BROWN, POTIERN, AND BLOT Fig. 2. Plot of percentage of occurrence for wind direction near the zinc plant (grid 5 4>. smoked per day was grouped into the following categories: nonsmoker: ight or I former smoker--smoked half a pack or less or quit smoking at least I > years prior to death; moderate--smoked more than half a pack but less than lw\ packs; and heavy--smoked two or more packs per day. Nineteen samples of the top 5 cm of soil were collected by Lehigh Un versity in October 1980 at locations in the three-county area where Environmen al Pro- j lection Agency (EPA) air monitors had been previously located. Except f- r three samples collected near the zinc plant, all were in areas where the soil w s "an- ' disturbed"; that is, where trees or large shrubs were growing wild indicr ing no disturbance for five or more years. Heavy metal contents were determ icd by ' atomic absorption spectrophotometry using two replicate samples, back round correction, and standard solutions (Grewling. 1976). Copper, lead, man mese, and zinc determinations were carried out at Lehigh University. Arsenic a d cad- , mium determinations were performed by the Department of Agronomy. ' ornell \ University. Limited air data for sulfur dioxide (SO;) were obtained for - itions , near the zinc and steel plants from the Commonwealth of Pennsyivani. s De- partment of Environmental Resources Bureau of Air Quality and Noise . mtroi (personal communication. 1981). Geologic survey maps of the three counties were used to assign each rc orted residence to a l-km: grid. Grid coordinates were also assigned to the zi c and steel plants and to areas where soil and air sample measurements were nade. Each grid from which a soil sample had been taken and the eight contiguou grids surrounding it were given a score of heavy or light based on the level >f the pollutant of interest. The grid of longest residence and the grid of residence uring the individual years 1920. 1930. 1940. 1950. I960, and 1970 were determine . and case-control differences according to distance from the zinc and the steel >iants I and by level of heavy metals present in soil samples were analyzed. The following model adapted from Stinnett ct al. (1981) was also used t< esti mate relative exposure (A',) according to distance {D.) from the two zinc hints IV VT LUNG CANCER IN RELATION TO POLLUTANTS 253 and fro: .! the steel plant. = -t - i wk I/>,t where Da is the distance from the Ath grid to the t'th subject. VV't is an emissions weight for the Ath grid, and P is number of grids in which the zinc or steel plants were located. The measure of nearnes . Xj. is based on the idea that ambient air conditions vary inversely with distance from the source of emissions and directly with emissions rate, and that f concent ations from several stacks are additive. Since sufficient data were not available for the steel plant to determine the relative emissions of specific pol lutants per grid, each grid was assigned an emissions weight of 1. The grids containing the east and west zinc smelters were given weights of 6 and 7 (cast plant), md 4 and 3 (west plant) to represent the relative emissions of zinc and cadmim i. respectively, as documented by Buchauer (1973). All zinc plant grids were gi en an emissions weight of i for the other heavy metals. Exposures from the stee and the zinc plants were not assessed simultaneously as they were over 25 km part and the environmental pollutants emitted from each source were differed. Relative exposures from each source to each grid of residence were lant Igrid 5t ismoker; ht or ! at least I years s than two acks; then grouped into three proximity categories--Far, Middle, or Near--based on the con oined distributions of cases and controls. Estimates of the relative risk (RR) of lung cancer associated with residence were cj culated using the odds ratio for 2 x 2 tables (Fleiss. 1981). Significance tests (c ;e-sided because of the a priori hypotheses of increased risk with expo sure) ai-.i corresponding 909c confidence intervals (Cl) were calculated according ehigh Urn .rsity to the Methods described by Gart (1970). Results reported elsewhere (Blot et til., vironmen Pro- 1983) re vealed increased risks associated with long-term employment in both the Except fc three steel arid zinc smelting industries, although cigarette smoking was the predomi he soil w; "un- ' nant ri:-< factor for lung cancer. Thus, wherever our numbers permitted. RR for indica ig no residen ud exposures were looked at controlling for these factors. A multiple etermi d by ` logistic model for disease incidence using stratified matching criteria [county of . back ound usual residence and five age groups (34-56. 57-62. 63-67. 68-72. and 73-79)) cad. mant tese. was applied in order to simultaneously control residential exposures for the in Arsenic at cad- fluence of smoking and usual occupation in steel manufacturing or zinc smelting xonomv, rnell ; (Lubin 1981). The four smoking categories--nonsmoker, light, moderate, and ined for s tions heavy- -were assigned scores of 0. 0.5. 1.25. and 2.25. respectively. insylvanic De al Noise ( ntrol RESULTS un each re 'tied to the zi and j ents were ade. contiguou rrids the level the residence ring determine and .1 the steel ants Inte. iews were completed for 345 cases (9690 and 340 controls (9490. Ten cases : nd eight controls were deleted from the analysis because of cither misdi agnose. of lung cancer determined by medical record review or incomplete ques tions--es. Analyses were conducted on the remaining 335 cases and 332 controls. The locations of the zinc and steel plants, the soil samples, and the major rivers and m >untains are shown in Fig. 1. Data from the soil samples are presented in Table for each heavy metal pollutant by grid where collected. Each pollutant was a signed a value of heavy (H) or light (L) for purposes of analyses and each soil sti nple grid was given a value of Far (F). Middle (M). or Near (N) to indicate zed. its rel. five exposure to the zinc smelter. Also presented is the range in unpolluted ilso used t esti- soil fi. cadmium (Rose et til.. 1979) and the mean in U.S. soils for the other two zinc ants heavy metals (Connor and Shacklette. 1975). Compared to the reported values. jn 'n.uLun ,iki g iid unbeaten w hether the pollutant level was considered to he heavy o r lig h t fo r (hat grid. cT h. V z' y. z iu - ni.i9 w ill joi ititiii .10 L mmu so m n.n.nMi.iii'1 LUNG CANCER IN RELATION TO POLLUTANTS 255 the levels lor the pollutants tended to be substantially elevated, especially near the zinc smelter (grids 5036. 5134. 5233). Even though the soil samples taken near the smelter were from 'disturbed" soil, they were similar in magnitude to values reported by Buchauer (1973). As can be seen from the table, it was impossible to separate out the effects of cadmium and zinc by an analysis of H and L grids because of the strong correlation between the two in this study. Similarly, a separate effect of arsenic could not be evaluated because grids H in cadmium and zinc, but L in arsenic contained only one person. From the available SO: data. 20 measurements in 1970 downwind of the zinc smelter (grid 5334) yielded a yearly mean of 416 ng/m3 and a 24-h maximum of 676 p.g/nv\ while 35 measurements in 1974 near the steel plant (grid 7154) yielded a yearly mean of 39 pg/m3 and a 24-hr maximum of 260 n.g/m\ The levels near the zinc smelter were higher than the National and Pennsylvania primary stan dards of 80 and 365 p.g/m3 for I-year mean and 24-hr maximum values, respec tively. The RR of lung cancer associated with usual residence in areas with H and L levels of the six heavy metal pollutants are presented in Table 2. Significant increases were seen for study subjects residing in grids with high levels of arsenic (RR = 23,/? = 0.03) and cadmium/zinc (RR = 2.0. P = 0.05) but not for copper, lead, or manganese. When these risks were adjusted for the effects of smoking and usual occupation in zinc and steel, the RR for lead and manganese were reduced to less than 1.0, while the RR for arsenic, cadmium/zinc, and copper were reduced to 1.6 (P - 0.21), 1.5 (P = 0.24). and 1.1. respectively. Figure 3 shows the grids of usual residence for a 20-km: area surrounding the zinc smelter. Grids of residence at proximities far. middle, or near to the smelter are shaded. The white grids indicate areas where none of the study population usually lived. The area surrounding the smelter is sparsely populated with large unsettled areas due to mountains. The RR. however, were above 1.0 for those TABLE 2 RR ok Lino Canter according ro Usual Residence in Grids with Light or Heavy Levels ok Various Pollutants Pollutant Level No. of cases No. of controls RR" (W* CIl Arsenic Light Heavy 57 16 74 1.0 __ 9 2.3 11.0-5.4) ' Cadmium (zinc) Copper Lead Manganese Light Heavy Light Heavy Light Heavy Light Heavy 57 16 43 30 19 54 5! 73 1.0 -- 10 2.0 (0.9-4.6) 56 1.0 _ 27 1.4 (0.8-2.7) >* 1.0 -- 61 1.0 (0.5-2.0) 60 1.0 -- 23 1.1 (0.6-2.1) Risk for each pollutant is relative to RR of 1.0 for light. 256 BROWN, POTTERN, AND BLOT PROXIMITY TO ZINC PLANT NEAR MIDDLE I-AH S3 NO STUOY SUBJECTS UV60 IN THESE AREAS Flo. 3. Proximity to the zinc smeller for grids of usual residence in a 20-km- area. study subjects who lived near the smelter (Table 3). This nonsignificant increase is apparent for both usual residence and residence in 1950. Similar patterns of risk were seen for residence in each decade year. When the RR were adjusted using a multiple logistic model which simultaneously controlled for the influence of cigarette smoking, and usual employment in steel manufacturing and zinc smelling, the RR for near proximity in 1950 was significantly elevated. For both usual and 1950 residence, there were trends of increasing risk with increasing proximity to the smelter. When the proximity patterns were redefined using the zinc and cadmium weighted exposure measurements, RR comparable to those presented for the unweighted exposures were found. When the near category was expanded to include the next three closest grids to the west of the smelter 15031, 5032, 5130) the RR were still elevated, although somewhat reduced. With the TABLE 3 RR of Lung Cancer According to Proximo y to the Zinc Smelter for Usual Resident s: and Residence in 1950 Residence Proximity No. of cases No. of controls RR" (90% CD RR* (90% Cl) Usual Far Middle Near 292 26 16 292 1.0 _ 1.0 . 29 0.9 (0.5-1.5) 1.2 (0.6-2.t) 10 1.6 (0.8-.3.4) 1.6 (0.6-4.31 1950 Far Middle Near 242 27 14 234 1.0 -- (.0 -- 24 1.1 (0.6-1.8) 2.3 (1.1-4.6) 7 1.9 (0.8-4.7) 5.0 (1.4-!7 " All risks relative to a RR of 1.0 for far proximity. * RR adjusted for smoking and usual employment in steel or zinc. RR T *5 JlNC PLVvT r- .'F !0-km; area. tnificant it rease milar patu ns of R were a< Jsted for the in! ence cturing at zinc evated. Ft both rith inci using ined us; the parable to hose car catego was ne smelter 031, duced. Wi the LUNG CANCER IN RELATION TO POLLUTANTS 257 prevailii -s winds at the smelter from the northwest and southwest, it was reason able to i. sume that residents upwind of the smelter would receive less pollution than tho ; downwind: therefore, it was felt justified to have the next three closest grids in he middle distance category. ' Many ->f the men who lived near the zinc smelter also worked there. Table 4 shows, however, that elevated risks were seen for those with 1950 residence near the smei er regardless of their employment status. Similar results were seen for usual revdence and residence in each decade year. The sh aded grids in Fig. 4 show the location of the usual residences of the study population for a 20-km: area surrounding the steel plant. The area surrounding the steel plant is more densely populated than that surrounding the smelter, but there at-: areas of sparseness due to topographical barriers. Unlike the zinc smelter, to elevated risks are seen for residence near the steel plant (Table 5). The risk for middle and near proximity become even further reduced when controlled for smoking, and usual employment in steel or zinc in the multiple logistic model. DISCUSSION . Air pollution has long been a suspected risk factor for lung cancer, although a clear link between the two has yet to emerge. To evaluate the issue, several recent investigations have focused on the lung cancer experience of populations living around point sources of pollution. Environmental exposures near arsenic-emitting industries were examined in studies by Pershagen et al. (1977). Lyon et al. (1977). Reed et :il. (1978). Greaves ct al. (1981). Matanoski et al. (1981). Rom et al. (1982). and Cordier et al. (1983). . Cordicr et al. (1983) found elevated rates of lung cancer in men living near a copper smelter. Matanoski et al. (1981) found elevated lung cancer risks in men living near a chemical plant which produced arscnicals. and Reed et al. (1978) found lu ig cancer rates to he highest in the residential areas closest to a nickel smelter. None of the other studies found significant associations, but the effects of air pollution were difficult to evaluate in these studies due to the inability to '.L KESIUHNi \N'D R" ISO'7 | 0 to.h-: II).6-4 !) > 11.1-4 0 1 1.4-1 1 RR of l TABLE 4 no Cancer According id Proximity of 1950 Resiid-nce mom me Zinc Smelier and Usual Employ mi-, vi in Zinc Smelting Usual , ic No. of No. of employr. ent Proximity cases controls RR' (90'7 Cl) RR" (VlKf Cll No Far 242 234 1.0 -- 1.0 -- Middle )* 19 (.1 tO.6-2.0) 2.4 II. 1-4.9) Near S 2.4 (0.5-14.5) 4.2 10.8-20.6) Yes Ear 0 Middle 5 Near 9 0-- -- -- -- 5 1.0 (0.3-3.3i 2.0 10.6-7.4) 5 1.7 IO.h-5.2i 5.6 11.7-18.7) " All ri'-Ls relative to RR of 1.0 lor no usual zinc employment and far proximity. ''RR tn listed for smoking and usual employment in steel. 258 BROWN. POTTERN, AND Bl.OT IrK). 4. Proximity u> the -.teeI plant for grids of usual residence in a 20-km: area adjust for confoundcrs such as cigarette smoking and occupation, and . ten the absence of pollutant data. The present study offered several advantages. Information obtained fr n com plele smoking and occupational histories enabled adjustment for these ir portam risk factors: and the collection of lifetime residential histories formed ? o basis for the estimation of pollutant exposure. Usual or longest residence, a .veil as residence at any point in time, was ascertained. Decade intervals were sed for convenience in the analyses and enabled examination of residential patte s prior to the development of cancer. Soil pollutant data were available, ant melter slack emission levels for certain metals could be used in developing the u amceexposure index. For most of the study area historic air quality data we either unavailable or were of limited coverage. The decision to measure the tv .rentra- TABl.P ? RR or t.i sv. Canci.k AcvoRniN'.i m Proximity from rut. .Sti.t-.t. Plant for Ust ai. Ri \nd Ri.stnt Ni r IN 195(1 No. of No. of Residence Proximity CilMS controls RR' 190''.' cn RR' (9 Usual 1450 Far Middle Near Kur Middle Near 261 44 24 251 40 2k 252 to t.O o.y 10.6- i.xi 0.6 to 26 0.9 t().5 - 1.51 11.7 to 146 1.0 .... 1.0 44 o.x tO.'- 1.2) 0 5 (0 25 1.0 (11.6- 1.7) OS 1(1 ' All risks rekilive to u RR of I.ll for far proximity. '' RR adjusted for smoking and usual employment in steel or /me. Ci! 0.91 1.51 il.Vt 1.61 r LUNG CANCER IN RELATION TO POLLUTANTS 259 tion of heavy metals in soil was made with the assumption that the levels of pollutants reported in current samples of undisturbed soil may reflect an accu .(tv ro steel plant mulation of airborne metals over several decades. The locations of our soil sam ples were not scientifically selected, but instead were chosen to correspond to CHE *AR R areas where EPA had previously set up air sampling monitors. It was our intention to correlate soil and air sample data; however, due to the extremely short sam pling period used by EPA their data were unrepresentative of yearly air pollution OY S'JSjECTS UVfO ,E AREAS means and therefore were not useful for comparison purposes. The case-control study revealed a higher relative risk of lung cancer in men who had lived near the zinc smelter. The soil sample analyses showed large variations in the concentration of all six metal pollutants across the study area, N l with the case-control comparisons showing an increased risk associated with residence in areas where arsenic and cadmium were high. The increases associ ated with nearness to the smelter and with the pollutant indices w<ere not ex plained by the effects of cigarette smoking or by employment in the zinc or steel industry. Thus, although the study was considerably restricted by relatively small numbers of cases and controls in the high-exposure areas and by limited pollutant data, there was some indication of an increased risk associated with particulate a ZO-km- area. metal pollution. One of the major problems in a study of this nature is discriminating among dlion. and of n the the effects of so many potential exposures. In the environment near the smelter residents were exposed to heavy concentrations of complex mixtures of metals obtained froi com- (Nordberg and Andersen, 1981) and elevated levels of SO:. A primary suspect for these im; irtant for an elevated lung cancer risk is inorganic arsenic, a recognized human lung es formed th' basis carcinogen. Arsenic has been implicated in the increased risk of lung cancer zsidence, as ell as among copper smelter workers in the United States. Japan, and Sweden (Lee and ls were u .^d for Fraumeni. 1969; Lubin et al., 1981; Enterline and Marsh. 1980; Rencher et al.. ial palterr. prior 1977; Kuratsune et al.. 1974; Axelson et al.. 1978). Zinc ores typically contain tilable. and nelter smaller amounts of arsenic than copper ores, with average arsenic stack emissions loping the dis mce- from American smelters estimated at 4.9 lb per ton of copper compared to 1.3 lb ity data wen. either per ton of zinc (Sula, 1978). Nevertheless, the soil analyses showed high levels lsure the con :ntra- for this substance in residential areas nearer to the smelter. One cannot, however, rule out the possibility that the other metals found in conjunction with arsenic may be involved in the development of disease. Indeed, an increased risk of lung ! OR USLAL RliSi ANCE cancer has been reported in a study of cadmium smelter workers (Lemen et al.. 1976). This is intriguing since the soil samples taken near the smelter had ex tremely high concentrations of cadmium, a finding in agreement with an earlier Kg'1 190' Cl) | survey of nonferrous smelters that reported significantly elevated cadmium levels in the hair of children living near smelters (Baker et al.. 1977). Also, while ex 1.0 0.6 (0.-?- 0.9) 0.7 IO.h t.3) posure to SO=. a lung irritant in sensitive persons at concentrations of 1-2 ppm (Welch et al.. 1978), has not been shown to be an independent risk factor for lung cancer, it may act synergistically with arsenic (Lubin et al.. 1981) or other metals. 1.0 0.5 (0.7 09) 0.8 UM !.6I No excess risk was found among men living near the steel facility, although long-term employment in the steel industry was associated with significantly ele vated lung cancer risk in another report (Blot et al.. 1983). Other studies of steel workers have found substantially increased risks of lung cancer associated with 1 260 BKOWN. POTTER N, AND BLOT coke oven operations and have implicated benzo(a)pyrene and other po. nuclear aromatic hydrocarbons (PAH) (Tanimtira. 1968). Adequate environment;, data on PAH concentrations in the vicinity of the steel plant were unavailable, he pre. vailing winds, however, were such that the majority of the emissions om the plants were probably distributed to very sparsely populated areas. The oncentrations of the six heavy metal pollutants were generally low in areas ear the steel plant. Of some concern in this study was the use of an exposure model whic. did not include meteorological or topographical factors. However, due to the t istering of the study population we believe that the model used was adequate * access whether cases resided closer to the index.plants than did controls. Also of concern was the use of deceased controls instead of a randoi sample of live controls. While it is possible that the associations may be inflm iced by the diseases among the controls that led to their death, the controls wet chosen from a variety of causes of death so that an association with any specifi> disease should be minimal. In conclusion, our results raise the possibility of an increased lung ca ;er risk associated with environmental exposure to high concentrations of heav metals, including arsenic and cadmium. The small sample size and limited envin mental data preclude causal interpretation, bur suggest the need for further res arch in other polluted areas of the United States. ACKNOWLEDGMENTS The authors thank the stall of Lehigh University's Center for Social Research for data . llection. t>r. Patricia Bradl and Ms. Diane Mumlt for technical assistance, and Dr. B. J. Stone foi omputer support. REFERENCES Axelson. O.. Dahlgren. K.. Jans.xon. C.-D.. and Rchnkmd. S. (). (1978). Arsenic exposure nd martality: A case-referent study from a Swedish copper smelter. Brit. J. hid. Med. 35. 8 '5. Buchauer. M. J. 1)973). Contamination of soil and vegetation near a zinc smelter by zinc, dmium, copper, and lead. Environ. Sti. Teelmol. 7, 131-135. Baker, ti. L.. Hayes. C. G.. Landrigan. P. J.. Handke. .1. I... I.eger. R. T.. Housworth. ' J.. and Harrington. J. M. (1077). A nationwide survey of heavy metal absorption in children I ng near primary copper, lead, and zinc smelters. Amer. J. Epidemiol. 107. 261-273. Blot. W. J.. and Praumetii. J. K. .!r. (1975). Arsenical air pollution and lung cancer. < tivei 2, 142- 146. Blot. W. J.. Brown. I... M.. Pottern. L.. M.. Stone. B. J.. and Kraumeni. J. h. Jr. 11983). t.u cancer n/nong long-term steel workers. Amer. J. Epidemiol. JI7, 7(16-716. Connor. J. J.. and Shacklelte. H. T. (1975). Background geochemistry of some soils, plants nd veg etables in the conterminous United States. 17..5. Geol. Sure. Prof. Pap. 574. 164. Cordier. S.. Theriault, (j.. and Iturra. H. (.1983). Moilahiy patterns in a population iivir near a copper smelter. Environ. Res. 31. 311 -322. Lnterline. P. K.. and Marsh. Ci. M. I t98()l. Mortality studies of smeller workers. Amer. J. / Med. 1, 251-259. Heiss. J. L. (1981). "Statistical Methods for Rates and Proportions." Wiley. New York. Gan. J. J. (1970). Poini and interval estimation of the common odds mtio in the combinaikv if 2 * 2 tables with fixed marginals. Hiommrikti 57. 4" 1-47*. Greaves. W. W.. Rom. W. N.. Lyon. J. L.. Varies. G.. Wright. D. D.. and Chiu. G. (I9sii. j laticn- V u r er polynuclear ronmentai data on vailable. The pre. nissions from the eas. The concenin areas near the \ LUNG CANCER IN RELATION TO POLLUTANTS 261 ( 1 ship between lung cancer and distance of residence from nonferrous smelter stack effluent. Amer. [ J. ind. Med. 2, 15-23. Orewling. T. (1976). Chemical analysis of plant tissue. Search 6(8), 18-35. Cornell University Agri- i culture Experiment Station and New York State College of Agriculture and Life Sciences, Ithaca, j N.Y. 1 gurntsune, M., Tokudome, S., Shiraklsa, T., Yoshida, M., Tokumitsu, Y., Hayano, T.t and Seita, M. ( (1974). Occupational lung cancer among copper smelters. Int. J. Cancer 13, 552-558. del which did not to the clustering 1equate to access ols. a random sample be influenced by } ; fandrigan, P. J. (1981). Arsenic--State of the art. Amer. J. Ind. Med. 2, 5-14. j Lee. A. M.. and Fraumeni, J. F., Jr. (1969). Arsenic and respiratory cancer in man: An occupational ; study. J. Natl. Cancer Inst. 42, 1045-1052. Lemen, R. A., Lee, J. S., Wagoner, J. K., and Blejer, H. P. (1976). Cancer mortality among cadmium production workers. Ann. N. Y. Acad. Sci. 271, 273-279. Lubin, J. H., (1981). A computer program for the analysis of matched case-control stadies. Comput. i . Biomed. Res. 14, 138-143. ' | jjibin. J. H., Pottem, L. M., Blot, W. J., Tbkudome, S., Stone, B. J., and Fraumeni. J. F., Jr. (198J). rols were chosen / specific disease lung cancer risk of heavy metals. :d environmental rther research in [:. .. Respiratory cancer among copper smelter workers: Recent mortality statistics. J. Occup. Med. i 23,779-784. I Lyon, J. L., Fillmore, J. L., and Klauber, M. R. (1977). Arsenical air pollution and lung cancer. [ . ' Lancet 2, 869. fjMatanoski, G. M., Landau, E., Tonascia, J., Lazar. C., Elliott, E. A.,McEnroe, W., and King, K. j (1981). Cancer mortality in an industrial area of Baltimore. Environ.Res. 25, 8-28. j fifordberg, G. F., and Andersen, O. (1981). Metal interactions in carcinogenesis: Enhancement, in I hibition. Environ. Health Perspec. 40, 65-81. 'QfBce of Management and Budget (OMB), (1972). Executive Office of the President. "Standard Industrial Classification Manual." U.S. Govt. Printing Office. Washington, D.C. ' 'ftrshagen, G., Elinder, C.-G., Bolander, A.-M. (1977). Mortality in a region surrounding an arsenic li for data collection, Stone for computer - emitting plant. Environ. Health Perspec. 19, 133-137. i Reed. D., Lessard, R., and Maheaux, B. (1978). Lung cancer on a nickle smelting island. Amer. J. | . Epidemiol. 108, 233. f Sencher, A. C., Carter, M. W,, and McKee, D. W. (1977). A retrospective epidemiologic study of 5 - mortality at a large western copper smelter. J. Occup. Med. 19, 754-758. Rom, W. N.. Variey, G., Lyon, J. L.. and Shopkow, S. (1982). Lung cancer among residents living near the El Paso smelter. Brit. J. Ind. Med. 39, 269-272. : exposure and moried. 35, 8-15. r by zinc, cadmium, uswortb, W. J., and children living near : cancer. Lancet 2, ( .'Rose, A. W., Hawkes, H\E., and Webb, J. S. (1979). "Geochemistry in Mineral Exploration," Ac ademic Press, New York/London. Stinnett, S. S., Buffler, P. A., and Eifler, C. W. (1981). A case-control method for assessing enviI . ronmentai risks from multiple industrial point sources. Environ. Res. 25, 62-74. |- .Sota, B. E. (1978). "Human Exposures to Atmospheric Arsenic." CRESS Report No. 50, p. 6. Menlo 7 Park, Calif. (1983). Lung cancer . Ihnimura. H. (1968). Benzo(a)pyrene in an iron and steel works. Arch. Environ. Health 17, 172-177. ' Welch, K. B., Haring, M. K., Morris, L. E., and Higgins, 1. T. T. (1978). "Health Effects of Sulfur jiis, plants and veg- ., 164. Oxides," Vol. I, "Literature Review." Department of Epidemiology, School of Public Health, University of Michigan, Ann Arbor, Mich. lation living near a Amer. J. ind. Med. v York. ! :-V rombination of 2 x 3. (1981). Retalion-