Document 3QoNd90NLp6nb89xdzYO7x1Q6

28 Occupational and Environmental Medicine 1995;52:28-3|ft \*fs Cancers related to exposure to arsenic at a copper | smelter Philip E Enterline, Richard Day, Gary' M Marsh I 9D/9 Department of Biostatistics, Graduate School of Public Health, University of Pittsburgh, PA P E Enterline R Day G M Marsh Correspondence to: Dr Philip E Enteriine, Denartmenr of Rinstatistics. Abstract Objective--This is an update of an earlier study on the relation between exposure to arsenic in air and deaths from respira tory cancer. The purpose was to verify earlier findings of a supralinear dose response relation and to examine rela tions with other cancers, particularly those reported in studies on drinking water. Methods--An earlier study of 2802 men who worked at a copper smelter for a year or more during the period 1940-64 and who were followed up for deaths dur ing the period 1941-76 was updated until 1986. Estimates of exposure for the period 1977-1984 were added. Results and conclusions--The additional follow up confirms the earlier finding that at low doses the increments in death rates for respiratory cancer for a given increment in dose are greater than at high doses. The additional follow up also shows significant increases in cancer of the large intestine and bone, and SMRs >150 for cancer of the buccal cavity and pharynx, rectal cancer, and kidney can cer. There was a positive relation between exposure to arsenic in air and kidney and bone cancer, but none for the other cancers, except respiratory'. (Occup Environ Med 1995;52:28-32} Keywords: arsenic; cancer; dose-response We (PE, GM) previously reported on the mortality experience of 2802 men who were exposed to arsenic at a copper smelter at Tacoma, WA after a year or more of employ ment during the period 1940-64.13 These reports showed an excess in deaths from res piratory cancer during the period 1941-76 with the excess related to cumulative airborne exposure to arsenic. In the most recent report, the respiratory cancer dose-response curve at lower exposures was much steeper than at higher exposures so that the dose-response curve was supralinear (concave downward). This was true for the relative as well as the absolute response, and for follow up that started at entry into the study, or follow up that started at the end of exposure.3 Thus, in tional 10 years of follow up, until 1986. The v purpose is to verify our previous findings and ?! to explore an additional question, and that is | whether inhaled arsenic is related to internal f cancers as well as to respiratory cancer. | The copper smelter where workers in this study were exposed to arsenic started opera- j: tions in 1913 and at one time was the largest in the United States, producing about 10% of;;, the refined copper in the entire country'. For many years it was the source of all of the; arsenic commercially produced in the United |V States. Further information about this smelter [ is contained in an earlier paper.- : Methods For the entire cohort of 2802 men who worked for a year or more during the period v 1940-64 it was possible to determine vital; status at the end of 1986 for 98-5%. An addi tional 522 deaths were identified, which increased the total from 1061 in the last report to 1583. Of 1583 known deaths, death certificates were obtained for 1530 or 96-6%. The additional 522 deaths were traced through the United States Social Security Administration, the United States Nation^ Death Index, and by personal phone contacts. Death certificates were obtained from state Health Departments and coded by a nosolo- gist to the underlying cause of death accord ing to the revision of international classification of causes of death in effect at the ft time of death. For malignant neoplasms expected numbers of deaths were based upon | the mortality of the population of the State off! Washington.4 These are available for the|f entire follow up period 1941-86 for malignantfft neoplasms but only from 1960 for otherfi: causes of death. Thus, for this report followj| up for non-malignant disease only starts ing| 1960 whereas follow up for malignant disease starts in 1941. . W In the calculation of expected rioarhc onlyjll theff:mortality of white men was used as all 2802 workers in the study were men and nearly all were white. Expected deaths were" also calculated with death rates for Piei County, the county where the smelter is; located.4 Results were about the same as when Washington state rates were used. Because larger numbers of deaths and more stabli rates only Washington state rates are used 1 c A Hie and it is trial this ira-t in i of For the ited Iter vho riod last .`ath 6%. ,ced `rity mal icts. rate olo>rdmal the sms pon e of the rant ther ilow s in ease only the and hen ;e of able d in ality Cancers related to exposure to arsenic at a copper smelter 29 Expected deaths are age-rime specific death rates of white men for the state of Washington multiplied by person-years lived by the cohort in five year age-rime intervals. The SMRs were calculated with the computer program OCMAP.5 Exposure to arsenic was estimated from departmental measurements of arsenic in air contained in company annual reports that started in 1938, and measurements of urinary arsenic identified by department and worker started in 1948. We combined these data to allow' an analysis of the relation between the concentrations of arsenic in air and various cancers. Measurements of arsenic in air were mostly confined to departments in which arsenic was thought to be a problem. Measurements of urinary7 arsenic, on the other hand, were offered to all workers, and they represent all departments of the smelter. The conversion of data of urinary arsenic to air arsenic was made by the identification of departments and years for which data from both air and urinary arsenic were available and by the determina tion of the mathematical relation between the two. Although a total of 34 pairs of data were available, six pairs were deleted, one pair because the value of arsenic in air was based on only one sample and the other five pairs because notations in the annual reports indi cated that the measurements of arsenic in air were not representative of normal working conditions. The remaining 28 pairs of data represented 11 of the 33 departments at the smelter. Average concentrations of arsenic in air were weighted to reflect as closely as possible actual exposures of workers. Before 1971, data of arsenic in air came from spot samples and "tape samples. As far as possible these were weighted for hours per shift at the sam ple location, numbers of men at that location per shift, and frequency of operation (for sam ples that reflected exposure from a particular operation). Starting in 1971, data for arsenic in air were based on readings from personal samples. These seem to be representative of personal exposure for particular departments. From these data an exposure matrix of arsenic in air was developed by department and year from 1938 up to the time the smelter closed in 1984. Combined with job histories for each worker cumulative exposure was cal culated and expressed as /ig/mVy. For years before 1938 the exposure data for 1938 were used. Additional details on the method used in estimations of exposure are contained in an earlier report.' RESULTS FOR CANCER Table 1 shows the mortality experienced for selected malignant neoplasms in total and by time since first exposure. Overall, there were significant excesses for all malignant neo plasms taken together, cancer of the large intestine, cancer of the respiratory system, and bone cancer. When there were <20 years since first exposure the only significant excess is for respiratory cancer. For > 20 years since first exposure SMRs are generally higher but those that are significant are the same as for the total cohort. As well as the cancers that were signifi cantly in excess in table 1 there are three sites for which there were 10 or more deaths and for which the SMRs w7ere >150 but were not significant. For cancer of the buccal cavity7 and pharynx the SMR was 168-7 based on 12 deaths. For rectal cancer the SMR was 176-0 based on 15 deaths, and for kidney cancer there were 11 deaths and the SMR was 163-5. Ten of the 11 deaths from kidney cancer occurred >20 years after first exposure. Table 1 Observed (O) and expected (E) deaths and SMRs for selected cancers by years since first exposure, Tacoma copper smelter Cancer site Total O E Exposure <20y SMr(ca)o SMR Exposure ^20 y OE All malignant neoplasms: Buccal cavity and pharynx Digestive organs and peritoneum: Oesophagus Stomach Large intestine Rectum Biliary passages and liver Pancreas All other digestive Respiratory system: Larynx Bronchus, trachea, lung All other respiratory Breast Prostate Testes and other male genital organs Kidney Bladder and other urinary organs Malignant melanoma of skin Eye Central nervous system Thyroid and other endocrine glands Bone AH lymphatic, haematopoietic: Lymphosarcoma and reticulosaroma Hodgkin's disease Leukemia and aleukemia All fif-twar limnnhonAipliri 395 276-06 143-1** 43 12 712 168-7 2 95 7908 120-1 11 7 608 115-2 0 18 16-43 109-6 2 38 23-48 161-8** 2 15 8-52 176-0 4 1 4-72 21-2 0 14 16-26 86-1 2 2 3-59 55-7 I 188 89-65 209-7** 17 3 3-17 94-7 0 182 85 03 214-1** 17 3 1-46 205-4 0 0 0-46 -- 0 28 25-62 109-3 2 1 1-31 76-6 ^1 11 6-73 163-5(-sT`i) 1 8 9-61 83-2 > i 1 2-56 39-1 0 1 0-28 361-6 0 5 8-15 61-4 0 1 1-00 99-9 0 5 110 455-6* 0 24 27-21 88-2 7 3 5-36 55-9 0 2 2-91 68-8 2 9 11-28 79-8 2 10 7-66 130-5 3 36-10 114 10-94 0-64 2-79 2-61 1 -33 0-58 2-19 0-80 9-66 0-42 9-04 0-20 0-10 1 -23 0-70 104 0-94 0-55 0-04 1-98 0-24 0-28 5-16 1 -21 1-20 2-04 0-70 * 00 00 119-1 175-8 100-6 -- 71-8 76-8 299-9 -- 91-5 125-0 176-0* -- -- -- 163-2 143-2 95-9 106-7 -- -- -- --" -- 135-5 -- 166*4 97-8 426-8 352 10 84 7 16 36 11 1 12 1 171 3 165 3 0 26 0 10 7 1 1 5 1 5 17 3 0 7 7 239-95 5-98 68 14 5-44 13-64 20-87 7-19 4-13 14-07 2-79 79-99 2-75 75-99 1-26 0-36 24-39 0-61 5-69 8-67 2-00 0-23 6-16 0-76 0-82 22-05 4-15 1-70 9-24 6-96 SMR 146-7** 167 3 123-3 128-8 117-3 172-5** 153-0 24-2 85-3 35-8 213-8** 109-2 217-1*' 238-1 106-6 175-9 (0 80-7 49-9 428-1 81-i 131-7 610-2** 77-1 72-3 75-8 100-6 32 populations. We looked for skin cancers in the smelter population of Tacoma and found no excess.-1 These are rarely fatal and could only be found at work, not very long after the onset of exposure. Skin cancer related to arsenic in drinking water does not usually appear until after the age of 30 and the incidence rises rapidly with age.- Thus, the latent period for skin cancer and perhaps other cancers related to ingested arsenic could be very long, so per haps this is the reason these were not seen in the smelter worker population of Tacoma. It may also be that, as noted by Bates et al, in comparison with populations exposed to arsenic in drinking water the exposure of our population, and perhaps other populations exposed to arsenic in air is relatively low.1" One cancer with an apparendy short latent period in this study is respiratory cancer. As table 1 shows, a significant excess (SMR = 203-3) occurred in <20 years. The SMR remained around 200 in successive 10 year periods and was 151-9 >50 years since first exposure. We are not aware of published studies of other workers from copper smelters that report increments in respiratory' or lung cancer by time from first exposure. Ott et al reported no deaths from respiratory cancer in a population occupationally exposed to insecti cides containing arsenic <15 years from first exposure (0-4 expected), but did report excesses for the subsequent time since first exposure categories.11 In a study by us (PE, rGM) of workers at a copper smelter at Salt Lake City, UT where there was exposure to arsenic the SMR for respiratory- cancer <20 years since first exposure was 170-0 (11 deaths) and >20 years 107-8 (39 deaths).1--3 Perhaps a short latent period is somehow related to the way exposure occurs in the lung as inhaled arsenic seems to be retained in the lung for long periods.34 On the other hand, perhaps the lung has a particular susceptibil ity. One curious finding in studies of arsenic in drinking water is a dose related excess in mortality from lung cancer--an excess sur passed only by an excess in urinary cancer.1'1 We are not aware of any other ingested sub stance that may cause lung cancer. The dose-response relation between air borne arsenic and respiratory cancer found in this study is unusual. As noted in our earlier report, this could be because air measure ments are not good measures of biological dose.3 The relation is not likely to be due to confounding, but this needs further study. Smoking can be an important confounder in respiratory cancer and data on the histories of the study population for smoking are being collected. These will be modelled along with other covariates in a future publication. The excess we found in bone cancer has not, to our knowledge, been reported in other studies. As arsenic is stored in the bone, how ever, this may be an important finding.25 u Of the five deaths coded by our nosologist to ----- i:ui~ ..... information on the death certificate indicated! that these may not have been the primaiy| sites. If we had recoded these, however, itj would have invalidated our comparison with! the external control population. j n This research was supported in part by the US EnvironmentalJ Protection Agency. (Agreement No CR811173.) Jj 1 Enterline PE, Marsh GM. Mortality study of copper gfe smelter workers. AmJInd Med 1980;1:251-9. If;. 2 Emerlinc PE, .Marsh GM. Cancer among workers exposed fe: to arsenic and other substances in a copper smelter. .-bnjfcfe Epidemiol 1982;116:895-91.1. f 3 Enterline PE, Hendrgson vL, Marsh GM. Exposure to gj?: arsenic and respiratory cancer--a reanalysis. Am Jfe Epidemiol 1987,125:929-38. -. g) 4 Marsh GM, Ehland J, Sefcik S. Mortality and population Jfe data system (MPDSJ. Pittsburgh: University otj|; Pittsburgh, Department of Biostatisrics, Technical jO- Report, 1987. 5 Marsh GM, Preininger ME. A user-oriemed occupational 'I;; - cohort mortality analysis program. American Statistician Xy 1980;34:245. 5, 6 Lee AM, Fraumeni JF Jr. Arsenic and respiratory cancer in man: an occupational study. J Natl Cancer Inst 1969; fe: 42:1045-52. % 7 Rencher AC, Carter MW, McKee DW. A retrospective ?; epidemiological study of mortality at a large western copper smelter. J Occitp .Med 1977;19:754-8. y 8 Tokudome S, Kuratsunc M. A cohort study on mortality fe. . from cancer and other causes among workers at a metal refinery. IntJ Cancer 1976;17:310-7. fe 9 Sandstrom AIM, Wall SGI, Taube A. Cancer incidence |fe and mortality among Swedish smelter workers. BrJ lm 0 Med 1989;46:82-9. pi. 10 Mabuchi K, Lilienfeld AM, Snell LM. Lung cancer among $ pesticide workers exposed to inorganic arscnicals. Ar:k ft Environ Health 1979;23:312-20. gfe 11 Ott MG, Holder BB, Gordon HI. Respiratory cancer and occupational exposure to arscnicals. Arch Environ HealtkW 1974;29:250-5. ?-; 12 Enterline PE, Marsh GM, Esmen NA, Henderson VL,|fi Callahan CM, Paik M. Some effects of cigarette smok- |J ing, arsenic, and SO_- on mortality among US copper fe smelter workers. J Occup Med 1987;29:831-8. p 13 Lee-Fcldstein A. Cumulative exposure to arsenic and itsg relationship to respiratory cancer among copper smeltct f employees. J Occup Med 1986;28:286-302. fe; 14 Higgins I, Welch K, Oh MS, Kryston KL, Burchfie! CM, fe Wilkinson NM. Arsenic exposure and respiratmy cancer, y Ann Arbor, MI: Department of Epidemiology, h University-of Michigan, 1986. 15 Jarup L, Pershagcn Cr, Wall S. Cumulative arsenic expo-!i sure and lung cancer in smelter workers: a dose-response E" study. Am3 bid Med 1989;15:31-41. f? 16 Hertz-Picciotto I, Smith AH. Observation on the dose-|fe response curve for arsenic exposure and lung cancer. i: ScandJ Work Environ Health 1993;19:217-26. fe 17 Jarup L, Pershagen G. Arsenic exposure, smoking and lung cancer in smelter workers--a case-control study. Am J Epidemiol 1991;134:545-51. ft; 18 Chen CJ, Chiang YC, Lin TM, Wen HY. Malignant neo-; plasms among residents of a Blackfoot disease-endemic area in Taiwan. Cancer Res 1985;45:5895-9. 19 Bates MN, Smith AH, Hopenhayn RC, Arsenic ingestion and internal cancers: a review. Ant J Epidemiol 1992J 135:462-76. 20 Gibb H, Chen C. Is inhaled arsenic carcinogenic for sites other than the lung? In: Mohr U, Bates DV, Dungwortbj DL, Lee PN, McClellan RO, Roe FJC, eds. Assessment inhalation hazards--integration and extrapolation t. diverse data. Berlin: Springer-Veriag, 1989. 21 Andelman JB, Barnett M. Feasibility study to resolve q. tions on the relationship of arsenic in drinking water skin cancer. Pittsburgh: Center for Environmem Epidemiology, University of Pittsburgh, I (Technical Report 84-8.) 22 US Environmental Protection Agency. Special report ingested inorganic arsenic. Washington: US EPA, 191 (EPA-625/3-87/013.) 23 Enterline PE, Marsh GM, Esmen Nr, Henderson V, . E. Mortality among copper and sine smelter workers in United Stales. Final report to the Smelter Envirvnmi Research Association. Pittsburgh, PA: University Pittsburg, 1986. 24 Gerhardson L, Brune D, Nordberg GF, Wester _ Multi-elemental assay of tissues of deceased smell workers and controls. Sci Total Environ 1988;74:97-11) 25 Kadowski K. Studies on the arsenic contents in oi * tissues of the normal Japanese. Osaka City Med 7 1 9:2083-8. 26 Saady JJ, Blanke RV, Poklis A. Estimation of the