Document vBE643YEL163rm5KvEOkqaaEZ

I Mortality in a Cohort of Orchard Workers Exposed to Lead Arsenate Pesticide Spray KRISTINE TOLLESTRUP Institute for Health and Population Research The Lovelace Institutes Albuquerque, New Mexico JANET R. DALING Fred Hutchinson Cancer Research Center Seattle, Washington JACK ALLARD Washington State Department of Health Seattle, Washington ABSTRACT. During the period from 1890 to 1940, lead arsenate was the major pesticide used in apple orchards to control the coddling moth. In the Wenatchee area of Washington State, lead arsenate spray was used for longer periods and in larger quantities than in other areas of the United States. In 1938, a cohort of 1 231 people who lived in this area was selected for a study to determine the effects of exposure to lead arsenate spray and residue. This same cohort was re-examined to determine whether there was excess mortality that could be attributed to the lead arsenate exposure. Three levels of exposure (i.e., orchardist, intermediate, consumer) were defined, based upon the use of lead arsenate pesticide spray before and during the 1938 apple growing season. Age-adjusted hazard ratios for all caus es of mortality were elevated for both male orchardists and male intermediates. The only significantly increased age-adjusted hazard ratio (1.94) was heart disease in male interme diates. No significantly elevated age-adjusted hazard ratios were observed for women in any exposure group. The lack of evidence that supported an increase in mortality from respira tory cancer in this cohort may have resulted from the lower cumulative concentration of arsenic exposure, the type of arsenical compound, and the small number of study subjects. BOTH LEAD AND ARSENIC are known toxic sub stances. In spite of their toxicity, occupational exposure to both these compounds has been common in agricul tural and smelting occupations. Many epidemiologic studies of workers exposed to lead and arsenic have documented serious health problems and an excess risk of cancer and mortality.1-16 Exposure to lead in various settings, including lead smelters, storage battery facto ries, and residential areas adjacent to lead smelters, has been associated with elevated blood lead levels, clini cal symptoms of lead poisoning, and impairment of central nervous system functioning.1-3 Elevated bloodlead levels have also been associated with increases in blood pressure, especially in men4; exposure to lead may also increase the risk of death from cerebrovascu May/June 1995 [Vol. 50 (No. 3)] lar disease.5 Several epidemiologic studies have sug gested a higher incidence of digestive and respiratory cancers in lead smelter and battery plant workers6-7 Numerous other studies of smelter and factory work ers have shown that occupational exposure to arsenical compounds is associated with excess cancers, particu larly respiratory cancers.8-17 A clear dose-response rela tionship between arsenic levels and respiratory cancers has been observed in smelter workers.8-14 Respiratory cancer rates are also higher in employees who are involved in the manufacture of arsenic-containing pes ticides.15-17 Several studies have also reported an asso ciation between arsenic exposure and nonrespiratory cancers. Studies of arsenic ingestion via drinking water in Taiwan and Japan suggest that arsenic is associated 221 with cancers of the bladder, kidney, lung, and liver.18-20 Elevated relative risks for non respiratory cancers (e.g., cancers of the urinary tract, gastrointestinal tract, and liver) have also been observed in smelter workers in the United States, Japan, and Sweden, as well as in Japan ese refinery and mine workers.,3 2-22 Arsenic is often used in (a) insecticides to control the cotton bollworm and tobacco pests; (b) organic herbi cides, such as MSMA and DMSA, to control weeds in cotton fields; and (c) desiccants, such as arsenic acid. During the period from 1890 to 1940, lead arsenate was the major pesticide used in apple orchards to con trol the coddling moth. Other compounds have since replaced lead arsenate, and it is no longer used widely. However, as a consequence of lead arsenate use, orchardists have been exposed to a potentially harmful substance over a long period of time. Results of studies that have examined exposure to arsenic in agricultural settings have varied.23-28 An asso ciation between exposure to arsenates in vineyards and respiratory cancer has been suggested in two stud ies.2324 Barthel25 reported a twofold excess of lung can cer in male German agricultural workers, but he was unable to show an association with any specific pesti cide. A study of orchardists in Washington State was also inconclusive as to whether exposure to lead arsen ate increased the risk of lung cancer.26 The long-term health effects of lead arsenate spray on individuals who lived in an apple-growing region in Washington State were examined in two additional studies.27 28 In 1938, a cohort of 1 231 individuals who lived in the Wenatchee area of Washington State was selected for a study that was designed to determine effects of inhalation or ingestion of lead arsenate spray and residue.27 This group was known as the Neal Study Cohort. The Wenatchee area was originally chosen because lead arsenate spray had been used there for a longer period of time (i.e., more than 25 y) and in larg er quantities than in other areas of the United States. People who lived in this area were also isolated from industrial sources of both lead and arsenic. This early study noted no significant increase in clinical symptoms of lead arsenate intoxication in individuals exposed to the pesticide during the 2 y of the study. A follow-up study in 1968 re-examined the Neal Study Cohort to determine whether there was excess mortality that could be attributed to lead arsenate exposure.28 There was no overall excess in mortality and no excess deaths from cancer, heart disease, or stroke. In 1968, however, only 37% of the cohort had died, and many of the study members had not reached high-risk mortality ages. Suf ficient time may not have elapsed since the exposure to lead arsenate for mortality differences to appear. The current study examined mortality risks in the Neal Study Cohort 22 y after the first follow-up mortal ity study and 52 y after the original morbidity study. Mortality effects, if they existed, should have been more evident than in the earlier follow-up study. The goals of this second follow-up study were to determine (a) whether there was an overall increased risk of dying as a result of exposure to lead arsenate and (b) whether 222 there was an increased risk of dying from lung and digestive cancers, which have been associated with exposure to arsenic. Method The study cohort included 1 225 individuals who had lived in the Wenatchee area of Washington State during the 1938 apple growing season and who had partici pated in the 1938 Neal study and the 1968 Nelson study.27 28 The cohort excluded 4 individuals who had died in 1938, prior to the end of the original recruit ment period for the Neal study, and 2 individuals for whom no names were available in the 1968 Nelson fol low-up study. The original assignment to an exposure group used in 1938 was retained. Three levels of exposure were defined, based upon use of lead arsenate pesticide spray before and during the apple-growing season in 1938. Consumers included individuals whose occupa tions did not bring them into contact with lead arsenate spray (e.g., school teachers, store clerks, housewives). Orchardists were individuals who prepared and applied lead arsenate spray during the 1938 growing season. Intermediates included individuals who had not used lead arsenate spray during 1938 or who had infrequent exposure to lead arsenate spray. This group included retired orchardists who had long histories of using the spray prior to 1938, as well as warehouse workers who had experienced infrequent exposures. In 1938, blood lead and urinary arsenic measure ments were taken for most of the participants. Howev er, by 1968, the individual measurements had been lost in a warehouse flood. Mean values for each exposure group were published by Neal et al.27 (Table 1). Arsenic and lead levels of consumers were lower than those for orchardists and intermediates. Levels in men were con sistently higher than those identified in women. Mean urinary arsenic levels for the consumers were similar to those published for other control groups studied in the 1940s and 1950s.8'9'29 Mean blood-lead levels for the consumers were higher than mean levels reported for the rural U.S. population in 1965 (i.e., men: 16 pg/dl; women: 10 pg/dl).30 Table 1.--Average Concentrations of Blood Lead and Urinary Arsenic in Selected Members of the Neal Study Cohort in 193827 Blood lead (pg/d) n X SD Urinary arsenic (|jg//.) nX SD Adult males Consumers Orchardists Intermediates Adult females Consumers Orchardists Intermediates 148 26.3 11.1 329 43.9 11.0 108 29.5 13.2 124 25.8 9.5 58 34.4 13.2 27 21.9 9.6 140 62.0 88.0 305 140.5 176.4 123 71.0 103.5 121 56.3 69.4 58 97.9 118.1 25 58.0 69.3 Archives of Environmental Health Table 2.--Arsenic and Lead Concentrations in the Orchard Air in 1938 during Various Orchard Activities19 Orchard activity Length of exposure h/d wk/y Mixing insecticide Burning containers Spraying orchard Thinning fruit Picking fruit Dumping fruit October December Sorting and packing October 1 NA 8-14 8-14 8-14 8-12 NA NA 8-12 *NA = data were not available. 8-12 NA 8-12 2-3 6-10 8-12 NA NA 4-12 Arsenic concentration X Range 18.5 166.7 1.4 0.8 8.8 0.6 0.1 0.06 0.2-110.7 48.6-261.2 0.4--4.8 0.1-3.2 2.6-19.0 0.1-1.9 0.02-0.2 0.03-0.08 Lead concentration X Range 57.4 35.8 4.5 3.0 29.3 0.9-467.3 10.2-76.5 1.3-14.3 0.4-1 7.0 7.7-75.2 1.9 0.4-6.9 0.3 0.01-1.1 0.16 0.07-0.22 In the original 1938 study, the ambient air levels of arsenic and lead were monitored during various orchard activities (Table 2). The usual hours per day and weeks per year spent in these activities were also cal culated for the orchardists. Although the original list of Neal Study members and their assignment to exposure group were available, extensive information was obtained from the 1968 fol low-up records. The 1968 records included name in 1938 and name in 1968; age at time of exposure in 1938; age in 1968; and birthdate, exposure group, and vital status in 1968. If the study member had been alive in 1968, an address at which the member resided in 1968 had been recorded. Copies of the death certifi cates were also available for most study members who had died during or prior to 1968. In the 1968 study, a questionnaire was used to gather information on dura tion of use of lead arsenate and other pesticides during the time period from 1938 to 1968. Unfortunately, this information had been lost for some of the study mem bers during the ensuing years. A list of the 777 study members who were not known to be deceased in 1968 was used to search the 1969-1990 death records of Washington State and to locate any deaths. The search included the 54 people who had been lost to follow up in 1968. Copies of death certificates were then obtained for individuals who had died in Washington State. The vital status of 507 individuals was still unknown after searching the death records. Letters that explained the purpose of the study and a questionnaire were mailed to these study members at their 1968 address or, if a current address could be found in the local tele phone directory, it was used. The questionnaire includ ed information on current name, current address, age, birthdate, and whether the study member was still alive. If deceased, a relative or close friend was asked to pro vide date and place of death. A total of 380 individuals were found to be alive. Death certificates were obtained for those study members who were identified from the questionnaire as having died. Death certificates were not obtained for 35 May/June 1995 [Vol. 50 (No. 3)] of the deaths, of which 5 resulted from World War II injuries; however, information provided on the ques tionnaires for the remaining 30 cases was not sufficient ly specific to allow the death certificates to be located. Twenty-one of the remaining 30 cases were reported to have died outside of Washington State. These names were matched to the National Death Index so that death certificates could be located. Information on cause of death, date of death, and birthdate was obtained from the death certificates. Regardless of the date of death, the cause of death was coded for all decedents, using the International Classification of Diseases (9th revision). Despite the fact that the ICD cause-of-death code was written on most death certificates, the cause of death for deaths that occurred between 1939 and 1975 was recoded, using the ICD-9 conventions. The Neal Study Cohort originally included 128 chil dren who were less than 18 y of age. The majority (81.3%) were classified as consumers. Given that the long-term effects of lead and arsenic exposure in chil dren may differ from those in adults, the children were removed from the analysis. Therefore, 1 097 study members who were 18 y of age or older were included in the survival analysis. Survival time was calculated in months; January 1, 1939, was the beginning of the study period, and December 31, 1984, served as the endpoint. Survival time for those who were lost to follow up in 1938 (i.e., not located after the original study period) was set at 1 mo. Survival time for all others was calculated as time (mo) (a) to death, (b) to date of loss to follow up, or (c) to the end of the study. Survival endpoints for the differ ent analyses included death from all causes; from all cancers (ICD-9 codes 140-208, 230-239); from diges tive cancers (ICD-9 codes 150-159); from stomach can cer (ICD-9 code 151); from pancreatic cancer (ICD-9 code 157); from respiratory cancers (ICD-9 codes 160-165); from lung cancer (ICD-9 code 162); from breast cancer (ICD-9 code 174); from lymphatic and hematopoietic cancers (ICD-9 codes 200-208); from coronary heart disease (ICD-9 codes 410-414); from other heart disease (ICD-9 390-398, 401-405, 223 415-417, 420-429); and death from stroke (ICD-9 codes 430-438). Crude mortality rates for cancers associated with arsenic exposure and other major causes of death were calculated for each sex and each exposure group. Per son-months of follow up or person-months to death were used as the denominator. Survival experiences of each of the two exposed groups (intermediates and orchardists) were compared with the unexposed group (consumers), using a Cox proportional hazards model that estimates the effects of explanatory variables (i.e., sex, age at exposure, expo sure category) on mortality rates. This model is appro priate for cohort survival analysis because it allows the use of censored observations (those individuals lost to follow up and not failing), provides for varying lengths of observations, and produces estimates of the hazard (mortality) rate ratio (exposed versus unexposed mortal ity rates). The consumers served as the internal baseline group. Results The follow-up status of the entire study cohort (1 225 members) is summarized and compared with the 1968 status in Table 3. At the conclusion of the study period, the vital status of 1 123 members (91.6%) was known: 380 were alive and 743 were deceased. There were four ways that the remaining study members were lost to fol low up: (1) 25 individuals were last contacted in 1938 and were not located in either of the two follow-up studies, (2) 24 individuals were last known to be alive at some point in time between 1938 and 1968, (3) 1 individual was last known to be alive between 1969 and 1984, and (4) 52 individuals were contacted in 1968 but were not located in 1984. The characteristics of the 102 study members lost to follow up are com pared with those of known vital status in Table 4. Study members lost to follow up were more likely to be female (in the consumer group) and were 18-34 y of age in 1938. The age and sex distributions of each of the exposure groups are shown in Table 5. Most of the children (81.8%) were consumers. Once the children were excluded, a lower percentage of orchardists and inter mediates were in the youngest adult age group (18-34 y), compared with consumers. The overall age distribu tions of the two sexes were similar. However, there were proportionately more male intermediates (37.4%) and orchardists (20.3%), compared with consumers (17.2%) in the oldest age group (55 y and older); and there were proportionately more female consumers (36.8%), compared with intermediates (26.9%) and orchardists (27.9%) who were 18-34 y of age. A large percentage of orchardists were men (88.8%), whereas more consumers were women (56.8%). There were approximately equal numbers of male and female inter mediates (males: 50.8%; females: 49.2%). Crude mortality rates for the causes of death of inter est were calculated per 100 000 mo of follow up (Tables 6 and 7). A total of 728 study members had died, of whom 84 men and 39 women died of some type of cancer. There was only 1 female lung cancer death that occurred in a consumer, and no other respi ratory cancer deaths occurred in women. Two women orchardists and 1 woman intermediate died of pancre atic cancer. The age-adjusted hazard ratios for deaths from select ed causes are also given in Tables 6 and 7. Risk of dying during the follow-up period was increased significantly for both male orchardists and intermediates (hazard ratios of 1.36 and 2.00, respectively). The only signifi cantly increased specific cause of death was coronary heart disease in the male intermediates (hazard ratio of 1.94). Overall mortality was not elevated for women. Discussion There is much experimental and epidemiologic evi dence that suggests that exposure to lead and arsenic has adverse effects on health and mortality.1'17 In this study of Wenatchee orchardists we found an increased risk of dying of all causes in male orchardists and inter mediates. Almost a twofold increased risk of dying of coronary heart disease was observed in men in the intermediate group, but this risk was not observed in the orchardist group. This is consistent with other studies that have shown a dose-response relationship between Table 3.--Follow-up Status of Neal Study Cohort in 1968 and in 1990 Follow-up status All ages 1968 1990 n% n% Alive Deceased Lost to follow up Not found after 1938 Last alive 1938-1967 Last alive 1969-1990 Alive 1968, not found in 1990 723 448 28 26 Total 1 225 59.0 36.6 2.3 2.1 380 743 25 24 1 52 1 225 31.0 60.6 2.0 2.0 0.1 4.2 1990 Adults n% Children n% 277 728 25 21 1 45 1 097 25.2 66.4 2.3 1.9 0.1 4.1 103 80.4 15 11.7 0 0.0 3 ?3 0 0.0 7 5.5 128 224 Archives of Environmental Health arsenic exposure and death from cardiovascular disease in smelter workers,11'13'31 as well with studies in which increases in cardiovascular mortality have been corre lated with exposure to lead.5 In smelter workers ex posed to arsenic at the Ronnskar smelter in Sweden, there was a twofold risk increase for death from cardio vascular disease.31 However, a more recent study of this cohort did not support the findings of the earlier study.15 Lead exposure may have also contributed to the excess in cardiovascular deaths. There is some evidence to suggest that elevated blood-lead levels are associated with hypertension and heart disease. Analyses of the NHANES II data have demonstrated an increase in sys tolic pressure in men exposed to lead.4'32 33 However, after important confounding factors have been adjusted for, other large-population studies have not demonstrat ed such an association.34-35 It is surprising that there was not an increase in mor tality from respiratory cancer in this cohort, given the amount of evidence that supports an association be tween arsenic exposure and lung cancer. Previous stud ies have documented increased standardized mortality ratios (SMRs) from lung cancer, ranging from 1.68 to Table 4.--Characteristics of the Study Members Lost to Follow Up Age group (y) 0-17 18-34 35-54 >55 Sex Male Female Exposure group Consumer Orchardists Intermediates Lost to follow up n% ii 10.8 43 42.2 40 39.2 8 7.8 43 42.2 59 57.8 41 40.2 28 27.4 33 32.4 Known vital status n% 117 10.4 295 26.3 456 40.6 255 22.7 757 67.4 366 32.6 292 26.0 535 47.6 296 26.4 3.43, in copper smelter, insecticide, and agricultural workers exposed to arsenic.6"17 The results of this fol low-up study, however, are consistent with the results of two other studies, as well with the earlier follow-up study in 1968. Barthel examined a cohort of male Ger man agricultural workers who were exposed to pesti cides between 1948 and 1972.25 Prior to 1960, arseni cal dusts were the primary pesticide used by these agricultural workers. The SMR for lung cancer for men exposed to arsenicals did not differ appreciably from the SMR for men who were not exposed to arsenicals. Wicklund et al. completed a case-control study of white male orchardists who died from respiratory cancer in Washington State during the time period 1968-1980.26 The presence, intensity, and duration of lead arsenate exposure did not differ among case and control sub jects. Information on smoking habits was also available and used in the analysis. Three factors might explain the absence of increased respiratory cancers in the Wenatchee orchard workers: (1) amount of exposure, (2) type of arsenical compound used, and (3) the low number of cancer deaths in the cohort. During certain time periods, orchardists and intermediates were exposed to high concentrations (i.e., up to 26 120 pg/m3) of lead arsenate as spray and residue. Nonetheless, the total amount of lead and arsenic that they encountered each year may have been less than that experienced by constantly employed industrial workers in smelters and insecticide factories. The concentrations of arsenic in the air during orchard operations were generally lower than those encoun tered in a pesticide factory. Ott et al. measured air arsenic levels that ranged from 180 to 40 800 pg/m3 in a pesticide factory.36 Body arsenic levels of the Neal Co hort intermediates (mean urinary arsenic of 71.0 pg/l) and orchardists (mean urinary arsenic of 140.5 pg/l) were also much lower than those of smelter workers (mean urinary arsenics ranging from 174 to 820 pg/l) in whom excess cancer neoplasms have been noted.9-29-37 Body arsenic levels in female orchardists and interme diates were even lower than those of men. Studies of Swedish copper smelter workers have suggested that, with respect to increased cancer risk, the intensity of the Table 5.--Age and Sex Distribution of Each Exposure Croup in the Neal Study Cohort Consumer n% Orchardist n% Croup Intermediate n% Age group 0-17 18-34 35-54 > 55 All ages Sex Males Females Males and females 104 31.2 84 24.9 95 28.5 50 15.0 333 143 42.9 190 57.1 333 20 3.6 168 29.8 266 47.2 109 19.4 563 500 88.8 63 11.2 563 4 1.2 86 26.1 135 41.0 104 31.6 329 167 50.8 162 49.2 329 All n% 128 338 496 263 1 225 810 415 1 225 10.4 27.6 40.5 21.5 66.1 33.9 May/June 1995 (Vol. 50 (No. 3)] 225 Table 6.--Crude Mortality Rates per 100 000 Person-Months of Follow Up and Cox Proportional Hazards Survival Analysis for Male Orchardists and Intermediates Who Were 18 y and Older in 1938, Compared with Male Consumers Cause of death Consumer Crude Crude mortality mortality rate rate Orchardist Hazard ratio 95% Cl All causes All cancers (140-208, 230-239) Digestive cancers (150-159) Pancreatic cancer (157) Respiratory cancers (160-165) Lung cancer (162) Lymphatic and hemato. cancers (200-208) Coronary heart diseases (410-414) Other heart diseases (390-398, 401-405, 415-417, 420-429) Stroke (430-438) Person-months of follow up No. deaths 143.78 26.96 5.99 3.00 11.98 11.98 3.00 50.92 14.98 14.98 33 385 48 192.28 30.91 13.16 4.01 8.01 6.30 2.28 64.67 1.36 1.20 2.23 1.42 0.69 0.59 0.78 1.23 1.01-1.85 0.59-2.44 0.52-9.52 0.17-11.60 0.23-2.16 0.19-1.85 0.08-7.00 0.73-2.06 17.74 1.40 0.54-3.62 24.61 1.65 0.65-4.20 174 744 336 Intermediate Crude mortality rate Hazard ratio 95% Cl 282.23 43.58 16.60 2.08 12.45 12.45 4.15 89.23 2.00 1.42-2.81 1.74 0.78-3.85 2.61 0.55-12.5 1.01 0.06-18.53 1.47 0.41-5.31 1.35 0.37-4.90 0.79 0.06-9.76 1.94 1.08-3.48 35.28 24.90 1.87 0.66-5.34 2.18 0.72-6.57 48 122 136 Table 7.--Crude Mortality Rates per 100 000 Person-Months of Follow Up and Cox Proportional Hazards Survival Analysis for Female Orchardists and Intermediates Who Were 18 y and Older in 1938, Compared with Female Consumers Cause of death Consumer Crude Crude mortality mortality rate rate Orchardist Hazard ratio 95% Cl All causes All cancers (140-208, 230-239) Digestive cancers (150-159) Pancreatic cancer (157) Respiratory cancers (160-165) Lung cancer (162) Breast cancer (157) Lymphatic and hemato. cancers (200-208) Coronary heart diseases (410-414) Other heart diseases (390-398, 401-405, 415--417, 420-429) Stroke (430-438) Person-months of follow up No. deaths 154.65 23.79 3.97 0.00 1.98 1.98 5.95 7.93 31.72 23.79 31.72 50 435 78 152.67 25.44 12.72 8.40 0.00 0.00 4.24 4.24 25.44 16.96 29.68 23 581 36 1.07 1.31 3.09 -- -- -- 0.75 0.55 0.80 0.71-1.61 0.48-3.61 0.52-18.79 -- -- 0.08-7.25 0.06-4.97 0.31-2.11 0.93 0.29-3.00 0.95 0.38-2.36 Intermediate Crude mortality rate Hazard ratio 95% Cl 164.79 36.81 12.27 1.75 0.00 0.00 14.02 0.00 31.56 1.03 0.76-1.39 1.63 0.79-3.34 2.92 0.61-14.16 ---- 2.38 0.62-9.00 0.98 0.49-1.94 15.78 73.96 57 043 94 0.63 1.02 0.26-1.54 % 0.52-1.98 226 Archives of Environmental Health arsenic exposure may be more important than its dura tion.14 The orchard workers in our cohort may not have been exposed to sufficiently intense levels of arsenic to produce an increase in lung cancer. The toxicity of^rsenic varies according to its chemi cal form.38 Lead arsenate is a pentavalent form, and arsenic trioxide is a trivalent form. The LD50 for lead arsenate is 100 mg/kg, whereas the LDS0 for arsenic tri oxide ranges from 15 to 110 mg/kg.39 Both the pentava lent and trivalent forms are well absorbed via inhala tion. In the body, arsenite is partially oxidized to arsenate, and arsenate is partially reduced to arsenite. The blood contains a mixture of As(+3) and As(+5). Smelter and sheep dip workers in whom excess lung cancers were noted were exposed to arsenic trioxide. Despite the fact that 66.3% of the cohort of adults had died, there may have been too few cancer deaths to detect any association between lead arsenate exposure and death resulting from a specific type of cancer. There was, however, a suggestion of a twofold increase in digestive cancers in both men and women in our expo sure categories. This observation deserves further follow up, given the results of other studies that have found excess digestive cancers in smelter, mine, and refinery workers.20,21 The observed increased risk of dying of all causes in the male orchardists and intermediates in this study contradicted the results of the earlier follow-up study by Nelson et al.,28 who found no overall excess mortality in either male or female orchardists and intermediates. This may have resulted from the increase in total num ber of deaths in the cohort, the many years that fol lowed exposure to lead arsenate, and from the choice of the comparison group. In 1968, only 37% of the study cohort had died, and most had not reached the age at which mortality begins to increase. In our study, 66.4% of the study members who were adults in 1938 had died. It has been suggested that arsenic has a laten cy period of up to 51 y between exposure and appear ance of cancers and other diseases.36 Hence, in the 1968 follow up, 30 y may not have been sufficient time to observe excess mortality. There were several factors that could not be con trolled for in this study. Ideally, the actual amount of exposure (i.e., blood-lead and urinary arsenic levels) for each individual should be known in order to assign each individual to an exposure group. In this study, however, participants were assigned to the exposure group, based on their use of lead arsenate spray prior to and during the 1938 apple-growing season. Blood-lead and urinary arsenic levels were measured after the par ticipants were assigned to the exposure group. Different amounts of lead and arsenic were encountered in the various jobs in the orchards (Table 2). Thus, an orchardist who spent relatively more time burning con tainers would have had greater exposure to lead arsen ate than an orchardist who spent less time in this activ ity. Nonetheless, both would have been classified in the same exposure group. A more accurate measure of exposure would have included the number of years each study member was engaged in each specific job as well as measures of blood lead and urinary arsenic lev els. The lack of this information resulted in misclassification from using such broadly defined occupations. It was not possible to control for additional exposure to lead and arsenic after 1938. Exposure to large con centrations of lead arsenate probably did not occur often after 1938. In 1938, the Department of Agricul ture recommended that other pesticides replace the use of lead arsenate. After the 1940s, the major routes of exposure to lead and arsenic would have been expect ed to be through the soil and water. Neal et al., howev er, determined that the local water supply did not pro vide an additional source of arsenic.27 The possibility that lead and arsenic were present in locally grown produce (other than apples) was not addressed. The healthy-worker effect can often explain reduced mortality risks in occupational studies that use the gen eral population as a comparison group. The healthyworker effect, however, disappears as the length of the follow-up time increases. In this study of orchardists, the follow-up time of 50 y was of sufficient duration that the healthy-worker effect did not affect the results. Smoking habits may also have been a modifying or confounding factor, because smoking is known to be associated with several of the diseases common to arsenic exposure. It was not possible to obtain an accu rate account of the smoking habits of this cohort for sev eral reasons. The amount of time since the participants were first identified was substantial (52 y), and a major ity of the participants were deceased. An informant or close relative would have been needed to provide infor mation about the decedent's smoking habits. Any infor mation on the smoking habits of an individual who died many years previously would most likely lack accuracy. In fact, even living individuals found it difficult to recall their own smoking habits for a period of similar dura tion (i.e., 52 y). Many of the study members who were alive and who were between the ages of 80 and 100 y had difficulty recalling past events. Wicklund et al. examined the smoking habits of orchardists who were living in this same area of Washington State26 (some of their cases were also members of the Neal Study). No differences in smoking habits between orchardists and nonorchardists were observed. This suggests that the smoking habits of the three groups of this study may not have been very different. If smoking were more com mon in the two exposed groups (i.e., orchardists and intermediates), a higher excess of lung cancer mortality should have been observed. Our ability to observe differences in mortality in this study was limited by the small number of study sub jects. Although the length of follow up was long, the power of the study remained limited because of the small sample size. Consistency in the coding of cause of death over such a long time period may also have affected our ability to detect differences in mortality in our study. However, in an effort to reduce inconsistencies in coding, all of the earlier death certificates were recoded, using ICD-9. Accuracy of the cause of death has always been depen dent upon the ability of the attending physician to May/June 1995 [Vol. 50 (No. 3)] 227 determine the reason(s) for the death, especially if an autopsy is not completed. There is no reason to believe the attending physicians for individuals in each expo sure group differed in their ability to determine cause of death. In summary, this study has demonstrated an increased risk of dying of all causes of death in male orchardists and intermediates, as well as an increased risk of dying of coronary heart disease in male interme diates who were exposed to lead arsenate pesticide spray. The study size was too small to determine if there were other elevated risks for specific causes. No increased risk of death from cancer was found in any of the groups exposed; however, our power to detect these differences was low. The study cohort was of lower occupational levels and exposed at different forms of arsenic than were workers in whom excess cancers have been noted. These two factors may explain why no overall increased risk of death from cancer was observed in the Neal Cohort. Authors' Note One reviewer suggested that we confer with a toxi cologist to calculate estimates of expected excess risk of lung cancer in this cohort. We did so, and we think that this approach would be the same as that of a standard mortality ratio analysis. We did not select this approach initially because the Cox proportional hazards analysis is a much more powerful method to use when follow up times vary considerably. We were also able to con trol for other explanatory variables, such as age at expo sure, using the proportional hazards method. Although we realize that many occupational studies use the SMR method, we think that more powerful analytic methods are available and should be used. The calculation of SMRs would also require an extensive amount of addi tional time and probably would not add much to the results of the study. Another reviewer suggested that we calculate cumulative exposure values. We were also unable to do this because we were unable to determine the number of years each individual spent at work in orchards. ********** Submitted for publication December 22, 1993; revised; accepted for publication August 11, 1994. Requests for reprints should be sent to Kristine Tollestrup, Ph.D., Institute for Health and Population Research, The Lovelace Institutes, 2425 Ridgecrest Drive, S.E., Albuquerque, NM 87108. ********** References 1. Agency for Toxic Substances and Disease Registry. Toxicological profile on lead. U.S. Department of Health and Human Services. Report No. TP-92/12, 1992; 307 p. 2. Seppalainen AM, Hornberg S. A follow-up study of nerve con duction velocities in lead exposed workers. Neurobehav Toxic Terat 1982;4:721-73. 3. Stollery BT, Banks HA, Broadbent DE, et al. Cognitive function ing in lead workers. Br J Ind Med 1989; 46:698-707. 228 4. Harlan WR, Landis JR, Schmouder RL, Goldstein NG, Harlan LC. Blood lead and blood pressure. JAMA 1985; 253:530-34. 5. Michaels D, Zoloth SR, Stern FB. Does low-level lead exposure increase risk of death? A mortality study of newspaper printers. Int J Epidemiol 1991; 20:978-83. 6. Cooper WC. Cancer mortality patterns in the lead industry. Ann NY Acad Sci 1976; 171 :250-59. 7. Copper WC, Gaffey WR. Mortality of lead workers. J Occup Med 1975; 17:100-07. 8. Pinto SS, Enterline PE, Henderson V, Varner MO. Mortality expe rience in relation to a measured arsenic trioxide exposure. Envi ron Health Persp 1977; 19:127-30. 9. Enterline PE, Marsh GM. Cancer among workers exposed to arsenic and other substances in a copper smelter. Am J Epidemi ol 1982; 116:895-911. 10. Enterline PE, Henderson V, Marsh GM. Exposure to arsenic and respiratory cancer--a reanalysis. Am J Epidemiol 1987; 125:929 38. 11. Lee-Feldstein A. Arsenic and respiratory cancer in humans: fol low-up of copper smelter employees in Montana. JNCI 1983; 70:601-09. 12. Lee-Feldstein A. Cumulative exposure to arsenic and its relation ship to respiratory cancer among copper smelter employees. J Occup Med 1986; 28:296-302. 13. Welch K, Higgins I, Oh M, et al. Arsenic exposure, smoking, and respiratory cancer in copper smelter workers. Arch Environ Health 1982; 37:325-35. 14. Jarup L, Pershagen G, Wall S. Cumulative arsenic exposure and lung cancer in smelter workers: a dose-response study. Am J Ind Med 1989; 15:31-41. 15. Ott MG, Holder BB, Gordon HL. Respiratory cancer and occu pational exposure to arsenicals. Arch Environ Health 1974; 29:250-55. 16. Sobel W, Bond GG, Baldwin CL, Ducommun DJ. An update of respiratory cancer and occupational exposure to arsenicals. Am J Ind Med 1988; 13:263-70. 17. Mabuchi K. Lung cancer among pesticide workers exposed to inorganic arsenicals. Arch Environ Health 1979; 34:312-20. 18. Chen CJ, Chuang SL, You TM, et al. A retrospective study on malignant neoplasms of bladder, lung, and liver in blackfoot dis ease endemic area in Taiwan. Br J Cancer 1986; 53:399-405. 19. Chen CJ, Wang CJ. Ecological correlation between arsenic level in well water and age-adjusted mortality from malignant neo plasms. Cancer Res 1990; 50:5470-74. 20. Tsuda T, Nagira T, Yamamoto M, et al. An epidemiological study on cancer in certified arsenic poisoning patients in Toroku. Ind Health 1990; 28:53-62. 21. Tokudome S, Kuratsune M. A cohort study on mortality from can cer and other causes among workers at a metal refinery. Int J Can cer 1976; 17:310-17. 22. Wall S. Survival and mortality pattern among Swedish smelter workers. Int J Epidemiol 1980; 9:73-87. 23. Roth F. The sequelae of chronic arsenic poisoning in Moselle vintners. Germ Med Monthly 1957; 2:172-75. 24. Braun W. Carcinoma of the skin and internal organs caused by arsenic. Germ Med Monthly 1958; 3:321-24. 25. Barthel E. Increased risk of lung cancer in pesticide-exposed agri cultural workers. J Toxicol Environ Health 1981; 8:1027-40. 26. Wicklund KG, Daling JR, Allard ), Weiss NS. Respiratory cancer among orchardists in Washington State, 1968-1980. J Occup Med 1988; 30:561-64. 27. Neal PA, Dreesen WC, Edwards Tl, et al. A study of the effect of lead arsenate exposure on orchardists and consumers of sprayed fruits. U.S. Public Health Bull 1941; 267:1-181, 28. Nelson WC, Lykins MH, Mackey J, et al. Mortality among orchard workers exposed to lead arsenate spray: a cohort study. J Chron Dis 1973; 26:105-18. 29. Pinto SS, Bennett BM. Effect of arsenic trio*ide exposure on mor tality. Arch Environ Health 1963;7: 583-91. 30. Boulos BM. Special problems of lead in women workers. Carnow BW, Ed. Health effects of occupational lead and arsenic expo sure. U.S. Department of Health, Education and Welfare, NIOSH (76-134); 1976, pp 76-134. Archives of Environmental Health 31. Axelson O, Dalhgren E, Jansson C-D, Rehnlund SO. Arsenic exposure and mortality: a case-referent study from a Swedish copper smelter. Br J Ind Med 1978; 35:8-15. 32. Landis JR, Flegal KM. A generalized Mantel-Haenszel analysis of the regression of blood pressure on blood lead using IMHANES II data. Environ Health Perspect 1988; 78:35-41. 33. Schwartz, J. Lead, blood pressure and cardiovascular disease. Needleman H, Ed. Human lead exposure. Boca Raton, FL: CRC Press, 1992; pp 223-31. 34. Elwood PC, Davey-Smith C, Oldham PD, et al. Two Welsh sur veys of blood lead and blood pressure. Environ Health Perspect 1988; 78: 119-21. 35. Neri LC, Hewitt D, Orser B. Blood lead and blood pressure: analysis of cross-sectional and longitudinal data from Canada. Environ Health Perspect 1988; 78:123-26. 36. Ott M, Holder B, Gordon H. Respiratory cancer and occupation al exposure to arsenicals. Arch Environ Health 1974; 29:250-55. 37. Pinto SS, Henderson V, Enterline PE. Mortality experience of arsenic exposed workers. Arch Environ Health 1978; 33:325-31. 38. National Research Council. Arsenic. Washington, D.C.: National Academy of Science, 1977; 332 pp. 39. Agency for Toxic Substances and Disease Registry. Toxicological profile on arsenic. U.S. Department of Health and Human Ser vices. Report No. TP-92/02; 1992, 175 p. LU CO Eri^RornnaiT DC O CO CD an ORDER FORM YES! I would like to order a one-year subscription to Environment, pub lished monthly (combined Jan./Feb., July/Aug.). I understand payment can be made to Heldref Publications or charged to my VISA/MasterCard (circle one). $35.00 individuals $70.00 institutions ACCOUNT#. .EXPIRATION DATE. SIGNATURE. NAME/INSTITUTION. ADDRESS___________ CITY/ST ATE/ZIP_ COUNTRY_______ ADD $16.00 FOR POSTAGE OUTSIDE THE U.S. ALLOW 6 WEEKS FOR DELIVERY OF FIRST ISSUE. SEND ORDER FORM AND PAYMENT TO: HELDREF PUBLICATIONS, ENVIRONMENT 1319 EIGHTEENTH STREET. NW, WASHINGTON, DC 20036-1802 PHONE (202) 296-6267 FAX (202) 296-5149 SUBSCRIPTION ORDERS 1 (800) 365-9753 Environment brings you today's most critical environ mental issues put in perspective by the world's foremost scien tists and policymakers. The re sult is a lively, well-respected magazine with readable, thought-provoking articles for the nonspecialist as well as the environmental professional. For more than 36 years, Environment has been the source for balanced, authorita tive discussions of the environ mental issues shaping the world. Regular features include news briefs of current issues, critical reviews of major government and institutional reports, listings of notable books, and commen taries that provide differing points of view on the articles. May/June 1995 JVol. 50 (No. 3)1 229