Archives of Environmental Health: An International Journal
ISSN: 0003-9896 (Print) (Online) Journal homepage: http://www.tandfonline.com/loi/vzeh20
Lung Cancer among Pesticide Workers Exposed to Inorganic Arsenicais Kiyohiko Mabuchl M.D. , Abraham M. Lilienfeld M.D. & Laura M. Snell M.P.H. To cite this article: Kiyohiko Mabuchl M.D. , Abraham M. Lilienfeld M.D. & Laura M. Snell M.P.H. (1979) Lung Cancer among Pesticide Workers Exposed to Inorganic Arsenicais, Archives of Environmental Health: An International Journal, 34:5, 312-320, DOI: 10.1080/00039896.1979.10667423 To link to this article: http://dx.doi.org/10.1080/00039896.1979.10667423
Published online: 17 Apr 2013.
Submit your article to this journal
Article views: 3
View related articles
Citing articles: 36 View citing articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=vzeh20 Download by: [University of California, San Diego]
Date: 11 April 2016, At: 16:06
Downloaded by [University of California, San Diego] at 16:06 11 April 2016
Lung Cancer among Pesticide Workers Exposed to Inorganic Arsenicals KIYOHIKO MABUCHI, M.D. Department of Epidemiology and Preventive Medicine University of Maryland School of Medicine Baltimore, Maryland
ABSTRACT. Cancer mortality was studied in 1,393 persons exposed to high air concentrations of inorganic arsenicals for varying lengths of time during the manufacture and packaging of pesticides at a plant in Baltimore, Maryland. Employees consisting of 1,050 males and 343 females were traced for the period 1946 to 1977, and vital status was determined for 86.9% of males and 66.8% of females. The observed number of deaths from all and selected causes was compared with the number expected from the Baltimore City mortality experience in terms of the Standardized Mortality Ratio (SMR). A significantly increased SMR was found for lung cancer and anemias in males. The SMR for lung cancer was especially high in males with presumed high exposure to arsenicals. A dose-response relationship was suggested by the SMR for lung cancer which increased with increasing duration of exposure to arsenicals, but no such relationship was evident for nonarsenicals. Although smoking habits could not be examined in the study subjects, the authors believe that the fmdings provide additional epidemiologic evidence on the respiratory carcinogenicity of occupational exposure to airborne inorganic arsenicals.
SINCE the observation by Paris in 1820 on cancer of the scrotum among smelter workers in Cornwall and Wales, l epidemiologic studies on the carcinogenic effect of occupational exposure to airborne arsenic have yielded varying results. Some studies conducted among sheep-dip factory workers/ copper smelter workers,3,4 and pesticide workers s who were exposed to a variety of inorganic arsenicals, suggested an increased risk of cancer, particularly of the respiratory tract, whereas others demonstrated no carcinogenic risk. 6 , 7
312
ABRAHAM M. LILIENFELD, M.D. LAURA M. SNELL, M.P.H. Department of Epidemiology Johns Hopkins University School of Hygiene and Public Health
Some of the negative results may be attributed to methodologic faults, i.e., incomplete ascertainment of deaths and the use of inappropriate comparison groups. 8 Another serious problem in many occupational studies stemmed from the wide variety of agents other than arsenic to which study subjects were concomitantly exposed. Thus, smelter workers were exposed not only to arsenical dusts but also to sulfur dioxide, benzpyrene, and other smelting dusts,3,4 and pesticide workers to a variety of both arsenical and non arsenical compounds. s However, it is unclear whether the observed excess cancer risk was due to arsenic alone, to other carcinogenic agents, or to a co.mbination of both. An opportunity arose for an epidemiologic study in an occupational group exposed to high airborne concentrations of inorganic arsenicals during the manufacture and packaging of pesticides in a plant in Baltimore, Maryland. A preliminary study among retirees of the plant suggested an excess mortality from cancer of the lung and the lymphatic tissues as compared with the mortality experience of the Baltimore City population. 9 Excess lymphatic cancer mortality had been reported by only one other study among another group of pesticide workers exposed to arsenicals. S These preliminary findings led the authors to undertake a comprehensive follow-up study of those workers who were employed during the 1946 to 1974 period. The specific objectives were: (a) to determine whether the preliminary findings would be substantiated among the total group of workers, and (b) to attempt to separate the effects of exposure to arsenicals and non arsenicals. The major findings on the cancer mortality experience of the workers are presented here. More detailed descriptions of the methods as well as worker mortality and morbidity will be reported separately.
Archives of Environmental Health
Downloaded by [University of California, San Diego] at 16:06 11 April 2016
The Plant
1961. Chlordecone (among other organic chemicals handled), was first packaged in 1959. Arsenic trioxide was the starting material used for production of arsenical pesticides. It was shipped into the plant by rail, unloaded, and stored at various locations surrounding the arsenic acid plant where it was reacted with nitric acid. The resulting arsenic acid liquid was stored for the subsequent production of arsenical compounds or packaged for sales. Various arsenical compounds were produced in a separate building ("Insecticide Building"), located adjacent to the arsenic acid plant. When manufacturing lead arsenate, arsenic acid was first mixed with litharge slurry. The precipitated lead arsenate was transported to drum dryers, and the product was then screened, milled, packaged, and bagged. Production procedures for other arsenicals were similar. Before 1952, hygienic control at the arsenic acid plant was poor, and workers involved in arsenic acid operations frequently developed keratoses and other toxic effects of arsenic.lO In 1952, the arsenic acid plant was reconstructed and improved hygienic practices, such as daily showers and clothing changes, were introduced. Measures were also taken to improve hygienic conditions in the insecticide building during the 1950s.
The plant under study, established in 1917, had three major production activities: (1) manufacture of arsenic acid from arsenic trioxide; (2) manufacture of various arsenical insecticides, rodenticides, and herbicides; and (3) blending and formulation of a wide variety of agricultural chemicals. 10 Figure 1 depicts the production of major arsenical and nonarsenical compounds during the 1946 to 1974 period for which quantitative information was available. Production of arsenical compounds, which peaked around 1950, included lead arsenate, calcium arsenate, sodium arsenite, zinc arsenite, magnesium arsenite, and copper acetoarsenite (Paris green). Before 1946 production of arsenical compounds was also high, and various other inorganic chemicals, including copper sulfate, were produced in substantial amounts. Among the organic chemicals manufactured or packaged, chlorinated hydrocarbons (e.g., DDT, aldrin, and toxaphene), organic phosphates, and carbamates were first produced between the mid-1940s and 1955, with specific compounds varying in different years, and organic herbicides (e.g., 2, 4-dichlorophenoxy acetic acid and monuron-TCA) were first produced between 1956 and
Arsenicals
oy Le~d arsenate
5,000
,y~i't
Arsenic trioxide
O
-
f,
Q)
O-f~\ 1 ' /011
Z
~
I
o
.1 '/-.i 1/ x _
I
I
1946
Calcium arsenate
/ x--~
\\
Sodium arsenite 0
\\........ ~-
. ./.,C
x--+=:.....
I
I
\~~/ "-%~ '" I
I
I
I
':f--, _-"-_ """Q=IIiI I I ~X-~--"'I. I I
1955
1960
10,000 Non-arsenicals
1965
"
!'\\ / \
DDT and I related compounds!
\/
~!
f
-
r
.,
]':_ A ? .... 'V I
o 1946
.X--_x...
V
,x..
I
\
....---J
_~ --x......,L.0~-..x-
X--~-:;-=i4~-?/~ 1950
Organic I herbicides
~
"'"......
Parathion/ \ Carbamate group
I
Aldrin /Toxaphere group
Z
\
\
!
5,000
Q)
........
I ~~ 'S!- ~.~ " - -."1-",r~.i:;;~? o_~=o-
"'I.'
"1(--
1950
~t~
I "'
"'V fA.'- 0 ,.,.
) 0
\\ I \~ I .~ .(
'V,
• ~
t·-·.·~ :.-~=~-r:x--x--x i
1955
1960
1965
Year Fig. 1. Production of selected arsenical and non-arsenical compounds by calendar year (DDT-related compounds include other diphenyl aliphatic chlorinated hydrocarbons).
September/October 1979
313
The only air sample obtained from the insecticide building in 1972 recorded the atmospheric concentration of 0.5 mg As/m 3 as a time-weighted average. The air concentration of arsenic in the insecticide building in the 1950s was suspected to be at least twice this concentration. The highest level of arsenic in the arsenic acid plant before its reconstruction in 1952 was suspected to be at least 5 mg/m 3 . Methods
Downloaded by [University of California, San Diego] at 16:06 11 April 2016
Study Population A list of the names and the first and last dates of employment were assembled for 3,141 individuals who had been employed either as hourly workers during 1946 to 1974 or as salaried workers during 1955 to 1974. Complete record files were not available for the hourly workers who had left the plant before 1946 and for salaried workers who had left the plant before 1955. Because some difficulties were anticipated in tracing temporary, short-term employees, 441 or a 20% simple random sample of the 2,189 workers with less than 4 months of employment interval (derived by subtracting the first from the last date of employment) were selected for the follow-up. All of the remaining 952 employees (employed 4 months or longer) were also included in the study. The total number of individuals in the follow-up study was, therefore, 1,393 (1,050 males and 343 females). All those employed before 1970 were white. Only a small proportion of workers employed after 1970 were black, but no specific information on race was available in the individual records for workers employed after 1970. Work Exposure During the period 1946 to 1976, there were three major departments: (1) manufacture or production; (2) maintenance (including construction, mechanic, and service shop); and (3) shipping and warehouse. Until the late 1960s the production department was further organized into various divisions according to the chemicals manufactured, including arsenic acid, insecticides, DDT, parathion, and herbicides. "High" exposure to arsenicals was assumed among arsenic acid operators and production workers in the insecticide division. Similarly, high exposure to nonarsenicals was assumed among production workers in the DDT, parathion, herbicide, and other manufacturing divisions. In many instances no records were available at the time of study regarding the specific chemicals handled in certain areas before 1930. After 1970, the division in which an employee worked was often not specified. "Possibly high" exposure to either arsenicals or non arsenicals was assumed among these unspecified production workers. Maintenance and shipping workers were assumed to have been exposed to both arsenicals and nonarsenicals at a "medium" level, and office workers at a "low" level. To characterize each worker's exposure experience, each person was allocated into one of the following six work exposure groups according to the highest degree of exposure to arsenicals and/or nonarsenicals: (1) arsenical
314
(AS) production; (2) predominantly arsenical (PA) production; (3) predominantly nonarsenical (PNA) production; (4) unspecified production; (5) maintenance/shipping; and (6) office workers. The number of male employees allocated to each group by varying degree of arsenical and non arsenical exposure is shown in Table 1. Of the 259 male PA workers, 69.5% had concomitant high exposure, 6.2% had possible high exposure, and 24.3% had medium or less exposure to nonarsenicals. Therefore, some PA workers were substantially exposed to nonarsenicals as well. Of the 100 PNA workers, only 2.7% had possibly high exposure and 17.3% had medium exposure to arsenicals, while the remaining 80.0% experienced low exposure. Follow-up The study employees were traced from January 1975 through August 1977 using company pension records, mail questionnaires, telephone contacts of employees and their family members, friends and other acquaintances, as well as other sources such as records of the Boards of Elections of Baltimore City and its two surrounding counties, the motor vehicle departments of Maryland and other states, the Social Security Administration, the Veterans Administration, and the Vital Statistics Office of the State of Maryland. For each person who was found to be deceased, a copy of the death certificate was requested. Of the 1,393 employees, 901 (64.7%) were found to be alive, 240 (17.2%) deceased, and 252 (18.1%) could not be located. By sex, 86.9% of 1,050 males and 66.8% of 343 females had been traced. The lower percentage for females was probably due, in part, to name changes because of marriage. Among males, the percentage of traced persons was lowest among short-term employees (84.3% for employees of less than 1 yr, 88.1% for 1 to 4 yr) but quite high in long-term employees (99.3% for 5 yr or longer). Observed and Expected Number of Deaths Underlying causes of death were ascertained from death certificates and coded according to the revision of the International Classification of Diseases (lCD) in effect at the time of death. l l The expected number of deaths from all or specific causes was computed as the sum of the product of the number of person-years of observation for each 5-yr age group and calendar period and the Baltimore City white mortality rate for the same age group and calendar periodY-14 In computing person-years, a person'was entered into observation in the calendar year of initial employment, except persons who were already employed at the beginning of the study were entered in the first study year. Computations and tabulations of observed and expected number of deaths were facilitated by a computer program provided by Monson. 1S Standardized Mortality Ratios (SMR) were computed as the ratio of observed to expected numbers of deaths. Statistical significance for the deviation of the SMR from unity was assessed under the assumption that the observed number of deaths was a Poisson variable. 16
Archives of Environmental Health
Table I.-Number of Persons by Work Exposme Group and the Highest Degree of Exposure among Males Highest Degree of Exposure to Work Exposure Group
Arsenicals
Nonarsenicals
Number of Persons
1) Arsenical (AS) production workers
High
Low
205
High High High High
High Possibly high Medium Low
180 16 62 1
TOTAL
259
Downloaded by [University of California, San Diego] at 16:06 11 April 2016
2) Predominantly arsenical (P A) production workers
3) Predominantly nonarsenical (rNA) production workers
4) Unspecified production workers
Possibly high Medium Low
High High High
3 19 88
TOTAL
110
Possibly high
Possibly high
130
Medium
Medium
266
Low
Low
77
5) Maintenance and shipping workers 6) Office workers
Because changes in the ICD rules for determining causes of death resulted in considerable loss of comparability of mortality statistics before and after 1950, cause-specific mortality rates for 1950 to 1954 were used for the 1946 to 1949 period. Rates for 1970 to 1974 were used for 1975 to 1977 during which time complete mortality statistics were not available. Since lung cancer mortality has markedly increased in the last few decades, the expected number of deaths derived by this procedure for the 1946 to 1949 period would be overestimated and that for the 1975 to 1977 period underestimated. The total effect for the entire study period will depend on the number of person-years of observation in each period. For males, the 1946 to 1949 period represented 3.7% of the total personyears and the 1975 to 1977 period 8.5%. It appears, therefore, that the overall effect of this procedure on the SMR for lung cancer is small. Results Total and Selected Causes of Death by Sex A total of 197 male and 43 female deaths from all causes were observed, while 203.8 and 43.6, respectively, were expected (Table 2). The overall SMR for each sex was close to unity. Among the selected causes of death, only lung cancer and anemias in males showed a statistically significant excess SMR (Table 2). The observed number of deaths from anemias was only two: one from pernicious anemia and the other from aplastic anemia.
September/October 1979
The earlier study of retirees revealed three deaths from lymphatic cancer (lymphosarcoma) which was significantly more than expected. 9 No additional deaths from lymphosarcoma were identified in this follow-up study but there was one death from multiple myeloma. Consequently, mortality from cancer of the lymphatic and other hematopoietic tissues (except leukemia) was excessive, though not significant. Further analysis of the mortality data by employment and exposure characteristics was focused on lung cancer among males. The number of deaths from anemias was too small for detailed analysis. Lung Cancer by Employment Characteristics The SMRs for lung cancer and aU other causes among males were analyzed by work exposure groups as well as by year of first employment and duration of employment Cfable 3). Significantly increased SMRs for lung cancer were found among PA workers, those who were first employed before 1946 and those who were employed for 25 yr or longer. When interpreting the data, it must be remembered that these employment characteristics are not independent. For example, a person who had been employed for 25 yr or longer must have been employed before 1952. When mortality d~ta are classified by a single characteristic, the effect of the others cannot be entirely ignored. To separate the effects of each characteristic, the SMRs for lung cancer among male PA workers were cross-
315
Table 2.-0bserved and Expected Number of Deaths and SMRs from All and Selected Causes by Sex Female
Male
Number of Deaths
Number of Deaths Cause of Death
Observed Expected (0) (E)
Observed Expected (E) (0) SMR
SMR
An causes
(000-999)
197
203.8
.97
43
43.6
.99
Infective and parasitic disease
(000-136)
7
4.4
1.58
2
.7
2.69
Malignant neoplasms, ill sites
Downloaded by [University of California, San Diego] at 16:06 11 April 2016
ICD Eight Revision Code
Buccal cavity and pharynx Esophagus Stomach Large intestine, except rectum Rectum and rectosigmoid Trachea, bronchus, and lung Skin Breast Prostate Leukemia Lymphatic and other hematopoietic tissues Other and unspecified tissues Endocrine, nutritional and metabolic diseases
(140-209)
47
39.4
1.19
9
12.9
.70
040-149) (150) (151) (152,153) (154) (162) (172,173) (174) (185) (204-207)
2 3 3 1 1 23 1 0 1 1
2.0 1.2 1.7 3.2 1.3 13.7 .6 .0 1.5
0 0 0 2 1 0 1 0 0
.3 .1 .4 1.3 .4 1.2 .2 3.1 .0 .3
.00 .00 .00 .79 5.41 .83 .00 .33
1.2
1.01 2.57 1.73 .31 .78 1.68* 1.68 .00 .66 .83
(200-203,208,209)
4
2.1
1.91
0
.6
.00
(155-161,163, 170171,180-184,186-189)
7
9.6
.73
4
3.6
1.12
(240-289)
4
4.7
.86
1
2.1
.48
1
---
.00
(280-285)
2
.2
0
.1
.00
Diseases of circulatory system
(390458)
82
97.8
.84
21
17.6
1.19
Diseases of respiratory system
(460-519)
12
10.4
1.15
2
1.6
1.24
3.3
.60
Anemias
Diseases of digestive system Accidents, suicide and homicide All other causes
10.37*
(520-577)
16
16.5
.97
2
(E800-E978)
24
20.8
1.16
5
2.7
1.87
(210-239,290-389, 580-796, E980-E999)
5
9.8
.51
1
2.7
.37
*P