internationalArchivesof
Int Arch Occup Environ Health (1990) 62:253-257
OCeUpaifloald
Environmental Health © Springer-Verlag 1990
Mortality and cancer morbidity in a cohort of Swedish glassworkers G Wingren1 and V Englander 2 'Department of Occupational Medicine, University Hospital, S-581 85 Linkoping, Sweden Department of Occupational Medicine, University Hospital, S-221 85 Lund, Sweden
2
Received December 8, 1989 / Accepted January 19, 1990
Summary A cohort of 625 male, art glassworkers with employment of more than one month for some time between 1964 and 1985 was studied with regard to causes of death and cancer incidence The number of total deaths was 97 versus 98 9 expected from national death rates and 82 9 expected from county death rates A moderate increase in total cancer deaths was seen (26 observed cases vs 22 3 and 18 9 expected from national and county death rates, respectively) as also associated with duration of exposure The excess number of cases was particularly noted for lung cancer, ( 6 observed cases vs 4 2 and 2 5 expected, respectively), colon cancer (4 observed cases vs 1 6 expected from both national and county death rates), cancer of the pharynx (2 observed cases vs 0 2 and 0 1 expected, respectively), and prostate (4 observed cases vs 3 0 and 2 4 expected, respectively). Most of these excesses occurred among men working in the foundry producing either heavy crystal glass or semicrystal glass, except for cancer of the prostate which occurred in the context of glass refinement Excess risks for deaths from ischemic heart disease (39 observed cases vs 32 1 and 30 9 expected, respectively) and cerebrovascular disease (11 observed cases vs 7 3 and 6 5 expected, respectively) were also found, although with no clear relationship to duration of exposure These latter risks occurred particularly among men working with the refinement of glass Approximative correction for smoking habits resulted in increased risk estimates for lung cancer and ischemic heart disease, indicating a negative confounding effect from smoking in this material. Key words: Glassworks Metals Cancer heart disease Cerebrovascular disease
Ischemic
Introduction Glassworkers are exposed to a variety of chemicals, mostly metallic compounds, but also silica dust and some Offprint requests to: G Wingren
asbestos Very little can be found in the literature regarding possible health hazards in this worker group, however An investigation of the hygienic situation in the Swedish glass industry was undertaken in 1976, and revealed some problems concerning climatic conditions, particularly heat, stress, noise, lighting, use of chemicals, ergonomics, and accidents l19 l. Two recent and partially overlapping Swedish casecontrol studies have shown increased risks for glassworkers to die from cancers of the stomach, colon and lung, as well as from cardiovascular disease l21, 22 l The excess mortality from these causes appeard especially among the glassblowers Furthermore, an Italian cohort study has shown an excess of lung cancer and cancer of the larynx among glassworkers l5 l An increased risk of lung cancer has also been noted for the Danish glass industry l14l, as well as among employees in glass manufacturing in Alameda county, USA l15 l, and in female glass product workers in Shanghai l12l. Although retrospective epidemiologic research only reflects occupational risks in the past, the production of handmade art glass is quite traditional and certain conditions of exposure have probably changed very little over time, especially the oral exposure likely to take place through the blow-pipe Therefore the exposure pattern may be rather much the same today as decades ago and epidemiologic findings may be relatively more relevant for assessing current risks in this industry than in many others. Against the background of the results in the two previous studies from the glass-producing area of southeast Sweden l21, 22 l, the present study was undertaken as an attempt to confirm earlier results on a cohort basis and to obtain further information about the possible health hazards in art glass production The intention was also to consider the influence of duration of exposure on mortality, which was not accounted for in the two earlier case-control studies l21, 22 l The existence of data on smoking in a group of glassworkers at the glassworks involved in this study would also permit some elucidation of the role of smoking as possibly confounding the results of the earlier performed case-control studies.
254
Materials and methods Detailed records of the employees, including information on work tasks, were obtained from one of the largest glassworks in Sweden. All men employed for more than one month at any time between January 1964 and December 1985 were included in a cohort of 666 individuals Their vital statistics were traced until December 1985 through the National Health Insurance and the taxation authorities The National Central Bureau of Statistics was consulted for information on dates of death in the cohort for the period 1964-1985 along with the registered causes of death Information on diagnosed cancers was available for 1964-1983 and obtained from the Swedisch Cancer Registry Thirty-six men were found to have emigrated and five (0 75 %) were lost to follow-up for other reasons, thus leaving 625 men to be included in the analyses. The expected number of deaths was calculated by means of the EPILIN package l2 l with and without latency requirements, i e by multiplying the person-years at observation during the respective calender years by the cause-, age and sex-specific national death rates and summarizing all the fractional contributions over the strata An additional analysis of the material was made by using the EPILUND package (Department of Occupational Medicine in Lund) permitting also a check against the local cause-, age and sex-specific death rates as well as the cancer morbidity rates from the county of Kronoberg to be applied, i e the rates of the county where the glassworks is situated. Standardized mortality ratios (SM Rs) for total deaths, total cancer deaths, specific cancer deaths and for deaths from ischemic heart disease and cerebrovascular disease were calculated as well as standardized incidence ratios (SI Rs) for various cancers along with the p-values and two-tailed confidence intervals based on the Poisson distribution As appearing in the tables, subcohorts were also defined according to different exposure categories and with regard to time period of exposure For economic reasons these subcohorts could only be analysed by the EPILIN package and hence be compared only to national death rates. The smoking habits for a group of 60 workers employed during the 1960S at this glassworks had previously been inquired about in a questionnaire survey, regarding work tasks, various exposures and disease symptoms besides smoking habits l 23 l The data from these 60 men were used to evaluate approximately the possible confounding effect from smoking with regard to lung cancer and ischemic heart disease l3l Reference data regarding smoking habits for Swedish men, 17 to 80-years-old at the appropriate point in time, were taken from an investigation conducted by an insurance company and an antismoking organisation l20l (Table 1) In an alternative analysis, resulting in Table 4, the expected number of cases of lung cancer and ischemic heart disease were adjusted so as to correspond with the smoking habits of the glassworkers l3l under the assumption of a 10-fold increase in lung cancer for exsmokers and a 20-fold increase for smokers For ischemic heart disease, the corresponding estimates were taken as 1 5 for exsmokers and 2 5 for smokers, which seems to be reasonable, judging from the literature l 18 l Furthermore, a requirement of an in-
Table 1 Smoking habits in a group of glassworkers Proportion in percent for each category
All glassworkers Foundry workers Others Reference population
Non smokers
Ex smokers
Smokers
Refused to answer
43 45 42
23 21 26
20 17 22 5
13 17 95
32
21
46
aThe reference population is taken from an investigation about "smoking habits in the Swedish population 1976"
duction-latency period of ten years was added in connection with a condition of 15 years of employment, which would make the analysis somewhat more sensitive to any existing risk. The main route of exposure for glassworkers is probably through inhalation of respirable airborne particles and/or through ingestion of compounds contaminating the blow-pipe Some measurements were undertaken in a few glassworks, both with regard to air contamination at the workplaces and to slag from blowpipes The air measurements showed concentrations of lead from less than 0 001 to 0 110 mg/m3, with the highest levels in the heavy crystal producing glassworks The slag from the blow-pipes contained lead, but also manganese and nickel were found l 1l. The overall definition of exposure in this study was taken as employment at the glassworks for more than one month A separation into six exposure categories (or subcohorts) was also made, however: (I) smelters and other oven workers; (II) glassblowers and other workers in the foundry manufacturing heavy crystal glass plus service men in this area; (III) glassblowers and other workers manufacturing semi-crystal glass; (IV) refinement workers i e grinders, etchers, polishers and controlers; (V) glasspainters; (VI) packers, transportation workers, storehouse workers and others This grouping of workers was based on an assumption of comparable metal exposures as decided by a representative of the glassworks, an industrial hygienist and an epidemiologist, all three with particular experience of the hygienic problems in the glass industry. Three employment periods were taken into account in the analyses, and since a person could belong to different exposure categories in different periods, one particular case can appear in more than one of the different subcohorts.
Results The final cohort for which the vital statistics were known consisted of 625 men, contributing 9151 person-years at observation Their average exposure (employment) time was 15 6 years and the earliest year of employment was 1907. There were a total of 97 observed deaths in the cohort vs 98 9 expected by national death rates and 82 9 expected by county death rates (Table 2) Only causes of death and cancer incidence involving more than two
Table 2 Total and cause-specific mortality in the entire cohort. Expected number of cases calculated from national and county death rates Causes of death
Total deaths All tumors Cancer of the pharynx Colon cancer Lung cancer Cancer of the prostate Ischemic heart disease Cerebrovascular disease
ICD
Obs
National death rates
County death rates
Exp
SMR
Exp
SMR
0 98 1 16
82 9 18 9
1 17 1 38
000-999 140-239
97 26
98 9 22 3
146-149 153 162-163
2 4 6
02 16 42
9 87 ** 0 1 2 45 16 1 44 25
185
4
30
1 34
24
1 65
410-414
39
32 1
1 21
30 9
1 26
430-438
11
73
1 50
65
1 68
*P
Int Arch Occup Environ Health (1990) 62:253-257
OCeUpaifloald
Environmental Health © Springer-Verlag 1990
Mortality and cancer morbidity in a cohort of Swedish glassworkers G Wingren1 and V Englander 2 'Department of Occupational Medicine, University Hospital, S-581 85 Linkoping, Sweden Department of Occupational Medicine, University Hospital, S-221 85 Lund, Sweden
2
Received December 8, 1989 / Accepted January 19, 1990
Summary A cohort of 625 male, art glassworkers with employment of more than one month for some time between 1964 and 1985 was studied with regard to causes of death and cancer incidence The number of total deaths was 97 versus 98 9 expected from national death rates and 82 9 expected from county death rates A moderate increase in total cancer deaths was seen (26 observed cases vs 22 3 and 18 9 expected from national and county death rates, respectively) as also associated with duration of exposure The excess number of cases was particularly noted for lung cancer, ( 6 observed cases vs 4 2 and 2 5 expected, respectively), colon cancer (4 observed cases vs 1 6 expected from both national and county death rates), cancer of the pharynx (2 observed cases vs 0 2 and 0 1 expected, respectively), and prostate (4 observed cases vs 3 0 and 2 4 expected, respectively). Most of these excesses occurred among men working in the foundry producing either heavy crystal glass or semicrystal glass, except for cancer of the prostate which occurred in the context of glass refinement Excess risks for deaths from ischemic heart disease (39 observed cases vs 32 1 and 30 9 expected, respectively) and cerebrovascular disease (11 observed cases vs 7 3 and 6 5 expected, respectively) were also found, although with no clear relationship to duration of exposure These latter risks occurred particularly among men working with the refinement of glass Approximative correction for smoking habits resulted in increased risk estimates for lung cancer and ischemic heart disease, indicating a negative confounding effect from smoking in this material. Key words: Glassworks Metals Cancer heart disease Cerebrovascular disease
Ischemic
Introduction Glassworkers are exposed to a variety of chemicals, mostly metallic compounds, but also silica dust and some Offprint requests to: G Wingren
asbestos Very little can be found in the literature regarding possible health hazards in this worker group, however An investigation of the hygienic situation in the Swedish glass industry was undertaken in 1976, and revealed some problems concerning climatic conditions, particularly heat, stress, noise, lighting, use of chemicals, ergonomics, and accidents l19 l. Two recent and partially overlapping Swedish casecontrol studies have shown increased risks for glassworkers to die from cancers of the stomach, colon and lung, as well as from cardiovascular disease l21, 22 l The excess mortality from these causes appeard especially among the glassblowers Furthermore, an Italian cohort study has shown an excess of lung cancer and cancer of the larynx among glassworkers l5 l An increased risk of lung cancer has also been noted for the Danish glass industry l14l, as well as among employees in glass manufacturing in Alameda county, USA l15 l, and in female glass product workers in Shanghai l12l. Although retrospective epidemiologic research only reflects occupational risks in the past, the production of handmade art glass is quite traditional and certain conditions of exposure have probably changed very little over time, especially the oral exposure likely to take place through the blow-pipe Therefore the exposure pattern may be rather much the same today as decades ago and epidemiologic findings may be relatively more relevant for assessing current risks in this industry than in many others. Against the background of the results in the two previous studies from the glass-producing area of southeast Sweden l21, 22 l, the present study was undertaken as an attempt to confirm earlier results on a cohort basis and to obtain further information about the possible health hazards in art glass production The intention was also to consider the influence of duration of exposure on mortality, which was not accounted for in the two earlier case-control studies l21, 22 l The existence of data on smoking in a group of glassworkers at the glassworks involved in this study would also permit some elucidation of the role of smoking as possibly confounding the results of the earlier performed case-control studies.
254
Materials and methods Detailed records of the employees, including information on work tasks, were obtained from one of the largest glassworks in Sweden. All men employed for more than one month at any time between January 1964 and December 1985 were included in a cohort of 666 individuals Their vital statistics were traced until December 1985 through the National Health Insurance and the taxation authorities The National Central Bureau of Statistics was consulted for information on dates of death in the cohort for the period 1964-1985 along with the registered causes of death Information on diagnosed cancers was available for 1964-1983 and obtained from the Swedisch Cancer Registry Thirty-six men were found to have emigrated and five (0 75 %) were lost to follow-up for other reasons, thus leaving 625 men to be included in the analyses. The expected number of deaths was calculated by means of the EPILIN package l2 l with and without latency requirements, i e by multiplying the person-years at observation during the respective calender years by the cause-, age and sex-specific national death rates and summarizing all the fractional contributions over the strata An additional analysis of the material was made by using the EPILUND package (Department of Occupational Medicine in Lund) permitting also a check against the local cause-, age and sex-specific death rates as well as the cancer morbidity rates from the county of Kronoberg to be applied, i e the rates of the county where the glassworks is situated. Standardized mortality ratios (SM Rs) for total deaths, total cancer deaths, specific cancer deaths and for deaths from ischemic heart disease and cerebrovascular disease were calculated as well as standardized incidence ratios (SI Rs) for various cancers along with the p-values and two-tailed confidence intervals based on the Poisson distribution As appearing in the tables, subcohorts were also defined according to different exposure categories and with regard to time period of exposure For economic reasons these subcohorts could only be analysed by the EPILIN package and hence be compared only to national death rates. The smoking habits for a group of 60 workers employed during the 1960S at this glassworks had previously been inquired about in a questionnaire survey, regarding work tasks, various exposures and disease symptoms besides smoking habits l 23 l The data from these 60 men were used to evaluate approximately the possible confounding effect from smoking with regard to lung cancer and ischemic heart disease l3l Reference data regarding smoking habits for Swedish men, 17 to 80-years-old at the appropriate point in time, were taken from an investigation conducted by an insurance company and an antismoking organisation l20l (Table 1) In an alternative analysis, resulting in Table 4, the expected number of cases of lung cancer and ischemic heart disease were adjusted so as to correspond with the smoking habits of the glassworkers l3l under the assumption of a 10-fold increase in lung cancer for exsmokers and a 20-fold increase for smokers For ischemic heart disease, the corresponding estimates were taken as 1 5 for exsmokers and 2 5 for smokers, which seems to be reasonable, judging from the literature l 18 l Furthermore, a requirement of an in-
Table 1 Smoking habits in a group of glassworkers Proportion in percent for each category
All glassworkers Foundry workers Others Reference population
Non smokers
Ex smokers
Smokers
Refused to answer
43 45 42
23 21 26
20 17 22 5
13 17 95
32
21
46
aThe reference population is taken from an investigation about "smoking habits in the Swedish population 1976"
duction-latency period of ten years was added in connection with a condition of 15 years of employment, which would make the analysis somewhat more sensitive to any existing risk. The main route of exposure for glassworkers is probably through inhalation of respirable airborne particles and/or through ingestion of compounds contaminating the blow-pipe Some measurements were undertaken in a few glassworks, both with regard to air contamination at the workplaces and to slag from blowpipes The air measurements showed concentrations of lead from less than 0 001 to 0 110 mg/m3, with the highest levels in the heavy crystal producing glassworks The slag from the blow-pipes contained lead, but also manganese and nickel were found l 1l. The overall definition of exposure in this study was taken as employment at the glassworks for more than one month A separation into six exposure categories (or subcohorts) was also made, however: (I) smelters and other oven workers; (II) glassblowers and other workers in the foundry manufacturing heavy crystal glass plus service men in this area; (III) glassblowers and other workers manufacturing semi-crystal glass; (IV) refinement workers i e grinders, etchers, polishers and controlers; (V) glasspainters; (VI) packers, transportation workers, storehouse workers and others This grouping of workers was based on an assumption of comparable metal exposures as decided by a representative of the glassworks, an industrial hygienist and an epidemiologist, all three with particular experience of the hygienic problems in the glass industry. Three employment periods were taken into account in the analyses, and since a person could belong to different exposure categories in different periods, one particular case can appear in more than one of the different subcohorts.
Results The final cohort for which the vital statistics were known consisted of 625 men, contributing 9151 person-years at observation Their average exposure (employment) time was 15 6 years and the earliest year of employment was 1907. There were a total of 97 observed deaths in the cohort vs 98 9 expected by national death rates and 82 9 expected by county death rates (Table 2) Only causes of death and cancer incidence involving more than two
Table 2 Total and cause-specific mortality in the entire cohort. Expected number of cases calculated from national and county death rates Causes of death
Total deaths All tumors Cancer of the pharynx Colon cancer Lung cancer Cancer of the prostate Ischemic heart disease Cerebrovascular disease
ICD
Obs
National death rates
County death rates
Exp
SMR
Exp
SMR
0 98 1 16
82 9 18 9
1 17 1 38
000-999 140-239
97 26
98 9 22 3
146-149 153 162-163
2 4 6
02 16 42
9 87 ** 0 1 2 45 16 1 44 25
185
4
30
1 34
24
1 65
410-414
39
32 1
1 21
30 9
1 26
430-438
11
73
1 50
65
1 68
*P
