|Home | About | Journals | Submit | Contact Us | Français|
T. Behrens1, W. Schill1, P. Wild2, R. Frentzel‐Beyme1, W. Ahrens1. 1Bremen Institute for Prevention Research and Social Medicine; 2Department of Occupational Epidemiology, Institut National de Recherche et Sécurité, Vandoeuvre
ObjectivesFumes from heated bitumen have been associated with various types of cancer, including cancer of the lung. However, epidemiological studies conducted so far yielded only equivocal results. To study the mortality of asphalt workers, the IARC assembled an international cohort of workers employed in the asphalt industry. Here we report the mortality data of the German cohort updated to 2004.
MethodsGerman asphalt workers with potential bitumen exposure, employed between 1965 and 1997 for at least 1 year, were included in the cohort. Standardised mortality ratios (SMR) and 95% CI were calculated, based on age and period‐specific reference data for the West German male population. Bitumen exposed and unexposed subjects were compared, calculating the relative risk by Poisson regression and adjusting for age and period.
Results7919 male workers were included in the follow‐up, contributing 132200 person years of observation. Based on company information, subjects were classified as exposed to bitumen (n=2535), to bitumen and (potentially) coal tar (n=832), and to neither tar or bitumen (n=2737). 1873 workers were classified as having unknown exposure. By the end of 2004, 835 cohort members had died (SMR 1.27; 95% CI 1.19 to 1.36). The SMR for deaths from lung cancer was 1.77; 95% CI 1.46 to 2.16). Head and neck cancers (defined as oral, pharyngeal, laryngeal or oesophageal cancer showed an SMR of 2.36; 95% CI 1.78 to 3.07). Significantly elevated SMRs were also found for all malignant tumours, alcoholism, and unnatural causes of death (including accidents). Lagged analyses of bitumen exposure delivered ambiguous results. When we stratified the cohort according to bitumen exposure, no clear mortality pattern emerged. Cancer mortality was significantly elevated among both the exposed and unexposed subjects. The internal comparison between bitumen‐exposed and unexposed workers revealed elevated, but non‐significant associations for lung cancer.
ConclusionThe cohort follow‐up confirmed previous results of an elevated cancer mortality for workers employed in the asphalt industry. However, we cannot rule out that our results were biased by confounders that were not collected in the cohort study, most notably smoking and alcohol consumption, as well as other occupational exposures within and outside the asphalt industry.
Key wordsasphalt workers; bitumen; cancer
Y. Iwatsubo, L. Benezet, O. Boutou‐Kempf, E. Chabault, J. Fevotte, L. Garras, M. Goldberg, D. Luce, E. Imbernon. Institut de Veille Sanitaire
ObjectivesIn December 2002, a kidney cancer cluster was reported among the workers of a chemical plant manufacturing nutritional supplements for animal feed. The estimated standardised incidence ratio was 13.1 (95% CI 6.28 to 24.10). Nine out of the ten index cases were diagnosed after their enrolment in an ultrasonic abdominal screening program conducted since 1986. To investigate this cluster, two epidemiological studies were conducted: a retrospective cohort study to analyse all causes and cause specific mortality and a nested case‐control study to examine the relation between exposure to occupational factors and the occurrence of renal cell cancer among the employees of this plant.
MethodsAll workers who had been employed at least 6 months in the plant between 1960 and 2003 were included in the cohort study. An extensive search for additional kidney cancer cases among the former or current employees of the plant was conducted in hospitals and pathological laboratories located in the region. Only the cases diagnosed between 1980 and 2003 were included in the case‐control study. Five controls per case matched on date of birth and gender were selected at random using the incidence density sampling method. A standardised questionnaire for work history and past medical history was completed by trained interviewers. Assessment of occupational exposure by industrial hygienists is still in progress.
ResultsThe cohort included 2523 subjects. For the 1968–2002 period, no significant excess mortality was observed for all causes of death, all cancers or kidney cancer. The case‐control study included 18 cases and 82 controls. Attending the ultrasonic abdominal screening program (OR 5.0, 95% CI 1.6 to 15.8), obesity (OR 4.8, 95% CI 1.0 to 23.3), hypertension (OR 3.2, 95% CI 1.1to 9.4) and tobacco consumption (OR 2.4, 95% CI 0.6 to 10.0) were associated with an increased risk of renal cell cancer. A non‐significant increase in OR with the length of employment in the plant was also observed.
ConclusionThe analyses of the relationship with time periods and the types of industrial process are still in progress for the case‐control study.
Key wordsrenal cell cancer; vitamin synthesis; nested case‐control study
B. Peplonska, U. Wilczynska, W. Sobala, N. Szeszenia‐Dabrowska. Nofer Institute of Occupational Medicine
ObjectivesTo evaluate the risk of death from cancer in a cohort of workers in a rubber tyre factory.
MethodsThe cohort of 17637 workers (11578 men and 6059 women) employed in a rubber tyre plant for at least 3 months during the years 1950–1995 was followed‐up until December 31, 2002. Deaths by causes were analysed using age and time period standardised mortality ratio (SMR) calculated by the person‐years method. The mortality pattern of the general population of Poland was used as the reference. Additionally, internal comparisons were performed according to the production area. We used the Cox proportional hazards model with age at risk as time variable to estimate the hazard ratio (HR).
ResultsTotal mortality in the cohort was significantly lower when compared to the general population (men: SMR 77; women: SMR 67); a similar result was found for mortality from all malignant neoplasms (men: SMR 73; women: SMR 65). The results of the internal comparisons indicated a significantly increased risk of deaths from all malignant neoplasms and from stomach cancer in workers employed at compounding and mixing (HR 1.5 and 2.5, respectively). We also recorded an increased risk of deaths from all cancers (HR 1.3), and from brain cancer in particular (HR 4.2), in workers employed at component building and assembly. The latter cancer was also increased in the storage of finished goods, packing and dispatch areas.
ConclusionSuggestive finding of an increased risk of death from stomach cancer in compounding and mixing has been recorded previously in other studies and might be linked to high dust (carbon black) exposure observed in these areas. The smaller than expected number of observed deaths in the cohort overall and from the specific cancer causes might be explained by the “healthy worker effect” (HWE). However, there is no indication that HWE influenced internal analyses.
Key wordscancer mortality; cohort; rubber
A. Thuret, B. Geoffroy‐Perez, D. Luce, M. Goldberg, E. Imbernon. Institut de Veille Sanitaire
ObjectivesTo study the cause‐specific mortality of a cohort of 12788 construction workers aged 20–64 years, derived from a representative sample of 1% of the French population established during the 1968 national census.
MethodsMortality was followed from 1974 through 1999 yielding 254281 person‐years, and 1908 deaths were observed. Standardised mortality ratios (SMR) and their 95% confidence intervals (95% CI) were calculated for all causes and specific causes using the French population as reference. SMR were calculated first for the overall cohort and then according to socioeconomic status and specific branches within the construction industry.
ResultsConstruction workers had an increased mortality for all causes (SMR 107, 95% CI 102 to 112), and all cancers (SMR 120, 95% CI 111 to 128). Among cancers, an excess of mortality was observed more specifically for cancers of the oral cavity and pharynx (SMR 130, 95% CI 105 to 159), of the digestive system (SMR 119, 95% CI 104 to 137), especially of the oesophagus (SMR 141, 95% CI 111 to 178), and for cancers of the respiratory system (SMR 143, 95% CI 128 to 159). Among non‐malignant diseases, excess mortality was found for cerebrovascular disease (SMR 130, 95% CI 106 to 159), diseases of the digestive system (SMR 130, 95% CI 113 to 149), especially for chronic liver disease and cirrhosis (SMR 131, 95% CI 111 to 153), and for accidental falls (SMR 158, 95% CI 105 to 125). The analysis of workers according to specific branches partly confirmed these results. These results were found essentially in blue‐collar workers. In skilled workers, bladder cancer mortality was increased (SMR 219, 95% CI 113 to 382). Mortality from diabetes mellitus (SMR 241, 95% CI 110 to 458) and mental disorders (SMR 154, 95% CI 113 to 204) was increased for unskilled workers, more specifically for unskilled masons.
ConclusionDespite potential limitations of the study and a probable role of lifestyle factors, elevated mortality was observed for several causes possibly related to occupational factors. Our results are consistent with those of other studies of construction workers. More detailed investigations should study specific occupations and take into account both occupational and non‐occupational risk factors.
Key wordsmortality study; construction workers; socioeconomic status
J. U. Won1, D. H. Koh2, J. H. Roh1, K. S. Kim2. 1Yonsei University College of Medicine; 2Korea Occupational Safety and Health Agency
ObjectivesThe purpose of this retrospective cohort study is to investigate the relationship between exposure in the refinery/petrochemical industry and lymphohaematopoietic cancer.
MethodsThe cohort consists of 19240 men who worked during 1992–2003 in a refinery/petrochemical complex. Standardised mortality ratios (SMR) were calculated for 1992–2003 based on the death rate of the Korean population. Standardised rate ratios (SRR) were calculated for a manufacturing/office group to find the excess risk. The standardised proportionate cancer mortality rate (SPCMR) for lymphohaematopoietic cancer was calculated for all workers.
ResultsThe overall mortality and most cause‐specific mortalities were lower than those for the general Korean population. But there were increased SMRs for lymphohaematopoietic cancer (8/6.1; SMR 133; 95% CI 57 to 262), leukaemias (5/3.7; SMR 137; 95% CI 44 to 319) and multiple myeloma (2/0.4; SMR 532; 95% CI 58 to 1992) for manufacturing workers. Manufacturing workers showed an increased standardised rate ratio (160, 95% CI 36 to 707) for lymphohaematopoietic cancer in comparison with office workers. There was increased SPCMR (225, 95% CI 108 to 413) for lymphohaematopoietic cancer.
ConclusionThe results support the relationship between exposure in the refinery/petrochemical industry and lymphohaematopoietic cancers. But the increased risks were insignificant due to the small number of lymphohaematopoietic cancers. A follow‐up study should be carried out to confirm the relationship in the future.
Key wordslymphohaematopoietic cancer; occupational cancer; cancer mortality
I. GusevaCanu1, G. Molina2, P. Collomb2, M. Goldberg3, P. Perez2, F. Paquet1, A. Acker2, M. Tirmarche1. 1Institute of Radioprotection and Nuclear Safety; 2AREVA NC, 3INSERM
ObjectivesDespite numerous publications on nuclear cycle workers, epidemiological results currently available provide limited evidence for an association between occupational exposure to internal radiation from uranium and cancer risk. Difficulties in the assessment of past exposures and potential confounders constitute the most important limitations to investigating this thoroughly. A pilot study was carried out at the AREVA NC Pierrelatte uranium conversion plant to develop a methodology for a comprehensive reconstruction of occupational exposure to internal radiation when few exposure measurement data are available for the risk assessment.
MethodsA period‐specific job exposure matrix (JEM) was designed. Job characteristics, including function, task and facility, according to time‐periods were used to define homogeneous occupational categories. A method derived from the Delphi technique was used to quantify occupational exposure.
ResultsThroughout the plant period 1964–2006, 316 “job‐periods” and 22 agents groups were defined. Six of them involved uranium compounds classified by their blood‐transferability and toxicity characteristics. Three hundred retired workers and 100 active workers were included as experts in the exposure assessment committee. A quantitative assignment of quantity and frequency of handling (both coded from 0 to 3) was performed for each agent group. A first validation of the JEM by experts in radioprotection and industrial hygiene showed an acceptable internal consistency.
ConclusionIn the context of missing past exposure measurement data, the plant‐ and period‐specific job exposure matrices may be considered as a valid alternative for exposure estimation. Linkage of the JEM with individual work histories allows computation of the cumulative value of exposure for each worker for future epidemiological studies.
Key wordsnuclear workers; uranium; job exposure matrix
C. Berriault1, N. Lightfoot1, M. Conlon2, R. Bissett3, B. Gottfred4. 1Epidemiology Research Unit, Regional Cancer Program, Sudbury Regional Hospital; 2Research Department, Regional Cancer Program, Sudbury Regional Hospital; 3Radiation Treatment, Regional Cancer Program, Sudbury Regional Hospital; 4Master of Public Health Program, Lakehead University
ObjectivesXstrata remains among the world's top producers of copper and nickel, with a large cohort of nickel workers employed in the Province of Ontario, Canada at Sudbury in northeastern Ontario. Operating since 1929, Sudbury operations have historically included underground working, milling, and smelting related to the production of nickel, copper, cobalt and other precious metals. Objectives for the study include: (1) examination of mortality for various diseases and cancer incidence in this occupational cohort; and (2) description of the challenges and benefits of undertaking an occupational cohort study in Canada and within a joint health and safety committee environment.
MethodsThis presentation will describe an occupational cohort study of mortality (1964–2001) and cancer incidence (1964–2001) conducted for Xstrata Nickel's Ontario male nickel workers (n=10253 males). The cohort was originally assembled by researchers at McMaster University, with responsibility transferred to the Northeastern Ontario Regional Cancer Centre of the Sudbury Regional Hospital in the late 1990s. Company nominal role and work history information were linked to provincial health data at Cancer Care Ontario. Given minimal out‐of‐province migration, using the Ontario population as a reference, standardised mortality ratios for a wide variety of diseases and standardised cancer incidence ratios were calculated overall, by work area, and by work area and duration of employment.
ResultsIn the cohort of male workers, there were 1984 (19.35%) deaths and 1127 (10.99%) incident cancers. A variety of SIR and SMR results will be provided.
ConclusionThis presentation will discuss the challenges and benefits faced by the study team (1) in undertaking this type of epidemiological research in light of recent developments in privacy legislation, (2) within a joint company‐union environment, and (3) related to the conduct of occupational cohort studies in Canada, as well as describing the impact and implications of the study results for occupational health and safety.
Key wordsoccupational; cohort; epidemiology
C. F. Robinson, J. P. Sestito, J. Wood, J. T. Walker. The National Institute for Occupational Safety and Health (NIOSH)
ObjectivesThe US manufacturing industry employed 16253000 workers in 2005. The objective of this study was to assess occupational risks in the US manufacturing industry. There are a limited number of studies of cohorts from specific manufacturing process settings. Our aim was to identify larger manufacturing sub‐sectors with elevated proportionate mortality due to cancer and occupational disease and to focus attention on preventable occupational disease.
MethodsWe undertook a proportionate mortality analysis of 1161405 white male manufacturing workers who died between 1984 and 1998 in 28 US states that code industry and occupation on death certificates. Underlying cause PMRs for cancer, heart disease, and other chronic disease and injury are presented. Multiple cause respiratory disease mortality among men who died in 23 states in 1990–1999 is described. Disease and injury rates from the US Bureau of Labor Statistics are reviewed in the context of elevated mortality in manufacturing workers.
ResultsMen younger than age 65 experienced significantly elevated proportionate mortality for cancer, asbestosis, silicosis, heart disease, stroke, industrial accidents, fall, and transport fatalities. The larger sub‐sectors accounted for many excess deaths during the 15‐year interval. Findings varied from sector to sector: gallbladder cancer (PMR 128) and cardiomyopathy (PMR 119) in the food manufacturing industry; malignant melanoma (PMR 140) and amytrophic lateral sclerosis (PMR 129) in the chemical manufacturing sector; leukaemia (PMR 167) and Alzheimer's disease (PMR 155) in the electronic computing equipment manufacturing industry; and stomach cancer (PMR 109) and other ischaemic heart disease (PMR 106) in the transportation equipment manufacturing industry.
ConclusionMen under age 65 usually employed in US manufacturing sectors experienced significant excess mortality. Despite lifestyle differences and other limitations of the study, the large numbers of excess deaths indicate the need for additional research and preventive action for manufacturing workers.
Key wordsmanufacturing industries; occupational disease; occupational disease surveillance
C. Brooks1, L. Linsell1, T. J. Keegan1, T. Langdon1, V. Beral1, P. Doyle2, T. Fletcher2, N. Maconochie2, M. J. Nieuwenhuijsen3, L. M. Carpenter1, K. M. Venables1. 1University of Oxford; 2London School of Hygiene and Tropical Medicine; 3Imperial College, London
ObjectivesResearch into the effects of chemical warfare agents on military capability at Porton Down started in 1916. We have assembled a cohort of military veterans who attended Porton Down in 1941–89 to study their long‐term health. At EPICOH 2005, we reported early progress with data assembly. We now report on data assembled for the complete cohort.
MethodsThe cohort includes two groups: military personnel who participated in the human volunteer programme (“PD veterans”) and similar military personnel who did not (“non‐PD veterans”). Non‐PD veterans were selected using the service number of the PD veteran to achieve comparability in key variables (eg, sex, branch of the military, military rank, and calendar period of military service). Exposure data were abstracted for PD veterans and vital status to the end of 2004 determined for both groups. The primary source for outcome data was the NHS central register, with additional information on deaths from the Commonwealth War Graves Commission and the Department for Work and Pensions.
ResultsUp to March 2007, personnel files had been located in the military archives for 18447 PD veterans, over 90% of those eligible for inclusion, and 18097 non‐PD veterans. The two groups were similar in age and other key military and demographic variables. However, the PD veterans had a longer duration of military service than the non‐PD veterans. The number of women was too few for a meaningful analysis. By the end of 2004, 42% of the cohort had died, 52% were presumed alive, and 6% were lost to follow‐up. 7% of deaths had occurred, or were assumed to have occurred, abroad.
ConclusionRetrieval of military personnel files was good, as was linkage to national mortality registers. Additional deaths could be determined from other sources. Deaths occurring abroad may reflect the military occupation or the emigration pattern from the UK after the Second World War. Although well‐matched on many key variables, the difference between the two groups in duration of military service may reflect underlying differences in health‐related variables and may require further consideration in the analysis.
Key wordschemical warfare agents; veterans; mortality