Airborne exposures and risk of gastric cancer: A prospective cohort study
Article first published online: 31 JAN 2007
Copyright © 2007 Wiley-Liss, Inc.
International Journal of Cancer
Volume 120, Issue 9, pages 2013–2018, 1 May 2007
How to Cite
Sjödahl, K., Jansson, C., Bergdahl, I. A., Adami, J., Boffetta, P. and Lagergren, J. (2007), Airborne exposures and risk of gastric cancer: A prospective cohort study. Int. J. Cancer, 120: 2013–2018. doi: 10.1002/ijc.22566
- Issue published online: 28 FEB 2007
- Article first published online: 31 JAN 2007
- Manuscript Accepted: 28 NOV 2006
- Manuscript Received: 13 SEP 2006
- Swedish Cancer Society
- risk factor
There is an unexplained male predominance among patients with gastric cancer, and many carcinogens are found in male-dominated dusty occupations. However, the relation between occupational exposures and risk of gastric cancer remains unclear. To investigate whether airborne occupational exposures might influence the risk of noncardia gastric cancer, we used a large, prospective cohort study of male Swedish construction workers. These workers were, during the period 1971–1993, regularly invited to health examinations by a nationwide occupational health service organization. Data on job titles and other variables were collected through self-administered questionnaires and forms completed by the health organization's staff. Industrial hygienists assessed 12 specific airborne occupational exposures for 200 job titles. Gastric cancer, death or emigration occurring during follow-up in 1971–2002 were identified by linkage to the Swedish registers of Cancer, Causes of Death and Total Population, respectively. Incidence rate ratios (IRR) and 95% confidence intervals (CI), adjusted for attained age, tobacco smoking, calendar period and body mass, were derived from Cox regression. Among 256,357 cohort members, contributing 5,378,012 person-years at risk, 948 noncardia gastric cancers were identified. Increased risk of this tumor was found among workers exposed to cement dust (IRR 1.5 [95% CI 1.1–2.1]), quartz dust (IRR 1.3 [95% CI 1.0–1.7]) and diesel exhaust (IRR 1.4 [95% CI 1.1–1.9]). Dose-response relations were observed for these exposures. No consistent positive associations were found regarding exposure to asbestos, asphalt fumes, concrete dust, epoxy resins, isocyanates, metal fumes, mineral fibers, organic solvents or wood dust. In conclusion, this study provides some support to the hypothesis that specific airborne exposures increase the risk of noncardia gastric cancer. © 2007 Wiley-Liss, Inc.
Despite the declining incidence of noncardia gastric cancer in developed countries during the past decades,1, 2 this malignancy remains a major health concern globally, as it is the fourth most common cancer and the second leading cause of cancer-related death worldwide.3 Advances in diagnostic and therapeutic procedures have not had much influence on the poor prognosis for gastric cancer patients,3 stressing the urgent need for research that can identify preventable risk factors. Infection with the bacteria Helicobacter pylori (H. pylori) is a main known causative agent of noncardia gastric cancer.4, 5 Moreover, tobacco smoking6 and low socioeconomic status2 have consistently been associated with moderately increased risks, and some dietary factors have also been linked with this cancer.7, 8 The recent etiologic research on gastric cancer has focused on H. pylori, but this infection is difficult to prevent or broadly eradicate, and it cannot explain the male predominance of patients with gastric cancer (2–3 to 1).2 Thus, yet unidentified environmental risk factors might be of relevance. In many male-dominated industries, the exposure to carcinogenic agents is high, and the reduction of occupational chemical hazards in industrialized countries during recent decades9 seems to have coincided with the falling incidence of gastric cancer, indicating a possible etiologic role of occupational exposures. In line with this, several occupations and “dusty” work environments have been implicated in the etiology of this disease,10, 11, 12 but unfortunately, the scientific evidence regarding an association between dust exposure and risk of gastric cancer is insufficient and there is a need for further investigations of specific agents.10 We hypothesized that specific airborne exposures, which often occur in the construction industry, such as dust, fumes and solvents, could be inhaled and swallowed and have a direct harmful effect on the gastric mucosa. With the aim of clarifying the relation between such exposures and risk of gastric cancer in a male-dominated industry, we used prospectively collected data for a large cohort of Swedish construction workers.13
Diverging incidence trends, marked geographic variation and different risk factor profiles indicate that gastric cardia and noncardia cancer represent separate disease entities.2 The present study was therefore restricted to noncardia gastric cancer. Hence, when we use the term gastric cancer in the following, cardia cancer cases are excluded. The methods used has been presented in detail in our study addressing airborne occupational exposures and risk of esophageal and gastric cardia cancer.14 In brief, the Swedish Construction Workers Cohort consists of almost 400,000 employees within the Swedish construction industry who between 1971 and 1993 were regularly invited to attend health examinations by a nationwide occupational health service organization with almost complete coverage of the construction industry.13 The participation rate among the invited persons was high (85–90%). Information on job titles and other variables, notably tobacco smoking and anthropometric measures, was obtained prospectively through self-administered questionnaires and forms completed by specially trained nurses within the health service organization. Since 95% of the cohort members were men, no women were included in the current study. The National Registration Number, a unique personal identifier assigned to all Swedish residents, was used to identify each cohort member and to link each member to the nationwide Swedish Cancer Register. By this means, all incident cases of gastric cancer occurring during follow-up of the cohort, in 1971 through 2002, were identified. The Swedish Cancer Register codes gastric tumors (ICD-7: 1510, 1518 and 1519) with an overall completeness of 98%.15, 16 For complete follow-up and for correct censoring of persons in whom death or emigration precluded the risk of gastric cancer, each cohort member was also linked to the nationwide Swedish registers of Causes of Death and the Total Population.
Assessment of airborne occupational exposures
The exposure assessment was based on job titles as described in previous reports.13, 14, 17, 18 Only the job title at each worker's first health examination was used, as we lacked information for constructing lifetime occupational histories. Between 1971 and 1976, industrial hygienists assessed the exposure patterns within more than 200 occupations specific for the Swedish construction industry. Each of these occupations was studied at visits to ∼5 different sites in different geographical regions in Sweden. The job-exposure matrix regarding airborne exposures included 12 agents: asbestos, asphalt fumes, cement dust, concrete dust, diesel exhaust, epoxy resins, isocyanates, mineral fibers, metal fumes, organic solvents, quartz dust and wood dust. Each of these exposures was graded on an ordinal scale from 0 to 5, where level 3 corresponded to the Swedish threshold limit value at the time of the study. When no such limits were applicable, level 3 corresponded to an exposure level considered to be “acceptable” at that time. No specific quantitative meaning was assigned to the other grades. The exposure level scales were categorized into no exposure (0), moderate exposure (0.5–1) and high exposure (2–5). We also examined exposure to “combined” dust (defined as exposure to any of the following: asbestos, cement dust, concrete dust, mineral fibers, quartz dust or wood dust) and fumes (defined as diesel exhaust, asphalt fumes or metal fumes).
The cohort members were followed up from the date of their first health examination through December 31, 2002, the date of death, date of emigration or date of a primary gastric cancer diagnosis, whichever occurred first. Cox regression19 was used to estimate incidence rate ratios (IRR) and 95% confidence intervals (CI), using time since entry into the cohort as the underlying timescale. Models were estimated using the PHREG procedure in SAS.20 In multivariable models, adjustments were made for attained age (classified into 5-year age groups), calendar period at entry into the cohort (in 3 categories: 1971–1975, 1976–1980 and 1981–1993), tobacco smoking status at entry (in 3 categories: never, previous and current) and body mass index (BMI) at entry (in 4 categories: ≤21.9 [low], 22.0–24.9 [normal], 25.0–29.9 [overweight] and ≥30 [obese]). Individuals with missing data for any of the covariates included in the models were excluded from the analyses. The overall effect of each covariate was assessed by a Wald test of homogeneity across all exposure strata.
Study participants and incidence rates of gastric cancer
From the original cohort of 384,147 members, we excluded all women (n = 19,224) and men with (i) a diagnosis of gastric cancer before their first visit (n = 31), (ii) incorrect death dates (n = 28) or (iii) missing or insufficient information on job title, smoking status (mainly due to lack of recording of smoking status during 1975–1978) and/or BMI (n = 108,507). Hence, 256,357 men constituted the final study cohort. Together, these study participants contributed 5,378,012 persons-years at risk of developing gastric cancer during the follow-up period. In total, 948 incident cases of gastric cancer were identified. Some characteristics of the study participants are presented in Table I. The total incidence rate (IR) of gastric cancer was 17.6 per 100,000 person-years. The IR of gastric cancer was higher for participants who attended for their first health examination during the earliest years of inclusion into the cohort. The IR was increased among previous or current smokers, and among those who had a BMI above 25 at entry into the cohort.
|Characteristic||No. of subjects, N(%)||Person-years||Gastric Cancer|
|Attainted age, yr|
|Calendar period at entry into the cohort|
|Tobacco smoking status at entry into the cohort|
|Body mass index2at entry into the cohort|
|≤21.9 low weight||65,962 (26)||1,387,301||174||12.5|
|22.0–24.9 normal weight||100,734 (39)||2,133,156||358||16.8|
|25.0–29.9 overweight||77,857 (30)||1,627,372||361||22.2|
|≥30.0 obese||11,804 (5)||230,183||55||23.9|
Airborne occupational exposures and risk of gastric cancer
The relative risk estimates based on the adjusted models are presented in Table II. There were positive associations, seemingly with dose-response relationships between exposure to cement dust, quartz dust and diesel exhaust and risk of gastric cancer. Statistically, significantly increased risks of gastric cancer were found among workers highly exposed to cement dust (IRR 1.5 [95% CI 1.1–2.1]), quartz dust (IRR 1.3 [95% CI 1.0–1.7]), diesel exhaust (IRR 1.4 [95% CI 1.1–1.9]) and among workers exposed to “combined” fumes (IRR 1.2 [95% CI 1.1–1.4]). A negative association was observed between exposure to organic solvents and risk of gastric cancer (IRR 0.6 [95% CI 0.5–0.9]). No consistent associations were found between exposure to any of the other studied specific agents or “combined” dust exposure and risk of gastric cancer (Table II).
|Occupational exposure||No. of subjects n(%)||Person-years||Gastric cancer|
|All cases||IRR1 (95% CI)||p value2|
|No exposure||245,872 (96)||5,134,108||920||1.0 (reference)|
|Moderate exposure||6,971 (3)||160,794||21||0.8 (0.5–1.2)|
|High exposure||3,514 (1)||83,111||7||0.7 (0.3–1.4)||0.33|
|No exposure||251,626 (98)||5,276,100||934||1.0 (reference)|
|High exposure||4,731 (2)||101,912||14||0.9 (0.5–1.5)||0.64|
|No exposure||234,419 (91)||4,927,675||812||1.0 (reference)|
|Moderate exposure||18,550 (7)||374,195||99||1.1 (0.9–1.4)|
|High exposure||3,388 (1)||76,142||37||1.5 (1.1–2.1)||0.03|
|No exposure||159,661 (62)||3,340,495||586||1.0 (reference)|
|Moderate exposure||48,065 (19)||969, 562||154||1.0 (0.8–1.1)|
|High exposure||48,631 (19)||1,067,955||208||0.9 (0.8–1.1)||0.74|
|No exposure||222,720 (87)||4,660,435||758||1.0 (reference)|
|Moderate exposure||27,889 (11)||591,487||146||1.3 (1.1–1.6)|
|High exposure||5,748 (2)||126,089||44||1.4 (1.1–1.9)||<0.01|
|No exposure||254,000 (99)||5,328,298||939||1.0 (reference)|
|Moderate exposure||2,357 (1)||49,714||9||0.7 (0.4–1.4)||0.30|
|No exposure||240,068 (94)||5,088,919||903||1.0 (reference)|
|Moderate exposure||15,431 (6)||271,963||41||1.2 (0.8–1.6)|
|High exposure||858 (<1)||17,131||4||1.6 (0.6–4.2)||0.46|
|No exposure||232,107 (91)||4,863,320||867||1.0 (reference)|
|Moderate exposure||1,092 (<1)||19,738||3||0.7 (0.2–2.2)|
|High exposure||23,158 (9)||494,955||78||1.0 (0.8–1.3)||0.83|
|No exposure||237,113 (92)||4,965,554||887||1.0 (reference)|
|Moderate exposure||12,122 (5)||261,002||50||1.1 (0.9–1.5)|
|High exposure||7,122 (3)||151,457||11||0.6 (0.3–1.0)||0.12|
|No exposure||205,286 (80)||4,277,595||690||1.0 (reference)|
|Moderate exposure||42,165 (16)||904,552||200||1.2 (1.0–1.4)|
|High exposure||8,906 (3)||195,865||58||1.3 (1.0–1.7)||0.03|
|No exposure||228,915 (89)||4,804,594||885||1.0 (reference)|
|Moderate exposure||7,014 (3)||150,690||18||0.7 (0.4–1.1)|
|High exposure||20,428 (8)||422,728||45||0.6 (0.5–0.9)||<0.01|
|No exposure||239,004 (93)||5,030,241||892||1.0 (reference)|
|Moderate exposure||16,796 (7)||334,135||53||0.9 (0.7–1.2)|
|High exposure||557 (<1)||13,636||3||1.2 (0.4–3.6)||0.65|
|Unexposed||114,226 (45)||2,374,208||367||1.0 (reference)|
|Exposed||142,131 (55)||3,003,804||581||1.0 (0.9–1.3)||0.59|
|Unexposed||199,250 (78)||4,161,867||686||1.0 (reference)|
|Exposed||57,107 (22)||1,216,145||262||1.2 (1.1–1.4)||<0.01|
This study indicates positive, dose-dependent associations between exposure to cement dust, quartz dust and diesel exhaust and risk of gastric cancer. No such increased risk was detected among workers exposed to asbestos, asphalt fumes, concrete dust, epoxy resins, isocyanates, mineral fibers, metal fumes, organic solvents or wood dust.
Strengths and limitations of the study methods deserve some attention. The statistical power is good, by virtue of the large number of participants, recruited through an organization with almost complete coverage of employees in the Swedish construction industry in 1971 through 1993.13 Other advantages include the prospectively collected information regarding job titles and the unbiased expert exposure assessment, the long and complete follow-up, and the availability of information on potential confounding by tobacco smoking and other variables. Moreover, each individual's job title was linked to a job-exposure matrix developed by industrial hygienists, and thus specific agents could be analyzed. However, this type of job-exposure matrices has some limitations,21e.g. possible exposure misclassification. We were unable to study lifetime occupational histories, i.e. duration of exposure, and could only use the job title at the first visit as information regarding previous occupations was insufficient before 1986 and not collected after 1986. Moreover, the exposures were based solely on job titles, and not on each individual's unique exposure. However, in a previous study based on this cohort, it was found that among workers examined before 1986 few persons had changed their work tasks, and 96.3% had the same exposure level for both previous and current job title, indicating that the construction industry has a stable work force.22 Moreover, the job-exposure matrix was based on detailed expert assessments of exposure patterns. Another potential weakness is a lack of data concerning some potential confounders, including H. pylori infection. But any association between the studied exposures and H. pylori infection in this cohort is not likely to be strong enough to cause appreciable confounding, if any. Moreover, the high socioeconomic homogeneity of the cohort reduces potential confounding associated with such infection or with lifestyle factors. Finally, any “healthy worker effect” was avoided, since workers were internally compared.23
There have been reports on positive associations between various occupational groups and risk of gastric cancer,10, 12 notably “dusty” occupations, e.g., coal and tin mining,24, 25, 26, 27, 28, 29 metal processing,30, 31, 32, 33, 34, 35 rubber manufacturing36, 37, 38, 39, 40 and carpentry or construction work.10, 12 Other dusty work environments have also been implicated in the etiology.41, 42, 43, 44 However, possible occupational exposures linked to a risk of gastric cancer have not been established, since the majority of previous studies have not addressed specific exposures, have not adjusted for potential confounders, or have revealed only weak associations without dose-response patterns.10
Our finding of an increased risk of gastric cancer among workers exposed to cement dust is interesting. However, the highly exposed persons in our cohort mainly consisted of storage-workers who may be generally less fit than other workers, which would mean that confounding cannot be ruled out. These storage-workers were often handling “fresh” cement, carrying on their backs large sacks of it (unpublished information). Therefore, it is likely that the most highly cement exposed workers in this cohort have been exposed to fresh cement in warehouses, and not to concrete-related cement. Our results are supported by a study of Lithuanian cement masons,45 and a study of US cement-producing workers,46 while no clear associations were found in other studies of cement workers in Sweden or the US.47, 48 The positive association between quartz dust and risk of gastric cancer found in our study is in line with reports on workers exposed to silica dust in Canada,49, 50, 51 Spain28 and Japan.52 The current large and prospective cohort study design that revealed positive, dose-response associations regarding both cement dust and quartz dust provides valid support in favor of true relations between these exposures and gastric cancer.
Diesel exhaust contains several carcinogenic chemicals, such as polycyclic aromatic hydrocarbons.18 Increased risks of gastric cancer have been found in studies of lorry drivers in London53 and professional drivers in Geneva.54 Our study adds some evidence of a true link between diesel exhaust and the risk of gastric cancer.
Asbestos is a well-known human carcinogen, causing for example pleural mesothelioma and lung cancer,9 but with regard to gastric cancer previous reports have been contradictory. In a recent Norwegian study, a possibly increased risk of gastric cancer was found among lighthouse keepers who drank water highly contaminated with asbestos,55 and similar weak positive associations with gastric cancer have been reported from other studies of workers exposed to asbestos.56, 57, 58, 59, 60 In other investigations, however, no such association has been found.35, 61, 62, 63 The lack of association between exposure to asbestos and risk of gastric cancer in our study further argues against influence of airborne asbestos in the etiology of gastric cancer.
In an earlier follow-up of this cohort studying only concrete workers, and not specific exposures, a significantly increased risk of gastric cancer was detected compared to the general Swedish population (SIR 1.39 [95% CI 1.22–1.58]).64 Our study did not confirm these findings, however. Concrete workers are to some extent also exposed to cement dust and quartz dust, and the positive finding in the previous study might be explained by exposure to cement dust or quartz dust, and not concrete dust.
In some previous studies, positive associations have been found between metal-related work and risk of gastric cancer.10 One Swedish study showed an excess risk of gastric cancer among metal industry workers that seemed to increase with longer duration of employment.35 Exposure to metal fumes is low in the construction industry as compared to the metal industry, and no association between exposure to metal fumes and risk of gastric cancer was detected in our study.
Painters form the dominating occupational group exposed to organic solvents, and in a previous investigation the risk of gastric cancer by occupational groups in Sweden 1971–1989, a decreased risk was detected among painters.12 This is in line with the results of the current study. However, no association between specific exposure to organic solvents and risk of gastric cancer was found in a Swedish case-control study.35 Furthermore, a negative association between exposure to organic solvents and risk of gastric cancer might not be biologically plausible, and the finding might be due to chance. Alternatively, this decreased risk may reflect a possibility that workers heavily exposed to organic solvents are less likely to be exposed to other risk factors.
We hypothesized that a mechanism by which airborne particles might increase the risk of gastric cancer could be that inhaled dust and fume particles are swallowed and thereby act directly as carcinogens on the gastric mucosa. Particular agents such as cement dust and quartz dust could have an abrasive effect on the gastric mucosa, thus acting as irritants.11, 41 An inflammatory milieu can promote mitogenesis and lead to increased mutagenesis.10, 65 It is believed that excessive and continual formation of reactive oxygen species from inflammatory cells play a key role in the primary and secondary genotoxicity of fibrous and nonfibrous particles.66 Furthermore, harmful occupational exposures most likely interact with numerous nonoccupational risk factors at various stages of gastric cancer development. Studies have shown a close linkage between low acid output and an increased concentration of nitrate and N-nitroso compounds in gastric juice.67 In addition to its potentially abrasive effect on the gastric mucosa, a possible effect of swallowed cement dust, which is composed of chalk and clay, may be an increase in the intragastric pH. It has further been argued that dusts and fumes could potentially also act as carriers delivering other carcinogens to the stomach.11, 66 The findings in our study are not likely to be explained by such a mechanism; however, since not all of the studied particles were associated with an increased risk of gastric cancer. Indeed, our results suggest that the increased risk depends on exposure to specific types of dust particles rather than to dusty environments in general.
Although the overall burden of cancer caused by occupational exposures is probably limited,9 involuntary exposures encountered in the working environment could have a substantial impact on the cancer risk in specific subgroups of the population, mainly blue-collar workers. But the limited strength of the associations found in this study and the low exposure prevalence in the population at large indicate that these exposures should not materially influence the overall incidence of gastric cancer or explain the male predominance. Nevertheless, many occupational factors with an adverse effect on human health have been successfully controlled in industrialized countries, and future preventive measures should continue to target modifiable risk factors, including airborne potentially carcinogenic occupational exposures.
In conclusion, this large, prospective cohort study provides some support to the hypothesis that quartz dust, cement dust and diesel exhaust are moderate risk factors for gastric cancer. However, the studied exposures should not substantially influence the overall IR or the sex difference of this cancer, even if these associations are indeed causal. Nevertheless, preventive measures might reduce the mortality from gastric cancer among workers in highly exposed occupations.
We are grateful to all persons who designed and organized the “Bygghälsan” program, in particular Drs. Anders Englund and Göran Engholm. We thank Jenny Carlsson for administration and organization of the database.
- 5IARC Working Group on the Evaluation of Carcinogenic Risks to Humans. Schistosomes, liver flukes and Helicobacter pylori. IARC Monogr Eval Carcinog Risks Hum 1994; 61: 1–241.
- 6IARC. Tobacco smoking and tobacco smoke. In: IARC monographs on evaluation of carcinogenic risks to humans, vol. 83. Lyon, France: International Agency for Research on Cancer, 2002.
- 8WCRF&AICR. Food, nutrition and prevention of cancer: a global perspective. Washington, DC: American Institute for Cancer Research, 1997.
- 19Statistical methods in cancer research, vol. 2: The design and analysis of cohort studies. IARC Scientific Publication, 1987., .
- 20Changes and enhancements through release 6.11. Cary, NC: SAS Institute, 1996..
- 21Exposure surrogates: job-exposure matrices, self-reports, and expert evaluations. In: NieuwenhuijsenMJ, ed. Exposure assessment in occupational and environmental epidemiology. New York: Oxford University Press, 2003..
- 23Modern epidemiology, 2nd edn. Philiadelphia: Lippincott-Raven, 1998., .
- 30IARC. IARC monographs for the evaluation of carcinogenic risks of chemicals to humans, vol. 83: Polynuclear aromatic compounds, Part 3: industrial exposures in aluminum production, coal gasification, coke production and iron and steel founding. Lyon, France: International Agency for Research on Cancer, 1984.
- 31IARC. Overall evaluation of carcinogenicity: an updating of IARC monographs 1–42. In: IARC monographs on the evaluation of carcinogenic risks of chemicals to humans. Lyon, France: International Agency for Research on Cancer, 1987.
- 36IARC. The rubber industry. In:IARC monographs for the evaluation of carcinogenic risks of chemicals to humans, vol. 28. Lyon, France: International Agency for Research on Cancer, 1982.
- 49Silica and cancer associations from a multicancer occupational exposure case-referent study. IARC Sci Publ 1990; 97: 29–42., , , , , .