Occupational Risk for Oral Cancer in Nordic Countries

Discussion

This study compared the risk of cancer of the mouth and pharynx in different occupations before and after adjustment for tobacco and alcohol use. In most cases, high risk (SIR over 1.5) decreased and low risk (SIR under 0.67) increased after adjustment as expected, because alcohol and tobacco are well-known risk factors for cancer of the mouth and pharynx (11, 13, 14). However, there were some occupations that showed high risk in different cancer sites after adjustment. These findings need further discussion.

An unexpected finding is the high and statistically significant RR of tongue cancer after adjustment for tobacco and alcohol use in male dentists. A similar finding can also be seen in tongue and pharyngeal cancer for female dentists, although not statistically significant. This high RR for tongue cancer could be an indication of occupational risk factors such as exposure to human papillomavirus (HPV) infection (15, 16) due to burr-generated aerosol and salivary contamination. It has been demonstrated that cancer-associated HPVs are especially resistant to surface active disinfectants (17). In addition, chemical compounds including organic solvents and plastic monomers (18) that are commonly used in dentistry, or a higher than average use of alcohol-based mouth rinse (19) may increase the risk for oral cancer in dentists.

Artistic workers had high RRs for tongue cancer in males and for pharyngeal cancer in both genders. Female artistic workers' RRs for tongue and oral cavity cancer were also elevated but the results were not statistically significant. Artistic workers are quite a heterogeneous group and include musicians, painters etc. and it is hard to find occupational risk factors that would be common to the majority of artistic occupations. An irregular lifestyle could be one uniting feature for most artistic workers and this could possibly lead to poor oral hygiene, known to be an important risk factor for oral cancer (20). This hypothesis needs further investigation because there is no evidence of poor oral hygiene in artistic workers.

Journalists had high RR, after adjustment for alcohol and tobacco use, in males for tongue and pharyngeal cancer and in females for oral cavity cancer. The RR was aIso high for pharyngeal cancer in females but was not statistically significant. Female journalists had exceptionally high mortality from liver cirrhosis and slightly elevated risk for lung cancer, indicating heavy drinking and moderate tobacco consumption. Male journalists instead had a low lung cancer incidence, moderately elevated mortality from liver cirrhosis and liver cancer incidence, indicating low consumption of tobacco and moderate consumption of alcohol. It is difficult to imagine any specific risk factors for cancer in journalists' work, but they might also have an irregular lifestyle like artistic workers.

Both male and female waiters had elevated RRs for oral cavity cancer and male waiters also for pharyngeal cancer. Irregularities in lifestyle and daily habits due to variable working hours may provide one explanation. The burden of liver diseases among both male and female waiters was found to be up to four-fold greater as compared to the general population and higher than for any other profession (Table V). This indicates significantly high alcohol consumption, which is a risk factor for these cancer types and liver diseases. It is possible that the adjustment method used in our analysis does not work as well with such outlier values as it works with the modest range of variation, and hence we should interpret the adjusted RRs for waiters with certain reservations.

Male hairdressers had a high RR for tongue cancer. Moreover, female hairdressers had a slightly elevated RR for tongue cancer, but this was not statistically significant. It is possible that an elevated RR for tongue cancer after adjustment for tobacco and alcohol use in hairdressers is due to occupational risk factors because there is evidence of occupation-related risk excess among hairdressers for other cancer types. For example, the risk of bladder cancer was elevated at least for hairdressers working before the 1980s when aromatic amines were used in hair dyes (21). Risk for laryngeal cancer, lung cancer and multiple myeloma were also elevated according to some studies (22). Hairdressers are exposed to many chemicals in hair products and the International Agency for Research on Cancer (IARC) has classified “occupational exposures of hairdressers and barbers” as probably carcinogenic to humans (23).

Male cooks and stewards had a high RR for oral cavity cancer and slightly elevated RR for tongue and pharyngeal cancer. Female cooks and stewards did not have statistically elevated RRs for any cancer of the mouth and pharynx. Cooks have been demonstrated as having higher risk of oral cancer. A possible explanation for elevated risk can be exposure to cooking oil fumes (24). In addition, continuous exposure to fermentable starch and sugars during tasting is a known risk factor for caries, which can lead to elevated acetaldehyde levels in the oral cavity. The IARC recently stated that consumption of very hot beverages increases the risk for esophageal cancer (25). Whether continuous tasting of hot foods also increases the risk for oral cancer remains to be investigated.

Seamen had on elevated RR for oral cavity cancer, and their RR for tongue and pharyngeal cancer was moderately elevated. At the same time, fishermen had a decreased RR, indicating that the working environment at sea is not an explanation, but instead the explanation for elevated risk among seamen may be an irregular lifestyle and numerous intimate relationships which could lead to higher HPV prevalence.

After adjustment for alcohol and tobacco use, male farmers had low RRs for tongue and pharynx cancer. The RR for oral cavity cancer was also quite low in this group. Farmers live a very regular life and probably eat vegetables and fruits more than average. In Finland, only office workers eat more fresh vegetables than farmers (26). High consumption of vegetables and fruits reduces the risk of oral and pharyngeal cancer (27).

Male forestry workers had a low RR for pharyngeal cancer and quite low RR for tongue and oral cavity cancer after adjustment for tobacco and alcohol use. There is no evidence of consumption of healthier food or other protective agents that could explain this finding, hence it needs further study.

Male chimney sweeps and miners had low RRs for oral and pharyngeal cancer after adjustment for tobacco and alcohol use. It is hard to believe that consumption of protective agents such as fruits and vegetables play a role in these occupations. However, there seem to be some occupational risk factors that cause lung cancer, other than smoking. Exposure to soot among chimney sweeps (28), radon, dust exposure (29) and diesel exhaust among miners (30) are possible risk factors for lung cancer. Accumulation of multiple risk factors for lung cancer can cause over-adjustment of the SIR for oral and pharyngeal cancer. Therefore in these professions, a low RR for oral and pharyngeal cancer is possibly a statistical artefact.

The low RR for tongue cancer for female launderers might be explained by a similar mechanism as above; higher rates of lung cancer due to occupational exposure to solvent vapors can cause over-adjustment of the SIR for oral and pharyngeal cancer (31). Similar elevated occupational risk for lung cancer may also explain shoe and leather workers' low RR of oral cavity cancer (32) since their working conditions or socio-economic status presumably does not significantly differ from those of clothing industry workers whose risk for tongue cancer is not changed.

The strengths of this study include the large number of cases during the 45 years of follow-up, reliable diagnoses with virtually full coverage of cancer cases (33, 34), and the high accuracy of occupational codes (12). The unique personal identity codes in all Nordic countries used as keys in linkage between the census data, mortality and emigration data, and cancer registries guarantee accurate linkages. Occupational information was based on national censuses from 1960 to 1990. Validity studies indicate that the classification by occupation in the Nordic censuses is reasonably accurate (12). The extent to which the first census occupation reflects the lifetime experience varies considerably across occupational categories. Occupational stability is considered most stable in the early decades of follow-up and it is higher among men than women and highest in occupational categories where a long education is required, such as for physicians, dentists and teachers (35, 36). A limitation of this study is the absence of individual data on risk behavior factors such as smoking and consumption of alcohol, as well as HPV infection or oral hygiene. The adjustment method using lung cancer and liver disease mortality and liver cancer incidence seems to work adequately for most occupations since the RRs are closer to 1.00 than SIRs. However, we cannot rule-out the possibility of residual confounding by risk factors such as smoking and alcohol, that could explain risk in some professional categories particularly exposed to those products.

The main result of this study is that after adjustment for alcohol and tobacco use, the only individuals that had elevated RRs were dentists, artistic workers, hairdressers, journalists, cooks and stewards, seamen and waiters. For dentists, we speculate that occupational exposure to HPV or chemical compounds including organic solvents and plastic monomers may have a greater role than expected. Identification of specific occupational risk factors for these professions requires separate in-depth studies.