Exposure to cutting oils and its relation to skin tumors and premalignant skin lesions on the hands and forearms.

P. Exposureto cutting oils and its relation to skin tumorsand premalignantskin lesionson the hands and forearms. Scand J Work Environ Health II (1985) 365-369. In a cohortstudythe incidenceof skin tumorson thehandsand forearmsof 682turnersexposed to cutting oilswas compared to theincidence of thegeneralmalepopulationand of officeworkers. Among the 682turners, five premalignant squamous celltumors and eightkeratoacanthomaswere found between 1960 and 1980. Fiveof this total of 13 turners withprimary skin tumors had another skin tumor as well. Among the 682men there were four casesof scrotal cancer. Among the 375office workers examined, no malignant or premalignant tumors or keratoacanthomas werefound on the hands or forearms. No primary skin tumors werefound among the turners after 1975, when the acid-refined mineraloils were replacedbysolvent-refinedoilscontaininga far lower concentration of polyaromatichydrocarbons, which are probably the cancerogenic agent in the oils. The study showsa greatlyincreased risk of skin tumors on the hands and forearms of workersexposed to acid-refinedmineral oils. Individual susceptibility also seems to be indicated.

Long-term exposure to mineral oil is a well-known cau se of skin cancer on the scrotum (14). How ever , the association between long-term exposure to cutting oils and skin tumors on other sites is far less clearly established. Cruickshank & Gourevitch (4) described 44 cases of skin cancer on the hands and forearms of men in the Birmingham area between 1941 and 1950. Thirty-seven of these individuals were available for interviews, and 18 reported a history of oil exposure. Mastromatteo (12) described five cases of skin cancer on the hands and forearms of workers in a mechanical engineering plant in 1955. Compared to national incidence rates , this ob servation suggested a possible relationship between the occurrence of squamous cell carcinoma and long occupation as a machine operator. Studies of workers exposed to oil mist have not revealed an increased incidence of skin cancer on the hands or forearms (5,6,8).
Mortality from squamous cell skin tumors is low.
Cruickshank & Gourevitch (4) have estimated the case mortality from squamous cell carcinomas on the hands and forearms to be 12 010. It is therefore difficult to detect an increased incidence of squamous cell skin carcinoma in mortality studies. Some of the tumors may even be treated as premalignant lesions, and subsequent morbidity studies based on cancer registrations ma y thus underestimate the risk . Ghadially et al (7) reviewed 238 cases of keratoacantoma in Sheffield between 1954 and 1959. They could not find any evidence of an association between oil exposure and keratoacanthomas, although they did find an association between this disease and exposure to soot and tar.
The objective of our study was to investigate the risk for malignant and premalignant skin lesions and keratoacantomas on the hands and forearms of workers exposed to cutting oils. The study was designed as a cohort investigation in which the incidence of these lesions among workers exposed to cutti ng oil is compared to that of office workers and that of the total male population of the same city.

Subjects and methods
The study was performed in an industry producing bearing rings . The company was founded in 1907 and today employs about 2 500 blue-collar workers and 1 000 white-collar workers. It is situated in Goteborg, Sweden, a city with about 500 000 inhabitants.
In the 1940s and 1950s a few cases of skin cancer on the scrotum were disco vered among workers in the lathe departments. In 1958 a dermatologist was hired to examine the turners regularly in order to detect scrotal cancer at an early stage. This screening was performed three or four times a year. The same dermatologist (KF) has made all the examinations since 1958. Almost every turner employed in that department participated in the examinations , which were made during the workday. The results were noted on special lists that also contained information for each indi vidual as to the occurrence of oil acne on the forearms and the number of years each person had worked in the lathe department. Between 1 January 1960and 31 December 1980 these lists contained the names of 685 workers (excluding foremen), 682 of whom could be identified.
Thi s study was designed as a cohort stud y in which the incidences of skin lesions on men exposed to oil and those of reference groups are compared. The 682 lathe operators constituted the exposed group. Their total exposure time, that is, their exposure time through 31 December 1980, is described in table 1.
The national cancer register was studied to fmd cases of malignant skin tumors among the turners. Th is incidence was compared to that of the male population of the same city. The expected incidence was calculated by multiplying the gender-specific incidence rates by the number of person-years; stratifying for calendar year, age (5-year age groups), and expo sure time or latency period; and summarizing across the cohort. The latency period was defined as the time having elapsed since the date of the individual' s employment in the turning department. Person-years were calculated through 31 December 1980 or to the date of death or emigration, if it occurred before 1980. Th e incidence rates for the city of Goteoorg were provided by the Cancer Registry.
There is no official register of premalignant skin lesions or keratoacanthomas. We used health records kept by the health care department of the company to establish the incidence among the turners. The expected incidences were estimated from interviews of office workers and examinations of statistics from the register of the Department of Pathology in Goteborg, The different procedures used for this process follow .
This industry has had a health care department since its establishment. Health records could be located for 544 of the 682 men. There were no compulsory examinations at the health care department, and the other 138 men had probably never used its services. No health records have been destroyed. One hundred and twenty (87 0/0) of the 138 men had exposure times of less than 10 years. The health records were reviewed for the occurrence of skin lesion s, interest being focused on malignant skin tumors, premalignant skin lesions (eg, Bowen's disease , actinic keratosis, dys-plasia), and keratoacanthoma. Onl y biopsied skin tumors verified by a pathologist's examination were accepted.
In all, 400 office workers in the same industr y were selected as a reference group. They were a random sample of all office workers employed by the company in 1983, in the lower range of the salary scale . Office workers are classified by salary on a scale from 1 through 8, where 1 is the highe st level. To obtain reasonably similar economic back grounds among the men in the exposed and reference groups, only men from levels 4 through 8 were included. A questionnaire was sent to the 400 office workers. It contained the question "Have you had an y skin tumors?" and, if the answer was yes, a request to state the location of the tumor, date of treatment, and the physician's name and address. Three hundred and seventy-five of the office workers answered the questionnaire (94 %).
The Department of Pathology of the University of Goteborg has a computerized register of diagnoses made since late 1982. The Department makes more than 95 % of the pathology examinations in the region. All cases of premalignant skin tumors and keratoacanthomas diagnosed in 1983 were retrieved from this register. The diagnoses are classified according to the SNOMed index (3) and the following tumors were con sidered (SNOMed number in parentheses): atypia UNS (M69700), senile keratosis (actinic keratosis; M72850), keratoacanthoma (M72860), dysplasia (M74000), cancer in situ (M80702).
The differences between the observed and expected incidences were tested according to a Poisson distribution. The figures for the p-values describe two-tailed tests (1).

Exposure
Straight mineral oils ha ve alwa ys been used in the turning department. Until 1975 acid-refined oils were used. Then there was a change to solvent-refined oils. Sulfur has been added as an extreme pressure agent in concentrations of less than 1 %. Until 1950 about 10 % animal or vegetable oils was added.
There are always difficulties involved in measuring exposure to the skin . The skin is mostly exposed through splashing and, to a far smaller extent, through oil mist. The level of oil mist in this plant has earlier been assessed and was about 5 mg/rn" or more before 1965. It is about 2 mg/rn! today. Oil acne may occur if the skin is heavily exposed to mineral oils. We have used the prevalence of oil acne on the forearm to measure exposure to the skin. The occurrence of oil acne was registered twice to four times a year by the same physician (KF) at the medical examinations . In the early 1960s a process was developed for removing the pieces of metal from the lathes mechanically instead of manually. This occurrence did not particularly influence the prevalence of oil acne. The change made in the turning method in the late 1960s, however, reduced the prevalence dramatically. With this new method, the piece of metal was turned in one operation, which greatly reduced the need for toolsetting. The turning machines were unchanged throughout the 1960s and 1970s, and there have only been minor changes in the splash guards. However, the splash guards could be used much more effectively once there was less need of tool -setting, beginning in the late 1960s. The change to solvent-refined oils and the introduction of workclothes supplied free of charge (ie, T-shirts and trousers) also coincides with a reduction in the prevalence of oil acne. The washing facilities have rema ined unaltered throughout this time period.

Results
Five of the 682 exposed men were reported as having malignant skin tumors in the cancer register between 1960and 1980(exclusive of basaliomas). All cases were squamous cell carcinoma (no melanoma). Four were situated on the scrotum and one on the face. The exposure time of these men was 13, 19,33 , 39, and 13 years, respectively. Compared to the male population of the same city 0.3 cases were expected (melanomas excluded) among men with at least five years of exposure time (p < 0.001). A total of 24 malignant tumors were reported to the registry between 1960 and 1980. One man had two tumors at different sites. Twentytwo of these cases occurred after a laten cy period of at least 10 years, the expected incidence was 29.2 on the basis of data on males in the same city as a reference group and an allowance for at least a lo-year latency period (5 901 person-years) (not significant).
Except for cases of skin cancer the sites varied greatly, with just one to three cases at each site. This finding means that a separate analysis by site would not be reward ing.
In the health records there were 13 patients with primary skin lesions on the hands and forearms , eight cases of keratoacanthoma and five cases of premalignant lesions (table 2). All were verified through biopsies. Two more keratoacanthomas on the hands were mentioned in the records , but there were no examined biopsies. The saved biopsies were reviewed in 1983.
All the cases had rather long exposure times before the occurrence of the lesions . In five cases secondary primary skin lesions also developed, two of which were cases of malignant squamous cell carcinoma on the scrotum. One of the 13 patients also developed a lung tumor.
Five cases of premalignant skin tumors or keratoacanthomas situated on the hands or forearms of males 20-69 years of age were registered in 1983in.the computerized register of the Department of Pathology. The cases in the register occurred in a population of about 156000 males 20-69 years of age. The number of cases was too few to make a reliable estimate of the expected number for the turners exposed to oil, but it is obvious that there is an increased incidence of these lesions among the workers exposed to oil. No premalignant skin lesions or keratoacanthomas situated on the hands or forearm s were reported among the office workers.

Discussion
When we began this study, we thought it might be the case that skin lesions on the hands or forearms were treated at a premalignant stage, and this hypothesis has been confirmed. There were five premalignant lesions on the hand s or forearms among the turners, but none were registered in the cancer register. (It is not compulsory to report premalignant lesions to the Swedish Cancer Registry.) This finding might explain why the association between tumors on these sites and mineral oil exposure is difficult to establish . The 13 cases of premalignant skin lesions or keratoacanthomas found in this study are only a minimum finding. Some persons may have visited other doctors for skin lesions and consequently not have been revealed in this study. However , the treatment at the company was free of charge, and the workers interviewed believed that almost all workers went to the company health care department if they had a skin lesion.
A comparison with the data from the computer register of the Department of Pathology makes it evident that there was an increased incidence of skin lesions on the hands and forearms of men exposed to oil. A possible bias from changes in diagnosing practices was excluded by a reexamination of the biopsies. The office workers reported no premalignant skin lesion or keratoacanthomas on their hands or forearms. Another possible source of error might be underreporting among these men. However, we reviewed all records where any skin desease on the questionnaire was mentioned (N = 69). Although there might ha ve been some underreporting, it would hardly account for most of the difference. Among the office workers skin lesions not treated by the health care department could have been reported in the questionnaires. This possibility would mean that the risk among office workers compared to men exposed to oil was overestimated. Another possible bias is the health surveillance of workers exposed to oil. Office workers may not have sought treatment for their premalignant skin lesions or keratoacanthomas, but this is improbable. We think that most people with easy access to medical care will see a doctor if they have such a lesion.
Thus there was found to be a greatly increased risk of premalignant skin lesions and keratoacanthomas 368 among workers with long exposure times to oil. At least some 15 years of exposure seem to be necessary for these lesions to develop (table 2). Thirteen (8 %) of the 155men with at least 15 years of exposure had such lesions.
No primary cases were found after 1975, a date which coincides with a sharp decrease in oil acne and a change to solvent-refined mineral oils that took place at that time. Solvent refining reduces the content of polycyclic aromatic hydrocarbons (PAH), which are probably the major carcinogenic compounds in the oil. It is well known from animal experiments that the carcinogenic potency of acid-refined oils is great compared to that of solvent-refined oils (2,9). These obser vations agree well with our findings.
Toda y, solvent-refined mineral oils are used in most cutting oils. New PAH may be formed because of high temperatures and pressures but not nearly to as great an extent in cutting oils as in other types of oil. In motor oils or quenching oils there may be far more than a tenfold increase of PAH (13). In cutting oils, the increase of benzo(a)pyrene is about twofold (II,13).
In 1975 workclothes began to be supplied at the plant free of charge, and this procedure may have decreased exposure to cutting oils. Instead of the long sleeveshirts which were used by some of the workers, short sleeve T-shirts began to be used. This change facilitated the cleaning of the forearms . The use of highly refined (solvent-refined) mineral oils and good work practices may therefore considerably decrease the risk of premalignant skin lesions and keratoacanthomas.
Some individuals seem to be especially prone to oilinduced skin lesions, as 6 of the 13 men with primary skin lesions on the hands or forearms also had other lesions, generally elsewhere on the skin. Some of these lesions developed after the change to solvent-refined or identical lathe machines, and we do not consider differences in exposure to be a completely satisfactory explanation for the much higher risk for some individuals. Mastromatteo (12) found that three of five men who were exposed to oil and had squam o us cell carcinoma on the hands or forearms developed other tumors. The theory of increased susceptibility of some individuals is also supported by the findings of Kipling & Waldron (10) . They found an increased incidence of tumors of the skin, lung, and stomach in 282 cases of primary scrotal cancer, as compared to the general population. However, no increased incidence of lung cancer has been found in recently published studies of workers exposed to oil mist (5,8). The cause of this individual susceptibility is unknown, but differences in the ability to induce aryl hydrocarbon hydroxyla se has been suggested, although it has not been proved in man (10). Differences in the ability to carry out the repair of deoxyribonucleic acid (DNA) ma y be an alternative explanation. This susceptibility may be primary or acquired before, during, or after the first skin lesion. As there is a strong indication of individ ual susceptibility, a person exposed to oil who has a malignant or a premalignant skin lesion or a keratoachantoma sho uld receive a recommendation to change to a job not requiring exposure to oil.
In summary, this study shows that workers exposed to oil run an increased risk of developing premalignant skin lesions and keratoachantomas on the hands and forearms and that there seems to be individual susceptibility to develop such lesions.