Occupation and risk of primary fallopian tube carcinoma in nordic countries

Page 1 of 21
International Journal of Cancer
Occupation and risk of primary Fallopian tube carcinoma in Nordic countries A Riska1, JI Martinsen2, K Kjaerheim2, E Lynge3, P Sparen4, L Tryggvadottir5, E Weiderpass2,4,6,7, E Pukkala8,9 1Department of Obstetrics and Gynaecology, Helsinki University Central Hospital, Helsinki, Finland 2Cancer Registry of Norway, Oslo, Norway 3University of Copenhagen, Denmark 4Department of Epidemiology and Biostatistics, Karolinska Institute, Stockholm, Sweden 5Icelandic Cancer Registry, Reykjavik, Iceland 6Samfundet Folkhälsan, Helsinki, Finland 7Department of Community Medicine, University of Tromsø, Tromso, Norway 8Finnish Cancer Registry, Institute for Statistical and Epidemiological Cancer Research, Helsinki, Finland 9School of Public Health, University of Tampere, Tampere, Finland John Wiley & Sons, Inc.
International Journal of Cancer
Page 2 of 21
Abstract
The aetiology of primary Fallopian tube carcinoma (PFTC) is poorly understood. Occupational exposures may contribute to PFTC risk. We studied incidence of PFTC in occupational categories in the Nordic female population aged 30-64 years during the 1960, 1970, 1980/1981 and/or 1990 censuses in Denmark, Finland, Iceland, Norway and Sweden. Standardized incidence ratios (SIRs) for the years following inclusion in the study up to 2005 were calculated for 53 occupations; the expected numbers of cases were based on PFTC incidence in the national populations. Altogether 2206 PFTC cases were detected during follow up via data linkages with the Nordic cancer registries. Significantly increased risks of PFTC were observed for smelting workers (SIR 3.99, 95% confidence interval 1.46- 8.68,Obs=6), artistic workers (2.64, 1.44-4.43,Obs=14), hairdressers (2.18, 1.41-3.22,Obs=25), packers (1.62, 1.11-2.29,Obs=32), nurses (1.49, 1.14- 1.92,Obs=60), shop workers (1.25, 1.07-1.46,Obs=159) and clerical workers (1.20, 1.07-1.35,Obs=271) and these sustained over times and different Nordic countries. There was a non-significant increased risk for PFTC among welders, printers, painters and chemical process workers. The risk was significantly and consistently low for women working in farming (0.68, 0.47-0.95,Obs=34) and among economically inactive women (0.88, 0.82- 0.94,Obs=833). The possible role of occupational exposures to the PFTC risks found in this study must be further evaluated in studies with a possibility to adjust for possible confounding factors, such as reproductive and life-style factors, which was not possible in our study. Key words: Primary Fallopian tube carcinoma; occupation; standardized incidence ratio; Nordic;
John Wiley & Sons, Inc.
Page 3 of 21
International Journal of Cancer
Introduction
Primary Fallopian tube carcinoma (PFTC) is a rare malignancy, comprising about 1% of all female genital tract cancers (1, 2). Studies from the United States indicated an incidence rate of 0.4/100,000 person-years in the 1980s; with the rate being higher in Caucasian women including Hispanics than in African Americans (3).The age-adjusted incidence of PFTC has been increasing during the recent decades in all Nordic countries, varying from 0.43/100,000 person-years in Sweden and Denmark to 0.65/100,000 person-years in Finland during 2003-2007 (Figure 1) (4). The 5-year survival rate has varied between 22% and 87% over times and regions (5-6). The aetiology of PFTC is largely unknown, although some risk factors have been identified: high parity decreases the risk, and previous sterilization may be protective (7-9), while chronic infections have not been confirmed as a predisposing factor for this carcinoma (11-12). In Finland, the incidence of PFTC is higher in urban areas than in rural, and almost twice as common among women in the higher social classes as compared to women in the lowest social classes (2). A Finnish study (2) suggested that in 1971-1995 women employed as hairdressers, nurses, private secretaries, book-keepers and accountants had elevated risks for PFTC, but the numbers of cases were small precluding any firm conclusions. The present study is an extension of the previous study on occupation and risk of PFTC (2), including data from all Nordic countries as well as more recent follow up. Subjects and Methods
The study cohort consists of women aged 30-64 years participating in the 1960, 1970, 1980/1981 and /or 1990 censuses in Denmark, Finland, Iceland, Norway and Sweden and still alive and living in the respective countries on January 1 in the year following the census. All persons were given unique personal identity codes in Sweden in 1947, in Iceland in 1953, in Norway in 1964, in Finland in 1967 John Wiley & Sons, Inc.
International Journal of Cancer
Page 4 of 21
and in Denmark in 1968. These codes were used to link census data with cancer data. National cancer registration started in 1943 in Denmark, in 1953 in Finland and Norway, in 1955 in Iceland, and in 1958 in Sweden. In Sweden and Norway, the computerized registration of census data by personal identity codes started in 1960, whereas in Denmark and Finland the first computerised data are available from 1970 census and in Iceland only for the year 1981. Census questionnaires included questions related to persons’ economic activity, occupation, and industry. All questionnaires were centrally coded and computerized in the national statistical offices. In Finland, Norway and Sweden, occupation was coded according to national adaptations of the International Standard Classification of Occupations (ISCO) (usually abbreviated to NYK; Nordisk Yrkesklassifikajon) from 1958, with more than 300 categories. Occupation was coded in Iceland according to a national adaptation of ISCO-68. In Denmark, occupation was coded according to a special national nomenclature with a distinction between self-employed persons, family workers, salaried employees, skilled workers, and unskilled workers; a total of 218 categories. Entities similar to those used in NYK were formed by combining these occupational codes with the 245 codes for industry. For the present study, the original national occupational codes were converted to a common classification with 53 relatively specific, but not too narrow, occupational categories, and an additional category of economically inactive persons (Table 1). Occupational classification in this study is based on the occupation recorded in the first census the person participated in the age range of 30-64 years. The female population that entered the cohort was followed-up for PFTC incidence during the period from January 1 of the year after the first available census through emigration, death, or to December 31 of the following years: in Denmark 2003, in Finland 2005, in Iceland 2004, in Norway 2003 and in Sweden 2005. The sources of data for death and emigration were the national population registers. John Wiley & Sons, Inc.
Page 5 of 21
International Journal of Cancer
The results are presented as standardized incidence ratios (SIRs) with the cancer incidence rates for the entire national study populations used as reference rates. SIR is counted as the ratio of observed (Obs) and expected numbers of cases. Exact 95% confidence intervals were defined for SIRs with less than 100 observed cases by assuming a Poisson distribution for the observed cases. When the number of observed cases was > 100, the confidence intervals were calculated based on a normal approximation to the Poisson distribution. Although SIR calculations were based on 5-year categories of both calendar periods and age, we have here combined the results to 15 year periods and broad age groups of 30-49, 50-64 and 65+ years. The study has been approved by the ethical committee and data inspections There were 7.5 million women included in the cohort; 3.4 million from Sweden, 1.7 million from Finland, 1.3 million from Norway, 1.0 million from Denmark and 60,000 from Iceland. The number of person-years accumulated by the end of follow up was 200 million. The mean length of follow-up was During the follow up, 2206 PFTC cases were detected: 884 in Sweden, 571 in Finland, 384 in Denmark, 358 in Norway and 9 in Iceland. The incidence of PFTC was highest among smelting workers (SIR 3.99, 95% CI 1.46-8.68), artistic workers (2.64, 1.44-4.43) and hairdressers (2.18, 1.41- 3.22). Significantly elevated risks for PFTC were also found among packers, nurses, shop workers and clerical workers (Table 1). Non-significantly elevated risks were detected among women working as welders, printers, painters and chemical process workers. Decreased SIRs were discovered among farmers (0.68, 0.47-0.95) and economically inactive women (0.88, 0.82-0.94). The SIR for PFTC sustained quite consistent over time periods among smelting workers, artistic workers, hairdressers, John Wiley & Sons, Inc.
International Journal of Cancer
Page 6 of 21
nurses, chemical process workers as also shop workers and clerical workers (Table 1). There were a slightly decreasing trend over time for the risk of PFTC among hairdressers, smelting workers, chemical process workers and shop and clerical workers but the differences between the period-specific rates were not statistically significant. The SIR among packers in 1991-2005 was significantly lower than before the 1990s. Among farmers the incidence of PFTC was lower than the population average during the period 1976-2005 and among the economically inactive women during the whole study period. We made stratification according to both age and period of diagnosis, but the numbers of cases were too small for identification of significant trends. We tried to evaluate the consistence of the risk for PFTC in different occupations across different Nordic countries (Table 2). The only occupations with consistent risk were hairdressers and nurses. Discussion
Primary Fallopian tube carcinoma is an uncommon cancer. We present the largest study published so far on occupational risk for PFTC. We found an increased risk for PFTC among smelting workers, artistic workers, hairdressers, nurses, printers, painters, chemical process workers, packers, shop workers and clerical workers. A decreased risk was observed for farmers and economically inactive The strengths of the present study are the relatively large number of cases – the largest study by now, with 45 years of follow-up, reliable diagnoses with virtually full coverage of cancer cases (13-17), and the high accuracy of the occupational codes (18). The linkage between the census data, the mortality and emigration data and the cancer incidence data was based on the unique personal identity codes used in all these registers in all Nordic countries. Since errors in the identifiers are rare (19), the method ensured a complete ascertainment of relevant events. John Wiley & Sons, Inc.
Page 7 of 21
International Journal of Cancer
In the present study, the information on occupations was based on national censuses from 1960-1990. As discussed in detail in Pukkala et al (18), the validity studies indicate that the classification by occupation in the Nordic censuses has been reasonably accurate, but that economic activity has been somewhat underestimated, especially among the women. Occupational classification in this study is based on the occupation recorded in the first census the person participated in the age range of 30-64 years. The extent to which the first census occupation reflects the lifetime experience varies considerably across occupational categories. The occupation at one point in time may not always correspond to the lifelong occupational history of a person. However, comparison with results of special occupational cancer studies indicates that the risk diluting effect of misclassification is small (20-22). The occupational stability proportion of individuals who had the same occupational category (of the 53 categories defined for the present study) in the first and second census available, i.e., 1960 and 1970 censuses in Norway and Sweden, and 1970 and 1980 censuses in Finland was estimated and is described in detail in Pukkala et al (18). In general, in that study stability was higher among men than women and highest in occupational categories where a long education is required such as physicians, dentists and teachers. Nurses were more likely to keep their occupation than assistant nurses, except in Sweden, where female assistant nurses also had a high rate of staying in the occupation. Hairdressers in all countries and of both genders tended to remain in the occupation. In the present analysis we were unable to take into account the role of potential confounding factors related to PFTC such as parity, breast feeding, sterilization history, previous use of contraceptive pills or hormone replacement therapy or family history of breast and ovarian cancer. Parity strongly protects against PFTC, and the protection increases with increasing number of deliveries (7, 9). Among women in occupations that require high education, mean age at first birth is higher than in women working in other occupational branches, and the total number of pregnancies and deliveries is probably smaller John Wiley & Sons, Inc.
International Journal of Cancer
Page 8 of 21
than, for example, among women working in occupations requiring shorted training periods, such as farming. In the Nordic countries the fertility rates in different socio-economic classes are quite alike. In studies from Norway, Finland, Denmark and Sweden, a similar pattern of lower fertility rates in occupations that require high education and a longer studying period has been detected (18, 23-25). The number of children was also lower among women with higher education and typically the age at first birth is higher. For example, in Norway 16% of physicians are nulliparous, whereas only 6.5% of farmers had no children. In Norway the age at first birth among physicians was 28 years of age and the number of children 2.1, whereas among farmers the corresponding numbers were 23 years of age and the number of children was 3.0 (18). The corresponding numbers for economically inactive women were 24 years of age and the number of children 2.5 (18). In a Finnish survey from 1985 including women aged 40-49, a similar fertility pattern was detected (23). In our study the risk for PFTC was not higher among women in occupations requiring high educations but a lowered risk among farmers was detected. This risk reduction may be in part explained by the effect of higher parity. In the study by Jordan et al. relationship with contraceptive use and breast feeding on the risk of PFTC was studied and a small inverse effect was seen (9). The only studies analyzing the risk of PFTC after a sterilization procedure suggest a non-significant risk reducing effect of this procedure (7,9). These confounding factors might have a minor effect on our results, but it is implausible that they would change our results. The only study analyzing the risk of PFTC after a sterilization procedure is from Finland and suggests a risk reducing effect of this procedure in univariate analysis, but the effect vanished in multivariate analysis (7). To evaluate the possible confounding effect of a previous sterilization procedure we did a search of the commonness of this procedure between the years 1996 and 2008 in different social classes in Finland (; www.stat.fi). The frequency of this procedure was more common in lower social classes. If an earlier sterilization would confound our results, the lowered risk of PFTC for farmers would be too low. As among hairdressers an earlier sterilization procedure has John Wiley & Sons, Inc.
Page 9 of 21
International Journal of Cancer
been more common, the real risk for PFTC would be even higher than was detected. The same effect would be among smelting workers. Among women in higher social classes the sterilization procedure has been rarer, thus the detected risk estimate for PFTC among nurses might have been too high. We were unable to take into account the confounding effect of family history of breast and ovarian cancer i.e. the effect of germline BRCA mutations among PFTC patients. A previous study of PFTC discovered that almost one third of PFTC patients have a BRCA germline mutation background, the highest proportions being among younger PFTC patients (26). It is not likely that the BRCA carriers would have been clustered in any specific occupational categories, and therefore we do not consider genetic factors to confound our results. Postmenopausal hormone therapy may be associated with an increase in the risk of PFTC (8, 9). The scarce data available on use of postmenopausal hormone therapy in the Nordic countries or other countries suggest that the selection between the population subcategories is nevertheless rather small and the variation between different educational levels has We observed an elevated PFTC risk for women working as smelters in particular in Finland and Sweden. Smelting workers may be exposed to nickel, aluminium, arsenic, lead, iron, cadmium or chromium. There are no previous studies of the risk of PFTC among smelting workers other than the one from Finland (2); the material of which is included in the present study, thus no comparison of our results with others is possible. An Italian study reported an increased risk of ovarian cancer among metal workers (31), while in the study by Pukkala et al a similar risk increase was not detected for ovarian cancer in Nordic countries (18). Our results may, thus, reflect a real occupational exposure as a risk especially for PFTC for female workers in smelting factories. John Wiley & Sons, Inc.
International Journal of Cancer
Page 10 of 21
The overrepresentation of PFTC among artistic workers in our study could be, at least partly, an effect of exposure to chemical agents as the trend was increasing by age. The dose-response risks related to specific exposure agents will be evaluated in our forthcoming study. The cancer risk among hairdressers has been widely examined, as hair dyes may contain a variety of carcinogenic agents such as aromatic amines (32). For female hairdressers an elevated risk for cancers of the pancreas, lung, cervix and in situ cancer of the skin was detected in a Swedish cohort study (33). For gynaecological cancers, elevated risks among hairdressers have been observed especially for

Source: http://astra.cancer.fi/NOCCA/NIVA/Riska_2011.pdf