Temporal changes in cumulative mortality risks of cancer, by occupation, in the working population of Japan from 1995 to 2020: a benchmark for radiation risk comparison

Data sorted by occupation in Japan for the 5 year period 1995–2020 were taken from the Report of Vital Statistics: Occupational and Industrial Aspects published by the Ministry of Health, Labor and Welfare [18]. Population and mortality by 5 year age groups from 15 to 74 were used for the analyses. The population data are consistent with the total population aged 15–74, including non-workers (students and the unemployed). Six occupational categories were used in the present study: professional; managerial; services; agriculture, forestry, and fishery (AFF); and transportation and manufacturing, as shown in table 1. The occupational classification categories in the vital statistics data have been modified and subdivided over a 25 year period; therefore, to examine temporal changes over this period and to be consistent with international occupational classifications, the classification categories in the vital statistics were merged for some categories. The analysis excluded the unemployed, including students, and those who were employed but whose occupation was unknown. The detailed occupations included in the six occupation categories are shown in table S1.

In this study, we focused on the same types of cancer among the effects induced by radiation, colorectal cancer, lung cancer, and breast cancer (females only), and calculated the cumulative mortality risk from ages 15–74. To examine the factors contributing to the differences in the cancer mortality risks by occupation, cumulative mortality risks for all causes of death except death of old age were also assessed by dividing causes of death into four categories: all cancers; cardiovascular disease; other diseases, such as diabetes and pneumonia; and external causes due to unintentional accidents and suicide. Table S2 shows the corresponding ICD-10 classification codes.

To calculate the cumulative mortality risk from age 15–74, we used the following formula for lifetime risk, based on Ulanowski et al's [10] work:

$$\begin{equation}{\text{lifetime risk}} = \int\limits_e^\infty \lambda \left( t \right){S_e}\left( t \right){\text{d}}t\end{equation} \tag{ 1 }$$

where λ(t) is the mortality rate from a certain cause of death at age t and S_e (t) is the survival rate from age e to age t, as in the following equation:

$$\begin{equation}{S_e}\left( a \right) = {\text{exp}}\left\{ { - \int\limits_e^a \mu \left( t \right){\text{d}}t} \right\}\end{equation} \tag{ 2 }$$

where μ(t) is the mortality rate from all causes at age t.

The actual calculation was discretized by age and approximated. Since the number of deaths from each cause of death is provided for each 5 year age group, the cumulative mortality risk was calculated for each age group from 15 to 74 years. Mortality rates (5 year average μ(t)) were obtained from the population and total deaths in each age group, and survival rates were calculated from equation (2). Mortality rates for the target causes of death were obtained from the population in each age group and the number of deaths for the target cause of death.

From 1995 to 2020, the population aged 15–74 years showed a decreasing trend; males decreased from 4.8 × 10⁷–4.5 × 10⁷ (6.3% reduction) and females decreased from 4.9 × 10⁷–4.5 × 10⁷ (8.5% reduction). Figure 1 shows the percentage of each occupational category in the total population, including the non-workers, aged 15–74 in each year. The proportion of male non-workers has increased over time, from approximately 20% in 1995%–35% in 2020. The proportion of female non-workers has not changed significantly, accounting for about half of the population. Based on the population changes over the past 25 years, the number of workers has decreased for both males and females.

Zoom In Zoom Out Reset image size

Regarding each occupational category, transportation and manufacturing accounted for the largest proportion of workers among males, 30%–25%, and showed a decreasing trend over time. The next-largest group was service workers, at 15%, followed by professional and clerical workers, at around 10% for each. The smallest number of workers were in AFF and managers, each with less than 5%.

The largest proportions of working females were service and clerical workers, each accounting for about 15% of the total population. Transportation and manufacturing and professional workers were next, at about 10% each, but while transportation and manufacturing have declined over the past 25 years, the number of professional workers has increased. AFF accounted for less than 5% of the total. The smallest group of workers was managers, accounting for less than 0.5%.

The cumulative mortality risks for each occupation for colorectal cancer, lung cancer, and breast cancer (females only) between 1995 and 2020 for people aged 15–74 are shown in figure 2 and table S3. The cumulative mortality risk for colorectal cancer for the average male worker ranged from 5.0 × 10^–3 to 1.1 × 10^–2, and for lung cancer it ranged from 8.9 × 10^–3 to 2.0 × 10^–2. Over the 25 years examined, the mortality risk showed a decreasing trend, with a reduction of 53.1% for colorectal cancers and 56.5% for lung cancers. By occupation, the highest risk of colorectal cancer in males was 1.6 × 10^–2, for professional workers in 2000, and the lowest risk was 2.1 × 10^–3, for clerical workers in 2020. The highest for lung cancer in males was 2.6 × 10^–2, for professional workers in 2000, and the lowest was 3.4 × 10^–3, for clerical workers in 2020.

Zoom In Zoom Out Reset image size

The average cumulative mortality risk for female workers for colorectal cancer ranged from 2.6 × 10^–3 to 6.1 × 10^–3, for lung cancer from 2.3 × 10^–3 to 5.2 × 10^–3, and for breast cancer from 3.3 × 10^–3 to 4.9 × 10^–3. As in males, the mortality risk showed a decreasing trend over the 25 year period; colorectal cancers decreased by 55.2%, lung cancers by 53.5%, and breast cancers by 26.1%. By occupation, the highest risk for colorectal cancer among females was 1.6 × 10^–2, for managers in 2000, and the lowest risk was 1.1 × 10^–3, for clerical workers in 2020. The highest for lung cancer in females was 1.3 × 10^–2, for managers in 2000, and the lowest was 9.4 × 10^–4, for clerical workers in 2020. For breast cancer, the highest value was 2.2 × 10^–2, for managers in 2000, and the lowest value was 1.4 × 10^–3, for transportation and manufacturing in 2005.

For males, compared with an average cumulative mortality risk of 1 for all workers aged 15–74 for both colorectal and lung cancer, the risks were about 0.5 for clerical and transportation and manufacturing workers, about average for service workers and managers, and up to 1.5–2 times the average for professionals (figure S1). For AFF workers, the risk was at average level in 1995, increased over time, and doubled in 2020. The difference between the maximum and minimum mortality risk was smallest in 1995, about 2-fold, and then increased and remained between 3- and 4-fold.

For females, the occupational differences in cumulative mortality risk were generally similar to those for males (figure S1). For all cancers, the risk was about half the workers' average for clerical and transportation and manufacturing workers, about average for service and AFF workers, and high for professional and managerial workers. The maximum risk was about twice as high for professionals, as it was for males, but the maximum risk was nearly five times the average for female managers. The difference between the maximum and minimum risk ranged from 4 to 10 times. Breast cancer showed a slightly decreasing trend, but colorectal and lung cancers did not change significantly over time.

We calculated the cumulative mortality risk from age 15–74 for each occupation for all deaths, except death of old age, among the working population by dividing the causes of death into four categories: all cancers, cardiovascular disease, other diseases, and external causes (figure 3, table S4). Over the 25 year period, the mortality risk from each cause for all male workers showed a downward trend. By occupation, the highest value for all cancers for males was 1.4 × 10^–1, for managers in 2015, and the lowest value was 1.6 × 10^–2, for clerical workers in 2020. The highest value for cardiovascular disease was 9.3 × 10^–2, for professionals in 2020, and the lowest value was 9.7 × 10^–3, for clerical workers in 2020. The highest value for other diseases was 8.9 × 10^–2, for managers in 2010, and the lowest value was 2.6 × 10^–3, for clerical workers in 2015. The highest value for deaths from external causes was 2.0 × 10^–1, for managers in 2005, and the lowest value was 8.7 × 10^–3, for clerical workers in 2020.

Zoom In Zoom Out Reset image size

Similar to males, females' mortality risks from each cause of death also showed a decreasing trend over the 25 year period. By occupation, the highest risk of all-cancer mortality was 1.4 × 10^–1, for managers in 2000, and the lowest was 1.0 × 10^–2, for clerical workers in 2020. The highest value for cardiovascular disease was 4.6 × 10^–2, for managers in 2000, and the lowest value was 3.3 × 10^–3, for clerical workers in 2020. The highest value for other diseases was 1.5 × 10^–1, for managers in 2015, and the lowest value was 4.9 × 10^–4 for clerical workers in 2020. The highest risk for external causes was 3.3 × 10^–1, for managers in 2020, and the lowest value was 2.3 × 10^–3, for clerical workers in 2015.

The cumulative mortality risk for lung cancer has decreased over the past 25 years. This may be due to a decrease in smoking rates in Japan, as smoking is the largest risk factor for lung cancer [19, 20]. According to Tanaka and Kobayashi [21], smoking rates among 25–64 year olds in Japan decreased from 56.0% to 38.4% for males and from 17.0% to 13.0% for females between 2001 and 2016.

Cao et al [22] reported on changes in international cancer mortality among 40–84-year-olds from 1981 to 2010 by grouping countries according to the human development index (HDI, a measure of a country's level of development based on health, education, and living standards) and found that lung cancers in females have been increasing in high HDI countries, a group that includes Japan. This this is different from the trend we have observed in Japan. The declining trends in cumulative mortality risk for lung cancer for males were consistent with international trends. In addition to lung cancer, the cumulative mortality risks from colorectal and breast cancers have also shown decreasing trends, which is consistent with international trends. Finally, the cumulative mortality risk from any cause of death, not just these three types of cancer, has declined over the past 25 years, consistent with the international trend of declining international mortality rates from 1981 to 2010 for all cancers, cardiovascular disease, and all causes of death for males and females aged 40–84 years.

Regarding differences in the cumulative mortality risk by occupation, two groups were below average, clerical workers and transport and manufacturing workers, while service workers, male managers, and female AFF workers were average. Professionals and female managers tended to be above average. The trends for the three cancer groups were generally similar for the four categories of all causes of death (all cancers, cardiovascular disease, other diseases, and external causes). This suggests that the causes of occupational differences in the three types of cancer mortality may be related to factors associated with cancer in general and cardiovascular disease, rather than to factors such as occupational exposure to carcinogens.

The high risks among female managers and both male and female professionals is expected to be due, in part, to prolonged sedentary behaviors, long working hours (and consequently lower levels of physical activity), and stress [15, 23–28]. Professional workers, such as those in medical occupations, can be expected to complete shift work and night shifts, a work style associated with a higher risk of cancer [13, 14], and these effects may be reflected in the present study's results. Report on smoking among adults in Japanese indicated that smoking rates among professionals and managers were average for females and below average for males [21], making it difficult to conclude that the high mortality risks for the three types of cancer and other causes of death are strongly influenced by smoking. Highly educated females are known to have higher risk of breast cancer [29, 30]. This is likely applicable to female managers in Japan. The high mortality risks of female managers were also for colorectal and lung cancers, as well as all other causes of death. These risks are thought to be associated with prolonged occupational sedentary behavior and long working hours, which are common characteristic of female managers [24, 25, 31].

The high risk of lung cancer mortality in male AFF workers and average level among female AFF workers suggests that smoking may be a factor for them; female AFF workers' rates were much lower than the average, while male AFF workers were among those with the highest smoking rates. In contrast, transportation and manufacturing, a low-risk group, showed a low risk, despite being the group with the highest smoking rates for both males and females. This may be because the occupations in these groups include, for example, construction workers, with high levels of occupational physical activity, which is a protective factor against cancer. Despite this, it is unclear what causes the higher risk among AFF workers, who have similarly high smoking rates and occupational physical activity levels, despite the lower risk in transportation and manufacturing workers. This trend was observed for all causes of death, not just the three types of cancer. Accordingly, there may be an influence of factors other than smoking and occupational physical activity, such as stress, which have associations with all diseases. Another low-risk occupation was clerical work, which it is essentially a sedentary behavior job, yet it is one of the occupational categories with the lowest smoking rates, which may contribute to its members' low risk. Clerical work is less associated with risk factors such as shift work, night shifts, and long working hours, which may also contribute to its low risk. Service workers were at the average level for both male and female workers, although smoking rates were higher than average in this group for both sexes.

Previously conducted international review studies have suggested that higher occupational socioeconomic status is associated with lower cancer mortality, except for breast cancer in females [9, 32]. For colorectal and lung cancer, the Japanese occupational trends identified in this study were inconsistent with these studies; occupations with higher socioeconomic status, such as managers and professionals, tended to have higher mortality risks. It has been reported that cardiovascular disease in Japan also does not follow the same socioeconomic gradient as confirmed mainly in studies conducted in Western countries [33]. This difference between the international studies and the present results, which are confined to Japan, may be due to differences in culture, social, and medical systems, which are also related to diet, lifestyle, and employment norms. This is supported by the fact that studies conducted in Asian and Latin American countries have reported inconclusive or positive socioeconomic gradients in lung cancer incidence [9].

Our study has several limitations. The demographic data used in this study do not have individual-level characteristics, so there is no information on cancer risk/prevention factors such as smoking or diet, which prevented us from identifying causes for the differences. In addition, the use of broad occupational classifications may limit the interpretation of the results. Such broad classifications would not be able to detect the effects of occupational exposure to carcinogens. Finally, as people's occupations can and do change, discussions using occupational classifications reported at the time of death would affect the results.

Over the past 25 years in Japan, the cumulative mortality risks for colorectal and lung cancers among the working population have declined by more than 50% in both males and females. Breast cancer has declined by about 30% in females. Internationally, there has also been a downward trend in the mortality risk from these types of cancers. Such a reduction in baseline cancer mortality risk could result in a non-negligible reduction in the associated estimate of radiation risk, and even more so if the excess cancer mortality from low-dose radiation were projected based on a multiplicative model. As the tolerability of cancer risk associated with low-dose radiation involves subjective considerations [2], the substantial downward trend in cancer mortality among the working population suggests that societal risk tolerability may also be changing over time.

The differences in mortality by occupation found in the present study have been confirmed internationally. However, high-risk/low-risk occupations vary by country and region, perhaps reflecting social systems and cultural differences. Accordingly, it may not be possible to draw a general conclusion about the tolerability of radiation-induced cancer mortality in relation to occupational types. At the same time, identifying the causes for differences in mortality risk among different groups in society, as in the present study, could be useful for addressing the relationship between society and tolerability. In addition, it is important to confirm whether there is a gap between international recommendations on tolerability and the social situation in one's own country, and it will be useful to organize data such as those features in this analysis. The results also can contribute to help workers exposed to radiation to put radiation risks into perspective with the background risk of cancer.