In Japan, as in most developed countries of the world, the stomach cancer mortality rate has shown a marked and stable downward trend over the last several decades (Correa and Chen, 1994;Spataro et al., 1994;Ansari, 1995;Levi et al., 1996;Tredaniel et al., 1997;Laheij et al., 1999). However, stomach cancer is still relatively common in Japan (Tsugane et al., 1999). According to recent data (Health and Welfare Statistics Association, 1999), stomach cancer is associated with the highest mortality rate in females and the second highest mortality rate in males for deaths caused by malignancies in 1997. Therefore, it is still an important public health problem in Japan and other countries (Pisani et al., 1990).
Against this background, we feel it is necessary to present a picture of the stomach cancer trend in Japan. It is clear that more work is needed to further characterize annual trends of stomach cancer. Analysis of the annual trends of mortality may provide useful evidence for the introduction or elimination of possible aetiologic factors.
In this study, we compared the annual mortality rates of stomach cancers with those related to cancers of digestive system and trachea, bronchus and lung (TBL), since the mortality rate associated with TBL cancer has increased in the last few decades compared with a decrease in stomach cancer death rate. Our aim was to provide a comprehensive epidemiological analysis of the impact of stomach cancer in Japan. Specifically, we examined similarities and differences in the epidemiology of stomach cancer and cancers of other parts of the digestive system and TBL cancers, and provided further information about the characteristics of stomach cancer.
Several studies have examined gender differences in mortality rates of stomach cancer. In all countries where such studies have been conducted, the rate in females is universally lower than in males, with a ratio ranging from 1.6:1 to 5:1 for male to female mortality rate (Spataro et al., 1994;Ansari, 1995;Herrero et al., 1996;Neugut et al., 1996;Niang et al., 1996;Aragones et al., 1997;Mikov et al., 1997). The impact of gender and age, including age distribution, needs to be studied further. The present study assessed gender differences in the epidemiology of stomach cancer and examined the implications of these differences for further public health interventions. We also analysed the contribution of stomach cancer deaths to the overall difference in life expectancy between males and females.
The data used for analysis were obtained from the Journal of Health and Welfare Statistics published by the Health and Welfare Statistics Association (1999). The collected data included cancers of the digestive system, TBL cancers and cancers of all organs (total cancer), which were used for the comparative study. Because the mortality data of several years were not available for colon and rectum cancer, the term ‘digestive system cancer’ used in this study refers to only cancers of the oesophagus, liver, pancreas and stomach.
Age-adjusted death rates for each cancer were calculated. The direct method of age adjustment was used in our analysis of the Japanese standard population, based on the 1985 National Census. The time trend of age-specific mortality rates was plotted on a graph to examine the effect of age. The overall male–female ratio was calculated from standardized mortality rates. Data from 1970 to 1995 were stratified into 5-year spans for age-specific sex impacts of stomach cancer, TBL cancer and total cancer. The contribution of stomach cancer to the overall change in life expectancy was calculated using cause-elimination life tables in the abridged form with ages grouped into 5-year intervals, except the first 5 years. All analyses were carried out separately for males and females.
Analysis of stomach cancer mortality rates in Japan was performed during a 25-year period extending from 1970 to 1995. The adjusted death rates for the population of all ages decreased from 88.9 to 45.4 per 100 000 in males and from 46.5 to 18.5 per 100 000 in females, representing a fall of 48.9% and 60.2% for males and females, respectively.
Age-specific stomach cancer-related death rates between 1970 and 1995 are summarized in Fig. 1
. It is clear that the mortality rates were higher in 1970 than in 1995 for every age group. The mortality rate increased progressively with age in both sexes and both study years.
shows the overall male to female ratio for stomach cancer mortality rates as well as those of other digestive system cancers for the period 1970–1995. Compared with other digestive system cancers, stomach cancer was associated with a relatively low sex ratio (range 1.9–2.5) compared with those of oesophageal cancer (3.9–7.8), liver cancer (2.0–3.6) and pancreatic cancer (1.6–1.8). Furthermore, during the period 1970–1995, the change in male–female ratio differed among various digestive system cancers. There was a gradual but steady rise in the ratio throughout the study period for both stomach and oesophageal cancers, although the speed of rise in the former was less than in the latter. Pancreatic cancer and liver cancer demonstrated a less marked time trend in changes related to male–female ratio. In all examined cancers, the male–female ratio was markedly higher than 1.0, indicating that more males than females died from cancers arising from the digestive system.
shows the effect of cancers of the digestive system on life expectancy by age and gender in 1995. Stomach cancer was the largest negative contributor to life expectancy, followed by liver cancer, pancreatic cancer and oesophageal cancer. The effect of oesophageal cancer on life expectancy was almost negligible, particularly in females; the negative effect was stable at young age and became smaller until later in life. The analysis further showed that the effect of stomach cancer on life expectancy was more pronounced in males (0.65 years) than females (0.42 years) in the younger age groups. The differential effect between the two sexes varied from 0.23 to 0.26 years in the age groups before 65, and was lower in the elderly. This is similar to the pattern seen in life expectancy tables with advanced age.
We also examined the time trends of life expectancy at birth for the effect of stomach cancer (Table 2
). In general, there was a progressive increase in the life expectancy with time. Further analysis showed that the negative contribution of stomach cancer on life expectancy was 41.8% of total cancer in 1970 but only 17.8% of total cancer in 1995 in males, and 34.5%, 15.9%, respectively, in females. We also compared the effect on life expectancy of stomach cancer with that of TBL cancers (Fig. 3
). A detailed time trend comparison of stomach cancer and TBL cancer curves indicated that a decrease in stomach cancer and increase in TBL cancer with calendar years contributed negatively to life expectancy. The change was higher in males than in females. By the end of the observation period for both cancers the effects in males was almost of the same magnitude although the difference in negative contribution to life expectancy in 1970 was more than 4 times. Females showed a trend similar to that seen in males.
Because incidence data are not available, our study only analysed the latest trends in stomach cancer mortality rates. The high ranking of cause of death and continuing decline in mortality rate for stomach cancer is a basic feature of this disease, not only in Japan but also in other countries (Marinelli et al., 1989). In our study, this feature was not different from that in other countries (Spataro et al., 1994;Ansari, 1995;Laheij et al., 1999). The observed changes were consistent with a continuing decline in stomach cancer-related mortality rate for both sexes and individual age groups over the study period. The overall mortality rate decreased in males and to a larger extent in females during the 25-year period of analysis. The reduced rate of stomach cancer-related mortality rate is higher than that reported in most countries in the world (Spataro et al., 1994;Levi et al., 1996;Laheij et al., 1999) during the observation years, but it is similar to that of Italy, a country where the stomach cancer-related mortality rate is high, as in Japan (Marinelli et al., 1989). Our analysis also showed a considerable variability in the mortality rate with regard to age; stomach cancer-related death in the elderly was worse than in the young. The age group associated with the highest mortality rate was older than that reported in Vojvodina, Yugoslavia (Mikov et al., 1997). In today's aged society, the characteristic of ‘a disease of the elderly’ becomes clearer and clearer. However, there is a large variability in this trend in different countries (Marinelli et al., 1989;Gil et al., 1990;Levi et al., 1996;Aragones et al., 1997).
Our study analysed the recent trends in male–female ratio with respect to stomach cancer-related mortality in order to understand different epidemiological aspects of stomach cancer. The variability in male–female ratio is high worldwide (Herrero et al., 1996;Neugut et al., 1996;Niang et al., 1996;Aragones et al., 1997;Mikov et al., 1997). In some countries, stomach cancer-related mortality rate in males is 5 times that in women (Neugut et al., 1996). In our study of cancer of the digestive system, the male–female ratio of stomach cancer was lower than that of oesophageal cancer. In time trend studies, the change in the ratio was also slower than that of oesophageal cancer in each calendar year. The male–female ratio in our study was slightly higher than that reported in Italy (Marinelli et al., 1989).
We also noted age-specific changes in the differences in male–female mortality rates related to stomach cancer. It is noted that in groups less than 40 years of age, the stomach cancer-related mortality rate was higher in females than in males. It is possible that the difference is due to physiological, endocrinological and immunological factors, which may determine the susceptibility to and invasiveness of stomach cancer. One way of interpreting this phenomenon is that male and female hormones may affect the ability of individuals of each sex to resist stomach cancer. In short, the present analysis clearly demonstrated the age-dependent contribution of determinants to male–female differences in mortality due to stomach cancer. It also suggested different age- and sex-related vulnerability to death from stomach cancer. In a search for the aetiological mechanisms of stomach cancer, therefore, it is important to investigate the relevant factors at different age groups as well as factors related to either sex.
A more accurate method of summarizing the overall impact of stomach cancer in different age and sex groups is to measure its effect on life expectancy. In this study, we examined the effect of stomach cancer-related deaths on life expectancy for different age groups and times. During the 25-year period of our study, the negative contribution of stomach cancer on the average life expectancy decreased by 0.27 years in males and 0.28 years in females in parallel with the trends of reduced mortality rate from stomach cancer. In cancers of the digestive system, the contribution of stomach cancer to the reduction in average life expectancy was the largest, but this effect diminished for females in young age groups compared with other digestive cancers in which deterioration commenced at 60 years in females as opposed to 40 years for stomach cancer. The negative impact of stomach cancer in males started to diminish later in life compared with females. This effect was different from that seen in other cancers of the digestive system. One reason for this observation is the high mortality rate related to stomach cancer in young females and the lower rate of increase with age than in males. On the other hand, compared with other cancers of the digestive system, the impact of stomach cancer on life expectancy in the elderly is relatively high; it was associated with a high mortality rate in the elderly.
Comparison of the mortality rates of stomach cancer and TBL cancer showed that the most striking difference was the opposite contribution trend between 1970 and 1995. Following elimination of stomach cancer, the gain in life expectancy was 0.65 years in males and 0.42 years in females while after elimination of TBL cancer, the gain in life expectancy was 0.65 and 0.27 years, respectively in 1995. Although the mortality of stomach cancer has decreased sharply and consistently while that of TBL cancer increased sharply in Japan and other countries (Stanley et al., 1988), the negative contribution of stomach cancer to average life expectancy is still the largest, followed by TBL cancer at the end of this observation period. This difference between stomach and TBL cancers refers to those of females only because the effects on life expectancy were identical for both cancers in males.
As for the time trend, the largest change in life expectancy is in TBL cancer in males. From 1970 to 1995, the absolute change in males was 0.44 years; in contrast, there was a smaller absolute change in females (0.17 years). In stomach cancer, as described above, the changes in both sexes were almost identical. Absolute change in life expectancy is 0.27 year in males and 0.28 year in females respectively between 1970 and 1995, though in general a higher mortality rate was noted in males.
Many studies have examined the underlying mechanisms of the observed reduction in stomach cancer-related mortality rate. The proposed mechanisms include widespread improvement in environmental factors (Laheij et al., 1999), living circumstances, diet (Kaaks et al., 1998), economic level, lifestyle habits (Watanabe et al., 1996), labour conditions (Steenland et al., 1991;Roscoe et al., 1992;Acquavella et al., 1993;Spataro et al., 1994;Plato et al., 1995), housing, transportation, and the health insurance system (Watanabe et al., 1996). Increased consumption of fresh fruits and vegetables (Verhoeven et al., 1996;Takezaki et al., 1999) and reduced use of salt are considered to be the most relevant factors in explaining the reduced temporal trend. Some studies in Japan showed that the correlations between stomach cancer and some foods (rice and green yellow vegetables) were in general agreement with most studies but did not find an inverse association between stomach cancer and fruit intake. Residents of Okinawa, the region with the lowest mortality from stomach cancer, consumed beef, cheese, and oolong tea most frequently and rice, fruit, pickled vegetables, mushrooms and soybean products least frequently (Tsubono et al., 1997). Green tea consumption decreases the odds ratio of stomach cancer (Hirohita and Kono, 1997).
On the other hand, several risk factors are known to predispose to stomach cancer-related deaths, including smoking (Akiba and Hirayama, 1990;Tredaniel et al., 1997;Pisani et al., 1999) and chronic gastritis caused by Helicobacter pylori (Forman, 1996;Sipponen, 1996;Palli, 1997;Tomaselli et al., 1999). The results of Tsugane et al. (1999) showed that the consumption frequency of pickled vegetables was positively associated with the prevalence of H. pylori infection. However, prevalence of H. pylori infection is very high even among populations at very low risk for stomach cancer (Hirohita and Kono, 1997). While a number of factors may account for this trend, the exact mechanisms of stomach carcinogenesis are not yet identified, and the above findings have been confirmed only in part; no strong direct association has been provided.
A declining trend of stomach cancer-related mortality rate has been reported almost worldwide, while differences and changes in the above factors, for example, diet (Kaaks et al., 1998) and alcohol consumption (Ekstrom et al., 1999), are very large in different countries and regions (Palli, 1997). Furthermore, the fall in mortality rate has been noted over a long period of time. In the last year of observation in our study, stomach cancer-related death was still the leading cause of cancer death in Japan while in the USA the leading cause of cancer death occurred in 1930 (Neugut et al., 1996). Therefore, general local factors and environmental hazards seem unlikely to have a major impact on the overall trend in mortality rate of stomach cancer. The reduction in mortality rate due to stomach cancer can reasonably be considered the result of a spontaneous trend because of a natural rule – the choice of demographic heterogeneity (Mecklin et al., 1988;Correa and Chen, 1994;Levi et al., 1996), namely the development and natural evolution of the environment over a long period of time.
1. Acquavella J , Leet T , Johnson G ( 1993 ). Occupational experience and mortality among a cohort of metal components manufacturing workers . Epidemiology 4 : 428 – 434 .
2. Akiba S , Hirayama T ( 1990 ). Cigarette smoking and cancer mortality risk in Japanese men and women–results from reanalysis of the six-prefecture cohort study data . Environ Health Perspect 87 : 19 – 26 .
3. Ansari MZ ( 1995 ). Gastric cancer in Tasmania (1978–1992) . Asia Pacific J Public Health 8 : 171 – 176 .
4. Aragones N , Pollan M , Rodero I , Lopez-Abente G ( 1997 ). Gastric cancer in the European Union (1968–1992): mortality trends and cohort effect . Ann Epidemiol 7 : 294 – 303 .
5. Correa P , Chen VW ( 1994 ). Gastric cancer . Cancer Surv 19/20 : 55 – 76 .
6. Ekstrom AM , Eriksson M , Hansson LE et al. ( 1999 ). Occupational exposures and risk of gastric cancer in a population-based case–control study . Cancer Res 59 : 5932 – 5937 .
7. Forman D (1996). Helicobacter pylori
and gastric cancer. Scand J Gastroenterol214(Suppl):
31–3; Discussion 40–3.
8. Gil MA , Castello BF , Sarthou CA , Fernandez GFJ , Birlanga CP ( 1990 ). Evolution of stomach cancer
mortality in Spain (1951–1985) . Rev Esp Enferm Dig 78 : 61 – 66 .
9. Health and Welfare Statistics Association . J Health Welfare Stat 46 : 53 , ( 1999 ).
10. Herrero J , Vergara M , Fraga X et al. ( 1996 ). Changes in the epidemiology of gastric cancer in the last 15 years . Rev Esp Enferm Dig 88 : 323 – 327 .
11. Hirohita T , Kono S ( 1997 ). Diet/nutrition and sc in Japan . Int J Cancer 10 ( Suppl ): 34 – 36 .
12. Kaaks R , Tuyns AJ , Haelterman M , Riboli E ( 1998 ). Nutrient intake patterns and gastric cancer risk: a case–control study in Belgium . Int J Cancer 78 : 415 – 420 .
13. Laheij RJ , Straatman H , Verbeek AL , Jansen JB ( 1999 ). Mortality trend from cancer of the gastric cardia in The Netherlands, 1969–1994 . Int J Epidemiol 28 : 391 – 395 .
14. Levi F , La Vecchia C , Lucchini F , Negri E ( 1996 ). Worldwide trends in cancer mortality in the elderly, 1955–1992 . Eur J Cancer 32A : 652 – 672 .
15. Marinelli M , Bianucci F , Leoni E ( 1989 ). Trends in the mortality from stomach tumors in Italy . Ann Ig 1 : 109 – 124 .
16. Mecklin JP , Nordling S , Saario I ( 1988 ). Carcinoma of the stomach and its heredity in young patients . Scand J Gastroenterol 23 : 307 – 311 .
17. Mikov MM , Burany B , Zdravkovic S ( 1997 ). Epidemiological characteristics of gastric cancer in Vojvodina . Eur J Epidemiol 13 : 523 – 525 .
18. Neugut AI , Hayek M , Howe G ( 1996 ). Epidemiology of gastric cancer . Semin Oncol 23 : 281 – 291 .
19. Niang A , Mbengue M , Diouf ML et al. ( 1996 ). Current aspects of gastric cancer in Senegal. Epidemiological and clinical study of 220 cases (1984–1991) . Dakar Med 41 : 99 – 103 .
20. Palli D ( 1997 ). Gastric cancer and Helicobacter pylori
: a critical evaluation of the epidemiological evidence . Helicobacter 2 ( Suppl 1 ): S50 – S55 .
21. Pisani P , Parkin DM , Bray F , Ferlay J ( 1990 ). Estimates of the worldwide mortality from 25 cancers in . Int J Cancer 83 : 18 – 29 .
22. Pisani P , Parkin DM , Bray F , Ferlay J ( 1999 ). Estimates of the worldwide mortality from 25 cancers in 1990 . Int J Cancer 83 : 18 – 29 .
23. Plato N , Westerholm P , Gustavsson P et al. ( 1995 ). Cancer incidence, mortality and exposure-response among Swedish man-made vitreous fiber production workers . Scand J Work Environ Health 21 : 353 – 361 .
24. Roscoe RJ, Steenland K, McCammon CS Jr, et al
. (1992). Colon and stomach cancer
mortality among automotive wood model makers. J Occup Med34:
25. Sipponen P ( 1996 ). Helicobacter pylori
gastritis – epidemiology . J Gastroenterol 32 : 273 – 277 .
26. Spataro V , Pedrinis E , Muller W ( 1994 ). Descriptive epidemiology of early and advanced gastric cancer in Ticino, Switzerland, with special emphasis on time trends
. Ann Oncol 5 : 954 – 956 .
27. Stanley K , Stjernsward J , Koroltchouk V ( 1988 ). Cancers of the stomach, lung and breast: mortality trends and control strategies . World Health Stat Q 41 : 107 – 140 .
28. Steenland K , Stayner L , Greife A et al. ( 1991 ). Mortality among workers exposed to ethylene oxide . N Engl J Med 324 : 1402 – 1407 .
29. Takezaki T , Gao CM , Ding JH et al. ( 1999 ). Comparative study of lifestyles of residents in high and low risk areas for gastric cancer in Jiangsu Province, China; with special reference to allium vegetables . J Epidemiol 19 : 297 – 305 .
30. Tomaselli G , Citarda F , Crespi M ( 1999 ). Helicobacter pylori
and gastric cancer . Clin Ter 150 : 221 – 224 .
31. Tredaniel J , Boffetta P , Buiatti E , Saracci R , Hirsch A ( 1997 ). Tobacco smoking and gastric cancer: review and meta-analysis . Int J Cancer 72 : 565 – 573 .
32. Tsubono Y , Kobayashi M , Tsugane S ( 1997 ). Food consumption and gastric cancer mortality in five regions of Japan . Nutr Cancer 27 : 60 – 604 .
33. Tsugane S , Fahey MT , Sasaki S , Baba S ( 1999 ). Alcohol consumption and all-cause and cancer mortality among middle-aged Japanese men: seven-year follow-up of the JPHC study Cohort I. Japan Public Health Center . Am J Epidemiol 150 : 1201 – 1207 .
34. Verhoeven DT , Goldbohm RA , van Poppel G , Verhagen H , van den Brandt PA ( 1996 ). Epidemiological studies on brassica vegetables and cancer risk . Cancer Epidemiol Biomarkers Prev 5 : 733 – 748 .
35. Watanabe S , Sobue T , Kinjyo Y ( 1996 ). Research activities of cancer epidemiology in Japan . Epidemiol 6 ( Suppl ): S19 – S29 .