18. Epidemiology and Biostatistics Division

The Epidemiology and Biostatistics Division has planned and conducted independent and collaborative studies on cancer etiology and prevention, with a special focus on dietary factors. In this respect, several epidemiological projects are currently in progress, including cross-sectional, case-control, cohort and intervention studies. In addition, collaborative studies in the field of clinical and basic sciences are conducted by providing statistical design, analysis and data processing support.

Etiological Roles of Foods and Nutrients in Cancer

Diet has been implicated as playing an etiological role in cancer occurrence and as contributing to the unique patterns of cancer incidence in Japan. However, epidemiological evidence regarding this issue has been limited. The Division has therefore initiated a population-based prospective study on diet and cancer, in collaboration with the Cancer Information and Epidemiology Division and the National Cardiovascular Center, in which approximately 140,000 individuals will be followed-up for 10 years. Among 55,000 population cohort members from 4 areas with a 7 year follow-up,1,540 deaths, 1,515 incident cancers and 2,881 moved-out had been documented as of December 31, 1996.

Geographical differences in stomach cancer mortality in 5 cohort areas has been investigated in relation to dietary habits (32 item food frequency questionnaire, 3-day dietary record and plasma level) for randomly-selected subjects by using an ecological study method. The frequency of salted food intake such as miso soup and pickled vegetables and sodium intake level were positively correlated with age-adjusted mortality rate of stomach cancer, while frequency of green and yellow vegetable intake, carotene intake level and plasma lycopene level were negatively correlated.^(215-217) All these findings should be examined in future cohort analyses. Certain dietary habits can be associated each other. A favorable effect of green tea intake on HDL cholesterol was diminished after adjusting other dietary factors⁽²¹⁸⁾ and dietary habits were significantly different among each category of green tea intake.⁽²¹⁹⁾ Within- and between-person variations in portion sizes⁽²²⁰⁾ and effect on blood pressure by type of alcohol beverages⁽²²¹⁾ were also examined in this cohort.

Environmental and Genetic Interaction in the Etiology of Human Cancer

Patterns of cancer incidence vary according to ethnicity and therefore change with migration which results in intermingling with other populations. These observations suggest an interaction between environmental and genetic factors in cancer occurrence. The division has been conducting several epidemiological studies in Brazil, a multi-ethnic nation with 1.2 million people of Japanese ancestry.⁽²²²⁾ A cross-sectional study has been conducted to obtain quantitative dietary data among Japanese residents in Sao Paulo. The mean daily proportions of energy from fat were 27% for male Japan-born residents and 30% for male Brazil-born residents.⁽²²³⁾ These figures were intermediate between Japan and Brazil.

An Epidemiological Evaluation of Cancer Prevention Methods

Practical prevention strategies are generally based on total evidence obtained from observational epidemiological studies, experimental animal studies and mechanistic interpretation. An evaluation based on intervention studies is, however, essential before recommending a strategy to the general public. The available evidence suggests that the risk of stomach cancer is increased in individuals with chronic atrophic gastritis and is decreased in those with a high consumption of vitamin C and beta-carotene. A randomized controlled trial to assess the effectiveness of vitamin C and beta-carotene supplementation on stomach cancer incidence was initiated for individuals with chronic atrophic gastritis in a district where the mortality rate is one of the highest in Japan. In the first year of the study, 439 (73%) out of 602 eligible subjects agreed to participate in the trial and started to take one of four possible regimens based on a two-by-two factorial design with two levels of vitamin C (50 or 500 mg/day) and beta-carotene (0 or 15 mg/day). Most participants continued to take the supplements till January 1996. However, after a press release by the US NCI regarding its unfavorable effect, beta-carotene supplementation was halted and approximately 300 remained in the vitamin C trial.⁽²²⁴⁾ The trial has progressed to the third year.

Epidemiological and Biostatistical Support for Basic and Clinical Research

Epidemiological and biostastical support have been provided for a randomized controlled trial of esophageal cancer,⁽²²⁵⁾ pancreas,⁽²²⁶⁾for the prognostic analysis of cancer of breast,^{(2, 205)} lung,⁽²²⁷⁾ and liver,⁽²²⁸⁾ and for the risk analysis for atherosclerosis.⁽²²⁹⁾