Youji Iitoi

Nutrient consumption and gastric cancer mortality in five regions of Japan

To explore the causes of the threefold variation in mortality rate from gastric cancer in Japan, we studied the geographic correlations between nutrient consumption and the disease in five Public Health Center districts including the regions with the highest and lowest mortality rates in the country. In the winters of 1989-1991, a three-day weighed food record was collected from 207 men and the wives of 165 of the men sampled from the five districts. The average daily consumption of selected nutrients was computed and correlated with the age-adjusted mortality rates from gastric cancer. Partial rank correlation coefficients adjusted for sex and other nutrients were 0.45, -0.80, -0.20, and -0.07 for sodium, carotene, ascorbic acid, and retinol, respectively. The results suggest that variation in gastric cancer mortality in Japan may be partly accounted for by the regional differences in consumption of sodium, carotene, and possibly ascorbic acid.

Interpopulation and intrapopulation variability of nutrient intake in five regions of Japan

Objectives: To determine the extent to which interpopulation (between-population) variance, relative to intrapopulation (within-population) variance, contribute to the total variability in nutrient intakes.Design: Cross-sectional study.Setting: Five Public Health Center districts in Japan.Subjects: Two hundred and seven men and 183 spouses.Interventions: A three-day weighed food record.Main outcome measures: The total variance in the consumption of 17 nutrient variables was partitioned by analysis of variance into its inter- and intrapopulation components separately for men and women.Results: The percentage contribution of the interpopulation to total variance differed according to the nutrient; it was notable (8–17%) for total energy, carbohydrates, phosphorus, and sodium in both men and women, but was negligible (less than 4%) for micronutrients such as retinol, carotene, thiamin, riboflavin, niacin and ascorbic acid. The ratio of intra- to interpopulation variance was estimated for 31 nutrients (17 in men and 14 in women). The point estimates of the ratio were larger than unity in all nutrients, and the lower limit of the 95% confidence intervals exceeded unity for all but 5 nutrients. Of the two sources of intrapopulation variation, intraindividual variance was larger than interindividual variance in most of the nutrient.Conclusions: The relative magnitude of interpopulation variation in dietary data can be used to quantify the range of exposure in ecological studies and to examine the heterogeneity of populations pooled for individual-based analysis.Sponsorship: This study was supported in part by Grants-in-Aid for Cancer Research from the Ministry of Health and Welfare, and for Scientific Research from the Ministry of Education, Science and Culture of Japan.