[Relationship between physical activity and health examination variables in male workers--new methods to assess physical activity and their applications to epidemiologic research].

Few epidemiological studies investigating the relationship between physical activity and cardiovascular diseases have been performed in Japan. To quantify the physical activity of individuals, a 24-hour activity record was developed to calculate 24-hour energy expenditure for use in an epidemiologic survey. Major physical activity was recorded every 15 minutes during a day and the relative metabolic rate for each physical activity was adopted from data of the Labor Science Institute in Japan. The 24-hour energy expenditure per body surface area was used as an index of physical activity. Reproducibility of this index was examined by retesting subsamples 3 months apart, and good results were confirmed. The association between physical activity and health examination variables was analyzed for the 319 male workers aged 40-49 in Osaka, Kochi, and Akita Prefecture. Twenty-four-hour energy expenditure per body surface area was positively correlated with maximal oxygen uptake and negatively correlated with skinfold thickness, resting heart rate, logarithm of triglyceride, hemoglobin and uric acid. In addition, a newly developed questionnaire on physical activity with 11 questions was administered in Osaka on 1819 male factory workers aged 40-49. From multiple regression analysis, three questions were shown to be significant predictors of 24-hour energy expenditure per body surface area--1. "What percentage of work time is spent on your feet?" 2. "When climbing three floors, which do you usually choose, the stairs or the escalator?", 3. "How long are you involved in heavy physical labor each day?" The daily physical activity score estimated using the regression equation showed good reproducibility in retesting a year apart. Controlling for age, body mass index, alcohol intake, and cigarette smoking, the score was negatively correlated to skinfold thickness, diastolic blood pressure, resting heart rate, total cholesterol, logarithm of triglyceride, hemoglobin, and uric acid; and was positively correlated to HDL-cholesterol. This cross-sectional study shows that in Japanese there is a significant relationship between physical activity and some health examination variables, and also that the survey methods to estimate physical activity can be useful for further epidemiologic studies in Japan.