We conducted a population-based prospective cohort study in Japan to examine the relationship between body mass index (BMI) and the risk of incidence of any cancer and of cancer at individual sites. Body mass index was calculated from self-administered body weight and height at baseline. Relative risks (RR) and 95% confidence intervals (CI) were calculated in multivariate proportional-hazards models. Among 27,539 persons (15,054 women and 12,485 men) aged 40 years or older who were free of cancer at enrollment in 1984, 1,672 (668 women and 1,004 men) developed cancer during 9 years of follow-up. In women, after adjustment for potential confounders, the RR of all cancers associated with different BMI, relative to a BMI of 18.5-24.9, were 1.04 (95% CI ؍ 0.85-1.27) for BMI ؍ 25.0 -27.4, 1.29 (1.00 -1.68) for BMI ؍ 27.5-29.9 and 1.47 (1.06 -2.05) for BMI >30.0 (p for trend ؍ 0.007). Higher BMI was also significantly associated with higher risk of cancers of the colorectum, breast (postmenopausal), endometrium and gallbladder in women. In men, we observed significantly increased all-cancer risk among only never-smokers. Overweight and obesity could account for 4.5% (all subjects) or 6.2% (never-smokers) of the risk of any cancer in women and ؊0.2% (all subjects) or 3.7% (never-smokers) in men. The value for women was within the range among women reported from Western populations (3.2%-8.8%). Our data demonstrate that excess weight is a major cancer risk among Japanese women. © 2004 Wiley-Liss, Inc. Key words: cancer; incidence; Japan; obesity; population attributable fraction Growing evidence indicates that excess body weight is associated with increased risk of cancer of the colon, 1-14 breast (in postmenopausal women), 1,3-9,15-18 endometrium 1-8,19 -21 and kidney. 1,3-5,8 -10,22-24 Adenocarcinoma of the esophagus, 1,4,5,10,[25][26][27][28] cancers of the gastric cardia, 1,5,10,[25][26][27][28] and gallbladder [1][2][3]7,8,10,29,30 have also been suggested to be associated with excess weight. Data for cancers of the prostate, 1,6 -10,31-42 pancreas 2,8,9,43-45 and ovary 2,8,9 are still inconclusive.Most of these reports come from Western populations. There is limited knowledge regarding the risk of excess weight among other populations, 46 -57 whose life-style characteristics and genetic backgrounds substantially differ from those in Western populations. Because variations in those factors could interfere with metabolic abnormalities or hormonal conditions, it is necessary to examine whether excess weight is also a risk factor for cancer among non-Western populations. 3,5,7,58 -61 In addition, there has been limited effort to estimate the total proportion of all incident cancers attributable to excess weight. 1,9,62 One meta-analysis from Europe estimated the proportions of all cancer cases attributable to excess weight: excess body weight accounted for 6% of incidence in women, ranging between 3.9% in Denmark and 8.8% in Spain; and 3% in men, ranging between 2.1% in Greece and 4.9% in Germany. 1 A rec...