Associations between anthropometric measures and cancer have been studied previously, but relatively few studies have had the opportunity to control for genetic and early shared environmental factors. In this study, we analyzed 2 twin cohorts from Sweden born 1886-1925 (n = 21,870) and 1926-1958 (n = 30,279) and 1 from Finland born 1880-1958 (n = 25,882) including in total 78,031 twins, and studied the association between BMI and height and risk of prostate, breast, ovarian, corpus uteri, colon and rectal cancer. The cohorts were both analyzed through a co-twin control method and as traditional cohorts. In co-twin control analyses, older obese (BMI > or = 30 kg/m(2)) subjects (median age 56 years at baseline) were at higher risk of cancer of the corpus uteri (OR = 3.0; 95% CI 0.9-10.6), colon (OR = 1.9; 95% CI 0.8-4.5) and breast (OR = 2.5; 95% CI 1.3-4.2). For younger obese women (median age 30 years at baseline), an inverse tendency was observed for breast cancer (OR = 0.6; 95% CI 0.3-1.5, p for trend = 0.05). The tallest women had an increased risk of breast (OR = 1.8; 95% CI 1.3-2.7) and ovarian cancer (OR = 1.7; 95% CI 0.8-3.5). No consistent associations were found for prostate cancer either for BMI or height. There are some suggestions in our study that uncontrolled genetic or early shared environmental factors may affect risk estimates in studies of anthropometric measures and cancer risk, but do not explain observations of increased cancer risks related to BMI or height.