Adiposity or gaining weight increases risk of breast cancer in older, postmenopausal women but may reduce risk in premenopausal or young women (Ballard-Barbash, 1994). A recent meta-analysis of 23 studies of current body mass index (BMI) and risk of premenopausal breast cancer concluded that, although there was substantial heterogeneity in the results of these studies, BMI was inversely related to risk (Ursin et al, 1995). Research on breast cancer in young women is inconsistent in demonstrating whether risk is reduced with greater adiposity in early adulthood (ages 18Ð25) (London et al, 1989;Lund et al, 1990) or with subsequent weight gain (Taioli et al, 1995;Mannisto et al, 1996). There has been little research on effects of adolescent body size (Brinton and Swanson, 1992;Ursin et al, 1994).In a case control study of breast cancer in women under the age of 45, we found that thin women had a greater risk of breast cancer than did heavy women . The inverse relation was confined to early-stage breast cancer, but it did not vary by method of breast cancer detection, indicating that it was not explained by earlier detection among thinner women. The purpose of this current study was to examine risk of breast cancer in these young women in relation to adiposity in early adulthood and to subsequent weight gain. In addition, we examined risk in relation to perceived relative height and weight during adolescence and to changes in these perceived relative heights and weights.
MATERIALS AND METHODSThis case control study was conducted in three geographically defined locations in the USA, the metropolitan areas of Atlanta, Georgia and Seattle, Washington, and five counties in central New Jersey (Brinton et al, 1995). The protocol was approved by human subjects review boards at each collaborating institution. All women aged 20Ð44 years newly diagnosed with a first primary in situ or invasive breast cancer in these areas during 1 May 1990 to 31 December 1992, were rapidly identified through frequent review of hospital records. Of 1940 eligible cases identified, we obtained interviews from 1668 (86.0%). Reasons for noninterview included subject refusal (6.7%), physician refusal (5.8%), death or illness (0.8%) or other reasons (0.7%). Eighty per cent of the cases were interviewed within 6 months of diagnosis.Control subjects were ascertained through 13 waves of random digit telephone dialing. Sampling probabilities were varied depending upon the age distribution of women in the areas and the projected distributions of cases. A short telephone screener to identify potentially eligible controls was completed at 90.5% of 16 254 telephone numbers assessed as residential, and a random sample stratified by 5-year age groups was obtained. Of 1989 controls selected, 82 were subsequently found ineligible, and interviews were completed on 1500 (78.7%). The overall response rate (screening times interview) was 71.2%. Summary Findings have been inconsistent on effects of adolescent body size and adult weight gain on risk of breast cancer ...