• BMI;
  • weight;
  • breast cancer;
  • methodological issue;
  • HRT use;
  • cohort study


The association between weight, BMI and breast cancer was analyzed on 94,805 women of the E3N cohort according to their menopausal status. Seven hundred eighty-six incident invasive premenopausal breast cancers and 1,522 incident invasive postmenopausal breast cancers occurred during a mean follow-up of 9.7 years. Weight and BMI were updated every 24 months and considered as time-dependent variables. Data were analyzed using multivariate Cox proportional hazards models. Trend RRs of premenopausal breast cancer were 0.97 (0.92–1.01) for a 5 kg increase in weight and 0.96 (0.91–1.01) for a 2 kg/m2 increase in BMI, adjusted for other known risk factors. Opposite trend RRs were found after menopause: 1.05 (1.02–1.08) for weight and 1.06 (1.02–1.09) for BMI, respectively, for similar increases. Women with a BMI of over 30 kg/m2 had a RR of premenopausal breast cancer of 0.66 (0.40–1.10) compared to those with a BMI of between 18.5 and 25 kg/m2. Postmenopausal women with a BMI of over 30 kg/m2 had a RR of breast cancer of 1.23 (1.00–1.59). The increase in risk of postmenopausal breast cancer with increased weight or BMI was similar whatever the HRT used, although the point estimates were higher in HRT users. We strongly recommend to use anthropometric measurements updated during follow-up to assess the effect of weight, BMI on breast cancer risk. © 2004 Wiley-Liss, Inc.