Experimental and epidemiological studies suggest that calcium-regulating hormones--parathyroid hormone (PTH) and vitamin D--may be associated with breast cancer risk. No prospective cohort study has investigated the association between pre-diagnostic calcium levels and subsequent risk of breast cancer. We have examined this in a cohort of 7,847 women where serum calcium levels and established risk factors for breast cancer had been assessed at baseline. During a mean follow-up of 17.8 years, 437 incident breast cancer cases were diagnosed. Incidence of breast cancer was calculated in different quartiles of serum calcium levels and a Cox's proportional hazards analysis was used to obtain corresponding relative risks (RR), with a 95% confidence interval (CI), adjusted for potential confounders. In premenopausal women, serum calcium levels were inversely associated with breast cancer risk in a dose-response manner. The adjusted RR (95% CI) of breast cancer was in the 2nd calcium quartile 0.91 (0.65-1.30), in the 3rd quartile 0.89 (0.60-1.31), and in the 4th quartile 0.56 (0.32-0.98), as compared to the 1st calcium quartile. In postmenopausal overweight women (BMI > 25), breast cancer risk was higher in calcium quartiles 2-4 as compared to the 1st quartile. Our findings may have implications for primary prevention of breast cancer and for the management of asymptomatic primary hyperparathyroidism.