Calcium's ability to prevent bone loss in early postmenopausal women is controversial. We used data on 394 women from the placebo group of the Early Postmenopausal Interventional Cohort study, a clinical trial of alendronate, to investigate the relation of calcium intake to bone loss. Calcium intake was recorded, and bone mineral density (BMD) (in the lumbar spine, total body, forearm, and hip) and biochemical markers of bone turnover (serum total alkaline phosphatase, serum osteocalcin, and urinary N-telopeptide crosslink levels) were measured at baseline and annually thereafter. Women whose baseline calcium intake was <500 mg/d were advised to increase their calcium intake. Mean (+/- SE) BMD decreased by 1.9% +/- 0.16% at the lumbar spine and 1.6% +/- 0.14% at the hip over the 24-month period. Despite wide variations in baseline calcium intake and changes in calcium intake, these measures were not significantly associated with changes in BMD or bone turnover. Even women whose total calcium intake was >1333 mg/d (the highest tertile of total calcium intake) showed a decline in BMD of almost 2%, similar to declines in the lower two tertiles of total calcium intake (<869 and 869-1333 mg/d, respectively). Increased calcium intake resulted in modest mean increases of approximately 200 mg/d. We were unable to demonstrate that increases of this magnitude or much greater (1 g/d) were protective against declines in BMD at any site, even in women who had the lowest calcium intake at baseline. In addition to adequate calcium intake, more effective therapy appears to be required when the therapeutic goal is to increase or maintain BMD.