The skeleton is the most common site of metastatic disease in breast cancer and the site of first distant relapse in almost one half of the cases. Bone metastases are the source of a considerable morbidity, including pain, long bone fractures in 10-20%, and hypercalcemia in 10-15% of the cases. The median survival after first relapse in bone is close to two years compared to three months after first relapse in liver. A review of endocrine and chemotherapy trials indicates that patients with metastatic bone disease have a lower response rate to antineoplastic therapy than patients with soft tissue or visceral metastases, but this probably reflects selection bias and the insensitivity of our current methods for evaluating bone response. Classical UICC criteria require radiological recalcification, implying not only tumor regression but also bone healing, which can take many months. Symptom evaluation, measurement of tumor markers and of biochemical parameters of bone turnover should be further investigated for early assessment of bone response. Pain relief could occur in more than half of the patients after radiotherapy, but uncertainty remains as to the relationship between radiotherapy dose or fractionation and the incidence duration of pain relief. Radioactive isotopes have been used successfully in patients with blastic bone metastases from prostate cancer, but controlled studies are lacking in breast cancer. The pathophysiology of metastatic bone destruction makes it logical to use osteoclast inhibitors. Bisphosphonates are potent inhibitors of bone resorption that have opened the way for a noncytotoxic medical treatment of bone metastases. Two large-scale studies in patients with breast cancer metastatic to the skeleton, one with clodronate and one with pamidronate, indicate that the prolonged administration of oral bisphosphonates, in addition to systemic antineoplastic therapy, can reduce the frequency of morbid skeletal events, including the incidence of hypercalcemic episodes and the need for radiotherapy, and probably the incidence of severe pain and of fractures. On the other hand, in patients with established tumor-induced osteolysis, intravenous pamidronate infusions can induce bone pain relief and an objective sclerosis of lytic lesions maybe in one-third and in one-fourth of the cases, respectively. These figures must, however, be taken with caution, and prospective placebo-controlled trials in large series of patients are needed.