In a retrospective study 2161 patients with histologically proven breast cancer following up maximum 6 1/2 years after initial treatment were reevaluated. Of all distant metastases, osseous metastases were most frequent with 11.5%. From these we found simultaneously non-osseous metastases in 55%, but only in 7% of patients without osseous metastases. Skeletal metastasis of 2107 patients with known tumor status increased with the primary tumor size (pT1-4) and the extent of the nodal involvement (pN0-3). Prognosis of patients without metastatic lymph nodes (pN0) was only slightly affected by the tumor size, the five-year disease free period decreased from 0.93 (pT1) to 0.84 (pT2), and 0.78 (pT3). However, the prognosis was poorer in patients with primarily lymph node involvement, in these cases, the tumor size was a prognostic discriminant. The five-year osseous metastases-free period decreased from 0.80 (pT1) to 0.68 (pT2), 0.46 (pT3), and 0.11 (pT4). Independent of the primary tumor size (pT1-4), the prognosis was deteriorated by the primary lymph node status, the metastases free period decreased from 0.88 (pT0) to 0.77 (pT1), 0.53 (pN2), and 0.35 (pN3). Thus, the lymph node status beside the tumor size is considered to be the most prognostic criterion with respect to patients developing metastatic bone disease. The routine follow-up by bone scintigraphy without clinical suspicion of osseous metastases is necessary regardless of tumor size if an extended lymph node involvement (pN2, pN3) or a large primary tumor (pT3, pT4) has been detected. Not required are scintigraphic follow-up studies in patients with a small tumor size (pT1) but without metastatic lymph nodes. A facultative use of bone scanning seems to be justified in stage pT1N1 and pT2N0.