Metastasis of breast cancers is a common and serious clinical problem, with up to one third of women with earlystage breast cancer eventually dying of the disease. Of all patients with distant spread of the cancer, 70% have bone metastases (1, 2), reflective of the observations of Paget (3) that breast cancers had a particular predilection to grow in bone. A special property required of cancer cells for growth as metastases in bone is the ability to promote bone resorption by inducing the formation and activity of osteoclasts (4). Parathyroid hormone-related protein (PTHrP) is produced by two thirds of primary breast cancers (5), promotes osteoclast formation and bone resorption, and has been implicated in site-specific metastasis of breast cancer to bone (6). The current study investigated the relationship among detection of PTHrP in primary breast cancers, development of skeletal complications, and overall survival. From December 1, 1989, through December 31, 1994, a total of 402 consecutively accrued women with operable (stages 1–3) invasive primary breast cancer were followed prospectively for a median period of 67 months (range, 3–120 months). Impalpable tumors detected by screening mammography accounted for 5% of the cancers. Patients with previous malignancy or preoperative therapy before open biopsy were excluded (35 patients). The timing and site of metastases and overall survival were recorded, as were routine prognostic factors. Patients were examined at least twice a year for 3 years and then annually. The study was approved by the Human Research Ethics Committee of St. Vincent’s Hospital (Fitzroy, Australia), and patients provided informed consent according to institutional guidelines.

There were 367 eligible patients, and the median age at diagnosis was 60 years (range, 27–92 years). The majority of these patients were postmenopausal (70%), the median tumor size was 26 mm (range, 4–180 mm), and most patients were axillary lymph node negative (60%)(Table 1). PTHrP was detected by immunohistochemistry in 265 (72%) of the primary breast tumors and was associated with estrogen receptor (ER) status (odds ratio [OR] 2.4; 95% confidence interval [CI] 1.4 to 4.0; P<. 001), progesterone receptor (PR) status (OR 1.9; 95% CI 1.1 to 3.2; P. 01), and presence of tumor calcification (OR 1.6; 95% CI 1.0 to 2.6; P. 05) but not with menopausal status, tumor size, lymph node status, tumor grade, or the presence of lymphatic/vascular invasion by tumor (7). The crude 5-year cancerspecific survival for all patients was 82%(95% CI 78% to 86%). Prognostic factors associated with survival included American Joint Committee on Cancer stage, tumor size, lymph node status, ER status, PR status, presence of tumor calcification, absence of tumor lymphatic/vascular invasion, and tumor grade but not menopausal status. Patients with tumors containing detectable PTHrP had improved survival compared with patients with PTHrP-negative tumors (Fig. 1, A)(8, 9). The crude 5-year cancer-specific survivals were 87%(95% CI 82% to 91%) for patients with PTHrP-positive tumors and 73%(95% CI 63% to 81%) for those with PTHrP-negative tumors (P. 002). Cox proportional hazards modeling (10) demonstrated that positive PTHrP status was independently associated with improved survival (P. 003), as were the number of axillary lymph nodes involved (P<. 001), tumor size (P<. 001), and tumor grade (P. 007). After adjustment for these independently predictive prognostic factors, the hazard rate for death from breast cancer in women with a PTHrP-positive tumor was 45%(95% CI 26% to 80%) of that in women with a PTHrP-negative tumor (P. 003 …