Background: Bone metastases are a common site of distant recurrence in breast cancer. Evidence from randomized trials, including a recent meta-analysis, suggests that adjuvant bisphosphonates can decrease recurrence and death. SWOG S0307 compares efficacy of 3 bisphosphonates in early stage breast cancer. Methods: Patients with stage I-III breast cancer receiving adjuvant systemic therapy were randomized to receive 3 years of clodronate (CLOD) (1600 mg po qd), ibandronate (IBAN) (50 mg po qd) or zoledronic acid (ZA) (4 mg IV q month x 6, then q3 months x 2.5 years). The primary endpoint is disease-free survival. Overall survival, sites of first recurrence, and adverse events are secondary endpoints. Results: Between Nov 2005 and Feb 2010 6,097 patients were enrolled. Survival data are maturing with 50% of the expected events occurring to date. Annual interim analyses are being conducted. 5,752 patients are assessable for toxicity. Median age was 53. 78% of tumors were ER positive, 16% HER2 positive, 16% triple negative. 34% were stage I, 45% stage II, and 21% stage III. Planned adjuvant treatment included chemotherapy in 80%, endocrine therapy in 76%, and both in 56%. 25 (0.4%) and 494 (8.6%) patients experienced grade 4 or 3 toxicities, respectively. The most common adverse events were musculoskeletal, pain, gastrointestinal, metabolic/laboratory (creatinine, calcium), and constitutional symptoms (acute phase reactions). There have been 40 reported cases of osteonecrosis of the jaw (ONJ): ZA 24/2094 (1.15%), CLOD 6/2151 (0.28%), and IBAN 10/1507 (0.66%) (p=0.003). Fractures have been reported in 4.5% of patients in ZA arm, 4.8% in CLOD arm, and 4.1% in IBAN arm (p= ns). Prior to randomization, 76% preferred oral medication versus 24% for intravenous if drugs proved equal in efficacy. Preferences changed little at completion of therapy, although some switched preference. Conclusions: Grade 3 and 4 toxicities were low in S0307. ONJ, a rare but serious complication, was statistically highest for ZA and lowest for CLOD. Fractures were equal across arms. The majority of patients indicated a preference for oral formulation. Clinical trial information: NCT00127205.