Background: Information on deaths occurring during oncological clinical trials has never been systematically assessed to describe patient, tumor, and treatment characteristics. Methods: Information on patients´ deaths during German Breast Group (GBG) led breast cancer (BC) trials was prospectively captured. In addition to the trial databases, data were derived from death narratives that included autopsy results if performed. All deaths were evaluated for possible causes, underlying conditions, treatment relatedness, time point and rate of autopsies. Results: From 12/1996 to 12/2014 23,570 patients were treated within 31 trials. Of those 75 (0.3%) died on therapy within 14 trials. 29/12,956 patients died in neoadjuvant (0.2%), 35/9,851 in adjuvant (0.3%) and 11/763 in metastatic studies (1.4%). Median age was 64 yrs (range 35-84), 63.5% of patients had an abnormal BMI (25.7% underweight, 32.4% overweight, 5.4% obese); 60% had 1-3 and 13.3% ≥ 4 comorbidities; 56% had 1-2 cardiovascular risk factors (CRFs); 45.3% took ≥ 3 drugs; 80% had an ECOG 0. Over 50% of patients had a stage III tumor at baseline with a luminal B-like BC subtype (HR+/Her2-/G3). More patients with advanced disease had a high BMI (p = 0.024), ≥ 3 comorbidities (p = 0.013) and CRFs (p = 0.001) compared to early stage patients. Main causes of death were infections (34.7%; febrile neutropenia 4%), cardiac (14.7%) and respiratory disorders (12%). 13% of patients (4 in metastatic BC, 6 in early BC) had a disease progression. 42.5% of patients received taxane (T)-based chemotherapy (CT) and the event mostly occurred in the first 4 cycles. Relatedness to chemotherapy was declared in 55% of patients, mainly when a T-based CT was given (51.2% for T-based CT, 14.6% for anthracycline (A)-based CT, 22% for A-T-based CT, 12.2% for other regimens (capecitabine, bevacizumab, lapatinib, CMF, celecoxib), none for hormone therapy); p = 0.001). An autopsy was performed in 13% of patients. Conclusions: Death during study treatment was mainly related to infections, and patients with advanced disease, high BMI, underlying comorbidities and CRFs. If considered for study participation these patients need careful monitoring due to their higher risk for death on study.