Fatal events during clinical trials: an evaluation of deaths during breast cancer studies

Background Information on deaths occurring during oncological clinical trials has never been systematically assessed. Here, we examine the incidence of death and the profile of patients who died during randomized clinical breast cancer (BC) trials. Methods Information on fatal events during German Breast Group (GBG) led BC trials was prospectively captured. Data were derived from the trial databases and death narratives. All deaths were evaluated for possible causes, underlying conditions, treatment relatedness, time point and rate of autopsies. Results From 12/1996 to 01/2017, 23,387 patients were treated within 32 trials. Of those 88 (0.4%) died on therapy within 17 trials. Median age was 64 [range 35–84] years, 63.2% of patients had a body mass index (BMI) ≥ 25 kg/m 2 ; 65.9% 1–3 and 22.7% ≥ 4 comorbidities; 61.4% 1–2 cardiovascular risk factors (CRFs); 26.4% took > 3 drugs; 81.7% had ECOG 0; 50.0% stage III, 76.7% luminal BC. The main causes of death were infection (38.6%; of those, 82.3% sepsis, 17.6% pneumonia), heart failure (14.8%), and pulmonary embolism (13.6%). Fatal events mainly occurred within the first 4 therapy cycles (55.7%), in the investigational arm (66.7%) and under anthracycline–taxane-based chemotherapy (51.1%). A relationship with the treatment was declared in 27.3% of the cases. An autopsy was performed in 13.6% of patients. Conclusions Death during study treatment was mainly related to infections, and patients with advanced disease, high BMI, underlying comorbidities, CRFs and concomitant medications. If considered for study participation these patients need careful monitoring due to their higher risk for death on study.