Treatment-related deaths (TRD) in investigator-initiated cancer cooperative group trials: From the datasets of 44 studies (4,964 patients) by the Japan Clinical Oncology Group (JCOG)

Abstract

6632 Background: The proportion of Serious Adverse Events (SAE) including TRDs has a critical impact on choosing a new standard therapy as well as the efficacy of treatment in clinical trials. The proportion of TRD (%TRD) has been reported in a single clinical trial, but an overall % TRD in plural clinical trials has merely been evaluated so far. Methods: The JCOG trials which met both of the following criteria were included in this analysis: i) started after 1995 by 2008, ii) primary analysis was completed. In JCOG, attribution of all reported SAE are classified by Data and Safety Monitoring Committee into 5 categories (definite, probable, possible, unlikely, not related) according to causality to protocol treatment, and the death with causality judged as definite, probable, or possible is defined as TRD. The total number of any deaths and TRD were counted in all included trials and the incidences (%) of TRD were calculated. The association between %TRD and the following trial profiles was evaluated: arm (standard vs. testing), phase, modality and median survival time of each arm. Results: In total, 44 JCOG trials (67 arms) with 4,964 patients were included in the analysis. The number of all deaths and TRD were 2,974 and 60. Thus, the overall %TRD for all enrolled patients was 1.2% (95%CI, 0.9–1.6%), and %TRD among all deaths was 2.0% (95%CI, 1.5–2.6%). Major results of the association between %TRD and trial profiles are shown in the table. The later phase trials showed the lower %TRD, and the trials with shorter MST tended to have higher %TRD. Conclusions: When the trial is in earlier phase and includes more advanced disease, the trial should be planned and performed more carefully about minimization of patient risk. [Table: see text] No significant financial relationships to disclose.