Aggregated adverse-events outcomes in oncology phase III reports: A systematic review.

Abstract

Randomised controlled trials (RCTs) represent a major source of information on treatment-related adverse events (AEs). In this study, we reviewed the use and the reporting methods of aggregated-AEs (A-AEs) outcomes in RCTs reports published in oncology and compared that to the expectations of European Organisation for Research and Treatment of Cancer (EORTC) membership. RCTs reports published between 2007 and 2011 were reviewed regarding the reporting of A-AEs-outcomes. A-AEs were defined as summary outcome combining several related AEs, usually grouped by organ system e.g. cardiac-AEs, dermatologic-AEs. Trial characteristics associated with the use of A-AEs outcomes were investigated. The expectation of EORTC members concerning A-AEs utilisation was queried through a survey. Among 325 RCTs published between 2007 and 2011, 94 (29%) included one or more A-AE outcomes. A clear description of the nature of AEs included in such aggregations was provided in 19 articles (20%). No description of A-AEs was conversely provided in the other 75 articles (80%). The most commonly used A-AEs-outcomes were dermatologic-AEs (45%) and cardiac-AEs (33%). In multivariate analysis, the use of A-AEs outcomes was more frequent when trials were conducted in Europe (p = 0.038) and in trials performed on colon/rectal cancers (p = 0.016). Finally, there is no consensus of EORTC members regarding the utilisation of A-AEs but a majority of them (88%) felt that a clear description of A-AEs should systematically be reported. The use of A-AEs is infrequent in oncology RCT manuscripts although their use is accepted by most clinicians. However, a clear definition of A-AEs is strongly recommended if they are to be used in order to avoid a loss of important details about drug toxicities that is useful to clinicians.