Adverse event reporting in publications compared with sponsor database for cancer clinical trials.

PURPOSE: Prospectively planned collection and analysis of adverse event (AE) data are essential parts of well-conducted clinical trials. The AE data in a trial sponsor's database should be comparable with what is stipulated in the protocol and with the AE data published. We examined whether the published AE data differ from those in the sponsor's database and from the data collection requirements stated in study protocols.
METHODS: We searched the National Cancer Institute (NCI) Clinical Data Update System (CDUS) for studies that used the Common Toxicity Criteria version 2.0 and for which a final study publication was available. We extracted from the protocols information pertaining to AE collection and reporting methods and compared it with the methods cited in the article. We also compared the AE data in the trial publication with the AE data submitted by the investigators to CDUS.
RESULTS: We identified 22 studies meeting the criteria for this review. There was considerable inconsistency between AE collection and reporting methods cited in the protocols versus final publications. AE data in the article and CDUS were not identical. Twenty-seven percent of article high-grade AEs could not be matched to agent-attributable AEs in the CDUS. Twenty-eight percent of CDUS high-grade AEs could not be matched to AEs in the corresponding article. In 14 of 22 articles, the number of high-grade AEs in CDUS differed from the number in the articles by 20% or more.
CONCLUSION: Lack of consistency in and reporting of AEs are associated with NCI database and trial publication AE data discrepancy.