Discrepancies in the results reported for multiple sclerosis clinical trials: A comparison between ClinicalTrials.gov and peer-reviewed journals.

Abstract

We aimed to compare the results of phase III and IV clinical trials examining drugs to treat multiple sclerosis (MS) registered at ClinicalTrials.gov to those published in peer-reviewed journals. After identifying trials registered at ClinicalTrials.gov, consecutive searches were conducted in PubMed, EMBASE and Google Scholar for matching publications. Information regarding participants and efficacy and safety results was extracted and compared. The degree of consistency was classified as 'concordant', 'discrepant' or 'not comparable'. The Kaplan-Meier method was used to model time to reporting. In total, 65 trials were appraised. The median time from completion to reporting was shorter for ClinicalTrials.gov (16.4 vs 27.3 months; p = 0.010). Information availability was generally higher in journals except for serious adverse events (SAEs) (86.2% vs 100.0%, p = 0.029) and their description (78.2% vs 100.0%, p < 0.001). However, 45 trials had at least one reporting discrepancy (69.2%). Three studies omitted one or more primary outcomes in the matching journal publication. Regarding safety results, the lowest consistencies were found for causes of death (60.0%) and description of SAEs (27.9%). Consulting both ClinicalTrials.gov and journals increases the accessibility to MS clinical trial results. Some data were frequently missing or disagreed between sources, raising concerns about transparency and generalizability of results.