Potential Negative Effects of all the REDUCE-IT trial statistical analyses being initially performed by company employees of Amarin Corporation.
The statistical analyses in the REDUCE-IT trial1 (CV Risk Reduction with Icosapent Ethyl for Hypertriglyceridemia) were initially performed by the company statisticians, and then validated by an independent statistician.
“All the data analyses were performed by the sponsor, and the primary, secondary, and tertiary adjudicated end-point analyses were validated by an independent statistician from the data and safety monitoring committee.”1
For a large randomized trial, one would expect that even under this arrangement, the statistical analyses would be accurately performed. However, it is somewhat different to validate, rather than to initially perform the analyses.
Furthermore, any post hoc analyses not specifically planned prior to the trial would be potentially influenced by who did the initial statistical analysis given the free form nature of data interrogation. (Even with oversight, post hoc analyses, including the decision of the data parameters for the analysis, would be strongly influenced by who does the primary data interrogation.)
Assigning the initial statistical analyses of data for a major randomized trial to the company owning the study medication does not appear optimal.
Conflict of Interest Information for the REDUCE-IT trial: Of the 12 primary authors listed for the scientific article providing the results for the REDUCE-IT trial, 5 of the 12 authors are employees with stock ownership in Amarin Corporation (Vascepa).1 An additional 3 of the 12 authors were consultants for the company, including outside the REDUCE-IT study.1
1. Bhatt DL, Steg PG, Miller M, et al. for the REDUCE-IT Investigators. Cardiovascular risk reduction with icosapent ethyl for hypertriglyceridemia. N Engl J Med 2019;380:11-22.
