Conflicts Of Interest In Pharmaceutical Sponsored Research On Erythropoietin Receptors In Cancer: An Updated Analysis

Abstract

BackgroundFinancial conflicts of interests in clinical studies have been a perennial subject of investigation because of their potential to harm the public's health. In 2010, we reported the first systematic analysis examining such conflicts in preclinical research, identifying that academic researchers without financial conflicts of interest significantly differed from academic researchers with financial conflicts of interest as well as scientists employed by manufacturers of erythropoiesis stimulating agents (ESAs). Those with no conflict of interest identified functional erythropoietin (Epo) receptors on tumor cells and that these receptors when activated carried out detrimental cell signaling and tumor promotion effects. The analysis was based on published studies from 1993 to 2008. In 2008, an NCI-sponsored conference with 14 academic scientists and 6 manufacturer employed-scientists restated these findings. We now review the status of the peer-reviewed published literature that has been disseminated following the NCI conference. MethodsArticles identified in MEDLINE and EMBASE databases (2008-2012) investigating preclinical findings were reviewed for information on Epo receptors, signaling events, cellular function, and study conclusions. Study findings were classified into 3 groups according to funding source. (1) Academic-based studies authored by investigators without manufacturer funding [13 studies], (2) academic-based studies authored by investigators with funding from ESA manufacturers [8 studies], and (3) industry-based studies authored by more than 75% of investigators employed by ESA manufacturers [2 studies]. ResultsStudies authored by investigators in groups 1 and 2 report: Epo-induced signaling events in tumor cells (84.6% and 100.0%, respectively) and conclude that ESAs could be clinically harmful (83.3% and 57.2%, respectively); while no investigators in group 3 report Epo-induced signaling events in tumor cells or conclude that ESAs could be clinically harmful. Using univariate optimal discriminate analysis, 6/9 studies authored by investigators in groups 2 and 3 found no harmful changes in cellular function and is statistically different from investigators in group 1 who found harmful changes in cellular function in 11/12 studies (p<.02). ConclusionsConvergence has occurred over time regarding the harmful conclusions of ESAs and cancer progression among academic scientists with versus without financial conflicts of interests. However scientists employed by ESA manufacturers still report no detrimental basic science findings of ESAs on tumor cells. This claim is of considerable importance to the ESA manufacturers, as FDA advisories have indicated that should detrimental preclinical effects of ESAs be found on certain tumor cell lines, then the advisory committees may consider particular caution in the use of ESAs among persons with these specific tumor types. Disclosures:Bennett:Amgen: Consultancy, Research Funding.