“Expression of concern”: publication bias for positive preclinical cardioprotection studies

Abstract

The present analysis reports on the robustness of preclinical cardioprotection studies with infarct size as endpoint which were published in Basic Research in Cardiology, Cardiovascular Research, and Circulation Research between January 2013 and December 2023. Only 26 out of 269 papers with technically robust analysis of infarct size by triphenyltetrazolium chloride staining, magnetic resonance imaging or single photon emission tomography applied a prospective power analysis. A retrospective power calculation revealed that only 75% of the reported data sets with statistically significant positive results from all these studies had a statistical power of ≥ 0.9, and an additional 9% had a statistical power ≥ 0.8. The remaining 16% of all significant positive data sets did not even reach the 0.8 threshold. Only 13% of all analyzed data sets were neutral. We conclude that neutral studies are underreported and there is indeed a significant lack of robustness in many of the published preclinical cardioprotection studies which may contribute to the difficulties of translating cardioprotection to patient benefit.