Article Impact Level: HIGH Data Quality: STRONG Summary of BMJ, 389, r1201. https://doi.org/10.1136/bmj.r1201 Dr. Peter Doshi et al.
Points
- A recent investigation into the ONSET/OFFSET and RESPOND trials for the heart drug ticagrelor has uncovered significant evidence of scientific misreporting, challenging the drug’s foundational evidence base.
- The RESPOND study’s primary endpoint was reported as statistically significant with a P-value of 0.005, while the original protocol analysis showed a non-significant P-value of 0.157.
- Over 60 platelet readings from a primary investigator’s machines were missing from FDA datasets, and these omitted measurements indicated significantly higher platelet activity than what was officially reported.
- Serious procedural flaws included a listed author who denied any participation in the research and a lack of confirmation that investigators received the necessary training for complex laboratory tests.
- These findings highlight a paradox where the drug’s purported superior platelet inhibition did not result in better early outcomes for specific patient groups in the larger PLATO clinical trial.
Summary
A recent investigation by The BMJ has raised significant concerns regarding the scientific integrity of two key pharmacodynamic studies that supported the approval and widespread use of ticagrelor. Building on a prior investigation that found data integrity problems in the landmark 18,000-patient PLATO trial, this analysis focused on the ONSET/OFFSET (123 patients) and RESPOND (98 patients) trials. These studies were pivotal in establishing ticagrelor’s mechanism of rapid and superior platelet inhibition compared to clopidogrel. The investigation, which analyzed trial data submitted to the U.S. Food and Drug Administration (FDA) and documents from equipment manufacturers, uncovered evidence of serious misreporting and procedural flaws that question the validity of the trials’ conclusions.
The analysis identified multiple critical data discrepancies. For the RESPOND trial, the primary endpoint was reported in a major cardiology journal as statistically significant (P = 0.005), whereas the protocol-specified analysis was non-significant (P = 0.157); this discrepancy was attributed to an undeclared change in the endpoint definition. In the ONSET/OFFSET trial, the published “intention-to-treat” analysis improperly excluded seven patients who lacked baseline data. Furthermore, a review of raw data from platelet analysis machines at one primary site revealed that over 60 readings (approximately 25% of the total) were absent from the datasets submitted to the FDA. These omitted readings showed significantly higher platelet activity than what was included in the final analyses.
Numerous procedural and methodological irregularities compound these findings. The investigation noted authorship discrepancies, including a listed author who denied participation and an active investigator who was not credited. It remains unclear if adequate, standardized training for the technically demanding, multi-site platelet aggregation assays was ever conducted. These issues amplify a paradox previously identified by an FDA reviewer: the purported superior platelet inhibition in these studies did not correlate with improved early clinical outcomes for patients undergoing percutaneous coronary intervention in the PLATO trial. Collectively, the documented irregularities challenge the reliability of the evidence base that has supported the use of ticagrelor for over a decade.
Link to the article: https://www.bmj.com/content/389/bmj.r1201
References Doshi, P. (2025). Ticagrelor doubts: Inaccuracies uncovered in key studies for AstraZeneca’s billion dollar drug. BMJ, 389, r1201. https://doi.org/10.1136/bmj.r1201
