Matthias Siepe, Editor in Chief, said:
Today, EJCTS has published the results of an independent review of the meta-analysis entitled, Radial artery versus saphenous vein versus right internal thoracic artery for coronary artery bypass grafting, that was originally published in EJCTS in June 2022.
The original paper by Gaudino et al. was based on a meta-analysis of individual patient data comparing outcomes using different grafts for coronary artery bypass surgery.
Following publication of the original paper by Gaudino et al, several concerns were raised with us, including the appropriateness of the methodology that was selected for the analysis. After careful consideration, we commissioned Professor Nick Freemantle’s Comprehensive Clinical Trials Unit at the Institute of Clinical Trials and Methodology, University College London, to conduct a further review.
Further analysis conducted by UCL showed that:
- The results were similar to those identified by Gaudino et al when using the same methodology, confirming that the data handling was numerically and statistically correct.
- Differences were noted between individual matched groups using propensity score matching across three groups. Supportive conventional analyses found was associated with better matching and didn’t find systematic difference on the outcome of mortality when comparing the different grafts used in CABG.
This is an outstanding example of open science: editorials, letters to the editor and social media questioned a paper. The authors were open and shared their entire dataset and code, in compliance with the author guidelines, for independent re-analysis. These supporting analyses give a clearer picture of the robustness of the initial findings, and paint a clearer picture of how to interpret the results.
We have published the additional analysis today in the interests of open science and to ensure as comprehensive a picture as possible is available publicly to help inform the best treatment options for patients.
The original authors should be applauded for sharing their entire dataset and codes with us for re-analysis, and we will welcome further responses on this issue for publication in EJCTS.