Fig. 3.

Metaregression-analysis on the composite endpoint of death, myocardial infarction (MI) and stroke. Panel a Meta-regression analysis of all included studies, showing a significant inverse interaction of the effect size with the proportion of diabetics included in single studies, indicating that CABG performed better than PCI in studies that included a larger number of diabetics. Panel b Meta-regression analysis of all included studies, showing a significant inverse interaction of the effect size with the proportion of female patients included in single studies, indicating that CABG performed better than PCI in studies that included a larger number of females. Panel c Meta-regression analysis of all included studies, showing a non-significant direct interaction of the effect size with mean patients’ age from single studies, suggesting that PCI may perform better than CABG in older patients. Panel d Meta-regression analysis of all included studies, showing a non-significant direct interaction of the effect size with mean left ventricular ejection fraction from single studies. Each study is represented by a circle that shows the effect size. The area of each circle is proportional to that study’s weight in the analysis