Coronary Artery Bypass Grafting Versus Percutaneous Coronary Intervention in Ischaemic Heart Failure

Can Reliable Treatment Decisions in High-risk Patients Be Based on Non-randomized Data?

Masafumi Ono; Scot Garg; Yoshinobu Onuma; Patrick W. Serruys


Eur Heart J. 2021;42(27):2665-2669. 

In This Article

Abstract and Introduction


Graphical Abstract: Treatment uncertainties. An observational study cohort includes not only the 'randomizable' patients who are eligible for both PCI and CABG based on contemporary clinical evidence, but also patients who are ineligible for either PCI or CABG. Patients who are ineligible for CABG (e.g. too frail to undergo surgery) have significantly higher mortality risk than those who are ineligible for PCI (e.g. high SYNTAX score). PCI, percutaneous coronary intervention; CABG, coronary artery bypass graft; LVEF, left ventricular ejection fraction; MVD, multivessel disease; LMCAD, left main coronary artery disease; SYNTAX, Synergy between PCI with Taxus and Cardiac Surgery; COPD, chronic obstructive pulmonary disease; PVD, peripheral vascular disease; MT, medical therapy.


Since the RITA trial,[1] the first randomized study comparing percutaneous with surgical revascularization for coronary artery disease (CAD), was published in 1993 controversy has endured over the optimal treatment strategy—be it pharmacological, or revascularization by percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG)—for various types of patients with CAD. Based on previous studies, several factors have been identified and proposed as key elements in deciding between the two, and in the context of precision or personalized medicine, some risk models incorporating these factors are presently advocated to predict and quantify the overall risk of one modality of revascularization vs. the other.[2–5] These models can therefore assist the Heart Team in deciding the most appropriate revascularization strategy.

Recently, we reported on the (re)development and validation of the SYNTAX score II 2020, which uses anatomical and clinical factors to help predict 10-year mortality after PCI or CABG. This tool can assist not only the Heart Team in selecting the appropriate type of revascularization, but also the individual who has most to gain or lose by the decision—the patient.[6,7] In the probabilistic formula of the SYNTAX score II 2020, left ventricular ejection fraction (LVEF) was identified as a key component in the prediction of the treatment benefit of CABG over PCI; however, the score was derived from, and validated in, randomized cohorts, which only enrolled few patients with severe LVEF (Table 1).[8–13] In fact, in the SYNTAX trial, only 1.3% of patients had severe LVEF. Hence, whether such a score can work appropriately in this specific impaired LVEF population is questionable, and still a matter of debate.[5]

In the current issue of the European Heart Journal, Völz et al. report the long-term mortality between PCI and CABG in patients with ischaemic heart failure with reduced ejection fraction [defined by International Classification of Diseases (ICD) codes] and multivessel disease (MVD; defined as coronary artery stenosis >50% in ≥2 vessels) or left main coronary artery disease (LMCAD) from the Swedish Coronary Angiography and Angioplasty Registry (SCAAR).[14] To minimize the selection bias and the other intrinsic limitations of observational studies, the authors used a propensity score-adjusted logistic and Cox proportional hazards regression and an instrumental variable model. Their key findings were that compared with PCI, CABG was associated with a significantly lower risk of mortality [odds ratio 0.62; 95% confidence interval (CI) 0.41–0.96; P = 0.031], with the beneficial survival effect mainly observed 3 years after revascularization. They concluded that their findings reinforce the current European and American guidelines, in which CABG is recommended as the first choice for revascularization in patients with impaired LVEF.[14]

As mentioned above, in this specific clinical setting, the decision between PCI and CABG is still controversial, and consequently the topic of the current study remains of true interest from a clinical point of view. Undoubtedly the high quality of data and sophisticated statistical methodology strengthen the accuracy and validity of the findings of Völz et al. However, it should be also noted that data derived from non-randomized sources such as registries cannot properly account for residual confounding associated with treatment decisions of PCI or CABG left to individual physicians, and that definitive declarations of superiority of one approach (in this case CABG) over another (such as PCI) cannot be made, given the inherent limitations of observational data.