Milan Milojevic

Chapter 1 is a general introduction to this thesis. This chapter gives an overview of the epidemiology and current trends in myocardial revascularization in Europe. Although both percutaneous coronary intervention (PCI) and coronary artery bypass grafting (CABG) are used in patients with left main (LM) disease and/or multivessel disease, the optimal patient selection and individualized medication strategies are crucial to ensure improved outcomes. Rigorously developed evidence-based recommendations in clinical guidelines can improve decision-making and quality of health-care. The studies in this thesis should inform clinicians about specifically choosing revascularization strategies for patients with coronary artery disease. The aims and outline of this thesis are described in chapter 2.

Part 1. Current Practice in Bypass Surgery
Chapter 3 describes the outcome and life expectancy of the first venous CABG procedures during 40 years of follow-up. The 10-, 20-, 30and 40-year survival for the 1041 patients who underwent CABG between 1971 and 1980 was 77%, 39%, 14% and 4%, respectively. Average life expectancy was 18 years while repeat revascularization was performed in 36% of patients. Factors associated with decreased late survival were the age at operation, diabetes mellitus (DM), multivessel disease and left ventricular ejection fraction under 50%. However, over the last four decades, surgical techniques have improved, increasing its safety and efficacy while at the same time reducing invasiveness and re-intervention demands. The proper use of surgical techniques is most closely linked to the revascularization success or failure. A critical evaluation of contemporary indications, techniques, and outcomes of bypass surgery are discussed in Chapter 4. This review suggests that, despite its improvements, several techniques for CABG surgery can be adopted more widely to further improve outcomes: use of intraoperative graft flow assessment, epiaortic scanning, more use of arterial conduits, and hybrid revascularization.

To steer the choice of the most optimal technique, clinical practice guidelines are one of the most valuable tools. One of the major advances has been a class IC recommendation for the multidisciplinary “Heart Team” decision-making process in the North American and European guidelines for revascularization. Evidence to support the Heart Team is scarce, and some clinicians fear that the requirement of a Heart Team discussion can delay treatments and be unsafe for patients. Chapter 5 provides evidence that real-world Heart Team meetings are feasible and safe in evaluating coronary artery disease complexity and additional comorbidities along with patient’s preferences to guide the most appropriate revascularization strategy. Approximately 90% of patients received treatment within 6 weeks, as recommended by the 2014 European Society of Cardiology (ESC) / European Association for Cardio-Thoracic Surgery (EACTS) Guidelines on myocardial revascularization. Delay was caused by the need for additional diagnostic tests to be performed, showing that logistics can be further improved.

Part 2. Bypass Surgery versus Stenting
Several multicenter randomized clinical trials (RCTs) have examined the clinical effects of PCI versus CABG across patients with multivessel and/or LM disease. One of these trials was the Synergy between Percutaneous Coronary Intervention with Taxus and Cardiac Surgery (SYNTAX) trial; a large randomized multicenter study comparing PCI with first-generation drug-eluting stents (DES) (TAXUS™ Express™, Boston Scientific) to CABG for patients with three-vessel and/or LM disease. A significant contribution of the SYNTAX trial was the development of the SYNTAX score, which has become a unique tool to score the complexity of coronary artery disease (CAD) and to help guide decision-making between PCI and CABG.

The RCTs that have been performed used various composite endpoints with a varying clinical impact to boost the statistical power, but no study has been adequately powered to examine the mortality differences as the primary outcome. Chapter 6 provides an individual patient-data pooled analysis of 11518 patients from 11 RCTs, which showed a significantly higher 5-year mortality rate after PCI than after CABG in patients with multivessel disease, especially in those with diabetes and higher coronary complexity according to the core laboratory SYNTAX scores. Furthermore, in patients with LM disease, no difference in mortality rate was seen between two treatment groups. In Chapter 7, the specific cause of mortality was examined based on data from the SYNTAX trial, demonstrating that cardiac death due to spontaneous myocardial infarction (MI) was markedly higher after PCI with TAXUS compared to CABG at 5-year of follow-up.

Stroke following PCI and CABG, although rare, can be a devastating complication, associated with high rates of mortality and reduced health-related quality of life. In Chapter 8, we found, among 11518 patients randomized to CABG or PCI with stents, that PCI was associated with significantly lower 30-day stroke rates compared to CABG (0.4% versus 1.1%, respectively), but no difference was found between two treatment groups beyond 30-day after the procedures. Comparing CABG with PCI, diabetes had a significant effect on the occurrence of stroke during 5 years of follow-up (2.6% versus 4.9%, P for Interaction = 0.004). The reason for the higher rates of stroke after CABG may be multifactorial, including the use of antifibrinolytics to reduce the risk of bleeding after surgery. Chapter 9 is a letter to the editor that stresses the importance of the correct intraoperative dose of antifibrinolytics-tranexamic acid (TXA) agent to prevent bleeding complications after CABG, but also highlights that the routine use of TXA may influence the perioperative stroke occurrence in CABG.

In chapter 10, additional analyses from the SYNTAX trial were conducted to explore the impact of repeat revascularization on the 5-year clinical outcome. Rates of repeat revascularization are higher after PCI compared with CABG at all time points. Our study reports that at 5-year follow-up, repeat revascularization rates were significantly higher after PCI compared to CABG (13.7% versus 25.9%, P<0.001), showing a significant correlation between any repeat revascularization after an initial PCI procedure and increase in the incidence of serious adverse events. Moreover, long-term results have also demonstrated a significantly higher need for multiple repeat revascularization after an initial PCI than after CABG (9.0% versus 2.8%, P=0.022; respectively). Independent predictors of repeat revascularization were diabetes, incomplete revascularization, the number of overlapping stents and absence of antiplatelet therapy among patients randomized to PCI while the treatment in the United States and the use of off-pump technique were reliable predictors of repeat revascularization in the CABG group. Apart from the individual endpoints of mortality, myocardial infarction, stroke, and repeat revascularization, clinical trials often use composite endpoints to increase the statistical power of the analyses. Individual endpoints in composites are weighted equally, while the different individual components have evident varying impacts on long-term prognosis. Therefore, several novel approaches to assess the results of composite endpoints have been introduced. In chapter 11, a win ratio approach is applied on the SYNTAX trial to provide additional clinical insights into the results of the primary composite endpoint of mortality, stroke, MI or repeat revascularization, also evaluating strengths and weaknesses of alternative methods for the analysis of composite endpoints. This study demonstrates that the critical advantage of CABG over multivessel PCI is the reduction of hard clinical endpoints such as mortality and MI. Moreover, this approach is readily applicable to analyze composite endpoints with multiple distinct events, while maintaining the integrity of the study results. Multiple factors can influence treatment decision-making. Patients with chronic kidney disease (CKD) and/or diabetes have a high prevalence of CAD and high risk of cardiovascular mortality. Whether the use of PCI or CABG would improve patient survival in patients with these associated diseases remains uncertain due to limited data from randomized comparisons and conflicting data from observational studies. To address this knowledge gap, the effect of CKD on 5-year outcome after PCI and CABG in the SYNTAX trial has been investigated in chapter 12. This subgroup analysis shows that CABG appears to be the favorable revascularization strategy over PCI, mainly supporting the more significant use of CABG among diabetic patients complicated by CKD. A randomized study of 1905 patients with LM disease, enrolled in the Evaluation of XIENCE versus Coronary Artery Bypass Surgery for Effectiveness of Left Main Revascularization (EXCEL) trial compared CABG with second-generation DES. Patients with diabetes had a significantly higher rate of the 3-year composite primary endpoint of death, stroke or MI (12.9% versus 20.0%, P<0.001) (chapter 13). However, there was no difference in the primary endpoint between PCI and CABG in diabetic patients (20.7% versus 19.3%, P=0.87) and non-diabetic patients (12.9% versus 12.9%, P=0.89), suggesting that in selected diabetic patients with LM, PCI may be a reasonable treatment approach beyond CABG. The globalization of clinical trials has emerged as a new phenomenon describing the movement of trial location to lower-income countries to decrease costs and accelerate recruitment of trial participants. One of the major concerns is that the imbalance between the quality of care, patient health levels, treatment choice, and hospital infrastructures may influence the overall generalization of the trial results. Chapter 14 focuses on the influence of practice patterns on outcomes in specific countries within the SYNTAX trial. We found that baseline characteristics of included patients and clinical practice patterns are substantially different between participating countries, resulting in a significant difference in clinical outcomes, for which specific treatment recommendations were provided. Furthermore, relevant aspects for future trial design are discussed. Part 3. Improving Outcomes in Cardiac Surgery Advances in the whole spectrum of hospital care have resulted in significant improvements of in-hospital outcomes in patients undergoing CABG. However, CAD is a chronic progressive process that requires intensive postoperative medication therapy to slow down the progression of the disease and reduce the risks of future cardiovascular events. Clinical guidelines are developed to help in decision-making by providing recommendations that are supported by the best available evidence. Along with increasing awareness of clinical outcomes between different treatment modalities, clinical trials can provide valuable information about the use of guideline-recommended medical therapy (GDMT) in daily practice. In chapter 15, based on an individual patient-data analysis of 7085 patients from 5 RCTs, we studied the compliance with GDMT after myocardial revascularization. The pooled analysis shows the suboptimal use of GDMT after CABG and significant correlation between the optimal use of medications and risks of adverse clinical outcomes at 5 years of follow-up. Moreover, in chapter 16, we provide the evidence-based recommendations for perioperative medical therapies in adult cardiac surgery. Chapter 17 is a letter to the editor that discusses the evidence used to answer the question regarding stopping or continuing acetylsalicylic acid (ASA) until the day of CABG. The main finding of this meta-analysis was opposite to our clinical guideline recommendations. Therefore, methodological comments on the inclusion and exclusion criteria for given meta-analysis and the most important trials in this field are put into perspective to substantiate the recommendations in our treatment guidelines.