CHAPTER 44: CORONARY ARTERY BYPASS GRAFTING AND PERCUTANEOUS INTERVENTIONS IN STABLE ISCHEMIC HEART DISEASE

Great advances have been made in the consideration of coronary revascularization for patients with stable ischemic heart disease (SIHD). Perhaps the greatest advance over the past decade has been the establishment of the heart team approach. This is when clinical cardiologists, interventional cardiologists, and cardiovascular surgeons, along with nursing staff, consider together the revascularization options for our patients with SIHD. The heart team approach facilitates better engagement of the patient in the informed consent and decision-making processes.

The diagnosis of SIHD is based on symptomatology and primarily obtained through the history of the patient (see Chap. 43). The traditional risk factors for the development of obstructive coronary artery disease (CAD) are ascertained to estimate its likelihood. Stable angina often presents with retrosternal heaviness and aching, which may or may not radiate to the jaw(s) or shoulder(s) and is provoked by exertion or emotional stress and relieved within 5 minutes of rest or nitroglycerine use. Patients with diabetes and other chronic conditions may present with atypical symptoms including the feeling of general malaise after exertion. In combination with the physical examination and 12-lead electrocardiogram (ECG), the important risk markers of hemoglobin A1C, lipid profile, renal and liver function tests, thyroid function tests, and hemoglobin should all be part of the initial evaluation.

The identification of ischemic heart disease has largely relied on the documentation of ischemia by noninvasive testing (see Chap. 43). Noninvasive testing largely pertains to patients who are at intermediate risk for establishment of SIHD. Because the workup of noninvasive testing will be the topic of another chapter, it is important to emphasize that noninvasive testing is reserved for men aged over 40 years and women aged over 60 years with a strong history of the presence of angina. With the advancement of noninvasive technology spanning into computed tomography (CT), stress methoxyisobutylisonitrile (MIBI), and echocardiography, the most reliable noninvasive test should be accessible and of high expertise. The choice of noninvasive testing is based on whether or not the ECG is normal or whether or not a left bundle branch block or ventricular paced rhythm is present.

All patients with a history of SIHD should be treated with optimal medical therapy (OMT; see Chap. 43), and it is important to follow evidence-based guidelines for OMT for all patients considered for coronary revascularization.¹ Routinely, patients would receive aspirin daily with or without additional P2Y₁₂ inhibitor antiplatelet therapy depending on the history of ischemic acute coronary syndrome. All patients with SIHD should be on a potent statin therapy and at high doses. All patients should be considered for an angiotensin-converting enzyme inhibitor or an angiotensin receptor blocker, in particular patients with low ejection fraction, chronic kidney disease, diabetes, or hypertension. β-Blocker therapy should be considered in patients with SIHD and especially those with an ejection fraction < 40%. The β-blocker dose should be titrated to a heart rate between 50 and 60 bpm.

It is important to establish the reason for consideration of coronary revascularization in SIHD. Coronary revascularization is indicated for the improvement of angina symptoms and quality of life. The evidence for reduction of major adverse coronary and cerebrovascular events and reduction in mortality is largely limited to surgical coronary revascularization (coronary artery bypass grafting [CABG]). When approaching a discussion with the patient, there are a number of factors that need to be considered from the outset.

Coronary angiography is largely reserved for two major groups of patients, namely those who have inadequate symptom relief or persistent reduction in quality of life despite OMT and those with noninvasive testing indicating risk for ischemic events. Currently, the National Heart, Lung, and Blood Institute is sponsoring the International Study of Comparative Health Effectiveness With Medical and Invasive Approaches (ISCHEMIA) trial to evaluate the benefit of coronary angiography in patients with high-risk features on noninvasive testing compared to OMT alone.² This pivotal trial carried out in 10 countries and in more than 200 centers will help us to better define which patients will be referred for consideration of coronary revascularization.

In the meantime, the criteria listed in Table 44–1 will be used to define patients who should be referred for coronary angiography,³ including specific findings on noninvasive testing.

The mode of coronary revascularization is an individualized decision-making process for each patient. Beyond the heart team deliberation, the patient’s own wishes and the balance of risk and benefits for each patient with regard to quality of life, duration of life, and potential adverse events should be considered. The consideration of technical issues, such as the potential for the completeness of revascularization, should also play a role.

The following review is largely based on recommendations of the 2014 American College of Cardiology (ACC)/American Heart Association (AHA)/American Association for Thoracic Surgery (AATS)/Preventive Cardiovascular Nurses Association (PCNA)/Society for Cardiovascular Angiography and Interventions (SCAI)/Society of Thoracic Surgeons (STS) focused update on the guidelines for the diagnosis and management of patients with SIHD.⁴ For the sake of discussion, we will be focusing on contemporary CABG using at least one arterial conduit and contemporary percutaneous coronary intervention (PCI) largely involving drug-eluting stents (DESs).

The extent and complexity of CAD are important parameters informing triage to PCI versus CABG in patients with stable multivessel CAD. The SYNTAX Score (SS), developed for use in the multicenter Synergy Between Percutaneous Coronary Intervention With Taxus and Cardiac Surgery (SYNTAX) trial provides a semiquantitative angiographic measure of disease complexity that is useful for this purpose.⁵ For a given patient, the SS represents a sum of scores for every lesion with > 50% diameter stenosis in vessels with a diameter > 1.5 mm. As outlined in Table 44–2, scores for individual lesions reflect both the anatomic “importance” of the stenosis (derived from population-based estimates of the proportion of myocardium subtended by a given vascular segment) and the presence of features influencing the technical complexity of PCI (chronic total occlusion, bifurcation features, ostial location, lesion length, calcification, tortuosity, and presence of thrombus).^2,6 Accordingly, although the SS does provide a rough index of disease burden in a given patient, it is more accurately interpreted as a metric of the expected complexity of PCI.

The SYNTAX trial demonstrated that such complexity is an important determinant of the comparative effectiveness of PCI versus CABG. Among 1800 patients with previously untreated three-vessel CAD, left main CAD, or both randomized to PCI with a first-generation paclitaxel-eluting stent or CABG, patients with the least complex disease (SS ≤ 22) experienced similar rates of a composite of death, stroke, myocardial infarction, and repeat revascularization with PCI and CABG.⁶ In contrast, patients with more complex disease (SS of 23-32 or ≥ 33) experienced a significantly higher rate of major adverse cardiovascular events with PCI as compared with CABG, driven by excesses of both repeat revascularization and myocardial infarction.

The use of PCI with stenting in the DES era has increased in absolute numbers, but the degree of angiographically documented disease has brought into question whether a threshold exists beyond which CABG is preferred. For each operator and center, this threshold will differ.

With recent advances in PCI platforms, the consideration of optimal revascularization of left main CAD has been the topic of a number of clinical trials. Left main CAD is found in approximately 4% of patients undergoing coronary interventions.

There are a number of features that have made the consideration of a percutaneous approach for left main CAD more attractive. Most important has been the early risk of coronary bypass surgery in this population combined with the significant improvement in stent platform and technology.

Until recently, the main limitation of early trials of PCI versus CABG for unprotected left main artery disease has been the lack of long-term follow up. Recently, the Premier of Randomized Comparison of Bypass Surgery Versus Angiogram Trial (PRECOMBAT) study using DESs in patients with left main CAD was published in the Journal of the American College of Cardiology.⁷ This trial followed over 600 patients who had at least a 50% lesion of the left main coronary artery who were randomized to PCI with DESs versus CABG with a left internal mammary artery graft. The 5-year results show that there was a nonsignificant trend toward a benefit in the bypass surgery group in the primary end point of all-cause mortality, myocardial infarction, stroke, and ischemia-driven target vessel revascularization. The PCI group had an event rate of 17.5% versus 14.3% in the CABG group, leading to a hazard ratio of 1.27 (95% confidence interval [CI], 0.84-1.90; P = .26). The primary driver of the difference between the two groups was ischemia-driven revascularization. Target vessel revascularization was almost double in the PCI group (11.4% vs 5.5%; hazard ratio, 2.11; P = .012). In a landmark analysis after the first year of entry into PRECOMBAT, a trend for more repeat revascularization procedures in the PCI group remained. Importantly, the PRECOMBAT trial enrolled not only patients with isolated unprotected left main CAD, but also patients with multivessel CAD. In the patients with left main CAD and triple-vessel disease, the results significantly favored bypass surgery. Overall, it appears in the PRECOMBAT era that patients with unprotected main CAD have similar major adverse cardiac event rates with PCI-DES and CABG. This is with the caveat that a repeat revascularization procedure may be needed at a rate of about 2% per year and that patients with more extensive disease, particularly those with diabetes, should undergo CABG as their front-line therapy. The preponderance of data calls for an individualized approach to the problem. We await the results of the landmark randomized Evaluation of Xience Prime Versus Coronary Artery Bypass Surgery for Effectiveness of Left Main Revascularization (EXCEL) trial to definitively inform the use of PCI in patients with left main disease (Fig. 44–1).⁸

The presence of diabetes at the time of coronary revascularization is one of the strongest predictors of outcome and is a major determinant of the optimal mode of revascularization.⁹ Over the last 20 years, a great deal of emphasis has been placed on the diabetic subgroup, and this has been further delineated in recent years by randomized trials, primarily conducted in the diabetic subpopulation.

The presence of diabetes is a marker of a greater extent of CAD; greater likelihood of important comorbidities such as renal dysfunction, thrombosis, and left ventricular dysfunction; and greater likelihood of heart failure. In addition, rates of in-stent restenosis and target vessel revascularization are higher in diabetic than nondiabetic patients.¹⁰ In the era of DESs, there have been four major clinical trials. These are summarized in Table 44–3. They include the CARDia trial, VA CARDS trial, the diabetic subgroup of the SYNTAX trial, and the FREEDOM trial.^10,11,12,13 Across the board, diabetic patients with multivessel CAD involving at least two major coronary artery beds have reduced major adverse cardiovascular and cerebrovascular events (MACCE) after CABG compared to multivessel PCI.

When evaluating the components of the MACCE outcome, it is important to emphasize that although each of the individual components may be considered separately, within a single trial, these are typically post hoc analyses that are underpowered and must be considered hypothesis generating only. For the purposes of powering a randomized controlled trial, a composite primary end point has typically been required due to feasibility of enrolling large sample sizes. By combining data between multiple trials, meta-analysis may provide a better assessment of the individual components of MACCE.

All-Cause Mortality

A number of meta-analyses have been conducted with these four clinical trials. The most comprehensive is the meta-analysis by Verma and colleagues,¹⁴ which demonstrated that CABG was associated with a 33% reduction in mortality at 5 years compared to PCI (relative risk, 0.67; 95% CI, 0.52-0.86; P = .002; Fig. 44–2).¹⁴ This was also matched by reductions in myocardial infarction and repeat revascularization.

The FREEDOM trial was designed to address the optimal coronary revascularization strategy in patients with diabetes and multivessel CAD in the absence of left main disease, prior CABG, prior stenting within 6 months, and current ST-segment elevation myocardial infarction. In FREEDOM, the overall 5-year Kaplan-Meyer estimates of the primary composite end point of death, nonfatal myocardial infarction, and nonfatal stroke indicated a 26.6% event rate in the first-generation DES group compared to 18.7% in the CABG group (P < .005; Fig. 44–3). Indeed, in FREEDOM, CABG was associated with a significantly reduced all-cause mortality compared to PCI at 5 years, and the mortality curves continued to diverge in favor of CABG throughout follow-up. Eight-year follow-up of the FREEDOM trial is pending but is likely to demonstrate a striking survival benefit for CABG compared to multivessel PCI in diabetic patients. The results of FREEDOM and the subsequent meta-analysis lead to a Class I recommendation in the 2014 ACC/AHA guidelines stating that CABG was associated with improved survival in diabetics with multivessel CAD.⁴ This was upgraded from a Class II recommendation in 2012. CABG remains the revascularization treatment of choice for diabetic patients with multivessel disease. This is particularly true if a left internal mammary artery (LIMA) graft to the left anterior descending artery was performed.

Stroke

For patients with diabetes undergoing CABG in the FREEDOM trial, there was an excess of early stroke compared to PCI. The difference accrued mostly over the first 30 days after follow-up. In a recent analysis from the FREEDOM CABG cohort, Domanski et al¹⁵ described the predictors of early stroke. These include a history of stroke, the use of warfarin, and whether CABG was performed outside North America. In long-term follow-up, the strongest predictor of stroke was the presence of chronic renal insufficiency. Over 5 years of follow-up, CABG was still associated with a significantly higher stroke rate compared to PCI (5.2% vs 2.4%; P = .03).

Perhaps the most complicated subpopulation is patients with chronic renal insufficiency. These patients have increased risk for death and MACCE. In addition, patients who are not yet dialysis dependent are also at risk for exacerbating renal insufficiency. Both CABG and PCI are associated with worsening renal insufficiency. Direct randomized comparisons of CABG versus PCI are limited because of small numbers of renal failure patients enrolled in the pivotal clinical trials. Consideration of the optimal revascularization strategies in patients with chronic kidney disease is made more complicated because there are essentially two subgroups of patients: those who have end-stage renal disease or who are on hemodialysis or peritoneal dialysis and those with chronic kidney disease not yet on dialysis. This distinction is quite important as we evaluate the evidence. In addition, the rates of utilization of cardiac catheterization and medical therapy are traditionally reduced in patients with chronic kidney disease, and this has been repeatedly recognized in multiple international registries.^16,17

For patients with end-stage renal disease, the ACC/AHA guidelines recommend that CABG should be carried out in patients with multivessel CAD.⁴ For patients with moderate-to-severe chronic kidney disease with multivessel CAD, the 2014 European Society of Cardiology guidelines also recommend the use of bypass surgery over PCI.¹⁸

The use of newer generation stents, however, has brought these recommendations into question. Recently, Bangalore and colleagues¹⁹ from the New York State Registry retrospectively studied over 5000 patients who were propensity-matched and reported that long-term mortality for PCI and CABG was equal (hazard ratio, 1.07; 95% CI, 0.92-1.24), as shown in Figure 44–4. PCI was associated with a higher risk of myocardial infarction and risk of repeat revascularization, but a lower risk of stroke (hazard ratio, 0.56; 95% CI, 0.41-0.76). In the same analysis, 243 matched pairs of patients with end-stage renal disease on hemodialysis were studied; PCI was associated with a higher risk of death (hazard ratio, 2.0; 95% CI, 1.40-2.93). Although there continues to be an ongoing debate as to the use of optimal PCI versus contemporary CABG in patients with chronic kidney disease, CABG remains the treatment of choice for renal failure patients with multivessel CAD. This will be the topic of ongoing discussions in the decade to come, and multiple groups are proposing a definitive cardiovascular randomized trial in patients with moderate-to-severe chronic kidney disease with renal outcomes as an important co-primary end point.

For the important subgroup of patients with left ventricular systolic dysfunction, there are few data from randomized trials suggesting the optimal revascularization strategy. This group has largely been dependent on the heart team approach due to multiple variables ranging from myocardial variability to clinical judgement to patient preference. A meta-analysis in The Lancet in 2009 suggests that CABG is the preferred mode of revascularization in patients with left ventricular dysfunction²⁰; however, this analysis was largely limited to patients in the bare metal stent era. A recent analysis from the APPROACH investigators in Alberta, Canada, identified 2925 patients with CAD with left ventricular dysfunction undergoing CABG (n = 1326) or PCI (n = 1599) between 1995 and 2008.²¹ In a Cox proportional hazards analysis of the propensity-matched subgroups, CABG was associated with lower rates of death and repeat revascularization when compared to PCI with up to 15 years of follow-up. These rather robust data appear to confirm recommendations for most of the major societies. Overall, because of the small numbers in randomized trials, a definitive trial of CABG versus PCI in patients with left ventricular systolic dysfunction may be difficult to perform. In a recent analysis from the Surgical Treatment for Ischemic Heart Failure (STICH) study, patients with ischemic cardiomyopathy and an ejection fraction of 35% or less were randomly assigned to CABG versus OMT alone and followed for 9.8 years. There were more cardiovascular deaths in the OMT group (297 deaths, 49.3%) compared to the CABG group (247 deaths, 40.5%; hazard ratio, 0.79; 95% CI, 0.66-0.93; P = .006).²²

The increasing use of PCI in patients with prior bypass surgery is largely due to patient preference to avoid redo CABG and the attendant risk of perioperative mortality and morbidity.²³ Indeed, the number of redo CABG procedures continues to fall in the United States.²⁴ There are no prospective randomized data to guide decision making; however, adjusted comparisons in the subpopulation of prior bypass surgery patients appear to be slightly better for PCI when compared to redo CABG and medical therapy alone in the APPROACH registry.²⁵ This will require further evaluation in the years to come.

1. The importance of co-intervention: The importance of OMT has been stressed on all fronts. One of the important weaknesses of many clinical trials is that patients are not being treated equally in the PCI and CABG arm with regard to OMT. For example, in the SYNTAX trial, only 75% of patients in the bypass arm were discharged with statin therapy compared to over 85% in the PCI arm.⁶ It is possible that more balanced administration of OMT might have led to even greater improvement in outcomes in the CABG group compared to the PCI group. The effective co-intervention is an important variable when evaluating results of clinical trials because many patients do not reach guideline-directed targets for major medical risk factors.²⁶

2. The importance of newer generation stenting: With the advent of newer generation stents and reductions in stent thrombosis, there has been consideration for the need of yet another PCI versus CABG trial with a newer platform of stents. The authors caution against this reasoning. Comparing the hazard ratio for mortality from the original BARI trial published in 1995 versus those after the advent of bare metal stents in the publications of the ARTS and SOS trials and the Hlatky meta-analysis versus those in the publications of the FREEDOM and SYNTAX trials and subsequent Verma meta-analysis, the hazard ratio for mortality has stayed relatively constant, in the range of approximately 0.70 for CABG versus PCI. With long-term follow-up, it is unlikely that PCI in any form would make up this difference without a significant impact on non–culprit lesion progression.

   The introduction of newer generation DESs has improved the safety profile for PCI, particularly by reducing rates of stent thrombosis. The recent TUXEDO trial from India demonstrates the superiority of the everolimus-eluting stent strategy over the paclitaxel stented group for reducing target vessel failure (P = .005), spontaneous myocardial infarction (1.2% vs 3.2%; P = .004), stent thrombosis (0.4% vs 2.1%; P = .002), and target vessel revascularization (1.2% vs 3.4%; P = .002) at 1 year in a diabetic population.²⁷

   The question of whether newer generation stenting will improve MACCE outcomes and reduce the difference in MACCE outcomes between PCI and CABG in advanced CAD remains unknown. It is highly unlikely that trials such as SYNTAX and FREEDOM will be performed with the newer stent platforms since even significant reductions in stent thrombosis are not going to make up the majority of the differences in cardiovascular outcomes. This again remains a topic for future investigation.

3. The importance of multiple arterial grafting in bypass surgery versus a single LIMA to left anterior descending artery (LAD): Models for the determination of the single pivotal variable that favors a mortality benefit of CABG over PCI indicate that implantation of the LIMA to the LAD graft appears to confer the most benefit.²⁸

   Data from the Mayo Clinic involving 8622 patients with isolated CABG from multivessel CAD from 1993 to 2009 evaluated the use of LIMA plus saphenous vein grafts versus multiple arterial grafting.²⁹ The operative mortality was not significantly different between the two groups. When evaluating late survival at 15 years, multiple arterial grafting revealed a significant survival advantage. Multi-arterial grafting remained a strong predictor of survival (hazard ratio, 0.79; 95% CI, 0.66-0.94). We await the 10-year results of the randomized trial of bilateral versus single internal mammary coronary artery grafting (the ART trial) in the years to come. The 1-year findings of ART suggest that there are similar clinical outcomes for single internal mammary versus bilateral internal mammary grafting, with a 1.3% increase in the need for sternal wound reconstruction with bilateral internal mammary surgery.³⁰ In a meta-analysis by Weiss et al³¹ of bilateral internal mammary artery (BIMA) versus LIMA grafting alone, there appears to be a long-term survival benefit associated with BIMA versus LIMA alone. This was based entirely on observational data. The long-term survival advantage was associated with a hazard ratio of 0.78 (95% CI, 0.72-0.84; P < .00001). This survival benefit was noted among both diabetic and nondiabetic patients in another large retrospective series, in which BIMA conferred a 35% reduction (95% CI, 12%-52%; P = .006) in the long-term hazard of death equally for nondiabetic and diabetic patients.³² Cardiologists and surgeons alike should advocate for the safe adoption of more multiarterial grafting in patients referred for surgical coronary revascularization.

4. Completeness of revascularization: The importance of completeness of revascularization cannot be overstated as a critical determinant of which revascularization strategy is selected in an individual patient. Patients with a very low likelihood of complete revascularization by PCI may be appropriately steered to CABG. The importance of the completeness of revascularization was emphasized in a report from the New York State Registry.³³ This showed that incomplete revascularization with stenting was associated with adverse impact on long-term outcome (hazard ratio, 1.15; 95% CI, 1.01-1.30) when compared with complete revascularization. Those at highest risk were patients with incomplete revascularization with residual total occlusion and/or a total of two or more vessels with incomplete revascularization.

   A subsequent analysis by Wu and colleagues³⁴ evaluated the association of complete revascularization and long-term mortality in 6511 pairs of propensity-matched patients and showed that 5-year survival was lower for patients after incomplete revascularization by PCI compared to those with complete revascularization (79.3% vs 81.4%; P < .0004). This underscores the importance of this variable in considering the optimal approach to our patients.

5. The importance of on-pump coronary artery bypass surgery compared to off-pump coronary artery bypass (OPCAB) surgery: The recent United Kingdom national adult cardiac audit registry demonstrates no difference in the 5-year survival in patients undergoing on-pump versus off-pump surgery.³⁵ This confirms the 1-year findings of the large international CORONARY trial in which there was no difference in the rate of death, myocardial infarction, stroke, or new renal failure, as well as no differences in quality of life and neurocognitive function between the off-pump and on-pump groups.³⁶ Overall, there was an influence of operator on outcomes, with a recommendation that OPCAB be performed by surgeons who were more experienced and who perform at least 10% of their surgeries as OPCAB.³⁷ Moreover, OPCAB has shown greater benefit in patients at higher preoperative risk, including women.³⁸ These findings have been confirmed in large retrospective analyses of the STS National Cardiac Database.³⁹ Regional differences for outcomes in surgery need to be further evaluated not only for off-pump surgery but also for stroke risk.

6. The risk of stroke associated with CABG: With the emergence of CABG as the preferred mode of revascularization for patients with diabetes, left ventricular dysfunction, renal failure, and greater comorbidities, it is important to evaluate ways to reduce stroke, particularly in the first 30 days after bypass. Strokes may be caused by multiple etiologies, including coincident cerebral vascular disease and, importantly, manipulation of the aorta at the time of surgery. In a large multicenter registry, off-pump bypass surgery did not result in a lower risk of stroke when stratified by the Northern New England Cardiovascular Disease Study Group (NNECDSG) strata for perioperative stroke, nor did it result in any significant improvement in long-term cognitive function when compared to on-pump grafting.⁴⁰ However, in a larger sample of the STS Database, OPCAB was strongly associated with a reduced risk of 30-day stroke (adjusted odds ratio, 0.66; 95% CI, 0.62-0.71; P < .0001; n = 876,081 patients).⁴¹ With regard to concomitant cerebrovascular disease, staging of procedures so that the most unstable territory is intervened upon first is the approach that has been most widely adopted. For example, patients having unstable carotid disease and SIHD may proceed to carotid intervention first. Other approaches, including concomitant procedures, have also been undertaken. A study from the Cleveland Clinic shows that in long-term follow-up with adjustment after the first year, staged carotid artery stenting–open heart surgery (OHS) was associated with a significantly lower composite end point of death, myocardial infarction, and stroke compared with both staged carotid endarterectomy (CEA)-OHS (hazard ratio, 0.33; 95% CI, 0.15-0.77; P = .01) and combined CEA-OHS (hazard ratio, 0.35; 95% CI, 0.18-0.70; P = .003).⁴²

   In the effort to reduce operative and perioperative risk of stroke with CABG, there are multiple important issues that need to be considered, including the following: (1) the preoperative screening for cerebrovascular disease; (2) the operative screening of the ascending aorta to individualize surgical techniques to avoid distal atherothrombotic emboli; (3) the influence of post-operative atrial fibrillation and appropriate anticoagulation; and (4) the importance of dual antiplatelet therapy in selected patients (particularly patients after acute coronary syndrome) and the disruption of dual antiplatelet therapy (DAPT) in the postoperative period. These issues and others will remain the subject of ongoing research undertaken to understand ways to reduce the risk of stroke after bypass.

7. Sternal wound infections: The serious complication of deep sternal wound infections can occur in 0.5% to 2% of patients and is commonly associated with diabetes and obesity.⁴³ In diabetic patients, the use of BIMA is linked to both significantly improved long-term survival and increased rates of perioperative deep sternal wound infection.⁴⁴ In particular, obese, female patients with poorly controlled diabetes are at elevated risk of deep sternal wound infections, and BIMA grafting should be avoided in this group.⁴⁵ Continuous intravenous insulin infusion should be implemented during surgery and early postoperatively in cardiac surgery patients whenever glucose levels are elevated in order to minimize risk of deep sternal wound infections.⁴⁶

Optimal Antiplatelet Therapy: Intensity, Duration, and Bleeding Balance

Antiplatelet therapy is the cornerstone of drug therapy to prevent recurrent ischemic events after coronary revascularization. DAPT, consisting of aspirin plus an inhibitor of the platelet P2Y₁₂ receptor, is considered standard of care after PCI. Once uniform after DES implantation, decision making about DAPT has grown increasingly complex as stent technologies have improved,^47,48 newer more-potent P2Y₁₂ inhibitors have come to market,^49,50 and a growing body of evidence has tested durations of therapy longer or shorter than the classic 12-month period.^51,52,53 Central to this complexity is the judgment of an individual patient’s dynamic balance of ischemic and bleeding risks.

Substantial overlap exists between risk factors for ischemic and bleeding events after PCI. Of note, acuity is an important determinant. Individuals undergoing PCI for unstable angina and acute myocardial infarction are, in general, at higher risk for both events. For the purposes of this review, we will restrict our discussion to patients with SIHD. The most important predictor of stent thrombosis, a sentinel ischemic event after PCI, is premature disruption of DAPT. Additional predictors of stent thrombosis include renal failure, diabetes mellitus, bifurcation lesions, calcified lesions, reduced left ventricular ejection fraction, small stent diameter and longer stent length, malignancy, prior stroke, dissection, and presence of intermediate coronary disease proximal and distal to a culprit lesion.^54,55,56,57 Additional predictors of bleeding include age, female sex, heart failure, renal failure, and peripheral vascular disease.⁵⁸ Predictors of post-discharge bleeding, in particular, include older age, lower baseline hemoglobin, lower platelet reactivity on clopidogrel, and use of chronic oral anticoagulant therapy.⁵⁹

Seminal trials establishing the anti-ischemic efficacy and supporting US Food and Drug Administration approval of the newer oral platelet P2Y₁₂ inhibitors prasugrel and ticagrelor were conducted in patients with acute coronary syndromes. The utility of these more-potent agents in patients with SIHD is less certain. Promising data from the PEGASUS-TIMI 54 trial indicate that DAPT with aspirin and ticagrelor, initiated greater than 1 year after myocardial infarction, is associated with a 15% reduction in the primary composite end point of death, myocardial infarction, and stroke (P < .01).⁶⁰

The role of DAPT after CABG is less well established. It is uncertain whether DAPT improves graft patency. In a single-center study of 249 consecutive patients randomized to aspirin 100 mg or aspirin 100 mg plus clopidogrel 75 mg daily after elective CABG, DAPT was associated with improved venous graft patency at 3 months.^37,61 In a second randomized study of aspirin 162 mg versus aspirin 162 mg plus clopidogrel 75 mg daily after CABG, there was no difference in the primary end point of intravascular ultrasound–determined saphenous vein graft intimal hyperplasia and no difference in freedom from major adverse cardiovascular events.^38,62

A separate potential benefit of DAPT may be added reduction in incidence of stroke after coronary revascularization.^39,63 Larger trials are required to assess efficacy and safety of DAPT after CABG with particular attention to stroke outcomes.

Bioresorbable Vascular Scaffolds: Successors to Stents?

Ongoing investigation and intense interest are presently focused on bioresorbable vascular scaffolds (BVS) as a prospective successor to stents. Although risks of ischemic complications, and in particular stent thrombosis, have progressively declined with evolution of DES technology, the best available DESs continue to pose several putative disadvantages. Permanent metallic stents, with or without drug elution, can impede future coronary instrumentation and contribute to chronic inflammation,^40,64 endothelial dysfunction,^41,65 and adverse remodeling.^42,66 In theory, a BVS might ameliorate these problems while providing a temporary period of radial support after PCI.

There is now experience with use of one first-generation BVS, the everolimus-eluting Absorb BVS (Abbott Vascular, Redwood City, CA), in over 100,000 patients worldwide. In ABSORB III, a large, multicenter randomized comparison of the Absorb BVS versus the everolimus-eluting Xience (Abbott Vascular) stent,^43,67 rates of target lesion failure (cardiac death, target vessel myocardial infarction, and ischemia-driven target lesion revascularization) were higher at 1 year in patients randomized to receive a BVS (7.8% vs 6.1%). Subjects randomized to receive a BVS also experienced a nonsignificantly higher rate of device thrombosis (1.5% vs 0.7%). The studied BVS platform satisfied prespecified criteria for noninferiority. Observed excesses in ischemic events may stem from a learning curve in optimal deployment of BVS, which appears to rely on meticulous lesion preparation. Validation of the proposed long-term benefits of a BVS for angina and late ischemic events awaits results of the ongoing ABSORB IV trial.⁶⁸

Hybrid Revascularization

The use of a hybrid approach to coronary revascularization is gaining popularity and may be an appropriate alternative for a carefully selected subset of patients with limited CAD involving the proximal or mid LAD and at least one other non-LAD coronary artery. In hybrid coronary revascularization (HCR), a minimally invasive surgical approach is used to bypass the LAD with a LIMA graft in combination with stenting of non-LAD targets. This is believed to leverage the advantages of surgical arterial revascularization of the LAD and substitution of advanced stent technology for saphenous vein grafts for the other targets. In a recent analysis from the STS Adult Cardiac Surgery Database, HCR is offered in about one-third of hospitals but represents less than 1% of CABG procedures in the United States.⁶⁹ The current body of evidence is limited to observational studies and a small pilot randomized trial (Fig. 44–5). Overall, pooling of the studies shows a trend for fewer MACCE (odds ratio, 0.69; 95% CI, 0.41-0.92; P = .18) in favor of HCR over conventional CABG. Although stroke was reduced by 80% (P = .039) in the HCR group, there was an excess of repeat revascularization (odds ratio, 4.05; P < .001).

The current ACC/AHA guidelines on coronary revascularization consider HCR only when PCI of the LAD is considered suboptimal, and therefore, it is given a Class IIa recommendation.⁴ A large prospective randomized trial may be necessary to define the benefit of this new alternative technique.

High-Risk Percutaneous Coronary Intervention

Advances in technique and technology have dramatically improved the safety of elective PCI. For a subset of patients, however, elective PCI continues to pose substantial risk of morbidity and mortality. Candidates for so-called “high-risk PCI” are typified by advanced age, multiple comorbidities, severe left ventricular dysfunction, unsuitability for surgery, and little room for error in the catheterization laboratory. Lesions are frequently complex and may be characterized by multivessel disease, unprotected left main disease, or involvement of a last remaining coronary vessel.

Assessment of a proposed PCI as high risk is a clinical judgment made by an experienced heart team. Risk models developed from analysis of large registries can help in predicting likelihood of in-hospital mortality after elective PCI. One such model is the Mayo Clinic Risk Score (MCRS), derived from a cohort of 7640 patients undergoing PCI between 2000 and 2005,^3,70 and subsequently validated in a data set of over 300,000 patients as a predictor of in-hospital mortality after PCI.⁷¹ The MCRS (Fig. 44–6) incorporates seven variables: age, serum creatinine, left ventricular ejection fraction, preprocedural shock, myocardial infarction within 24 hours, heart failure on presentation, and peripheral vascular disease. Patients at high risk for elective PCI are often also at high risk for surgery, and indeed, the MCRS correlates strongly with the STS score for risk of death after CABG.⁷²

Selected cases of high-risk PCI may benefit from adjunctive use of mechanical circulatory support. An expanding armamentarium of percutaneous devices for this purpose today includes intra-aortic balloon pump (IABP), centrifugal flow (TandemHeart; CardiacAssist, Pittsburgh, PA) and axial flow (Impella; Abiomed, Danvers, MA) ventricular assist devices, and extracorporeal membrane oxygenation (CardioHelp; Maquet, Wayne, NJ). Recent studies of device support have helped to define current concepts of high-risk elective PCI.

The Balloon Pump-Assisted Coronary Intervention Study (BCIS-1) randomized 301 patients with multivessel CAD and left ventricular ejection fraction ≤ 30% to PCI with or without IABP support. MACCE were equally common in the hospital (15.2% vs 16.0%; P = .85).⁷³ Mortality in this cohort was high at 5-year follow-up (33%) and, surprisingly, significantly lower in IABP-treated patients (hazard ratio, 0.66; 95% CI, 0.44-0.98; P = .039).⁷⁴

The PROTECT II trial, which was underpowered due to early termination, randomized 452 symptomatic patients with complex three-vessel CAD or unprotected left main CAD and left ventricular ejection fraction ≤ 30% to PCI with IABP or Impella support.⁷⁵ Both IABP and Impella-supported patients experienced significant improvements in functional status and left ventricular ejection fraction after PCI. At 30 days, there was no difference in a broad primary composite end point of major adverse events (40.1% vs 35.1%; P = .23). After discharge, however, there was a lower rate of death, stroke, myocardial infarction, and repeat revascularization in patients randomized to Impella.

In considering these data, a few salient points emerge. First, patients with ischemic cardiomyopathy and severe left ventricular dysfunction are indeed at high risk, with substantial short-term jeopardy of morbidity and mortality. Second, these patients benefit from revascularization, with improvement in symptoms and left ventricular function. Third, and perhaps most interestingly, the benefits of adjunctive mechanical circulatory support may not appear in the hospital but rather accrue with passage of time. The explanation for this final observation remains uncertain but may relate to differences in completeness or quality of PCI under conditions of mechanical circulatory support.

The Impact of Fractional Flow Reserve

Major trials^10,11,12,13 informing current management of multivessel CAD have relied predominantly on angiographic criteria for study inclusion and lesion selection. Yet recent studies of physiologic lesion assessment using fractional flow reserve (FFR) have underscored the limitations of anatomy-guided PCI.

FFR is a lesion-specific index of stenosis severity defined as the ratio of maximum flow in the presence of stenosis to normal maximum flow.⁷⁶ In practice, FFR is estimated as the ratio of coronary artery pressure distal to a lesion of interest to aortic pressure, averaged over the entire cardiac cycle, with variation in microvascular resistance minimized by pharmacologic induction of maximal hyperemia. Dedicated fiberoptic pressure wire systems have helped make FFR measurement simple, safe, and widely available in clinical practice.

FFR is useful to reclassify the functional severity of angiographically indeterminate coronary lesions⁷⁷ and to define the utility of PCI. When a lesion is functionally insignificant, as defined in the DEFER trial by an FFR ≥ 0.75, PCI may be safely deferred without excess risk of cardiac death or myocardial infarction.⁷⁸ Use of FFR guidance improves both efficacy and appropriateness of PCI. In the FAME study, which randomized 1005 patients with multivessel CAD to PCI guided by angiography alone or angiography plus FFR, a strategy of FFR guidance was associated with a reduced rate of death, nonfatal myocardial infarction, and repeat revascularization at 1 year (13.2% vs 18.3%; P = .02) as well as a reduced usage of stents per patient (1.9 ± 1.3 vs 2.7 ± 1.2; P < .001).⁷⁹ Indeed, when a lesion is functionally significant, as defined by an FFR ≤ 0.80, revascularization is beneficial. In the FAME 2 trial, which randomized 888 patients with at least one functionally significant stenosis to PCI plus medical therapy or medical therapy alone, PCI conferred a significant reduction in death, nonfatal myocardial infarction, and urgent revascularization at 2 years.⁸⁰

Whether the advantages of FFR guidance for PCI may be extrapolated to CABG remains uncertain. Physiologic differences in the mechanism of CABG invoke separate considerations, including risk of conduit failure and potential for lesion progression. Furthermore, the standard of care for CABG has been founded on a principle of complete anatomic revascularization, rendering the notion of omitting bypass of FFR-negative intermediate stenosis as exploratory. In a retrospective study of 627 patients undergoing CABG, of whom 198 underwent grafting or deferral of grafting of at least one intermediate stenosis based on an FFR cut point of 0.80, the use of FFR guidance was associated with a lower number of graft anastomoses, a lower rate of angina, and no excess of death, myocardial infarction, or target vessel revascularization at 36 months.⁸¹ Results are awaited from randomized clinical trials currently under way to evaluate the role of FFR guidance for CABG planning.

Based on presently available data, the authors advocate a central role for FFR guidance in determining appropriateness for PCI (Fig. 44–7).

Presently, PCI is the most common treatment for limited CAD with SSs less than 22. More complex disease (SS > 32) is more effectively managed with CABG. Intermediate disease may be technically amenable to either strategy, but results in patients with intermediate disease are better with CABG in higher risk groups. Diabetic patients with multivessel disease and patients with low ejection fraction or renal failure have better intermediate and late outcomes with CABG, and CABG remains the treatment of choice in those subgroups. New techniques for minimizing aortic manipulation during CABG and increasing the use of multiple arterial conduits are available and should be routinely applied to improve clinical outcomes after CABG. Similarly, technical improvements in PCI devices and an evolving understanding of the role of FFR in selecting vessels and patients will contribute to improve outcomes with PCI in the future. Reducing stroke after bypass surgery and optimizing medical co-intervention with the newer low-density lipoprotein–lowering drugs in all patients will potentially change the coronary revascularization landscape as well. Together, the heart team approach can ensure that each patient with coronary disease enjoys the best possible outcomes.