Will Stent Revascularization Replace Coronary Artery Bypass Grafting?

Thank you for the invitation to Thoracic Park. As a cardiologist speaking to cardiovascular surgeons, I will not argue that stenting will replace surgery in the treatment of coronary artery disease (CAD). However, we have made substantial advances in engineering, technique, and pharmacology that make stenting a very good option for revascularization in many patients. These patients are not only those with a single coronary lesion, but also those with multivessel CAD.

When we discuss revascularization outcomes, we are talking about 3 major endpoints: death, myocardial infarction, and symptom control. With respect to death, we know that revascularization benefits patients who have severe multivessel disease and left ventricular dysfunction or other physiologic indicators of high risk. That proof comes from 3 seminal trials performed in the 1970s and 1980s^1–3 and from many observational studies. Of note, 2-vessel disease with proximal left anterior descending coronary artery (LAD) stenosis has been accepted as an indication for revascularization, even though the supporting data come from a small subgroup in a single trial.⁴ In that subgroup, the difference in outcomes had a P value of only 0.06. I concede that the only evidence to support this therapy for patients with stable CAD comes from this single surgical trial. I also concede that we, the cardiologists, have been unsuccessful in proving that endovascular treatment has a positive impact on stable CAD. This proof is important, because we leave the native arteries relatively intact. The alternative, surgery, remains largely dependent upon the venous autograft, the patency of which deteriorates rapidly after only a few years (Fig. 1).⁵ Attempts to improve graft performance beyond the relatively spectacular performance of the pedicled internal mammary artery (IMA) graft to the LAD have been disappointing.

Fig. 1 Graph of graft patency shows deterioration rates over 10 years and the comparative superiority of using the internal mammary artery (IMA) instead of the saphenous vein (SVG).

Percutaneous Transluminal Coronary Angioplasty

When angioplasty was introduced, the hope was for a method of revascularization that would rival coronary artery bypass grafting. However, the results were mixed. Angioplasty worked well in patients with no major risk factors, such as diabetes mellitus, but failed miserably in diabetic patients.^6,7 In fact, the Bypass Angioplasty Revascularization Investigation (BARI) trial⁶ taught us this: if revascularization is needed, regardless of physiologic markers of high risk, the use of percutaneous coronary intervention (PCI) is potentially harmful in comparison with an IMA bypass for the LAD.

Stents and Short-Term Outcomes

Along came the bare-metal stent—a metallic buttress to solve all the problems of angioplasty—and, thereafter, the drug-eluting stent (DES), which “cured” restenosis (Fig. 2). Indeed, the stent was a godsend for those of us who do not use cardioplegia. The use of stents drastically reduced the probability of emergent surgery after attempted PCI^8,9; however, the probability of new lesion formation or restenosis after intervention did not decrease.¹⁰

Fig. 2 Diagrams show the calculated success (after percutaneous revascularization) of A) percutaneous transluminal coronary angioplasty (PTCA), and B) bare-metal and C) drug-eluting stenting in patients with 3-vessel coronary artery disease (CAD). Success refers to the initial procedure, restenosis risk, and reintervention rates that have been published. Closed boxes contain actual values from randomized trials; other figures are estimates from published data.

CAB = coronary artery bypass; PCI = percutaneous coronary intervention

At the same time, surgeons got better. Myocardial preservation techniques improved, and the use of the pedicled IMA graft changed the game. As a result, successful revascularization, meaning long-term success, became the domain of the surgeon. We at the Texas Heart Institute/St. Luke's Episcopal Hospital (THI/SLEH) examined our long-term outcomes after stenting or surgery, and we initially reported that stenting was just as beneficial as surgery.¹¹ This was in accord with the results of several trials (albeit of limited follow-up duration)^12,13: whenever placing a stent was feasible, stent therapy and surgery had the same outcome.

Stents and Long-Term Outcomes

Later, when we looked at longer-term follow-up data and the effects of multiple procedures, this picture began to change. Stented patients underwent more procedures. When the risk of one surgical procedure was compared with that of multiple endovascular procedures, the outcomes became more similar,^14,15 especially in patients with bifurcation lesions or lesions with severe calcification. Drug-eluting stents, with their promise of no restenosis, substantially increased interventional cardiologists' reach, but not their grasp. In patients with multivessel disease and high-risk lesions, DES placement was almost as risky as surgery and did not yield the same long-term benefit.^16,17

Nevertheless, we found locally that the introduction of the DES, with its lower risk of restenosis, was treated as a blessing to proceed with stenting (Table I). This did not follow the data, but cardiologists continued anyway, given the promise of less restenosis. Early risk was discounted, glycoprotein IIb/IIIa inhibitor use declined overnight, and the rate of endovascular procedural complications rose to meet that of surgery without the promise of an IMA graft in our future.

Table I. Independent Predictors of 30-Day Major Adverse Cardiac Events and 3-Year Survival after Drug-Eluting Stent Placement

Comparing Stenting and Surgery

For decades, methods have been sought to quantify lesion complexity in order to compare the early and late risks associated with stenting versus surgery. Although no perfect system has been devised, the SYNTAX score¹⁸ was an important step forward. The SYNTAX score is a simple, computer-based tool for evaluating the risk of complications or failure after PCI. And there are other tools for estimating the same complications after surgery.¹⁹ These estimates enable cardiologists to give patients objective advice regarding the revascularization method that has the best short- and long-term probability of success.

In the patient with non-life-threatening disease (that is, not left main or severe multivessel CAD with left ventricular dysfunction or severely impaired function), stent revascularization has become a reasonable, although not ideal, alternative to surgical revascularization. However, this is true only if stenting is confined to patients whose anatomy and physiology are suited to it—considerations that are well quantified in the SYNTAX score. Whenever questions arise as to the most appropriate therapy, the SYNTAX score should be weighed against clinical characteristics that affect surgical risk. This will guide discussions between the cardiologist, cardiovascular surgeon, patient, and treating physician.

I think that our THI risk score²⁰ is more useful than the other available scores. It uses simple clinical data and can be easily calibrated to the geographic location of its use. Other scores require data that might not be available at the time of clinical decision-making or at all—making such predictions hazardous, at best.

Conclusion

With regard to the chosen mode of revascularization, it is perhaps safe to say that the decision goes beyond the individual physician and must become collective. When a patient has multivessel disease, a reasoned approach must be taken, using these predictive tools and considering the patient's wishes. Treatment decisions should include all interested parties: the patient, cardiologist, cardiovascular surgeon, and anesthesiologist. The time of ad hoc angioplasty for the patient with multivessel CAD has passed.