The Left Main Facts: Faced, Spun, But Alas Too Few

Counseling Patients on Revascularization Options

It is true that interventionalists do not routinely suggest a surgical consultation to patients they feel are candidates for PCI. The majority of such patients are of low to moderate risk, the procedural success rates are high, acute complication rates are low, and this approach is widely considered justifiable given patient preference, the ad-hoc nature of most procedures, and the logistical demands of all involved. For high-risk nonemergent patients, however, it would be in the best interests of all to have an inclusive team approach to ensure that the patient and family are well informed as to therapeutic options. Although interventionalists might consider themselves capable of presenting an even-handed discussion of both PCI and CABG, the perception among many of our surgical colleagues is that a conflict of interest is inherent in such an approach. Interestingly, in a survey of attitudes of interventionalists toward LM PCI, 48% consulted a surgeon in all potential LM PCI, 48% did so in selected patients, and only 4% thought that a surgeon should not be involved in the process. Only 19% thought that LM PCI should be offered to patients who were good surgical candidates; however, 38% said that is was appropriate to do the procedure in patients who request it (20).

CABG Versus PCI for Multivessel Disease

The contention that CABG is superior to PCI for multivessel disease is arguable. The paper by Hoffman et al. (21) is a meta-analysis consisting of 8 randomized trials spanning the transition from balloon angioplasty to stenting, with only 35% of the total receiving stents. There was no survival difference at 1 and 3 years (in which stented patients were represented), but there was a difference at 5 and 8 years favoring CABG; these data, however, were derived from the early studies that did not include stents. Hannan et al. (22) analyzed data from the New York State CABG and PCI registries (pre-DES), showing better survival with the former. This is an important observation, although it suffers the limitations of a nonrandomized retrospective study. The long-term results of 2 pre-DES randomized trials are available: ERACI II (Argentine Randomized Trial of Coronary Angioplasty With Stenting Versus Coronary Bypass Surgery in Patients With Multiple Vessel Disease) (23) and ARTS (Arterial Revascularization Therapies Study) (24) show CABG and PCI to be equivalent in infarct-free survival. Surgery, however, has been consistently shown to be associated with a lessened need for revascularization. Over the 5-year follow-up of ARTS and ERACI II, the BMS group required revascularization about 3 and one-half times more frequently; however, these studies were performed with BMS. At the 10-year follow-up report of the BARI (Bypass Angioplasty Revascularization Investigation) trial comparing balloon angioplasty with CABG, revascularization had been performed in 77% of the PCI group and 20% of the CABG patients with no difference in survival between the 2 arms (25). Although restenosis is a significant disadvantage of PCI, its specific association with infarction and death has not been emphasized but rather bundled with the overall PCI risk. The potential to reduce the incidence of restenosis has refocused attention on its associated 10% risk of myocardial infarction and 0.7% risk of death (26). Although the risk of restenosis may not be considered a compelling reason to favor CABG over PCI as the initial approach to multivessel disease, it becomes a more disturbing shortcoming in the LM, where the consequences may be more ominous. A further consideration, which remains uncertain, is whether the risk of CABG for PCI restenosis is increased (27,28).

Influence of DES on PCI

The introduction of DES has reduced the incidence of restenosis and target lesion revascularization compared with that seen with BMS without evidence thus far of the anticipated increase in infarct-free survival in patients with multivessel disease (29). The beneficial effects on the restenotic process, however, extend across a wide range of patient subgroups including diabetic patients and have led to more aggressive PCI (30) as well as a decrease and change in the population of patients going to CABG who are now older and sicker (31). Whereas the adverse event rate for DES is increased in “off-label” compared with “on-label” application, this is not true for all such applications (32,33), nor is it clear how much “off-label” risk is related to clinical substrate as opposed to stent type because there is a paucity of controlled studies comparing BMS or CABG with “off-label” DES use.

There has been concern that the current generation of DES may prevent or delay protective endothelium from covering the stent and, hence, increase the risk of ST. Although the incidence of ST in the early post-implant period is no different from that of BMS, there appears to be a continued risk with DES at a rate of 0.6 per year (34) for an as yet undefined duration. The interventional community, at first skeptical of an increased risk of late ST, has now accepted this and argues that a favorable tradeoff exists between a small risk of ST and a large reduction in restenosis. Whether the incidence of ST is further increased in “off-label” use is uncertain. The role of dual antiplatelet therapy in preventing ST is emphasized, and extension of the labeled duration of this regimen to 12 months advocated (35). Many of us, however, tend to prolong clopidogrel well beyond this period, a strategy that would seem prudent for DES placed at critical sites such as the LM. Because premature cessation of antiplatelet therapy has serious consequences, expectation of bleeding complication or need for noncardiac surgery within a year of implant argues against placing DES, especially in an unprotected LM.

Suitability of the LM for PCI

There are studies suggesting that distal LM disease is not well dealt with by PCI. Valgimigli et al. (36) found PCI to have twice the incidence of death or myocardial infarction in distal LM disease compared with disease limited to the ostium or shaft. This appeared to be independent of whether a single or bifurcation stenting strategy was used. Serial surveillance angiography in a cohort of LM PCI patients with primarily bifurcation stenting revealed a 38% incidence of restenosis requiring revascularization. Most restenosis was focal and located at the orifice of the circumflex artery, thus lending itself to repeat PCI. Although the incidence of death in this group was low (2%), acute ST occurred in 2 patients (37). In contradistinction to these studies, Chieffo et al. (38) have reported quite favorable results for PCI not involving the left main bifurcation. In this group followed for about 2.5 years, there was an all-cause death rate of 3.4%. Revascularization was necessary in 5.4% of patients, but only 1 (0.9%) had LM restenosis.

Further data suggest that distal LM disease may be reasonably approached by PCI. In a single-center registry comparing PCI and CABG outcomes at 1 year, there was no difference in the composite end point of death, stroke, myocardial infarction, and revascularization. Distal LM disease was present in 81% of the PCI group, three-quarters of which received stenting to both branches. Repeat revascularization was performed in 20% of the PCI (one-half of which was clinically driven) and 5% of the CABG patients (39). Other studies using primarily a single stent DES approach to patients with distal disease demonstrated even lower rates of revascularization (40,41).

High-risk patients for CABG might be considered at increased risk for PCI as well. Lee et al. (42) compared outcomes of high-risk patients receiving DES (18% turned down for surgery) with outcomes of CABG patients (Parsonnet scores 18.3 and 13.7, respectively). Distal LM disease was present in 60% of DES patients, and bifurcation stenting was performed in two-thirds of these. The estimated major adverse cardiac and cerebral vascular event-free survival at 6 months was 89% for DES and 83% for CABG. Freedom from revascularization at 1 year was estimated to be 95% for CABG and 87% for PCI (42). Other reports by Migliorini et al. (43) and Palmerini et al. (44) of high-risk patients, the majority of which had bifurcation disease, reveal results similar to CABG except for a more frequent need for repeat revascularization. Thus, PCI with DES has demonstrated acceptable short- and intermediate-term survival results compared with those seen with CABG, even in patients with complex disease.

The Bologna registry (45) provides interesting data that help provide some perspective for comparison of CABG and PCI. Patients with unprotected LM disease were treated with either CABG or PCI in accordance with physician and patient choice. The PCI group was older and had higher risk scores. Retrospective analysis by physician and surgeon revealed that 68% of the total group were appropriate for either PCI or CABG; 19% for CABG only, and 13% for PCI only. There was no difference in death or myocardial infarction at 14 months between PCI and CABG; however, the incidence of repeat revascularization was higher in the PCI group, 25.5% versus 2.6% for CABG. The cohort receiving DES (60% of PCI) had a 22% incidence of revascularization. This was thought to be related to the high incidence of bifurcation stenting as well as routine surveillance angiography resulting in the detection and treatment of lesions, 63% of which were asymptomatic. Over 50% of the patients had associated multivessel disease with complete revascularization achieved equally in both groups. Outcome was related to pre-procedural risk assessment: patients deemed good for surgery had equally good long-term survival with either CABG or PCI (3.3% and 3.5%, respectively), whereas poor CABG candidates had 18.1% and 18.8% mortality, respectively.

Durability of CABG

If ST is the Achilles heel of DES, the saphenous vein offers a similar frailty for the surgeon. As many as 7% of these grafts may occlude within 2 weeks of surgery (46). At 1-year post-bypass 12% to 20% of the grafts are occluded, and this number approaches 50% at 10 years (47). Because the placement of a graft tends to accelerate upstream disease (48) and its progression to complete occlusion (49), a subsequent graft occlusion might render PCI more difficult or impossible. We find that 25% of all PCI at Johns Hopkins Hospital is performed in patients who have had a prior bypass. Although internal thoracic arterial (ITA) grafts have a greater durability, 1.3% are occluded and 8% narrowed by ≥50% within 2 weeks of surgery (50); the patency rates at 1 and 5 years are 93% and 88%, respectively (51). Although there have been reports of good results with bilateral ITA grafting, this is not frequently performed. Data from the Society of Thoracic Surgery suggest that considerably <20% of patients receive this procedure at a first isolated CABG (52). At my institution, about 13% of patients undergoing CABG for left main disease and just under 20% of all initial CABG procedures receive bilateral ITA grafts. The SYNTAX (Synergy Between Percutaneous Coronary Intervention With Taxus and Coronary Surgery) run-in study cites all arterial grafting being performed in only 12% of patients operated on for LM and triple-vessel disease (18). A greater consensus among surgeons on the importance of bilateral ITA grafts, radial arterial grafts, and off-pump bypass might be helpful in the design of the best comparator to LM PCI.

Clinical Trials for LM Disease

One would agree that a randomized controlled trial of a therapy in patients with a life-threatening illness should offer enrollees a reasonable chance of an equal or better result than does the current standard of care. It is disappointing that Taggert et al. (19) suggest that a randomized controlled trial of DES versus CABG might be unethical because there is no equipoise between the 2 arms. This impugns the work of the many surgeons, physicians, and other professionals who have designed, approved, and are participating in the large, ongoing randomized trials exploring this question. Equipoise is an interesting term. Who should be equally poised—the scientific community? Surgeon? Cardiologist? Subject? A significant minority of patients with unprotected LM stenosis are currently receiving PCI as the initial revascularization procedure. It would seem important for the wider medical community to know whether this is a reasonable option. Equipoise is currently a controversial topic among the ethicists; it would seem that the most important ethical consideration lies with the subject, and his/her willingness to give “… an adequately informed, free, and unexploited consent…equipoise is irrelevant” (53).

I do not know whether the randomized trials will demonstrate a broadened role for PCI in LM disease or not; perhaps that is the greatest endorsement for their performance. My guess is, however, that the trials will help to identify clinico-anatomical subgroups that do better with one or the other revascularization procedure as well as a group that does well with either. The term “well” is relative in that both CABG and PCI are likely to improve over time and provide better long-term results. The tradeoff of repeated interventions for less invasivity (assuming no difference in infarct-free survival or quality of life) may be a reasonable patient preference. Taggert et al. (19) raise important issues, however. Not the least of these is the perception that interventionalists are conflicted in the advice they give patients with regard to revascularization therapy and do not provide an opportunity for the surgeon to participate in the counseling process. The role of commercial interests in influencing interventional practice is also relevant. We owe it to our patients and our colleagues to practice in an open manner that is above suspicion. As for the current approach to LM disease, we can agree that percutaneous intervention is an acceptable option to CABG for patients with a protected LM as well as select patients with unprotected LM, including those with emergent presentation, those at very high risk for surgery, and perhaps those with disease limited to the ostium or shaft of the LM. Bifurcation disease is an unsettled question even with DES and should await the results of the ongoing randomized trials. We have little long-term data with DES for LM and almost none on the outcome of repeated PCI for LM restenosis. Optimal duration of antiplatelet therapy is not yet determined for DES use in the LM, although it would seem prudent at this time to favor prolonged dual antiplatelet therapy. Bypass surgery carries with it 3 intuitively beneficial aspects: 2 grafts to the left system are almost always used providing some protective redundancy; grafts are usually placed at sites that also protect against future disease in the proximal LAD and circumflex systems; and complete revascularization is more likely to be achieved by surgery in patients with LM accompanied by multivessel disease. In common with the many other editorials recently appearing on this topic, I end with a caution that in the absence of definitive data to the contrary, bypass surgery remains the standard of care for patients with unprotected LM coronary artery disease who are reasonable candidates for surgery (54–56).