American College of Cardiology

BRAUNWALD ET AL., MANAGEMENT OF PATIENTS WITH UNSTABLE ANGINA AND NON-ST-SEGMENT ELEVATION MYOCARDIAL INFARCTION UPDATE
http://www.acc.org/clinical/guidelines/unstable/incorporated/index.htm

ACC/AHA 2002 Guideline Update for the Management of Patients With Unstable Angina and Non-ST-Segment Elevation Myocardial Infarction

A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee on the Management of Patients With Unstable Angina)

IV. Coronary Revascularization

A. General Principles

As discussed in Section III, coronary angiography is useful for defining the coronary artery anatomy in patients with UA/NSTEMI and for identifying subsets of high-risk patients who may benefit from early revascularization. Coronary revascularization (PCI or CABG) is carried out to improve prognosis, relieve symptoms, prevent ischemic complications, and improve functional capacity. The decision to proceed from diagnostic angiography to revascularization is influenced not only by the coronary anatomy but also by a number of additional factors, including anticipated life expectancy, ventricular function, comorbidity, functional capacity, severity of symptoms, and quantity of viable myocardium at risk. These are all important variables that must be considered before revascularization is recommended. For example, patients with distal obstructive coronary lesions or those who have large quantities of irreversibly damaged myocardium are unlikely to benefit from revascularization, particularly if they can be stabilized with medical therapy. Patients with high-risk coronary anatomy are likely to benefit from revascularization in terms of both symptom improvement and long-term survival (Figure 12). The indications for coronary revascularization in patients with UA/NSTEMI are similar to those for patients with chronic stable angina and are presented in greater detail in the ACC/AHA/ACP-ASIM Guidelines for the Management of Patients With Chronic Stable Angina ⁽²⁶⁾, as well as in the ACC/AHA Guidelines for Coronary Artery Bypass Graft Surgery ⁽²⁷⁴⁾.

Plaque rupture with subsequent platelet aggregation and thrombus formation is most often the underlying pathophysiological cause of UA ^(1,18). The management of many patients with UA/NSTEMI often involves revascularization of the underlying CAD with either PCI or CABG. Selection of the appropriate revascularization strategy often depends on clinical factors, operator experience, and extent of the underlying CAD. Many patients with UA/NSTEMI have coronary disease that is amenable to either form of therapy. However, some patients have high-risk features, such as reduced LV function, that places them in a group of patients who experience improved long-term survival rates with CABG. In other patients, adequate revascularization with PCI may not be optimal or even possible, and CABG may be the better revascularization choice.

Findings in large registries of patients with CAD suggest that the mode of clinical presentation should have little bearing on the subsequent revascularization strategy. In a series of 9,263 patients with CAD, an admission diagnosis of UA (vs. chronic stable angina) had no influence on 5-year survival rates after CABG, percutaneous transluminal coronary angioplasty (PTCA), or medical treatment ⁽²⁸⁸⁾. An initial diagnosis of UA also did not influence survival 3 years after either CABG or PTCA in 59,576 patients treated in the state of New York ⁽²⁸⁹⁾. Moreover, long-term survival rates after CABG are similar for UA patients who present with rest angina, increasing angina, new-onset angina, or post-MI angina ⁽²⁹⁰⁾. These observations suggest that published data that compare definitive treatments for patients who initially present with multiple clinical manifestations of CAD can be used to guide management decisions for patients who present with UA/NSTEMI. Consequently, the indications for coronary revascularization in patients with UA/NSTEMI are, in general, similar to those for patients with stable angina. The principal difference is that the impetus for some form of revascularization is stronger in patients with UA/NSTEMI by the very nature of the presenting symptoms ⁽²⁹⁰⁾.

Recommendations for Revascularization with PCI and CABG in Patients with UA/NSTEMI (see Table 20)

Class I

1. CABG for patients with significant left main CAD. (Level of Evidence: A)
2. CABG for patients with 3-vessel disease; the survival benefit is greater in patients with abnormal LV function (EF less than 0.50). (Level of Evidence: A)
3. CABG for patients with 2-vessel disease with significant proximal left anterior descending CAD and either abnormal LV function (EF less than 0.50) or demonstrable ischemia on noninvasive testing. (Level of Evidence: A)
4. PCI or CABG for patients with 1- or 2-vessel CAD without significant proximal left anterior descending CAD but with a large area of viable myocardium and high-risk criteria on noninvasive testing. (Level of Evidence: B)
5. PCI for patients with multivessel coronary disease with suitable coronary anatomy, with normal LV function and without diabetes. (Level of Evidence: A)
6. Intravenous platelet GP IIb/IIIa inhibitor in UA/NSTEMI patients undergoing PCI. (Level of Evidence: A)

Class IIa

1. Repeat CABG for patients with multiple saphenous vein graft (SVG) stenoses, especially when there is significant stenosis of a graft that supplies the LAD. (Level of Evidence: C)
2. PCI for focal SVG lesions or multiple stenoses in poor candidates for reoperative surgery. (Level of Evidence: C)
3. PCI or CABG for patients with 1- or 2-vessel CAD without significant proximal left anterior descending CAD but with a moderate area of viable myocardium and ischemia on noninvasive testing. (Level of Evidence: B)
4. PCI or CABG for patients with 1-vessel disease with significant proximal left anterior descending CAD. (Level of Evidence: B)
5. CABG with the internal mammary artery for patients with multivessel disease and treated diabetes mellitus. (Level of Evidence: B)

Class IIb

PCI for patients with 2- or 3-vessel disease with significant proximal left anterior descending CAD, with treated diabetes or abnormal LV function, and with anatomy suitable for catheter-based therapy. (Level of Evidence: B)

Class III

1. PCI or CABG for patients with 1- or 2-vessel CAD without significant proximal left anterior descending CAD or with mild symptoms or symptoms that are unlikely due to myocardial ischemia or who have not received an adequate trial of medical therapy and who have no demonstrable ischemia on noninvasive testing. (Level of Evidence: C)
2. PCI or CABG for patients with insignificant coronary stenosis (less than 50% diameter). (Level of Evidence: C)
3. PCI in patients with significant left main coronary artery disease who are candidates for CABG. (Level of Evidence: B)

B. Percutaneous Coronary Intervention

In recent years, technological advances coupled with high acute success rates and low complication rates have increased the use of percutaneous catheter procedures in patients with UA/NSTEMI. Stenting and the use of adjunctive platelet GP IIb/IIIa inhibitors have further broadened the use of PCI by improving both the safety and durability of these procedures.

Percutaneous coronary revascularization (intervention) strategies are referred to in these guidelines as "PCI." This term refers to a family of percutaneous techniques, including standard balloon angioplasty (PTCA*), intracoronary stenting, and atheroablative technologies (e.g., atherectomy, thrombectomy, laser). The majority of current PCIs involve balloon dilatation and coronary stenting. Stenting has contributed greatly to catheter-based revascularization by reducing the risks of both acute vessel closure and late restenosis. Although stenting has become the most widely used percutaneous technique, and in 1998 it was used in approximately 525,000 of 750,000 PCIs, other devices continue to be used for specific lesions and patient subsets. Although the safety and efficacy of atheroablative and thrombectomy devices have been demonstrated, limited outcome data are available that describe the use of these new strategies specifically in patients with UA/NSTEMI ⁽²⁹¹⁾.

*PTCA is used to refer to studies in which this was the dominant form of PCI, before the widespread use of stenting.

In the absence of active thrombus, rotational atherectomy is useful to debulk arteries that contain large atheromatous burdens and to modify plaques in preparation for more definitive treatment with adjunctive balloon angioplasty or stenting. This approach is particularly well suited for use in hard, calcific lesions, in which it preferentially ablates inelastic tissue. Rotational atherectomy, even in patients with stable angina, may result in the release of CK-MB isoenzymes after seemingly uncomplicated procedures. This often reflects distal embolization of microparticulate matter and platelet activation, and the clinical outcome has been correlated with the magnitude of the enzyme elevation ⁽²⁹²⁾. The magnitude and frequency of postprocedural myocardial necrosis reflected in CK-MB enzyme rises can be reduced with concomitant treatment with a platelet GP IIb/IIIa inhibitor ^(293,294).

Other new techniques and devices, such as the use of Angiojet thrombectomy and extraction atherectomy (transluminal extraction catheter), are being tested for the treatment of thrombi that are visible within a coronary artery ⁽²⁹⁵⁾. In addition, there is some evidence that extraction atherectomy can be used to treat SVG disease through the removal of degenerated graft material and thrombus ⁽²⁹⁶⁾. In this situation, it often is used as an adjunct to more definitive therapy with balloon angioplasty and stents.

The reported clinical efficacy of PCI in UA/NSTEMI has varied. This is likely attributable to differences in study design, treatment strategies, patient selection, and operator experience. Nevertheless, the success rate of PCI in patients with UA/NSTEMI is often quite high. In TIMI IIIB, for example, angiographic success was achieved in 96% of patients with UA/NSTEMI who underwent balloon angioplasty. With clinical criteria, periprocedural MI occurred in 2.7%, emergency bypass surgery was required in 1.4%, and the death rate from the procedure was 0.5% ^(4,19,297).

The use of balloon angioplasty has been evaluated in several other trials of patients with UA vs. stable angina ^(298-303). A large retrospective study compared the results of angioplasty in patients with stable angina with that in patients with UA ⁽²⁹⁹⁾. After an effort to control patients with UA with medical therapy, PTCA was carried out an average of 15 days after hospital admission. In comparison with patients with stable angina, UA patients showed no significant differences with respect to primary clinical success (92% for UA vs. 94% for stable angina), in-hospital mortality rates (0.3% vs. 0.1%), or the number of adverse events at 6-month follow-up ⁽²⁹⁹⁾. These findings suggest that PTCA results in immediate and 6-month outcomes that are comparable in patients with stable angina and UA. In addition, in a retrospective analysis, the results in UA patients were similar regardless of whether the procedure was performed early (less than 48 h) or late (greater than 48 h) after hospital presentation ⁽²⁹⁸⁾.

Although other earlier studies (predominantly from the 1980s) have suggested that patients with UA who undergo balloon PTCA have higher rates of MI and restenosis compared with patients with stable angina ^(300-304), contemporary catheter revascularization often involves coronary stenting and adjunctive use of platelet GP IIb/IIIa receptor inhibitors, which are likely to affect not only immediate- but also long-term outcome ⁽²⁴⁶⁾. Historically, PTCA has been limited by acute vessel closure, which occurs in approximately 5% of patients, and by coronary restenosis, which occurs in approximately 35% to 45% of treated lesions during a 6-month period. Coronary stenting offers an important alternative to PTCA because of its association with both a marked reduction in acute closure and lower rates of restenosis. By preventing acute or threatened closure, stenting reduces the incidence of procedure-related STEMI and need for emergency bypass surgery and may also prevent other ischemic complications.

In a comparison of the use of the Palmaz-Schatz coronary stent in patients with stable angina and patients with UA, no significant differences were found with respect to in-hospital outcome or restenosis rates ⁽³⁰⁵⁾. Another study found similar rates of initial angiographic success and in-hospital major complications in stented patients with UA compared with those with stable angina ⁽³⁰⁶⁾. Major adverse cardiac events at 6 months were also similar between the 2 groups, whereas the need for repeat PCI and target vessel revascularization was actually less in the UA group. On the other hand, other recent data have suggested that UA increases the incidence of adverse ischemic outcomes in patients undergoing coronary stent deployment despite therapy with ticlopidine and ASA, which suggests the need for more potent antiplatelet therapy in this patient population ^(307,308).

1. Platelet Inhibitors and Percutaneous Revascularization
An important advance in the treatment of patients with UA/NSTEMI who are undergoing PCI has been the introduction of platelet GP IIb/IIIa receptor inhibitors (see Section III. B) ^{(10,18,21,244-246,309-311)}. This therapy takes advantage of the fact that platelets play an important role in the development of ischemic complications that may occur in patients with UA/NSTEMI or during coronary revascularization procedures. Currently, 3 platelet GP IIb/IIIa inhibitors are approved by the Food and Drug Administration based on the outcome of a variety of clinical trials: abciximab (ReoPro), tirofiban (Aggrastat), and eptifibatide (Integrilin). The Evaluation of c7E3 for the Prevention of Ischemic Complications (EPIC), Evaluation of PTCA and Improve Long-term Outcome by c7E3 GP IIb/IIIa receptor blockade (EPILOG), CAPTURE, and Evaluation of Platelet IIb/IIIa Inhibitor for STENTing (EPISTENT) trials investigated the use of abciximab; the PRISM, PRISM-PLUS, and Randomized Efficacy Study of Tirofiban for Outcomes and REstenosis (RESTORE) trials evaluated tirofiban; and the Integrilin to Minimize Platelet Aggregation and Coronary Thrombosis (IMPACT) and PURSUIT trials studied the use of eptifibatide (Figures 13 and 14). All 3 of these agents interfere with the final common pathway for platelet aggregation. All have shown efficacy in reducing the incidence of ischemic complications in patients with UA (Figure 10, Table 16).

In the EPIC trial, high-risk patients who were undergoing balloon angioplasty or directional atherectomy were randomly assigned to 1 of 3 treatment regimens: placebo bolus followed by placebo infusion for 12 h; weight-adjusted abciximab bolus (0.25 mg per kg) and 12-h placebo infusion; or weight-adjusted abciximab bolus and 12-h infusion (10 mcg per min) ^(244,309). In this trial, high risk was defined as severe UA, evolving MI, or high-risk coronary anatomy defined at cardiac catheterization. The administration of bolus and continuous infusion of abciximab reduced the rate of ischemic complications (death, MI, revascularization) by 35% at 30 days (12.8 vs. 8.3%, p = 0.0008), by 23% at 6 months, and by 13% at 3 years ^{(244,309,310)}. The favorable long-term effect was mainly due to a reduction in the need for bypass surgery or repeat PCI in patients with an initially successful procedure.

The administration of abciximab in the EPIC trial was associated with an increased bleeding risk and transfusion requirement. In the subsequent EPILOG trial, which used weight-adjusted dosing of concomitant heparin, the incidence of major bleeding and transfusion associated with abciximab and low-dose weight-adjusted heparin (70 U per kg) was similar to that seen with placebo ⁽²⁴⁵⁾. The cohort of patients with UA undergoing PCI in the EPILOG trial demonstrated a 64% reduction (10.1% to 3.6%, p = 0.001) in the composite occurrence of death, MI, or urgent revascularization to 30 days with abciximab therapy compared with placebo (standard-dose weight-adjusted heparin).

The RESTORE trial was a randomized double-blind study that evaluated the use of tirofiban vs. placebo in 2,139 patients with UA or AMI, including patients with non-Q-wave MI who underwent PCI (balloon PTCA or directional atherectomy) within 72 h of hospitalization ⁽³¹²⁾. The trial was designed to evaluate both clinical outcomes and restenosis. Although the infusion of tirofiban (bolus of 10 mcg per kg followed by a 36-h infusion at 0.15 mcg · kg^-1 · min^-1) had no significant effect on the reduction in restenosis at 6 months, a trend was observed for a reduction in the combined clinical end point of death/MI, emergency CABG, unplanned stent placement for acute or threatened vessel closure, and recurrent ischemia compared with placebo at 6 months (27.1% vs. 24.1%, p = 0.11).

The clinical efficacy of tirofiban was further evaluated in the PRISM-PLUS trial, which enrolled patients with UA/NSTEMI within 12 h of presentation ⁽²¹⁾ (see Section III). Among patients who underwent PCI, the 30-day incidence of death, MI, refractory ischemia, or rehospitalization for UA was 15.3% in the group that received heparin alone compared with 8.8% in the tirofiban/heparin group. After PCI, death or nonfatal MI occurred in 10.2% of those receiving heparin vs. 5.9% of tirofiban-treated patients.

Eptifibatide, a cyclic heptapeptide GP IIb/IIIa inhibitor, has also been administered to patients with ACS. In the PURSUIT trial, nearly 11,000 patients who presented with an ACS were randomized to receive either UFH and ASA or eptifibatide, UFH, and ASA ⁽¹⁰⁾. In patients undergoing PCI within 72 h of randomization, eptifibatide administration resulted in a 31% reduction in the combined end point of nonfatal MI or death at 30 days (17.7 vs. 11.6%, p = 0.01).

The EPISTENT trial was designed to evaluate the efficacy of abciximab as an adjunct to elective coronary stenting ^(246,313). Of the nearly 2,400 patients who were randomized, 20% of the stented patients had UA within 48 h of the procedure. Patients were randomly assigned to either stent deployment with placebo, stent plus abciximab, or PTCA plus abciximab. Nineteen percent of the PTCA group had provisional coronary stent deployment for a suboptimal angioplasty result. All stented patients in this trial received oral ASA (325 mg) and oral ticlopidine (250 mg twice daily for 1 month). The adjunctive use of abciximab was associated with a significant reduction in the composite clinical end point of death, MI, or urgent revascularization. The 30-day primary end point occurred in 10.8% of the stent-plus-placebo group, 5.3% of the stent-plus-abciximab group, and 6.9% of the PTCA-plus-abciximab group. Most of the benefit from abciximab were related to a reduction in the incidence of moderate to large MI (CK greater than 5 times the upper limit of normal or Q-wave MI); these reductions occurred in 5.8% of the stent-plus-placebo group, 2.6% of the balloon-plus-abciximab group, and 2.0% of the stent-plus-abciximab group.

At 1 year of follow-up, stented patients who received bolus and infusion abciximab had reduced mortality rates compared with patients who received stents without abciximab (1.0% vs. 2.4%, representing a 57% risk reduction; p = 0.037) ⁽³¹⁴⁾. In diabetics, target vessel revascularization at 6 months was markedly and significantly reduced (51%, p = 0.02) in stented patients who received abciximab compared with those who did not. Although a similar trend was also observed in nondiabetic patients, it did not reach statistical significance.

The Enhanced Suppression of Platelet Receptor GP IIb/IIIa Using Integrilin Therapy (ESPRIT) trial was a placebo-controlled trial designed to assess whether eptifibatide improved the outcome of patients undergoing stenting ⁽⁵⁴²⁾. Fourteen percent of the 2064 patients enrolled in ESPRIT had UA/NSTEMI. The primary end point (the composite of death, MI, target-vessel revascularization, and "bailout" GP IIb/IIIa inhibitor therapy) was reduced from 10.5% to 6.6% with treatment (p = 0.0015). There was consistency in the reduction of events in all components of the composite end points and in all major subgroups, including patients with UA/NSTEMI. Major bleeding occurred more frequently in patients who received eptifibatide (1.3%) than in those who received placebo (0.4%; p = 0.027). However, no significant difference in transfusion occurred. At 1-year follow-up, death or MI occurred in 12.4% of placebo-track patients and 8.0% of eptifibatide-treated patients (p = 0.001) ⁽⁵⁴³⁾.

In the only head-to-head comparison of 2 GP IIb/IIIa inhibitors, the Tirofiban and Reopro Give Similar Efficacy Outcomes Trial (TARGET) randomized 5,308 patients to tirofiban or abciximab before undergoing PCI with the intent to perform stenting ⁽⁵⁴⁴⁾. The primary end point, a composite of death, nonfatal MI, or urgent target-vessel revascularization at 30 days, occurred less frequently in those receiving abciximab than tirofiban (6.0% vs. 7.6%, p = 0.038). There was a similar direction and magnitude for each component of the end point. The difference in outcome between the 2 treatment groups may be related to a suboptimal dose of tirofiban resulting in inadequate platelet inhibition.

Eptifibatide has not been compared directly to either abciximab or tirofiban.

In summary, data from both retrospective observations and randomized clinical trials indicate that PCI can lead to angiographic success in most patients with UA/NSTEMI (Figures 13 and 14). The safety of these procedures in these patients is enhanced by the addition of intravenous platelet GP IIb/IIIa receptor inhibitors to the standard regimen of ASA, heparin, and anti-ischemic medications.

C. Surgical Revascularization

Two randomized trials conducted in the early years of CABG compared medical and surgical therapy in UA. The National Cooperative Study Group randomized 288 patients at 9 centers between 1972 and 1976 ⁽³¹⁷⁾. The Veterans Administration (VA) Cooperative Study randomized 468 patients between 1976 and 1982 at 12 hospitals (269,319-321). Both trials included patients with progressive or rest angina accompanied by ST-T-wave changes. Patients greater than 70 years old or with a recent MI were excluded; the VA study included only men. In the National Cooperative Study, the hospital mortality rate was 3% for patients undergoing medical therapy and 5% after CABG (p = NS). Follow-up to 30 months showed no differences in survival rates between the treatment groups. In the VA Cooperative Study, survival rates to 2 years were similar after medical therapy and CABG overall and in subgroups defined by the number of diseased vessels. A post hoc analysis of patients with depressed LV function, however, showed a significant survival advantage with CABG regardless of the number of bypassed vessels ⁽³²¹⁾.

All randomized trials of CABG vs. medical therapy (including those in stable angina) have reported improved symptom relief and functional capacity with CABG. However, long-term follow-up in these trials has suggested that by 10 years, there is a significant attenuation of both the symptom relief and survival benefits previously conferred by CABG, although these randomized trials reflect an earlier era for both surgical and medical treatment. Improvements in anesthesia and surgical techniques, including internal thoracic artery grafting to the LAD, and improved intraoperative myocardial protection with cold potassium cardioplegia, are not reflected in these trials. In addition, the routine use of heparin and ASA in the acute phase of medical therapy and the range of additional therapeutic agents that are now available (e.g., LMWH, GP IIb/IIIa inhibitors) represent significant differences in current practice from the era in which these trials were performed.

A meta-analysis was performed on the results of 6 trials conducted between 1972 and 1978 to compare long-term survival in CAD patients treated medically or with CABG ⁽¹⁴²⁾. A clear survival advantage was documented for CABG in patients with left main and 3-vessel coronary disease that was independent of LV function. No survival difference was documented between the 2 therapies for patients with 1- or 2-vessel coronary disease.

Pocock et al. ⁽³²²⁾ performed a meta-analysis on the results of 8 randomized trials completed between 1986 and 1993 and compared the outcomes of CABG and PTCA in 3,371 patients with multivessel CAD before widespread stent use. Many of these patients presented with UA. At 1-year follow-up, no difference was documented between the 2 therapies in cardiac death or MI, but a lower incidence of angina and need for revascularization was associated with CABG.

The Bypass Angioplasty Revascularization Investigation (BARI) trial is the largest randomized comparison of CABG and PTCA in 1,829 patients with 2- or 3-vessel CAD ^(323,324). UA was the admitting diagnosis in 64% of these patients, and 19% had treated diabetes. A statistically significant advantage in survival without MI independent of the severity of presenting symptoms was observed in the entire group for CABG over PCI 7 years after study entry (84.4% vs. 80.9%, p = 0.04) ⁽³²⁵⁾. However, subgroup analysis demonstrated that the survival benefit seen with CABG was confined to diabetic patients treated with insulin or oral hypoglycemic agents. At 7 years, the survival rate for diabetics was 76.4% with CABG compared with 55.7% among patients treated with PTCA (p = 0.001). In patients without diabetes, survival rates were virtually identical (CABG vs. PTCA, 86.4% vs. 86.8%, p = 0.71). Subsequent analysis of the Coronary Angioplasty versus Bypass Revascularisation Investigation (CABRI) trial results also showed a survival benefit for the use of CABG in comparison with PTCA in diabetic patients with multivessel CAD ⁽³²⁶⁾. These observations have been confirmed in a study from Emory University, which showed that with correction for baseline differences, there were improved survival rates for insulin-requiring patients with multivessel disease who were revascularized with CABG rather than with PTCA ⁽³²⁷⁾ (see Section VI. C).

Other nonrandomized analyses have compared CABG, PTCA, and medical therapy. With statistical adjustment for differences in baseline characteristics of 9,263 consecutive CAD patients entered into a large registry, the 5-year survival rates were compared for patients who were treated medically and those who underwent PTCA and CABG between 1984 and 1990 ⁽²⁸⁸⁾. Patients with 3- or 2-vessel disease with a proximal severe (greater than or equal to 95%) LAD stenosis treated with CABG had significantly better 5-year survival rates than did those who received medical treatment or PTCA. In patients with less severe 2-vessel CAD or with 1-vessel CAD, either form of revascularization improved survival relative to medical therapy. The 2 revascularization treatments were equivalent for patients with nonsevere 2-vessel disease. PTCA provided better survival rates than CABG in patients with 1-vessel disease except for those with severe proximal LAD stenosis, for whom the 2 revascularization strategies were equivalent. However, in patients with 1-vessel disease, all therapies were associated with high 5-year survival rates, and the differences among the treatment groups were very small.

Hannan et al. ⁽²⁸⁹⁾ compared 3-year risk-adjusted survival rates in patients undergoing revascularization in the state of New York in 1993. The 29,646 CABG patients and 29,930 PTCA patients had different baseline and angiographic characteristics evaluated with Cox multivariable models. The anatomic extent of disease was the only variable that interacted with the specific revascularization therapy that influenced long-term survival. Although the limitations of such observational studies must be recognized, it is of interest that UA or diabetes did not result in treatment-related differences in long-term survival rates. Patients with 1-vessel disease not involving the LAD or with less than 70% LAD stenosis had statistically significant higher adjusted 3-year survival rates with PTCA (95.3%) than with CABG (92.4%). Patients with proximal LAD stenosis of greater than or equal to 70% had statistically significant higher adjusted 3-year survival rates with CABG than with PTCA regardless of the number of diseased coronary vessels. Patients with 3-vessel disease had statistically significant higher adjusted 3-year survival rates with CABG regardless of proximal LAD disease. Patients with other 1- or 2-vessel disease had no treatment-related difference in survival rates.

Thus, large cohort trials with statistical adjustment showed that survival differences between CABG and PTCA were related to the anatomic extent of disease, in contrast to the randomized trials of multivessel disease that showed no differences. This difference may be due to the smaller numbers of patients in the randomized trials and, hence, their lower power and to the fact that a broad range of angiographic characteristics were not included in the randomized trials in comparison with the patient cohort studies. The location of a coronary stenosis in the LAD, especially if it is severe and proximal, appears to be a characteristic associated with higher mortality rates and, therefore, with a more favorable outcome with CABG. As already noted, the finding in the BARI and CABRI randomized trials that diabetes appeared to identify a subset of patients who had a better outcome with CABG than with PTCA was not confirmed in the 2 cohort studies ^{(323,324,326)}. Analysis of the diabetic subgroup was not proposed at the time of trial design in either the BARI or CABRI trial. Moreover, this treatment-related effect was not reproduced in the BARI registry population ⁽³²⁸⁾. A reasonable explanation is that in the cohort studies, physicians may be able to recognize characteristics of coronary arteries of diabetic patients that will permit them to more safely undergo one or another of the revascularization therapies. However, when all diabetic patients are randomly assigned to therapies without the added insight of clinical judgment, a treatment advantage is apparent for CABG. Until further studies that compare newer percutaneous devices (in particular, stents) and surgical techniques can more clearly resolve these differences, it is reasonable to consider CABG as the preferred revascularization strategy for most patients with 3-vessel disease, especially if it involves the proximal LAD and patients with multivessel disease and treated diabetes or LV dysfunction. Alternatively, it would be unwise to deny the advantages of PCI to a patient with diabetes and less severe coronary disease on the basis of the current information.

An important consideration in a comparison of different revascularization strategies is that none of the large randomized trials reflect the current practice of interventional cardiology that includes the routine use of stents and the increasing use of platelet receptor inhibitors. Coronary stenting improves procedural safety and reduces restenosis in comparison with PTCA. The adjuvant use of platelet inhibitors, particularly in high-risk patients, is also associated with improved short- and intermediate-term outcomes. Although the effects of coronary stenting and platelet GP IIb/IIIa inhibitors would have likely improved the PCI results observed, their added benefit relative to CABG cannot be assessed on the basis of the previously reported randomized trials or large registries. Refinement of surgical management with right internal mammary artery grafts, radial artery grafts, retroperfusion, and less invasive methodology may reduce the morbidity rates for CABG, but no recent advance has been shown to influence long-term survival more favorably than the current standard operative technique. Therefore, decisions regarding appropriate revascularization strategies in the future will have to be made on the basis of information that compares long-term outcome for these 2 techniques and the effects of adjunctive pharmacotherapy.

D. Conclusions

In general, the indications for PCI and CABG in UA/NSTEMI are similar to those for stable angina ^{(324,329-333)}. High-risk patients with LV systolic dysfunction, patients with diabetes mellitus, and those with 2-vessel disease with severe proximal LAD involvement or severe 3-vessel or left main disease should be considered for CABG (Figure 12). Many other patients will have less-severe CAD that does not put them at high risk for cardiac death. However, even less-severe disease can have a substantial negative impact on the quality of life. Compared with high-risk patients, low-risk patients will receive negligibly or very modestly increased chances of long-term survival with CABG. Therefore, in low-risk patients, quality of life and patient preferences are given more weight than are strict clinical outcomes in the selection of a treatment strategy. Low-risk patients whose symptoms do not respond well to maximal medical therapy and who experience a significant negative impact on their quality of life and functional status should be considered for revascularization. Patients in this group who are unwilling to accept the increased short-term procedural risks to gain long-term benefits or who are satisfied with their existing capabilities should be managed medically at first and followed carefully as outpatients. Other patients who are willing to accept the risks of revascularization and who want to improve their functional status or to decrease symptoms may be considered appropriate candidates for early revascularization.