In evaluating the effect of therapy in these two broad groups of patients, the metrics of treatment must be taken into consideration (Table 23.3). Selection of the specific metrics of evaluation has important implications for trial design and interpretation as well as for societal recommendations and patient expectations. Some trials have used single discrete endpoints; others have used a composite. The specific endpoint chosen has great implications for both trial design and interpretation. For example as in the SYNTAX trial [24, 25] if the primary endpoint of the trial is negative, then all other analyses can only be considered hypothesis generating. This makes subsequent analyses of subsets of patients and lesions in these trials considerably less strong scientifically and restricts their interpretation and inclusion in Guideline documents. Another example relates to the use of composite endpoints which is very common in trial design in an attempt to decrease the sample size required. For example if the composite endpoint of one trial is death and/or MI and urgent revascularization compared with a composite of death/MI and any repeat revascularization or compared with only a composite of death and/or MI in another trial, it is very difficult to compare trial conclusions [26, 27]. This latter is of particular importance in trials comparing CABG and PCI.

In an era increasingly focused on patient centered outcomes, these metrics have taken on added importance because each patient may have a unique idea on the risk benefit ratio of a procedure in terms of a specific endpoint for their own situation. Hard endpoints include death, myocardial infarction, stroke and need for cardiovascular surgery. Stent thrombosis may be included in this because it often results in either death or myocardial infarction although there may be other reasons for the latter two adverse outcomes than stent thrombosis [28–34].

As an endpoint, the definition of myocardial infarction has important implications (vide infra) [35–37]. A periprocedural elevation of a biomarker such as troponin in an asymptomatic patient, while it is sometimes used as an endpoint, has different clinical consequences than a STEMI which occurs spontaneously. The former may not be considered important by the patient even though it may be used as an endpoint in trial design while the latter is of great importance. The relative importance of each of these hard endpoints varies from patient to patient.

For many patients, stroke is the most important concern [38–40]. A variety of cerebrovascular endpoints have been used. In some studies, a completed stroke is the specific endpoint, while in others, the combination of stroke or TIA may be used. Recently, the use of Diffusion Weighted MRI imaging has been used to identify asymptomatic abnormalities in the cerebrovascular system. Because of the enhanced sensitivity of this diagnostic modality, up to 80 % of patients undergoing some procedures, for example TAVR, are found to have cerebral abnormalities associated with the procedure despite the fact that a clinical event is low [40]. If DWMRI changes become used as an endpoint for either follow-up or trial design, that will have important implications for trial results and conclusions.

As previously mentioned, an important endpoint for revascularization strategies is myocardial infarction [35–37]. This has assumed increasing importance in patients being treated with PCI but also has implications for cardiac surgical procedures. Treatment of more complex lesions and application of high sensitivity analytic techniques for cardiac biomarkers has led to a marked increase in the rate of diagnosis of periprocedural infarctions. The clinical consequences of these elevated biomarkers depends on the setting (e.g. PCI complicated by transient acute closure or side branch occlusion vs PCI without any procedural complications) as well as the level of elevation of the specific biomarker for the detection algorithm. The 2010 Third Universal Definition of Myocardial Infarction has been developed to bring clarity to this situation [37]. This document defines myocardial infarction associated with PCI by the presence of two findings: 1. Elevation of cTn values >5 × 99th percentile upper reference limit in patients with normal baseline values or a >20 % rise in cTn if the baseline values are elevated but stable or falling and 2. Symptoms suggestive of myocardial ischemia or new ischemia ECG changes or procedural evidence of abnormal flow in a major coronary artery or new regional wall motion abnormalities. Given that myocardial infarction is a typical endpoint in PCI trials, understanding of these definitions is of great importance. In addition the timing and circumstances of the biomarker elevation is also important. If the endpoint of any trial of PCI versus medical therapy includes asymptomatic biomarker elevation, the results of the comparison would be very different than if the trial endpoint was death/myocardial infarction rather than only biomarker elevation.

Other endpoints are also used including repeat revascularization either target vessel revascularization (TVR) or target vessel failure (TVF) which may be more difficult to assess. While the need for repeat PCI is important, provided that it is successful and uncomplicated, it may not have the same patient significance as the need for CABG.

Finally, softer endpoints include functional status, presence and severity of angina, need for nitrates, quality of life (QOL) and cost effectiveness [41–44]. Although grading schemes for quantitating some of these endpoints have been developed, these remain more subjective even though they are of fundamental importance to patients. Consideration of these endpoints is of great importance in patient selection for PCI [42]. In the era of Comparative Effectiveness research, increasingly patient centered outcomes will occupy a more central position even though they are more subjective.

The treatment of chronic stable angina continues to evolve. This evolution in some measure began with the publication of the multicenter randomized clinical trial – COURAGE, which has had a dramatic impact on the field [27]. This trial has been the focus of numerous subsequent analyses, in large part because of its profound implications [45–54]. The underlying hypothesis of COURAGE was that a strategy of PCI with optimal medical therapy would reduce the risk of death and non-fatal reinfarction compared with optimal medical therapy alone. There are several important considerations concerning this trial: (1) All patients had undergone angiography but only a small minority (6.4 %) were enrolled, (2) The trial was aimed at patients with stable symptoms, not an acute coronary syndrome. A substantial portion of patients (~40 %) had either no angina or only Class I angina, (3) 30 % had single vessel disease and the ejection fractions was well preserved at ~60 %, (4) Drug-eluting stents (DES) were not routinely used, (5) The need for repeat revascularization was not part of the composite primary endpoint, (6) There was superb adherence to medical therapy in both limbs. In such a trial with these patient characteristics, the chances of the endpoints of either mortality or non-procedural related MI being different between the two limbs is extremely small.

The findings from the COURAGE Trial documented several important points: (1) the primary endpoint of death and MI was significantly increased in the PCI + optimal medical therapy limb. This was related in large part to the inclusion of procedural related infarction as part of the composite endpoint (2) PCI resulted in improvement in symptoms and a decrease in the need for subsequent revascularization. Approximately 40 % of medically treated COURAGE patients crossed over to PCI.

These findings which occurred on the background of documented large variability in rates of performance of PCI have led to a decrease in PCI for chronic stable angina and more emphasis on optimal medical therapy as the initial therapeutic strategy [13]. However, there have been concerns because of the low risk patient groups included in the trial as well as the lack of significant ischemia at baseline. These concerns have led to several new important trials.

The most recent trial, FAME II [27], was a logical extension of the COURAGE trial given the emphasis on the importance of borderline ischemia in longer-term clinical outcomes. In trials of revascularization versus medical therapy, the importance of the presence and degree of ischemia at baseline had not been well studied. As part of this, some patients in former trials and registries had not undergone stress evaluation prior to the procedure to either identify or quantitate ischemia and this may be of great importance in determining longer term outcome. FFR has been validated as a marker of ischemia in multiple clinical and randomized trials [55–62]. Accordingly, the FAME-II investigators aimed at studying the clinical outcomes of FFR guided contemporary PCI plus optimal medical therapy (OMT) versus OMT alone. Using FFR at baseline, this trial could therefore focus on the two different strategies in patients all of whom had ischemia.

FAME II included stable patients with one, two or three vessel CAD scheduled for elective stenting with DES. For trial entry, at least one stenosis had to have a measured FFR ≤0.80. Patients in whom no lesion had an FFR ≤0.80 were entered into a registry. Six-hundred ninety-one patients were randomized to either PCI + OMT (n = 352) or OMT alone (339). Overall, both groups were similar in baseline clinical and angiographic characteristics. Particularly relevant comparisons using FFR guidance included the mean number of stenoses/patient (1.50 ± 0.8 vs 1.42 ± 0.7). On a per patient basis 1 stenotic vessel was present in 51.0 % vs 54.0 %, 2 stenotic vessels in 38.6 % vs 32.8 % and proximal or mid LAD stenosis in 33.9 % vs 34.8 %, for PCI + OMT vs OMT alone respectively. The average FFR 0.69 was the same for each group. The study was stopped prematurely by the DSMB for safety considerations before total patient enrollment was completed. Overall, the composite combined endpoint of death, myocardial infarction or urgent revascularization occurred in 4.3 % in the PCI group vs 12.7 % in the medical group (p < 0.001) (Fig. 23.1). This safety consideration was the marked imbalance between the two groups in the frequency of urgent target vessel revascularization. In the PCI group, 1.6 % required urgent revascularization as compared to 11.1 % in the medical therapy group (HR 0.13, 95 % CI 0.06–0.30, p < 0.001). Nonurgent revascularization was also increased in the OMT alone group vs the patients treated with PCI + OMT. There was, however, no significant difference in death or MI. Accordingly, a strategy of treating ischemia + OMT in patients with stable CAD should be considered standard not by virtue of improving survival or preventing infarction but in dramatically decreasing the need for subsequent revascularization which is often required for urgent symptoms.

Similar to the COURAGE trial, controversy exists and concerns with the FAME II trial have been raised including stopping the trial early (although that was the unanimous recommendation of the DSMB) as well as inclusion of urgent revascularization in the composite endpoint. Another important consideration is the fact that the results of angiography were known to both the patient and physician; such information may have potentially affected the judgement about performing subsequent revascularization in the patients being treated with OMT. For example, knowledge that there is an untreated severe LAD lesion may affect subsequent care suggestions. This issue will be addressed and excluded from consideration in the forthcoming ISCHEMIA trial [63].

The ISCHEMIA Trial [63] (International Study of Comparative Effectiveness with Medical and Invasive Approaches) has been designed to study management strategies in patients who have stable coronary artery disease but documented moderate to severe ischemia. This trial has the advantage compared to the COURAGE and FAME 2 trials that patients will be randomized on the basis of at least moderate ischemia on stress testing documented but prior to angiography. After the stress test, randomization will be used to select patients to be scheduled for coronary angiography followed by revascularization plus optimal medical therapy or to a strategy of optimal medical therapy alone with catheterization and revascularization performed only if the patient develops refractory symptoms or an acute coronary syndrome or reaches a primary endpoint. The primary endpoint for the trial will be the time to CV death, myocardial infarction, hospitalization for unstable angina, resuscitated cardiac death or heart failure. This randomization strategy will eliminate the biases seen in trials where angiography is performed and the results known before randomization. Recruitment however is estimated to be a lengthy process and power calculations include the need to randomize approximately 8,000 patients.

Given the information that is currently available, the goals of therapy in patients with stable angina treated with PCI are improvement in symptoms, reduction in ischemia and prevention of recurrent angina. Softer endpoints such as QOL, and reduction in medications required for symptom control may also be improved. In these patients, however, there is not apt to be a survival benefit or a reduction in development of subsequent myocardial infarction. This later prediction is a result of the fact that lesions associated the development of subsequent infarction are often only mild to moderate in angiographic severity, being more closely related to the underlying characteristics of the plaque such as the presence of a thin cap fibroatheroma. Reduction in anginal frequency, however, is an important goal. In a recent study of patients with angina followed as outpatients by primary care physician, approximately 30 % had weekly angina [64]. Angina occurring weekly, even if the pattern was stable, was associated with worse quality of life. Accordingly, PCI could improve quality of life by reducing the frequency of angina.

Patient selection criteria include consideration of the degree and severity of CAD, LV function, specific lesion morphology and co-morbid conditions which could impact on the outcome. The specific technical approach varies depending on details of the coronary anatomy. Based on the FAME-II trial [27], efforts should be focused on ischemia producing lesions. In some patients, with positive non-invasive tests in the distribution of single vessel CAD with a severe stenosis, FFR does not need to be performed. In other patients with MVD, FFR is very important for documenting ischemia and selecting lesions requiring treatment.

In patients with SVD, chronic stable angina but persistent ischemia, PCI is the preferred strategy for revascularization if the lesion is suitable. This has important implications if the patient has SVD with a single CTO but viability and persistent ischemia. In this case, intense efforts should be undertaken to achieve a successful intervention. Selection of the specific device to be used is based upon several clinical and angiographic factors. Given that the goal is symptom improvement and reduction in need for repeat revascularization, DES’s are preferred because of their effectiveness in reducing restenosis. In the past, there was substantial concern about stent thrombosis which may occur late, even after 1 year and result in death and/or MI [28–34, 65–70]. Much of the information on stent thrombosis with DES was obtained using first generation DES’s. With current generation DES, the results are improved with less stent thrombosis and improved TVR [3, 71–82] (Fig. 23.2). This is the result of improvement in all three components of DES – the metallic backbone which has migrated toward cobalt chromium and thinner stent struts, improved polymers some of which are biodegradable to enhance normal healing and iterations of the specific anti proliferative drug which is now usually from the “limus” family. DES continue to be used much more frequently than BMS because of improved restenosis rates. The most recent data would indicate that DES is used in 80–90 % of all PCI procedures.

Selection of the specific DES is dependent on several factors. All the third generation devices are very deliverable and have excellent success rates. Head to head comparisons have documented some differences in longer term outcome [3, 71–82]. Compared with first generation paclitaxel eluting stents, everolimus-eluting stents has been found to be superior and associated with decreased adverse outcomes including repeat revascularization, myocardial infarction and stent thrombosis [3, 71, 72, 75–77, 82]. The everolimus stents may also be associated with decreased myocardial infarction and stent thrombosis compared with sirolimus eluting stents [73]. Other stents with different drugs such as the Endeavor zotarolimus eluting stents are also significantly more effective and safer than first generation paclitaxel eluting stents [78, 79]. The results of trials comparing the zotarolimus eluting stents compared with sirolimus eluting stents have been variable [3, 78, 79]. In actual practice, typically the choice of specific stent depends upon operator preference and device availability. Most catheterization laboratories do not stock product from all manufacturers; instead the selection may be based on hospital inventory or specific contractural relationships with manufacturers.

Procedural performance also varies widely. The optimal long-term results are obtained with matching the stent size to the vessel size and then optimizing stent deployment. Evaluation of vessel size may be problematic particularly in the setting of diffuse disease. IVUS is particularly important in this regard because it allows for more accurate vessel sizing and then can be used to assess deployment and evaluate the need for post stent deployment with a larger balloon or a noncompliant balloon inflated to higher pressures. OCT has also been used for this purpose and is excellent at evaluating stent tine apposition [83–88]. Persistent malapposition has been associated with restenosis.

Bare metal stents are used less frequently although they remain an important strategy. The cost effectiveness has been studied and because of the increased cost of DES, BMS may be chosen electively in large vessels with short lesions in which restenosis rates can be predicted to be low. In this regard, Steinberg et al. [89] evaluated the effectiveness of bare metal versus DES in 466 patients undergoing revascularization with coronary vessels ≥3.5 mm. Target lesion revascularization and target vessel revascularization rates were low and comparable during follow-up. Major adverse cardiac events at 1 year were documented in 8.5 % of the 233 patients treated with DES versus 7.7 % of the BMS patients. Importantly, there were no episodes of either subacute or late stent thrombosis in either group. A more common reason to select a BMS relates to the need for DAPT. The issue of DAPT has occupied a central place in interventional cardiology. This is the result of the issue of stent thrombosis with its associated morbidity and mortality. The specific role of antiplatelet therapy following PCI has been the subject of multiple trials. Early studies evaluated antiplatelet and anti-thrombotic agents either alone or in combination [90–94]. The initial studies focused on ticlopidine and documented that stent thrombosis rates were significantly lower with the combination of ASA plus ticlopidine compared to ASA plus warfarin or ASA alone. Subsequent studies focused on clopidogrel rather than ticlopidine because of its improved side effect profile and clopidogrel became the standard of care for DAPT along with ASA.

The whole field of interventional cardiology has benefitted by the universal acceptance of the Academic Research Consortium (ARC) definitions for stent thrombosis [34]. Prior to that, because multiple definitions were used, it was difficult to evaluate and make comparisons. Although stent thrombosis had been had been well documented with BMS [29, 30], there was heightened scrutiny after the introduction and wide spread use of DES particularly late stent thrombosis which was related to discontinuation of DAPT. There have been numerous studies on the issues of specific drugs and dosages and durations [95–108]. Information about the pharmacogenomics has led to new agents e.g. Prasugrel and Ticagrelor as alternatives to clopidogrel [109–114]. The optimal role of either for widespread use as well as the role of pharmacogenomic testing continues to evolve. In contrast to the almost universal shift from ticlopidine to clopidogrel, such a general move in use to one of these new agents has not occurred. This relates to cost among other considerations such as side effects which may vary considerably. The major concerns in the field have been the specific drug and duration of therapy [115]. The tradeoff between bleeding and stent thrombosis has very important implications [115–117, 122]. Although risk factors for stent thrombosis have been identified, they are imperfect. This has to be considered on the background that stent thrombosis may occur very late several years after the index procedure in which case patients at potential risk for stent thrombosis will need very prolonged DAPT. Subsequent guidelines [116] were developed mainly on the basis of expert consensus opinion with the recommendation that patients with DES should receive DAPT for 1 year after the index stent placement. Such recommendations have important implications because of the issues of bleeding and cost. These are particularly important in patients requiring warfarin for other indications such as atrial fibrillation or those undergoing triple therapy for optimization of PCI [103, 117–122]. In these patients with triple therapy, bleeding rates are markedly increased. This will occur irrespective of whether the patient is being treated with warfarin or one of the newer anticoagulant drugs. In this setting, some practices prefer to use BMS because of the shortened need for both ASA and a thienopyridine. Other practices try to minimize the long term use of oral AC – e.g. in patients with atrial fibrillation in whom warfarin is administered for stroke prevention, patients may be offered a LAA occlusion device. Recently, a small RCT has addressed this issue [121]. In this trial 573 patients who would have received triple therapy were randomized to either warfarin + DAPT or warfarin + clopidogrel. The authors found that warfarin plus clopidogrel was associated with equivalent protection for ST and MACCE but significantly less bleeding. In the patients receiving double therapy bleeding episodes were seen in 19.4 % versus 44.6 % in the triple therapy group (HR 0.36, 95 % CI 0.26–0.50, p < 0.001). In addition, the cumulative incidence of death, myocardial infarction, stroke, TVR and stent thrombosis was also improved (Fig. 23.3). At the present time, there are no completed RCT’s with stent thrombosis as the primary endpoint to definitively determine the risk benefit ratio of prolonged DAPT and at the present time, practice at least in Europe is influenced by trials using composite endpoints such as death, MI and revascularization. These trials although not adequately powered for stent thrombosis have resulted in shorter duration of DAPT [103]. In contrast in the United States, 1 year of DAPT is still recommended routinely. Ongoing trials such as DAPT will shed light on this [122].

The issue of DAPT and stent thrombosis plays out against a background of increasing information on the variable incidence of stent thrombosis depending on the specific stent. There are new multiple meta-analyses and registry experiences [71, 74–76, 123]. A recent meta-analysis of 11 randomized clinical trials compared EES with other DES in 16,775 patients [123]. The comparator to EES included PES in 5 trials of 7,133 patients, SES in 5 trials of 7,370 patients and a single trial of the resolute zotarolimus eluting stent (n = 2,292). The investigators documented a 62 % relative risk reduction in ST with EES and a 0.6 % absolute risk reduction. Other studies have also studied differences in ST between different stent types. These studies vary depending upon the specific comparator. In general, the newer generation DES are much improved compared to the initial PES. An important finding has been that ST rates with newer generation DES stents are very low. New technology with biogradable polymers may even result in further improvements in both clinical and angiographic outcome (Table 23.4). This data will have important implications on the issue of duration of DAPT.

Another consideration in recommendation for patients with chronic angina is the degree and severity of CAD. The importance of this has been highlighted in the multicenter randomized clinical trial of PCI with DES versus CABG in patients with extensive two or three vessels or LMCA disease (SYNTAX) [24, 25, 124]. In this study, a score of angiographic severity of disease was developed (SYNTAX SCORE). Tertiles of severity were evaluated. At 5 years of follow up using the composite endpoint of MACCE, patients in the lowest tertile had outcomes which were not statistically significantly different between PCI and CABG while patients in the middle and upper tertile had improved outcome with CABG including a reduction in mortality in the highest tertile [24, 25] (Fig. 23.4).

Patients with diabetes have presented some specific challenges in the selection of revascularization strategy. They have been intensively studied since the early days of coronary intervention beginning with the BARI trial [126–130]. Controversy has existed concerning the relative merits of PCI versus CABG [126–130]. The recent FREEDOM Trial [131] of PCI with DES versus CABG in patients with diabetes and multivessel coronary artery disease provides additional information. Only approximately 30 % of the total group of (1,900 patients) had a history of a recent ACS while the remainder had stable CAD. The composite endpoint was all cause mortality, myocardial infarction and stroke. At 5 years, patients randomized to CABG had an improved composite endpoint (18.7 % versus 26.6 %: p < 0.005) (Table 23.5). There was some heterogeneity in the individual components of the composite endpoint: both all-cause mortality and myocardial infarction were reduced with CABG; in contrast there was an increase in stroke with CABG (5.2 % versus 2.4 %: p + .03). The benefit of CABG versus PCI was consistent across all prespecified sub groups (Fig. 23.5).