Multivessel CAD is not a homogenous anatomic or prognostic entity and includes individuals at both high and low risk for cardiovascular death, MI, or stroke (1). Consequently, merely establishing the presence of multivessel disease is not sufficient to define proper management. Rather, the efficient estimation of the patient’s short- and long-term risk of adverse cardiac events (especially death and MI) is the principal foundation on which treatment decisions are made. It is conceptually helpful to perceive the patient’s prognosis as the sum of the risks attributable to the patient’s current disease state and the risk that the patient’s disease will progress to a higher or lower risk state.

The risks associated with the patient’s current condition can be estimated from four major types of prognostic measures (Table 2.1). The strongest individual prognostic indicator in coronary disease is the extent of left ventricular dysfunction. In most instances, the ejection fraction is the variable used to estimate left ventricular function (2). However, because it is a ratio (stroke volume divided by left ventricular end-diastolic volume), the compensatory responses to left ventricular damage that serve to maintain cardiac output (e.g., Frank-Starling mechanism) may cause the ejection fraction to overestimate left ventricular contractility (3). More recent studies therefore have focused directly on ventricular volumes as indices of myocardial systolic dysfunction (4). Even the observation at left ventricular angiography of a dilated left ventricle (an informal ventricular volume assessment) indicates a higher risk state for any given ejection fraction value compared with a nondilated ventricle. Cardiomegaly on the plain chest
radiograph is a similar measure that has been shown repeatedly to have independent prognostic value (5, 6, 7). Another predictor is the presence and severity of congestive heart failure symptoms (6,8). For any given ejection fraction value, symptomatic heart failure indicates a patient at substantially higher risk than a similar patient without congestive heart failure symptoms (9).

Ischemic mitral regurgitation is recognized as an important and often underdiagnosed problem in multivessel coronary disease patients (6,10,11). Overall, approximately 20% of CAD patients presenting for diagnostic cardiac catheterization have some degree of mitral regurgitation, and 3% have severe regurgitation. Pathophysiologically, there are three major forms of this disorder, each with somewhat different prognostic implications. The most common form is papillary muscle dysfunction, which is typically due to posterior wall infarction caused by the occlusion of the circumflex or right coronary arteries. Such infarcts result in posteromedial papillary muscle dysfunction and restriction of the posterior mitral valve leaflet. Mitral regurgitation resulting from papillary muscle dysfunction may be associated with a favorable long-term prognosis if it is caused by a culprit lesion in the arterial supply to the papillary muscle that can be revascularized, and if the overall ventricular function is preserved. The second type of mitral regurgitation seen in ischemic heart disease is that resulting from global left ventricular dilation with secondary disruption of the function of the mitral valve apparatus. Dilation of the left ventricle caused by ischemic damage will move the papillary muscles out of proper alignment, with a resulting incomplete systolic coaptation of the mitral leaflets and varying degrees of regurgitation. In addition, long-standing left ventricular dilation may result in secondary dilation of the mitral annulus, also disrupting proper valvular function. This form of mitral regurgitation is associated with a poor prognosis, largely because of the severity of underlying left ventricular dysfunction. The final and least common type of ischemic mitral regurgitation is papillary muscle rupture, which typically occurs as a consequence of acute MI and is observed in less than 1% of such patients. Patients may present with hemodynamic deterioration with acute pulmonary edema unless the disorder is promptly recognized and aggressively treated. In studies from the Duke Database for Cardiovascular Diseases, the presence of at least 1 + mitral regurgitation is a significant adverse prognostic factor, and severe regurgitation is a major independent determinant of impaired survival in CAD patients (11, 12, 13). Because mitral regurgitation provides afterload reduction to the left ventricle, the combination of severe ventricular dysfunction and moderate or severe mitral regurgitation likely indicates that the true systolic performance of the ventricle is worse than the measured ejection fraction implies.

Aside from left ventricular function, the extent and severity of coronary atherosclerosis remain the most important determinants of prognosis. Traditionally, the extent of disease is measured as “the number of diseased vessels.” In this system, the coronary tree is divided into three distributions: the left anterior descending (including diagonal branches), the left circumflex (including marginal branches), and the right coronary artery. A coronary obstruction estimated angiographically ≥70% stenosed is considered hemodynamically significant. Although this classification is very widely used, it remains inadequate as the singular prognostic factor for clinical decision making, because it does not account for other factors that include clinical presentation (e.g., chronic stable angina versus unstable acute coronary syndrome), anatomic lesion location (e.g., proximal left anterior descending versus distal right coronary), or assessment of left ventricular function.

Over the last two decades, many alternative diseased vessels classification systems have been proposed, yet none has been adopted routinely in clinical practice. The one system that has achieved some use in research studies is a coronary artery jeopardy score that has been independently validated in subsequent studies (14). The score divides a stereotypic coronary tree into six major segments and assigns two points to each segment with a ≥70% stenosis. The score values thus range from 0 (no significant CAD) to 12 (significant left main and right CAD) and may
therefore provide greater prognostic accuracy than a more simplified number of diseased vessels classification. However, the score has several important limitations. First, significant lesions are recognized as prognostically equivalent without consideration for the extent of viable myocardium at risk or the possible varying risk associated with different degrees of “significant” coronary stenosis. Further, the score neither accounts for the presence of serial lesions or collateral vessels, nor the morphologic and pathophysiologic characteristics of the atherosclerotic plaques, such as the presence of thrombus.

Novel approaches in this area have utilized computerized coronary tree programs that enable the processing of much more detailed information than could be expected in routine clinical practice. One alternative method is the Duke CAD Severity Index (Table 2.2), which includes detailed information from the coronary angiogram, but does not require a quantitative analysis of the coronary tree or computer processing to create the score (1,15,16). Specifically, the index is a hierarchical model that assigns each patient to the worst applicable category based on information regarding lesion severity and location. Prognostic weights have been assigned using Cox regression analyses and a linear transformation so that the score ranges from 0 (no CAD) to 100 (≥95% left main disease). This new index can identify important anatomic subsets of patients with multivessel disease who derive particular benefit from percutaneous coronary intervention (PCI) or from coronary artery bypass graft (CABG) that were not evident using the overall number of diseased vessels classification (17).

Work in quantitative coronary angiography has challenged the primacy of the long-accepted visually determined, “significant” coronary stenosis (18). Although the percent diameter stenosis assessed by quantitative coronary angiography appears more accurate and consistent than the visual determination, it is as yet unclear whether this measure provides incremental benefit in clinical decision making or prognostic risk stratification. Investigators in the Angioplasty Compared to Medicine (ACME) study found that visual stenosis measurement actually correlated better with exercise capacity on the treadmill than did stenosis measurements with quantitative angiography or hand-held calipers (19).

The occurrence of transient ischemia provides another important marker relative to the severity of CAD. Although many investigators describe asymptomatic episodes of ischemia as “silent,” the use of this term emphasizes an artificial dichotomy among ischemic episodes that is probably no longer relevant. The original reason for making such a distinction was based on Cohn’s hypothesis that silent ischemia reflected a “defective anginal warning system,” placing patients who manifested this phenomenon at a particularly increased risk of adverse prognostic events relative to their symptomatic counterparts (20). For the most part, the defective anginal warning system theory has not been borne out by the evidence. It is now well established that many CAD patients have a majority of their ischemic events without symptoms (21, 22, 23). Increasing evidence also suggests that ischemia occurs on a continuum, and that the frequency and extent of transient ischemic episodes (both symptomatic and silent) correlate strongly with the severity of underlying coronary disease. What remains unsettled is the extent to which transient ischemia provides independent prognostic information regarding the patient’s disease beyond that available from an examination of the coronary arteriogram. It is possible, for example, that transient ischemia during exercise testing or ambulatory monitoring helps to differentiate otherwise similar-looking coronary lesions with differing “functional” importance (21).

In addition to the extent of CAD and left ventricular function, another important predictor of early and late outcomes is clinical presentation. In most patients presenting with acute coronary syndromes, a common denominator is the occurrence of coronary atherosclerotic plaque rupture with overlying thrombus (24, 25, 26). Coronary plaque rupture appears to occur most commonly in high-risk or vulnerable plaques, which are those characterized by a thin, fibrous cap overlying a lipid-rich core of cholesterol esters and intense inflammatory reaction (26,27). In particular, the presence of inflammatory cells and metalloproteinases may facilitate atherosclerotic plaque rupture (28). The clinical manifestations of plaque rupture vary considerably, ranging from asymptomatic status to acute ST-elevation MI and sudden cardiac death. It is noteworthy that whereas most of the focus in the treatment of coronary disease has been on plaques judged to be “significant” by
coronary angiography (i.e., at least 70% stenosis), those plaques that do not appear angiographically significant are now believed to be associated with the greatest risk for rupture and associated thrombosis. The “insignificant” plaques that are noted with varying frequency on coronary angiography and that are not suitable for treatment with either PCI or CABG are now believed to be a significant source of thrombotic events for many patients (24,25).

Clinically, the principal marker of an unstable coronary plaque is a change in the patient’s symptom pattern, typically manifesting as a sudden increase in the frequency, severity, or ease with which ischemic attacks are provoked. In a detailed evaluation of the prognostic information available from the patient’s history, the presence of increasingly progressive symptoms over the preceding 6 weeks and a greater frequency of symptoms were both identified as strong predictors of outcome, even when information regarding left ventricular ejection fraction and coronary disease severity from cardiac catheterization was considered (29).

The fourth major domain of CAD risk relates to the electrical stability of the myocardium. Whether atrial and ventricular arrhythmias, interventricular conduction delays, and ventricular after-potentials are merely markers of a significantly damaged myocardium or instead are independent predictors of risk for patients with coronary disease remains uncertain (30). Atrial fibrillation among individuals with CAD has been reported to be as low as 0.6% in the Coronary Artery Surgery Study (CASS) registry (31) and has high as 31.2% in the Goteborg study (32). Investigators have reported that atrial fibrillation in coronary disease correlates particularly with the presence of ischemic mitral regurgitation, heart failure, and stroke (31, 32, 33). Even after accounting for these factors in the Framingham study, the presence of atrial fibrillation independently increased the risk of death in men (odds ratio [OR] 1.5; 95% confidence interval [CI], 1.2 to 1.8) and women (OR 1.9; 95% CI, 1.5 to 2.2), compared with sinus rhythm (34). Similar observations have been described regarding interventricular conduction disturbances, particularly left bundle branch block or incomplete conduction defects (35).

Several recent studies have evaluated the relationship between various forms of ventricular arrhythmias and prognosis in coronary disease. In general, the findings from these studies conclude that malignant ventricular arrhythmias (e.g., sustained ventricular tachycardia, ventricular fibrillation) are significant adverse prognostic markers except when they occur in the earliest phase (e.g., first 48 hours) of acute MI. The significance of lesser degrees of ventricular arrhythmias, such as frequent premature ventricular contractions or nonsustained ventricular tachycardia, remains more controversial. Whether ventricular arrhythmias are markers for myocardial electrical instability or instead are a consequence of left ventricular dysfunction and scarring and do not convey independent prognostic information remains uncertain. This debate has been complicated by the findings of the Cardiac Arrhythmia Suppression Trial (CAST), which showed that antiarrhythmic drugs that were quite effective in suppressing ventricular arrhythmias actually increased mortality in a cohort of post-MI patients (36). In contrast, β-blockers and coronary bypass surgery, two therapies whose primary impact is on ischemia rather than on arrhythmias, may decrease the risk of sudden cardiac death (37,38).

The measurement of late potentials on the signal-averaged ECG has been used to identify patients at risk for sudden cardiac death. Late potentials are believed to indicate the electrophysiologic substrate for reentrant ventricular tachycardia, and numerous studies have reported that late potentials are powerful adverse prognostic findings that are independent of the results of ambulatory monitoring for ventricular arrhythmias and left ventricular ejection fraction (39). However, the clinical utility of this measure continues to be debated, and studies examining heart rate variability have proposed another marker for high risk that is of uncertain pathophysiologic or therapeutic significance (40, 41, 42). Heart rate variability is presumed to reflect the net effects of the parasympathetic and sympathetic nervous systems, both of which have been shown to be important in affecting the threshold for ventricular fibrillation. However, neither late potentials nor heart rate variability has yet been accepted as part of the standard risk assessment for CAD.

Many available therapies have been proposed for the treatment of CAD. However, only few agents have demonstrated improved survival in pivotal randomized trials.