Identification of the coronary endoluminal lesion(s) responsible for the process now called acute coronary syndrome (ACS) has become a central focus of both non-invasive and invasive treatment modalities in patients with coronary heart disease. Impressive reductions in cardiovascular mortality seen with lipid lowering agents1–3 and angiotensin-converting enzyme inhibitors4,5 suggest a critical role for pharmacologic alteration of plaque and vessel wall dynamics. Increasingly aggressive invasive approaches to the management of patients with coronary heart disease have suggested that catheter-based therapies directed toward a “culprit plaque” may diminish the risk of subsequent ACS.6 The former approach utilizes systemic therapy to treat what is both a local process and a systemic disease while the latter approach posits that a directed, local approach is also efficacious. How did such a dichotomous approach to the prevention of myocardial infarction arise? What are the assumptions surrounding each approach? What are the data? We will attempt to answer these questions and pose additional ones that will hopefully focus our thinking and highlight those areas in need of more information.

Atherosclerosis, primarily a disorder of the vascular intima, is characterized by both atherosis and sclerosis. Post mortem studies indicate that the former is typically comprised of soft, pultaceous material while the latter is typically fibrotic.7–10 Importantly, most plaques contain varying amounts of both “hard” and “soft” components with a minority characterized as predominantly one or the other.11–13 The progression of atherosclerosis has been classified into five phases (Table 1),14 which represent the dynamic and progressive nature of atherosclerotic disease.

A classification scheme of vascular injury has recently been suggested and describes three distinct stages of injury.15 Notably, considerable overlap exists between this classification scheme and that outlined in Table 1. The first stage of vascular injury is characterized by the functional alteration of endothelial cells without detectable patho-morphologic changes. This stage is recognized by the accumulation of monocytes and lipids in the extracellular space and the subsequent appearance of lipid-laden macrophages (foam cells). The second stage is characterized by endothelial denudation and intimal damage with, however, an intact internal elastic lamina. This stage is also characterized by migration and proliferation of smooth muscle cells and platelet adhesion. The third stage is characterized by a deeper injury to the media with associated thrombus formation. Small thrombi may organize and further contribute to the growth of the atherosclerotic plaque,16–20 while large thrombi may result in an ACS (myocardial infarction, unstable angina, sudden cardiac death).16,19–21
This important classification scheme of the severity of the atherosclerotic lesion and the attendant changes in vessel architecture does not, however, provide insight into the molecular biological mechanisms underlying the observed changes. The recognition of the importance of acute and chronic inflammation in both the formation and destruction of the culprit plaque(s) in ACS provides an important framework for much of the intensive research in this area today.22 It was not appreciated at the outset that the vascular changes described in Table 1 were the result of the release of growth factors, chemokines and cytokines released by endothelial cells, macrophages, leukocytes and platelets. Nor was the critical balance of smooth muscle cell function and death in determining the stability of the fibrous cap surrounding the lipid core appreciated. However, it quickly became clear that the explosive nature of ACS stood in marked contrast to the steady, inexorable course of stable coronary heart disease.

The risk of developing an ACS is, of course, directly related to the extent of disease.23–25 Numerous post mortem studies support the extensive nature of atherosclerotic changes in the coronary circulation in patients dying from ACS.24,25 However, coronary atherosclerosis is necessary but not sufficient in patients with fatal ACS. A striking increase in the prevalence of ruptured or fissured plaques, irrespective of the presence of (fresh) thrombus, is a hallmark of fatal ACS.26,27 These findings have been confirmed in numerous studies of survivors of ACS who underwent coronary angiography28,29 and are supported by the striking correlation between angiographically demonstrable lesion complexity and that noted on histopathologic examination.30 Nevertheless, the fundamental theme of the patho-anatomy of ACS has been the role of thrombus.31 As noted earlier, thrombus is essentially the final common pathway for ACS. Plaque disruption/rupture with consequent exposure of the highly prothrombotic lipid core to the bloodstream generally results in complete thrombotic occlusion (Q-wave myocardial infarction) or high-grade, intermittent obstruction (non-Q wave myocardial infarction, unstable angina, sudden cardiac death).

However, the wealth of information on the diagnosis, natural history and prognosis in coronary heart disease provided by coronary angiography has also led to confusing (and erroneous) interpretations of the relationship between the severity of coronary obstruction and the likelihood of subsequent myocardial infarction. Prospective angiographic studies have shown that the frequency of progression to coronary occlusion is significantly and positively related to the severity of the stenosis at baseline.32,33 However, retrospective angiographic studies, in selected populations, have concluded that stenoses considered non-critical (< 50% diameter reduction) were more frequently associated with subsequent myocardial infarction than their critical counterparts. Table 2 summarizes the data from 4 published studies.34–37 Note that this database has relatively few patients and that there is astonishing variation in the time interval between defining angiograms. Both of these features alone raise the concern of bias — both in ascertainment and follow-up. How, then, can one conclude that while high-grade stenoses progress more frequently (to occlusion), milder stenoses are responsible for the majority of events now termed ACS? Figure 1 illustrates the key distinction between a lesion-based analysis and a patient-based analysis. The left-sided y-axis shows the percent of lesions occluded at the 5-year follow-up angiogram in the CASS study,32 while the right-sided y-axis shows the number of coronary segments analyzed. The x-axis categorizes the baseline stenoses as none, mild (5–49%), moderate (50–80%) and severe (81–95%).

It is immediately apparent that the number of mild and moderate stenoses far outnumber the number of severe stenoses. Nevertheless, severe lesions are more likely to progress to occlusion over time. However, when the outcome variable (myocardial infarction) is expressed on a per patient basis (Figure 2), then mild to moderate stenoses are more frequently found on the initial angiogram. Unfortunately, neither approach (lesion-based or patient-based) to the analysis of the relationship between the severity of a plaque (at one point in time) and the likelihood that it may rupture (at another point in time) can answer the following question: What is the likelihood that any given plaque will rupture and result in an ACS? The Poisson distribution is an appropriate statistical model for the occurrence of discrete, infrequent events. Accepting that there are many more mild and moderately obstructive plaques than severely obstructive plaques in the coronary circulation of an adult, the Poisson distribution can provide a reasonable likelihood of a given plaque rupture resulting in an ACS. If we posit an average plaque rupture rate of 10 per a given time interval (e.g., Table 2) per individual, Table 3 indicates the probability of 0, 1, 2 or 10 events. If we posit an average plaque rupture rate of 20 per given time interval per individual (a very unlucky individual), it can be seen in Table 4 that the probability of 0, 1, 2 or 20 events is many orders of magnitude less.

It appears, then, that the risk of a plaque rupture is, from a statistical standpoint, a rare event. However, not all plaques are created equal and, from a stochastic perspective, the risk of plaque rupture is the multiplicative product of anatomic, hemodynamic, “exogenous” and “endogenous” covariates. In fact, a reliable “risk stratification” approach to the identification of the culprit plaque prior to its clinical manifestation as an ACS is urgently needed. Current interventional modalities, of necessity, treat only the sequelae of plaque rupture and not its risk. In contrast, contemporary pharmacologic approaches have achieved impressive reductions in the risk of initial and subsequent myocardial infarctions. How, then, can we improve our ability to predict which of the thousands of plaques in the coronary circulation is likely to progress to the final stage of Table 1? A re-examination of old concepts and assumptions and a new look at the risk factors for plaque rupture is a good start.

We have seen how the concept that angiographically insignificant plaque is more likely to result in acute thrombotic occlusion was developed. While distinctions between individual risk and pooled risk were not sufficiently emphasized in the above-cited literature, a more fundamental set of observations will aid in our goal of improved prediction algorithms. The argument that arteries are dynamic structures that actively respond to adverse hemodynamic conditions was convincingly supported by the studies of Glagov et al.38 These investigators advanced the concept of remodeling by demonstrating changes in total arterial size in direct proportion to the endoluminal plaque burden. It is of more than passing interest that this compensatory behavior — that is, the preservation of lumen area — becomes exhausted as the relative plaque area approaches 40–50% of the endoluminal surface. At this point, further accretion in plaque mass compromises lumen dimension. This fundamental observation serves to clarify the confusing situation with regard to the culprit plaque-ACS question. Clearly, the angiographic assessment of disease severity is compromised by the inability to quantify the extent of remodeling and the underlying plaque burden. It is precisely in those situations where lumen dimension is relatively preserved or compromised by < 50% that plaque volume is substantial. As discussed below, plaque volume is a critical covariate for the risk of plaque rupture. Further in vivo substantiation of both the concept of remodeling and its relationship to plaque mass can be seen using intravascular ultrasound.39 As seen in Figures 3–5, a relatively mild stenosis at angiography in a patient with ACS corresponds, on IVUS, to severe, complex disease with significant plaque burden. A number of studies have consistently shown that the plaque mass/volume associated with culprit lesions in patients with ACS is consistently greater than the plaque mass in lesions in patients with more clinically stable presentations.40–42 Thus, although it is now known that the culprit plaque is frequently more severe in character than suggested by angiography, the requirement that we can predict which plaque will rupture (not has ruptured) led to the concept of the vulnerable, culprit plaque. In other words, which of the many plaques in the coronary circulation are likely to rupture and under what circumstances are they likely to rupture? While traditional thinking has held that plaque ruptures (and myocardial infarctions) are seemingly random events, the identification of structural, hemodynamic and identifiable exogenous factors that act alone or in concert will contribute significantly to the ability to model plaque rupture. Table 5 outlines our current understanding of the “variables” associated with plaque rupture.

Until recently, the natural history of coronary heart disease was viewed as a linear process interspersed with random, albeit catastrophic, events. Information derived from coronary angiographic studies initially supported this clinical concept by demonstrating a finite rate of disease progression. More recent data, both angiographic and clinical, now suggest that the progression of coronary atherosclerosis is erratic, often unpredictable and multi-focal in nature.60 Histopathologic studies have taught us that the progression of disease varies widely within and between individuals, that repetitive, sub-clinical episodes of plaque rupture are not infrequent and that the composition of a given plaque has a closer relationship to the likelihood of plaque rupture than the severity of the endoluminal obstruction. Cardiologists in general, and interventional cardiologists in particular, have become expert in the recognition and management of acute coronary syndromes. The larger challenge remains, as in cardiovascular disease in general, the ability to predict and modify the risk of plaque rupture progressing to ACS. Insights from prospectively designed studies of primary and/or secondary prevention in which detailed clinical, biochemical and endoluminal anatomy is obtained will enable us to do so. In their absence, the ability to predict plaque rupture will rest on unfounded assumptions, incomplete data and preconceived bias.