Normal Artery Wall

The arterial lumen is lined by a monolayer of endothelial cells that overlies smooth muscle cells (Figure 6-2). The inner layer of smooth muscle cells, known as the intima, is circumscribed by the internal elastic lamina. Between the internal elastic lamina and external elastic lamina is another layer of smooth muscle cells, the media. Outside the external elastic lamina is an adventitia that is sparsely populated by cells and microvessels of the vasa vasorum.

Figure 6-2 Normal human coronary artery of a 32-year-old woman.

The intima (i) and media (m) are composed of smooth muscle cells. The adventitia (a) consists of a loose collection of adipocytes, fibroblasts, vasa vasorum, and nerves. The media is separated from the intima by the internal elastic lamina (open arrow) and the adventitia by the external elastic lamina (closed arrow). (Movat’s pentachrome-stained slide, original magnification, ×6.6.)

Intima

Traditionally, the intima has been considered the most important layer of the artery wall.⁷ The intima can vary from a single endothelial layer to a more complex structure of an endothelium overlying a patchwork of extracellular matrix and vascular smooth muscle cells. As part of the normal development of many large arteries, smooth muscle cells populate this space and form a neointima. This diffuse form of intimal thickening consists of layers of smooth muscle cells and connective tissue, the thickness of which may vary considerably. For convenience, the intima/media ratio is often measured, and the reference range is 0.1 to 1.0. How this benign intima forms is not well understood. Presumably, the intima represents a physiologic adaptation to changes in arterial flow and wall tension. The intima is made up of two distinct layers.⁸ As seen by electron microscopy, the inner layer subjacent to the luminal endothelium contains an abundance of proteoglycan ground substance. Smooth muscle cells found in this layer are usually distributed as isolated cells in a sea of matrix, rather than in contiguous layers. A few macrophages also may be found in this layer underneath the endothelial monolayer. The outer, musculoelastic layer of the intima is adjacent to the internal elastic lamina and contains smooth muscle cells and elastic fibers.

Media

In normal adult arteries, several smooth muscle cell subpopulations with distinct lineages exist within the media.⁹ These diverse cell populations likely fulfill different functions to maintain homeostasis in the artery wall. For example, in response to pressure elevations, increases in smooth muscle cell mass and extracellular matrix may be required. Alternatively, for arteries to be able to stretch both longitudinally and circumferentially, smooth muscle cells with variable orientations of cytoskeletal fibers must be present. These distinct cell types may be important not only in health but in disease. In certain experimental models of neointimal formation, proliferation and inward migration of subpopulations of medial smooth muscle cells occur.¹⁰ The biologic determinants of medial smooth muscle cell diversity are unknown.¹¹

Adventitia

The adventitia, the outermost layer of the artery wall, normally consists of a sparse collection of fibroblasts, microvessels (vasa vasorum), nerves, and few inflammatory cells. The majority of the vasa vasorum that nourish the inner layers of the artery wall originate in the adventitia. Traditionally, the adventitia has been ignored and is not thought to play a role in vascular lesion formation. However, more recent studies have elucidated the role of the adventitia as not only a source of inflammatory cells in the development of atherosclerosis, but a hub for paracrine signaling that can maintain vascular homeostasis in a variety of vascular diseases. ¹²

Transmembrane and Transcellular Communication

Blood vessels respond to a multitude of neural, humoral, and mechanical stimuli in fulfilling their role in homeostasis. When norepinephrine, released from adrenergic nerve terminals in the adventitia, binds to receptors on the vascular smooth muscle cell membrane, a series of events takes place, culminating in a change in vessel diameter. Much progress has been made in understanding this transmembrane signaling since the discovery of cyclic adenosine monophosphate (cAMP) in the late 1950s. Hormones circulating in the blood must interact with receptors on endothelial cells before the message reaches the vascular smooth muscle cell. The mechanism of communication between cells has been one of the central themes of biologic research in the past decades. Future understanding of cardiovascular disease will likely be based on identification of abnormalities of the molecules involved in transmembrane and transcellular communication. A brief introduction to these topics is provided here.

Figure 6-3 illustrates examples of pathways of transmembrane signaling. Up to five components can be involved: receptor, G protein, effector producing a second messenger, phosphorylation of regulator protein, and the consequent change in cell behavior. G proteins (guanine nucleotide-binding regulatory proteins) are made up of three subunits (α, β, γ) and float in the cell membrane. On contact with a ligand-receptor complex, guanosine diphosphate (GDP) on the α subunit is replaced by guanosine triphosphate (GTP). The activated α subunit then dissociates from the beta-gamma complex and can interact with several membrane targets (see Figure 6-3B). For example, β-receptor activation results in the activation of G_s (s = stimulate), which will stimulate the synthesis of cAMP by adenylyl cyclase. Muscarinic receptor activation activates a G_i (i = inhibit) protein that inhibits adenylyl cyclase. A single G protein can interact with more than one effector. In this way, the G protein can be a branch point for the regulation of multiple effectors in response to a single signal. These proteins already have been implicated in human disease; cholera toxin covalently modifies G_s so that it becomes persistently active in stimulating adenylyl cyclase in intestinal epithelial cells, likely causing the severe diarrhea of cholera.

Figure 6-3 Steps in the process whereby hormone-receptor binding results in a change in cell behavior.

In this example, the final result is the opening of an ion channel. A, A hormone or ligand (L) binds to a receptor (R) embedded in the cell membrane. The receptor-ligand complex interacts with G protein (G) floating in the membrane, resulting in activation of the α subunit (Gα). The activated α subunit can then follow different pathways (B). Effector enzymes in the membrane (E), such as adenylyl cyclase, cyclic guanosine monophosphate (cGMP), phospholipase C, or phospholipase A₂, change the cytoplasmic concentration of their “messengers”: cyclic adenosine monophosphate (cAMP), cGMP, diacylglycerol (DAG), and inositol 1,4,5-triphosphate (IP₃). These soluble molecules activate protein kinase A or C (PKA or PKC), or release Ca⁺⁺ from sarcoplasmic reticulum (SR). Subsequently, cell behavior is changed by phosphorylation of an ionic channel on the cell membrane (CHAN) or by release of Ca⁺⁺ from SR. B, Several pathways coupling receptor activation to final effect are illustrated. It is likely that multiple pathways are activated concomitantly, both facilitatory and inhibitory. In this way, the final response can be determined by the sum of the effects of several stimuli.

(A, B, From Brown AM, Birnbaumer L: Ionic channels and their regulation by G-protein subunits. Annu Rev Physiol 52:197, 1990.)

Several second-messenger systems have been characterized. G_s can directly enhance conductance through calcium channels, with the increased intracellular calcium acting as second messenger. The cyclic nucleotides, cAMP and guanosine monophosphate (GMP), act as second messengers. Their intracellular action is terminated when they are cleaved by phosphodiesterase enzymes, which, in turn, are regulated by stimuli and second messengers. The breakdown products of membrane phosphoinositide constitute another, more recently recognized set of second messengers.¹³ In response to agonists such as vasopressin, G protein is activated, leading to activation of the membrane-associated enzyme phospholipase C. This enzyme cleaves phosphatidylinositol 4,5-biphosphate on the inner leaflet of the plasma membrane, producing inositol 1,4,5-triphosphate (IP₃) and diacylglycerol (DAG). Both are second messengers. IP₃ diffuses through the cytoplasm and mobilizes calcium from intracellular stores. DAG remains within the plasma membrane and activates protein kinase C, which modulates cellular activity by phosphorylating intracellular proteins. In many cell types, activation of the same receptors that control phosphoinositide breakdown also results in the liberation of arachidonate and/or eicosanoids (prostaglandins, leukotrienes, and thromboxanes). The resultant change in cell behavior can be the opening of an ion channel, contraction or relaxation of smooth muscle, secretory activity, or initiation of cell division (see Chapter 7).

Endothelium

Although the vascular endothelium was once thought of as an inert lining for blood vessels, it is more accurately characterized as a very active, distributed organ with many biologic functions. It has synthetic (Table 6-1) and metabolic (Table 6-2) capabilities, and contains receptors for a variety of vasoactive substances (Table 6-3). Functions of the endothelium that may play an important role in the pathophysiology of ischemic heart disease are discussed.

TABLE 6-1 Substances Produced by Vascular Endothelium

From Bassenge E, Busse R: Endothelial modulation of coronary tone. Prog Cardiovasc Dis 30:349, 1988.

TABLE 6-2 Vasoactive Substances Processed by Vascular Endothelium

ACE, angiotensin-converting enzyme.

From Bassenge E, Busse R: Endothelial modulation of coronary tone. Prog Cardiovasc Dis 30:349, 1988.

TABLE 6-3 Stimulators of Endothelium-Mediated Vasodilation

Modified from Bassenge E, Busse R: Endothelial modulation of coronary tone. Prog Cardiovasc Dis 30:349, 1988.

Endothelium-Derived Relaxing Factors

The first vasoactive endothelial substance to be discovered was prostacyclin (PGI₂), a product of the cyclooxygenase pathway of arachidonic acid metabolism (Figure 6-4; Box 6-1).¹⁴ The production of PGI₂ is activated by shear stress, pulsatility of flow, hypoxia, and a variety of vasoactive mediators. On production it leaves the endothelial cell and acts in the local environment to cause relaxation of the underlying smooth muscle or to inhibit platelet aggregation. Both actions are mediated by the stimulation of adenylyl cyclase in the target cell to produce cAMP.

Figure 6-4 The production of endothelium-derived vasodilator substances.

Prostacyclin (PGI₂) is produced via the cyclooxygenase pathway of arachidonic acid (AA) metabolism, which can be blocked by indomethacin (Indo) and aspirin. PGI₂ stimulates smooth muscle adenylyl cyclase and increases cyclic adenosine monophosphate (cAMP) production, which cause relaxation. Endothelium-derived relaxing factor (EDRF), now known to be nitric oxide (NO), is produced by the action of NO synthase on l-arginine in the presence of reduced nicotinamide adenine dinucleotide phosphate (NADPH), oxygen (O₂), and calcium and calmodulin. This process can be blocked by arginine analogs like N^G-monomethyl-l-arginine (l-NMMA). NO combines with guanylate cyclase in the smooth muscle cell to stimulate production of cyclic guanosine monophosphate (cGMP), which results in relaxation. Less well characterized is an endothelium-derived factor, which hyperpolarizes the smooth muscle membrane (EDHF) and probably acts via activation of potassium (K⁺) channels. 5-HT, serotonin; ACh, acetylcholine; ADP, adenosine diphosphate; M, muscarinic receptor; P, purinergic receptor; T, thrombin receptor.

(From Rubanyi GM: Endothelium, platelets, and coronary vasospasm. Coron Artery Dis 1:645, 1990.)

BOX 6-1 Endothelium-Derived Relaxing and Contracting Factors

Healthy endothelial cells have an important role in modulating coronary tone by producing the following factors:

Vascular Muscle Relaxing Factors

• Prostacyclin

• Nitric oxide

• Hyperpolarizing factor

Vascular Muscle Contracting Factors

• Prostaglandin H₂

• Thromboxane A₂

• Endothelin

In 1980, Furchgott and Zawadzki¹⁵ observed that the presence of an intact endothelium was necessary for acetylcholine-induced vasodilation. Since that time it has been shown that many physiologic stimuli cause vasodilation by stimulating the release of a labile, diffusible, nonprostanoid molecule termed endothelium-derived relaxing factor (EDRF; see Figure 6-4), now known to be nitric oxide (NO). NO is the basis of a widespread paracrine signal transduction mechanism whereby one cell type can modulate the behavior of adjacent cells of different type.^16,17 NO is a very small lipophilic molecule that can readily diffuse across biologic membranes and into the cytosol of nearby cells. The half-life of the molecule is less than 5 seconds, so that only the local environment can be affected. NO is synthesized from the amino acid l-arginine by nitric oxide synthase (NOS). In vascular endothelium, the enzyme (endothelial NOS or NOS3) is always present (constitutive) and resides in the cytoplasm. Its function depends on the presence of Ca⁺⁺ and calmodulin, as well as tetrahydrobiopterin. Serine phosphorylation is important for prolonged activity. The enzyme is activated in response to receptor occupancy or physical stimulation (see Table 6-3). When NO diffuses into the cytosol of the target cell, it binds with the heme group of soluble guanylate cyclase, resulting in a 50- to 200-fold increase in production of cyclic GMP, its second messenger. If the target cells are vascular smooth muscle cells, vasodilation occurs; if the target cells are platelets, adhesion and aggregation are inhibited. In vascular smooth muscle, cyclic GMP leads to activation of protein kinase G, which phosphorylates various intracellular target proteins, including the myosin light-chain regulatory subunit and proteins that control intracellular calcium.¹⁸

It is likely that NO is the final common effector molecule of nitrovasodilators (including sodium nitroprusside and organic nitrates such as nitroglycerin). The cardiovascular system is in a constant state of active vasodilation that is dependent on the generation of NO. The molecule is more important in controlling vascular tone in veins and arteries compared with arterioles. When the microcirculation dilates in response to metabolic myocardial demand (e.g., exercise), increased flow through epicardial coronary arteries increases shear stress at the endothelium. This leads to release of NO, which causes vascular smooth muscle relaxation and dilation of the conductance vessels, thereby facilitating the increase in flow. The importance of the loss of this mechanism in atherosclerosis is underlined by the fact that, in this situation, more than 50% of the resistance to flow in the coronary circulation resides in vessels larger than 100 μm in diameter (see Figure 6-1). Abnormalities in the ability of the endothelium to produce NO likely plays a role in diseases such as diabetes, atherosclerosis, and hypertension.^19,20 The venous circulation of humans appears to have a lower basal release of NO and an increased sensitivity to nitrovasodilators when compared with the arterial side of the circulation.²¹

Many agents, such as acetylcholine and norepinephrine, can cause contraction when applied directly to the vascular smooth muscle membrane instead of relaxation, which occurs when it is applied to the intact endothelium (Figure 6-5). The net effect of neural or humoral stimuli depends on a combination of direct effects mediated by binding to vascular smooth muscle receptors and indirect effects because of the ligand binding to endothelial receptors causing NO release from the endothelium. In the presence of healthy endothelium, vasodilation usually predominates. When the endothelium is absent (injured vessel) or diseased (atherosclerosis), vasoconstriction may be the net effect. NO has important roles in neurohumoral regulation of vascular tone, in preventing intravascular platelet aggregation, and in the structural adaptation of blood vessels to the demands of blood flow and pressure. Knowledge of its role in inflammation and atherosclerosis is rapidly expanding.

Figure 6-5 Role of endothelium in the control of coronary tone.

Intact endothelium has an important modulatory role in the effect of numerous factors on vascular smooth muscle. In the absence of a functional endothelium (mechanical trauma, atherosclerosis), many factors act directly on smooth muscle to cause constriction (left). Under normal conditions (right), the release of nitric oxide (NO; endothelium-derived relaxing factor [EDRF]) and prostacyclin (PGI₂) stimulated by these same factors can attenuate constriction or cause dilation. PGI₂ release is predominantly into the lumen, whereas EDRF release is similar on both the luminal and abluminal sides. Substances in parentheses elicit only vasodilation. 5-HT, serotonin; A, adenosine; ACh, acetylcholine; ADP, adenosine monophosphate; AII, angiotensin II; ATP, adenosine triphosphate; Bk, bradykinin; CGRP, calcitonin gene–related peptide; ET, endothelin; NA, norepinephrine; PAF, platelet-activating factor; SP, substance P; VIP, vasoactive intestinal polypeptide; VP, vasopressin.

(From Bassenge E, Heusch G: Endothelial and neurohumoral control of coronary blood flow in health and disease. Rev Physiol Biochem Pharmacol 116:77, 1990.)

In addition to PGI₂ and NO, another less-well-understood pathway for receptor-mediated or mechanically induced endothelium-derived vasodilation exists that is associated with smooth muscle hyperpolarization. Both epoxyeicosatrienoic acid (a metabolite of cytochrome P450) and H₂O₂ have been suggested as possible endothelium-derived hyperpolarizing factors (EDHFs).^22,23 Smooth muscle relaxation is a result of hyperpolarization of the myocyte, which leads to decreased intracellular calcium concentration. EDHF-mediated vasodilation can be blocked by inhibition of calcium-dependent potassium channels. EDHF may have an important vasodilator role in the human coronary microcirculation.²⁴

Endothelium-Derived Contracting Factors

Contracting factors produced by the endothelium include prostaglandin H₂, thromboxane A₂ (TxA₂; via cyclooxygenase), and the peptide endothelin (ET). ET is a potent vasoconstrictor peptide (100-fold more potent than norepinephrine)²⁵ with remarkable similarities to the toxin of the burrowing asp. Both have potent coronary constrictor activity to which the strong cardiac toxicity and lethality of the toxin are attributed.²⁶ Three closely related 21 amino acid peptides have been identified: endothelin-1 (ET-1), ET-2, and ET-3. The primary product of vascular endothelium is ET-1, which is synthesized from prepro-ET-1 within vascular endothelial cells by the action of ET-converting enzyme. It is not stored but rapidly synthesized in response to stimuli such as ischemia, hypoxia, and shear stress, and released predominantly abluminally (toward the underlying smooth muscle).²⁷ In vascular smooth muscle cells, ET-1 binds to specific membrane receptors (ET_A) and, via phospholipase C, induces an increase in intracellular calcium resulting in long-lasting contractions.²⁸ It is also linked via a G_i protein to voltage-operated calcium channels. This peptide has greater vasoconstricting potency than any other cardiovascular hormone, and in pharmacologic doses can abolish coronary flow, leading to ventricular fibrillation and death.²⁹ Another receptor subtype, ET_B, is expressed by both smooth muscle and endothelium and binds ET-1 and ET-3 equally well (Figure 6-6). When isolated vessels are perfused with ET-1, there is an initial NO-mediated vasodilation because of binding with ET_B receptors on the endothelial cells, followed by contraction because of binding of ET-1 to ET_A receptors on the vascular smooth muscle membrane. Studies utilizing bosentan, a combined ET_A– and ET_B-receptor antagonist, have demonstrated that ET exerts a basal coronary vasoconstrictor tone in humans.³⁰ There is evidence that ET may play a role in the pathophysiology of pulmonary and arterial hypertension, atherosclerosis, myocardial ischemic syndromes, and heart failure.³¹ Clinical trials of bosentan in patients with congestive heart failure³² and hypertension have shown promise, but hepatic side effects have limited the dose to less than 500 mg daily, with the primary indication being severe pulmonary hypertension.³³

Figure 6-6 Endothelin (ET) released abluminally interacts with ET_A and ET_B receptors on vascular smooth muscle to cause contraction. Activators of ET_B receptors on endothelial cells cause vasodilation. cAMP, cyclic adenosine monophosphate; cGMP, cyclic guanosine monophosphate; ECE, endothelin-converting enzyme; NO, nitric oxide; PGI₂, prostacyclin.

(From Luscher TF: Do we need endothelin antagonists? Cardiovasc Res 29:2089, 1997; reproduced by permission of Elsevier Science-NL, Sara Burgerhartstraat 25, 1055 KV Amsterdam, the Netherlands.)

Endothelial Inhibition of Platelets

A primary function of endothelium is to maintain the fluidity of blood. This is achieved by the synthesis and release of anticoagulant (e.g., thrombomodulin, protein C), fibrinolytic (e.g., tissue-type plasminogen activator), and platelet inhibitory (e.g., PGI₂, NO) substances (Box 6-2).³⁴ Mediators released from aggregating platelets stimulate the release of NO and PGI₂ from intact endothelium, which act together to increase blood flow and decrease platelet adhesion and aggregation (Figure 6-7), thereby flushing away microthrombi and maintaining the patency of the vessel.

BOX 6-2 Endothelial Inhibition Of Platelets

Healthy endothelial cells have a role in maintaining the fluidity of blood by producing:

• Anticoagulant factors: protein C and thrombomodulin

• Fibrinolytic factor: tissue-type plasminogen activator

• Platelet inhibitory substances: prostacyclin and nitric oxide

Figure 6-7 Inhibition of platelet adhesion and aggregation by intact endothelium.

Aggregating platelets release adenosine diphosphate (ADP) and serotonin (5-HT), which stimulate the synthesis and release of prostacyclin (PGI₂) and endothelium-derived relaxing factor (EDRF; nitric oxide [NO]), which diffuse back to the platelets and inhibit further adhesion and aggregation, and can cause disaggregation. PGI₂ and EDRF act synergistically by increasing platelet cyclic adenosine monophosphate (cAMP) and cyclic guanosine monophosphate (cGMP), respectively. By inhibiting platelets and also increasing blood flow by causing vasodilation, PGI₂ and EDRF can flush away microthrombi and prevent thrombosis of intact vessels. P_2y, purinergic receptor.

(From Rubanyi GM: Endothelium, platelets, and coronary vasospasm. Coron Artery Dis 1:645, 1990.)

With vital roles in modulating the tone of vascular smooth muscle, inhibiting platelets, and processing circulating chemicals, it seems clear that endothelial cell dysfunction would cause or contribute to ischemic syndromes. There is evidence of endothelial dysfunction in atherosclerosis, hyperlipidemia, diabetes, and hypertension.³⁵ Procedures such as coronary artery surgery and angioplasty disrupt the endothelium. The role of endothelium in the pathophysiology of myocardial ischemia is discussed later (see Dynamic Stenosis section).

Perfusion Pressure and Myocardial Compression

Coronary blood flow is proportional to the pressure gradient across the coronary circulation (Box 6-3). This gradient is calculated by subtracting downstream coronary pressure from the pressure in the root of the aorta. The determination of downstream pressure is complicated because the intramural coronary vessels are compressed with each heartbeat.

During systole, the heart throttles its own blood supply. The force of systolic myocardial compression is greatest in the subendocardial layers, where it approximates intraventricular pressure. Resistance caused by extravascular compression increases with blood pressure, heart rate, contractility, and preload. Because it is difficult to measure intramyocardial pressure, the relative importance of these factors is controversial.^36,37 Flow is impeded both by direct compression and by shear caused by twisting of vessels as the heart contracts. Myocardial extravascular compression is less in the right ventricle, where pressures are lower and coronary perfusion persists during systole (Figure 6-8). In pathologic conditions associated with pulmonary hypertension, right coronary flow assumes a phasic pattern similar to left coronary flow. Under normal conditions, extravascular compression contributes only a small component (10% to 25%) to total coronary vascular resistance. When the coronary vessels are dilated by pharmacologic agents such as dipyridamole or during ischemia, the effects of extravascular compression on myocardial perfusion become more important (see Transmural Blood Flow section later in this chapter).

Figure 6-8 Blood flow in the left and right coronary arteries.

The right ventricle is perfused throughout the cardiac cycle. Flow to the left ventricle is largely confined to diastole.

(From Berne RM, Levy MN: Special circulations. In Berne RM, Levy MN [eds]: Physiology. St. Louis: CV Mosby, 1988, pp 540–560.)

With each contraction, the intramural vessels are squeezed and blood is expelled forward into the coronary sinus and retrograde into the epicardial arteries. The large coronary arteries on the epicardial surface act as capacitors, charging with blood during systole and expelling blood into the coronary circulation during diastole.³⁸ Coronary capacitance likely explains the findings of Bellamy,³⁹ who reported that flow in the proximal left anterior descending coronary artery of the dog ceased when arterial pressure decreased to less than 45 mm Hg. It was suggested that flow throughout the coronary circulation stopped at pressures far in excess of the pressure at the coronary sinus. This pressure at which flow stopped was termed critical closing pressure or zero-flow pressure (P_zf). This had important implications in the calculation of coronary resistance because the effective downstream pressure would be P_zf and not the much lower coronary venous pressure. This is analogous to a stream with a waterfall, where flow rate over the waterfall depends on the drop from the source to the waterfall edge and is unaffected by the distance to the bottom of the falls. It was later suggested that flow through the intramural coronary vessels continues after coronary inflow near the ostia (measured by Bellamy) has ceased.^40,41 There is evidence that antegrade movement of red blood cells in 20-μm arterioles continues until coronary pressure is only a few millimeters of mercury greater than coronary sinus pressure.⁴² The concept of a critical closing pressure greatly in excess of coronary sinus pressure is probably not valid in the coronary circulation.

Although the true downstream pressure of the coronary circulation is likely close to the coronary sinus pressure, other choices may be more appropriate in clinical circumstances. In patients with CAD, the subendocardial layers of the left ventricle are at greatest risk for ischemia and necrosis (see Transmural Blood Flow section later in this chapter). Because this layer is perfused mostly when the aortic valve is closed, the most appropriate measure of the driving pressure for flow here is the average pressure in the aortic root during diastole. This can be approximated by aortic diastolic or mean pressure. Pressures monitored in peripheral arteries by routine methods in clinical settings can differ from central aortic readings. This is due to distortion of the pressure waveform as it is propagated through the arterial tree and inaccuracies associated with the hydraulic and electronic components of the monitoring system. Under these conditions, the mean arterial pressure may be the most reliable measure of coronary driving pressure. The true downstream pressure of the left ventricular subendocardium is the left ventricular diastolic pressure, which can be estimated by pulmonary artery occlusion pressure. When the right ventricle is at risk for ischemia (e.g., severe pulmonary hypertension), right ventricular diastolic pressure or central venous pressure may be more appropriate choices for downstream pressure.

Myocardial Metabolism

Myocardial blood flow, like flow in the brain and skeletal muscle, is primarily under metabolic control. Even when the heart is cut off from external control mechanisms (neural and humoral factors), its ability to match blood flow to its metabolic requirements is almost unaffected.³⁵ Because coronary venous oxygen tension is normally 15 to 20 mm Hg, there is only a small amount of oxygen available through increased extraction. A major increase in cardiac oxygen consumption (Mvo₂) can occur only if oxygen delivery is increased by augmentation of coronary blood flow. Normally, flow and metabolism are closely matched so that over a wide range of oxygen consumption, coronary sinus oxygen saturation changes little.⁴³

Despite intensive research over the last several decades, the mediator or mediators linking myocardial metabolism so effectively to myocardial blood flow are still unknown. Hypotheses of metabolic control propose that vascular tone is linked either to a substrate that is depleted, such as oxygen or adenosine triphosphate (ATP), or to the accumulation of a metabolite such as CO₂ or hydrogen ion (Box 6-4). Adenosine has been proposed in both categories. Feigl⁴³ has proposed six criteria for a chemical transmitter between the cardiac myocyte and the coronary vascular smooth muscle cell:

Many potential mediators of metabolic regulation have been proposed.⁴⁴ Although NO has a role in many coronary vasoregulatory pathways, it does not fulfill the role of metabolic regulator because blockade of NOS does not alter the increase in myocardial blood flow associated with an increase in myocardial oxygen demand.⁴⁵ The arguments for oxygen, carbon dioxide, and adenosine are briefly examined.

Neural and Humoral Control

Autoregulation

Autoregulation is the tendency for organ blood flow to remain constant despite changes in arterial perfusion pressure.¹⁰⁴ Autoregulation can maintain flow to myocardium served by stenotic coronary arteries despite low perfusion pressure distal to the obstruction. This is a local mechanism of control and can be observed in isolated, denervated hearts. If Mvo₂ is fixed, coronary blood flow will remain relatively constant between mean arterial pressures of 60 to 140 mm Hg. Figure 6-9 illustrates that, at a given cardiac workload, the level of flow (determined by metabolic regulation) is maintained constant over a broad range of pressure by autoregulation.

Figure 6-9 Autoregulation at two levels of myocardial oxygen consumption.

Pressure in the cannulated left circumflex artery was varied independently of aortic pressure. When pressures were suddenly increased or decreased from 40 mm Hg, flow instantaneously increased with pressure (steep line, green triangles). With time, flow decreases to the steady-state level determined by oxygen consumption (purple and red circles). The vertical distance from the steady-state (autoregulating) line to the instantaneous pressure-flow line is the autoregulatory flow reserve.

(From Mosher P, Ross J Jr, McFate PA, Shaw RF: Control of coronary blood flow by an autoregulatory mechanism. Circ Res 14:250, 1964.)

Coronary perfusion pressure must be varied while holding Mvo₂ constant to study autoregulation. This is difficult in the heart because changing aortic pressure changes both the perfusion pressure for the coronary arteries and the afterload of the left ventricle. Thus, changes in aortic pressure inevitably change Mvo₂. This problem is overcome by cannulating the coronary arteries and perfusing them with a pump. However, even when heart rate and aortic pressure are held constant, Mvo₂ changes with changing coronary pressure. This is because myocardial contractility and metabolism increase when coronary pressure is increased to more than the normal autoperfused level. This phenomenon is known as the Gregg effect and may be explained by the “garden hose” hypothesis of Lochner, whereby engorgement of the coronary vasculature elongates the myocardial sarcomere length during diastole and contractile strength is increased because of the Frank–Starling mechanism (for a detailed review, see Feigl⁴³ and Gregg¹⁰⁵).

In addition to the Gregg effect, two other issues complicate studies of autoregulation: collateral flow and myocardial oxygen extraction. If pressure is lowered in the left coronary artery and not in the right, there will be a pressure gradient for flow from the right to left coronary artery via collateral vessels. Flow measured proximally in the left coronary artery will then underestimate flow reaching the myocardium. Normal coronary sinus oxygen tension (CSo₂) is less than 20 mm Hg. Dole¹⁰⁶ observed that autoregulation was effective when CSo₂ was less than 25 mm Hg, but was completely lost when CSo₂ exceeded 32 mm Hg. Autoregulation can be intensified by vasoconstriction (increased oxygen extraction) and attenuated by vasodilation (decreased oxygen extraction).¹⁰⁷ The degradation of autoregulation with α-receptor blockade suggests a benefit of adrenergic coronary vasoconstriction.¹⁰⁸

Early reports indicated that autoregulation is less effective in the right ventricle than the left. More recently, it has been suggested that increases in right coronary pressures may produce large changes in Mvo₂, perhaps because of an exaggerated Gregg effect. When changes in myocardial metabolism are taken into account, autoregulation in the right and left ventricle is similar.^109,110

Quantitation of the degree of autoregulation must involve a comparison of the observed change in vascular resistance to the change in resistance that would have occurred in the absence of flow autoregulation. Some degree of autoregulation exists when the relative change in flow (ΔF/F) is less than the relative change in pressure (ΔP/P). From these definitions, Dole¹⁰⁷ has derived an autoregulation index that can be used to quantify the effects of different agents on coronary autoregulation.¹¹¹

Three theories have been proposed to explain coronary autoregulation: the tissue pressure theory, the myogenic theory, and the metabolic theory.¹¹² The tissue pressure hypothesis proposes that changes in perfusion pressure result in directionally similar changes in capillary filtration and, therefore, tissue pressure. In this way, extravascular resistance would oppose changes in flow with changes in perfusion pressure. Experimental evidence has shown, however, that there is no relation between the degree of autoregulation and the magnitude of change in tissue pressure. Arterial smooth muscle contracts in response to augmented intraluminal pressure; this is known as the myogenic response. Recently, this response has been demonstrated in coronary arterioles in the presence and absence of functioning endothelium.¹¹³ The argument for myogenic regulation of coronary flow is that myocardial metabolic changes are not rapid enough to explain large decreases in resistance after coronary occlusions for one or two heartbeats. However, myocardial metabolic events have been shown to occur during the course of a single cardiac contraction.¹¹⁴ The metabolic theory of autoregulation proposes that coronary arteriolar tone is determined by the balance of myocardial oxygen supply and demand. An increase in flow above the requirements of metabolism would wash out metabolites or cause accumulation of substrates, and this would be the signal for an appropriate change in coronary tone. Although metabolic regulation and autoregulation are separate phenomena, they may, therefore, have a common underlying mechanism. Metabolic regulation is discussed earlier (see Myocardial Metabolism). For an instructive, three-dimensional, graphic analysis of the interrelations among coronary artery pressure, myocardial metabolism, and coronary blood flow, see Feigl et al.¹¹⁵

Coronary Reserve

Myocardial ischemia causes intense coronary vasodilation. After a 10- to 30-second coronary occlusion, restoration of perfusion pressure is accompanied by a marked increase in coronary flow. This large increase in flow, which can be five or six times resting flow in the dog, is termed reactive hyperemia. Figure 6-10 illustrates that the repayment volume is greater than the debt volume. There is, however, no overpayment of the oxygen debt because oxygen extraction declines during the hyperemia.¹¹⁶ The presence of high coronary flows when coronary venous oxygen content is high suggests that mediators other than oxygen are responsible for this metabolically induced vasodilation.⁴³ The difference between resting coronary blood flow and peak flow during reactive hyperemia represents the autoregulatory coronary flow reserve: the further capacity of the arteriolar bed to dilate in response to ischemia. In Figure 6-9, the flow reserve is the vertical distance from the autoregulating pressure-flow curve (purple or red circles) to the nonautoregulating curve (triangles). The reserve is greater at higher perfusing pressure and lower Mvo₂. Unlike cannula-perfused preparations in which these data are obtained, in the clinical setting, increases in pressure increase both perfusing pressure and Mvo₂. Reactive hyperemia responses have been used in animals and humans to estimate coronary reserve in conditions such as obstructive coronary disease, aortic stenosis, and left ventricular hypertrophy^117–119. The myocardial fractional flow reserve (FFR) is calculated by dividing the pressure in a coronary vessel distal to a stenosis during maximal pharmacologic dilation by the aortic root pressure. This ratio (FFR) easily can be measured in the angiography suite and has been recommended as a useful index of the functional severity of coronary stenoses of intermediate morphologic severity on angiography, as well as a measure of residual obstruction after interventions.¹²⁰ Indeed, the relevance of a reduction in the FFR is highlighted in a recent randomized, controlled study that demonstrated improved clinical outcomes in FFR-guided percutaneous coronary interventions as opposed to angiography alone.¹²¹

Figure 6-10 Schematic diagram of the reactive hyperemic response to a 10-second coronary occlusion.

(From Marcus ML: Metabolic regulation of coronary blood flow. In Marcus ML [ed]: The coronary circulation in health and disease. New York: McGraw-Hill, 1983, pp 65–92. Reproduced by permission of McGraw-Hill Companies.)

It has been generally accepted that the coronary resistance vessels are maximally dilated when coronary perfusion pressure is reduced sufficiently to cause myocardial ischemia. In fact, agents such as adenosine, carbochromene, and dipyridamole can cause further increases in coronary flow in the presence of intense ischemia, when autoregulatory reserve is believed to be exhausted. This pharmacologic vasodilator reserve is greater than the autoregulatory vasodilator reserve. If flow to ischemic myocardium can be increased by pharmacologic dilation of resistance vessels, the use of these agents should reverse ischemic dysfunction and metabolism. Arteriolar dilators have, in general, not been found to be beneficial during myocardial ischemia. Coronary blood flow in the different layers of the ventricle must be reviewed to understand why (Box 6-6).

Transmural Blood Flow

It is well-known that, when coronary perfusion pressure is inadequate, the inner one third to one fourth of the left ventricular wall is the first region to become ischemic or necrotic.¹²² This increased vulnerability of the subendocardium may be caused by an increased demand for perfusion or a decreased supply, compared with the outer layers. There has been extensive study of the transmural distribution of oxygen consumption, use of oxidizable substrates, activity of glycolytic and mitochondrial enzymes, tissue contents of endogenous substrates, high-energy phosphates, lactate, isoforms of contractile proteins, and fiber stress and fiber shortening. In general, these studies indicate that if such differences exist between the layers of the left ventricle, they are unlikely to exceed 10% to 20%.^43,123 It is likely that preferential underperfusion of the subendocardium is the primary determinant of its increased vulnerability.

Regional blood flow in the myocardium is usually determined using radioactive microspheres. These plastic beads, labeled with a radioisotope, are injected into the bloodstream. The assumption is that they will mix uniformly with blood and be distributed in proportion to blood flow, as if they were red blood cells.¹²⁴ Because they are rigid and larger than red cells (9- or 15-μm diameters are usually chosen), they are trapped in the microcirculation. At the end of an experiment, the heart can be divided into small blocks and the amount of radioactivity in each piece measured in a gamma counter. The blood flow to each block of tissue will be proportional to the number of microspheres in each piece, which can be determined from its radioactivity. By using different radioisotopes as labels, several sets of microspheres can be injected during an experiment, giving “snapshots” of what flow was at the time of each injection. It is difficult to reduce the variability of the technique below 10%.^43,125 Subendocardial blood flow is found using this technique to be about 10% greater than subepicardial blood flow under normal circumstances. This gives a normal subendocardial/subepicardial or inner/outer (I/O) blood flow ratio of 1.10. This ratio is maintained at normal perfusing pressures even at heart rates greater than 200 beats/min.

If coronary pressure is gradually reduced, autoregulation is exhausted and flow decreases in the inner layers of the left ventricle before it begins to decrease in the outer layers (Figure 6-11). This indicates that there is less flow reserve in the subendocardium than in the subepicardium. The limits of autoregulation will depend on the level of cardiac work (see Autoregulation section earlier in this chapter) and on the experimental conditions. In conscious dogs, the mean coronary pressure at which evidence of subendocardial ischemia appeared was 38 mm Hg at a heart rate of 100 beats/min, and increased to 61 mm Hg at 200 beats/min. Subepicardial flow during tachycardia did not decline even at pressures as low as 33 mm Hg.¹²⁶ Because subepicardial flow is rarely inadequate, a subendocardial/subepicardial blood flow ratio close to 1.0 indicates adequate subendocardial flow and an appropriate matching of myocardial oxygen supply to oxygen demand. For this reason, the I/O ratio is often used as a measure of the adequacy of myocardial blood flow.

Figure 6-11 Pressure-flow relations of the subepicardial and subendocardial thirds of the left ventricle in anesthetized dogs.

In the subendocardium, autoregulation is exhausted and flow becomes pressure dependent when pressure distal to a stenosis declines to less than 70 mm Hg. In the subepicardium, autoregulation persists until perfusion pressure declines to less than 40 mm Hg. Autoregulatory coronary reserve is less in the subendocardium.

(Redrawn from Guyton RA, McClenathan JH, Newman GE, Michaelis LL: Significance of subendocardial ST segment elevation caused by coronary stenosis in the dog. Am J Cardiol 40:373, 1977.)

Three mechanisms have been proposed to explain the decreased coronary reserve in the subendocardium: differential systolic intramyocardial pressure, differential diastolic intramyocardial pressure, and interactions between systole and diastole. Because the force of systolic myocardial compression is greatest in the inner layers of the ventricle and is low at the subepicardium, it was believed that the outer layers of the heart were perfused throughout the cardiac cycle, whereas the subendocardium was perfused only during diastole. The subendocardium would have to obtain its entire flow during only a portion of the cycle and, therefore, would have to have a lower resistance. Recent studies, suggesting that there may be little systolic flow even to the outer layers, argue against this explanation.¹²⁷ The second mechanism is based on the high coronary pressures observed when coronary flow has ceased during a long diastole, P_zf (see Perfusion Pressure and Myocardial Compression earlier in this chapter).³⁹ The shape of the pressure-flow relation during a long diastole suggests that P_zf is higher in the subendocardium. This would mean that perfusion pressure for the subendocardium is lower in diastole compared with the outer layers of myocardium. Available evidence suggests that P_zf is not high in any layer and is unlikely to be more than 2 to 3 mm Hg greater in the subendocardium than in the subepicardium.¹²⁷ Hoffman^109,127 proposed an interaction between systole and diastole as the explanation for the increased vulnerability of the subendocardium to ischemia. During systole, intramyocardial pressure is high enough throughout most of the ventricular wall to squeeze blood out of the intramural vessels and into the extramural coronary veins and arteries. Because the compressive force is greatest in the subendocardium, vessels here are the narrowest at end systole. At the beginning of diastole, blood will be directed first to vessels with the lowest resistance, the larger vessels in the subepicardium, and last to the most narrowed vessels in the subendocardium. In this way, should the duration of diastole or the diastolic perfusion pressure be reduced, the subendocardial muscle would receive the least flow. Spaan³⁶ presents an interesting analysis of the interaction between arterial pressure and force of contraction as an intramyocardial pump. Although this theory is compatible with existing evidence, support for it will remain indirect until it becomes possible to measure phasic pressures and flows in separate layers of myocardium.

When the left ventricle hypertrophies in response to a pressure load (aortic stenosis, hypertension), myofibrillar growth outstrips the capillary network, resulting in decreased capillary density and increased diffusion distances. The net effect is to reduce coronary autoregulatory reserve.¹²⁸ The transmural gradient of reserve is exaggerated as well, so the subendocardium is at increased risk for ischemia in the hypertrophied heart compared with normal.¹²⁹

In addition to the transmural gradient of coronary reserve from outer to inner layer of the left ventricle, there is also marked variation of reserve between small regions of myocardium within a layer.¹³⁰ This heterogeneity of flow reserve may explain why pharmacologic reserve exceeds autoregulatory reserve (see Coronary Reserve section earlier in this chapter). During hypoperfusion, regional myocardial blood flow is decreased, but in all layers some small pieces of muscle will have no flow reserve left, whereas adjacent pieces can have substantial reserve. Fewer pieces will retain reserve in the subendocardium than in the subepicardium. The increase in flow in response to an infusion of adenosine is due to increased flow in the small regions with reserve, with no change in the adjacent fully dilated regions.^131–133 These findings suggest that ischemia causes maximal coronary vasodilation, and that increases in flow with adenosine or dipyridamole are due to dilation of vessels in nonischemic regions. Contrasting evidence is provided by Duncker and Bache,⁵⁴ who used a balloon occluder to simulate a coronary stenosis in exercising dogs. The occluder was adjusted to maintain distal coronary pressure constant at 43 mm Hg. During exercise, an intracoronary infusion of adenosine increased blood flow to all myocardial layers and improved regional systolic segment shortening. Although this is evidence of vasodilator reserve in ischemic myocardium, the constant distal pressure preparation does not faithfully mimic a coronary stenosis because it makes transmural steal impossible (see Coronary Steal later in this chapter). In general, pharmacologic dilation of resistance vessels has the potential to worsen ischemia by producing coronary steal. Dilation of larger penetrating vessels (50 to 500 μm in diameter) with nitrovasodilators could preferentially decrease resistance to blood flow to the subendocardium, and this may, in addition to favorable effects on the systemic circulation, explain the usefulness of nitrates in the treatment of angina.⁵⁴

Atherogenesis

Certain human arteries are more prone to develop atherosclerosis than others. For example, the coronary, renal, and internal carotid arteries, as well as some areas of the aorta, are known to be common sites for lesion formation.¹³⁴ In the Pathobiological Determinants of Atherosclerosis in Youth (PDAY) study, the aorta and right coronary arteries of 1378 young people aged 15 to 34 who died as a result of trauma were studied.¹³⁵ Two-dimensional maps of lipid-laden fatty streaks, as well as fibrous plaques, were made for each vessel. Although atherosclerosis is usually clinically silent until middle age or later, these investigators found that the disease process begins in adolescence or childhood. Moreover, fatty streaks and fibrous plaques do not occur randomly in the circulation but follow a well-defined distribution pattern. For example, in the right coronary artery, fatty streaks were found with the highest probability in the proximal 2 cm of this vessel, which closely parallels the distribution of raised fibrous lesions. However, in the abdominal aorta, where aortic lesions are commonly found, the high prevalence of fatty streaks did not always correlate with the prevalence of raised fibrous lesions. Therefore, at least in the aorta, the role of childhood fatty streaks in the development of adult fibrous lesions is uncertain.

The atherogenic stimuli that promote the progression of early lesions to clinically relevant stenoses are not known. Currently, the development of CAD is associated with various risk factors. Dyslipidemias, hypertension, diabetes mellitus, cigarette smoking, and a family history of premature CAD are known to correlate with premature vascular disease. The association between lipid disorders and atherogenesis is best understood and is discussed later in this chapter. Unfortunately, little is known about how the remaining risk factors may contribute to lesion development.

Historically, there are two classical theories of atherogenesis. According to von Rokitansky’s¹³⁶ thrombogenic (or encrustation) theory, fibrin is the initiating factor in lesion development. Later, Duguid¹³⁷ expanded on this theory by suggesting that atherosclerosis is the result of altered fibrinolysis, whereas more recent studies have documented the overexpression of prothrombotic factors, such as plasminogen activator inhibitor-1, in atherosclerotic plaques.¹³⁸ Alternatively, in 1856, Virchow’s imbibition (or insudation) theory proposed that atherosclerotic lesions were the result of altered vessel wall permeability.¹³⁹ Variations of this theory have been suggested by others, and all support the concept that the accumulation of various plasma components, including lipoproteins, may be important during lesion formation. For example, Ross and Glomset¹⁴⁰ blended the concepts of these original hypotheses into the “response-to-injury” hypothesis in which both lipid infiltration and thrombus formation play important roles in atherogenesis. Similarly, Schwartz and colleagues¹⁴¹ compared arterial narrowing in atherosclerotic arteries with the process of wound healing. This perspective has advantages, as it allows a multifactorial process such as atherosclerosis to be broken down into components of a more completely understood process such as the biology of a skin wound. For example, wound healing of any form begins with the formation of a clot (fibrin- and fibronectin-containing gel) that fills the wound and provides a provisional matrix for inflammatory cells, fibroblasts, and newly formed microvessels.^142,143 This is followed by the proliferation and migration of fibroblasts into the wound.¹⁴⁴ By day 7 after injury, microvessels grow into the base of the wound and form granulation tissue. As the wound matures and undergoes contracture, these blood vessels regress and fibroblasts disappear. After resorption of microvessels, tissue hypoxia develops and likely plays a role in the completion of the final scarring process.¹⁴⁵ As discussed later, there is now ample evidence to suggest that many similar events take place during arterial wound healing; however, because atherosclerosis is a chronic process, it is likely that vascular lesion formation involves indolent levels of inflammation with ongoing cycles of injury and repair over many years.^142,146,147

Arterial Wall Inflammation

A number of studies have demonstrated the presence of monocytes/macrophages and T lymphocytes in the arteries of not only advanced lesions but early atherosclerotic lesions of young adults.^148,149 Moreover, in experimental atherosclerosis, leukocyte infiltration into the vascular wall is known to precede smooth muscle cell hyperplasia.¹⁵⁰ Once inside the artery wall, mononuclear cells may play several important roles in lesion development. For example, monocytes may transform into macrophages and become involved in the local oxidation of low-density lipoproteins (LDLs) and accumulation of oxidized LDLs. Alternatively, macrophages in the artery wall may act as a rich source of factors that, for example, promote cell proliferation, migration, or the breakdown of local tissue barriers. The latter process of local tissue degradation may be important for the initiation of acute coronary artery syndromes because loss of arterial wall integrity may lead to plaque fissuring or rupture.¹⁵¹

Normally, the endothelium exhibits a low affinity for circulating leukocytes. Therefore, the transmigration of leukocytes into the artery wall must occur as a facilitated process. The release of proinflammatory cytokines such as interleukin-1 may promote the expression of leukocyte adhesive molecules.¹⁵² For simplicity, the interaction between leukocytes and the endothelium can be considered to involve three steps.¹⁵³ First, leukocytes in the bloodstream must loosely associate and roll along the endothelium—a process that is mediated by selectins expressed on endothelial cells.¹⁵² Second, firm adhesion of these leukocytes to endothelial cells occurs via the interaction between integrins, such as α₄β₁ (also known as very late antigen-4 [VLA-4]), expressed on leukocytes, and counter-receptors, such as vascular cell adhesion molecule-1 (VCAM-1), on endothelial cells.¹⁵⁴ Finally, the transmigration of leukocytes into the subendothelial space is mediated by various migration-inducing factors, such as monocyte chemoattractant protein-1 (MCP-1).¹⁵⁵

Dysfunction, discontinuity, or injury of the endothelial cell monolayer has been postulated to play a significant role in facilitating the transmigration of leukocytes into the intima and the development of intimal hyperplasia. However, the premise that regrowth of a healthy endothelium will limit neointimal accumulation is inconsistent with the results of several independent lines of investigation. For example, in experimental models, smooth muscle cell proliferation is not increased in arterial regions devoid of an endothelium.¹⁵⁶ Moreover, restoration of the endothelium, as might be achieved by seeding endothelial cells back into a denuded artery, does not decrease neointimal accumulation after vascular interventions.¹⁵⁷ Therefore, the presence of an endothelium in the central lumen of an artery and resistance to intimal growth do not appear to be inextricably linked. Finally, it is important to note that the endothelium is not restricted to the central lumen, as the artery wall is also invested with a rich supply of microvessels (i.e., vasa vasorum).^158–161 The vasa vasorum are likely another portal of entry for inflammatory cells into the artery wall, particularly because the expression of certain adhesion molecules is more abundant in the endothelium lining these microvessels than that of the central arterial lumen.¹⁶²

Role of Lipoproteins in Lesion Formation

The clinical and experimental evidence linking dyslipidemias with atherogenesis is well established and need not be reviewed here. However, the exact mechanisms by which lipid moieties contribute to the pathogenesis of atherosclerosis remain elusive. Although the simple concept of cholesterol accumulating in artery walls until flow is obstructed may be correct in certain animal models, this theory is not correct for human arteries.

Much of the pioneering work in understanding cholesterol metabolism is based on seminal observations by Brown and Goldstein.¹⁶³ The work of these two investigators focused on LDL, the so-called bad form of cholesterol, and the absence (or deficient forms) of the LDL receptor that are seen in familial hypercholesterolemia (FH). Patients with FH have high levels of LDL cholesterol and suffer from accelerated forms of atherosclerosis as cholesterol moieties enter the cell via an alternate route. In the absence of a functional LDL receptor, LDL cholesterol is oxidized and taken up by scavenger receptors of monocytes and macrophages resident within the artery wall. Steinberg¹⁶⁴ and others have integrated these data into a theory of atherogenesis that highlights the central role of LDL oxidation and the formation of lipid-laden monocytes in fatty streaks.

One of the major consequences of cholesterol accumulation in the artery wall is thought to be the impairment of endothelial function. The endothelium is more than a physical barrier between the bloodstream and the artery wall. Under normal conditions, the endothelium is capable of modulating vascular tone (e.g., via nitrous oxide), thrombogenicity, fibrinolysis, platelet function, and inflammation. In the presence of traditional risk factors, particularly dyslipidemias, these protective endothelial functions are reduced or lost. Notably, the loss of these endothelial-derived functions may occur in the presence or absence of an underlying atherosclerotic plaque and may simply imply that atherogenesis has begun. Aggressive attempts to normalize atherosclerotic risk factors (e.g., diet and lipid-lowering therapies) may markedly attenuate endothelial dysfunction—even in the presence of extensive atherosclerosis. A number of clinical studies now demonstrate dramatic improvements in endothelial function, as well as cardiovascular morbidity and mortality, with the use of inhibitors of 3-hydroxy-3-methylglutaryl coenzyme A (HGM-CoA) reductase, or “statins.”^165–167 Future studies may help clarify the exact mechanisms by which dyslipidemias (and other risk factors) alter endothelial function.

Smooth Muscle Cell Proliferation, Migration, and Arterial Remodeling

The dominant cell type in atherosclerotic lesions is the smooth muscle cell, and as lesions progress, the number of smooth muscle cells in the artery wall tends to increase. Therefore, smooth muscle replication must occur at some time during atherogenesis. Perhaps the first line of evidence that cell replication occurs in human arteries is from the observation that atherosclerotic plaques contain monoclonal cell populations. Elegant studies by Benditt and Benditt¹⁶⁸ demonstrated that groups (or clones) of cells that arise from a single progenitor cell are present in tissue from atherosclerotic coronary arteries of women who were deficient in glucose-6-phosphate dehydrogenase (G-6-PD). Because G-6-PD is an X-chromosome–linked enzyme that has two isoforms, cells would express only one isoform, with the other isoform being suppressed on the inactivated X chromosome. Therefore, groups of cells in an atherosclerotic plaque that contain only one isoform of G-6-PD are likely the result of proliferation of a single progenitor cell. More recently, Murry and colleagues¹⁶⁹ studied the monoclonality of atherosclerotic plaques using X-chromosome inactivation patterns. Using the polymerase chain reaction, they examined the monoclonality of plaques according to the methylation pattern of the human androgen receptor gene, a highly polymorphic locus on the X chromosome for which 90% of women are heterozygous. These investigators note that diseased and normal arteries contain monoclonal populations (or patches) of cells. Therefore, they speculate that the monoclonality of plaques might be caused by expansion of a preexisting monoclonal patch of cells, rather than mutation or selection of individual cells in the artery wall.

Little is known about when and why cells proliferate in the artery wall. However, it is known that early in life there is a rapid expansion in neointimal smooth muscle cell mass. Sims and Gavin^170,171 describe the accumulation of intimal smooth muscle cells in the left anterior descending coronary artery of neonates. Using electron microscopy, these investigators demonstrate interruptions in the internal elastic lamina in coronary arteries where a neointima had formed.¹⁷¹ These interruptions in the internal elastic lamina are not present in all human arteries. Indeed, the internal mammary artery, which typically is devoid of atherosclerosis, has an intact internal elastic lamina. Therefore, it has been suggested that medial smooth muscle cells migrate inward through breaks in the internal elastic lamina to expand and form a neointima. The frequency and degree of smooth muscle cell replication in adult coronary arteries have been examined by various investigators. The majority of these studies have demonstrated very low replication rates in tissue from both normal and diseased arteries.^172–175 Whether these low cell replication rates are sufficient to gradually result in advanced lesions, or whether sporadic bursts of replication occur in response to injury, is unknown. Finally, it is recognized that programmed cell death, or apoptosis, occurs in the artery wall.¹⁷⁶ Therefore, the accumulation of cells in the artery wall is a function of not only cell proliferation, but also apoptosis.

The role of smooth muscle cell migration in adult CAD is poorly understood. It has been suggested, however, that like fibroblasts that migrate into the base of a wound, arterial wall smooth muscle cells migrate inward to expand plaque mass. Smooth muscle cell migration into the intima has been studied in various animal models of neointimal formation (e.g., rat carotid artery model).¹¹ The majority of these models demonstrate the inward migration of medial smooth muscle cells after normal arteries are subjected to balloon injury. A number of growth factors (e.g., platelet-derived growth factor) have been shown to play an important role in facilitating smooth muscle cell migration in these models.^177–179 Unfortunately, the clinical relevance of these experimental observations remains to be clarified, as the milieu for cell migration in complex human lesions appears to be very different from that of normal animal arteries that are subjected to injury. More information is required regarding the factors that regulate smooth muscle cell migration, as well as why smooth muscle cells differ in their propensity to migrate after injury.

Finally, it is important to point out that the buildup of atherosclerotic plaque does not always translate into the formation of arterial obstructions.¹⁴² For example, Glagov¹⁸⁰ notes that human vessels can accumulate massive amounts of atherosclerotic plaque without encroaching on the central arterial lumen. Instead, abluminal expansion of the artery wall may occur until 40% of the area encompassed by the internal elastic lamina is occupied by plaque. Thereafter no further enlargement may occur, and luminal narrowing may ensue. Although this form of compensatory enlargement is referred to as “remodeling,” the term is confusing because it holds different meaning in different contexts (Figures 6-13 and 6-14).^180–182 For example, remodeling has also been used to describe the arterial response to changes in blood flow (e.g., during pregnancy or in the neonatal period) or pressure (e.g., hypertension).¹⁸³ In addition, remodeling has been invoked as a key component of the response to arterial injury—however, with quite a different meaning.¹⁸⁴ In animal models of arterial injury, as well as studies of human coronary arteries that have undergone angioplasty (or percutaneous transluminal coronary angioplasty [PTCA]), “shrinkage” or constrictive remodeling of the artery wall is a major determinant of lumenal narrowing, whereas neointimal formation plays a minor role in this process.^185–192

Figure 6-13 Arterial remodeling.

Serial sections, proximal (A), mid (B), and distal (C), of an atherosclerotic human left circumflex coronary artery (Movat’s pentachrome-stained slide, original magnification ×40). There is narrowing of the central arterial lumen in the mid and distal sections; however, the total arterial area of these sections is also larger than that of the proximal section. The ability of arteries to undergo compensatory enlargement is referred to as arterial remodeling.

Figure 6-14 Model of atherogenesis in human coronary arteries.

A, Normal coronary artery. B, Infiltration of the intima by particles containing low-density lipoproteins (LDLs) stimulates expression of adhesion molecules on the luminal surface of the endothelium. C, Monocyte/macrophage translocation into the intima. Once translocation is complete, uptake of LDL via scavenger receptors on macrophages creates foam cells that secrete proinflammatory cytokines, such as interleukin-1. D, Further inflammation promotes division and migration of medial and/or adventitial smooth muscle cells with ongoing accumulation of foam cells and extracellular matrix, resulting in intimal thickening. E, Local cell signaling via paracrine factors can lead to regional apoptosis and development of a necrotic core with accumulation of cholesterol deposits.

How arterial wall constriction is accomplished or why some, but not all, arteries undergo compensatory dilatation to preserve lumen area is incompletely understood.^193–195 Blood flow and shear stress are known to play a critical role in remodeling. The response of arteries to chronic alterations in blood flow are endothelium dependent.^196,197 For example, Langille and O’Donnell¹⁹⁶ demonstrated in rabbit carotid arteries that decreased blood flow results in narrowing of the vessel diameter that is unchanged with papaverine and likely caused by structural changes in the artery wall. However, when the endothelium is removed from these vessels, the response to reduced blood flow is abolished. In atherosclerotic arteries that contain a rich network of endothelial cell-lined microvessels or vasa vasorum, the role of the endothelium in regulating remodeling may be important.^198–200

Assessment of Atherosclerosis by Intravascular Ultrasonography

A detailed description of both diagnostic and therapeutic procedures performed in the cardiac catheterization laboratory is provided in Chapter 3. However, given that an integral understanding of the anatomy of the coronary artery and the atherosclerotic lesion is necessary for the appropriate interpretation and use of these technologies, a brief review of new developments in invasive assessment of atherosclerosis by intravascular ultrasonography (IVUS) is included here.

Standard coronary angiography gives operators a two-dimensional representation of the lumen. By examining arteries in multiple views, the operator estimates coronary stenoses by comparison of the lumen diameter at the point of maximal narrowing to adjacent disease-free segments. However, as discussed earlier, development of the atherosclerotic plaque results not only in luminal encroachment but arterial remodeling,²⁰¹ meaning significant disease may be overlooked on traditional angiography. Thus, technologies, such as IVUS, are becoming more prevalent in the assessment of CAD.

IVUS of coronary arteries was first popularized in the 1990s, ²⁰² with subsequent refinement of catheter delivery systems and commercialization making it now commonplace in the modern catheterization laboratory. Compatible with most guiding catheters, IVUS probes are delivered to the coronary arteries via standard angiography techniques, and both manual and mechanical pullback of the IVUS probe allows operators to assess real-time cross-sectional images. Complementary software can then allow users to generate either longitudinal or three-dimensional reconstructions of the interrogated vessel.

IVUS images demonstrate remarkable fidelity to cross-sectional histologic specimens and permit accurate visualization and measurement of the intima, media, and, in some instances, the adventitia (Figure 6-15). Arterial remodeling with significant intimal hyperplasia but relatively intact lumen diameter can thus identify occult disease not otherwise appreciated on standard angiography. One landmark article has noted that even if only minor luminal irregularities exist, atherosclerotic disease can be demonstrated throughout most other vessels in the coronary circulation, suggesting luminograms may be simply the tip of the atherosclerotic iceberg. ²⁰³ Indeed, IVUS is now commonly used to more accurately quantify lesions in cases of intermediate severity lesions,²⁰⁴ or in regions that are otherwise difficult to assess on standard angiography, such as left main disease. As well, IVUS allows operators to assess arteries after percutaneous intervention for not only adequate stent deployment but also for complications, such as arterial dissection, which can be missed on standard angiography.²⁰⁵ However, IVUS is not limited to simply documenting and quantifying atherosclerotic burden. Plaque composition also can be assessed qualitatively and classified based on acoustic impedance, allowing differentiation among fibromuscular “soft” lesions, dense “fibrous” lesions, and “calcified” hyperechoic lesions.²⁰⁶ Although not yet reliably predictable, ongoing studies are aimed at identifying which plaques are “vulnerable” or susceptible to rupture, thus causing acute vessel closure and myocardial infarction.^207,208

Figure 6-15 Intravascular ultrasonography (IVUS) for assessment of human coronary arteries.

A, A right anterior oblique projection of the left coronary circulation. The left main artery (LM), left anterior descending artery (LAD), circumflex artery (Cx), and obtuse marginal artery (OM) are seen. A severe distal LM stenosis is seen bifurcating into the LAD and Cx. B, Postpercutaneous intervention angiogram demonstrates no residual stenosis. Drug-eluting stents were placed in the left main, LAD, and Cx. The tip of an aortic balloon pump can be seen (arrow). Letters indicate location of IVUS images for correlating panels. C–E, IVUS images of LM, LM bifurcation, and LAD, respectively. In, intima; IP, IVUS probe; M, media. Asterisk indicates guidewire; arrows indicate stent struts.

Coronary Artery Stenoses and Plaque Rupture

Coronary atherosclerosis is a chronic disease that develops over decades, remaining clinically silent for prolonged periods (Box 6-8). Clinical manifestations of CAD occur when the atherosclerotic plaque mass encroaches on the vessel lumen and obstructs coronary blood flow, causing angina. Alternatively, cracks or fissures may develop in the atherosclerotic lesions and result in acute thromboses that cause unstable angina or myocardial infarction.

On angiography, patients with stable angina typically have lesions with smooth borders. Only a minority of coronary lesions are concentric, most having a complex geometry varying in shape over their length. Eccentric stenoses, with a remaining pliable, musculoelastic arc of normal wall, can vary in diameter and resistance in response to changes in vasomotor tone or intraluminal pressure. Most human coronary stenoses are compliant.²⁰⁹ The intima of the normal portion of the vessel wall is often thickened, making endothelial dysfunction probable (see Dynamic Stenosis section later in this chapter). In contrast, patients with unstable angina usually have lesions characterized by overhanging edges, scalloped or irregular borders, or multiple irregularities. These complicated stenoses likely represent ruptured plaque or partially occlusive thrombus, or both.²¹⁰ On angiography, these lesions may appear segmental, confined to a short segment of an otherwise normal proximal coronary artery. At autopsy, however, the most common pathologic finding is diffuse vessel involvement with superimposed segmental obstruction of greater severity.²¹¹ In a diffusely narrowed vessel, even modest progression of luminal narrowing can be significant. In such a patient, rating the significance of the obstruction by the percentage of diameter reduction relative to adjacent vessel segments will underestimate its physiologic importance.^212,213 Therefore, understanding the characteristics of atherosclerotic plaques is of central importance to the management of acute coronary artery syndromes.

The intuitive notion that the severity of coronary artery stenoses should correlate with the risk for complications from CAD has been disproved by several key investigations. Ambrose and colleagues²¹⁴ reviewed the coronary angiograms of 38 patients who had had a Q-wave myocardial infarct in the interval between serial studies. On the preinfarct angiograms, the mean percentage stenosis at the coronary segment that was later responsible for infarction was only 34%. Similarly, Little and colleagues²¹⁵ reviewed the coronary angiograms of 42 patients who also had this procedure performed at an interval before and after myocardial infarction. Total occlusion of a previously patent artery was observed in 29 patients; yet, for 19 of these occluded arteries, the degree of stenosis was less than 50% on the initial angiogram. Therefore, although the revascularization of arteries with critical stenoses in target lesions is appropriately indicated to reduce symptoms and myocardial ischemia, the risk for further cardiac events remains because atherosclerosis is a diffuse process and mild or modest angiographic stenoses are more likely to result in subsequent myocardial infarction than are severe stenoses.

With this background comes the question of predicting which arterial segments with minimal angiographic disease will later develop new critical stenoses. Clues to the answer for this question are emerging from careful pathologic studies of lesions by Davies and Thomas.²¹⁶ Superficial intimal injury (plaque erosions) and intimal tears of variable depth (plaque fissures) with overlying microscopic mural thrombosis are commonly found in atherosclerotic plaques. In the absence of obstructive luminal thrombosis, these intimal injuries do not cause clinical events. However, disruption of the fibrous cap, or plaque rupture, is a more serious event that typically results in the formation of clinically significant arterial thromboses. From autopsy studies, it is known that rupture-prone plaques tend to have a thin, friable fibrous cap.²¹⁷ The site of plaque rupture is thought to be the shoulder of the plaque, where substantial numbers of mononuclear inflammatory cells are commonly found.²¹⁸ The mechanisms responsible for the local accumulation of these cells at this location in the plaque are unknown; presumably, monocyte chemotactic factors, the expression of leukocyte cell adhesion molecules, and specific cytokines are involved.^162,219 Moreover, macrophages in plaques have been shown to express factors such as stromelysin, which promote the breakdown of the extracellular matrix, and thereby weaken the structural integrity of the plaque.¹⁵¹ Currently, no effective strategies have been designed to limit the possibility of plaque rupture; however, as discussed later, aggressive lipid-lowering therapy may be a helpful preventative measure.

Hemodynamics

If accurate angiographic assessment of the geometry of a coronary stenosis is made, hydrodynamic principles can be used to estimate the physiologic significance of the obstruction.²²⁰ Energy is lost when blood flows through a stenosis because of entrance effects, frictional losses in the stenotic segment, and separation losses caused by turbulence as blood exits the stenosis (Figure 6-16). The equation relating stenosis geometry to hemodynamic severity is:

Figure 6-16 Sources of energy loss across a stenosis.

Equations that (accurately) predict the pressure gradient across a stenosis usually ignore entrance effects. Frictional losses are proportional to blood velocity but are usually not important except in very long stenoses. Separation losses, caused by turbulence as blood exits the stenosis, increase with the square of blood velocity and account for more than 75% of energy loss. F, friction coefficient (Poiseuille); S, separation coefficient; V, blood velocity.

(From Marcus ML: The physiologic effects of a coronary stenosis. In Marcus ML [ed]: The coronary circulation in health and disease, New York: McGraw-Hill, 1983, pp 242–269. Reproduced by permission of McGraw-Hill Companies.)

where ΔP is the pressure decline across the stenosis, Q is the volume flow of blood, f is a factor accounting for frictional effects, and s accounts for separation effects. Based on the Poiseuille law for laminar flow:

where π is the blood viscosity, L is stenosis length, A_n is the cross-sectional area of the normal vessel, and A_s is the cross-sectional area of the stenosis. The separation or turbulence factor is:

where ρ is blood density, and k is an experimentally determined coefficient. Thus, frictional losses are directly proportional to the first power of stenosis length but are inversely proportional to the square of the area (or fourth power of diameter). Separation losses are particularly prominent because they increase with the square of flow. Even at resting flows, more than 75% of energy loss is due to this turbulence when blood exits the stenosis. Except for very long stenoses, the frictional term can be neglected.²⁰⁹ Thus, the amount of energy loss or pressure decline across the obstruction increases exponentially as flow rate increases. For this reason, exercise, anemia, and arteriolar vasodilator drugs (e.g., dipyridamole) are poorly tolerated in the presence of a severe stenosis. Figure 6-17 illustrates that, although resting flow is unaffected until coronary diameter is reduced by more than 80%, maximal flow begins to decline when diameter is reduced by 50%.

Figure 6-17 Effect of increasing stenosis severity at resting and maximal coronary flows.

At rest, lumen diameter must be reduced by more than 80% before flow decreases (green line). Because pressure drop across a stenosis increases exponentially with blood velocity, maximal coronary flow is restricted by a 50% diameter reduction (purple line).

(From Gould KL, Lipscomb K: Effect of coronary stenoses on coronary flow reserve and resistance. Am J Cardiol 34:48, 1974.)

Resting flow in Figure 6-17 remains constant as lumen diameter decreases because the coronary arterioles progressively dilate, thereby reducing the resistance of the distal coronary bed sufficiently to compensate for the resistance of the stenosis. As the severity of the stenosis increases further, the arteriolar bed can no longer compensate and flow begins to fall. This is an example of autoregulation: As stenosis severity increases, distal perfusion pressure decreases, arterioles dilate to maintain flow until autoregulation is exhausted (in the subendocardium first) and flow becomes pressure dependent. As illustrated in Figure 6-9, the distal pressure (or stenosis diameter) at which flow becomes pressure dependent is lower at low levels of myocardial metabolism (Mvo₂). The interpretation of normal resting flow can be difficult in the presence of CAD. A coronary artery supplying blood through collaterals to a large mass of myocardium will require high resting flow rates and even a mild stenosis may be flow limiting.

The term critical stenosis is frequently used. This is usually defined as a coronary constriction sufficient to prevent an increase in flow over resting values in response to increased myocardial oxygen demands.²²¹ This is a greater degree of obstruction than an angiographically significant stenosis, which is usually defined as a reduction in cross-sectional area of 75%, which is equivalent to a 50% decrease in the diameter of a concentric stenosis.²¹² A critical stenosis is demonstrated experimentally by blunting or abolishing reactive hyperemia (see Autoregulation section earlier in this chapter). This is evidence that autoregulation has been exhausted in at least the inner layer of myocardium (see Transmural Blood Flow section earlier in this chapter). Notably, the critical nature of the stenosis is relative to the resting Mvo₂. If oxygen demand decreases, some coronary autoregulatory reserve will be recovered and the stenosis will no longer be critical. The failure to recognize this fact has led to misinterpretation of studies designed to demonstrate coronary steal (see later).

Coronary Collaterals

Coronary collaterals are anastomotic connections, without an intervening capillary bed, between different coronary arteries or between branches of the same artery. In the normal human heart, these vessels are small and have little or no functional role. In patients with CAD, well-developed coronary collateral vessels may play a critical role in preventing death and myocardial infarction. Individual differences in the capability of developing a sufficient collateral circulation is a determinant of the vulnerability of the myocardium to coronary occlusive disease.²²² There is great interspecies variation in the ability of the collateral circulation to support myocardial perfusion after acute coronary occlusion; pigs and rats have little collateral circulation and infarct almost all the area at risk, whereas dogs and cats with better collateralization will infarct less than 75% of the area at risk.²²³ In the guinea pig, collaterals are so well developed that coronary occlusion does not even decrease myocardial blood flow. There are also differences in the location of collateral vessels; in dogs, collaterals develop in a narrow subepicardial zone, at the border of the potentially ischemic region, whereas in pigs, a dense subendocardial plexus develops in response to coronary occlusion. In the presence of coronary disease, humans exhibit a small number of large epicardial collateral vessels and numerous small subendocardial vessels.

In response to coronary occlusion, native coronary collateral vessels do not passively stretch but undergo an active growth process that within 8 weeks, in the dog, can restore perfusion sufficient to support normal myocardial function, even during exercise. Human collaterals have a tortuous corkscrew-like pattern visible on angiography. This may be because of an embryonal pattern of vascular development in which longitudinal growth of smooth muscle cells occurs at the same time as radial growth. In the nongrowing adult heart, this increase in length results in tortuosity.²²⁴ There is much interest in discovering the factors that control collateral vessel growth in the hope of providing therapy for patients who cannot be revascularized otherwise. Arteriogenesis refers to the transformation of preexisting collateral arterioles into functional arteries with a thick muscular coat and the acquisition of viscoelastic and vasomotor properties.²²⁵ Fujita and Tambara²²⁶ provide an overview of the process: A high-grade coronary stenosis decreases distal intra-arterial pressure, resulting in an increased pressure gradient across the preexisting collateral network. Increased collateral blood flow results in increased shear stress at the endothelium, which upregulates cell adhesion molecules. This leads to adherence of monocytes, which transform into macrophage, and the production and release of growth factors such as granulocyte macrophage colony-stimulating factor (GM-CSF), MCP-1, and basic fibroblast growth factor (bFGF). Angiogenesis is not directly related to collateral vessel development but refers to the proliferation, migration, and tube formation of capillaries in the central area of ischemic regions.²²⁷ The development of a treatment to promote collateral growth in patients with intractable CAD is currently a subject of intense investigation.

Evidence in dogs suggests that mature coronary collaterals respond differently to neurohumoral stimulation than normal coronary arteries. Collaterals do not constrict in response to α-receptor activation, but do dilate in response to β₁– or β₂-agonists. They constrict in response to prostaglandin F_2α (PGF_2α) and AII, but less so than normal vessels. Remarkably, collateral vessels constrict in response to vasopressin to a much greater extent than normal vessels. In vivo studies in dogs indicate that levels of vasopressin present during stress (hemorrhage, cardiopulmonary bypass) can diminish flow to collateral-dependent myocardium. This is likely due to constriction of collateral vessels, as well as enhanced vasoconstriction of the resistance vessels in the collateral-dependent myocardium.⁷⁹ It is possible that the endothelial cells of both types of vessel are dysfunctional.²²⁸ Relaxation in response to nitroglycerin was enhanced. This is a further mechanism for the beneficial effects of nitroglycerin in CAD. The deleterious effects of coronary arteriolar dilators such as adenosine and dipyridamole are discussed later (see Coronary Steal).

It has been estimated that, in humans, perfusion via collaterals can equal perfusion via a vessel with a 90% diameter obstruction.²²⁹ Although coronary collateral flow can be sufficient to preserve structure and resting myocardial function, muscle dependent on collateral flow usually becomes ischemic when oxygen demand increases above resting levels.²³⁰ It is possible that evidence from patients with angina underestimates collateral function of the population of all patients with CAD. Perhaps individuals with coronary obstructions but excellent collateralization remain asymptomatic and are not studied.

Pathogenesis of Myocardial Ischemia

Ischemia is the condition of oxygen deprivation accompanied by inadequate removal of metabolites consequent to reduced perfusion.²³¹ Clinically, myocardial ischemia is a decline in the blood flow supply/demand ratio resulting in impaired function. There is no universally accepted “gold standard” for the presence of myocardial ischemia. In practice, symptoms, anatomic findings, and evidence of myocardial dysfunction must be combined before concluding that myocardial ischemia is present.²³² Conclusive evidence of anaerobic metabolism in the setting of reduced coronary blood flow (relative to demand) would be convincing. Such evidence is extremely difficult to obtain, even in experimental preparations.

Determinants of Myocardial Oxygen Supply/Demand Ratio

An increase in myocardial oxygen requirement beyond the capacity of the coronary circulation to deliver oxygen results in myocardial ischemia (Box 6-9). This is the most common mechanism leading to ischemic episodes in chronic stable angina and during exercise testing. Intraoperatively, the anesthesiologist must measure and control the determinants of Mvo₂ and protect the patient from “demand” ischemia. The major determinants of Mvo₂ are heart rate, myocardial contractility, and wall stress (chamber pressure × radius/wall thickness). Shortening, activation, and basal metabolic requirements are minor determinants of Mvo₂ (Figure 6-18).

Figure 6-18 Relative importance of variables that determine myocardial oxygen consumption (Mvo₂).

Each line roughly approximates the effect of manipulating one variable without changing the others. Most interventions cause changes in several of the variables at the same time. The importance of contractility, which is difficult to monitor in practice, is apparent.

(From Marcus ML: Metabolic regulation of coronary blood flow. In Marcus ML [ed]: The coronary circulation in health and disease. New York: McGraw-Hill, 1983, pp 65–92. Reproduced by permission of McGraw-Hill Companies.)

An increase in heart rate can reduce subendocardial perfusion by shortening diastole. Coronary perfusion pressure may decline because of reduced systemic pressure or increased left ventricular end-diastolic pressure. With the onset of ischemia, perfusion may be further compromised by delayed ventricular relaxation (decreased subendocardial perfusion time) and decreased diastolic compliance (increased left ventricular end-diastolic pressure). Anemia and hypoxia also can compromise delivery of oxygen to the myocardium. Several indices of myocardial oxygen supply/demand ratio have been proposed to guide therapy. The rate-pressure product (heart rate × systolic blood pressure) gives a good estimate of Mvo₂ but does not correlate well with ischemia. A patient with a systolic pressure of 160 and heart rate of 70 has a much lower likelihood of ischemia than a patient with a pressure of 70 and rate of 160, although both have a rate-pressure product of 11,200. The ratio of the diastolic pressure–time index (DPTI) to the systolic pressure–time index (SPTI) was devised to estimate subendocardial perfusion and takes into account determinants of oxygen delivery (Figure 6-19).^233–235 When blood oxygen content was included, the index became a good predictor of endocardial flow in animals with normal coronary arteries. More recently, the ratio of mean arterial pressure/heart rate has been proposed as a correlate of myocardial ischemia.²³⁶ In dogs with moderate-to-severe coronary stenoses, systolic shortening was best with high pressures and low heart rate, and worst with low pressure and high heart rate. None of these indices has proved to be reliable in the clinical setting. Their major value is to bring attention to the important variables determining the supply/demand ratio. These variables should be measured (or estimated) and controlled individually.

Figure 6-19 Three indices, proposed to predict the adequacy of subendocardial perfusion in normal dogs, illustrate the variables determining myocardial oxygen supply and demand. The systolic pressure-time index (SPTI) relates to oxygen demand. The diastolic pressure-time index (DPTI) relates to the supply of coronary blood flow (CBF) to the inner layers of the left ventricle. Arterial oxygen content (O₂ content) is important when there are large changes in hematocrit. Ao, aortic pressure; ENDO, subendocardial layer of left ventricle; EPI, subepicardial layer of the left ventricle; LV, left ventricular pressure.

(From Hoffman JIE, Buckberg GD: Transmural variations in myocardial perfusion. In Yu PN, Goodwin JF [eds]: Progress in cardiology. Philadelphia: Lea & Febiger, 1976, p 37.)

Dynamic Stenosis

Patients with CAD can have variable exercise tolerance during the day and between days. Ambulatory monitoring of the electrocardiogram has demonstrated that ST-segment changes indicative of myocardial ischemia, in the absence of changes in oxygen demand, are common.²³⁷ These findings are explained by variations over time in the severity of the obstruction to blood flow imposed by coronary stenoses.

Although the term hardening of the arteries suggests rigid, narrowed vessels, in fact, most stenoses are eccentric and have a remaining arc of compliant tissue (Figure 6-20). A modest amount (10%) of shortening of the muscle in the compliant region of the vessel can cause dramatic changes in lumen caliber.²⁰⁹ This was part of Prinzmetal’s original proposal to explain coronary spasm. Maseri et al²³⁸ suggest that the term spasm should be reserved for “situations where coronary constriction is both focal, sufficiently profound to cause transient coronary occlusion, and is responsible for reversible attacks of angina at rest” (i.e., variant angina). Although this syndrome is rare, lesser degrees of obstruction in response to vasoconstrictor stimuli are common among patients with CAD.

Figure 6-20 Drawings and incidence of the various types of structure of stenoses observed in human coronary artery specimens.

In almost three quarters of vessels with greater than 50% narrowing, the residual arterial lumen was eccentric and partially circumscribed by an arc of normal arterial wall. In such lesions, a decline in intraluminal pressure or an increase in vasomotor tone can cause lumen diameter to decrease further and sufficiently to precipitate myocardial ischemia.

(From Brown BG, Bolson EL, Dodge HT: Dynamic mechanisms in human coronary stenosis. Circulation 70:917, 1984; redrawn from Freudenberg H, Lichtlen PR: The normal wall segment in coronary stenoses-a postmortem study. Z Kardiol 70:863, 1981.)

Sympathetic tone can be increased by the cold pressor test (immersing the arm in ice water) or isometric handgrip testing. In response to this maneuver, coronary resistance decreased in normal subjects but increased in patients with coronary disease, some of whom experienced angina.²³⁹ This increase in resistance appears to be mediated by α receptors because it can be prevented by phentolamine. Studies using quantitative coronary angiography have documented reductions in caliber in diseased vessels in contrast with dilation of vessels in healthy subjects.^98,240 Zeiher et al²⁴⁰ showed that the same vessel segments that constricted with the cold pressor test also constricted in response to an infusion of acetylcholine. Because the normal, dilatory response to acetylcholine is dependent on intact endothelium, these findings suggest that the abnormal response of stenotic coronary arteries is due to endothelial dysfunction.

Animal models of coronary vasospasm demonstrate that enhanced vascular smooth muscle reactivity also may underlie vasospasm.²⁴¹ Rho, a GTP-binding protein, sensitizes vascular smooth muscle cells to calcium by inhibiting myosin phosphatase activity through an effector protein called Rho-kinase. Upregulation of this pathway may be a mechanism of coronary vasospasm. Interestingly, Rho-kinase inhibitors have been shown to block agonist-induced vasoconstriction of internal thoracic artery segments from patients undergoing coronary artery surgery.²⁴²

It also has been noted that, in patients with coronary disease, some of the angiographically normal-appearing segments also respond abnormally.²⁴³ It appears likely that, during the development of coronary atherosclerosis, endothelial dysfunction precedes the appearance of visible stenoses. In patients with angiographically smooth coronary arteries, Vita et al²⁴⁴ found that an abnormal response to acetylcholine was correlated with serum cholesterol, male sex, age, and family history of coronary disease. The normal dilation of epicardial coronary arteries in response to increased blood flow (shear stress) has been shown to be absent in atherosclerotic vessels.²⁴⁵ As well, it has been demonstrated that patients with coronary disease respond to 5-HT with coronary vasoconstriction instead of the normal vasodilatory response.^246,247 The concentrations of 5-HT used were within the range found in coronary sinus blood of patients with coronary disease. Very high concentrations of 5-HT may be found on the endothelium at the site of aggregating platelets.²⁴⁶ All these findings point to the central role of endothelial dysfunction in the abnormal coronary vasomotion of patients with atherosclerosis²⁴⁸ (see Endothelium section earlier in this chapter).

Coronary Steal

Steal occurs when the perfusion pressure for a vasodilated vascular bed (in which flow is pressure dependent) is lowered by vasodilation in a parallel vascular bed, both beds usually being distal to a stenosis. Two kinds of coronary steal are illustrated: collateral and transmural (Figure 6-21).

Figure 6-21 Conditions for coronary steal between different areas of the heart (collateral steal [A]) and between the subendocardial and the subepicardial layers of the left ventricle (transmural steal [B]). P₁, aortic pressure; P₂, pressure distal to the stenosis; R₁, stenosis resistance; R₂ and R₃, resistance of autoregulating and pressure-dependent vascular beds, respectively.

(From Epstein SE, Cannon RO, Talbot TL: Hemodynamic principles in the control of coronary blood flow. Am J Cardiol 56:4E, 1985.)

Collateral steal in which one vascular bed (R₃), distal to an occluded vessel, is dependent on collateral flow from a vascular bed (R₂) supplied by a stenotic artery is shown (see Fig 6-21A). Because collateral resistance is high, the R₃ arterioles are dilated to maintain flow in the resting condition (autoregulation). Dilation of the R₂ arterioles will increase flow across the stenosis, R₁, and decrease pressure, P₂. If R₃ resistance cannot further decrease sufficiently, flow there will decline, producing or worsening ischemia in the collateral-dependent bed. The values of all the resistances, including collaterals, and the baseline myocardial metabolic state will determine how powerful the vasodilator stimulus must be to produce ischemia in the collateral bed. Failure to recognize this has confounded studies of vasodilator drugs. If collateral vessels are very well developed or Mvo₂ is low, sufficient autoregulatory reserve may remain in the collateral-dependent bed to maintain adequate myocardial blood flow even with the administration of a moderately powerful vasodilator.

Transmural steal is also illustrated (see Figure 6-21B). Normally, vasodilator reserve is less in the subendocardium (see Transmural Blood Flow section earlier in this chapter). In the presence of a stenosis, flow may become pressure dependent in the subendocardium, whereas autoregulation is maintained in the subepicardium. This is illustrated in Figure 6-11, where at a perfusion pressure of 50 mm Hg, flow has declined in the subendocardium, whereas the subepicardium retains autoregulatory reserve. Dilation of the subepicardial arterioles, R₂, will then increase flow across the stenosis (R₁) causing P₂ to fall and resulting in decreased flow to the subendocardium as subepicardial flow increases.

The term steal is most appropriate when the vasodilation is caused by a pharmacologic agent (adenosine, dipyridamole) producing “luxury” flow (beyond metabolic requirements) in the vascular bed with coronary reserve (R₂). The same redistribution of blood flow also occurs during exercise in response to metabolically mediated vasodilation. The study of coronary steal demonstrates well the complex interrelations among the determinants of myocardial blood flow.