Grossman’s Cardiac Catheterization, Angiography, & Intervention
7th Edition

The desired catheter is inserted into the femoral sheath and advanced to the level of the left mainstem bronchus over the guidewire. Alternatively, some operators prefer to advance the tip of the coronary angiography catheter around the arch and into the ascending aorta before the guidewire is removed. Although this may reduce snagging of the catheter tip on aortic wall plaques and irregularities, it places greater emphasis on the precision of initial catheter flushing. After removal of the guidewire, the catheter is attached to a specially designed manifold system that permits the maintenance of a closed system during pressure monitoring, catheter flushing, and contrast agent administration (Fig. 11.2). The catheter is immediately double-flushed—blood is withdrawn and discarded, after which heparinized saline flush is injected through the catheter lumen. Difficulty in blood withdrawal suggests apposition of the catheter tip to the aortic wall, which can be rectified by slight withdrawal or rotation of
the catheter until free blood aspiration is possible. The lumen of the introducing sheath should also be flushed immediately before and after each catheter insertion and every 5 minutes thereafter to prevent the encroachment of blood into the sheath. Alternatively, the side arm of the sheath may be connected to a 30 mL/hour continuous flow regulator (Intraflo II).

Once the catheter has been flushed with saline solution, tip pressure should be displayed on the physiologic monitor at all times (except during actual contrast injections). Recording this baseline pressure before contrast administration serves as an important baseline reference point. Next, the catheter lumen should be gently filled with contrast agent under fluoroscopic visualization, avoiding selective contrast administration into small branches such as the lumbar arteries if filling is performed in the descending aorta. Filling with contrast results in slight attenuation of high-frequency components in the aortic pressure waveform, whose new shape should be carefully noted. Any subsequent alteration in that waveform during coronary angiography (see damping and ventricularization, below) may signify an ostial coronary stenosis or an unfavorable catheter position within the coronary artery. Once these measures are completed, the coronary angiographic catheter is advanced into the aortic root in preparation for selective engagement of the desired coronary ostium.

A fall in overall catheter tip pressure (damping) or a fall in diastolic pressure only (ventricularization, Fig. 11.3) during catheter engagement in a coronary ostium indicates obstruction of the catheter tip or interference with coronary inflow. The catheter tip may have been inserted across a proximal coronary stenosis or may have an adverse catheter lie that places it against the coronary wall. If either
of these phenomena is observed, the catheter should be withdrawn into the aortic root immediately until the operator can analyze the situation further. The catheter may be re-engaged and a cautious small-volume contrast injection made to further clarify the situation. This may disclose a proximal occlusion of the vessel, against which the tip of the coronary catheter is resting, in which case a cine run should be performed to document this finding. The test injection may also indicate ostial stenosis with absent reflux into the aortic root or retention of the injected contrast in the proximal and mid vessel. Lack of reflux indicates that the catheter tip is severely restricting or occluding ostial inflow and mandates that only a gentle injection be performed followed by immediate removal of the catheter at the end of the cine run to restore antegrade flow. Actually, continuing to inject and film as the catheter is removed from the ostium may capture the few frames that show the ostial lesion clearly.

Another approach to evaluating such ostial lesions is to perform a nonselective injection into the sinus of Valsalva in an appropriate view (that displays the ostium of the vessel in question with no overlap by the sinus of Valsalva). Or the standard end-hole diagnostic catheter may be exchanged for an end- and side-hole angioplasty guiding catheter to overcome damping by preserving antegrade flow into the side holes, through the lumen of the catheter, and into the coronary artery, even though the catheter tip may be obstructing entry of blood into the ostium itself (see discussion of cannulation of the right coronary ostium, below). Vigorous injection despite a damped or ventricularized pressure waveform should be avoided, however, since it predisposes to ventricular fibrillation or dissection of the proximal coronary artery with major ischemic sequelae. Such a dissection is manifest by tracking of contrast down the vessel over the course of the injection and failure of contrast to clear on fluoroscopy after the injection is terminated. Prompt consideration of repair by catheter-based intervention or bypass surgery should be considered if creation of such a dye stain is associated with impeded antegrade coronary flow and signs of myocardial ischemia.

Engagement of the left coronary ostium is usually quite easy with the Judkins technique. As Judkins himself has stated, “No points are earned for coronary catheterization—the catheters know where to go if not thwarted by the operator” (23). If a left Judkins catheter with a 4-cm curve (commonly referred to as a JL4) is simply allowed to remain en face as it is advanced down into the aortic root, it will engage the left coronary ostium without further manipulation in 80 to 90% of patients (Fig. 11.4). Engagement should take place with the arm of the catheter traversing the ascending aorta at an angle of approximately 45°, the tip of the catheter in a more or less horizontal orientation, and with no change in the pressure waveform recorded from the catheter tip.

In patients with a widened aortic root owing to aortic valve disease or long-standing hypertension, the 4-cm left Judkins curve may be too short to allow successful engagement: The catheter arm may lie nearly horizontally across the aortic root with the tip pointing vertically against the roof of the left main artery, or the catheter may even refold into its packaged shape during advancement into the aortic root (Fig. 11.4D). In this case, a left Judkins catheter with a larger (JL4.5, JL5, or even JL6) curve should be selected. In the long run, changing to a larger catheter under these circumstances may end up saving time compared with trying to make an unsuitable catheter work.

In the occasional patient with a short or narrow aortic root (usually a younger female, particularly if of short stature), even the 4-cm Judkins curve may be too long. When brought down into the aortic root, the catheter arm may lie nearly vertically with the tip pointing downward below the left coronary ostium. The left ostium may still be engaged despite this somewhat unfavorable situation by pushing the catheter down into the left sinus of Valsalva for approximately 10 seconds to tighten the tip angle and then withdrawing the catheter slowly. Having the patient take a deep breath during this maneuver also helps by pulling the heart into a more vertical position to assist in engagement of the left ostium. The most satisfactory approach, however, is to exchange for a JL3.5 catheter with a shorter curve.

On rare occasions, the left coronary ostium lies out of plane (typically high and posterior), as seen in the right anterior oblique (RAO) projection where the ostium is seen to be posterior to the catheter tip. In this case, limited counterclockwise rotation of the left Judkins catheter may help orient the catheter’s tip posteriorly and facilitate engagement. Too much rotation of this catheter, however, may result in a refolded catheter that requires guidewire reinsertion to straighten. In that case, it may be helpful to step up to the next larger Judkins curve. Alternately, some operators prefer to switch to a left Amplatz shape (Fig. 11.1; (available in progressively larger curves—1, 2, 3, 4). Amplatz catheters (24) are more tolerant of rotational maneuvering and allow easy engagement of left coronary ostia that lie out of the conventional Judkins plane, as well as subselective engagement of the left anterior descending and circumflex coronary arteries in patients with short left main coronary segments or separate left coronary ostia. The left Amplatz is advanced around the arch oriented toward the left coronary ostium (Fig. 11.5). The tip of the catheter usually comes to rest in the sinus of Valsalva below the coronary ostium. As the catheter is advanced farther, however, the Amplatz shape causes the tip of the catheter to ride up the wall of the sinus until it engages the ostium. At that point,
slight withdrawal of the catheter causes deeper engagement of the coronary ostium, whereas further slight advancement causes paradoxic retraction of the catheter tip.

The Judkins technique for engaging the right coronary ostium requires slightly more catheter manipulation than cannulation of the left coronary ostium (16,23). After being flushed and filled with contrast in the descending aorta (with the catheter tip directed anteriorly to avoid injection into the intercostal arteries), the right Judkins catheter with a 4-cm curve (JR4) is brought around the aortic arch with the tip facing inward until it comes to lie against the right side of the aortic root with its tip aimed toward the left coronary ostium (Fig. 11.4). In a left anterior oblique (LAO) projection, the operator slowly and carefully rotates the catheter clockwise by nearly 180° to engage the right coronary artery. The tip of the right Judkins catheter tends to drop more deeply into the aortic root when the catheter is rotated, as the tertiary curve of the right Judkins shape aligns with the top of the aortic arch. To compensate for this effect, the operator must either begin the rotational maneuver with the tip 2 to 3 cm above the coronary ostium or withdraw the catheter slowly during rotation. Care must be taken to avoid overrotation of the catheter, which tends to cause the catheter tip to engage too deeply into the right coronary artery. To avoid this common technical error, the operator should be prepared to apply a small amount of counterclockwise torque immediately as the tip of the catheter enters the ostium. Catheters with smaller (3.5-cm) or larger (5- or 6-cm) Judkins curves or right Amplatz catheters (AR1 or AR2) may be of value if aortic root configuration and proximal right coronary anatomy make engagement difficult.

Sometimes, the right coronary ostium lies high and anterior above the commissure of the left and right aortic valve leaflets rather than in the middle of the right sinus. If it has not been possible to engage the right coronary with the approach described above, a nonselective injection should be performed into the right sinus of Valsalva. This will show the high-anterior origin and trigger a change to a left Amplatz (either AL0.75 or AL1) as required to make contact with the aortic wall at that ostium location.

Damping and ventricularization are far more common in the right coronary artery than in the left. The cause may be (a) the generally smaller caliber of the vessel (particularly in nondominant vessels; see below), (b) ostial spasm around the catheter tip, (c) selective engagement of the conus branch, or (d) true ostial stenosis. These problems in right coronary engagement can usually be elucidated by nonselective injections into the right sinus of Valsalva or cautious injections in the damped position with immediate postinjection withdrawal of the catheter. As mentioned above, a 6 or 7F angioplasty guiding catheter with side holes near the tip may be used to allow uninterrupted coronary perfusion between contrast injections, if necessitated by true ostial or proximal right coronary disease.

Despite the high initial rate of anginal relief following bypass surgery, 3 to 12% of saphenous vein grafts occlude within the first month. Additional veins occlude between 1 month and 1 year after surgery due to exaggerated neointimal hyperplasia. By far the dominant failure mode of saphenous vein graft failure beyond 1 year, however, is diffuse graft atherosclerosis, which accounts for a 50% graft closure rate by 7 years (25). Free arterial grafts (free radial or free internal mammary) are sometimes used instead of saphenous vein grafts, and these have an intermediate long-term patency between that of saphenous vein grafts and pedicled internal mammary grafts (see next page). An
increasing number of patients thus develop recurrent angina after prior bypass surgery owing to vein graft or progressive native vessel disease, and these patients account for more than 20% of the diagnostic procedures in our laboratory.

The proximal anastomosis of a vein graft or free arterial graft is usually placed on the right or left anterior aortic surface, several centimeters above the sinuses of Valsalva. Because many surgeons still resist the practice of placing radiopaque markers on the proximal graft (26), the operator must generally rely on the surgeon’s operative report or diagram, as well as knowledge of usual surgical practice in her or his own institution. The operative report always should be obtained before elective angiography on any patient with prior bypass surgery, but is absolutely essential on patients who underwent their operation at another medical center (where local preference may include practices like proximal anastomosis to the right posterior surface of the aorta, see below), or even proximal anastomosis to the descending aorta in patients with aortic root disease. It may thus be quite frustrating to embark on coronary angiography in a patient with prior bypass surgery without a detailed graft map, operative note, or prior detailed catheterization report/films in hand.

Most commonly, grafts to the left coronary arise from the left anterior surface of the aorta, with grafts to the circumflex system usually placed somewhat higher on the aorta than those to the left anterior descending or diagonal branches. Alternatively, some surgeons prefer to route grafts to the circumflex through the transverse sinus behind the heart, in which case the circumflex graft may originate from the posterior surface of the aorta. Grafts to the right coronary (or the distal portions of a dominant circumflex) usually originate from the right anterior surface of the aorta, above and somewhat behind the plane of the native right coronary ostium. We generally use the right Judkins (JR4) or Amplatz (AL1) catheter to engage anterior (i.e., left) coronary grafts. Special left coronary bypass, internal mammary, or hockey stick catheters may be required for left grafts that originate with an upward trajectory (Fig. 11.6). For downward-pointing right coronary grafts, we prefer a soft catheter with no primary curve (a multipurpose, Wexler, or JR3.5 short-tip catheter), which provides better alignment with the proximal portion of the graft and thus better opacification. The Wexler catheter can also be used for grafts originating from the left or posterior surface of the aorta. Since its tip remains in contact with the aortic wall, the shaft of this catheter can be rotated or the tip can be flexed to bring it into alignment with the proximal graft once the ostium has been engaged.

If no markers have been provided, the catheter tip should be oriented against the appropriate aortic wall and slowly advanced and then withdrawn until its tip catches in a graft ostium. The graft is injected in multiple projections that show its origin, shaft, distal anastomosis, and the native vessels beyond the anastomosis. This process must then be repeated until all graft sites have been identified. Grafts should not be written off as occluded unless a clear stump is demonstrated. If the myocardial territory supplied by a graft assumed to be occluded is still contracting, and there is no evident native or collateral blood supply to that territory, there may be a missed graft—the myocardium cannot function without a visible means of support! In that case, it may be valuable to perform an aortogram in an appropriate view to try to demonstrate flow in and locate the origin of such a missed graft. The emergence of effective therapies for focal lesions in vein grafts has placed a premium on being able to find and fix such diseased grafts before they occlude (Fig. 11.7; see Chapters 23,24 and 25).

Based on their superior demonstrated 10-year patency, the pedicled left and right internal mammary arteries (IMAs, also known as internal thoracic arteries [ITAs]) have become the conduits of choice. The proximal end of this graft remains attached to the subclavian artery (supplying
the nutritional needs of the graft itself), as the vessel is freed up from its lower sternal attachments and anastomosed to the target coronary artery (usually the left anterior descending). More than 90% of current elective bypass procedures involve placement of at least one internal mammary graft.

Successful cannulation (27) requires knowledge of the left subclavian and brachiocephalic trunk as well as the right subclavian arteries, as shown in. Fig. 11.8A. It is also important to understand some of the common anatomic variants in the internal mammary artery, including more proximal origin in the vertical portion of the subclavian, or origin as a common vessel with the thyrocervical trunk.

Although uncommon, these grafts can develop significant lesions, making it important to evaluate such grafts during any postbypass catheterization. In patients with early recurrence of angina (within the first 6 months after surgery), the most common lesion is located at the distal mammary-coronary anastomosis. It is usually due to local intimal hyperplasia rather than atherosclerosis and responds well to balloon angioplasty (see Chapter 22).
Flow-limiting kinks may also be present in the midgraft, and ostial lesions at the origin of the internal mammary from the subclavian may also occur. In patients years postbypass, significant lesions may develop in the native coronary artery beyond the internal mammary touchdown. In addition to establishing the patency of the internal mammary itself, it may also be important to look for large nonligated side branches that may divert flow from the coronary circulation and whose occlusion (in the occasional patient) may be required for angina relief (28). It is also important to look for stenoses in the subclavian artery before the takeoff of the internal mammary that may compromise the inflow to the graft and thereby cause myocardial ischemia (Fig. 11.9). Such lesions may require construction of a carotid-to-subclavian graft, or more commonly stent placement (29) to restore normal flow to the internal mammary and vertebral branches of the subclavian artery (see Chapters 14 and 26).

Although mammary grafts can be studied easily from the ipsilateral brachial approach, we prefer the femoral approach using a soft-tip preformed internal mammary catheter, which resembles a right Judkins catheter except for a tighter primary curve. This used to be a time-consuming (up to 20 minutes for some operators) process, but has been reduced to less than 3 minutes in our laboratory by adoption of a systematic strategy (see Fig. 11.8B; 27). In the LAO projection, cannulation of the left internal mammary artery begins by advancement of this catheter into the aortic arch until it lies just inside the left edge of the wedgelike density formed by the shadow of the upper mediastinum against the lung fields. With 1 to 2 cm of J guidewire protruding from its tip, the mammary catheter is rotated counterclockwise until it falls into the subclavian artery origin. From there, the wire can be advanced well out into the axillary artery. The mammary catheter is then advanced over the wire, into the midsubclavian, where the guidewire is then removed and the catheter is flushed and filled with
contrast. A low-osmolar contrast agent should be used to avoid causing CNS toxicity by reflux of hyperosmolar ionic contrast up the vertebral arteries. Switching to the straight AP projection, the catheter is rotated counterclockwise slightly (to make the tip point slightly anteriorly) as it is withdrawn slowly until the internal mammary is engaged. Intermittent gentle puffs of contrast will help localize the mammary origin during this withdrawal. Great care should be taken to avoid catheter tip trauma/dissection of the relatively delicate mammary vessel.

If selective cannulation is difficult because of tortuosity or anatomic variations, a variety of superselective or nonselective techniques can be used to permit angiographic evaluation. Nonselective injections into the subclavian will generally allow adequate opacification to see that the internal mammary is open, but generally not to provide detailed information about the distal native vessel. Inflation of a blood pressure cuff on the ipsilateral arm may help reduce runoff through the axillary artery and improve opacification of the internal mammary in cases where selective cannulation is difficult. When selective cannulation proves difficult, we sometimes attach a Y connector to the hub of the diagnostic internal mammary catheter and advance a 0.014 soft-tipped coronary angioplasty guidewire into the mammary to serve as a support for catheter advancement.

Cannulation of the right internal mammary artery may be slightly more difficult because of the need to avoid the right carotid before entering the right subclavian itself. Again in the LAO projection, the upper mediastinal wedge is identified. The mammary catheter with protruding J wire is taken to the right edge of this shadow and rotated counterclockwise until it falls into the brachiocephalic trunk. The wire is then advanced toward the right subclavian artery. Predilection for the wire to advance into the right carotid artery may require removing the guidewire and performing a nonselective contrast injection in the brachiocephalic trunk to identify the origin of the subclavian branch. The RAO-caudal projection often gives the best spatial resolution of the right carotid and right subclavian origins, after which steerable Wholey guidewire (Mallinckrodt) can be used to cannulate the subclavian. Once the wire is firmly out of the subclavian artery, the mammary catheter is advanced as described above. For cannulation of the right internal mammary artery, however, the catheter is rotated slightly clockwise during withdrawal to point its tip anteriorly.

Taken together, the left and right internal mammary arteries can be used to revascularize most lesions in the left anterior descending, proximal circumflex, and proximal right coronary arteries. Even with sequential distal anastomoses, however, the fact that there are only two internal mammary arteries means that most revascularization procedures still suffer the long-term limitations associated with the use of saphenous veins. Free segments of radial artery have also been used as bypass conduits, either from the ascending aorta (like a saphenous vein) or from the descending thoracic aorta (30) in some patients undergoing repeat bypass surgery. Although the radial artery may have slight benefit over the saphenous vein, it is prone to
spasm in the early postop period, and does not match the long-term patency record of the internal mammary artery (because it does not retain its blood supply and innervation when used as a free graft). The effort to perform all arterial bypass has brought back the right gastroepiploic artery (as an arterial pedicle graft) for anastomosis to the posterior descending or other vessels on the inferior surface of the heart (31,32). The right gastroepiploic normally supplies most of the greater curvature of the stomach, but can be dissected free from that organ and tunneled through the diaphragm to reach the inferior wall of the heart. Angiography of this vessel is possible using standard visceral angiographic catheters (e.g., Cobra) which are designed to enter visceral arteries such as the celiac axis (33). From there, the catheter can be advanced into the common hepatic (as opposed to the splenic) artery and then turned downward into the gastroduodenal artery (Fig. 11.10). A 0.025-inch Glidewire (Terumo) can then be used to cannulate the right gastroepiploic (as opposed to the superior pancreatoduodenal artery) if more selective injection is desired.

The coronary angiographer must develop a detailed familiarity with normal coronary arterial anatomy and its common variants. For individuals just learning coronary anatomy, the main coronary trunks can be considered to lie in one of two orthogonal planes (Fig. 11.14). The anterior descending and posterior descending coronary arteries lie in the plane of the interventricular septum, whereas the right and circumflex coronary trunks lie in the plane of the atrioventricular valves. In the 60° left anterior oblique (LAO) projection, one is looking down the plane of the interventricular septum, with the plane of the AV valves seen en face; in the 30° right anterior oblique (RAO) projection, one is looking down the plane of the AV valves, with the plane of the interventricular septum seen en face. The major segments and branches have each been assigned a numerical identification in the BARI modification (45) of the CASS nomenclature (Fig. 11.15).

Accurate coronary diagnosis requires coronary injections in multiple views to be sure that all coronary segments are seen clearly without foreshortening or overlap. The angulation of each view is given in two terms. The first term denotes rotation, i.e., the term right anterior oblique (RAO) designates a view where the image intensifier is located over the patient’s right anterior chest wall, and left anterior oblique (LAO) designates a view where the image intensifier is located over the patient’s left anterior chest wall. The second term denotes skew, i.e., the amount of angulation toward the patient’s head (cranial) or foot (caudal). Although the full nomenclature of skew specifies first the source of the beam and then the location of the imaging device (e.g., caudocranial, to denote that the x-ray tube is toward the patient’s feet while the image intensifier is located toward the patient’s head), in practice this is simplified to give just the location of the imaging device. The term RAO caudocranial is thus stated as RAO-cranial.

When cradle systems were in use in the 1970s, these views were usually limited to different degrees of left or right anterior obliquity in the transverse plane, including the classic 60° LAO and 30° RAO projections (see Fig. 11.14). To allow concurrent cranial angulation of the x-ray beam, cradle systems were then modified by propping the patient’s shoulders up on a foam wedge—hence the name sit-up view—to provide compound LAO-cranial projection. In the 1980s, cradle systems were abandoned in favor of parallelogram or rigid U-arm systems supported by a rotating pedestal (see Chapter 2) that allow compound beam angulation in any combination of conventional transverse (LAO, RAO) with skew (cranial, caudal) angulation up to
45°. Although these views place increased demands on the generator and increase the scattered radiation, there is no doubt that they have improved our ability to define coronary anatomy (Fig. 11.18; 52,53,54).

It is not necessary to perform all potential views in a given patient to constitute an adequate study. Rather, a series of screening views should be used as the foundation of the study, adjusted or supplemented by one or more additional views selected to more completely define suspicious areas. This requires the operator to interpret the coronary anatomy as each injection is made or by digital review—it is not acceptable to simply shoot a series of routine views and hope that the study will prove adequate when reviewed later. Although some laboratories rely on a technician to set up shots and pan the table during coronary angiography, I (DB) believe that each operator should know how to do this himself or herself to develop a good understanding of how the choice of gantry angulation influences the projected coronary anatomy. One valuable training tool in this respect is a simple wire model of the coronary anatomy, which is viewed as it is moved into different angles (Fig. 11.19; 55). A computer program that simulates the effect of changing angles on the projected coronary anatomy is provided on the companion DVD-ROM to this text book. Although there is no substitute for this type of hands-on learning, the discussion below is provided as a rough introduction.

To quantify a coronary stenosis accurately, it must be seen in profile, free from artifact related to foreshortening or obfuscation by a crossing vessel. Multiple views are important, because many lesions have a markedly eccentric (elliptical rather than round) lumen (56). When seen across its major axis, the width of the lumen may appear nearly normal, but a clue to the presence of a severe degree of narrowing in the other axis may be marked lucency caused by thinning of the contrast column. Any such suspicious lesions must be examined in a variety of other projections to reveal their true severity and to distinguish the lucency caused by eccentric stenosis from a similar lucency that may be seen adjacent to an area of denser contrast (caused by tortuosity or overlapping vessels in the absence of any true abnormality at the site) through a perceptual artifact known as the Mach effect (57).

The ability of coronary angiography to quantify the degree of stenosis at different points in the coronary circulation is fundamentally limited by the fact that it consists of a “lumen-o-gram,” in which each stenosis can be evaluated only by comparison to an adjacent reference segment that is presumed to be free of disease. In fact, both intravascular ultrasound (56; see Chapter 19) and pathologic examination (58) show that even segments that appear smooth on angiography may harbor substantial plaque. It is thus important to have a sense of the normal caliber of the major coronary arteries (4.5 ± 0.5 mm for the left main, 3.7 ± 0.4 mm for the left anterior descending, 3.4 ± 0.5 mm for a nondominant versus 4.2 ± 0.6 mm for a dominant circumflex, and 3.9 ± 0.6 mm for a dominant versus 2.8 ± 0.5 mm for a nondominant right coronary artery (59). By comparing the diameter of a presumably disease-free segment of coronary artery to the size of the diagnostic catheter (6F equals 2 mm), the operator can identify vessels that fall below these normal size ranges and may thus be diffusely diseased.

In addition to the difficulty in finding a disease-free reference segment, a major problem in the interpretation of a coronary angiogram is deciding the severity of any given stenosis. Both animal data (60) and human data (61) show that a stenosis that reduces the lumen diameter by 50% (and hence cross-sectional area by 75%) is hemodynamically significant in that it reduces the normal threefold to fourfold flow reserve of a coronary bed (Fig. 11.20), whereas a 70% diameter stenosis (90% cross-sectional area) eliminates virtually any ability to increase flow above its resting level (see Chapter 18). Stenoses that reduce the lumen diameter by 90%, however, rarely exist without reducing antegrade flow (i.e., TIMI [Thrombolysis in Myocardial Infarction] grade 1 or 2, rather than TIMI grade 3 normal flow).

Instead of the subjective TIMI flow grading system, Gibson et al. (62) have created norms for the number of cine frames (at 30 frames per second [fps]) required for contrast to leave the catheter tip and reach standardized distal landmarks in each coronary artery (the LAD “mustache,” the first posterolateral branch of the right coronary). Contrast normally reaches these points in 20 frames for the RCA and 36 frames for the LAD, with TIMI 2 (partial) flow corresponding to more than a doubling of those frame counts. Of course, even more precise data about hemodynamic lesion significance can be determined by performance of flow or pressure gradient measurements, at rest and during arteriolar vasodilation (e.g., after adenosine administration) to calculate the coronary flow or fractional flow reserve (63). Lesions that permit a flow increase of more than twofold, or have a ratio of distal pressure to aortic pressure >0.75 in the setting of peak flow after adenosine injection, are generally considered to be hemodynamically insignificant. Angiographically borderline (40 to 60%) lesions for which there is no clear objective evidence of ischemia (i.e., exercise test, perfusion scan) should thus be interrogated further with intravascular ultrasound or pressure wire measurements before considering intervention (see Chapters 18 and 19).

In clinical practice, however, the degree of lesion stenosis is usually just estimated visually from the coronary angiogram. The operator must thus develop a sense of what constitutes a 50, 70, and 90% diameter stenosis (see Fig. 11.21). Although the process of visually estimating the degree of coronary stenosis may seem straightforward, it is subject to significant operator variability (the standard deviation for repeat estimates is ≤18%: 64) as well as a systematic
form of “stenosis inflation” that causes operators to estimate a diameter stenosis that is roughly 20% higher than that measured by quantitative coronary angiography (QCA; 65). A stenosis that measures 50% will thus typically be called 70%, whereas a stenosis that measures 70% will be called 90%.

Tools are available to resolve this problem. The simplest is to project the coronary image on a wall-mounted viewing screen and to use inexpensive digital calipers (available from machinist supply houses) to measure the relative diameters of the stenotic and reference segments (66). Percent stenosis then can be calculated as 100 × [1-(stenosis diameter/reference diameter)] to provide a more accurate estimate of stenosis. This technique also reduces the standard deviation for diameter stenosis to 6 to 8% (64,66). Even greater precision can be obtained by using computer-assisted algorithms to perform automated edge detection on digitally acquired images to measure the coronary lumen with a standard deviation <5% (67,68). The amount of variation in diameter stenosis readings for one study (69) using these different methods concurrently is shown in Fig. 11.22.

The good news is that angiographers who have trained their eye in actual stenosis quantification (by using digital calipers or computer-assisted quantitative coronary angiography) can then actually give visual estimates much closer to true measurements. (70) This would allow angiographers to be more uniform in their visual estimates and move away from reporting physiologically impossible findings like a 95% stenosis with normal distal flow. Until there is a stenosis reading reform, so that those of us who call such lesions accurately (e.g., 70%) will not be accused of intervening on mild lesions, there will be no substitute to seeing the films yourself before making any clinical decisions!

It has also become important to evaluate lesion morphology more accurately from the coronary angiogram. Features such as eccentricity, ulceration, and thrombus may be associated with unstable clinical patterns (71,72), whereas features such as calcification, eccentricity, or thrombus may influence the choice of catheter intervention. Many of these features can be recognized from careful study of high-quality cineangiograms, although angiography is clearly not as sensitive to these features as intravascular ultrasound (73) or angioscopy (for thrombus or dissection; 74). Angiographers may also have trouble predicting the physiologic significance of a coronary lesion, in which case angiography may need to be supplemented by other techniques such as direct flow or distal pressure measurements (63). Finally, the absence of lesions narrowing the coronary lumen by >50% does not necessarily confer immunity from subsequent coronary events, since it is frequently a less severe stenotic lesion that
has the large lipid core and thin fibrous cap that may predispose to subsequent plaque rupture and the resulting coronary thrombosis (75). Despite these recognized limitations in quantification and morphology assessment, contrast coronary angiography remains the clinical standard by which lesions are evaluated and decisions are made regarding the need for (and best mode of providing) revascularization in the patient with ischemic heart disease.

In reviewing the coronary angiogram, one basic principle is that there should be evident blood supply to all portions of the left ventricle. Previously occluded vessel branches are usually manifest as truncated stumps, but a stump may not be evident if there has been a flush occlusion at the origin of the involved vessel. These occluded or severely stenotic vessels will frequently be seen to fill late in the injection by antegrade (so-called bridging) collaterals or collaterals that originate from the same (intracoronary) or an adjacent (intercoronary) vessel, which are reviewed in an excellent paper by Levin (76) and illustrated in Figs. 11.23,11.24 and 11.25. Finally, coronary occlusion may present in some patients simply as an angiographically arid area to which there is no evidence of either antegrade or collateral flow and no evident vascular stump. If such an area fails to show regional hypokinesis on the left ventriculogram, however, the operator should search carefully for blood supply by way of anomalous vessels or unopacified collaterals (i.e., a separate origin conus branch that was not opacified during the main right coronary injections), because the myocardium cannot continue to function normally with no visible means of support. Functioning collaterals, however, can maintain a coronary wedge pressure that averages nearly 40% of mean aortic pressure (77,78), thereby maintaining myocardial viability in the collateral-fed distribution. Along with other measures of retained or augmentable wall motion, redistributing defects on perfusion imaging, and positron emission tomographic (PET) evidence of ongoing glucose metabolism, the angiographic presence of collateral flow to an area in the distribution of an occluded coronary artery is one of the strongest evidences of ongoing myocardial viability and an important factor in determining the best revascularization strategy.

Although it is uncommon, what appears as a network of collaterals may be the vascular supply to an organized thrombus (in the left ventricle or left atrium) or a cardiac tumor. Those entities should be suspected when filling of an apparent collateral network is seen in the absence of occlusion or severe stenosis of the normal supply to a myocardial territory.

Vasospasm of an epicardial coronary artery typically presents as variant (or Prinzmetal) angina in which episodes of rest pain occur despite well-preserved effort tolerance at other times (94). An electrocardiogram recorded during an episode of spontaneous pain usually shows ST elevation in the territory supplied by the vasospastic artery. Absence of a significant coronary lesion in such a patient confirms the diagnosis of variant angina owing to focal coronary spasm (Fig. 11.28). In these patients, coronary angiography is performed mainly to look at the extent of underlying atherosclerosis (95). Provocative maneuvers to initiate spasm were once common to confirm the diagnosis and evaluate drug therapy (96). It is now used mostly when the diagnosis of variant angina is uncertain and a patient with troublesome chest pain fails to manifest sufficient disease to explain its cause.

If provocative testing for coronary spasm is contemplated, the patient should be withdrawn from calcium channel blockers for at least 24 hours and long-acting nitrates for at least 12 hours before the study and should not be premedicated with either atropine or sublingual nitroglycerin. Ongoing therapy with any of these agents may render provocative tests falsely negative (96). Although various provocative tests have been used (methacholine, epinephrine and propranolol, hyperventilation and tris-buffer, cold pressor), the most commonly used provocative agent has been ergonovine or methylergonovine maleate (Methergine, Sandoz, East Hanover, NJ; 97,98,99). These agents are stimulants of the α-adrenergic and serotonin receptors in coronary vascular smooth muscle.

Testing for coronary spasm should be performed only after baseline angiographic evaluation of both the left and
right coronary arteries. It should not be performed in patients with severe hypertension or severe anatomic cardiac pathology (left ventricular dysfunction, left main or multivessel disease, or aortic stenosis). As an example, our protocol for using methylergonovine calls for a total of 0.4 mg (400 mg equals 2 ampules) to be diluted to a total volume of 8 mL in a 10-mL syringe that is appropriately labeled. The provocative test consists of graded intravenous administration of 1 mL (0.05 mg), 2 mL (0.10 mg), and 5 mL (0.25 mg) of this mixture at 3- to 5-minute intervals. Parenteral nitroglycerin (100 to 200 mg/mL) must be premixed and loaded in a labeled syringe before the testing is begun. It is also advisable to have an intracoronary calcium channel blocker (verapamil 100 μg/mL, diltiazem 250 μg/mL) or nitroprusside (100 μg/mL) close at hand in case nitroglycerin-refractory spasm develops. Temporary pacing and defibrillator equipment should also be available to treat the bradyarrhythmias or tachyarrhythmias that sometimes accompany coronary spasm. At 1 minute before each ergonovine dose, the patient is interrogated about symptoms similar to those of her or his clinical complaint, and a 12-lead electrocardiogram is recorded. After each electrocardiogram, coronary angiography is performed, looking either at both arteries or only at the artery of highest clinical suspicion for vasospasm. In the absence of clinical symptoms, electrocardiographic changes, or focal coronary vasospasm, the next ergonovine dose is administered, and the cycle is repeated until the total dose of 0.4 mg has been given. The provocative test should be considered positive only if focal spasm (>70% diameter stenosis) occurs and is associated with clinical symptoms and/or electrocardiographic changes. Even if there are no symptoms or electrocardiographic changes, both coronary arteries should be opacified at the end of the provocative test, and any generalized vasoconstrictor effect should be terminated by administration of nitroglycerin to document the resolution of spasm and the extent of underlying atherosclerotic stenosis. Note that coronary artery spasm may occur in two vessels simultaneously (Fig. 11.29), and visualization of only one vessel may fail to adequately assess the magnitude of the vasospastic response.

Some operators have used an intracoronary methylergonovine administration protocol, in which a 4-minute intracoronary infusion (10 μg/minute in the right and 16 μg/min in the left coronary) is administered. Alternatively, discrete doses of 5 to 10 μg may be given into a coronary artery, waiting 3 minutes and imaging before a second dose is given (maximal total dose 50 μg per vessel). These intracoronary protocols may be advantageous in that they produce less systemic effect (hypertension, esophageal
spasm). The other intracoronary provocative test for coronary spasm uses acetylcholine (serial doses of 20-50-100 μg injected into the left coronary, and 20-50-80 μg injected into the right coronary). Some investigators have also used hyperventillation as a provocative test for spasm (100, Figure 11.29). The same caveats regarding ready availability of potent intracoronary vasodilators to treat spasm also apply to any of these prevocational protocols.

Several additional comments about ergonovine are in order. Ergonovine testing should be avoided in patients with severe atherosclerotic stenosis (≥80%), in whom spasm is not required to explain the clinical symptoms. In these patients, however, we frequently do repeat coronary angiography of the stenotic vessel after the intracoronary administration of 200 mg of nitroglycerin to exclude the possibility that spontaneous focal vasospasm is contributing to the appearance of severe atherosclerotic stenosis. Second, the operator should be aware that the positivity rate depends strongly on which patients are studied; the test is almost always positive in patients with known variant angina (if their disorder is active and medications have been withheld) and is positive in approximately one third of patients with clinically suspected variant angina, but it is positive in <5% of patients whose symptoms do not suggest variant angina (99). The Duke group (101) reported ergonovine testing in 3,447 patients without significant coronary disease or variant angina, with an overall positivity rate of 4% in such patients. There were two independent predictors of a positive test: mild to moderate disease on the angiogram (spasm often takes place at the point of such disease) and a history of smoking, whose presence increased the positivity rate to 10%.

Since finding spasm is so uncommon now that the syndrome is detected clinically in most patients and is treated so effectively by calcium channel blockers, the risk of ergonovine testing to evaluate patients with atypical symptoms and minimal fixed coronary disease is remarkably low. In the Duke study, significant complications occurred in only 11 patients (0.03%), including myocardial infarction in four patients and ventricular tachycardia or fibrillation (VT or VF) in seven patients (101). When provocative testing produces clinical symptoms but no angiographic evidence of vasospasm in either coronary artery, there may still be scintographic evidence of myocardial ischemia due to microvascular spasm. Both multivessel epicardial and microvascular spasm have been implicated in tako-tsubo syndrome where extreme emotional stress is followed by chest pain, ST elevation, and a particular pattern of apical hypokinesis extending beyond the usual single coronary territory. If there are no signs of myocardial ischemia, an alternative diagnosis such as esophageal dysmotility (93), which can also be provoked by methylergonovine, should be considered.

It is also important to distinguish the intense focal spasm seen in patients with variant angina from the normal mild (15 to 20%) diffuse coronary narrowing seen as a pharmacologic response to ergonovine in normal patients (102). True coronary spasm must also be distinguished from spasm induced by mechanical interventions such as rotational atherectomy (see Chapter 28) or catheter-tip spasm (Fig. 11.30). Catheter-tip spasm is most common in the right coronary artery, is not associated with clinical symptoms or electrocardiographic changes, and does not indicate variant angina (103). It should be recognized as such, however, and treated by withdrawal of the catheter, administration of nitroglycerin, and nonselective or cautious repeat selective opacification of the involved vessel to avoid mistaking catheter-tip spasm for an atherosclerotic lesion. Spasm should also be distinguished from a “pleating” artifact that may occur when a curved artery is straightened out by a stiff guidewire (Fig. 11.31), causing folds of the vessel wall to impinge on the lumen. Pleating is refractory to nitroglycerin but resolves immediately when the stiff guidewire is withdrawn (104).

Evidence has been accumulating that the patient group with angina and angiographically normal coronary arteries may contain a subgroup of patients who have myocardial ischemia on the basis of abnormal vasodilator reserve. Despite angiographic normality, intravascular
ultrasound examination may show normal vessel wall architecture, intimal thickening, or atheromatous plaque (84). In these patients, coronary blood flow (as described in Chapter 18) may fail to rise normally with pacing tachycardia or exercise, and the coronary vascular resistance is increased abnormally (105). Also, many of these patients show an abnormal rise in left ventricular end diastolic pressure following pacing tachycardia and show less lactate consumption than normal subjects in response to pacing tachycardia (106). A failure of small vessel coronary vasodilation, inappropriate vasoconstriction at the arteriolar level, or functional abnormalities of capillary endothelial cells in releasing endothelium-derived relaxing factor (EDRF) have been postulated to account for these findings. Many patients with so-called syndrome X respond at least partially to treatment with a calcium channel blocker. Disordered small vessel vasoconstriction has also been implicated in the tako-tsubo syndrome where patients with angiographically normal epicardial arteries may present with chest pain, anterior ST elevation, and a unique pattern of apical akinesis (107,108; see also Chapter 12).

There is no standard number of projections that will always provide complete information. Each major vessel must be viewed in an isolated fashion as it stands apart from other vessels. Usually, the angulated views discussed earlier in this chapter are necessary to visualize clearly the anatomy of the proximal left anterior descending and circumflex arteries.

The inexperienced operator or assistant has a tendency to hold back on the volume and force of injection into the coronary circulation. This results in inadequate or intermittent, pulsatile opacification of the coronary arterial tree as contrast flow fails short of peak coronary flow during diastole. Because there is inadequate mixing of contrast agent and blood, pockets of nonradiopaque blood in such inadequate injections may even give the appearance of arterial narrowing.

It is not uncommon to catheterize the left anterior descending or circumflex coronary artery superselectively, especially when the left main coronary artery is short and its bifurcation is early. To the inexperienced operator, this may give the impression of total occlusion of the nonvisualized vessel (e.g., if only the circumflex artery is opacified, the operator may conclude that the left anterior descending artery is occluded). If adequate filling of the noncannulated vessel cannot be achieved by reflux, selective cannulation of the LAD may be obtained by counterclockwise rotation or use of the next-smaller Judkins catheter (e.g., JL3.5), whereas selection cannulation of the circumflex may be obtained by clockwise rotation or use of the next-larger Judkins catheter (e.g., JL5). With the right coronary artery, superselective injection may occur if the catheter tip is too far down the vessel, leading to failure to visualize the conus and sinus node arteries. Because these are important sources of collateralization
of the left coronary system, important information may be missed (see Fig. 11.13). Adequate injection to give a continuous (nonpulsatile) reflux of contrast agent back into the sinus of Valsalva will help the operator to recognize vessels that originate proximally to the catheter tip and thus avoid the interpretation error of superselective injection.

Selective cannulation of a coronary artery may also fail to detect significant ostial stenosis, particularly if the catheter tip lies beyond the lesion and adequate contrast reflux is not produced. If ostial stenosis is suspected (e.g., if there is partial ventricularization or damping), we have found it helpful to perform a final injection during withdrawal of the catheter from the ostium (Fig. 11.32).

Coronary artery spasm may be related to the catheter itself, possibly caused by mechanical irritation and a myogenic reflex (see Fig. 11.30). It is seen most commonly when the right coronary artery is engaged selectively, although it may occur rarely in the left anterior descending artery as well. Although catheter-tip spasm can occur with either the brachial or femoral approach, it is probably more common with the right Judkins catheter, especially if the catheter tip enters the right coronary ostium at an angle and produces tenting of the proximal vessel. If coronary narrowing suggests the occurrence of spasm to the operator, sublingual, intravenous, or intracoronary nitroglycerin should be given and the injection repeated.

This topic has been discussed earlier in this chapter, but it bears re-emphasis. Variation in origin and distribution of the coronary artery branches may confuse the operator and cause him or her to mistakenly diagnose coronary
occlusion. For example, a small right coronary artery that terminates in the AV groove well before the crux may be interpreted as an abnormal or occluded artery, whereas it is a normal finding in 7 to 10% of human hearts. Double ostia of the right coronary artery or origin of the circumflex artery from the right coronary artery may be similarly confusing and lead to misdiagnosis.

As discussed earlier, coronary arteries occasionally dip below the epicardial surface under small strips of myocardium. During systole, the segment of the artery surrounded by myocardium is narrowed and appears as a localized stenosis. These myocardial bridges occur most commonly in the distribution of the left anterior descending artery and its diagonal branches. The key to the recognition of these bridges is that the apparent localized stenosis returns to normal during diastole. Recent studies using the flow wire (see Chapter 18) show clear derangement in phasic flow dynamics in muscle bridge segments and their normalization by stent placement. Although some severe muscle bridges can thus cause true myocardial ischemia under certain circumstances, they are seen in at least 5% of normal angiograms obtained in patients with no evidence of ischemia in the LAD territory.

If a coronary artery or branch is totally occluded at its origin, it may not be visualized, and the occlusion may be missed. If the occlusion is flush with the parent vessel, no stump will be seen. Such occlusions are primarily recognized by visualization of the distal segment of the occluded vessel by means of collateral channels or by noting the absence of the usual vascularity seen in a particular portion of the heart.