Death as a complication of diagnostic catheterization has declined progressively over the last 30 years. Whereas a 1% mortality was seen with diagnostic catheterization in the 1960s,³ the first Society for Cardiac Angiography registry of 53,581 diagnostic catheterizations performed in 1979-1981 showed a 0.14% procedure-related mortality.⁴ By the second registry of 222,553 patients catheterized in 1984-1987,⁵ procedure-related mortality for diagnostic catheterization had fallen further, to 0.1% (i.e., 1 in 1,000). This small reduction in mortality, however, belies the fact that the second registry included many more patients who fell into a high-risk subgroup for the procedure. Based on variables identified from the 218 deaths in the second registry (age older than 60 years, New York Heart Association (NYHA) functional class IV, left ventricular ejection fraction <30%, or left main disease), the mortality for such patients fell by half between the first and second registry.⁶ A third registry of 58,332 patients studied in 1990 showed an even lower overall mortality of 0.08%, with a 1.5% incidence of any major complication.⁷ A number of baseline variables (including NYHA class, multivessel disease, congestive heart failure, and renal insufficiency) were identified in this registry, whose presence predicted an up to eightfold increase in major complication rates (from 0.3% in patients with none of these factors to 2.5%).⁸ Several of the major factors are discussed below.

Although there has been a progressive reduction in the overall mortality of diagnostic cardiac catheterization over the last 25 years, patients with severe left main coronary disease remain at increased risk. Their mortality was 6% in the 1976 report by Bourassa,⁹ and 2.8% in the study by Boehrer and others performed between 1978 and 1992 (compared with a mortality of 0.13% in patients without such disease).¹⁰ Although the mortality of such patients had fallen to 0.86% in the first Society for Cardiac Angiography registry, this was still more than 20 times higher than the 0.03% mortality seen in patients with single-vessel disease.⁴

Because roughly 7% of patients undergoing coronary angiography have significant left main disease, the protocol used for coronary angiography (see Chapter 15) should always begin with careful catheter entry into the left coronary ostium to facilitate early recognition of ostial left main disease through catheter pressure damping or performance of a test “puff” immediately after engagement. Even without these early warnings of left main disease, we routinely perform the first left coronary injection in the right anterior oblique (RAO) projection with caudal angulation to screen for mid and distal left main disease and get the maximal anatomic information on the first injection. If ostial left main stenosis is suspected, a straight anterior (AP) injection may be performed. If severe left main disease is present, the only other left coronary injection needed is an RAO projection with cranial angulation (to see the left anterior descending and its diagonal branches). If angiography shows a borderline lesion (30% to 50%), additional diagnostics including intravascular ultrasound (IVUS; Chapter 25) or pressure wire (Chapter 24) can be performed after completing diagnostic coronary angiography to inform the subsequent management decision. However, performing a large number of superfluous contrast injections in a patient with a critical left main disease offers little more in the way of important anatomical information and increases the risk of triggering the vicious cycle of ischemia/hypotension/more ischemia that may lead to irreversible collapse.

Careful attention to all other aspects of technique is essential, since even an otherwise minor complication (e.g., a vasovagal reaction or arrhythmia) may have fatal consequences in this situation. If a patient with severe left main disease exhibits any significant instability during the procedure, we usually opt to place an intraaortic balloon pump (see Chapter 27) and arrange for prompt coronary artery bypass graft surgery. When the hemodynamics are markedly compromised and the patient is a poor surgical candidate, emergency coronary stenting can be performed if a trained operator and the necessary equipment are available (see Chapters 27 and 28). A similar consideration regarding the use of hemodynamic support applies to any patient with an unstable ischemic syndrome or acute myocardial infarction who behaves in a brittle fashion under the stresses of catheter placement and contrast injection.

Patients with cardiogenic shock in the setting of acute myocardial infarction or severe chronic left ventricular dysfunction (ejection fraction <30%) also have a several-fold increased risk of procedural morbidity and mortality,⁵ particularly when reduction in ejection fraction is associated with a baseline pulmonary capillary wedge pressure >25 mmHg and a systolic arterial pressure <100 mmHg. An effort should be made to bring such congestive heart failure under control before cardiac catheterization is attempted.

Although right heart catheterization is no longer routine (see Chapter 6), it should be performed before angiography in a patient with poor ejection fraction, because it provides valuable data about baseline hemodynamic status and allows ongoing monitoring of pulmonary artery pressure as an early warning about hemodynamic decompensation before frank pulmonary edema ensues. If the baseline pulmonary capillary wedge pressure is >30 mmHg, every effort should be made to improve hemodynamic status before angiography is attempted. This may entail administration of a potent intravenous diuretic (furosemide), supplemental oxygen, a vasodilator (intravenous nitroglycerine or sodium nitroprusside) when the mean arterial pressure is >65 mmHg, or a positive inotrope (dobutamine, milrinone) when the mean
arterial pressure is <65 mmHg, or when severe congestive heart failure hemodynamics persist despite vasodilator treatment (see below). When frank cardiogenic shock is present or develops during a cardiac catheterization, prompt placement of an intraaortic balloon pump or of alternative percutaneous left ventricular assist devices in the contralateral groin may be required to get the patient safely though the procedure (see Chapter 27). Low-osmolar contrast agents produce less myocardial depression than traditional high-osmolar agents. They have replaced high-osmolar contrast agents for the performance of coronary and vascular angiography and they have greatly enhanced our ability to perform necessary angiography without precipitating hemodynamic decompensation in such unstable patients (see Chapter 2).

Despite the preponderance of coronary artery disease as the indication for diagnostic cardiac catheterization patients with severe valvular heart disease are also at increased risk for dying during cardiac catheterization. The VA Cooperative Study on Valvular Heart Disease¹¹ showed a 0.2% mortality among 1,559 preoperative catheterizations performed in patients with valvular heart disease, with one death in a patient with mitral regurgitation and two deaths in patients with aortic stenosis. With current noninvasive methods for assessing the severity of valvular lesions, there has been debate about whether it is necessary to cross severely stenotic valves in the course of preoperative cardiac catheterization.¹² According to the most recent ACC/AHA update to the 2006 Guidelines for the Management of Patients with Valvular Heart Disease, cardiac catheterization is not recommended for the assessment of severity of aortic stenosis before aortic valve replacement “when noninvasive tests are adequate and concordant with clinical findings (Level of Evidence: C).”¹³

Patients who have previously undergone coronary bypass surgery make up a growing subgroup of diagnostic and interventional catheterizations. They are typically 5 years older, have more diffuse coronary and generalized atherosclerosis, worse left ventricular function, and require a lengthier and more complex procedure to image both native coronary arteries and all grafts. Despite these adverse risk factors, the Post CABG Trial¹⁴ looked at 2,635 diagnostic angiograms performed in stable patients and found 0% mortality, with major complications in 0.7% (myocardial infarction 0.08%, stroke 0.19%, vascular trauma requiring transfusion or surgery 0.4%).

Pediatric patients may be at higher risk (see Chapter 9). One review of 4,952 patients (median age 2.9 years) studied at the

Hospital for Sick Children in Toronto¹⁵ found a mortality of 1.2% confined to patients younger than age 5 years (half in critically ill neonates <30 days of age). Although the risk was lower for diagnostic than for electrophysiologic or interventional procedures, there were three deaths (0.1%) among the 3,149 diagnostic procedures.

Because they involve the use of more aggressive catheters, superselective cannulation of diseased coronary arteries, and brief interruption of coronary or even systemic flow (see Chapter 28), interventional procedures tend to carry higher mortality than purely diagnostic catheterizations. In the first 1,500-patient coronary angioplasty registry sponsored by the National Heart, Lung, and Blood Institute (NHLBI) from 1979 to 1982, the mortality of elective angioplasty was 1.1%.¹⁶ This was relatively unchanged at 1.0% in the second NHLBI registry of 1,802 patients treated at 15 centers between 1984 and 1987, mainly because the second registry included more patients with adverse features (advanced age, poor ventricular function, multivessel disease, prior bypass surgery, and so on).¹⁷ In fact, the mortality for single-vessel procedures fell from 1.3% to 0.2% between the first and second registry.

With the introduction of newer devices (e.g., stents, atherectomy, laser) to treat high-risk lesions preemptively or reverse abrupt closure following attempted conventional balloon angioplasty, the overall mortality for elective coronary intervention has fallen further (Chapter 28), but the extension of intervention to other high-risk subsets, including patients with acute myocardial infarction undergoing primary angioplasty, has kept overall mortality close to 1%¹⁸,¹⁹— roughly 10-fold higher than purely diagnostic catheterization (i.e., 1% vs. 0.1%). Several multivariable models that predict procedural mortality have been developed based on age, ejection fraction, treatment for acute myocardial infarction/shock, urgent/emergent priority, and so on¹⁸, ¹⁹, ²⁰, ²¹, ²², ²³, ²⁴ (see also Chapter 28 and 30). In general, there is a wide variation in risk of death in the course of coronary intervention—based on patient comorbidities, clinical indication, and procedure type. While an “average” risk can be quoted to “average” patients, patients with one or more adverse risk factors should be told candidly during the informed consent process that their expected risks are higher than these averages (Figures 4.1 and 4.2). A similar approach can be used when discussing prognosis following PCI for acute myocardial infarction or cardiogenic shock.

Coronary interventions may produce myocardial infarction by a variety of mechanisms that include dissection, abrupt vessel closure, “snowplow” occlusion of side branches, spasm of the epicardial or arteriolar vessels (no reflow), thrombosis, or distal embolization (see Chapters 28 and 29). Q-wave myocardial infarction was reported in 4.8% of patients in the first NHLBI registry and 3.6% in the second NHLBI registry.¹⁷ This includes roughly half of the 6% of percutaneous transluminal coronary angioplasty (PTCA) patients who were sent for emergency bypass surgery owing to abrupt vessel closure.¹⁶ Over the last decade, experience with coronary stenting has led to marked reduction in the need for emergency bypass surgery (to roughly 0.2%; see Chapters 28 and 31), and accordingly the incidence of Q-wave infarction has fallen to <1%.

Largely as a spin-off of trials conducted with the platelet glycoprotein IIb/IIIa receptor blockers in the mid-1990s, however, the official definition of periprocedural myocardial infarction has now been broadened to include non-Q-wave infarctions (more properly called non-ST-elevation myocardial infarctions) detected by elevation of cardiac biomarkers above the 99th percentile of upper limit of normal.²⁶ According to the updated universal definition of myocardial infarction,²⁷ PCI-related myocardial infarction (type 4a) is defined as a cardiac CK-MB or troponin elevation of >3× the upper reference limit in patients with normal baseline levels (see Chapter 28 for a more extensive discussion). Patients with these low-level enzyme elevations are more likely to have some degree of chest discomfort due to occlusion of small side branches or distal microembolization²⁸,²⁹ (Figure 4.3), but this finding is common also in patients without enzyme elevation, where it presumably represents stimulation of adventitial pain receptors by local stretching at the treatment site.³⁰

Several studies have evaluated the relationship between elevations of CK-MB and long-term mortality. Although
elevation above five or eight times normal corresponds to a significant amount of myocardial necrosis and carries the same adverse impact on long-term prognosis as a Q-wave infarction (Figure 4.4A), long-term follow-up of patients from several multicenter trials³² has shown that patients with even low-level (one to three times normal) elevation of postprocedural CK after PCI have a greater incidence of late adverse outcomes (Figure 4.4B). Similar results have been shown by analysis evaluating the relationship between troponin elevations and long-term mortality.³³,³⁴ Whether any such relationship is cause and effect, or simply an association of both periprocedural cardiac biomarker elevation and late events with a common confounding variable (such as the diffuse underlying atherosclerosis) remains to be determined.³⁵,³⁶ However, it has also been suggested that due to the higher sensitivity of troponin, a larger percentage of patients will meet the definition of 4a myocardial infarction when troponin is used when compared to the same patients when CK-MB is used. Some of these patients will not have any evidence of myocardial necrosis even when a very sensitive imaging modality such as contrast-enhanced CMR is used.³⁷

Femoral artery thrombosis can occur in patients with a small common femoral artery lumen (peripheral vascular disease, diabetes, female gender), in whom a large-diameter catheter or sheath (e.g., an intraaortic balloon pump) has been placed, particularly when the catheter dwell time is long or when prolonged postprocedure compression is applied. Such patients have a white painful leg with impaired distal sensory and motor function, as well as absent distal pulses. If this develops during the catheterization procedure and is not corrected promptly by sheath removal, a flow-obstructing dissection or thrombus at the femoral artery puncture site or a distal arterial embolus should be suspected. This requires urgent attention via vascular surgery consultation (for exploration and correction of any local dissection or plaque avulsion and Fogarty embolectomy of the distal vessel as needed to restore distal pulses). Alternatively, operators skilled in peripheral intervention may be able to puncture the contralateral femoral artery, cross over the aortic bifurcation, and address a common femoral occlusion percutaneously⁴⁹ (Figure 4.5). Either way, failure to restore limb flow within 2 to 6 hours may result in extension of thrombosis into smaller distal branches, with muscle necrosis requiring fasciotomy or even amputation, and predispose to the development of renal failure.

Femoral venous thrombosis and pulmonary embolism are rare complications of diagnostic femoral catheterization (Figure 4.6). A small number of clinical cases have been reported, however, particularly in the setting of venous compression by a large arterial hematoma, sustained mechanical compression (see Chapter 6), or prolonged procedures with multiple venous lines (e.g., electrophysiologic studies).⁵⁰ However, the actual incidence of thrombotic and pulmonary embolic complications may be substantially under-reported, since most are not evident clinically. Asymptomatic lung scan abnormalities have thus been described in up to 10% of patients after diagnostic catheterization.⁵¹

Although thrombotic complications do occur, poorly controlled bleeding from the arterial puncture site is a more common problem after cardiac catheterization by the femoral approach.⁵² Uncontrollable free bleeding around the sheath suggests laceration of the femoral artery. If such free bleeding does not respond to replacement with the next larger diameter sheath, the bleeding should be controlled by manual compression around the sheath until the procedure is completed. Anticoagulation may be reversed, and an attempt made to remove the sheath and control bleeding with prolonged (30- to 60-minute) compression or to place a femoral closure device (see Chapter 6). The vascular surgeons should be consulted regarding operative repair should the bleeding continue.

Formation of a hematoma—a collection of blood within the soft tissues of the upper thigh—is more common than free bleeding. It tends to cause a tender mass the size of a baseball or softball. If ongoing bleeding stops with manual compression, the hematoma will usually resolve over 1 to 2 weeks as the blood gradually spreads and is reabsorbed from the soft tissues. Larger hematomas may require transfusion, but surgical repair of a hematoma (as opposed to a false aneurysm, see later section) is generally not required. Given the discomfort caused by large hematomas, and the potential of such hematomas to evolve into false aneurysms, accurate puncture and puncture site compression or closure technique to minimize hematoma formation are essential parts of good catheterization technique.

The level of anticoagulation and antiplatelet therapy as well as increased sheath size, female gender, and advanced age all increase the risk of hemorrhagic complications. During the era (1990-1996) when uninterrupted transition from intravenous heparin to oral warfarin was used for stenting, vascular complications were as high as 10%.⁵³ Even before the switch to less aggressive anticoagulant protocols (aspirin and ticlopidine or clopidogrel), second-generation stents that permit use of smaller 6F sheaths, and the widespread use of punctureclosure devices, the incidence of hemorrhagic access site complications after stenting remains at 1% to 2% (see Chapter 6). The tendency of platelet glycoprotein IIb/IIIa blockers to increase local hemorrhagic complications, however, has been tempered by lower levels of heparinization and the growing use of bivalirudin as an alternative antithrombotic agent associated with a lower risk of bleeding.⁵⁴ Various approaches for collagen plugging or percutaneous suture-mediated closure of the femoral arterial puncture site have been introduced in the last several years (see Chapter 6). Although these devices avoid the discomfort of prolonged manual or mechanical compression and allow earlier or even immediate ambulation, clinical trials have failed to demonstrate significant reduction of major vascular complications compared with those caused by compression.⁵⁵ It is possible that this class of devices will
improve sufficiently to make closure of the femoral artery puncture site so reliable as to eliminate the 1% to 2% incidence of complication. Until that time, operators must be prepared to recognize and repair them when they occur or to work from other access sites such as the radial artery (where hemorrhagic complications are unheard of, and thrombosis [with a negative Allen test] is usually inconsequential) (see Chapter 7), in patients at high risk for a femoral complication.