Premature ventricular complexes (PVCs) are a common cause of palpitations, and are also often detected incidentally on electrocardiography (ECG), ambulatory monitoring, or inpatient telemetry. At the cellular level, ventricular myocytes spontaneously depolarize to create an extra systole that is “out of sync” with the cardiac cycle.

Although nearly everyone has some PVCs from time to time, people vary widely in their frequency of PVCs and their sensitivity to them.^1,2 Some patients are exquisitely sensitive to even a small number of PVCs, while others are completely unaware of PVCs in a bigeminal pattern (ie, every other heartbeat). This article will review the evaluation and management of PVCs with a focus on clinical aspects.

DIAGNOSTIC EVALUATION

Personal and family history

Symptoms. The initial history should establish the presence, extent, timing, and duration of symptoms. Patients may use the word “palpitations” to describe their symptoms, but they also describe them as “hard” heartbeats, “chest-thumping,” or as a “catch” or “skipped” heartbeat. Related symptoms may include difficulty breathing, chest pain, fatigue, and dizziness.

The interview should determine whether the symptoms represent a minor nuisance or a major quality-of-life issue to the patient, and whether there are any specific associations or triggers. For example, it is very common for patients to become aware of PVCs at night, particularly in certain positions, such as lying on the left side. Patients often associate PVC symptoms with emotional stress, exercise, or caffeine or stimulant use.

Medication use. An accurate and up-to-date list of prescription medications should be screened for alpha-, beta-, or dopamine-receptor agonist drugs. Similarly, any use of over-the-counter sympathomimetic medications and nonprescription supplements should be elicited, including compounded elixirs or beverages. Many commercially available products designed to treat fatigue or increase alertness contain large doses of caffeine or other stimulants. It is also important to consider the use of illicit substances such as cocaine, amphetamine, methamphetamine, and their derivatives.

The patient’s medical and surgical history should be queried for any known structural heart disease, including coronary artery disease, myocardial infarction, congestive heart failure, valvular heart disease, congenital heart disease, and heritable conditions such as hypertrophic cardiomyopathy, prolonged QT syndromes, or other channel disorders. Pulmonary disorders such as sarcoidosis, pulmonary hypertension, or obstructive sleep apnea are also relevant. Similarly, it is important to identify endocrine disorders, including thyroid problems, sex hormone abnormalities, or adrenal gland conditions.

A careful family history should include any instance of sudden death in first-degree relatives, any heritable cardiac conditions, or coronary artery disease at an early age.

Physical examination

The physical examination should focus on findings that suggest underlying structural heart disease. Findings suggestive of congestive heart failure include elevated jugular venous pressures, abnormal cardiac sounds, pulmonary rales, abnormal arterial pulses, or peripheral edema. A murmur or a pathologic heart sound should raise suspicion of valvular or congenital heart disease when present in a young patient.

Inspection and palpation of the thyroid can reveal a related disorder. Obvious skin changes or neurologic findings can similarly reveal a systemic and possibly related clinical disorder that can have cardiac manifestations (eg, muscular dystrophy).

Electrocardiography, Holter monitoring, and other monitoring

Assessment of the cardiac rhythm includes 12-lead ECG and ambulatory Holter monitoring, typically for 24 or 48 hours.

Holter monitoring provides a continuous recording, usually in at least two or three leads. Patients are given a symptom journal or are asked to keep a diary of symptoms experienced during the monitoring period. The monitor is worn underneath clothing and is returned for download upon completion. Technicians process the data with the aid of computer software, and the final output is reviewed and interpreted by a cardiologist or cardiac electrophysiologist.

Holter monitoring for at least 24 hours is a critical step in assessing any patient with known or suspected PVCs, as it can both quantify the total burden of ventricular ectopy and identify the presence of any related ventricular tachycardia. In addition, it can detect additional supraventricular arrhythmias or bradycardia during the monitoring period. The PVC burden is an important measurement; it is expressed as the percentage of heartbeats that were ventricular extrasystoles during the monitoring period.

Both ECG and Holter monitoring are limited in that they are only snapshots of the rhythm during the period when a patient is actually hooked up. Many patients experience PVCs in clusters every very few days or weeks. Such a pattern is unlikely to be detected by a single ECG or 24- or 48-hour Holter monitoring.

A 30-day ambulatory event monitor (also known as a wearable loop recorder) is an important diagnostic tool in these scenarios. The concept is very similar to that of Holter monitoring, except that the device provides a continuous loop recording of the cardiac rhythm that is digitally stored in clips when the patient activates the device. Some wearable loop recorders also have auto-save features for heart rates falling outside of a programmed range.

Mobile outpatient cardiac telemetry is the most comprehensive form of noninvasive rhythm monitoring available. This is essentially the equivalent of continuous inpatient cardiac telemetry, but in a patient who is not hospitalized. It is a wearable ambulatory device providing continuous recordings, real-time automatic detections, and patient-activated symptom recordings. It can be used for up to 6 weeks. Advantages include detection and quantification of asymptomatic events, and real-time transmissions that the physician can act upon. The major disadvantage is cost, including coverage denial by many third-party payers.

This test is rarely indicated as part of a PVC evaluation and is typically ordered only by a cardiologist or cardiac electrophysiologist.

Surface echocardiography is indicated to look for overt structural heart disease and can reliably detect abnormalities in cardiac chamber size, wall thickness, and function. Valvular heart disease is concomitantly identified by two-dimensional imaging as well as by color Doppler. The finding of significant structural heart disease in conjunction with PVCs should prompt a cardiology referral, as this carries significant prognostic implications.^3–5

Exercise treadmill stress testing is appropriate for patients who experience PVCs with exercise or for whom an evaluation for coronary artery disease is indicated. The expected finding would be an increase in PVCs or ventricular tachycardia with exercise or in the subsequent recovery period. Exercise testing can be combined with either echocardiographic or nuclear perfusion imaging to evaluate the possibility of myocardial ischemia. For patients unable to exercise, pharmacologic stress testing with dobutamine or a vasodilator agent can be performed.

Advanced noninvasive cardiac imaging— such as computed tomography, magnetic resonance imaging, or positron-emission tomography—should be reserved for specific clinical indications such as congenital heart disease, suspected cardiac sarcoidosis, and infiltrative heart disease, and for specific cardiomyopathies, such as hypertrophic cardiomyopathy and arrhythmogenic right ventricular cardiomyopathy. For example, frequent PVCs with a left bundle branch block morphology and superior axis raise the concern for a right ventricular disorder and may prompt cardiac magnetic resonance imaging for either arrhythmogenic right ventricular cardiomyopathy or sarcoidosis.

PVCs WITHOUT STRUCTURAL HEART DISEASE

Outflow tract PVCs and ventricular tachycardia

The right or left ventricular outflow tracts, or the epicardial tissue immediately adjacent to the aortic sinuses of Valsalva are the most common sites of origin for ventricular ectopy in the absence of structural heart disease.^6–9 Affected cells often demonstrate a triggered activity mechanism due to cyclic adenosine monophosphate-mediated and calcium-dependent delayed after-depolarizations.^7,8

Most of these foci are in the right ventricular outflow tract, producing a left bundle branch block morphology with an inferior axis (positive R waves in limb leads II, III, and aVF) and typical precordial R-wave transition in V₃ and V₄ (Figure 1). A minority are in the left ventricular outflow tract, producing a right bundle branch block with an inferior axis pattern, or in the aortic sinuses with a left bundle branch block pattern but with early precordial R transition in V₂ and V₃.

A study in 122 patients showed that right and left outflow tract arrhythmias had similar electrophysiologic properties and pharmacologic sensitivities, providing evidence for shared mechanisms possibly due to the common embryologic origin of these structures.⁹

Such arrhythmias are typically catecholamine-sensitive and are sometimes inducible with burst pacing in the electrophysiology laboratory. The short ventricular coupling intervals can promote intracellular calcium overload in the affected cells, leading to triggered activity.

Therefore, outflow tract PVCs and ventricular tachycardia are commonly encountered clinically during exercise and, to an even greater extent, in the postexercise cool-down period. Similarly, they can be worse during periods of emotional stress or fatigue, when the body’s endogenous catecholamine production is elevated. However, it is worthwhile to note that there are exceptions to this principle in which faster sinus rates seem to overdrive the PVCs in some patients, causing them to become paradoxically more frequent at rest, or even during sleep.

Outflow tract PVCs can be managed medically with beta-blockers, nondihydropyridine calcium channel blockers (verapamil or diltiazem), or, less commonly, class IC drugs such as flecainide. They are also highly curable by catheter ablation (Figure 2), with procedure success rates greater than 90%.^9.10

However, a subset of outflow tract PVCs nested deep in a triangle of epicardial tissue between the right and left endocardial surface and underneath the left main coronary artery can be challenging. This region has been labeled the left ventricular summit, and is shielded from ablation by an epicardial fat pad in the adjacent pericardial space.¹¹ Ablation attempts made from the right and left endocardial surfaces as well as the epicardial surface (pericardial space) sometimes cannot adequately penetrate the tissue deep enough to reach the originating focus deep within this triangle. While ablation cannot always fully eliminate the PVC, ablation from more than one of the sites listed can generally reduce its burden, often in combination with suppressive medical therapy (Figure 3).

Fascicular PVCs

Fascicular PVCs originate from within the left ventricular His-Purkinje system¹² and produce a right bundle branch block morphology with either an anterior or posterior hemiblock pattern (Figure 4). Exit from the posterior fascicle causes an anterior hemiblock pattern, and exit from the anterior fascicle a posterior hemiblock pattern. Utilization of the rapidly conducting His-Purkinje system gives these PVCs a very narrow QRS duration, sometimes approaching 120 milliseconds or shorter. This occasionally causes them to be mistaken for aberrantly conducted supraventricular beats. Such spontaneous PVCs are commonly associated with both sustained and nonsustained ventricular tachycardia and are usually sensitive to verapamil.¹³

Special issues relating to mapping and catheter ablation of fascicular arrhythmias involve the identification of Purkinje fiber potentials and associated procedural diagnostic maneuvers during tachycardia.¹⁴

Other sites for PVCs

Other sites of origin for PVCs in the absence of structural heart disease include ventricular tissue adjacent to the aortomitral continuity,¹⁵ the tricuspid annulus,¹⁶ the mitral valve annulus, ¹⁷ papillary muscles,¹⁸ and other Purkinje-adjacent structures such as left ventricular false tendons.¹⁹ An example of a papillary muscle PVC is shown in Figures 5 and 6.

Curable by catheter ablation

Any of these PVCs can potentially be cured by catheter ablation when present at a sufficient burden to allow for activation mapping in the electrophysiology laboratory. The threshold for offering ablation varies among operators, but is generally around 10% or greater. Pacemapping is a technique applied in the electrophysiology laboratory when medically refractory symptomatic PVCs occurring at a lower burden require ablation.

PVCs WITH AN UNDERLYING CARDIAC CONDITION

Coronary artery disease

Tissue injury and death caused by acute myocardial infarction has long been recognized as a common cause of spontaneous ventricular ectopy attributed to infarct border zones of ischemic or hibernating myocardium.^20,21

Suppression has not been associated with improved outcomes, as shown for class IC drugs in the landmark Cardiac Arrhythmia Suppression Trial (CAST),²² or in the amiodarone treatment arm of the Multicenter Automatic Defibrillator Implantation Trial II (MADIT-II).²³ Therefore, treatment of ventricular ectopy in this patient population is usually symptom-driven unless there is hemodynamic intolerance, tachycardia-related cardiomyopathy, or a very high burden of PVCs in a patient who may be at risk of developing tachycardia-related cardiomyopathy. Antiarrhythmic drug treatment, when required, usually involves beta-blockers or class III medications such as sotalol or amiodarone.

Nonischemic dilated cardiomyopathy

This category includes patients with a wide variety of disease states including valvular heart disease, lymphocytic and other viral myocarditis, cardiac sarcoidosis, amyloidosis and other infiltrative diseases, familial conditions, and idiopathic dilated cardiomyopathy (ie, etiology unknown). Although it is a heterogeneous group, a common theme is that PVCs in this patient cohort may require epicardial mapping and ablation.²⁴ Similarly, epicardial PVCs and ventricular tachycardia cluster at the basal posterolateral left ventricle near the mitral annulus, for unclear reasons.²⁵

While specific criteria have been published, an epicardial focus is suggested by slowing of the initial QRS segment, pseudo-delta waves, a wider overall QRS, and Q waves in limb lead I.²⁶

Treatment is symptom-driven unless the patient has a tachycardia-related cardiomyopathy or a high burden associated with the risk for its development. Antiarrhythmic drug therapy, when required, typically involves a beta-blocker or a class III drug such as sotalol or amiodarone. Sotalol is used in this population but has limited safety data and should be used cautiously in patients without an implantable cardioverter-defibrillator.

Spontaneous ventricular ectopy and tachycardia are common, if not expected, in patients with this heritable autosomal dominant disorder. This condition is progressive and associated with the risk of sudden cardiac death. Criteria for diagnosis were established in 2010, and patients with suspected arrhythmogenic right ventricular cardiomyopathy often undergo cardiac magnetic resonance imaging.²⁷ Diagnostic findings include fibro-fatty tissue replacement, which usually starts in the right ventricle but can progress to involve the left ventricle. PVCs and ventricular tachycardia can involve the right ventricular free wall and are often epicardial.

Catheter ablation is usually palliative, as future arrhythmias are expected. Many patients with this condition require an implantable cardioverter-defibrillator for prevention of sudden cardiac death, and some go on to cardiac transplantation as the disease progresses and ventricular arrhythmias become incessant.

Other conditions

Spontaneous ventricular ectopy is common in other heritable and acquired cardiomyopathies including hypertrophic cardiomyopathy and in infiltrative or inflammatory disorders such as cardiac amyloidosis and sarcoidosis. While technically falling under the rubric of nonischemic heart disease, the presence of spontaneous ventricular ectopy carries specific prognostic implications depending on the underlying diagnosis. Therefore, an appropriate referral for complete cardiac evaluation should be considered when a heritable disorder or other acquired structural heart disease is suspected.

TACHYCARDIA-RELATED CARDIOMYOPATHY

Tachycardia-related cardiomyopathy refers to left ventricular systolic dysfunction that is primarily caused by arrhythmias. This includes frequent PVCs or ventricular tachycardia but also atrial arrhythmias occurring at a high burden that directly weaken myocardial function over time. Although much research has been devoted to this condition, our understanding of its etiology and pathology is incomplete.

PVCs and ventricular ectopy burdens in excess of 15% to 20% have been associated with the development of this condition.^28,29 However, it is important to note that cardiomyopathy can also develop at lower burdens.³⁰ One study found that a burden greater than 24% was 79% sensitive and 78% specific for development of tachycardia-related cardiomyopathy.³¹ Additional studies have demonstrated specific PVC morphologic features such as slurring in the initial QRS segment and also PVCs occurring at shorter coupling intervals as being associated with cardiomyopathy.^32–34

For these reasons, both quantification of the total burden and careful evaluation of available electrocardiograms and rhythm strips are important even in asymptomatic patients with frequent PVCs. Similarly, unexplained left ventricular dysfunction in patients with PVC burdens in these discussed ranges should raise suspicion for this diagnosis. Patients with tachycardia-related cardiomyopathy usually have at least partially reversible left ventricular dysfunction when identified or treated early.^29,35

MEDICAL AND ABLATIVE TREATMENT

Available treatments include medical suppression and catheter ablation. One needs to exercise clinical judgment and incorporate all of the PVC-related data to make treatment decisions.

Little data for trigger avoidance and behavioral modification

Some patients report a strong association between palpitations related to PVCs and caffeine intake, other stimulants, or other dietary triggers. However, few data exist to support the role of trigger avoidance and behavioral modification in treatment. In fact, an older randomized trial in 81 men found no benefit in a program of total abstinence from caffeine and smoking, moderation of alcohol intake, and physical conditioning.³⁶

Nonetheless, some argue in favor of advising patients to make these dietary and lifestyle changes, given the overall health benefits of aggressive risk-factor modification for cardiovascular disease.³⁷ Certainly, a trial of trigger avoidance and behavioral modification seems reasonable for patients who have strongly associated historical triggers in the absence of structural heart disease and PVCs occurring at a low to modest burden.

Beta-blockers are the mainstay

Beta-blockers are the mainstay of medical suppression of PVCs, primarily through their effect on beta-1 adrenergic receptors to reduce intracellular cyclic adenosine monophosphate and thus decrease automaticity. Blocking beta-1 receptors also causes a negative chronotropic effect, reducing the resting sinus rate in addition to slowing atrioventricular nodal conduction.

Cardioselective beta-blockers include atenolol, betaxolol, metoprolol, and nadolol. These drugs are effective in suppressing PVCs, or at least in reducing the burden to more tolerable levels.

Beta-blockers are most strongly indicated in patients who require PVC suppression and who have concomitant coronary artery disease, prior myocardial infarction, or other cardiomyopathy, as this drug class favorably affects long-term prognosis in these conditions.

Common side effects of beta-blockers include fatigue, shortness of breath, depressed mood, and loss of libido. Side effects can present a significant challenge, particularly for younger patients. Noncardioselective beta-blockers are less commonly prescribed, with the exception of propranolol, which is an effective sympatholytic drug that blocks both beta-1 and beta-2 receptors.

Many patients with asthma or peripheral arterial disease can tolerate these drugs well despite concerns about provoked bronchospasm or claudication, respectively, and neither of these conditions is considered an absolute contraindication. Excessive bradycardia with beta-blocker therapy can lead to dizziness, lightheadedness, or overt syncope, and these drugs should be used with caution in patients with baseline sinus node dysfunction or atrioventricular nodal disease.

Nondihydropyridine calcium channel blockers are particularly effective for PVC suppression in patients without structural heart disease by the mechanisms previously described involving intracellular calcium channels. In particular, they are highly effective and are considered the drugs of choice in treating fascicular PVCs.

Verapamil is a potent drug in this class, but it also commonly causes constipation as a side effect. Diltiazem is less constipating but can cause fatigue, drowsiness, and headaches. Both drugs reduce the resting heart rate and slow atrioventricular nodal conduction. Patients predisposed to bradycardia or atrioventricular block can develop dizziness or overt syncope. Calcium channel blockers are also used cautiously in patients with congestive heart failure, given their potential negative inotropic effects.

Overall, calcium channel blockers are a very reasonable choice for young patients without structural heart disease who need PVC suppression.

Other antiarrhythmic drugs

Sotalol merits special consideration because it has both beta-blocker and class III antiarrhythmic properties, blocking potassium channels and prolonging cardiac repolarization. It can be very effective in PVC suppression but also creates some degree of QT prolongation. The QT-prolonging effect is accentuated in patients with baseline QT prolongation or abnormal renal function. Rarely, this can lead to torsades de pointes. As a safety precaution, some patients are admitted to the hospital when they start sotalol therapy so that they can be monitored with continuous telemetry and ECG to detect excessive QT prolongation.

Amiodarone is a versatile drug with mixed pharmacologic properties that include a predominantly potassium channel-blocking class III drug effect. However, this effect is balanced by its other pharmacologic properties that make QT prolongation less of a clinical concern. Excessive QT prolongation may still occur when used concomitantly with other QT-prolonging drugs.

Amiodarone is very effective in suppressing PVCs and ventricular arrhythmias but has considerable short-term and long-term side effects. Cumulative toxicity risks include damage to the thyroid gland, liver, skin, eyes, and lungs. Routine thyroid function testing, pulmonary function testing, and eye examinations are often considered for patients on long-term amiodarone therapy. Short-term use of this drug does not typically require such surveillance.

Catheter ablation

As mentioned in the previous sections, catheter ablation is a safe and effective treatment for PVCs. It is curative in most cases, and significantly reduces the PVC burden in others.

Procedure. Patients are brought to the electrophysiology laboratory in a fasted state and are partially sedated with an intravenous drug such as midazolam or fentanyl, or both. Steerable catheters are placed into appropriate cardiac chambers from femoral access sites, which are infiltrated with local anesthesia. Sometimes sedative or analgesic drugs must be limited if they are known to suppress PVCs.

Most operators prefer a technique called activation mapping, in which the catheter is maneuvered to home in on the precise PVC origin within the heart, which is subsequently ablated. This technique has very high success rates, but having enough spontaneous PVCs to map during the procedure is essential for the technique to succeed. Conversely, not having sufficient PVCs on the day of the procedure is a common reason that ablation fails or cannot be performed at all.

Pace-mapping is an alternate technique that does not require a continuous stream of PVCs. This involves pacing from different candidate locations inside the heart in an effort to precisely match the ECG appearance of the clinical PVC and to ablate at this site. Although activation mapping generally yields higher success rates and is preferred by most operators, pace-mapping can be successful when a perfect 12–12 match is elicited. In many cases, the two techniques are used together during the same procedure, particularly if the patient’s PVCs spontaneously wax and wane, as they often do.

Risks. Like any medical procedure, catheter ablation carries some inherent risks, including rare but potentially serious events. Unstable arrhythmias may require pace-termination from the catheter or, rarely, shock-termination externally. Even more rare is cardiac arrest requiring cardiopulmonary resuscitation. Uncommon but life-threatening complications also include pericardial effusion or cardiac tamponade requiring percutaneous drainage or, rarely, emergency surgical correction. Although such events are life-threatening, death is extremely rare.

Complications causing permanent disability are also very uncommon but include the risk of collateral injury to the conduction system requiring permanent pacemaker placement, injury to the coronary vessels requiring urgent treatment, or diaphragmatic injury affecting breathing. Left-sided cardiac ablation also carries a small risk of stroke, which is mitigated by giving intravenous heparin during the procedure.

More common but generally non-life-threatening complications include femoral vascular events such as hematomas, pseudoaneurysms, or fistulas that sometimes require subsequent treatment. These complications are generally treatable but can significantly prolong the recovery period.

Catheter ablation procedures are typically 2 to 6 hours in duration, depending on the chambers involved, PVC frequency, and other considerations. Postprocedure bed rest is required for a number of hours. A Foley catheter is sometimes used for patient comfort when a prolonged procedure is anticipated. This carries a small risk of urinary tract infection. Epicardial catheter ablation that requires access to the surface of the heart (ie, the pericardial space) is uncommon but carries some unique risks, including rare injury to coronary vessels or adjacent organs such as the liver or stomach.

Overall, both endocardial and epicardial catheter ablation can be performed safely and effectively in the overwhelming majority of patients, but understanding and explaining the potential risks remains a crucial part of the informed consent process.

TAKE-HOME POINTS

• PVCs are a common cause of palpitations but are also noted as incidental findings by ECG, Holter monitoring, and inpatient telemetry.
• The diagnostic evaluation includes an assessment for underlying structural heart disease and quantification of the total PVC burden.
• Patients without structural heart disease and with low-to-modest PVC burdens may not require specific treatment. PVCs at greater burdens, typically 15% to 20%, or with specific high-risk features carry a risk of tachycardia-related cardiomyopathy and may require treatment even if they are asymptomatic. These high-risk features include initial QRS slurring and PVCs occurring at shorter coupling intervals.
• Treatment involves medical therapy with a beta-blocker, a calcium channel blocker, or another antiarrhythmic drug, and catheter ablation in selected cases.
• Catheter ablation can be curative but is typically reserved for drug-intolerant or medically refractory patients with a high PVC burden.

Premature ventricular complexes (PVCs) are a common cause of palpitations, and are also often detected incidentally on electrocardiography (ECG), ambulatory monitoring, or inpatient telemetry. At the cellular level, ventricular myocytes spontaneously depolarize to create an extra systole that is “out of sync” with the cardiac cycle.

Although nearly everyone has some PVCs from time to time, people vary widely in their frequency of PVCs and their sensitivity to them.^1,2 Some patients are exquisitely sensitive to even a small number of PVCs, while others are completely unaware of PVCs in a bigeminal pattern (ie, every other heartbeat). This article will review the evaluation and management of PVCs with a focus on clinical aspects.

DIAGNOSTIC EVALUATION

Personal and family history

Symptoms. The initial history should establish the presence, extent, timing, and duration of symptoms. Patients may use the word “palpitations” to describe their symptoms, but they also describe them as “hard” heartbeats, “chest-thumping,” or as a “catch” or “skipped” heartbeat. Related symptoms may include difficulty breathing, chest pain, fatigue, and dizziness.

The interview should determine whether the symptoms represent a minor nuisance or a major quality-of-life issue to the patient, and whether there are any specific associations or triggers. For example, it is very common for patients to become aware of PVCs at night, particularly in certain positions, such as lying on the left side. Patients often associate PVC symptoms with emotional stress, exercise, or caffeine or stimulant use.

Medication use. An accurate and up-to-date list of prescription medications should be screened for alpha-, beta-, or dopamine-receptor agonist drugs. Similarly, any use of over-the-counter sympathomimetic medications and nonprescription supplements should be elicited, including compounded elixirs or beverages. Many commercially available products designed to treat fatigue or increase alertness contain large doses of caffeine or other stimulants. It is also important to consider the use of illicit substances such as cocaine, amphetamine, methamphetamine, and their derivatives.

The patient’s medical and surgical history should be queried for any known structural heart disease, including coronary artery disease, myocardial infarction, congestive heart failure, valvular heart disease, congenital heart disease, and heritable conditions such as hypertrophic cardiomyopathy, prolonged QT syndromes, or other channel disorders. Pulmonary disorders such as sarcoidosis, pulmonary hypertension, or obstructive sleep apnea are also relevant. Similarly, it is important to identify endocrine disorders, including thyroid problems, sex hormone abnormalities, or adrenal gland conditions.

A careful family history should include any instance of sudden death in first-degree relatives, any heritable cardiac conditions, or coronary artery disease at an early age.

Physical examination

The physical examination should focus on findings that suggest underlying structural heart disease. Findings suggestive of congestive heart failure include elevated jugular venous pressures, abnormal cardiac sounds, pulmonary rales, abnormal arterial pulses, or peripheral edema. A murmur or a pathologic heart sound should raise suspicion of valvular or congenital heart disease when present in a young patient.

Inspection and palpation of the thyroid can reveal a related disorder. Obvious skin changes or neurologic findings can similarly reveal a systemic and possibly related clinical disorder that can have cardiac manifestations (eg, muscular dystrophy).

Electrocardiography, Holter monitoring, and other monitoring

Assessment of the cardiac rhythm includes 12-lead ECG and ambulatory Holter monitoring, typically for 24 or 48 hours.

Holter monitoring provides a continuous recording, usually in at least two or three leads. Patients are given a symptom journal or are asked to keep a diary of symptoms experienced during the monitoring period. The monitor is worn underneath clothing and is returned for download upon completion. Technicians process the data with the aid of computer software, and the final output is reviewed and interpreted by a cardiologist or cardiac electrophysiologist.

Holter monitoring for at least 24 hours is a critical step in assessing any patient with known or suspected PVCs, as it can both quantify the total burden of ventricular ectopy and identify the presence of any related ventricular tachycardia. In addition, it can detect additional supraventricular arrhythmias or bradycardia during the monitoring period. The PVC burden is an important measurement; it is expressed as the percentage of heartbeats that were ventricular extrasystoles during the monitoring period.

Both ECG and Holter monitoring are limited in that they are only snapshots of the rhythm during the period when a patient is actually hooked up. Many patients experience PVCs in clusters every very few days or weeks. Such a pattern is unlikely to be detected by a single ECG or 24- or 48-hour Holter monitoring.

A 30-day ambulatory event monitor (also known as a wearable loop recorder) is an important diagnostic tool in these scenarios. The concept is very similar to that of Holter monitoring, except that the device provides a continuous loop recording of the cardiac rhythm that is digitally stored in clips when the patient activates the device. Some wearable loop recorders also have auto-save features for heart rates falling outside of a programmed range.

Mobile outpatient cardiac telemetry is the most comprehensive form of noninvasive rhythm monitoring available. This is essentially the equivalent of continuous inpatient cardiac telemetry, but in a patient who is not hospitalized. It is a wearable ambulatory device providing continuous recordings, real-time automatic detections, and patient-activated symptom recordings. It can be used for up to 6 weeks. Advantages include detection and quantification of asymptomatic events, and real-time transmissions that the physician can act upon. The major disadvantage is cost, including coverage denial by many third-party payers.

This test is rarely indicated as part of a PVC evaluation and is typically ordered only by a cardiologist or cardiac electrophysiologist.

Surface echocardiography is indicated to look for overt structural heart disease and can reliably detect abnormalities in cardiac chamber size, wall thickness, and function. Valvular heart disease is concomitantly identified by two-dimensional imaging as well as by color Doppler. The finding of significant structural heart disease in conjunction with PVCs should prompt a cardiology referral, as this carries significant prognostic implications.^3–5

Exercise treadmill stress testing is appropriate for patients who experience PVCs with exercise or for whom an evaluation for coronary artery disease is indicated. The expected finding would be an increase in PVCs or ventricular tachycardia with exercise or in the subsequent recovery period. Exercise testing can be combined with either echocardiographic or nuclear perfusion imaging to evaluate the possibility of myocardial ischemia. For patients unable to exercise, pharmacologic stress testing with dobutamine or a vasodilator agent can be performed.

Advanced noninvasive cardiac imaging— such as computed tomography, magnetic resonance imaging, or positron-emission tomography—should be reserved for specific clinical indications such as congenital heart disease, suspected cardiac sarcoidosis, and infiltrative heart disease, and for specific cardiomyopathies, such as hypertrophic cardiomyopathy and arrhythmogenic right ventricular cardiomyopathy. For example, frequent PVCs with a left bundle branch block morphology and superior axis raise the concern for a right ventricular disorder and may prompt cardiac magnetic resonance imaging for either arrhythmogenic right ventricular cardiomyopathy or sarcoidosis.

PVCs WITHOUT STRUCTURAL HEART DISEASE

Outflow tract PVCs and ventricular tachycardia

The right or left ventricular outflow tracts, or the epicardial tissue immediately adjacent to the aortic sinuses of Valsalva are the most common sites of origin for ventricular ectopy in the absence of structural heart disease.^6–9 Affected cells often demonstrate a triggered activity mechanism due to cyclic adenosine monophosphate-mediated and calcium-dependent delayed after-depolarizations.^7,8

Most of these foci are in the right ventricular outflow tract, producing a left bundle branch block morphology with an inferior axis (positive R waves in limb leads II, III, and aVF) and typical precordial R-wave transition in V₃ and V₄ (Figure 1). A minority are in the left ventricular outflow tract, producing a right bundle branch block with an inferior axis pattern, or in the aortic sinuses with a left bundle branch block pattern but with early precordial R transition in V₂ and V₃.

A study in 122 patients showed that right and left outflow tract arrhythmias had similar electrophysiologic properties and pharmacologic sensitivities, providing evidence for shared mechanisms possibly due to the common embryologic origin of these structures.⁹

Such arrhythmias are typically catecholamine-sensitive and are sometimes inducible with burst pacing in the electrophysiology laboratory. The short ventricular coupling intervals can promote intracellular calcium overload in the affected cells, leading to triggered activity.

Therefore, outflow tract PVCs and ventricular tachycardia are commonly encountered clinically during exercise and, to an even greater extent, in the postexercise cool-down period. Similarly, they can be worse during periods of emotional stress or fatigue, when the body’s endogenous catecholamine production is elevated. However, it is worthwhile to note that there are exceptions to this principle in which faster sinus rates seem to overdrive the PVCs in some patients, causing them to become paradoxically more frequent at rest, or even during sleep.

Outflow tract PVCs can be managed medically with beta-blockers, nondihydropyridine calcium channel blockers (verapamil or diltiazem), or, less commonly, class IC drugs such as flecainide. They are also highly curable by catheter ablation (Figure 2), with procedure success rates greater than 90%.^9.10

However, a subset of outflow tract PVCs nested deep in a triangle of epicardial tissue between the right and left endocardial surface and underneath the left main coronary artery can be challenging. This region has been labeled the left ventricular summit, and is shielded from ablation by an epicardial fat pad in the adjacent pericardial space.¹¹ Ablation attempts made from the right and left endocardial surfaces as well as the epicardial surface (pericardial space) sometimes cannot adequately penetrate the tissue deep enough to reach the originating focus deep within this triangle. While ablation cannot always fully eliminate the PVC, ablation from more than one of the sites listed can generally reduce its burden, often in combination with suppressive medical therapy (Figure 3).

Fascicular PVCs

Fascicular PVCs originate from within the left ventricular His-Purkinje system¹² and produce a right bundle branch block morphology with either an anterior or posterior hemiblock pattern (Figure 4). Exit from the posterior fascicle causes an anterior hemiblock pattern, and exit from the anterior fascicle a posterior hemiblock pattern. Utilization of the rapidly conducting His-Purkinje system gives these PVCs a very narrow QRS duration, sometimes approaching 120 milliseconds or shorter. This occasionally causes them to be mistaken for aberrantly conducted supraventricular beats. Such spontaneous PVCs are commonly associated with both sustained and nonsustained ventricular tachycardia and are usually sensitive to verapamil.¹³

Special issues relating to mapping and catheter ablation of fascicular arrhythmias involve the identification of Purkinje fiber potentials and associated procedural diagnostic maneuvers during tachycardia.¹⁴

Other sites for PVCs

Other sites of origin for PVCs in the absence of structural heart disease include ventricular tissue adjacent to the aortomitral continuity,¹⁵ the tricuspid annulus,¹⁶ the mitral valve annulus, ¹⁷ papillary muscles,¹⁸ and other Purkinje-adjacent structures such as left ventricular false tendons.¹⁹ An example of a papillary muscle PVC is shown in Figures 5 and 6.

Curable by catheter ablation

Any of these PVCs can potentially be cured by catheter ablation when present at a sufficient burden to allow for activation mapping in the electrophysiology laboratory. The threshold for offering ablation varies among operators, but is generally around 10% or greater. Pacemapping is a technique applied in the electrophysiology laboratory when medically refractory symptomatic PVCs occurring at a lower burden require ablation.

PVCs WITH AN UNDERLYING CARDIAC CONDITION

Coronary artery disease

Tissue injury and death caused by acute myocardial infarction has long been recognized as a common cause of spontaneous ventricular ectopy attributed to infarct border zones of ischemic or hibernating myocardium.^20,21

Suppression has not been associated with improved outcomes, as shown for class IC drugs in the landmark Cardiac Arrhythmia Suppression Trial (CAST),²² or in the amiodarone treatment arm of the Multicenter Automatic Defibrillator Implantation Trial II (MADIT-II).²³ Therefore, treatment of ventricular ectopy in this patient population is usually symptom-driven unless there is hemodynamic intolerance, tachycardia-related cardiomyopathy, or a very high burden of PVCs in a patient who may be at risk of developing tachycardia-related cardiomyopathy. Antiarrhythmic drug treatment, when required, usually involves beta-blockers or class III medications such as sotalol or amiodarone.

Nonischemic dilated cardiomyopathy

This category includes patients with a wide variety of disease states including valvular heart disease, lymphocytic and other viral myocarditis, cardiac sarcoidosis, amyloidosis and other infiltrative diseases, familial conditions, and idiopathic dilated cardiomyopathy (ie, etiology unknown). Although it is a heterogeneous group, a common theme is that PVCs in this patient cohort may require epicardial mapping and ablation.²⁴ Similarly, epicardial PVCs and ventricular tachycardia cluster at the basal posterolateral left ventricle near the mitral annulus, for unclear reasons.²⁵

While specific criteria have been published, an epicardial focus is suggested by slowing of the initial QRS segment, pseudo-delta waves, a wider overall QRS, and Q waves in limb lead I.²⁶

Treatment is symptom-driven unless the patient has a tachycardia-related cardiomyopathy or a high burden associated with the risk for its development. Antiarrhythmic drug therapy, when required, typically involves a beta-blocker or a class III drug such as sotalol or amiodarone. Sotalol is used in this population but has limited safety data and should be used cautiously in patients without an implantable cardioverter-defibrillator.

Spontaneous ventricular ectopy and tachycardia are common, if not expected, in patients with this heritable autosomal dominant disorder. This condition is progressive and associated with the risk of sudden cardiac death. Criteria for diagnosis were established in 2010, and patients with suspected arrhythmogenic right ventricular cardiomyopathy often undergo cardiac magnetic resonance imaging.²⁷ Diagnostic findings include fibro-fatty tissue replacement, which usually starts in the right ventricle but can progress to involve the left ventricle. PVCs and ventricular tachycardia can involve the right ventricular free wall and are often epicardial.

Catheter ablation is usually palliative, as future arrhythmias are expected. Many patients with this condition require an implantable cardioverter-defibrillator for prevention of sudden cardiac death, and some go on to cardiac transplantation as the disease progresses and ventricular arrhythmias become incessant.

Other conditions

Spontaneous ventricular ectopy is common in other heritable and acquired cardiomyopathies including hypertrophic cardiomyopathy and in infiltrative or inflammatory disorders such as cardiac amyloidosis and sarcoidosis. While technically falling under the rubric of nonischemic heart disease, the presence of spontaneous ventricular ectopy carries specific prognostic implications depending on the underlying diagnosis. Therefore, an appropriate referral for complete cardiac evaluation should be considered when a heritable disorder or other acquired structural heart disease is suspected.

TACHYCARDIA-RELATED CARDIOMYOPATHY

Tachycardia-related cardiomyopathy refers to left ventricular systolic dysfunction that is primarily caused by arrhythmias. This includes frequent PVCs or ventricular tachycardia but also atrial arrhythmias occurring at a high burden that directly weaken myocardial function over time. Although much research has been devoted to this condition, our understanding of its etiology and pathology is incomplete.

PVCs and ventricular ectopy burdens in excess of 15% to 20% have been associated with the development of this condition.^28,29 However, it is important to note that cardiomyopathy can also develop at lower burdens.³⁰ One study found that a burden greater than 24% was 79% sensitive and 78% specific for development of tachycardia-related cardiomyopathy.³¹ Additional studies have demonstrated specific PVC morphologic features such as slurring in the initial QRS segment and also PVCs occurring at shorter coupling intervals as being associated with cardiomyopathy.^32–34

For these reasons, both quantification of the total burden and careful evaluation of available electrocardiograms and rhythm strips are important even in asymptomatic patients with frequent PVCs. Similarly, unexplained left ventricular dysfunction in patients with PVC burdens in these discussed ranges should raise suspicion for this diagnosis. Patients with tachycardia-related cardiomyopathy usually have at least partially reversible left ventricular dysfunction when identified or treated early.^29,35

MEDICAL AND ABLATIVE TREATMENT

Available treatments include medical suppression and catheter ablation. One needs to exercise clinical judgment and incorporate all of the PVC-related data to make treatment decisions.

Little data for trigger avoidance and behavioral modification

Some patients report a strong association between palpitations related to PVCs and caffeine intake, other stimulants, or other dietary triggers. However, few data exist to support the role of trigger avoidance and behavioral modification in treatment. In fact, an older randomized trial in 81 men found no benefit in a program of total abstinence from caffeine and smoking, moderation of alcohol intake, and physical conditioning.³⁶

Nonetheless, some argue in favor of advising patients to make these dietary and lifestyle changes, given the overall health benefits of aggressive risk-factor modification for cardiovascular disease.³⁷ Certainly, a trial of trigger avoidance and behavioral modification seems reasonable for patients who have strongly associated historical triggers in the absence of structural heart disease and PVCs occurring at a low to modest burden.

Beta-blockers are the mainstay

Beta-blockers are the mainstay of medical suppression of PVCs, primarily through their effect on beta-1 adrenergic receptors to reduce intracellular cyclic adenosine monophosphate and thus decrease automaticity. Blocking beta-1 receptors also causes a negative chronotropic effect, reducing the resting sinus rate in addition to slowing atrioventricular nodal conduction.

Cardioselective beta-blockers include atenolol, betaxolol, metoprolol, and nadolol. These drugs are effective in suppressing PVCs, or at least in reducing the burden to more tolerable levels.

Beta-blockers are most strongly indicated in patients who require PVC suppression and who have concomitant coronary artery disease, prior myocardial infarction, or other cardiomyopathy, as this drug class favorably affects long-term prognosis in these conditions.

Common side effects of beta-blockers include fatigue, shortness of breath, depressed mood, and loss of libido. Side effects can present a significant challenge, particularly for younger patients. Noncardioselective beta-blockers are less commonly prescribed, with the exception of propranolol, which is an effective sympatholytic drug that blocks both beta-1 and beta-2 receptors.

Many patients with asthma or peripheral arterial disease can tolerate these drugs well despite concerns about provoked bronchospasm or claudication, respectively, and neither of these conditions is considered an absolute contraindication. Excessive bradycardia with beta-blocker therapy can lead to dizziness, lightheadedness, or overt syncope, and these drugs should be used with caution in patients with baseline sinus node dysfunction or atrioventricular nodal disease.

Nondihydropyridine calcium channel blockers are particularly effective for PVC suppression in patients without structural heart disease by the mechanisms previously described involving intracellular calcium channels. In particular, they are highly effective and are considered the drugs of choice in treating fascicular PVCs.

Verapamil is a potent drug in this class, but it also commonly causes constipation as a side effect. Diltiazem is less constipating but can cause fatigue, drowsiness, and headaches. Both drugs reduce the resting heart rate and slow atrioventricular nodal conduction. Patients predisposed to bradycardia or atrioventricular block can develop dizziness or overt syncope. Calcium channel blockers are also used cautiously in patients with congestive heart failure, given their potential negative inotropic effects.

Overall, calcium channel blockers are a very reasonable choice for young patients without structural heart disease who need PVC suppression.

Other antiarrhythmic drugs

Sotalol merits special consideration because it has both beta-blocker and class III antiarrhythmic properties, blocking potassium channels and prolonging cardiac repolarization. It can be very effective in PVC suppression but also creates some degree of QT prolongation. The QT-prolonging effect is accentuated in patients with baseline QT prolongation or abnormal renal function. Rarely, this can lead to torsades de pointes. As a safety precaution, some patients are admitted to the hospital when they start sotalol therapy so that they can be monitored with continuous telemetry and ECG to detect excessive QT prolongation.

Amiodarone is a versatile drug with mixed pharmacologic properties that include a predominantly potassium channel-blocking class III drug effect. However, this effect is balanced by its other pharmacologic properties that make QT prolongation less of a clinical concern. Excessive QT prolongation may still occur when used concomitantly with other QT-prolonging drugs.

Amiodarone is very effective in suppressing PVCs and ventricular arrhythmias but has considerable short-term and long-term side effects. Cumulative toxicity risks include damage to the thyroid gland, liver, skin, eyes, and lungs. Routine thyroid function testing, pulmonary function testing, and eye examinations are often considered for patients on long-term amiodarone therapy. Short-term use of this drug does not typically require such surveillance.

Catheter ablation

As mentioned in the previous sections, catheter ablation is a safe and effective treatment for PVCs. It is curative in most cases, and significantly reduces the PVC burden in others.

Procedure. Patients are brought to the electrophysiology laboratory in a fasted state and are partially sedated with an intravenous drug such as midazolam or fentanyl, or both. Steerable catheters are placed into appropriate cardiac chambers from femoral access sites, which are infiltrated with local anesthesia. Sometimes sedative or analgesic drugs must be limited if they are known to suppress PVCs.

Most operators prefer a technique called activation mapping, in which the catheter is maneuvered to home in on the precise PVC origin within the heart, which is subsequently ablated. This technique has very high success rates, but having enough spontaneous PVCs to map during the procedure is essential for the technique to succeed. Conversely, not having sufficient PVCs on the day of the procedure is a common reason that ablation fails or cannot be performed at all.

Pace-mapping is an alternate technique that does not require a continuous stream of PVCs. This involves pacing from different candidate locations inside the heart in an effort to precisely match the ECG appearance of the clinical PVC and to ablate at this site. Although activation mapping generally yields higher success rates and is preferred by most operators, pace-mapping can be successful when a perfect 12–12 match is elicited. In many cases, the two techniques are used together during the same procedure, particularly if the patient’s PVCs spontaneously wax and wane, as they often do.

Risks. Like any medical procedure, catheter ablation carries some inherent risks, including rare but potentially serious events. Unstable arrhythmias may require pace-termination from the catheter or, rarely, shock-termination externally. Even more rare is cardiac arrest requiring cardiopulmonary resuscitation. Uncommon but life-threatening complications also include pericardial effusion or cardiac tamponade requiring percutaneous drainage or, rarely, emergency surgical correction. Although such events are life-threatening, death is extremely rare.

Complications causing permanent disability are also very uncommon but include the risk of collateral injury to the conduction system requiring permanent pacemaker placement, injury to the coronary vessels requiring urgent treatment, or diaphragmatic injury affecting breathing. Left-sided cardiac ablation also carries a small risk of stroke, which is mitigated by giving intravenous heparin during the procedure.

More common but generally non-life-threatening complications include femoral vascular events such as hematomas, pseudoaneurysms, or fistulas that sometimes require subsequent treatment. These complications are generally treatable but can significantly prolong the recovery period.

Catheter ablation procedures are typically 2 to 6 hours in duration, depending on the chambers involved, PVC frequency, and other considerations. Postprocedure bed rest is required for a number of hours. A Foley catheter is sometimes used for patient comfort when a prolonged procedure is anticipated. This carries a small risk of urinary tract infection. Epicardial catheter ablation that requires access to the surface of the heart (ie, the pericardial space) is uncommon but carries some unique risks, including rare injury to coronary vessels or adjacent organs such as the liver or stomach.

Overall, both endocardial and epicardial catheter ablation can be performed safely and effectively in the overwhelming majority of patients, but understanding and explaining the potential risks remains a crucial part of the informed consent process.

TAKE-HOME POINTS

• PVCs are a common cause of palpitations but are also noted as incidental findings by ECG, Holter monitoring, and inpatient telemetry.
• The diagnostic evaluation includes an assessment for underlying structural heart disease and quantification of the total PVC burden.
• Patients without structural heart disease and with low-to-modest PVC burdens may not require specific treatment. PVCs at greater burdens, typically 15% to 20%, or with specific high-risk features carry a risk of tachycardia-related cardiomyopathy and may require treatment even if they are asymptomatic. These high-risk features include initial QRS slurring and PVCs occurring at shorter coupling intervals.
• Treatment involves medical therapy with a beta-blocker, a calcium channel blocker, or another antiarrhythmic drug, and catheter ablation in selected cases.
• Catheter ablation can be curative but is typically reserved for drug-intolerant or medically refractory patients with a high PVC burden.

Premature ventricular complexes (PVCs) are a common cause of palpitations, and are also often detected incidentally on electrocardiography (ECG), ambulatory monitoring, or inpatient telemetry. At the cellular level, ventricular myocytes spontaneously depolarize to create an extra systole that is “out of sync” with the cardiac cycle.

Although nearly everyone has some PVCs from time to time, people vary widely in their frequency of PVCs and their sensitivity to them.^1,2 Some patients are exquisitely sensitive to even a small number of PVCs, while others are completely unaware of PVCs in a bigeminal pattern (ie, every other heartbeat). This article will review the evaluation and management of PVCs with a focus on clinical aspects.

DIAGNOSTIC EVALUATION

Personal and family history

Symptoms. The initial history should establish the presence, extent, timing, and duration of symptoms. Patients may use the word “palpitations” to describe their symptoms, but they also describe them as “hard” heartbeats, “chest-thumping,” or as a “catch” or “skipped” heartbeat. Related symptoms may include difficulty breathing, chest pain, fatigue, and dizziness.

The interview should determine whether the symptoms represent a minor nuisance or a major quality-of-life issue to the patient, and whether there are any specific associations or triggers. For example, it is very common for patients to become aware of PVCs at night, particularly in certain positions, such as lying on the left side. Patients often associate PVC symptoms with emotional stress, exercise, or caffeine or stimulant use.

Medication use. An accurate and up-to-date list of prescription medications should be screened for alpha-, beta-, or dopamine-receptor agonist drugs. Similarly, any use of over-the-counter sympathomimetic medications and nonprescription supplements should be elicited, including compounded elixirs or beverages. Many commercially available products designed to treat fatigue or increase alertness contain large doses of caffeine or other stimulants. It is also important to consider the use of illicit substances such as cocaine, amphetamine, methamphetamine, and their derivatives.

The patient’s medical and surgical history should be queried for any known structural heart disease, including coronary artery disease, myocardial infarction, congestive heart failure, valvular heart disease, congenital heart disease, and heritable conditions such as hypertrophic cardiomyopathy, prolonged QT syndromes, or other channel disorders. Pulmonary disorders such as sarcoidosis, pulmonary hypertension, or obstructive sleep apnea are also relevant. Similarly, it is important to identify endocrine disorders, including thyroid problems, sex hormone abnormalities, or adrenal gland conditions.

A careful family history should include any instance of sudden death in first-degree relatives, any heritable cardiac conditions, or coronary artery disease at an early age.

Physical examination

The physical examination should focus on findings that suggest underlying structural heart disease. Findings suggestive of congestive heart failure include elevated jugular venous pressures, abnormal cardiac sounds, pulmonary rales, abnormal arterial pulses, or peripheral edema. A murmur or a pathologic heart sound should raise suspicion of valvular or congenital heart disease when present in a young patient.

Inspection and palpation of the thyroid can reveal a related disorder. Obvious skin changes or neurologic findings can similarly reveal a systemic and possibly related clinical disorder that can have cardiac manifestations (eg, muscular dystrophy).

Electrocardiography, Holter monitoring, and other monitoring

Assessment of the cardiac rhythm includes 12-lead ECG and ambulatory Holter monitoring, typically for 24 or 48 hours.

Holter monitoring provides a continuous recording, usually in at least two or three leads. Patients are given a symptom journal or are asked to keep a diary of symptoms experienced during the monitoring period. The monitor is worn underneath clothing and is returned for download upon completion. Technicians process the data with the aid of computer software, and the final output is reviewed and interpreted by a cardiologist or cardiac electrophysiologist.

Holter monitoring for at least 24 hours is a critical step in assessing any patient with known or suspected PVCs, as it can both quantify the total burden of ventricular ectopy and identify the presence of any related ventricular tachycardia. In addition, it can detect additional supraventricular arrhythmias or bradycardia during the monitoring period. The PVC burden is an important measurement; it is expressed as the percentage of heartbeats that were ventricular extrasystoles during the monitoring period.

Both ECG and Holter monitoring are limited in that they are only snapshots of the rhythm during the period when a patient is actually hooked up. Many patients experience PVCs in clusters every very few days or weeks. Such a pattern is unlikely to be detected by a single ECG or 24- or 48-hour Holter monitoring.

A 30-day ambulatory event monitor (also known as a wearable loop recorder) is an important diagnostic tool in these scenarios. The concept is very similar to that of Holter monitoring, except that the device provides a continuous loop recording of the cardiac rhythm that is digitally stored in clips when the patient activates the device. Some wearable loop recorders also have auto-save features for heart rates falling outside of a programmed range.

Mobile outpatient cardiac telemetry is the most comprehensive form of noninvasive rhythm monitoring available. This is essentially the equivalent of continuous inpatient cardiac telemetry, but in a patient who is not hospitalized. It is a wearable ambulatory device providing continuous recordings, real-time automatic detections, and patient-activated symptom recordings. It can be used for up to 6 weeks. Advantages include detection and quantification of asymptomatic events, and real-time transmissions that the physician can act upon. The major disadvantage is cost, including coverage denial by many third-party payers.

This test is rarely indicated as part of a PVC evaluation and is typically ordered only by a cardiologist or cardiac electrophysiologist.

Surface echocardiography is indicated to look for overt structural heart disease and can reliably detect abnormalities in cardiac chamber size, wall thickness, and function. Valvular heart disease is concomitantly identified by two-dimensional imaging as well as by color Doppler. The finding of significant structural heart disease in conjunction with PVCs should prompt a cardiology referral, as this carries significant prognostic implications.^3–5

Exercise treadmill stress testing is appropriate for patients who experience PVCs with exercise or for whom an evaluation for coronary artery disease is indicated. The expected finding would be an increase in PVCs or ventricular tachycardia with exercise or in the subsequent recovery period. Exercise testing can be combined with either echocardiographic or nuclear perfusion imaging to evaluate the possibility of myocardial ischemia. For patients unable to exercise, pharmacologic stress testing with dobutamine or a vasodilator agent can be performed.

Advanced noninvasive cardiac imaging— such as computed tomography, magnetic resonance imaging, or positron-emission tomography—should be reserved for specific clinical indications such as congenital heart disease, suspected cardiac sarcoidosis, and infiltrative heart disease, and for specific cardiomyopathies, such as hypertrophic cardiomyopathy and arrhythmogenic right ventricular cardiomyopathy. For example, frequent PVCs with a left bundle branch block morphology and superior axis raise the concern for a right ventricular disorder and may prompt cardiac magnetic resonance imaging for either arrhythmogenic right ventricular cardiomyopathy or sarcoidosis.

PVCs WITHOUT STRUCTURAL HEART DISEASE

Outflow tract PVCs and ventricular tachycardia

The right or left ventricular outflow tracts, or the epicardial tissue immediately adjacent to the aortic sinuses of Valsalva are the most common sites of origin for ventricular ectopy in the absence of structural heart disease.^6–9 Affected cells often demonstrate a triggered activity mechanism due to cyclic adenosine monophosphate-mediated and calcium-dependent delayed after-depolarizations.^7,8

Most of these foci are in the right ventricular outflow tract, producing a left bundle branch block morphology with an inferior axis (positive R waves in limb leads II, III, and aVF) and typical precordial R-wave transition in V₃ and V₄ (Figure 1). A minority are in the left ventricular outflow tract, producing a right bundle branch block with an inferior axis pattern, or in the aortic sinuses with a left bundle branch block pattern but with early precordial R transition in V₂ and V₃.

A study in 122 patients showed that right and left outflow tract arrhythmias had similar electrophysiologic properties and pharmacologic sensitivities, providing evidence for shared mechanisms possibly due to the common embryologic origin of these structures.⁹

Such arrhythmias are typically catecholamine-sensitive and are sometimes inducible with burst pacing in the electrophysiology laboratory. The short ventricular coupling intervals can promote intracellular calcium overload in the affected cells, leading to triggered activity.

Therefore, outflow tract PVCs and ventricular tachycardia are commonly encountered clinically during exercise and, to an even greater extent, in the postexercise cool-down period. Similarly, they can be worse during periods of emotional stress or fatigue, when the body’s endogenous catecholamine production is elevated. However, it is worthwhile to note that there are exceptions to this principle in which faster sinus rates seem to overdrive the PVCs in some patients, causing them to become paradoxically more frequent at rest, or even during sleep.

Outflow tract PVCs can be managed medically with beta-blockers, nondihydropyridine calcium channel blockers (verapamil or diltiazem), or, less commonly, class IC drugs such as flecainide. They are also highly curable by catheter ablation (Figure 2), with procedure success rates greater than 90%.^9.10

However, a subset of outflow tract PVCs nested deep in a triangle of epicardial tissue between the right and left endocardial surface and underneath the left main coronary artery can be challenging. This region has been labeled the left ventricular summit, and is shielded from ablation by an epicardial fat pad in the adjacent pericardial space.¹¹ Ablation attempts made from the right and left endocardial surfaces as well as the epicardial surface (pericardial space) sometimes cannot adequately penetrate the tissue deep enough to reach the originating focus deep within this triangle. While ablation cannot always fully eliminate the PVC, ablation from more than one of the sites listed can generally reduce its burden, often in combination with suppressive medical therapy (Figure 3).

Fascicular PVCs

Fascicular PVCs originate from within the left ventricular His-Purkinje system¹² and produce a right bundle branch block morphology with either an anterior or posterior hemiblock pattern (Figure 4). Exit from the posterior fascicle causes an anterior hemiblock pattern, and exit from the anterior fascicle a posterior hemiblock pattern. Utilization of the rapidly conducting His-Purkinje system gives these PVCs a very narrow QRS duration, sometimes approaching 120 milliseconds or shorter. This occasionally causes them to be mistaken for aberrantly conducted supraventricular beats. Such spontaneous PVCs are commonly associated with both sustained and nonsustained ventricular tachycardia and are usually sensitive to verapamil.¹³

Special issues relating to mapping and catheter ablation of fascicular arrhythmias involve the identification of Purkinje fiber potentials and associated procedural diagnostic maneuvers during tachycardia.¹⁴

Other sites for PVCs

Other sites of origin for PVCs in the absence of structural heart disease include ventricular tissue adjacent to the aortomitral continuity,¹⁵ the tricuspid annulus,¹⁶ the mitral valve annulus, ¹⁷ papillary muscles,¹⁸ and other Purkinje-adjacent structures such as left ventricular false tendons.¹⁹ An example of a papillary muscle PVC is shown in Figures 5 and 6.

Curable by catheter ablation

Any of these PVCs can potentially be cured by catheter ablation when present at a sufficient burden to allow for activation mapping in the electrophysiology laboratory. The threshold for offering ablation varies among operators, but is generally around 10% or greater. Pacemapping is a technique applied in the electrophysiology laboratory when medically refractory symptomatic PVCs occurring at a lower burden require ablation.

PVCs WITH AN UNDERLYING CARDIAC CONDITION

Coronary artery disease

Tissue injury and death caused by acute myocardial infarction has long been recognized as a common cause of spontaneous ventricular ectopy attributed to infarct border zones of ischemic or hibernating myocardium.^20,21

Suppression has not been associated with improved outcomes, as shown for class IC drugs in the landmark Cardiac Arrhythmia Suppression Trial (CAST),²² or in the amiodarone treatment arm of the Multicenter Automatic Defibrillator Implantation Trial II (MADIT-II).²³ Therefore, treatment of ventricular ectopy in this patient population is usually symptom-driven unless there is hemodynamic intolerance, tachycardia-related cardiomyopathy, or a very high burden of PVCs in a patient who may be at risk of developing tachycardia-related cardiomyopathy. Antiarrhythmic drug treatment, when required, usually involves beta-blockers or class III medications such as sotalol or amiodarone.

Nonischemic dilated cardiomyopathy

This category includes patients with a wide variety of disease states including valvular heart disease, lymphocytic and other viral myocarditis, cardiac sarcoidosis, amyloidosis and other infiltrative diseases, familial conditions, and idiopathic dilated cardiomyopathy (ie, etiology unknown). Although it is a heterogeneous group, a common theme is that PVCs in this patient cohort may require epicardial mapping and ablation.²⁴ Similarly, epicardial PVCs and ventricular tachycardia cluster at the basal posterolateral left ventricle near the mitral annulus, for unclear reasons.²⁵

While specific criteria have been published, an epicardial focus is suggested by slowing of the initial QRS segment, pseudo-delta waves, a wider overall QRS, and Q waves in limb lead I.²⁶

Treatment is symptom-driven unless the patient has a tachycardia-related cardiomyopathy or a high burden associated with the risk for its development. Antiarrhythmic drug therapy, when required, typically involves a beta-blocker or a class III drug such as sotalol or amiodarone. Sotalol is used in this population but has limited safety data and should be used cautiously in patients without an implantable cardioverter-defibrillator.

Spontaneous ventricular ectopy and tachycardia are common, if not expected, in patients with this heritable autosomal dominant disorder. This condition is progressive and associated with the risk of sudden cardiac death. Criteria for diagnosis were established in 2010, and patients with suspected arrhythmogenic right ventricular cardiomyopathy often undergo cardiac magnetic resonance imaging.²⁷ Diagnostic findings include fibro-fatty tissue replacement, which usually starts in the right ventricle but can progress to involve the left ventricle. PVCs and ventricular tachycardia can involve the right ventricular free wall and are often epicardial.

Catheter ablation is usually palliative, as future arrhythmias are expected. Many patients with this condition require an implantable cardioverter-defibrillator for prevention of sudden cardiac death, and some go on to cardiac transplantation as the disease progresses and ventricular arrhythmias become incessant.

Other conditions

Spontaneous ventricular ectopy is common in other heritable and acquired cardiomyopathies including hypertrophic cardiomyopathy and in infiltrative or inflammatory disorders such as cardiac amyloidosis and sarcoidosis. While technically falling under the rubric of nonischemic heart disease, the presence of spontaneous ventricular ectopy carries specific prognostic implications depending on the underlying diagnosis. Therefore, an appropriate referral for complete cardiac evaluation should be considered when a heritable disorder or other acquired structural heart disease is suspected.

TACHYCARDIA-RELATED CARDIOMYOPATHY

Tachycardia-related cardiomyopathy refers to left ventricular systolic dysfunction that is primarily caused by arrhythmias. This includes frequent PVCs or ventricular tachycardia but also atrial arrhythmias occurring at a high burden that directly weaken myocardial function over time. Although much research has been devoted to this condition, our understanding of its etiology and pathology is incomplete.

PVCs and ventricular ectopy burdens in excess of 15% to 20% have been associated with the development of this condition.^28,29 However, it is important to note that cardiomyopathy can also develop at lower burdens.³⁰ One study found that a burden greater than 24% was 79% sensitive and 78% specific for development of tachycardia-related cardiomyopathy.³¹ Additional studies have demonstrated specific PVC morphologic features such as slurring in the initial QRS segment and also PVCs occurring at shorter coupling intervals as being associated with cardiomyopathy.^32–34

For these reasons, both quantification of the total burden and careful evaluation of available electrocardiograms and rhythm strips are important even in asymptomatic patients with frequent PVCs. Similarly, unexplained left ventricular dysfunction in patients with PVC burdens in these discussed ranges should raise suspicion for this diagnosis. Patients with tachycardia-related cardiomyopathy usually have at least partially reversible left ventricular dysfunction when identified or treated early.^29,35

MEDICAL AND ABLATIVE TREATMENT

Available treatments include medical suppression and catheter ablation. One needs to exercise clinical judgment and incorporate all of the PVC-related data to make treatment decisions.

Little data for trigger avoidance and behavioral modification

Some patients report a strong association between palpitations related to PVCs and caffeine intake, other stimulants, or other dietary triggers. However, few data exist to support the role of trigger avoidance and behavioral modification in treatment. In fact, an older randomized trial in 81 men found no benefit in a program of total abstinence from caffeine and smoking, moderation of alcohol intake, and physical conditioning.³⁶

Nonetheless, some argue in favor of advising patients to make these dietary and lifestyle changes, given the overall health benefits of aggressive risk-factor modification for cardiovascular disease.³⁷ Certainly, a trial of trigger avoidance and behavioral modification seems reasonable for patients who have strongly associated historical triggers in the absence of structural heart disease and PVCs occurring at a low to modest burden.

Beta-blockers are the mainstay

Beta-blockers are the mainstay of medical suppression of PVCs, primarily through their effect on beta-1 adrenergic receptors to reduce intracellular cyclic adenosine monophosphate and thus decrease automaticity. Blocking beta-1 receptors also causes a negative chronotropic effect, reducing the resting sinus rate in addition to slowing atrioventricular nodal conduction.

Cardioselective beta-blockers include atenolol, betaxolol, metoprolol, and nadolol. These drugs are effective in suppressing PVCs, or at least in reducing the burden to more tolerable levels.

Beta-blockers are most strongly indicated in patients who require PVC suppression and who have concomitant coronary artery disease, prior myocardial infarction, or other cardiomyopathy, as this drug class favorably affects long-term prognosis in these conditions.

Common side effects of beta-blockers include fatigue, shortness of breath, depressed mood, and loss of libido. Side effects can present a significant challenge, particularly for younger patients. Noncardioselective beta-blockers are less commonly prescribed, with the exception of propranolol, which is an effective sympatholytic drug that blocks both beta-1 and beta-2 receptors.

Many patients with asthma or peripheral arterial disease can tolerate these drugs well despite concerns about provoked bronchospasm or claudication, respectively, and neither of these conditions is considered an absolute contraindication. Excessive bradycardia with beta-blocker therapy can lead to dizziness, lightheadedness, or overt syncope, and these drugs should be used with caution in patients with baseline sinus node dysfunction or atrioventricular nodal disease.

Nondihydropyridine calcium channel blockers are particularly effective for PVC suppression in patients without structural heart disease by the mechanisms previously described involving intracellular calcium channels. In particular, they are highly effective and are considered the drugs of choice in treating fascicular PVCs.

Verapamil is a potent drug in this class, but it also commonly causes constipation as a side effect. Diltiazem is less constipating but can cause fatigue, drowsiness, and headaches. Both drugs reduce the resting heart rate and slow atrioventricular nodal conduction. Patients predisposed to bradycardia or atrioventricular block can develop dizziness or overt syncope. Calcium channel blockers are also used cautiously in patients with congestive heart failure, given their potential negative inotropic effects.

Overall, calcium channel blockers are a very reasonable choice for young patients without structural heart disease who need PVC suppression.

Other antiarrhythmic drugs

Sotalol merits special consideration because it has both beta-blocker and class III antiarrhythmic properties, blocking potassium channels and prolonging cardiac repolarization. It can be very effective in PVC suppression but also creates some degree of QT prolongation. The QT-prolonging effect is accentuated in patients with baseline QT prolongation or abnormal renal function. Rarely, this can lead to torsades de pointes. As a safety precaution, some patients are admitted to the hospital when they start sotalol therapy so that they can be monitored with continuous telemetry and ECG to detect excessive QT prolongation.

Amiodarone is a versatile drug with mixed pharmacologic properties that include a predominantly potassium channel-blocking class III drug effect. However, this effect is balanced by its other pharmacologic properties that make QT prolongation less of a clinical concern. Excessive QT prolongation may still occur when used concomitantly with other QT-prolonging drugs.

Amiodarone is very effective in suppressing PVCs and ventricular arrhythmias but has considerable short-term and long-term side effects. Cumulative toxicity risks include damage to the thyroid gland, liver, skin, eyes, and lungs. Routine thyroid function testing, pulmonary function testing, and eye examinations are often considered for patients on long-term amiodarone therapy. Short-term use of this drug does not typically require such surveillance.

Catheter ablation

As mentioned in the previous sections, catheter ablation is a safe and effective treatment for PVCs. It is curative in most cases, and significantly reduces the PVC burden in others.

Procedure. Patients are brought to the electrophysiology laboratory in a fasted state and are partially sedated with an intravenous drug such as midazolam or fentanyl, or both. Steerable catheters are placed into appropriate cardiac chambers from femoral access sites, which are infiltrated with local anesthesia. Sometimes sedative or analgesic drugs must be limited if they are known to suppress PVCs.

Most operators prefer a technique called activation mapping, in which the catheter is maneuvered to home in on the precise PVC origin within the heart, which is subsequently ablated. This technique has very high success rates, but having enough spontaneous PVCs to map during the procedure is essential for the technique to succeed. Conversely, not having sufficient PVCs on the day of the procedure is a common reason that ablation fails or cannot be performed at all.

Pace-mapping is an alternate technique that does not require a continuous stream of PVCs. This involves pacing from different candidate locations inside the heart in an effort to precisely match the ECG appearance of the clinical PVC and to ablate at this site. Although activation mapping generally yields higher success rates and is preferred by most operators, pace-mapping can be successful when a perfect 12–12 match is elicited. In many cases, the two techniques are used together during the same procedure, particularly if the patient’s PVCs spontaneously wax and wane, as they often do.

Risks. Like any medical procedure, catheter ablation carries some inherent risks, including rare but potentially serious events. Unstable arrhythmias may require pace-termination from the catheter or, rarely, shock-termination externally. Even more rare is cardiac arrest requiring cardiopulmonary resuscitation. Uncommon but life-threatening complications also include pericardial effusion or cardiac tamponade requiring percutaneous drainage or, rarely, emergency surgical correction. Although such events are life-threatening, death is extremely rare.

Complications causing permanent disability are also very uncommon but include the risk of collateral injury to the conduction system requiring permanent pacemaker placement, injury to the coronary vessels requiring urgent treatment, or diaphragmatic injury affecting breathing. Left-sided cardiac ablation also carries a small risk of stroke, which is mitigated by giving intravenous heparin during the procedure.

More common but generally non-life-threatening complications include femoral vascular events such as hematomas, pseudoaneurysms, or fistulas that sometimes require subsequent treatment. These complications are generally treatable but can significantly prolong the recovery period.

Catheter ablation procedures are typically 2 to 6 hours in duration, depending on the chambers involved, PVC frequency, and other considerations. Postprocedure bed rest is required for a number of hours. A Foley catheter is sometimes used for patient comfort when a prolonged procedure is anticipated. This carries a small risk of urinary tract infection. Epicardial catheter ablation that requires access to the surface of the heart (ie, the pericardial space) is uncommon but carries some unique risks, including rare injury to coronary vessels or adjacent organs such as the liver or stomach.

Overall, both endocardial and epicardial catheter ablation can be performed safely and effectively in the overwhelming majority of patients, but understanding and explaining the potential risks remains a crucial part of the informed consent process.

TAKE-HOME POINTS

• PVCs are a common cause of palpitations but are also noted as incidental findings by ECG, Holter monitoring, and inpatient telemetry.
• The diagnostic evaluation includes an assessment for underlying structural heart disease and quantification of the total PVC burden.
• Patients without structural heart disease and with low-to-modest PVC burdens may not require specific treatment. PVCs at greater burdens, typically 15% to 20%, or with specific high-risk features carry a risk of tachycardia-related cardiomyopathy and may require treatment even if they are asymptomatic. These high-risk features include initial QRS slurring and PVCs occurring at shorter coupling intervals.
• Treatment involves medical therapy with a beta-blocker, a calcium channel blocker, or another antiarrhythmic drug, and catheter ablation in selected cases.
• Catheter ablation can be curative but is typically reserved for drug-intolerant or medically refractory patients with a high PVC burden.

Although we still have no evidence from randomized trials that patients have better outcomes if we measure the calcification in their coronary arteries, a growing body of evidence shows that we can estimate risk more accurately than with a risk model score alone if we also score coronary artery calcification in asymptomatic patients, especially those at intermediate risk.

See related editorial

Current guidelines¹ recommend using the Framingham Risk Score or a similar tool to estimate coronary risk in asymptomatic patients, but these tools have only modest accuracy. Calcification scoring is accurate, inexpensive, quick, widely available, low-risk, and does not appear to increase medical costs afterward. In addition to improving risk stratification, it may also encourage patients to adhere better to drug therapy and lifestyle modification.

HOW IS CORONARY ARTERY CALCIFICATION MEASURED?

Calcification of the coronary arteries is synonymous with atherosclerosis. It can easily be detected with computed tomography without contrast (Figure 1), and the amount can be quantified with a scoring system such as the volumetric score or the Agatston score. The latter, which is more commonly used, is based on the product of the area of the calcium deposits and the x-ray attenuation in Hounsfield units.

Scores can be roughly categorized (with some overlap owing to data from different studies) as:

• Low risk: 0 Agatston units (AU)
• Average risk: 1–112 AU
• Moderate risk: 100–400 AU
• High risk: 400–999 AU
• Very high risk: 1,000 AU.²

The actual test takes only a few seconds, and the patient can usually be out the door in 15 minutes or less. It does not require iodinated contrast and the radiation dose is minimal, usually less than 1 mSv, equivalent to fewer than 10 chest radiographs.3

The cost is typically between $200 and $500. The test is usually not covered by health insurance, but this differs by insurer and by state; for example, coverage is mandated in Texas, and the test is covered by United Healthcare.

WHAT IS THE EVIDENCE IN FAVOR OF CALCIFICATION SCORING?

Cohort studies with long-term follow-up show that calcification scoring has robust prognostic ability. A pooled analysis of several of these studies² showed that a higher score strongly correlated with a higher risk of cardiac events over 3 to 5 years. Compared with the risk in people with a score of 0, the risk was twice as high in those with a score of 1 to 112, four times as high with a score of 100 to 400, seven times as high with a score of 400 to 499, and 10 times as high with a score greater than 1,000.²

A cohort study of more than 25,000 patients had similar conclusions about the magnitude of risk associated with coronary calcification.⁴ It also found that the 10-year risk of death was 0.6% in patients with a score of 0, 3.4% with a score of 101 to 399, 5.3% with a score of 400 to 699, 6.1% with a score of 700 to 999, and 12.2% with a score greater than 1,000.

Although progression of coronary artery calcification may predict the risk of death from any cause,⁵ the clinical utility of serial measurements is not yet apparent, especially since statin therapy—our front-line treatment for coronary disease—has not been shown to slow the progression of calcification.

If a patient’s 10-year coronary risk is intermediate (10% to 20%), calcification scoring can reclassify the risk as low or high in about 50% of cases and can improve the accuracy of risk prediction.^6–8

For example, Elias-Smale et al⁶ evaluated the effect of calcification scoring in 2,028 asymptomatic patients, with median follow-up of 9.2 years and 135 coronary events observed. Adding the calcification score to the Framingham model significantly improved risk classification, with a net reclassification improvement (NRI) of 0.14 (P < .01). (NRI is a measure of discriminatory performance for a diagnostic test; higher is better.⁹) Reclassification was most robust in those at intermediate risk, 52% of whom were reclassified, with 30% reclassified to low risk and 22% reclassified to high risk.

Erbel et al⁷ reported data from the Heinz Nixdorf Recall study, which used calcification scoring to estimate the NRI in 4,129 patients followed for 5 years. During this time there were 93 coronary deaths and non-fatal myocardial infarctions. The addition of the calcification score to the Framingham risk model resulted in an NRI of 0.21 (P = .0002) for patients with a risk of 6% to 20% and 0.31 (P < .0001) for those with a risk of 10% to 20%. Erbel et al also estimated the C statistic (area under the receiver operating characteristic curve; the maximum value is 1.0 and the higher the value the better) for the addition of the calcification score to the Framingham risk model and to the Adult Treatment Panel (ATP) III algorithm. They reported a significant increase of 0.681 to 0.749 with the Framingham model and 0.653 to 0.755 with the ATP III algorithm.

Polonsky et al⁸ studied a cohort of 5,878 participants from the Multi-Ethnic Study of Atherosclerosis (MESA) and estimated the event risk using a model based on Framingham risk characteristics. When the calcification score was added to the prediction model, 26% of the sample was reclassified to a new risk category. In intermediate-risk patients, 292 (16%) were reclassified as high risk, and 712 (39%) were reclassified as low risk, achieving an NRI of 0.55 (95% confidence interval 0.41 to 0.69; P < .001). In addition, the C statistic for the prediction of cardiovascular events was 0.76 for the model based on Framingham risk characteristics and increased to 0.81 (P < .001) with the addition of calcification scoring.

Improving adherence and care

Knowing that a patient has a higher calcification score, physicians are more likely to prescribe lipid-lowering and antihypertensive drugs (Table 1),^10–12 and patients with a higher score are also more often adherent to recommendations regarding diet and exercise.¹³

Rozanski et al,¹⁴ in a randomized controlled trial, showed that measuring coronary artery calcification did not increase downstream medical spending. A modest improvement in systolic blood pressure (P = .02), serum low-density lipoprotein level (P = .04), and waist circumference (P = .01) was observed in patients who had their calcification measured. Patients with the highest scores had the greatest improvement in coronary risk factors, including blood pressure, cholesterol, weight, and regular exercise.

On the other hand, other analyses have suggested that imaging tests are not effective for motivating behavioral changes. This topic deserves more research.¹⁵

Less utility in symptomatic disease

Coronary artery calcification scoring has less clinical utility in patients who already have coronary symptoms. Villines et al¹⁶ described a cohort of 10,037 patients with coronary symptoms who underwent calcification scoring and computed tomographic coronary angiography and found that stenosis of greater than 50% was present in 3.5% of those who had a score of 0 and in 29% of those with a score higher than 0. Therefore, a score of 0 does not rule out obstructive coronary heart disease if the patient has symptoms. Conversely, these patients may still have coronary artery calcification even if perfusion stress imaging is normal,^17,18 and calcification scoring may have a role in the evaluation of equivocal stress tests.¹⁹

CALCIFICATION SCORING GUIDELINES

In their most recent (2010) joint guidelines for assessing risk of coronary heart disease in asymptomatic patients,²⁰ the American College of Cardiology and the American Heart Association say coronary artery calcification scoring:

• Is recommended for asymptomatic patients at intermediate 10-year risk (10% to 20%) of coronary heart disease (class IIa recommendation, level of evidence B)
• May be acceptable for asymptomatic patients at low to intermediate risk (6% to 10%) (class IIb recommendation)
• Is discouraged for those at low risk (< 6%) (class III recommendation).

The most recent (2010) criteria for the appropriate use of cardiac computed tomography²¹ provide similar recommendations. Specifically, coronary artery calcification scoring with noncontrast computed tomography was rated as appropriate for patients at intermediate risk (10% to 20%) of coronary heart disease and for the specific subset of patients who are at low risk (6% to 10%) but who have a family history of premature coronary heart disease.

These recommendations are based on multiple lines of evidence that calcification scoring is a robust risk-predictor, can enhance risk estimates beyond traditional scoring strategies, and may—in theory—improve outcomes.

CALCIFICATION SCORING’S LIMITATIONS

The images used for measuring coronary calcification do predict risk of cardiovascular events, but they are not adequate to assess the severity of coronary stenosis. Further, calcification scoring often leads to incidental findings, which can cause anxiety and possibly lead to more imaging, entailing more radiation exposure and expense. And as noted, there are no randomized trial data demonstrating a reduction in cardiovascular events with the use of calcification scoring.

Although we still have no evidence from randomized trials that patients have better outcomes if we measure the calcification in their coronary arteries, a growing body of evidence shows that we can estimate risk more accurately than with a risk model score alone if we also score coronary artery calcification in asymptomatic patients, especially those at intermediate risk.

See related editorial

Current guidelines¹ recommend using the Framingham Risk Score or a similar tool to estimate coronary risk in asymptomatic patients, but these tools have only modest accuracy. Calcification scoring is accurate, inexpensive, quick, widely available, low-risk, and does not appear to increase medical costs afterward. In addition to improving risk stratification, it may also encourage patients to adhere better to drug therapy and lifestyle modification.

HOW IS CORONARY ARTERY CALCIFICATION MEASURED?

Calcification of the coronary arteries is synonymous with atherosclerosis. It can easily be detected with computed tomography without contrast (Figure 1), and the amount can be quantified with a scoring system such as the volumetric score or the Agatston score. The latter, which is more commonly used, is based on the product of the area of the calcium deposits and the x-ray attenuation in Hounsfield units.

Scores can be roughly categorized (with some overlap owing to data from different studies) as:

• Low risk: 0 Agatston units (AU)
• Average risk: 1–112 AU
• Moderate risk: 100–400 AU
• High risk: 400–999 AU
• Very high risk: 1,000 AU.²

The actual test takes only a few seconds, and the patient can usually be out the door in 15 minutes or less. It does not require iodinated contrast and the radiation dose is minimal, usually less than 1 mSv, equivalent to fewer than 10 chest radiographs.3

The cost is typically between $200 and $500. The test is usually not covered by health insurance, but this differs by insurer and by state; for example, coverage is mandated in Texas, and the test is covered by United Healthcare.

WHAT IS THE EVIDENCE IN FAVOR OF CALCIFICATION SCORING?

Cohort studies with long-term follow-up show that calcification scoring has robust prognostic ability. A pooled analysis of several of these studies² showed that a higher score strongly correlated with a higher risk of cardiac events over 3 to 5 years. Compared with the risk in people with a score of 0, the risk was twice as high in those with a score of 1 to 112, four times as high with a score of 100 to 400, seven times as high with a score of 400 to 499, and 10 times as high with a score greater than 1,000.²

A cohort study of more than 25,000 patients had similar conclusions about the magnitude of risk associated with coronary calcification.⁴ It also found that the 10-year risk of death was 0.6% in patients with a score of 0, 3.4% with a score of 101 to 399, 5.3% with a score of 400 to 699, 6.1% with a score of 700 to 999, and 12.2% with a score greater than 1,000.

Although progression of coronary artery calcification may predict the risk of death from any cause,⁵ the clinical utility of serial measurements is not yet apparent, especially since statin therapy—our front-line treatment for coronary disease—has not been shown to slow the progression of calcification.

If a patient’s 10-year coronary risk is intermediate (10% to 20%), calcification scoring can reclassify the risk as low or high in about 50% of cases and can improve the accuracy of risk prediction.^6–8

For example, Elias-Smale et al⁶ evaluated the effect of calcification scoring in 2,028 asymptomatic patients, with median follow-up of 9.2 years and 135 coronary events observed. Adding the calcification score to the Framingham model significantly improved risk classification, with a net reclassification improvement (NRI) of 0.14 (P < .01). (NRI is a measure of discriminatory performance for a diagnostic test; higher is better.⁹) Reclassification was most robust in those at intermediate risk, 52% of whom were reclassified, with 30% reclassified to low risk and 22% reclassified to high risk.

Erbel et al⁷ reported data from the Heinz Nixdorf Recall study, which used calcification scoring to estimate the NRI in 4,129 patients followed for 5 years. During this time there were 93 coronary deaths and non-fatal myocardial infarctions. The addition of the calcification score to the Framingham risk model resulted in an NRI of 0.21 (P = .0002) for patients with a risk of 6% to 20% and 0.31 (P < .0001) for those with a risk of 10% to 20%. Erbel et al also estimated the C statistic (area under the receiver operating characteristic curve; the maximum value is 1.0 and the higher the value the better) for the addition of the calcification score to the Framingham risk model and to the Adult Treatment Panel (ATP) III algorithm. They reported a significant increase of 0.681 to 0.749 with the Framingham model and 0.653 to 0.755 with the ATP III algorithm.

Polonsky et al⁸ studied a cohort of 5,878 participants from the Multi-Ethnic Study of Atherosclerosis (MESA) and estimated the event risk using a model based on Framingham risk characteristics. When the calcification score was added to the prediction model, 26% of the sample was reclassified to a new risk category. In intermediate-risk patients, 292 (16%) were reclassified as high risk, and 712 (39%) were reclassified as low risk, achieving an NRI of 0.55 (95% confidence interval 0.41 to 0.69; P < .001). In addition, the C statistic for the prediction of cardiovascular events was 0.76 for the model based on Framingham risk characteristics and increased to 0.81 (P < .001) with the addition of calcification scoring.

Improving adherence and care

Knowing that a patient has a higher calcification score, physicians are more likely to prescribe lipid-lowering and antihypertensive drugs (Table 1),^10–12 and patients with a higher score are also more often adherent to recommendations regarding diet and exercise.¹³

Rozanski et al,¹⁴ in a randomized controlled trial, showed that measuring coronary artery calcification did not increase downstream medical spending. A modest improvement in systolic blood pressure (P = .02), serum low-density lipoprotein level (P = .04), and waist circumference (P = .01) was observed in patients who had their calcification measured. Patients with the highest scores had the greatest improvement in coronary risk factors, including blood pressure, cholesterol, weight, and regular exercise.

On the other hand, other analyses have suggested that imaging tests are not effective for motivating behavioral changes. This topic deserves more research.¹⁵

Less utility in symptomatic disease

Coronary artery calcification scoring has less clinical utility in patients who already have coronary symptoms. Villines et al¹⁶ described a cohort of 10,037 patients with coronary symptoms who underwent calcification scoring and computed tomographic coronary angiography and found that stenosis of greater than 50% was present in 3.5% of those who had a score of 0 and in 29% of those with a score higher than 0. Therefore, a score of 0 does not rule out obstructive coronary heart disease if the patient has symptoms. Conversely, these patients may still have coronary artery calcification even if perfusion stress imaging is normal,^17,18 and calcification scoring may have a role in the evaluation of equivocal stress tests.¹⁹

CALCIFICATION SCORING GUIDELINES

In their most recent (2010) joint guidelines for assessing risk of coronary heart disease in asymptomatic patients,²⁰ the American College of Cardiology and the American Heart Association say coronary artery calcification scoring:

• Is recommended for asymptomatic patients at intermediate 10-year risk (10% to 20%) of coronary heart disease (class IIa recommendation, level of evidence B)
• May be acceptable for asymptomatic patients at low to intermediate risk (6% to 10%) (class IIb recommendation)
• Is discouraged for those at low risk (< 6%) (class III recommendation).

The most recent (2010) criteria for the appropriate use of cardiac computed tomography²¹ provide similar recommendations. Specifically, coronary artery calcification scoring with noncontrast computed tomography was rated as appropriate for patients at intermediate risk (10% to 20%) of coronary heart disease and for the specific subset of patients who are at low risk (6% to 10%) but who have a family history of premature coronary heart disease.

These recommendations are based on multiple lines of evidence that calcification scoring is a robust risk-predictor, can enhance risk estimates beyond traditional scoring strategies, and may—in theory—improve outcomes.

CALCIFICATION SCORING’S LIMITATIONS

The images used for measuring coronary calcification do predict risk of cardiovascular events, but they are not adequate to assess the severity of coronary stenosis. Further, calcification scoring often leads to incidental findings, which can cause anxiety and possibly lead to more imaging, entailing more radiation exposure and expense. And as noted, there are no randomized trial data demonstrating a reduction in cardiovascular events with the use of calcification scoring.

Although we still have no evidence from randomized trials that patients have better outcomes if we measure the calcification in their coronary arteries, a growing body of evidence shows that we can estimate risk more accurately than with a risk model score alone if we also score coronary artery calcification in asymptomatic patients, especially those at intermediate risk.

See related editorial

Current guidelines¹ recommend using the Framingham Risk Score or a similar tool to estimate coronary risk in asymptomatic patients, but these tools have only modest accuracy. Calcification scoring is accurate, inexpensive, quick, widely available, low-risk, and does not appear to increase medical costs afterward. In addition to improving risk stratification, it may also encourage patients to adhere better to drug therapy and lifestyle modification.

HOW IS CORONARY ARTERY CALCIFICATION MEASURED?

Calcification of the coronary arteries is synonymous with atherosclerosis. It can easily be detected with computed tomography without contrast (Figure 1), and the amount can be quantified with a scoring system such as the volumetric score or the Agatston score. The latter, which is more commonly used, is based on the product of the area of the calcium deposits and the x-ray attenuation in Hounsfield units.

Scores can be roughly categorized (with some overlap owing to data from different studies) as:

• Low risk: 0 Agatston units (AU)
• Average risk: 1–112 AU
• Moderate risk: 100–400 AU
• High risk: 400–999 AU
• Very high risk: 1,000 AU.²

The actual test takes only a few seconds, and the patient can usually be out the door in 15 minutes or less. It does not require iodinated contrast and the radiation dose is minimal, usually less than 1 mSv, equivalent to fewer than 10 chest radiographs.3

The cost is typically between $200 and $500. The test is usually not covered by health insurance, but this differs by insurer and by state; for example, coverage is mandated in Texas, and the test is covered by United Healthcare.

WHAT IS THE EVIDENCE IN FAVOR OF CALCIFICATION SCORING?

Cohort studies with long-term follow-up show that calcification scoring has robust prognostic ability. A pooled analysis of several of these studies² showed that a higher score strongly correlated with a higher risk of cardiac events over 3 to 5 years. Compared with the risk in people with a score of 0, the risk was twice as high in those with a score of 1 to 112, four times as high with a score of 100 to 400, seven times as high with a score of 400 to 499, and 10 times as high with a score greater than 1,000.²

A cohort study of more than 25,000 patients had similar conclusions about the magnitude of risk associated with coronary calcification.⁴ It also found that the 10-year risk of death was 0.6% in patients with a score of 0, 3.4% with a score of 101 to 399, 5.3% with a score of 400 to 699, 6.1% with a score of 700 to 999, and 12.2% with a score greater than 1,000.

Although progression of coronary artery calcification may predict the risk of death from any cause,⁵ the clinical utility of serial measurements is not yet apparent, especially since statin therapy—our front-line treatment for coronary disease—has not been shown to slow the progression of calcification.

If a patient’s 10-year coronary risk is intermediate (10% to 20%), calcification scoring can reclassify the risk as low or high in about 50% of cases and can improve the accuracy of risk prediction.^6–8

For example, Elias-Smale et al⁶ evaluated the effect of calcification scoring in 2,028 asymptomatic patients, with median follow-up of 9.2 years and 135 coronary events observed. Adding the calcification score to the Framingham model significantly improved risk classification, with a net reclassification improvement (NRI) of 0.14 (P < .01). (NRI is a measure of discriminatory performance for a diagnostic test; higher is better.⁹) Reclassification was most robust in those at intermediate risk, 52% of whom were reclassified, with 30% reclassified to low risk and 22% reclassified to high risk.

Erbel et al⁷ reported data from the Heinz Nixdorf Recall study, which used calcification scoring to estimate the NRI in 4,129 patients followed for 5 years. During this time there were 93 coronary deaths and non-fatal myocardial infarctions. The addition of the calcification score to the Framingham risk model resulted in an NRI of 0.21 (P = .0002) for patients with a risk of 6% to 20% and 0.31 (P < .0001) for those with a risk of 10% to 20%. Erbel et al also estimated the C statistic (area under the receiver operating characteristic curve; the maximum value is 1.0 and the higher the value the better) for the addition of the calcification score to the Framingham risk model and to the Adult Treatment Panel (ATP) III algorithm. They reported a significant increase of 0.681 to 0.749 with the Framingham model and 0.653 to 0.755 with the ATP III algorithm.

Polonsky et al⁸ studied a cohort of 5,878 participants from the Multi-Ethnic Study of Atherosclerosis (MESA) and estimated the event risk using a model based on Framingham risk characteristics. When the calcification score was added to the prediction model, 26% of the sample was reclassified to a new risk category. In intermediate-risk patients, 292 (16%) were reclassified as high risk, and 712 (39%) were reclassified as low risk, achieving an NRI of 0.55 (95% confidence interval 0.41 to 0.69; P < .001). In addition, the C statistic for the prediction of cardiovascular events was 0.76 for the model based on Framingham risk characteristics and increased to 0.81 (P < .001) with the addition of calcification scoring.

Improving adherence and care

Knowing that a patient has a higher calcification score, physicians are more likely to prescribe lipid-lowering and antihypertensive drugs (Table 1),^10–12 and patients with a higher score are also more often adherent to recommendations regarding diet and exercise.¹³

Rozanski et al,¹⁴ in a randomized controlled trial, showed that measuring coronary artery calcification did not increase downstream medical spending. A modest improvement in systolic blood pressure (P = .02), serum low-density lipoprotein level (P = .04), and waist circumference (P = .01) was observed in patients who had their calcification measured. Patients with the highest scores had the greatest improvement in coronary risk factors, including blood pressure, cholesterol, weight, and regular exercise.

On the other hand, other analyses have suggested that imaging tests are not effective for motivating behavioral changes. This topic deserves more research.¹⁵

Less utility in symptomatic disease

Coronary artery calcification scoring has less clinical utility in patients who already have coronary symptoms. Villines et al¹⁶ described a cohort of 10,037 patients with coronary symptoms who underwent calcification scoring and computed tomographic coronary angiography and found that stenosis of greater than 50% was present in 3.5% of those who had a score of 0 and in 29% of those with a score higher than 0. Therefore, a score of 0 does not rule out obstructive coronary heart disease if the patient has symptoms. Conversely, these patients may still have coronary artery calcification even if perfusion stress imaging is normal,^17,18 and calcification scoring may have a role in the evaluation of equivocal stress tests.¹⁹

CALCIFICATION SCORING GUIDELINES

In their most recent (2010) joint guidelines for assessing risk of coronary heart disease in asymptomatic patients,²⁰ the American College of Cardiology and the American Heart Association say coronary artery calcification scoring:

• Is recommended for asymptomatic patients at intermediate 10-year risk (10% to 20%) of coronary heart disease (class IIa recommendation, level of evidence B)
• May be acceptable for asymptomatic patients at low to intermediate risk (6% to 10%) (class IIb recommendation)
• Is discouraged for those at low risk (< 6%) (class III recommendation).

The most recent (2010) criteria for the appropriate use of cardiac computed tomography²¹ provide similar recommendations. Specifically, coronary artery calcification scoring with noncontrast computed tomography was rated as appropriate for patients at intermediate risk (10% to 20%) of coronary heart disease and for the specific subset of patients who are at low risk (6% to 10%) but who have a family history of premature coronary heart disease.

These recommendations are based on multiple lines of evidence that calcification scoring is a robust risk-predictor, can enhance risk estimates beyond traditional scoring strategies, and may—in theory—improve outcomes.

CALCIFICATION SCORING’S LIMITATIONS

The images used for measuring coronary calcification do predict risk of cardiovascular events, but they are not adequate to assess the severity of coronary stenosis. Further, calcification scoring often leads to incidental findings, which can cause anxiety and possibly lead to more imaging, entailing more radiation exposure and expense. And as noted, there are no randomized trial data demonstrating a reduction in cardiovascular events with the use of calcification scoring.

To try to identify and treat people who are at highest risk of cardiovascular events, including death, we use comprehensive risk-prediction models. Unfortunately, these models have limited accuracy and precision and do not predict very well.

See related article

Attractive, then, is the idea of using a noninvasive imaging test to measure coronary atherosclerosis before it causes trouble and thereby individualize the risk assessment. Noncontrast computed tomography (CT) can measure the amount of calcification in the coronary arteries, and therefore it can estimate the coronary atherosclerotic burden. It seems like an ideal test, and calcification as a marker of subclinical atherosclerosis has been extensively investigated.

However, despite more than 2 decades of use and data from hundreds of thousands of patients, the test remains poorly understood. Many physicians seem to use it solely as a means of placating “worried well” patients and do not truly appreciate its implications. Others proceed to ordering CT angiography, a more expensive test that involves the added risks of using higher x-ray doses and iodinated contrast, even when a correctly interpreted calcification score would provide ample information.

In this issue of the Cleveland Clinic Journal of Medicine, Chauffe and Winchester review the utility of coronary artery calcification scoring in current practice. We wish to supplement their review by suggesting some considerations to take into account before ordering this test:

• Does the patient have symptoms of coronary artery disease, and what is his or her risk-factor profile? Baseline patient characteristics are important to consider if we are to use this test appropriately.
• How should the result be interpreted, and does the ordering physician have the confidence to accept the result?

BEST USED IN ASYMPTOMATIC PATIENTS AT INTERMEDIATE RISK

Many large retrospective and prospective registries have demonstrated the predictive value of coronary artery calcification in diverse cohorts of patients without symptoms.

In three prospective registries—the Multi-Ethnic Study of Atherosclerosis¹ (MESA) with 6,722 patients, the Coronary CT Angiography Evaluation for Clinical Outcomes² (CONFIRM) with 7,590 patients, and the Heinz Nixdorf Recall (NHR) study³ with 4,129 patients—most of the patients who had heart attacks had a calcification score greater than 100. And conversely, data from more than 100,000 people show that the absence of calcification (ie, a score of 0) denotes a very low risk (< 1% over 5 years).^1–6

The pretest probability of coronary artery disease needs to be considered. The data clearly indicate that a Bayesian approach is warranted and that coronary artery calcification scoring should mainly be done in patients at intermediate or low-intermediate risk. Trials have shown that calcification scoring will reclassify more than 50% of intermediate-risk patients into the high-risk or low-risk category.³

The implications of these findings were eloquently assessed in the Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER). In this trial, it was estimated that for patients with no calcification who would otherwise fulfill the criteria for treatment with a statin, 549 patients would need to be treated to prevent one coronary event, compared with 24 similar patients with a calcification score greater than 100.⁷

Although such analyses have potential shortcomings, in this era of greater concern about how to allocate finite resources, using a simple, inexpensive test to individualize long-term treatment is an attractive idea. Further, measuring calcification does not appear to increase testing “downstream” and indeed reduces it as compared with no calcification scoring. It also results in better adherence to drug therapy and lifestyle changes.

Because calcification scoring provides additional prognostic data and accurately discriminates and reclassifies risk, the American College of Cardiology and the American Heart Association have awarded it a class IIa recommendation for asymptomatic patients at intermediate risk, meaning that there is conflicting evidence or a divergence of opinion about its usefulness, but the weight of evidence or opinion favors it.⁸

ITS ROLE IS MORE CONTROVERSIAL IN SYMPTOMATIC PATIENTS

Perhaps a less established and more controversial use of coronary artery calcification scoring is in patients who are having coronary symptoms. In patients at high cardiovascular risk, this test by itself may miss an unacceptable number of those who truly have significant stenoses.⁹ However, when the appropriate population is selected, there is substantial evidence that it can be an important means of risk stratification.

In patients at low to intermediate risk, the absence of calcification indicates a very low likelihood of significant coronary artery stenosis, as demonstrated in the Coronary CT Angiography Evaluation for Clinical Outcomes: An International Multicenter (CONFIRM) registry.¹⁰ In the 10,037 symptomatic patients evaluated, a score of 0 had a 99% negative predictive value for excluding stenosis greater than 70% and was associated with a 2-year event rate less than 1%. These data were supported by a meta-analysis of nearly 1,000 symptomatic patients with a score of 0, in whom the 2-year event rate was less than 2%.⁴

Taken together, these data suggest that the absence of coronary calcification in people at low to intermediate risk indicates a very low likelihood of significant stenotic coronary artery disease and foretells an excellent prognosis.

These data have already been incorporated into the British National Institute for Health and Clinical Excellence (NICE) guidelines, in which calcification scoring is an integral part of the management algorithm in patients with chest pain who are at low risk.

WHY NOT JUST DO CT ANGIOGRAPHY?

But why bother with coronary artery calcification scoring when we can do CT angiography instead? The angiography scanners we have today can cover the entire heart in a single gantry rotation. Dual-source scanners provide temporal resolution as low as 75 ms, and sequential, prospective electrocardiographic gating and iterative reconstruction can routinely achieve scans with doses of radiation as low as 3 mSv that provide coronary artery images of exquisite quality.

On the other hand, calcification scoring is fast and easy to perform and poses less potential harm to the patient, since it uses lower doses of radiation and no contrast agents. In addition, the quantification is semi-automated, so the results can be interpreted quickly and are reproducible.

In the CONFIRM trial, prediction by CT angiography was no better than calcification scoring in asymptomatic patients, so it is not recommended in this population.² In symptomatic patients, the CONFIRM trial data suggest that almost 1,000 additional CT angiography procedures would need to be done to identify one myocardial infarction and more than 1,500 procedures to identify one patient at risk of death missed by calcification scoring of 0 in patients at low to intermediate risk.¹¹

Chauffe and Winchester nicely summarize the limitations of calcification scoring. However, we would emphasize the potential implications of the above findings. Appropriately utilized, calcification scoring is safe, reproducible, and inexpensive and helps individualize treatment in asymptomatic patients at low to intermediate risk, thereby avoiding under- and overtreatment and potentially reducing downstream costs while improving compliance.

In patients at low to intermediate risk who present with chest pain, documenting the absence of calcification can rationalize downstream testing and reliably, quickly, and safely permit patient discharge from emergency departments. In a time of increasing costs and patient demands and finite resources, clinicians should remain cognizant of the usefulness of evaluating coronary artery calcification.

To try to identify and treat people who are at highest risk of cardiovascular events, including death, we use comprehensive risk-prediction models. Unfortunately, these models have limited accuracy and precision and do not predict very well.

See related article

Attractive, then, is the idea of using a noninvasive imaging test to measure coronary atherosclerosis before it causes trouble and thereby individualize the risk assessment. Noncontrast computed tomography (CT) can measure the amount of calcification in the coronary arteries, and therefore it can estimate the coronary atherosclerotic burden. It seems like an ideal test, and calcification as a marker of subclinical atherosclerosis has been extensively investigated.

However, despite more than 2 decades of use and data from hundreds of thousands of patients, the test remains poorly understood. Many physicians seem to use it solely as a means of placating “worried well” patients and do not truly appreciate its implications. Others proceed to ordering CT angiography, a more expensive test that involves the added risks of using higher x-ray doses and iodinated contrast, even when a correctly interpreted calcification score would provide ample information.

In this issue of the Cleveland Clinic Journal of Medicine, Chauffe and Winchester review the utility of coronary artery calcification scoring in current practice. We wish to supplement their review by suggesting some considerations to take into account before ordering this test:

• Does the patient have symptoms of coronary artery disease, and what is his or her risk-factor profile? Baseline patient characteristics are important to consider if we are to use this test appropriately.
• How should the result be interpreted, and does the ordering physician have the confidence to accept the result?

BEST USED IN ASYMPTOMATIC PATIENTS AT INTERMEDIATE RISK

Many large retrospective and prospective registries have demonstrated the predictive value of coronary artery calcification in diverse cohorts of patients without symptoms.

In three prospective registries—the Multi-Ethnic Study of Atherosclerosis¹ (MESA) with 6,722 patients, the Coronary CT Angiography Evaluation for Clinical Outcomes² (CONFIRM) with 7,590 patients, and the Heinz Nixdorf Recall (NHR) study³ with 4,129 patients—most of the patients who had heart attacks had a calcification score greater than 100. And conversely, data from more than 100,000 people show that the absence of calcification (ie, a score of 0) denotes a very low risk (< 1% over 5 years).^1–6

The pretest probability of coronary artery disease needs to be considered. The data clearly indicate that a Bayesian approach is warranted and that coronary artery calcification scoring should mainly be done in patients at intermediate or low-intermediate risk. Trials have shown that calcification scoring will reclassify more than 50% of intermediate-risk patients into the high-risk or low-risk category.³

The implications of these findings were eloquently assessed in the Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER). In this trial, it was estimated that for patients with no calcification who would otherwise fulfill the criteria for treatment with a statin, 549 patients would need to be treated to prevent one coronary event, compared with 24 similar patients with a calcification score greater than 100.⁷

Although such analyses have potential shortcomings, in this era of greater concern about how to allocate finite resources, using a simple, inexpensive test to individualize long-term treatment is an attractive idea. Further, measuring calcification does not appear to increase testing “downstream” and indeed reduces it as compared with no calcification scoring. It also results in better adherence to drug therapy and lifestyle changes.

Because calcification scoring provides additional prognostic data and accurately discriminates and reclassifies risk, the American College of Cardiology and the American Heart Association have awarded it a class IIa recommendation for asymptomatic patients at intermediate risk, meaning that there is conflicting evidence or a divergence of opinion about its usefulness, but the weight of evidence or opinion favors it.⁸

ITS ROLE IS MORE CONTROVERSIAL IN SYMPTOMATIC PATIENTS

Perhaps a less established and more controversial use of coronary artery calcification scoring is in patients who are having coronary symptoms. In patients at high cardiovascular risk, this test by itself may miss an unacceptable number of those who truly have significant stenoses.⁹ However, when the appropriate population is selected, there is substantial evidence that it can be an important means of risk stratification.

In patients at low to intermediate risk, the absence of calcification indicates a very low likelihood of significant coronary artery stenosis, as demonstrated in the Coronary CT Angiography Evaluation for Clinical Outcomes: An International Multicenter (CONFIRM) registry.¹⁰ In the 10,037 symptomatic patients evaluated, a score of 0 had a 99% negative predictive value for excluding stenosis greater than 70% and was associated with a 2-year event rate less than 1%. These data were supported by a meta-analysis of nearly 1,000 symptomatic patients with a score of 0, in whom the 2-year event rate was less than 2%.⁴

Taken together, these data suggest that the absence of coronary calcification in people at low to intermediate risk indicates a very low likelihood of significant stenotic coronary artery disease and foretells an excellent prognosis.

These data have already been incorporated into the British National Institute for Health and Clinical Excellence (NICE) guidelines, in which calcification scoring is an integral part of the management algorithm in patients with chest pain who are at low risk.

WHY NOT JUST DO CT ANGIOGRAPHY?

But why bother with coronary artery calcification scoring when we can do CT angiography instead? The angiography scanners we have today can cover the entire heart in a single gantry rotation. Dual-source scanners provide temporal resolution as low as 75 ms, and sequential, prospective electrocardiographic gating and iterative reconstruction can routinely achieve scans with doses of radiation as low as 3 mSv that provide coronary artery images of exquisite quality.

On the other hand, calcification scoring is fast and easy to perform and poses less potential harm to the patient, since it uses lower doses of radiation and no contrast agents. In addition, the quantification is semi-automated, so the results can be interpreted quickly and are reproducible.

In the CONFIRM trial, prediction by CT angiography was no better than calcification scoring in asymptomatic patients, so it is not recommended in this population.² In symptomatic patients, the CONFIRM trial data suggest that almost 1,000 additional CT angiography procedures would need to be done to identify one myocardial infarction and more than 1,500 procedures to identify one patient at risk of death missed by calcification scoring of 0 in patients at low to intermediate risk.¹¹

Chauffe and Winchester nicely summarize the limitations of calcification scoring. However, we would emphasize the potential implications of the above findings. Appropriately utilized, calcification scoring is safe, reproducible, and inexpensive and helps individualize treatment in asymptomatic patients at low to intermediate risk, thereby avoiding under- and overtreatment and potentially reducing downstream costs while improving compliance.

In patients at low to intermediate risk who present with chest pain, documenting the absence of calcification can rationalize downstream testing and reliably, quickly, and safely permit patient discharge from emergency departments. In a time of increasing costs and patient demands and finite resources, clinicians should remain cognizant of the usefulness of evaluating coronary artery calcification.

To try to identify and treat people who are at highest risk of cardiovascular events, including death, we use comprehensive risk-prediction models. Unfortunately, these models have limited accuracy and precision and do not predict very well.

See related article

Attractive, then, is the idea of using a noninvasive imaging test to measure coronary atherosclerosis before it causes trouble and thereby individualize the risk assessment. Noncontrast computed tomography (CT) can measure the amount of calcification in the coronary arteries, and therefore it can estimate the coronary atherosclerotic burden. It seems like an ideal test, and calcification as a marker of subclinical atherosclerosis has been extensively investigated.

However, despite more than 2 decades of use and data from hundreds of thousands of patients, the test remains poorly understood. Many physicians seem to use it solely as a means of placating “worried well” patients and do not truly appreciate its implications. Others proceed to ordering CT angiography, a more expensive test that involves the added risks of using higher x-ray doses and iodinated contrast, even when a correctly interpreted calcification score would provide ample information.

In this issue of the Cleveland Clinic Journal of Medicine, Chauffe and Winchester review the utility of coronary artery calcification scoring in current practice. We wish to supplement their review by suggesting some considerations to take into account before ordering this test:

• Does the patient have symptoms of coronary artery disease, and what is his or her risk-factor profile? Baseline patient characteristics are important to consider if we are to use this test appropriately.
• How should the result be interpreted, and does the ordering physician have the confidence to accept the result?

BEST USED IN ASYMPTOMATIC PATIENTS AT INTERMEDIATE RISK

Many large retrospective and prospective registries have demonstrated the predictive value of coronary artery calcification in diverse cohorts of patients without symptoms.

In three prospective registries—the Multi-Ethnic Study of Atherosclerosis¹ (MESA) with 6,722 patients, the Coronary CT Angiography Evaluation for Clinical Outcomes² (CONFIRM) with 7,590 patients, and the Heinz Nixdorf Recall (NHR) study³ with 4,129 patients—most of the patients who had heart attacks had a calcification score greater than 100. And conversely, data from more than 100,000 people show that the absence of calcification (ie, a score of 0) denotes a very low risk (< 1% over 5 years).^1–6

The pretest probability of coronary artery disease needs to be considered. The data clearly indicate that a Bayesian approach is warranted and that coronary artery calcification scoring should mainly be done in patients at intermediate or low-intermediate risk. Trials have shown that calcification scoring will reclassify more than 50% of intermediate-risk patients into the high-risk or low-risk category.³

The implications of these findings were eloquently assessed in the Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER). In this trial, it was estimated that for patients with no calcification who would otherwise fulfill the criteria for treatment with a statin, 549 patients would need to be treated to prevent one coronary event, compared with 24 similar patients with a calcification score greater than 100.⁷

Although such analyses have potential shortcomings, in this era of greater concern about how to allocate finite resources, using a simple, inexpensive test to individualize long-term treatment is an attractive idea. Further, measuring calcification does not appear to increase testing “downstream” and indeed reduces it as compared with no calcification scoring. It also results in better adherence to drug therapy and lifestyle changes.

Because calcification scoring provides additional prognostic data and accurately discriminates and reclassifies risk, the American College of Cardiology and the American Heart Association have awarded it a class IIa recommendation for asymptomatic patients at intermediate risk, meaning that there is conflicting evidence or a divergence of opinion about its usefulness, but the weight of evidence or opinion favors it.⁸

ITS ROLE IS MORE CONTROVERSIAL IN SYMPTOMATIC PATIENTS

Perhaps a less established and more controversial use of coronary artery calcification scoring is in patients who are having coronary symptoms. In patients at high cardiovascular risk, this test by itself may miss an unacceptable number of those who truly have significant stenoses.⁹ However, when the appropriate population is selected, there is substantial evidence that it can be an important means of risk stratification.

In patients at low to intermediate risk, the absence of calcification indicates a very low likelihood of significant coronary artery stenosis, as demonstrated in the Coronary CT Angiography Evaluation for Clinical Outcomes: An International Multicenter (CONFIRM) registry.¹⁰ In the 10,037 symptomatic patients evaluated, a score of 0 had a 99% negative predictive value for excluding stenosis greater than 70% and was associated with a 2-year event rate less than 1%. These data were supported by a meta-analysis of nearly 1,000 symptomatic patients with a score of 0, in whom the 2-year event rate was less than 2%.⁴

Taken together, these data suggest that the absence of coronary calcification in people at low to intermediate risk indicates a very low likelihood of significant stenotic coronary artery disease and foretells an excellent prognosis.

These data have already been incorporated into the British National Institute for Health and Clinical Excellence (NICE) guidelines, in which calcification scoring is an integral part of the management algorithm in patients with chest pain who are at low risk.

WHY NOT JUST DO CT ANGIOGRAPHY?

But why bother with coronary artery calcification scoring when we can do CT angiography instead? The angiography scanners we have today can cover the entire heart in a single gantry rotation. Dual-source scanners provide temporal resolution as low as 75 ms, and sequential, prospective electrocardiographic gating and iterative reconstruction can routinely achieve scans with doses of radiation as low as 3 mSv that provide coronary artery images of exquisite quality.

On the other hand, calcification scoring is fast and easy to perform and poses less potential harm to the patient, since it uses lower doses of radiation and no contrast agents. In addition, the quantification is semi-automated, so the results can be interpreted quickly and are reproducible.

In the CONFIRM trial, prediction by CT angiography was no better than calcification scoring in asymptomatic patients, so it is not recommended in this population.² In symptomatic patients, the CONFIRM trial data suggest that almost 1,000 additional CT angiography procedures would need to be done to identify one myocardial infarction and more than 1,500 procedures to identify one patient at risk of death missed by calcification scoring of 0 in patients at low to intermediate risk.¹¹

Chauffe and Winchester nicely summarize the limitations of calcification scoring. However, we would emphasize the potential implications of the above findings. Appropriately utilized, calcification scoring is safe, reproducible, and inexpensive and helps individualize treatment in asymptomatic patients at low to intermediate risk, thereby avoiding under- and overtreatment and potentially reducing downstream costs while improving compliance.

In patients at low to intermediate risk who present with chest pain, documenting the absence of calcification can rationalize downstream testing and reliably, quickly, and safely permit patient discharge from emergency departments. In a time of increasing costs and patient demands and finite resources, clinicians should remain cognizant of the usefulness of evaluating coronary artery calcification.

A 74-year-old man presented to the emergency department in December 2011 with a 1-week history of worsening abdominal pain, nausea with emesis, and decreased appetite. The pain was dull, diffuse, and not related to oral intake or bowel movements. He denied any bloody stools, melena, or hematemesis, but he had not had a bowel movement in the past week.

He was already known to have stage IV colon cancer with metastases to the lungs and liver. He had undergone a partial colectomy in 2009 and was receiving chemotherapy at the time of admission.

He also had an infrarenal abdominal aortic aneurysm that had been repaired in 2003 with endovascular placement of a Gore Excluder stent graft. This was complicated by a type II endoleak, treated with coil embolization. The same endoleak later recurred and was treated with injection of Onyx liquid embolic agent.

His medical history also included hypertension, type 2 diabetes mellitus, and hyperlipidemia. He had undergone a laparoscopic cholecystectomy in 2007.

He denied any fevers, chills, headache, lightheadedness, or change in vision. He had no respiratory, cardiac, or urinary symptoms. He had been constipated for the past few weeks and had recently been started on a bowel regimen, with mild relief. There had been no other changes to his medications.

His temperature on presentation was 97.5°F (36.4°C), blood pressure 120/64 mm Hg, pulse 96, respiratory rate 22, and oxygen saturation 95% on room air. He was awake, alert, oriented, and in no acute distress. His mucous membranes were dry. His lungs were clear to auscultation, and his heart sounds were normal. His bowel sounds were hyperactive and his abdomen was slightly tender diffusely, but there was no abdominal distention, rebound tenderness, guarding, or palpable masses. His joints, muscle strength, and muscle tone were normal. Table 1 shows his initial laboratory values.

Given the patient’s history of colon cancer, the emergency department physician ordered computed tomography (CT) of the abdomen to assess the state of his disease and to evaluate for bowel obstruction. The scan revealed a large abdominal aortic aneurysm with foci of gas within the aneurysmal sac. Metastases in the liver, lung, and retroperitoneum appeared stable; abundant colonic stool suggested constipation (Figure 1).

CAUSES OF PERIAORTIC GAS AFTER ANEURYSM REPAIR

1. What is the most common cause of periaortic ectopic gas in a patient with a repaired abdominal aortic aneurysm?

• Endoleak
• Stent graft infection
• Retroperitoneal fibrosis
• Aortoenteric fistula

Endoleak

Endoleak, a complication of endovascular abdominal aortic aneurysm repair, is defined as blood flow within the aneurysm sac but outside the endoluminal graft.¹ It occurs in up to 15% of patients after endograft placement in the first month alone, and in up to 47% of patients eventually.² It can lead to aneurysm enlargement and rupture. Endoleaks are classified into five types, each with different causes and management options.^3,4 Contrast-enhanced CT is the most commonly used diagnostic tool.⁵

Endoleak cannot be ruled out in our patient, since CT was done without contrast. However, gas within the aneurysm is not consistent with this diagnosis.

Infection has been reported in 1% to 6% of patients receiving a stent graft for aortic aneurysm.⁶ They occur most commonly in the first year after placement; one study showed that 42% of patients diagnosed with graft infection presented within 3 months of endovascular repair.⁷

The leading cause of graft infection is contamination during the original procedure, but secondary infection from hematologic seeding and contamination from adjacent bowel are also possible.⁸ In our patient, who underwent graft placement followed by endovascular repairs of endoleaks, bacterial seeding of his aortic aneurysm from the procedures should be considered.⁹

The most common organisms are staphylococcal species, with Staphylococcus aureus more common in early infection and coagulase-negative staphylococci more common in late infection.¹⁰ Methicillin-resistant S aureus has been reported in as many as 25% of cases of graft infection. Diphtheroids and gram-negative enteric organisms should also be considered.¹¹

CT is the most effective imaging test for graft infection. Perigraft soft tissue, fluid, and gas are the major CT findings.¹²

Given that our patient presented with abdominal pain, leukocytosis, and the CT finding of perigraft gas, graft infection should be high on our list differential diagnoses.

Retroperitoneal fibrosis

Retroperitoneal fibrosis is most often idiopathic, although many believe it is due to an exaggerated local inflammatory reaction to aortic atherosclerosis or is a manifestation of a systemic autoimmune disease.¹³ Secondary retroperitoneal fibrosis may be due to drugs, infection, or malignancy.

Pathologic findings include sclerotic plaques, typically around the abdominal vessels and ureters. Clinical presentations are often nonspecific, with early symptoms that include back or abdominal pain, malaise, anorexia, edema, and hematuria.^14,15 Progressive ureteral obstruction can occur in later stages. CT with contrast is the imaging test of choice to visualize the extent of disease, with the fibrosis exhibiting attenuation similar to that of muscle.¹⁶

Initial treatment of idiopathic retroperitoneal fibrosis is with a glucocorticoid or other immunosuppressive agent, whereas treatment of secondary retroperitoneal fibrosis is aimed at the underlying cause.¹⁷ Late stages complicated by ureteral obstruction typically require surgery.¹⁸

Our patient did have some nonspecific complaints that could be due to retroperitoneal fibrosis. He also had an intra-abdominal malignancy, which could lead to secondary retroperitoneal fibrosis. However, his CT findings of periaortic gas are not consistent with this diagnosis.¹⁹

Aortoenteric fistula

Aortoenteric fistulas can be either primary or secondary.

Primary aortoenteric fistulas occur de novo in patients who have never undergone any surgery or procedure in the aorta. This type of fistula usually results from pressure erosion of an atherosclerotic abdominal aortic aneurysm into the gastrointestinal tract. They are rare, with an annual incidence of 0.04% to 0.07% in the general population.^20,21

Secondary aortoenteric fistulas are complications of aortic reconstructive therapy. After open repair, a perianastomotic or pseudoaneurysmal fistula can develop into the gastrointestinal tract.⁴ Endovascular repair leaves the aortic wall intact with no exposed suture lines, but an aortoenteric fistula can still develop²² and in fact occur in 0.4% to 3.1% of recipients of stent grafts for abdominal aortic aneurysm repair.²³ In such cases, it is commonly thought that graft infection can lead to formation of an aortoenteric fistula, but a penetrating gastrointestinal ulcer, tumor invasion, radiation therapy, and trauma have also been implicated.^19,24–26 An aortoenteric fistula can present several months to several years after either open or endovascular abdominal aortic aneurysm repair.^4,23

One of the main CT signs of an aortoenteric fistula is periaortic ectopic gas at least 3 to 4 weeks after surgery or endovascular repair.¹⁹ Gas around the stent graft is most commonly caused by infection, but an aortoenteric fistula must also be considered in our patient, as roughly one-third of graft infections present as aortoenteric fistula.²⁷ Our patient denied having any gastrointestinal bleeding, but his hemoglobin concentration at presentation was 8.9 g/dL.

Highlight point. Perigraft gas after abdominal aortic aneurysm repair can be seen in graft infection and aortoenteric fistula.

SIGNS AND SYMPTOMS OF AORTOENTERIC FISTULA

2. What is the most common clinical sign or symptom of an aortoenteric fistula?

• Gastrointestinal bleeding
• Sepsis
• Abdominal pain
• Back pain

Gastrointestinal bleeding occurs in 80% of patients who have an aortoenteric fistula, sepsis in 40%, abdominal pain in 30%, and back pain in 15%.¹⁹ The classic triad of symptoms is gastrointestinal bleeding, abdominal pain, and a pulsatile abdominal mass. However, symptoms can vary widely, and the classic triad is present in fewer than 25% of cases.²⁸ Sepsis may be the predominant clinical manifestation, particularly in the early stages of fistula formation. Unexplained fever is an underrecognized early manifestation.²⁴

Highlight point. The classic triad of symptoms of an aortoenteric fistula (gastrointestinal bleeding, abdominal pain, and a pulsatile abdominal mass) is seen in fewer than 25% of cases.

Case continued: The patient develops frank bleeding

The vascular surgery service was consulted because of concern for an aortic graft infection, since surgical removal of the infected material is recommended.¹⁰ The patient was deemed to be a poor surgical candidate, given his stage IV colon cancer, so he was treated conservatively with broad-spectrum antibiotics.

Over the next 2 days, he had two episodes of dark, bloody bowel movements, but he remained hemodynamically stable. He subsequently developed frank bleeding per rectum with symptoms of lightheadedness, and his hemoglobin concentration fell to 6.9 g/dL. He was given a total of 3 units of packed red blood cells, which raised his hemoglobin level, but only to 8.3 g/dL. The gastroenterology service was consulted to evaluate for the source of the bleeding.

Comment. In a situation like this, an aortoenteric fistula is high on our list of differential diagnoses as the cause of bleeding, but other causes of frank bleeding per rectum such as diverticulosis, arteriovenous malformation, hemorrhoids, or a rapid upper-gastrointestinal bleed cannot be ruled out.

Upper-gastrointestinal endoscopy is the most commonly used diagnostic test for aortoenteric fistulas. It can also find other possible sources of gastrointestinal bleeding. CT with contrast is another option. It can depict the fistula itself or reveal signs of infection, such as gas or liquid surrounding the graft. In an emergency, when there is not enough time for diagnostic testing and an aortoenteric fistula is strongly suspected on clinical grounds, emergency surgical exploration is warranted.^4,24

In our patient, the gastrointestinal service elected to first perform endoscopy to look for an aortoenteric fistula.

WHERE DO AORTOENTERIC FISTULAS OCCUR?

3. In which part of the gastrointestinal tract is an aortoenteric fistula most commonly located?

• Esophagus
• Stomach
• Duodenum
• Jejunum

Aortoenteric fistulas can occur at any of these locations, but 80% of cases of secondary aortoenteric fistula are in the duodenum, most often in the third or fourth (horizontal or ascending) part.¹⁹ Endoscopic visualization of a pulsatile bleeding mass in this area is diagnostic. However, even if no fistula is seen, upper endoscopy cannot rule out an aortoenteric fistula because the lesion can be located more distal than the scope can reach, which is typically no farther than the first or second parts.^4,24

Case continued: What endoscopy showed

The esophagus was normal. There was old clotted blood in the stomach, but no lesions or ulcers. The duodenal bulb and second portion of the duodenum were normal. Three ulcers were noted in the third and fourth portions of the duodenum. The largest and deepest ulcer had an adherent blood clot, and the bowel wall was pulsatile in this region (Figure 2). These findings revealed the source of the gastrointestinal bleeding and were consistent with an aortoenteric fistula.

The patient’s initial bloody bowel movements were herald bleeds, ie, transient and self-limited episodes resulting from necrosis and mucosal ulceration. Herald bleeds can precede a massive gastrointestinal hemorrhage resulting from a true aortoenteric communication.¹⁹

Highlight point. Herald bleeds are self-limited and precede hemorrhage that results from a true aortoenteric communication.

TREATMENT OF AORTOENTERIC FISTULA

4. How are aortoenteric fistulas treated?

• Surgery
• Antibiotics
• Endoscopic intervention

Surgery is the definitive treatment. The traditional procedure is open surgical resection of the affected portion of the aorta followed by extra-anatomic (axillobifemoral) bypass or in situ aortic reconstruction using an antibiotic-impregnated prosthetic graft, autogenous femoral vein graft, or cryopreserved allograft.^9,29 There have been cases of successful endovascular repair of aortoenteric fistulas, but this approach is generally used as a palliative bridge to definitive surgery.³⁰

Antibiotics should be used if graft infection is suspected, ie, in most cases. However, surgery is still needed to repair the fistula and remove the source of infection. Cultures taken during surgical repair can help guide the choice of antibiotic after surgery.

Endoscopy can aid in diagnosing an aortoenteric fistula, as in the case of our patient. However, vascular surgery is necessary to close the communication between the aorta and the gastrointestinal tract.

Case continued: The patient declines treatment

In view of the patient’s enteroscopic findings, the vascular surgery service was again consulted for surgical correction of the aortoenteric fistula. Treatment was discussed with the patient and his family, but they declined any intervention in view of the high risk of morbidity and death that surgery would entail. Nearing the end of life with advanced cancer and a newly diagnosed aortoenteric fistula, the patient preferred comfort measures with hospice care.

Take-home points

Abdominal pain is the reason for 5% to 10% of emergency department visits, and between 35% to 41% of patients admitted to the hospital because of abdominal pain do not have a definitive diagnosis.³¹ It is crucial to think about an aortoenteric fistula in such patients who have a history of abdominal aortic aneurysm repair and gastrointestinal bleeding. Timely diagnosis and intervention are necessary to manage this otherwise-fatal condition.

A 74-year-old man presented to the emergency department in December 2011 with a 1-week history of worsening abdominal pain, nausea with emesis, and decreased appetite. The pain was dull, diffuse, and not related to oral intake or bowel movements. He denied any bloody stools, melena, or hematemesis, but he had not had a bowel movement in the past week.

He was already known to have stage IV colon cancer with metastases to the lungs and liver. He had undergone a partial colectomy in 2009 and was receiving chemotherapy at the time of admission.

He also had an infrarenal abdominal aortic aneurysm that had been repaired in 2003 with endovascular placement of a Gore Excluder stent graft. This was complicated by a type II endoleak, treated with coil embolization. The same endoleak later recurred and was treated with injection of Onyx liquid embolic agent.

His medical history also included hypertension, type 2 diabetes mellitus, and hyperlipidemia. He had undergone a laparoscopic cholecystectomy in 2007.

He denied any fevers, chills, headache, lightheadedness, or change in vision. He had no respiratory, cardiac, or urinary symptoms. He had been constipated for the past few weeks and had recently been started on a bowel regimen, with mild relief. There had been no other changes to his medications.

His temperature on presentation was 97.5°F (36.4°C), blood pressure 120/64 mm Hg, pulse 96, respiratory rate 22, and oxygen saturation 95% on room air. He was awake, alert, oriented, and in no acute distress. His mucous membranes were dry. His lungs were clear to auscultation, and his heart sounds were normal. His bowel sounds were hyperactive and his abdomen was slightly tender diffusely, but there was no abdominal distention, rebound tenderness, guarding, or palpable masses. His joints, muscle strength, and muscle tone were normal. Table 1 shows his initial laboratory values.

Given the patient’s history of colon cancer, the emergency department physician ordered computed tomography (CT) of the abdomen to assess the state of his disease and to evaluate for bowel obstruction. The scan revealed a large abdominal aortic aneurysm with foci of gas within the aneurysmal sac. Metastases in the liver, lung, and retroperitoneum appeared stable; abundant colonic stool suggested constipation (Figure 1).

CAUSES OF PERIAORTIC GAS AFTER ANEURYSM REPAIR

1. What is the most common cause of periaortic ectopic gas in a patient with a repaired abdominal aortic aneurysm?

• Endoleak
• Stent graft infection
• Retroperitoneal fibrosis
• Aortoenteric fistula

Endoleak

Endoleak, a complication of endovascular abdominal aortic aneurysm repair, is defined as blood flow within the aneurysm sac but outside the endoluminal graft.¹ It occurs in up to 15% of patients after endograft placement in the first month alone, and in up to 47% of patients eventually.² It can lead to aneurysm enlargement and rupture. Endoleaks are classified into five types, each with different causes and management options.^3,4 Contrast-enhanced CT is the most commonly used diagnostic tool.⁵

Endoleak cannot be ruled out in our patient, since CT was done without contrast. However, gas within the aneurysm is not consistent with this diagnosis.

Infection has been reported in 1% to 6% of patients receiving a stent graft for aortic aneurysm.⁶ They occur most commonly in the first year after placement; one study showed that 42% of patients diagnosed with graft infection presented within 3 months of endovascular repair.⁷

The leading cause of graft infection is contamination during the original procedure, but secondary infection from hematologic seeding and contamination from adjacent bowel are also possible.⁸ In our patient, who underwent graft placement followed by endovascular repairs of endoleaks, bacterial seeding of his aortic aneurysm from the procedures should be considered.⁹

The most common organisms are staphylococcal species, with Staphylococcus aureus more common in early infection and coagulase-negative staphylococci more common in late infection.¹⁰ Methicillin-resistant S aureus has been reported in as many as 25% of cases of graft infection. Diphtheroids and gram-negative enteric organisms should also be considered.¹¹

CT is the most effective imaging test for graft infection. Perigraft soft tissue, fluid, and gas are the major CT findings.¹²

Given that our patient presented with abdominal pain, leukocytosis, and the CT finding of perigraft gas, graft infection should be high on our list differential diagnoses.

Retroperitoneal fibrosis

Retroperitoneal fibrosis is most often idiopathic, although many believe it is due to an exaggerated local inflammatory reaction to aortic atherosclerosis or is a manifestation of a systemic autoimmune disease.¹³ Secondary retroperitoneal fibrosis may be due to drugs, infection, or malignancy.

Pathologic findings include sclerotic plaques, typically around the abdominal vessels and ureters. Clinical presentations are often nonspecific, with early symptoms that include back or abdominal pain, malaise, anorexia, edema, and hematuria.^14,15 Progressive ureteral obstruction can occur in later stages. CT with contrast is the imaging test of choice to visualize the extent of disease, with the fibrosis exhibiting attenuation similar to that of muscle.¹⁶

Initial treatment of idiopathic retroperitoneal fibrosis is with a glucocorticoid or other immunosuppressive agent, whereas treatment of secondary retroperitoneal fibrosis is aimed at the underlying cause.¹⁷ Late stages complicated by ureteral obstruction typically require surgery.¹⁸

Our patient did have some nonspecific complaints that could be due to retroperitoneal fibrosis. He also had an intra-abdominal malignancy, which could lead to secondary retroperitoneal fibrosis. However, his CT findings of periaortic gas are not consistent with this diagnosis.¹⁹

Aortoenteric fistula

Aortoenteric fistulas can be either primary or secondary.

Primary aortoenteric fistulas occur de novo in patients who have never undergone any surgery or procedure in the aorta. This type of fistula usually results from pressure erosion of an atherosclerotic abdominal aortic aneurysm into the gastrointestinal tract. They are rare, with an annual incidence of 0.04% to 0.07% in the general population.^20,21

Secondary aortoenteric fistulas are complications of aortic reconstructive therapy. After open repair, a perianastomotic or pseudoaneurysmal fistula can develop into the gastrointestinal tract.⁴ Endovascular repair leaves the aortic wall intact with no exposed suture lines, but an aortoenteric fistula can still develop²² and in fact occur in 0.4% to 3.1% of recipients of stent grafts for abdominal aortic aneurysm repair.²³ In such cases, it is commonly thought that graft infection can lead to formation of an aortoenteric fistula, but a penetrating gastrointestinal ulcer, tumor invasion, radiation therapy, and trauma have also been implicated.^19,24–26 An aortoenteric fistula can present several months to several years after either open or endovascular abdominal aortic aneurysm repair.^4,23

One of the main CT signs of an aortoenteric fistula is periaortic ectopic gas at least 3 to 4 weeks after surgery or endovascular repair.¹⁹ Gas around the stent graft is most commonly caused by infection, but an aortoenteric fistula must also be considered in our patient, as roughly one-third of graft infections present as aortoenteric fistula.²⁷ Our patient denied having any gastrointestinal bleeding, but his hemoglobin concentration at presentation was 8.9 g/dL.

Highlight point. Perigraft gas after abdominal aortic aneurysm repair can be seen in graft infection and aortoenteric fistula.

SIGNS AND SYMPTOMS OF AORTOENTERIC FISTULA

2. What is the most common clinical sign or symptom of an aortoenteric fistula?

• Gastrointestinal bleeding
• Sepsis
• Abdominal pain
• Back pain

Gastrointestinal bleeding occurs in 80% of patients who have an aortoenteric fistula, sepsis in 40%, abdominal pain in 30%, and back pain in 15%.¹⁹ The classic triad of symptoms is gastrointestinal bleeding, abdominal pain, and a pulsatile abdominal mass. However, symptoms can vary widely, and the classic triad is present in fewer than 25% of cases.²⁸ Sepsis may be the predominant clinical manifestation, particularly in the early stages of fistula formation. Unexplained fever is an underrecognized early manifestation.²⁴

Highlight point. The classic triad of symptoms of an aortoenteric fistula (gastrointestinal bleeding, abdominal pain, and a pulsatile abdominal mass) is seen in fewer than 25% of cases.

Case continued: The patient develops frank bleeding

The vascular surgery service was consulted because of concern for an aortic graft infection, since surgical removal of the infected material is recommended.¹⁰ The patient was deemed to be a poor surgical candidate, given his stage IV colon cancer, so he was treated conservatively with broad-spectrum antibiotics.

Over the next 2 days, he had two episodes of dark, bloody bowel movements, but he remained hemodynamically stable. He subsequently developed frank bleeding per rectum with symptoms of lightheadedness, and his hemoglobin concentration fell to 6.9 g/dL. He was given a total of 3 units of packed red blood cells, which raised his hemoglobin level, but only to 8.3 g/dL. The gastroenterology service was consulted to evaluate for the source of the bleeding.

Comment. In a situation like this, an aortoenteric fistula is high on our list of differential diagnoses as the cause of bleeding, but other causes of frank bleeding per rectum such as diverticulosis, arteriovenous malformation, hemorrhoids, or a rapid upper-gastrointestinal bleed cannot be ruled out.

Upper-gastrointestinal endoscopy is the most commonly used diagnostic test for aortoenteric fistulas. It can also find other possible sources of gastrointestinal bleeding. CT with contrast is another option. It can depict the fistula itself or reveal signs of infection, such as gas or liquid surrounding the graft. In an emergency, when there is not enough time for diagnostic testing and an aortoenteric fistula is strongly suspected on clinical grounds, emergency surgical exploration is warranted.^4,24

In our patient, the gastrointestinal service elected to first perform endoscopy to look for an aortoenteric fistula.

WHERE DO AORTOENTERIC FISTULAS OCCUR?

3. In which part of the gastrointestinal tract is an aortoenteric fistula most commonly located?

• Esophagus
• Stomach
• Duodenum
• Jejunum

Aortoenteric fistulas can occur at any of these locations, but 80% of cases of secondary aortoenteric fistula are in the duodenum, most often in the third or fourth (horizontal or ascending) part.¹⁹ Endoscopic visualization of a pulsatile bleeding mass in this area is diagnostic. However, even if no fistula is seen, upper endoscopy cannot rule out an aortoenteric fistula because the lesion can be located more distal than the scope can reach, which is typically no farther than the first or second parts.^4,24

Case continued: What endoscopy showed

The esophagus was normal. There was old clotted blood in the stomach, but no lesions or ulcers. The duodenal bulb and second portion of the duodenum were normal. Three ulcers were noted in the third and fourth portions of the duodenum. The largest and deepest ulcer had an adherent blood clot, and the bowel wall was pulsatile in this region (Figure 2). These findings revealed the source of the gastrointestinal bleeding and were consistent with an aortoenteric fistula.

The patient’s initial bloody bowel movements were herald bleeds, ie, transient and self-limited episodes resulting from necrosis and mucosal ulceration. Herald bleeds can precede a massive gastrointestinal hemorrhage resulting from a true aortoenteric communication.¹⁹

Highlight point. Herald bleeds are self-limited and precede hemorrhage that results from a true aortoenteric communication.

TREATMENT OF AORTOENTERIC FISTULA

4. How are aortoenteric fistulas treated?

• Surgery
• Antibiotics
• Endoscopic intervention

Surgery is the definitive treatment. The traditional procedure is open surgical resection of the affected portion of the aorta followed by extra-anatomic (axillobifemoral) bypass or in situ aortic reconstruction using an antibiotic-impregnated prosthetic graft, autogenous femoral vein graft, or cryopreserved allograft.^9,29 There have been cases of successful endovascular repair of aortoenteric fistulas, but this approach is generally used as a palliative bridge to definitive surgery.³⁰

Antibiotics should be used if graft infection is suspected, ie, in most cases. However, surgery is still needed to repair the fistula and remove the source of infection. Cultures taken during surgical repair can help guide the choice of antibiotic after surgery.

Endoscopy can aid in diagnosing an aortoenteric fistula, as in the case of our patient. However, vascular surgery is necessary to close the communication between the aorta and the gastrointestinal tract.

Case continued: The patient declines treatment

In view of the patient’s enteroscopic findings, the vascular surgery service was again consulted for surgical correction of the aortoenteric fistula. Treatment was discussed with the patient and his family, but they declined any intervention in view of the high risk of morbidity and death that surgery would entail. Nearing the end of life with advanced cancer and a newly diagnosed aortoenteric fistula, the patient preferred comfort measures with hospice care.

Take-home points

Abdominal pain is the reason for 5% to 10% of emergency department visits, and between 35% to 41% of patients admitted to the hospital because of abdominal pain do not have a definitive diagnosis.³¹ It is crucial to think about an aortoenteric fistula in such patients who have a history of abdominal aortic aneurysm repair and gastrointestinal bleeding. Timely diagnosis and intervention are necessary to manage this otherwise-fatal condition.

A 74-year-old man presented to the emergency department in December 2011 with a 1-week history of worsening abdominal pain, nausea with emesis, and decreased appetite. The pain was dull, diffuse, and not related to oral intake or bowel movements. He denied any bloody stools, melena, or hematemesis, but he had not had a bowel movement in the past week.

He was already known to have stage IV colon cancer with metastases to the lungs and liver. He had undergone a partial colectomy in 2009 and was receiving chemotherapy at the time of admission.

He also had an infrarenal abdominal aortic aneurysm that had been repaired in 2003 with endovascular placement of a Gore Excluder stent graft. This was complicated by a type II endoleak, treated with coil embolization. The same endoleak later recurred and was treated with injection of Onyx liquid embolic agent.

His medical history also included hypertension, type 2 diabetes mellitus, and hyperlipidemia. He had undergone a laparoscopic cholecystectomy in 2007.

He denied any fevers, chills, headache, lightheadedness, or change in vision. He had no respiratory, cardiac, or urinary symptoms. He had been constipated for the past few weeks and had recently been started on a bowel regimen, with mild relief. There had been no other changes to his medications.

His temperature on presentation was 97.5°F (36.4°C), blood pressure 120/64 mm Hg, pulse 96, respiratory rate 22, and oxygen saturation 95% on room air. He was awake, alert, oriented, and in no acute distress. His mucous membranes were dry. His lungs were clear to auscultation, and his heart sounds were normal. His bowel sounds were hyperactive and his abdomen was slightly tender diffusely, but there was no abdominal distention, rebound tenderness, guarding, or palpable masses. His joints, muscle strength, and muscle tone were normal. Table 1 shows his initial laboratory values.

Given the patient’s history of colon cancer, the emergency department physician ordered computed tomography (CT) of the abdomen to assess the state of his disease and to evaluate for bowel obstruction. The scan revealed a large abdominal aortic aneurysm with foci of gas within the aneurysmal sac. Metastases in the liver, lung, and retroperitoneum appeared stable; abundant colonic stool suggested constipation (Figure 1).

CAUSES OF PERIAORTIC GAS AFTER ANEURYSM REPAIR

1. What is the most common cause of periaortic ectopic gas in a patient with a repaired abdominal aortic aneurysm?

• Endoleak
• Stent graft infection
• Retroperitoneal fibrosis
• Aortoenteric fistula

Endoleak

Endoleak, a complication of endovascular abdominal aortic aneurysm repair, is defined as blood flow within the aneurysm sac but outside the endoluminal graft.¹ It occurs in up to 15% of patients after endograft placement in the first month alone, and in up to 47% of patients eventually.² It can lead to aneurysm enlargement and rupture. Endoleaks are classified into five types, each with different causes and management options.^3,4 Contrast-enhanced CT is the most commonly used diagnostic tool.⁵

Endoleak cannot be ruled out in our patient, since CT was done without contrast. However, gas within the aneurysm is not consistent with this diagnosis.

Infection has been reported in 1% to 6% of patients receiving a stent graft for aortic aneurysm.⁶ They occur most commonly in the first year after placement; one study showed that 42% of patients diagnosed with graft infection presented within 3 months of endovascular repair.⁷

The leading cause of graft infection is contamination during the original procedure, but secondary infection from hematologic seeding and contamination from adjacent bowel are also possible.⁸ In our patient, who underwent graft placement followed by endovascular repairs of endoleaks, bacterial seeding of his aortic aneurysm from the procedures should be considered.⁹

The most common organisms are staphylococcal species, with Staphylococcus aureus more common in early infection and coagulase-negative staphylococci more common in late infection.¹⁰ Methicillin-resistant S aureus has been reported in as many as 25% of cases of graft infection. Diphtheroids and gram-negative enteric organisms should also be considered.¹¹

CT is the most effective imaging test for graft infection. Perigraft soft tissue, fluid, and gas are the major CT findings.¹²

Given that our patient presented with abdominal pain, leukocytosis, and the CT finding of perigraft gas, graft infection should be high on our list differential diagnoses.

Retroperitoneal fibrosis

Retroperitoneal fibrosis is most often idiopathic, although many believe it is due to an exaggerated local inflammatory reaction to aortic atherosclerosis or is a manifestation of a systemic autoimmune disease.¹³ Secondary retroperitoneal fibrosis may be due to drugs, infection, or malignancy.

Pathologic findings include sclerotic plaques, typically around the abdominal vessels and ureters. Clinical presentations are often nonspecific, with early symptoms that include back or abdominal pain, malaise, anorexia, edema, and hematuria.^14,15 Progressive ureteral obstruction can occur in later stages. CT with contrast is the imaging test of choice to visualize the extent of disease, with the fibrosis exhibiting attenuation similar to that of muscle.¹⁶

Initial treatment of idiopathic retroperitoneal fibrosis is with a glucocorticoid or other immunosuppressive agent, whereas treatment of secondary retroperitoneal fibrosis is aimed at the underlying cause.¹⁷ Late stages complicated by ureteral obstruction typically require surgery.¹⁸

Our patient did have some nonspecific complaints that could be due to retroperitoneal fibrosis. He also had an intra-abdominal malignancy, which could lead to secondary retroperitoneal fibrosis. However, his CT findings of periaortic gas are not consistent with this diagnosis.¹⁹

Aortoenteric fistula

Aortoenteric fistulas can be either primary or secondary.

Primary aortoenteric fistulas occur de novo in patients who have never undergone any surgery or procedure in the aorta. This type of fistula usually results from pressure erosion of an atherosclerotic abdominal aortic aneurysm into the gastrointestinal tract. They are rare, with an annual incidence of 0.04% to 0.07% in the general population.^20,21

Secondary aortoenteric fistulas are complications of aortic reconstructive therapy. After open repair, a perianastomotic or pseudoaneurysmal fistula can develop into the gastrointestinal tract.⁴ Endovascular repair leaves the aortic wall intact with no exposed suture lines, but an aortoenteric fistula can still develop²² and in fact occur in 0.4% to 3.1% of recipients of stent grafts for abdominal aortic aneurysm repair.²³ In such cases, it is commonly thought that graft infection can lead to formation of an aortoenteric fistula, but a penetrating gastrointestinal ulcer, tumor invasion, radiation therapy, and trauma have also been implicated.^19,24–26 An aortoenteric fistula can present several months to several years after either open or endovascular abdominal aortic aneurysm repair.^4,23

One of the main CT signs of an aortoenteric fistula is periaortic ectopic gas at least 3 to 4 weeks after surgery or endovascular repair.¹⁹ Gas around the stent graft is most commonly caused by infection, but an aortoenteric fistula must also be considered in our patient, as roughly one-third of graft infections present as aortoenteric fistula.²⁷ Our patient denied having any gastrointestinal bleeding, but his hemoglobin concentration at presentation was 8.9 g/dL.

Highlight point. Perigraft gas after abdominal aortic aneurysm repair can be seen in graft infection and aortoenteric fistula.

SIGNS AND SYMPTOMS OF AORTOENTERIC FISTULA

2. What is the most common clinical sign or symptom of an aortoenteric fistula?

• Gastrointestinal bleeding
• Sepsis
• Abdominal pain
• Back pain

Gastrointestinal bleeding occurs in 80% of patients who have an aortoenteric fistula, sepsis in 40%, abdominal pain in 30%, and back pain in 15%.¹⁹ The classic triad of symptoms is gastrointestinal bleeding, abdominal pain, and a pulsatile abdominal mass. However, symptoms can vary widely, and the classic triad is present in fewer than 25% of cases.²⁸ Sepsis may be the predominant clinical manifestation, particularly in the early stages of fistula formation. Unexplained fever is an underrecognized early manifestation.²⁴

Highlight point. The classic triad of symptoms of an aortoenteric fistula (gastrointestinal bleeding, abdominal pain, and a pulsatile abdominal mass) is seen in fewer than 25% of cases.

Case continued: The patient develops frank bleeding

The vascular surgery service was consulted because of concern for an aortic graft infection, since surgical removal of the infected material is recommended.¹⁰ The patient was deemed to be a poor surgical candidate, given his stage IV colon cancer, so he was treated conservatively with broad-spectrum antibiotics.

Over the next 2 days, he had two episodes of dark, bloody bowel movements, but he remained hemodynamically stable. He subsequently developed frank bleeding per rectum with symptoms of lightheadedness, and his hemoglobin concentration fell to 6.9 g/dL. He was given a total of 3 units of packed red blood cells, which raised his hemoglobin level, but only to 8.3 g/dL. The gastroenterology service was consulted to evaluate for the source of the bleeding.

Comment. In a situation like this, an aortoenteric fistula is high on our list of differential diagnoses as the cause of bleeding, but other causes of frank bleeding per rectum such as diverticulosis, arteriovenous malformation, hemorrhoids, or a rapid upper-gastrointestinal bleed cannot be ruled out.

Upper-gastrointestinal endoscopy is the most commonly used diagnostic test for aortoenteric fistulas. It can also find other possible sources of gastrointestinal bleeding. CT with contrast is another option. It can depict the fistula itself or reveal signs of infection, such as gas or liquid surrounding the graft. In an emergency, when there is not enough time for diagnostic testing and an aortoenteric fistula is strongly suspected on clinical grounds, emergency surgical exploration is warranted.^4,24

In our patient, the gastrointestinal service elected to first perform endoscopy to look for an aortoenteric fistula.

WHERE DO AORTOENTERIC FISTULAS OCCUR?

3. In which part of the gastrointestinal tract is an aortoenteric fistula most commonly located?

• Esophagus
• Stomach
• Duodenum
• Jejunum

Aortoenteric fistulas can occur at any of these locations, but 80% of cases of secondary aortoenteric fistula are in the duodenum, most often in the third or fourth (horizontal or ascending) part.¹⁹ Endoscopic visualization of a pulsatile bleeding mass in this area is diagnostic. However, even if no fistula is seen, upper endoscopy cannot rule out an aortoenteric fistula because the lesion can be located more distal than the scope can reach, which is typically no farther than the first or second parts.^4,24

Case continued: What endoscopy showed

The esophagus was normal. There was old clotted blood in the stomach, but no lesions or ulcers. The duodenal bulb and second portion of the duodenum were normal. Three ulcers were noted in the third and fourth portions of the duodenum. The largest and deepest ulcer had an adherent blood clot, and the bowel wall was pulsatile in this region (Figure 2). These findings revealed the source of the gastrointestinal bleeding and were consistent with an aortoenteric fistula.

The patient’s initial bloody bowel movements were herald bleeds, ie, transient and self-limited episodes resulting from necrosis and mucosal ulceration. Herald bleeds can precede a massive gastrointestinal hemorrhage resulting from a true aortoenteric communication.¹⁹

Highlight point. Herald bleeds are self-limited and precede hemorrhage that results from a true aortoenteric communication.

TREATMENT OF AORTOENTERIC FISTULA

4. How are aortoenteric fistulas treated?

• Surgery
• Antibiotics
• Endoscopic intervention

Surgery is the definitive treatment. The traditional procedure is open surgical resection of the affected portion of the aorta followed by extra-anatomic (axillobifemoral) bypass or in situ aortic reconstruction using an antibiotic-impregnated prosthetic graft, autogenous femoral vein graft, or cryopreserved allograft.^9,29 There have been cases of successful endovascular repair of aortoenteric fistulas, but this approach is generally used as a palliative bridge to definitive surgery.³⁰

Antibiotics should be used if graft infection is suspected, ie, in most cases. However, surgery is still needed to repair the fistula and remove the source of infection. Cultures taken during surgical repair can help guide the choice of antibiotic after surgery.

Endoscopy can aid in diagnosing an aortoenteric fistula, as in the case of our patient. However, vascular surgery is necessary to close the communication between the aorta and the gastrointestinal tract.

Case continued: The patient declines treatment

In view of the patient’s enteroscopic findings, the vascular surgery service was again consulted for surgical correction of the aortoenteric fistula. Treatment was discussed with the patient and his family, but they declined any intervention in view of the high risk of morbidity and death that surgery would entail. Nearing the end of life with advanced cancer and a newly diagnosed aortoenteric fistula, the patient preferred comfort measures with hospice care.

Take-home points

Abdominal pain is the reason for 5% to 10% of emergency department visits, and between 35% to 41% of patients admitted to the hospital because of abdominal pain do not have a definitive diagnosis.³¹ It is crucial to think about an aortoenteric fistula in such patients who have a history of abdominal aortic aneurysm repair and gastrointestinal bleeding. Timely diagnosis and intervention are necessary to manage this otherwise-fatal condition.

A 45-year-old woman with no chronic medical problems presented to the emergency room with a 1-day history of cramps and paresthesias in both hands and feet, mainly involving the fingers and toes. She said that after an argument with her daughter she began feeling anxious, and this was accompanied by shortness of breath and palpitations as well as generalized weakness, fatigue, and body aches. She also reported nausea and repeated vomiting but no abdominal pain, distention or change in bowel movements. She had had no loss of consciousness, confusion, incontinence, headache, dizziness, diplopia, or facial paresthesia.

She is a cigarette smoker, is alcohol-dependent, but does not use illicit drugs and is not on any medications.

Examination revealed a temperature of 37.1°C (98.8°F), blood pressure 150/75 mm Hg, heart rate 105 bpm, respiratory rate 24 breaths per minute, and oxygen saturation 97% on room air. She appeared very fatigued, thin, and in mild distress due to her cramps. Her mucous membranes were dry, but she had no orthostatic changes. She had noticeable carpopedal spasms (Figure 1), reproducible by inflating a blood-pressure cuff placed on her arm (Trousseau sign) (Figure 2). Also noted was the Chvostek sign—contraction of the ipsilateral facial muscles when the facial nerve is tapped just in front of the ear. The rest of the systemic evaluation was normal. Laboratory investigations were as listed in Table 1. Electrocardiography showed a prolonged QTc interval (0.5 sec). The chest radiograph was normal.

HYPERVENTILATION AND TETANY

The presumptive diagnosis was latent tetany caused by an electrolyte derangement, in this case a combination of hypocalcemia, hypomagnesemia, and hypokalemia as the result of alcohol abuse, repeated vomiting, and hyperventilation brought on by a severe attack of anxiety.

Tetany results from increased excitability of nerves and muscles, leading to painful muscle cramps.^1,2 Typical symptoms include circumoral and distal paresthesias, stiffness, clumsiness, myalgia, carpopedal spasm, laryngospasm, bronchospasm, and generalized seizure. The Chvostek and Trousseau signs help to confirm the diagnosis of tetany.^3,4

The differential diagnosis of carpopedal spasm includes other conditions of involuntary muscle contraction, such as myotonic disorders; myokymia from Isaac syndrome (writhing movements of the muscles under the skin visualized by continuous “rippling” movements of the muscle); stiff-man syndrome (an autoimmune-antiglutamic acid decarboxylase antibody-associated muscle rigidity that waxes and wanes with concurrent spasms); and snake envenomation.

In addition, our patient’s symptoms were probably brought on by hyperventilation. In general, patients with hyperventilation syndrome are young females who display various manifestations of underlying anxiety and can develop tetany even after a brief episode of hyperventilation. At the time of presentation, our patient was found to have mixed respiratory and metabolic alkalosis. The anxiety-induced hyperventilation likely contributed to the respiratory alkalosis. She had no other symptoms or signs to suggest an acute organic respiratory illness such as pulmonary embolism, pneumothorax, or infection. Vomiting likely caused the metabolic alkalosis and hypokalemia.

Tetany is usually triggered by acute hypocalcemia and is uncommon unless the serum ionized calcium concentration falls below 4.3 mg/dL (1.1 mmol/L), which usually corresponds to a serum total calcium concentration of 7.0 to 7.5 mg/dL (1.8 to 1.9 mmol/L). Patients with a gradual onset of hypocalcemia tend to have fewer symptoms.^3,4

Although alkalosis alone can cause tetany, it also enhances tetany by reducing the level of ionized calcium in the serum. Alkalemia causes hypocalcemia by an intravascular chelative mechanism in which the decrease in concentration of hydrogen ions leaves the negatively charged binding sites on albumin available to bind ionized calcium.³

The same happens to the magnesium, a cation with the same size and valence. Significant hypomagnesemia is common in tetanic patients with hyperventilation attacks and may, by itself or in combination with hypocalcemia, cause tetany.^2,5,6 Hypokalemia can develop in patients with hypomagnesemia or metabolic alkalosis and may lead to tetany.^6,7 Furthermore, our patient was dependent on alcohol, and this is known to cause hypomagnesemia from the excessive urinary excretion of magnesium. This defect of alcohol-induced tubular dysfunction is reversible within 4 weeks of abstinence. Magnesium depletion can cause hypocalcemia by producing resistance to parathyroid hormone or by decreasing its secretion, and this occurs in severe hypomagnesemia, ie, when the serum magnesium concentration falls below 1.0 mg/dL (0.4 mmol/L).

IDENTIFY AND TREAT THE UNDERLYING CAUSE

The management of tetany consists of identifying and treating the underlying cause. Infusion of calcium or magnesium is effective as acute therapy for tetany, and, if appropriate, vitamin D supplementation should also be provided.^3,4,7 However, if associated hyperventilation syndrome is present, patients benefit from reassurance and treatment for underlying psychological stress. The traditional treatment of rebreathing into a paper bag is no longer recommended because of the potential risk of hypoxia. Sedatives and antidepressants should be reserved for patients who have not responded to conservative treatment.

Our patient was given an explanation of the condition together with breathing exercises. She received lorazepam and was immediately treated with intravenous hydration, along with intravenous infusion of magnesium, calcium, and potassium. These interventions led to an immediate resolution of her symptoms.

Her low level of intact parathyroid hormone may also have been caused by hypomagnesemia. She was given oral magnesium, potassium, calcium, and vitamin D to continue at home. In addition, she was advised to avoid excessive alcohol consumption and to see us or her primary care doctor should the symptoms recur. As expected, all the laboratory values normalized within 1 month of abstinence from alcohol, and she has been well since.

A 45-year-old woman with no chronic medical problems presented to the emergency room with a 1-day history of cramps and paresthesias in both hands and feet, mainly involving the fingers and toes. She said that after an argument with her daughter she began feeling anxious, and this was accompanied by shortness of breath and palpitations as well as generalized weakness, fatigue, and body aches. She also reported nausea and repeated vomiting but no abdominal pain, distention or change in bowel movements. She had had no loss of consciousness, confusion, incontinence, headache, dizziness, diplopia, or facial paresthesia.

She is a cigarette smoker, is alcohol-dependent, but does not use illicit drugs and is not on any medications.

Examination revealed a temperature of 37.1°C (98.8°F), blood pressure 150/75 mm Hg, heart rate 105 bpm, respiratory rate 24 breaths per minute, and oxygen saturation 97% on room air. She appeared very fatigued, thin, and in mild distress due to her cramps. Her mucous membranes were dry, but she had no orthostatic changes. She had noticeable carpopedal spasms (Figure 1), reproducible by inflating a blood-pressure cuff placed on her arm (Trousseau sign) (Figure 2). Also noted was the Chvostek sign—contraction of the ipsilateral facial muscles when the facial nerve is tapped just in front of the ear. The rest of the systemic evaluation was normal. Laboratory investigations were as listed in Table 1. Electrocardiography showed a prolonged QTc interval (0.5 sec). The chest radiograph was normal.

HYPERVENTILATION AND TETANY

The presumptive diagnosis was latent tetany caused by an electrolyte derangement, in this case a combination of hypocalcemia, hypomagnesemia, and hypokalemia as the result of alcohol abuse, repeated vomiting, and hyperventilation brought on by a severe attack of anxiety.

Tetany results from increased excitability of nerves and muscles, leading to painful muscle cramps.^1,2 Typical symptoms include circumoral and distal paresthesias, stiffness, clumsiness, myalgia, carpopedal spasm, laryngospasm, bronchospasm, and generalized seizure. The Chvostek and Trousseau signs help to confirm the diagnosis of tetany.^3,4

The differential diagnosis of carpopedal spasm includes other conditions of involuntary muscle contraction, such as myotonic disorders; myokymia from Isaac syndrome (writhing movements of the muscles under the skin visualized by continuous “rippling” movements of the muscle); stiff-man syndrome (an autoimmune-antiglutamic acid decarboxylase antibody-associated muscle rigidity that waxes and wanes with concurrent spasms); and snake envenomation.

In addition, our patient’s symptoms were probably brought on by hyperventilation. In general, patients with hyperventilation syndrome are young females who display various manifestations of underlying anxiety and can develop tetany even after a brief episode of hyperventilation. At the time of presentation, our patient was found to have mixed respiratory and metabolic alkalosis. The anxiety-induced hyperventilation likely contributed to the respiratory alkalosis. She had no other symptoms or signs to suggest an acute organic respiratory illness such as pulmonary embolism, pneumothorax, or infection. Vomiting likely caused the metabolic alkalosis and hypokalemia.

Tetany is usually triggered by acute hypocalcemia and is uncommon unless the serum ionized calcium concentration falls below 4.3 mg/dL (1.1 mmol/L), which usually corresponds to a serum total calcium concentration of 7.0 to 7.5 mg/dL (1.8 to 1.9 mmol/L). Patients with a gradual onset of hypocalcemia tend to have fewer symptoms.^3,4

Although alkalosis alone can cause tetany, it also enhances tetany by reducing the level of ionized calcium in the serum. Alkalemia causes hypocalcemia by an intravascular chelative mechanism in which the decrease in concentration of hydrogen ions leaves the negatively charged binding sites on albumin available to bind ionized calcium.³

The same happens to the magnesium, a cation with the same size and valence. Significant hypomagnesemia is common in tetanic patients with hyperventilation attacks and may, by itself or in combination with hypocalcemia, cause tetany.^2,5,6 Hypokalemia can develop in patients with hypomagnesemia or metabolic alkalosis and may lead to tetany.^6,7 Furthermore, our patient was dependent on alcohol, and this is known to cause hypomagnesemia from the excessive urinary excretion of magnesium. This defect of alcohol-induced tubular dysfunction is reversible within 4 weeks of abstinence. Magnesium depletion can cause hypocalcemia by producing resistance to parathyroid hormone or by decreasing its secretion, and this occurs in severe hypomagnesemia, ie, when the serum magnesium concentration falls below 1.0 mg/dL (0.4 mmol/L).

IDENTIFY AND TREAT THE UNDERLYING CAUSE

The management of tetany consists of identifying and treating the underlying cause. Infusion of calcium or magnesium is effective as acute therapy for tetany, and, if appropriate, vitamin D supplementation should also be provided.^3,4,7 However, if associated hyperventilation syndrome is present, patients benefit from reassurance and treatment for underlying psychological stress. The traditional treatment of rebreathing into a paper bag is no longer recommended because of the potential risk of hypoxia. Sedatives and antidepressants should be reserved for patients who have not responded to conservative treatment.

Our patient was given an explanation of the condition together with breathing exercises. She received lorazepam and was immediately treated with intravenous hydration, along with intravenous infusion of magnesium, calcium, and potassium. These interventions led to an immediate resolution of her symptoms.

Her low level of intact parathyroid hormone may also have been caused by hypomagnesemia. She was given oral magnesium, potassium, calcium, and vitamin D to continue at home. In addition, she was advised to avoid excessive alcohol consumption and to see us or her primary care doctor should the symptoms recur. As expected, all the laboratory values normalized within 1 month of abstinence from alcohol, and she has been well since.

A 45-year-old woman with no chronic medical problems presented to the emergency room with a 1-day history of cramps and paresthesias in both hands and feet, mainly involving the fingers and toes. She said that after an argument with her daughter she began feeling anxious, and this was accompanied by shortness of breath and palpitations as well as generalized weakness, fatigue, and body aches. She also reported nausea and repeated vomiting but no abdominal pain, distention or change in bowel movements. She had had no loss of consciousness, confusion, incontinence, headache, dizziness, diplopia, or facial paresthesia.

She is a cigarette smoker, is alcohol-dependent, but does not use illicit drugs and is not on any medications.

Examination revealed a temperature of 37.1°C (98.8°F), blood pressure 150/75 mm Hg, heart rate 105 bpm, respiratory rate 24 breaths per minute, and oxygen saturation 97% on room air. She appeared very fatigued, thin, and in mild distress due to her cramps. Her mucous membranes were dry, but she had no orthostatic changes. She had noticeable carpopedal spasms (Figure 1), reproducible by inflating a blood-pressure cuff placed on her arm (Trousseau sign) (Figure 2). Also noted was the Chvostek sign—contraction of the ipsilateral facial muscles when the facial nerve is tapped just in front of the ear. The rest of the systemic evaluation was normal. Laboratory investigations were as listed in Table 1. Electrocardiography showed a prolonged QTc interval (0.5 sec). The chest radiograph was normal.

HYPERVENTILATION AND TETANY

The presumptive diagnosis was latent tetany caused by an electrolyte derangement, in this case a combination of hypocalcemia, hypomagnesemia, and hypokalemia as the result of alcohol abuse, repeated vomiting, and hyperventilation brought on by a severe attack of anxiety.

Tetany results from increased excitability of nerves and muscles, leading to painful muscle cramps.^1,2 Typical symptoms include circumoral and distal paresthesias, stiffness, clumsiness, myalgia, carpopedal spasm, laryngospasm, bronchospasm, and generalized seizure. The Chvostek and Trousseau signs help to confirm the diagnosis of tetany.^3,4

The differential diagnosis of carpopedal spasm includes other conditions of involuntary muscle contraction, such as myotonic disorders; myokymia from Isaac syndrome (writhing movements of the muscles under the skin visualized by continuous “rippling” movements of the muscle); stiff-man syndrome (an autoimmune-antiglutamic acid decarboxylase antibody-associated muscle rigidity that waxes and wanes with concurrent spasms); and snake envenomation.

In addition, our patient’s symptoms were probably brought on by hyperventilation. In general, patients with hyperventilation syndrome are young females who display various manifestations of underlying anxiety and can develop tetany even after a brief episode of hyperventilation. At the time of presentation, our patient was found to have mixed respiratory and metabolic alkalosis. The anxiety-induced hyperventilation likely contributed to the respiratory alkalosis. She had no other symptoms or signs to suggest an acute organic respiratory illness such as pulmonary embolism, pneumothorax, or infection. Vomiting likely caused the metabolic alkalosis and hypokalemia.

Tetany is usually triggered by acute hypocalcemia and is uncommon unless the serum ionized calcium concentration falls below 4.3 mg/dL (1.1 mmol/L), which usually corresponds to a serum total calcium concentration of 7.0 to 7.5 mg/dL (1.8 to 1.9 mmol/L). Patients with a gradual onset of hypocalcemia tend to have fewer symptoms.^3,4

Although alkalosis alone can cause tetany, it also enhances tetany by reducing the level of ionized calcium in the serum. Alkalemia causes hypocalcemia by an intravascular chelative mechanism in which the decrease in concentration of hydrogen ions leaves the negatively charged binding sites on albumin available to bind ionized calcium.³

The same happens to the magnesium, a cation with the same size and valence. Significant hypomagnesemia is common in tetanic patients with hyperventilation attacks and may, by itself or in combination with hypocalcemia, cause tetany.^2,5,6 Hypokalemia can develop in patients with hypomagnesemia or metabolic alkalosis and may lead to tetany.^6,7 Furthermore, our patient was dependent on alcohol, and this is known to cause hypomagnesemia from the excessive urinary excretion of magnesium. This defect of alcohol-induced tubular dysfunction is reversible within 4 weeks of abstinence. Magnesium depletion can cause hypocalcemia by producing resistance to parathyroid hormone or by decreasing its secretion, and this occurs in severe hypomagnesemia, ie, when the serum magnesium concentration falls below 1.0 mg/dL (0.4 mmol/L).

IDENTIFY AND TREAT THE UNDERLYING CAUSE

The management of tetany consists of identifying and treating the underlying cause. Infusion of calcium or magnesium is effective as acute therapy for tetany, and, if appropriate, vitamin D supplementation should also be provided.^3,4,7 However, if associated hyperventilation syndrome is present, patients benefit from reassurance and treatment for underlying psychological stress. The traditional treatment of rebreathing into a paper bag is no longer recommended because of the potential risk of hypoxia. Sedatives and antidepressants should be reserved for patients who have not responded to conservative treatment.

Our patient was given an explanation of the condition together with breathing exercises. She received lorazepam and was immediately treated with intravenous hydration, along with intravenous infusion of magnesium, calcium, and potassium. These interventions led to an immediate resolution of her symptoms.

Her low level of intact parathyroid hormone may also have been caused by hypomagnesemia. She was given oral magnesium, potassium, calcium, and vitamin D to continue at home. In addition, she was advised to avoid excessive alcohol consumption and to see us or her primary care doctor should the symptoms recur. As expected, all the laboratory values normalized within 1 month of abstinence from alcohol, and she has been well since.

Severe acute pancreatitis has been known since the time of Rembrandt, with Nicolaes Tulp—the physician credited as first describing it—immortalized in the famous painting, The Anatomy Lesson. However, progress in managing this disease has been disappointing. Treatment is mainly supportive, and we lack any true disease-modifying therapy. But we are learning to recognize the disease and treat it supportively better than in the past.

The early hours of severe acute pancreatitis are critical for instituting appropriate intervention. Prompt fluid resuscitation is key to preventing immediate and later morbidity and death. This article focuses on identifying and managing the most severe form of acute pancreatitis—necrotizing disease—and its complications.

NECROTIZING DISEASE ACCOUNTS FOR MOST PANCREATITIS DEATHS

The classification and definitions of acute pancreatitis were recently revised from the 1992 Atlanta system and published early in 2013.¹ In addition, the American Pancreatic Association and the International Association of Pancreatology met in 2012 to develop evidence-based guidelines on managing severe pancreatitis.

An estimated 210,000 new cases of acute pancreatitis occur each year in the United States. About 20% of cases of severe acute pancreatitis are necrotizing disease, which accounts for nearly all the morbidity and death associated with acute pancreatitis.

The clinical spectrum of acute pancreatitis ranges from mild to life-threatening, reflecting interstitial (death rate < 1%) to necrotizing histology (the latter associated with a 25% risk of death if the pancreatitis becomes infected and a 10% risk if it is sterile). When death occurs early in the disease course, it tends to be from multiorgan failure; when death occurs later in the course, it tends to be from infection. Appropriate early treatment may prevent death in both categories.

DIAGNOSING ACUTE PANCREATITIS AND PREDICTING ITS SEVERITY

The diagnosis of acute pancreatitis requires two of the following three criteria:

• Clinical presentation—epigastric pain, nausea, vomiting
• Biochemical—amylase level more than three times the upper limit of normal, or lipase more than three times the upper limit of normal
• Evidence from computed tomography (CT), ultrasonography, or magnetic resonance imaging.

Although the biochemical criteria are variably sensitive for detecting acute pancreatitis (55%–100%), the specificity is very high (93% to 99%).

Recently, urinary trypsinogen-2, measured by dipstick, has also been used to aid diagnosis. It has a reasonable sensitivity (53%–96%) and specificity (85%) if positive (> 50 ng/mL).

Speed is critical

Over the years, many clinical prediction rules have been used for predicting the severity of acute pancreatitis. The Ranson criteria,² from 1974, and the Acute Physiology and Chronic Health Evaluation (APACHE) II system³ are cumbersome and require waiting up to 48 hours after the onset of acute pancreatitis to obtain a complete score. The Imrie-Glasgow score is another predictor.

The systemic inflammatory response syndrome (SIRS) is currently the most important indicator of prognosis.⁴ Originally adopted for predicting the development of organ failure with sepsis, it requires at least two of the following criteria:

• Heart rate > 90 beats/min
• Core temperature < 36°C or > 38°C
• White blood cells < 4,000 or > 12,000/mm³
• Respirations > 20/min.

The advantages of this system are that it identifies risk very early in the course of the disease and can be assessed quickly in the emergency department.

The Bedside Index for Severity of Acute Pancreatitis (BISAP) score is another simple, easy-to-perform prognostic index,^5,6 calculated by assigning 1 point for each of the following if present within the first 24 hours of presentation:

• Blood urea nitrogen > 25 mg/dL
• Abnormal mental status (Glasgow coma score < 15)
• Evidence of systemic inflammatory response syndrome
• Age > 60 years
• Pleural effusion seen on imaging study.

A score of 3 points is associated with a 5.3% rate of hospital death, 4 points with 12.7%, and 5 points with 22.5%.

At its most basic, severe acute pancreatitis is defined by organ failure (at least one organ from the respiratory, renal, or cardiovascular system) lasting for more than 48 hours. Failure for each organ is defined by the Marshall scoring system.¹

EARLY MANAGEMENT IS KEY TO OUTCOME

The window of opportunity to make a significant difference in outcome is within the first 12 to 24 hours of presentation. Volume resuscitation is the cornerstone of early management. By the time of presentation for severe acute pancreatitis, the pancreas is already necrotic, so the aim is to minimize the systemic inflammatory response syndrome with the goals of reducing rates of organ failure, morbidity, and death. Necrotizing pancreatitis is essentially an ischemic event, and the goal of volume resuscitation is to maintain pancreatic and intestinal microcirculation to prevent intestinal ischemia and subsequent bacterial translocation.⁷

Early resuscitation with lactated Ringer’s solution recommended

The evidence supporting a specific protocol for fluid resuscitation in severe acute pancreatitis is not strong, but a few studies provide guidance.

Wu et al⁸ randomized 40 patients with acute pancreatitis to one of four arms: “goal-directed fluid resuscitation” with either lactated Ringer’s solution or normal saline, or standard therapy (by physician discretion) with either lactated Ringer’s solution or normal saline. Goal-directed therapy involved a bolus of 20 mL/kg given over 30 to 45 minutes at presentation followed by infusion with rates dependent on an algorithm based on change in blood urea nitrogen level at set times. Patients receiving either goal-directed or standard therapy had significantly lower rates of systemic inflammatory response syndrome at 24 hours than at admission. Most striking was that treatment with lactated Ringer’s solution was associated with dramatically improved rates, whereas normal saline showed no improvement.

In a retrospective study of patients with acute pancreatitis, Warndorf et al⁹ identified 340 patients who received early resuscitation (more than one-third of the total 72-hour fluid volume within 24 hours of presentation) and 90 patients who received late resuscitation (less than one-third of the total 72-hour fluid volume within 24 hours of presentation). Patients who received early resuscitation developed less systemic inflammatory response syndrome and organ failure, and required fewer interventions.

Monitoring for optimum fluid resuscitation

Fluid resuscitation should be carefully managed to avoid administering either inadequate or excessive amounts of fluid. Inadequate fluid resuscitation can result in renal failure, progression of necrosis, and possibly infectious complications. Excessive resuscitation—defined as more than 4 L in the first 24 hours—is associated with respiratory failure, pancreatic fluid collections, and abdominal compartment syndrome.

Optimum resuscitation is controlled fluid expansion averaging 5 to 10 mL/kg per hour, with 2,500 to 4,000 mL given in the first 24 hours.

Adequate volume resuscitation can be evaluated clinically with the following goals:

• Heart rate < 120 beats per minute
• Mean arterial pressure 65–85 mm Hg
• Urinary output > 1 mL/kg per hour
• Hematocrit 35%–44%.

EARLY CT IS JUSTIFIED ONLY IF DIAGNOSIS IS UNCLEAR

The normal pancreas takes up contrast in the same way as do the liver and spleen, so its enhancement on CT is similar. If there is interstitial pancreatitis, CT shows the pancreas with normal contrast uptake, but the organ appears “boggy” with indistinct outlines. With necrotizing pancreatitis, only small areas of tissue with normal contrast may be apparent.

Peripancreatic fat necrosis may also be visible on CT. Obese patients tend to have a worse clinical course of necrotizing pancreatitis, probably because of the associated peripancreatic fat that is incorporated into the pancreatic necrosis.

For clear-cut cases of acute pancreatitis, time is wasted waiting to obtain CT images, and this could delay fluid resuscitation. Results from immediate CT almost never change the clinical management during the first week of acute pancreatitis, and obtaining CT images is usually not recommended if the diagnosis of acute pancreatitis is clear. CT’s sensitivity for detecting necrosis is only 70% in the first 48 hours of presentation, so it is easy to be fooled by a false-negative scan: frequently, a scan does not show necrotizing pancreatitis until after 72 hours. In addition, evidence from animal studies indicates that contrast agents might worsen pancreatic necrosis.

Immediate CT is justified if the diagnosis is in doubt at presentation, such as to evaluate for other intra-abdominal conditions such as intestinal ischemia or a perforated duodenal ulcer.

Contrast-enhanced CT is recommended 72 to 96 hours after presentation, or earlier if the patient is worsening despite treatment. Specific CT protocols will be included in new management guidelines, expected to be published soon.

PREVENTING INFECTIOUS COMPLICATIONS

Risk of infection is associated with the degree of pancreatic necrosis. Patients with less than 30% necrosis have a 22.5% chance of infection, whereas those with more than 50% necrosis have a 46.5% risk of infection.¹⁰

Infection can develop from a variety of sources:

Bacterial translocation from the colon and small bowel is thought to be one of the major sources of infection in necrotic pancreatitis. Volume resuscitation and maintaining gut integrity with early enteral nutrition are believed to minimize the risk of bacterial translocation.

Hematogenous spread of bacteria is another suspected source of infection into the pancreas. Again, enteral nutrition also reduces the risk by minimizing the need for central catheters.

Biliary sources may also play a role. Bile duct stones or gall bladder infection can lead to infected pancreatic necrosis.

ANTIBIOTICS NOT ROUTINELY RECOMMENDED

Treating acute pancreatitis with antibiotics has fallen in and out of favor over the past decades. From being standard practice in the 1970s, it dropped off in the 1980s and 1990s and then became more common again.

Current recommendations from the American Pancreatic Association and the International Association of Pancreatology are not to routinely use intravenous antibiotics to prevent infection in necrotizing pancreatitis because of lack of evidence that it changes overall outcome. Antibiotic usage may be associated with more bacterial resistance and the introduction of fungal infections into the pancreas.

Selective gut decontamination, involving oral and rectal administration of neomycin and other antibiotics, was shown in a single randomized trial to reduce the incidence of infection, but it is very cumbersome and is not recommended for acute pancreatitis.

Treatment with probiotics is also not recommended and was shown in one study to lead to a worse outcome.¹¹

ENTERAL BETTER THAN TOTAL PARENTERAL NUTRITION

Enteral tube feeding with either an elemental diet or a polymeric enteral formulation is the first-line therapy for necrotizing pancreatitis. Compared with total parenteral nutrition, it reduces infection, organ failure, hospital length of stay, the need for surgical intervention, and the risk of death. Total parenteral nutrition should be considered only for patients who do not tolerate enteral feeding because of severe ileus.

Conventional thinking for many years was to provide enteral feeding with a tube passed beyond the ligament of Treitz, thinking that it reduced stimulation to the pancreas. However, recent studies indicate that nasogastric feeding is equivalent to nasojejunal feeding in terms of nutrition, maintaining gut integrity, and outcome.

INTRA-ABDOMINAL HYPERTENSION AND ABDOMINAL COMPARTMENT SYNDROME

Movement of fluid into the intracellular space (“third-spacing”) occurs in acute pancreatitis and is exacerbated by fluid resuscitation. Intra-abdominal hypertension is associated with poor outcomes in patients with severe acute pancreatitis. Especially for patients with severe pancreatitis who are on mechanical ventilation, pressure should be monitored with transvesicular bladder measurements.

Intra-abdominal hypertension is defined as a sustained intra-abdominal pressure of more than 12 mm Hg, with the following grades:

• Grade 1: 12–15 mm Hg
• Grade 2: 16–20 mm Hg
• Grade 3: 21–25 mm Hg
• Grade 4: > 25 mm Hg.

Abdominal compartment syndrome is defined as a sustained intra-abdominal pressure of more than 20 mm Hg. It is associated with new organ dysfunction or failure. It should first be managed with ultrafiltration or diuretics to try to reduce the amount of fluid in the abdomen. Lumenal decompression can be tried with nasogastric or rectal tubes for the stomach and bowels. Ascites or retroperitoneal fluid can be drained percutaneously. In addition, analgesia and sedation to reduce abdominal muscle tone can help the patient become better ventilated. Neuromuscular blockade can also relax the abdomen.

Open abdominal decompression is the treatment of last resort to relieve abdominal compartment syndrome. The abdominal wall is not closed surgically but is allowed to heal by secondary intention (it “granulates in”).¹²

IDENTIFYING INFECTION

Fine-needle aspiration if clinical and imaging signs are not clear

Untreated infected pancreatitis is associated with a much higher risk of death than sterile pancreatic necrosis. Unfortunately, it can be difficult to determine if a patient with necrotizing pancreatitis has an infection because fever, tachycardia, and leukocytosis are usually present regardless. It is important to determine because mechanically intervening for sterile necrosis does not improve outcome.

Fine-needle aspiration, either guided by CT or done at the bedside with ultrasonography, with evaluation with Gram stain and culture, was widely used in the 1990s in cases of necrotizing pancreatitis to determine if infection was present. There has been a shift away from this because, although it can confirm the presence of infection, the false-negative rate is 15%. Clinical and imaging signs can be relied on in most cases to determine the presence of infection, and it is now recognized that fineneedle aspiration should be used only for select cases. Clinical studies have not shown that fine-needle aspiration improves outcomes.

Clinical scenarios typical of infected pancreatic necrosis include patients who have obvious signs of infection with no identifiable source, such as those who stabilize after acute severe acute pancreatitis, and then 10 to 14 days later become worse, with a dramatically higher white blood cell count and tachycardia. Such a patient likely needs an intervention regardless of the results of fine-needle aspiration.

On the other hand, a patient with a continually up-and-down course that never stabilizes over 3 weeks, with no identifiable source of infection, and with no peripancreatic gas apparent on imaging would be a good candidate for fine-needle aspiration.

If peripancreatic gas is seen on imaging, fine-needle aspiration is unnecessary. Peripancreatic gas is traditionally attributed to gasforming bacteria within the pancreas, but in my experience, it is usually from a fistula from the necrosis to the duodenum or the colon, the fistula being caused as the necrosis erodes at the hepatic flexure, the transverse colon, or the splenic flexure.

MECHANICAL INTERVENTIONS FOR INFECTIVE NECROSIS

Late, minimally invasive procedures preferred

Conventional management has shifted away from removing the necrosis with early surgical debridement of the pancreas. Experience with myocardial infarction shows that it is not necessary to remove a sterile necrotic organ, and studies with sterile pancreatic necrosis have found that surgical intervention is associated with a higher risk of death than medical management.

Documented infection has traditionally been considered a definite indication for debridement, but even that is being called into question as more studies are emerging of infected necrosis treated successfully with antibiotics alone.

Sterile necrosis with a fulminant course is a controversial indication for surgery. It was traditionally felt that surgery was worth trying for such patients, but this is no longer common practice.

For cases in which debridement was deemed advisable, surgery was done more frequently in the past. Now, a minimally invasive approach such as with endoscopy or percutaneous catheter is also used. Waiting until at least 4 weeks after the onset of acute pancreatitis is associated with a better outcome than intervening early.

WALLED-OFF NECROSIS

Watchful waiting or minimally invasive intervention

Patients who survive multiorgan failure but are still ill more than 4 weeks after the onset of pancreatitis should be suspected of having walled-off necrosis, formerly referred to as a pancreatic phlegmon. This term was abandoned after the 1992 Atlanta symposium.¹³ In the mid to late 1990s, the process was referred to as organized pancreatic necrosis. It is characterized by a mature, encapsulated collection of pancreatic or peripancreatic necrosis that contains variable amounts of amylase-rich fluid from pancreatic duct disruption.

Walled-off pancreatic necrosis (WOPN) is often confused with pancreatic pseudocyst; these may appear similar on CT, and higherdensity solid debris may be visible in walled-off necrosis within an otherwise homogenous-appearing collection. Magnetic resonance imaging defines liquid and solid much better than CT.

The best way to distinguish WOPN from pseudocyst is by clinical history: a patient with a preceding history of clinically severe acute pancreatitis almost always has necrotizing pancreatitis that evolves to walled-off necrosis, usually over 3 to 4 weeks.

Endoscopic removal and other minimally invasive approaches, such as aggressive percutaneous interventions, have replaced open necrosectomy for treatment, which was associated with high morbidity and mortality rates.^14–16

Intervening for sterile walled-off necrosis is still a controversial topic: although systemically ill, the patient is no longer having life-threatening consequences, and watchful waiting might be just as expedient as intervention. Evidence to support either view is lacking. Most experts believe that intervention should be done if the patient has gastric outlet obstruction and intractable pain and is unable to eat 4 to 6 weeks after the onset of pancreatitis with WOPN. Infected WOPN is considered an indication for drainage.

Severe acute pancreatitis has been known since the time of Rembrandt, with Nicolaes Tulp—the physician credited as first describing it—immortalized in the famous painting, The Anatomy Lesson. However, progress in managing this disease has been disappointing. Treatment is mainly supportive, and we lack any true disease-modifying therapy. But we are learning to recognize the disease and treat it supportively better than in the past.

The early hours of severe acute pancreatitis are critical for instituting appropriate intervention. Prompt fluid resuscitation is key to preventing immediate and later morbidity and death. This article focuses on identifying and managing the most severe form of acute pancreatitis—necrotizing disease—and its complications.

NECROTIZING DISEASE ACCOUNTS FOR MOST PANCREATITIS DEATHS

The classification and definitions of acute pancreatitis were recently revised from the 1992 Atlanta system and published early in 2013.¹ In addition, the American Pancreatic Association and the International Association of Pancreatology met in 2012 to develop evidence-based guidelines on managing severe pancreatitis.

An estimated 210,000 new cases of acute pancreatitis occur each year in the United States. About 20% of cases of severe acute pancreatitis are necrotizing disease, which accounts for nearly all the morbidity and death associated with acute pancreatitis.

The clinical spectrum of acute pancreatitis ranges from mild to life-threatening, reflecting interstitial (death rate < 1%) to necrotizing histology (the latter associated with a 25% risk of death if the pancreatitis becomes infected and a 10% risk if it is sterile). When death occurs early in the disease course, it tends to be from multiorgan failure; when death occurs later in the course, it tends to be from infection. Appropriate early treatment may prevent death in both categories.

DIAGNOSING ACUTE PANCREATITIS AND PREDICTING ITS SEVERITY

The diagnosis of acute pancreatitis requires two of the following three criteria:

• Clinical presentation—epigastric pain, nausea, vomiting
• Biochemical—amylase level more than three times the upper limit of normal, or lipase more than three times the upper limit of normal
• Evidence from computed tomography (CT), ultrasonography, or magnetic resonance imaging.

Although the biochemical criteria are variably sensitive for detecting acute pancreatitis (55%–100%), the specificity is very high (93% to 99%).

Recently, urinary trypsinogen-2, measured by dipstick, has also been used to aid diagnosis. It has a reasonable sensitivity (53%–96%) and specificity (85%) if positive (> 50 ng/mL).

Speed is critical

Over the years, many clinical prediction rules have been used for predicting the severity of acute pancreatitis. The Ranson criteria,² from 1974, and the Acute Physiology and Chronic Health Evaluation (APACHE) II system³ are cumbersome and require waiting up to 48 hours after the onset of acute pancreatitis to obtain a complete score. The Imrie-Glasgow score is another predictor.

The systemic inflammatory response syndrome (SIRS) is currently the most important indicator of prognosis.⁴ Originally adopted for predicting the development of organ failure with sepsis, it requires at least two of the following criteria:

• Heart rate > 90 beats/min
• Core temperature < 36°C or > 38°C
• White blood cells < 4,000 or > 12,000/mm³
• Respirations > 20/min.

The advantages of this system are that it identifies risk very early in the course of the disease and can be assessed quickly in the emergency department.

The Bedside Index for Severity of Acute Pancreatitis (BISAP) score is another simple, easy-to-perform prognostic index,^5,6 calculated by assigning 1 point for each of the following if present within the first 24 hours of presentation:

• Blood urea nitrogen > 25 mg/dL
• Abnormal mental status (Glasgow coma score < 15)
• Evidence of systemic inflammatory response syndrome
• Age > 60 years
• Pleural effusion seen on imaging study.

A score of 3 points is associated with a 5.3% rate of hospital death, 4 points with 12.7%, and 5 points with 22.5%.

At its most basic, severe acute pancreatitis is defined by organ failure (at least one organ from the respiratory, renal, or cardiovascular system) lasting for more than 48 hours. Failure for each organ is defined by the Marshall scoring system.¹

EARLY MANAGEMENT IS KEY TO OUTCOME

The window of opportunity to make a significant difference in outcome is within the first 12 to 24 hours of presentation. Volume resuscitation is the cornerstone of early management. By the time of presentation for severe acute pancreatitis, the pancreas is already necrotic, so the aim is to minimize the systemic inflammatory response syndrome with the goals of reducing rates of organ failure, morbidity, and death. Necrotizing pancreatitis is essentially an ischemic event, and the goal of volume resuscitation is to maintain pancreatic and intestinal microcirculation to prevent intestinal ischemia and subsequent bacterial translocation.⁷

Early resuscitation with lactated Ringer’s solution recommended

The evidence supporting a specific protocol for fluid resuscitation in severe acute pancreatitis is not strong, but a few studies provide guidance.

Wu et al⁸ randomized 40 patients with acute pancreatitis to one of four arms: “goal-directed fluid resuscitation” with either lactated Ringer’s solution or normal saline, or standard therapy (by physician discretion) with either lactated Ringer’s solution or normal saline. Goal-directed therapy involved a bolus of 20 mL/kg given over 30 to 45 minutes at presentation followed by infusion with rates dependent on an algorithm based on change in blood urea nitrogen level at set times. Patients receiving either goal-directed or standard therapy had significantly lower rates of systemic inflammatory response syndrome at 24 hours than at admission. Most striking was that treatment with lactated Ringer’s solution was associated with dramatically improved rates, whereas normal saline showed no improvement.

In a retrospective study of patients with acute pancreatitis, Warndorf et al⁹ identified 340 patients who received early resuscitation (more than one-third of the total 72-hour fluid volume within 24 hours of presentation) and 90 patients who received late resuscitation (less than one-third of the total 72-hour fluid volume within 24 hours of presentation). Patients who received early resuscitation developed less systemic inflammatory response syndrome and organ failure, and required fewer interventions.

Monitoring for optimum fluid resuscitation

Fluid resuscitation should be carefully managed to avoid administering either inadequate or excessive amounts of fluid. Inadequate fluid resuscitation can result in renal failure, progression of necrosis, and possibly infectious complications. Excessive resuscitation—defined as more than 4 L in the first 24 hours—is associated with respiratory failure, pancreatic fluid collections, and abdominal compartment syndrome.

Optimum resuscitation is controlled fluid expansion averaging 5 to 10 mL/kg per hour, with 2,500 to 4,000 mL given in the first 24 hours.

Adequate volume resuscitation can be evaluated clinically with the following goals:

• Heart rate < 120 beats per minute
• Mean arterial pressure 65–85 mm Hg
• Urinary output > 1 mL/kg per hour
• Hematocrit 35%–44%.

EARLY CT IS JUSTIFIED ONLY IF DIAGNOSIS IS UNCLEAR

The normal pancreas takes up contrast in the same way as do the liver and spleen, so its enhancement on CT is similar. If there is interstitial pancreatitis, CT shows the pancreas with normal contrast uptake, but the organ appears “boggy” with indistinct outlines. With necrotizing pancreatitis, only small areas of tissue with normal contrast may be apparent.

Peripancreatic fat necrosis may also be visible on CT. Obese patients tend to have a worse clinical course of necrotizing pancreatitis, probably because of the associated peripancreatic fat that is incorporated into the pancreatic necrosis.

For clear-cut cases of acute pancreatitis, time is wasted waiting to obtain CT images, and this could delay fluid resuscitation. Results from immediate CT almost never change the clinical management during the first week of acute pancreatitis, and obtaining CT images is usually not recommended if the diagnosis of acute pancreatitis is clear. CT’s sensitivity for detecting necrosis is only 70% in the first 48 hours of presentation, so it is easy to be fooled by a false-negative scan: frequently, a scan does not show necrotizing pancreatitis until after 72 hours. In addition, evidence from animal studies indicates that contrast agents might worsen pancreatic necrosis.

Immediate CT is justified if the diagnosis is in doubt at presentation, such as to evaluate for other intra-abdominal conditions such as intestinal ischemia or a perforated duodenal ulcer.

Contrast-enhanced CT is recommended 72 to 96 hours after presentation, or earlier if the patient is worsening despite treatment. Specific CT protocols will be included in new management guidelines, expected to be published soon.

PREVENTING INFECTIOUS COMPLICATIONS

Risk of infection is associated with the degree of pancreatic necrosis. Patients with less than 30% necrosis have a 22.5% chance of infection, whereas those with more than 50% necrosis have a 46.5% risk of infection.¹⁰

Infection can develop from a variety of sources:

Bacterial translocation from the colon and small bowel is thought to be one of the major sources of infection in necrotic pancreatitis. Volume resuscitation and maintaining gut integrity with early enteral nutrition are believed to minimize the risk of bacterial translocation.

Hematogenous spread of bacteria is another suspected source of infection into the pancreas. Again, enteral nutrition also reduces the risk by minimizing the need for central catheters.

Biliary sources may also play a role. Bile duct stones or gall bladder infection can lead to infected pancreatic necrosis.

ANTIBIOTICS NOT ROUTINELY RECOMMENDED

Treating acute pancreatitis with antibiotics has fallen in and out of favor over the past decades. From being standard practice in the 1970s, it dropped off in the 1980s and 1990s and then became more common again.

Current recommendations from the American Pancreatic Association and the International Association of Pancreatology are not to routinely use intravenous antibiotics to prevent infection in necrotizing pancreatitis because of lack of evidence that it changes overall outcome. Antibiotic usage may be associated with more bacterial resistance and the introduction of fungal infections into the pancreas.

Selective gut decontamination, involving oral and rectal administration of neomycin and other antibiotics, was shown in a single randomized trial to reduce the incidence of infection, but it is very cumbersome and is not recommended for acute pancreatitis.

Treatment with probiotics is also not recommended and was shown in one study to lead to a worse outcome.¹¹

ENTERAL BETTER THAN TOTAL PARENTERAL NUTRITION

Enteral tube feeding with either an elemental diet or a polymeric enteral formulation is the first-line therapy for necrotizing pancreatitis. Compared with total parenteral nutrition, it reduces infection, organ failure, hospital length of stay, the need for surgical intervention, and the risk of death. Total parenteral nutrition should be considered only for patients who do not tolerate enteral feeding because of severe ileus.

Conventional thinking for many years was to provide enteral feeding with a tube passed beyond the ligament of Treitz, thinking that it reduced stimulation to the pancreas. However, recent studies indicate that nasogastric feeding is equivalent to nasojejunal feeding in terms of nutrition, maintaining gut integrity, and outcome.

INTRA-ABDOMINAL HYPERTENSION AND ABDOMINAL COMPARTMENT SYNDROME

Movement of fluid into the intracellular space (“third-spacing”) occurs in acute pancreatitis and is exacerbated by fluid resuscitation. Intra-abdominal hypertension is associated with poor outcomes in patients with severe acute pancreatitis. Especially for patients with severe pancreatitis who are on mechanical ventilation, pressure should be monitored with transvesicular bladder measurements.

Intra-abdominal hypertension is defined as a sustained intra-abdominal pressure of more than 12 mm Hg, with the following grades:

• Grade 1: 12–15 mm Hg
• Grade 2: 16–20 mm Hg
• Grade 3: 21–25 mm Hg
• Grade 4: > 25 mm Hg.

Abdominal compartment syndrome is defined as a sustained intra-abdominal pressure of more than 20 mm Hg. It is associated with new organ dysfunction or failure. It should first be managed with ultrafiltration or diuretics to try to reduce the amount of fluid in the abdomen. Lumenal decompression can be tried with nasogastric or rectal tubes for the stomach and bowels. Ascites or retroperitoneal fluid can be drained percutaneously. In addition, analgesia and sedation to reduce abdominal muscle tone can help the patient become better ventilated. Neuromuscular blockade can also relax the abdomen.

Open abdominal decompression is the treatment of last resort to relieve abdominal compartment syndrome. The abdominal wall is not closed surgically but is allowed to heal by secondary intention (it “granulates in”).¹²

IDENTIFYING INFECTION

Fine-needle aspiration if clinical and imaging signs are not clear

Untreated infected pancreatitis is associated with a much higher risk of death than sterile pancreatic necrosis. Unfortunately, it can be difficult to determine if a patient with necrotizing pancreatitis has an infection because fever, tachycardia, and leukocytosis are usually present regardless. It is important to determine because mechanically intervening for sterile necrosis does not improve outcome.

Fine-needle aspiration, either guided by CT or done at the bedside with ultrasonography, with evaluation with Gram stain and culture, was widely used in the 1990s in cases of necrotizing pancreatitis to determine if infection was present. There has been a shift away from this because, although it can confirm the presence of infection, the false-negative rate is 15%. Clinical and imaging signs can be relied on in most cases to determine the presence of infection, and it is now recognized that fineneedle aspiration should be used only for select cases. Clinical studies have not shown that fine-needle aspiration improves outcomes.

Clinical scenarios typical of infected pancreatic necrosis include patients who have obvious signs of infection with no identifiable source, such as those who stabilize after acute severe acute pancreatitis, and then 10 to 14 days later become worse, with a dramatically higher white blood cell count and tachycardia. Such a patient likely needs an intervention regardless of the results of fine-needle aspiration.

On the other hand, a patient with a continually up-and-down course that never stabilizes over 3 weeks, with no identifiable source of infection, and with no peripancreatic gas apparent on imaging would be a good candidate for fine-needle aspiration.

If peripancreatic gas is seen on imaging, fine-needle aspiration is unnecessary. Peripancreatic gas is traditionally attributed to gasforming bacteria within the pancreas, but in my experience, it is usually from a fistula from the necrosis to the duodenum or the colon, the fistula being caused as the necrosis erodes at the hepatic flexure, the transverse colon, or the splenic flexure.

MECHANICAL INTERVENTIONS FOR INFECTIVE NECROSIS

Late, minimally invasive procedures preferred

Conventional management has shifted away from removing the necrosis with early surgical debridement of the pancreas. Experience with myocardial infarction shows that it is not necessary to remove a sterile necrotic organ, and studies with sterile pancreatic necrosis have found that surgical intervention is associated with a higher risk of death than medical management.

Documented infection has traditionally been considered a definite indication for debridement, but even that is being called into question as more studies are emerging of infected necrosis treated successfully with antibiotics alone.

Sterile necrosis with a fulminant course is a controversial indication for surgery. It was traditionally felt that surgery was worth trying for such patients, but this is no longer common practice.

For cases in which debridement was deemed advisable, surgery was done more frequently in the past. Now, a minimally invasive approach such as with endoscopy or percutaneous catheter is also used. Waiting until at least 4 weeks after the onset of acute pancreatitis is associated with a better outcome than intervening early.

WALLED-OFF NECROSIS

Watchful waiting or minimally invasive intervention

Patients who survive multiorgan failure but are still ill more than 4 weeks after the onset of pancreatitis should be suspected of having walled-off necrosis, formerly referred to as a pancreatic phlegmon. This term was abandoned after the 1992 Atlanta symposium.¹³ In the mid to late 1990s, the process was referred to as organized pancreatic necrosis. It is characterized by a mature, encapsulated collection of pancreatic or peripancreatic necrosis that contains variable amounts of amylase-rich fluid from pancreatic duct disruption.

Walled-off pancreatic necrosis (WOPN) is often confused with pancreatic pseudocyst; these may appear similar on CT, and higherdensity solid debris may be visible in walled-off necrosis within an otherwise homogenous-appearing collection. Magnetic resonance imaging defines liquid and solid much better than CT.

The best way to distinguish WOPN from pseudocyst is by clinical history: a patient with a preceding history of clinically severe acute pancreatitis almost always has necrotizing pancreatitis that evolves to walled-off necrosis, usually over 3 to 4 weeks.

Endoscopic removal and other minimally invasive approaches, such as aggressive percutaneous interventions, have replaced open necrosectomy for treatment, which was associated with high morbidity and mortality rates.^14–16

Intervening for sterile walled-off necrosis is still a controversial topic: although systemically ill, the patient is no longer having life-threatening consequences, and watchful waiting might be just as expedient as intervention. Evidence to support either view is lacking. Most experts believe that intervention should be done if the patient has gastric outlet obstruction and intractable pain and is unable to eat 4 to 6 weeks after the onset of pancreatitis with WOPN. Infected WOPN is considered an indication for drainage.

Severe acute pancreatitis has been known since the time of Rembrandt, with Nicolaes Tulp—the physician credited as first describing it—immortalized in the famous painting, The Anatomy Lesson. However, progress in managing this disease has been disappointing. Treatment is mainly supportive, and we lack any true disease-modifying therapy. But we are learning to recognize the disease and treat it supportively better than in the past.

The early hours of severe acute pancreatitis are critical for instituting appropriate intervention. Prompt fluid resuscitation is key to preventing immediate and later morbidity and death. This article focuses on identifying and managing the most severe form of acute pancreatitis—necrotizing disease—and its complications.

NECROTIZING DISEASE ACCOUNTS FOR MOST PANCREATITIS DEATHS

The classification and definitions of acute pancreatitis were recently revised from the 1992 Atlanta system and published early in 2013.¹ In addition, the American Pancreatic Association and the International Association of Pancreatology met in 2012 to develop evidence-based guidelines on managing severe pancreatitis.

An estimated 210,000 new cases of acute pancreatitis occur each year in the United States. About 20% of cases of severe acute pancreatitis are necrotizing disease, which accounts for nearly all the morbidity and death associated with acute pancreatitis.

The clinical spectrum of acute pancreatitis ranges from mild to life-threatening, reflecting interstitial (death rate < 1%) to necrotizing histology (the latter associated with a 25% risk of death if the pancreatitis becomes infected and a 10% risk if it is sterile). When death occurs early in the disease course, it tends to be from multiorgan failure; when death occurs later in the course, it tends to be from infection. Appropriate early treatment may prevent death in both categories.

DIAGNOSING ACUTE PANCREATITIS AND PREDICTING ITS SEVERITY

The diagnosis of acute pancreatitis requires two of the following three criteria:

• Clinical presentation—epigastric pain, nausea, vomiting
• Biochemical—amylase level more than three times the upper limit of normal, or lipase more than three times the upper limit of normal
• Evidence from computed tomography (CT), ultrasonography, or magnetic resonance imaging.

Although the biochemical criteria are variably sensitive for detecting acute pancreatitis (55%–100%), the specificity is very high (93% to 99%).

Recently, urinary trypsinogen-2, measured by dipstick, has also been used to aid diagnosis. It has a reasonable sensitivity (53%–96%) and specificity (85%) if positive (> 50 ng/mL).

Speed is critical

Over the years, many clinical prediction rules have been used for predicting the severity of acute pancreatitis. The Ranson criteria,² from 1974, and the Acute Physiology and Chronic Health Evaluation (APACHE) II system³ are cumbersome and require waiting up to 48 hours after the onset of acute pancreatitis to obtain a complete score. The Imrie-Glasgow score is another predictor.

The systemic inflammatory response syndrome (SIRS) is currently the most important indicator of prognosis.⁴ Originally adopted for predicting the development of organ failure with sepsis, it requires at least two of the following criteria:

• Heart rate > 90 beats/min
• Core temperature < 36°C or > 38°C
• White blood cells < 4,000 or > 12,000/mm³
• Respirations > 20/min.

The advantages of this system are that it identifies risk very early in the course of the disease and can be assessed quickly in the emergency department.

The Bedside Index for Severity of Acute Pancreatitis (BISAP) score is another simple, easy-to-perform prognostic index,^5,6 calculated by assigning 1 point for each of the following if present within the first 24 hours of presentation:

• Blood urea nitrogen > 25 mg/dL
• Abnormal mental status (Glasgow coma score < 15)
• Evidence of systemic inflammatory response syndrome
• Age > 60 years
• Pleural effusion seen on imaging study.

A score of 3 points is associated with a 5.3% rate of hospital death, 4 points with 12.7%, and 5 points with 22.5%.

At its most basic, severe acute pancreatitis is defined by organ failure (at least one organ from the respiratory, renal, or cardiovascular system) lasting for more than 48 hours. Failure for each organ is defined by the Marshall scoring system.¹

EARLY MANAGEMENT IS KEY TO OUTCOME

The window of opportunity to make a significant difference in outcome is within the first 12 to 24 hours of presentation. Volume resuscitation is the cornerstone of early management. By the time of presentation for severe acute pancreatitis, the pancreas is already necrotic, so the aim is to minimize the systemic inflammatory response syndrome with the goals of reducing rates of organ failure, morbidity, and death. Necrotizing pancreatitis is essentially an ischemic event, and the goal of volume resuscitation is to maintain pancreatic and intestinal microcirculation to prevent intestinal ischemia and subsequent bacterial translocation.⁷

Early resuscitation with lactated Ringer’s solution recommended

The evidence supporting a specific protocol for fluid resuscitation in severe acute pancreatitis is not strong, but a few studies provide guidance.

Wu et al⁸ randomized 40 patients with acute pancreatitis to one of four arms: “goal-directed fluid resuscitation” with either lactated Ringer’s solution or normal saline, or standard therapy (by physician discretion) with either lactated Ringer’s solution or normal saline. Goal-directed therapy involved a bolus of 20 mL/kg given over 30 to 45 minutes at presentation followed by infusion with rates dependent on an algorithm based on change in blood urea nitrogen level at set times. Patients receiving either goal-directed or standard therapy had significantly lower rates of systemic inflammatory response syndrome at 24 hours than at admission. Most striking was that treatment with lactated Ringer’s solution was associated with dramatically improved rates, whereas normal saline showed no improvement.

In a retrospective study of patients with acute pancreatitis, Warndorf et al⁹ identified 340 patients who received early resuscitation (more than one-third of the total 72-hour fluid volume within 24 hours of presentation) and 90 patients who received late resuscitation (less than one-third of the total 72-hour fluid volume within 24 hours of presentation). Patients who received early resuscitation developed less systemic inflammatory response syndrome and organ failure, and required fewer interventions.

Monitoring for optimum fluid resuscitation

Fluid resuscitation should be carefully managed to avoid administering either inadequate or excessive amounts of fluid. Inadequate fluid resuscitation can result in renal failure, progression of necrosis, and possibly infectious complications. Excessive resuscitation—defined as more than 4 L in the first 24 hours—is associated with respiratory failure, pancreatic fluid collections, and abdominal compartment syndrome.

Optimum resuscitation is controlled fluid expansion averaging 5 to 10 mL/kg per hour, with 2,500 to 4,000 mL given in the first 24 hours.

Adequate volume resuscitation can be evaluated clinically with the following goals:

• Heart rate < 120 beats per minute
• Mean arterial pressure 65–85 mm Hg
• Urinary output > 1 mL/kg per hour
• Hematocrit 35%–44%.

EARLY CT IS JUSTIFIED ONLY IF DIAGNOSIS IS UNCLEAR

The normal pancreas takes up contrast in the same way as do the liver and spleen, so its enhancement on CT is similar. If there is interstitial pancreatitis, CT shows the pancreas with normal contrast uptake, but the organ appears “boggy” with indistinct outlines. With necrotizing pancreatitis, only small areas of tissue with normal contrast may be apparent.

Peripancreatic fat necrosis may also be visible on CT. Obese patients tend to have a worse clinical course of necrotizing pancreatitis, probably because of the associated peripancreatic fat that is incorporated into the pancreatic necrosis.

For clear-cut cases of acute pancreatitis, time is wasted waiting to obtain CT images, and this could delay fluid resuscitation. Results from immediate CT almost never change the clinical management during the first week of acute pancreatitis, and obtaining CT images is usually not recommended if the diagnosis of acute pancreatitis is clear. CT’s sensitivity for detecting necrosis is only 70% in the first 48 hours of presentation, so it is easy to be fooled by a false-negative scan: frequently, a scan does not show necrotizing pancreatitis until after 72 hours. In addition, evidence from animal studies indicates that contrast agents might worsen pancreatic necrosis.

Immediate CT is justified if the diagnosis is in doubt at presentation, such as to evaluate for other intra-abdominal conditions such as intestinal ischemia or a perforated duodenal ulcer.

Contrast-enhanced CT is recommended 72 to 96 hours after presentation, or earlier if the patient is worsening despite treatment. Specific CT protocols will be included in new management guidelines, expected to be published soon.

PREVENTING INFECTIOUS COMPLICATIONS

Risk of infection is associated with the degree of pancreatic necrosis. Patients with less than 30% necrosis have a 22.5% chance of infection, whereas those with more than 50% necrosis have a 46.5% risk of infection.¹⁰

Infection can develop from a variety of sources:

Bacterial translocation from the colon and small bowel is thought to be one of the major sources of infection in necrotic pancreatitis. Volume resuscitation and maintaining gut integrity with early enteral nutrition are believed to minimize the risk of bacterial translocation.

Hematogenous spread of bacteria is another suspected source of infection into the pancreas. Again, enteral nutrition also reduces the risk by minimizing the need for central catheters.

Biliary sources may also play a role. Bile duct stones or gall bladder infection can lead to infected pancreatic necrosis.

ANTIBIOTICS NOT ROUTINELY RECOMMENDED

Treating acute pancreatitis with antibiotics has fallen in and out of favor over the past decades. From being standard practice in the 1970s, it dropped off in the 1980s and 1990s and then became more common again.

Current recommendations from the American Pancreatic Association and the International Association of Pancreatology are not to routinely use intravenous antibiotics to prevent infection in necrotizing pancreatitis because of lack of evidence that it changes overall outcome. Antibiotic usage may be associated with more bacterial resistance and the introduction of fungal infections into the pancreas.

Selective gut decontamination, involving oral and rectal administration of neomycin and other antibiotics, was shown in a single randomized trial to reduce the incidence of infection, but it is very cumbersome and is not recommended for acute pancreatitis.

Treatment with probiotics is also not recommended and was shown in one study to lead to a worse outcome.¹¹

ENTERAL BETTER THAN TOTAL PARENTERAL NUTRITION

Enteral tube feeding with either an elemental diet or a polymeric enteral formulation is the first-line therapy for necrotizing pancreatitis. Compared with total parenteral nutrition, it reduces infection, organ failure, hospital length of stay, the need for surgical intervention, and the risk of death. Total parenteral nutrition should be considered only for patients who do not tolerate enteral feeding because of severe ileus.

Conventional thinking for many years was to provide enteral feeding with a tube passed beyond the ligament of Treitz, thinking that it reduced stimulation to the pancreas. However, recent studies indicate that nasogastric feeding is equivalent to nasojejunal feeding in terms of nutrition, maintaining gut integrity, and outcome.

INTRA-ABDOMINAL HYPERTENSION AND ABDOMINAL COMPARTMENT SYNDROME

Movement of fluid into the intracellular space (“third-spacing”) occurs in acute pancreatitis and is exacerbated by fluid resuscitation. Intra-abdominal hypertension is associated with poor outcomes in patients with severe acute pancreatitis. Especially for patients with severe pancreatitis who are on mechanical ventilation, pressure should be monitored with transvesicular bladder measurements.

Intra-abdominal hypertension is defined as a sustained intra-abdominal pressure of more than 12 mm Hg, with the following grades:

• Grade 1: 12–15 mm Hg
• Grade 2: 16–20 mm Hg
• Grade 3: 21–25 mm Hg
• Grade 4: > 25 mm Hg.

Abdominal compartment syndrome is defined as a sustained intra-abdominal pressure of more than 20 mm Hg. It is associated with new organ dysfunction or failure. It should first be managed with ultrafiltration or diuretics to try to reduce the amount of fluid in the abdomen. Lumenal decompression can be tried with nasogastric or rectal tubes for the stomach and bowels. Ascites or retroperitoneal fluid can be drained percutaneously. In addition, analgesia and sedation to reduce abdominal muscle tone can help the patient become better ventilated. Neuromuscular blockade can also relax the abdomen.

Open abdominal decompression is the treatment of last resort to relieve abdominal compartment syndrome. The abdominal wall is not closed surgically but is allowed to heal by secondary intention (it “granulates in”).¹²

IDENTIFYING INFECTION

Fine-needle aspiration if clinical and imaging signs are not clear

Untreated infected pancreatitis is associated with a much higher risk of death than sterile pancreatic necrosis. Unfortunately, it can be difficult to determine if a patient with necrotizing pancreatitis has an infection because fever, tachycardia, and leukocytosis are usually present regardless. It is important to determine because mechanically intervening for sterile necrosis does not improve outcome.

Fine-needle aspiration, either guided by CT or done at the bedside with ultrasonography, with evaluation with Gram stain and culture, was widely used in the 1990s in cases of necrotizing pancreatitis to determine if infection was present. There has been a shift away from this because, although it can confirm the presence of infection, the false-negative rate is 15%. Clinical and imaging signs can be relied on in most cases to determine the presence of infection, and it is now recognized that fineneedle aspiration should be used only for select cases. Clinical studies have not shown that fine-needle aspiration improves outcomes.

Clinical scenarios typical of infected pancreatic necrosis include patients who have obvious signs of infection with no identifiable source, such as those who stabilize after acute severe acute pancreatitis, and then 10 to 14 days later become worse, with a dramatically higher white blood cell count and tachycardia. Such a patient likely needs an intervention regardless of the results of fine-needle aspiration.

On the other hand, a patient with a continually up-and-down course that never stabilizes over 3 weeks, with no identifiable source of infection, and with no peripancreatic gas apparent on imaging would be a good candidate for fine-needle aspiration.

If peripancreatic gas is seen on imaging, fine-needle aspiration is unnecessary. Peripancreatic gas is traditionally attributed to gasforming bacteria within the pancreas, but in my experience, it is usually from a fistula from the necrosis to the duodenum or the colon, the fistula being caused as the necrosis erodes at the hepatic flexure, the transverse colon, or the splenic flexure.

MECHANICAL INTERVENTIONS FOR INFECTIVE NECROSIS

Late, minimally invasive procedures preferred

Conventional management has shifted away from removing the necrosis with early surgical debridement of the pancreas. Experience with myocardial infarction shows that it is not necessary to remove a sterile necrotic organ, and studies with sterile pancreatic necrosis have found that surgical intervention is associated with a higher risk of death than medical management.

Documented infection has traditionally been considered a definite indication for debridement, but even that is being called into question as more studies are emerging of infected necrosis treated successfully with antibiotics alone.

Sterile necrosis with a fulminant course is a controversial indication for surgery. It was traditionally felt that surgery was worth trying for such patients, but this is no longer common practice.

For cases in which debridement was deemed advisable, surgery was done more frequently in the past. Now, a minimally invasive approach such as with endoscopy or percutaneous catheter is also used. Waiting until at least 4 weeks after the onset of acute pancreatitis is associated with a better outcome than intervening early.

WALLED-OFF NECROSIS

Watchful waiting or minimally invasive intervention

Patients who survive multiorgan failure but are still ill more than 4 weeks after the onset of pancreatitis should be suspected of having walled-off necrosis, formerly referred to as a pancreatic phlegmon. This term was abandoned after the 1992 Atlanta symposium.¹³ In the mid to late 1990s, the process was referred to as organized pancreatic necrosis. It is characterized by a mature, encapsulated collection of pancreatic or peripancreatic necrosis that contains variable amounts of amylase-rich fluid from pancreatic duct disruption.

Walled-off pancreatic necrosis (WOPN) is often confused with pancreatic pseudocyst; these may appear similar on CT, and higherdensity solid debris may be visible in walled-off necrosis within an otherwise homogenous-appearing collection. Magnetic resonance imaging defines liquid and solid much better than CT.

The best way to distinguish WOPN from pseudocyst is by clinical history: a patient with a preceding history of clinically severe acute pancreatitis almost always has necrotizing pancreatitis that evolves to walled-off necrosis, usually over 3 to 4 weeks.

Endoscopic removal and other minimally invasive approaches, such as aggressive percutaneous interventions, have replaced open necrosectomy for treatment, which was associated with high morbidity and mortality rates.^14–16

Intervening for sterile walled-off necrosis is still a controversial topic: although systemically ill, the patient is no longer having life-threatening consequences, and watchful waiting might be just as expedient as intervention. Evidence to support either view is lacking. Most experts believe that intervention should be done if the patient has gastric outlet obstruction and intractable pain and is unable to eat 4 to 6 weeks after the onset of pancreatitis with WOPN. Infected WOPN is considered an indication for drainage.

The US Preventive Services Task Force (USPSTF) recently published a clinical guideline on the use of calcium and vitamin D supplements to prevent fractures in adults.¹ This agency “strives to make accurate, up-to-date, and relevant recommendations about preventive services in primary care,”² and within those parameters they generally succeed. But I am confused about the value of this specific guideline, and apparently I am not alone.

The task force came to several major conclusions about calcium and vitamin D supplementation to prevent fractures:

• There is insufficient evidence to offer guidance on supplementation in premeno-pausal women or in men
• One should not prescribe supplementation of 400 IU or less of vitamin D₃ or 1 g or less of calcium in postmenopausal women
• The data are insufficient to assess the harm and benefit of higher doses of supplemental vitamin D or calcium.

The task force stuck to their rules and weighed the data within the constraints of the specific question they were charged to address.

A challenge to clinicians attempting to apply rigidly defined, evidence-based conclusions is that the more precisely a question is addressed, the more limited is the answer’s applicability in clinical practice. Thus, Dr. Robin Dore, in this issue of the Journal, says that she believes there are benefits of vitamin D and calcium supplementation beyond primary prevention of fractures, and the benefits are not negated by the magnitude of potential harm (stated to be “small” by the USPSTF).

We are bombarded by clinical practice guidelines, and we don’t know which will be externally imposed as a measure of quality by which our practice performance will be assessed. In the clinic, we encounter a series of individual patients with whom we make individual treatment decisions. Like the inhabitants of Lake Wobegon, few of our patients are the “average patient” as derived from cross-sectional studies. Some have occult celiac disease, others are on proton pump inhibitors, some are lactose-intolerant, and some are on intermittent prednisone. For these patients, should the USPSTF guidelines warrant the extra effort and time to individually document why the guidelines don’t fit and why we made the clinical judgment to not follow them? Additionally, how many patients in the clinical studies used by the USPSTF fit into these or other unique categories and may have thus contaminated the data? I don’t see in these guidelines recommendations on how best to assess calcium and vitamin D intake and absorption in our patients in a practical manner. After all, supplementation is in addition to the actual intake of dietary sources.

For me, further confusion stems from trying to clinically couple the logic of such carefully analyzed, accurately stated, and tightly focused guidelines with what we already know (and apparently don’t know). We know that severe vitamin D deficiency clearly causes low bone density and fractures from osteomalacia, and the Institute of Medicine has previously stated that adequate vitamin D is beneficial and so should be supplemented.³ Vitamin D deficiency is a continuum and is very unlikely to be defined by the quantity of supplementation. Additionally, the USPSTF has previously published guidelines on supplementing vitamin D intake to prevent falls—falls being a major preventable cause of primary fractures. There seems to be some conceptual incongruence between these guidelines.

While epidemiologic studies have incorporated estimates of dietary and supplemental intake of calcium and vitamin D, what likely really matters is the absorption and the achieved blood levels and tissue incorporation. As shown in the examples above, many variables influence these in individual patients. And most troublesome is that there is no agreement as to the appropriate target level for circulating vitamin D. I agree with two-thirds of the task force’s conclusions—we have insufficient evidence. Are these really guidelines, or a plea for the gathering of appropriate outcome data?

The US Preventive Services Task Force (USPSTF) recently published a clinical guideline on the use of calcium and vitamin D supplements to prevent fractures in adults.¹ This agency “strives to make accurate, up-to-date, and relevant recommendations about preventive services in primary care,”² and within those parameters they generally succeed. But I am confused about the value of this specific guideline, and apparently I am not alone.

The task force came to several major conclusions about calcium and vitamin D supplementation to prevent fractures:

• There is insufficient evidence to offer guidance on supplementation in premeno-pausal women or in men
• One should not prescribe supplementation of 400 IU or less of vitamin D₃ or 1 g or less of calcium in postmenopausal women
• The data are insufficient to assess the harm and benefit of higher doses of supplemental vitamin D or calcium.

The task force stuck to their rules and weighed the data within the constraints of the specific question they were charged to address.

A challenge to clinicians attempting to apply rigidly defined, evidence-based conclusions is that the more precisely a question is addressed, the more limited is the answer’s applicability in clinical practice. Thus, Dr. Robin Dore, in this issue of the Journal, says that she believes there are benefits of vitamin D and calcium supplementation beyond primary prevention of fractures, and the benefits are not negated by the magnitude of potential harm (stated to be “small” by the USPSTF).

We are bombarded by clinical practice guidelines, and we don’t know which will be externally imposed as a measure of quality by which our practice performance will be assessed. In the clinic, we encounter a series of individual patients with whom we make individual treatment decisions. Like the inhabitants of Lake Wobegon, few of our patients are the “average patient” as derived from cross-sectional studies. Some have occult celiac disease, others are on proton pump inhibitors, some are lactose-intolerant, and some are on intermittent prednisone. For these patients, should the USPSTF guidelines warrant the extra effort and time to individually document why the guidelines don’t fit and why we made the clinical judgment to not follow them? Additionally, how many patients in the clinical studies used by the USPSTF fit into these or other unique categories and may have thus contaminated the data? I don’t see in these guidelines recommendations on how best to assess calcium and vitamin D intake and absorption in our patients in a practical manner. After all, supplementation is in addition to the actual intake of dietary sources.

For me, further confusion stems from trying to clinically couple the logic of such carefully analyzed, accurately stated, and tightly focused guidelines with what we already know (and apparently don’t know). We know that severe vitamin D deficiency clearly causes low bone density and fractures from osteomalacia, and the Institute of Medicine has previously stated that adequate vitamin D is beneficial and so should be supplemented.³ Vitamin D deficiency is a continuum and is very unlikely to be defined by the quantity of supplementation. Additionally, the USPSTF has previously published guidelines on supplementing vitamin D intake to prevent falls—falls being a major preventable cause of primary fractures. There seems to be some conceptual incongruence between these guidelines.

While epidemiologic studies have incorporated estimates of dietary and supplemental intake of calcium and vitamin D, what likely really matters is the absorption and the achieved blood levels and tissue incorporation. As shown in the examples above, many variables influence these in individual patients. And most troublesome is that there is no agreement as to the appropriate target level for circulating vitamin D. I agree with two-thirds of the task force’s conclusions—we have insufficient evidence. Are these really guidelines, or a plea for the gathering of appropriate outcome data?

The US Preventive Services Task Force (USPSTF) recently published a clinical guideline on the use of calcium and vitamin D supplements to prevent fractures in adults.¹ This agency “strives to make accurate, up-to-date, and relevant recommendations about preventive services in primary care,”² and within those parameters they generally succeed. But I am confused about the value of this specific guideline, and apparently I am not alone.

The task force came to several major conclusions about calcium and vitamin D supplementation to prevent fractures:

• There is insufficient evidence to offer guidance on supplementation in premeno-pausal women or in men
• One should not prescribe supplementation of 400 IU or less of vitamin D₃ or 1 g or less of calcium in postmenopausal women
• The data are insufficient to assess the harm and benefit of higher doses of supplemental vitamin D or calcium.

The task force stuck to their rules and weighed the data within the constraints of the specific question they were charged to address.

A challenge to clinicians attempting to apply rigidly defined, evidence-based conclusions is that the more precisely a question is addressed, the more limited is the answer’s applicability in clinical practice. Thus, Dr. Robin Dore, in this issue of the Journal, says that she believes there are benefits of vitamin D and calcium supplementation beyond primary prevention of fractures, and the benefits are not negated by the magnitude of potential harm (stated to be “small” by the USPSTF).

We are bombarded by clinical practice guidelines, and we don’t know which will be externally imposed as a measure of quality by which our practice performance will be assessed. In the clinic, we encounter a series of individual patients with whom we make individual treatment decisions. Like the inhabitants of Lake Wobegon, few of our patients are the “average patient” as derived from cross-sectional studies. Some have occult celiac disease, others are on proton pump inhibitors, some are lactose-intolerant, and some are on intermittent prednisone. For these patients, should the USPSTF guidelines warrant the extra effort and time to individually document why the guidelines don’t fit and why we made the clinical judgment to not follow them? Additionally, how many patients in the clinical studies used by the USPSTF fit into these or other unique categories and may have thus contaminated the data? I don’t see in these guidelines recommendations on how best to assess calcium and vitamin D intake and absorption in our patients in a practical manner. After all, supplementation is in addition to the actual intake of dietary sources.

For me, further confusion stems from trying to clinically couple the logic of such carefully analyzed, accurately stated, and tightly focused guidelines with what we already know (and apparently don’t know). We know that severe vitamin D deficiency clearly causes low bone density and fractures from osteomalacia, and the Institute of Medicine has previously stated that adequate vitamin D is beneficial and so should be supplemented.³ Vitamin D deficiency is a continuum and is very unlikely to be defined by the quantity of supplementation. Additionally, the USPSTF has previously published guidelines on supplementing vitamin D intake to prevent falls—falls being a major preventable cause of primary fractures. There seems to be some conceptual incongruence between these guidelines.

While epidemiologic studies have incorporated estimates of dietary and supplemental intake of calcium and vitamin D, what likely really matters is the absorption and the achieved blood levels and tissue incorporation. As shown in the examples above, many variables influence these in individual patients. And most troublesome is that there is no agreement as to the appropriate target level for circulating vitamin D. I agree with two-thirds of the task force’s conclusions—we have insufficient evidence. Are these really guidelines, or a plea for the gathering of appropriate outcome data?