A 59-Year-Old Man Considering Implantation of a Cardiac Defibrillator

DR REYNOLDS: Mr M is a 59-year-old man with a history of myocardial infarction (MI) and a low ejection fraction who is considering implantable cardioverter-defibrillator (ICD) placement.

Mr M was well until 7 years ago, when he had an acute anterior MI. He was initially treated with thrombolytics and his pain resolved. A cardiac catheterization revealed a tight lesion of the proximal left anterior descending coronary artery; Mr M underwent successful stent placement across the lesion. His course was complicated by congestive heart failure (CHF); a post-MI echocardiogram showed a left ventricular ejection fraction (LVEF) of 25%. Mr M has done well since; he is currently without chest pain, shortness of breath, or palpitations. He has never experienced syncope but does note diminished exercise capacity. His most recent echocardiogram, performed in 2004, showed a stable ejection fraction of 25% with anteroseptal, anterior, and apical akinesis.

Mr M's primary care physician now wonders whether Mr M should undergo implantation of a cardioverter-defibrillator for the prevention of sudden cardiac death.

Mr M is being treated for hyperlipidemia; he has no other significant past medical history. His medications include metoprolol XL (50 mg once daily), lisinopril (10 mg once daily), isosorbide mononitrate (30 mg once daily), atorvastatin (20 mg once daily) clopidogrel (75 mg once daily), and aspirin (325 mg once daily).

Mr M works as a car salesman; he lives alone. He drinks alcohol rarely and does not smoke. His family history is significant for his father's death at 53 years of recurrent MI.

At a recent physical examination, Mr M appeared well. His blood pressure measured 120/70 mm Hg; his jugular veins were not distended; there were no carotid bruits. His chest was clear; his cardiovascular examination showed a normal S1 and S1 without murmurs or gallops. His extremities were without edema; his pulses were intact.

A recent electrocardiogram showed normal sinus rhythm with normal intervals and QRS axis. There were anterior Q waves with T-wave inversions, consistent with his large anterior MI.

I don't have the same exercise tolerance at all, not even close. I mean, prior to the heart attack I was able to do 5 miles at a stretch; I did about 20 miles a week, running, and I can't come close to equaling those standards right now. Unfortunately, I wish I could, that's my goal. But I think I’ve pretty much put that off, unfortunately. It's not going to happen.

I’ve been talked to about putting a device in my chest that is supposed to regulate the heart. I’ve talked to a few different doctors on it, and I thought I had a grasp on the situation. But my understanding of it is that at the present time it's not a necessary thing. I’m 6 years removed from having my original procedure; I want to know why it's important to do it now. Right now, for me, right at this particular moment, there's no reason for me to do it. That's why I’m having such a hard problem as far as making a decision on it.

When I talk with somebody about buying a car, I’m trying to get them to commit to spending a sum of money to do that, $30 000, $35 000. A lot of times after the fact, I have to sit down and kind of resolve in my own mind that I do the same thing every day. I talk to people every day that want to buy something, and I try to put them in a position where they can make a decision on it.

The same with having this put in. I don't know that the people that I’m speaking to really have thought about it. Have they sat down and said, “Why would this person put himself in a position to have something, an operation that is not necessary at the present time, and what empirical evidence is there to make you do it?”

You know, no matter what you do in life, there's always somebody that wants you to do something because they think it is the most important thing. You bring your car in, you get the oil changed; you go to a dentist; everything's important to the person that's doing it. But is it important to me? That's the question. I’d like to see something in black and white, basically. The fact that if I don't do it, I have this life expectancy. If I do it, I’m going to live another X amount of time.

What are the current indications for ICD placement? How much does ICD placement cost and for what groups are ICDs cost-effective? Given that Mr M had an MI 7 years ago, is his risk of sudden death as high as that of someone who just had an MI? What is Mr M's life expectancy without and with an ICD? Can an electrophysiology study (EPS) help identify whether patients will benefit from an ICD? What are the complications of ICD placement? How do recipients feel? Are there any special precautions for patients with ICDs (magnetic resonance imaging [MRI], metal detectors at airports, battery changes, driving if it discharges, and the like)? What do you recommend for Mr M?

DR ZIMETBAUM: Mr M is a 59-year-old man with a history of MI 7 years prior to this discussion. As a consequence of his MI, he has a cardiomyopathy with LVEF of 25%. He is receiving a comprehensive and data-driven medical regimen including aspirin, metoprolol, lisinopril, and atorvastatin. He has no angina and mild symptoms consistent with class II CHF in the form of diminished exercise capacity. He has been presented with the option for the implantation of an ICD as prophylactic therapy to reduce his risk of sudden arrhythmic death.

Mr M is understandably resistant to having an invasive procedure performed for a risk that seems hypothetical. He notes that had this option been presented to him in 1999 when he had his MI and was already hospitalized, he might have considered it. Now after 7 years of feeling well without a hospitalization, it doesn't make sense to him to electively choose to have a procedure done that will expose him to risk of complication and a lifelong requirement for maintenance of the implanted device. Mr M also wonders whether this recommendation is really necessary and in his best interests, or is instead a reflection of market-driven forces and the specialists' desire to perform operations for which they are well reimbursed.

The discussion of the prophylactic placement of an ICD represents one of the more confusing and poorly understood consultations cardiologists perform. Much of the difficulty understanding and adapting the current recommendations for the primary prevention of sudden death comes from the rapid development of data in the last decade and the expected delay in the dissemination and acceptance of new information in medical practice.^{1 - 2} It also arises from the daunting societal costs of these therapies and the conceptual difficulty with embracing an invasive therapy to reduce the risk of an event often with no associated prior symptoms. These factors are further complicated by the ultimate consequence of being wrong (ie, death).

The process of understanding the potential benefit of ICD implantation begins with understanding the risk of death without ICD. Cardiovascular death is traditionally characterized as sudden arrhythmic death and death due to pump failure. Arrhythmic death is the target for prevention by ICD implantation (Figure 1), and the process of selecting patients at greatest potential need of these devices remains a significant challenge for medical science.

Approximately 450 000 people in the United States die of sudden cardiac death each year.² The vast majority of these deaths occur in patients without discernible risk factors. One group of patients, those like Mr M with coronary artery disease (CAD) resulting in impaired ventricular function, have been identified as particularly at risk for sudden death.^{3 - 4} The risk of sudden death is increased at least 2-fold among patients with CAD, nonsustained ventricular tachycardia (NSVT), and reduced (<40%) ventricular function,^{3 - 5} due primarily to scar-mediated sustained ventricular arrhythmias. Efforts to decrease this mortality with antiarrhythmic drugs proved at best ineffective and at worst dangerous.^{6 - 7} In the mid 1990s, 2 trials, the Multicenter Automatic Defibrillator Implantation Trial (MADIT 1)⁸ and Multicenter Unsustained Tachycardia Trial (MUSTT)⁹ identified patients with prior MI, LVEF less than 40%, NSVT, and ventricular tachycardia induced during EPS, and randomized them to receive standard medical therapy compared with antiarrhythmic therapy (primarily ICDs). Differences existed between these studies, but the remarkable finding was a 23% reduction in absolute mortality in the patients treated with the ICD during 2 to 3 years of follow-up. In other words, approximately 4 patients with these risk factors would need to have an ICD implanted to save 1 life. Subsequent trials (Multicenter Automatic Defibrillator Implantation Trial 2 [MADIT 2]¹⁰ and Sudden Cardiac Death Heart Failure Trial [SCD-HeFT]¹¹ ) have studied slightly sicker patients (LVEF<30% in MADIT 2 or <35% with CHF in SCD-HeFT) and have eliminated the requirement to demonstrate ambient ventricular ectopy or sustained ventricular arrhythmia at invasive electrophysiologic testing. The reduction in absolute mortality has been more modest in these trials (5%-7%) but still significant, with approximately 11 to 17 patients treated to save 1 life (Table). Mr M would have met criteria to be enrolled in the later studies; based on those results, he could gain an average 6% reduction in his risk of sudden cardiac death (Table).

Mr M's belief that he should be safe given the 7 years that have passed since his MI makes intuitive sense. At the time of an MI, nearly half a patient's ensuing mortality risk will be due to sudden arrhythmic death. The Valsartan in Acute Myocardial Infarction (VALIANT) study enrolled 14 000 patients similar to Mr M within 6 days of MI; they all had left ventricular dysfunction (≤40%) and/or heart failure.¹² These patients were randomized to receive treatment with an angiotensin-converting enzyme inhibitor, an angiotensin II receptor blocker, or both. In a substudy, these patients were found to have the greatest risk of sudden death during the first month following MI. This risk decreased thereafter and reached a constant rate at 1 year. In this analysis, risk was greatest (2.5%) in the first month in those with an ejection fraction of 30% or less. The rate of sudden death then plateaued at 0.15% per month thereafter. Interestingly, a separate study evaluating the role of ICDs in the early postinfarction period (Defibrillator in Acute Myocardial Infarction Trial [DINAMIT]) found no mortality benefit associated with early post-MI ICD implantation (6-40 days following MI) for patients with an ejection fraction of 35% or less.¹³ Multiple factors including recovery of left ventricular function, sudden death due to recurrent ischemia not mitigated by an ICD, and possible excess mortality related to the ICD in the early post-MI period may have contributed to this negative result. Currently, guidelines require that decisions regarding ICD implantation be delayed beyond 40 days following an MI.¹⁴

Mr M is beyond the early post-MI phase, but is well-represented by 2 of the previously cited trials. The control groups of the MUSTT and MADIT 2 trials enrolled patients an average of 3 and 7 years following MI, respectively. These trials included patients with a high incidence of CHF and a modest rate of β-blocker and ACEI use and found that during a mean follow-up of 3 years, mortality rates were more than twice those documented in VALIANT.¹² A substudy from MADIT 2 evaluated the influence of time post-MI on benefit derived from ICD implantation.¹⁵ In this analysis the reduction in mortality associated with ICD implantation persisted even in patients enrolled longer than 10 years following an MI. Therefore, current practice is to wait at least 6 weeks post-MI but not to limit how long after MI an ICD will be considered. In summary, Mr M can assume he continues to be at elevated risk for arrhythmia despite the 7 years that have passed since his MI, and that an ICD could reduce that risk.

Another important initiative to improve use of ICD therapy is to more effectively select patients who will benefit. Currently, noninvasive predictors of risk include the degree of LV dysfunction as well as electrocardiographic tests that identify abnormal electrical function. Programmed stimulation of the right ventricle is a minimally invasive procedure that historically has been the gold standard for identifying arrhythmia risk.

Invasive Electrophysiologic Testing. The role of electrophysiologic testing for risk stratification was most directly tested in MUSTT,¹⁶ which required a positive study for enrollment. Patients were then randomized to receive standard medical therapy or antiarrhythmic therapy (largely ICD). Patients who did not have a sustained ventricular arrhythmia induced at EPS were followed in a registry. Patients who had a ventricular arrhythmia induced and were in the control group had a 2-year incidence of sudden death of 18%, whereas patients who had no arrhythmia induced and were followed in the registry had a 2-year incidence of sudden death of 12%. The significant risk of mortality in the noninducible group in the MUSTT study supported the elimination of EPS as an enrollment requirement in MADIT 2. A substudy of the MADIT 2 trial in which an EPS was performed demonstrated that a positive EPS, found in 593 of the 742 ICD-treated patients, correlated with an increase in the risk of subsequent ventricular tachycardia but not ventricular fibrillation.¹⁷ Currently, EPS is not required for risk stratification in patients with CAD and an LVEF of less than 30%, and this study would add little predictive value for Mr M. For patients with an ejection fraction of between 30% and 40%, EPS may add some predictive value, but as demonstrated in the MUSTT registry, still misses a significant proportion of patients at risk for lethal ventricular arrhythmias.

Left Ventricular Dysfunction. The degree of left ventricular dysfunction, particularly when less than 40%, remains one of the strongest predictors of sudden death and all-cause mortality. This association has been consistently demonstrated both before and since the advent of primary reperfusion therapies and aggressive medication use post-MI.^{3 - 5} This association is particularly strong when the ejection fraction is below 30%, as is the case for Mr M.¹² Unfortunately, reduced ejection fraction is not by itself a sufficient stratifier of risk of sudden death. Specifically, a significant group of patients with CAD and an ejection fraction above 30% remain at risk of sudden death and require other markers to help identify them.

Congestive Heart Failure. The presence of CHF in patients with CAD and depressed ejection fraction increases the risk of sudden death and all-cause mortality.^{1 ,11 ,18} The Metoprolol CR/XL Randomized Intervention Trial in Chronic Heart Failure (MERIT-HF) study of β-blocker use in New York Heart Association class II to IV CHF found that the rates of sudden and all-cause mortality increase with increasing class of heart failure.¹⁹ However, the relative proportion of sudden to all-cause mortality was substantially greater in patients with less severe heart failure (class II with a rate of sudden death of 64% compared with 59% and 33% for class III or IV, respectively). Therefore, the presence of CHF, even when relatively compensated, increases the risk of sudden death. Mr M's complaints of exercise intolerance suggest class II CHF. Current guidelines support the implantation of ICDs in patients with class II and III CHF or patients with coronary or noncoronary-related cardiomyopathy and an ejection fraction of 35% or less^{2 ,14 ,20} ( Article ).

*Based on 2006 American College of Cardiology/American Heart Association/European Society of Cardiology guidelines.¹⁴

Ventricular Arrhythmia. The first 2 trials of ICD use for primary prevention of sudden death (MADIT 1⁸ and MUSTT⁹ ) required the presence of NSVT at least 1 month post-MI for inclusion. A more recent evaluation of 2130 patients with acute MI, of whom 70% had undergone coronary revascularization and 94% were treated with β-blockers, found that NSVT was a more robust predictor of sudden death in patients with more preserved ventricular function than in those with LVEF below 35%.²¹ This finding supports the elimination of NSVT as a requirement in the MADIT 2 study where ejection fraction alone qualified patients for enrollment.¹⁰ In summary, NSVT is a relevant finding to identify risk in patients with CAD and an ejection fraction between 30% and 40% but not for those with ejection fraction below 30%. The timing and method of identifying NSVT remain unstandardized.

Prolonged QRS Duration. In the MUSTT study,⁹ QRS prolongation with a left bundle branch block but not a right bundle branch block was associated with an increase in sudden and all-cause mortality.²² Patients with depressed LV function due to CAD are considered candidates for ICD regardless of QRS morphology, but patients with left bundle branch block often have dysynchronous activation of the left ventricle and should be considered for cardiac resynchronization with a biventricular ICD.²⁰

T-Wave Alternans. T-wave alternans (TWA) refers to microscopic alternans of the T-wave amplitude (T-wave timing or morphology), which may predict arrhythmic risk in patients with abnormal cardiac substrate.²³ This test is performed with specialized analysis of the electrocardiogram during a stress test. A recent randomized controlled trial of TWA in patients with CAD and LVEF less than 40% demonstrated that a negative test was associated with a 97.5% 2-year survival, supporting the excellent negative predictive value of this marker.²⁴ The value of a positive test is of modest value (15% rate of death at 2 years), and its performance is limited in patients who cannot exercise and reach a target heart rate of more than 105 beats/min. The widespread use of TWA for risk stratification will depend on its performance in larger randomized clinical trials. Nonetheless, a negative TWA study would provide some reassurance for Mr M should he remain reluctant to have an ICD implanted.

The previously described markers for sudden death remain inadequate to fully capture the entirety of the “at risk” population and to exclude those who do not require primary prevention. Many of these markers have not yet been tested thoroughly enough to allow their inclusion in clinical guidelines. Studies are ongoing to evaluate the utility of some of these newer markers alone and in combination to improve the specificity of our recommendations. The characteristics of patients currently receiving ICDs for the primary prevention of sudden death and their outcomes will be tracked by hospitals participating in the American College of Cardiology's National Cardiovascular Data Registry (ACC-NCDR) ICD registry.²⁵

Current guidelines for the primary prevention of sudden death ( Article ) would identify Mr M as having a class I indication for ICD implantation.¹⁴

Mr M is appropriately cautious about undergoing a surgery. Implantable cardioverter-defibrillators are implanted using local anesthesia with conscious sedation. The acute risks of implantation include cardiac or vascular perforation, pneumothorax, and lead dislodgement. These complications are relatively uncommon (<3%).^{26 - 27} Implantable cardioverter-defibrillator pocket hematomas occur frequently, particularly in patients receiving anticoagulation.²⁶

The longer-term morbidity associated with ICDs include inappropriate therapies (shocks) delivered most often when a supraventricular tachyarrhythmia is mistaken for a ventricular tachyarrhythmia.²⁸ The frequency of these inappropriate therapies has decreased with the development of increasingly sophisticated device algorithms that differentiate supraventricular tachycardia from ventricular tachycardia, but still occur in 15% of patients over the lifetime of their device.^{28 - 30} “Storms” of inappropriate or appropriate ICD discharges are unusual but can result in severe anxiety and depression.^{31 - 33} Lead fracture or insulation breach can occur, particularly in physically active individuals who engage in repetitive upper extremity exercises. In one series of patients with device-related complications, 2% to 7% of patients had lead fractures or insulation defects.²⁶ Infections of the device pocket or intravascular lead systems occur in 0.5% to 2% of cases, particularly in the setting of generator replacements, and may require the removal of the entire system.³⁴ Removal of these systems can be hazardous and is associated with a 1% mortality rate.³⁴ Finally, mechanical dysfunction due to manufacturing error can lead to system recalls.³⁵ Mechanical dysfunction can involve all components of the ICD system including the ICD generator (eg, battery, capacitor, screw mechanism that secures the leads to the generator) as well as all components of the pacing and defibrillating leads (Figure 2). Some of these malfunctions involve components of the ICD system, which require mandatory replacement, while others can be fixed with noninvasive programming.³⁶

Mr M is concerned that the physician's economic incentive to implant ICDs may both influence their recommendations and make them less sensitive to a patient's reluctance to have a device implanted. This concern is understandable and is partly why physicians are compelled to practice evidence-based selection for ICD implantation according to published criteria.¹⁴ Ultimately, if Mr M is concerned that his physician does not have his best interests in mind, he should pursue a second opinion.

The health care system as a whole should be concerned with the expenditures related to ICDs. The above noted trials have confirmed the superiority of the ICD compared with standard medical therapy and have resulted in a startling increase in the number of ICDs implanted in the United States.³⁷ According to the Massachusetts Medicare fee schedule, the Medicare diagnosis related group payment to the hospital for ICD implantation is currently $45 000, plus the additional physician fee of approximately $1000.³⁸ A cost-based analysis that included the above-described trials found that ICD implantation for primary prevention added 1 to 3 quality-adjusted life years (QALYs). The calculated cost per QALY saved ranged from $34 900 to $70 200.³⁹ Typically, costs within the range of $50 000 to $100 000 per QALY gained are acceptable in the United States for preventive treatments. Conversely, a recently published economic analysis of the MADIT 2 study that showed a significant but lower reduction in risk of death found an unfavorable cost of $235 000 per year of life saved over the short duration (3.5 years) of the study.⁴⁰ Mr M's characteristics place him in this group. The authors estimated that costs would decline to $80 000 to $110 000 per year of life saved over 10 to 12 years of follow-up.

Implantable cardioverter-defibrillators should be interrogated in the ICD clinic 2 to 3 times a year.^{41 - 42} Many physicians recommend bringing patients into the hospital once a year to induce ventricular fibrillation and make certain the device detects and treats it appropriately.^{41 - 42} The device is “interrogated” through application of a programming magnet over the skin. This is performed regularly to make certain that the ICD system is functioning appropriately and to document any abnormal heart rhythms that may have occurred between office visits. New technology will allow much of this device interrogation to occur wirelessly.⁴³

The battery life is typically 5 years, and indicators of the end of battery life begin to occur within a year of battery depletion. Battery or generator replacement is a simple procedure with minimal morbidity.²⁶ System malfunctions are indicated by a beeping tone that is audible to the patient and should trigger a visit to the ICD clinic. Recommendations for driving once an ICD is implanted vary by state.⁴⁴ If the device is placed for primary prevention (no symptomatic arrhythmia), driving is allowed 2 weeks postprocedure. In general, patients should not drive for at least 6 months following a symptomatic arrhythmia including an appropriate ICD discharge.⁴⁴ Patients should be warned that an ICD discharge is painful and generally feels like a strong punch to the chest (Figure 1A).⁴² Arrhythmias that are terminated by programmed rapid (antitachycardia) pacing (Figure 1B) are generally asymptomatic and are only detected during routine ICD evaluation.⁴²

Current ICDs are well insulated against microwaves and are unaffected by cell phones.^{45 - 46} Exposure to the magnetic field of an MRI or arc welder can damage the ICD system and render it ineffective.^{45 - 46} Airport security screening magnets can temporarily suspend the device's ability to detect tachyarrhythmias.⁴⁷ The placement of a magnet over the device generator will also suspend the device's ability to detect tachyarrhythmias. Magnet application is useful during surgery where electrocautery can cause signals that might otherwise be misinterpreted as ventricular tachyarrhythmia by the device.⁴⁵

Mr M meets criteria for an ICD based on the findings of the MADIT 2 study without need for a screening EPS. Despite these recommendations, Mr M feels well and would like to avoid invasive procedures, including ICD implantation. The fact that he is being treated with an optimal medical regimen and has a normal QRS duration and minimal (class II) CHF may improve his prognosis. His risk of sudden death could be further characterized in a noninvasive way with a negative TWA test, potentially offering reassurance of a reduced risk of arrhythmic death. Nonetheless, Mr M must understand that should he die from a cardiac cause in the next few years, it would likely be from a lethal ventricular arrhythmia that might be prevented by an ICD.

The final and perhaps most important part of the discussion for Mr M should relate to how the facts he has learned about ICD implantation relate to his personal views of how he lives his life. For some patients, the security of having an ICD to terminate a lethal arrhythmia provides great personal comfort for themselves and their families. In other patients, the physical presence of a device and the stress associated with its firing both appropriately and inappropriately significantly impairs quality of life. Some patients, particularly those with advanced CHF, may find a death due to a ventricular arrhythmia more appealing than that due to recurrent pulmonary edema or low output failure. Similarly, an elderly patient may not wish to have life prolonged and may find a lethal ventricular arrhythmia preferable to other possible causes of death. Many patients find discussions with ICD support groups helpful to further inform decisions about implantation of the ICD as well as clarify unresolved questions about life with a defibrillator.

Ultimately, physicians have a responsibility to not just present the data, but also to have honest and personal discussions with patients about the implications of their decisions. If the data are presented with clarity and compassion, patients can make the decisions that are right for them.

QUESTION: A recent JAMA article suggests that women are at much lower risk of sudden cardiac death than men.⁴⁸ And I wonder if you can comment on whether any of these studies included women, and if so, if they’ve been investigated as a subset.

DR ZIMETBAUM: The recently published article reported a low incidence of sudden death occurring during exercise in women compared with men. This risk was made even smaller with regular (>4 hours per week) aerobic exercise. In general, women were relatively underrepresented in the randomized trials of ICD for primary prevention (Table). Population-based studies have, however, noted relatively equivalent rates of sudden death in women compared with men.⁴⁹ Until further study of women suggests otherwise, I believe clinicians should treat men and women alike in following the guidelines mentioned.

QUESTION: A lot of this decision hangs on the question of what the patient's ejection fraction is, which is notoriously subjective. Has there been any sort of push toward standardizing calculation of ejection fraction? Because easily in the community you could see people pushing toward, “Oh, it's less than 30.”

DR ZIMETBAUM: That is a very important question. In most clinical situations left ventricular function is determined by a fairly subjective evaluation of the contractile function, ie, a rough estimation of ejection fraction from an echocardiogram. We base our decisions on the echocardiographic ejection fraction given that this was the modality used in most studies. A more quantitative estimate of the ejection fraction can be obtained by MRI, which we occasionally employ if the echocardiographic windows are suboptimal. I am hopeful that future studies of new technologies will allow us to make decisions regarding ICD placement based on multiple risk predictors which will include but not be solely dependent upon the ejection fraction.