Complete author affiliations and disclosures are at the end of this activity.

Sudden cardiac death (SCD) strikes an estimated 325,000 Americans each year, and about 85% to 90% of these deaths are due to a first-time arrhythmic event.^[1] Efforts to prevent this tragic circumstance in patients known or believed to be at heightened risk for SCD have focused on the use of implantable cardioverter defibrillators (ICDs) to automatically detect and interrupt life-threatening arrhythmias. Recently, the use of ICDs has expanded to include primary prevention in high-risk heart failure (HF) patients with ventricular dysfunction and either ischemic or nonischemic heart disease. Several large studies demonstrated conclusively that ICDs are a life-saving therapy in these patient populations, and this led to recent changes in reimbursement that greatly increased the eligible patient pool and the number of devices being implanted.

However, not all has been smooth sailing. The rapid uptake of ICDs for primary prevention has raised important cost concerns and prompted numerous discussions about proper patient selection criteria. Moreover, recent device recalls and advisories have caused some eligible patients, and their physicians, to eschew ICD implantation altogether, potentially leaving a substantial number of people at risk of unprotected, life-threatening arrhythmic events. As with any medical therapy, ICDs are not 100% effective, nor are current methods of selecting potential candidates for the therapy. However, patient selection criteria are continuously being refined, and new tools are emerging that may help physicians target ICD therapy to those who will benefit most, thus bringing us closer to the ultimate goal of simultaneously reducing unnecessary costs while optimizing patient safety and outcomes.

SCD Risk in HF Patients

With an estimated 5 million Americans currently living with HF and predictions that prevalence will increase 2- to 3-fold by the year 2010,^[2] the challenge of appropriately managing these patients is substantial. HF is a complex, age-related syndrome precipitated by a number of underlying cardiovascular disorders, including chronic hypertension, valvular disease, congenital heart defects, and ischemic heart disease/myocardial infarction (MI) -- all of which set in motion a complicated and interrelated set of compensatory molecular and cellular injury responses in the myocardium.

These maladaptive compensatory responses alter cellular expression and repair processes and eventually lead to anatomic remodeling of the heart, ventricular enlargement, and a chronic decline in the heart's ability to pump blood throughout the body. Of importance in terms of SCD risk, the syndrome also results in compensatory activation of neurohormonal pathways and anti-inflammatory mediators that can result in disturbances in the heart's electrical activity, which increases the risk of potentially serious ventricular arrhythmias and SCD. In fact, the available evidence suggests that the incidence of SCD is 6 to 9 times higher in people with HF than in the general population (Figure 1).^[3]

Figure 1. Incidence of sudden death in specific populations and the annual numbers of sudden deaths in those populations. Source: Huikuri H, et al. N Engl J Med. 2001;345:1473-1482. Permission pending.

Although there have been significant advances over the years in both drug and device therapy for HF that have transformed the disease into a largely treatable chronic condition, HF still takes a significant toll in terms of treatment costs and quality of life. US hospitalizations for HF have continued to increase, with hospital discharges for HF totaling 399,000 in 1979 and growing to 1,093,000 in 2003, an increase of 174% (Figure 2).^[3] Moreover, the total economic cost (both direct and indirect) now approaches $30 billion per year.^[3]

Figure 2. Hospital discharges for congestive heart failure by gender, United States, 1970-2002. Permission pending.

Deaths from HF are also on the rise as the proportion of elderly in the population grows. Although the HF death rate declined 2% between 1993 and 2003, the absolute number of deaths during that time span increased 20.5%.^[3] HF now causes or contributes to over 286,000 deaths each year in the United States.^[3]

Even with the treatment advances of the past several decades, less than half of HF patients survive more than 5 years following their diagnosis, and 1 in 5 dies within 1 year of being diagnosed.^[3] Notably, most who do survive longer will experience a dramatic decline in functional capacity and quality of life, a fact that illustrates medicine's continued shortcomings in adequately managing this patient population. These advanced HF patients typically die of pump failure due to progressive dilated cardiomyopathy, while earlier-stage HF patients are more likely to succumb to SCD due to arrhythmic changes in the heart. Thus, although SCD risk increases as HF progresses, there is some evidence to suggest that SCD risk should be addressed earlier rather than later in the progression of the disease (see below).

Preventing SCD is paramount, since survival following sudden cardiac arrest is dismal, with only about 5%-6% of patients surviving their first episode of cardiac arrest.^[3,4] Each year, an estimated 335,000 people in the United States die of heart disease outside the hospital, which accounts for about half of all deaths due to coronary heart disease.^[3] The bulk of these out-of-hospital events are sudden deaths due to cardiac arrest, and up to 90% are caused by ventricular tachyarrhythmias leading to ventricular fibrillation (VF).^[5]

Since HF patients are at a significantly increased risk of SCD, and because this risk increases as ventricular function declines, the use of ICDs has emerged as an important therapy for primary prevention in the HF population. ICDs are designed to detect ventricular tachycardia (VT) that can lead to VF and deliver a shock to restore a normal rhythm. They may also deliver antitachycardiac pacing to terminate VT. The benefit of ICDs in selected HF patients has been demonstrated conclusively by several studies, most notably the 2 landmark trials, MADIT II (Multicenter Automatic Defibrillator Implantation Trial)^[6] and SCD-HeFT (Sudden Cardiac Death in Heart Failure Trial).^[7] A summary of major ICD studies is shown in Figure 3.

Figure 3. Landmark primary and secondary prevention ICD trials. Hazard ratios (vertical line) and 95% confidence intervals (horizontal lines) for death from any cause in the ICD group compared with the non-ICD group. *Includes only ICD and amiodarone patients from CASH. CABG = coronary artery bypass graft surgery; EP = electrophysiologic study; LVD = left ventricular dysfunction; LVEF = left ventricular ejection fraction; MI = myocardial infarction; N = number of patients; NICM = nonischemic cardiomyopathy; NSVT = nonsustained ventricular tachycardia; PVCs = premature ventricular complexes; SAECG = signal-averaged electrocardiogram.
Source: 2005 ACC/AHA Guidelines; permission pending.

MADIT II^[6] was the first large study to evaluate the use of ICDs in patients with ischemic heart disease who had suffered an MI and had impaired left ventricular (LV) function (ejection fraction [EF] ≤ 30%; mean EF = 23%) but no other known risk factors for SCD. About 70% of the patients enrolled in MADIT II were in New York Heart Association (NYHA) class I or II HF. The study enrolled 1232 patients who had experienced an MI at least 1 month prior to study entry and randomized them in a 3:2 ratio to receive either optimal medical therapy (OMT) plus an ICD (n = 742) or OMT therapy alone (n = 490). After an average follow-up of 20 months, the ICD group demonstrated a significant 31% relative reduction in mortality (P =.016; hazard ratio [HR] 0.69) compared with the OMT-only group. However, the trial also showed an unexplainable, slightly higher rate of hospitalization for HF in the ICD group, a finding that is still under debate. The researchers suggested several potential explanations for this result, including the possibility that hospitalization was more frequent in this group simply because more ICD patients survived to develop advanced HF. It may also be possible that ICD shocks have a deleterious effect on LV function that could impact hospitalization, or that there are negative effects from chronic right ventricular pacing, which has been linked in other studies to ventricular dyssynchrony, HF hospitalization, and death.^[8] Despite this unresolved issue, the substantial survival benefit associated with ICD therapy that was observed in MADIT II is compelling and outweighs concerns about increased hospitalizations.

A similar survival benefit with ICD therapy was demonstrated in the subsequent SCD-HeFT study,^[7] which, unlike MADIT II, enrolled HF patients with both ischemic and nonischemic disease. The aim of SCD-HeFT was to determine whether treatment with the antiarrhythmic drug amiodarone or an ICD could reduce all-cause mortality in patients with NYHA class II/III HF and reduced ventricular function (EF ≤ 35%). The study enrolled 2521 patients and randomized them to receive OMT plus placebo (n = 847), OMT plus shock-only (ventricular backup pacing) ICD therapy (n = 829), or OMT plus amiodarone (n = 845). Patients were followed for a median of 3.8 years, during which time 29% of the patients in the placebo group died, compared with 22% in the ICD group, which amounts to a 23% relative reduction in all-cause mortality in the ICD group (HR 0.77; P =.007). There was no survival benefit observed in the patients who received amiodarone.

The reduction in mortality associated with ICD therapy in the SCD-HeFT study was similar regardless of the etiology of the disease (ischemic HR = 0.79; P = .05; or nonischemic HR = 0.73; P = .06); however, there were significant differences based on a NYHA class subgroup analysis. NYHA class II patients who received ICD therapy demonstrated a significant reduction in mortality, while those in class III HF did not. The relative risk reduction for the class II ICD patients was 46% with an absolute mortality reduction of 11.9%, compared with no statistically significant reduction in risk in the NYHA class III patients. The finding suggesting that class II patients may benefit more from ICD therapy than their more impaired class III counterparts is also supported by an analysis of the SCD-HeFT data. When the SCD-HeFT results were evaluated by 6-minute-walk testing and from the Duke Activity Status Index, the patients with the most functional disability did not demonstrate a reduction in mortality. The implication is that the reduction in arrhythmic mortality is balanced by an increase in HF-related mortality.

Since NYHA class can vary up or down 1 class within the span of a few weeks or months, and because class designation depends to a certain degree upon who is assessing the patient, only the primary study results should be used to guide physician practice. In addition, most of the NYHA III patients are receiving biventricular pacing devices to deliver cardiac resynchronization therapy (CRT). When the sudden death rate in the control and CRT pacemaker arms of the CARE-HF (Cardiac Resynchronization -- Heart Failure)^[9] trial and the COMPANION trial are compared with the CRT-ICD limb of the COMPANION (Comparison of Medical Therapy, Pacing, and Defibrillation in Heart Failure)^[10] trial, it is clear that NYHA III patients have a significantly increased rate of SCD unless they also have an ICD implanted. Therefore, both class II and class III patients should be considered appropriate candidates for ICD implantation.

The striking reduction in mortality observed in the MADIT II and SCD-HeFT trials had a major impact on reimbursement for ICD primary prevention in the United States, which in turn accelerated the adoption of this device therapy. In 2002, ICD product labeling was expanded by the US Food and Drug Administration (FDA) to include the MADIT II patient population. However, the Centers for Medicare and Medicaid Services (CMS) initially restricted Medicare reimbursement to a subset of patients with QRS duration > 120 ms, which automatically excluded over half of otherwise eligible patients and made it difficult for physicians to adequately prescribe the therapy. Moreover, the original coverage criteria also limited reimbursement to patients with an EF ≤ 30%, which further limited the therapy.

This changed in 2005 following publication of the SCD-HeFT results, which prompted CMS to expand ICD primary prevention reimbursement criteria to include a wider range of patients. Coverage is now provided for both ischemic and nonischemic HF patients with an EF ≤ 35%, with no stipulation with regard to QRS duration (Table 1).^[11,12] (Of note, however, coverage is routinely provided for the use of conservatively programmed, single-lead ICDs (similar to those employed in SCD-HeFT); whereas dual-chamber devices are covered only if justified in the patient's medical record. In addition, patients receiving ICDs for primary prevention outside of an IDE-approved clinical trial must be entered into a national registry database in order to be eligible for Medicare coverage.)

Table 1. Centers for Medicare & Medicaid Services (CMS) Coverage Requirements for Implantable Cardioverter-Defibrillator Implantation for Primary Prevention*

*For all groups, patients must not have (1) cardiogenic shock or symptomatic hypotension while in a stable baseline rhythm; (2) undergone CABG surgery or PTCA within the past 3 months; (3) experienced an acute MI within the past 40 days; (4) clinical symptoms or findings that would make them a candidate for coronary revascularization; (5) irreversible brain damage from preexisting cerebral disease; or (6) any disease, other than cardiac disease, associated with a likelihood of survival less than 1 year.
^†The CMS has determined that the evidence, though less compelling at this time, is adequate to conclude that an ICD is reasonable and necessary for this subset of patients.
CABG = coronary artery bypass graft; CRT = cardiac resynchronization therapy; EF = ejection fraction; IDCM = ischemic dilated cardiomyopathy; MI = myocardial infarction; NIDCM = nonischemic dilated cardiomyopathy; NYHA = New York Heart Association; PTCA = percutaneous transluminal coronary angioplasty.

This expanded coverage increased the number of Medicare beneficiaries eligible for ICD implantation to over 500,000 and spurred rapid acceptance and implantation of these devices for primary prevention in the United States.^[13] However, cost issues have raised numerous debates of late about the appropriate use of ICD therapy for primary prevention, and safety concerns that have arisen over the past 18 months have resulted in a slight pull-back in ICD implantations in recent months. Still, the number of ICD implants performed worldwide is expected to continue to rise in the coming years (Figure 4). See below for further discussion of cost and safety concerns regarding ICD therapy.

Figure 4. Exponential increase of ICD use between 1996 and 2007E.

Concerns about ensuring the appropriate use of ICD therapy and avoiding the costs and risks associated with unnecessary implantations and inappropriate shocks have put the issues of proper patient selection and risk stratification in the spotlight. Revised physician practice guidelines jointly released in 2005 by the American Heart Association (AHA) and American College of Cardiology (ACC)^[14] reflect the new data amassed by landmark studies over the past several years, and the guideline recommendations closely follow study inclusion criteria. The guidelines assign ICD therapy a class I designation for the primary prevention of SCD in HF patients with ischemic or nonischemic disease who have an EF ≤ 30% to 40%, are NYHA class II or III, are receiving chronic OMT, and have a reasonable expectation of survival with good functional status for more than 1 year. A class I designation was also given to patients with EF ≤ 35% with inducible VT on electrophysiologic (EP) study, including those with coronary artery disease and dilated nonischemic cardiomyopathy. In addition, ICD therapy was listed as "reasonable" (class IIa designation) for primary prevention in NYHA class I HF patients with LV dysfunction due to prior MI (EF 30% to 35%) who are on OMT and have a survival expectation of more than 1 year, and in patients with recurrent stable VT and a normal or near-normal EF. The revised recommendations for ICD therapy are summarized in Table 2.

Table 2. Selected Treatment Guideline Recommendations for Primary ICD Prevention, Released Jointly by AHA/ACC in 2005

*All eligible patients must also have reasonable expectation of more than 1-year survival with good functional status.
DCM = dilated cardiomyopathy; EF = ejection fraction; HF = heart failure; ICD = implantable cardioverter defibrillator; LV = left ventricular; MI = myocardial infarction; NYHA = New York Heart Association; OMT = optimal medical therapy; SCD = sudden cardiac death; VT/VF = ventricular tachycardia/ventricular fibrillation

Closely following the recommendations set forth by the ACC/AHA, the Heart Failure Society of America (HFSA)^[15] also released revised guidelines this year highlighting the new role of device-based treatment as a first-tier therapy in HF patients. According to the HFSA document, prophylactic ICD implantation should be considered in patients with or without concomitant coronary artery disease who have mild to moderate HF symptoms (NYHA class II/III) and reduced ventricular function (EF ≤ 30%). In addition, prophylactic ICD implantation may also be considered in patients with EF between 31% and 35%, and concomitant ICD implantation should also be considered in NYHA class III/IV patients undergoing CRT device implantation. The guidelines recommend against ICD placement in chronic, severe, refractory HF patients when there is no reasonable expectation for improvement, although the authors note that ICD implantation is under investigation as a "bridge to transplant" in NYHA class IV patients. In newly diagnosed HF patients, the guidelines recommend 3-6 months of treatment with beta-blockers and angiotensin converting enzyme (ACE) inhibitors before considering ICD implantation.

As was the case with the landmark studies, the revised guidelines emphasize the use of impaired ventricular function, as determined by EF, as the primary method of identifying patients at elevated risk for SCD who may be candidates for prophylactic ICD implantation. Although EF is the most widely used tool for risk stratification, and the best tool we have today, it is far from ideal, as many patients with a low EF will not experience life-threatening ventricular arrhythmias. In fact, over 80% of patients in the SCD-HeFT trial with an EF ≤ 35% did not receive an appropriate shock over the follow-up period. Despite its shortcomings, EF will remain the gold standard for assessing risk for the foreseeable future, as there are currently no methods available to definitively identify HF patients who will have an ICD-mediated event in the future or, for that matter, to rule out those who won't. However, new technologies are hopefully bringing us closer to this goal.

In addition to EF, there are several other noninvasive risk stratification tools that can be employed to help physicians decide which patients are the most appropriate ICD candidates. These are used most often in cases where patients have borderline indications for an ICD, or for those who may be appropriate candidates but are reluctant to undergo device implantation. Methods in use today include exercise treadmill testing, Holter monitoring, signal-averaged electrocardiogram testing, heart rate variability, baroreceptor sensitivity testing, and QT dispersion. In addition, spontaneous ventricular arrhythmias in post-MI patients may have value for identifying high-risk patients following MI, although there is no evidence that suppression of these arrhythmias improves survival in this group.

While these methods can sometimes be helpful, they all have drawbacks to routine use -- either they require significant additional time and/or expertise to conduct, or they lack the specificity, reproducibility, and/or accuracy to be applied to a wide range of patients.

Microvolt T-Wave Alternans

Recently, microvolt T-wave alternans (MTWA) has emerged as a promising new tool for risk stratification in potential primary prevention patients. MTWA is a noninvasive, ECG-based test that can be performed in a physician's office with modifications to currently available exercise testing equipment. The test measures beat-to-beat microvolt variations in the shape, amplitude, or timing of the ECG T-wave that have been linked with the development of clinical ventricular arrhythmias.^[16] Studies suggest the test may be able to play an important role in helping to rule out patients who are at low risk and may not benefit from ICD implantation (the test's positive predictive value for identifying high-risk patients for implantation is less compelling).

An MTWA test that utilizes a proprietary spectral analysis algorithm to identify and assess the ECG T-wave variations is cleared by the FDA for use in stratifying risk of ventricular tachyarrhythmias and SCD. In March 2006, CMS issued a national coverage decision that established nationwide Medicare reimbursement for this test to aid in the selection of heart disease patients for ICD therapy when used in conjunction with other risk-assessment tests.^[17]

The CMS decision was based on the results of a large, prospective, multicenter trial that was published in the Journal of the American College of Cardiology in early 2006,^[18] demonstrating the benefits of MTWA for ICD risk stratification in heart disease patients without a history of prior arrhythmic events. The study enrolled 549 patients with EF ≤ 40% and no history of sustained ventricular arrhythmias. Patients with atrial fibrillation, unstable coronary artery disease, or NYHA class IV HF were excluded from the trial. All participants received an MTWA exercise stress test at baseline. The test result was classified as "abnormal" (high risk) if the heart rate at onset of T-wave variations was ≤ 110 beats/min or the result was indeterminate, and "normal" (low risk) if the maximum negative heart rate was ≥ 105 beats/min. The primary outcome was all-cause mortality or nonfatal sustained ventricular arrhythmias.

About one third of the patients were classified as low risk based on MTWA testing; the rest were classified as high risk. After a mean follow-up of 20 months, the 2-year actuarial event rate was 15% in the high-risk group, compared with 2.5% in the low-risk group (hazard ratio 6.5, P < .001). The high-risk patients experienced a total of 47 events over the follow-up period, compared with only 4 events among the low-risk patients, and the low-risk group had a 2-year survival rate of 97.5%. Moreover, the results were the same regardless of disease etiology (ischemic or nonischemic disease), ventricular function, age, gender, or severity of HF. According to the study authors, the MTWA test may be able to rule out ICD therapy in up to one third of current patients. Thus, using this strategy to narrow patient selection, only about 7 ICDs would have to be implanted to save 1 life, compared with at least double that number with current patient selection criteria.

These results support outcomes observed in previous, smaller studies of MTWA, and they suggest that heart disease patients with a normal MTWA test, regardless of disease etiology, likely have a < 5% risk of experiencing either death or sustained ventricular arrhythmia over the next 2 years.

ABCD trial. The role of MTWA testing became a bit clearer with the results of the prospective, multicenter Alternans Before Cardioverter Defibrillator (ABCD) study, presented at this year's AHA meeting.^[19] The trial looked at MTWA testing in a series of 566 patients with coronary artery disease, EF ≤ 40%, and a history of unsustained VT but no sustained ventricular arrhythmias. The majority of the patients (70%) were in NYHA class II/III HF, and over 80% were on OMT. Patients received both MTWA testing and an invasive EP study, and those with a positive result on either test were required to receive an ICD. In those with a negative result on both tests, ICD implantation was recommended, but not required and the decision to implant was left up to the individual physician. Altogether, 70% of the latter group had ICDs implanted. Patients were evaluated over a 2-year follow-up period for arrhythmic event rate (including shocks for VT/VF, SCD, or antitachycardia pacing for VT).

Of the patients enrolled in the study, 46% had a positive MTWA test, 25% had an indeterminate result, and 29% had a negative test result, which is similar to the test outcomes observed in the previous MTWA study discussed above. Moreover, 39% had a positive EP study. Over the course of follow-up, 65 patients in the trial experienced VT/VF, SCD, or antitachycardia pacing for VT termination, resulting in an event rate of 7.5% at 1 year and 14% at 2 years.

Of interest, the event rate was lowest among patients who had both a negative MTWA and a negative EP test and was highest in patients who had positive results on both tests (Table 3). This suggests that neither test may be sufficiently predictive when used alone, but the greatest value for risk stratification may be achieved when the tests are combined synergistically, perhaps in a sequential fashion, to identify those at lowest risk. This would most likely involve performing the noninvasive MTWA test first and confirming those results selectively with an invasive EP test before ruling out patients for ICD therapy. The study results suggest that this approach may be effective in ruling out about 17% of patients for ICD therapy with an acceptably low risk of an unprotected event in this group.

Table 3. Results of the ABCD Test: Primary Endpoint Event Rates^†

*Includes shocks for VT/VT, SCD, or antitachycardia pacing for VT.
^†Comparison of Kaplan-Meier event rates for MTWA abnormal/EPS positive vs MTWA normal/EPS negative yielded a P value of .017.
EPS = electrophysiologic study; MTWA = microvolt T-wave alternans

However, using this approach, it may be necessary to repeat MTWA testing on an annual basis, since the study also found that the predictive value of MTWA may change over time. Although the initial positive and negative predictive values for both the MTWA and EP tests were similar (about 8%-10% and 96%, respectively), the EP test's predictive value remained stable over follow-up, while the predictive value of the MTWA test dissipated after the first year.

Although these results suggest that MTWA testing is a promising new risk-stratification tool, questions still remain with regard to what level of risk physicians and patients are willing to accept in those who have a normal MTWA test result and decide not to undergo ICD implantation. In the ABCD trial, the risk of an unprotected event in low-risk patients who were ruled out for ICD therapy using MTWA testing alone was calculated at about 1.2% in 12 months, which is fairly low, but is still not zero. Whether this is an acceptable level of risk remains to be determined. Moreover, MTWA has other drawbacks that limit its usefulness; for example, it cannot be used in patients with atrial fibrillation, who represent a substantial proportion of potential ICD candidates. Thus, it appears that MTWA testing can certainly play a role in risk stratification, most likely as part of an approach that combines several methodologies, but this test is not likely to provide the ultimate answer to the risk-stratification dilemma.

We await the results of the ongoing MASTER I (Microvolt T Wave Alternans Testing for Risk Stratification of Post MI Patients) trial to help further define the utility of MTWA to risk-stratify patients for SCD.^[20]

Risk-stratification tools such as MTWA may not only help reduce the number of unnecessary ICD implantations, but could also address lingering concerns about the cost-effectiveness of ICD therapy. A recently published cost analysis from the SCD-HeFT trial^[21] found that patients must survive nearly a decade following ICD implantation in order for the therapy to be considered cost-effective based on currently established benchmarks. For this reason, the study concluded that NYHA class II HF patients were a more economically attractive group for prophylactic ICD implantation than sicker class III patients. According to the study authors, the use of conservatively programmed, single-lead ICDs for primary prevention is an "economically efficient" way to improve patient outcomes, providing the devices are implanted in stable, moderately symptomatic patients (particularly NYHA class II) who survive at least 8 years following implantation.

This was one of the most comprehensive studies to date on the topic of ICD cost-effectiveness, employing statistical modeling methods to estimate treatment-specific costs up to 14 years following implantation. On the basis of this analysis, a 5-year survival would yield an ICD cost-effectiveness ratio (expressed as the incremental lifetime cost to save a life-year relative to placebo) of $127,503 per life-year -- this declined to $88,657 at 8 years, and $58,510 at 12 years. When NYHA class II patients were analyzed separately, the analysis yielded better results, producing a discounted ICD cost-effectiveness ratio of $29,872, due to the greater survival benefit in this group.

The analysis also revealed no evidence that ICD patients had significantly more hospitalizations, major cardiac procedures, or outpatient visits over the first 5 years of follow-up than patients in the placebo arm. Taking into account projected life expectancy, the lifetime cost for patients in the ICD arm was $158,840, compared with $79,028 for the placebo arm. Thus, the lifetime incremental cost of ICD implantation was estimated at $79,812 ($62,420 with 3% discounting). It remains to be seen, however, whether these results will apply to the real-world population of patients, who are older and generally sicker than patients included in the landmark trials such as SCD-HeFT. Nevertheless, cost-effectiveness of ICD prophylaxis will likely be a concern for years to come unless a practical and effective method for stratifying risk becomes widely available.

ICD safety also has emerged over the past 18 months as a topic of great interest among both physicians and patients, particularly in light of recent high-profile advisories and recalls from various ICD manufacturers. Several analyses performed earlier this year^[22-24] confirmed that clinically significant ICD device malfunctions are not rare and are significantly more common than malfunctions associated with pacemakers. In Maisel's meta-analysis of device registries,^[22] the malfunction rate associated with ICDs was 20-fold higher than that associated with pacemakers, totaling 26.5 malfunctions per 1000 person-years. Although device malfunction rates apparently declined in the latter years of the study (2003 and 2004), it is unclear whether this is a continuing trend. Maisel and colleagues also looked at data reported to the FDA^[23] and found 61 confirmed deaths documented in conjunction with device malfunction (30 in pacemaker patients and 31 in ICD patients) among 2.7 million devices implanted. Perhaps most troubling is that there is often a time delay of more than 3 years in the life cycle of a device model before problems are uncovered, underscoring the need for continuing and vigilant monitoring and patient follow-up.

While the first inclination for physicians and patients when faced with an ICD recall is often to replace the device, a study from Gould and Krahn^[24] illustrates the trade-offs encountered and the difficulty that a device replacement decision entails. In their analysis, the researchers surveyed 17 ICD implanting centers in Canada to assess complication rates as a result of ICD replacements due to advisories that occurred between October 2004 and October 2005. Of 2915 patients with advisory devices, 533 underwent elective ICD replacement. Over a 2.7-month follow-up, complications directly attributable to elective advisory device replacement occurred in 8.1% of these patients and major complications that required an additional operation occurred in 5.8%. Major complications included 2 deaths after device extraction for pocket infection. According to the study authors, the data suggest that 6000 ICD replacements would lead to 100 infections requiring extraction and 1 extraction-related death.

In an attempt to address the safety issue head-on, the Heart Rhythm Society (HRS)^[25] recently released new policy guidelines on device safety that were drafted by a 15-member expert Task Force. The guidelines provide detailed recommendations to industry, physicians, and the FDA aimed at improving the safety of implantable cardiac rhythm devices as well as boosting the effectiveness and transparency of current postmarket surveillance and reporting systems. Among other things, the guidelines urge physicians to improve communication with their patients, informing them not only of the risks and benefits of the implantation procedure, but also of expected device performance, battery life, and potential malfunction rates. Physicians are also urged to return all explanted devices to the manufacturer for analysis. The document also assists physicians in deciding when device replacement is warranted, and it urges physicians to consider potential alternatives to device replacement whenever possible in order to minimize patient risk.

The HRS postmarket device safety guidelines are the first step in the right direction, but we must not lose sight of the fact that patient-care decisions should be based on individual patient needs and that each case is unique. Although future technological advances hopefully will improve product reliability, and new methodologies for risk stratification may vastly improve our patient selection capabilities, physicians must still consider the best interests of each patient when deciding whether to implant an ICD for primary prevention or whether to replace a device under recall. This can only be accomplished if the patients and their family fully understand the benefits and risks involved and have full confidence that the final decision is the right one for them. The significant life-saving attributes of ICDs in the primary-prevention population will continue to outweigh the risks as long as these criteria are met.

Supported by an independent educational grant from Medtronic.

As an organization accredited by the ACCME, Medscape, LLC requires everyone who is in a position to control the content of an education activity to disclose all relevant financial relationships with any commercial interest. The ACCME defines "relevant financial relationships" as financial relationships in any amount, occurring within the past 12 months, including financial relationships of a spouse or life partner, that could create a conflict of interest.

Medscape, LLC encourages Authors to identify investigational products or off-label uses of products regulated by the US Food and Drug Administration, at first mention and where appropriate in the content.