Over the past decade, there has been a concerted effort by health professionals, industry, and the government to reduce sudden cardiac death (SCD) caused by ventricular arrhythmias. One of the major milestones in cardiovascular disease has been the realization that an implantable cardioverter defibrillator (ICD) can significantly reduce mortality in a select group of patients.4,11 We now know from multiple randomized controlled trials that patients with a reduced ejection fraction (EF), prior history of cardiac arrest, and heart failure symptoms are at highest risk from a fatal ventricular arrhythmia.5,6,12 Furthermore, the duration of disease increases the risk of sudden cardiac death.12,13 Unfortunately, the method of risk stratifying patients for ICDs is neither 100% sensitive nor specific. As a result, some patients may have ICD implantation without any subsequent life-threatening arrhythmias. Thus, they are at risk for inappropriate shocks and complications relating to implantation without ever needing the life-saving benefits of the ICD. The greatest challenge in treating sudden cardiac death within this high-risk group has been identifying exactly who benefits the most from an ICD and when the right time to implant is. Prior research designed to answer this question have typically divided enrolled patients by the cause of their cardiomyopathy. Patients may have low ejection fractions from either coronary disease (ischemic cardiomyopathy) or a second category of illnesses called nonischemic cardiomyopathy. This second category is comprised of a wide array of patients who have sustained viral illness, rapid sustained heart rhythms, inherited diseases, pregnancy-related disease or unknown causes (termed idiopathic), which all ultimately lead to low ejection fractions. Patients with either ischemic or nonischemic cardiomyopathy have been studied in large randomized controlled trials and have concluded that both groups benefit from ICDs. What makes these groups different is the issue of when to implant. Prior trials have required patients with ischemic cardiomyopathy to be three months past a coronary revascularization procedure or 40 days after a myocardial infarction. There may be exceptions where a patient with ischemic cardiomyopathy may be eligible for an ICD within these timeframes, such as cardiac arrest or sustained ventricular tachycardia, but in general, these time windows are fairly well agreed upon. When considering patients with non-ischemic cardiomyopathy, the exact timing of when to implant an ICD has been less clear. Early trials have suggested that patients with NIMCP who carry the diagnosis for less than nine months duration are not afforded any benefit from an ICD.14,15 These trials have been criticized for the small number of enrollees. More recently, two larger trials have investigated patients with NICMP who have had symptoms of heart failure as early as 1 week and have suggested that these patients may benefit from an ICD.7,8 Many clinicians have been cautious with patients who have been symptomatic for such a small period of time. It is well known that when newly diagnosed cardiomyopathy patients are placed on standard heart failure medications, a portion of them improve with an increase in their EF.16,19 A higher EF reduces the risk of SCD, which in turn, lessens the importance of implanting an ICD. The puzzling task has been identifying which patients improve their EF on medications, and which patients can take the risk of waiting to be treated without succumbing to a lethal arrhythmia. In June of 2005, the Centers for Medicare and Medicaid Services issued a proposed reimbursement guideline suggesting that patients with NICMP should have symptoms for at least nine months before qualifying for reimbursement. It is important to realize that these guidelines provide a framework to determine which patients will receive federal funding through insurance for medical procedures, and does not dictate how any one patient should be clinically assessed. Unfortunately, the devices are costly and if they are not reimbursed by insurance, they are not implanted. This makes the CMS reimbursement guideline a critical document, as it ultimately determines which patients go on to have an ICD implanted. A discussion regarding the clinical benefit versus societal cost of an ICD has been previously published.20 Recently, CMS has scrutinized which patients with NICMP would be eligible for reimbursement after an ICD is implanted. In January 2005, CMS issued a draft proposal stating patients with NICMP for less than nine months duration would not be reimbursed for an ICD procedure. This decision was based on two small trials, AMIOVIRT and CAT, showing no benefit was afforded to the study patients.14,15 The trials were largely inconclusive because of the small study size. Conversely, the DEFINITE and SCD-HeFT trials included patients with NICMP for less than nine months, and did show a mortality benefit from an ICD, although the trials did not specifically address the question of heart failure duration and risk from SCD.7,8 In an effort to clarify how we should treat recently diagnosed NICMP patients, we conducted a retrospective review of our own patient population at the Tufts University-New England Medical Center.21 We investigated over 100 consecutive patients with NICMP referred to our center for evaluation. When we compared patients with recently diagnosed NICMP (less than nine months duration) to patients with established disease (greater than nine months), we discovered that both groups had the same risk of ventricular arrhythmias. In other words, the duration of NICMP was a poor discriminator of SCD risk. A re-analysis of the DEFINITE trial also showed no significant difference in survival in patients with symptoms of NICMP for 3, 9, and greater than 9 months in a randomized population.22 Although both analyses have limitations, it raises uncertainty as to whether prior beliefs regarding when to implant ICDs are justified. CMS released their final reimbursement guidelines and revised the duration of NICMP required to become eligible for an ICD from nine months to three months. They do require patients with NICMP for three to nine months to be identified in a national ICD registry. There are many unanswered questions with regards to risk stratifying patients with NICMP. The task of predicting which patients with NICMP might recover left ventricular function with heart failure treatment is paramount to risk stratification algorithms. Clinicians currently use the etiology of dilated cardiomyopathy to help risk stratify patients using the assumption that certain forms of NICMP may respond to treatment (e.g., pregnancy-related and sustained tachycardias), and therefore should not have ICD implants. Unfortunately, the exact etiology of NICMP is unknown in many cases. Furthermore, the fortunate patients who do have reversible dilated cardiomyopathies may be exposed to the risk of SCD while waiting for their ventricular function to improve. The approach to SCD prevention has changed considerably from the days of antiarrhythmics and electrophysiology testing to the development of the ICD. It is now widely accepted that the ICD can successfully treat lethal arrhythmias. Questions remain regarding how best to implement this technology in a cost-conscious healthcare environment while providing optimal care to our patients. To further define the role of the ICD in SCD management, the national ICD Registry, developed in partnership with the Heart Rhythm Society, has been created; research using this registry may assist in further clarifying which patients benefit and when device implantation should be carried out. Until then, patients who fall outside the scope of prior ICD studies should be profiled using a combination of available trial data and individualized risk stratification to make decisions on the appropriateness of the ICD.