The transvenous implantable cardioverter-defibrillator (T-ICD) is a lifesaving device that has proven its safety and efficacy during 3 decades of use in over 1 million patients. Multiple clinical trials have established its benefit in diverse patient populations for the primary and secondary prevention of sudden cardiac death. Certainly lead problems and persistent concerns with regard to costs and appropriate use are real issues. Nevertheless, we currently have devices that are reasonably reliable, cost-effective, and lifesaving. Now, a new treatment option is on the horizon, the subcutaneous implantable cardioverter-defibrillator (S-ICD) (1).

The proposed indication for an S-ICD is for patients who are candidates for ICD insertion on the basis of current guidelines and who do not have symptomatic bradycardia or spontaneous, recurring ventricular tachycardia (VT) that is reliably terminated with anti-tachycardia pacing (ATP). The S-ICD consists of an electrically active pulse generator, which is implanted near the left mid-axillary line (Figure 69_gr1), and a subcutaneous lead, consisting of sensing electrodes and a shocking coil, which is tunneled 1 to 2 cm to the left of the mid-sternal line. The S-ICD weighs 145 grams and has a lithium battery with a projected life of 5 years; therapy consists of 80-Joule (J) biphasic transthoracic shocks and 30 s of post-shock pacing. Given the available clinical data, is the potentially simpler and safer S-ICD ready for widespread use? Specifically, should patients want one?

In April 2012, the U.S. Food and Drug Administration (FDA) Cardiovascular Device panel voted to approve the S-ICD. The FDA will make its decision later in 2012 or early 2013. The panel based its approval on the results of a 180-day prospective, single-arm, 330 patient multicenter investigational device exemption (IDE) study and data from previous studies and registries (2). The primary IDE safety and efficacy endpoints were met, but the efficacy endpoint did not test the ability of the S-ICD to terminate spontaneous ventricular fibrillation (VF); rather, the efficacy endpoint was based on successful detection and termination of induced VF. This surrogate might be appropriate for transvenous ICDs, but its applicability to S-ICDs is unknown. Thus, by design, this study did not demonstrate the efficacy of S-ICD in ambulatory patients. The most frequent adverse event was inappropriate therapy, which affected 11.8% of patients, who received unnecessary 80-J shocks.

Compared with S-ICDs, single chamber T-ICDs weigh approximately one-half as much, have a longer battery life, and they defibrillate transvenously with ≤40 J. Unlike the S-ICD, the tiered therapy T-ICD provides 3 functions, namely defibrillation, ATP, and bradycardia pacing. A large proportion of ICD patients have VT that can be terminated painlessly by ATP. Twenty years ago, a pivotal study of tiered therapy ICDs concluded that the “most important advance in device therapy is the option to treat monomorphic ventricular tachycardia with antitachycardia pacing maneuvers” (3). Without ATP, a conscious patient will receive painful 80-J shocks for VT. But it is often difficult to determine which primary prevention patients will develop VT. In the MADIT-II (Multicenter Automatic Defibrillation Trial), 25% of patients—who were not inducible in the electrophysiological testing arm—experienced a clinical VT episode within the first 3 years after implant (4). Indeed, in the S-ICD IDE study, even though patients with known pace-terminable VT were excluded, 18 of 28 patients (64%) who had a spontaneous VT/VF event during follow-up received an 80-J shock for monomorphic VT. Moreover, although controversial, current data suggest that all T-ICD shocks, both appropriate and inappropriate, might be associated with reduced longevity and quality of life (5). Thus, ATP might confer important advantages that are not available in shock-only devices.

But there is another, even larger issue: the S-ICD has not yet been shown to be safe and effective in a diverse patient population, and it has not yet been shown to be non inferior to T-ICDs. Implantable defibrillators are lifesaving devices; T-ICDs perform this function reliably. Any new device introduced for a similar indication must perform comparably to current ICD systems, because the consequences of inferior performance are not tolerable. Although the results of the IDE study are encouraging, the critical question is: can we inform patients during the consent process on the basis of current knowledge that the S-ICD and T-ICD are equivalent therapies? Or, is the S-ICD the better choice? The answer is there are no data showing equivalency—much less superiority—in patients at high risk for sudden cardiac death.

Unless critical questions with regard to safety and efficacy in primary and secondary prevention are addressed, the S-ICD should be confined to certain subgroups. These include individuals without venous access and who are poor candidates for thoracotomy that might be suitable for an S-ICD. Similarly, patients who are at high risk for developing sepsis with an intravascular device might be best served with an S-ICD. It has been suggested that young patients who have hereditary heart disease might be candidates for an S-ICD, because transvenous leads are prone to failure in children and active adults; this would be a legitimate argument if we could be confident the S-ICD can provide lifesaving therapy for such patients, most of whom have a good prognosis if sudden cardiac death can be avoided.

Beyond these discrete scenarios, the immediate use of S-ICDs ought to be limited. The need for more data with this promising device should be paramount at this time. We should seize the opportunity to do the appropriate comparative effectiveness research in multiple patient cohorts. Although it can be argued that such an approach might stifle innovation, there is no reason to lower the bar for an FDA Class 3 device (devices that support or sustain human life, are of substantial importance in preventing impairment of human health, or which present a potential, unreasonable risk of illness or injury). The requirement for data supporting safety and efficacy in Class 3 devices should remain absolute when a new device like the S-ICD is intended for patients who can be treated with an approved device (e.g., the T-ICD) that has been shown to be safe and effective.

History has taught us that there should be no shortcuts in bringing life-supporting or life-sustaining products to market. If the FDA approves the S-ICD on the basis of the IDE study results, there will be no incentive for the manufacturer or independent investigators to sponsor or support a trial where the S-ICD is compared with the T-ICD. However, the FDA should consider conditional approval, which would allow the S-ICD to be used in the context of clinical trials that are designed to assess its safety and efficacy.

The S-ICD is a promising technology that could fill a gap in the treatment of ventricular tachyarrhythmias by extending this vital therapy to patients in countries where the facilities to implant T-ICDs are not available. Indeed, if this new device that is based on simplicity and ease of implantation can be deployed at less cost and with non inferior clinical outcomes, it could become a breakthrough therapy. But it is a new technology and, as such, requires the scrutiny of a comparative effectiveness trial. Only then can we tell patients that it is as effective as existing ICDs. Otherwise, a fully informed patient who is offered this device today should not want one.