The PRAETORIAN Trial: Guarded Approach to Subcutaneous ICD Best

John Mandrola, MD


September 10, 2020

Years ago, during training, I remember a colleague dissing an early model of an implantable cardioverter-defibrillator (ICD) because it was primitive. All it did was shock ventricular tachycardia (VT) or ventricular fibrillation (VF).  

Almost three decades later, a similarly primitive device has garnered favor among many electrophysiologists. Boston Scientific's subcutaneous ICD (S-ICD) uses a lead placed outside the thorax, close enough to the heart to sense cardiac activity. The generator lies in the left lateral chest, and the vector between the lead and generator senses and terminates VT or VF.

While randomized controlled trials (RCTs) such as MADIT-2 and SCD-HeFT have shown that a standard transvenous (TV) ICD saves lives compared with medical therapy, complications accrue over time. Most complications relate to the lead—dislodgement, infection, threshold rises, sensing problems, and manufacturer recalls. Since most lead issues occur late, after scar tissue has grown over the lead, removal requires a technically challenging and potentially risky procedure.

The idea behind the S-ICD is simple—if there are no leads in the veins and heart, there will be fewer long-term issues: defibrillation support without the downsides.

Yet, there are serious tradeoffs with having no hardware in the heart.

The S-ICD has no pacing support. This means no backup for bradycardia and no ability to painlessly terminate regular tachycardias—which occur often, especially in patients with structural heart disease.

Sensing is a challenge. A transvenous lead fixated in the myocardium easily "sees" a cardiac signal; a subcutaneous lead outside the ribs struggles to do this. With the S-ICD, a change in heart activation, say from a new bundle-branch block or an ablation procedure, can lead to oversensing of R and T waves and cause inappropriate shocks.

The Trial

Until Dutch-led investigators reported the PRAETORIAN trial, good intentions, industry-funded registries and marketing supported the popularity of the S-ICD. This RCT is welcome.

The trial used a noninferiority design to compare the S-ICD to standard TV-ICDs. The investigators chose a safety primary endpoint—a composite of device-related complications and inappropriate shocks. The agreed-upon margin was that the S-ICD could be 45% worse and be noninferior (hazard ratio [HR], 1.45).

Nearly 900 patients with a standard ICD indication were randomly assigned to one of the two devices. Patients were not enrolled if they had an indication for pacing or cardiac resynchronization, or if they had  VT at less than 170 beats/min.

The primary endpoint occurred in an identical number of patients in each group (n = 68) for an HR of 0.99 (95% CI, 0.71 - 1.39; P = .01), which fell within the noninferiority margin.

The two components of the primary endpoint diverged. Device-related complications favored the S-ICD (5.9% vs 9.8%). But inappropriate shocks occurred more often with the S-ICD: 41 vs 29 patients in the transvenous ICD group (HR, 1.43; 95% CI, 0.89 - 2.30).

The trial was not powered for efficacy outcomes, but appropriate shocks were more common in the S-ICD group (19.2% vs 11.5%). This is expected because an S-ICD can only terminate a VT with a shock whereas a standard TV-ICD can painlessly terminate VT with overdrive pacing. Painless termination occurred in 55% of treated VT episodes in the TV-ICD arm.

Death from any cause occurred in 83 patients (16.4%) in the S-ICD arm vs 68 patients (13.1%) in the TV-ICD arm, a difference that was not statistically significant. Sudden death and cardiovascular death were similar in both arms.

The authors concluded that if you have an indication for an ICD but no need for pacing, the S-ICD proved noninferior to the TV-ICD.  

Critical Appraisal

First the global concerns. Why did the PRAETORIAN study assess safety rather than efficacy? We know from the seminal trials that the standard TV-ICD reduces mortality compared with medical therapy. The S-ICD may be a defibrillator, but it is not that similar to a TV-ICD.

Proponents of this device argue that because registries have shown high efficacy of defibrillation at the time of initial implant of the S-ICD, the device will reduce mortality over time, similar to a TV-ICD. This may or may not be true. For instance, I have seen many patients with a TV-ICD who developed heart block years after implant and were likely saved by the backup pacing. There is no backup pacing with the S-ICD.

Defibrillation efficacy of the S-ICD over time is also uncertain. Something as simple as weight gain may impede the flow of electrons from a subcutaneous generator to leads in the chest. A 2019 German study found that 20% of patients with S-ICDs had failure of defibrillation at the time of generator change (just 5 years in this paper). Pause here to consider the plain language translation of that finding: one in five patients whose sudden death risk was high enough to warrant having an ICD had a device that may not reliably defibrillate VF. The authors reported a trend toward higher body mass index in patients with ineffective defibrillation.

Of course, patients with transvenous ICDs also gain weight. But we have long-term efficacy data with these devices and decades-long experience with reassuring defibrillation threshold testing at the time of generator changes. Proponents of the S-ICD speculate that implant techniques using deeper generator placement (submuscular) may help reduce later problems with defibrillation.

PRAETORIAN provides no reassurance of long-term efficacy of an S-ICD.

Another concern is the asymmetry of the safety outcomes. While any adverse outcome is best avoided, a pneumothorax, lead perforation, or lead repositioning at the time of an initial hospital stay is arguably less traumatic than one or more wide-awake high-voltage inappropriate shocks. That's my impression from seeing the post-traumatic stress that shocks cause. I've never seen a patient haunted by a lead repositioning.

S-ICD proponents will argue that most patients in PRAETORIAN (78%) were enrolled before an improvement in a sensing algorithm designed to reduce inappropriate shocks. Iterative improvements are always welcome, but even if this software update reduces inappropriate shocks, it remains uncontested that appropriate shocks were higher with the S-ICD—because the device cannot pace terminate VT.

Limitations and Conclusion

Recall that the upper bound of the confidence interval for the primary endpoint was 1.39, close to the threshold for noninferiority of 1.45. The authors acknowledge that the magnitude of this noninferiority margin is debatable on clinical grounds.

This is an important point because clinical-events committee members were aware of treatment assignments, which could have biased their adjudication of safety events.

Another limitation is that nearly 5% of the patients (n = 38) were lost to follow-up, a substantial number in a trial that just met noninferiority. And more patients crossed over from the S-ICD group to the TV-ICD group than vice versa (14 vs 5), raising cost concerns.

PRAETORIAN's 4-year follow-up cannot support the notion that the S-ICD is a way to avoid long-term lead complications.

Finally, on a practical note, compared with a transvenous device, the S-ICD device requires general anesthesia for implantation, has a larger generator and worse battery longevity, and costs more.

The evidentiary support for the standard ICD is a high bar. While PRAETORIAN is a start, I strongly believe we need a lot more data before accepting this new technology for the treatment of patients at risk for dying suddenly. 

John Mandrola practices cardiac electrophysiology in Louisville, Kentucky, and is a writer and podcaster for Medscape. He espouses a conservative approach to medical practice. He participates in clinical research and writes often about the state of medical evidence. 

Follow John Mandrola on Twitter

Follow | Medscape Cardiology on Twitter

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.