By Joshua D. Moss, MD
Associate Professor of Clinical Medicine, Cardiac Electrophysiology, Division of Cardiology, University of California, San Francisco
Dr. Moss reports he is a consultant for Biosense Webster and Abbott.
SYNOPSIS: In a randomized trial of the wearable cardioverter-defibrillator prescribed to patients with left ventricular dysfunction after acute myocardial infarction, overall compliance rates were low. The authors observed no significant improvement in arrhythmic death rates; yet, overall mortality was lower.
SOURCE: Olgin JE, Pletcher MJ, Vittinghoff E, et al. Wearable cardioverter-defibrillator after myocardial infarction. N Engl J Med 2018;379:1205-1215.
The risk of sudden cardiac death (SCD) in patients with reduced left ventricular ejection fraction (EF) after acute myocardial infarction (MI) is high. However, randomized trials of implantable cardioverter-defibrillators (ICDs) placed immediately after MI have not shown long-term mortality benefit. Olgin et al sought to test the hypothesis that a wearable cardioverter-defibrillator (WCD) would reduce arrhythmic death in the acute period after MI without the attendant risks of permanent ICD implant.
Patients who were hospitalized with acute MI and had an EF ≤ 35% when assessed ≥ 8 hours after MI, ≥ 8 hours after percutaneous coronary intervention (PCI), or ≥ 48 hours after coronary artery bypass grafting were enrolled within seven days of hospital discharge. In the final analysis, 2,302 patients were included from 108 centers in four countries, randomized in a 2:1 ratio to either a WCD plus medical therapy for three months (1,524 patients) or guideline-directed medical therapy alone (778 patients). Mean EF was 28%, and 83.6% of patients underwent PCI during their index hospitalization. Although the original primary outcome was meant to be death from any cause at 60 days, slow recruitment necessitated a change to “arrhythmic death” (sudden cardiac death and nonsudden death due to ventricular tachyarrhythmia) at 90 days. Overall mortality became a secondary endpoint. Importantly, cause of death was adjudicated by an independent panel blinded to group assignment and, therefore, without any data from the WCD.
Over a mean follow-up of about three months, 3.1% of the device group and 4.9% of the control group died, a significant difference. However, there was not a significant difference in the rate of “arrhythmic death” (1.6% in the device group and 2.4% in the control group). Twenty-one patients received appropriate shocks from the device, 15 of whom survived. Nine patients in the device group received at least one inappropriate shock, and three patients were hospitalized for aborted or inappropriate shock.
Device compliance declined precipitously over the course of the trial. By the end of 90 days, only 41% of patients in the device group were wearing the WCD on any given day. Median wear time was 23.4 hours immediately after randomization down to zero hours per day (interquartile range, 0-22.2 hours) at 90 days. Of the 48 deaths in the device group, 36 occurred in patients not wearing their WCD at the time of death. Patients who wore the device typically did so for > 20 hours per day. The authors performed an as-treated analysis, showing a rate of arrhythmic death of 0.37 per 100 person-months of wearing the device vs. 0.86 per 100 person-months of not wearing the device (P = 0.03). Reduction in overall mortality also was highly significant in the as-treated analysis. The authors concluded that in immediate post-MI patients with an EF ≤ 35%, a WCD did not reduce arrhythmic death significantly.
Despite the known higher risk of SCD after acute MI among patients with reduced EF, the VEST trial can in one sense be added to the list of randomized studies with “negative” primary endpoint results for defibrillators (now, both implanted and wearable). However, the extensive data collected in this trial, much of it included in a lengthy appendix, are worth a close look. It may be difficult to justify routine use of the WCD after acute MI given its cost and these results, but that does not mean it has no utility in this population. Several broad questions are worth considering:
Why was there an overall mortality benefit in the absence of significant reduction in arrhythmic death? Aside from the possibility of chance, several factors may have played a role. After the reduction in target sample size from 4,506 patients to 2,300 patients, the trial simply may have been underpowered to detect a significant reduction in arrhythmic death. Additionally, true arrhythmic death is notoriously difficult to adjudicate accurately. Of 12 patients who died while wearing a device, nine were adjudicated as having had arrhythmic death (without the benefit of data from the device itself, as the panel needed to remain blinded to therapy arm). However, more than half of that subgroup actually had no tachyarrhythmias recorded by the device; they had sinus or idioventricular rhythms that became asystole or complete heart block. As to the reduction in total mortality, in addition to potential misadjudication, it also is possible that wearing or even just keeping the device in the home offered unexpected benefits (like an additional reminder to be vigilant about medication use and symptom follow-up).
Why was device compliance so poor? Would better compliance have changed the primary outcome? The WCD can be cumbersome, both physically and emotionally. For example, there was a greater than one in 10 chance that on any given day, a patient wearing the device would have to deal with at least one arrhythmia alarm. However, overall compliance was even lower than that seen in prior registry studies of the device. The authors offered an interesting potential explanation. A randomized trial requires presumed clinical equipoise for the proposed treatment arms, as clinicians and patients must be comfortable with the possibility of randomization to either group. Patients who receive a WCD in that setting, knowing they might just as well have received nothing, may not feel as compelled to wear the device as patients prescribed a WCD as part of “standard” clinical practice. The counterargument to this is the often-higher rates of compliance seen in trials vs. clinical practice. Regardless of the reasons for low compliance, the as-treated analysis convincingly showed reduction in both arrhythmic death and overall mortality when participants wore the device. Such an analysis can be criticized as not representing “real world” use, but that does not mean a WCD cannot be a good thing. It may mean that the technology must continue to evolve and become less cumbersome to achieve optimal results. The 15 survivors who received appropriate shocks from the device probably do not regret dealing with inconveniences.
How do we better choose patients who might derive benefit from the WCD? The editorial that accompanied the article astutely suggested that such a trial, which cost millions of dollars and lasted many years, is unlikely to ever be repeated. It seems quite likely that a lower-cost device that is less burdensome to wear and more accurate in arrhythmia detection could prove beneficial in a similar broad cohort of patients. In the meantime, until such a device is developed, the best clinicians can do is choose a subset of patients for whom the benefits will justify the cost and inconvenience. Motivated patients who become invested in understanding the potential strengths and weaknesses of the device will be more likely to wear it and fare better. Hopefully, further analyses of the extensive data collected by the VEST investigators will help identify clinical factors for risk stratification.