Does the trend mean that ICDs are indicated less often?

The incidence of sudden death in heart failure patients with reduced ejection fraction, in clinical trials, has declined significantly in the last 20 years, according to a new study published in the New England Journal of Medicinethat may impact inclination to recommend implantable cardioverter-defibrillators (ICDs) for them.

The take-home message for physicians is that they should “take time to get medical therapy right and only put an ICD in if the patient’s LVEF [left ventricular ejection fraction] remains persistently low despite this treatment,” said lead author John McMurray (University of Glasgow).

“Many patients will have an improvement in LVEF above 35% and not need an ICD,” he added. “There is no need to rush. Early implantation may mean use of an expensive and unnecessary device with only the harms and without the benefits.”

McMurray led a group of heart failure clinical trialists who analyzed patient-level data from 40,195 people who had heart failure with reduced ejection fraction and who were enrolled in large clinical trials between 1994 and 2015. The rate of sudden death dropped over that period by 44%, from a 6.5% annual rate in the earliest trial to 3.3% in the latest trial. The authors wrote that the decline in sudden death “paralleled the increasing use of evidence-based pharmacotherapies that are known to reduce the incidence of sudden death. The contemporary cumulative incidence of sudden death (according to the three most recent trials in our study) is approximately 1% by 3 months and 2% or less by 6 months among patients treated with an ACE inhibitor or angiotensin-receptor blocker [ARB], a beta-blocker, and a mineralocorticoid-receptor antagonist [MRA].”

The lower rate of sudden death “suggests that it may be difficult to show a significant benefit of ICD implantation for primary prevention in most patients with heart failure with reduced ejection fraction in the current era.” ICDs, the authors noted, are expensive, may cause infections, inappropriate shocks, and may malfunction.

The authors wrote that the lower rate of sudden death means “that new efforts are needed to find a high-risk subgroup of patients who benefit from ICD implantation and in whom it is cost effective.” They raised the idea that 3 months “may be too short a period to wait to see whether there is sufficient recovery of left ventricular function to obviate the need for an ICD.”

Commenting on the study, Larry Allen (University of Colorado) said it would be a mistake to interpret the study as meaning that ICDs are not relevant. “The rate of SCD declined, but it was still 1% at 90 days, so not trivial. What is the threshold at which an ICD no longer provides value? That’s not really addressed and is a related but different discussion.”

The findings add to and support last year’s DANISH study, which found no benefit for ICD therapy in heart failure patients with a nonischemic etiology. “If the background rate of SCD goes down a lot, then the absolute benefit of SCD prevention goes down, and meanwhile there is a lot of other stuff going on with these patients,” said Allen. “So, at some point, a reduction in SCD crosses a poorly-defined threshold where the value of an ICD becomes too small.”

McMurray said that “there is a big debate over ICDs, which just got bigger following the results of the recent DANISH trial.” The 3-month waiting period “was a compromise length of time – long enough to get some medical treatment on board but not too long a wait during which the patient might die suddenly.

“The question many doctors have, and which we have tried to answer is: ‘Can I risk taking the time to initiate and up-titrate an ACE inhibitor/ARB/sacubitril-valsartan [Entresto], a beta-blocker and an MRA (and consider CRT [cardiac resynchronization therapy]) before re-checking the LVEF, or do I need to act now and implant an ICD in case my patient dies suddenly while I’m doing all this?’ A lot of ICDs are implanted before treatment is optimized and after less than 3 months treatment. Our data suggest that the risk of sudden death has dropped to as low as 1% at 90 days, i.e., that it is generally safe to wait and optimize medical therapy.”

Despite their potential life-saving benefits, McMurray noted that “ICDs are expensive and ICD implantation is not without complications and very few patients ever have an appropriate discharge, i.e., ever use their device. Indeed, the DANISH trial has now suggested that ICDs may not reduce mortality in patients with non-ischemic cardiomyopathy receiving excellent pharmacological and device therapy. So I think there is now a lot of debate about the value of ICDs at all in the era of modern pharmacological and other device therapy (CRT). What we really need to do is to figure out who are the people who remain at high risk of sudden death in the current era and target ICDs to them.”

The study emphasizes “what can be achieved with optimal medical therapy,” commented Mathew Maurer (Columbia University). “The decision regarding ICD implantation for primary prevention in patients with HFrEF [heart failure with reduced ejection fraction] is complex and the data presented will assist patients, their family members, and clinicians in the process of shared decision-making to define who wishes to pursue such therapy in the current era.”

“There is a compelling need for better SCD risk discrimination than ejection fraction and/or disease etiology alone,” said Clyde Yancy, Northwestern University.  “Given the certain outcome of an untreated SCD event, any new screening algorithm needs to over-estimate risk and reduce the false negatives to nil. This is the challenge.”

This article has been republished from Larry’s blog CardioBrief as part of a licensing agreement between Everyday Health and the limbic.