Primary Prevention ICDs Still Render Better Survival, Even in Modern HFrEF Patients

PARIS, France—Use of implantable cardioverter-defibrillators (ICDs) for the primary prevention of sudden cardiac death in patients with heart failure with reduced ejection fraction (HFrEF) is associated with a lower mortality risk, a new propensity-matched analysis shows, easing some of the concern that the devices would not have as much benefit in contemporary practice.

ICD use was associated with an absolute 4.2% lower risk of all-cause mortality at 1 year and an absolute 2.1% lower risk at 5 years, Benedikt Schrage, MD (Karolinska Institutet, Stockholm, Sweden), reported recently at the European Society of Cardiology Congress 2019. Similar results were observed for cardiovascular mortality, although the difference did not reach statistical significance at 5 years.

The findings, published simultaneously online in Circulation, “support the current guideline recommendations and call for better implementation of ICD use in clinical practice,” Schrage concluded during his presentation.

Addressing Doubts

International guidelines advise use of ICD therapy for the primary prevention of sudden cardiac death in select HFrEF patients, but those recommendations are based on randomized trials that were initiated more than two decades ago, Schrage noted. Thus, he said, the results may not be applicable to contemporary patients, who are treated with improved heart failure therapies on the background of declining rates of sudden death.

Moreover, recent studies, particularly the DANISH study in patients with nonischemic systolic heart failure, have raised some doubts about the utility of ICD therapy for primary prevention.

To explore the issue, Schrage and his colleagues examined data from the ongoing Swedish Heart Failure Registry (SwedeHF). The analysis included 16,702 patients (mean age 73 years; 28% women) enrolled between 2000 and 2016 who met guideline-recommended criteria for ICD implantation for primary prevention—ie, they had ejection fraction less than 40%, heart failure duration of at least 3 months, and NYHA class II or higher symptoms. Schrage explained that the EF cutoff of 40% was used instead of the guideline-recommended threshold of 35% because of the way that information is collected in the registry. There were high rates of use of guideline-recommended heart failure therapies in the registry.

Overall, 10% of patients were treated with an ICD. After propensity-score matching, there were 1,305 pairs of patients who either did or did not receive an ICD.

In this comparison, patients who received the device had lower rates of all-cause mortality at 1 year (12.7% vs 16.9%; HR 0.73; 95% CI 0.60-0.90) and 5 years (47.4% vs 49.5%; HR 0.88; 95% CI 0.78-0.99). The findings were consistent across various subgroups.

ICD-treated patients also had lower risks of cardiovascular mortality at 1 year (10.1% vs 13.9%; HR 0.71; 95% CI 0.57-0.90) and 5 years (36.6% vs 39.5%; HR 0.88; 95% CI 0.77-1.02), with the latter difference falling shy of statistical significance.

The investigators tested for the presence of potential residual confounding with a negative control analysis of noncardiovascular hospitalization, which should not be affected by primary prevention ICD use. There were no differences between groups at 1 year or 5 years, supporting the reliability of the main findings, Schrage said.

Guidelines Bolstered

Commenting for TCTMD, E. Kevin Heist, MD, PhD (Massachusetts General Hospital, Boston, MA), said there has been some uncertainty about whether there would still be a benefit from primary prevention ICD use in the current era. Despite the fact that this is an observational study and cannot establish causal relationships, it still bolsters existing recommendations.

“It’s not a randomized study, and I think that if the randomized studies had not been done, this study by itself would not support ICD indications,” Heist said. “But inasmuch as these observational data support what the randomized studies have largely shown, I think they support continuing with the guideline [based on] the older randomized studies.”

Of note, the DANISH study published 3 years ago did not show a significant benefit of ICD use for primary prevention in nonischemic patients. The current study, however, showed a consistent—and possibly even a larger—positive relationship between ICD use and lower mortality in the nonischemic subgroup.

“I think this observational study supports where the guidelines have left things, which is that there is still a nonischemic indication for defibrillators,” Heist said.

Another interesting aspect of the study, he pointed out, was that the apparent mortality benefit with ICD use was evident within the first year, after which survival curves remained roughly parallel. That’s the opposite of what was seen in the MADIT II trial, in which there was no discernable difference in mortality at 1 year but a growing benefit of ICD use over time.

That could just be a statistical anomaly from one study to another, “but I think it at least raises the possibility that there may be some early benefit to defibrillators, that you don’t have to wait until patients are beyond a year to expect any detectable mortality benefit,” Heist said.

Some Limitations

Schrage acknowledged that the latest study had some limitations, including the potential impact of residual and unmeasured confounders, the assessment of ICD use at baseline only, the lack of data on sudden cardiac death or antiarrhythmic drug use, and small sample size in the matched cohort, which hampered the ability to detect differences between subgroups.

Gerhard Hindricks, MD (Heart Center Leipzig, Germany), who served as a discussant following Schrage’s presentation, highlighted some of the same shortcomings, but said the study was a “solid analysis of registry data” and was “well performed with appropriate statistical methods.”

Even so, “the study leaves me with more questions than answers,” Hindricks said.

In particular, the narrowing of the mortality curves over time between the ICD and no-ICD groups “is difficult to explain to me as other studies have clearly shown if there is an effect it should be consistent over time,” he said. Also, the fact that the results differ from the DANISH study by showing an even larger benefit in nonischemic patients “leaves me with an ‘I don’t understand that’ feeling.”

To get a more conclusive answer to whether primary prevention ICD therapy is still useful in the modern era, a randomized trial is needed, Hindricks said, pointing out that the European PROFID project, which is reevaluating sudden cardiac death after MI, will culminate in a 4,000-patient clinical trial.