Vintage ICD Trials in Heart Failure Upheld by Contemporary Registry Data

September 04, 2019

PARIS — The survival benefits of primary-prevention device therapy seen in seminal trials reported in the early 2000s still apply today, suggests a registry study from Sweden that also saw gross implantable cardioverter-defibrillator (ICD) underuse in its sample of eligible patients.

The analysis, based on propensity-matched pairs of such patients from the Swedish Heart Failure (HF) Registry who did and did not receive ICDs, showed a significant 27% drop in death from any cause and 29% decrease in cardiovascular (CV) mortality over 1 year. Declines in both endpoints over 5 years reached 12% for the ICD recipients.

The findings counter recent doubts regarding the survival advantage of primary-prevention ICDs, established from patients who entered trials like MADIT-2 and SCD-HeFT decades ago, on a background of contemporary medical and device management of heart failure with reduced ejection fraction (HFrEF), Benedikt Schrage, MD, lead author of the analysis, told | Medscape Cardiology.

Medical therapy in those older trials less often included the array of current guideline-recommended agents available now, and cardiac resynchronization therapy (CRT) wasn't yet an option.

In the total Swedish HF Registry cohort of 16,702 ICD-eligible predominantly male patients, most whom were seen since 2010, virtually everyone was on renin–angiotensin–system inhibitors (RASi), well over 90% were on beta blockers, and more than half were taking aldosterone inhibitors.

Indeed, the registry analysis shows excellent use of guideline-recommended therapies, in a major difference between HFrEF patients managed in recent years compared with 20 years ago, said Schrage, University Heart Center, Hamburg, Germany and Karolinska Institutet, Stockholm, Sweden.

Even so, "We still see this association between ICD use and reduced mortality."

That was true consistently in a subgroup analysis that showed such a benefit regardless of age or sex, year of entry into the registry, whether the HFrEF was ischemic or nonischemic, and whether or not the patient had a CRT device.

The findings therefore support guidelines on both sides of the Atlantic that recommend ICD therapy for HFrEF based on the older trials, even in patients on state-of-the-art medical therapy, said Schrage, who presented them here at ESC Congress 2019 in conjunction with their publication yesterday in the journal Circulation.

They indeed dispute "current arguments that ICDs seem less effective over time given advances in heart failure therapy," and that "they still do have a survival effect independent of ischemic or nonischemic etiology," agreed Sanjay Sharma, MD, St. George's, University of London, United Kingdom, not associated with the analysis, for | Medscape Cardiology.

Ischemic vs Nonischemic HFrEF

The finding in nonischemic patients is noteworthy, he said, even though it is consistent with guideline recommendations for primary prevention ICDs in symptomatic HFrEF patients without coronary disease.

That's because of recent data questioning the mortality-reducing effectiveness of ICDs in such nonischemic patients on contemporary medical therapy. The DANISH study "did an almost 180-degree flip" in making that finding, and in seeing preferential benefit from the devices in younger compared with older patients, Sharma said.

"Although the DANISH trial showed no survival benefit of primary prevention ICDs in the aggregate of patients with nonischemic dilated cardiomyopathy, many do not agree with the argument that such patients should no longer receive an ICD," states an accompanying editorial from Sana M. Al-Khatib, MD, MHS, Duke University Medical Center, Durham, North Carolina, and Fred M. Kusumoto, MD, Mayo Clinic, Jacksonville, Florida.

"Understanding the nuances of the DANISH trial is critically important to defining its implications for clinical practice," they write. For example, DANISH allowed CRT in patients with or without ICDs; and by enrolling only patients with elevated natriuretic peptides, "it likely selected patients who are more likely to die from advanced heart failure than sudden cardiac death," in whom an ICD would be less expected to improve survival.

Moreover, subgroup analyses can be statistically misleading or puzzling, as in the difference in survival outcomes by ICD therapy between ischemic and nonischemic patients in the current analysis, observed Gerhard Hindricks, MD, PhD, University of Leipzig, Germany, invited discussant for Schrage's presentation of the study.

For example, he pointed out, "the benefit of the ICD was more pronounced in the nonischemics," a 50% decrease in 5-year, all-cause mortality hazard ratio (95% confidence interval [CI], 0.28 - 0.91) compared with 22% in patients with coronary disease (95% CI, 0.63 - 0.98).

The analysis, Hindricks noted, also suggests "that the benefit of the ICD is limited to the first 12 months. There is at later stages a decrease in the effect on total mortality, which is difficult to explain. As other studies have clearly shown, if there is an effect it should be consistent over time."

Indeed, Hindricks said, "The study leaves me with more questions than answers." He proposed its propensity matching was hampered by "a selection bias that could not be cleaned up by the matching process."

ICD "Underuse," Selection Bias

Any bias might also be related to another significant finding of the analysis, he said, which suggests the ICD recipients were highly selected in some way. Only 1599 of the patients, less than 10%, had been implanted with an ICD despite their eligibility for one under the European guidelines.

That rate falls well under that of North America and other European countries of comparable affluence, possibly because they differ in HFrEF referral patterns. Most patients with HFrEF in Sweden are cared for by primary care physicians or other generalists, not cardiologists who may be more familiar with guidelines, Schrage observed.

The implantation rate also reflects ICD use at the study's baseline, and so doesn't reflect their use in patients who may have received one later, Schlage said.

Of the study's 16,702 patients fulfilling criteria for primary prevention ICDs, 9.6% received a device; they were followed for a median of 2.7 years.

The researchers created 1305 propensity-matched pairs of patients, each consisting of 1 ICD recipient and 1 who didn't get an ICD.


Table. Hazard Ratio (95% CI) for Clinical Outcomes, Propensity-Matched ICD-Eligible Cohorts of ICD Use vs No ICD Use


1-year HR (95% CI)

5-year HR (95% CI)

Death from any cause

0.73 (0.60 - 0.90)

0.88 (0.78 - 0.99)

Death from CV causes

0.71 (0.57 - 0.90)

0.88 (0.77 - 1.02)


In an analysis of the overall cohort adjusted for the same criteria that made up the propensity scores, the HR for CV mortality at 1 year was 0.81 (95% CI, 0.67 - 0.98) and at 5 years was 0.91 (95% CI, 0.81 - 1.02), favoring ICD recipients.

Among the subgroup who entered the registry from 2012 to 2016, when state-of-the-art heart-failure management was more as it is today, the HR for all-cause mortality was 0.69 (95% CI, 0.50 - 0.95).

"That management of HFrEF has improved drastically is irrefutable; however, the argument that this improvement in management has dampened the effect of ICDs on survival is not supported by strong data," state Al-Khatib and Kusumoto in their editorial.

Despite the limitations of the current observational study, they write, it "adds to the mounting evidence that ICDs are associated with improved survival in contemporary patients with HFrEF and highlights the need for strategies to improve their utilization."

The analysis was supported by Boston Scientific and the EU/EFPIA European Medicines Initiative. Schrage discloses receiving fees for speaking from AstraZeneca; the other authors have disclosed no relevant financial relationships. Al-Khatib, Kusumoto, and Sharma have disclosed no relevant financial relationships. Hendricks discloses receiving research contracts with Abbott, Biosense Webster, Biotronik, the Volkswagen Foundation, the German Innovation Fund, and the European Commission.

Circulation. Published online September 3, 2019. Abstract, Editorial

European Society of Cardiology (ESC) Congress 2019, World Congress of Cardiology


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.