IRIS Confirms: No Survival Gains From Early Post-MI Primary-Prevention ICDs

October 08, 2009

October 8, 2009 (Boston, Massachusetts) — Another randomized, controlled trial has shown no overall survival benefit from starting implantable cardioverter defibrillator (ICD) therapy in the early weeks after an acute MI in patients with features that put them at increased risk for events [1].

In the Immediate Risk Stratification Improves Survival (IRIS) trial, which randomized almost 900 patients with acute MI, overall mortality was unaffected by ICD therapy started within a month of the event on top of medical therapy, compared with medical therapy alone. Over a mean of about three years, the ICD group's hazard ratio for death compared with the control group was 1.04 (p=0.78).

The patients getting ICDs showed a 45% lower risk of sudden death (p=0.049), countered by an almost-doubled risk of non–sudden-cardiac death (p=0.001), reported IRIS investigators in the October 8, 2009 issue of the New England Journal of Medicine.

Although the trial had been preliminarily presented at the American College of Cardiology 2009 Scientific Sessions and reported then by heartwire , the data had been configured differently from how they appear in the published report. The investigators' conclusions are the much the same.

The IRIS population and results mirror those from the smaller Defibrillator in Acute Myocardial Infarction Trial (DINAMIT), published in 2004 [2] and the basis for current guideline proscriptions against early post-MI implantation of a primary-prevention ICD.

"The main results are remarkably similar," lead author Prof Gerhard Steinbeck (Ludwig-Maximilians University, Munich, Germany) told heartwire . "With implantation of an ICD early after MI, you are able to reduce sudden cardiac death, but at the expense of an increase of non-sudden death."

It's very encouraging that the results of IRIS were basically identical to DINAMIT; it's nice to see consistency.

But while IRIS confirms DINAMIT, he said, differences in the enrollment criteria and size--DINAMIT randomized only 674 patients--also broaden what has been learned about the effect of ICD therapy "in the first weeks and months after myocardial infarction, when mortality and also the risk of sudden cardiac death are highest." More patients in IRIS means greater statistical power, Steinbeck observed; also, its follow-up was slightly longer, and its patients weren't quite as sick. "And our patients were more frequently treated according to modern guidelines."

The mean LVEF in IRIS was 35%, compared with 28% in DINAMIT. There was an IRIS subgroup that had an LVEF >40%; they didn't show any survival benefit from ICDs either, according to Steinbeck.

More Alike Than Different

IRIS didn't follow DINAMIT per se: the two trials ran concurrently for several years, but the IRIS entry criteria that defined increased mortality risk pushed so many screened patients out of the running it was eight years before the trial reached target enrollment, the IRIS authors note.

Entry to IRIS required patients to have had an MI (with or without ST-segment elevation) within a month prior to enrollment and have at least one of two sets of risk factors: the first, a heart rate of >90 bpm within 48 hours of the infarction with an LVEF <40%; and the second, episodes of nonsustained ventricular tachycardia (NSVT) at Holter monitoring with a heart rate of >150 bpm after the subacute phase of the infarction (with no LVEF requirement).

Patients in DINAMIT had an MI within the last 40 days, an LVEF <35%, and either reduced heart-rate variability or a heart rate of >80 bpm.

"It's very encouraging that the results [of IRIS] were basically identical to DINAMIT; it's nice to see consistency," Dr Alfred E Buxton (Brown University, Providence, RI), who wasn't a coauthor with either trial, told heartwire . "But I don't think it's terribly surprising."

The populations of the two trials were more alike than different, according to Buxton. "Both groups had features that put them at an increased risk of death, but not necessarily of sudden death as opposed to non-sudden death. They just identify sicker patients."

According to Dr Alan M Garber and Dr Mark A Hlatky (Stanford University, Palo Alto, CA) in an accompanying editorial [3], IRIS helps alleviate any concerns that DINAMIT might have been a "false-negative" trial as a result of limited sample size or protocol issues.

"ICDs benefit selected patients, but the IRIS trial confirmed that they do not benefit all patients at high risk for sudden cardiac death. This result is worth knowing," they write. "It can direct efforts away from an expensive yet ineffective procedure toward either new or established alternatives. Confirmatory studies such as this one have a circumscribed but important role in comparative-effectiveness research."

IRIS Details

IRIS randomized 898 patients who met one or both sets of risk-stratification criteria, were within a month of their MI (mean 13 days), on optimal medical therapy, and didn't have sustained ventricular arrhythmias, NYHA class 4 heart failure, or severe comorbidities to receive (n=445) or not receive (n=453) an ICD.

No significant difference was seen between cumulative mortality curves over three years, which put the rates of death at one year at 10.6% in the ICD group and 12.5% for controls; at two years, 15.4% and 18.2%, respectively; and at three years, 22.4% and 22.9%, respectively.

IRIS: Hazard Ratios (95% CI) for Clinical End Points Among ICD Recipients vs Controls Over a Mean Follow-up of 37 Months

End point HR (95% CI) p
All-cause mortality* 1.04 (0.81–1.35) 0.78
Sudden cardiac death 0.55 (0.31–1.00) 0.049
Non–sudden cardiac death 1.92 (1.29–2.84) 0.001
*Primary end point

Nor were there differences in all-cause mortality between ICD recipients and controls by risk-stratification criteria for study entry. A reduced rate of sudden cardiac death in the ICD group was offset by a higher rate of non–sudden cardiac death, regardless of which enrollment criteria for increased risk patients had shown.

With the current IRIS report limited to the study's primary and secondary end points, Steinbeck said, there is much more to be learned from its patients, including insights about the mechanisms of non-cardiac death and whether ICDs themselves may have increased their mortality risk.

After having performed about 6000 interrogations of ICDs from patients in the trial, he said, the group will be looking at whether amount of pacing, appropriate and inappropriate shocks, antitachycardia pacing in lieu of shocks, defibrillation-threshold testing, and generator replacements may be contributing to an increased non–sudden-cardiac-death risk with early post-MI device therapy.

IRIS: Secondary Mortality End Points, ICD vs No ICD, By Risk-Stratification Criteria for Study Entry, Mean Follow-up 37 Months

Mortality end points ICD group No ICD p
Sudden cardiac death (%)      
Criteria with LVEF requirement a 6.7 14.9 0.003
Criteria without LVEF requirement b 3.0 6.4 0.25
Both sets of criteria met 8.5 19.5 0.08
Non–sudden cardiac death (%)      
Criteria with LVEF requirement a 14.7 8.9 0.02
Criteria without LVEF requirement b 13.1 5.5 0.06
Both sets of criteria met 23.4 14.6 0.39
a. Heart rate >90 bpm with LVEF <40%

b. Heart rate >150 bpm with nonsustained VT

IRIS was sponsored by Medtronic and AstraZeneca. Steinbeck reports receiving consulting fees from Medtronic, lecture fees from Medtronic, St Jude Medical, and Boston Scientific and grant support from Medtronic and St Jude Medical; disclosures for the other authors are in the report. Garber reports serving on the Committee on Comparative Effectiveness Research Prioritization of the Institute of Medicine. Hlatky reports receiving consulting fees from GE Healthcare. Buxton reports receiving honoraria for speaking from Medtronic.

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....