SCD-HeFT: Experts look for closure on ICD reimbursement debate after MADIT-II

March 02, 2004

New Orleans, LA - With the Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT) primary clinical results scheduled for release on March 8, 2004 at the American College of Cardiology scientific sessions here, arrhythmia specialists are hoping to nail down reimbursable indications for one of their most potentially effective therapies in patients with ischemic heart disease.

"The SCD-HeFT trial is the largest of the trials that have involved [implantable] defibrillator therapy, with a longer patient follow-up than any of the previous studies," Dr Kerry L Lee (Duke University, Durham, NC), the principal investigator, told heartwire . The trial has compared all-cause mortality in >2500 patients with NYHA class 2-3 HF and an LVEF of <35% who were randomized to receive an implantable cardioverter-defibrillator (ICD), amiodarone, or placebo on top of standard medical therapy. About half the patients had CAD and the remainder had nonischemic cardiomyopathy, according to Lee. The trial was conducted at more than 100 centers in North America.

The question is, will it beat amiodarone in the nonischemic group?

Prospective economic and quality-of-life analyses of SCD-HeFT data will be reported at a future meeting, Lee said.

Amiodarone is unlikely to be better than placebo at prolonging survival in patients with CAD and left ventricular dysfunction (LVD), according to the literature, Dr Eric N Prystowsky (St Vincent Hospital, Indianapolis, IN) said to heartwire . "The ICD has a very good chance of beating the control group in both ischemic and nonischemic cardiomyopathy. The question is, will it beat amiodarone in the nonischemic group?" said Prystowsky, who is on the SCD-HeFT executive committee but unaware of the final results.

SCD-HeFT in context: The controversy

Because of its large CAD population, SCD-HeFT is seen partly as an attempt to replicate the controversial results of the second Multicenter Automatic Defibrillator Implantation Trial (MADIT-II)[2]. In that trial, the ICD was associated with a significant SCD relative-risk reduction of 31% over 20 months in post-MI patients with an LVEF<30%, as compared with best medical therapy. No amiodarone group was included; that drug's value in CAD patients with LVD had been questioned in several major trials[2,3,4].

For some clinicians, MADIT-II fell short of showing that ICDs should be used routinely in all patients who meet its entry criteria[5,6]. It was, after all, only one study, and it didn't attempt to narrow the population of ICD recipients with some form of risk stratification, such as arrhythmia inducibility or ECG features. Others took the MADIT-II results as solid enough to take right into clinical practice.

"It's hard to argue with the [MADIT-II] data. They are clean and were appropriately analyzed," said Dr Douglas Zipes (Indiana University School of Medicine, Indianapolis) when interviewed by heartwire .

Of the patients in SCD-HeFT, only about one fifth meet the MADIT-II entry criteria, according to Lee. But there is substantially greater overlap among the SCD-HeFT patients with CAD. Although many MADIT-II patients were asymptomatic, about 63% were in NYHA class 2 or higher at randomization.

"Although you never know, my expectation is that SCD-HeFT will confirm MADIT-II. It was a very good trial, and I can't imagine why SCD-HeFT would be different," Dr Albert Waldo (Case Western Reserve, Cleveland, OH) told heartwire.

My expectation is that SCD-HeFT will confirm MADIT-II.



The US FDA acknowledged MADIT-II by approving an expanded ICD indication consistent with the trial's entry criteria. Suddenly, the ICD was eligible for primary SCD prevention in an expanded population of patients with CAD and LVD without the previously included requirement of VT inducibility.

The ACC, American Heart Association, and the North American Society for Pacing and Electrophysiology (NASPE) also responded by giving qualified approval to ICD use in patients meeting MADIT-II criteria[7]. Their class-2a recommendation for using the device in such cases recognized that, despite "conflicting evidence and/or a divergence of opinion, the weight of evidence/opinion" was in favor of it.

With total initial ICD implantation costs reaching tens of thousands of dollars per patient, these developments rattled third-party payers.

SCD-HeFT won't increase ICD use?

"The SCD-HeFT data are unlikely to have any significant influence on the growth rate of the ICD market," according to a February 19, 2004 release to the trade press, posted on the web by Newratings, which cited an industry observer's report[8]. According to the release, SG Cowen analyst Matt Dodds claimed that SCD-HeFT is "estimated to show" an absolute mortality reduction with the ICD to a degree that would make it "difficult to rationalize referrals and reimbursement." The release did not elaborate, nor did it attribute his information sources. Dodds did not respond to heartwire 's attempts to reach him by telephone.

Lee told heartwire the results of the trial are a guarded secret and he is unaware of any information leaks.

The Centers for Medicare & Medicaid Services (CMS) updated its ICD policies in June 2003, to account for the MADIT-II findings. It already permitted federal reimbursement for ICD implantation in CAD patients with an LVEF<35% who had EP-inducible sustained VT or VFcriteria validated by the first MADIT trial[9] and the Multicenter Unsustained Tachycardia Trial (MUSTT)[10]. The agency exacerbated the indications controversy, according to some clinicians, by adding the reimbursement-eligibility requirement that MADIT-II-like patients also have a wide QRS complex (>120 ms). That ECG screening variable emerged in a MADIT-II post-hoc analysis as marginally associated with increased risk but is not present in many patients who fit the MADIT-II entry criteria.

I think they are wrong for doing that from a scientific standpoint . . . and I think they are wrong for doing that from an ethical standpoint.

"It is very inappropriate that the CMS is playing doctor and making decisions that should be left to physicians," Zipes said. "They used . . . data from a secondary analysis that is of questionable statistical significance to further narrow the individuals who would qualify for an ICD. I think they are wrong for doing that from a scientific standpoint . . . and I think they are wrong for doing that from an ethical standpoint."

The CMS created a harsh dilemma for clinicians. Hospital administrators didn't want to allow unreimbursed ICD implantations, even though many candidates fit society guidelines for receiving the devices, Zipes observed. That left the choice of either jumping through the Medicare hoop by performing an unnecessary invasive EP study to show VT/VF inducibility, he said, or being vulnerable to charges that contemporary treatment guidelines were not followed. "Both of those situations are totally untenable."

Zipes told heartwire he nonetheless believes most specialists currently use MADIT-II criteria to select patients for ICDs. But most EP groups are not trying to apply them to their patients from the pre-MADIT-II era, he said. If the SCD-HeFT data from ischemic patients confirm MADIT-II, "it might galvanize us into doing that: scouring our clinics for patients who would fit those criteria [but] have not been so treated."

It's all about the money

If routine ICD use were not so potentially costly, many believe, the CMS wouldn't balk at setting unqualified MADIT-II criteria as its requirement for reimbursement eligibility.

Death is always cheaper. I'm not saying that as a cynic. I always choose life.

Observed Waldo, "People are afraid of the expense of the ICD, because there are an awful lot of patients out there who are MADIT-II-like patients. . . . Death is always cheaper. I'm not saying that as a cynic. I always choose life." Society, he said, must decide whether or not it wants to support the use of medical devices that save lives.

After MADIT-II, any ICD survival benefit that may emerge in SCD-HeFT's CAD patients wouldn't have to be dramatic to get the CMS to drop the QRS reimbursement requirement, Prystowsky said. "If [SCD-HeFT shows] at least a 1% survival advantage, I think they will seriously consider it, especially because it would be the second trial to show it."

Implications for nonischemic cardiomyopathy

Whether ICDs should be routinely used for primary prevention in patients with nonischemic cardiomyopathy is also unsettled. The Defibrillators in Nonischemic Cardiomyopathy Treatment Evaluation (DEFINITE), reported at the November 2003 AHA meeting, provided some clues but no firm answers.

DEFINITE followed 458 patients with nonischemic cardiomyopathy who were maintained on standard medical therapy and randomized to receive or not receive an ICD. For entry, the patients had been required to have an LVEF <35% and a history of spontaneous premature ventricular contractions (PVCs) or nonsustained ventricular tachycardia (NSVT). The ICD was associated with a 34% reduction in the primary end point of all-cause mortality, which did not reach statistical significance. However, a 74% relative-risk reduction in arrhythmic mortality, a secondary end point, was significant.

"SCD-HeFT has more nonischemic cardiomyopathy patients than did the DEFINITE trial," said Lee. "I think SCD-HeFT will enrich . . . and greatly add to the information that came forth from DEFINITE."

Despite a negative finding for the primary end point in DEFINITE, according to Prystowsky, the ICD remains highly promising in such a population. "If they had had enough patients in that study," he said, "the defibrillator would have won."

In SCD-HeFT, however, "it wouldn't surprise me if there were some slight survival advantage with amiodarone. [Still,] in my opinion, the defibrillator won't lose to anything. It never has, and it's almost impossible to think that it would."

Description of the SCD-HeFT patients and methodology

Population, characteristics

>2500 patients with NYHA class 2-3 HF for at least 3 months; LVEF <35% documented in the prior 6 months; 52% with CAD and 48% with nonischemic cardiomyopathy; all on optimal drug therapy, including ACE inhibitors, for at least one month

Major exclusion criteria

Documented sustained VT/VF, recent history of unexplained syncope

Treatment groups

Conventional therapy plus: single-lead ICD vs amiodarone (double-blinded) vs placebo

Mean follow-up time

41 months (all randomized patients)

End points

Primary: all-cause mortality over 2.5 years
Secondary: arrhythmic mortality, nonarrhythmic mortality, morbidity (all-cause mortality or HF rehospitalization), health-related quality-of-life, cost of care
The ICD's long journey

The ICD had been available for more than a decade in 1991 when a randomized trial shook things up by showing that a number of commonly prescribed antiarrhythmic drugs might do more harm than good. Before the Cardiac Arrhythmia Suppression Trial (CAST)[11]. ICDs had generally been limited to survivors of cardiac arrest and other very-high-risk groups. After the CAST trial, the highly invasive ICD started to look more attractive compared with most antiarrhythmic drug therapies.

Over the next dozen years, a series of randomized primary- and secondary-prevention trials took the ICD from niche status to a far-reaching and controversial FDA approval indication based on the MADIT-II entry criteria.

Selected published ICD trials in patients with ischemic heart disease and LVD/HF


Population, features

Randomization groups

Follow-up time



n=196, prior MI, NYHA 1-3, LVEF <35%, on standard therapy, documented asymptomatic unsustained VT, inducible nonsuppressible VT ICD, n=95; no ICD, n=101 Mean 27 months All-cause mortality: ICD 15.8%; no ICD 38.6%; 54% RR reduction with ICD, p=0.009


n=704, prior MI, LVEF <40%, on standard therapy, spontaneous asymptomatic unsustained VT, inducible nonsuppressible VT EP-guided antiarrhythmic therapy including amiodarone, with or without ICD, n=351; no antiarrhythmic therapy, n=353 Median 39 months Cardiac arrest or arrhythmic death: EP-guided therapy 25%; no EP-guided therapy 32%, p<0.001
Same end point solely among patients randomized to EP-guided therapy: ICD 9%; no ICD 37%; 76% RR reduction with ICD, p<0.001


n=900, scheduled for CABG, LVEF <36%, abnormal signal-averaged ECG ICD, n=446; no ICD, n=454 Mean 32 months All-cause mortality: ICD 27%; no ICD 24%, p=0.64
 MADIT-II[13] n=1232, prior MI, LVEF <30%, on standard therapy 3:2 randomization: ICD, n=742; no ICD, n=490 Mean 20 months All-cause mortality: ICD 14.2%; no ICD 19.8%, 31% RR reduction with ICD, p=0.016


Selected ICD trials in survivors of SCD/malignant VT or VF


Population, features

Treatment groups

Follow-up time



n=1016, resuscitated VF or cardioversion of symptomatic sustained VT ICD, n=507; amiodarone or other class-3 antiarrhythmic, n=509 Mean 18.2 months All-cause mortality: ICD 15.8%; drug therapy 24.0%, p=0.02


n=659, resuscitated VT/VF or unobserved syncope ICD, n=328; amiodarone, n=331 Mean 3 years All-cause mortality: ICD 8.3%/year; amiodarone 10.2%/year, p=0.142
Arrhythmic mortality: ICD 3.0%/year; amiodarone 4.5%/year, p=0.094


n=288, resuscitated cardiac arrest, history of documented sustained VT ICD, n=99; amiodarone, n=92; metoprolol, n=97 Mean 57 months All-cause mortality: ICD 36.4%; amiodarone and metoprolol groups combined 44.4%, p=0.081


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.