Both trials "very strongly supportive"

Extracted from The Heart.org

“T-wave-alternans researcher Dr Paul S Chan (St Luke's Mid America Heart Institute, Kansas City, MO), also not connected with the MASTER or SCD-HeFT analyses, agreed that they "really raise questions as to whether the T-wave alternans test performs as well as we thought it did.

"We now have a lack of clarity about how best to use the test," Chan told heartwire. He points out that the overall SCD-HeFT trial randomized patients to receive or not receive ICD therapy, "but therapy was not randomized based on T-wave-alternans status. To date, we still do not have a randomized trial looking at T-wave alternans."

…

I walk away with a different interpretation. I think both trials are very strongly supportive of T-wave alternans in terms of demonstrating its potential in risk stratification," Dr David S Rosenbaum (Case Western Reserve University, Cleveland, OH) said to heartwire. "When you look at the SCD-HeFT substudy data, they strongly suggest that alternans is actually tracking the sudden-death phenotype quite nicely." Rosenbaum was principal investigator for the ABCD trial of mTWA.

It's exactly the result you would expect if T-wave alternans was indeed tracking susceptibility to sudden cardiac death.

In his editorial accompanying the SCD-HeFT substudy paper in print [3], Rosenbaum asserts that—because mTWA testing was done in a subset of 490 patients among the overall SCD-HeFT randomized population of 2521 patients, some of whom received ICDs and some of whom did not—it could be shown that the SCD risk prediction of mTWA is time dependent. Yet, he observes, the negative conclusions of Gold et al are based on arrhythmic risk during the overall follow-up period.

In SCD-HeFT, he contends, the separation point in the survival curves of patients with and without ICD therapy marks the emergence of the "SCD phenotype," because the only way ICDs can prevent death is by preventing SCD. The curves separated after 18 to 24 months, he observed, about the same time mTWA results became significantly predictive of arrhythmic events.

"It's exactly the result you would expect if T-wave alternans was indeed tracking susceptibility to sudden cardiac death." The SCD-HeFT substudy's authors missed the time-dependent effect because "they focused on the overall event rate for the entire follow-up period," according to Rosenbaum. "There's a clear relationship between the T-wave-alternans signal and the arrhythmia phenotype."

Chow said for heartwire that the MASTER results are "very robust, in that there was no difference between mTWA-negative and -nonnegative patients with respect to the primary end point of ventricular tachyarrhythmic events" (VTEs), as documented by the implanted devices. "And I think those data are very consistent with the SCD-HeFT trial."

But MASTER also found that the mTWA findings predicted all-cause mortality—intriguingly, because it has been long assumed that T-wave alternans reflects an abnormality of myocardial electrical substrate, suggesting that any prognostic value would be limited to arrhythmic events.

The alternans test is just like any other test. . . . I think we're expecting too much from it, sometimes.

In MASTER, "if you had a nonnegative alternans test, you were about twice as likely to die, although you were not necessarily more likely to die from an arrhythmia and you weren't necessarily likely to die even from a cardiovascular cause," according to Chow. "It's possible that an abnormal alternans test is associated with higher-risk clinical features. . . . The heart will often fail when the rest of the body fails, so it's not inconceivable that the alternans test may be a barometer of just generally sicker patients, but I think this needs to be explored."

Also pointing to the mortality finding in MASTER, Rosenbaum said the trial showed mTWA "was in fact a marker of a clinically relevant end point. So the issue here may not be so much the marker but the end points we're using in some of these trials. To say these are negative trials for alternans, I don't think so. When you look at them objectively and carefully, they suggest a very interesting signal that needs to be pursued."