HFSA: Was MADIT-CRT was underpowered for mortality?
Cleland, who is from the department of cardiology at the University of Hull in Kingston-upon-Hull, England, noted that MADIT-CRT is a substantial and well-constructed study, which achieved its primary endpoint, and is consistent in size and direction with similar landmark studies (COMPANION, REVERSE and CARE-HF). As a result of the findings of these studies, he recommended a change should be made to the guidelines, and drop the NYHA II criteria, at least when it comes to choosing interventions.
He reiterated that other multicenter, randomized trials, such MIRACLE, REVERSE, COMPANION and CARE-HF, came to similar conclusions. In REVERSE, the control arm was randomized against CRT or CRT-D. As a result, many of these patients received CRT only, especially in the European cohort, but achieved the same endpoints. In fact, Cleland said that REVERSE even had a larger reduction in primary outcomes of death, heart failure or hospitalization in the European cohort—with a 62 percent reduction, as well as a significant reduction in mortality.
In the COMPANION trial, the control arm was randomized against CRT and CRT-D. The primary outcome of death, heart failure or hospitalization, Cleland pointed out, was no difference between CRT and CRT-D.
Cleland was the principal investigator of CARE-HF, which he said is often misperceived as having a quite sick patient population, as only 21 percent of the patients had Class I-II heart failure, and had identical outcomes to those with Class IV heart failure. He said this highlights the prognostic benefits of CRT, and symptoms are not a useful guide.
At 18-24 months, the remodeling evidence of MADIT-CRT, CARE-HF and REVERSE are very consistent. “Of course, the consistency should give us the reassurance about the proper approach to this patient population,” Cleland said.
He mentioned that certain people have questioned whether larger events are exhibited in REVERSE and MADIT-CR, than in patients with sicker heart failure. However, Cleland pointed out that in CARE-HF, patients with a high NT-proBMP had a worse prognosis, and benefited from CRT, but the absolute reduction in mortality was similar and the relative reduction in mortality was greater in patients with less elevated NT-proBMP.
However, Cleland noted that there also are some caveats to MADIT-CRT, including QRS. In fact, he questioned: “Is it really QRS that matters?” Based on the trials, he said that the outcomes were different:
- COMPANION: Perhaps;
- CARE HF: No; and
- REVERSE: Perhaps, in terms of ventricular remodeling.
He also questioned whether QRS is “simply a surrogate for something else,” such as a better prognosis. “It may be that patients with narrow QRS have less severe dysfunction, and therefore, more difficulty showing a difference in outcome. Or, it’s also possible the patient may have more dilated cardiomyopathy, and therefore, a better prognosis.” Despite these possibilities and others, Cleland also speculated that there is not enough power to assess its interaction with mortality.
“It’s possible that the reason we are seeing less events in these patients with shorter QRS in MADIT-CRT is because it may be a surrogate for one of these other measures,” Cleland said, adding that they need to be cautious of surrogates.
Most importantly, Cleland questioned the high rates of all-cause mortality in MADIT-CRT, especially in ischemic patients. Therefore, he said that the primary endpoint of MADI-CRT is “deceptive,” adding that the study may have been underpowered for mortality. Therefore, he suggested that he and his colleagues should “hesitate before we put too much weight on the QRS story.”
Also, MADIT-CRT’s failure to show a reduction in mortality is inconsistent with the results of the previous randomized trials, CARE HF, COMPANION and REVERSE—all of which showed a reduction in mortality. As a result of this finding, he questioned whether there was a less than positive effect of CRT plus ICD.
However, to take a “positive” position of the trial, he acknowledged that there were beneficial effects of remodeling, which he said may be a surrogate for symptoms. He also said that it is possible that the morbidity benefits would reveal themselves as mortality benefits in longer follow up. Also, he said that if it is the best method, patients should be initially treated with CRT-D, because “we don’t want to go around upgrading our patients after they have deteriorated.”
To present an opposing view, “more negative” view of the trial, Cleland said remodeling is just a surrogate, and suggested until some of these questions are answered, it may not be harmful to wait for worsening heart failure to upgrade, especially because “newer technologies are making this simpler.”
He concluded that ICD is the “minority player” in the impact of CRT on outcome. In fact, he noted that 75 percent of the benefits of CRT-D come from CRT only. As a result, “there is no good evidence of benefit of adding ICD to CRT," while adding the question, "should we be planning to CRT only, with an occasional upgrade?”