COAPT at 3 Years: Crossover Patients See Benefits From MitraClip

SAN FRANCISCO, CA (UPDATED)—New, 3-year follow-up data from the COAPT trial provide a solid signal that even after a 2-year delay, patients initially randomized to guideline-directed medical therapy (GDMT) derive benefit by undergoing mitral valve repair using the MitraClip (Abbott).

Such crossovers were permitted after the 2-year mark in COAPT, explained Michael Mack, MD (Baylor Scott and White Heart Hospital, Plano, TX), who presented the 3-year data during a late-breaking clinical trial session here at TCT 2019.

As Mack showed here, the Kaplan-Meier curves for the primary endpoint of death/heart failure hospitalization for the intention-to-treat groups continued to separate at 3 years, reaching 88.1% for the GDMT group and 58.8% in the MitraClip group (HR 0.48; 95% CI 0.39-0.59), yielding a number needed to treat of 4.5.

More striking, perhaps, was an analysis that zoomed in on the total of 58 patients who crossed over to MitraClip either before 24 months (five patients) or after 24 months (53 patients). Among these patients, the 12-month rate of the primary endpoint was just 28%, nearly matching the rate seen among the group originally treated with the MitraClip at 1 year.

“GDMT only patients who crossed over and received a MitraClip experienced fewer HF hospitalizations and deaths or heart failure hospitalizations within 12 months than those who did not cross over, with rates comparable to patients originally assigned to the MitraClip,” Mack said here.

Also telling in these longer follow-up, however, is just how sick patients were in this study: even among the patients initially randomized to the MitraClip, almost half had died by the end of 36 months.

Diverse Reactions

“These are incredible results,” said Marie-Claude Morice, MD (Cardiovascular European Research Center, Massy, France), the discussant at an early morning press conference, raising hopes for patients who “have a very bad prognosis,” despite having their medications optimized within the trial. “So yes, the technology works and save lives and that’s very clear,” she stressed. “That's an extremely important result.”

The technology works and save lives and that’s very clear. Marie-Claude Morice

The fact that the crossover patients seem to derive a benefit so markedly similar to patients initially treated with the clip supports the hypothesis that “biologically, this actually does make sense,” said Dharam Kumbhani, MD (UT Southwestern Medical Center, Dallas, TX), who also spoke at the press conference. A key question, he continued, is whether the patients who crossed over were in some way fundamentally different from those who did not. In all, only 38% of surviving patients initially randomized to GDMT crossed over to receive the MitraClip.

Mack, in response, said he and his co-investigators have not yet delved into these analyses but that the bottom line is patients couldn’t cross over if they were dead. As such, “one [explanation] is that there is some survivorship bias, and what this also means is that, well, if these patients benefited at 2 years, why not just wait to do [the clip procedure]?” The answer, he continued, is that “there is a mortality cost of waiting, because a significant number of patients did die prior to the 2 years, and there is some selection of patients who did well enough to survive 2 years and were still in good enough shape to receive the MitraClip.” 

But survival bias cannot be ruled out, Mack stressed in his main arena presentation. “A significant number of patients died before they had the opportunity to cross over so there, in all likelihood, is some selection bias both in terms of survivorship and the decision on the local heart team as to having these patients crossed over.”

So while the groups look “comparable,” he continued, “this is a select group of patients and needs to be interpreted cautiously with that in mind.”

Co-principal investigator Gregg Stone, MD (Icahn School of Medicine at Mount Sinai, New York, NY), also weighed in saying, “we really don’t know” why eligible patients, still alive, did not end up getting therapy. Everyone was offered MitraClip at no cost, he noted. Stone said it is likely “a combination” of reasons: “Some were too sick, but others were doing quite well and so they didn't feel there was a reason to cross over.” Others may simply have been unaware that this was an opportunity to them, having participated in the trial.

Sanjay Kaul, MD (Cedars-Sinai Medical Center, Los Angeles, CA), who commented on the analysis for TCTMD, said he wasn’t quite sure what to make of the conclusions, adding that he, too, would like more information on what distinguished the crossovers from the patients who didn’t end up getting a MitraClip.

“I am not sure what clinically meaningful information can be garnered from this analysis,” he told TCTMD. “It is essentially an observational data set, and we don’t know . . . whether there are differences among those who crossed over versus those who did not.”

Kaul continued: “It is reassuring to see that patients do benefit even if they are offered the device after 2 years of optimized GDMT.” But as additional slides in Mack’s presentation made clear, the driver of the benefit was reduced hospitalizations, not fewer deaths.

Martin Leon, MD (NewYork-Presbyterian/Coumbia University Irving Medical Center, New York, NY), and Nir Uriel, MD (University of Chicago Medical Center, IL), both discussants in today’s session, also pointed to the mortality rates at 3 years. “Even at 3 years,” said Leon, “you have a 42.8% mortality and still an annual heart failure hospitalization rate of 35.5% in the MitraClip patients.”

Uriel, picking up on this thread, pointed out that one prevailing explanation is that these patients were undergoing MitraClip implantation “too late,” although he acknowledged this theory is muddied somewhat by the fact that the patients in whom treatment was delayed by 2 years seemed to fare roughly as well as the patients treated with MitraClip at the outset.

“I think this calls for a very aggressive approach in mitral regurgitation patients,” he said. “For heart failure patients with mitral regurgitation, within 3 years, 50% will die, even if we treat the mitral regurgitation,” Uriel said. “So I think we need to address it before it is moderate-to-severe, probably, and we need to see whether or not addressing it earlier can prolong life and maybe [lead to] less hospitalizations and improved quality of life.”

Of note, the French MITRA-FR trial, which did not show a benefit of the MitraClip versus optimal medical therapy and so prompted 12 months of explanations and debate, is not permitting crossovers in that trial until the final follow-up at 5 years. COAPT investigators, asked by TCTMD if they whether this sounded ethical, given the current findings from COAPT, stressed—as they have in the past—that the trials enrolled different types of patients.

“The MITRA-FR patients were a very different patient population and did not show any collective benefit: not immediately, not at 1 year, not at 2 years. I'm sure, in that population, there are some who might be COAPT-eligible, said Stone. “And if there were some patients, probably the 10 or 20% who might meet the COAPT eligibility criteria, I would strongly recommend that they get treated, but I suspect it’s a minority of patients.”

Investigators for both trials have agreed to pool their two datasets to try to get a clearer sense of the differences between the two studies.