Shaky FUTURE? FFR Takes a Hit With Mortality Signal in French Study

NEW ORLEANS, LA—A European study testing the utility of functional testing prior to coronary revascularization is puzzling proponents of fractional flow reserve (FFR), with the study halted early due to an increased risk of death at 1 year among patients randomized to a treatment strategy guided by functional assessment.

In the FUTURE study, 17 patients in the FFR-guided strategy died compared with seven patients in the angiography-guided control arm. At the time the trial was stopped by the data and safety monitoring board (DSMB), the difference in mortality among 836 patients included in the analysis was significantly higher for those randomized to FFR (2% vs 4%; P = 0.02).

Presenting the results at the American Heart Association Scientific Sessions 2016, Gilles Rioufol, MD (Hospices Civils de Lyon, France), said the study leaves more questions than answers, mainly because its early end precludes firm conclusions on the primary endpoint of all-cause mortality, MI, repeat revascularization, and stroke. Moreover, it’s possible the mortality signal is a spurious finding, he said.  

Speaking with the media during a late-breaking clinical trials press conference, Rioufol presented an analysis of clinical events for the 797 patients who reached the 1-year endpoint in the trial. Within this group, death from any cause occurred in 1.8% of the 398 patients in the control arm compared with 3.9% of the 399 patients given the FFR-guided strategy, a difference that was no longer statistically significant (HR 1.98; 95% CI 0.85-4.60).

Given the surprising mortality signal, experts grappled with the results, but for the most part they do not appear to think this signal is a legitimate or concerning finding. To TCTMD, Justin Davies, MD (Imperial College, London, England), who was not involved in the study, said he was disappointed the trial was stopped early by the DSMB, saying he believes the results are likely a play of chance.

“Obviously, with all these studies, like FAME-2, it’s a shame it was stopped prematurely,” said Davies. “I think looking at the reasons for being stopped early—the increased death rate between the two arms [in FUTURE]—the numbers are small, so you wonder if it’s just a statistical chance. As time plays out, we may well find out the two [event] curves start to come back together.”

C. Michael Gibson, MD (Beth Israel Deaconess Medical Center, Boston, MA), said FFR is an excellent tool to assess lesion function but that it does not assess the stability. For example, an acute coronary syndrome with plaque rupture might “not be all that tight but yet may cause problems down the road because it’s a very active lesion.” In FUTURE, approximately 46% of patients in both the FFR- and angiography-guided treatment arms had ACS, with one in five having STEMI.

“It might be an acute coronary syndrome issue, but my suspicion is that if the trial had run its course and had the full statistical bandwidth, there would be very little difference between the two groups at this stage,” said Davies.

Gazing Into FUTURE

FUTURE was a multicenter, randomized study conducted at 31 centers in France with a planned enrollment of 1,721 patients with multivessel disease. The hypothesis, said Rioufol, was that FFR would help guide treatment—either with PCI, CABG surgery, or optimal medical therapy—and improve clinical prognosis compared with traditional angiography-guided treatment.

Speaking during the press conference, Rioufol said registry data have shown that FFR modifies the PCI revascularization strategy in approximately 40% of patients, with the percentage even higher in patients without prior noninvasive testing. In addition, the SYNTAX score, as assessment of coronary artery disease complexity, is frequently downgraded in patients with multivessel disease when functional testing with FFR is utilized.

In FUTURE, the therapeutic management of patients was altered with FFR, as would be expected based on the study hypothesis. In the control and FFR arms, there was no difference in the number of patients who underwent CABG surgery, but the number of patients treated with optimal medical therapy alone was significantly higher among those who underwent FFR, while the use of PCI was significantly less. In the FFR group, 17% received medical therapy, 12% received surgery, and 71% underwent PCI. In the angiography-guided arm, 9% received medical therapy, 12% underwent surgery, and 78% underwent PCI.

For Herbert Aronow, MD (Cardiovascular Institute of Rhode Island Hospital, Miriam Hospital, and Newport Hospital, Providence, RI), the results are surprising, and like Davies, he chalks them up to chance. “If there really was a mortality excess due to FFR, it would have to happen because of the FFR procedure itself or the way in which the information we get from FFR changed decisions,” he said. “We know there were very few complications from the FFR procedure and even if they had all been fatal, which they were not, it wouldn’t have accounted for the mortality excess.”

Aronow does not believe the increased mortality risk was the result of differential treatment resulting from the FFR results. In the FFR-guided arm, use of medical therapy was greater and use of PCI less, so the argument one could make would be that more PCI would have led to a lower mortality rate. However, in stable patients with multivessel disease, “there is no evidence to demonstrate that mortality is reduced with PCI,” he said.

Most importantly, 13 of the 17 patients who died in the FFR arm underwent PCI. “We know it wasn’t the change in treatment decision that led to this excess,” said Aronow.

David Taggart, MD (University of Oxford, England), pointed out that when FFR shifts the SYNTAX score downward, it’s possible patients might be deprived of an optimal revascularization strategy with CABG. Aronow agreed with the rationale, but said the FUTURE data do not bear that out, noting that the rate of CABG was similar in both the FFR- and angiography-guided treatment strategies.

Despite concerns over the trial being stopped early, Aronow, Davies, Taggart, and Rioufol all acknowledged the DSMB is in a bind when it comes to stopping trials prematurely, particularly given the mortality signal. As Taggart pointed out, many trials are confounded by short-term follow-up. “SYNTAX, NOBLE, or EXCEL, if we’d taken the primary outcome measure at 1, 2, or 3 years, we’d have had it totally wrong answer as to what the outcome was,” he said.