October 12, 2011

October 11, 2011 (Lisbon, Portugal) — Four-year data from the SYNTAX trial show for the first time a divergence in death rates between the patients who were treated with bypass surgery and those who got PCI with a Taxus stent. All-cause mortality and cardiac death were both significantly higher in the PCI group compared with the surgery arm, as was MI, and the excess rate of strokes initially observed in the CABG arm has now leveled out, reported Dr Patrick Serruys (Erasmus University Medical Center, Rotterdam, the Netherlands) at the European Association of Cardio-Thoracic Surgery 2011 Annual Meeting last week.

But the surrogate safety end point of death/stroke/MI is not significantly different between the two arms of the study, and subgroup analysis by SYNTAX score shows that the key message remains unchanged from that of a year ago, said Serruys.

"CABG remains the standard of care for patients with complex disease and an intermediate or high SYNTAX score. However, PCI may be an acceptable alternative revascularization method to CABG when treating patients with less complex diseases (SYNTAX score <22), including left main." To put it another way, 75% of patients with left main or three-vessel disease are still best treated with CABG, but for the remaining 25%, "PCI is an alternative to surgery, at least out to four years," he commented.

Who Won SYNTAX? And Does Anyone Use the Score?

Discussion followed about who had "won" the SYNTAX trial. "The surgeon would say the longer you go in follow-up, the better surgery looks. We knew this was going to look good four to five years out," said chair of the session, Dr Michael Mack (Medical City Dallas Hospital, TX), a surgeon. "But as an interventionalist, you might say, 'We don't use Taxus stents any more, we use the everolimus-eluting Xience, and the results wouldn't be the same.'"

The surgeon would say the longer you go in follow-up, the better surgery looks.

In fact, observed another surgeon: "We have to congratulate you. The surgeons did not realize we had lost after SYNTAX." Serruys, for his part, merely remarked: "There is no winner or loser. There is just knowing more about what we do. One trial generates another trial, and it has been like that for the past 60 years, and it will go on."

Another attendee wondered how often the SYNTAX score is actually used in practice. Serruys said it varies: "I see practitioners who say, 'I don't have time to do it.' " He believes "the future is an automated version, which will be able to provide us with a perfect SYNTAX score, including the calcium and the fractional flow reserve [FFR]," both obtained from using noninvasive fractional flow reserve computed tomography. Mack agreed: "I would encourage everyone to check out heartflow.com; this technology is absolutely fascinating."

Slow Catch-up for Surgery in Terms of Stroke; No Difference Now

SYNTAX was an 1800-patient trial randomizing patients with left main coronary disease and/or three-vessel disease to either CABG or PCI using the Taxus drug-eluting stent (DES). Serruys reminded people that the study was designed for "all comers" and that the practice in Europe around the time SYNTAX was started, in 2004, was for two-thirds of these complex left main stem/three-vessel disease patients to go to surgery and for the other third to go to PCI.

At one year, PCI failed to meet the prespecified margin of noninferiority against CABG, after the primary end point (major adverse cardiac and cerebral events [MACCE]) occurred significantly more in the PCI arm than in the CABG one, driven by repeat procedures in the PCI group. For the "harder" end point of death/stroke/MI, rates were almost identical between the two groups; the stroke rate was higher in the CABG-treated patients. Two-year SYNTAX results showed MACCE rates continuing to diverge, still driven by higher repeat-revascularization rates and a signal of increased MI among PCI-treated patients. Last year, the three-year results made it crystal clear that patients at intermediate risk (SYNTAX score 23–32) are better off with open-heart surgery, as are those at highest risk.

Remember the commotion created by different stroke rates in the first years? There is no difference now.

The four-year SYNTAX results consist of follow-up in 819 CABG patients (91.3%) and 879 PCI patients (97.3%), Serruys said. Of note was the fact that the stroke rate evened out between the two groups.

"Remember the commotion created by different stroke rates in the first years?" he reminded attendees, noting there had been a "slow catch-up" for surgery, and the difference--initially significant and in favor of PCI--"is now no longer significant. There is no difference now," he observed.

And while deaths did become significantly greater in the PCI group, and the higher rate of MIs in the stent group also continued unabated, the "surrogate for safety" end point of death/stroke/MI remains not that dissimilar between the two arms, he noted.

SYNTAX 4-Year Cumulative Results

Outcome CABG surgery, n=819 (%) PCI, n=879 (%) p
MACCE 23.6 33.5 <0.001
Death/stroke/MI 14.6 18 0.07
All-cause mortality 8.8 11.7 0.048
Cardiac death 4.3 7.6 0.004
Stroke 3.7 2.3 0.06
MI 3.8 8.3 <0.001
Repeat revascularization 11.9 23 <0.001

Commenting on the discrepancy in follow-up between the two arms, whereby 78 CABG patients and 24 PCI patients were lost to follow-up, he noted that two different sensitivity analyses were performed to try to take this into account. The first assumed that all nonevaluable patients had died, "and when you do that, there is a still a [significant] difference in MACCE (p=0.04), but there is no difference in death/stroke/MI." The second assumed that all nonevaluable patients were event-free, he said, "and when you assume this, there is very close to no difference from what I have already presented."

Subgroup Analysis by SYNTAX Score; More to Come at TCT

Serruys also presented a subanalysis of the new results by SYNTAX score. "Subgroups have been criticized, but they are useful to dissect," he observed. He reminded attendees that the "original goal of the SYNTAX score was to provide guidance on optimal and detailed analysis of the coronary angiography anatomy; it was not yet a prognostic code."

Subgroups have been criticized, but they are useful to dissect.

Now, at four years, there is no difference in MACCE between CABG and PCI in those with a SYNTAX score of 0 to 22, he noted (26.1% vs 28.6%; p=0.57). This is "pretty good," he said, "and would legitimize the use of PCI in this kind of patient."

But for those with an intermediate SYNTAX score of 23 to 32, "you see immediately a highly significant difference" in MACCE rate (21.5% for CABG vs 32% for PCI; p=0.006). And for those with a high SYNTAX score (>33), "mortality is double in the PCI group compared with CABG (16.1% vs 8.4%; p=0.04) and MI is two to three times higher with PCI than with CABG (9.3% vs 3.9%; p=0.01)," he observed.

In this highest-risk group, even the end point of death/stroke/MI becomes significantly higher with PCI, Serruys added (22.7% vs 14.6%; p=0.01), and MACCE were much higher (40.1% vs 23.6%; p<0.001), driven in large part by a 17% higher rate of revascularization in this high-risk group at four years.

Four-year results in the subgroup of left main disease will be presented at TCT 2011 next month.

EXCEL Results Awaited; in the Meantime, "We Have a Clear Message"

Commenting on how things might change now that many are using the newer-generation stents such as the Xience, Serruys observed that some have gone so far as to say it is now "unethical" to implant Taxus and older-generation stents because of the difference in stent thrombosis and MACCE between the two products. "I wouldn't go that far," he noted, "but certainly the technology has improved."

In EXCEL, we are going for the irreversible end points--I lost my life, I lost my brain, or I lost part of my myocardium.

To that end, the results of the EXCEL trial are eagerly awaited, he said. EXCEL is a 2600-patient study comparing patients with left main disease randomized to bypass surgery or PCI with the Xience stent and followed for at least three years. The primary end point is death, stroke, and MI; repeat revascularization is a secondary end point.

"In EXCEL, we are going for the irreversible end points--I lost my life, I lost my brain, or I lost part of my myocardium. Revascularization is not benign, but it's less irreversible," Serruys pointed out.

In the meantime, he said, "I think we got a clear message."

He adds that it is vital for patients to be clearly informed of the numbers from SYNTAX and for there to be a proper discussion with a heart team composed of surgeons, interventionalists, and cardiologists, as indicated by the latest European guidelines.


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.