SYNTAX Year Two: Significantly Higher MI Rate in PCI Arm, No Increase in CABG Strokes After 12 Months

Shelley Wood

September 02, 2009

September 2, 2009 (Barcelona, Spain) — Two-year results from SYNTAX show what many people had predicted: that event rates in patients treated with either CABG or PCI for complex coronary disease would continue to diverge, with CABG cementing its status as the preferred strategy. There's also a new signal of increased MI after one year among PCI-treated patients that had heads in the audience snapping to attention here at the European Society of Cardiology 2009 Congress.

But presenting the results, Dr A Pieter Kappetein (Erasmus Medical Center, Rotterdam, the Netherlands), who provided no details on the type or cause of MI, cautioned that it will be important to wait for the full five-year follow-up before drawing conclusions.

"At two years, [major adverse coronary and cerebrovascular event] MACCE rates were significantly higher for PCI than CABG, mainly driven by higher repeat revascularization in the PCI arm," he said. "The two-year results suggest that CABG remains the standard of care for patients with complex disease; however, PCI may be an acceptable alternative revascularization method to CABG when treating patients with less complex disease."

The surgeons will like this.

Speaking with heartwire after the presentation, Dr Alfred Bove (Temple University Medical Center, Philadelphia, PA) first commented: "The surgeons will like this.

"I think this supports the data that complex coronary disease probably ought to be treated surgically," Bove said. He pointed out that patients in the trial are typically getting four or five stents. "If you go in and try to stent all those lesions, knowing that the probability of restenosis is 5% or 6% in any given stent, multiple that by four or five, and you're going to have a pretty high recurrence rate," he calculated. "From the standpoint of the patient at the time, it doesn't sound good to have a sternotomy and all that stuff, but if you look out three or four years, they may forget the sternotomy a year later, and their long-term outcomes from a standpoint of recurrence might be better."

The SYNTAX Trial

As previously reported by heartwire , SYNTAX was an 1800-patient trial conducted in Europe and the US, randomizing patients to either CABG or PCI using the Taxus DES. The one-year results showed that the primary end point (MACCE) occurred significantly more often among PCI-treated patients than among CABG-treated patients, a 7.7% difference driven by repeat procedures in the PCI group. As such, PCI did not meet the prespecified margin of noninferiority of 6.6%. For the composite, "harder" safety end point of death/cerebrovascular events/MI, rates were almost identical between the two groups, whereas the stroke rate, by contrast, was higher in the CABG-treated patients. Of note, however, stroke rates were analyzed on an intention-to-treat basis, and almost half of the strokes in the CABG arm actually occurred prior to surgery, but postrandomization, while patients were awaiting a surgery date.

Almost 96% of the original trial numbers were included in the two-year analysis: 836 in the CABG arm and 885 in the PCI arm. At two years, MACCE rates were significantly different between the two groups, driven by a repeat revascularization rate in PCI-treated patients that was more than double that of the CABG-treated group. The significantly higher rate of strokes seen in CABG-treated patients at one year was also seen by two years, but the difference appeared to be a carryover from the first 12 months, since very few strokes occurred between the one- and two-year mark in either group. For the hard end point of death/stroke/MI, there were no significant differences between the two groups.

Two-Year Outcomes for SYNTAX

End point CABG (%) PCI (%) p
All-cause death 4.9 6.2 0.24
All stroke 2.8 1.4 0.03
Stroke before 1 y 2.2 0.6 0.003
Stroke after 1 y 0.6 0.7 0.82
MI 3.3 5.9 0.01
MI before 1 y 3.3 4.8 0.11
MI after 1 y 0.1 1.2 0.008
All-cause death, stroke, MI 9.6 10.8 0.44
Repeat PCI 8.6 17.4 <0.001
MACCE 16.3 23.4 <0.001

Subgroups at Two Years

Investigators also stratified the two-year findings according to SYNTAX score, noting that rates of MACCE were no different between the two revascularization strategies for patients who were low risk by SYNTAX score at baseline (17.4% for CABG, 19.4% for PCI; p=0.63). But as that risk rose, so too did the curves begin to separate: in patients with intermediate risk by SYNTAX, MACCE rates were 16.4% for CABG-treated patents and 22.8% for PCI-treated patients, just missing statistical significance (p=0.06). In high-risk patients, CABG was clearly the winner, with MACCE rates of 15.4% vs 28.2% in the PCI-treated group (p<0.001).

As with the one-year results, the two-year outcomes differed according to whether the patients were enrolled in the study for treatment of three-vessel disease or for left main stenting. Kappetein cautioned that the subset analysis included low numbers and was not appropriately powered, so it had to be considered only hypothesis-generating. But at least for the primary MACCE end point, event rates were significantly lower in CABG patients with three-vessel disease--14.4% vs 23.8% (p<0.001)--but were no different, statistically, between the groups for patients with left main disease--19.3% for CABG, 22.9% for PCI (p=0.27)

More Details, More Follow-Up Needed

Commenting on the study for heartwire , Dr Roxana Mehran (Columbia University, New York, NY) said, "I would like to know what types of MIs are seen--are these related to new procedures or stent thrombosis? This was not presented."

Also commenting, Dr William Wijns called the presentation "not breaking news in my opinion. . . . By and large, the results do not change much with respect to one year."

The efficacy is more or less the same, he noted, and the safety end points "are still okay, even though there is fear for late stent thrombosis in the PCI group. There seems to be a signal indicating that the initial good results with PCI will erode progressively, especially with infarction rates. We seem to recognize the signal because we may be expecting late stent thrombosis. However, so far numbers remain--fortunately--small."

Everyone, including Dr Manuel J Antunes (University Hospital, Coimbra, Portugal), the discussant for the presentation, echoed calls to withhold judgment until longer-term follow-up is in. He made the additional point that what's best for patients may depend on the specific expertise--interventional or surgical--at individual centers.

Overall, he said: "The two-year follow-up now presented to us confirms all the results and trends shown by the one-year report. The differences that were statistically different remain so and the differences that were not significant continue the trends toward significance, which, all appears to indicate, [they] will reach with time."

To access complete presentation slides (pdfs), click here.


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.