"Progress" is the word Dr Patrick W Serruys (Erasmus University
Medical Center, Rotterdam, the Netherlands), coˇVprincipal
investigator for the Synergy Between PCI With Taxus and Cardiac
Surgery (SYNTAX) trial, uses to summarize the landmark study,
published in the February 19, 2009 issue of the New England Journal
of Medicine (NEJM) [1].

Reflecting on over almost 30 years of angioplasty-vs-surgery trials,
Serruys says ever more patients are heading to the cath lab instead
of the OR, and SYNTAX expands the circle still further, suggesting
that at least some patients with left main or three-vessel disease
can safely and feasibly be treated with PCI.

"With every trial, we're taking a step forward. . . . I think we keep
making progress," he said.

It's a somewhat rosy take on what was, after all, a negative trial.
As previously reported by heartwire when the study was presented at
the ESC 2008 meeting, one-year results from SYNTAX showed that PCI
with drug-eluting stents (DES) was statistically inferior to CABG, at
least for the primary composite end point of major adverse cardiac or
cerebrovascular events (MACCE). But for combined "hard" end points--
death, MI, and stroke--event rates were no different between the two
trial groups, and secondaryˇVend-point findings indicated a
statistically higher risk of stroke with CABG and a statistically
higher risk of revascularization with PCI. Further details of the
trial, released at TCT 2008, suggested that in low- or intermediate-
risk patients with left main disease, PCI and CABG are both
reasonable options, at least for hard events at one year. In three-
vessel disease, however, only the lowest-risk patients should be
considered as candidates for PCI instead of surgery. These results,
however, while prespecified analyses, were merely "hypothesis-
generating," since the primary trial results were negative.

Patients deemed during the screening process to be unsuitable for
randomization were instead enrolled in one of two "nested" registries
(1077 to CABG; 198 to PCI). As previously reported by heartwire, CABG-
treated patients did extremely well (even better than CABG patients
in the randomized trial), while PCI-treated patients fared worse than
PCI-treated patients in the randomized arm.

The Essence of SYNTAX

Now published, the SYNTAX results are nearly identical to what
Serruys, with co-PI Dr Friedrich W Mohr (University of Leipzig,
Germany), previously presented (see related links). In the paper, the
authors cautiously conclude that the results of SYNTAX "show that
CABG as compared with PCI is associated with a lower rate of MACCE at
one year among patients with three-vessel or left main coronary
artery disease (or both) and should therefore remain the standard of
care for such patients."

But speaking with heartwire about the SYNTAX publication, Serruys
bluntly stated that he and his coauthors had grappled with that final
sentence: "At the end of the day, we have a study in the NEJM with a
last sentence that's been changed six or seven times during the
writing, because the NEJM wanted something more conservative."

In Serruys's opinion, that final sentence "is not the essence of the
trial."

The essence, he says, is that among the 1800 patients deemed by both
the surgeon and the interventionalist to be eligible for either
treatment, one-third of those patients can safely be treated with
either PCI or CABG. Furthermore, he said, the development of the
SYNTAX score--the other major contribution of the trial--allows
surgeons and cardiologists on the "heart team" to agree on a patient-
by-patient basis as to which therapy is appropriate. Anyone with a
cutoff score of 22, says Serruys, could "legitimately" be treated
with either therapy.

"When you do a trial, you have the rules of engagement, so you apply
one statistical rule to decide whether you have a positive or
negative result," he said. Yes, PCI did not meet the test of
noninferiority against CABG, he affirmed, "but that's why we created
the SYNTAX score, because we had the strong feeling that we had to
differentiate between patients on a case-by-case basis. That's what I
hope we have introduced to the medical field."

Serruys says he has increasingly been approached by cardiologists
interested in learning more about the SYNTAX score; he and his
colleagues will be launching a web-based tool and CD-program at
EuroPCR, he added.

Debating the Message

Mohr, for his part, agrees with Serruys about the critical role of
the "heart team" in the decision-making process, but is content with
the paper's wording, calling it an "honest analysis of the current
practice."

"The study clearly shows that about 65% to 70% of the study patients
with more complex three-vessel and/or left main disease . . . should
be treated by CABG, because the data show superior results: better
survival, lower myocardial infarction, etc," Mohr told
heartwire. "This is a very clear message, also demonstrated by the
8.8% MACCE rate at 12 months in the CABG registry." Moreover, MACCE
rates in the randomized arm will continue to diverge over time in
favor of CABG, Mohr notes.

He also points out that while rates of symptomatic bypass-graft
occlusion and stent occlusion were the same, at 3.5%, they have very
different clinical consequences--high mortality and MI rates for
stent occlusion, vs no mortality with bypass-graft occlusion. "This
will have a continuous impact in the future and puts PCI patients at
an imminent risk," Mohr said.

Informing the Decision

In an accompanying editorial, Drs Richard A Lange and L David Hillis
(University of Texas Health Science Center, San Antonio) describe how
the SYNTAX results can be applied in practice [2]. Patients who
cannot or will not take clopidogrel long term or in whom complete
revascularization is more easily accomplished with surgery "should be
encouraged to undergo CABG," they write. On the flip side, patients
with serious coexisting conditions or vessels not suitable to grafts
are better off with PCI. Patients in whom either strategy is
appropriate "should be presented with the advantages and
disadvantages of each procedure and allowed to choose between them."

According to interventionalist Dr Paul Teirstein (Scripps Clinic, La
Jolla, CA), who commented on the study for heartwire, the notion of
choice centers on the increased risk of stroke with CABG and the
increased risk of repeat procedures with PCI.

"Using the SYNTAX data to advocate CABG as the standard of care for
patients with three-vessel and left main disease is, in my opinion, a
case of 'good data, bad interpretation,' " he said. "If you talk to
most patients who undergo PCI, they tell you they would rather have
three, four, or even five PCI procedures instead of one bypass
surgery."

Crunching the numbers, Teirstein says that the small absolute
difference in revascularization procedures in SYNTAX translates into
a "number needed to prevent" of 14. "Thus, while our patients say
they would rather have several PCIs instead of one bypass, SYNTAX
teaches us we need to do 14 bypasses to prevent one repeat PCI, at a
cost of four times as many strokes. This argument makes no sense at
all and is just the opposite of what our patients want."

The argument is even more compelling in the left main subset, he
points out. "We have to do 19 CABGs to prevent just one repeat left
main PCI. This means 18 of every 19 CABGs we do for left main
patients are completely unnecessary! Certainly, with these new data,
patients with left main disease should be strongly considered for PCI
instead of CABG. I believe the SYNTAX data should be used to
highlight the fact that many bypass surgeries are currently being
performed unnecessarily."

Dr Grayson Wheatley (Arizona Heart Institute, Phoenix), a surgeon,
also commented on SYNTAX for heartwire, noting that the study
reaffirms the "excellence of treatment when a CABG is performed in
the correct patient."

That said, he continued, "PCI is eroding the 'CABG business,' and our
definition of the 'correct patient' for CABG is changing as new PCI
technology evolves. It is only a matter of time before there will be
another iteration of PCI technology that will prove to be superior to
DES."

Wheatley says he's seen an increase in use of PCI for three-vessel
and left main disease in recent months, but likely driven by safety
data supporting DES and not necessarily the SYNTAX results. "There
have been increased discussions between cardiac surgeons and
interventional cardiologists about the meaning of the SYNTAX data,
but little has changed with regard to treatment patterns," he said.
The increased dialogue, he said, is "always a good thing" and
underscores the need to "pause" between the diagnostic angiogram and
the treatment choice. "All too often there is no pause and an
angiogram automatically leads to PCI," Wheatley said. "Although such
a simple step, the introduction of a team discussion with the
patient, would better help define appropriate patient selection."
Everyone who commented on the study for heartwire also pointed to the
need for longer follow-up to tease out the impact of repeat
revascularization. "How significant is it really to reintervene in a
patient with PCI?" Wheatley asked. "The patient has still not had a
sternotomy and may only have had a few days of recovery vs a
significant recovery for CABG. Patients and interventional
cardiologists may feel that reintervention is an 'acceptable' event,
while cardiac surgeons may view reintervention as a failure. The
answer is not as simple as either side views it, and more attention
needs to be devoted to working through the metrics of reintervention."


Serruys P, Morice MC, Kappetein P, et al. Percutaneous coronary
intervention versus coronary-artery bypass grafting for severe
coronary artery disease. N Engl J Med 2009; 360:961-972.
Lange R, Hillis L. Coronary revascularization in context. N Engl J
Med 2009; 360:1024-1026.