Asymptomatic Carotid Narrowing: What CREST-2 Tells Us

The short answer#

For someone with severe but symptom-free narrowing of a carotid artery, CREST-2 asked whether a procedure still adds protection once medication is doing its job well. The answer, published in the New England Journal of Medicine on November 21, 2025, was nuanced: adding carotid stenting to intensive medical therapy lowered the combined risk of stroke or death, while adding surgery did not reach statistical significance. The deeper message is that the medical baseline has improved so much that the case for routine procedures is no longer what it was a generation ago.

Key points#

• The carotid arteries feed the brain, and plaque can narrow them silently for years.
• CREST-2 compared two procedures, each against the same modern, goal-driven medical regimen.
• Stenting plus medication beat medication alone; surgery plus medication did not reach significance.
• These were two separate trials, not a head-to-head contest between the two procedures.
• Event rates were low in every arm, which is the finding that reframes the whole debate.

An old assumption meets a new baseline#

Plaque in a carotid artery can build for years without any warning sign. For a long time the argument for operating on that narrowing rested on trials from the 1990s and 2000s, notably ACAS and ACST-1. Those studies showed that carotid endarterectomy, a surgery that scrapes plaque from the artery wall, modestly reduced future stroke in people with tight narrowing and no symptoms.

The problem was the yardstick. Those trials compared surgery against the medical care of their day, when blood pressure targets were looser, high-potency statins were not yet routine, and antiplatelet strategy was less refined. Any preventive procedure is only as valuable as the margin by which it beats the alternative, and the alternative has moved. If medication alone now prevents many of the strokes that surgery used to prevent, the bar a procedure must clear rises with it. CREST-2, registered as NCT02089217, was built to test that exact question against a current medical standard rather than a dated one.

What "intensive medical management" actually meant#

The medication arm is the part of CREST-2 worth studying closely, because it was not a token comparison. It aimed for systolic blood pressure below 130 mm Hg and LDL cholesterol below 70 mg/dL, with structured attention to blood sugar, smoking, weight, and physical activity, supported by health coaching. In other words, it was an actively managed program with numeric goals, not a prescription handed over and forgotten.

CREST-2 ran as two parallel, observer-blinded randomized trials across 155 centers in five countries, enrolling roughly 2,485 participants with narrowing of 70 percent or greater and a mean age near 70. One trial paired carotid-artery stenting with that medical program and compared it to the program alone. The other paired carotid endarterectomy with the same program and made the same comparison. Because both trials shared an identical medical foundation, their results can be read side by side.

Two trials, not a duel#

Here is the point that is easiest of all to get wrong. In the stenting trial, the primary outcome occurred in 6.0 percent of the medication-only group versus 2.8 percent of those who also received a stent, a difference that reached significance (P = 0.02), roughly a number needed to treat of 31. In the endarterectomy trial, the figures were 5.3 percent for medication alone versus 3.7 percent with added surgery, a narrower gap that did not reach significance (P = 0.24). These numbers appear in the NEJM report and are summarized in the American College of Cardiology's journal scan.

It is tempting to line those two P values up side by side and conclude that stenting works and surgery does not. That reading fuses two separate experiments into a rivalry they were never designed to settle. Each procedure was measured against the shared medical benchmark, not against the other one. A fair summary is that one procedure cleared the bar in its trial and the other did not, in populations and operators that were not identical.

Reading the numbers without overreading them#

The primary endpoint was deliberately built to capture both sides of the trade-off: any stroke or death within 44 days of randomization, plus ischemic stroke on the same side over the following four years. That structure prices in the upfront hazard of a procedure alongside the longer-term protection it is meant to deliver.

The early window is where procedures pay their tax. In the first 44 days, the stenting group carried a stroke-or-death rate near 1.3 percent against no events in its medical arm, and the surgical group ran about 1.5 percent against 0.5 percent on medication. Every procedure front-loads risk, so the later benefit has to be large enough to repay that opening cost. Commentators have also noted that the stenting result leaned on experienced operators and careful patient selection, which means it may not carry over cleanly to lower-volume settings.

Why the low event rates are the real story#

The most consequential figure in CREST-2 is one that did not make headlines: stroke rates on medication alone came in well below what the older trials would have predicted. That single fact tells you the modern regimen is now doing much of the protective work surgery once had to supply. When the baseline is that strong, the space for a procedure to add value shrinks, and small differences in how a trial is run can decide which side of significance a result lands on.

The commentary in Stroke: Vascular and Interventional Neurology frames the trial as offering clarity precisely because it re-anchors the decision around today's medical standard. For many people who can tolerate an intensive, goal-driven regimen, routine surgery for symptom-free narrowing is harder to justify than it used to be. Absolute differences of a few percentage points, rather than relative claims, are what you should carry away.

What this changes for a patient conversation#

None of this replaces a clinician who knows your artery, the degree of narrowing, the rest of your vascular risk, and what you want. What CREST-2 provides is a clearer map: the value of any intervention is measured against the best available alternative, and that alternative has become genuinely better. So the conversation now starts with whether your medication plan is truly optimized, and only then weighs whether a procedure adds enough to be worth its early risk.