Asymptomatic Carotid Narrowing: What CREST-2 Tells Us

The short answer

CREST-2 asked a question that had gone unanswered for a generation: when someone has severe narrowing of a carotid artery but no symptoms, does a procedure still add anything once modern medication is doing its job? The two parallel trials, published in the New England Journal of Medicine on November 21, 2025, found that carotid stenting plus intensive medical management lowered the combined risk of stroke or death compared with intensive medical management alone, while adding carotid endarterectomy did not reach statistical significance. The real lesson is less about which procedure wins and more about how far the medical baseline has moved since the older trials that made surgery routine.

This article is educational and not medical advice.

Why the question needed reopening

The carotid arteries carry blood to the brain, and plaque can narrow them over years without causing any warning. For decades, the case for operating on such narrowing rested on trials from the 1990s and 2000s, chiefly ACAS and ACST-1. Those studies showed that carotid endarterectomy, a surgery that removes plaque from the artery wall, modestly reduced future stroke in people with tight asymptomatic stenosis. The catch is that they ran in an era when blood pressure control was looser, potent statins were not yet standard, and antiplatelet strategy was less refined. The comparison arm in those trials received medical care that no clinician today would consider adequate.

That gap matters because the benefit of any preventive procedure is measured against whatever the alternative achieves. If the non-surgical alternative has improved dramatically, the margin a procedure has to beat rises with it. CREST-2, registered as NCT02089217, was designed to test carotid revascularization against a genuinely contemporary medical standard rather than an outdated one.

How the trial was built

CREST-2 was not one study but two parallel, observer-blinded randomized trials run across 155 centers in five countries, enrolling roughly 2,485 patients with asymptomatic stenosis of 70 percent or greater and a mean age near 70. One trial compared carotid-artery stenting plus intensive medical management against intensive medical management alone. The other compared carotid endarterectomy plus intensive medical management against intensive medical management alone. Both used the same rigorous medical protocol as the shared foundation, which is what makes the two results directly comparable.

The medical management arm is the part worth dwelling on. It targeted systolic blood pressure below 130 mm Hg and LDL cholesterol below 70 mg/dL, with structured attention to glucose, smoking, weight, and physical activity, plus health coaching. This was not medication in name only; it was an actively managed, goal-driven regimen. The primary endpoint combined any stroke or death within 44 days of randomization with ipsilateral ischemic stroke over the following four years, a design that captures both the upfront risk of a procedure and the longer-term stroke protection it is supposed to buy.

What the numbers showed

In the stenting trial, the primary outcome occurred in 6.0 percent of patients on medical management alone versus 2.8 percent of those who also received stenting, a difference that reached statistical significance (P = 0.02), corresponding to a number needed to treat around 31. In the endarterectomy trial, the figures were 5.3 percent for medical management alone versus 3.7 percent with added surgery, a smaller gap that did not reach significance (P = 0.24). These figures are reported in the NEJM primary publication and summarized in the American College of Cardiology's journal scan.

Two features deserve emphasis before anyone reads the stenting result as a verdict that stenting is superior to surgery. First, these were separate trials, not a head-to-head comparison, so the correct reading is that each procedure was tested against the same medical benchmark, not against each other. Second, the periprocedural window tells its own story: in the first 44 days, the stenting group carried a stroke-or-death rate near 1.3 percent against zero events in its medical arm, and the surgical group ran about 1.5 percent against 0.5 percent medically. Every procedure front-loads risk, and the long-term benefit has to be large enough to repay that early cost.

The finding hiding behind the headline

The most consequential number in CREST-2 may be the one that did not make headlines: how low the event rates were across every arm. Stroke rates on medical management alone came in well below what the older trials would have predicted, which tells you the modern regimen is doing much of the protective work that surgery once had to supply. When the baseline gets that good, the room for a procedure to add value shrinks, and small differences in trial execution can decide whether a result lands on one side of significance or the other.

That framing helps explain the divergence. The stenting benefit was real in this trial, but commentators have noted it depended on experienced operators and careful patient selection, so the result may not transfer cleanly to lower-volume settings. The endarterectomy finding, meanwhile, is consistent with a broader message: for many patients who can tolerate intensive medical therapy, routine surgery for asymptomatic narrowing is harder to justify than it once was. The commentary in Stroke: Vascular and Interventional Neurology frames CREST-2 as providing clarity precisely because it reanchors the decision around today's medical standard.

How to read results like these

CREST-2 is a useful case study in evidence literacy. A single P value of 0.02 next to a P value of 0.24 invites the conclusion that stenting works and surgery does not, but that reading collapses two separate experiments into a rivalry they were never structured to settle. The more defensible takeaways are narrower and more durable: intensive medical management has become a powerful stroke-prevention tool in its own right; any procedure must now clear a higher bar; and periprocedural risk, operator experience, and patient selection shape whether that bar is cleared. Absolute differences of a few percentage points, not relative claims, are what a reader should carry away.

For an individual, none of this substitutes for a clinician who knows the specific anatomy, the degree of narrowing, other vascular risk, and personal preference. What CREST-2 offers is a clearer map of the terrain, and a reminder that the value of an intervention is never fixed. It is always measured against the best available alternative, and that alternative has quietly gotten a great deal better.