A new deficit in recovery is two different problems wearing the same face. Where it sits — inside one arterial territory, or on the seam between two — narrows the mechanism faster than how bad it looks.
A deficit that fills one arterial territory to its edges points at that vessel: something closed it. A deficit that sits on the border between two territories points at pressure and delivery — the variables you were holding. The pattern narrows the mechanism, but it does not close it, because the border zone is also where emboli that were never washed through end up. Both patterns are worth recognizing on sight, and neither is a reason to stop looking.
The brain is fed by two systems that meet only at the base of the skull. The anterior circulation is the two internal carotid arteries and everything that derives from them; the posterior circulation is the two vertebral arteries, joining to form the basilar. Each ends in territories that are, functionally, end-arterial — an anterior cerebral strip along the midline, a middle cerebral territory over most of the convexity, a posterior cerebral territory at the back. Close one of those vessels and the deficit takes the shape of what it fed.
Between the two systems sits the circle of Willis, a hexagonal ring lying in the subarachnoid space around the pituitary gland. It is easy to over-read. In the normal state there is essentially no net flow through the communicating arteries at all, only a to-and-fro movement that keeps them from thrombosing shut. They carry blood when a pressure difference appears across them, and not before.
How reliable that reserve is depends on which question you ask. On MR angiography of 150 healthy volunteers, an entirely complete circle — every segment present — was found in 42%. Cottrell’s stricter reading, requiring components that are not just present but well developed and symmetrical, puts the figure at 18–20%. Those are not rival numbers; they answer different questions. Present and adequate are different properties, and a 0.5 mm communicating artery satisfies the first while failing the second at the moment it is needed.
The second tier of collateral is more diffuse. Pial arteries on the surface of the brain connect across the major territorial boundaries — anterior to middle, middle to posterior, anterior to posterior — through what are variously called pial-to-pial anastomoses or leptomeningeal pathways. These are what protect the border zones. They are also slow: collateral pathways are most effective in chronic ischemia, where they enlarge over weeks, and in the acute setting it is frequently necessary to raise the blood pressure to drive any flow across them at all. Since inadequate collateral supply is a normal anatomic variant, induced hypertension is not guaranteed to work.
The hydraulics are simple once vasodilation runs out. Distal to a stenosis, or below the lower limit of autoregulation, the arteriolar bed is already near-maximally dilated to keep resting flow normal — which means resting flow can look entirely normal while the reserve behind it is gone. From that point flow is pressure-passive, and the lowest pressures are found in the tissue farthest from its arterial input. That tissue is the border zone. It is the region most likely to become ischemic during systemic hypotension, and the reason is geometric rather than metabolic.
The more precise name for the border zone is the equal pressure boundary: the line at which the pressure arriving from one territory equals the pressure arriving from the next, so nothing crosses. That definition carries a consequence people rarely draw out. The boundary is not a fixed anatomic landmark. Its position varies considerably between individuals, and it moves when the pressures on either side change — during a fall in blood pressure, after a stenosis progresses, after an embolization alters the vascular architecture.
The teaching version says: one territory means an embolus, the seam means the pressure fell. The first half holds up. The second half is where honest uncertainty starts, and it is worth knowing about before someone senior points it out.
Momjian-Mayor and Baron reviewed the perfusion imaging in internal carotid artery disease and found hemodynamic compromise and embolism to be cooperating mechanisms rather than competing ones. The link is mechanical: a territory with no perfusion reserve is a territory with slow flow, and slow flow cannot wash a small embolus through to where it will do less harm. Emboli lodge preferentially where clearance is worst, and clearance is worst exactly where perfusion pressure is lowest — the border zone.
So the pattern is genuinely informative and genuinely incomplete. A border-zone deficit should send you to the blood pressure record first. It should not stop you from asking where an embolus might have come from.
No randomized trial addresses this.
The anatomy is settled and the hydraulics are settled. Beyond the point of maximal vasodilation, flow becomes pressure-passive and the tissue farthest from its arterial input is the tissue that fails first. That is why the border zones are the regions most likely to become ischemic during systemic hypotension, and it is not in dispute.
What is not settled is the inference residents are usually taught to make from it — that a border-zone infarct means the pressure fell. Momjian-Mayor and Baron's review of the perfusion imaging in internal carotid artery disease found the honest picture to be a joint one: hemodynamic compromise and embolism are not competing explanations for border-zone infarction so much as cooperating ones, because a bed with no perfusion reserve is also a bed that cannot wash a small embolus through. The pattern raises the probability of a pressure problem. It does not establish one.
No trial has randomized anything here, and none is likely to. This day rests on vascular anatomy, on perfusion imaging series, and on the consensus that has grown around them.
The characteristic failure on this day is not misnaming a territory. It is accepting "still sleepy" as an account of a patient who has a new focal sign, and letting the first hour go by without anyone examining them.
This is a real exposure rather than a theoretical one. In the NeuroVISION cohort, 1 in 14 patients over 65 having inpatient noncardiac surgery had an acute brain infarct on postoperative MRI, and the overwhelming majority were covert — no deficit obvious enough that anyone in the room called it a stroke. Whatever else that finding means, it establishes that perioperative cerebral infarction is common enough that a focal sign in recovery deserves a real examination rather than a reassuring adjective.
The mirror-image error is reflexive: a new deficit appears and the response is to raise the blood pressure and wait. For a border-zone pattern that instinct is defensible. For a territory filled to its edges it is a way of doing something while the treatment that matters — reopening the vessel — is not happening.
Full text · Preston goes through LibKey to whatever Preston holds — the PDF where there is one — and works on a phone. The doi link always reaches the publisher's record, whether or not we hold a subscription. If you are reading on a computer, LibKey Nomad adds the same one-click PDF button to PubMed and publisher pages everywhere else; it is a desktop browser extension, so it does nothing on a phone or tablet.
A 71-year-old with treated hypertension and a known 60% left internal carotid stenosis has an uneventful four-hour shoulder arthroplasty in the beach-chair position. In recovery he is awake and oriented but has bilateral proximal arm weakness with the legs comparatively spared. Which interpretation best fits?
The border zone is sometimes described more precisely as an "equal pressure boundary." Which statement best captures why that phrasing matters?
Under normal physiologic conditions, what is happening in the anterior and posterior communicating arteries of the circle of Willis?
No returning items before day 4. The retrieval window opens once there is something far enough back to be worth retrieving.