# Arteriovenous Malformation: What the Tangle Is, How It Is Graded, and Why Watching Is a Real Option URL: https://buzzle.com/health-fitness/arteriovenous-malformation Category: Health & Fitness Published: 2026-08-27T18:43:27 Updated: 2026-08-27T18:43:27 Image: https://buzzle.com/_astro/arteriovenous-malformation.C0nxRdHF_Z1gT1M3.webp An arteriovenous malformation is a knot of blood vessels in which arteries drain straight into veins. The capillary bed that normally sits between them is missing, so nothing slows the blood, drops its pressure, or hands oxygen to the tissue. In the brain, where these lesions get most of their attention, that missing step is the whole problem. Veins built for low pressure end up carrying arterial flow. Over years they stretch, thin, and sometimes tear. Most brain AVMs are thought to be present long before anyone knows about them. Most are found by accident, on a scan ordered for a head injury or a headache that turned out to be unrelated. Most never bleed. _This article explains a medical topic in general terms. It is not medical advice, it is not reviewed by a clinician, and it cannot account for your individual case. Decisions about your own diagnosis or treatment belong with your doctor._ What follows is the reference version. First, what the tangle does to the tissue around it, why it forms, and how it is found and confirmed. Then the grading scale neurosurgeons use before they decide anything, and the four standard responses, one of which involves no operation at all. Last, where the evidence on bleed risk is solid and where it is still being argued in the journals. ## What Is an Arteriovenous Malformation? A normal circuit runs artery, arteriole, capillary bed, venule, vein. The capillary bed is where the useful work happens, and it is also a brake. Blood entering it under arterial pressure passes through vessels roughly one red cell wide, gives up oxygen and takes on carbon dioxide, and leaves at a fraction of the pressure it arrived with. An arteriovenous malformation removes that brake. Feeding arteries connect through a tangle of abnormal channels, called the nidus, into draining veins, and the blood takes the shortcut. The speed of that shortcut is visible on an angiogram, which records the head over a couple of seconds rather than in a single frame. Contrast injected into a feeding artery rushes through the nidus and lights up the draining veins within a beat or two, well ahead of the surrounding brain, which is still waiting for its own blood to arrive. That early venous filling is one of the signatures radiologists look for, and it is a direct picture of the shunt at work. Several consequences follow from the missing capillary bed, and they explain most of what an AVM eventually does. - **The draining veins arterialize.** Carrying pressures they were not built for, they dilate and their walls remodel. The abnormal vessels of the nidus itself lack the orderly layers of smooth muscle and elastic tissue that let a healthy artery hold pressure. - **Aneurysms form on the circuit.** High flow through feeding arteries encourages flow-related aneurysms along them, and small aneurysms can develop inside the nidus. These are relevant because they are among the features associated with a higher chance of bleeding. - **Outflow can narrow.** If a draining vein develops a stenosis, a narrowing of its channel, pressure inside the nidus rises, which is a plausible mechanism for rupture and one reason drainage pattern matters so much in the assessments described below. - **The neighbouring tissue changes.** Brain adjacent to a long-standing AVM often shows gliosis, the scarring response of nervous tissue, and this is thought to underlie the seizures that bring some people to attention. The idea that an AVM starves nearby brain by diverting blood away from it, usually called steal, is older than the evidence for it. It is intuitive, it appears in a great many patient explanations, and the imaging studies that have looked for it have produced mixed results. It is fair to describe steal as a proposed mechanism rather than an established one. Size varies more than most people expect. Some are under a centimetre; some occupy a substantial part of a lobe. They occur elsewhere in the body too, in the spine, lungs, liver, and limbs, and a lung AVM behaves differently enough that it is really a separate subject. Within the skull, an AVM also has to be distinguished from its relatives, which have different risks and different management: cavernous malformations (low-flow, angiographically invisible), developmental venous anomalies (usually benign anatomical variants), capillary telangiectasias, and dural arteriovenous fistulas, which shunt into the venous sinuses rather than through a tissue nidus. When a report mentions any of these, the name is doing real work, and it is worth checking which one has actually been described. ## What Causes an AVM to Form? The traditional answer is that brain AVMs are congenital, present from birth, and arise during the development of the fetal vascular system. That answer is mostly right and has been getting more precise, and also somewhat more complicated. The precision came from genetics. Activating somatic mutations in the KRAS gene have been found in the endothelial cells lining sporadic brain AVMs, which places these lesions alongside other conditions driven by a mutation acquired in a single cell line rather than inherited from a parent. Somatic means the change is present in the abnormal tissue and not in the rest of the body, which has two consequences worth stating plainly: it is not passed to children, and it is not something anyone did or failed to do. There is no diet that produces an AVM, no injury that creates one, no stress that grows one. The KRAS change was identified in tissue removed at operation, and it sits in the lining of the malformed vessels rather than in the blood, which is why it does not show up on the germline genetic panels a family would be offered. The complication is that a small number of AVMs have been documented appearing on scans of people whose earlier imaging was clean, and others have regrown after apparently complete treatment. Neither is common. Both sit awkwardly with a strictly congenital account, and the current picture allows for lesions that develop or evolve after birth rather than simply enlarging from something already there. A minority of cases are inherited, through defined syndromes rather than a general family tendency. - **Hereditary hemorrhagic telangiectasia,** also called Osler-Weber-Rendu syndrome, is autosomal dominant and involves mutations in genes including ENG and ACVRL1. It produces recurrent nosebleeds, small telangiectasias on the lips, tongue, and fingers, and arteriovenous malformations in the lungs, liver, and brain. Brain involvement affects a minority of people with the syndrome, but the syndrome is worth identifying because the lung and liver lesions carry their own consequences and because relatives can be screened. - **Capillary malformation-arteriovenous malformation syndrome,** associated with [RASA1 and EPHB4 mutations](https://www.ncbi.nlm.nih.gov/books/NBK52764/), links characteristic skin capillary stains with arteriovenous lesions elsewhere. This is one of the places where the honest answer is thinner than people want. Genetics explains what kind of error occurred. It does not yet explain why this vessel, in this part of this brain, at this moment in development, and no risk factor has been identified that predicts who will have one. ## What Symptoms Does an AVM Cause? Increasingly, none. The share of brain AVMs discovered incidentally has risen over the past few decades, as MRI became widely available and routine for headaches, dizziness, minor head trauma, and research studies. Scans ordered for one thing turn up another. A lesion that produces no symptoms produced no symptoms in 1975 either. Nobody knew it was there. Among those that do announce themselves, the presentations fall into a few groups: - **Hemorrhage.** Historically the most common symptomatic presentation. Bleeding from a brain AVM is more often into the brain tissue itself than into the subarachnoid space, which distinguishes it from the pattern of a ruptured berry aneurysm. Onset is abrupt. - **Seizures.** A substantial share of symptomatic cases present with a first seizure, focal or generalized, thought to arise from the irritable, gliotic tissue bordering the nidus. Cortical and superficial lesions are more often implicated than deep ones. - **Focal neurological deficit.** Progressive weakness, numbness, speech disturbance, or visual field loss, developing without a bleed. Less common than the first two. - **Headache.** The most contested category. Headache is common in the general population and common in people with AVMs, and separating the two has proved difficult. The frequently repeated claim that AVMs cause a characteristic migraine-like headache on the same side as the lesion is repeated with more confidence than the evidence supports. In infants and very young children a large AVM can present differently, through high-output cardiac strain from the volume of blood shunted, or through an enlarging head circumference. Vein of Galen malformations, a distinct entity of the newborn period, are the classic version of this. The symptoms of a bleed inside the skull are worth stating once, in their standard published form, because they are a medical emergency regardless of the cause: a sudden, severe headache unlike any previous headache, often described as reaching full intensity within seconds; vomiting; a stiff neck; sudden weakness, numbness, or drooping on one side; sudden difficulty speaking or understanding speech; sudden loss of vision; a first seizure; or loss of consciousness. These warrant emergency services, not an appointment. That is the standard guidance for any suspected intracranial hemorrhage and it is not specific to AVMs. ## How Is an AVM Diagnosed? The imaging sequence is ordered by what each test rules out fastest, not by how much detail it produces. The most detailed study comes last, because it is also the only one that involves threading a catheter into someone’s arteries. ### CT and CT angiography In an emergency, a non-contrast CT comes first, and it answers exactly one question quickly: is there fresh blood inside the skull. It is fast, widely available, and good at that job. It is poor at showing an unruptured AVM, which may be invisible or show up only as a vague area of increased density. Adding contrast for a CT angiogram brings the vessels out and can demonstrate the nidus, the larger feeding arteries, and the draining veins within minutes of arrival. ### MRI and MR angiography MRI is where anatomy gets settled. Rapidly flowing blood in the abnormal vessels produces flow voids, dark serpentine channels on standard sequences, which are close to diagnostic in the right context. Gradient echo and susceptibility sequences pick up hemosiderin, the residue of old blood, which can reveal a previous bleed the person never knew about. Above all, MRI shows exactly where the lesion sits relative to the parts of the brain that carry function, and that relationship drives every subsequent decision. Functional MRI and diffusion tensor imaging are sometimes added to map language areas and white matter tracts before an operation. MR angiography is non-invasive but its resolution does not match what comes next. ### Catheter angiography Digital subtraction angiography remains the reference standard. It is done by passing a catheter from the femoral or radial artery up to the vessels of the neck, then injecting contrast while a rapid series of X-ray images is taken. Its advantage is time. It shows flow moving, in sequence. That lets the team identify each feeding artery on its own, see how the nidus fills, count and characterize the draining veins, find small aneurysms on the feeders or within the nidus, and time how fast blood crosses the shunt. None of that is fully available from a static picture, and all of it feeds the grading and treatment discussion. The trade-off is that it is an invasive procedure with a small but real risk of stroke, along with the more common minor problems of bruising or bleeding at the puncture site and reactions to contrast. It is not ordered casually, and it is generally reserved for the point at which its answers will change what happens next. One practical wrinkle: a large hematoma can compress an AVM enough to hide it, and an initial angiogram after a bleed can come back negative when a lesion is genuinely present. Repeating the study weeks later, once the clot has resorbed, is a recognized part of the workup rather than a sign that something was missed. ## How Do Doctors Grade an AVM Before Deciding on Treatment? The Spetzler-Martin scale, in use since the mid-1980s, is the common language. It assigns points across three features and adds them up. - **Size of the nidus:** under 3 cm scores 1 point, 3 to 6 cm scores 2, over 6 cm scores 3. - **Eloquence of adjacent brain:** 0 if the surrounding tissue is not eloquent, 1 if it is. Eloquent in this context has a defined list rather than a general meaning, covering the sensorimotor, language, and visual cortex, the hypothalamus and thalamus, the internal capsule, the brainstem, the cerebellar peduncles, and the deep cerebellar nuclei. - **Venous drainage:** 0 if drainage is entirely superficial, 1 if any component drains into the deep venous system. The total runs from 1 to 5. A separate designation, often written as grade 6, marks a lesion considered not operable at all. What the scale predicts is narrower than most people assume on first reading it. Spetzler-Martin was built to estimate one thing: the likelihood that surgical removal [leaves the person with a new neurological deficit](https://pubmed.ncbi.nlm.nih.gov/12657169/). Low grades correspond to lesions that surgeons remove with comparatively low rates of new deficit. High grades correspond to lesions where the operation itself carries substantial risk. The grade says nothing directly about how likely the AVM is to bleed if left alone. That estimate comes from natural-history cohorts, which follow untreated lesions over years and record what happens to them. The scale was derived from surgical series and validated against surgical results. Two different bodies of evidence answer the two questions, and a grade carries only the first. A supplementary scale, developed later, adds further variables that turned out to carry predictive weight: the person’s age, whether the lesion presented with a hemorrhage, whether the nidus is compact or diffuse, and whether it draws supply from deep perforating arteries. A number of centres now score it alongside the original. Diffuseness in particular is a surgical reality that the 1986 scale did not capture: a nidus with poorly defined borders is harder to separate from working brain than a well-circumscribed one of identical size. Two caveats belong with any description of these scales. Neither is fully objective, since judging whether adjacent tissue counts as eloquent, or whether a nidus is compact, involves interpretation on which experienced readers do not always agree. And applying either one to a specific lesion is the work of the team looking at that person’s angiogram, with the rest of their medical history in hand. The scales exist here as reference, so that a grade quoted at an appointment means something concrete rather than sounding like a verdict. ## What Are the Treatment Options for an AVM? Four approaches are in standard use: microsurgical resection, endovascular embolization, stereotactic radiosurgery, and observation. They are not steps in a sequence, and the fourth is a chosen course rather than a failure to choose one. Many cases combine two of them. ### Microsurgical resection An operation through a craniotomy, an opening made in the skull, in which the surgeon works around the outside of the nidus, closing off the feeding arteries in turn, separating the tangle from the brain around it, and dividing the main draining vein last. The order is not arbitrary. Taking the outflow before the inflow is closed would pressurize a lesion that has nowhere to drain. The attraction of surgery is that it is immediate and verifiable. A post-operative angiogram either shows the lesion gone or it does not, and if it is gone, the bleeding risk is gone with it. The cost is the operation: new neurological deficit, bleeding, infection, and the general risks of a craniotomy, weighted by exactly the factors the grading scales measure. ### Endovascular embolization Performed through the same catheter route as diagnostic angiography, embolization delivers a material into the vessels of the malformation to block flow. Liquid embolic agents that solidify on contact with blood are the common tools; coils and particles are used in some situations. It is most often an adjunct rather than a cure, done before surgery to reduce blood loss, or before radiosurgery to shrink a nidus, sometimes across several staged sessions weeks apart. Cure by embolization alone happens, but it is largely confined to small lesions with a simple arterial supply. Its risks are those of working inside abnormal vessels: rupture during the procedure, embolic material travelling somewhere unintended, and stroke in territory supplied by a vessel that also feeds normal brain. ### Stereotactic radiosurgery Focused radiation delivered in one session or a small number of sessions, using a Gamma Knife unit, a linear accelerator, or a proton beam. The radiation does not remove anything. It injures the walls of the abnormal vessels in a way that causes them to thicken and gradually close over a period usually quoted as two to four years, with the range varying between series. That delay is the defining feature of the modality, and it cuts both ways. The lesion is not treated in any protective sense until obliteration occurs, so whatever bleeding risk it carried continues through the latency period. Against that, radiosurgery reaches lesions a scalpel cannot reach without crossing functioning brain, which makes it a common choice for small, deep, compact malformations. Obliteration rates are considerably better for smaller volumes than for large ones. Complications include radiation-induced changes in the surrounding tissue, swelling, and, less often, radiation necrosis. When a follow-up angiogram at three years shows residual filling, repeat treatment is an option. ### Observation Observation means surveillance imaging on a defined schedule, treatment of seizures with medication if they occur, attention to general vascular health, and no intervention on the lesion itself. It is a real option with a real evidence base behind it, and the reason it is taken seriously today is largely one trial. ARUBA randomized people with unruptured brain AVMs to intervention or to medical management alone, and the medically managed group had [fewer strokes and deaths](https://www.thelancet.com/journals/lancet/article/PIIS0140-6736\(13\)62302-8/abstract). It changed practice, and it has been argued about ever since. Three objections come up most often. Its follow-up was short relative to a lesion someone may carry for fifty years. The intervention arm was weighted toward embolization rather than the surgery many centres would have chosen. And only a small fraction of screened patients were enrolled, which raises the question of whom the results actually describe. Both the finding and the criticisms are legitimate, and the honest summary is that the case for treating an unruptured AVM is genuinely unsettled and depends heavily on the individual lesion. Across all four, the factors that move a case toward one approach or another are consistent: the size, depth, and eloquence of the surrounding tissue, the venous drainage pattern, whether the lesion has already bled, any associated aneurysms, the person’s age and other medical conditions, and the experience of the centre. That last one is not a courtesy. Outcomes for complex cerebrovascular work correlate with how many such cases a team handles. The largest fork in the road is whether the AVM has ruptured. A lesion that has bled is generally treated, because rupture identifies it as one that ruptures. The whole argument sits on the unruptured side. ## What Is the Outlook for Someone With an AVM? The figure most often quoted for an unruptured brain AVM is an annual bleed risk of roughly 2 to 4 percent, and that range has held up reasonably well across long-running cohort studies. It is the number a reader will meet at an appointment, and it deserves three qualifications. First, it is an average across a heterogeneous group of lesions. The same studies also identified features that move an individual well off the average, in both directions. A previous hemorrhage, deep location, drainage only into the deep venous system, and associated aneurysms all appear repeatedly as markers of higher risk. A superficial lesion with superficial drainage and no history of bleeding sits at the low end of that band, and a deep lesion that has already bled sits considerably above it. Second, risk is not constant over time. After a rupture the chance of bleeding again is elevated for roughly the first year, reported in various series as several times the baseline rate, before declining back toward it. This is one of the reasons a ruptured AVM is usually treated and an unruptured one is discussed. Third, an annual percentage compounds. In clinic, cumulative risk is usually framed against how many years a person is likely to be living with the lesion, rather than against any single year. The same annual rate accumulates to a very different total across forty years than across ten, which is one reason age carries weight in the supplementary grading scale, and one reason two people quoted the same 2 to 4 percent can leave their appointments with different recommendations. When an AVM does bleed, outcomes are on average less severe than those of a ruptured berry aneurysm, partly because the bleeding is usually into brain tissue rather than into the subarachnoid space, and partly because the vessels involved have been under high pressure for years and are less prone to the catastrophic arterial rupture pattern. Less severe on average still covers a wide spread. Death and lasting neurological deficit are both well-documented outcomes. They occur in a minority of bleeds, but not a negligible minority, and recovery from a survived hemorrhage typically takes months of rehabilitation rather than weeks. Seizures, where they were the presenting problem, often remain manageable with medication whether or not the lesion is treated. Several things people reasonably want to know have thinner answers than they should. Whether pregnancy raises the risk of AVM rupture has been studied repeatedly with conflicting results, and centres differ in what they advise. Whether ordinary physical exertion triggers bleeding has never been demonstrated convincingly, which has not stopped the restriction from being handed out in some places and not others. The evidence in both cases is weaker than the confidence with which the guidance is usually delivered, and a reader who has been told two different things by two different clinicians is not necessarily hearing an error. A population average describes what happens to a group over a year. It cannot say what a particular person’s year will hold, and no honest quotation of it pretends otherwise. ## What Questions Are Worth Asking a Doctor About an AVM? The questions below are the ones that tend to change the shape of the conversation, phrased more or less as a person would actually say them: - Has this one bled? Is there any sign on the scans of an old bleed I never knew about? - What grade is it, and which of the three components drove that grade: the size, the location, or the way it drains? - Where does it drain, superficially or into the deep venous system, and does that change what you would consider? - Are there any aneurysms on the feeding arteries or inside the nidus? - Which of the four approaches are genuinely on the table here, and which are ruled out by the anatomy rather than by preference? - If radiosurgery is being considered, what obliteration rate is realistic for a lesion this size, and what happens during the years before it closes? - If we watch it, what does watching consist of: which scans, how often, and what finding would change the plan? - How many AVMs of roughly this grade does this centre treat in a year? - What restrictions, if any, are you recommending, and is that based on evidence or on local practice? The state of knowledge, stated flatly: the anatomy is well described and the imaging is excellent, the mechanism of formation is better understood than it was a decade ago, and the grading of surgical risk is mature and reproducible enough to be trusted. What remains open is the question that matters most to someone holding an unruptured lesion, which is whether treating it does more good than harm over a lifetime rather than over three or five years. ARUBA answered part of that and provoked the rest. The trials and registries that would settle it are still accumulating follow-up, and until they report, the decision is made lesion by lesion, in a room with the angiogram on the screen.