An arteriovenous malformation (AVM) is an abnormal connection of arteries and veins without the usual capillary network between them. MedlinePlus. Management depends on its location, symptoms, previous bleeding and treatment risks. An AVM in the brain raises different questions from one in a limb or an inherited vascular syndrome. Confidence: moderate for these distinctions and specialist assessment; low for a universally superior procedure or targeted medicine from financially cleared evidence.
- AVM is a specific vascular diagnosis; not every birthmark, vein abnormality or vascular tumour is an AVM.
- A thunderclap headache or sudden stroke symptoms require emergency care.
- Observation can be active care, with an agreed review and warning-sign plan.
- The ARUBA trial concerns selected adults with unruptured brain AVMs, not all vascular malformations.
- No supplement replacement or universal drug, procedure or screening schedule is established here.
Table of contents
- Evidence summary: the question differs for each AVM
- What are AVMs, fistulas and other vascular anomalies?
- Flow, genes and inherited vascular syndromes
- Treatment: observation, surgery, embolization and radiosurgery
- Supplements and targeted medicines: what is not established?
- Practical support during surveillance and after bleeding
- Safety: sudden headache, neurological changes and bleeding
- Medicines, procedures, pregnancy and coexisting disease
- Diagnosis: location, imaging and the possibility of a syndrome
- Deciding on a plan and keeping follow-up meaningful
- Animal and molecular research: promise without a proven cure
- Funding and source roles
- Frequently asked questions
- Sources and funding notes
Evidence summary: the question differs for each AVM
In the 2020 ARUBA follow-up report, medical management alone had fewer deaths or symptomatic strokes than intervention in the selected adults studied with unruptured brain AVMs. This is an important finding, not a reason to declare every intervention unnecessary. The trial stopped early, used varied interventions and had finite follow-up; lifelong comparative risk remains uncertain.
Read the population before adopting a treatment claim. Was the lesion previously ruptured? Was the study about a brain, lung or limb lesion? Were children included? Was the outcome complete closure on an image, neurological function, bleeding, pain or daily activity? These are different questions. A successful image or pooled series cannot by itself settle the balance of benefit and harm for an individual.
The February 2026 VASCERN pathway is expert and patient-representative opinion requiring further validation. It distinguishes central-nervous-system, visceral and peripheral lesions and supports specialist multidisciplinary review. It is attributed guidance, not an independently cleared comparison of all available procedures.
What are AVMs, fistulas and other vascular anomalies?
The ISSVA classification places AVMs among fast-flow vascular malformations. An arteriovenous fistula is a direct connection rather than the same complex tangle; venous and lymphatic malformations belong to different categories. Vascular tumours such as haemangiomas are also distinct. The terminology matters because drugs, imaging and procedures cannot be transferred automatically between these conditions.
Ask the clinician to name the lesion precisely in the report, including its site and whether it is part of a syndrome. “Vascular abnormality” is often an initial description rather than a final diagnosis. A visible mark or warm swelling can prompt investigation, but appearance alone cannot show the deeper anatomy or determine that treatment is needed.
For brain and spinal lesions, NINDS describes seizures, headaches and possible neurological symptoms; spinal involvement can produce back or lower-limb pain. Some lesions are discovered incidentally. Symptoms and an imaging finding need interpretation together: a headache alone cannot locate or diagnose a malformation.
Flow, genes and inherited vascular syndromes
The VASCERN pathway describes peripheral lesions with warmth, pulsation, swelling, pain, ulcers, bleeding or impaired function. Substantial shunting can contribute to high-output heart failure. Organ-specific problems can include coughing blood or gastrointestinal bleeding. These are possible presentations, not a checklist that diagnoses AVM or establishes the cause of every such symptom.
The older MedlinePlus CM–AVM description explains the RASA1-related inherited example. It is not a complete current gene list. The ISSVA glossary distinguishes RASA1-related CM–AVM1 from EPHB4-related CM–AVM2. Genetic counselling can clarify a result; a familiar birthmark does not identify a gene on its own.
Hereditary haemorrhagic telangiectasia (HHT) is a distinct inherited disorder with multisite vessel abnormalities. Recurrent nosebleeds and vascular lesions can be relevant clues, but nosebleeds are common and do not diagnose HHT. A confirmed syndrome can change the family and organ-assessment questions; those decisions need a syndrome-specific clinical plan.
The 2026 JCI review also discusses somatic variants: changes present in affected tissue rather than necessarily inherited throughout the body. Molecular findings can inform research without proving that a matching medicine is effective or safe.
Treatment: observation, surgery, embolization and radiosurgery
For neurological lesions, NINDS education describes surgery, catheter embolization and focused radiosurgery. Each has different aims and limitations. Embolization may be part of staged care; radiosurgery has a delayed effect, and bleeding risk may persist while vessels close. Smaller access incisions do not eliminate neurological or radiation injury risks. No modality is automatically preferable for every location.
A treatment discussion should compare the proposed intervention with the actual alternative, including surveillance when appropriate. Ask what is being treated now, what can remain afterwards, and whether more than one stage is expected. Separate a technical endpoint from the outcome that matters to you. Ask how recurrence or a residual lesion will be assessed, and how uncertainty changes the recommendation.
The urgency and aim of intervention after a haemorrhage differ from a planned decision about an unruptured lesion. Likewise, treating an organ complication and treating the entire malformation are not necessarily the same task. Request a clear explanation of the sequence and who coordinates it; this guide supplies no personal timing or technique prescription.
Supplements and targeted medicines: what is not established?
No supplement replacement for AVM assessment or treatment is established by the sources reviewed. A product advertised for “vascular health” needs evidence in the relevant human lesion, not only a general blood-flow claim. Do not treat improvement in an unrelated laboratory marker as proof that an AVM has closed or become safer.
The dated NCCIH supplement-safety guidance warns about medicine interactions and surgical concerns. Tell the team what you actually take, including herbs, powders and nonprescription products. A surgeon or anaesthetist should supply any procedure-specific changes. This is a disclosure precaution, not an instruction to stop all supplements yourself.
For an experimental medicine, ask whether its use is approved for this condition locally, part of a trial or an individualized off-label decision. Request the human outcome and adverse-event evidence. A medicine’s established role in another disease does not establish benefit in an AVM. The review here does not select a targeted drug or give a dose.
Practical support during surveillance and after bleeding
Observation should have a written purpose: what will be monitored, how appointments are arranged, and which changes require an earlier call. Keep the specialist’s contact details accessible. Ask how travel, pregnancy planning, work and activities affect your particular plan, rather than adopting a blanket restriction from another patient’s experience.
After bleeding on the brain, the NHS recovery guidance describes support for mobility, communication, cognition and mood. Rehabilitation needs vary and recovery can take time. Family members may need help understanding fatigue or cognitive changes. Ask which rehabilitation service is involved and how new deterioration differs from an expected recovery difficulty.
For a visible or painful lesion, tell the team about clothing, footwear, work, sleep and embarrassment as well as any change in size. Those effects belong in the care discussion. The useful outcome may include safer daily function or symptom relief; it should not be reduced to whether a photograph looks different.
Safety: sudden headache, neurological changes and bleeding
A sudden extremely painful headache that does not go away, especially with confusion, collapse, seizure or neck stiffness, requires emergency assessment. NHS subarachnoid-haemorrhage guidance. Do not drive yourself or wait for a routine AVM appointment. Use local emergency services outside the UK; knowing about an AVM does not establish the cause at home.
Sudden face or arm weakness, speech difficulty, loss of vision or other stroke symptoms need emergency help even if they resolve. NHS stroke-symptom guidance. A known lesion does not make a new neurological deficit an expected symptom to watch without assessment.
Ask your own team about urgent warnings specific to the lesion’s site, including significant bleeding or rapid clinical deterioration. If an acute problem is severe, seek emergency care rather than trying to classify the malformation yourself. Bring a medication list and tell emergency staff about the known diagnosis and previous procedures when possible.
Medicines, procedures, pregnancy and coexisting disease
Anticoagulants are not routine treatment for every AVM. If prescribed for a separate indication, bleeding precautions matter, as described in NHS anticoagulant guidance. Do not independently stop a necessary medicine or start aspirin because the lesion involves blood vessels. Ask the prescriber and AVM team to agree on the indication and procedure plan.
When contrast or a procedure is proposed, discuss kidney function, allergies, prior reactions and the full medicine list. NHS acute kidney injury information explains why renal illness can change medicine planning. Tell the team about a possible pregnancy and other current illnesses; an imaging or anaesthesia plan should reflect those circumstances.
Ask which clinician will provide instructions before and after each stage. The instructions should be specific to the actual medicines and procedure, including what to do if a new symptom occurs. An online protocol from a different lesion or country is not an adequate substitute.
Diagnosis: location, imaging and the possibility of a syndrome
The MedlinePlus summary describes history, examination and selected imaging such as ultrasound, CT, MRI or angiography. Testing depends on the location and clinical question. Ask what information the next test adds and whether it changes treatment selection. No single visible feature or home measurement supplies a complete diagnosis.
A family pattern, multiple lesions or features of an inherited disorder should be described to the team. Bring earlier reports if available and ask whether genetics or other organ specialists should be involved. A negative or uncertain genetic result needs interpretation; it should not be treated as a universal guarantee that family or organ risks are absent.
The distinction between a malformation and a tumour should be resolved before borrowing a treatment intended for haemangioma or another anomaly. Record the diagnostic name, important imaging findings and the team’s remaining uncertainty. That makes future consultations more useful than repeating only that a scan was “abnormal.”
Deciding on a plan and keeping follow-up meaningful
A useful consultation ends with a shared account of the lesion, current concerns and available choices. Ask the specialist to explain the reasons for recommending intervention or surveillance, the evidence population and the consequences of delaying a decision. Where uncertainty is substantial, a multidisciplinary review or second specialist opinion can help clarify the trade-off.
For a procedure, ask about neurological or organ injury, bleeding, residual disease, repeat treatment and follow-up. Ask which risks are measured in comparable patients and which are inferred. The hospital’s experience is relevant, but a small selected series cannot supply a precise personal probability. Request an explanation of how complications will be recognized and managed.
For surveillance, agree on who reviews results and what happens if you move or miss an appointment. Keep prior images available when transferring care. A stable report needs interpretation in the context of symptoms and function, rather than a conclusion that future review is unnecessary.
Animal and molecular research: promise without a proven cure
The May 2026 JCI review describes molecular pathways, patient-tissue research and animal models informing possible precision therapies. Such work can identify hypotheses but does not establish safe human treatment or a universal cure. Resected tissue and models have selection and translation limits; a proposed diagnostic biopsy or pathway-targeting drug needs its own clinical evidence.
This guide excludes laboratory, genetic-mechanism and manufacturer-funded efficacy claims from its independent treatment verdict. A patent, plausible biological pathway or published case is not a randomized demonstration of durable benefit. A trial invitation should explain whether results exist, the comparison, outcome measures, foreseeable harms and funding.
Funding and source roles
Who paid for the evidence?
Follow named sources to the funding and relationships disclosed in this article. Numbered article disclosures preserve notes where a source is not identified. A public or university name alone does not establish independence.
View 19 more funding disclosures
This graphic reorganizes the article's disclosures; it is not a new financial audit or independence classification. Highlighted notes show different funding relationships where available. Review funding is separate from underlying trial funding. Disclosure is not proof of falsehood, and no declared conflict is not proof of complete independence.
AVMs have no single owner; devices, medicines and procedures have commercial and provider interests. The original ARUBA disclosure bundle identifies public funding, later private support and author industry fees; no commercial study funding is declared, but the wider private chain remains unresolved. ISSVA’s congress supporters establish an institutional industry route, not proof its taxonomy was purchased. Public agency budgets and network funding are context. Source roles below keep clinical findings, expert guidance, institutional disclosures and emerging mechanisms separate; no maker-funded efficacy selects a winner.
Tier describes financial proximity; A–D describes credibility for the stated source role. Neither is a clinical certainty grade. Unknown finances remain unknown. Manufacturer- and sponsor-funded efficacy is excluded from the independent verdict; attributed clinical guidance is identified as guidance.
| Source | Funding / backers | Country / jurisdiction | Independence | Credibility / incentives / gaps |
|---|---|---|---|---|
| MedlinePlus: AVM summary, December 2023 | NLM/NIH public information; congressional budget route and welcomed donations checked. Page-specific gifts, author and cited-study chains unresolved. | United States; NLM Bethesda, Maryland | Tier 1 provisional for education | B for public educational role; agency mission/budget incentives and untraced underlying studies. |
| NINDS: original AVM education | Federal NIH institute; current fiscal/grant-policy route checked separately. Page writers, gifts and underlying device-study chains unresolved. | United States; federal NIH/NINDS jurisdiction | Tier 1 provisional for educational role | B for bounded procedure context; C for comparative efficacy. Simplified/datable guidance and incomplete study finances. |
| ISSVA: original 2025 classification and glossary | Society’s 2026 congress names pharmaceutical/device sponsors and exhibitors. Classification-specific receipts and individual author chains unresolved. | United States; Milwaukee society contact, international authors | Tier 2 provisional — institutional industry route | B for terminology; C for treatment inference. Specialty interests; sponsorship is not proof the taxonomy was purchased. |
| VASCERN VASCA: original February 2026 AVM pathway | Network reports EU4Health and French-government support; member institutions also receive national/regional funds. Individual author/employer and document-specific extra payer chains unreported. | European network; Paris, France coordination; international clinical authors | Tier 2 provisional — author and exact project finances unresolved | C for attributed expert guidance; specialty/referral and public-program interests. No independent comparative efficacy clearance. |
| ARUBA: original published 2020 extended-follow-up abstract | NINDS grants and later Vital Projects Fund gift documented in original manuscript bundle. An author reports industry-paid institutional fees; private donor chain incomplete. | International 39-centre trial; US NINDS/Columbia and New York private fund | Tier 3 — commercially connected author; mixed trial support | C provisional for attributed efficacy; randomized design considered separately. Early stopping, selected adults, varied interventions and finite follow-up. |
| ARUBA: original manuscript and disclosure bundle | NINDS U01NS051483/U01NS051566 (2007–13); Vital Projects Fund gift (2014–15). Van der Worp reports institutional fees from Boehringer Ingelheim, Bayer and LivaNova. | United States public/private funding; international authors and employers | Tier 3 — commercially connected author; mixed trial support | B for actual disclosures; no commercial study funding declared. Public funder helped design/interpretation/writing; private fund no listed role. Donor/employer chains unresolved. |
| JCI: original May 2026 precision-therapy review | NIH grants and several charities/private awards. Hale/Kahle AVM diagnostic patent; other authors disclose pharma consulting/grants and Prime Medicine financial interest. Complete donor chains unresolved. | United States and South Africa author affiliations; international corporate/charity backers | Tier 3 — materially connected authors, patents and commercial interests | C for emerging mechanisms; academic/IP incentives, selected tissue/model evidence and no cleared drug comparison. |
| MedlinePlus Genetics: CM–AVM syndrome, August 2011 | NLM/NIH public information; congressional budget route and welcomed donations checked. Page-specific gifts, author and cited-study chains unresolved. | United States; NLM Bethesda, Maryland | Tier 1 provisional for education | C for dated genetics; B for bounded description. Public accountability; simplification and unresolved underlying finances. |
| MedlinePlus Genetics: HHT, January 2020 | NLM/NIH public information; congressional budget route and welcomed donations checked. Page-specific gifts, author and cited-study chains unresolved. | United States; NLM Bethesda, Maryland | Tier 1 provisional for education | C for dated genetics; B for bounded description. Public accountability; simplification and unresolved underlying finances. |
| NHS: subarachnoid haemorrhage, September 2025 | DHSC-funded NHS website; its policy rejects advertising and corporate sponsorship. Page-author disclosures and all underlying trial finances not supplied. | United Kingdom; England public patient information | Tier 1 provisional for educational role | B — clinical sign-off and public-service accountability; policy is not an audit of underlying trials. |
| NHS: stroke symptoms, September 2024 | DHSC-funded NHS website; its policy rejects advertising and corporate sponsorship. Page-author disclosures and all underlying trial finances not supplied. | United Kingdom; England public patient information | Tier 1 provisional for educational role | B — clinical sign-off and public-service accountability; policy is not an audit of underlying trials. |
| NHS: acute kidney injury, March 2026 | DHSC-funded NHS website; its policy rejects advertising and corporate sponsorship. Page-author disclosures and all underlying trial finances not supplied. | United Kingdom; England public patient information | Tier 1 provisional for educational role | B — clinical sign-off and public-service accountability; policy is not an audit of underlying trials. |
| NHS: anticoagulant side effects, September 2024 | DHSC-funded NHS website; its policy rejects advertising and corporate sponsorship. Page-author disclosures and all underlying trial finances not supplied. | United Kingdom; England public patient information | Tier 1 provisional for educational role | B — clinical sign-off and public-service accountability; policy is not an audit of underlying trials. |
| NCCIH: supplement safety (January 2019) | NIH federal education; page-specific allocation, staff interests and underlying trials not fully cleared. | United States; NIH/NCCIH, Bethesda | Tier 1 provisional for safety context | C — dated public education; no condition-specific product efficacy assessment. |
| ISSVA: actual 2026 congress support | Palvella platinum; Novartis and Relay gold; OnKure bronze. Other named exhibitors include medical-product firms. No classification allocation provided. | United States society; global commercial supporters | Tier 3 for congress self-disclosure | B for named supporters; promotion/relationship incentives and unresolved full receipts. |
| ISSVA: current contact | Society contact record; membership, congress and complete financial receipts not audited. | United States; Milwaukee, Wisconsin | Tier 3 for institutional identity | B for current own address; not a financial audit or clinical endorsement. |
| NINDS: actual current fiscal-policy route | NIH institute funding/grant-management policy; appropriations and application decisions distinguished. Private receipts and individual trial allocation not fully audited. | United States; public NIH/NINDS grant jurisdiction | Tier 1 for fiscal provenance | B — primary governmental policy; budget/mission interests and incomplete project ledger. |
| VASCERN: own funding and partnerships | Names EU4Health, French-government and member national/regional contributions. Complete accounts, private receipts and document allocations not audited. | France coordination; European/national funding jurisdictions | Tier 3 for institutional self-description | B for named routes; self-report and unresolved full receipts/allocations. |
| European Commission: ERN funding route | EU4Health finances ERN coordination and auxiliary activities; specific pathway allocation and other provider resources not itemized. | European Union; Commission/HaDEA public programme jurisdiction | Tier 1 provisional for governmental context | B for named programme; budget/policy incentives and allocation gaps. |
| NLM: institutional identity and donations | Federal NIH library welcomes donations/bequests; Friends coalition includes corporations/foundations. A coalition membership is not a proven page donation. | United States; Bethesda, Maryland | Tier 1 provisional for institutional context | B — own explicit routes; self-report and no donor/page ledger. |
| NLM: congressional justification index | Federal budget publications; 2026 consolidation described as proposed. Budget requests are not enacted appropriations. | United States; federal NIH/NLM budget process | Tier 1 for public budget provenance | B — accountable institutional records; political/budget interests and no page-specific grant allocation. |
| NHS website: content and funding policy | DHSC funding; website states no advertising or corporate sponsorship. Full staff disclosure register not retrieved. | United Kingdom; NHS England website | Tier 1 provisional for institution | B — explicit editorial safeguards; institutional self-report does not clear every cited trial. |
Frequently asked questions
Is every vascular birthmark an AVM?
No. Different malformations and vascular tumours have distinct anatomy and treatment questions. A specialist diagnosis matters.
Does ARUBA mean all AVMs should be left untreated?
No. Its comparison concerns selected adults with unruptured brain lesions and finite follow-up. Other populations and circumstances need separate assessment.
Can radiosurgery remove bleeding risk immediately?
Its effect is delayed. Follow-up and a warning-sign plan remain important while the clinical team assesses the result.
Does an AVM always run in families?
No. Inherited syndromes are one part of the spectrum. Discuss the actual diagnosis and family features with the team.
Are molecular medicines or supplements an established cure?
No universal cure or supplement replacement is established here. Experimental or off-label treatment needs condition-specific human evidence and safety review.
Which symptoms need emergency care?
A thunderclap headache or sudden stroke symptoms require emergency assessment, including when neurological symptoms resolve. Significant bleeding or severe deterioration also needs urgent attention.
Sources and funding notes
Full official ISSVA taxonomy/glossary and ten-page VASCERN pathway were opened; only brief terminology is summarized. ARUBA’s published original abstract and a separate original submission/disclosure bundle were read; the mirror bundle is not the publisher final article. NINDS education was available in official indexed form, with direct access limitations retained. The full JCI publisher text and financial statements, NHS/MedlinePlus bodies and each named institutional funding route were checked. Dated genetic education is not treated as a complete current gene catalogue. No pooled rupture probability, universal screening schedule, device ranking, medicine dose or personal intervention threshold is supplied.
- MedlinePlus: AVM summary, December 2023 — Definition and selected investigations; no individual rupture probability.
- NINDS: original AVM education — Attributed neurological treatment and delayed radiosurgery effect; no universal modality winner.
- ISSVA: original 2025 classification and glossary — Fast-flow AVM, fistula and other lesion distinctions only.
- VASCERN VASCA: original February 2026 AVM pathway — Site-specific specialist assessment and surveillance; attributed expert opinion, not a validated treatment comparison.
- ARUBA: original published 2020 extended-follow-up abstract — Medical-management finding applies to trial population; not every AVM or lifetime outcome.
- ARUBA: original manuscript and disclosure bundle — Funding and methods provenance; not independent replication.
- JCI: original May 2026 precision-therapy review — Molecular hypotheses and laboratory limits only; no targeted-drug recommendation.
- MedlinePlus Genetics: CM–AVM syndrome, August 2011 — Dated RASA1/family example; not a complete current gene list.
- MedlinePlus Genetics: HHT, January 2020 — Inherited multisite vascular disease and nosebleed context; not a personal screening protocol.
- NHS: subarachnoid haemorrhage, September 2025 — Thunderclap headache emergency and recovery support; aneurysm clipping/coiling not transferred to AVMs.
- NHS: stroke symptoms, September 2024 — Sudden neurological symptoms and emergency care even if symptoms resolve.
- NHS: acute kidney injury, March 2026 — Renal/contrast and medicine planning only.
- NHS: anticoagulant side effects, September 2024 — Bleeding precautions only for actually prescribed anticoagulants; not routine AVM therapy.
- NCCIH: supplement safety (January 2019) — Disclosure and interaction precautions only.
- ISSVA: actual 2026 congress support — Institutional commercial route only; not proof of a specific clinical recommendation.
- ISSVA: current contact — Headquarters/country trace only.
- NINDS: actual current fiscal-policy route — Public funder identity, not clearance of all ARUBA backers.
- VASCERN: own funding and partnerships — Network funding only; not author financial clearance.
- European Commission: ERN funding route — Public network-finance corroboration, not treatment evidence.
- NLM: institutional identity and donations — NLM-specific gift route and headquarters.
- NLM: congressional justification index — Public finance route; no inferred reorganization or trial independence.
- NHS website: content and funding policy — Website funding and editorial safeguards only.
Last reviewed: October 4, 2026. Educational information; no personal diagnosis, medication dose or supplement regimen is supplied. Local approval, product labels and clinical circumstances may differ.
Have a question — or want us to cover something?
Ask about anything on this page, or request the next deep dive: an ingredient, a supplement, or a health concern. We use published research, evidence syntheses, and regulatory guidance, with clear source links.
One daily research roundup
Get the topics, key findings and links from our new articles in one email. At most one digest a day, only when there is something new.
