Rendu–Osler–Weber Disease (Hereditary Haemorrhagic Telangiectasia)

Autosomal dominant vascular dysplasia of the BMP9/BMP10–ALK1–endoglin–SMAD4 pathway, with recurrent epistaxis, mucocutaneous telangiectases and pulmonary, cerebral, hepatic and gastrointestinal arteriovenous malformations; includes ENG- and ACVRL1-related forms, SMAD4-related combined juvenile polyposis–hereditary haemorrhagic telangiectasia syndrome (with thoracic aortopathy) and GDF2-related vascular anomaly.

Literature and regulatory information updated: 6 October 2026.

0. Initial Reference Table

1. Introduction

Hereditary haemorrhagic telangiectasia is a systemic vascular dysplasia in which capillary beds are replaced by direct connections between arteries and veins [1, 26]. Small-calibre malformations are called telangiectases, lesions of a few millimetres formed by focal dilatations of postcapillary venules, with an endothelial-only wall and no elastic tissue; larger malformations, which may measure several centimetres and are located mainly in the lungs, liver and brain, are called arteriovenous malformations (AVMs) [1, 12, 26]. The clinical picture results from two distinct consequences of these lesions: bleeding, mainly nasal and gastrointestinal, with chronic iron-deficiency anaemia, and shunting through AVMs, which in the lung allows emboli, air and bacteria to enter the systemic circulation and in the liver creates high-cardiac-output states [1, 16].

The first description is attributed to Sutton in 1864, but the eponym commemorates Osler in the United States, Parkes Weber in the United Kingdom and Rendu in France, so French-language literature uses the order Rendu–Osler–Weber and English-language literature Osler–Weber–Rendu [1]. Smith's textbook dates recognition as a distinct entity to Osler's 1901 communication, entitled On a family form of recurring epistaxis, associated with multiple telangiectases of skin and mucous membrane, and the name “hereditary haemorrhagic telangiectasia” was proposed by one of Osler's house officers [1, 26]. OMIM clarifies the chronology: Osler (1849–1919) found only Rendu's 1896 report in the literature; Parkes Weber described cases later through his interest in angiomas; and Hanes, a house officer at Johns Hopkins Hospital, published a comprehensive discussion in 1909 with colour illustrations of lip, tongue and facial lesions and named the disease hereditary hemorrhagic telangiectasia — a date differing from the 1912 given by GeneReviews [1, 4]. The same text notes that the eponymous sequence, although not chronological, is justified by each author's contribution to nosology, and that, strictly, an individual lesion is a telangiectasis (telangiectasis) and the process telangiectasia [4].

The molecular era began in 1994, when McAllister and colleagues identified endoglin, an endothelial transforming growth factor beta (TGF-β)-binding protein, as the gene underlying HHT type 1 [27]. In 1996, Johnson and colleagues described variants in the activin receptor-like kinase 1 gene (ACVRL1, ALK1) in HHT type 2 [28]. In 2004, Gallione and colleagues showed that families with combined juvenile polyposis and HHT had SMAD4 (then MADH4) variants and none in ENG or ACVRL1, defining the combined syndrome [29]. In 2013, Wooderchak-Donahue and colleagues described variants in the bone morphogenetic protein 9 gene (BMP9, encoded by GDF2) in a vascular anomaly syndrome overlapping phenotypically with HHT [30]. Clinical standardisation arrived in 2000 with the Curaçao criteria, the first international guideline was published in 2011 and the second, in 2020, added recommendations on anaemia, paediatrics and pregnancy [15, 16, 21].

The disease is underdiagnosed and the delay between first symptoms and diagnosis is long: in the American CHORUS registry, most patients had typical manifestations by age 13, but 63% were diagnosed only in middle or late adulthood, with a mean interval exceeding two decades [16, 31]. This delay matters because the most serious complications — ischaemic stroke and brain abscess from pulmonary AVMs, intracranial haemorrhage from cerebral AVMs and pulmonary haemorrhage during pregnancy — are largely preventable through screening and treatment [1, 16, 31].

This article addresses the four genetic forms of HHT and incorporates SMAD4-associated thoracic aortopathy. Juvenile polyposis, gastrointestinal cancer risk and the corresponding endoscopic surveillance of SMAD4 carriers, as well as heritable pulmonary arterial hypertension, are addressed only in the aspects specific to HHT.

2. Molecular Genetics and Pathophysiology

2.1. The BMP9/BMP10–ALK1–endoglin–SMAD4 pathway

All known HHT genetic defects affect TGF-β superfamily proteins [1]. ACVRL1 encodes activin receptor-like kinase 1 (ALK1), a type I serine–threonine kinase receptor predominantly expressed in endothelium, whose ligands are BMP9 (the GDF2 product) and BMP10; ligand binding assembles two type II and two type I receptors, type II receptors phosphorylate ALK1 and this activates SMAD transcriptional mediators [1, 6]. ENG encodes endoglin, a homodimeric transmembrane glycoprotein of vascular endothelium acting as a complex coreceptor, required for BMP9 signalling through SMAD1 and endothelial responses to blood flow [1, 6]. BMP9 is a circulating angiogenesis inhibitor signalling exclusively through ALK1 with endoglin as coreceptor — the same axis whose disruption by ACVRL1 or ENG variants causes HHT, explaining the phenotypic overlap of GDF2-related disease [6]. SMAD4 encodes a 552-amino-acid intracellular protein, the common mediator (co-SMAD) of TGF-β/BMP signal transduction [1, 6]. The Oxford Desk Reference summarises this axis as the TGF-β superfamily cell-surface receptor pathway, including BMP9, mediating vascular remodelling through effects on extracellular matrix production [12].

The disease mechanism is loss of function: essentially all known pathogenic variants are null alleles, and ClinGen dosage sensitivity curation assigns ACVRL1 and SMAD4 a haploinsufficiency score of 3 (sufficient evidence) [1, 6]. Endothelial endoglin or ALK1 haploinsufficiency impairs signalling that normally restrains endothelial activation and directs flow-dependent vascular remodelling, and capillary beds are replaced by direct arteriovenous connections — telangiectases on a small scale and AVMs on a large scale [6]. Because these lesions have thin walls, tortuous courses, proximity to mucosal surfaces and no contractile elements, they rupture with minimal trauma and bleed rapidly and uncontrollably [1].

Hepatic disease illustrates shunt pathophysiology. Three types of intrahepatic communication, often simultaneous — hepatic artery–portal vein, hepatic artery–hepatic vein and portal vein–hepatic vein — produce overlapping presentations of high-output cardiac failure, portal hypertension, encephalopathy, biliary ischaemia and mesenteric ischaemia [16]. Blood diversion through lesions is associated with hepatocellular regenerative activity; focal nodular hyperplasia is approximately one hundred times more prevalent than in the general population, and hepatic nodularity interpreted on imaging as cirrhosis usually represents nodular regenerative hyperplasia — HHT does not cause hepatocellular failure [1, 16].

2.2. Genes, proportions and loci without an identified gene

Sources: GeneReviews (Tables 1 and A), McDonald and colleagues (2020), OMIM, GeneGuidance and Smith's textbook (GDF2 proportion, below 1% of patients) [1, 3, 5, 6, 11, 26].

The table's proportions derive from a referral-centre cohort in which probands strictly met Curaçao criteria: ENG variants occurred in 67 of 152 (44.1%; 95% CI 36.0–52.4), ACVRL1 variants in 79 of 152 (52.0%; 95% CI 43.7–60.1) and SMAD4 variants in 2 of 152 (1.3%), with only 4 of 152 (2.6%) remaining without a variant [11]. Older sources give lower yields — approximately 85% in the Oxford Desk Reference and approximately 90% in Smith's textbook — a difference the cohort authors attribute to heterogeneity in the experience of clinicians submitting samples to laboratories [11, 12, 26]. The 2020 guideline reports 97% detection in patients with a definite clinical diagnosis [16]. Smith's textbook attributes 1–2% of cases to SMAD4 and Orphanet 1–3%; the guideline mentions approximately 3% of patients with HHT and an SMAD4 variant [2, 16, 26].

One locus without an identified gene remains, within a 7 Mb region at 7p14 (HHT type 4), mapped by linkage analysis (maximum lod 3.60) in a single family with eight affected and four suspected individuals, in whom epistaxis and telangiectases were less frequent and less numerous than in other forms and pulmonary and cerebral AVMs occurred; candidate genes BMPER, CCM2, RALA and INHBA were excluded, and the family remained without a causal variant following genome sequencing and deletion/duplication analysis performed up to 2024 [1, 5]. The HHT type 3 locus, mapped in 2005 to a 5.4 cM interval at 5q31.3–q32 in a four-generation pedigree with pulmonary involvement, lost its basis: in 2024, genome sequencing in the 100,000 Genomes project found a pathogenic frameshift ENG variant in this family, and at least one person meeting three Curaçao criteria was unaffected [4, 33]. The family was reclassified as HHT type 1 [4]. Smith's textbook also mentions RASA1 variants rarely described in patients with suspected HHT, now included in the differential diagnosis of capillary malformation–arteriovenous malformation syndrome [1, 26].

2.3. Gene–phenotype correlations

Pulmonary and cerebral AVMs are more frequent with ENG variants, and hepatic AVMs with ACVRL1 variants; nevertheless, both genes cause multisystem vascular dysplasia and most manifestations occur with either, so the causal gene cannot be reliably predicted clinically [1]. In the French–Italian network (93 patients with HHT type 1 and 250 with HHT type 2), symptomatic pulmonary AVMs were more frequent in HHT type 1 (34.4% versus 5.2%), as were brain abscesses (7.5% versus 0.8%) and screening-detected asymptomatic pulmonary AVMs (54% versus 12.8%); gastrointestinal bleeding was more frequent in HHT type 2 (16.4% versus 6.5%), epistaxis began later with incomplete penetrance in HHT type 2, and symptomatic hepatic involvement occurred only in HHT type 2 [34]. The international guideline places pulmonary AVMs at approximately 50% in HHT type 1 and 10% in HHT type 2, and cerebral vascular malformations at 8–16% and 1–2%, respectively; hepatic AVMs present more severely in HHT type 2 and an ACVRL1 variant predicts progression [16]. In an Italian series of 142 genetically diagnosed patients published in 2026, age was independently associated with hepatic AVMs and ENG variants with lower hepatic AVM likelihood, while age, ENG and truncating variant status independently predicted cerebral AVMs [35].

Pulmonary hypertension without severe systemic shunting, which is rare, occurs mainly in carriers of ACVRL1 variants; the Oxford Desk Reference describes it as a rare complication in some families with ACVRL1 variants and recalls that most familial forms of pulmonary arterial hypertension are due to BMPR2 variants, another receptor of the same superfamily [1, 12]. The GeneReviews chapter on heritable pulmonary arterial hypertension includes ACVRL1, ENG and GDF2 among the twelve genes with definitive evidence, each accounting for fewer than 1% of cases in the PAH Biobank (2,572 patients), with “PAH-HHT”, isolated PAH and PAH with AVM phenotypes [36].

There is no absolute correlation between specific variants or variant types and phenotype [1]. In the French–Italian series, truncating ACVRL1 variants were associated with more frequent epistaxis and telangiectases [34]. Pulmonary and hepatic AVMs cluster in HHT type 1 and type 2 families, respectively [34].

2.4. SMAD4: combined juvenile polyposis–HHT syndrome and aortopathy

Loss-of-function SMAD4 variants cause a distinct syndrome combining juvenile polyposis and HHT features [1, 29]. Historically, carriers were identified in gastroenterology or hereditary cancer clinics without directed vascular assessment, and series suggested that only a minority had HHT signs — the Oxford Desk Reference reports 15–22% [1, 37]. With systematic vascular assessment, approximately 95% exhibit HHT signs and 75–80% meet diagnostic criteria, so any pathogenic SMAD4 variant should be considered to cause the combined syndrome, with risk from both components [1]. Evidence suggests pulmonary, cerebral and hepatic AVMs may be more prevalent in SMAD4-related disease than in ENG- or ACVRL1-related forms [1].

Thoracic aortopathy completes the SMAD4-associated phenotype. The heritable thoracic aortic disease chapter of Clinical Cardiogenetics divides the genes of heritable thoracic aortic disease into those encoding extracellular matrix components and those affecting the ability to remodel the aortic wall in response to mechanical load, and includes SMAD4 among TGF-β pathway genes alongside TGFBR1, TGFBR2, TGFB2, TGFB3, SMAD2 and SMAD3 [32]. In the table ordering each gene's entities from most syndromic to non-syndromic, SMAD4 is associated with juvenile polyposis syndrome, HHT, Myhre syndrome and isolated heritable thoracic aortic disease [32]. Several studies showed thoracic aortopathy prevalence of 31–38% in the combined syndrome, and one found some connective tissue disease manifestation in two-thirds of patients — mainly joint laxity, followed by micrognathia and chest deformities [1]. The juvenile polyposis GeneReviews chapter adds aortic root dilatation, aortic aneurysm and dissection and mitral valve dysfunction, and reports of retinitis pigmentosa, retinal detachment, joint laxity and marfanoid habitus whose place within the phenotype remains unestablished [38]. The Oxford Desk Reference therefore recommends screening SMAD4 carriers for aortic root dilatation alongside juvenile polyp screening [12].

2.5. Allelic disorders and related entities

No allelic phenotypes of germline ACVRL1 or ENG variants are known beyond those described in this article [1]. ENG variants were identified in two young patients with juvenile polyposis without an SMAD4 or BMPR1A variant; neither yet had signs of HHT, but neither had reached the usual age of manifestation [38]. In SMAD4, heterozygous gain-of-function variants affecting amino acids 496 or 500 cause Myhre syndrome, a progressive multisystem connective tissue disorder, through a gain-of-function or dominant-negative mechanism opposite to that of HHT [1, 6, 38]. Smith's textbook places the first molecular description in 2011, with missense variants at the same Ile500 codon in 11 patients, and notes that only variants at codons 496 and 500 of the MH2 domain have been identified since; the picture combines progressive thickening of the skin and serosae, muscular hypertrophy, joint limitation and cardiovascular disease [39]. The bone dysplasia atlas describes Myhre syndrome as autosomal dominant due to de novo variants, with a tendency towards advanced paternal age, and attributes LAPS syndrome (laryngotracheal stenosis, arthropathy, prognathism and short stature) to the same gene [40]. GDF2 has a recessive form: three homozygotes have been described — one with pulmonary arterial hypertension and linear facial telangiectases, one with multiple pulmonary AVMs and one with diffuse small-calibre pulmonary AVMs and epistaxis — and biallelic loss abolishes circulating BMP9 and BMP10 [1, 6].

2.6. Variant nomenclature

Nomenclature should use MANE Select transcripts: NM_000020.3 (ACVRL1, protein NP_000011.2, 503 amino acids), NM_001114753.3 (ENG, NP_001108225.1, 658 amino acids), NM_005359.6 (SMAD4, NP_005350.1, 552 amino acids) and NM_016204.4 (GDF2, NP_057288.1, 429 amino acids) [10]. For ENG, GeneGuidance and some literature use NM_000118, which is not MANE Select; positions in the two transcripts should not be manually converted but confirmed in VariantValidator [6, 10, 41]. The ClinGen HHT Variant Curation Expert Panel published ACVRL1 and ENG specifications for ACMG/AMP (American College of Medical Genetics and Genomics/Association for Molecular Pathology) standards on 20 March 2024 [1], and ClinVar records dozens of variants in both genes reviewed by this panel [42].

In juvenile polyposis, the Oxford Desk Reference describes an SMAD4 “hotspot” termed “1372–1375delACAG” in historical numbering, not converted here because correspondence could not be confirmed in a reference resource [37]. A variant of uncertain significance in ACVRL1, ENG or SMAD4 neither establishes nor excludes diagnosis [1].

3. Clinical Spectrum and Manifestations by System

3.1. Overall frequency of manifestations

The Oxford Desk Reference gives similar but broader ranges from older series: gastrointestinal AVMs 11–40%, pulmonary AVMs up to 50% depending on locus, asymptomatic hepatic AVMs 11–40%, central nervous system malformations 5–11%, cutaneous telangiectases 13–89% and nasal telangiectases above 90% [12]. Smith's textbook reports pulmonary arteriovenous fistulae in 30% and cerebral vascular abnormalities in 10%, adding cavernous angiomas, port-wine stains and duodenal ulcer as occasional findings [26].

The most recent data come from the prospective American CHORUS registry of 15 centres, with unselected, confirmed-diagnosis patients: in the first 600 participants (median age 53 years; 60% female), recurrent spontaneous epistaxis occurred in 95%, chronic gastrointestinal bleeding in 30% and heavy menstrual bleeding in 35% of postmenarchal women, with moderate-to-severe mucosal bleeding in 76% [31]. Iron deficiency or anaemia was diagnosed in 68%, 41% required intravenous iron and 25% transfusions; intracranial haemorrhage occurred in 3%, pulmonary haemorrhage in 2%, venous thromboembolism in 7%, arterial thromboembolism in 11%, heart failure in 7% and pulmonary hypertension in 7% [31]. These figures describe a specialist-centre cohort and should not be read as population incidences.

3.2. Epistaxis

Epistaxis is the most common manifestation and the one most often prompting medical consultation [1]. Mean onset is around age 12; one-third have epistaxis by age 10, approximately 80% by 20 and 90% before 30 [1]. Smith's textbook reports a mean of 12 years and 90% by age 21 [26]. Bleeding may be spontaneous or triggered by minimal trauma — blowing the nose, touching it while bathing, dry air — and frequency ranges from a few episodes yearly to several daily, lasting less than a minute to several hours; many patients, however, never have epistaxis severe enough to cause anaemia or prompt consultation [1]. Turbulent airflow dries mucosa and fissures telangiectases of the septum and inferior turbinate [12, 16]. Nocturnal epistaxis is particularly suggestive [12]. Some patients notice worsening after alcohol consumption [1].

3.3. Mucocutaneous telangiectases

Telangiectases occur mainly on the lips, tongue, face, fingers and oral, gastrointestinal and nasal mucosa; they also occur on conjunctivae, ears, trunk, extremities and nail beds, but those at other sites are not HHT-specific [1, 12, 26]. They usually appear after epistaxis — 5 to 20 years later according to the Oxford Desk Reference — although childhood onset occurs: 30% notice their first before age 20 and two-thirds before 40 [1, 12]. These macules range from pinpoint to pinhead size and in older people may become larger, elevated and supplied by several draining venules; even in adults, only 10 to 20 telangiectases are typically found on careful examination, and they are rarely numerous or uncountable at one site [1]. They differ from petechiae and angiomas by blanching under pressure and rapidly refilling; skin lesions bleed much less than nasal lesions [1]. Finger transillumination helps detect lesions invisible at the surface [26]. Facial telangiectases were relatively rare in the Afro-Caribbean population of the Netherlands Antilles because of skin pigmentation [14]. Figure 1 in Smith's textbook chapter shows small telangiectases of the labial mucosa [26].

3.4. Anaemia and iron deficiency

Iron-deficiency anaemia affects approximately half of patients, is severe in up to 25% and results from epistaxis or, particularly in older patients, gastrointestinal bleeding; underrecognised epistaxis causes iron deficiency more often than gastrointestinal bleeding [1, 16]. Anaemia is usually diagnosed in adulthood and rarely in children [16]. Beyond fatigue, exertional dyspnoea and palpitations, iron deficiency may cause pica, angular cheilitis, koilonychia, exercise intolerance, restless legs syndrome, hair loss, myalgia and attention impairment, even without anaemia; anaemia increases cardiac output and worsens hepatic AVM high-output states [16]. Bleeding complications are the most frequent cause of hospitalisation [16]. One series also described folate deficiency and haemolysis as additional anaemia causes [16]. Smith's textbook states that 10% never bleed and approximately one-third require hospitalisation for bleeding [26].

3.5. Pulmonary arteriovenous malformations

Pulmonary AVMs occur in approximately 50% of patients and create high-flow right-to-left shunts [1]. They may cause dyspnoea and hypoxaemia, but even in asymptomatic patients the main risk is passage of air, thrombi or bacteria through the AVM without pulmonary filtration, causing stroke, transient ischaemic attack and brain or other-organ abscess; less often they present with haemoptysis, haemothorax or intrapulmonary haemorrhage [1]. Migraine and polycythaemia are additional complications [1]. The British Thoracic Society statement estimates that more than one in four patients with pulmonary AVMs will have a paradoxical embolic stroke, abscess or myocardial infarction, and that life-threatening haemorrhage affects one in 100 women during pregnancy [44]. Pulmonary AVMs are mostly congenital but may enlarge over time [1]; severe complications rarely occur before mid-to-late adolescence, although symptomatic childhood lesions have been reported [12]. Severe paediatric complications occurred mainly in hypoxaemic children with saturation of 93–96% or lower [1]. Approximately 80–95% of people with a pulmonary AVM have HHT [1], and in a British series of 83 probands with pulmonary AVMs and genetically confirmed HHT, 24% had few or no other manifestations [45].

In Smith's textbook, neurological complications — peaking in the third decade — resulted from pulmonary arteriovenous fistulae in 60%, cerebral vascular malformations in 28%, spinal malformations in 8% and portosystemic encephalopathy in 3% [26]. The Oxford Desk Reference emphasises that numerically more neurological events arise from pulmonary than cerebral AVMs [12]. In a British population-based primary-care study (675 patients and 6,696 controls), HHT was associated with increased stroke risk (OR 1.8; 95% CI 1.2–2.6), brain abscess (OR 30.0; 95% CI 3.1–288) and migraine (OR 1.7) [46].

3.6. Hepatic vascular malformations

Hepatic vascular abnormalities occurred in 74% in a systematic computed tomography (CT) study and 41% in an ultrasound study; only a minority were symptomatic (8% in the CT study) [1]. Mean diagnostic age is approximately 48, with female predominance of 4.5 to 1, and lesions are usually diffuse and small, rarely discrete AVMs [16]. Although only 8–14% are symptomatic initially, prospective follow-up showed death incidence of 1.1% and morbidity of 3.6% per person-year; high-output cardiac failure and complicated portal hypertension occurred at comparable rates (1.4 and 1.2 per 100 person-years), and atrial fibrillation at 1.6 per 100 person-years in high-output states [16]. Encephalopathy, mesenteric angina and ischaemic cholangitis with bilomas or “hepatic disintegration” are rarer presentations [16]. One-third of patients with hepatic AVMs have anicteric cholestasis correlated with severity [16]. Hepatic AVMs may present with hepatomegaly, right upper quadrant pain, a pulsatile mass, thrill or bruit [26].

The French hepatic PNDS (2019) describes hepatic AVMs in 32–84% of patients with HHT type 2, with female predominance and symptoms mainly after 50 years of age, symptomatic in about 10%; three types of shunt coexist — arteriosystemic (hepatic artery to hepatic veins; more than 70% of symptomatic forms, with high cardiac output, then high-output heart failure and pulmonary hypertension that is usually postcapillary, and more atrial fibrillation), arterioportal (portal hypertension) and portosystemic (encephalopathy) — and all can cause mesenteric steal and, because the bile ducts depend solely on the hepatic artery, biliary ischaemia and necrosis (cholestasis, strictures, bilomas, cholangitis) [24]. Heterogeneous perfusion favours focal nodular hyperplasia, with a prevalence of 2.9% in HHT, about 100 times that of the general population (0.03%), and nodular regenerative hyperplasia; diagnosis is by imaging and biopsy should be avoided because of the bleeding risk [24]. Biochemically, transaminases are usually normal and GGT raised; hepatocellular failure outside cardiac decompensation requires another cause to be sought [24]. In a series of 154 patients followed for a median of 44 months there were 8 deaths (5.2%), 2 among the 13 symptomatic and 6 among the 141 asymptomatic patients, all with Buscarini grade 4 AVMs [24]. The 2024 PNDS summarises the overall frequency: pulmonary AVMs in about 50% and hepatic AVMs in 40–70%, symptomatic in fewer than 10% [23].

Table 3.6a — Doppler ultrasound grading of hepatic AVMs (Buscarini), according to the French hepatic PNDS [24]

Doppler ultrasound has a sensitivity of 86–97% and a specificity of 100% for hepatic AVMs, with no studies defining its positive predictive value; in screening, prevalence was 32–72% on ultrasound and 67–78% on CT; CT and magnetic resonance imaging are reserved for when invasive treatment is being considered (bevacizumab, transplantation), and echocardiography with cardiac output and cardiac index (normal < 4.0 L/min/m²) is warranted in any hepatic involvement [24].

3.7. Cerebral and spinal vascular malformations

Cerebral AVMs are usually present at birth and occur in approximately 10% of patients [1]. “Cerebral vascular malformation” encompasses three principal HHT lesions with different haemorrhage risks: AVM with a nidus, cerebral arteriovenous fistula and capillary vascular malformation; spontaneous risk is very low in capillary malformations, intermediate in AVMs with a nidus and high in fistulae, which are overrepresented in children [16]. They may bleed without preceding symptoms, including before age 12 months [1, 16]. An American paediatric group estimates childhood rupture risk at approximately 0.7% per lesion per year, additive across lesions and over life [47]. Spinal vascular malformations have approximately 0.5% prevalence, are generally perimedullary and thoracic and may present with paralysis [1, 16].

3.8. Digestive tract

Gastrointestinal telangiectases are common in adults, predominantly in the stomach and duodenum; approximately one-quarter eventually experience gastrointestinal bleeding, usually after age 50, presenting mainly as iron-deficiency anaemia rather than acute haemorrhage [1]. The guideline places gastrointestinal bleeding at approximately 30%, in the fifth and sixth decades; symptomatic patients have gastric telangiectases in 46–75%, small-bowel telangiectases in 56–91% and colonic lesions in up to 30% [16]. No food, activity or medicine has been identified as a trigger [1]. Patients without juvenile polyposis have a colorectal cancer risk similar to the general population [16].

3.9. Pulmonary hypertension and cardiac involvement

Pulmonary hypertension may result from increased cardiac output due to shunting through hepatic AVMs (less often from severe anaemia) or be clinically and histologically indistinguishable from idiopathic pulmonary arterial hypertension [1]. Right heart catheterisation distinguishes precapillary pulmonary arterial hypertension, with elevated pulmonary vascular resistance and normal wedge pressure, from postcapillary forms associated with high output [16].

3.10. Combined SMAD4-related syndrome

In SMAD4 carriers, signs of HHT may appear in early childhood and sometimes before the polyps; pulmonary AVMs with digital clubbing and epistaxis are particularly frequent, while telangiectases are less consistent, and anaemia, migraine and exercise intolerance have also been described [38]. In a multicentre review of 34 carriers from 20 families, 76% had some manifestation of HHT [38]. The digestive component includes colonic polyps in 31 of 32 carriers and an aggressive gastric phenotype [38]. OMIM brings together observations that reinforce this profile: in the series of Gallione and colleagues (2004), 9 of 14 patients had telangiectases, 9 epistaxis and 7 pulmonary AVMs, one diagnosed at 9 months, and 4 had hepatic AVMs, one at 9 years — earlier ages than usual in HHT — and digital clubbing and hypertrophic osteoarthropathy regressed in some patients after embolisation of pulmonary AVMs; among 30 patients with HHT without an ENG or ACVRL1 variant, three had an SMAD4 variant and all had colonic polyps, one with colorectal cancer [8]. In a family with polyposis and the truncating variant NM_005359.6:c.1333C>T p.(Arg445Ter), three of five affected individuals had aortic dilatation and three mitral regurgitation, without telangiectases, and in a review of 26 patients with HHT from 15 families the six with aortopathy — from mild dilatation to dissection — were all SMAD4 carriers [8].

3.11. Pregnancy

Pregnancy carries a risk of pulmonary AVM enlargement and rupture: cardiac output increases 30–50% and blood volume 40% by 28 weeks, while progesterone increases venous distensibility [16]. The Oxford Desk Reference reports complications in up to 50% of women with pulmonary AVMs, and the guideline gives an overall complication risk of 1% in pregnant women with HHT and, with untreated pulmonary AVMs, estimates of approximately 17% non-fatal complications and 2% mortality; risk is greatest in unscreened women and those with known untreated pulmonary AVMs and feeding arteries above 2–3 mm [12, 16]. In a British tertiary-centre series (262 pregnancies in 111 women with HHT and pulmonary AVMs and 222 pregnancies in affected relatives), 1.0% of pregnancies involved severe pulmonary AVM haemorrhage and 1.2% maternal death; all deaths occurred in women considered healthy, 74% unaware of their diagnosis, and prior knowledge was associated with better survival [48]. Epistaxis may decrease or cease in some pregnant women and increase in others, with new telangiectases [12]; Smith's textbook reports that pregnancy generally worsens bleeding [26].

4. Clinical Recognition and Warning Signs

HHT should be considered with any single Curaçao criterion: recurrent epistaxis (typically four or more episodes yearly), multiple telangiectases at characteristic sites, a visceral AVM at typical sites or a first-degree relative diagnosed by the same criteria [1]. History-taking should include a three-generation pedigree with directed questions about epistaxis, telangiectases, stroke, unexplained anaemia, gastrointestinal bleeding, headache, epilepsy, intracranial haemorrhage, transient ischaemic attack, brain abscess, exertional dyspnoea, haemoptysis and haemothorax [12]. For epistaxis, establish whether it is spontaneous, age at onset and frequency; nocturnal bleeding is particularly suspicious [12]. Examination carefully seeks telangiectases on lips, mouth, tongue, fingertips and nail beds — potentially subtle and visible only close up — clubbing and cyanosis, liver size and a hepatic bruit [12].

Recognition is difficult in children: epistaxis, telangiectases and visceral AVM symptoms are often absent, so clinical criteria frequently fail to identify affected children and an at-risk child cannot be considered unaffected on clinical grounds alone [1, 6]. In adults older than 40, absence of even mild recurrent epistaxis and subtle telangiectases at characteristic sites on careful examination is reassuring [1]. Pregnancy-related telangiectases may resemble HHT but regress postpartum; chronic liver disease lesions are almost always “spider naevi” on the face, chest and periumbilical region; and healthy adults may develop one or two cutaneous telangiectases with age, sometimes accompanied by cherry angiomas, unrelated to HHT [1].

5. Diagnosis: Criteria, Tests and Algorithms

5.1. Curaçao criteria

Sources: Shovlin and colleagues (2000), Oxford Desk Reference, 2020 guideline and GeneReviews [1, 12, 15, 16].

The criteria were published in 2000 and retained in both international guidelines [15, 16, 21]. The Oxford Desk Reference adds that within HHT families a firm diagnosis can be based on two distinct visceral manifestations [12]. The update proposal incorporated by GeneReviews clarifies that AVMs outside typical sites are rare and do not suggest HHT, that telangiectases outside characteristic sites are non-specific and that any family history of manifestations is suggestive, but only a first-degree relative diagnosed by the criteria fulfils the family criterion [1]. The first guideline, whose recommendations remain applicable, recommended diagnosing HHT by Curaçao criteria or identification of a causal variant (G1), considering diagnosis in anyone with one or more criteria (G2) and considering asymptomatic children of an affected parent as possible HHT until excluded by genetic testing (G3) [16].

Specificity of strictly fulfilled criteria is high — 96.1% of probands with definite HHT had an ENG or ACVRL1 variant — but sensitivity is limited, particularly in young people and patients with few signs [11]. In a series of 152 unrelated adults with genetically confirmed HHT, only 104 (68%) met three criteria, and genetic yield was similar with three criteria (76%) or one to two (68%) [45]. Imaging detection of hepatic AVMs may add a visceral criterion and clarify diagnosis [16]. Criteria diagnose HHT rather than the gene and do not distinguish ENG-, ACVRL1-, SMAD4- or GDF2-related forms [6]. They may also err in the opposite direction: in the family in which HHT type 3 was mapped, at least one relative fulfilling three criteria — recurrent epistaxis and telangiectases at characteristic sites, alongside kinship — lacked the familial ENG variant, illustrating overdiagnosis risk in relatives with non-specific signs and the advantage of familial variant testing [4].

5.2. Genetic testing: indications, strategy and coverage

The 2011 guideline recommends genetic testing referral to identify the causal variant in a clinically diagnosed family, establish or exclude diagnosis in asymptomatic or mildly symptomatic relatives and those wanting prenatal testing, and support diagnosis in people not meeting criteria; in patients negative for ENG and ACVRL1 coding sequences, it recommends considering SMAD4 testing [16]. The Oxford Desk Reference recommends ENG testing particularly in pulmonary AVM families, and ACVRL1 and SMAD4 testing [12]. GeneReviews proposes a multigene panel including ACVRL1, ENG, SMAD4 and differential diagnosis genes, or exome or genome sequencing; suspected HHT warrants a panel including all HHT-associated genes, capillary malformation–arteriovenous malformation syndrome genes (EPHB4, RASA1) and BMPR2 [1]. Molecular testing is unnecessary to distinguish HHT from ataxia-telangiectasia or von Willebrand disease [1].

Coverage depends on method. For ACVRL1 and ENG, sequencing detects approximately 90% of pathogenic variants and targeted deletion/duplication analysis (MLPA, multiplex ligation-dependent probe amplification, quantitative PCR or targeted microarray) the remaining 10%; exome or genome sequencing may detect deletions through read depth, with lower sensitivity [1]. For SMAD4, sequencing detects more than 99%, and only one large deletion has been described in a patient with juvenile polyposis and HHT [1]. Non-coding variants and large deletions/duplications have been described in ENG [36]. Rare somatic mosaicism may escape leucocyte analysis [1]. The proportion of variants detectable by a method is not clinical test sensitivity: even comprehensive testing leaves approximately 3% of definite clinical HHT without a molecular diagnosis, and a negative result does not exclude disease with a convincing phenotype [1, 11]. DNA banking should be considered for probands without a molecular diagnosis [1].

The report should state the gene, versioned transcript (MANE Select), DNA and protein HGVS nomenclature, zygosity, ACMG/AMP classification — with ClinGen expert panel specifications for ACVRL1 and ENG — and method limitations [1, 10]. Pathogenic and likely pathogenic variants are equivalent for clinical decisions; a variant of uncertain significance neither confirms nor excludes diagnosis, and partial familial segregation is research rather than automatic reclassification [1]. ClinGen's HHT actionability assessment (document AC107) classified it as moderate in adults and strong in children, consistent with the failure of clinical criteria in childhood [6].

The 2024 French PNDS offers any patient with a definite or possible clinical diagnosis (at least 2 criteria) a sequencing panel including ACVRL1, ENG, SMAD4 and GDF2, preferably extended to EPHB4 and RASA1, and, if the panel is negative in a family with several individuals with a definite diagnosis, whole-genome sequencing, to look for balanced rearrangements and deep intronic variants; the familial variant should be offered to first-degree relatives, even if asymptomatic, with the result communicated by the prescriber or the genetic counsellor, and the patient is informed of the obligations regarding informing relatives [23]. In the asymptomatic child, as clinical expression is weak, only genetic testing excludes the disease; the PNDS accepts presymptomatic testing after discussion with the legal representatives and the child, with a psychological interview, suggesting 7–10 years as the most appropriate age, and reserves cord-blood testing, which is not systematic, for rare families in which cerebral and spinal MRI between 1 and 4 months is intended; preimplantation or prenatal diagnosis is discussed only in exceptionally severe situations, in a multidisciplinary prenatal diagnosis committee [23].

5.3. AVM screening and characterisation

Contrast echocardiography is more sensitive than radiography with oximetry and avoids radiation but requires venous access; quantitative shunt grading increases specificity and guides CT decisions [16]. Both strategies are considered equivalent for initial paediatric assessment: the “Dutch protocol” of oximetry and radiography does not exclude small AVMs, so antibiotic prophylaxis is recommended for everyone, and the Dutch cohort suggests it suffices to prevent complications [16]. Hepatic Doppler ultrasound had 97–99% sensitivity and specificity in one series and permits severity grading (0.5 to 4), which predicts evolution; multiphase CT does not correlate with severity; magnetic resonance imaging is as accurate as CT without ionising radiation [16]. Endoscopic diagnostic yields in adults with definite HHT were 67–82% for upper endoscopy and 76–91% for capsule, which assesses the stomach poorly and may miss up to 50% of significant gastric lesions [16].

In the 2024 French PNDS, in children, screening for pulmonary AVMs is systematic if there are respiratory symptoms or low saturation (low-dose volumetric CT without contrast, with MIP and MPR reconstructions and sagittal reading); in the asymptomatic child it is done before adolescence by contrast echocardiography, without CT if this is negative, and repeated after adolescence; hepatic Doppler is not justified routinely in children; cerebral and spinal MRI may be offered, without demonstrated benefit and with divergence between European and international recommendations, before 4 months of age or when the child can cooperate, especially in families with early neurological complications; gastrointestinal angiodysplasias are investigated only in the presence of warning signs, and in SMAD4 carriers gastrointestinal endoscopy starts at 12–15 years [23]. In adults, systematic full blood count with ferritin, contrast echocardiography as first line (CT only if it cannot be performed or if positive, or with suggestive clinical features; ultra-low-dose CT according to the centre's experience), hepatic Doppler with measurement of calibres and echocardiography with cardiac output, cardiac index and probability of pulmonary hypertension, cerebral and spinal MRI offered after information, and gastrointestinal endoscopy (gastroscopy, colonoscopy and, if needed, capsule) only in the presence of bleeding or unexplained iron deficiency; spinal MR angiography before the first pregnancy is no longer recommended [23].

5.4. Practical algorithm

6. Differential Diagnosis

Recurrent epistaxis may manifest a bleeding diathesis, including von Willebrand disease, but also occurs in healthy people [1]. Telangiectases occur in inherited diseases of known and unknown cause, CREST syndrome (calcinosis, Raynaud phenomenon, oesophageal dysmotility, sclerodactyly and telangiectases), hereditary benign telangiectasia, pregnancy and chronic liver disease [1]. AVMs occur in other vascular dysplasias; absence of telangiectases specifically on lips, oral cavity and fingers best distinguishes these from HHT [1]. Most cerebral AVMs are isolated, although families with dominantly inherited cerebral AVMs without other HHT signs exist; multiple cerebral AVMs strongly suggest HHT [1]. A substantial minority of isolated pulmonary AVMs are sporadic [1]. Orphanet also includes gastrointestinal angiodysplasia and isolated sporadic pulmonary, hepatic and cerebral AVMs [2].

Smith's textbook emphasises the cutaneous distinction between the lesions of the two most often confused entities: in HHT, telangiectases are located mainly in the oral cavity, lips, hands, face, fingers and palms, have no halo and are less numerous than the capillary malformations of RASA1-related disorders, and mucosal and visceral AVMs are much more frequent in HHT [50]. In EPHB4-related CM-AVM, series recruited from HHT clinics show more epistaxis and telangiectases (6 and 8 of 10 patients) than those recruited for capillary malformations [49]. In the first three patients with GDF2 missense variants (identified by exome sequencing in 38 patients without variants in the three known genes and by sequencing another 153), the skin lesions were less concentrated on the hands and mouth and some resembled the atypical capillary malformations of CM-AVM; in 2022, a British family with the variant NM_016204.4:c.1282T>C p.(Cys428Arg), in which the proband had diffuse pulmonary AVMs from the age of 8 and the mother and sister mild forms, was the first with GDF2-associated HHT meeting the consensus criteria [9]. Uncertainty persists over the status of GDF2: it is not known whether the related disease will be considered a type of HHT or a distinct, overlapping entity, but the cases described — three with pulmonary AVMs and two with cerebral AVMs — suggest management similar to that of the ACVRL1- and ENG-related forms [1]. In children with polyposis, the differential diagnosis includes juvenile polyposis, Peutz–Jeghers syndrome, PTEN hamartoma tumour syndrome and adenomatous polyposis.

7. Treatments and Risk Management Strategies

7.1. Organisation of care and reference documents

Management follows the 2020 Second International Guidelines, developed using GRADE (Grading of Recommendations Assessment, Development and Evaluation) by experts from 15 countries, methodologists, patient representatives and people with HHT, with 36 recommendations in six areas — epistaxis, gastrointestinal bleeding, anaemia and anticoagulation, hepatic AVMs, paediatrics and pregnancy — requiring at least 80% consensus; first-guideline (2011) recommendations on diagnosis, cerebral AVMs and pulmonary AVMs were not reviewed and remain applicable [16, 21]. Recommendation strength and evidence quality are voted separately, so a strong recommendation may rest on low evidence, as with several paediatric and obstetric recommendations [16]. GeneReviews, revised in February 2026, considers guidelines complementary to expert advice, usually at an HHT centre of excellence; before any visceral AVM treatment, patients and physicians should contact their nearest multidisciplinary HHT clinic, located through Cure HHT in the United States or VASCERN in Europe [1, 16]. VASCERN has published its own position statements distinguishing specialist from non-specialist care and including definitions, practical recommendations, outcome measures, anticoagulation, cerebral screening and antiangiogenic treatments [22, 25]. In 2025, an international consensus standardised terminology, definitions and outcome criteria for bleeding and anaemia in HHT [51].

7.2. Epistaxis

Moisturisation is fundamental: in a controlled topical-therapy trial, saline used as placebo reduced the Epistaxis Severity Score (ESS) at 12 and 24 weeks [16, 52]. GeneReviews adds humidification and haemostatic products (gauze, sponges, powders) and recommends additional treatment when epistaxis causes anaemia or interferes with activities [1]. Ablative therapies provide temporary partial improvement, with generally minor adverse effects; septal perforation is a recognised complication of all techniques, bipolar electrosurgery is preferred over monopolar and radiofrequency operates at lower temperatures [16]. Outpatient sodium tetradecyl sulphate foam sclerotherapy under local anaesthesia reduced ESS in a crossover trial of 17 patients (difference 0.95 points), but may cause septal perforation, transient dizziness, blurred vision and, rarely, permanent blindness [16]. The Oxford Desk Reference considers laser preferable to cauterisation, and most ENT specialists experienced in HHT advise against electrical or chemical cauterisation and transcatheter embolisation for recurrent epistaxis [12, 38].

Surgery is an alternative equivalent to systemic therapies, to be decided with the patient, and may be preferable when a comorbidity such as atrial fibrillation requires anticoagulation or antiplatelet treatment [16]. In the largest septodermoplasty series, 86% reported better quality of life after mean follow-up of 3.75 years, at the cost of worsened sinusitis (30%), reduced smell (58%), crusting and nasal obstruction [16]. With nasal closure, 83% of 43 patients in one series and 94% of 50 with pre- and post-procedure data in another series of 100 stopped bleeding, without reversal requests in the first [16]. GeneReviews reserves septodermoplasty and nasal closure for failure of topical measures, tranexamic acid, ablation and/or antiangiogenic agents, performed by surgeons regularly treating HHT [1].

Older foundational sources mention treatments now of secondary importance. The Oxford Desk Reference mentions tamoxifen as potentially reducing epistaxis frequency and severity and bevacizumab as still “under investigation”, and Smith's textbook describes occasional success with oral oestrogens and septodermoplasty [12, 26]. The 2020 guideline states that early studies suggested hormonal treatment benefit, but antifibrinolytic and antiangiogenic agents have better benefit–risk profiles, and evidence for selective oestrogen receptor modulators (tamoxifen, raloxifene, bazedoxifene) is limited to small numbers [16]. Thalidomide reduced epistaxis severity and transfusion needs in small series (67 patients across six studies), but neuropathy frequently caused discontinuation and persisted in two-thirds after stopping, while teratogenicity prevents use in women of childbearing age [16].

7.3. Gastrointestinal bleeding

Endoscopy should take place at an experienced centre because unusual complications such as massive epistaxis may occur and pulmonary AVMs require precautions [16]. Endoscopic findings (telangiectasia number, size and distribution) correlate poorly with anaemia severity, so grading is based on haematological support [16]. In one series, oral tranexamic acid reduced endoscopic treatment requirements from 80% to 40% (trend, p=0.07), without significant anaemia improvement in epistaxis studies; its gastrointestinal bleeding effect is probably weak [16]. Pharmacological gastrointestinal bleeding treatment is usually necessary only when iron fails to maintain normal haemoglobin [1]. In the PATH-HHT extension, pomalidomide was less effective for gastrointestinal bleeding than epistaxis, particularly with substantial transfusion requirements [53].

7.4. Anaemia and iron

Assessment includes full blood count and ferritin; if there is anaemia with non-reduced ferritin, serum iron, total iron-binding capacity and transferrin saturation are measured and a haematology referral is considered [16]. Asymptomatic children do not need routine laboratory screening [16]. The initial oral dose proposed by the guideline is 35–65 mg of elemental iron once daily, 2 hours before or 1 hour after meals, switching to alternate days or another preparation if there is intolerance, and increasing the daily dose or giving it twice daily if correction is insufficient; a haemoglobin rise of less than 1 g/dL after one month is an inadequate response [16]. This regimen is based on evidence that hepcidin rises after each dose and reduces fractional absorption, a dose of 40–80 mg per dose being optimal [16]. For comparison, the Portuguese SmPC for prolonged-release ferrous sulphate (Ferro-Gradumet®, 105 mg of elemental iron per tablet) indicates, in adults, 1 tablet daily, on an empty stomach, with more than one tablet daily in severe cases and a duration of up to 6 months, except with continued losses such as those of chronic iron-deficiency anaemia [54]; the recommended dose is that of the guideline, lower than that of the label because of the absorption rationale described. Antacids, H2 antagonists, proton pump inhibitors, aluminium-containing phosphate binders, calcium and colestyramine should not be taken at the same time as iron, nor should tea, coffee or milk [16]. Intravenous iron can be calculated from the total deficit (Ganzoni formula) or given as an initial total dose of 1 g, in one or several infusions; in chronic bleeding, scheduled infusions every few months are often needed [16]. For comparison, the labels set the limits per administration: ferric carboxymaltose (Ferinject® in Portugal) with a total requirement of 500–2,000 mg according to weight and haemoglobin, a maximum single dose of 20 mg/kg per infusion (15 mg/kg per injection) and of 1,000 mg, and a maximum of 1,000 mg per week from 14 years of age; from 1 to 13 years, 15 mg/kg and 750 mg per administration and per week [55]; in the USA (Injectafer®), 750 mg in two doses at least 7 days apart, or 15 mg/kg up to 1,000 mg as a single dose in adults weighing ≥50 kg, and 15 mg/kg in two doses below 50 kg [56]; ferumoxytol (Feraheme®) 510 mg followed by 510 mg 3–8 days later [57]; ferric derisomaltose (Monoferric®) 1,000 mg at ≥50 kg or 20 mg/kg below that weight [58]. Repeated doses of ferric carboxymaltose have been associated with hypophosphataemia in more than 20% of patients, and after ferumoxytol magnetic resonance imaging should be avoided for at least 4 weeks; ferritin is unreliable in the 2 weeks following an infusion [16]. GeneReviews reserves transfusion for the situations of recommendation C3, including inability to maintain haemoglobin above 7–8 g/dL despite frequent iron infusions [1].

7.5. Anticoagulation and antiplatelet treatment

HHT is not an absolute contraindication to prophylactic or therapeutic anticoagulation or antiplatelet treatment: the guideline recommends these when indicated, considering individual bleeding risk (C5; low evidence, strong recommendation), and avoiding dual antiplatelet therapy and combined antiplatelet–anticoagulant treatment whenever possible (C6; consensus, weak recommendation) [16]. When anticoagulating, unfractionated heparin, low-molecular-weight heparin and vitamin K antagonists are preferred over direct oral anticoagulants, less well tolerated in HHT; left atrial appendage closure may be considered for atrial fibrillation with anticoagulant intolerance [1, 16]. HHT does not protect against thrombosis: in a large series, thrombotic event prevalence was 6%, above the age-matched general population, independent of comorbidities and antihaemorrhagic treatments and correlated with iron deficiency and elevated factor VIII, with increased thrombotic stroke risk also described; patients should therefore receive pharmacological thromboprophylaxis in usual risk situations such as prolonged immobilisation or major surgery [16]. When a non-steroidal anti-inflammatory is needed, one not inhibiting platelet aggregation, such as celecoxib, may be preferred [1]. CHORUS recorded venous thromboembolism in 7% and arterial thromboembolism in 11% [31].

7.6. Systemic bleeding treatment: drugs, doses and safety

No targeted treatment is FDA-approved specifically for HHT; the following four options are used off-label, with doses reviewed by an HHT expert, and paediatric experience with bevacizumab, pomalidomide and pazopanib is limited without recommended dosing [1].

In InHIBIT-Bleed (238 patients from 12 centres, median treatment 12 months), bevacizumab increased mean haemoglobin by 3.2 g/dL (8.6 to 11.8 g/dL) and reduced ESS by 3.4 points in the first year; in the first 6 months, transfused red cell units decreased 82% and iron infusions 70% relative to the preceding 6 months, with similar results regardless of gene; continuous maintenance performed better than “as-needed” maintenance with greater exposure, and venous thromboembolism occurred in 2% [69]. In European VASCERN experience (151 patients, mainly Denmark, France and the Netherlands), haemoglobin increased by a mean 2.9 g/dL, transfusions decreased and cardiac output improved in 85% treated for high-output heart failure; there were no serious adverse effects, but 70 had at least one [70]. In an earlier European survey, 69 bevacizumab-treated patients (63.8 person-years) had grade 1–3 adverse effects at 50 per 100 person-years, more often in women, and one possibly related fatal adverse effect; thalidomide incidence was 45.3 per 100 person-years, with three possibly related deaths [62]. Topical administration does not work: bevacizumab nasal spray (25, 50 or 75 mg per administration) did not reduce epistaxis duration and the trial stopped for futility, while NOSE found bevacizumab 1%, oestriol 0.1% and tranexamic acid 10% sprays no better than saline [52, 71].

The American bevacizumab label requires withholding for at least 28 days before elective surgery and not administering within 28 days after major surgery and until healing [61]. In HHT, the guideline estimates an optimal liver transplantation window between 2 and 6 months after the last dose [16].

7.7. Pivotal and proof-of-concept clinical trials

Comparison is limited across these studies. Outcomes differ — duration, frequency, ESS, haemoglobin, transfusions, cardiac output — and only PATH-HHT and ATERO showed primary efficacy outcome benefit, while Geisthoff's epistaxis reduction was secondary and haemoglobin, the primary outcome, unchanged; engasertib's primary outcome was safety, and epistaxis reductions are secondary findings not directly comparable with placebo results in other trials [17, 65, 72, 73]. Placebo effects are substantial in HHT: in NOSE all groups including saline improved ESS, and in PATH-HHT analysis placebo effects were smaller with more severe baseline disease [52, 77]. In that analysis, pomalidomide response was greater with higher baseline ESS and concomitant antifibrinolytics, and smaller with age and, with borderline significance, ACVRL1 variants [77]. Bevacizumab data derive from non-randomised series subject to selection bias and regression to the mean; improved mean haemoglobin is a group average, not an individual response prediction [69, 70]. A French phase 3 intravenous bevacizumab trial for severe bleeding (BABH, primary outcome: transfusion number at 6 months) is registered as completed, without results located in consulted sources [78]. Negative trials — nintedanib's primary outcome and nasal sprays — show that biological plausibility alone does not suffice [52, 71, 75].

7.8. Regulatory status of drugs used

In the European Union there are seven orphan designations for HHT treatment, all positive in October 2026, none corresponding to an authorised medicine [20]:

At present, no therapy is licensed for HHT in any country, as stated by the engasertib trial authors and the sponsor's August 2026 release [17, 18]. The only identified European authorisation application, bevacizumab (Orblid®) for adults with HHT, was withdrawn by the applicant on 21 May 2026 and published the following day; withdrawal does not cancel orphan designation, which remains positive [19, 20]. Orphan designations — raloxifene, bazedoxifene, bevacizumab, thalidomide, etamsylate, a small-molecule inhibitor and the ALK1/BMPRII agonist antibody — imply neither authorisation nor availability and do not demonstrate efficacy [20, 84]. Access to off-label drugs in Portugal and Brazil was not documented in consulted sources.

7.9. Pulmonary arteriovenous malformations

Transcatheter embolisation is the treatment of choice and should be considered for any pulmonary AVM with a feeding artery above 2–3 mm; it is also indicated for dyspnoea, exercise intolerance and hypoxaemia, but its principal aim is prevention of pulmonary haemorrhage and neurological complications (brain abscess and stroke), even with normal pulmonary function and saturation [1, 6]. Occasionally a small lesion is inaccessible because of location or vessel calibre [1]. In children, the guideline recommends treating large AVMs and those associated with desaturation (E3; moderate evidence, strong recommendation), considering feeding arteries of 3 mm or more suitable for embolisation; no data justify treating small AVMs with normal saturation [16]. Smith's textbook gives a lower threshold (feeding artery above 1.0 mm), an earlier recommendation superseded by current thresholds [1, 26]. The first guideline, still applicable, recommends screening all patients with possible or confirmed HHT (I1), initial contrast echocardiography (I2), embolisation treatment (I3), lifelong counselling for patients with treated or untreated pulmonary AVMs (I4) and long-term follow-up to detect growth of untreated AVMs and reperfusion of treated lesions (I5) [16]. CT is usually performed 6–12 months after embolisation and, without recanalisation or new AVMs, every 3 years [1].

When there is a pulmonary shunt on contrast echocardiography, even without an AVM visible on CT, antibiotic prophylaxis following the American Heart Association protocol is recommended for dental treatment and other procedures with a risk of bacteraemia, because the risk is abscess — especially cerebral — and not endocarditis; avoidance of air bubbles in intravenous lines (with an air filter when available and compatible) and avoidance of scuba diving are also recommended [1]. The Oxford Desk Reference states that patients with pulmonary AVMs, even after embolisation, should receive antibiotic prophylaxis and gives amoxicillin–clavulanic acid 375 mg before and after dental appointments as an example of British practice [12]. Prophylaxis is not among the label indications: in the US label (Augmentin®), 375 mg corresponds to the tablet of 250 mg amoxicillin and 125 mg clavulanic acid, whose therapeutic posology is one tablet every 8 hours in less severe infections, so the regimen cited is use outside the label conditions [85]. The British Thoracic Society statement, focused on adults over 16 years of age, covers embolisation, ancillary medical management, screening and follow-up after diagnosis, intervention or negative screening [44].

7.10. Hepatic vascular malformations

Asymptomatic hepatic AVMs should not be treated [16]. First-line management includes aggressive anaemia treatment, salt restriction and diuretics for high-output cardiac failure, antibiotics for cholangitis, and usual portal hypertension and encephalopathy management (lactulose, rifaximin); non-selective beta-blockers in severe heart failure require cardiological supervision, transjugular intrahepatic portosystemic shunting may precipitate heart failure, and endoscopic retrograde cholangiopancreatography with stenting is not an option [16]. First-line response was complete in 63%, partial in 21% and absent with progression to death in 14% [16]. Progression predictors are Doppler grade 4 and ACVRL1 variants, and significant disease predictors are presentation after age 47, female sex, haemoglobin below 8 g/dL and alkaline phosphatase above 300 IU/L; other concerning features include mean pulmonary arterial pressure of 25 mmHg or more, elevated bilirubin, weight loss, gastrointestinal bleeding, biliary ischaemia, atrial fibrillation, high transfusion requirements, right upper quadrant pain and sepsis [16]. Bevacizumab improved cardiac output or symptoms in 80% in small series, but relapse after withdrawal makes it a bridging treatment rather than a transplant substitute [16, 74]. Survival 5–10 years after transplantation is 82–92%, graft recurrence is rare, late and asymptomatic, transplantation is possible with pulmonary hypertension if pulmonary vascular resistance is below 3 Wood units, and the guideline cites suggested MELD (Model for End-Stage Liver Disease) exception points of 40 for acute biliary necrosis and 22 for high-output cardiac failure [16]. Hepatic embolisation and hepatic artery ligation carry high rates of serious complications, including death and hepatic infarction, and remain last resorts; liver biopsy should be avoided [1, 16].

7.11. Cerebral and spinal vascular malformations

Cerebral AVMs are treated by neurosurgery, embolisation and/or stereotactic radiosurgery according to size, location and symptoms, with magnetic resonance and angiographic follow-up because recurrence may occur, particularly in children [1]. The first guideline recommends magnetic resonance screening for cerebral AVMs in adults with possible or definite HHT (H1; weak recommendation), definitive treatment at a neurovascular-experienced centre for adults with acute cerebral AVM haemorrhage (H2), referral of other adults with cerebral AVMs to such centres for individual decisions (H3), and deferral of definitive treatment of asymptomatic AVMs discovered during pregnancy until after delivery (H4) [16]. In children, the second guideline recommends cerebral vascular malformation screening at diagnosis or presentation (E5; low evidence, strong recommendation) and treatment of lesions with high-risk features (E6) — feeding-artery or nidus aneurysms, draining venous stenosis, deep venous drainage, previous haemorrhage, intralesional microhaemorrhage and pial fistulae [16].

This is the greatest area of divergence between sources. The guideline itself acknowledges practices ranging from magnetic resonance screening in early childhood to no routine screening, and that patient representatives advocated screening [16]. The VASCERN position (2020) considers current evidence unfavourable to treating unruptured cerebral AVMs — with low HHT haemorrhage risk — and thus unsupportive of universal asymptomatic screening; it proposes informed discussion for all patients, explaining potential harm from detecting or intervening on lesions that would never cause problems [22]. The Oxford Desk Reference reports little screening benefit evidence and cites ARUBA, in which conservative management outperformed intervention after 3 years in adults with sporadic unruptured cerebral AVMs [12]. Conversely, an American multidisciplinary group argues that ARUBA does not apply to children with HHT, who predominantly have high-flow fistulae and a lifetime of additive risk, and considers waiting for haemorrhage unacceptable [47]. In North American practice, 25 of 28 centres of excellence rescreen children, mostly with magnetic resonance imaging, but only 4 of 30 rescreen adults [86]. GeneReviews, reflecting US practice, recommends magnetic resonance imaging as early as possible, preferably in the first year, and repetition by age 18–20 if initially performed in childhood, without further repetition if normal at that age [1]. Spinal malformations do not require routine screening [16].

7.12. Pulmonary hypertension

Pulmonary arterial hypertension is treated by cardiologists, respiratory physicians and other specialists, and patients with high-output heart failure and pulmonary hypertension should be co-managed by the HHT centre and a cardiologist experienced in HHT or a pulmonary hypertension clinic [1, 16].

7.13. Specific component of SMAD4-related syndrome

In patients with confirmed or suspected SMAD4-related HHT, the guideline recommends colonoscopy from 15 years of age, repeated every 3 years if there are no polyps or annually, with upper endoscopy, if there are colonic polyps; other patients follow general-population colorectal cancer screening [16]. The juvenile polyposis GeneReviews chapter recommends considering transthoracic echocardiography at diagnosis in SMAD4 carriers, with no defined starting age or frequency, and management of aortopathy and valve disease by a cardiologist and a cardiothoracic surgeon [38]. The Oxford Desk Reference recommends screening for aortic root dilatation in these patients, and the authors of the report of two cases with aortopathy proposed screening of the thoracic aorta to prevent death from dissection [12, 43]. Orphanet summarises that SMAD4-related HHT requires polyposis screening and aortic follow-up [2].

7.14. Agents and circumstances to avoid

With significant epistaxis, avoid forceful nose-blowing, heavy lifting, straining at stool and nose-picking [1]. Anticoagulants, acetylsalicylic acid and anti-inflammatory drugs affecting coagulation should be avoided without an indication, although HHT is not an absolute contraindication when indicated (section 7.5) [1, 16]. Scuba diving should be avoided unless contrast echocardiography within the preceding 5 years was negative for right-to-left shunt, and liver biopsy should be avoided, particularly with documented hepatic AVMs [1].

7.15. Developments in the last 12 months and ongoing research

Other agents remain under investigation: ALN-6400, an interfering RNA reducing plasminogen, as a haemostatic agent (phase 1/2 InsigHHT trial, primary safety outcome, centres in Europe, North America and Australia); nintedanib (a second placebo-controlled trial in Switzerland and France, recruiting); tacrolimus and sirolimus, acting through FKBP12 and SMAD1/5/8 or mTOR; and pazopanib, whose randomised placebo-controlled phase 2/3 trial completed its main phase in November 2025, with results expected in 2026 according to GeneReviews but still unpublished in consulted sources [1, 68, 90, 91]. Registrations and releases do not establish efficacy; trial-initiation announcements and orphan designations are not favourable opinions or authorisations.

7.16. French PNDS (2024 and 2019): complementary recommendations

Table 7.16 — Recommendations of the French PNDS that add detail to the international guidelines [23, 24]

Table 7.16a — Divergences between the French PNDS and the 2020 international guidelines [1, 16, 23, 24]

8. Reproductive Options and Family Counselling

8.1. Inheritance and risks

HHT follows autosomal dominant inheritance: each child of an affected individual has a 50% probability of inheriting the pathogenic variant [1, 12]. Most probands have an affected parent; rarely, disease results from a de novo variant or postzygotic mosaicism [1]. If the proband's variant is absent from both parents' leucocyte DNA after parentage confirmation, a de novo variant or parental gonadal (or somatic and gonadal) mosaicism is considered — leucocyte DNA testing does not exclude all mosaicism — and sibling risk is slightly above the general population [1]. A proband may appear uniquely affected because relatives' disease is unrecognised, penetrance reduced, a parent died early or onset was late; a carrier parent may not manifest until after age 40, and signs may remain subtle until late adulthood [1].

Penetrance depends on age and individual manifestation. The Oxford Desk Reference reports 62% symptomatic by age 16, 83% by 26 and 97% by 35, with complete penetrance at 40 [12]; GeneReviews describes almost 100% penetrance by late adulthood, and Orphanet indicates most have symptoms before 50 [1, 2]. Intrafamilial variability is considerable, and parental severity does not predict a child's [1, 2]. In SMAD4-related disease, a carrier's child inherits risk of both vascular and digestive components and aortopathy [1, 38].

8.2. Predictive testing and assessment of relatives

Apparently asymptomatic at-risk relatives, younger or older, should be assessed to identify early those benefiting from surveillance and treatment [1]. If the familial variant is known, molecular testing is performed; non-carriers are released from screening [1, 16]. If unknown and disease cannot be excluded, the relative follows affected-person surveillance; after age 40, directed history and examination may suffice, but up to age 40 cerebral and pulmonary AVM assessment should be added because clinical assessment fails to identify affected young people [1]. The 2020 guideline advises offering diagnostic genetic testing to asymptomatic children of an HHT parent (E1; high evidence, strong recommendation), after testing an affected relative to identify the variant, and discussing benefits, alternatives, advantages and disadvantages with the child or parents; testing value may be viewed differently depending on national screening protocols [16]. Predictive testing should be offered before age 15 in SMAD4 carriers because vascular surveillance starts earlier than polyp surveillance [38]. Pre-test counselling should discuss socioeconomic consequences, prolonged follow-up and test limitations [1].

8.3. Reproductive options

The guideline recommends discussing preconception and prenatal diagnostic options, including preimplantation genetic testing, with affected individuals (F1; very low evidence, strong recommendation) [16]. With a known familial variant, options are preimplantation genetic testing with transfer of unaffected embryos, chorionic villus sampling or amniocentesis (small miscarriage risks of approximately 1% and below 0.5%, respectively), or cord blood testing after delivery; choices depend on local legislation on preimplantation diagnosis and pregnancy termination, and if termination is not contemplated, invasive testing may be reserved for other indications [16]. Orphanet emphasises that prenatal diagnosis is unnecessary for pregnancy and delivery management and that routine ultrasound by sonographers aware of family history detects most major AVMs [2]. The optimal time to discuss genetic risk is before pregnancy, and counselling should be offered to affected or at-risk young adults [1].

8.4. Pregnancy and delivery

Pulmonary AVMs should ideally be screened and treated before pregnancy, and the Oxford Desk Reference suggests repeat screening between pregnancies [12, 16]. In asymptomatic pregnant women, CT and embolisation are deferred until organogenesis ends (12 weeks); symptomatic women (haemorrhagic or neurological complications) undergo immediate CT and embolisation at any gestational age [16]. Embolisation should minimise fetal radiation — avoid fluoroscopy over abdomen and pelvis, use pulsed or low-dose fluoroscopy, minimise angiographic series and collimate — and lead abdominal shielding does not help and may increase scatter [16]. The calibre threshold for embolising asymptomatic pregnancy AVMs is unestablished but should follow the general-population 2–3 mm threshold [16]. No evidence supports treating asymptomatic unruptured cerebral AVMs in pregnancy, and blood pressure management is essential with cerebral malformations; caesarean delivery may be considered for high-risk or symptomatic lesions [16]. In two large HHT pregnancy studies (92 spinal or epidural anaesthetics in 185 deliveries in one, 484 pregnancies in another), no spinal haemorrhages occurred, and the theoretical risk should be explained to pregnant women, who should consult an anaesthetist early in the third trimester [16]. Iron is preferred for anaemia, with transfusion potentially necessary for refractory symptomatic anaemia [1]. Bevacizumab is contraindicated in pregnancy and pomalidomide is teratogenic [16, 63].

The 2024 French PNDS recommends more frequent full blood count and ferritin monitoring during pregnancy, with oral iron first and, if this fails, intravenous ferric carboxymaltose or iron sucrose, whose pharmacovigilance data are reassuring; spinal MRI is not needed to authorise epidural analgesia; screening for and treatment of pulmonary AVMs before pregnancy are strongly advised, because they can grow and endanger maternal or fetal life in 1–2% of pregnancies; in an asymptomatic pregnant woman not yet investigated, screening is done by contrast echocardiography or low-dose CT, ideally at the start of the second trimester, and at any gestational age if there is hypoxaemia; haemoptysis or chest pain should suggest rupture of a pulmonary AVM, and seizures or stroke paradoxical embolism [23].

8.5. Practical counselling aspects

Counselling should be provided by genetics professionals and include the pedigree, disease nature, inheritance and implications for patient and family [1]. Extended-family cascade screening is important because serious complications are preventable [12]. Without an identified molecular cause, consider proband DNA banking [1]. In an approach preceding paediatric guidelines, the Oxford Desk Reference limited pulmonary AVM screening to after puberty and those with recurrent epistaxis, telangiectases or a known variant, proposing childhood clinical assessment for cyanosis and clubbing and oximetry; current recommendations advocate pulmonary and cerebral screening at diagnosis, even in asymptomatic children [12, 16].

9. Special Considerations: Population Diversity and Epidemiology

9.1. Frequency and what each figure measures

Prevalence estimated from clinically diagnosed patients is approximately 1 in 5,000, but experts consider true prevalence higher [1, 16]. An estimate using gnomAD (v4.1) allele frequencies of predicted pathogenic ENG and ACVRL1 variants through three methods placed prevalence between 2.1 and 11.9 per 5,000, two to twelve times clinical estimates, similar across ancestries and with missense variants contributing most — a genomic estimate dependent on in silico classification rather than diagnosed disease [13]. Orphanet gives 1–5 per 10,000 and approximately 1 in 6,000, and the Oxford Desk Reference approximately 1 in 10,000 [2, 12]. Smith's textbook reports incidence of approximately 1 in 50,000, an outdated figure; OMIM records the evolution from 1–2 per 100,000 in an 18-family series (1964) to estimates of 1 in 50,000 to 100,000 considered underestimates by Guttmacher and colleagues in 1994 [1, 4, 26]. VASCERN estimates HHT affects approximately 85,000 European citizens [25].

The 2024 French PNDS uses a frequency of about 1 in 6000 births; the 2019 hepatic PNDS reported 1 in 5000 to 1 in 10,000, with about 10,000 patients in France distributed among about 2000 known families and wide regional variation [23, 24].

9.2. Population data and founder populations

Founder effects alter the local ACVRL1–ENG ratio and therefore regional predominant phenotype; recognising them may simplify diagnosis using local genetic algorithms [93]. In the Netherlands Antilles, skin pigmentation made facial telangiectases relatively rare, relevant when examining darker skin [14]. The disease has wide ethnic and geographical distribution [1]. No Portuguese or Brazilian epidemiological data were found in consulted sources; patient 10.1 originated from Argentina and was treated in Spain [94]. Absence of these data in consulted sources does not demonstrate absence of published studies.

9.3. Special populations

In children, clinical criteria fail and genetic testing has a greater role than in adults; pulmonary AVM prevalence resembles that in adults and cerebral fistulae are overrepresented [16]. Pregnant women have the greatest pulmonary haemorrhage risk (section 8.4) [16, 48]. Women of childbearing age require strict pregnancy prevention with thalidomide and pomalidomide, and bevacizumab is contraindicated during pregnancy [16, 63]. Women predominate in hepatic AVMs (4.5:1) and experience more bevacizumab adverse effects [16, 62]. Gastrointestinal bleeding and symptomatic hepatic AVMs predominate in older people [1, 35]. Before diving or invasive procedures, patients with pulmonary shunts require the precautions in section 7.9 [1].

9.4. Reference networks and patient organisations

In Europe, HHT belongs to the European Reference Network on Rare Multisystemic Vascular Diseases (VASCERN), whose working group was established in 2016 by eight founding centres from six countries, following more than 10,000 patients at launch and incorporating patient representatives [25]. Recognised expert centres can be located through Cure HHT (curehht.org) in the United States and VASCERN (vascern.eu) [16]. Patient organisations include Cure HHT (formerly HHT Foundation International), HHT Europe, which participated in developing the 2020 guideline, the United Kingdom's Telangiectasia Self Help Group and Germany's Osler-Selbsthilfe [1, 12, 16, 73, 81]. The American CHORUS registry and institutional registries in Argentina and Europe follow prospective cohorts [31, 95].

In France, follow-up is organised in a national reference centre (Hospices Civils de Lyon) and a network of competence centres of the FAVA-Multi filière, with a consultation, or teleconsultation, at least every 2 years in stable patients; the attending physician refers for confirmation, collaborates in follow-up, oversees preventive measures (antibiotic prophylaxis, pregnancy), supports the applications for long-term condition No. 31 and for transport and offers psychological support; the patient association AMRO-HHT-France supports and informs patients and created the ROSE app for recording epistaxis [23].

10. Illustrative Clinical Case

10.1. Brain abscess as the first recognised pulmonary AVM complication in an ENG-related HHT family

Hernández-de Sosa and colleagues at Hospital de la Santa Creu i Sant Pau, Barcelona, described in 2023 a 37-year-old non-smoking man from Argentina admitted in a postictal state after his first tonic–clonic seizure, preceded by perioral paraesthesia [94]. He had thalassaemia minor, recurrent epistaxis and microcytic anaemia from recurrent gastrointestinal angiectasia bleeding treated with fulguration and oral iron; his mother also had thalassaemia minor and clinically diagnosed HHT with epistaxis and cutaneous and gastrointestinal telangiectases [94]. He fulfilled three of four Curaçao criteria [94]. Examination showed finger and toe clubbing without cyanosis and right central facial palsy; blood gases showed PaO2 63.8 mmHg (saturation 93.4%) without dyspnoea [94].

Chest radiography showed nodular lesions in both lower lobes and cranial CT three space-occupying lesions, two frontal and one right parietal; levetiracetam and dexamethasone were started [94]. Differential diagnoses included malignancy and infection; blood, urine and cerebrospinal fluid cultures and HIV, syphilis, cysticercosis and mycobacterial serology were negative [94]. Magnetic resonance and positron emission tomography suggested brain abscesses (article figure 1: https://pmc.ncbi.nlm.nih.gov/articles/PMC9978536/figure/f1-amjcaserep-24-e938548/) [94]. Contrast-enhanced thoracoabdominal CT revealed bilateral pulmonary AVMs measuring 25, 30 and 26 mm (right upper, right lower and left lower lobes), and saline transoesophageal echocardiography showed delayed right-to-left bubble passage without endocarditis [94]. With presumed septic embolic abscesses, he received 8 weeks of ceftriaxone and metronidazole; neurological symptoms resolved and lesions disappeared on follow-up CT [94]. AVMs larger than 2 cm were selectively embolised (article figure 2: https://pmc.ncbi.nlm.nih.gov/articles/PMC9978536/figure/f2-amjcaserep-24-e938548/) and chronic hypoxaemia normalised (PaO2 87 mmHg); three months after discharge there were no cerebral sequelae [94]. Genetic testing identified ENG exon 5 variants whose HGVS nomenclature is not given in the article [94].

Three observations contextualise the case. First, it illustrates the trajectory guidelines aim to prevent: a patient with epistaxis, gastrointestinal anaemia, an affected mother and clubbing was diagnosed only after a serious neurological complication despite fulfilling Curaçao criteria — precisely the CHORUS delay and the group in whom pulmonary AVM screening (I1) and embolisation (I3) would be preventive [16, 31]. Second, the article does not describe a feeding-artery-calibre embolisation protocol, and “AVMs larger than 2 cm” refers to malformation size, not the guideline feeding-artery threshold of 2–3 mm; residual AVMs justify CT follow-up and lifelong antibiotic prophylaxis [1, 16]. Third, the authors' discussion cites pulmonary AVM frequencies of 5–15% in HHT, below reference sources (30–50%) [1, 94]. Genetic testing now enables predictive offspring testing and assessment of at-risk relatives [16, 94].

10.2. Thoracic aortopathy in two SMAD4 carriers

Teekakirikul and colleagues described two unrelated patients with juvenile polyposis and SMAD4 variants with thoracic aortic dilatation [43]. The first, a White boy, had anaemia and massive haematochezia at age 6, with multiple juvenile polyps showing low-grade dysplasia, and underwent total colectomy at age 10; he had an apparently de novo 28-base-pair duplication in exon 10 (published as c.1340_1367dup28 in NM_005359.5; normalised as NM_005359.6:c.1349_1376dup, p.(Ala460GlyfsTer43), pathogenic/likely pathogenic in ClinVar) [41–43]. At age 11, without epistaxis or marfanoid signs, echocardiography showed mild-to-moderate aortic annular dilatation (2.40 cm; z-score 4.19) and root dilatation (3.29 cm; z-score 4.04) and mild sinotubular junction and ascending aortic dilatation; cardiac CT confirmed dilatation and revealed two small pulmonary AVMs (2 and 3 mm), while brain magnetic resonance imaging was normal; shortly thereafter spontaneous epistaxis occurred, and by age 12 aortic dilatation had not progressed [43].

The second patient, aged 34, had colonic polyps at age 6, partial colectomy and multiple polypectomies; HHT was diagnosed at age 28 following a pulmonary AVM and epistaxis, and hepatic focal nodular hyperplasia subsequently appeared; she also had a percutaneously closed patent foramen ovale, transient ischaemic attacks, migraine and myoclonic epilepsy [43]. She had marfanoid habitus (32° scoliosis, arachnodactyly, Beighton 9/9) and met clinical Marfan syndrome criteria; aortic root diameter was 3.8 cm (upper limit 3.7 cm), progressively increasing from 2.9 to 3.8 cm over 10 years [43]. Her tall father, with marfanoid signs, colonic polyps and clinically diagnosed HHT, died suddenly aged 60 from ascending aortic dissection and rupture, with a preceding 5.4 cm aneurysm [43]. She carried recurrent variant c.1245_1248delCAGA (NM_005359.5), now NM_005359.6:c.1245_1248del, p.(Asp415GlufsTer20), pathogenic in ClinVar; TGFBR1 and TGFBR2 sequencing was negative and FBN1 was not sequenced at her request [41–43].

The article has no figures; clinical data appear in tables I and II [43]. The authors proposed thoracic aorta screening in SMAD4 carriers to prevent dissection deaths, a position now reflected in the Oxford Desk Reference, GeneReviews and Orphanet, without established starting age or follow-up interval; including SMAD4 in the Clinical Cardiogenetics heritable thoracic aortic disease gene table accords with these observations [2, 12, 32, 38, 43]. As Marfan syndrome was not molecularly excluded in the second patient, the isolated contribution of SMAD4 to her skeletal phenotype cannot be asserted confidently.

11. Prognosis and Long-Term Follow-up

11.1. Prognosis

Telangiectases and AVMs tend to increase in number and size with age [12]. In the Danish population study, overall mortality was slightly increased but doubled below age 60 at entry, an excess entirely explained by severe HHT manifestations [92]. In the British primary-care study, HHT was associated with death HR 2.0 (95% CI 1.6–2.6) and median age at death 3 years below controls [46]. Smith's textbook states that fewer than 10% die from complications [26]. According to Orphanet, life expectancy is reduced in unscreened patients but comparable with the general population after pulmonary AVM assessment and treatment at an HHT centre [2]. Hepatic AVM mortality is 1.1% per person-year and post-transplant survival 82–92% at 5–10 years [16]. Bleeding is the principal cause of hospitalisation and impaired quality of life [16].

11.2. Follow-up plan

Repeat intervals are not trial-based and centre practice varies: among North American centres of excellence, 26 of 30 rescreen adults for pulmonary AVMs, mostly every 5 years [86]. Transition to adult care must preserve pulmonary, cerebral and, for SMAD4 carriers, gastrointestinal and aortic surveillance [1, 16].

12. Summary and Final Considerations

Hereditary haemorrhagic telangiectasia is an autosomal dominant vascular dysplasia caused by loss of function of BMP9/BMP10–ALK1–endoglin–SMAD4 pathway components: ACVRL1 (52%) and ENG (44%) explain almost all cases, SMAD4 approximately 1% with combined juvenile polyposis syndrome, and GDF2 a rare form with still uncertain status [1, 11]. Clinical prevalence of approximately 1 in 5,000 is probably underestimated, and diagnostic delay exceeds two decades [1, 13, 31].

Clinical diagnosis rests on Curaçao criteria, highly specific but poorly sensitive in children and those with few signs; an isolated pulmonary AVM should prompt genetic testing, and familial variant testing clarifies the status of children of an affected parent [15, 16, 45]. Panels should include differential diagnosis genes, particularly EPHB4, RASA1, GDF2 and BMPR2, with deletion/duplication analysis [1].

Management combines bleeding control — nasal moisturisation, tranexamic acid, ablation, antiangiogenic agents (bevacizumab, pomalidomide, pazopanib), nasal surgery, oral or intravenous iron — and AVM complication prevention through pulmonary AVM screening and embolisation, individualised cerebral AVM decisions and assessment of complicated hepatic AVMs with transplantation for refractory cases [1, 16]. No medicine is authorised for HHT; tranexamic acid and pomalidomide have randomised trials showing epistaxis benefit, engasertib entered phase 3 and several new-mechanism agents began trials in 2026 [17, 65, 72, 82]. Real divergences persist, particularly in asymptomatic cerebral screening between international guidelines, American practice and VASCERN's position [16, 22, 47].

In SMAD4 carriers, HHT, juvenile polyposis and aortopathy coexist and require simultaneous vascular, endoscopic and aortic surveillance [1, 32, 38]. Pregnancy is the period of greatest risk of pulmonary haemorrhage and requires prior screening and specialist management [16, 48]. The organisation of care in reference centres and networks such as VASCERN, with the participation of patient associations, is part of the prevention strategy [16, 25].

13. References

1. McDonald J, Gossage JR, Stevenson DA. Hereditary hemorrhagic telangiectasia. 2000 Jun 26 [updated 2026 Feb 19]. In: Adam MP, Bick S, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. PMID: 20301525. Available at: https://www.ncbi.nlm.nih.gov/books/NBK1351/
2. Orphanet. Hereditary hemorrhagic telangiectasia (ORPHA:774) [Internet]. Summary updated January 2019; definition and classification of 2026 Jul 2. Accessed 2026 Oct 7. Available at: https://www.orpha.net/en/disease/detail/774
3. Online Mendelian Inheritance in Man (OMIM). Entry identifiers 131195 (ENG), 601284 (ACVRL1), 600993 (SMAD4), 605120 (GDF2) and 139210 (Myhre syndrome) [Internet]. Baltimore (MD): Johns Hopkins University. Titles confirmed through NCBI E-utilities on 2026 Oct 7. Available at: https://omim.org/
4. Online Mendelian Inheritance in Man (OMIM). #187300 Telangiectasia, hereditary hemorrhagic, type 1; HHT1 [Internet]. Baltimore (MD): Johns Hopkins University. Accessed 2026 Oct 7 (full text). Available at: https://omim.org/entry/187300
5. Online Mendelian Inheritance in Man (OMIM). %610655 Telangiectasia, hereditary hemorrhagic, type 4; HHT4 [Internet]. Baltimore (MD): Johns Hopkins University. Accessed 2026 Oct 7. Available at: https://omim.org/entry/610655
6. GeneGuidance. Genetic curation database: ENG, ACVRL1, SMAD4 and GDF2 gene records [local database]. Version of 2026 Sep 26 (identifiers, clinical curation, ClinGen, frequencies and listed references). Accessed 2026 Oct 7.
7. Online Mendelian Inheritance in Man (OMIM). #600376 Telangiectasia, hereditary hemorrhagic, type 2; HHT2 [Internet]. Baltimore (MD): Johns Hopkins University. Accessed 2026 Oct 7. Available at: https://omim.org/entry/600376
8. Online Mendelian Inheritance in Man (OMIM). #175050 Juvenile polyposis/hereditary hemorrhagic telangiectasia syndrome; JPHT [Internet]. Baltimore (MD): Johns Hopkins University. Accessed 2026 Oct 7. Available at: https://omim.org/entry/175050
9. Online Mendelian Inheritance in Man (OMIM). #615506 Telangiectasia, hereditary hemorrhagic, type 5; HHT5 [Internet]. Baltimore (MD): Johns Hopkins University. Accessed 2026 Oct 7. Available at: https://omim.org/entry/615506
10. National Center for Biotechnology Information. NCBI Datasets: product report for genes ENG (2022), ACVRL1 (94), SMAD4 (4089) and GDF2 (2658), transcripts MANE Select [Internet]. Accessed 2026 Oct 7. Available at: https://www.ncbi.nlm.nih.gov/datasets/gene/
11. McDonald J, Bayrak-Toydemir P, DeMille D, Wooderchak-Donahue W, Whitehead K. Curaçao diagnostic criteria for hereditary hemorrhagic telangiectasia is highly predictive of a pathogenic variant in ENG or ACVRL1 (HHT1 and HHT2). Genet Med. 2020;22(7):1201-5. doi:10.1038/s41436-020-0775-8. PMID: 32300199.
12. Firth HV, Hurst JA. Hereditary haemorrhagic telangiectasia (HHT). In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 446–7.
13. Anzell AR, White CM, Diergaarde B, Carlson JC, Roman BL. Hereditary hemorrhagic telangiectasia prevalence estimates calculated from gnomAD allele frequencies of predicted pathogenic variants in ENG and ACVRL1. Circ Genom Precis Med. 2025;18(5):e005061. doi:10.1161/CIRCGEN.124.005061. PMID: 40964703.
14. Westermann CJ, Rosina AF, De Vries V, de Coteau PA. The prevalence and manifestations of hereditary hemorrhagic telangiectasia in the Afro-Caribbean population of the Netherlands Antilles: a family screening. Am J Med Genet A. 2003;116A(4):324-8. doi:10.1002/ajmg.a.10002. PMID: 12522784.
15. Shovlin CL, Guttmacher AE, Buscarini E, Faughnan ME, Hyland RH, Westermann CJ, et al. Diagnostic criteria for hereditary hemorrhagic telangiectasia (Rendu-Osler-Weber syndrome). Am J Med Genet. 2000;91(1):66-7. doi:10.1002/(sici)1096-8628(20000306)91:1<66::aid-ajmg12>3.0.co;2-p. PMID: 10751092.
16. Faughnan ME, Mager JJ, Hetts SW, Palda VA, Lang-Robertson K, Buscarini E, et al. Second International Guidelines for the Diagnosis and Management of Hereditary Hemorrhagic Telangiectasia. Ann Intern Med. 2020;173(12):989-1001 (including Supplement 1, full guideline and additional tables, and the Table of first-guideline recommendations remaining applicable). doi:10.7326/M20-1443. PMID: 32894695.
17. Al-Samkari H, Hessels J, Riera-Mestre A, Dupuis-Girod S, Van Zele T, Gómez Del Olmo V, et al. Engasertib versus placebo for bleeding in hereditary hemorrhagic telangiectasia. N Engl J Med. 2025;393(21):2131-41. doi:10.1056/NEJMoa2504411. PMID: 41297007.
18. Vaderis Therapeutics. Vaderis Therapeutics announces oversubscribed $152 million Series B financing and initiation of the global Phase 3 HEROIC study of engasertib for hereditary hemorrhagic telangiectasia [press release]. 2026 Aug 11. Accessed 2026 Oct 7. Available at: https://www.placera.se/pressmeddelanden/vaderis-therapeutics-announces-oversubscribed-152-million-series-b-financing-and-initiation-of-the-global-phase-3-heroic-study-of-engasertib-for-hereditary-hemorrhagic-telangiectasia-20260811
19. European Medicines Agency. Orblid (bevacizumab), EMEA/H/C/006392: withdrawn application — treatment of adult patients with hereditary haemorrhagic telangiectasia; applicant Laboratoires Delbert; withdrawn on 2026 May 21 (published 2026 May 22; updated 2026 Aug 4) [Internet]. Accessed 2026 Oct 7. Available at: https://www.ema.europa.eu/en/medicines/human/EPAR/orblid
20. European Medicines Agency. Medicines output – orphan designations (list of designations) and individual pages for designations EU/3/10/730, EU/3/14/1367, EU/3/14/1390, EU/3/17/1845, EU/3/18/2087, EU/3/23/2766 and EU/3/25/3098 for the treatment of HHT [Internet]. Accessed 2026 Oct 7. Available at: https://www.ema.europa.eu/en/medicines/human/orphan-designations/eu-3-14-1390
21. Faughnan ME, Palda VA, Garcia-Tsao G, Geisthoff UW, McDonald J, Proctor DD, et al. International guidelines for the diagnosis and management of hereditary haemorrhagic telangiectasia. J Med Genet. 2011;48(2):73-87. doi:10.1136/jmg.2009.069013. PMID: 19553198.
22. Eker OF, Boccardi E, Sure U, Patel MC, Alicante S, Alsafi A, et al. European Reference Network for Rare Vascular Diseases (VASCERN) position statement on cerebral screening in adults and children with hereditary haemorrhagic telangiectasia (HHT). Orphanet J Rare Dis. 2020;15(1):165. doi:10.1186/s13023-020-01386-9. PMID: 32600364.
23. Haute Autorité de Santé. Protocole National de Diagnostic et de Soins (PNDS): Maladie de Rendu-Osler. Centre de référence de la maladie de Rendu-Osler (Hospices Civils de Lyon); filière de santé maladies rares FAVA-Multi. Saint-Denis La Plaine: HAS; 2024 update (published 9 Aug 2024). Available at: https://www.has-sante.fr/jcms/c_875183
24. Silvain C, Thévenot T, Colle I. Atteintes hépatiques au cours de la maladie de Rendu-Osler: recommandations AFEF 2018 (maladies vasculaires du foie). Hépato-Gastro et Oncologie digestive. 2019. National diagnosis and care protocol drawn up by the reference centre for vascular liver diseases; HAS page published 28 Jan 2019: https://www.has-sante.fr/jcms/c_2900212; text: https://www.filfoie.com/wp-content/uploads/2019/01/Atteintes-h%C3%A9patiques-au-cours-de-la-maladie-de-Rendu-Osler.pdf
25. Shovlin CL, Buscarini E, Sabbà C, Mager HJ, Kjeldsen AD, Pagella F, et al. The European Rare Disease Network for HHT Frameworks for management of hereditary haemorrhagic telangiectasia in general and speciality care. Eur J Med Genet. 2022;65(1):104370. doi:10.1016/j.ejmg.2021.104370. PMID: 34737116.
26. Jones KL, Jones MC, del Campo M. Hereditary hemorrhagic telangiectasia (Osler hemorrhagic telangiectasia). In: Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 744–5.
27. McAllister KA, Grogg KM, Johnson DW, Gallione CJ, Baldwin MA, Jackson CE, et al. Endoglin, a TGF-beta binding protein of endothelial cells, is the gene for hereditary haemorrhagic telangiectasia type 1. Nat Genet. 1994;8(4):345-51. doi:10.1038/ng1294-345. PMID: 7894484.
28. Johnson DW, Berg JN, Baldwin MA, Gallione CJ, Marondel I, Yoon SJ, et al. Mutations in the activin receptor-like kinase 1 gene in hereditary haemorrhagic telangiectasia type 2. Nat Genet. 1996;13(2):189-95. doi:10.1038/ng0696-189. PMID: 8640225.
29. Gallione CJ, Repetto GM, Legius E, Rustgi AK, Schelley SL, Tejpar S, et al. A combined syndrome of juvenile polyposis and hereditary haemorrhagic telangiectasia associated with mutations in MADH4 (SMAD4). Lancet. 2004;363(9412):852-9. doi:10.1016/S0140-6736(04)15732-2. PMID: 15031030.
30. Wooderchak-Donahue WL, McDonald J, O'Fallon B, Upton PD, Li W, Roman BL, et al. BMP9 mutations cause a vascular-anomaly syndrome with phenotypic overlap with hereditary hemorrhagic telangiectasia. Am J Hum Genet. 2013;93(3):530-7. doi:10.1016/j.ajhg.2013.07.004. PMID: 23972370.
31. Al-Samkari H, Friday C, Kasthuri RS, Gossage JR, Murphy CG, Hountras P, et al. Clinical spectrum of hereditary hemorrhagic telangiectasia: data from the Comprehensive HHT Outcomes Registry of the US (CHORUS). Blood. 2026;148(4):417-32. doi:10.1182/blood.2026033112. PMID: 41843464.
32. Mulder BJM, van de Laar IMBH, De Backer J. Heritable thoracic aortic diseases: syndromal and isolated (F)TAAD. In: Baars HF, van der Smagt JJ, Doevendans PAFM, van Tintelen JP, editors. Clinical Cardiogenetics. 3rd ed. Cham: Springer; 2020. p. 309–11 (sections 19.1–19.2 and Table 19.1). doi:10.1007/978-3-030-45457-9_19.
33. Cole SG, Begbie ME, Wallace GM, Shovlin CL. A new locus for hereditary haemorrhagic telangiectasia (HHT3) maps to chromosome 5. J Med Genet. 2005;42(7):577-82. doi:10.1136/jmg.2004.028712. PMID: 15994879.
34. Lesca G, Olivieri C, Burnichon N, Pagella F, Carette MF, Gilbert-Dussardier B, et al. Genotype-phenotype correlations in hereditary hemorrhagic telangiectasia: data from the French-Italian HHT network. Genet Med. 2007;9(1):14-22. doi:10.1097/gim.0b013e31802d8373. PMID: 17224686.
35. Palermo M, Cocilovo F, Trevisi G, Lucci Cordisco E, Di Martino L, Sonnini E, et al. Genotype-driven cerebrovascular risk across age and sex in hereditary hemorrhagic telangiectasia. Neurol Sci. 2026;47(5):428. doi:10.1007/s10072-026-09018-z. PMID: 41989642.
36. Geiger H, Tinker RJ, Talbert J, Phillips JA, Austin ED. Heritable pulmonary arterial hypertension overview. 2002 Jul 18 [updated 2026 Apr 9]. In: Adam MP, Bick S, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. PMID: 20301658. Available at: https://www.ncbi.nlm.nih.gov/books/NBK1485/
37. Firth HV, Hurst JA. Juvenile polyposis syndrome (JPS). In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 578–80.
38. Larsen Haidle J, MacFarland SP, Howe JR. Juvenile polyposis syndrome. 2003 May 13 [updated 2022 Feb 3]. In: Adam MP, Bick S, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. PMID: 20301642. Available at: https://www.ncbi.nlm.nih.gov/books/NBK1469/
39. Jones KL, Jones MC, del Campo M. Myhre syndrome. In: Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 692–7.
40. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Myhre syndrome (MIM 139210). In: Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford: Oxford University Press; 2018. p. 490–3.
41. Freeman PJ, Hart RK, Gretton LJ, Brookes AJ, Dalgleish R. VariantValidator: accurate validation, mapping, and formatting of sequence variation descriptions. Hum Mutat. 2018;39(1):61-8. doi:10.1002/humu.23348. PMID: 28967166. Service accessed 2026 Oct 7 (GRCh38, MANE Select: NM_005359.6:c.1245_1248del; NM_005359.6:c.1340_1367dup normalised to c.1349_1376dup). Available at: https://rest.variantvalidator.org
42. National Center for Biotechnology Information. ClinVar [Internet]: VCV000142253, NM_005359.6(SMAD4):c.1245_1248del (pathogenic; multiple submitters, no conflicts; last evaluated 2025 Dec 15); VCV000233900, NM_005359.6(SMAD4):c.1349_1376dup (pathogenic/likely pathogenic; last evaluated 2025 Apr 1); ENG and ACVRL1 records reviewed by the HHT expert panel. Accessed 2026 Oct 7. Available at: https://www.ncbi.nlm.nih.gov/clinvar/variation/142253/
43. Teekakirikul P, Milewicz DM, Miller DT, Lacro RV, Regalado ES, Rosales AM, et al. Thoracic aortic disease in two patients with juvenile polyposis syndrome and SMAD4 mutations. Am J Med Genet A. 2013;161A(1):185-91. doi:10.1002/ajmg.a.35659. PMID: 23239472. Full text: https://pmc.ncbi.nlm.nih.gov/articles/PMC3535513/
44. Shovlin CL, Condliffe R, Donaldson JW, Kiely DG, Wort SJ; British Thoracic Society. British Thoracic Society Clinical Statement on Pulmonary Arteriovenous Malformations. Thorax. 2017;72(12):1154-63. doi:10.1136/thoraxjnl-2017-210764. PMID: 29141890.
45. Anderson E, Sharma L, Alsafi A, Shovlin CL. Pulmonary arteriovenous malformations may be the only clinical criterion present in genetically confirmed hereditary haemorrhagic telangiectasia. Thorax. 2022;77(6):628-30. doi:10.1136/thoraxjnl-2021-218332. PMID: 35165143.
46. Donaldson JW, McKeever TM, Hall IP, Hubbard RB, Fogarty AW. Complications and mortality in hereditary hemorrhagic telangiectasia: a population-based study. Neurology. 2015;84(18):1886-93. doi:10.1212/WNL.0000000000001538. PMID: 25862798.
47. Beslow LA, White AJ, Krings T, Hammill AM, Lang SS, Baba A, et al. Current practice: rationale for screening children with hereditary hemorrhagic telangiectasia for brain vascular malformations. AJNR Am J Neuroradiol. 2024;45(9):1177-84. doi:10.3174/ajnr.A8195. PMID: 38816017.
48. Shovlin CL, Sodhi V, McCarthy A, Lasjaunias P, Jackson JE, Sheppard MN. Estimates of maternal risks of pregnancy for women with hereditary haemorrhagic telangiectasia (Osler-Weber-Rendu syndrome): suggested approach for obstetric services. BJOG. 2008;115(9):1108-15. doi:10.1111/j.1471-0528.2008.01786.x. PMID: 18518871.
49. Bayrak-Toydemir P, Stevenson DA. Capillary malformation-arteriovenous malformation syndrome. 2011 Feb 22 [updated 2019 Sep 12]. In: Adam MP, Bick S, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. PMID: 21348050. Available at: https://www.ncbi.nlm.nih.gov/books/NBK52764/
50. Jones KL, Jones MC, del Campo M. Parkes Weber syndrome (capillary malformation-arteriovenous malformation syndrome, RASA1 related disorders). In: Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 728–9.
51. Al-Samkari H, Kasthuri RS, Mager HJ, Zhou JY, Serra MM, Samuelson-Bannow BT, et al. Standardization of terminology, definitions, and outcome criteria for bleeding in hereditary hemorrhagic telangiectasia: international consensus report. Am J Hematol. 2025;100(10):1813-27. doi:10.1002/ajh.70011. PMID: 40662351.
52. Whitehead KJ, Sautter NB, McWilliams JP, Chakinala MM, Merlo CA, Johnson MH, et al. Effect of topical intranasal therapy on epistaxis frequency in patients with hereditary hemorrhagic telangiectasia: a randomized clinical trial. JAMA. 2016;316(9):943-51. doi:10.1001/jama.2016.11724. PMID: 27599329.
53. Zhang E, Kasthuri RS, Parambil J, Prasad V, Iyer VN, Whitehead KJ, et al. Pomalidomide for hereditary hemorrhagic telangiectasia: after trial longitudinal assessment study (PATH-HHT ATLAS). Blood Adv. 2026;10(5):1799-808. doi:10.1182/bloodadvances.2025018382. PMID: 41512167.
54. Teofarma S.r.l. Ferro-Gradumet 329,7 mg comprimidos de libertação prolongada (sulfato ferroso): Resumo das Características do Medicamento [Internet]. Approved by INFARMED on 2022 Mar 11. INFARMED — Infomed. Accessed 2026 Oct 26. Available at: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=3387&tipo_doc=rcm
55. Vifor France. Ferinject 50 mg ferro/ml dispersão injetável ou para perfusão (carboximaltose férrica): Resumo das Características do Medicamento [Internet]. INFARMED — Infomed. Accessed 2026 Oct 26. Available at: https://extranet.infarmed.pt/INFOMED-fo/
56. American Regent, Inc. INJECTAFER (ferric carboxymaltose injection): prescribing information [Internet]. Effective version 2026 Aug 19 (openFDA/DailyMed). Accessed 2026 Oct 26. Available at: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=517b4a19-45b3-4286-9f6a-ced4e10447de
57. AMAG Pharmaceuticals, Inc. FERAHEME (ferumoxytol injection): prescribing information [Internet]. Effective version 2026 May 31 (openFDA/DailyMed). Accessed 2026 Oct 26. Available at: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=32b0e320-a739-11dc-a704-0002a5d5c51b
58. Pharmacosmos Therapeutics Inc. MONOFERRIC (ferric derisomaltose injection): prescribing information [Internet]. Effective version 2025 Nov 25 (openFDA/DailyMed). Accessed 2026 Oct 26. Available at: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=55859d2d-0456-4fa9-b41f-f535accc97db
59. Pharma Bavaria Internacional (PBI) Portugal, Unipessoal Lda. Tafixyl 500 mg comprimidos revestidos por película (ácido tranexâmico): Resumo das Características do Medicamento [Internet]. Text revised 2026 Jul. INFARMED — Infomed. Accessed 2026 Oct 26. Available at: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=606502&tipo_doc=rcm
60. Nordic Pharma, Inc. Tranexamic acid tablets USP 650 mg: prescribing information [Internet]. Effective version 2025 Sep 29 (openFDA/DailyMed). Accessed 2026 Oct 7. Available at: https://dailymed.nlm.nih.gov/dailymed/search.cfm?query=tranexamic+acid+tablets
61. Genentech, Inc. AVASTIN (bevacizumab) injection: prescribing information [Internet]. Effective version 2025 Jan 6 (openFDA/DailyMed). Accessed 2026 Oct 7. Available at: https://dailymed.nlm.nih.gov/dailymed/search.cfm?query=AVASTIN
62. Buscarini E, Botella LM, Geisthoff U, Kjeldsen AD, Mager HJ, Pagella F, et al. Safety of thalidomide and bevacizumab in patients with hereditary hemorrhagic telangiectasia. Orphanet J Rare Dis. 2019;14(1):28. doi:10.1186/s13023-018-0982-4. PMID: 30717761.
63. Celgene Corporation. POMALYST (pomalidomide) capsules: prescribing information [Internet]. Effective version 2025 Feb 26 (openFDA/DailyMed). Accessed 2026 Oct 7. Available at: https://dailymed.nlm.nih.gov/dailymed/search.cfm?query=POMALYST
64. European Medicines Agency. Imnovid (pomalidomide): EPAR [Internet]. Authorised 2013 Aug 5. Accessed 2026 Oct 7. Available at: https://www.ema.europa.eu/en/medicines/human/EPAR/imnovid
65. Al-Samkari H, Kasthuri RS, Iyer VN, Pishko AM, Decker JE, Weiss CR, et al. Pomalidomide for epistaxis in hereditary hemorrhagic telangiectasia. N Engl J Med. 2024;391(11):1015-27. doi:10.1056/NEJMoa2312749. PMID: 39292928.
66. Novartis Pharmaceuticals Corporation. VOTRIENT (pazopanib) tablets: prescribing information [Internet]. Effective version 2025 Dec 16 (openFDA/DailyMed). Accessed 2026 Oct 7. Available at: https://dailymed.nlm.nih.gov/dailymed/search.cfm?query=VOTRIENT
67. European Medicines Agency. Votrient (pazopanib): EPAR [Internet]. Authorised 2010 Jun 14. Accessed 2026 Oct 7. Available at: https://www.ema.europa.eu/en/medicines/human/EPAR/votrient
68. Cure HHT. Effects of pazopanib on hereditary hemorrhagic telangiectasia related epistaxis and anemia (phase 2/3, randomised, placebo-controlled). ClinicalTrials.gov NCT03850964 [Internet]. Updated 2026 Mar 19. Accessed 2026 Oct 7. Available at: https://clinicaltrials.gov/study/NCT03850964
69. Al-Samkari H, Kasthuri RS, Parambil JG, Albitar HA, Almodallal YA, Vázquez C, et al. An international, multicenter study of intravenous bevacizumab for bleeding in hereditary hemorrhagic telangiectasia: the InHIBIT-Bleed study. Haematologica. 2021;106(8):2161-9. doi:10.3324/haematol.2020.261859. PMID: 32675221.
70. Haahr PD, Kjeldsen AD, Fialla AD, Fargeton AE, Guilhem A, Buscarini E, et al. Availability, use, efficacy and safety of bevacizumab in European hereditary haemorrhagic telangiectasia centres. Br J Clin Pharmacol. 2026;92(2):535-44. doi:10.1002/bcp.70273. PMID: 40985322.
71. Dupuis-Girod S, Ambrun A, Decullier E, Fargeton AE, Roux A, Bréant V, et al. Effect of bevacizumab nasal spray on epistaxis duration in hereditary hemorrhagic telangectasia: a randomized clinical trial. JAMA. 2016;316(9):934-42. doi:10.1001/jama.2016.11387. PMID: 27599328.
72. Gaillard S, Dupuis-Girod S, Boutitie F, Rivière S, Morinière S, Hatron PY, et al. Tranexamic acid for epistaxis in hereditary hemorrhagic telangiectasia patients: a European cross-over controlled trial in a rare disease. J Thromb Haemost. 2014;12(9):1494-502. doi:10.1111/jth.12654. PMID: 25040799.
73. Geisthoff UW, Seyfert UT, Kübler M, Bieg B, Plinkert PK, König J. Treatment of epistaxis in hereditary hemorrhagic telangiectasia with tranexamic acid – a double-blind placebo-controlled cross-over phase IIIB study. Thromb Res. 2014;134(3):565-71. doi:10.1016/j.thromres.2014.06.012. PMID: 25005464.
74. Dupuis-Girod S, Ginon I, Saurin JC, Marion D, Guillot E, Decullier E, et al. Bevacizumab in patients with hereditary hemorrhagic telangiectasia and severe hepatic vascular malformations and high cardiac output. JAMA. 2012;307(9):948-55. doi:10.1001/jama.2012.250. PMID: 22396517.
75. Hermann R, Grobost V, Le-Guillou X, Lavigne C, Parrot A, Rivière S, et al. Effect of oral nintedanib vs placebo on epistaxis in hereditary hemorrhagic telangiectasia: the EPICURE multicenter randomized double-blind trial. Angiogenesis. 2024;28(1):9. doi:10.1007/s10456-024-09962-4. PMID: 39718659.
76. Hessels J, Kroon S, Boerman S, Nelissen RC, Grutters JC, Snijder RJ, et al. Efficacy and safety of tacrolimus as treatment for bleeding caused by hereditary hemorrhagic telangiectasia: an open-label, pilot study. J Clin Med. 2022;11(18):5280. doi:10.3390/jcm11185280. PMID: 36142926.
77. Al-Samkari H, Thomas SM, Marsh D, Kasthuri RS, Iyer VN, Pishko AM, et al. Characteristics associated with clinical response to pomalidomide in hereditary hemorrhagic telangiectasia. Blood Adv. 2026;10(9):2967-76. doi:10.1182/bloodadvances.2025019484. PMID: 41719457.
78. Hospices Civils de Lyon. BABH study: efficacy and safety of bevacizumab on severe bleedings associated with hemorrhagic hereditary telangiectasia (phase 3). ClinicalTrials.gov NCT03227263 [Internet]. Updated 2025 Aug 6. Accessed 2026 Oct 7. Available at: https://clinicaltrials.gov/study/NCT03227263
79. U.S. Food and Drug Administration. Drugs@FDA (openFDA drugsfda): BLA 125085 (Avastin), NDA 204026 (Pomalyst), NDA 022465 (Votrient), NDA 022430 (Lysteda) — original approval dates [Internet]. Accessed 2026 Oct 7. Available at: https://www.accessdata.fda.gov/scripts/cder/daf/
80. European Medicines Agency. Avastin (bevacizumab): EPAR – medicine overview and product information [Internet]. Authorised 2005 Jan 12. Accessed 2026 Oct 7. Available at: https://www.ema.europa.eu/en/medicines/human/EPAR/avastin
81. U.S. Food and Drug Administration. Search Orphan Drug Designations and Approvals: pazopanib, treatment of hereditary hemorrhagic telangiectasia (designated 2019 Oct 9; sponsor HHT Foundation International d/b/a Cure HHT) [Internet]. Accessed 2026 Oct 7. Available at: https://www.accessdata.fda.gov/scripts/opdlisting/oopd/detailedIndex.cfm?cfgridkey=705819
82. Vaderis Therapeutics AG. A trial to assess the efficacy and safety of engasertib in participants with moderate to severe hereditary hemorrhagic telangiectasia (HHT) (phase 3). ClinicalTrials.gov NCT07743671 [Internet]. Updated 2026 Sep 17. Accessed 2026 Oct 7. Available at: https://clinicaltrials.gov/study/NCT07743671
83. Diagonal Therapeutics, Inc. DIAG723 in adults with hereditary hemorrhagic telangiectasia (DIAMOND; phase 1/2). ClinicalTrials.gov NCT07623525 [Internet]. Updated 2026 Sep 8. Accessed 2026 Oct 7. Available at: https://clinicaltrials.gov/study/NCT07623525
84. European Medicines Agency. EU/3/25/3098 – orphan designation for treatment of hereditary haemorrhagic telangiectasia (human bispecific monoclonal antibody targeting ALK1 and BMPRII; designated 2025 Jul 18) [Internet]. Accessed 2026 Oct 7. Available at: https://www.ema.europa.eu/en/medicines/human/orphan-designations/eu-3-25-3098
85. AUGMENTIN (amoxicillin/clavulanate potassium) tablets and powder for oral suspension: prescribing information [Internet]. NDA 050575. Effective version 2024 May 16 (openFDA/DailyMed). Accessed 2026 Oct 26. Available at: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=04a31a42-d79e-bbb4-e063-6394a90a63ee
86. Beslow LA, Kim H, Hetts SW, Ratjen F, Clancy MS, Gossage JR, et al. Brain and lung arteriovenous malformation rescreening practices for children and adults with hereditary hemorrhagic telangiectasia. Orphanet J Rare Dis. 2024;19(1):421. doi:10.1186/s13023-024-03402-8. PMID: 39522006.
87. Atavistik Bio, Inc. Harmony-HHT: ATV-1601 in participants with hereditary hemorrhagic telangiectasia (phase 1/2). ClinicalTrials.gov NCT07601425 [Internet]. Updated 2026 Oct 5. Accessed 2026 Oct 7. Available at: https://clinicaltrials.gov/study/NCT07601425
88. Tectonic Operating Company, Inc. Phase 1b study of TX002100 in participants with hematologic support-dependent hereditary hemorrhagic telangiectasia. ClinicalTrials.gov NCT07863609 [Internet]. Updated 2026 Oct 7. Accessed 2026 Oct 7. Available at: https://clinicaltrials.gov/study/NCT07863609
89. Columbia University. Topical TOR-582 treatment of epistaxis in HHT (sirolimus ointment; phase 1). ClinicalTrials.gov NCT07667413 [Internet]. Updated 2026 Aug 21. Accessed 2026 Oct 7. Available at: https://clinicaltrials.gov/study/NCT07667413
90. Alnylam Pharmaceuticals. InsigHHT: a study to evaluate ALN-6400 in healthy volunteers and patients with hereditary hemorrhagic telangiectasia (phase 1/2). ClinicalTrials.gov NCT06659640 [Internet]. Updated 2026 Sep 16. Accessed 2026 Oct 7. Available at: https://clinicaltrials.gov/study/NCT06659640
91. Lazor R. Efficacy of nintedanib for treatment of epistaxis in hereditary hemorrhagic telangiectasia (phase 2). ClinicalTrials.gov NCT04976036 [Internet]. Updated 2026 Mar 11. Accessed 2026 Oct 7. Available at: https://clinicaltrials.gov/study/NCT04976036
92. Kjeldsen AD, Vase P, Green A. Hereditary haemorrhagic telangiectasia: a population-based study of prevalence and mortality in Danish patients. J Intern Med. 1999;245(1):31-9. doi:10.1046/j.1365-2796.1999.00398.x. PMID: 10095814.
93. Major T, Gindele R, Balogh G, Bárdossy P, Bereczky Z. Founder effects in hereditary hemorrhagic telangiectasia. J Clin Med. 2021;10(8):1682. doi:10.3390/jcm10081682. PMID: 33919892.
94. Hernández-de Sosa N, Cardona Corrales R, Casademont J. A 37-year-old man with familial hereditary hemorrhagic telangiectasia and pulmonary arteriovenous malformations presenting with seizures and a diagnosis of brain abscesses. Am J Case Rep. 2023;24:e938548. doi:10.12659/AJCR.938548. PMID: 36841939. Full text: https://pmc.ncbi.nlm.nih.gov/articles/PMC9978536/
95. ClinicalTrials.gov. Observational HHT registries: NCT06259292 (CHORUS, Cure HHT), NCT01761981 (Hospital Italiano de Buenos Aires) and NCT00230685 (Imperial College London) [Internet]. Accessed 2026 Oct 7. Available at: https://clinicaltrials.gov/search?cond=hereditary%20hemorrhagic%20telangiectasia