Hereditary Hemorrhagic Telangiectasia

Michelle Letarte

Claire L. Shovlin

Hereditary hemorrhagic telangiectasia (HHT), also known as Rendu-Osler-Weber syndrome, was first recognized as a genetic disorder in the 19th century. HHT is inherited as an autosomal dominant trait, and its prevalence is now estimated at 1 in 5,000 people. Arteriovenous malformations (AVMs), representing direct connections between arteries and veins, are the distinguishing features of HHT. Small AVMs, or telangiectasia, give rise to frequent nosebleeds in most patients and to gastrointestinal (GI) bleeds in older individuals; they can also be present on lips, tongue, ears, and fingers and increase with age. Larger AVMs are found in internal organs, primarily in lungs, liver, and brain in a large number of patients, and can give rise to serious complications. Two genes are responsible for more than 80% of cases of HHT: endoglin (ENG) for HHT1 (OMIM #187300) and ACVRL1 (activin receptor-like kinase 1, or ALK1) for HHT2 (OMIM #600376). A third gene, SMAD4 is mutated in 2% to 3% of HHT cases, co-presenting with juvenile polyposis (JP), a condition referred to as JPHT, OMIM #175050. These three genes code for members of the transforming growth factor-beta (TGF-β ) superfamily involved in regulating multiple cellular functions. Endoglin and ALK1 are transmembrane proteins primarily expressed in the vascular endothelium, where they mediate the effects of specific TGF-β superfamily growth factors on endothelial function and vasomotor tone. Reduced functional expression of these specialized receptors leads to dysregulation of pathways essential for vascular homeostasis and the generation of AVMs.

CLINICAL MANIFESTATIONS AND DIAGNOSIS OF HHT

Clinical Manifestations of HHT

The spectrum of HHT encompasses multiple organ systems and within these, numerous forms of disease. Patients presenting to hematologists usually have iron deficiency anemia due to sustained and repeated nasal and/or GI bleeding, and are recognized by characteristic mucocutaneous telangiectasia (FIGURE 70.1A). The majority of HHT patients are also affected by pulmonary (˜50%, FIGURE 70.1B) and/or hepatic (30% to 70%) AVMs, with cerebral and pancreatic AVMs affecting up to 20%.¹^,² These AVMs are characteristically silent unless detected by screening protocols. Non telangiectatic/AVM pathology in HHT includes pulmonary arterial hypertension³ and pulmonary hypertension in the context of high-output cardiac failure secondary to hepatic AVMs,² JP,⁴ and a prothrombotic state associated with elevated plasma levels of factor VIII.⁵ Reduction in life expectancy has been observed in younger individuals.^6a^,^6b

The spectrum of diseases associated with HHT has been reviewed recently.² In addition to age-related increases in the number of telangiectasia,²^,⁷ there are specific times in patients’ lives when their HHT pathologies become more hazardous, the most important of which is pregnancy that results in a 1.0% risk of maternal death per pregnancy (due to pulmonary AVM [PAVM] hemorrhage, cerebral hemorrhage, and thrombotic complications), with all maternal deaths occurring in women previously considered well.^8a

One of the emerging themes in HHT is the occurrence of pathologic thromboemboli in a condition that was formerly considered a purely hemorrhagic disorder.^8b HHT patients are not protected from general venous thromboembolism (VTE) risks, and thromboembolic strokes affect a high proportion of HHT patients with PAVMs. Provision of anticoagulant/antiplatelet therapies in the setting of chronic HHT hemorrhage may therefore be required.

Diagnosis of HHT

The international consensus diagnostic criteria for HHT are nosebleeds, mucocutaneous telangiectasia, visceral involvement, and a family history⁹ (Table 70.1). Anemia is not a criterion.⁹ An individual has a clinical diagnosis of “definite HHT” if three criteria are present; “suspected HHT” if two are present; and “unlikely HHT” if only one is present.⁹

Clinical diagnosis is more difficult in younger individuals and those who are paucisymptomatic. Clinical diagnosis may be assisted by detection of visceral involvement either for management of clinical problems or as a result of screening protocols (see below). Outside of HHT centers, it is critical that non-HHT specialists be aware of potential pitfalls in diagnoses, particularly in inappropriate designation of a visceral vascular malformation that is not HHT associated as a diagnostic criterion (Table 70.1). It is also important that when patients receive a diagnosis of visceral AVMs, that they receive appropriate information about recommended management to reduce potential complications that extend far beyond hemorrhagic considerations: provision of such information may be challenging outside of specialized HHT units.

After passing through a clinical diagnostic screen, many patients will be left with a diagnosis of suspected or unlikely HHT. If the HHT mutation has been identified in their family, genetic testing can confirm or refute the diagnosis for the individual. Where a clinical diagnosis of HHT has been made, DNA testing is not required to “confirm” the diagnosis.

If a mutation is not found for a family where individuals have definite clinical HHT, the absence of a detected mutation should not affect the clinical diagnosis. Mutations are not found in approximately 15% to 20% of HHT families.¹⁰^,¹¹^,¹² In addition, an estimated 10% to 20% of families have genetic variants of
uncertain significance, where a single-nucleotide change predicts an amino acid substitution that is not certain to sufficiently modify protein function to cause disease. Such “missense” changes have the potential to lead to misdiagnosis, as evidenced by families in which the alternate “real” mutation is also identified.¹³^,¹⁴

FIGURE 70.1 Clinical HHT. A: Tongue demonstrating typical florid telangiectasia. Note that these lesions increase with age and are often subtle or absent in younger individuals. B(i): Selective left pulmonary angiograms demonstrating large left lower lobe PAVM pretreatment, when it was associated with severe hypoxemia (SaO₂ 85% at rest) and normal pulmonary artery pressure (27/9, mean 16 mm Hg). B(ii): Following embolization of separate feeding arteries using four Amplatzer plugs, the SaO₂ rose to 95% at rest, but the pulmonary artery pressure was unchanged (28/8, mean 15 mm Hg). Angiograms courtesy of Dr James Jackson.

Organ-Specific Screening Issues in HHT

Symptomatic patients will usually first undergo investigations performed by the relevant specialists. From a hematologic perspective, patients are often referred with iron deficiency anemia due to recurrent nasal and/or GI bleeding. In addition, a major component of management for the individual with suspected or proven HHT is screening, that is the investigation for a manifestation of HHT for which they were not previously aware, in order to improve their long-term health.

The international guidelines recommend clinical screens for pulmonary, cerebral, and hepatic AVMs, and genetic tests.¹⁵ It is particularly in this context that overall management within, or in association with an HHT center, is recommended.¹⁵ As discussed recently,² there are complex issues involved in screening for certain manifestations of HHT, particularly cerebral AVMs in which investigations and treatments carry substantial risks, and hepatic AVMs for which no treatments are currently recommended for asymptomatic patients. Assumptions that screening tests will be negative are misguided, and detailed prescreening discussions are advised.

Pulmonary

PAVM screening is recommended for all patients with possible or confirmed HHT, based on evidence of long-term efficacy of PAVM embolization resulting in improved oxygenation¹⁵ and more recent evidence that PAVM embolization reduces stroke risk.¹⁶ For sufficiently sensitive screening tests, the choice lies between thoracic CT scans and contrast echocardiography (CE), each of which, in dedicated HHT screening units, provides very high sensitivity as a first-line screen. The international guidelines committee recommended CE as the initial test, with 96% agreement,¹⁵ but this may be reevaluated as the latest data manuscript calls into question the cost effectiveness of CE as the majority of contrast echocardiograms in HHT1 patients are positive, so the need for a subsequent CT scan may not be reduced.^17a Conversely, in less experienced centers, if a negative echocardiographic study may be used to withhold a CT scan for an HHT patient undergoing PAVM screening, it is essential that methodologic reasons for a possible false negative are not overlooked.^17b Where CT scans are ordered for PAVM screening, the rationale should be discussed with local radiologists, as use of standard protocols (CT-pulmonary angiogram and highresolution CT scan) may miss clinically important PAVMs. This is because CT-pulmonary angiogram protocols do not image the lower part of the lungs, while high-resolution CT scans have substantial gaps between the thin slices imaged, and both can miss PAVMs.

Screening for pulmonary hypertension is not generally recommended, but for HHT patients whose impaired exercise tolerance and dyspnea appear to be out of proportion to the level of their anemia, consideration of this possible pathology is appropriate. First-line screening is by echocardiography.

Cerebral

The international guidelines recommend screening adults and children with possible or definite HHT by magnetic resonance imaging, followed by referral to centers with neurovascular expertise for invasive imaging.¹⁵ This recommendation was made while recognizing the lack of evidence of treatment for asymptomatic individuals, and the risks of screening, stated as a 0.5% risk of permanent stroke per diagnostic angiogram.¹⁵ However, the interpretation of the limited evidence base in HHT (as in the general population) differs between health care systems. In some countries therefore, screening is only recommended for patients who are considered at higher risk of cerebral hemorrhage, for example, in the setting of neurologic symptoms, or a family history of cerebral hemorrhage.

Hepatic

The guidelines recommend hepatic AVM screening using Doppler studies to assist diagnosis when there are fewer than three diagnostic criteria, and clinical genetics is unhelpful.¹⁵
New data suggest that where screening investigations have already identified abnormal liver function tests in an HHT family, Doppler studies may also be of value to predict a group with more severe disease who may require different follow-up regimes.¹⁸

Table 70.1 HHT diagnostic criteria

Curaçao Criterion		Diagnostic Issues	Specificity for HHT
1.	Epistaxes (nosebleeds)	“Occur spontaneously on more than one occasion”⁹	Low¹²⁷ though likely to be enhanced if of higher severity¹⁰⁰
2.	Mucocutaneous telangiectasia	“Multiple, at characteristic sites—lips, oral cavity, fingers, nose”⁹	Not to be confused with actinic changes; spider naevi; Campbell de Morgan spots; or other cutaneous vascular lesions.^98,128
3.	Visceral lesions
	• GI telangiectasia	At endoscopy. With or without bleeding	Potential confusion with von Willebrands Disease, or other forms of angiodysplasia
	• Pulmonary AVM	Diagnosis usually by CT or CE	Single PAVM may occur sporadically, and 8% of controls have a positive contrast echo¹²⁹ but HHT accounted for >90% of PAVM diagnoses in one center¹³⁰
	• Hepatic AVM		Full discussion of diagnostic issues in¹³¹
	• Cerebral AVM		Does not include cavernous malformations or aneurysms (see^21,132 for spectrum)
	• Spinal AVM		Spectrum of non-HHT disease discussed¹³³
4.	Family history	“A first degree relative with HHT according to these criteria”⁹

Gastrointestinal

HHT patients with a SMAD4 mutation, or a family history of GI cancer, require endoscopic investigations and management according to standard guidance for GI cancer screening in the non-HHT population.¹⁹

Specialized Screening Circumstances

Pregnancy

HHT centers have recommended for many years PAVM screening and treatment before pregnancy. Radiation assessment studies indicated that embolization may be undertaken during pregnancy,²⁰ and routine embolization in the later stages of pregnancy is performed in some centers where prepregnancy treatment was not feasible.¹⁵ Spinal AVM screens form part of HHT pregnancy management in many units in order to facilitate the use of epidural analgesia if required during labor.^8a No means of distinguishing a group of HHT women more likely to have significant pregnancy-related complications has been identified, and there are no data that specific screening or treatment modalities reduce pregnancy-associated risks. Importantly, prior awareness of the HHT diagnosis improves the outcome for women experiencing a life-threatening event, most likely by enhanced patients’ awareness of significant symptoms, and obstetric preplanning in the event of major complications.^8a

Children

The issues regarding screening of children are also controversial, as evidenced by the low rate of agreement for dedicated pediatric guideline recommendations.¹⁵ There are clearly tragic cases of HHT-related deaths and disability in youngsters. Children with symptoms require specialist investigation and treatment guided by knowledge of HHT pathology. The most concerning pathology in childhood is a particular type of rare cerebrospinal arteriovenous fistula that appears to have a very high risk of hemorrhage, and high treatment-associated risks.²¹ Current North American practice includes screening for these lesions from the age of 6 months, and repeating screens to capture AVMs that develop later in childhood.²¹ Pediatric screening remains the matter of intense discussion in other health care systems.²

GENETICS OF HHT AND MOLECULAR DIAGNOSIS

HHT1 and HHT2

Mutations

ENG and ACVRL1 mutations are responsible for >80% of cases of HHT. ENG, located on chromosome 9q34 and coding for the protein endoglin, was the first gene identified as mutated in HHT, resulting in HHT type 1 (HHT1).²² ACVRL1, located on chromosome 12q31 encoding the ALK1 receptor, was then reported as mutated in HHT type 2 (HHT2).²³^,²⁴ Clinically, it is not possible to determine if a patient has HHT1 or HHT2, as there is extensive phenotypic heterogeneity among patients, even within the same family. A molecular diagnosis was developed and is now available in several countries, generally in association with
an HHT center, allowing for genetic counseling of the families, coordinated with patient care by expert physicians.

Results of research studies and diagnostic laboratories have led to the identification of hundreds of mutations in HHT patients.¹⁰^,²⁵^,²⁶^,²⁷^,²⁸^,²⁹^,³⁰^,³¹^,³²^,³³^,³⁴^,³⁵ The HHT mutation database (http://www.hhtmutation.org) currently lists 311 ENG mutations and 249 ACVRL1 mutations. The ENG gene has 15 exons (including exons 9a and 9b), and mutations have been reported in all exons but 14, the latter coding for the intracellular region. All types of mutations are present in the ENG gene, with deletions and insertions (from single nucleotides to multiple exons, inframe, or causing a frameshift) accounting for the largest proportion. Single-nucleotide substitutions including missense, splice site, and nonsense mutations are also well represented (FIGURE 70.2). The ACVRL1 gene has 10 exons, and the mutation frequency is highest in exons 3, 7, and 8. Single-nucleotide variants account for the majority with missense mutations being more numerous than stop codon and splice site mutations. Deletions and insertions are less frequent than in the ENG gene (FIGURE 70.2). Most mutations are unique but some have been reported in several families, such as in arginine codons 144, 374, 411, and 479 of ACVRL1. Therefore, mutation analysis requires the complete sequencing of both genes as well as the analysis of exon copy number.

FIGURE 70.2 HHT mutations. Distribution of different types of mutations in ENG, ACVRL1, and SMAD4 genes. PTC: premature termination codon generated directly, or by frameshift mutations. Note the higher proportion of missense single-nucleotide substitutions in ACVRL1. The graph illustrates the distribution of SMAD4 mutations reported for HHT patients (red lines) and JP alone (green lines). Whereas for ENG and ACVRL1, mutations occur throughout the respective genes,⁹⁶ note the limited and highly comparable SMAD4 mutational distribution for the two pathologies, for all mutations (solid lines), and missense substitutions affecting exons 10 to 13 (dotted lines). (Data derived from hht.mutation.org and Wooderchak WL, Zacary S, Crockett DK, et al. Repository of SMAD4 mutations: reference for genotype/phenotype correlation. J Data Mining Genom Proteomics 2010;1:101.) Ref.¹²⁶.

Mechanism of Disease

The underlying mechanism of HHT1 and HHT2 is haploinsufficiency, characterized by a lack of expression of the nonfunctional allele, as determined by protein quantification using metabolic labeling.³⁶^,³⁷ For example, in individuals with a molecular diagnosis of HHT1, the median level of endoglin was close to 50%, representing the normal allele only (in umbilical vein endothelial cells, median 45%, n = 30 newborns; peripheral blood-activated monocytes, median 48%, n = 109 patients).³¹ In many cases, the absent protein will reflect absence of mRNA bearing the mutation due to nonsense-mediated decay of transcripts with premature termination codons.³⁸^,³⁹ The haploinsufficiency model is also supported by the distribution of mutations throughout both genes, deletions of the whole allele, and severity of the clinical phenotype not correlating with the type of mutation.

Genotype/Phenotype Correlations

Phenotypic studies in large groups of patients with identified HHT1 and HHT2 mutations indicate that the gene, rather than a particular mutation, influences the clinical phenotype.¹⁰^,⁴⁰^,⁴¹^,⁴²^,⁴³^,⁴⁴ The major findings in the multiple studies are that PAVMs are more frequent in HHT1 (ranging from 49% to 75%) than HHT2 patients (5% to 48%). Cerebral AVMs, despite an overall much lower occurrence, are also higher in HHT1 (8% to 20%) than HHT2 patients (0% to 2%). On the contrary, hepatic involvement, assessed by the presence of asymptomatic AVMs, is much higher in HHT2 than HHT1 patients.

JPHT

Mutations in the SMAD4 gene, located on chromosome 18q21, are present in a subgroup of HHT patients (2% to 3%). Patients with SMAD4 mutations, including those with HHT, are susceptible to JP.⁴^,⁴⁵^,⁴⁶ Both JP and HHT are inherited as autosomal dominant traits with nonoverlapping clinical features, JP predisposing to GI tumors. SMAD4 mutations associated with JPHT had been found mostly in the last four exons.⁴^,⁴⁶ However, analysis of a larger number of patients has revealed mutations in other regions (FIGURE 70.2).⁴⁷ Thus, any JP patient with a SMAD4 mutation is at risk for visceral manifestations of HHT and any HHT patient with such a mutation is at risk for early-onset GI cancer. Therefore, all patients with a SMAD4 mutation should be monitored clinically for both HHT and JP.

Only gold members can continue reading. Log In or Register to continue