The previous guidelines regarding diagnosis and management of HHT(5) focused on screening and treatment of adults. While some manifestations such as telangiectasia and epistaxis manifestations are age dependent and may be absent in young children with HHT, potentially serious and even life-threatening complications of visceral AVMs can occur at any age. Currently, the literature about diagnosis and management in children with HHT is limited, but protocols for screening and treatment of children with HHT have been developed in HHT centers around the world. Complications described in the literature are mostly due to pulmonary arteriovenous malformations (AVMs) and brain vascular malformations (VMs). Therefore, the focus of the pediatric HHT guidelines is on screening and management of pulmonary AVMs and brain VMs. 

Since establishing the diagnosis of HHT based on clinical criteria is less reliable in children than in adults(168), a different approach is required in this age group, with genetic testing playing a more important role than in adults(77-79). HHT is an autosomal dominant disease with age-related but high penetrance; therefore, every child of a parent with HHT has a 50% chance of inheriting the disease. Genetic testing in children is usually performed in a stepwise approach in which the affected parent is tested first (see overall Background section). If a pathogenic variant has been identified in the index case or in other affected member of the family(5), genetic testing can be used to establish the diagnosis in children prior to screening for visceral AVMs. Equally important, genetic testing can identify non-affected children who can be released from follow-up.

The prevalence of pulmonary AVMs varies with the type of HHT: pulmonary AVMs are found in about 50% of patients with HHT1 and in about 10% of patients with HHT2(120, 156). While these estimates are based on studies in adults, data suggest that the prevalence of pulmonary AVM is comparable in children(80-83). This is supported by one study that found a similar prevalence of pulmonary AVM in children with HHT1 as in their parents suggesting that the vast majority of pulmonary AVMs are present early in life(84). This has important implications for screening as the yield in genetically confirmed cases is high. Pulmonary AVMs are found in children with all types of HHT and at any age. Pulmonary AVMs associated with low oxygen saturations (< 96% at sea level), as well as large pulmonary AVMs, can cause serious, sometimes life-threatening complications, including hemorrhage, brain abscess and stroke(82, 83, 85). For that reason, screening children with HHT or at risk for HHT is indicated after birth, or at the time of presentation. Two screening protocols have been studied in children (Supplement Table 9); at present both are seen as equivalent. The first screening approach (“Dutch protocol”) uses a conservative screening strategy of oximetry and chest X-ray. As small pulmonary AVMs cannot be excluded in this setting, procedural antibiotic prophylaxis is recommended to all subjects. Evidence from the Dutch cohort suggests that this protocol is sufficient to prevent pulmonary AVM related complications(90). Transthoracic contrast echocardiography (TTCE) is used in the second screening protocol and has a higher sensitivity as a screening test for pulmonary AVMs(86, 87). It requires an intravenous access and has not clearly been shown to detect additional pulmonary AVMs that would cause complications in childhood. TTCE has the advantage of being a non-radiating test. The use of a quantitative scoring system for analysis of TTCE can increase the specificity of the test and can be used to determine whether a CT-scan should be performed(88), as the diagnostic confirmatory test(88, 169).

Embolotherapy of pulmonary AVMs has a high success rate in children(85) (Supplement Table 10). There are however no data to suggest that small pulmonary AVMs associated with normal oxygen saturation need to be treated in children. In rare cases, larger pulmonary AVMs with normal saturation can occur and treatment can be considered, especially in the case of symptoms. Growth of pulmonary AVMs over time has also been documented in children(91); therefore follow-up of children is important to capture these changes.

Brain VM is a general term that encompasses three principal types of vascular lesions in HHT: nidus brain AVM, brain arteriovenous fistula (AVF), and capillary vascular malformation (CVM)(170). These vascular malformations are thought to have significantly different natural history risk for spontaneous brain hemorrhage, ranging from extremely low in CVM, to intermediate in brain AVM (as can be further risk-stratified with detailed angio-architectural information, see CR6 below), to high in AVF. Overall, brain VM are less common than pulmonary AVMs in HHT. The prevalence in children is not well defined; data from studies in adults suggest that brain VMs are found in 8-16% of patients with HHT1 and 1-2% of patients with HHT2(171-173), though the AVF type appears to be over-represented in children(92, 174). Brain VMs can be present from birth and there are often no warning signs or symptoms prior to hemorrhage of a brain VM(175, 176). Clinical symptoms are subtle or absent in children and case series from different centers have described brain hemorrhage in children prior to diagnosis or screening procedures(93, 94, 175). The purpose of imaging screening of children with HHT is to identify if a brain VM is present and, to the extent possible, differentiate between the three common subtypes of brain VM. The most sensitive and specific non-invasive imaging modality to identify brain VM is MRI(96-98).

Observational studies suggest that treatment of brain VM is successful and can prevent brain hemorrhage(92, 100) (Supplement Table 10). Brain VM with relatively high natural history risk for rupture include pial AVFs as well as nidus brain AVMs with specific angio-architectural features or evidence of prior hemorrhage(95, 101, 102, 170, 177). High risk features for future nidus brain AVM rupture, sometimes identifiable on MRI but more reliably identified on digital subtraction angiography (DSA), include but are not limited to: feeding artery aneurysms, nidus aneurysms, venous outflow stenoses, and deep venous drainage. Intra-lesional microhemorrhage seen on brain MRI is an independent risk factor for future nidus brain AVM rupture(102).

It is important to appreciate that while the recommendations below are based on consensus of experts in the field, different approaches regarding pre-symptomatic genetic testing and screening procedures are used in different countries. Whenever possible, these different strategies are mentioned in the recommendations.