Pulmonary arteriovenous malformations (PAVMs) were first described in the late 19th century; Churton1 reported the autopsy findings in a young boy with cyanosis in 1897. Based on the correlation of physical with postmortem findings, the triad of cyanosis, clubbing, and polycythemia was identified with PAVM in 1932.2 Hereditary hemorrhagic telangiectasia (HHT) was first connected to PAVM in 1938.3 As described below (Causes and Disease Associations), HHT is often intimately related to PAVMs—a fact that prompts the following discussion of the history of HHT.
Hereditary epistaxis was first described in 1864,4 though neither that nor Babbington’s description a year later report an association with telangiectasia.5 These reports were not generally recognized; nor were subsequent descriptions of telangiectasia, hereditary transmission, and epistaxis by Legg6 in 1876, or a similar kindred reported by Chiari in 1887.7 The first widely recognized connection of epistaxis to telangiectasia was made by Rendu in 1896.8 Osler9 added three cases, and recognized familial occurrence in 1901. Weber10 elucidated the familial nature and lack of coagulation abnormality, and thus earned his eponymic association. By precedence of description, this eponym should be Rendu–Osler–Weber, even though Osler–Weber–Rendu is the most common usage. Hanes11 was responsible for naming the syndrome HHT, the designation now most often preferred, in 1909.
The genetic basis, if any, of isolated PAVMs remains unknown. HHT is an autosomal dominant disease. Its frequency was believed until relatively recently to be less than 3 per 100,000 people.12 Newer studies suggest a much higher prevalence. The highest frequency reported, 1:1331, occurs in the Afro-Caribbean population of the Netherlands Antilles, presumably due to a founder effect.13 Other estimates vary geographically; 1:6410 in Denmark,14 1:8000 in Japan,15 and 1:16,500 in Vermont.16 Phenotypic variation is extreme, ranging from asymptomatic to severely symptomatic, and from cases with no or few mucocutaneous lesions to those with diffuse cutaneous telangiectasias. For many patients, the disease remains undiagnosed by their primary care physicians, suggesting that disease frequency may be greater than reported, and that some patients with “isolated” PAVMs may actually have HHT.
A gene for HHT was first localized to chromosome 9, region q33–34 (9q33–34).17–19 Investigation revealed the protein product to be endoglin, which associates with the transforming growth factor-beta (TGF-β) bone morphogenetic protein (BMP) receptor complex and binds TGF-β-1 and -3.20 The same work showed the disease to be genetically heterogeneous, with multiple mutations in the responsible gene. It rapidly became clear that there were other chromosomal mutations resulting in the same syndrome, and the endoglin mutation disease was designated HHT-I; it was noted to be associated more often with PAVMs than were those with non-9q3 mutations.21,22 A haploinsufficient mouse model also demonstrated phenotypic heterogeneity which was very dependent on ...