Hepatic vascular malformations in hereditary hemorrhagic telangiectasia: In search of predictors of significant disease


  • Guadalupe Garcia-Tsao M.D.,

    Corresponding author
    1. Section of Digestive Diseases, Yale University School of Medicine, VA-CT Healthcare System, New Haven, CT
    • Section of Digestive Diseases, Yale University School of Medicine, VA-CT Healthcare System, 333 Cedar Street, 1080 LMP, New Haven, CT 06520
    Search for more papers by this author
    • fax: 203-785-7273.

  • Karen L. Swanson D.O.

    1. Division of Pulmonary and Critical Care Medicine, Mayo Clinic, 200 First Street SW, Rochester, MN
    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

  • See Article on Page 1570

Hereditary hemorrhagic telangiectasia (HHT) is an uncommon autosomal dominant disease caused by mutations in two major genes: endoglin (ENG) and activin receptor-like kinase type 1 (ALK-1 or ACVRL1). HHT is characterized by multisystemic angiodysplastic lesions (telangiectases, arteriovenous malformations [AVMs]) that may occur in the skin, lungs, brain, gastrointestinal tract, and liver. Diagnostic criteria for HHT include a history of epistaxis, mucocutaneous involvement, visceral involvement, and a positive family history. Screening procedures to identify pulmonary and cerebral AVMs are recommended in patients with HHT given the availability of successful preventative treatment options. There is a risk of substantial morbidity and mortality in unscreened, untreated patients. Pulmonary AVMs may cause brain abscess, stroke, or hypoxemia due to right-to-left shunting, whereas cerebral AVMs may cause brain hemorrhage and sudden death.


AVM, arteriovenous malformation; DUS, Doppler ultrasonography; HHT, hereditary hemorrhagic telangiectasia; HVM, hepatic vascular malformation; RHC, right heart catheterization.

Hepatic vascular malformations (HVMs) have been described in up to 84% of patients with HHT1; however, they are mostly asymptomatic, with symptoms occurring in only about 8% of them.2, 3 The symptomatic phase has been well described and is due to shunting of blood from hepatic artery to hepatic vein (resulting in high-output cardiac failure and biliary ischemia), shunts from the hepatic artery to portal vein (resulting in portal hypertension and biliary ischemia), and/or shunts from portal vein to the hepatic vein (that contribute to high-output heart failure and may rarely cause hepatic encephalopathy).4, 5 Because hepatic artery to hepatic vein shunts are the most common in HHT, the majority of these patients have a hyperdynamic circulation with high cardiac output that will eventually lead to heart failure, the most common clinical presentation. Therapy depends on the specific clinical presentation, although good results have been obtained with liver transplantation in cases that do not respond to medical treatment.6

Identifying the subpopulation of patients with HVMs that will progress to a symptomatic phase is important, because preventative efforts can then be targeted to an at-risk population. The study by Gincul and colleagues in the current issue of HEPATOLOGY represents a step forward in this direction.7 The investigators prospectively evaluated the presence of HVMs and early cardiac consequences in 102 patients with HHT who had not been previously screened and who had no symptoms suggestive of HVMs. They used Doppler ultrasonography (DUS) to determine the presence and severity of HVMs and also to assess cardiac output. The investigators found HVMs in 56 of 102 (55%) of the cases. Those with abnormal liver tests and ALK-1 mutation were more likely to have HVMs.

Previous studies had shown an association between ALK-1 mutation and HVMs8–11 (Table 1); overall, 77% (217 of 281) of HVMs and all but one of 31 symptomatic HVMs have been described in patients with ALK-1 mutation (HHT type 2). Therefore, this is an important association, particularly now that the use of genetic testing is more widespread.

Table 1. Frequency of Hepatic Vascular Malformations (HVMs) in Patients with HHT Who Were Screened for Their Presence
StudyHVMs in HHT1HVMs in HHT2N with HVMsALK-1-Related HVMsSymptomatic HVMsSymptomatic HVMs by HHT1/HHT2
  • Screened only patients with liver bruit and/or liver test abnormalities and/or symptoms;

  • **

    routine screening in all patients with HHT;

  • *

    prevalence significantly higher than in HHT-1; NS = not stated; HHT1 = type of HHT related to ENG mutation; HHT2 = type of HHT related to ALK-1 mutation.

Bayrak-Toydemir81/59 (2%)13/47 (28%)*1413/14 (93%)61/5
Letteboer911/144 (8%)13/32 (41%)*2413/24 (54%)NSNS
Sabba10**27/45 (60%)64/77 (83%)*9164/91 (70%)70/7
Lesca11**20/46 (44%)87/151 (58%)10787/107 (81%)190/19
Gincoul7**5/20 (25%)40/62* (64%)4540/45 (89%)0

The 55% HVM prevalence described in this study is similar to another series using DUS3 but is lower than in other studies in which perhaps minimal alterations without a hemodynamic consequence were considered HVMs.1 The study by Gincul et al. defines HVMs by the presence of hepatic artery, portal vein, and/or hepatic vein abnormalities. By finding that even patients with “early changes” have a higher rate of “clinical” (physical examination), “biological” (liver test), and “cardiac” (cardiac index) abnormalities, they have potentially identified a subgroup of patients with clinically significant HVMs. The independent value of a larger hepatic artery diameter and the presence of focal nodular hyperplasia in predicting a high cardiac index is interesting. A larger hepatic artery is most probably related to greater arteriovenous shunting, whereas focal nodular hyperplasia, a focal mass consisting of regenerating hepatocytes, would also result from a more abnormal liver blood flow with a greater blood supply in the nodule as compared to the surrounding parenchyma.12

The most important association in the study is that of DUS and cardiac index abnormalities. However, cardiac index was calculated from the cardiac output obtained at echocardiography. Right heart catheterization (RHC) was not performed. Echocardiography provides only an estimate of cardiac output and, in a hyperdynamic state, may correlate poorly with measurements obtained during RHC. Therefore, the correlations between DUS findings and cardiac index require confirmation by RHC. Other factors that may also influence cardiac index include significant anemia due to epistaxis or gastrointestinal bleeding, frequent complications in HHT, and thyroid disease. These parameters need to be incorporated in future statistical models.

Furthermore, the added predictive value of other DUS parameters and other cutoff values previously identified through more rigorous methodology should be examined.13 The intraobserver and interobserver agreement of DUS findings should also be addressed, particularly in light of a recent study that found a high degree of agreement among ultrasonographers regarding the presence HVM but not regarding the staging of severity of HVMs.14

In order to better understand the natural history of liver involvement by HHT, the investigators will hopefully continue to follow these patients over time to provide insight into the longer term consequences of DUS abnormalities. It will be important to identify findings that predict the development of symptoms, not only of heart failure but of biliary ischemia and portal hypertension, and predict unfavorable outcomes due to late-stage complications. It will be even more important to determine whether any therapeutic maneuver (e.g., salt restriction, beta-blockers) in “high-risk” patients may exert an effect in delaying or preventing the development of symptoms and/or a poor outcome.

The fundamental purpose of screening is early diagnosis and treatment of the individual so as to prevent later complications and a poor outcome (such as occurs with pulmonary and cerebral AVMs). Because no standard therapy for asymptomatic HVMs exists to date, and as recently established by consensus,15 screening (and surveillance) using Doppler ultrasonography in patients with HHT should remain a research tool or be used to establish a diagnosis of “definite” HHT in patients who meet only one or two of the diagnostic criteria and in whom genetic testing is unavailable or inconclusive.