Liver VM's
Liver VMs occur in approximately 75% of HHT patients. Liver VMs occur more commonly in women and often present in the 5th decade. These seven recommendations - which include a recommendation from the First International Guidelines - cover screening, diagnosis and management. Full background can be found after the recommendations.
Recommendations from the Second International HHT Guidelines (2020)
0%
AgreementSeveral cross-sectional diagnostic studies demonstrated high yield and accuracy of Doppler ultrasound, multiphase contrast computed tomography (CT) and magnetic resonance imaging (MRI) for detection of liver VMs (5, 60-68) (Supplement Table 7), with Doppler ultrasound severity grading predictive of outcomes(69). Anicteric cholestasis, reported in one third of liver VMs patients, correlated with severity of liver VMs and complications(69-71).
0%
AgreementClinical considerations: The rationale for screening is based on the concept that awareness of liver VMs could improve subsequent patient management. In some cases, documenting presence of liver VMs can help to clarify the diagnosis of HHT by establishing an additional Curaçao criterion. The imaging test of choice for liver VM screening in HHT is the Doppler ultrasound due to its accuracy, safety, tolerability, low costs and operating characteristics. However, depending on local expertise and availability of Doppler ultrasound testing, as well as patient preference, patients may be screened clinically (history, physical and blood work) or alternative imaging may be considered, such as multiphase abdominal CT or MRI.
0%
AgreementSeveral cross-sectional diagnostic studies demonstrated high yield and accuracy of Doppler ultrasound, multiphasic contrast CT and MRI for diagnosis of liver VMs (5, 60-68) (Supplement Table 7)
0%
AgreementClinical considerations: The choice of imaging modality should be informed by the risk/benefit balance, local expertise and availability/cost. Contrast studies (CT and MRI) should be avoided if kidney dysfunction. Echocardiography provides additional information about the hemodynamic impact of liver VMs. These tests will be most informative when performed in a center with HHT expertise, in the context of a clinical assessment at an HHT Center of Excellence.
0%
AgreementQuality of Evidence: Moderate (Agreement 88%)
One large series demonstrated moderate response to first-line therapy, tailored to liver VM complication(69). Expert consensus supported the recommendation for specialized center management.
0%
AgreementStrength of Recommendation: Strong (Agreement 88%)
Clinical considerations: Typically, patients with symptomatic liver VMs are managed by an expert team at an HHT Center of Excellence, with at least annual follow-up.
0%
AgreementThree observational studies have identified clinical predictors of complications from liver VMs (69, 70, 72).
0%
AgreementClinical considerations: Clinicians should plan monitoring for patients with liver VMs patients based on estimated prognosis.
0%
AgreementSmall uncontrolled series of IV bevacizumab have demonstrated improved cardiac output or clinical symptoms in 80% of patients with severe liver VMs, primarily in those with HOCF(16) (Supplement Table 8). AE rate was reported at 50 per 100 person-years, including 1 fatal event probably related to bevacizumab(73).
0%
AgreementPrescribing and safety monitoring guidance for IV bevacizumab is detailed in Supplement Table 4.
0%
AgreementSmall uncontrolled series of orthotopic liver transplantation (OLT) for liver VMs in HHT demonstrated excellent 5-10 year survival (82-92%) (74, 75) with asymptomatic rare and late recurrence of liver VMs after OLT (76).
0%
AgreementClinical considerations: Timing for listing a symptomatic patient for OLT should be based on prognostic predictors and the severity of liver VMs complications, including pulmonary hypertension. Liver transplant can be undertaken in the presence of pulmonary hypertension if pulmonary vascular resistance, estimated by right heart catheterization, is < 3 Woods Units. Portal pressure measurement with hepatic venous pressure gradient is reserved for selected patients with complicated liver VMs when evaluated for OLT(60)
Currently Recommended from the First International HHT Guidelines (2009)
0%
AgreementStrength of Recommendation: Strong
0%
AgreementStrength of Recommendation: Strong
Background
Liver VMs occur in 41–74% of HHT patients(61, 155), occurring in all genotypes, but the clinical presentation is typically more severe in patients with ACVRL1 mutation (HHT2)(69, 120, 156). The mean age of patients at diagnosis of liver VMs is 48 years(61, 69, 120) with a female predominance of 4.5 to 1. Liver VMs in HHT typically present as diffuse small lesions throughout the liver, and rarely as discrete large AVMs. Three different and often concomitant types of intrahepatic shunting (hepatic artery to portal vein, hepatic artery to hepatic vein and/or portal vein to hepatic vein) can lead to different and potentially overlapping clinical features, including high-output cardiac failure (HOCF), portal hypertension, encephalopathy, biliary ischemia and mesenteric ischemia(60, 157). Liver VMs in HHT may be associated with either diffuse or partial hepatocellular regenerative activity(158); the prevalence of focal nodular hyperplasia in patients with HHT is 100-fold greater than in general population(159).
HHT liver involvement is not associated with liver insufficiency(60, 157). Whereas only 8 to 14% of patients with liver VMs are symptomatic at baseline(61, 70, 155), prospective study has shown significant development of morbidity and mortality. The incidence of fatal outcome and of morbidity was 1.1% and 3.6% per person-years, respectively(69, 70). HOCF represents the predominant reported complication associated with HHT, but complicated portal hypertension occurs at a rate comparable to that of HOCF (1.4 and 1.2, per 100 person-years, respectively)(69). In patients with a high-output cardiac state due to liver VMs, the incidence of atrial fibrillation is1.6 per 100 person-years(46, 69). Much rarer presentations of liver VMs in HHT include encephalopathy, mesenteric angina and ischemic cholangitis that can cause bilomas or more ominously lead to a catastrophic complication termed “hepatic disintegration”(5, 60, 74, 160, 161).
The suspicion of liver involvement in HHT comes from history, physical examination, laboratory assessment of liver function tests, echocardiographic evaluation (with measurement of cardiac index and estimation of pulmonary hypertension)(162), and screening for signs, symptoms and biomarkers of heart failure. Anicteric cholestasis is observed in one third of patients with liver VMs, with a direct correlation with the severity of VMs and their complications(69-71). Doppler ultrasound has been proposed as the preferred first-line investigation for the assessment of liver VMs due to its safety, tolerability, low costs and accuracy for the detection of liver VMs(5, 60-64) and very good interobserver agreement for the presence/absence of liver VMs (Kappa = 0.85-0.93)(65). Doppler ultrasound also allows grading of severity of liver VMs (from 0.5 to 4) which correlates with patient outcome and has been shown to be a predictor of clinical outcome(69). Abdominal computed tomography (CT) with a standardized protocol (multiphasic contrast-enhanced) provides detailed anatomic assessment and has the potential for reproducible results across centers, with excellent accuracy(155) (Supplement Table 7). However, CT findings do not correlate however with liver VMs severity(163) or clinical presentation(66), although CT has been recommended previously when expertise in Doppler US is lacking for diagnosing liver VMs(60). Magnetic resonance imaging (MRI) of the liver provides excellent accuracy with both multiphase anatomic assessment and hemodynamic characterization of liver VMs(68). The abnormalities are better depicted on MR angiograms and dynamic MRI images, providing a map of anomalous vessels and analysis of filling kinetics; MRI has been proven to be as accurate as CT for liver VMs, and involves no ionizing radiation(67). Moderate to good interobserver reproducibility for MR imaging has been demonstrated. In the case of pregnant patients, US is preferred to avoid ionizing radiation or gadolinium exposure to the fetus. We continue to recommend against liver biopsy, as we did in the first International HHT Guidelines(5) (Table), as a major and unnecessary bleeding risk.
Echocardiographic evaluation is recommended at the time of liver VM diagnosis, to evaluate of the impact liver VMs on cardiac function and morphology, particularly cardiac index and pulmonary artery pressures, and to provide a baseline for comparisons over time(60, 164, 165). In those with signs or symptoms of heart failure and an intermediate or high probability of pulmonary hypertension, right-heart catheterization should be performed to accurately assess cardiac and pulmonary hemodynamics(60, 164, 165). Right heart catheterization is also essential for diagnosing different forms of pulmonary hypertension, for example pre-capillary pulmonary arterial hypertension characterized by high pulmonary vascular resistance and normal pulmonary artery wedge pressure which can be associated with HHT(166).
In patients diagnosed with liver VMs, follow-up with ultrasound Doppler and echocardiography should help identify complications and disease progression. The assessment of prognosis of symptomatic liver VMs using available outcome predictors can assist in decision-making. Identified disease progression predictors include: stage 4 liver VMs at baseline and ACVRL1 mutation(69). Clinical factors that can be used to predict low, moderate and high risk categories for significant disease from liver VMs include: age at presentation >47 years, female gender, hemoglobin level at presentation < 8 g/dL (or < 5 mmol/L) and alkaline phosphatase level at presentation > 300 UI/L(70). A retrospective cohort (72) has demonstrated other worrisome features including mean pulmonary artery pressure (≥25 mmHg at catheterization), elevated bilirubin, weight loss, GI bleeding and any biliary ischemia, atrial fibrillation, high blood transfusion requirement, right upper quadrant pain, and sepsis.
Presently, no treatment is recommended for asymptomatic liver VMs. An intensive therapeutic approach, tailored to the type of complication present, is recommended for symptomatic liver involvement in HHT(60). Patients with HOCF should have care supervised by a specialist experienced in managing HOCF; treatments include aggressive treatment of anemia, salt restriction and the use of diuretics, as needed. Management of atrial fibrillation in HOCF follows the same principles as in the general population. Anticoagulation for stroke prevention should be considered based on individualized risk assessment, as discussed in the Anemia and Anticoagulation section. Patients with pulmonary hypertension should be evaluated and treated by a physician with expertise.
Antibiotic treatment is administered in HHT patients with liver VMs and cholangitis. Endoscopic retrograde cholangiopancreatography (ERCP) with stenting is not an option as large duct obstruction is usually not present and ERCP may increase the risk of infection, in ischemic ducts. Necrotizing cholangitis with hepatic necrosis is an ominous complication of liver VMs, requiring emergent liver transplantation. Management of portal hypertension follows the same principles as in patients without HHT. The use of non-selective beta-blockers in patients with severe HOCF should be supervised by a cardiologist. Transjugular intrahepatic portosystemic shunt placement may worsen hyperdynamic circulation and precipitate cardiac failure. Management of encephalopathy follows the same principles as in patients without HHT who have cirrhosis, including the use of lactulose and rifaximin.
The reported response to first-line treatment in patients with symptomatic liver VMs in HHT is complete in 63%, partial in 21% and absent (with progression to death) in 14%(69). These data support the recommendation to consider aggressive options only for otherwise intractable complications, after the assessment of response to first line treatment has been made, after 6-12 months(60). Outcomes of orthotopic liver transplantation (OLT) (Supplement Table 7 for liver VMs in HHT are excellent with 82-92% survival(74, 75). Liver VMs in HHT are included in MELD (Model for End Stage Liver Disease) exceptions: suggested MELD exception points for HHT include a score of 40 to patients with acute biliary necrosis and 22 to patients with HOCF(60). Potential morbidity and mortality rates associated with OLT are a cause for concern and the optimal timing for OLT in HHT with symptomatic liver involvement should be supported by predictors of outcome(69, 70, 72). Recurrence of liver VMs after OLT has been demonstrated in only a small number of cases, many years post-OLT, and has been asymptomatic(76). Other surgical or interventional options for treating complicated liver VMs such as hepatic embolization and/or banding of the hepatic arteries are associated with a high rate of serious complications including death and cholangiopathy and should be reserved as a last resort when medical therapies fail and OLT is not an option(5, 60, 167).
There is growing evidence for the role of intravenous bevacizumab in patients with severe liver VMs (Supplement Table 8), primarily in those with HOCF(16). However, potential adverse events (AE) related to bevacizumab need careful consideration: in 69 HHT patients who received bevacizumab treatment for a total of 63.8 person-years treatment, an average AE incidence rate of 50 per 100 person-years, including 1 fatal event probably related to bevacizumab, have been described(73). Furthermore, rates of non or partial response to bevacizumab(16), and recurrence of symptoms/signs after drug withdrawal make this drug unsuitable to replace OLT for complicated liver VMs in HHT. Bevacizumab may offer a potential “bridging” role to OLT, and if a response is obtained with resolution/improvement of the liver VM complication, the option of OLT should be re-assessed. Bevacizumab complicates wound healing and transplant teams should closely coordinate with HHT providers so that bevacizumab can be stopped long enough prior to OLT to minimize complications, while still minimizing the time off of therapy. The optimal OLT window is likely between 2 and 6 months after the last dose of bevacizumab.