Keywords
portal hypertension - fibrosis - hepatocellular carcinoma - liver transplantation
- combined heart–liver transplantation
Fontan-associated liver disease (FALD) is a unique form of liver disease which occurs
in patients who have undergone a Fontan surgery to manage certain congenital heart
defects characterized by absence of a ventricle. Liver fibrosis occurs in all patients
who have undergone a Fontan surgery. However, FALD can lead to more serious complications
including cirrhosis, hepatocellular carcinoma, and portal hypertension. Failure of
the Fontan circulation, including FALD, is leading to an increased need for combined
heart–liver transplantation for patients with Fontan physiology. As survival after
the Fontan surgery improves, there is an increasing need for more nuanced studies
of the pathophysiology of FALD, collaboration among centers, and multidisciplinary
management to improve care of this medically complex patient population.
Fontan-associated liver disease (FALD) is a unique form of congestive hepatopathy
(CH) which occurs in patients who have undergone Fontan palliation for functional
single ventricle congenital heart defects (CHD). FALD represents a spectrum of disease,
ranging from liver fibrosis to cardiac cirrhosis, portal hypertension (PHTN), and
hepatocellular carcinoma (HCC). The improved survival of patients with Fontan physiology
into adulthood has led to the recognition of the deleterious impact of Fontan physiology
on the liver and the need for an improved understanding of its pathophysiology.
Background
In 1971, Francis Fontan described the first successful total right heart bypass, which
was performed to relieve volume overload in a patient with tricuspid atresia.[1] The Fontan circulation achieves complete separation of the pulmonary and systemic
circulations by directing systemic venous return to the pulmonary arteries without
an intervening, subpulmonic ventricle.[2] In this circulation, filling of the pulmonary circulation is passive and is largely
due to negative intrathoracic pressure generated through the act of breathing.[3] As a result, transpulmonary flow and pulmonary vascular resistance (PVR) are critical
regulators of cardiac output (CO) in this patient population.[2] While the Fontan circulation improves arterial saturation, intracardiac mixing,
and chronic volume overload, it leads to chronic systemic congestion, sustained elevation
in central venous pressure (CVP), and decreased CO.[4] Such chronic systemic congestion leads to negative extracardiac consequences over
time. A recent matched retrospective cohort study assessed the risk of systemic disease
associated with the Fontan circulation.[5] Age- and gender-matched patients with isolated ventricular septal defects comprised
the control group. In this study, patients with Fontan physiology were significantly
more likely to develop systemic disease, including respiratory disease, liver disease,
and renal dysfunction, compared with the control group.
The Fontan circulation may ultimately fail, culminating in a complex syndrome termed
“Fontan failure” (FF).[6] FF has multiple potential causes, including ventricular dysfunction, increased PVR,
Fontan pathway obstruction, or other end-organ dysfunction.[2]
[7]
[8] Patients presenting with heart failure or end-organ consequences, such as protein-losing
enteropathy (PLE), require evaluation of their Fontan pathways with prompt intervention
if needed. PLE is a rare but devastating complication of Fontan physiology which leads
to increased enteric protein losses.[9] Patients with PLE may present with diarrhea, abdominal pain, pleural and pericardial
effusions, and ascites. The pathophysiology of PLE in the setting of Fontan physiology
is multifactorial and due in part to chronic elevation in systemic venous pressure
(SVP), lymphatic congestion, and impaired perfusion of the gastrointestinal mucosa
in the setting of low CO.[10]
[11] PLE is associated with significant morbidity and mortality in this patient population.
The Fontan procedure is now considered definitive palliation for patients with functional
single ventricle disease. Approximately 1,000 Fontan operations are performed annually
in the United States.[12] With surgical modifications and implementation of multidisciplinary management,
survival after the Fontan operation is improving, and some studies suggest that the
20-year survival exceeds 90% for patients who underwent extracardiac or lateral tunnel
Fontan operations.[13] As a result, the population of patients with Fontan physiology is expected to grow
worldwide, with a concomitant increase in the median age of the Fontan population.[14]
[15]
[16] As patients with Fontan physiology experience prolonged survival, extracardiac sequelae
of the Fontan circulation, including FALD, are increasingly recognized and pose a
multidisciplinary clinical challenge.
Histological Characteristics of Fontan-Associated Liver Disease
Histological Characteristics of Fontan-Associated Liver Disease
FALD encompasses a breadth of structural and functional changes in the liver secondary
to hemodynamic changes. A uniform, comprehensive definition of FALD is lacking to
date[17] and likely leads to an underestimation and underrecognition of its incidence and
scope. Liver fibrosis was first identified in a patient with Fontan physiology in
1981.[18] FALD is now recognized as a complication of the Fontan circulation which occurs
in all patients who have undergone Fontan palliation.
Several studies have attempted to delineate the incidence of, extent, and risk for
liver fibrosis in patients with Fontan physiology. A retrospective study described
liver histology in 12 patients who had a prior atriopulmonary Fontan operation, but
who underwent liver biopsy prior to conversion to an extracardiac Fontan. The authors
identified sinusoidal fibrosis in all 12 patients and cardiac cirrhosis in 7 of the
12 (58%) patients.[15] In this study, the severity of fibrosis was found to correlate with the duration
of Fontan circulation (r = 0.75, p = 0.013). Histology was otherwise notable for characteristic changes of hepatic congestion,
including sinusoidal dilatation and parenchymal atrophy. Another retrospective study
of liver histology in patients who had undergone Fontan palliation identified centrilobular
and sinusoidal fibrosis in all and portal fibrosis in 93.2% of patients.[16] In this study, portal fibrosis was attributed to PHTN in the setting of sustained
elevation in CVP. A robust relationship between cardiac hemodynamics and the severity
of FALD has not been established to date, however.[17]
While the duration of Fontan physiology has been positively correlated with fibrosis
and cirrhosis in multiple studies,[19]
[20]
[21]
[22]
[23]
[24] liver fibrogenesis likely starts prior to the Fontan surgery in patients with functional
single ventricle physiology.[25] Retrospective autopsy studies have identified both portal and sinusoidal fibrosis
in patients who passed away shortly after the Fontan procedure.[18]
[26] This suggests that pre-Fontan single ventricle defects instigate liver fibrosis.
In a retrospective study of liver histology in patients with Fontan physiology at
autopsy, 12 of 14 patients who passed away within 1 month of the Fontan surgery demonstrated
significant portal fibrosis.[26] In these patients, markers of pre-Fontan morbidity, including length of hospitalization
(LOH) after pre-Fontan cardiac interventions and pre-Fontan mean right atrial pressure
(RAP), were significantly associated with degree of portal fibrosis. The authors of
this study speculate that LOH after cardiac interventions may be associated with acceleration
of liver fibrosis due to repeated episodes of desaturation. Elevation in RAP may correlate
with ventricular diastolic dysfunction, which predisposes to higher hepatic venous
(HV) pressures and compromised CO, factors which may also instigate fibrogenesis in
these patients prior to the Fontan operation.[26] The development of cardiac cirrhosis is thought to be a time-dependent phenomenon
in patients after the Fontan surgery. A retrospective study that utilized radiological,
clinical, and laboratory characteristics reported 10-, 20-, and 30-year freedom from
cirrhosis in 99, 94, and 57% of patients with Fontan physiology, respectively,[22] confirming increased prevalence of cirrhotic-stage disease with longer duration
of Fontan physiology.
The Fontan surgery has undergone several revisions since its conception in 1971. The
original approach consisted of the atriopulmonary connection (or the classic Fontan),
and it resulted in a direct connection of the right atrium to the right pulmonary
artery.[27] The lateral tunnel approach was devised in 1988. This approach utilizes a patch-created
intracardiac tunnel in the right atrium to connect the superior vena cava (SVC) and
inferior vena cava (IVC) to the pulmonary arteries.[28] The extracardiac cavopulmonary connection (ECC) comprises the most recent iteration
of the Fontan surgery. The ECC creates a direct anastomosis between the SVC and right
pulmonary artery through the use of an extracardiac vascular prosthesis.[29] A recent retrospective study examined outcomes among 332 patients who underwent
lateral tunnel and ECC Fontan surgeries between 1989 and 2021 at a tertiary center.
In this study, cirrhosis was diagnosed on the basis of either (1) histology or (2)
compatible imaging in combination with examination by a hepatologist. This study identified
earlier onset of histologically diagnosed cirrhosis in patients after ECC compared
with the lateral tunnel Fontan. While the ECC Fontan has some advantages over earlier
iterations, including a lower incidence of arrhythmias, decreased risk of reoperation,
and decreased thromboembolic risk,[30]
[31]
[32] recent reports have raised concern that this approach is associated with acceleration
of liver fibrosis.[33]
[34]
[35] While these reports are small and representative of single-center experience to
date, additional prospective studies are needed to assess the impact of the surgical
modality on the risk of liver disease progression.
Clinical Presentation and Evaluation of Fontan-Associated Liver Disease
Clinical Presentation and Evaluation of Fontan-Associated Liver Disease
The clinical presentation of FALD is often indolent and can lead to an underestimation
of disease severity and progression. However, in early stages of the disease, FALD
may be undetectable on physical examination (PE). A retrospective study was performed
of 74 patients with Fontan circulation with a median time of 15.7 years from the Fontan
surgery to identify patterns of fibrosis and their association with clinical parameters
and presentation.[36] While all patients had liver fibrosis on biopsy, only a minority of patients had
abnormalities suggestive of FALD on PE, including hepatomegaly in 29.8%, splenomegaly
in 8.5%, and ascites in 4.2%. While retrospective studies have identified splenomegaly
in as many as 84.8% of patients with Fontan physiology,[36] splenomegaly may be due in part to posthepatic PHTN in the setting of cardiac dysfunction.
In addition, it is important to note that hepatic congestion can lead to ascites,
jaundice, and hepatomegaly without advanced fibrosis.[37]
Laboratory Abnormalities and Lab-Based Scoring Systems
Laboratory abnormalities are common in the setting of FALD but do not strictly correlate
with disease stage.[38] Predominantly indirect hyperbilirubinemia and disproportionate prolongation of the
prothrombin time compared with other coagulation parameters are among the most common
lab abnormalities noted.[37] A retrospective study identified elevation of gamma-glutamyl transferase in all
patients with FALD, while elevation in the either ALT or total bilirubin were present
in 15%.[36] The international normalized ratio (INR) was elevated in half, while thrombocytopenia
was present in approximately 25% of patients.
Despite prevalent lab abnormalities, studies have identified conflicting results regarding
the correlation of serum-based scoring systems with fibrosis stage in FALD.[39] The AST-to-platelet ratio index (APRI) and Fibrosis-4 (FIB-4) scores can be calculated
from routine laboratory studies, and their capacity to predict advanced fibrosis have
been validated via meta-analyses.[40]
[41]
[42]
[43] A retrospective study of 106 pediatric patients with Fontan physiology identified
the modest discriminatory power of the APRI and FIB-4 scores in predicting advanced
fibrosis.[44] Other studies have revealed a limited capacity to predict fibrosis stage in FALD.[45] A recent study assessed the correlation between hemodynamics, laboratory, imaging,
and pathology data in 159 adult patients with Fontan physiology.[39] In this study, APRI and FIB-4 scores did not correlate with fibrosis stage on histology
but were independently associated with overall mortality. The MELD score excluding
INR (MELD-XI) has been studied in FALD given the common need for anticoagulation in
this patient population. While the MELD-XI score correlates with histologically proven
fibrosis,[46] a specific cutoff value predictive of advanced fibrosis has not yet been identified.
Collectively, these studies highlight the complexity inherent in the noninvasive assessment
and staging of FALD.
Liver Biopsy in Fontan-Associated Liver Disease
Liver biopsy is a cornerstone in the diagnosis and staging of most liver diseases.[47] However, there are several caveats to pursuing and interpreting liver biopsy in
patients with FALD. Procedural risk associated with liver biopsy is of concern in
patients with Fontan physiology given their cardiac and liver dysfunction and elevated
CVP.[48] In addition, patients with Fontan physiology are commonly maintained on antiplatelet
and/or anticoagulants. A retrospective study reported risk associated with 68 percutaneous
liver biopsies performed in 67 patients with Fontan physiology. The median Fontan
SVP, measured by cardiac angiography immediately prior to biopsy, was 13.8 mm Hg.
Bleeding complications were reported in 5 of the 68 biopsies (7.4%) and were minor
in 4 (5.9%) and major in 1 (1.5%). In contrast, the bleeding rate associated with
liver biopsy in the non-Fontan population ranges from 0.6 to 0.7%.[49]
[50]
There has been concern that transjugular liver biopsies may lead to an overestimation
of disease stage due to the likelihood of sampling the perivenous area, which tends
to be more severely affected, although this has not been proven in all studies.[51] Therefore, liver biopsies obtained percutaneously may be more representative of
disease stage in patients with FALD. However, recent guidelines from the European
Association for the Study of the Liver (EASL) recommend transjugular over percutaneous
liver biopsy in patients with Fontan physiology due to decreased bleeding and the
possibility of performing cardiac and HV catheterization at the same time.[17]
FALD is a notably heterogeneous disease, which increases the risk of sampling error.
A retrospective biopsy series identified areas of stage 4 fibrosis/cirrhosis and nonfibrotic
parenchyma within the same specimen.[52] Another retrospective study compared liver biopsies obtained during pretransplant
evaluations with explant specimens posttransplant. In 40% of cases, the liver biopsy
underestimated the stage of fibrosis, likely secondary to its patchy, heterogeneous
nature.[51] EASL guidance therefore recommends obtaining at least two different passes when
performing liver biopsy to mitigate the risk of sampling error.[17]
In contrast to most etiologies of liver disease, fibrosis in FALD is predominantly
centrizonal.[53] More severe FALD may be characterized by extension of fibrosis to the portal area.[18]
[19]
[21]
[52]
[54]
[55]
[56] Due to the centrizonal prominence of fibrosis in FALD, traditional histological
scoring systems may underestimate the severity of fibrosis. The congestive hepatic
fibrosis score was developed to ensure adequate fibrosis staging by including the
severity of centrizonal fibrosis ([Table 1]).[53]
[57] This has been updated to create the modified Ishak congestive fibrosis score, which
incorporates the coexistent central and portal fibrosis common in FALD ([Table 2]).[58]
Table 1
The congestive hepatic fibrosis score that is commonly used to stage histological
disease severity in patients with Fontan-associated liver disease
|
Stage
|
Histology
|
|
0
|
No fibrosis
|
|
1
|
Central zone fibrosis
|
|
2a
|
Central zone and mild portal fibrosis, with accentuation at the central zone
|
|
2b
|
At least moderate portal fibrosis and central zone fibrosis, with accentuation at
the portal zone
|
|
3
|
Bridging fibrosis
|
|
4
|
Cirrhosis
|
Table 2
The modified Ishak congestive hepatic fibrosis score
|
Score
|
Histology
|
|
0
|
No fibrosis
|
|
1
|
Central zone fibrosis
|
|
2a
|
Central zone and mild portal fibrosis, with accentuation at the central zone
|
|
2b
|
Portal and central zone fibrosis, with accentuation of fibrosis in the portal zone
|
|
3
|
Fibrous expansion of most portal areas with occasional portal to portal or portal
to central bridging
|
|
4
|
Fibrous expansion of most portal areas with marked portal to portal or portal to central
bridging
|
|
5
|
Marked bridging with occasional nodules or incomplete cirrhosis
|
|
6
|
Cirrhosis
|
Imaging in Fontan-Associated Liver Disease
Imaging in Fontan-Associated Liver Disease
The Fontan circulation leads to several changes in the morphology of the liver, which
are appreciated on radiographic studies. Thus far, studies suggest that morphological
abnormalities seen on imaging do not correlate well with fibrosis grade noted on histopathology.[59] For example, in the absence of advanced fibrosis, hepatic congestion can impart
a grossly nodular appearance to the liver suggestive of cirrhosis, leading to an overestimation
of disease severity.[37] Interpretation of imaging in FALD therefore requires an experienced radiologist
with expertise in imaging performed in the setting of hepatic congestion.
Ultrasonography in Fontan-Associated Liver Disease
Similar to other etiologies of hepatic congestion, ultrasound (US) in FALD often reveals
anatomic abnormalities including hepatomegaly, dilation of the IVC, and dilation of
the HVs.[60]
[61] Venovenous shunts may be visualized between dilated HVs in more severely congested
livers.[62] Absence of normal triphasic HV waveforms and abnormal portal venous (PV) velocities
may be appreciated on waveform studies. Parenchymal abnormalities are also common
in FALD, with frequent identification of heterogeneous parenchymal echotexture, which
may be secondary to congestion, steatosis, and/or fibrosis.[63] A recent retrospective study assessed imaging changes that were seen in US and cross-sectional
imaging in 131 patients with FALD.[64] This study identified heterogeneous parenchyma, lobar redistribution (describing
atrophy of the posterior right hepatic lobe with compensatory hypertrophy of the left
lateral segment and caudate lobe), and surface nodularity as the most common abnormalities
noted on US. Notably, these imaging changes showed no significant association with
mortality in this study. Surface nodularity may correlate with duration of the Fontan
circulation and severity of fibrosis in FALD.[65]
Cross-Sectional Imaging in Fontan-Associated Liver Disease
Cross-sectional imaging studies in FALD may also show hepatomegaly, dilation of the
IVC and HVs, lobar redistribution, surface nodularity, and heterogeneous parenchyma.
Changes suggestive of PHTN, such as varices and recanalization of the umbilical vein,
may be noted more commonly on cross-sectional imaging studies compared with US.[64] On noncontrast-enhanced computed tomography (CT) studies, patchy areas of low attenuation
may be seen and likely represent areas of edema and/or fibrosis.[37] Studies performed after injection of intravenous (IV) contrast may reveal reflux
of contrast into the IVC or HVs.[62] It is important to note that the abnormal circulation in these patients may impact
the phase of enhancement obtained. Following the injection of IV contrast, CT images
are obtained at certain time points to optimize visualization of different vascular
structures, tissues, and organs.[66] Arterial phase images are typically acquired 20 to 40 seconds after the administration
of IV contrast to enhance visualization of arterial and hypervascular structures.
In contrast, venous phase images are acquired 60 to 70 seconds after administration
of contrast to allow for passage of contrast to the venous system. Venous phase imaging
facilitates visualization of venous structures and venous thromboses. However, these
standard phases of imaging are often altered in FALD. For example, due to the protracted
systemic circulation times, arterial phase images may be acquired if standard venous
phase imaging is used.[62] As a result, contrast bolus gating strategies are recommended to ensure acquisition
of the proper phase images.
CT and magnetic resonance imaging (MRI) studies may both reveal heterogeneous liver
parenchymal enhancement following administration of contrast agents in the PV phase,
likely due to slow PV circulation.[37] These agents equilibrate through the liver parenchyma over time, leading to a more
homogenous appearance on delayed phase images acquired 3 to 5 minutes after administration
of contrast.[62] On MRI, hyperintense areas may be noted on T2-weighted or diffusion-weighted images
and are most often appreciated in the periphery and right lobe.[37] This abnormal enhancement typically falls into two patterns: zonal or reticular.
Zonal enhancement describes an irregular pattern of poor enhancement in the liver
periphery. In contrast, a reticular pattern describes diffuse patchy enhancement,
although changes are often more prominent in the periphery.[62] Reticular enhancement has been associated with higher HV pressures, more advanced
fibrosis, and time from the Fontan procedure.[19] While abnormal parenchymal enhancement is most conspicuous in the right hepatic
lobe, perfusional abnormalities in the left hepatic lobe have been associated with
more severe PHTN.[67]
Elastography in Fontan-Associated Liver Disease
Elastography has been extensively studied as a noninvasive screening tool in FALD.
However, the presence of hepatic congestion confounds stiffness readings obtained
by elastography and may lead to overestimation of fibrosis stage. A prospective study
examined the relationship between hemodynamics as assessed via cardiac catheterization,
liver stiffness quantified by transient elastography (TE), and liver histology in
45 patients with Fontan physiology, 10 of whom had liver biopsies.[68] In this study, liver stiffness measurements (LSM) demonstrated a significant association
with more severe centrilobular fibrosis. However, LSM did not demonstrate a significant
association with the degree of portal fibrosis or the presence of bridging fibrosis.
The authors then assessed the validity of LSM normative values that have been validated
in predicting fibrosis in other etiologies of chronic liver disease.[69] In this study, the fibrosis stage predicted by TE overestimated the fibrosis stage
seen on histology by at least one stage in 70% of subjects. Fibrosis stage by TE overestimated
fibrosis stage by at least two stages in 50% of subjects, and TE did not underestimate
fibrosis stage in any subjects. Other studies have examined the relationship between
LSM and fibrosis stage on biopsy and have not identified a significant correlation
to fibrosis stage.[46]
[70]
Studies thus far highlight a complex interplay between LSM obtained by elastography
with fibrosis and Fontan pressures. Images obtained via magnetic resonance elastography
(MRE) often reveal peripheral areas of increased stiffness, which correspond with
congested areas.[37] Elastography-based LSM demonstrates a significant inverse correlation with cardiac
index and ejection fraction[71]
[72] and a positive correlation with CVP[73] and Fontan pressures,[72] confirming the sensitivity of LSM to hemodynamic status. LSM has been shown to correlate
with radiological PHTN (as defined by the presence of splenomegaly, ascites, and/or
gastrointestinal varices)[72]
[74] and poorer clinical outcomes. A retrospective study identified a relationship between
progression of LSM obtained via MRE with adverse outcomes including death, listing
for heart–liver transplant, engagement of palliative care, hospitalization, and need
for paracentesis.[75] Recent EASL guidelines suggest a role for longitudinal assessment of LSM in monitoring
disease progression and predicting clinical outcomes.[17] Further studies are needed to identify a more nuanced role for LSM in the clinical
care of patients with FALD.
A prospective study was recently completed to elucidate the role of noninvasive diagnostic
tools in identifying fibrosis stage in FALD.[76] This study led to the identification of the FonLiver risk score, which incorporates
LSM obtained via TE with platelet count. This score has an area under the receiver
operating characteristic curve of 0.81 in the identification of patients with severe
liver fibrosis. These results suggest that elastography, in combination with other
laboratory and clinical assessments, may have a role in monitoring disease progression,
predicting adverse clinical outcomes, and identifying patients with severe liver fibrosis.
Liver Masses in Fontan-Associated Liver Disease
Liver Masses in Fontan-Associated Liver Disease
Liver masses, both benign and malignant, are common in patients with Fontan physiology
and comprise another challenge in the evaluation and management of FALD. The reported
incidence of liver masses in patients with FALD ranges from 20 to 35% in retrospective
studies, although a recent prospective multicenter study reported an incidence of
48%.[19]
[71]
[77]
[78]
[79]
[80]
[81]
[82]
[83] Benign hypervascular nodules are the most common focal liver lesion detected in
FALD, the vast majority of which represent focal nodular hyperplasia (FNH). In the
setting of FALD, these lesions are termed “FNH-like lesions.” The pathogenesis of
these lesions is not fully understood but is thought to be related to compensatory
arterialization in response to HV outflow obstruction.[71]
[82] FNH-like lesions in FALD demonstrate many imaging characteristics of FNH lesions
identified in other patient populations. For example, on CT imaging, FNH-like lesions
appear as homogenous, well-circumscribed lesions, which may possess a hypoattenuating
central scar ([Table 3]). They demonstrate homogenous arterial phase hyperenhancement following the administration
of IV contrast ([Fig. 1]).[84] However, in FALD, FNH-like lesions may demonstrate hypointensity, or washout, in
the delayed phase (3–5 minutes after injection of contrast), possibly due to delayed
circulation of contrast in the setting of hepatic congestion.[80] This atypical finding is associated with HCC in patients with noncardiac etiologies
of liver disease[85] and often leads to diagnostic uncertainty.
Table 3
Imaging characteristics of focal nodular hyperplasia-like lesions and hepatocellular
carcinoma in Fontan-associated liver disease on computed tomography and magnetic resonance
imaging
|
CT
|
MRI
|
|
FNH-like lesions
|
• Homogenous lesions
• Well-circumscribed
• Hypoattenuating central scar
• With IV contrast: homogenous arterial phase hyperenhacement
• Delayed phase washout
• Stable in size on follow-up
|
• T1: isointense to slightly hypointense
• T2: isointense to slightly hyperintense
• Hepatobiliary contrast: homogenous or peripheral uptake
• Delayed phase washout
• Stable in size on follow-up
|
|
HCC
|
• Large and irregular
• Arterial phase hyperenhacement
• Pseudocapsule appearance
• Interval growth (50% in <6 mo, or 100% in >6 mo)
• Portal venous phase washout
• Delayed phase washout
|
• Large and irregular
• Interval growth
• Hyperenhancing in arterial phase
• Portal venous phase washout
• Delayed phase washout
• T1: high T1-weighted signal intensity with signal drop on T1 opposed phase
• T2: hypointense or heterogeneous signal on T2-weighted images
• Hepatobiliary contrast: lack of enhancement or heterogeneous pattern of uptake
|
Abbreviations: CT, computed tomography; FNH, focal nodular hyperplasia; HCC, hepatocellular
carcinoma; IV, intravenous; MRI, magnetic resonance imaging.
Fig. 1 Imaging of an FNH-like lesion in a 29-year-old patient with Fontan physiology. MRI
reveals a well-circumscribed lesion with (A) arterial phase hyperenhancement, (B) retention of hepatobiliary contrast, (C) slight hypointensity on T2-weighted images, and (D) isointensity on diffusion-weighted images. FNH, focal nodular hyperplasia; MRI,
magnetic resonance imaging.
As noted above, patients with Fontan physiology are also at increased risk of HCC.
While the Liver Imaging Reporting and Data System (LIRADS) criteria are used to noninvasively
diagnosis HCC in most patients, these criteria guidelines stipulate that noninvasive
criteria cannot be relied on to diagnose HCC in the setting of CH and other vascular
liver diseases due to overestimation of the probability of malignancy.[85] Many key LIRADS criteria (arterial phase hyperenhancement, pseudocapsule appearance,
washout, and interval growth) apply to HCC in the setting of FALD ([Fig. 2]). While washout in the delayed phase may also be seen in FNH-like lesions in FALD,
PV phase washout is more closely associated with HCC in the setting of FALD. LIRADS
ancillary features may also be observed in HCC in FALD and are associated with malignancy,
including restricted diffusion, lipid content, a heterogeneous pattern of enhancement,
and moderate T2 signaling intensity of MRI. With hepatobiliary contrast agents, most
HCCs demonstrate either a lack of enhancement or a heterogeneous pattern of uptake.[86] Lesions suspicious for HCC should be biopsied and reviewed in a multidisciplinary
discussion.[17] Quantification of α-fetoprotein is also helpful in the evaluation of liver masses
in patients with Fontan physiology, and elevation should raise concern for malignancy.
Fig. 2 MRI of biopsy-proven HCC in a 29-year-old patient with Fontan physiology. (A) Diffusion-weighted imaging reveals a 16-mm mass with restricted diffusion. This
mass also shows (B) moderate T2 hyperintensity and (C) lack of retention of hepatobiliary contrast. MRI was repeated in 3 months and revealed
interval growth of this lesion to 24.9 mm as noted on T2-weighted imaging (D), diffusion-weighted imaging (E), and on hepatobiliary phase imaging (F). HCC, hepatocellular carcinoma; MRI, magnetic resonance imaging.
Pathophysiology of Fontan-Associated Liver Disease
Pathophysiology of Fontan-Associated Liver Disease
FALD is a heterogeneous disease that represents the consequence of a spectrum of underlying
CHD and variable approaches and stages to repair. As a result, the etiology and pathogenesis
of FALD are complex and multifactorial. However, recent advances in basic and translational
studies have identified a few central contributing factors, which are discussed below.
Hemodynamic Alterations
The hemodynamic alterations associated with Fontan physiology impact the development
and severity of FALD. The Fontan surgery creates a total cavopulmonary anastomosis
which induces increased pressure associated with nonpulsatile flow within the IVC.
This nonpulsatile pressure transmits from the IVC to the HVs and sinusoids. As a result,
IVC pressure and hemodynamics impact the progression of FALD. A prospective study
of 33 patients with Fontan physiology investigated the relationship between systemic
and pulmonary hemodynamics and liver fibrosis in FALD. While several hemodynamic parameters
were studied, a significant, positive correlation was identified between the IVC flow
rate and collagen deposition as quantified by Sirius red staining performed on percutaneous
biopsy.[87] This suggests that hemodynamics within the IVC play a role in the pathogenesis of
FALD.
PV inflow is regulated in part by the pressure gradient between the portal and hepatic
veins. Sustained elevation in PV pressure therefore minimizes the transhepatic pressure
gradient and can compromise flow in the HVs.[88] A recent study identified imaging and flow characteristics in 22 patients with Fontan
physiology obtained via four-dimensional phase-contrast flow magnetic resonance imaging
(4D PC flow MRI) and revealed slower flow than normal in the superior mesenteric,
splenic, and portal veins.[89] Due to compromised PV inflow, the liver in the Fontan circulation is more reliant
on hepatic arterial flow and is particularly susceptible to changes in the hepatic
arterial circulation. However, chronic elevation in CVP in the Fontan circulation
may impede the effectiveness of the hepatic arterial buffer response (HABR).[88]
[90] The HABR describes a compensatory response whereby compromised PV inflow induces
a proportional increase in hepatic arterial flow. While this constitutes an adaptive
response in FALD, the HABR is finite and may not fully compensate for the altered
hemodynamics associated with the Fontan circulation.
Relative hypoxia can also instigate a fibrogenic response in the setting of FALD.
A recent retrospective review of 32 adults with Fontan physiology who underwent exercise
catheterization confirmed an inverse relationship between oxygen delivery indices
and serologic liver fibrosis scores (APRI and Fib-4).[91] In this study, APRI and Fib-4 scores correlated with resting and exercise Fontan
pressures. While serologic liver fibrosis scores have demonstrated poor correlation
with fibrosis grade on biopsy, they have been associated with all-cause mortality
in patients with Fontan physiology.[39] These results therefore suggest that HV hypertension and oxygen delivery may contribute
to the progression of FALD.
Recruitment of Immune Cells by Mechanocrine Signaling in Liver Sinusoidal Endothelial
Cells
CH and FALD lack the dense inflammatory infiltrates that characterize other liver
diseases, such as alcohol-associated liver disease and viral hepatitis. Imaging mass
cytometry analysis of liver tissue obtained from patients with FALD and CH has in
fact revealed a decrease in a majority of adaptive immune cell populations.[92] Regardless, studies reveal that recruitment of certain innate immune cell populations
by liver sinusoidal endothelial cells (LSECs) comprises a critical component of the
pathophysiology of congestive fibrosis and PHTN.
LSECs are the initial sensors of disturbances in the portal and hepatic venous circulations
and are subjected to various abnormal mechanical forces in the setting of FALD, including
stretch, static blood flow, and increased stiffness related to congestion. Furthermore,
the mechanical forces that LSECs experience in the setting of FALD are likely to vary
according to location in the sinusoid. For example, LSECs proximal to the central
vein are most likely to be impacted by the elevated CVP induced by the Fontan circulation.
They are therefore exposed to greater magnitudes of stretch and stasis than periportal
LSECs experience. This variable mechanical environment impacts the expression of zonated
genes in LSECs. While expression of the cytokine CXCL9 is more prominent in periportal
LSECs in normal liver and in other etiologies of liver disease,[93] CXCL9 is greatest in the pericentral subpopulation of LSECs in the setting of hepatic
congestion. CXCL9 induces the recruitment of macrophages, which contribute to congestive
fibrosis and PHTN. This pericentral expression of CXCL9 with subsequent macrophage
recruitment may contribute to the prominence of fibrosis around the central vein in
this disease.
The Fontan circulation represents a thrombophilic state.[94]
[95]
[96]
[97]
[98]
[99]
[100] Microvascular thromboses have previously been identified as critical mediators in
the process of parenchymal extinction while ultimately leads to fibrosis.[101] LSEC responses to mechanical force also contribute to the formation of microvascular
sinusoidal thromboses in the setting of CH and FALD.[37] Furthermore, microvascular thromboses contribute to PHTN through pressure and volume
effects in liver sinusoids. When subjected to mechanical forces that recapitulate
hepatic congestion, LSECs secrete the neutrophil chemotactic cytokine CXCL1. CXCL1
leads to the recruitment and accumulation of neutrophils in the liver sinusoids where
they form complexes with platelets. Neutrophil–platelet interactions instigate the
formation of prothrombotic neutrophil extracellular traps (NETs) in liver sinusoids,
which lead to the development of sinusoidal microvascular thromboses. Inhibition of
NET formation decreases congestive fibrosis and PHTN by preventing the formation of
sinusoidal thromboses.[102]
Hepatocyte Communications with Hepatic Stellate Cells
In addition to pericentral LSECs, the pericentral hepatocyte population is impacted
in Fontan physiology. A recent study employed single-cell multiomics to characterize
the transcriptomic and epigenomic changes that characterize FALD.[103] This study included liver biopsy specimens obtained from four patients with early-stage
FALD and two healthy age-matched healthy controls. The population of pericentral hepatocytes
demonstrated significant transcriptional and epigenomic changes leading to metabolic
reprogramming of hepatocytes. The authors then examined interactions between hepatocytes
and hepatic stellate cells (HSCs) to identify mechanisms of fibrogenesis mediated
by hepatocytes. They identified more than 100 ligand–receptor pairs between pericentral
hepatocytes and HSCs, some of which showed increased expression in FALD. The activin
family of ligands (INHBA, INHBB, and INHBC) in hepatocytes and their corresponding receptors in HSCs (ACVR1B, ACVR2A, and ACVR2B) are increased in FALD. The authors postulate that hepatocyte crosstalk with HSCs
contributes to fibrogenesis in the setting of FALD and suggest that approaches targeting
the metabolic dysregulation in pericentral hepatocytes hold therapeutic potential
in FALD.
Gut Dysbiosis and the Microbiome
Gut dysbiosis has been implicated in the pathophysiology of several chronic diseases,
including obesity, type 2 diabetes mellitus, and heart failure.[104]
[105] Cardiac dysfunction impacts the microbiome through intestinal edema and impaired
gut perfusion, which can induce bacterial translocation. A recent study examined the
microbiome in 155 patients with Fontan circulation through examination of fecal samples
and correlated this information with hemodynamic characteristics, assessment of liver
disease severity, and evidence of other end-organ dysfunction.[106] This study revealed significant differences in phyla, including a higher Bacteroidetes/Firmicutes
ratio in patients with Fontan physiology compared with healthy controls. The α-diversity,
representing mean species diversity, was low in patients with Fontan physiology, with
further reductions in patients with FF. Reduced α-diversity was associated with higher
APRI scores. Patients with a low α-diversity were found to have higher risk of hospitalization
for heart failure. Natural log-transformed C-reactive protein (lnCRP) levels correlated
with α-diversity in Fontan patients and with risk of hospitalization for heart failure.
The authors postulate that gut dysbiosis contributes to systemic inflammation and
progression of complications associated with Fontan physiology. The exact role of
gut dysbiosis in the progression of liver fibrosis and complications associated with
FALD requires further study.
Clinical Consequences and Management of Fontan-Associated Liver Disease
Clinical Consequences and Management of Fontan-Associated Liver Disease
Portal Hypertension
PHTN represents a heterogeneous syndrome in FALD. Patients with Fontan physiology
may have posthepatic PHTN secondary to cardiac dysfunction. Alternatively, patients
with more advanced fibrosis or cirrhosis may also have sinusoidal PHTN. The clinical
manifestations of PHTN in these patients are accordingly variable, and studies suggest
that more advanced PHTN portends a worse outcome in patients with Fontan physiology.[107] A retrospective study found that sequelae of PHTN included in the Varices, Ascites,
Splenomegaly, or Thrombocytopenia (VAST) score, with one point allocated each for
varices, ascites, splenomegaly, or thrombocytopenia, were associated with major adverse
events, including death, need for transplant, or diagnosis of HCC.[107] This highlights the salience of PHTN in outcomes of patients with Fontan physiology.
Varices and Variceal Bleeding
The true prevalence and risk associated with varices in the setting of Fontan physiology
remains poorly understood. Esophageal varices can occur independent of liver fibrosis
or cirrhosis in patients with Fontan physiology. In this context, “downhill” esophageal
varices can occur in the upper esophagus due to chronic elevation in central venous
and pulmonary pressures. This leads to diffuse dilation of the periesophageal venous
plexus and eventually culminates in the formation of venovenous communications.[17] In contrast, varices located in the lower third of the esophagus may be secondary
to PHTN.
The prevalence of varices in Fontan physiology appears to be dependent on age and
duration of Fontan circulation.[17] The reported incidence of esophageal varices in adults with Fontan physiology ranges
from 19 to 43%, with an annual incidence of 9% per year.[108] In contrast, an incidence of 9% has been reported in children.[53] Despite the widespread presence of varices in patients with Fontan physiology, the
incidence of variceal bleeding is low and ranges from 5 to 6% in retrospective studies.[109] In a retrospective study of the incidence of varices and variceal bleeding among
149 patients with Fontan physiology, higher HV wedge pressure and HV pressure gradient
were associated with bleeding complications.[108] In contrast, the rate of variceal bleeding among patients with decompensated cirrhosis
due to other etiologies of liver disease is approximately 10 to 15% per year.[110] The reason for the decreased risk of variceal bleeding in patients with Fontan physiology
is unclear.
Recent EASL guidelines recommend screening patients with Fontan physiology for varices.[17]
[107]
[111] Clinical decision-making pertaining to screening for varices must also take into
account the procedural and sedation risk associated with Fontan physiology. Management
of variceal bleeding and secondary prophylaxis thereafter in the setting of Fontan
physiology is similar to other etiologies of liver disease.[112] However, transjugular intrahepatic portosystemic shunt is not indicated in patients
with FALD due to the absence of a subpulmonic ventricle.[17]
Ascites
Ascites in patients with Fontan physiology may be multifactorial and requires a thorough
evaluation. Retrospective studies identify ascites in 4 to 58% of patients with Fontan
physiology.[36]
[46]
[113]
[114]
[115]
[116]
[117] Patients with new-onset ascites should undergo paracentesis with calculation of
a serum albumin ascites gradient (SAAG) to determine if the ascites is of cardiac
or hepatic etiology. A SAAG > 1.1 suggests that ascites is due to PHTN. An ascites
fluid total protein > 2.5 g/dL with a concomitant SAAG > 1.1 suggests that ascites
is posthepatic and likely due to cardiac dysfunction. A low ascites fluid total protein
suggests the possibility of hepatic dysfunction. It is important to note that PLE
can also present with ascites. Therefore, thorough evaluation of cardiac function
and comorbid symptoms is warranted in patients with Fontan physiology who develop
ascites.[17] After a thorough evaluation of cardiac function, ascites due to PHTN in the setting
of FALD can be managed with sodium restriction, diuretics, and paracentesis as needed.
Hepatocellular Carcinoma
Patients with Fontan physiology are at increased risk of HCC compared with the general
population. A retrospective study of 103 patients with Fontan physiology identified
cumulative incidence rates of HCC of 0, 7, and 13% at 10, 20, and 30 years after Fontan
surgery, respectively.[118] Another retrospective study identified 339 patients who underwent Fontan surgery
between 2005 and 2019. In this cohort, the annual incidence of HCC was 0.14% in the
second decade after the Fontan procedure, 0.43% in the third decade, and 8.83% in
the fourth post-Fontan decade.[119] While the median duration of time from Fontan surgery to a diagnosis of HCC is 22
years in published studies, the youngest patient identified with HCC in the setting
of FALD was 12 years old.[120]
[121] The pathophysiology of HCC in FALD is unclear, and case series report that advanced
fibrosis and cirrhosis are present in only approximately 50% of patients with FALD
who are diagnosed with HCC.[122] In contrast, in other chronic liver diseases, HCC occurs in the absence of cirrhosis
in approximately 20% of cases.[123]
A recent multicenter, retrospective case–control study examined clinical characteristics
associated with the development of HCC in patients with Fontan physiology 18 years
of age and older.[122] A total of 58 cases out of 3251 patients with Fontan physiology were identified
in this study. While prior retrospective studies report an increase in the incidence
of HCC with longer duration after the Fontan surgery,[119] neither age at Fontan nor time since Fontan surgery were different between cases
and controls in this study. In addition, there was no difference between cases and
controls in age, body mass index, sex, or underlying CHD. In this study, cases were
more likely to have older versions of the Fontan surgery (atriopulmonary or atrioventricular
connections). Patients with Fontan physiology who were diagnosed with HCC had more
cardiac comorbidities as well, including arrhythmia, desaturation, heart failure,
lymphatic disorders, and a history of prior Fontan revision or valve surgery. Other
studies have identified situs inversus anatomy and lack of anticoagulation with warfarin
as risk factors for HCC in multivariate analysis.[118]
Transplantation in Patients with Fontan-Associated Liver Disease
Transplantation in Patients with Fontan-Associated Liver Disease
The improved survival of patients with Fontan physiology into adulthood has led to
increased need for isolated heart transplantation (IHT) or combined heart–liver transplantation
(CHLT) in this patient population.[124]
[125] While older studies reported high perioperative mortality for patients with CHD
undergoing heart transplant,[126]
[127]
[128] more recent studies report long-term survival rates of 80.3 and 71.2% after heart
transplantation in patients with Fontan physiology.[129] The approach to transplant evaluation and decision-making in patients with Fontan
physiology is nuanced and highlights the high-risk nature of transplant in this patient
population. In particular, indications for IHT versus CHLT in this patient population
remain unclear.[37]
Indications for Transplantation
Heart transplantation currently comprises the only definitive, curative treatment
strategy for patients with failure of the Fontan circulation.[129] The recommendation to proceed with evaluation for heart transplantation in patients
with Fontan physiology may be instigated due to the onset of FF, PLE,[130] thromboembolic events, cyanosis, and refractory arrhythmia with decompensations.[37] While clinically significant PHTN or HCC are widely considered contraindications
to IHT,[37] the degree of liver dysfunction that precludes proceeding safely with IHT alone
remains unclear. This has led to significant intercenter variability pertaining to
recommendations for IHT versus CHLT for patients with Fontan physiology.
In addition, transplant decision-making varies in pediatric and adult populations.
Some retrospective studies suggest that children undergoing transplant within a few
years of the Fontan surgery may be considered for IHT alone. For example, a retrospective
study of nine adolescent patients with Fontan physiology assessed liver function before
and after IHT.[131] All patients undergoing IHT in this series had pretransplant MELD scores < 15, MELD-XI
scores < 16, and VAST scores of 2 or lower. No significant posttransplant decompensations
were noted, and VAST scores demonstrated significant improvement one year after IHT.
Pre- and posttransplant liver biospies were available in 4/9 patients and did not
demonstrate significant change 1-year posttransplant. Another study described outcomes
after IHT in 19 children and adolescents with Fontan physiology (median age at IHT
was 12 years).[132] Two patients had imaging evidence of cirrhosis prior to IHT. Overall survival for
this cohort was 76% at a median follow-up of 4.3 years posttransplant. None of the
patients included in this study developed perioperative or posttransplant liver failure.
While these studies are limited by their small size and retrospective design, they
suggest that IHT may be feasible in select children and adolescents with FALD.
The Fontan Outcomes Study to improve Transplant Experience and Results (FOSTER) Collaboration
includes patients seen at 17 adult congenital heart disease and transplant centers
in the United States and Canada[133]
[134] and has attempted to clarify the risks and outcomes associated with transplant in
adult patients with Fontan physiology. This study included 131 patients, 40 of whom
underwent CHLT. Among patients undergoing IHT alone, several predictors of 1- and
5-year posttransplant mortality were identified, including time from the diagnosis
of FF (defined as progressive volume overload, declining functional status, need for
hospitalization, and refractory arrhythmia) to transplant evaluation, presence of
venovenous collaterals and bilateral lower extremity varicosities, cardiopulmonary
bypass time, need for posttransplant renal replacement therapy, need for reoperation
for bleeding, and need for mechanical circulatory support.
In the event of posttransplant liver decompensation, liver transplantation (LT) may
be pursued following IHT. Reports of outcomes of sequential LT after IHT are rare.
A retrospective analysis of Organ Procurement and Transplantation Network data identified
six adults who were listed for LT after IHT for CHD.[135]
[136] Four of the six patients identified in this analysis died on the waiting list for
LT. This suggests that patients with CHD listed for LT due to post-IHT liver decompensation
may experience high waitlist mortality. Further studies of this transplant population
are needed.
Combined Heart–Liver Transplantation in Patients with Fontan Physiology
Recent studies have attempted to clarify the risks and outcomes among adult patients
with FALD undergoing IHT versus CHLT[133]
[134] through creation of the FOSTER Collaboration. This study group devised the FALD
score to stratify liver disease severity, which includes a history of two or more
paracenteses, evidence of cirrhosis on imaging or pathology, splenomegaly, and identification
of esophageal varices. This study included 131 patients, 40 of whom underwent CHLT.
Patients undergoing CHLT were older at the time of transplant and had longer mean
time from Fontan surgery to transplant. CHLT recipients were also more likely to have
imaging evidence of cirrhosis and PHTN and also had a higher FALD score. In this study,
survival rates did not differ 1-year posttransplant, but 5-year survival was better
among patients who underwent CHLT (52 vs. 86%). FALD scores of 2 or above were associated
with worse survival among all IHT and CHLT recipients. While other single-center studies
suggest no significant survival difference among patients with Fontan physiology who
undergo IHT versus CHLT,[137]
[138]
[139]
[140] the authors concluded from this study that CHLT may confer a survival advantage
in patients with evidence of advanced FALD. In addition, patients with clinical sequelae
of FALD, including decompensations related to PHTH, demonstrated inferior outcomes
compared with those patients with less severe liver disease. The authors speculated
that earlier referral for transplant among patients with FF may lead to improved outcomes
after transplantation.
Concluding Remarks
FALD is a complex disease that poses unique challenges to researchers and clinicians.
In the realm of clinical evaluation and monitoring of FALD, additional studies are
needed to clarify the utility and role of noninvasive testing in diagnosing and staging
FALD. Given the relatively small size of the population of patients with Fontan physiology,
multicenter, collaborative efforts are needed to ascertain study populations of sufficient
size and to allow for more confident conclusions that can impact clinical practice.
Basic and translational studies have improved our understanding of the unique pathophysiology
of this complex disease. The advent and rapid development of spatial transcriptomic
and proteomic studies offers an opportunity to advance of our understanding of FALD
given the uniquely heterogeneous nature of this disease. While small animal models
of CH have been studied,[141] existing murine models are not exactly replicative of Fontan physiology. While larger
animal models in sheep have been developed,[142] these models are more expensive and labor-intensive. As a result, the development
of a more scalable animal model is needed to advance our understanding and therapeutic
innovation in FALD.
