Fontan‐Associated Liver Disease: Trends Over Time and Relationship With Outcomes Authors: Mikyla Janzen, Parm Khakh, Defen Peng, Gnalini Sathananthan, Marla Kiess, Santabhanu Chakrabarti, Jasmine Grewal Published: 2025 ABSTRACT AbstractBackgroundFontan associated liver disease is a poorly understood complication of long‐term single ventricle palliation. This study sought to describe serial liver function changes in patients with a Fontan circulation from 2014 to 2020 and determine if these findings can be used to predict adverse clinical outcomes.MethodsWe conducted a retrospective chart review of Fontan patients age≥18 who were followed by a specialist clinic between 2014 and 2020. Annual bloodwork, including liver enzymes and AFP (alpha‐fetoprotein), and liver imaging were analyzed. Common liver function scores were calculated. Trends over time and their correlation to clinical events were assessed. Events were defined as congestive heart failure, cardiac hospitalization, thrombosis, protein losing enteropathy, and death. The study assessed possible associations with the composite event of adverse outcomes.ResultsOne hundred twelve Fontan patients were included with a mean age of 24.9±8.1 years. Age‐adjusted AFP and creatinine increased over the study period, whereas alkaline phosphatase decreased. There was no significant change in other liver biomarkers over the study period. The proportion of patients with higher liver imaging scores increased over the study period.Over a follow‐up time of 6.1 years (95% control limit, 5.5–6.9 years), 44 (39.3%) patients experienced at least 1 event, most commonly cardiac hospitalization. In multivariable analysis, age was associated with a greater risk of events (hazard ratio, 1.035 [95% CI, 1.008–1.063]; P=0.010); no other biomarkers or scores were significant.ConclusionsRoutine liver biomarkers do not increase over time despite increasing proportion of patients with higher liver imaging scores over the same time. Although our sample size was limited, current modalities to assess progression of liver dysfunction on bloodwork and imaging were not related to clinical outcomes.Keywords: Fontan‐associated liver disease, Fontan procedure, liver disease, congenital heart diseaseSubject Categories: Congenital Heart Disease Nonstandard Abbreviations and Acronyms alpha‐etoprotein Fontan‐associated liver disease hepatocellular carcinoma Clinical Perspective What Is New? Although we routinely measure liver biomarkers and composite risk scores, there is no significant change in these over the mid‐term despite progression of Fontan‐associated liver disease as per liver imaging scores. Although we routinely measure liver biomarkers and composite risk scores, there is no significant change in these over the mid‐term despite progression of Fontan‐associated liver disease as per liver imaging scores. Routine liver biomarker and imaging modalities are not related to clinical outcomes in the Fontan population, however, this study has a limited sample size. Routine liver biomarker and imaging modalities are not related to clinical outcomes in the Fontan population, however, this study has a limited sample size. What Are the Clinical Implications? There is no absolute reliable method of assessing Fontan‐associated liver disease, however, regular liver biomarker assessments are likely not clinically additive to imaging. There is no absolute reliable method of assessing Fontan‐associated liver disease, however, regular liver biomarker assessments are likely not clinically additive to imaging. The Fontan operation directs systemic venous blood directly to the pulmonary circulation, thereby preventing intracardiac mixing, in individuals with single ventricle physiology. As a result, arterial oxygen saturation is increased and volume load is decreased on the single functioning ventricle. 1 The Fontan procedure is completed in early childhood and was first described only 5 decades ago; as such, the first large cohort of patients is currently surviving well into adulthood. 2 The tradeoff of this palliative surgery is that patients experience chronic high venous pressures and decreased cardiac output, which increases the risk of extracardiac organ damage, such as Fontan‐associated liver disease (FALD). 3 The pathophysiology of FALD is not well documented aside from the state of chronic passive congestion disturbing the liver's blood supply and drainage. 4 FALD puts individuals at risk of progressive liver fibrosis, cirrhosis, and hepatocellular carcinoma. 4 , 5 , 6 , 7 The prevalence of FALD is highly variable across studies from 41% to 86%, although experts agree that some degree of FALD is inevitable given the chronic circulatory changes. 5 , 8 , 9 , 10 Effective screening methods are required to ensure patients with progressive FALD are identified to facilitate early appropriate interventions to prevent or mitigate clinical complications. 1 , 7 At the current time, routine noninvasive screening for FALD is recommended and includes serum liver enzymes and liver imaging. There are, however, many unknowns with proposed surveillance protocols. 4 , 5 The extent and trend in liver enzymes are not known to correlate strongly with fibrosis and hence their utility is unknown. 3 The threshold for an abnormal AFP (alpha‐fetoprotein), a tumor biomarker for hepatocellular carcinoma (HCC) in the general population, is not known in the Fontan population. 11 In the general population, thresholds between 20 and 400 ng/mL or an increase of ≥10% over 6 months have been used as clinically relevant cutoffs for HCC prediction and diagnosis. 12 , 13 , 14 Liver imaging by ultrasound, computed tomography, or magnetic resonance imaging (MRI) can reveal advanced liver disease. However, there is substantial inter‐ and intraobserver variability in image assessment and reduced sensitivity in patients with earlier onset liver disease. 15 , 16 The objectives of this study were as follows: (1) to determine changes in liver biomarkers and imaging parameters in Fontan patients over time and (2) to determine the relationship of these liver parameters with clinical outcomes. METHODS This is a single‐center retrospective cohort study of individuals with prior Fontan surgery being followed through the Yasmin and Amir Virani Provincial Adult Congenital Heart Program at St. Paul's Hospital in Vancouver, Canada. Inclusion criteria were as follows: history of Fontan surgery, ≥18 years old, and followed by our specialized adult congenital heart disease clinic from 2014 to 2020. Patients were excluded if they did not undergo any imaging studies over the time period. Data collected from chart review included the following annually from 2014 to 2020: bloodwork as outlined next, abdominal imaging results (ultrasound, computed tomography, or MRI), and clinical variables. Clinical variables included surgical and medical history, medications and symptoms at last follow‐up, and adverse outcomes. Data, analytic methods, and study materials are not available to other researchers. The study was approved by the University of British Columbia Research Ethics Board; as it was retrospective and observational, participant consent was not required. Bloodwork Assessment Comprehensive bloodwork was collected annually including liver enzymes (alkaline phosphatase, gamma‐glutamyl transferase, aspartate transaminase [AST], alanine transaminase), AFP, and other hematologic results (hemoglobin, platelets, bilirubin, albumin, creatinine, international normalized ratio). For patients who had multiple measurements in a given year, the most comprehensive analysis was taken. If multiple complete bloodwork panels were completed, the test closest to 12 months apart from other years was used. Normal ranges for the liver enzymes were defined as alkaline phosphatase 30 to 105 U/L, gamma‐glutamyl transferase <50 U/L, AST 15 to 45 U/L, and alanine transaminase 15 to 55 U/L, as per standard local laboratory limits. Normal ranges of AFP are defined as <8 ng/L as per standard laboratory limits, and a value for high‐normal range of 5 ng/L was determined arbitrarily by specialists. Composite scoring systems which have been validated in patients with other forms of liver disease were calculated to assess the liver in the Fontan population. These included Model for End Stage Liver Disease Excluding International Normalized Ratio, aspartate transaminase to platelet ratio index, fibrosis 4 score, and varices, ascites, splenomegaly, and thrombocytopenia (VAST). 6 , 17 , 18 , 19 Abdominal Imaging Abdominal imaging reports were assessed for markers of FALD. In patients who had multiple images completed in a given year, MRI was prioritized over computed tomography, which was prioritized over ultrasound. The presence of ascites or varices was recorded. Spleen size was categorized as normal, splenomegaly (length>13 cm), or absent in asplenic patients. Liver appearance was qualitatively reported; to facilitate analysis, this was quantified into 4 categories as previously described by Lindsay et al. to determine the “liver imaging score.” 15 A score of 1 was assigned to normal imaging, 2 to minor changes including early fibrosis and arterial nodules, 3 for cirrhosis, and 4 for cirrhosis with enhancing nodules. Outcomes Adverse outcomes included congestive heart failure (CHF), protein losing enteropathy, thromboembolism, hospitalization for cardiac or hepatic reasons, cardiac intervention or surgery, and death. Hospitalizations were adjudicated separately from clinical outcomes; for example, hospitalization for decompensated CHF was recorded as both a CHF event and a hospitalization event. CHF was defined as New York Heart Association Class III to IV with reduced systemic ventricular ejection fraction or pulmonary/peripheral edema requiring diuresis; no patients had documented heart failure with preserved ejection fraction. Protein losing enteropathy was defined as symptoms of fluid overload (eg, peripheral edema) with hypoalbuminemia and elevated fecal alpha‐1 antitrypsin. Thrombosis was considered a confirmed thromboembolic event (eg, transient ischemic attack or stroke), visualized thrombosis on imaging, or clinically suspected thrombosis acutely resolved with anticoagulation. Cardiac thrombosis included any thombus in the atria, ventricle, Fontan circuit, or shunt material. Cardiac interventions included Fontan revision, fenestration closure, balloon dilatation, stent insertion, valve replacement, or cardiac catheterization. Cause of death was recorded from clinical records where available. New liver imaging scores in each year were reported in this study. A correlation study was conducted to explore the associations with the adverse outcomes. Statistical Analysis Missing data at baseline were infrequent (<1% for most variables). All the calculations and analyses were conducted with complete data. Continuous variables were reported as mean±SD or median (25th to 75th interquartile range), except for survivals, which were reported as mean±SE. Categorical variables were presented as frequency (percentage). The formulas used to calculate Model for End Stage Liver Disease Excluding International Normalized Ratio, aspartate transaminase to platelet ratio index, fibrosis 4 score, and VAST in this study are as follows: Model for End Stage Liver Disease Excluding International Normalized Ratio=5.11×ln (serum bilirubin in mg/dL)±11.76×ln (serum creatinine in mg/dL)±9.44 Model for End Stage Liver Disease Excluding International Normalized Ratio=5.11×ln (serum bilirubin in mg/dL)±11.76×ln (serum creatinine in mg/dL)±9.44 Aspartate transaminase to platelet ratio index =[(AST/40)×100]/platelet count Aspartate transaminase to platelet ratio index =[(AST/40)×100]/platelet count Fibrosis 4 score =(Age in years×AST in U/L)/([platelet count in 109/L]×[ alanine transaminase in U/L]1/2) Fibrosis 4 score =(Age in years×AST in U/L)/([platelet count in 109/L]×[ alanine transaminase in U/L]1/2) VAST=1 point each for varices, ascites, splenomegaly (>13 cm), thrombocytopenia (<150) VAST=1 point each for varices, ascites, splenomegaly (>13 cm), thrombocytopenia (<150) The mean changes of longitudinal measures were tested using 1‐way repeated measures ANOVA or nonlinear mixed model and illustrated with trend graphs. Kaplan–Meier methods were used to examine survival by the categorical factors studied. Categorical predictors of outcomes were individually tested for equality of survival with log‐rank test. Risk factors associated with the composite outcome was first analyzed with Cox proportional hazards regression model testing each determinant of covariate variable (demographics, clinical and interventions, history of cardiac events, medications, and bloodwork at baseline) individually. Next, in the multivariable Cox regression analyses, we included potential independent risk factors with all determinants from the first analysis, excluding those missing multiple variables and those having low cell numbers (such as direct oral anticoagulant, protein losing enteropathy, angiotensin receptor blocker and novel oral anticoagulant). Model fit was assessed by means of the Akaike information criterion. The proportional hazards assumption of the Cox regressions was tested based on Schoenfeld residuals. The conventional 5% level of significance was used as a nominal reporting level. All statistical analyses were performed using SAS software version 9.4 and R software version 4.4.1. 20 , 21 RESULTS One hundred twelve Fontan patients were included with a mean age of 24.9±8.1 years and 67 (59.8%) were male (Table 1). Average time since Fontan was 24.9±6.0 years; there were 4 patients who underwent their Fontan procedure over the age of 20. Overall, in keeping with standard practice, the majority of patients underwent their Fontan procedure by the age of 5. The most common single‐ventricle diagnosis was tricuspid atresia (31.3%) followed by double inlet left ventricle (21.4%), and the majority had a morphological left ventricle (73.0%). This was a contemporary Fontan population and the majority had an extracardiac (45.5%) or lateral tunnel (33.9%) Fontan. Medication use was common with any antiarrhythmic therapy in 46 (41.2%), anticoagulation in 44 (39%), and antiplatelets in 46 (41.1%). History of cardiac events included arrhythmias (34.8%) and thromboembolism (21.4%) (Table 1). Forty‐eight (42.9%) patients had at least 1 previous cardiac intervention post Fontan. Fenestration closure was the most common (n=21; 18.8%). Table 1Baseline CharacteristicsVariablesSummary (total n=112)Demographics and clinicalAge, y24.9±8.1 yY since Fontan24.9±6.0 yMale sex67 (59.8%)Primary diagnosisHypoplastic left heart syndrome8 (7.1%)Double inlet left ventricle24 (21.4%)Tricuspid atresia35 (31.3%)Pulmonary atresia19 (17.0%)Atrioventricular septal defect5 (4.5%)Other21 (18.8%)Single ventricle morphologyLeft81 (73.0%)Right28 (25.2%)Mixed2 (1.8%)Type of FontanClassic20 (17.9%)Lateral tunnel38 (33.9%)Extracardiac51 (45.5%)Other3 (2.7%)MedicationsAngiotensin‐converting enzyme inhibitor/angiotensin receptor blocker46 (41.2%)Antiarrhythmic medication (excluding beta blocker)11 (9.8%)Beta blocker22 (19.6%)Digoxin12 (10.7%)Diuretic13 (11.6%)Warfarin42 (37.5%)Direct oral anticoagulant2 (1.8%)Antiplatelet46 (41.1%)Other cardiac medication22 (19.6%)History of cardiac eventsCongestive heart failure3 (2.7%)Arrhythmia39 (34.8%)Protein losing enteropathy4 (3.6%)Hepatocellular carcinoma3 (2.7%)Thromboembolism24 (21.4%)Transient ischemic attack/stroke13 (11.6%)Cardiac thrombosis6 (5.4%)Deep vein thrombosis/pulmonary embolism5 (4.5%)History of cardiac interventionNumber of surgeries before Fontan<236 (32.1%)≥276 (67.9%)Fenestration closure21 (18.8%)Fontan revision3 (2.7%)Pulmonary artery (left or right) stent/balloon dilatation/angioplasty6 (5.4%)Collateral vessel coil occlusion3 (2.7%)Ablation6 (5.4%)Pacemaker17 (15.2%)Other1 (0.9%) Baseline Characteristics Longitudinal bloodwork data and composite scores are presented in Figure 1 and Table 2. There was no difference in these results before or after adjusting for age. There was limited AFP data from 2014 to 2016. However, when analyzing 2017 to 2021 data there was a significant increase over time (Table 2). Although this change was statistically significant, neither the absolute values nor rate of change is likely to be clinically significant regarding HCC risk. 12 , 13 , 14 Of note, 3 large outlier values of AFP were removed for the trend analysis. Two AFP values of 520 ng/mL and 1870 were from 1 patient who had active progressing HCC and passed away in 2019. The other (AFP=130 ng/mL) was from a patient who was pregnant at the time and had no other AFP values available outside of pregnancy. Of note, the AFP levels in the other 2 patients with HCC were 6.6 ng/mL and 3.3 ng/mL at year of diagnosis, respectively. Creatinine also increased (P<0.001) and alkaline phosphatase decreased (P=0.002) over the study period; all other liver enzymes and hematological data remained stable (Figure 1A, Table 2). Composite liver scores did not significantly change over time (Figure 1B, Table 2). However, there was an increase in the proportion of patients with higher liver imaging scores over time (Figure 1C, Table 2). With the exception of alkaline phosphatase, there was no difference in the majority of biomarkers between those with higher versus lower liver imaging scores (Table 3). There was a trend to higher VAST score in those with a higher imaging score. Notaby AST, alanine transaminase, bilirubin, and AFP were not increased in those with higher imaging scores (Table 3). Trends plot of longitudinal variables. A, Biomarkers (zoomed in 30 for INR and AFP); B, Composite liver scores (zoomed out 10 for MELD‐XI); C, Liver imaging score. AFP indicates alpha‐feto protein; ALP, alkaline phosphatase; ALT, alanine transaminase; AST, aspartate aminotransferase; GGT, gamma‐glutamyl transferase; INR, international normalized ratio; and MELD‐XI, model for end‐stage liver disease without INR. Table 2Summary of Longitudinal VariablesVariable20142015201620172018201920202021 P valueHemoglobin, g/L152.6±18.5151.3±19.9152.2±20.5152.8±20.2150.6±23.3154.5±20.8152.3±23.8152.4±19.40.77Platelets173.1±62.3179.4±61.4171.2±59.0171.4±58.3166.9±54.3170.2±53.9169.1±55.6167.7±55.40.79Creatinine, μmol/L72.0±18.273.8±18.272.8±18.774.5±23.374.8±16.472.4±15.575.7±17.674.1±15.0<0.001INR1.7±0.71.8±0.81.9±1.01.7±0.71.8±0.91.8±0.71.9±0.82.0±1.00.06Alkaline phosphatase94.4±51.489.4±38.189.3±34.885.6±30.686.0±30.983.7±26.280.2±26.481.1±25.50.002Alanine aminotransferase38.3±45.034.2±17.632.0±14.231.7±14.535.1±47.831.6±15.729.1±15.731.6±16.60.28AST34.9±33.731.3±13.930.3±10.429.5±10.434.5±51.632.2±29.928.4±11.627.1±7.70.50Gamma‐glutamyl transferase97.5±105.092.2±70.8100.2±87.495.1±92.4100.8±84.086.9±60.682.2±52.684.6±57.40.49AFP* NANANA3.6±2.52.6±1.93.7±3.13.7±3.04.1±3.2<0.001Total bilirubin, μmol/L17.9±12.318.6±14.619.4±11.717.6±11.219.7±11.119.2±11.617.7±10.921.2±14.30.78Model for End‐Stage Liver Disease Excluding INR #20.6±4.121.0±4.121.1±4.421.3±3.721.7±3.421.2±4.021.2±4.021.9±3.10.25Fibrosis‐41.0±0.71.0±0.81.0±0.91.1±1.11.2±1.01.2±1.31.1±0.81.1±0.80.16AST‐to‐platelet ratio index0.6±0.80.5±0.40.5±0.30.5±0.40.6±1.00.5±0.50.5±0.30.4±0.20.55VAST† 0.9±1.00.6±0.80.8±0.80.8±0.90.9±0.90.8±0.90.8±1.00.7±1.00.32Liver imaging score<0.001Score 115 (46.9)10 (43.5)13 (33.3)12 (27.9)12 (22.2)11 (19.0)6 (14.3)4 (11.8)Score 216 (50.0)12 (52.2)24 (61.5)28 (65.1)36 (66.7)40 (69.0)29 (69.0)24 (64.9)Score 31 (3.1)1 (4.3)1 (2.6)3 (7.0)5 (9.3)4 (6.9)5 (11.9)7 (18.9)Score 40 (0.0)0 (0.0)1 (2.6)0 (0.0)1 (1.9)3 (5.1)2 (4.8)2 (5.4)Values shown as mean±SD or n (%). Calculation based on complete data.*3 extreme values of AFP were removed. † Values for VAST in 2020 were imputed by taking the averages of the values in 2019 and 2021.AFP indicates alpha‐feto protein; AST, aspartate aminotransferase; INR, international normalized ratio; and VAST, varices, ascites, splenomegaly, and thrombocytopenia. Summary of Longitudinal Variables Values shown as mean±SD or n (%). Calculation based on complete data. 3 extreme values of AFP were removed. Values for VAST in 2020 were imputed by taking the averages of the values in 2019 and 2021. AFP indicates alpha‐feto protein; AST, aspartate aminotransferase; INR, international normalized ratio; and VAST, varices, ascites, splenomegaly, and thrombocytopenia. Table 3Biomarkers and Composite Liver Scores Stratified by Liver ImagingVariables* All n=108Score 1 or 2 n=84Score 3 or 4 n=24 P valueHemoglobin153.1±20.8151.8±19.8157.7±24.10.25Platelets170.5±55.7170.8±56.7169.5±52.90.93Creatinine76.2±23.674.4±14.182.5±42.20.15International normalized ratio1.9±1.11.8±1.12.0±1.10.52Alkaline phosphatase83.1±28.478.7±27.198.9±27.90.003Alanine aminotransferase30.3±13.530.4±14.029.7±12.20.81AST29.2±10.728.6±9.831.2±13.50.35Gamma‐glutamyl transferase93.5±83.182.6±55.4131.1±137.60.19Alpha‐feto protein3.8±2.94.1±3.22.6±1.20.046Total bilirubin20.2±11.620.4±12.319.6±9.10.76Model for End‐Stage Liver Disease Excluding INR22.2±3.522.1±3.522.4±3.80.75Fibrosis‐41.1±0.81.0±0.81.3±0.80.26AST‐to‐platelet ratio index0.5±0.30.5±0.30.5±0.20.87Varices, ascites, splenomegaly, and thrombocytopenia0.8±0.90.7±0.81.1±1.10.07AST indicates aspartate aminotransferase; and INR, international normalized ratio.*All the values listed in this table were most recent ones. Biomarkers and Composite Liver Scores Stratified by Liver Imaging AST indicates aspartate aminotransferase; and INR, international normalized ratio. All the values listed in this table were most recent ones. Over a median follow‐up time of 6.1 (95% control limit, 5.5–6.9) years, 44 patients (39.3%) experienced at least 1 event. The most common events were 64 cardiac‐ or liver‐related hospitalizations followed by 24 CHF events (Table 4). Three patients were diagnosed with HCC. Four patients died: 1 due to complications of metastatic HCC, 1 due to Fontan failure, 1 due to a traumatic brian injury, and 1 due to a motor vehicle accident. Table 4Summary of EventsVariables20142015201620172018201920202021TotalHospitalizations391291393664Arrhythmia0052431217CHF exacerbation1344231220Chest pain012120028PLE/hepatic cause0402230011Obstetric001010002Other210020106CHF1253353224PLE020211006Thromboembolism2221231013Death000112004CHF indicates congestive heart failure; and PLE, protein losing enteropathy. Summary of Events CHF indicates congestive heart failure; and PLE, protein losing enteropathy. Overall event‐free survival was 89.3%±2.9% at 1 year follow‐up, 63.8%±4.7% at 5 years, and 56.7%±5.5% at 7 years (Figure 2A). Age and single diuretic use were independently associated with adverse outcomes (composite of CHF, protein losing enteropathy, thromboembolism, and death) (Table 5). The multivariable hazard ratio for age was 1.035 (95% CI, 1.008–1.063; P=0.010) and for diuretic use was 2.993 (95% CI, 1.524–5.877; P=0.002). Liver imaging parameters were not associated with adverse outcomes. Patients taking diuretics were more likely to experience an event during follow‐up compared with those not taking diuretics (Figure 2B; P<0.001). Survival curves. A, Overall survival curve. B, Survival curves by single diuretic. APRI indicates aspartate transaminase‐to‐platelet ratio index; FIB‐4, fibrosis‐4; MELD‐XI, Model for End‐Stage Liver Disease Excluding International Normalized Ratio; and VAST, varices, ascites, splenomegaly, and thrombocytopenia. Table 5Factors Associated With Adverse OutcomesVariableUnivariate analysis P valueMultivariable analysisHazard ratio (95% CI)Hazard ratio (95% CI) P valueDemographicsAge at 2014, y1.047 (1.016–1.079)0.0031.035 (1.008–1.063)0.010Sex (female)0.992 (0.540–1.820)0.98Clinical and interventionsType of Fontan* Lateral tunnel0.534 (0.247–1.156)0.11Extracardiac0.514 (0.251–1.051)0.07Years since Fontan1.045 (0.995–1.098)0.08Single ventricle morphology (right vs left)1.098 (0.565–2.134)0.78Number of surgeries before Fontan (>=2 vs <2)0.630 (0.344–1.152)0.13Surgery post‐Fontan (any vs none)1.240 (0.637–2.416)0.53History of cardiac eventsHistory of congestive heart failure1.011 (0.139–7.360)0.99Arrhythmia1.132 (0.613–2.090)0.69Protein losing enteropathy3.050 (0.931–9.988)0.07Thrombosis0.746 (0.346–1.610)0.46Intervention1.218 (0.643–2.306)0.54Ablation2.167 (0.665–7.060)0.20Antiarrhythmic device0.932 (0.425–2.045)0.86MedicationsAngiotensin‐converting enzyme inhibitor0.997 (0.546–1.820)0.99Antiarrhythmic medication2.750 (0.848–8.917)0.09Antiplatelet1.122 (0.615–2.045)0.71Angiotensin receptor blocker2.374 (0.325–17.33)0.39Beta blocker1.709 (0.879–3.320)0.11Diuretic3.553 (1.787–7.065)<0.0012.993 (1.524–5.877)0.002Novel oral anticoagulant1.973 (0.270–14.42)0.50Warfarin1.056 (0.575–1.941)0.86Other cardiac medications1.276 (0.644–2.528)0.48Other noncardiac medications2.520 (1.315–4.830)0.005Bloodwork at baselineHemoglobin (per 5 unit)1.040 (0.946–1.144)0.42Platelets (per 20 unit)0.959 (0.857–1.074)0.47Creatinine (per 10 unit)1.151 (0.985–1.346)0.08INR0.860 (0.529–1.399)0.54Alkaline phosphatase (per 5 unit)1.007 (0.980–1.034)0.62Alanine transaminase (per 5 unit)1.076 (1.036–1.118)<0.001AST (per 5 unit)1.127 (1.069–1.189)<0.001Gamma‐glutamyl transferase (per 5 unit)1.010 (0.999–1.021)0.07Total bilirubin1.033 (1.006–1.060)0.015Model for End‐Stage Liver Disease Excluding INR1.121 (1.019–1.233)0.019Fibrosis‐42.011 (1.390–2.911)<0.001AST‐to‐platelet ratio index4.474 (2.139–9.361)<0.001Varices, ascites, splenomegaly, and thrombocytopenia1.280 (0.809–2.024)0.29Liver imaging scoreScore 2 vs 12.214 (0.763–6.425)0.14Score 3/4 vs 12.887 (0.574–14.51)0.20*Reference=Classic Fontan.AST indicates aspartate aminotransferase; and INR, international normalized ratio. Factors Associated With Adverse Outcomes Reference=Classic Fontan. AST indicates aspartate aminotransferase; and INR, international normalized ratio. DISCUSSION FALD is a poorly understood consequence of the lifelong circulatory compromise in patients following a Fontan palliation. To inform our current approach to screening for FALD, it is important to better understand expected trends in routine liver marker and imaging parameters over time and whether these are associated with adverse outcomes. Overall, we found that there was no significant change in routinely measured liver biomarkers and composite liver scores over the 7‐year study period in our adult Fontan population. However, the liver imaging score increased over time, suggesting progressive FALD, which was evident on imaging. The stability of liver biomarkers over time despite progressive changes in imaging, support the complexity of assessing FALD and that suggested routine biomarker assessment is not clinically additive at least in the midterm. AFP is a widely used tumor marker, including for HCC. Patients with both chronic liver disease and elevated AFP may have a further increased risk of developing HCC; however, we found that AFP levels were consistent across liver imaging scores. 22 , 23 , 24 Monitoring AFP in populations with increased risk of HCC (such as the Fontan population) has increased in recent years, although its utility as a screening or diagnostic tool is not supported due to low sensitivity and specificity: some studies suggest only half of people with HCC will have elevated AFP at time of diagnosis. 25 , 26 Further, current literature is not robust in describing the change in AFP over time in populations both with and without known liver disease. 27 This study observed an increase in AFP over time although after removing outliers the values remained within normal range. Because HCC remains a rare occurrence in the Fontan population, AFP upper limit values have not been well defined within the Fontan population. 11 There was 1 patient with HCC, in whom the AFP values were well above normal. In another, the AFP value was elevated in the context of pregnancy as expected. However, the 2 other patients with HCC had AFP values within normal range (<8 ng/mL), with 1 high‐normal value of 6.6 ng/L at diagnosis. Although important to note that this is a small cohort, it appears that AFP may not be a reliable measure to aid in HCC diagnosis. A study by Toyoda et al. (2017) found that AFP had high prognostic value in early HCC; however, longer‐term prognosis was more accurately predicted by composite liver scores such as aspartate transaminase to platelet ratio index and fibrosis 4 score. 28 Similar to the general population, AFP should not be used in isolation in the assessment of FALD and HCC. Current guidelines recommend routine yearly liver biomarker assessment. We found that liver biomarkers and composite scores remained relatively stable over the study period. Similarly, there was no difference in biomarkers or composite scores between those with lower versus higher imaging scores. There was a trend to higher VAST score in those with a higher imaging score, and could be considered in assessing the presence of FALD. The utility of routine liver biomarker is likely limited and the cost of doing routine liver biomarkers not insignificant. Given this, it is important to identify other biomarkers that are associated with the presence of FALD. Other biomarkers such as hyaluronic acid, procollagen‐III‐peptide, and type IV collagen have been suggested as potential FALD markers. However, they are generally not collected as part of routine care and do not appear to offer consistent, accurate insight into the development of FALD. 25 , 29 Further study is required to identify relevant biomarkers as our current approach is clinically limited and not cost effective. The liver imaging score used was previously described in a Fontan population by Lindsay et al. (2015) and scores parenchymal appearance on a scale of 1 to 4 depending on the degree of nodularity or cirrhosis. 15 Each imaging modality (ultrasound, computed tomography, and MRI) has varying sensitivity and specificity to identify fibrotic changes in the Fontan population. 30 Given the retrospective nature of this study, the imaging modality available at each time point was used, the majority of which was ultrasound. It is possible that consistent use of higher fidelity imaging, such as MRI, may identify fibrotic changes earlier. However, even with the current approach we noted an increase in the proportion of individuals with a higher imaging score over time, in line with current literature indicating that time from Fontan procedure is correlated with worsening liver fibrosis. 31 Worsening fibrosis has also been linked to higher Fontan pressures, lower ventricular ejection fraction, and upregulated compensatory gene expression. 31 , 32 , 33 However, in the same studies, no significant correlation between laboratory biomarkers and Fontan outcomes was found; however, trends over time were not described. Given the complexity of the Fontan population, it is unlikely that a single imaging modality will prove superior in all cases. Certainly, the stability of routinely measured biomarkers in the face of increasing liver imaging scores further emphasizes the importance of regular liver imaging to assess the status of FALD. Event‐free survival declined steadily over the study period, with almost 40% of patients experiencing at least 1 event. Interestingly, age and diuretic use were the only 2 factors significantly associated with an increased risk for adverse events by multivariable analysis. It is the case that more complex, sicker individuals are more likely to be taking diuretics, and thus more likely to experience an adverse event. However, other common heart failure medications or a history of heart failure were not associated with increased events, suggesting a proportion of symptomatic edema requiring diuretic therapy may be secondary to Fontan circulatory dysfunction or FALD. Liver biomarkers and composite scores were associated with adverse outcomes in univariate modeling but not in the final multivariate model, similar to a previous retrospective study of right heart catheter data and outcomes. 34 As discussed, review of the current literature supports these findings of potential superiority in imaging compared with biomarker analysis. 31 , 32 This could be due to the sample size, but it also suggests that in addition to not being diagnostically relevant, our current approach to FALD assessment is not prognostically relevant. The limitations in this study included those associated with most single‐center retrospective chart reviews. Echocardiography parameters were somewhat inconsistently reported, with a mix of qualitative and quantitative values. As such, we were unable to accurately and clearly include echocardiography findings in our results or in some cases characterize heart failure diagnoses (ie, reduced versus preserved ejection fraction). Second, fibroscan and liver biopsy were rarely used in the study and not collected as part of our data set. Although these modalities are often seen as superior or even gold standard, during the study period they were not standard of practice at the study site and as such were not completed regularly. However, this speaks to the importance of this research because currently, definitive confirmation of cirrhosis would usually not change management in the Fontan population compared with suspicion on imaging. Lastly, specific medications related to abnormal liver function were not specifically collected. The most common of these would likely be short‐term medications such as antibiotics or acetaminophen; however, the impact of prolonged or high‐dose use would be important to assess for the risk of FALD development. CONCLUSIONS This single‐center study focuses on the current routine assessment of FALD and relationship with outcomes. It is apparent, that at least in the mid term, there is no significant change in routinely measured liver biomarkers and composite scores over time. This is despite increasing liver imaging scores suggestive of progressive FALD over the same time period. There is no one reliable method of assessing FALD; however, regular liver biomarker assessment is likely not clinically additive to imaging. Furthermore, we were not able to identify a FALD‐related marker associated with adverse clinical outcomes. This could be a reflection of study size or that we have yet to identify a reliable marker of FALD and adverse outcomes. Sources of Funding None. Disclosures None. Acknowledgments We are indebted to our patients and clinical support staff for supporting research in the Yasmin and Amir Virani Provincial Adult Congenital Heart Clinic. Dr Jasmine Grewal is a recipient of the Michael Smith Health Research Award. The abstract of this work was presented at the Canadian Cardiovascular Congress, Oct 24–27, 2024, in Vancouver, BC. This article was sent to John L. Jefferies, MD, MPH, Guest Editor, for review by expert referees, editorial decision, and final disposition. For Sources of Funding and Disclosures, see page 10. References