Liver stiffness measured by transient elastography is not predictive of mortality or adverse events in adult patients with Fontan associated liver disease Authors: Elettra Pomiato, Giovanni Tritto, Pauline Velasco, Catherine Head, Natali Chung, Alessandra Frigiola Published: 2026 ABSTRACT BackgroundFontan associated liver disease (FALD) is a known complication in adult patients with total cavo-pulmonary connection. The relationship between the severity of FALD and cardiovascular disease progression is yet to be clarified. We investigated the role of liver stiffness measured by transient elastography (TE) in predicting adverse events (AEs) in adult patients with Fontan circulation.Material and methodsWe retrospectively analysed TE data from our cohort of Fontan patients from September 2014 to December 2023. Baseline characteristics were recorded. The association of TE, liver function tests, and pro-collagen3 with AEs were analysed.Results167/207 patients (median age 20.0 years; 60% male) for a total of 613 TE from September 2014 to December 2023 (median follow up 56.0 months; 1.893 patient-year) were included in the study. TE values remained stable over time and did not show any correlation with baseline characteristics. TE values < 23.9 kPa could discriminate patients with advanced FALD (Se 50%; Sp 79%; NPV 97%). Forty AEs occurred (0.06 events per patient-year). Albumin (HR 0.91, p = 0.010) and APRI (HR 1.92, p = 0.034) were independent predictors of morbidity at multivariate analysis.ConclusionsIn this relatively young cohort, with low mortality and no HCC during follow-up, no significant association was found between TE and AEs. Nevertheless, TE appeared to be useful in discriminating advanced FALD. As no convincing risk factor related to FALD severity or progression has been yet identified, a multidisciplinary approach to risk assessment and surveillance is advised to reduce the impact of FALD and prevent progression to HCC. Introduction The Fontan operation represents the palliative strategy of choice in patients born with functionally univentricular hearts. First performed in 1968, this surgical technique allows the survival of 95% of babies at 5 years and 90% at 10 years from the Fontan completion at present time, with 85% of patients reaching adulthood [1]. However, the Fontan circulation does not restore physiologic haemodynamics and presents several drawbacks that negatively affect mid and long-term morbidity and mortality [2], [3], [4]. Amongst other drawbacks, the passive filling of the pulmonary arteries causes chronic systemic venous congestion and relatively low cardiac output. These conditions predispose patients to significant end-organ damage that contributes to premature death and diminished quality of life, with the liver being one of the affected organs [5]. Fontan associated liver disease (FALD) is a well-documented complication in patients who have undergone the eponymous palliative surgery. The first report was published in 1983, when Lemmer et al. described a case of severe liver fibrosis five years after Fontan surgery [6]. Since this initial observation, numerous clinical studies have described FALD, with reported incidence rates varying from 36% to 86%, largely depending on the diagnostic criteria and methods employed [7]. The true extent of FALD is likely underestimated, as many patients remain asymptomatic or subclinical for extended periods, while others develop liver complications, including hepatocellular carcinoma (HCC) even after a successful heart transplant. However, the relationship between the severity of FALD and the progression of cardiovascular disease remains unclear. Transient elastography (TE) measures the velocity of low frequency mechanical waves through the liver tissue and obtains a value for liver stiffness in kPa, which correlates with liver fibrosis. Although TE is a validated tool for non-invasive assessment of liver fibrosis in many forms of liver disease, studies in Fontan patients have shown poor correlation with fibrosis grade [8], [9], [10]. Nevertheless, longitudinal assessment of liver stiffness has been recommended by the European Association for the Study of the Liver (EASL-ERN) in a recent position paper, as a potentially useful tool to monitor patients and predict clinical outcomes [7], [11]. Shear -wave elastography (SWE) is an alternative tool that allows the assessment of liver stiffness during liver ultrasound. Although SWE is less used than TE the two methods show good correlation [12]. This study investigated the role of transient elastography (TE) and other non-invasive tests in predicting the severity of FALD and the occurrence of adverse events in adult patients with Fontan circulation. Methods Patient assessment and follow up Since September 2014, all adult patients with Fontan circulation followed up at our centre have been screened for FALD and subsequently followed up with regular non-invasive tests at variable intervals depending on disease severity. The results were reviewed monthly by a multidisciplinary dedicated team. In this longitudinal retrospective observational cohort study, we analysed collected data from first assessment and during subsequent follow up. Patients aged <16 or without any recorded TE value were excluded. Age, gender, systemic ventricle morphology, oxygen saturation and age at Fontan completion were recorded as baseline characteristics. TE (FibroScan®, EchoSens, Paris, France) was performed at baseline and subsequently at yearly follow up appointments, along with clinical examination, liver ultrasound and/or cross-sectional imaging to detect early signs of cardiac or liver decompensation. For the purpose of the study, we considered the TEs performed on joining the adult service and each subsequent follow up. If more than four serial TEs were undertaken, we analysed the first 3 and the last value. A liver ultrasound (US) with shear wave elastography (SWE) was carried out on the same day as TE whenever possible. Data on liver parenchyma (homogeneous versus coarse liver), liver nodules (presence, number and dimension), hepatic venous flow, spleen dimension, ascites, SWE were recorded. FALD was defined by US criteria as the presence of hepatomegaly, liver surface nodularity, parenchymal heterogeneity, hyperechoic lesions, splenomegaly or collaterals. Advanced FALD was defined according to EASL-ERN criteria definition as presence of oesophageal varices, portosystemic shunts, ascites or splenomegaly. Patients underwent contrast enhanced liver MRI scan when clinically recommended by the multidisciplinary team. A blood test panel to rule out other causes of liver disease was carried out at baseline (Hepatitis B and C, autoimmune panel, ferritin, transferrin saturation, alpha1 antitrypsin, caeruloplasmin). Serial routine blood tests were carried out at follow up appointments, MELD-XI score [12] [13] and AST to Platelet Ratio Index (APRI) were calculated [13]. The study received Institutional Ethics waiver as only anonymized and clinically driven data were used. Adverse events Adverse events (AEs) included: death, cardiac transplant or transplant referral, Fontan failure-related hospital admission (heart failure, supraventricular arrhythmias, bradyarrhythmia, lymphatic failure), need for surgical/interventional procedures, hepatocellular carcinoma (HCC). Mortality analysis included death and cardiac transplantation. Morbidity analysis included transplant referral, Fontan failure-related hospital admission, need for surgical/interventional procedures, hepatocellular carcinoma (HCC). Data collection within the National Health Service is performed without explicit consent for provision of healthcare, administrative, and clinical audit purposes (local or national) and is performed under the auspices of European Law—General Data Protection Regulation (GDPR) Act (May 2018): article 6(1)(e); 9(2)(h); special category 9(2)(i). These data were extracted retrospectively for assessment of healthcare quality, delivery, and outcomes, and therefore separate institutional review board approval was not required for this work. Data analysis Continuous variables were presented as median and interquartile range (IQR). Shapiro-Wilk normality test was used to test characteristics of distribution. Categorical variables are presented as total numbers and percentages. Differences in repeated measurements were evaluated by the Wilcoxon rank test. Differences between independent groups were evaluated by independent sample non-parametric tests. Differences in proportion were assessed by chi-squared test. Correlations among continuous variables were assessed by Pearson or Spearman coefficient as appropriate. Association with adverse events was investigated by univariate Cox analysis. Multivariate Cox analysis was performed if appropriate. A p value < 0.05 was considered statistically significant. Analyses were performed using Jamovi v 2.3.28. Results A total of 167/209 consecutive patients (median age 20.0; IQR 18.0; 25.5 years, 60% male) were included in the study. The remaining were excluded due to insufficient data. They underwent 613 TEs from September 2014 to December 2023 (median follow up 56.0; IQR 32.0; 74.5 months; 1.893 patient-year) as shown in Fig. 1.Fig. 1Distribution of TE over the observation period. Distribution of TE over the observation period. Baseline characteristics and TE data are shown in Table 1. There was no difference in TE values according to age, gender, BMI, NYHA Class, morphology of the systemic ventricle, oxygen saturation at rest nor the age at the Fontan completion.Table 1Characteristics of 167 Fontan patients.Table 1TOT (167 Pts)Age (years)20 (18; 25.5)Sex (male, %)100 (60)BMI23.5 (20.6; 26.4)SpO2 (%)94 (92; 96)NYHA ClassI136 (81)II21 (13)III8 (5)IV2 (1)Systemic ventricleDomLV80 (48)DomRV61 (36.5)BiVe25 (15)Undet1 (0.5)Age at Fontan4 (3; 7)TE first (kPa)18.0 (14.0; 22.8)TE Var time (last-first)−0.3 (−4.9; +5.0)TE Var max-min5.8 (3.2; 9.7)SpO2: oxygen saturation, DomLV: dominant left ventricle, DomRV: dominant right ventricle, BiVe: two adequately developed ventricles, Undet: single ventricle of undetermined morphology, TE: transient elastography, var time: variation in liver stiffness over time, var value: variation in liver stiffness absolute value (max value – min value). Continuous variables are presented as median (IQR), discrete variables as number (%). Table 1Characteristics of 167 Fontan patients.Table 1TOT (167 Pts)Age (years)20 (18; 25.5)Sex (male, %)100 (60)BMI23.5 (20.6; 26.4)SpO2 (%)94 (92; 96)NYHA ClassI136 (81)II21 (13)III8 (5)IV2 (1)Systemic ventricleDomLV80 (48)DomRV61 (36.5)BiVe25 (15)Undet1 (0.5)Age at Fontan4 (3; 7)TE first (kPa)18.0 (14.0; 22.8)TE Var time (last-first)−0.3 (−4.9; +5.0)TE Var max-min5.8 (3.2; 9.7)SpO2: oxygen saturation, DomLV: dominant left ventricle, DomRV: dominant right ventricle, BiVe: two adequately developed ventricles, Undet: single ventricle of undetermined morphology, TE: transient elastography, var time: variation in liver stiffness over time, var value: variation in liver stiffness absolute value (max value – min value). Continuous variables are presented as median (IQR), discrete variables as number (%). Characteristics of 167 Fontan patients. SpO2: oxygen saturation, DomLV: dominant left ventricle, DomRV: dominant right ventricle, BiVe: two adequately developed ventricles, Undet: single ventricle of undetermined morphology, TE: transient elastography, var time: variation in liver stiffness over time, var value: variation in liver stiffness absolute value (max value – min value). Continuous variables are presented as median (IQR), discrete variables as number (%). Transient elastography Median TE value in our population was 18.0 (IQR 14.0; 22.8) kPa on initial assessment and remained stable over time. Median intra-patient temporal TE values variation during follow up was −0.3 (IQR -4.9; +5.0) kPa, p = 0.893. Median intra-patient fluctuation in TE values was 5.8 (IQR 3.2; 9.7) kPa, p = 0.001, with no consistent trend over time. Abdominal ultrasound (US) An abdominal ultrasound with SWE was performed along with TE in 140 patients. In 77/140 cases, liver parenchyma was defined as coarse or inhomogeneous, in 22/140 patients one or multiple liver nodules were found. The median spleen diameter was 12.3 (IQR 10.9; 13.6) cm and ascites was detected in 7/140 patients. SWE showed a median value of liver stiffness of 9.9 kPa (IQR 7.9; 12.4). There was a weak but statistically significant positive correlation between TE and SWE (r 0.290, p = 0.004). SWE values were not predictive of mortality nor morbidity. Most patients (95/140; 67.9%) met the criteria for FALD. 8/140 (5.7%) patients met the criteria for advanced FALD. TE values were significantly higher in patients with advanced FALD compared to patients with no or non-advanced FALD, liver stiffness being 22 (IQR 19.7; 28.3) kPa in patients with advanced FALD and 18 (IQR 14; 22.1) kPa in patients with no or non-advanced FALD, p = 0.020, in Fig. 2. A cut-off score of 23.9 kPa demonstrated limited sensitivity (0.500) but good specificity (0.792), with a high negative predictive value (97%), in excluding the advanced FALD.Fig. 2TE values in patients with and without advanced FALD, liver stiffness in kPa. TE values in patients with and without advanced FALD, liver stiffness in kPa. Blood tests Haemoglobin, platelet count, renal and most liver function tests remained stable over time and within normal limits. Some values showed a minor but statistically significant increase over time. Bilirubin increased from 15 (IQR 11; 21) to 18 (IQR 13; 25) μmol/l, p 0.001; GGT from 50 (IQR 40; 85) to 77 (IQR 51; 115) IU/l, p 0.003; ALT from 26 (IQR 19; 34) to 30 (24; 39) IU/l, p < 0.001. No patient tested positive for HBV or HCV, 40/167 patients had previous HAV infection. Blood tests results performed at the time of the first and last TE are shown in Table 2. There was a positive, although weak, correlation between TE and P3NP (r 0.196, p = 0.037) and TE and APRI (r 0.391, p < 0.001) and a negative correlation between TE and platelet count (r −0.288, p = 0.001).Table 2Blood tests.Table 2First TELast TEp-valueHb (g/l)158 (146; 169)158 (147; 166)0.115PTL (x10^9/l)173 (133; 213)162 (132; 195)0.291Crea (μmol/l)76 (64; 87)78 (69; 88)0.381Na (mmol/l)139 (138; 141)140 (139; 141)0.476Bil (μmol/l)15 (11; 21)18 (13; 25)0.001ALT (U/l)26 (19; 34)30 (24; 39)< 0.001AST (U/l)31 (25; 33)29 (24; 33)0.862ALP (U/l)86 (71; 108)86 (70; 105)0.787GGT (U/l)50 (40; 85)77 (51; 115)0.003Alb (g/l)49 (46; 52)46 (44; 50)< 0.001APRI0.48 (0.37; 0.64)0.52 (0.39; 0.71)0.322MELD-XI33.3 (30.0; 36.2)34.7 (31.8; 36.5)0.017AFP (ng/l)2.1 (1.7; 3.3)2.3 (1.7; 3.3)0.080P3NP (μg/L)6.9 (5.4; 10.3)6.6 (5.3-8.0)0.192Hb: haemoglobin, PTL: platelet count, INR: International Normalised Ratio, Crea: creatinine, Na: sodium, Bil: total bilirubin, ALT: alanine transaminase, AST: aspartate aminotransferase, GGT: gamma-glutamyl transferase, ALP: alkaline phosphatase, Alb: albumin, APRI: AST-to-platelet ratio index, MELD-XI Model for End-stage Liver Disease excluding INR, AFP: alpha-fetoprotein, P3NP: pro-collagen 3. Continuous variables are presented as median (IQR). Table 2Blood tests.Table 2First TELast TEp-valueHb (g/l)158 (146; 169)158 (147; 166)0.115PTL (x10^9/l)173 (133; 213)162 (132; 195)0.291Crea (μmol/l)76 (64; 87)78 (69; 88)0.381Na (mmol/l)139 (138; 141)140 (139; 141)0.476Bil (μmol/l)15 (11; 21)18 (13; 25)0.001ALT (U/l)26 (19; 34)30 (24; 39)< 0.001AST (U/l)31 (25; 33)29 (24; 33)0.862ALP (U/l)86 (71; 108)86 (70; 105)0.787GGT (U/l)50 (40; 85)77 (51; 115)0.003Alb (g/l)49 (46; 52)46 (44; 50)< 0.001APRI0.48 (0.37; 0.64)0.52 (0.39; 0.71)0.322MELD-XI33.3 (30.0; 36.2)34.7 (31.8; 36.5)0.017AFP (ng/l)2.1 (1.7; 3.3)2.3 (1.7; 3.3)0.080P3NP (μg/L)6.9 (5.4; 10.3)6.6 (5.3-8.0)0.192Hb: haemoglobin, PTL: platelet count, INR: International Normalised Ratio, Crea: creatinine, Na: sodium, Bil: total bilirubin, ALT: alanine transaminase, AST: aspartate aminotransferase, GGT: gamma-glutamyl transferase, ALP: alkaline phosphatase, Alb: albumin, APRI: AST-to-platelet ratio index, MELD-XI Model for End-stage Liver Disease excluding INR, AFP: alpha-fetoprotein, P3NP: pro-collagen 3. Continuous variables are presented as median (IQR). Blood tests. Hb: haemoglobin, PTL: platelet count, INR: International Normalised Ratio, Crea: creatinine, Na: sodium, Bil: total bilirubin, ALT: alanine transaminase, AST: aspartate aminotransferase, GGT: gamma-glutamyl transferase, ALP: alkaline phosphatase, Alb: albumin, APRI: AST-to-platelet ratio index, MELD-XI Model for End-stage Liver Disease excluding INR, AFP: alpha-fetoprotein, P3NP: pro-collagen 3. Continuous variables are presented as median (IQR). Liver MRI scan Liver MRI scan was performed in 33/140 patients to further investigate suspicious liver nodules. In 17 cases at least one liver nodule was characterised as focal nodular hyperplasia (FNH). Other focal lesions detected included one hepatic cyst, one haematoma associated with sclerosed haemangioma, one arteriovenous shunt, one unspecified benign lesion. Compared to liver MRI scan, liver US had low sensitivity (47%) and reasonable specificity (75%) in detecting liver nodules, with an overall accuracy of 61%. Adverse events (AEs) Forty AEs occurred during the follow up period (23.7%, 0.06 events per patient-year). No cases of HCC were observed. Five patients died, 5 were referred for heart transplantation, 1 needed vv-ECMO support, 25 patients needed hospital admission (17/25 for supraventricular arrhythmias, 3/25 for signs of heart failure, 3/25 for bradyarrhythmia's, 1/25 with relapse of PLE, 1/25 for haemoptysis), 3 patients needed interventional procedures (namely coiling of collaterals), 1 was admitted for elective aortic valve replacement and personalized external aortic root support (PEARS). Mortality Five patients died during the follow up and were included in the mortality analysis. Neither TE nor its variation over time were predictive of death at univariate Cox analysis. At the same analysis, APRI (HR 5.23; CI95 1.36; 20.11, p = 0.016), AST (HR 1.06, CI95 1.01; 1.11; p = 0.014), creatinine (HR 1.03, CI95 1.01-1.05, p = 0.012), SpO2 (HR 0.81, CI95 0.80; 1.00, p = 0.050) and albumin (0.90, CI95 0.82; 0.99, p = 0.029) were predictors of death (see Table 3). Survival plots are displayed in Fig. 3, Fig. 4, Fig. 5, Fig. 6. Due to the small number of major events, multivariate analysis was not performed.Table 3Mortality Cox survival analysis (univariate).Table 3HRP-valueCI95TE first (kPa)1.060.2280.97; 1.16TE min (kPa)1.140.0541.00; 1.30TE max (kPa)1.050.3360.95; 1.17TE Var value (max-min, kPa)0.920.4710.72; 1.16TE Var time (last-first, kPa)1.000.9550.89; 1.14TE SWE, (kPa)1.000.9870.82; 1.22Age at Fontan (yrs)0.900.5400.63; 1.28SpO2 (%)0.810.0500.66; 1.00Albumin (g/l)0.890.0010.82; 0.95AST (U/l)1.060.0141.01; 1.11APRI5.230.0161.36; 20.11Crea (ug/l)1.030.0121.01;1.05TE transient elastography, with fibroscan. First: oldest value; Min minimum value; Max highest value; var value (highest – lowest TE value; var time: recent – oldest TE value. TE SWE transient elastography with shear wave elastography, SpO2: oxygen saturation, AST: aspartate aminotransferase, APRI: AST-to-platelet ratio index, crea: serum creatinine.Fig. 3Survival according to APRI at univariate Cox regression analysis. Time in months.Fig. 4Survival according to Albumin level at univariate Cox regression analysis. Time in months.Fig. 5Survival according to oxygen saturation at univariate Cox regression analysis. Time in months.Fig. 6Survival according to creatinine level at univariate Cox regression analysis. Time in months. Table 3Mortality Cox survival analysis (univariate).Table 3HRP-valueCI95TE first (kPa)1.060.2280.97; 1.16TE min (kPa)1.140.0541.00; 1.30TE max (kPa)1.050.3360.95; 1.17TE Var value (max-min, kPa)0.920.4710.72; 1.16TE Var time (last-first, kPa)1.000.9550.89; 1.14TE SWE, (kPa)1.000.9870.82; 1.22Age at Fontan (yrs)0.900.5400.63; 1.28SpO2 (%)0.810.0500.66; 1.00Albumin (g/l)0.890.0010.82; 0.95AST (U/l)1.060.0141.01; 1.11APRI5.230.0161.36; 20.11Crea (ug/l)1.030.0121.01;1.05TE transient elastography, with fibroscan. First: oldest value; Min minimum value; Max highest value; var value (highest – lowest TE value; var time: recent – oldest TE value. TE SWE transient elastography with shear wave elastography, SpO2: oxygen saturation, AST: aspartate aminotransferase, APRI: AST-to-platelet ratio index, crea: serum creatinine. Mortality Cox survival analysis (univariate). TE transient elastography, with fibroscan. First: oldest value; Min minimum value; Max highest value; var value (highest – lowest TE value; var time: recent – oldest TE value. TE SWE transient elastography with shear wave elastography, SpO2: oxygen saturation, AST: aspartate aminotransferase, APRI: AST-to-platelet ratio index, crea: serum creatinine. Survival according to APRI at univariate Cox regression analysis. Time in months. Survival according to Albumin level at univariate Cox regression analysis. Time in months. Survival according to oxygen saturation at univariate Cox regression analysis. Time in months. Survival according to creatinine level at univariate Cox regression analysis. Time in months. Morbidity Morbidity analysis included all other listed adverse events. Neither TE nor its variation over time were predictive of morbidity at univariate Cox analysis. At the same analysis, APRI (HR 2.33; CI95 1.28; 4.24, p = 0.006), AST (HR 1.03, CI95 1.01; 1.05; p = 0.005), SpO2 (HR 0.87, CI95 0.80; 0.95, p = 0.002) and albumin (0.92, CI95 0.88; 0.96 p < 0.001) were predictors of minor adverse events (see Table 4). At multivariate regression analysis, albumin (HR 0.91; CI95 0.85; 0.98, p = 0.010) and APRI (HR 1.92; CI95 1.05; 3.50, p = 0.034) remained significant predictors of events (Fig. 7).Table 4Morbidity Cox survival analysis (univariate).Table 4HRP-valueCI95TE first (kPa)1.020.3920.98; 1.06TE min (kPa)1.050.0701.00; 1.11TE max (kPa)1.040.0601.00; 1.08TE Var value (max-min, kPa)1.030.4360.96; 1.10TE Var time (last-first, kPa)1.040.1110.99; 1.09TE SWE (kPa)0.860.0590.73; 1.01Age at Fontan (years)1.000.9900.94; 1.06SpO2 (%)0.870.0020.80; 0.95Albumin (g/l)0.92<0.0010.88; 0.96AST (U/l)1.030.0051.01; 1.05APRI2.330.0061.28; 4.24Crea (ug/l)0.980.1070.96; 1.00TE transient elastography, with fibroscan. First: oldest value; Min minimum value; Max highest value; var value (highest – lowest TE value; var time: recent – oldest TE value. TE SWE transient elastography with shear wave elastography, SpO2: oxygen saturation, AST: aspartate aminotransferase, APRI: AST-to-platelet ratio index, Crea: serume creatinine.Fig. 7Multivariable Cox proportional hazards regression analysis for morbidity. Table 4Morbidity Cox survival analysis (univariate).Table 4HRP-valueCI95TE first (kPa)1.020.3920.98; 1.06TE min (kPa)1.050.0701.00; 1.11TE max (kPa)1.040.0601.00; 1.08TE Var value (max-min, kPa)1.030.4360.96; 1.10TE Var time (last-first, kPa)1.040.1110.99; 1.09TE SWE (kPa)0.860.0590.73; 1.01Age at Fontan (years)1.000.9900.94; 1.06SpO2 (%)0.870.0020.80; 0.95Albumin (g/l)0.92<0.0010.88; 0.96AST (U/l)1.030.0051.01; 1.05APRI2.330.0061.28; 4.24Crea (ug/l)0.980.1070.96; 1.00TE transient elastography, with fibroscan. First: oldest value; Min minimum value; Max highest value; var value (highest – lowest TE value; var time: recent – oldest TE value. TE SWE transient elastography with shear wave elastography, SpO2: oxygen saturation, AST: aspartate aminotransferase, APRI: AST-to-platelet ratio index, Crea: serume creatinine. Morbidity Cox survival analysis (univariate). TE transient elastography, with fibroscan. First: oldest value; Min minimum value; Max highest value; var value (highest – lowest TE value; var time: recent – oldest TE value. TE SWE transient elastography with shear wave elastography, SpO2: oxygen saturation, AST: aspartate aminotransferase, APRI: AST-to-platelet ratio index, Crea: serume creatinine. Multivariable Cox proportional hazards regression analysis for morbidity. Discussion In the present study, we retrospectively investigated the role of TE, liver US with SWE and other clinical data in the assessment of FALD in adult patients with Fontan circulation and their association with AEs. The prevalence of FALD in patients with Fontan circulation remains uncertain and likely under-recognized [14], [15]. Autopsy studies on patients who died shortly after the surgical procedure indicate that liver damage may start even before Fontan circulation, as a consequence of cumulative hemodynamic insults from birth [5], [16]. Several techniques have been used to diagnose FALD, namely ultrasound, computed tomography, and magnetic resonance imaging. Although previous studies have tried to implement standard of practice for monitoring development and progression of FALD, widely agreed screening and surveillance protocols and comprehensive studies are lacking [17]. In agreement with previous studies, liver stiffness was elevated in our adult patients with Fontan circulation compared to general population. Wilson and colleagues [18], reported liver stiffness assessed by TE ≥ 15 kPa in 75/133 (56%), and ≥20 kPa in 41/133 (31%) patients with FALD. Similarly, in our cohort, liver stiffness was 18 kPa, a value that would be suggestive of advanced liver disease in non-congestive aetiologies [19]. Elective liver biopsy findings revealed that all patients with Fontan circulation exhibit some degree of liver fibrosis. In this population, chronic liver disease is likely multifactorial and related to the altered vascular dynamics: chronic systemic venous congestion leads to elevated hepatic venous pressures, liver congestion and subsequent fibrosis [20]; low cardiac output and cyanosis cause reduced oxygen delivery, which contributes to visceral and hepatic hypoxia and ischemic injury [21], [22]; the multifactorial prothrombotic state observed in Fontan patients increases the risk of hepatic sinusoidal thrombosis and intrahepatic micro-thromboses with subsequent liver damage [23]; elevated hepatic lymphangiogenesis associated with impaired drainage can lead to hepatic lymphatic congestion and fibrosis and ultimately to an increased intestinal permeability [24]; subtle chronic inflammation hypothesised in Fontan patients may induce fibrogenesis in kidneys, liver, and myocardium [25], while other risk factors such as chronic viral hepatitis or hepatotoxic medication like amiodarone, can contribute to the development and progression of FALD. FALD was highly prevalent in our population (≈68%) with 7.4% of patients meeting the EASL criteria of advanced FALD [7]. Although the prevalence of FALD can vary according to the diagnostic criteria, our findings are similar to previous studies. In a recent paper by Jarasvaraparn et al. [26], the prevalence of FALD was 55% in adult and 68% in children after Fontan completion. The role of TE in assessing FALD has been widely debated. In our study TE correlated with biomarker of liver fibrosis such as APRI (p < 0.001) and P3NP (p 0.037), and a cut-off score of 23.9 kPa at TE had a specificity of 79% and a negative predictive value a 97% when discriminating advanced FALD. In another study TE at a cut-off 18.5kPA had a sensitivity of 81.3%, but low specificity at 45% in detecting advanced liver disease in Fontan patients compared to liver biopsy [26], while Bütikofer and colleagues did not find significant differences in TE values and the degree of liver fibrosis, although higher TE values were associated with Fontan failure and advanced centrilobular hepatic fibrosis [27]. Notably, in the subgroup of patients requiring a liver MRI for further characterisation, most patients were diagnosed with focal nodular hyperplasia (FNH). However, in two cases precancerous lesions were diagnosed. Importantly, when compared to liver MRI, liver US had low sensitivity (47%) yet reasonable specificity (75%) in detecting liver nodules, with an accuracy of 61%. A study by Wang et al. investigating ultrasound and MRI agreement in a cohort of Fontan patients of similar age found a lower sensitivity (34%) but higher specificity (100%) of liver ultrasound in detecting lesion of any size compared to liver MRI [28]. Although some studies have investigated the impact of FALD [29], [30], a standardised diagnostic protocol for FALD is still lacking. To overcome liver US limitations, other diagnostic modalities, such as liver MRI might be considered at the time of transition to adult services and the findings used to tailor subsequent follow up. Adverse events were frequent in our population (23.7%, 0.06 events per patient-year). In adult Fontan patients, the incidence of hepatocellular carcinoma (HCC) is elevated compared to general population and varies from 3% to 15% [31]. Interestingly, although two cases of HCC were diagnosed prior to starting a regular surveillance protocol in our Fontan patients, no cases of HCC were detected during the observational period, which may have negatively affected the prognostic value of TE. TE values were not associated with mortality nor morbidity in our cohort. In contrast, low albumin and APRI emerged as the sole independent predictor of morbidity on multivariate Cox regression analysis, and were associated with mortality on univariate analysis. Data on albumin as a risk factor are scarce, however albumin levels were positively associated with exercise capacity at univariate analysis [32]. Interestingly, among the 11 patients with Albumin <40 g/l, only 5 had a clinically evident PLE. Considering that liver function indices were within the normal range in our cohort, these findings may suggest a potential role of inflammation in the pathogenesis of hypoalbuminemia, suggesting even more complex mechanisms underlying the Fontan failure. However, further studies are required to confirm this hypothesis. Overall, our findings suggest that the prognostic assessment in FALD cannot rely on a single parameter but must instead be multifactorial, including clinical, biochemical and morphological data, along with the assessment of liver stiffness. Study limitations Our study has some limitations. Although the follow up was longitudinal, data on the first TE and blood tests were collected retrospectively and this could have led to information bias, however as the electronic chart has been available in our institution since before 2013, we believe the loss of data has been minimal. With regard to adverse events (AEs), the limited number of deaths and the absence of FALD-related events (including hepatocellular carcinoma) suggest that the study may be underpowered, and therefore some associations may not have reached statistical significance. In addition, the low number of deaths precluded multivariable modelling, allowing only univariate analyses to be performed for mortality. Conclusion TE values are elevated in adult patients with Fontan circulation, and, despite frequent fluctuations, they remain overall stable in the mid-term follow-up (≈5 years). Although TE is an easy and non-invasive technique that can be used in outpatient setting to screen patients with advanced FALD, in this relatively young cohort, with low mortality and no HCC during follow-up, no significant association was found between TE and adverse clinical events. Therefore, the follow up of FALD should encompass both biochemical and morpho-functional liver assessment. Moreover, as the sensitivity and accuracy of liver ultrasound in detecting liver nodules is suboptimal, a baseline liver MRI should be considered at the time of transition to adult services and the findings discussed with a liver specialist to tailor subsequent hepatic follow up. A multidisciplinary approach to risk assessment and surveillance is advised to optimise the management of FALD and prevent its complications. Funding This research received no external funding. CRediT authorship contribution statement Elettra Pomiato: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing. Giovanni Tritto: Conceptualization, Formal analysis, Methodology, Supervision, Writing – review & editing. Pauline Velasco: Data curation, Investigation. Catherine Head: Methodology, Supervision, Writing – review & editing. Natali Chung: Methodology, Supervision, Writing – review & editing. Alessandra Frigiola: Conceptualization, Methodology, Supervision, Writing – review & editing. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.