RESUMO
BACKGROUND & AIMS: Fontan-type surgery is used as a palliation for congenital heart disease with univentricular physiology but may, in the long term, lead to advanced chronic liver disease. This study assessed the accuracy of conventional non-invasive models in assessing liver fibrosis and introduces a new risk score employing non-invasive tools. METHODS: A prospective, cross-sectional, observational study was conducted across five European centers and encompassing all consecutive adult patients with Fontan circulation, liver biopsy and non-invasive tests (elastography, APRI, FIB-4, Fibrosis score, Doha, GUCI, and AAR). The primary outcome was the identification of severe liver fibrosis on biopsy. Multivariable logistic regression identified non-invasive predictors of severe fibrosis, leading to the development and internal validation of a new scoring model named the FonLiver risk score. RESULTS: In total, 217 patients (mean [standard deviation] age, 27.9 [8.9] years; 50.7% males) were included. Severe liver fibrosis was present in 47.9% (95% CI 41.2%-54.5%) and correlated with a lower functional class, protein-losing enteropathy, and compromised cardiopulmonary and systemic hemodynamics. The final FonLiver risk score incorporated liver stiffness measurement using transient elastography and platelet count and demonstrated strong discrimination and calibration (area under the receiver operating curve [AUROC] of 0.81). The FonLiver risk score outperformed conventional prediction models (APRI, FIB-4, Fibrosis score, Doha, GUCI, and AAR), which all exhibited worse performance in our cohort (AUROC < 0.70 for all). CONCLUSION: Severe liver fibrosis is prevalent in adults following Fontan-type palliation and can be effectively estimated using with the novel FonLiver risk score. This scoring system can be easily incorporated into the routine assessment of patients with Fontan circulation.
RESUMO
BACKGROUND & AIMS: Fontan surgery is used to treat a variety of congenital heart malformations, and may lead to advanced chronic liver disease in the long-term. This study examines the prevalence, characteristics and predictors of liver nodules in patients following Fontan surgery. METHODS: This was a prospective, cross-sectional, observational study conducted at 8 European centres. Consecutive patients who had undergone Fontan surgery underwent blood tests, abdominal ultrasonography (US), transient elastography (Fibroscan®), echocardiography, haemodynamic assessments, and abdominal MRI/CT scan. The primary outcome measure was liver nodules detected in the MRI/CT scan. Predictors of liver nodules were identified by multivariate logistic regression. RESULTS: One hundred and fifty-two patients were enrolled (mean age 27.3 years). The mean time elapsed from surgery to inclusion was 18.3 years. Liver nodule prevalences were 29.6% (95% CI 23-37%) on US and 47.7% (95% CI 39-56%) on MRI/CT. Nodules were usually hyperechoic (76.5%), round-shaped (>80%), hyperenhancing in the arterial phase (92%) and located in the liver periphery (75%). The sensitivity and specificity of US were 50% (95% CI 38-62%) and 85.3% (95% CI 75-92%), respectively. Inter-imaging test agreement was low (adjusted kappa: 0.34). In the multivariate analysis, time since surgery >10 years was the single independent predictor of liver nodules (odds ratio 4.18; p = 0.040). Hepatocellular carcinoma was histologically diagnosed in 2 of the 8 patients with hypervascular liver nodules displaying washout. CONCLUSION: While liver nodules are frequent in Fontan patients, they may go unnoticed in US. Liver nodules are usually hyperechoic, hypervascular and predominantly peripheral. This population is at risk of hepatocellular carcinoma, the diagnosis of which requires confirmatory biopsy. LAY SUMMARY: Fontan surgery is the standard of care for many patients with univentricular congenital cardiopathies. Recent advances have improved the survival of Fontan patients, and nowadays most of them reach adulthood. In this setting, Fontan-associated liver disease (FALD) is increasingly recognised, and has become a significant prognostic factor. Liver nodules are considered a component of FALD yet their prevalence, imaging features and predictors have hardly been evaluated. In this study, we observed that liver nodules are frequent, typically hyperechoic, hypervascular and predominantly peripheral in patients with FALD. This population is at risk of hepatocellular carcinoma, the diagnosis of which must be confirmed by biopsy.