Hepatocellular carcinoma in survivors after Fontan operation: a case–control study

Abstract

Liver injury is universal in patients who have undergone Fontan palliation for complex congenital heart disease and results in fibrosis and cirrhosis.1 Fontan-associated liver disease (FALD) can be rarely complicated by hepatocellular carcinoma (HCC),2–4 risk factors for which are not fully elucidated. We aim to identify clinical characteristics associated with the development of HCC in adult single ventricle patients with Fontan circulation. This is a multi-centre, retrospective case–control study of Fontan patients ≥ 18 years old across 18 North American centres of the Alliance for Adult Research in Congenital Cardiology between 2005 and 2021 and approved by all research oversight boards. Cases were included based on histologic or imaging-based diagnosis of HCC [computed tomography (CT) or magnetic resonance imaging (MRI) that fulfil Liver Imaging Reporting and Data System® (LI-RADS)-5 criteria,5 plus alpha-fetoprotein (AFP) > 20 ng/mL or evidence of lesion growth ≥ 50% within 6 months or > 100% in a period of over 6 months on serial imaging studies]. Controls were randomly selected in a 3:1 ratio from the centre in which the case was derived. Inclusion criteria were blood work and abdominal imaging (ultrasound, CT, or MRI) within 18 months of the last clinic visit, the latter without a focal liver lesion or if present, imaging evidence of stable size over at least 24 months. Those without imaging or labs, AFP ≥ 10 ng/dL at any time, or history of hepatitis B or C were excluded. Clinical characteristics and laboratory values were collected by chart review. AST to Platelet Ratio Index (APRI) and Fibrosis-4 Index (FIB-4) (non-invasive methods that correlate with degree of fibrosis on biopsy in other chronic liver disease states)6 and Model for End-stage Liver Disease excluding INR (MELD-XI) scores were calculated and collapsed into binary (low/high) variables using cut-offs of < 0.5 and ≥ 0.5, < 1.45 and ≥ 1.45, and < 11 and ≥ 11, respectively. Radiology and liver biopsy reports were reviewed by a board-certified hepatologist. Bivariate and multivariate analyses were performed to examine association between variables and HCC. We included characteristics having a significant association with HCC risk (P <0.05) in the multivariate logistic regression model after evaluating collinearity and area under the curve (AUC) of the receiver operating characteristic (ROC). There were 58 HCC cases (43% female and 79% White) and 174 controls (n = 232). Out of a total of 3251 adult Fontan patients followed across all 18 centres, the estimated prevalence of HCC was 1.8%. Median age of HCC diagnosis was 31 (interquartile range 26–38) years. One-year survival after HCC diagnosis was 81%. There was no difference between cases and controls regarding age, sex, body mass index, underlying congenital heart defect, age at Fontan operation, or time since Fontan; those with HCC more commonly had undergone right atrial to pulmonary artery or right atrial to right ventricle Fontan repair, had higher burden of cardiac comorbidities, including arrhythmia, desaturation (oxygen saturation < 90%), heart failure, and lymphatic disorders, and were more likely to have undergone prior Fontan revision and valve surgery, despite comparable echocardiographic and historical catheterization haemodynamic parameters. Just over half of HCC cases occurred in patients with FIB-4 < 1.45, a threshold which has a 90% negative predictive value for advanced fibrosis in other liver disease populations.7 Similarly, nearly half of HCC cases occurred in patients with MELD-XI < 11 and with APRI < 0.5. Bivariate analyses examining the relationship between select variables and HCC were performed (Table 1). There was no relationship between age at Fontan or time since Fontan operation with HCC. Bivariate analyses for the odds of hepatocellular carcinoma P-values <0.05 are noted in bold. APRI, AST to Platelet Ratio Index; AST, aspartate transaminase; CI, confidence interval; FIB-4, Fibrosis-4; MELD-XI, Model for End-stage Liver Disease eXcluding International normalized ratio; OR, odds ratio. We created two multivariable models using select variables from the bivariate analyses. Model A showed desaturation [adjusted odds ratio (aOR) 2.4, 95% confidence interval (CI) 1.1–5.2; P = .02], history of Fontan revision (aOR 2.3, 95% CI 1.1–4.8; P = .02), and thrombocytopenia (aOR 2.3, 95% CI 1.2–4.5; P = .02) were associated with increased odds of HCC. Model B showed desaturation (aOR 2.2, 95% CI 1.0–4.7; P = .05) and FIB-4 ≥ 1.45 (aOR 3.9, 95% CI 2.0–8.0; P < .001) as associated with higher odds of developing HCC. Both models had ROC AUC = 0.72. We hypothesized that HCC is time-dependent and that those affected would be older than controls, but the data do not suggest that length of exposure to Fontan physiology is the main driver of HCC. Instead, we found that advanced FALD and cardiac comorbidities were more prevalent in HCC cases. Progression of liver fibrosis in FALD is not uniform8 and is likely driven by circulatory factors, which hypothetically could explain why length of exposure to Fontan physiology alone was not found to be significant. It is important to emphasize that non-invasive fibrosis scores, such as APRI and FIB-4, do not have adequate discrimination for the detection of HCC and should be interpreted in the context of other clinical factors. Our results support the concept that worse FALD (by liver function tests and imaging) is the primary substrate for the development of HCC in the Fontan liver and suggest that ‘sicker’ Fontan patients—those with desaturation and need for cardiac surgery beyond the Fontan operation itself—may be at highest risk. The numbers studied were relatively small but represent the largest collection of HCC in the Fontan population published to date. This was a study on Fontan patients of adult age and therefore results may not be applicable to the paediatric population. We also acknowledge that the imaging-based inclusion criteria as described are not standard, as LI-RADS is not applicable in ‘cardiac congestion’. By additionally requiring either elevated AFP or evidence of rapid lesion growth, we attempted to improve the diagnostic accuracy of LI-RADS in the Fontan liver. Hepatocellular carcinoma is diagnosed in up to 2% of adult Fontan patients. Those with more advanced FALD represented by abnormal liver biochemistries and signs of circulatory failure may be at increased risk. Until more clearly defined risk factors are identified, we recommend abdominal ultrasounds and AFP every 6 months for adult Fontan patients in accordance with guidelines for HCC screening.9 We would like to acknowledge the Alliance for Adult Research in Congenital Cardiology for its support rendered to this study. All authors declare no disclosure of interest for this contribution. No data were generated or analysed for or in support of this paper. This study was supported by a grant from the Children’s Hospital of Philadelphia Cardiac Center. Ethical approval was not required. None supplied.

Topics & Concepts