Treatment With Direct-acting Antivirals for HCV Decreases But Does Not Eliminate the Risk of Hepatocellular Carcinoma

Federico Piñero; Manuel Mendizabal; Ezequiel Ridruejo; Fernando Herz Wolff; Beatriz Ameigeiras; Margarita Anders; María Isabel Schinoni; Virginia Reggiardo; Ana Palazzo; María Videla; Cristina Alonso; Luisa Santos; Adriana Varón; Sebastián Figueroa; Cecilia Vistarini; Raúl Adrover; Nora Fernández; Daniela Perez; Federico Tanno; Nelia Hernández; Marcela Sixto; Silvia Borzi; Andres Bruno; Daniel Cocozzella; Alejandro Soza; Valeria Descalzi; Claudio Estepo; Alina Zerega; Alexandre de Araujo; Hugo Cheinquer; Marcelo Silva; LALREAN


Liver International. 2019;39(6):1033-1043. 

In This Article

Materials and Methods

This was a prospective cohort study performed between 1 May 2016 and 1 June 2018 in different regional Liver Units from Argentina, Uruguay, Chile, Colombia and Brazil. Sites were instructed to enrol all eligible patients on a prospectively consecutive basis. Study data were registered into a web-based electronic system. Central revision and resubmission was requested in cases of missing data.

Consecutive adult patients (>18 years of age) with chronic HCV infection were included in the study. HCV infection was confirmed with quantitative real chain polymerase reaction and genotyping/subgenotyping was carried out. Subjects were included after they signed an informed consent. Patients with any degree of liver fibrosis treated with all-oral DAA were eligible. All patients who received at least one pill of DAAs were included in the study as part of an intention to treat analysis. In addition, we included a small group of patients from a registry supported by LALREAN who received DAA therapies as part of different compassionate use programs during 2015. Patients with prior solid organ transplantation or previous diagnosis of HCC were excluded. All patients were under strict HCC surveillance in order to exclude HCC diagnosis prior to DAA initiation, according to international guidelines.[16,17]

Baseline variables recorded in included subjects were demographic data, relevant past medical history and co-infection with chronic hepatitis B virus (HBV) or human immunodeficiency virus (HIV). Evaluation of liver disease stage included: liver fibrosis stage assessed by liver biopsy or non-invasive methods including transient elastography or serum biomarkers (Fibrotest®) or by clinical diagnosis of cirrhosis (the presence of gastro-esophageal varices or ascites). Liver stiffness measurement (LSM) recorded as a continuous variable in kPa, was performed using liver elastography[18] and were categorized according to international recommendations in all patients;[19] kPa values more than 9.5 and more than 12.5 were categorized as F3 and F4 respectively.[18,19] Severity of liver disease was evaluated according to Child-Turcotte Pugh (CTP) classification. Clinically significant portal hypertension (CSPH) was defined as either the presence of ascites or portosystemic encephalopathy or gastro-esophageal varices on endoscopy.

Evaluation prior to DAA treatment included: year of HCV diagnosis, infection route and prior treatment with IFN-based therapies. The particular DAA regimen used was based on physicians' criteria according to available regimens in each country following international guidelines.[5] Laboratory values prior and following DAA therapy including HCV viral loads (IU/mL) and other laboratory values were registered. We assessed response to DAA therapy documenting SVR, defined as an undetectable HCV RNA (HCV viral load <15 UI/mL) 12 weeks after completion of therapy. Treatment failure was defined as the presence of detectable HCV RNA after discontinuation of antiviral regimen. Laboratory operators were blinded from baseline patient characteristics, treatment regimens and main events or outcomes.

Main Event: Definition and Diagnosis of Hepatocellular Carcinoma

The primary end-point analysed was cumulative incidence of de novo HCC. All patients with fibrosis stage 3 or 4 underwent HCC surveillance with abdominal ultrasound (US) with or without serum alpha-foetoprotein (AFP) before and every 6 months during and after DAAs, as recommended by international guidelines.[16,17] In order to avoid potential selection bias, all patients had at least two consecutive previous US with a negative test for HCC screening. Although surveillance for HCC is not recommended in patients with F1–2 fibrosis grades, an abdominal US every 6 months was recommended in the study protocol in this group of patients. Operators who performed abdominal ultrasounds at each centre were blinded from treatment regimens and achievement of SVR. Time of follow-up was recorded since DAA initiation until death or end of study. Patients were followed at least every 4 weeks during and until 12 weeks after treatment and every 6 months thereafter.

If during the semi-annual screening with an abdominal US a liver nodule ≥10 mm was identified, an abdominal triphasic-contrast image was performed either with computed axial tomography scan (CT) or magnetic resonance imaging (MRI). HCC diagnosis was defined as arterial enhancement (not rim-like lesion) and venous washout with or without capsule appearance on triphasic CT or MRI.[16,17] Tumour burden at diagnosis was classified according to Barcelona Clinic Liver Cancer criteria (BCLC).[20]

Secondary end-points analysed were patients' survival and adverse events related to DAAs and were classified according to common toxicity criteria.[21] Causes of death were recorded and categorized as complications of portal hypertension, cancer-related (HCC related), sepsis, cardiovascular events and other causes.

Procedures were conducted in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.[22] The Austral University Ethics Committee and each Ethical Committee (CIE) from all the participating centres approved the study protocol. All procedures followed ethical standards (institutional and national) as well as the Helsinki Declaration of 1975, as revised in 2008. The study protocol was recorded and registered in the CIE 16–023, had a public access database for all the co-investigators on a web link ( and received a Research National Grant from Argentina from the National Institute of Cancer (INC).

Statistical Analysis

A total sample size to enrol a minimum of 420 patients under DAAs was calculated to include a cohort with an annual risk of HCC 1%-6% as previously mentioned (α type I error of 5%, type II error or β of 0.20). Categorical data were compared using Fisher's exact test (2-tailed) or Chi-square (χ 2) test as appropriate. Continuous variables are shown with mean (±standard deviation, SD) or median (Interquartile ranges 25%-75%, IQR) and were compared with Student's T or Mann-Whitney U tests according to their distributions.

Kaplan-Meier survival and cumulative incidence curves were compared using the log-rank test. Among patients with or without SVR, the absolute and relative risks for de novo HCC and the number needed to treat (NNT) to prevent one case of HCC were calculated. A multivariable Cox regression analysis with hazard ratios (HR) and 95% confidence intervals (CI) was performed to evaluate independent variables associated with HCC development. Proportional hazards assumption was assessed through graphic and statistical evaluation (Schoenfeld residuals test) and calibration of the model by comparison of observed and predicted curves and by Harrell's c-statistic index. To further re-adjust the effect of SVR on the risk for de novo HCC, a competing risk regression analysis was carried out (subhazard ratios, SHR) considering death before HCC occurrence as a competing event. Finally, since the exposure of interest was an intervention not randomly assigned, we also used a propensity score matching from a logistic regression analysis, to reduce confounding bias. Therefore, we estimated the probability of achieving SVR as a function of different variables and used that score as a single matching covariate.[23] Collected data were analysed using STATA 13.0.