Background: North American data are lacking on the effect of nucleos(t)ide analogues (NA) in preventing chronic hepatitis B (CHB)-related hepatocellular carcinoma (HCC).
Aim: To determine the incidence of HCC in NA-treated patients and compare this risk with that predicted without treatment based on the REACH-B model.
Methods: In this retrospective study, the incidence of HCC was determined in CHB patients initiated on NA from 1999 to 2012. Pre-treatment data utilised in the REACH-B model were used to predict the annual HCC risk. The standardised incidence ratio (SIR) for HCC was calculated by comparing the observed to expected number of cases, and HCC risk factors determined by Cox proportional hazards regression.
Results: Five hundred and forty nine initiated NA (14% lamivudine, 5% adefovir, 1.5% telbivudine, 39% entecavir, 41% tenofovir). Over a median follow-up of 3.2 years (IQR 1.9-4.6), 11 (3.2%) were diagnosed with HCC. Among 322 with data to calculate the REACH-B model, the median age at treatment initiation was 46 years (IQR 38-55), 65% were male, 32% HBeAg positive and 20% had cirrhosis. The median pre-treatment ALT was 71 U/L (IQR 41-127) and HBV DNA was 6.48 log10 copies/mL (4.95-8.04). The observed annual HCC incidence (0.9%; 95% CI 0.5-1.7) was significantly lower than predicted without treatment by the REACH-B model (SIR 0.46; 95% CI 0.23-0.82); this risk was reduced after 4 years of therapy (SIR 0.49; 95% CI 0.2-1.00).
Conclusions: In this Canadian study of nucleos(t)ide analogues-treated patients with chronic hepatitis B, the incidence of HCC was lower than expected, suggesting that NA reduce the risk of chronic hepatitis B-related HCC.
© 2014 John Wiley & Sons Ltd.