We aimed to identify determinants of hepatocellular carcinoma (HCC) in cirrhotic patients who received nucleos(t)ide analogues for chronic hepatitis B (CHB).Patients and methods:
This retrospective-prospective study screened all patients (n = 1630) who received antiviral therapy for CHB between 1 September 2007 and 31 March 2013 at the E-Da Hospital and enrolled 210 consecutive cirrhotic patients with pretreatment viral DNA >2000 IU/mL. Those who developed HCC within 3 months of treatment were excluded. All participants were observed until occurrence of HCC, death or 1 January 2014. The incidence and determinants of HCC were estimated using competing risk analyses adjusted for mortality.Results:
Thirty-five (16.7%) patients developed HCC during a median follow-up of 25.2 months (IQR, 16.3-37.3 months), with a cumulative incidence of 24.1% (95% CI, 16.3%-32.0%) at 5 years. Multivariate-adjusted analyses identified age >55 years [adjusted hazard ratio (HR), 2.19; 95% CI, 1.03-4.66], male gender (adjusted HR, 3.07; 95% CI, 1.05-9.02), model for end-stage liver disease (MELD) score >12 points (adjusted HR, 2.16; 95% CI, 1.10-4.23) and diabetes mellitus (DM; adjusted HR, 3.49; 95% CI, 1.54-7.91) as independent risk factors after adjusting for multiple covariates, including antidiabetes medication. A scoring formula that used information on age, gender, MELD score, DM and antidiabetes regimen significantly discriminated patients at high or low risk of HCC, with sensitivity and specificity of 82.9% and 62.3%, respectively.Conclusions:
Age, gender, hepatic dysfunction, DM and medication for DM are baseline factors that stratify the risk of HCC in cirrhotic patients who receive nucleos(t)ide analogues for CHB.