To conduct a cost-effectiveness analysis of currently available nucleos(t)ide antiviral treatments (lamivudine, telbivudine, entecavir, and tenofovir) for chronic hepatitis B in Canada.
Markov modeling was used to project the lifetime health benefits and costs associated with the antiviral treatments. The hypothetical patient population was hepatitis B e antigen–positive chronic hepatitis B–infected patients aged 34 years. Quality-adjusted life-years were used as a measure of effectiveness. Long-term cumulative incidence of liver complications was also projected. Treatment effectiveness data were derived from the literature; meta-analysis was conducted when there was a large variance in reported effectiveness data. Costs were obtained from a cost analysis of treating chronic hepatitis B–related complications in Canada. Stochastic parameter uncertainty was examined in probabilistic sensitivity analysis by using second-order Monte Carlo simulation. Alternative modeling assumptions were assessed in scenario analysis. One-way sensitivity analysis was used to explore each parameter's impact on the uncertainty of the results.
In the base-case analysis, telbivudine was dominated by entecavir and tenofovir. Tenofovir strictly dominated lamivudine, telbivudine, and entecavir. Over the 72-year period of the model, the expected life expectancy (undiscounted) of lamivudine, telbivudine, entecavir, and tenofovir was 35.71, 36.94, 37.65, and 37.99 years, respectively. Tenofovir had the highest expected quality-adjusted life-years at 11.86 (discounted) in all comparisons. Scenario and sensitivity analyses proved the robustness of the base-case results. The projected 10-year cumulative incidence of cirrhosis and hepatocellular carcinoma was 11.40% and 3.05%, respectively, for tenofovir, which is significantly lower than that for lamivudine.
Tenofovir generated the best results compared with all other therapies under evaluation.