LAUSR

Background Antiviral therapy for chronic hepatitis B (CHB) is effective and can substantially reduce the risk of progressive liver disease and hepatocellular carcinoma but is often administered for an indefinite duration. Adherence has been shown in clinical trials to maximize the benefit of therapy and prevent the development of resistance, however the optimal threshold for predicting clinical outcomes has not been identified. The aim of this study was to analyse adherence using the medication possession ration (MPR) and its relation to virological outcomes in a large multi-centre hospital outpatient population, and guide development of an evidence-based threshold for optimal adherence. Methods Pharmacy and pathology records of patients dispensed CHB antiviral therapy from 4 major hospitals in Melbourne between 2010 and 2013 were extracted and analysed to determine their MPR and identify instances of unfavourable viral outcomes. Viral outcomes were classified categorically, with unfavourable outcomes including HBV DNA remaining detectable after 2 years treatment or experiencing viral breakthrough. The association between MPR and unfavourable outcomes was assessed according to various thresholds using ROC analysis and time-to-event regression. Results Six hundred forty-two individuals were included in the analysis. Median age was 46.6 years, 68% were male, 77% were born in Asia, and the median time on treatment was 27.5 months. The majority had favourable viral outcomes (91.06%), with most having undetectable HBV DNA at the end of the study period. The most common unfavourable outcome was a rise of < 1 log in HBV DNA (6.54% of the total), while 2.49% of participants experienced viral breakthrough. Adherence was linearly associated with favourable outcomes, with increasing risk of virological breakthrough as MPR fell. Decreasing the value of MPR, at which a cut-point was taken, was associated with a progressively larger reduction in the rate of unfavourable event; from a 60% reduction under a cut-point of 1.00 to a 79% reduction when the MPR cut-point was set at 0.8. Conclusion Lower adherence as measured using the MPR was strongly associated with unfavourable therapeutic outcomes, including virological failure. Optimising adherence is therefore important for preventing viral rebound and potential complications such as antiviral resistance. The evidence of dose-response highlights the need for nuanced interventions.