How can we optimise shortening of Hepatitis C treatment? Observational and interventional approaches

Although direct-acting antivirals (DAAs) are shorter and better tolerated than older treatments for Hepatitis C (HCV), many HCV-infected patients who are not yet cured will find adhering to standard treatment lengths challenging. However, which patients might achieve cure with short-course treatment is unclear. This thesis used data from two trials investigating short-course HCV treatment, STOP-HCV-1 and SEARCH-1, to explore factors associated with cure, time to treatment failure and with the kinetics of viral load (VL) rebound to help understand which patients might be most suitable for short-course treatment. Quality of life (QoL) with varying treatment durations and ribavirin was investigated to determine acceptability of treatment options. The statistical aspects of VIETNARMS, a HCV treatment trial with an early stopping mechanism, were assessed. Baseline VL was strongly associated with all outcomes. The probability of cure was highest in those with HCV genotype 1b, CC IL28B genotype, no DAA resistance and no current substance abuse at baseline. Treatment failure happened quicker in older participants, those with genotype 1a and with shorter treatment lengths. Peak rebound VL was higher in those taking ribavirin, no DAA resistance and a higher Fibroscan score, and faster in those with IL28B genotype CC. There were no clear differences in QoL between those taking different lengths of treatment and no lasting impact on QoL of taking ribavirin. A statistically rigorous approach to monitoring the VIETNARMS design was developed: the probability of incorrectly or correctly stopping a group during follow-up was appropriate and overall power at the final analysis will be sufficiently high. Overall, baseline VL was the main determinant of suitability for short-course treatment, with HCV genotype, IL28B genotype and pre-existing DAA resistance also key factors. Treatment lengths and ribavirin use can be tailored to these patient characteristics without compromising QoL and could be tested in well-designed trials