Prioritization of HCV treatment in the direct-acting antiviral era: An evaluation.

Background & Aims :We determined the optimal HCV treatment prioritization strategy for interferon-free (IFN-free) HCV direct-acting antivirals (DAAs) by disease stage and risk status incorpo-rating treatment of people who inject drugs (PWID). Methods :A dynamic HCV transmission and progression model compared the cost-effectiveness of treating patients early vs. delaying until cirrhosis for patients with mild or moderate fibrosis, where PWID chronic HCV prevalence was 20, 40 or 60%. Treatment duration was 12 weeks at £3300/wk, to achieve a 95% sustained viral response and was varied by genotype/stage in alternative scenarios. We estimated long-term health costs (in £UK = € 1.3 = $1.5) and outcomes as quality adjusted life-years (QALYs) gained using a £20,000 willingness to pay per QALY threshold. We ranked strategies with net monetary benefit (NMB); negative NMB implies delay treatment. Results :The most cost-effective group to treat were PWID with moderate fibrosis (mean NMB per early treatment £60,640/£23,968 at 20/40% chronic prevalence, respectively), followed by PWID with mild fibrosis (NMB £59,258 and £19,421, respectively) then ex-PWID/non-PWID with moderate fibrosis (NMB £9,404). Treatment of ex-PWID/non-PWID with mild fibrosis could be delayed (NMB -£3,650). In populations with 60% chronic HCV among PWID it was only cost-effective to prioritize DAAs to ex-PWID/non-PWID with moderate fibrosis. For every one PWID in the 20% chronic HCV setting, 2 new HCV infections were averted. One extra HCV-related death was averted per 13 people with moderate disease treated. Rankings were unchanged with reduced drug costs or varied sustained virological response/duration by genotype/fibrosis stage. chronic HCV prevalence settings among PWID (20%, 40% and 60%) at baseline. This corresponded to baseline incidences of infection/reinfection among PWID of 4% (2.5–97.5% Confidence Interval (95% CI) 3–5%), 9% (95% CI 7–13%), and 21% (95% CI 15–30%), in the 20%, 40%, and 60% chronic prevalence scenarios, respectively. The model was open, with PWIDs entering the population on initiation of injecting and were tracked after permanent cessation of injecting when they were assumed to be no longer at risk of reinfection or transmission. The model was a deterministic, compartmental model which was stratified by risk status (PWID, former-PWID), HCV genotype (genotype 1 and 4, genotype 2, and genotype 3) and infection status and disease stage (never infected or infected and spontaneously cleared, mild HCV, moderate HCV, compensated cirrhosis [CC], decompensated cirrhosis [DC], hepatocellular carcinoma [HCC], liver transplant, post-transplant). For simplicity, we assumed an individual had one dominant genotype strain which affected treatment SVR, and that an individual’s risk of acquiring a specific genotype was related to the circulating prevalence of each genotype. Additionally, for those stages eligible for antiviral treatment (mild HCV, moderate HCV, and compensated cirrhosis), the model was further stratified by treatment status (never treated, on treatment, SVR, non-SVR). Those who achieved SVR were at risk of reinfection; we assumed no change in risk behavior after treatment, so each individual’s risk of reinfection was equal to that of pri- mary infection. We assumed that those with mild or moderate fibrosis who achieved SVR were at no risk of further liver disease progression unless they were reinfected. Based on clinical evidence, we assumed that those with compensated cirrhosis who achieved SVR remained at elevated risk of disease progression due to existing liver damage [20,21]. Individuals who did not attain SVR proceeded in base- case), behavior after treatment (50% reduction in risk after treatment compared to no reduction), average injecting duration (5 or 20 years, compared to a mean of 11 years in base-case). Additionally, because future costs are uncertain and likely to reduce over time, we simulated an additional scenario where DAA costs reduced by 75% in 10 years (from £3300 to £825/week (approximately $1200 USD/week), comparable to the price of sofosbuvir/ledipasvir available in some low/middle income countries.