Venue: The Fuqua School of Business, Duke University, 1 Towerview Drive, Durham, NC 27708-0120
Presentation
Natural History of Chronic HCV Infection Obtained Through Injection Drug Use: A Bayesian Meta-Analysis
Background: Chronic Hepatitis C viral (HCV) infection can result in fibrosis of the liver and cirrhosis. In North America, more than 70% of new infections occur in injection drug users (IDUs), therefore understanding the natural history of chronic HCV in this population is essential to inform future health care policy.
Aim: The purpose of this study was to estimate the rate of progression to cirrhosis for those infected with HCV through injection drug use.
Methods: Systematic review of the literature identified articles with information on the mean duration of infection and the prevalence of cirrhosis for those who obtained HCV infection through injection drug use. Data on mean age, mean alanine aminotransferase (ALT) enzyme levels, proportion male, proportion with HIV co-infection, proportion with alcohol abuse, and study setting (academic liver clinic, community-based clinic or addiction therapy) were abstracted. Summary progression rates were estimated using random effects Poisson meta-regression, fitted with WinBUGS software. Uninformative prior distributions were used. The impact of study co-variates on the progression rate was assessed by estimating the posterior probability that the relative risk (RR) exceeds 1.0.
Results: Systematic review identified 5,225 abstracts. Abstract review identified 459 relevant articles and a total of 41 articles met the inclusion criteria. Each of the 41 studies had a retrospective study design. The progression rate estimate (adjusted for all co-variates) was 11.2 per 1000 person-years (95% Credible Region, 4.9 to 27.2 per 1000 person-years) corresponding to a 20-year cirrhosis prevalence of 20.2% for patients with chronic HCV infection, in a community-based clinic/addiction therapy setting, in which patients at advanced stage of disease are excluded. The progression rate derived from studies in which chronic infection (HCV-RNA testing) was not established was 0.5 times lower. Faster progression was associated with a greater proportion male and a greater proportion with alcohol abuse, but not a greater proportion co-infected with HIV (probability RR>1 = 0.90, 0.84 and 0.56, respectively). Two studies (one with a large sample size) that demonstrated no difference in prognosis associated with HIV co-infection may explain this counterintuitive result.
Conclusion: The study supports the hypothesis that source of infection does not impact on the rate of progression to cirrhosis. Similar progression rate estimates have been derived from studies conducted in post-transfusion or liver clinic cohorts.