icon-folder.gif   Conference Reports for NATAP  
 
  American Association for the Study of Liver Diseases 2003 Conference
Boston, MA
Oct 24-28, 2003
Back grey_arrow_rt.gif
 
 
 
Natural History of Fibrosis in Chronic Hepatitis C
 
 
  Reported by Jules Levin
54th AASLD
Oct 24-28, 2003
Boston, MA
 
Poster abstract #56. John B Wong, Tufts-New England Medical Center, Boston, MA; Thierry Poynard, Groupe Hospitalier Pitie Salpetriere, Paris, France
 
This study examines HCV monoinfection. Natural history of HCV/HIV co-infection needs to be addressed. In the post HAART era studies find HIV accelerates HCV fibrosis 1.5 to 2 times. One study suggests HAART might slow progression but other studies find differently. One study finds HIV accelerates decompensated HCV-cirrhosis 5 fold.
 
The natural history of hepatitis C (HCV) remains controversial with little data beyond the first 2 decades and conflicting estimates based on study design and population characteristics, such as age and gender. Using 2313 liver biopsies from untreated patients including some with biopsies 20 to 40 years post infection, we translated a Cox proportional hazards model into a fibrosis-based Markov model.
 
The study aim is to compare cohort simulation projections to published outcomes and to predict future outcomes. Cox proportional hazards models estimated the likelihood of developing Metavir fibrosis stages F1, F2, F3 or F4 over time. For all models, covariates included age, gender, alcohol consumption (>50 gm/day), injection drug use, and Metavir inflammation A2 or A3. These Cox models and recent UNOS, SEER and NIH data were used to modify a previously published computer cohort simulation (Wong, JAMA 1998). Summary patient characteristics from the retrospective-prospective Kenny-Walsh (NEJM 1999) and Thomas (JAMA 2000) studies were then applied to the fibrosis-based Markov model to estimate the observed outcomes and to project future outcomes.
 
For the Kenny-Walsh study (n=376), the mean age of the Irish women when they received HCV-contaminated anti-D immune globulin was 28 years. After a mean of 17 years, liver biopsy revealed F0 in 49%, F1 in 34%, F2 in 10%, F3 in 5% and F4 in 2%. Markov model projections were 50% F0, 36% F1, 10% F2, 2% F3 and 1.6% F4. After 27 years of follow-up, the model predicted 20% F2, 4% F3 and 7% F4; and after 37 years of follow-up, 27% F2, 8% F3 and 20% F4. To examine the impact of selected cohort factors, the 17-year incidence of cirrhosis (base-case estimate 1.6%) was 3% if all of the women had instead acquired HCV through injection drug use and 6% if they drank >50 gm alcohol per day. If the women had been older, the 17-year cirrhosis incidence rose to 5% for 38 year-olds and 14% for 48 year-olds. Thus, age and duration of infection are the primary determinants of fibrosis progression. Despite the slow fibrosis progression, the projected life expectancy for this cohort was 41.8 years and 37.3 quality-adjusted life years compared to an expected 51.7 years for 28 year-old women. The projected lifetime medical care costs were $52,000 without antiviral treatment.
 
For the Thomas study (n=1667), the median age at first injection drug use was 20 years. With follow-up exceeding 15 years in over 75% of patients, cirrhosis was found in 2.4%. The Markov model predicted a 2.5% incidence of cirrhosis after 15 years. Future cirrhosis predictions were 11% after 25 years and 30% after 35 years. Again, to examine the impact of selected cohort factors, the 15-year incidence of cirrhosis (base-case estimate 2.4%) was 1.3% if all of the patients had instead acquired HCV through transfusion and 7% if they drank >50 gm alcohol per day. If the patients had been older, the 15-year cirrhosis incidence rose to 8% for 30 year-olds, 21% for 40 year-olds, and 46% for 50-year olds. Thus, age and duration of infection are the primary determinants of fibrosis progression. The projected life expectancy for this cohort was 30.4 years and 26.2 quality-adjusted life years compared to an expected 56.3 years for 20 year-olds. The projected lifetime medical care costs were $66,359 without antiviral treatment.
 
Our fibrosis-based Markov model predictions matched observed community cohort HCV outcomes well. The model suggests rapid fibrosis progression beyond 20 years and very rapid progression beyond 30 years. Despite the slow early progression, hepatitis C may still significantly reduce life expectancy and quality of life and induce substantial medical care costs because of the normally long life expectancy in these community cohorts and this delayed rapid progression. Although men, injection drug use, alcohol use and active hepatic inflammation all raise the likelihood of progression, increasing duration of infection and older age at infection onset are the primary determinants of fibrosis progression.