Predicting Progression to Cirrhosis in Chronic Hepatitis C Virus Infection

A. J. Freeman, M. G. Law, J. M. Kaldor, G. J. Dore

Disclosures

J Viral Hepat. 2003;10(4) 

In This Article

Methods

Published studies of people with chronic HCV infection who had undergone assessment for stage of chronic liver disease were identified as described previously.[7] The studies were divided into four groups: cross-sectional series of people referred to specialist liver clinics, longitudinal studies of people with post-transfusion non-A non-B hepatitis subsequently diagnosed with chronic HCV infection, community-based studies and cross-sectional series of people diagnosed at blood donor screening. The prevalence of cirrhosis in the population studied, and data related to factors proposed as influencing progression to cirrhosis, was extracted. The factors examined included: patient ethnicity, sex distribution, alcohol consumption, age at infection, mode of HCV acquisition, serum transaminase levels, histological grade of inflammation, HBV status, HIV co-infection, HCV genotype and viral load.

An ecologic analysis was used to estimate relative risk of cirrhosis for factors identified as influencing fibrosis progression. For each study, mean duration of HCV infection was abstracted when available. In those studies not reporting mean duration of HCV infection, it was estimated as described previously.[7] For each study, cirrhosis prevalence was standardized to a duration of chronic HCV infection of 20 years based on a linear fibrosis progression rate. For example, a study with an estimated mean duration of HCV infection of 10 years and 5% cirrhosis prevalence would have an estimated 20 year cirrhosis prevalence of 10%. The impact of continuous variables on progression to cirrhosis was graphically assessed by plotting cirrhosis prevalence after 20 years against the study mean for each factor. For categorical data, the proportion of subjects positive for the factor in question was plotted against cirrhosis prevalence after 20 years. Age at infection was analysed as a continuous and as a categorical variable (< 30 and ≥ 30 years). Only studies that provided information regarding the factor in question were included. To allow estimation of relative risk of cirrhosis using linear regression, a transformation was performed: y = ln {- ln [1 - cirrhosis(20)]}. Then, for each study category with a sufficient number of eligible studies, least squares linear regression lines were fitted: y = µstudy category + xβfactor. Confidence intervals were based on the standard error of the slope of the regression line. Where regression line slope was not significantly different between study categories, it was assumed that the modifying effect of the factor in question was the same regardless of the population examined. A mean linear regression line slope, weighted according to the number of studies in each category, was generated. For factors identified as influencing disease progression, mean regression line slope was significantly different from zero and it was used to estimate the relative risk of cirrhosis: RRfactor = exp(βfactor). Stepwise multiple linear regression was performed to adjust forcolinearity between the factors studied.

With regard to histological inflammation, the inflammatory component of the histological activity index (HAI,[9,10] maximum score 12 excluding confluent necrosis) was extracted where available. Alternative grading scores (Scheuer,[11] inflammatory component, maximum score 8; Metavir,[12] maximum score 3) were proportionally adjusted to a score out of 12. Where grade of liver disease was not scored, patients with chronic persistent hepatitis (CPH) were allocated a score of 3; patients with mild, moderate and severe chronic active hepatitis (CAH) were allo-cated a score of 6, 9 and 12 respectively; and an estimated mean HAI was calculated. In six studies that reported mean HAI and described grade of liver disease in terms of Scheuer score or CPH and CAH, the calculated HAI was not significantly different and the two values correlated (Spearman r = 0.94, P = 0.02).[13,14,15,16,17,18] Analyses were performed including and excluding studies in which HAI had to be calculated. HAI was analysed as a continuous and as a categorical variable.

Cirrhosis prevalence beyond 20 years chronic HCV infection is uncertain. A Weibull distribution was chosen to estimate progression to cirrhosis among subjects with chronic HCV infection: cirrhosis (T) = 1 - exp (- exp (α + Σβ) Tλ). The equation was solved for α and λ using estimates of cirrhosis prevalence after 20 and 40 years chronic HCV infection of 6.5% (3.5-9.5%)[7] and 20% (10-40%)[19] respectively. To predict progression to cirrhosis based on risk factor profile, data was normalized to the community-based studies as these would appear to be the most representative of disease progression among the majority of people with chronic HCV infection in industrialized countries.

Data was analysed using Stata 7.0 (Stata Corporation, TX, USA). A P value less than 0.05 was considered significant. Data is presented with 95% confidence intervals.

Comments

3090D553-9492-4563-8681-AD288FA52ACE
Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as:

processing....