Background: There are few published data characterizing patterns of liver stiffness measurements (LSMs) among HCV-infected persons and their potential impact on clinical decisions (for example, deferring treatment and hepatocellular carcinoma surveillance).
Methods: A total of 591 HCV-infected injection drug users in a community-based cohort had four LSMs. We used semi-parametric latent class growth modelling to identify patterns, which then became a gold standard against which we characterized validity of information from the initial measurements.
Results: Median age was 49, 68% were male, 92% African-American and 33% HIV-coinfected. The median LSM at visit 1 was 6.7 kPa (IQR 5.3-8.8). Over a median 1.75 years, LSM measures were stable; median change between visits was 0 kPa (IQR -1.4-1.7). Only 3% had evidence of fibrosis progression. Other groups included stable patterns of no fibrosis (59%), moderate fibrosis (21%), severe fibrosis (7%) and cirrhosis (9%). Individuals with fibrosis progression were more likely to be HIV-infected than those with stable low fibrosis (P<0.001). The diagnostic accuracy of the first LSM for identification of need for cancer surveillance (cirrhosis ≥12.3 kPa) was high (positive predictive value =97%). Although no single low LSM had high negative predictive value for significant fibrosis (metavir <2), individuals with two or more low results rarely had progression.
Conclusions: These data underscore the stability of liver fibrosis in a cohort of predominantly African-American HCV-infected persons over 1.75 years, support using LSMs to monitor untreated persons at risk for progression and assess need for hepatocellular carcinoma surveillance.