This journal club reviewed a recent Hepatology article on the predictive value of serum ferritin on the stage of fibrosis in NASH. Data was colleted from a large registry that excluded other forms of liver disease. This is important, because we recognize that ferritin, as well as iron stores, are influenced by other forms of chronic liver disease including viral hepatitis, alcoholic liver disease, cirrhosis and, of course, all of the diseases of iron dysmetabolism.
The groups were divided into those with normal ferritin, ferritin > 1.5 X the ULN, and ferritin > 2.5 X the UNL. Odds ratios were used to calculate the likelihood of advanced fibrosis. Sure enough, there appeared to be a correlation between ferritin level and fibrosis level. The authors concluded that ferritin levels can be used as a prognostic indicator of disease severity.
We had two major problems with this conclusion. First of all, it is not clear whether there were different ferritin assays (and different cutoffs for ULN) used at the various medical centers that collected data; if so, the absolute level cannot be compared from one to the next. Secondly, the supposedly predictive values were not presented in AUROC graphs. Therefore, what value should be considered predictive of advanced fibrosis? A ferritin of 250? 700?
Nevertheless, ferritin may be useful one day as a component in a larger panel of serum markers in gauging severity of NASH, similar to the panels we currently use to stage hepatitis C (ie. Hepascore, Fibrotest). For the full article, please click on the link below: