Reply

Abstract

We would like to thank Manco and colleagues for their interest in our article.1 We agree that stratification of our study cohort according the presence of diabetes and HFE (hereditary hemochromatosis) genotype status could be of interest. We have conducted additional analyses with regard to diabetes status in our cohort. HFE genotyping was not routinely performed as part of the Nonalcoholic Steatohepatitis Clinical Research Network (NASH CRN) Database study; this is currently in progress in an ancillary study. Among the 849 adult subjects in our study, 221 were enrolled in the PIVENS (Pioglitazone or Vitamin E for NASH Study) clinical trial,2 which excluded diabetic patients; therefore, to prevent any potential bias, we report here the relationship of diabetes and iron deposition in the remaining 628 subjects. As shown in Table 1, diabetic NASH CRN subjects were less likely to have hepatic iron deposition. In particular, diabetic patients had significantly less hepatocellular but not reticuloendothelial system iron deposition. There are several possible explanations for these findings. The iron regulatory hormone hepcidin is expressed in adipose tissue and has been shown to correlate to body mass index.3 It is possible that increased visceral fat in our diabetic subjects could lead to greater circulating hepcidin levels, reduced iron absorption, and less hepatocellular iron deposition. Recently, alcohol has been shown to down-regulate hepcidin gene expression.4 Because diabetic subjects in our study consumed less alcohol than subjects without diabetes, this could also contribute to higher hepcidin levels in these patients. Diabetic subjects were also more likely to be women, and physiologic blood loss through menstruation and childbirth may have resulted in decreased body iron stores. Additional studies are warranted to define the mechanism for decreased hepatocellular iron content in diabetic patients with NAFLD. In contrast to Manco et al.,5 who have reported hepatic iron deposition in 29% of their Italian pediatric subjects, which was not associated with measurements of insulin resistance, the prevalence of iron deposition in the NASH CRN pediatric subjects was only 4%. It is likely that, as in adults, differences in the Italian and U.S. cohorts, including body mass index (25.2 in Manco et al. versus 33.0 in the NASH CRN pediatric subjects) account for the observed differences.6 James Nelson Ph.D.*, Kris Kowdley M.D.*, * Benaroya Research Institute, Virginia Mason Medical Center, Seattle, WA.