Introduction
Hereditary hemochromatosis (HH [MIM: 235200]) is the most common genetic disorder identified in those of European ancestry and is characterized by an inappropriate increased absorption of dietary iron. If left untreated, HH can lead to morbidity and mortality, including liver cirrhosis, hepatocellular carcinoma, diabetes, and heart disease.1 If treatment with regular phlebotomy is initiated before organ damage develops, these complications can be prevented, and people with HH can have a normal life expectancy.2 Variants in HFE (MIM: 613609; GenBank: NM_000410.3) are associated with the majority of adult-onset HH cases; the pathogenic variant with the highest penetrance is c.845G>A, a missense mutation (minor allele frequency [MAF] = 4% in the 1000 Genomes European population) that results in the substitution of tyrosine for cysteine in the protein product (p.Cys282Tyr) and accounts for 80%–85% of individuals with HH.3,4 A more common HH variant (MAF = 17% in the 1000 Genomes European population) is c.187C>G, which results in the substitution of aspartate for histidine at amino acid position 63 (p.His63Asp); this variant causes a milder degree of iron overabsorption and is most relevant to disease when paired with the allele for p.Cys282Tyr in compound heterozygotes.3 To improve clarity, we will henceforth use the protein change as a surrogate for genotype. Several other hemochromatosis-associated pathogenic variants in HFE have been described, as well as variants in genes other than HFE (e.g., HJV [MIM: 608374], HAMP [MIM: 606464], and TFR2 [MIM: 604720]), but these are rare.5,6
Estimates of penetrance for HH-related variants vary widely depending on the signs or symptoms used in disease assessment.7 For instance, in p.Cys282Tyr homozygotes from a racially diverse cohort, the prevalence of elevated serum ferritin (>300 ng/ml for males and >200 ng/ml for females) and transferrin saturation levels (>50% in men and >45% in women) was 40%–60% in females and 75%–100% in males.8–12 In contrast, in another cohort of northern European descent, the prevalence of iron-overload-related disease (defined as liver fibrosis, elevated transaminases, hepatocellular carcinoma, arthropathy, or physician-diagnosed HH in individuals with high ferritin or transferrin saturation) in p.Cys282Tyr homozygotes was 1.2% (95% confidence interval [CI] = 0.03–6.5) in females and 28.4% (95% CI = 18.8–40.2) in males.13 The lower clinical penetrance in women has been suggested to result from iron loss through menstrual bleeding and childbirth, although evidence is lacking.5
These variations in estimates of penetrance have resulted in insufficient evidence for confidently projecting the impact, or estimating the benefit, of widespread genetic screening for HH.14 Whereas newborn screening for HH has not been adopted because of its adult onset and incomplete penetrance,15 the utility of using existing genetic data for screening for HH risk has not been carefully addressed. The American College of Medical Genetics and Genomics (ACMG) does not currently include HH among the gene-disease pairs it recommends evaluating and returning as incidental findings for genome-scale sequencing,16 although other recommendations have been proposed.17 Development of such recommendations would be informed by the penetrance and rate of diagnosis of HH in unselected samples.
The primary objective of this study was to determine the frequency of diagnosis of HFE-related HH and to estimate the penetrance of clinically related variables in the Electronic Medical Records and Genomics (eMERGE) Network, a national consortium organized by the National Human Genome Research Institute to develop, disseminate, and apply approaches to research by combining DNA biorepositories with electronic-medical-record systems for high-throughput genetic research. The network’s goals include returning genomic testing results to individuals in a clinical setting.18,19 This cohort was generally ascertained independently of HH diagnosis. By identifying all participants carrying the p.[Cys282Tyr];[Cys282Tyr] and p.[Cys282Tyr];[His63Asp] genotypes and reviewing their medical records, we obtained a minimally biased estimate of general population frequency of HH diagnosis and related signs in those at genetic risk. A secondary objective of this study was to serve as a proof of principle to determine the efficacy of this multicenter consortium to estimate the penetrance of common phenotypes associated with relatively uncommon genetic variants.