Results
In the cohort of approximately 39,000 individuals with genotype data available in the eMERGE Network, we identified 618 individuals with the p.Cys282Tyr homozygous or p.[Cys282Tyr];[His63Asp] compound-heterozygous genotype; of these, 538 individuals had corresponding data from electronic medical records. Demographic characteristics of these individuals are summarized in Table 1. The mean age of participants was 66.4 years (±15.8), and the average age at HH diagnosis was 59.6 years (±12.5). The subsequent analysis included a total of 495 individuals of European ancestry (92%), 98 with the p.[Cys282Tyr];[Cys282Tyr] and 397 with the p.[Cys282Tyr];[His63Asp] genotype. We excluded six Northwestern participants who were ascertained from a liver clinic from the calculation of HH penetrance; 95 p.Cys282Tyr homozygotes and 392 compound heterozygotes with information on HH diagnosis were included in this analysis.
The frequency of HH diagnosis was 24.4% in male p.Cys282Tyr homozygotes and 3.5% in male p.[Cys282Tyr];[His63Asp] compound heterozygotes (p < 0.001), whereas the diagnostic rate was 14.0% in female p.Cys282Tyr homozygotes and 2.3% in female compound heterozygotes (p < 0.001). A summary of the differences in diagnostic rate and relevant phenotypes of HH across genotypes is shown in Table 2, and a comprehensive list of phenotypes is shown in Table S3. The diagnostic rates by genotype and study site are shown in Table S4. The Kaplan-Meier curve (Figure 1) demonstrates the frequency of HH diagnoses by age and sex for p.Cys282Tyr homozygotes and p.[Cys282Tyr];[His63Asp] compound heterozygotes separately.
As expected, for many signs of HH, the penetrance was higher in p.Cys282Tyr homozygotes than in p.[Cys282Tyr];[His63Asp] compound heterozygotes, although some signs of HH did not differ by genotype. Iron studies were significantly different between genotypes in males only, where transferrin saturation above 50% was more common in homozygotes than in compound heterozygotes (100% versus 37.5%; p = 0.003), and serum ferritin higher than 300 ng/ml was more frequent in homozygotes (77.8% versus 33.3%; p = 0.006). No differences in iron studies were found across genotypes in females.
The overall prevalence of liver disease in males was 34.3% in homozygotes and 24.4% in compound heterozygotes (p = 0.279), and in females the prevalence was 29% for both genotypes. The rate of liver biopsy was also significantly different between homozygotes and compound heterozygotes for males (10.9% versus 1.8%; p = 0.013) and females (9.1% versus 2%; p = 0.035). No significant genotype differences were found for other liver phenotypes, which included the presence of any liver disease, liver cirrhosis, other chronic liver diseases, hepatocellular carcinoma, elevated transaminases, or physical findings consistent with liver disease (e.g., hepatomegaly or ascites).
Differences between genotypes were found in the proportion of individuals with a history of phlebotomy in males (19.6% in homozygotes versus 2.9% in compound heterozygotes; p = < 0.001) and females (8% versus 0.5%; p = 0.005), as expected. Males, but not females, had differences between genotypes in the rates of diabetes (44.7% versus 28%; p = 0.03) and family history of HH (8.1% versus 0%; p = 0.006), whereas females had genotype differences in the proportion of hand X-ray for evaluation of arthritis (24.5% versus 11.5%; p = 0.023). No significant differences were found between genotypes for either sex in the rates of congestive heart failure, cardiomyopathy, osteoarthritis, hypogonadism, history of alcohol or tobacco abuse, use of over-the-counter medication for arthritis, diabetes medication (including insulin), proportion of individuals with imaging studies (e.g., abdominal ultrasound or echocardiogram), presence of arthralgia, pain on palpation of proximal interphalangeal or metacarpophalangeal joints, or skin hyperpigmentation.
In one of our study sites (Vanderbilt University), it was noted that among 41 p.Cys282Tyr homozygotes, 7 (17%) were receiving iron, and none of these individuals had a diagnosis of HH. Out of these seven individuals, three had iron labs tested: two had normal studies, and one underwent Roux-en-Y gastric bypass surgery, after which iron was prescribed (iron saturation was initially normal at 37% but then jumped to 71% after 6 months of iron-replacement therapy).