Hemochromatosis Case Definition: Out of Focus?
Although hemochromatosis was first described in the 19 century, John Sheldon's 1934 description of 300 cases remains a landmark publication in defining the clinical characteristics of the disease. Since that time there have been many advances, including the discovery of the hemochromatosis (HFE) gene in 1996, mutations of which (e.g. 845G·A [C282Y]) are found in the majority of patients with typical hemochromatosis. Expert opinion is, however, varied regarding what constitutes the ideal case definition for hemochromatosis. This has led to confusion amongst clinicians and patients.
Why do we need a case definition for any medical condition? Has the subclassification of any group of clinically similar diseases improved life for the patient with a single subclass of those diseases? A precise case definition is useful to study the natural history of a disease, direct appropriate treatment, and judge response to therapy. Furthermore, a consistent case definition is particularly important when studies are compared and outcomes are combined, such as in a meta-analysis.
Most of the debate regarding hemochromatosis centers on its phenotypic or genotypic diagnosis. Whether a clinically asymptomatic patient with elevated iron levels should be classified as having hemochromatosis is currently an issue of debate. Some investigators have proposed a case definition for hemochromatosis comprising a sustained elevation in serum ferritin levels (>200 µg/l in women and >300 µg/l in men) and transferrin saturation (>45% in women and >50% in men) in the absence of any other risk factors for iron overload. This approach has the advantage of simplicity, but it would probably identify a number of patients who do not have iron overload. A screening study demonstrated that a transferrin saturation of >55% had a positive predictive value of only 4.4% for identifying male 845G·A (C282Y) homozygotes.
Other researchers insist that iron overload should be documented and quantified by liver biopsy. The pattern of iron distribution might be helpful in defining the cause of disease. The hepatic iron index (hepatic iron concentration divided by patient age) was devised before genetic testing, and originally used to differentiate patients with alcoholic siderosis from those with genetic hemochromatosis. Later it was extrapolated for use in the case definition of hemochromatosis. This introduced the concept that a quantitative measure of iron overload should be used to define the disease. Subsequent studies using genetic testing, however, have shown that many patients with early hemochromatosis do not have an elevated hepatic iron index (>1.9), and that patients with other types of iron overload do have an elevated hepatic iron index. Liver biopsy is unappealing to patients with mild abnormalities (serum ferritin levels <1,000 µg/l), and MRI has a poor sensitivity for the detection of mild iron overload.
Quantitative phlebotomy is another method to measure iron overload. The removal of 5 g of mobilizable iron by phlebotomy (approximately 20 × 500 ml phlebotomies) followed by a fall in serum ferritin levels to under 50 µg/l has been considered as a case definition.
The creation of arbitrary quantitative thresholds for any of these variables (serum ferritin levels, transferrin saturation, liver iron, iron removed) in order to define hemochromatosis has obvious limitations. A patient just below a threshold might have a clinical diagnosis as strong as another patient just above a threshold.
Another perspective on the case definition for hemochromatosis is that patients should have a constellation of clinical signs and symptoms. The association of any symptoms of hemochromatosis with elevated iron levels or 845G·A (C282Y) homozygosity has, however, come under increased scrutiny. Patients with bronze diabetes (a type of diabetes associated with hemochromatosis) probably represent <1% of all hemochromatosis patients, and studies have shown that the prevalence of diabetes is similar in both 845G·A (C282Y) homozygous and control populations. Other symptoms associated with hemochromatosis, such as fatigue and arthralgia, are nonspecific and very common in the general population. Of all of the putative symptoms of hemochromatosis, liver disease is the most consistently identified abnormality. Many patients with liver disease are, however, asymptomatic-even those with hepatic fibrosis.
A purely phenotypic case definition of hemochromatosis seems to combine apples and oranges (HFE-related and nonHFE-related iron overload) and it defeats the original purpose of having a case definition: to characterize the disease prognosis and natural history and, most importantly, guide its management.
Following the discovery of the HFE gene, a simple genetic test became available for the diagnosis of hemochromatosis. In tertiary referral centers for hemochromatosis in Canada, Australia and France, over 95% of typical hemochromatosis patients were homozygous for 845G·A (C282Y). By contrast, reports continue to be published that suggest only 60–80% of patients with hemochromatosis are homozygous for 845G·A (C282Y); however, the phenotypic case definition for hemochromatosis in these studies might not have been as stringent, and family studies are not often performed. The use of the genetic test as a 'gold standard' test for hemochromatosis has not, therefore, been widely accepted. Furthermore, it has been recognized that approximately 20% of male and 50% of female 845G·A (C282Y) homozygotes have normal serum ferritin levels, making iron overload in these individuals an unlikely possibility. If iron overload is the cornerstone of the diagnosis, how can these patients be classified as having hemochromatosis? There was also concern that the identification of a nonexpressing 845G·A (C282Y) homozygote could lead to anxiety and genetic discrimination, but this has not been verified in large international studies.
Previous studies have suggested that the classification of hemochromatosis should be based on genetic testing for HFE, ferroportin, transferrin receptor 2, hemojuvelin and hepcidin. Genetic testing for nonHFE-related iron overload might never be broadly available clinically because of the rarity of these diseases. Furthermore, genetic sequencing will often be required for diagnosis in many cases because of the heterogeneity of the mutations involved. For most clinicians there are only two types of hemochromatosis: typical 845G·A-related (C282Y-related) and atypical. The term nonHFE-related iron overload is preferred to the term nonHFE hemochromatosis.
I favor the case definition for hemochromatosis as being 845G·A (C282Y) homozygosity with an elevated serum ferritin level. An 845G·A (C282Y) homozygote with normal iron tests has a genetic predisposition to develop iron overload-related symptoms. A compound heterozygote (845G·A [C282Y]/187C·G [H63D]) or 187C·G (H63D) homozygote with elevated serum ferritin and no other risk factors probably has HFE-related iron overload. As patients with these genotypes usually have much milder degrees of iron overload, they should be considered separately in clinical studies. A disadvantage of a mixed genotypic/phenotypic case definition is that it will not be suitable for studies comparing the effect of genotype on phenotype. Until rapid total genome sequencing is introduced into clinical practice, a small category of patients will remain with nonHFE-related familial iron overload that cannot be classified without substantial genetic sequencing efforts. Most experts in the area agree to disagree on a case definition for hemochromatosis. It is most important for individual studies to have precise case definitions. Will a standard case definition be developed by the 100 anniversary of John Sheldon's landmark monograph, in 2034? Definitely maybe.
CLICK HERE for subscription information about this journal.
Although hemochromatosis was first described in the 19 century, John Sheldon's 1934 description of 300 cases remains a landmark publication in defining the clinical characteristics of the disease. Since that time there have been many advances, including the discovery of the hemochromatosis (HFE) gene in 1996, mutations of which (e.g. 845G·A [C282Y]) are found in the majority of patients with typical hemochromatosis. Expert opinion is, however, varied regarding what constitutes the ideal case definition for hemochromatosis. This has led to confusion amongst clinicians and patients.
Why do we need a case definition for any medical condition? Has the subclassification of any group of clinically similar diseases improved life for the patient with a single subclass of those diseases? A precise case definition is useful to study the natural history of a disease, direct appropriate treatment, and judge response to therapy. Furthermore, a consistent case definition is particularly important when studies are compared and outcomes are combined, such as in a meta-analysis.
Most of the debate regarding hemochromatosis centers on its phenotypic or genotypic diagnosis. Whether a clinically asymptomatic patient with elevated iron levels should be classified as having hemochromatosis is currently an issue of debate. Some investigators have proposed a case definition for hemochromatosis comprising a sustained elevation in serum ferritin levels (>200 µg/l in women and >300 µg/l in men) and transferrin saturation (>45% in women and >50% in men) in the absence of any other risk factors for iron overload. This approach has the advantage of simplicity, but it would probably identify a number of patients who do not have iron overload. A screening study demonstrated that a transferrin saturation of >55% had a positive predictive value of only 4.4% for identifying male 845G·A (C282Y) homozygotes.
Other researchers insist that iron overload should be documented and quantified by liver biopsy. The pattern of iron distribution might be helpful in defining the cause of disease. The hepatic iron index (hepatic iron concentration divided by patient age) was devised before genetic testing, and originally used to differentiate patients with alcoholic siderosis from those with genetic hemochromatosis. Later it was extrapolated for use in the case definition of hemochromatosis. This introduced the concept that a quantitative measure of iron overload should be used to define the disease. Subsequent studies using genetic testing, however, have shown that many patients with early hemochromatosis do not have an elevated hepatic iron index (>1.9), and that patients with other types of iron overload do have an elevated hepatic iron index. Liver biopsy is unappealing to patients with mild abnormalities (serum ferritin levels <1,000 µg/l), and MRI has a poor sensitivity for the detection of mild iron overload.
Quantitative phlebotomy is another method to measure iron overload. The removal of 5 g of mobilizable iron by phlebotomy (approximately 20 × 500 ml phlebotomies) followed by a fall in serum ferritin levels to under 50 µg/l has been considered as a case definition.
The creation of arbitrary quantitative thresholds for any of these variables (serum ferritin levels, transferrin saturation, liver iron, iron removed) in order to define hemochromatosis has obvious limitations. A patient just below a threshold might have a clinical diagnosis as strong as another patient just above a threshold.
Another perspective on the case definition for hemochromatosis is that patients should have a constellation of clinical signs and symptoms. The association of any symptoms of hemochromatosis with elevated iron levels or 845G·A (C282Y) homozygosity has, however, come under increased scrutiny. Patients with bronze diabetes (a type of diabetes associated with hemochromatosis) probably represent <1% of all hemochromatosis patients, and studies have shown that the prevalence of diabetes is similar in both 845G·A (C282Y) homozygous and control populations. Other symptoms associated with hemochromatosis, such as fatigue and arthralgia, are nonspecific and very common in the general population. Of all of the putative symptoms of hemochromatosis, liver disease is the most consistently identified abnormality. Many patients with liver disease are, however, asymptomatic-even those with hepatic fibrosis.
A purely phenotypic case definition of hemochromatosis seems to combine apples and oranges (HFE-related and nonHFE-related iron overload) and it defeats the original purpose of having a case definition: to characterize the disease prognosis and natural history and, most importantly, guide its management.
Following the discovery of the HFE gene, a simple genetic test became available for the diagnosis of hemochromatosis. In tertiary referral centers for hemochromatosis in Canada, Australia and France, over 95% of typical hemochromatosis patients were homozygous for 845G·A (C282Y). By contrast, reports continue to be published that suggest only 60–80% of patients with hemochromatosis are homozygous for 845G·A (C282Y); however, the phenotypic case definition for hemochromatosis in these studies might not have been as stringent, and family studies are not often performed. The use of the genetic test as a 'gold standard' test for hemochromatosis has not, therefore, been widely accepted. Furthermore, it has been recognized that approximately 20% of male and 50% of female 845G·A (C282Y) homozygotes have normal serum ferritin levels, making iron overload in these individuals an unlikely possibility. If iron overload is the cornerstone of the diagnosis, how can these patients be classified as having hemochromatosis? There was also concern that the identification of a nonexpressing 845G·A (C282Y) homozygote could lead to anxiety and genetic discrimination, but this has not been verified in large international studies.
Previous studies have suggested that the classification of hemochromatosis should be based on genetic testing for HFE, ferroportin, transferrin receptor 2, hemojuvelin and hepcidin. Genetic testing for nonHFE-related iron overload might never be broadly available clinically because of the rarity of these diseases. Furthermore, genetic sequencing will often be required for diagnosis in many cases because of the heterogeneity of the mutations involved. For most clinicians there are only two types of hemochromatosis: typical 845G·A-related (C282Y-related) and atypical. The term nonHFE-related iron overload is preferred to the term nonHFE hemochromatosis.
I favor the case definition for hemochromatosis as being 845G·A (C282Y) homozygosity with an elevated serum ferritin level. An 845G·A (C282Y) homozygote with normal iron tests has a genetic predisposition to develop iron overload-related symptoms. A compound heterozygote (845G·A [C282Y]/187C·G [H63D]) or 187C·G (H63D) homozygote with elevated serum ferritin and no other risk factors probably has HFE-related iron overload. As patients with these genotypes usually have much milder degrees of iron overload, they should be considered separately in clinical studies. A disadvantage of a mixed genotypic/phenotypic case definition is that it will not be suitable for studies comparing the effect of genotype on phenotype. Until rapid total genome sequencing is introduced into clinical practice, a small category of patients will remain with nonHFE-related familial iron overload that cannot be classified without substantial genetic sequencing efforts. Most experts in the area agree to disagree on a case definition for hemochromatosis. It is most important for individual studies to have precise case definitions. Will a standard case definition be developed by the 100 anniversary of John Sheldon's landmark monograph, in 2034? Definitely maybe.
CLICK HERE for subscription information about this journal.
SHARE
