Data Synthesis
Key Question 1. What Is the Risk for Developing
Clinical Hemochromatosis among Those with a
Homozygous C282Y Genotype?
Of 134 full-text studies examined, we excluded 120
studies for reasons specified by our inclusion and exclusion
criteria (Appendix Table 4). We eliminated all studies that
combined outcome measures for C282Y homozygotes
from more than one population source (for example, from
family, clinical, or healthy population screening) since disease
expression potentially differs among these groups. We
eliminated studies that did not report data on morbid conditions
associated with clinical hemochromatosis (or at
least iron overload) among participants.
We had 2 other
main categories for study exclusion:
- Studies that involved groups of homozygotes that did not derive from any definable population—particularly one that could be subject to screening.
- Studies with data reported in ways that did not conform to our hemochromatosis-related definitions.
One study was identified, but not yet published,
at the time we prepared this manuscript (Appendix Table 11). Two studies supplied
data that did not meet requirements for our final data
synthesis;69,70 3 studies on genotyping in blood donors
71-73 were not relevant to this paper but are included in
our full evidence report.74
Table 2 summarizes the findings for this key question.
The best evidence is from 2 fair- to good-quality longitudinal
studies reporting the risk for developing disease in
initially nondiseased C282Y homozygotes.46,47 Although
neither was done in an inception cohort, these retrospective
cohort studies from Australia.46 and Denmark
47 reported on disease expression (penetrance) of 33
C282Y homozygotes (22 women and 11 men) over 17 to
25 years of followup. Participants' average age at the end
of observation was 47 to 63 years. Most, but not all,
C282Y homozygotes (61% to 75%) developed some elevations
in serum iron measures during followup. When
compared with other age- and sex-matched genotypes,
C282Y homozygotes tended to have higher mean transferrin
saturation and serum ferritin levels, and average measures
generally increased with age among all genotypes
47 However, C282Y homozygotes also showed more individual
variation in serum iron measures than other genotypes,
and many individuals did not show steady increases
in these measures over time.46,47
For example, neither
blood loss nor donation explained the substantial decreases
in serum ferritin levels over 17 years seen in 2 of 10 C282Y
homozygotes.46 The Australian study.46 objectively
evaluated iron overload using liver biopsy in the 6 of 10
participants who developed serum ferritin levels greater
than 500 µg/L. At least moderate iron overload (select
for definition) was detected in 5 patients who underwent
biopsy (representing 5 of 10 total study participants).
Two of the patients who underwent biopsy had hepatic fibrosis, while the single patient with cirrhosis reported
alcohol intake greater than 6 drinks per day. In
contrast, none of the 23 Danish patients had liver disease
detectable by clinical examination.47 Thus, when both
studies were considered together, liver disease developed in
3 of 33 C282Y homozygotes. Similarly, 2 of 33 C282Y
homozygotes developed diabetes and 6 of 33 developed
arthralgias. No participant developed cardiomyopathy or
hypogonadism.
These retrospective cohort studies have 2 potential
limitations. The first limitation relates to whether these
data accurately represent lifelong disease expression in
C282Y homozygotes. Despite the long followup period of
17 to 25 years, 8 women were 50 years of age or younger at
final followup. Thus, 8 of 33 (24%) of those studied may
not yet have reached the age at which clinical expression
would be likely.
Second, selective mortality bias resulting
from followup only for survivors could have influenced
these findings to represent the experience of healthier
C282Y homozygotes. In the Australian study, however, the
prevalence of C282Y homozygotes (5.3 per 1000) was
within the population range expected, and complete data
were available on 83% of the cohort.46 In the Danish
study, selective mortality bias may be more likely since
35% of the original cohort did not have genotyping and 3
of the 23 C282Y homozygotes died before they could be
examined.47 We calculated the upper bound for disease
penetrance as follows to determine the potential impact of selective mortality bias on this study. If all 3 C282Y homozygotes
who died were counted as developing hemochromatosis,
the proportion developing clinical disease
would still be about one quarter (4 of 23). If the 35% of
the cohort lost to followup had the usual population prevalence
of C282Y homozygosity (5 per 1000), then about
25 C282Y homozygotes would have been lost to followup.
If all 25 homozygotes developed clinical disease, the
estimate for disease penetrance would be 60% (29 of 48)
after 25 years of followup.
While cross-sectional studies were more plentiful, they
provided an estimate of disease expression only at the time
of genotype identification. Twelve papers.32,48-58 report
cross-sectional genotypic and selected phenotypic and
disease expression results from 9 screening studies (Appendix Table 8).
C282Y homozygotes were identified at 2
health clinics.32,48-51 through mass screening,52
through voter rolls or employment screening,53-56 or
through family screening.57,58 We combined health
clinics, mass screening, voter rolls, and employment
screening results to represent "general population" screening
based on the similarity of findings for C282Y prevalence
and phenotypic expression between settings. A total
of 282 C282Y homozygotes were identified from screening
67,771 patients in these general population settings, and
426 C282Y homozygotes were identified from genotyping
in an unspecified number of family members of probands.
The prevalence of C282Y homozygosity was 4.2 per 1000
screened in the general population and 161 per 1000 family
members screened (based on the single family screening
study that reported the number of family members
screened).57 Transferrin saturation levels were elevated
in 75% or more of male C282Y homozygotes identified
from general population screening, and the majority (58%
to 76%) had elevated serum ferritin levels. Elevations of
transferrin saturation and serum ferritin levels were more
variable or less common among female homozygotes from
the general population than among male homozygotes.
Transferrin saturation and serum ferritin elevations in family
members were very common (88% to 96%).
Among C282Y homozygotes identified from general
population genetic screening, 38% of those undergoing
further evaluation met criteria for iron overload, 25% had
liver fibrosis, and 6% had cirrhosis. Data could not be
reported reliably for males and females separately. These
iron overload and disease estimates could be too high if the
C282Y homozygotes who were not evaluated further are
less likely to be penetrant. Assuming that all the untested
C282Y homozygotes were unaffected, the prevalence of
iron overload, hepatic fibrosis, and cirrhosis among newly
screening-identified C282Y homozygotes would be 24%,
6%, and 1.4%, respectively. These estimates, however,
should be viewed with caution because they are based on
very small numbers. We also cannot be sure of the likelihood
of disease penetrance (same, higher, or lower) in the
large proportion of untested screening-identified C282Y
homozygotes.
Data from genotyping of family members of probands
may indicate that a higher proportion of C282Y homozygotes'
relatives have evidence of iron overload, but not necessarily
of clinical disease, at the time of screening compared
with homozygotes identified through population
screening. Among male first-degree relatives, 74% of those
further evaluated have iron overload, 23% have fibrosis,
and 6% have cirrhosis. Among female first-degree relatives,
62% of those further evaluated have iron overload, 4%
have fibrosis, and 3% have cirrhosis. If we assume that all
those not further tested were unaffected, estimates of the
prevalence of iron overload, fibrosis, and cirrhosis in male
C282Y homozygotes identified through family screening
are 41%, 13%, and 4%. The respective prevalences for
females are 23%, 2%, and 1%. Iron overload and disease
expression at the time of identification were reported only
for the limited number of C282Y homozygotes undergoing
further evaluation for clinical reasons. Not all studies reported
these measures and, within studies, variably selected
participants received disease evaluations because of differences
in the participants' clinical presentation, in their willingness
to be tested, and in clinical practice norms. Estimates
across studies cannot be easily compared because of
potential detection bias and likely between-group differences
in important factors in penetrance (such as age and
sex) between C282Y homozygotes, particularly those identified
from general population screening compared with
those identified through family screening.
Key Question 2. Does Earlier Therapeutic Phlebotomy of
Individuals with Primary Iron Overload Due to
Hereditary Hemochromatosis Reduce Morbidity and
Mortality Compared with Treatment after Diagnosis in
Routine Clinical Care?
We found no controlled studies of phlebotomy treatment
in patients with hemochromatosis due to any cause,
nor any studies that allowed a valid comparison of early
versus delayed treatment. Four fair-quality case series of
patients with hemochromatosis reported objective measures
before and after, or simply after, treatment.25,58-61 in 7 publications.22,23,25,58-60,75 One retrospective
observational survey.76 reported recalls of
changes in symptoms after treatment among patients with
hemochromatosis identified through multiple outreach
mechanisms (Appendix Table 9). We excluded 61 full-text
articles, primarily because of study quality, small size (<20
patients), or lack of primary data or relevant outcomes
(Appendix Table 5).
Table 3 summarizes the findings for this key question.
Altogether, treatment studies of patients from referral centers,
who were identified and treated over a 50-year period,
report on the survival experience of 447 patients over a
mean duration of 8.1 (SD, 6.8) to 14.1 (SD, 6.8) years,
and the reduction in morbidity after treatment of 370 patients with hemochromatosis.25,58-60 Only 105 of
these patients had genetically confirmed hereditary hemochromatosis,25,58 and, of these, source of detection
(clinical detection or family screening) was available for 85
patients (56% were probands and 44% were family members).25 Fewer patients with confirmed hereditary
hemochromatosis had cirrhosis at diagnosis (3.4%58 to
32%25), compared with reports from patients whose
condition was not genetically confirmed (57%60 to 79%59); these findings are consistent with strong secular
trends in disease severity at diagnosis.60 Secular trends
in survival were also apparent, since survival improved over
10 years of followup in patients in whom hemochromatosis
was diagnosed in 1982 to 1991, compared with 2
groups who received the diagnosis earlier (P < 0.05, log-rank
test).60 For patients whose hemochromatosis was
diagnosed during this later time (1982 to 1991), cumulative
survival was not significantly reduced from rates
expected for an age- and sex-matched population.60
Similarly, patients with genetically confirmed hemochromatosis
who did not have cirrhosis at diagnosis experienced
the same survival as population controls.25
Among treated patients with hereditary hemochromatosis,
cirrhosis at diagnosis appeared to confer a worse
prognosis (adjusted relative risk for death, 5.54 [CI, 1.76
to 17.47]).25 However, comparisons of survival differences
between cirrhotic and noncirrhotic patients, between
other patient subgroups (for example, diabetic vs. nondiabetic
patients60 or between all patients and historical
controls59) are not completely reliable because of potential
confounding by uncontrolled and unmeasured factors,
such as era of diagnosis, age at diagnosis, sex, excessive
alcohol use, concomitant hepatitis, and dietary factors.
In the best available evidence on the effects of phlebotomy
treatment, pretreatment and post-treatment liver
biopsies in 260 patients who received a diagnosis through
routine clinical practice suggest some reversibility of hepatic
disease, with 7% to 23% showing improvement and
1% to 3% showing worsening.59,60 Improvement in
histologic characteristics was more common (32.6%) in
patients with less severe, precirrhotic liver disease than in
patients with cirrhosis (14.8% improved).60 In a highly
selected subgroup of family (and health check) screening-detected
patients (n = 25) who underwent a second biopsy
after treatment for persistently elevated liver enzyme levels
or uncertainty about cirrhosis on first biopsy, 19 of 20
showed improvement in hepatic fibrosis scores after treatment;
the only case with baseline cirrhosis was unchanged.58 These findings are not clearly generalizable because of
the selected nature of the patient group and because biopsy
results in 5 cases with high alcohol intake were not reported.
Several studies suggest that some, but not all, other
disease process and symptoms will respond to phlebotomy
treatment. In 183 primarily male symptomatic patients
(57% of whom had cirrhosis) who received a diagnosis
before 1991, 41% of those with type 1 diabetes mellitus
reduced their daily dosage; 73% with elevated levels of liver
enzymes (alanine aminotransferase or aspartate aminotransferase)
showed improvement; and symptoms such as
weakness, lethargy, or abdominal pain improved in more
than half.60 Improvements in arthralgias (30%) and potency
(19%) were less prominent. A total of 2851 primarily
male patients with hemochromatosis, most of whom
received a diagnosis after 1990 through family screening or
an abnormal laboratory test finding, were asked to recall
their experience before and after treatment. They reported
comparable improvements in extreme fatigue (50%), abdominal
pain (22%), impotence (13%), and joint pain
(9%). Many patients also recalled improvement in depression
(41%), but many (33%) also recalled onset of new
symptoms after treatment.76 This study is weakened by
its reliance on recall and the absence of controls to compare
nonspecific symptom prevalence and changes over
time.
Key Question 3. Are There Groups at Increased Risk for
Developing Hereditary Hemochromatosis That Can Be
Readily Identified before Genetic Screening?
We examined 55 full-text articles and excluded 47
studies from this question for various reasons (Appendix Table 6), such as not reporting relevant measures or results,
addressing the wrong population, not using C282Y genotype
to define the family risk group, using an ineligible study
design, or having poor quality. One fair- to good-quality
cross-sectional study of family members of genotyped probands
57 and 6 fair- to good-quality cross-sectional studies
(in 7 publications).51,61-66 of patients with signs or
symptoms consistent with iron overload or hemochromatosis
met our inclusion criteria.
Table 4 summarizes the findings for this key question.
Potential high-risk groups were examined for a higher
prevalence of C282Y homozygosity, including 150 family
members of probands and 42,636 patients with fatigue or
increased liver enzyme levels from primary care or hepatology,
endocrinology, and rheumatology specialty settings.
Family screening identified the highest prevalence of undetected
C282Y homozygotes (23% overall), particularly
among siblings of probands (33% homozygosity). Among
symptomatic patients selected from primary care, rheumatology,
endocrinology, or referral medicine clinics, 0% to
5.8% were C282Y homozygotes, compared with 0.2% of a
random sample of persons attending a health appraisal
clinic.27
Overall, the prevalence of C282Y homozygosity
did not differ between patients in the health appraisal
clinic and primary care patients with an index sign or
symptom. Compared with controls, C282Y homozygosity
was significantly more prevalent only in hospitalized diabetic
patients from an endocrinology clinic (5.8%) and in
patients from a referral medicine clinic with chronic fatigue
and arthralgias (5.7%). Three other studies confirm or extend
these results. Males, but not females, with chronic fatigue symptoms visiting a health appraisal clinic had a
slightly higher (0.85%) prevalence of C282Y homozygosity
than patients without symptoms (0.14%).51 The prevalence
of C282Y homozygosity in patients from a rheumatology
clinic was similar to that in the general population.65 In patients with a history of coronary heart disease,
prevalence of C282Y homozygosity was the same as, or
lower than, that of patients without symptoms (0.17% to
0.28%).62 Findings may not be conclusive in comparisons
based on fewer than 300 patients, given the population
prevalence of C282Y homozygotes (3 to 5 per 1000
white persons).
Some studies restricted genotyping to symptomatic patients
who also had some laboratory abnormality. The
prevalence of C282Y homozygosity was somewhat increased
in a range of patients with hemochromatosis-compatible
signs and symptoms and elevated iron measures
(Table 4. Among 667 patients from a liver clinic who had
elevated iron measures, 7.1% were homozygous for C282Y.63 For hospitalized patients with diabetes and patients
with chronic fatigue or arthralgias who were referred to
specialists, C282Y homozygosity was higher in patients
with transferrin saturation greater than 0.40 or serum ferritin
level greater than 300 µg/L than in patients with
disease but without elevated iron measures (6.6% to 17.3%
compared with 5.7% to 5.8%).61 The sensitivity of
transferrin saturation greater than 0.40 for detecting
C282Y homozygosity in diabetic patients hospitalized for
disease-related complications was 100%, but the specificity
was 13%. In diabetic patients, the sensitivity of a serum
ferritin level greater than 300 µg/L was 86% and the specificity
was 56%. For patients referred for arthralgias and
unexplained fatigue, transferrin saturation greater than
0.40 and a serum ferritin level greater than 300 µg/L were
about equally sensitive and specific for C282Y homozygosity
(100% sensitive and 65% to 67% specific). In patients
from a health appraisal clinic who had elevated liver enzyme
levels, the prevalence of C282Y homozygosity appeared
the same (in women), or slightly higher (0.57% vs.
0.28%, in men), compared with those with normal enzyme
levels.51
Return to Contents
Discussion
We have data on the risk for developing signs or symptoms
of iron overload and hemochromatosis in 33 C282Y homozygote adults monitored over 17 to 25 years and on
the burden of disease at the time of identification for an
additional 228 newly identified C282Y homozygote adults
from the general population. Taken together, these data
suggest that up to 38% to 50% of C282Y homozygotes
develop iron overload according to our criteria and up to
10% to 33% develop definite disease (fibrosis, cirrhosis, or
diabetes). Much lower estimates are also compatible with
available data. Findings from a large case series on the
disease expression of 271 patients with hereditary hemochromatosis
identified through genetic testing of those
with elevated serum iron levels detected at health appraisal
screening complement our review.58 Although these patients'
disease expression would represent only C282Y homozygotes
already exhibiting iron accumulation by definition,
rates of cirrhosis (6.3%), fibrosis (10.7%), diabetes
(3.6%), or any combination of these (20.6%) were similar
to or marginally higher than limited results from general
population screening found in our review.
Available data
remain too limited to clearly establish estimates of disease
penetrance, since so few people have been studied in depth
(only 10 C282Y homozygotes were evaluated per our criteria
for iron overload or hemochromatosis in longitudinal
studies), and in those studied over time, disease could still
develop with longer followup. Indeed, 8 of 33 of those
followed longitudinally were women age 50 years or
younger at last followup, in whom disease may not have
yet developed. Also, while a higher proportion clearly develop
iron overload, its clinical significance is less clear than
that of clinical hemochromatosis. Finally, data reported
here (and elsewhere) clearly articulate that a subgroup of
untreated homozygotes—perhaps even 40% 58—do not
exhibit any or progressive iron accumulation over years of
followup, thus complicating any message that would be
given to asymptomatic screening-detected individuals.
Family members of individuals with hereditary hemochromatosis
are noted to be at higher risk for being homozygous,
and family screening has been established as a
standard of care based on HLA-typing studies of family
members of probands.77,78 We found 1 U.S. study and
1 Australian study using HFE genotyping to determine risk
in probands and family members that support this practice.
A high proportion of tested biological relatives (23%) were
also C282Y homozygotes. Similarly, compared with the
general population, a higher proportion (49% to 86%) of
C282Y homozygotes identified from family screening met
iron overload criteria, although the proportion with fibrosis
and cirrhosis did not clearly differ. Direct comparisons
in disease penetrance between these different types of
screening-detected C282Y homozygotes have very limited
value, however, because these groups may differ with regard
to who receives more in-depth clinical work-up (selection
bias), as well as other ways important to disease
expression. For example, a recently published study reporting
on C282Y homozygous persons identified over many
years through family screening and through phenotypic
followed by genotypic screening found significant differences
in baseline characteristics between the 2 groups that
could affect disease expression.58 In addition, even if it
is considered the standard of care, approaches to family
screening also need to consider other associated ethical,
legal, social, and psychological issues.78
Studies examining survival are limited to 4 case series
reporting on a total of 447 patients who received a diagnosis
between 1937 and 1989. Disease severity at diagnosis
and survival showed pronounced secular trends. Patients
with a more recent diagnosis are less severely affected, and
with treatment they have 10-year survival rates similar to
those of age- and sex-matched controls. These trends may
be due to earlier diagnosis from increased clinical suspicion
or enhanced family screening due to recognition of hemochromatosis
as a hereditary disease leading to earlier diagnosis,
or to increases in adequate treatment after diagnosis.
Liver biopsies before and after treatment suggest arresting
disease progression in most individuals and a possible
reduction in the severity of hepatic fibrosis, particularly
in less severely affected patients. Available data are
consistent with improvements in some, but not all, hemochromatosis-related morbid conditions after treatment.
None of these data come from controlled trials, however,
and studies do not generally ensure minimally valid measures
of treatment response. No studies reported harms,
limiting the ability to determine net risks and benefits of
treatment.
Given these caveats, treatment may result in
reduced insulin doses in patients with type 1 diabetes and
decreases in elevated liver enzyme levels. Symptoms such as
extreme fatigue, abdominal pain, and lethargy improve in
most patients, while arthralgia and impotence do not.
Some have suggested a targeted approach to screening
by identifying persons with signs or symptoms consistent
with undiagnosed, early-stage hemochromatosis. Primary
care patients selected for symptoms or signs consistent with
hemochromatosis did not have a higher prevalence of
C282Y homozygosity than healthy controls, and neither
did selected symptomatic or diseased patients from rheumatology
or other specialty clinics. A slightly higher proportion
of C282Y homozygotes could be identified by conducting
genotyping only in patients from a liver clinic
prescreened to have transferrin saturation greater than 0.45
(7.7% C282Y/C282Y) or by targeting diabetic patients
hospitalized for poor control or complications (5.5%) or
patients referred to specialists for chronic fatigue and arthralgias
(5.7%). While biochemical screening with transferrin
saturation and serum ferritin further enriched this
patient pool, calculated specificity remained low (56% to
67%).
Overall Evidence
The quantity of evidence that met quality and relevance
criteria for the focused key questions posed by this
review was small, despite a very large published literature
(Table 5). A great deal was published before the
availability of HFE genotyping for hereditary hemochromatosis. After
reviewing 1,886 abstracts and 256 full-text articles, we located
only 23 fair- to good-quality studies that were relevant
to some aspect of our 3 key questions on disease
burden, benefits of early treatment, and high-risk groups.
Some articles cited to support screening and treatment
benefits in this field did not meet minimal quality or diagnostic
criteria for our review, as was true of often-cited data
within the studies we could include. All the reviewed evidence,
including treatment studies, was observational,
much of it representing the experience of a small number
of relatively selected individuals, and much of it without
data to allow comparisons with an unaffected or an untreated
population. The published research was often difficult to interpret consistently
and accurately given incompleteness and extreme variability in reporting standards.
While more recent reports are of higher quality with clearer
case definitions, authors still fail to acknowledge the impact
that selection bias probably has on their estimates of
disease expression in C282Y homozygotes; thus, the applicability
of their findings to the evaluation of general population
screening is limited.58
In reviewing this field, others have included a larger
range of study designs, such as modeling the expected frequency
of genotyping in older populations, autopsy studies,
and other circumstantial approaches. Our focused key
questions did not allow incorporation of this type of evidence
into our review, but it is unlikely that their inclusion
would be of great use to the USPSTF given its evidence
hierarchy and requirement of at least fair-quality evidence
for making its recommendations.67
Limitations
The articles we included required substantial interpretation
for data abstraction and synthesis. For individual
articles, we typically reviewed all tables for possibly relevant
data and checked text calculations. We made every effort to
report data only on adult populations relevant to screening,
which required careful reading and data dissection in studies
that combined cases from many sources. We excluded
studies with serious discrepancies or those in which outcomes
could not be related back to a sample or population
source we were addressing. Many articles required further
hand calculations to extract data in the most comparable form in order to allow cross-study comparisons, and inconsistencies
between tables and text in many articles complicated
this process. The number of calculations and interpretation
from descriptive data raise a concern about data
errors. Overall, the difficulties in understanding and interpreting
this literature posed challenges to meeting our
usual standards of comprehensiveness and consistency.
We primarily focused on hereditary hemochromatosis
as the condition of interest for this screening review and,
within that, on the most common associated HFE genotype
in the United States (C282Y homozygosity), which
accounts for 85% to 90% of cases in white persons. We
did not examine other hereditary causes or the impact of
HFE heterozygosity that may account for 3% to 5% of
patients with hereditary hemochromatosis. While we did
not review evidence on phenotypic screening in primary
care, others have recently done so,79 and the evidence
has been found insufficient for phenotypic screening for
hereditary hemochromatosis in the general population.80
Return to Contents
Conclusions
On the basis of this focused evidence review, research
regarding screening for hereditary hemochromatosis remains
very limited. Despite the availability of new studies
in response to calls for improved research,18,40,81 not
enough is known to allow a confident projection of the
benefit from widespread genotypic screening for hereditary
hemochromatosis. Data are beginning to be reported for
targeted high-risk population screening approaches (for example,
high-risk identification followed by phenotypic
screening followed by genotypic screening), which may
prove to be useful.
Recent studies suggest that disease expression or penetrance
is certainly less than 100% in C282Y homozygotes
identified through some method of screening. How much
less than 100%, and for whom, remains uncertain. In the
next year or two, the HEIRS followup should provide
information on short-term disease expression based on
clinical examinations of C282Y homozygotes; those with
elevated iron measures at the time of screening, regardless
of genotype; and a sample of controls. However, only self-reported
disease expression data will be available on all
99,000 (genotyped and phenotyped) primary care patients,
and followup beyond 1 to 2 years is not planned. If funding
is provided, this study could be a rich resource of prospective
information on disease development, as well as
observational data on treatment response in contemporarily
diagnosed patients with clear disease definition. Without
other data, such as might come from the HEIRS study,
the literature on treatment remains quite small, consisting
of dated case series in fewer than 500 patients (few of
whom have hereditary hemochromatosis documented by
genotype). Controlled treatment trials will probably never
be undertaken for ethical reasons, so higher-quality observational
treatment data would be very useful.
The literature on genotyping family members of
C282Y/C282Y probands is also of limited quantity because
of the relatively recent availability of HFE testing
(1996), but there is a large body of HLA-based literature
on which family screening of probands has been established.
Research needs in this area remain high.79
From Oregon Evidence-based Practice Center, Center for Health Research,
Kaiser Permanente, Portland, Oregon.
Return to Contents
Proceed to Next Section
540 Gaither Road Rockville, MD 20850