Soluble Transferrin Receptors and Iron Deficiency, a Step
beyond Ferritin. A Systematic Review
Anastasios Koulaouzidis1, Elmuhtady Said2, Russell Cottier3, Athar A Saeed4
1) Centre for Liver and Digestive Disorders, Royal Infirmary of Edinburgh, Medical Day Case & Endoscopy Unit,
Edinburgh; 2) Barnsley General Hospital, Gawber Road, Barnsley; 3) Warrington General Hospital, Lovely Lane,
Warrington, Cheshire; 4) Queen Elizabeth Hospital, Gateshead, Tyne & Wear, UK
Abstract
Iron deficiency is a common disorder; it can also be the
first indicator of significant gastrointestinal pathology. Total
iron stores are evaluated by ferritin measurement, but this
is often difficult, as coexistent disease can obscure ferritin
results. Bone marrow (BM) examination was previously
felt to be superior to all known serological markers of iron
status, but it has a number of disadvantages. The validity
of measurement of soluble transferrin receptors (sTfR) as a
surrogate marker of BM iron stores has been the subject of
various studies so far. Aim: To critically review the use of
sTfR as a marker for the evaluation of iron stores. Methods: A
systematic computerised literature search, in order to identify
studies that compared sTfR measurement against BM iron
stain. Results: Twenty prospective studies were identified, of
which nine fulfilled the inclusion criteria (sTfR measurement
in anaemic adults alongside BM iron staining). For a total
of 979 patients, a different sTfR reference range was used,
but levels of 2.5mg/L (29.5nmol/L) were consistently used
as the threshold for iron-deficiency anaemia resulting in
good specificity and sensitivity. Conclusion: The use of
sTfR improves the clinical diagnosis of iron deficiency
anaemia, especially in the presence of coexisting chronic
disease or gastrointestinal malignancies. The safety and costeffectiveness of a ferritin/sTfR-based approach to exclude
gastrointestinal cancer in the presence of iron deficiency
has to be proven with a prospective, well standardised,
multicentre trial or a meta-analysis.
Keywords
Iron deficiency – bone marrow – soluble transferrin
receptors – ferritin – endoscopy.
Received: 22.03.2009Accepted: 16.05.2009
J Gastrointestin Liver Dis
September 2009 Vol.18 No 3, 345-352
Address for correspondence:
A Koulaouzidis MD, MRCP(UK)
Centre for Liver and Digestive Disorders
Royal Infirmary of Edinburgh
Edinburgh, EH16 4SA, UK
Email: [email protected]
Introduction
Iron deficiency (ID) is a common disorder, affecting
almost 1.2 billion people worldwide [1]. In 1985, the
World Health Organization reported that 15% to 20% of the
world’s population had iron deficiency anaemia (IDA) [2].
In developed countries more than 500 million people are
iron deficient [3-5]. Iron deficiency anemia is a particularly
important problem for men and postmenopausal women, as it
can be the first presenting feature of an occult gastrointestinal
malignancy [6-8].
It is important to remember that for most clinicians
ID is associated with anaemia; the truth is that it is a
continuous process evolving in three stages. The first phase
is the depletion of storage iron (stage I), where total body
iron is decreased but haemoglobin (Hb) synthesis and red
cell indices remain unaffected. Both these indexes change
when the supply of iron to bone marrow (BM) becomes
problematic (iron deficient erythropoiesis – IDE, or stage II).
In stage III the iron supply is insufficient to maintain a normal
Hb concentration, and eventually IDA develops [9].
Evaluation of iron deficiency
Evaluation of iron deficiency is performed by various
methods. Red cell indices [mean corpuscular volume (MCV)
and mean corpuscular haemoglobin (MCHC)] may confirm
a microcytic, hypochromic picture, with rerum ferritin
often low. Unfortunately both these tests are less reliable
with coexistent illness. Anaemia of chronic disease (ACD)
is a syndrome associated with chronic inflammatory and
autoimmune conditions, malignancies, or infections (acute
or chronic), and causes a state of ‘functional’ ID. Recently,
hepcidin (a liver-produced peptide) was found to be a link
between the immune system and the movement of iron to
and from the storage depots. Hepcidin binds to ferroportin,
a transmembrane iron exporter present on macrophages,
enterocytes and hepatocytes. It has been demonstrated that
in vitro hepcidin induces the internalization and degradation
of ferroportin. By regulating the number of iron exporters,
hepcidin is involved in the control of iron uptake and release
from enterocytes or macrophages. It is conceivable that
346
processes that affect these regulatory mechanisms can result
in defective iron mobilisation without deficiency in the iron
stores per se [10].
The two entities (IDA & ACD) can coexist and so the
detection of ID in the presence of chronic disease can be an
important diagnostic challenge. Efficient management of
IDA requires, on most occasions, not only administration
of iron therapy to replenish iron stores, but also invasive
investigations (i.e. endoscopies) to find and treat an
underlying cause [11-12]. Endoscopies are known to have
potential life-threatening complications [13], whereas
inappropriate administration of iron in ACD may aggravate
the underlying disease.
The British Society of Gastroenterology guidelines
recommend confirmation of ID before embarking on
endoscopic tests [12]. Measurements of peripheral blood
indices (i.e. MCV, MCH and ferritin) can prove inadequate
in certain patient groups because of known confounders
(infection, inflammation or cancer). Haemoglobin (Hb)
determination, the most widespread screening method for
ID, has serious limitations when used as the only laboratory
marker for iron deficiency because of its low specificity and
sensitivity. This is because any individual must loose a large
portion of total body iron for iron stores to decrease and the
Hb to fall below normal levels [14].
Despite that, when hereditary anaemias are excluded,
in the majority of cases the presence of hypochromia and
microcytosis in the face of low ferritin makes the diagnosis of
ID fairly straightforward [15]. The difficulty is in interpreting
low markers in isolation, particularly when the Hb is normal
[16]. The work of Jacobs, Walters and Bentley established
serum ferritin as an accurate indicator of storage iron [17-21].
Numerous studies since have demonstrated the superiority
of ferritin over other measurements in identifying ID, and
thus ferritin has been the most useful laboratory marker over
the past 30 years [15].
In their review of 55 studies, Guyatt et al. reported that
the mean area of the receiver-operator characteristic (ROC)
curves for ferritin was 0.95 ± 0.1, while that of MCV was only
0.76 [22]. A ferritin level of ≤12μg/L is considered a highly
specific indicator of ID (98% specificity). The sensitivity
for this is only 25% [23, 24], and can be improved to 92%
if a cut-off ferritin level of 30μg/L is used for ID diagnosis,
thus resulting in a positive predictive value (PPV) of 92%
[24, 25]. However it is known that ferritin is an acute phase
reactant and can be elevated irrespective of iron status in
patients with concurrent chronic inflammatory or infectious
diseases, recurrent acute infections and malignancies [26].
The ‘gold standard’
The examination of bone marrow (BM) aspirate,
stained with Prussian blue for iron, is still considered as
the best method for evaluating iron status in patients with
indeterminate laboratory findings. Its drawbacks are that it is
invasive, expensive, and that results are operator dependant.
Barron et al. studied a cohort of 108 consecutive unselected
Koulaouzidis et al
patients in whom the BM examinations were reported as iron
deficient, and found that reports of absent BM haemosiderin
may be inaccurate in more than 30% of cases – even when
accurate, this may not necessarily signify the presence of ID
[27]. Measurement of serum ferritin levels is still required
to confirm a clinical diagnosis. Furthermore, up to 35% of
BM aspirate may be inadequate for interpretation [20]. In
a small retrospective study, Ganti et al. concluded that the
PPV of absent BM iron stores for the diagnosis of IDA was
only 50% [28]. Another recent study showed that patients
who had stainable iron in their BM had in fact ‘functional’
ID [29].
The invasive nature of BM aspiration and biopsy has
made it tempting to identify a peripheral blood test which
could act as surrogate marker of BM iron stores.
Witte et al. examined a cohort of 97 consecutive anaemic
patients and devised a graph relating serum ferritin with
erythrocyte sedimentation rate (ESR); they suggested that it
could allow accurate confirmation or exclusion of ID [30].
Two years later, they reported that a two-dimensional linear
graphic relationship between ferritin and ESR provided a
PPV of 97%, thus aiding the confirmation or exclusion of ID
in community hospital practice [31]. Others have proposed
that a ferritin level of 70μg/L was the necessary safety limit
for exclusion of IDA [32]. Guyatt et al. showed that 40% of
patients with ferritin levels between 18μg/L and 100μg/L
had ID on BM examination [33].
There are two main categories of laboratory tests for
identifying ID. Screening measurements (Hb, transferrin
saturation, MCV, MCH, zinc protoporphyrin and reticulocytes
Hb) that detect iron deficient erythropoiesis (IDE) by
demonstrating either a reduced supply of serum iron or poor
haemoglobination of red cells and definitive measurements
that evaluate tissue iron status by measuring proteins derived
from either the iron store compartment in macrophages
(ferritin), or the iron utilization compartment in red cell
precursors [34]. Of note, red cell distribution width (RDW)
usually precedes the other markers of IDE [35]. Soluble
transferrin receptors (sTfR) can detect stages II and III of
ID, and are considered as a ‘dual’ index.
Soluble transferrin receptors
The first description of the presence of transferrin
receptors (TfR) on the surface of reticulocytes was made
in 1963 [36]; they were further characterised as cell surface
glycoproteins in 1981 [37]. The overall dynamics of these
receptors was studied using the mouse erythroleukemia K562
cell line [38]. Ward et al. found that in these cells, TfR mRNA
synthesis was increased when iron-deficient serum, or ironchelating agents were added to the culture medium [39, 40].
It was later proposed that, the up-regulation of receptors must
be mediated (at least in part) by the transferrin levels.
The transferrin receptor mediates cellular uptake of iron
by binding the iron carrier-protein transferrin (Tf). Following
internalisation of the iron-transferrin-TfR complex, iron
is released from its binding sites in the acidic milieu of
Soluble transferrin receptors and iron deficiency
the endosomes (acidosomes), and the Tf-TfR complex is
returned back to the cell surface, where apo-transferrin is
released again [41].
The transferrin receptor is comprised of two identical
subunits, each composed of 760 amino-acids with a molecular
mass of 95 kDaltons. Each polypeptide subunit contains a
trans-membrane segment with 21 amino acid residues, an
N-terminal cytoplasmic domain with 61 residues, and a large
C-terminal extracellular domain with 671 amino acids. The
soluble form of TfR in serum was recognised since 1986
[42]. Kohgo et al. described its use as a surrogate marker of
BM erythropoiesis and red cell production [43].
Humans use 80% of their body iron for erythropoiesis,
and almost the same proportion of TfR in the body is found
in erythroid progenitor cells. Reticulocytes entering the
peripheral bloodstream carry a high surface concentration
of the receptor; as the cells mature the receptors are shed
into the circulation [44]. Release of TfR has been shown
to be mediated by an integral membrane proteinase, and
inhibited by a matrix metalloproteinase and the TNF-N
protease inhibitor-2 [41].
As the idea behind Witte’s work remained attractive,
serum or soluble TfR (sTfR) was the next to be scrutinised
with regard to the validity of its use as a marker of BM
iron. We critically reviewed the existing evidence in order
to evaluate if sTfR measurement or any calculated values
(such as the sTfR/Log10ferritin ratio or sTfR–F index) could stand as a surrogate marker of BM iron.
Review criteria
A comprehensive computerized literature search of
English-language studies was performed using the [Mesh]
terms “receptors, transferrin” and ‘bone marrow’’ in the
MEDLINE database with limits in humans, in the English
Language, and over the age of 19 years. We checked Medline
1966 – May 2007 and we tracked appropriate references. A
total of 31 references were found. Twelve more references
were tracked from a manual review of related articles.
From the total 43 articles, twenty relevant references (sTfR
measurement and BM iron staining) were identified; nine
of which fulfilled the inclusion criteria and were included
in our analysis.
Inclusion Criteria: Studies were included only if they
satisfied the following criteria:
1. Examined anaemic adult patients who had sTfR
measurements documented.
2. Bone marrow aspiration and iron stain was the
reference standard for detection of iron status.
In some studies, only a subgroup of patients fulfilled the
inclusion criteria. Therefore, included studies ought to have
at least 50% of patients subjected to BM examination.
Exclusion criteria: letters, comments, and articles
without original data and conference abstracts were excluded.
We also excluded 2 studies because recruited patients had
malignant lymphomas and myelodysplastic syndromes [45,
46]. Haematological malignancies are known to interfere
347
with the level of sTfR and qualify as exclusion criteria for
most of the other studies. Two more studies were not included
in our final analysis as they contained preliminary results
reported later in a larger study by Punnonen et al [47-49].
Assessment of methodological quality: the following
criteria were used to assess methodological quality [22]
(Table I).
Table I. Quality assessment of the studies evaluated
Study (year)
Population
Intervention
Outcomes
Total
score
Punnonen et al. 1997
3
2
1
6
Rimon et al. 2002
3
2
1
6
Means et al. 1999
2
2
2
6
Hanif et al. 2005
2
2
1
5
Joosten et al. 2002
2
2
1
5
Junca et al. 1998
1
2
2
5
Lee et al. 2002
2
2
1
5
Suominen et al. 2000
1
2
1
4
Fitzsimons et al. 2002
1
2
1
4
Quality assessment
No. studies (%)
Population
Ideal
2 (22.2)
Best
4 (44.5)
Acceptable
3 (33.3)
Intervention
Ideal
Acceptable
9 (100)
- (0)
Outcome measures
Ideal
- (0)
Best
2 (22.2)
Acceptable
7 (77.8)
A. Population
Ideal - prospective study, consecutive anaemic patients
(with explicit definition of anaemia), exclusion criteria;
marked as 3
Best - prospective study, not consecutive but with
exclusion criteria; marked as 2
Acceptable - any other prospective study; marked as 1
B. Intervention
Ideal - specified method of testing (i.e. how laboratory
tests were done); marked as 3
Acceptable - anything else; marked as 1
C. Outcome – Bone marrow
Ideal - BM examined by 2 or more blinded (to each other
and the lab tests) readers; marked as 3
Best - BM examined by 1 person blinded to the lab tests,
or 2 readers not blinded; marked as 2
Acceptable - anything else; marked as 1
Data Extraction: Data was extracted independently
from each report (by AK and ES) using a predefined review
form, and disagreement was resolved by consensus. The
348
Koulaouzidis et al
extractors were aware of the site of origin of publication,
journal and year of publication. The following data was
recorded from each article: author and year of publication,
sample size, number of anaemic patients, number of bone
marrow examinations performed, classification of patients
according to bone marrow iron status and sensitivity, and
specificity of sTfR and/or sTfR–F index if that had been
calculated (Table II).
Studies breakdown and discussion
Punnonen et al (1997) studied 129 consecutive anaemic
patients. All underwent BM aspiration to define the type of
anaemia and determine iron stores [49]. All blood samples
were obtained before any blood transfusions, and patients on
iron therapy were excluded. Haematological malignancies,
haemolytic anaemias, B12 or folic acid deficiencies were
also considered as exclusion criteria [50-52].
Bone marrow aspirations were performed from sternal
or iliac crest sites. Patients were assigned to three groups
according to their BM iron. Forty-eight belonged to the group
of ID (no stainable iron), 64 were classified as having ACD
(stainable iron present), whilst 17 with both no stainable
iron in the BM plus a concurrent inflammatory process
or non-haematological malignancy, were placed in the
combined (COMBI) anaemia group. Measurement of sTfR
was performed using a polyclonal antibody in a sandwich
Enzyme ImmunoAssay (EIA). The reference distribution
levels were 0.85 to 3.5 mg/L. The calculated areas under the
receiver operator characteristics (AUCROC) for sTfR were
0.98 for the whole patient population. The same value was
obtained for the subpopulation of uncomplicated IDA and
ACD patients, as well as the one containing the ACD and
COMBI anaemias. The logarithmic transformation of the
ferritin values, and calculation of the sTfR–F Index provided
an outstanding indicator of iron depletion (AUCROC:
1.00).
The efficiency curves for distinction between IDA and
ACD suggest cut-off limits for sTfR and for sTfR–F index
of 2.7 and 1.5 mg/L respectively. Such a cut-off level for
sTfR (2.7 mg/L) had sensitivity, specificity, PPV, negative
predictive value (NPV) and efficiency of 0.94 each. The
equivalent values for the sTfR–F index were 0.98, 1.00,
1.00, 0.98, 0.99.
A year later (1998), Juncà et al. reported on their study
of 37 anaemic patients (10 of them had hypoferritinaemia
due to chronic blood losses and were used as controls
of the sTfR technique) with concurrent infectious or
inflammatory pathology [53]. Those from the group (n=27)
with hyperferritinaemia underwent BM aspiration. Twelve
were found to have low BM iron and fifteen (n=15) normal or
increased BM iron content. The BM smears were evaluated
by two pathologists and marrow iron grade was recorded as
0 (absent), + (reduced), ++ (normal) or +++ (increased).
They simplified the results into absent or low vs.
normal or high BM iron. sTfR was measured with an
immunoenzymometric assay and the reference range was
3.1– 4.5 mg/L. Applying multivariate logistic regression
analysis they found that sTfR was the first variable to be
excluded from the model as non-significant, as it did not
improve its predictive value at any cut-off level. At a level
of 4.5 mg/L the sensitivity was only 50% and the specificity
Table II. Review results
IDA
ACD
Combi
anaemia
sTfR
sTfR-F Index
Name of
study
IDA on
BM
sTfR
mean
ACDtotal
sTfR
mean
no
sTfR
mean
sTfR
cut-off
level
sens%
spec%
PPV%
NPV%
Index
level
sens%
spec%
Punnonen
1997
48
6.2 ±
3.5
64
1.8 ±
0.6
17
5.1 ±
2.0
2.7
94
94
94
94
1.5
98
100
Junca1998
n/a
n/a
15
3.39 ±
1.9
12
5.63 ±
3.3
2.8
84
47
n/a
n/a
n/a
n/a
n/a
Means 1999
24
52.7 ±
39.9*
121
23.7 ±
12.3*
n/a
n/a
71
74
n/a
n/a
n/a
n/a
n/a
Suominen
2000
19
2.92 ±
0.76
11
1.81 ±
0.48
n/a
n/a
2.3
n/a
n/a
n/a
n/a
13.5
n/a
n/a
Joosten
2002
27
34±11
38
25.5 ±
10
7
n/a
28
68
61
n/a
n/a
n/a
n/a
n/a
Rimon
2002
49
n/a
14
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
1.5
88
93
Lee 2002
31
n/a
48
n/a
41
n/a
1.8
97
88
n/a
n/a
1.36
100
98
Fitzsimons
2002
6
n/a
12
23.9*
15
40.2*
28.1*
93
92
n/a
n/a
n/a
n/a
n/a
Hanif 2005
86
9.68 ±
2.48
90
2.96 ±
1.28
n/a
n/a
n/a
83.3
100
100
87
n/a
n/a
n/a
IDA: iron deficiency anaemia, sTfR: soluble Transferrin receptor, ACD: anaemia of chronic disease, Combi anaemia: ACD+IDA,
no: number, sens: sensitivity, spec: specificity, PPV: positive predictive value, NPV: negative predictive value, *sTfR levels in
nmo/L (sTfR levels for the rest expressed in mg/L)
Soluble transferrin receptors and iron deficiency
74%. This study involved a small number of patients and the
sTfR–F index was not estimated.
A year later (1999), Means et al. studied the predictive
value of sTfR for the BM iron results in a heterogeneous
group of patients [54]. This was a large, three-centre
study that included 145 anaemic patients undergoing BM
examination for diagnostic purposes (no BM aspiration
was performed solely for study purposes). The investigators
examined the stained marrow smears in a blinded manner,
i.e. without any information about the patients and reported
the BM iron content in a semi-quantitative scale. Subjects
receiving iron or erythropoietin therapy were excluded.
Of the 216 individuals recruited, 71 were excluded for
various reasons [15 were excluded later due to concomitant
treatment with iron, some had BM biopsy for cancer staging
or haematological malignancies (leukaemia/myeloma), 11
had B12 deficiency or other haemolytic states, while 45
patients had unsatisfactory or uninterruptible BM samples].
The remaining 145 entered the final analysis, including
some with malignancies and liver disease (n=87). Soluble
TfR was measured by enzyme-linked immunosorbent assay
(ELISA) and the reference range was 8.8 – 28.1nmol/L for
white and 10.6 – 29.9nmol/L for coloured subjects. Mean
sTfR concentration for patients with stainable iron on the
BM aspirate was 23.7 ± 12.3nmol/L (mean ± SD), while the
value for those with absent BM iron were 52.7 ± 39.9nmol/L.
Means et al. tested a sequential algorithm (predictor of BM
iron) for both ferritin alone and ferritin/sTfR in combination.
The ferritin algorithm correctly identified 42% of the patients
with absent BM iron (sensitivity) and 70% of those with BM
stainable for iron (specificity).
The authors found that sTfR was the only laboratory
test used (amongst serum iron, Tf, Tf saturation, ferritin,
MCV) which showed mean values in the normal range for
patients with detectable BM iron and mean values outside the
normal range in those without detectable BM iron. Despite
this, the sTfR assay in this study failed to identify 29% of
the patients with absent BM iron. The authors explained
this to be inherent to the study design as only 7% of the
recruited subjects underwent BM examination primarily for
the evaluation of anaemia.
Suominen et al. (2000) studied a group of 30 anaemic
patients with rheumatoid arthritis (RA) and performed BM
examination in all of them [55]. The investigators determined
both sTfR and sTfR–F index and according to their protocol
they supplemented patients with diminished or exhausted
BM stores with iron. In one more study from Turku, Finland,
it was reported that an sTfR level of 2.3mg/L and an sTfR–F
index of 1.35 was the most efficient in discriminating
between absent and replete BM iron stores.
In 2002, four more studies on that subject were published.
Joosten et al. prospectively studied elderly hospitalised
patients (mean age 83 years) in order to check the validity of
sTfR in discriminating IDA from ACD [56]. After excluding
cases of haematological malignancies, haemolytic anaemia,
B12 or folate deficiency, moderate renal failure, recent
surgery or therapy with iron/transfusion, they recruited 83
349
anaemic patients who all underwent BM aspiration. Thirtyfour had no stainable BM iron and were classified as IDA;
38 had iron replete iron stores from either acute infections
or chronic inflammatory conditions, and were classified
as ACD; 11 were excluded from further analysis as they
had normal BM iron and no evidence of chronic/acute
infection, inflammation or malignancy. Their results showed
that at a level of 2.8mg/L for the detection of ID, the sTfR
measurement had a sensitivity of 68% and specificity of 61%.
The authors concluded that sTfR measurement provided
results similar to that of ferritin in their study population,
partly because of the broader spectrum of coexisting
underlying diseases.
Rimon et al. also focused on elderly inpatients and
studied a population of 49 anaemic and 14 non-anaemic
individuals (n=63) [57], all of whom had BM examinations.
As well as measuring the usual haematological parameters
and sTfR values, they also calculated the sTfR–F index.
This was a well-conducted, prospective study with stringent
inclusion and exclusion criteria and a good study size. The
mean age of the cohort was 83 years. Using a sTfR–F index
level of 1.5 they came up with a specificity of 92.9% and
sensitivity of 87.8% for the detection of IDA (PPV of 97.7%
and NPV of 68.4%).
Fitzsimons et al. (2002) examined BM from 29 anaemic
patients with rheumatoid arthritis (RA), 6 patients with
simple IDA, and healthy volunteers. They employed 125ITf and 59Fe-Tf in order to assess the in vitro erythroblast
surface TfR expression, and the uptake of high purity
erythroblast fractions prepared from the BM samples. Serum
TfR values were measured only for the RA anaemia group,
which was subdivided as either RA-ACD (marrow iron
present) or RA-IDA (marrow iron absent), on the basis of
visible reticuloendothelial BM iron stores. They concluded
that sTfR levels in RA-ACD remain within the normal range.
RA erythroblasts, however are still able to respond to any
additional worsening of the iron supply caused by absent
reticuloendothelial iron stores. This additional response
causes the highly significant increase in serum sTfR values
seen between RA-IDA and RA-ACD [58].
Later that year, Lee et al. reported their study of the use
of sTfR and sTfR–F index in anaemic patients with nonhaematological malignancies and chronic inflammation [59].
The authors recruited 120 anaemic adults (mean age 54 years)
who underwent BM examination for anaemia and 81 nonanaemic controls. The group was divided into 4 subgroups
according to BM results: IDA (n=31), ACD (n=48), COMBI
with chronic inflammatory disease (n=15) and COMBI with
malignancy (n=26). Exclusion criteria were similar to those
of previous studies and sTfR was measured with the use of
an immunoturbidimetric method. At an sTfR–F Index level
of 1.36, the sensitivity for detection of IDA was 100% and
the specificity 98%, while the same parameters for a sTfR
level of 1.8 mg/L were 97% and 88%, respectively.
Hanif et al. (2005) conducted a study wherein 176
patients were subjected into BM examination [60]. The
authors reported excellent PPV for the test (sTfR) in both
350
Koulaouzidis et al
IDA and ACD, but the NPV for the latter state was only
74.1%. This study had no clear sTfR level used for defining
IDA, and the reported PPV and NPV in different subgroups
of IDA and ACD appear confusing.
The studies we examined are heterogeneous as to the
population of recruited patients (Table III) and the exclusion
criteria used (Table IV). Most of them, with four notable
exceptions [49, 54, 59, 60], were limited by the number
of BM examinations performed, and were restricted to a
maximum of 3 centres [54]. Furthermore reference ranges
for the measurement of sTfR are blighted as there are
several commercial assays available, and the technique is
not standardized [61, 62]. This leads to confusion in the
transferability of results from one study to another, both
nationally and internationally; however this problem is not
unique within the remit of diagnostic testing. There must
be increased pressure to establish international guidelines
for the calibration of sTfR, as first highlighted by Skikne
in 1998 [62].
Currently, there is no National Quality Assessment
Scheme so it still remains impractical to compare across
different laboratory analyses, which is necessary if sTfR
is to be utilized on a wider basis. The sTfR assay is also
claimed to be cost-prohibitive in comparison to established
‘traditional’ laboratory indices of iron assessment, but with
the continued proliferation of high-throughput analyzers
and utilisating micro-sampling techniques, costs have fallen
significantly (unpublished authors’ data). Of note is the
Table III. Populations, characteristics
immunoturbidimetric method, which guarantees results in
11 minutes.
STfR measurements appear to have high sensitivity, but
suffer from low specificity when it comes to differentiating
iron deficiency from other causes of increased erythropoiesis
(e.g. haemolytic states such as ineffective erythropoiesis
from B12/folate deficiency and thalassaemia). This can be
partially corrected by incorporating ferritin in the calculated
index.
On the basis of this review we conclude that if sTfR or the
calculated sTfR–F index is to be used on a broader scale for
the assessment of BM iron stores, a prospective, multicenter
study is required. Such a study would need to utilize the
existing ‘gold standard’, i.e. BM iron stain, as a crossreference and apply uniform exclusion and inclusion criteria.
A sTfR level of 2.5mg/L (29.5nmol/L) appears a reasonable
threshold for identification of iron-deficiency anaemia (with
good specificity and sensitivity) (Fig. 1) [63].
With the integration of Clinical Chemistry and
Haematology departments, the assessment of iron status is
beginning to enjoy a more unified approach. The use of sTfR
as sTfR–F index can improve the diagnosis of iron deficiency
anaemia, especially in the presence of coexisting chronic
disease, but further research is required [64, 65].
Table IV. Exclusion criteria per study
Study
(year)
Exclusion
criteria
Type of exclusion criteria
Punnonen
et al. 1997
yes
Oral iron therapy, haematological
malignancies, haemolytic anaemia, B12/
folate deficiency
Junca et al.
1998
no
n/a
Means et al.
1999
yes
Iron and/or erythropoietin therapy,
haemolysis & B12 deficiency
Suominen
et al. 2000
n/a
n/a
Joosten et
al. 2002
yes
Haematological malignancies,haemolysi,
B12/folate deficiency, creatinine>200μg/
L, iron therapy, transfusion or surgery
Rimon et
al. 2002
yes
Not consented, iron therapy, B12/folate
deficiency, acute GI bleed, malignancies,
renal or liver failure
Lee et al.
2002
yes
Haematological malignancies,haemolysis
,B12/folate deficiency, iron therapy, bone
marrow hypercellularity
Study
(year)
No
Mean age
(SD)
Anaemia
cause
sTfR reference
values
Punnonen
et al. 1997
129
n/a
n/a
0.85-3.05‡
Junca et al.
1998
37
hypo-F: 65
IDA, GI
loss
3.1-4.5‡
hyper-F: 65.7
ACD
Means et al.
1999
216
IDA: 43(12)
IDA
various
White: 8.8-28.1*
Suominen
et al. 2000
30
range:
26 - 79
RA
n/a
Joosten et
al. 2002
83
83
various
to 28.1*
Rimon et
al. 2002
63
83(2.8)
various
0.85-3.05‡
Lee et al.
2002
201
54
various
n/a
Fitzsimons
et al. 2002
no
n/a
Fitzsimons
et al. 2002
44
IDA:
59.8(19.3)
IDA
8.8-28.1*
Hanif et al.
2005
yes
Thallasaemia, sideroblastic anaemia
Hanif 2 et
al. 005
ACD: 51(18)
176
ACD: 63.1
RA
n/a
various
Black: 10.6-29.9*
1.3-3.33‡
IDA: Iron Deficiency Anaemia; ACD: Anaemia of Chronic
Disease; RA: Rheumatoid Arthritis; n/a: not applicable; hypoF: hypoferritinaemic; hyper-F: hyperferritinaemic; GI loss:
gastrointestinal blood loss; sTfR reference values:
*nmol/L, ‡ mg/L
Conclusion
The use of sTfR improves the clinical diagnosis of
iron deficiency anaemia, especially in the presence of
coexisting chronic disease or gastrointestinal malignancies.
The safety and cost-effectiveness of a ferritin/sTfR-based
approach to exclude gastrointestinal cancer in the presence
Soluble transferrin receptors and iron deficiency
Fig 1. sTfR cut-off level and mean with its confidence intervals.
of iron deficiency has to be proven with a prospective, well
standardised, multicentre trial or a meta-analysis.
Acknowledgements
We thank Mageed Abdalla, MSc for his invaluable help
with Figure 1, A Karagiannidis, MD for his help with the
tables, and Shivaram Bhat, MBBCh & Sarah Douglas, BSc
for their help with proof reading.
References
1. Grosbois B, Decaux O, Cador B, Cazalets C, Jego P. Human iron
deficiency. Bull Acad Natl Med 2005; 189: 1649-1663.
2. DeMaeyer E, Adiels-Tegman M. The prevalence of anaemia in the
world. World Health Stat Q 1985; 38: 302-316.
3. Cook JD. Iron-deficiency anaemia. Baillieres Clin Haematol 1994;
7: 787-804.
4. Looker AC, Dallman PR, Carroll MD, Gunter EW, Johnson CL.
Prevalence of iron deficiency in the United States. JAMA 1997; 277:
973-976.
5. Hercberg S, Preziosi P, Galan P. Iron deficiency in Europe. Public
Health Nutr 2001; 4: 537-545.
6. Rockey DC, Cello JP. Evaluation of the gastrointestinal tract in
patients with iron-deficiency anemia. N Engl J Med 1993; 329:
1691-1695.
7. McIntyre AS, Long RG. Prospective survey of investigations in
outpatients referred with iron deficiency anaemia. Gut 1993; 34:
1102-1107.
8. Ioannou GN, Rockey DC, Bryson CL, Weiss NS. Iron deficiency
and gastrointestinal malignancy: a population-based cohort study.
Am J Med 2002; 113: 276-280.
9. Metzgeroth G, Adelberger V, Dorn-Beineke A, et al. Soluble
transferrin receptor and zinc protoporphyrin--competitors or efficient
partners? Eur J Haematol 2005; 75: 309-317.
10. Kemna EH, Tjalsma H, Willems HL, Swinkels DW. Hepcidin:
from discovery to differential diagnosis. Haematologica 2008; 93:
90-97.
11. Logan EC, Yates JM, Stewart RM, Fielding K, Kendrick D.
Investigation and management of iron deficiency anaemia in
general practice: a cluster randomized controlled trial of a simple
management prompt. Postgrad Med J 2002; 78: 533-537.
12. Goddard AF, James MW, McIntyre AS, et al. Guidelines for the
Management of Iron Deficiency Anaemia. BSG 2005; Clinical
Practice – Guidelines. www.bsg.org.uk/pdf_word_docs/iron_def.
351
pdf (accessed 28th February 2008)
13. Green J. Guidelines on Complications of Gastrointestinal Endoscopy.
BSG 2006; Clinical Practice – Guidelines. www.bsg.org.uk/pdf_
word_docs/complications.pdf (accessed 28th February 2008)
14. Cook JD. Diagnosis and management of iron-deficiency anaemia.
Best Pract Res Clin Haematol 2005; 18: 319-332.
15. Brugnara C. Iron deficiency and erythropoiesis: new diagnostic
approaches. Clin Chem 2003; 49: 1573-1578.
16. Wright CM, Kelly J, Trail A, Parkinson KN, Summerfield G. The
diagnosis of borderline iron deficiency: results of a therapeutic trial.
Arch Dis Child 2004; 89: 1028-1031.
17. Jacobs A, Worwood M. The clinical use of serum ferritin estimation.
Br J Haematol 1975; 31: 1-3.
18. Walters GO, Miller FM, Worwood M. Serum ferritin concentration
and iron stores in normal subjects. J Clin Pathol 1973; 26: 770772.
19. Bentley DP, Williams P. Serum ferritin concentration as an index
of storage iron in rheumatoid arthritis. J Clin Pathol 1974; 27: 786788.
20. Krause JR, Stolc V. Serum ferritin and bone marrow iron stores. I.
Correlation with absence of iron in biopsy specimens. Am J Clin
Pathol 1979; 72: 817-820.
21. Mazza J, Barr RM, McDonald JW, Valberg LS. Usefulness of the
serum ferritin concentration in the detection of iron deficiency in a
general hospital. Can Med Assoc J 1978; 119: 884-886.
22. Guyatt GH, Oxman AD, Ali M, Willan A, McIlroy W, Patterson C.
Laboratory diagnosis of iron-deficiency anemia: an overview. J Gen
Intern Med 1992; 7: 145-153.
23. Ali MA, Luxton AW, Walker WH. Serum ferritin concentration and
bone marrow iron stores: a prospective study. Can Med Assoc J 1978;
118: 945-946.
24. Mast AE, Blinder MA, Gronowski AM, Chumley C, Scott MG.
Clinical utility of the soluble transferring receptor and comparison
with serum ferritin in several populations. Clin Chem 1998; 44: 4551.
25. Weiss G, Goodnough LT. Anemia of chronic disease. N Engl J Med
2005; 352: 1011-1023.
26. Lipschitz DA, Cook JD, Finch CA. A clinical evaluation of serum
ferritin as an index of iron stores. N Engl J Med 1974; 290: 12131216.
27. Barron BA, Hoyer JD, Tefferi A. A bone marrow report of absent
stainable iron is not diagnostic of iron deficiency. Ann Hematol 2001;
80: 166-169.
28. Ganti AK, Moazzam N, Laroia S, Tendulkar K, Potti A, Mehdi SA.
Predictive value of absent bone marrow iron stores in the clinical
diagnosis of iron deficiency anemia. In Vivo 2003; 17: 389-392.
29. Ervasti M, Kotisaari S, Romppanen J, Punnonen K. In patients
who have stainable iron in the bone marrow an elevated plasma
transferrin receptor value may reflect functional iron deficiency. Clin
Lab Haematol 2004; 26: 205-209.
30. Witte DL, Kraemer DF, Johnson GF, Dick FR, Hamilton H. Prediction
of bone marrow iron findings from tests performed on peripheral
blood. Am J Clin Pathol 1986; 85: 202-206.
31. Witte DL, Angstadt DS, Davis SH, Schrantz RD. Predicting bone
marrow iron stores in anemic patients in a community hospital using
ferritin and erythrocyte sedimentation rate. Am J Clin Pathol 1988;
90: 85-87.
32. Coenen JL, van Dieijen-Visser MP, van Pelt J, et al. Measurements of
serum ferritin used to predict concentrations of iron in bone marrow
in anemia of chronic disease. Clin Chem 1991; 37: 560-563.
33. Guyatt GH, Patterson C, Ali M, et al. Diagnosis of iron-deficiency
anemia in the elderly. Am J Med 1990; 88: 205-209.
352
34. Cook JD, Flowers CH, Skikne BS. The quantitative assessment of
body iron. Blood 2003; 101: 3359-3364.
35. Oski FA. Iron deficiency in infancy and childhood. N Engl J Med
1993; 329: 190-193.
36. Jandl JH, Katz JH. The plasma-to-cell cycle of transferrin. J Clin
Invest 1963; 42: 314-326.
37. Trowbridge IS, Omary MB. Human cell surface glycoprotein related
to cell proliferation is the receptor for transferrin. Proc Natl Acad
Sci U S A 1981; 78: 3039-3043.
38. van Renswoude J, Bridges KR, Harford JB, Klausner RD. Receptormediated endocytosis of transferrin and the uptake of fe in K562
cells: identification of a non-lysosomal acidic compartment. Proc
Natl Acad Sci U S A 1982; 79: 6186-6190.
39. Ward JH, Kushner JP, Ray FA, Kaplan J. Transferrin receptor function
in hereditary hemochromatosis. J Lab Clin Med 1984; 103: 246254.
40. Kohgo Y, Niitsu Y, Nishisato T, et al. Externalization of transferrin
receptor in established human cell lines. Cell Biol Int Rep 1987; 11:
871-879.
41. Kaup M, Dassler K, Weise C, Fuchs H. Shedding of the transferrin
receptor is mediated constitutively by an integral membrane
metalloprotease sensitive to tumor necrosis factor alpha protease
inhibitor-2. J Biol Chem 2002; 277: 38494-38502.
42. Kohgo Y, Nishisato T, Kondo H, Tsushima N, Niitsu Y, Urushizaki
I. Circulating transferrin receptor in human serum. Br J Haematol
1986; 64: 277-281.
43. Kohgo Y, Niitsu Y, Kondo H, et al. Serum transferrin receptor as a
new index of erythropoiesis. Blood 1987; 70: 1955-1958.
44. Shih YJ, Baynes RD, Hudson BG, Cook JD. Characterization and
quantitation of the circulating forms of serum transferrin receptor
using domain-specific antibodies. Blood 1993; 81: 234-238.
45. Bjerner J, Amlie LM, Rusten LS, Jakobsen E. Serum levels of soluble
transferrin receptor correlate with severity of disease but not with
iron stores in patients with malignant lymphomas. Tumour Biol 2002;
23: 146-153.
46. Matsuda A, Misumi M, Shimada T, et al. Soluble transferrin receptor
and its ratio to erythroblasts in bone marrow may be a new diagnostic
tool to distinguish between aplastic and refractory anemia. Acta
Haematol 2004; 111: 138-142.
47. Punnonen K, Irjala K, Rajamäki A. Iron-deficiency anemia is
associated with high concentration of transferrin receptor in serum.
Clin Chem 1994; 40: 774-776.
48. Suominen P, Punnonen K, Rajamaki A, Irjala K. Evaluation of
new immunoenzymometric assay for measuring soluble transferrin
receptor to detect iron deficiency in anemic patients. Clin Chem
1997; 43: 1641-1646.
49. Punnonen K, Irjala K, Rajamaki A. Serum transferrin receptor and
its ratio to serum ferritin in the diagnosis of iron deficiency. Blood
1997; 89: 1052-1057.
50. Huebers HA, Beguin Y, Pootrakul P, Einspahr D, Finch CA.
Intact transferrin receptors in human plasma and their relation to
erythropoiesis. Blood 1990; 75: 102-107.
Koulaouzidis et al
51. Carmel R, Skikne BS. Serum transferrin receptor in the megaloblastic
anemia of cobalamin deficiency. Eur J Haematol 1992; 49: 246250.
52. Beguin Y, Lampertz S, De Groote D, Igot D, Malaise M, Fillet G.
Soluble CD23 and other receptors (CD4, CD8, CD25, CD71) in
serum of patients with chronic lymphocytic leukemia. Leukemia
1993; 7: 2019-2025.
53. Juncà J, Fernández-Avilés F, Oriol A, et al. The usefulness of the
serum transferrin receptor in detecting iron deficiency in the anemia
of chronic disorders. Haematologica 1998; 83: 676-680.
54. Means RT Jr, Allen J, Sears DA, Schuster SJ. Serum soluble
transferrin receptor and the prediction of marrow aspirate iron results
in a heterogeneous group of patients. Clin Lab Haematol 1999; 21:
161-167.
55. Suominen P, Mottonen T, Rajamaki A, Irjala K. Single values of
serum transferrin receptor and transferrin receptor ferritin index can
be used to detect true and functional iron deficiency in rheumatoid
arthritis patients with anemia. Arthritis Rheum 2000; 43: 10161020.
56. Joosten E, Van Loon R, Billen J, Blanckaert N, Fabri R, Pelemans
W. Serum transferrin receptor in the evaluation of the iron status in
elderly hospitalized patients with anemia. Am J Hematol 2002; 69:
1-6.
57. Rimon E, Levy S, Sapir A, et al. Diagnosis of iron deficiency anemia
in the elderly by transferrin receptor-ferritin index. Arch Intern Med
2002; 162: 445-449.
58. Fitzsimons EJ, Houston T, Munro R, Sturrock RD, Speekenbrink
AB, Brock JH. Erythroblast iron metabolism and serum soluble
transferrin receptor values in the anemia of rheumatoid arthritis.
Arthritis Rheum 2002; 47: 166-171.
59. Lee EJ, Oh EJ, Park YJ, Lee HK, Kim BK. Soluble transferrin
receptor (sTfR), ferritin, and sTfR/log ferritin index in anemic
patients with nonhematologic malignancy and chronic inflammation.
Clin Chem 2002; 48: 1118-1121.
60. Hanif E, Ayyub M, Anwar M, Ali W, Bashir M. Evaluation of serum
transferrin receptor concentration in diagnosing and differentiating
iron deficiency anaemia from anaemia of chronic disorders. J Pak
Med Assoc 2005; 55: 13-16.
61. Kuiper-Kramer EP, Van Raan J, Van Eijk HG. A new assay for soluble
transferrin receptors in serum: time for standardisation. Eur J Clin
Chem Clin Biochem 1997; 35: 793.
62. Skikne BS. Circulating transferrin receptor assay—coming of age.
Clin Chem 1998; 44: 7-9.
63. Koulaouzidis A, Saeed AA, Abdallah M, Said EM. Transferrin
receptor level as surrogate peripheral blood marker of iron deficiency
states. Scand J Gastroenterol 2009; 44: 126-127.
64. Koulaouzidis A, Bhat S. Investigating iron status in microcytic
anaemia: zincprotoporphyrin and soluble transferrin receptor have
a role. BMJ 2006; 333: 972.
65. Koulaouzidis A, Cottier R, Bhat S, Said E, Linaker BD, Saeed AA.
A ferritin level >50 microg/L is frequently consistent with iron
deficiency. Eur J Intern Med 2009; 20: 168-170.