OUP user menu

Serum transferrin receptor assay in iron deficiency anaemia and anaemia of chronic disease in the elderly

E. Chua, J.E. Clague, A.K. Sharma, M.A. Horan, M. Lombard
DOI: http://dx.doi.org/10.1093/qjmed/92.10.587 587-594 First published online: 1 October 1999


The most common cause of anaemia in the elderly is anaemia of chronic disease (ACD). However, iron deficiency anaemia (IDA) may coexist, and can be difficult to diagnose. The serum transferrin receptor (sTfR) blood test may be a better indicator of iron status as it is not affected by inflammation nor by advancing age. We evaluated it in four groups (10 males, 10 females each): `young' controls, `elderly' controls, IDA and ACD. All patients in the IDA group had elevated sTfR levels (mean±SD 65.2±17.7 nmol/l). All `young' controls had normal sTfR (22.3±7.3 nmol/l) and ferritin levels (92.7±61.1 μg/l). Although all subjects in the `elderly' controls and ACD group had normal, and raised or normal serum ferritin, respectively (88±62.3 μg/l; 631.2±509.5 μg/l), three (15%) `elderly' controls and four (20%) ACD patients had raised sTfR levels, suggesting depleted iron stores. Bone-marrow aspirates were available in 3/4 ACD patients with raised sTfR. Haemosiderin was absent in two. The sTfR blood test is comparable to serum ferritin in diagnosing IDA in the elderly but also seems capable of differentiating ACD from IDA. Its potential as a non-invasive test of iron status, especially in elderly anaemic patients, deserves further evaluation.


The diagnosis of iron deficiency anaemia (IDA) from a full blood count and serum ferritin can be difficult in the elderly, owing to the presence of multiple pathologies. The typical blood film appearance of a hypochromic microcytic state is suggestive of iron deficiency, but these features may also occur in malignancy, chronic infection, and renal failure.1,,2 Furthermore, in early iron deficiency, the blood film appearance is usually normochromic and normocytic. In addition, a full blood count can be normal even with no iron stores in the bone marrow, provided erythropoeisis is unaffected.3

Serum ferritin is regarded as the best single test for iron deficiency, as its concentration is roughly proportional to total iron stores.4 However, its interpretation in the elderly is complicated by its tendency to rise with age,5 chronic sepsis and malignancy.6 The reference range for serum ferritin is wide (15–300 μg/l). A previous study revealed that a serum ferritin of <45 μg/l in elderly subjects (>65 years) is suggestive of IDA, compared to a value of 12 μg/l in younger adults.7 A more recent study on elderly patients (serum ferritin 12–45 μg/l), revealed that 84% of patients (27/32) had no haemosiderin iron in their bone marrow,8 bone-marrow analysis being taken as the definitive test for IDA. In these patients, haemoglobin and mean corpuscular volume (MCV) were within the reference range. As inflammatory disorders and ageing may influence the traditional parameters of iron status, a parameter that is not influenced by these conditions would be very useful. A promising candidate is the serum transferrin receptor (sTfR).

Transferrin receptor (TfR) is a disulphide-linked dimmer of two identical sub-units, 95 kDa each,9 found on most cells and especially those with a high requirement for iron, such as immature erythroid and malignant cells. Its function is to internalize absorbed iron into target cells. The mechanisms by which cells acquire iron is complex. Most iron uptake occurs via a receptor-mediated endocytosis route with diferric transferrin (the main iron-transport protein), bound to the TfR located on the cell surface, forming a ligand complex. This is followed by internalization of the complex and in a series of reactions, iron is released from transferrin into the cytosol; the apotransferrin-TfR complex is returned to the cell surface.10

Interest has been directed towards TfR expression in the proximal small intestine, the principal site of iron absorption in man. Intestinal TfR expression (protein and mRNA) is also higher in cells that have a higher iron requirement, e.g. in proliferating crypt cells.11,,12 An inverse relationship occurs between TfR and body iron stores, such that when body iron stores are low, receptor expression is enhanced to acquire more iron. Conversely, when body iron stores become replete, down-regulation of receptor expression occurs.13,,14

TfR is susceptible to proteolysis at the Arg100–Leu101 site15,,16 producing a serum transferrin receptor (sTfR), a 75 kDa monomer that circulates in plasma at a concentration which is proportional to the total cell mass TfR.17 Serum TfR is also increased in iron deficiency, unaffected by chronic diseases and is comparable to bone-marrow aspirate iron-status estimation.18,,19 However, such studies have mainly focused on a younger population. We have therefore investigated if the sTfR assay could be used to diagnose IDA in elderly subjects, to differentiate IDA from anaemia of chronic disease (ACD) and to diagnose mixed anaemia (combined IDA and ACD).


We studied 40 anaemic patients admitted to the care of the elderly wards at the University Hospital, Aintree. The Joosten20 criteria for anaemia were used (haemoglobin <11.5 g/dl) as there is a tendency for haemoglobin to be reduced even in healthy elderly individuals. Subject details are shown in Table 1. All patients with a MCV <80 fl were considered to be microcytic. Study patients were classified as having IDA (haemoglobin <11.5 g/dl, serum ferritin <15 μg/l and MCV <80 fl) or ACD (haemoglobin <11.5 g/dl, MCV >80 fl, CRP>5 mg/l and a normal or high serum ferritin). Each group contained 10 male and 10 female patients. In the IDA group, patients were found to have colonic malignancy (3), benign gastric ulcer (2), grade III or IV oesophagitis (3); in the remainder (12), there was no definite diagnosis and these patients were designated anaemia ?cause. In the ACD group, patients were found to have chronic inflammatory joint disease (3), polymyalgia rheumatica (1), anaemia with congestive heart failure (1), septicaemia (1), metastatic prostatic carcinoma (1) and chronic sepsis (13). Two control groups (10 males and 10 females in each) were also studied: a `young' control group of doctors and nurses with no previous medical illness, and an `elderly' control group. The latter group was recruited from a panel of healthy elderly subjects at the University Department of Geriatric Medicine, Hope Hospital, Salford. Subjects were health-status-defined elderly who had fulfilled the criteria of the SENIEUR protocol for immunogerontological studies.21

View this table:
Table 1 

Characteristics of patients

A venous sample of blood was obtained from all subjects for full blood count (STKS Coulter Counter, Stack S model), serum iron (calorimetric ferrozine-ascorbic acid reductase assay, Boehringer Mannheim), serum transferrin, ferritin, C-reactive protein (immunoturbidimetric assay, Boehringer Mannheim), and erythrocyte zinc protoporphyrin (ZPP) (haematofluorometer, Helena Laboratories). All blood samples were obtained prior to any blood transfusions in anaemic patients. Anaemic patients on oral iron therapy, and patients with haematological malignancies, haemolytic anaemia, or vitamin B12/folic acid deficiency were excluded from the study, as these have been associated with elevated sTfR regardless of iron status.22,,23 Urea and electrolytes, calcium and phosphate, blood glucose were measured and thyroid and liver function tests also performed, depending on clinical presentation; a reticulocyte count was recorded in the ACD group.

The sTfR was measured using an enzyme-linked immunosorbent assay (R and D Systems Europe), this assay was undertaken in accordance with the manufacturer's instructions. Absorbance (duplicate samples) was recorded at 450 nm (corrected at 540 nm) using a microplate manager spectrophotometer (Bio-Rad).

Statistical analysis

Data are expressed as means±SD (or median and range if not normally distributed). All calculation procedures were carried out using the Arcus-Pro II statistical package (Medical Computing).


Neither of the two control groups had a microcytic blood picture. There was no statistical difference in age between healthy `elderly' and the IDA and ACD groups. Median (range) CRP in the `young' and `elderly' controls was 5 (5–7) and 5 (5–16), respectively (Table 1). Median CRP in the IDA group was 17 (5–75), suggesting an inflammatory component. However, these patients were allocated to the IDA group as all subjects had a low serum ferritin (<15 μg/l). All patients in the ACD group had CRP>5 mg/l, median 85 (18–178). All `young' controls and 19 (95%) `elderly' controls had serum ferritin concentrations within the reference range (Figure 1a and b)

Figure 1.

Serum ferritin in females (a) and males (b) (upper and lower reference range). All IDA patients had low ferritin, all ACD patients had normal or raised ferritin, all `young' controls had ferritin within reference range and one subject in the `elderly' controls had low ferritin.

Nineteen (95%) and 18 (90%) of the `young' and `elderly' controls, respectively, had serum iron within the reference range (Figure 2a). Only 11 (55%) of the IDA and five (25%) of the ACD patients had low serum iron.

Figure 2.

Serum iron (a), serum transferrin (b) and transferrin saturation (c) (upper and lower reference range). Serum iron is not a useful reflector of iron status. In IDA, only 55% of patients had low serum iron. This is improved by measuring serum transferrin (80%) or transferrin saturation (70%). Note that one subject in the `elderly' control group had raised transferrin and low transferrin saturation, two ACD patients also had low transferrin saturation.

All `young' controls and 19 (95%) `elderly' controls had serum transferrin within the reference range (Figure 2b). Sixteen (80%) IDA patients and one (5%) ACD patient had a raised serum transferrin. All subjects in both control groups had transferrin saturation within the reference range (Figure 2c). Fifteen (75%) IDA patients and two (10%) ACD patients had low transferrin saturation.

None of the `young' control subjects had elevated sTfR levels (Figure 3). Contrary to expectation, three subjects (15%) in the `elderly' control group had raised sTfR levels. All patients in the IDA group and four patients (20%) in the ACD group had raised sTfR levels. Bone-marrow aspirates were available in three ACD patients with raised sTfR and haemosiderin was absent in two patients (Table 2).

View this table:
Table 2 

Characteristics of patients with raised sTfR not in the IDA group

Figure 3.

Serum transferrin receptor (upper and lower reference range). All patients in IDA had raised sTfR. Three elderly controls and four ACD patients also had raised sTfR.

All `young' controls had normal ZPP levels, and all IDA patients had elevated ZPP levels (Table 1). In the `elderly' control group, of the three subjects with elevated sTfR levels, two also had elevated ZPP levels (Table 2). In the ACD group, 15 (75%) patients had elevated ZPP levels.


Iron metabolism is complex, but can be viewed in the context of three pools: metabolic, transit and storage. Over 80% of red blood cells (RBC) and precursor cells constitute the metabolic pool. Erythrocyte haem iron is measured as blood haemoglobin. If haemoglobin is normal, then the patient is unlikely to be iron-deficient, but if the haemoglobin is normal, iron deficiency is one of many possibilities. The transit pool can be measured directly as serum iron, serum transferrin and, more recently, sTfR.

The storage pool is measured by the concentration of serum ferritin, which is roughly proportional to the total iron stores.4 Ferritin concentration below the normal range indicates depletion of iron and iron deficiency. However, ferritin, being an acute-phase protein, may be elevated in inflammatory disorders.6 Furthermore, ferritin concentration increases with age, especially in men.5 Therefore, in elderly patients with chronic disease, serum ferritin is not a reliable indicator of iron deficiency. This is important because the most common cause of anaemia in the elderly is ACD. IDA is the diagnosis in only 10–15% of anaemic patients,24 although worldwide, it is the most common cause of anaemia. Furthermore, the two conditions may coexist. Differeniation can be difficult but important, because although appropriate iron therapy may relieve the symptoms of anaemia, inappropriate therapy may lead to undesirable side-effects. Furthermore, iron deficiency may be a marker of more sinister gastrointestinal pathology.

The `experimental' gold standard for assessing iron status is quantitative phlebotomy. This involves removing up to 500 ml of blood weekly until anaemia develops such that most of the iron used for haemoglobin synthesis is obtained from iron stores rather than from intestinal absorption.25 However, this is a laborious procedure, and clinically impractical. The clinical `gold standard' for assessing iron status is staining haemosiderin iron with Perls' stain in bone-marrow aspirate. However, bone-marrow aspiration is invasive, expensive, and time-consuming, and access to this test may be limited for many physicians. A new parameter, free from the aforementioned constraints, would be desirable.

Serum iron is a poor indicator of iron status; only 55% of IDA patients had low serum iron concentration in this study. Serum iron fluctuates throughout the day4 and falls with advancing age.26,,27 However, the `diagnostic yield' was improved in this study by determining serum transferrin (80%) and percentage transferrin saturation (75%).

ZPP also appears useful in diagnosing IDA in the elderly. However, raised ZPP is only an indicator of disturbed iron metabolism, which occurs in sideroblastic disorders such as myelodysplasia, neoplastic disease and chronic inflammation, as well as in iron deficiency.28 This study confirms the observation that ZPP is raised in inflammatory disorders.

Allocation of patients to the IDA group was based on low serum ferritin and low MCV. All IDA patients also had elevated sTfR levels. Allocation of patients to the ACD group was based on a raised CRP, MCV >80 and raised ferritin. The ACD group may contain mixed anaemia, as bone-marrow aspiration was not performed on all patients in this group. Indeed, within this group, four patients had elevated sTfR levels (see Table 2). This observation can be interpreted in a number of ways. Serum TfR could be elevated in inflammatory conditions, as are ZPP and ferritin, but this is unlikely since animal and studies in man show lower serum receptor levels with inflammation.17,18,34,,35 The pathophysiology of ACD, although incompletely understood, involves impaired iron utilisation from senescent red cells, a shortened life span of erythrocytes and reduced erythropoeitin stimulation.29,,30 The latter was supported by a study showing no significant change in sTfR in ACD patients but reduced serum erythropoeitin levels, compared to a significant rise in sTfR and serum erythropoeitin levels in IDA patients.31 We found no evidence of haemolysis in our ACD patients. Furthermore, two of the four patients with raised sTFR also had low transferrin saturation (Table 2), suggesting that a more likely explanation for raised sTfR is combined iron deficiency and ACD. Bone-marrow studies revealed absent haemosiderin deposits in two ACD patients (Table 2).

In three `elderly' control subjects with raised sTfR, one had low serum ferritin (this patient developed IDA several months later), and two had serum ferritin concentrations within the reference range but below 45 μg/l (Table 2). Iron deficiency may have been present, since absent haemosiderin iron in bone-marrow aspirates from `elderly' patients with normal serum ferritin levels (<45 μg/l) have been reported.7,,8 This is further supported by raised ZPP in two of the `elderly' controls and raised serum transferrin (low transferrin saturation) in one `elderly' control (Table 2). Two out of three `elderly' subjects with raised sTfR would also fulfil the `MCV model' (MCV, ZPP and transferrin saturation) and `ferritin model' criteria for IDA (ferritin, transferrin saturation and ZPP) which require any two of the three tests to be abnormal.32 In studies where iron deficiency was induced in healthy subjects by serial phlebotomies, sTfR was a sensitive and highly quantitative guide to early iron deficiency.33 We included a `young' control group in our study to illustrate that interpretation of serum assays of iron status are not problematic by comparison to the `elderly'.

This study indicates that the sTfR blood test has a comparable ability to serum ferritin in diagnosing IDA, and is also able to differentiate ACD from IDA in an elderly population group. Although this finding has been reported previously, such studies were focused on a younger population.18,31,34,,35 Serum TfR may have an additional role of identifying the development of iron deficiency in patients with mixed anaemia. The two groups of patients with chronic inflammation and ACD who are particularly prone to develop IDA are patients with rheumatoid arthritis and patients with inflammatory bowel disease, where IDA occurs due to chronic blood loss. Several studies with bone-marrow aspirates34,,35 to address this issue have used the TfR : ferritin ratio. Elevated ratios were observed in patients with mixed anaemia (sensitivity of 87%) but not in patients with `pure' ACD (specificity of 100%), an overall diagnostic accuracy of 92%. It would appear that the mixed anaemia group may be identified in most cases by the sTfR or the sTfR : ferritin ratio.36 Again, these studies were conducted in a younger population.

Differentiation between IDA and ACD is particularly important in the elderly. Inappropriate gastrointestinal investigations, which are often unpleasant, and exhausting for elderly patients, may be avoided if IDA is not supported by a non-invasive blood test. Such investigations are expensive, elderly patients often have to be admitted for bowel preparation, and colonoscopy complication rates tend to be higher in the elderly. The sTfR appears to meet the criteria for that elusive parameter that reflects iron status which is non-invasive, easily accessible and independent of inflammatory conditions and ageing. However, this parameter should be used in conjunction with the `traditional' parameters of body iron status.


The authors would like to acknowledge Dr Mike Dennis, Specialist Registrar in Haematology, Royal Liverpool University Hospital, for his help and interpretation of the bone-marrow studies.


View Abstract