Clinical Chemistry 48:7
1066 –1076 (2002)
Biochemical Markers and Hematologic Indices in
the Diagnosis of Functional Iron Deficiency
Christian Thomas and Lothar Thomas*
Background: The hypochromic red cell is a direct indi-
cator of functional iron deficiency (ID) in contrast to the
majority of biochemical markers, which measure func-
tional ID indirectly via iron-deficient erythropoiesis.
The aim of this study was to evaluate the extent to which
these biochemical markers can distinguish ID from
anemia of chronic disease (ACD) as well as from the
combined state of functional ID/ACD, using red cell
hemoglobinization as the gold standard.
Methods: We studied 442 patients with various disease-
specific anemias and 154 nonanemic patients. As indi-
cators of red cell hemoglobinization, we measured the
reticulocyte hemoglobin content (CHr) and the propor-
tion of hypochromic red cells (HYPO), using an Advia
120 hematology analyzer. Ferritin, transferrin, transferrin
saturation, and the concentration of the soluble transferrin
receptor (sTfR) were determined by ELISA and immuno-
turbidimetric assay. The sTfR/log ferritin ratio (sTfR-F
index) was used as an additional marker for biochemical
identification of iron-deficient erythropoiesis.
Results: In a control group (n ⴝ 71), the 2.5 percentile
values were 28 pg for CHr and 5% for HYPO. These
values were used to indicate unimpaired red cell hemo-
globinization and absence of functional ID. In patients
with deficient red cell hemoglobinization but no acute-
phase response (APR), iron-deficient erythropoiesis was
indicated by serum ferritin and sTfR-F index values
<20.8 ␮g/L and >1.5, respectively. Corresponding val-
ues in patients with APR were <61.7 ␮g/L and >0.8,
respectively. The positive likelihood ratios for the bio-
chemical markers and the sTfR-F index for identifying
iron-restricted erythropoiesis in patients with and with-
out APR were 2.6 – 6.9 and 4.3–16.5, respectively.
Conclusion: In APR patients, biochemical markers
demonstrate weaknesses in the diagnosis of functional
ID as defined by hematologic indices. Use of diagnostic
Laboratoriumsmedizin, Krankenhaus Nordwest, Steinbacher Hohl 2-26,
60488 Frankfurt/Main, Germany.
*Author for correspondence. Fax 49-69787390; e-mail TH-books@t-online.de.
Received November 27, 2001; accepted April 11, 2002.
1066
Hematology
plots to illustrate the relationship between the sTfR-F
index and CHr allows the progression of ID to be
identified, regardless of whether an APR is present.
© 2002 American Association for Clinical Chemistry
Iron balance is fundamentally regulated by the rate of
erythropoiesis and the size of the iron stores (1 ). Iron
deficiency (ID)1 is one of the most common nutritional
deficiencies worldwide and is the leading cause of ane-
mia, especially in children and adult women (2 ). Clinical
interest therefore focuses on (a) early recognition of sub-
clinical ID to prevent the systemic complications of this
disease (3 ); (b) identification of iron-deficient anemia
(IDA) caused by a lack of iron in the diet, by iron
malabsorption, or by increased bleeding as the underlying
reason for anemia, because this condition responds
promptly to therapy (4 ); (c) distinction between ID and
other entities in the differential diagnosis of anemia,
especially anemia that accompanies infection, inflamma-
tion, and cancer, commonly termed anemia of chronic
disease (ACD) (5 ), to distinguish ACD from the combined
state of functional ID/ACD.
ACD is characterized by inadequate production of
erythropoietin, inhibition of the proliferation of erythroid
progenitor cells in the bone marrow, and disturbances in
iron distribution (6 ). ACD results from activation of the
immune and inflammatory systems. There is increased
production of inflammatory cytokines, which in turn in-
creases the concentration of C-reactive protein (CRP) (7 ). In
ACD, like IDA, functional ID is the limiting factor of
erythrocyte hemoglobinization. Functional ID is defined as
an imbalance between the iron needs of the erythroid
1
Nonstandard abbreviations: ID, iron deficiency; IDA, iron-deficient ane-
mia; ACD, anemia of chronic disease; CRP, C-reactive protein; Hb, hemoglo-
bin; TfS, transferrin saturation; sTfR, soluble transferrin receptor; APR, acute-
phase response; ␤-thal, ␤-thalassemia; HYPO, proportion of hypochromic red
cells; CHr, reticulocyte hemoglobin content; RDW, red cell distribution width;
CRI, corrected reticulocyte index; MCV, mean cellular volume; CH, red cell
hemoglobin content; sTfR-F index, sTfR/log ferritin ratio; AUC, area under the
ROC curve; CRA, cancer-related anemia; and rHuEPO, recombinant human
erythropoietin.
 Clinical Chemistry 48, No. 7, 2002
marrow and the iron supply, which is not maintained at a
rate sufficient to allow normal hemoglobinization of the
erythrocytes. This leads to reduced reticulocyte and eryth-
rocyte cellular hemoglobin (Hb) content. In IDA, the iron
supply depends on the amount of the iron stores, whereas in
ACD, the supply depends on its rate of mobilization. In
ACD, functional ID may occur even in the presence of large
iron stores when iron release is impaired (8 ).
The hemoglobinization of erythrocytes is used to detect
functional ID because the Hb content of reticulocytes and
erythrocytes provides an evaluation of the bone marrow
activity, reflecting the balance between iron and erythro-
poiesis (3 ). However, biochemical markers measure iron
supply for the bone marrow and are only indirect indica-
tors of the balance between iron and erythropoiesis (9 ).
Traditionally, the standard biochemical markers of
iron metabolism have been serum or plasma iron, trans-
ferrin (Tf), transferrin saturation (TfS), ferritin, and more
recently, measurement of serum soluble Tf receptor
(sTfR). The diagnosis of IDA is based on the presence of
anemia and erythrocyte morphology (hypochromia, mi-
crocytosis) in conjunction with low serum ferritin, de-
creased TfS, or increased sTfR.
Diagnosis of ID associated with serum ferritin concen-
trations within established reference intervals can be
difficult in anemia that accompanies diseases in which
there is an acute-phase response (APR). Ferritin is an
acute-phase reactant and Tf a negative one (10, 11 ). In-
creased sTfR is also a useful indicator of ID. This marker
may be increased in patients who display hyperprolifera-
tive erythropoiesis (12, 13 ). Finally, heterozygous
␤-thalassemia (␤-thal trait) may mimic the usual hemato-
logic indicators of ID by producing hypochromia and
microcytosis. These diagnostic difficulties have led to
efforts to develop clinical laboratory tests that are capable
of measuring the functional iron availability at the site of
Hb synthesis in the erythron, especially the erythrocyte
and its precursors. Modern hematology analyzers, which
measure individual erythrocytes rather than mean cell
indices of the whole erythrocyte population, are able to
provide acceptable identification of small subpopulations
of iron-deficient erythrons (14, 15 ).
An hematologic index that has gained merit in assess-
ing functional iron status is the measurement of the
proportion of hypochromic red cells (HYPO). Because
erythrocytes have a lifespan of ⬃120 days, the HYPO
index is able to provide information over a several-month
period and is a late indicator of iron-restricted erythro-
poiesis (16 ). A HYPO value ⬍10% in association with low
serum ferritin is assumed to indicate that the iron supply
for erythropoiesis is maintained at a rate sufficient for
normal red cell hemoglobinization, although the body’s
iron stores may have been considerably depleted (17 ).
Reticulocyte Hb content (CHr) is an early marker of
functional ID, as reticulocytes exist in the circulation for
only 1–2 days. The usefulness of this index in monitoring
erythropoietic function is indicated by studies of the
1067
evaluation of iron status in hemodialysis patients (18, 19 ),
in the diagnosis of ID in children (3 ), and in the diagnosis
and treatment of various hematologic disorders (20 ).
The intent of this study was to evaluate the clinical
efficacy of the majority of standard biochemical markers
for ID in anemic patients to determine the best marker for
distinguishing ID from ACD and the combined state of
functional ID/ACD. The diagnosis of functional ID for all
patients was based on an examination of erythrocyte
hemoglobinization using the CHr test and the HYPO test
as the gold standard (3, 15–17, 21 ).
Materials and Methods
selection of patients and controls
Among all adult patients who were admitted to the Univer-
sity Hospital Nordwest, we selected one nonanemic and
three anemic patients every day from May to December
2000. Depending on the Hb content of the red cells, the CRP
concentration, and the physical examinations, patients were
selected for a control group, for groups of anemic patients
with and without inflammation, and for disease-specific
subgroups. We studied 602 patients who were admitted to
the following departments: Internal Medicine, Oncology,
Surgery, Gynecology and Obstetrics, Urology, and Neurol-
ogy. Patient characteristics are shown in Table 1. Diagnoses
were unknown to the analysts until the laboratory findings
were completed. Blood specimens were collected from the
patients within 24 h of their hospitalization. Of the 602
patients selected, 596 (240 men and 356 women; mean ⫾ SD
age, 64 ⫾ 15 years for the men and 54 ⫾ 22 years for the
women) were investigated for complete blood cell count, the
standard biochemical markers of iron metabolism, and CRP.
Six patients were excluded from the study because their
specimens did not fulfill the preanalytical criteria of our
clinical chemistry laboratory. Of the 596 patients, 442 (74%)
were anemic and 154 (26%) were not anemic.
It is important to apply inclusion criteria very carefully
to individuals to ensure that reference intervals for the
hematologic indices and biochemical markers of iron
metabolism are correct. Reference intervals may not be
appropriate for hospitalized patients if only entirely
healthy individuals are enrolled. The controls for this
study were selected by the following procedures:
• To establish reference intervals for CHr and HYPO, we
used samples from 71 patients (30 men and 41 women;
mean age, 55 ⫾ 19 years) who exhibited no abnormal
hematologic findings in their complete blood cell count.
Values within the reference intervals are indicative of
normal red cell Hb content (CH) and are taken to
exclude functional ID. The criteria used were Hb ⱖ140
g/L for men and ⱖ123 g/L for women (22 ); normocytic,
normochromic red cells; a red cell distribution width
(RDW) ⱕ15%; a corrected reticulocyte index (CRI)
ⱕ2.6%; a leukocyte count ⱕ10 000/␮L; and a CRP
concentration ⱕ5 mg/L (23 ).
• To establish gender-specific reference intervals for the
1068 Thomas and Thomas: Markers of Iron Deficiency

consisted of 227 patients with normal red cell hemoglo-

Table 1. Patient characteristics.
binization and 369 patients with reduced red cell hemo-
Men/women, n 240/356
globinization. The accuracy of the biochemical markers
Mean age (men/women), years 64/54
was determined according to gender and to whether the
Ethnicity Caucasian
concentration of the acute-phase reactant CRP was within
Cause of admittance to the hospital according to
medical departments, n the reference interval or increased. The most informative
Oncology biochemical markers and hematologic indices of ID were
Solid tumors 219 selected to divide the 442 anemic patients into diagnostic
Plasmacytoma, non-Hodgkin lymphoma, Hodgkin 11 plot quadrants for classifying the advancing ID phases.
lymphoma Because APR is an important influencing factor on iron
Neurology metabolism and the proliferation and hemoglobinization of
Diseases of central nervous system 50 red cells, the anemic patients were primarily divided into
Urology groups with normal and increased CRP, and not as is
Prostate adenoma 8 usually done, into IDA and ACD. Most of the anemic
Renal stone disease 8 patients with APR had either ACD or the ID/ACD combi-
Urinary tract infection 6
nation.
Heterogeneous urologic afflictions of only one cause 7
EDTA-blood specimens were collected for the complete
Gynecology and obstetrics
blood cell count and measurement of CHr and HYPO. The
Pregnancies, last trimester 47
specimens were measured within 8 h of venipuncture.
Myomatosis 12
Heterogeneous gynecologic afflictions of only one 11
Plasma specimens were used to determine the standard
cause biochemical markers of iron metabolism and CRP. Because
Surgery sTfR is considered a sensitive analyte for detection of ID
Hot abscess 8 (24 ), the reliability of this marker was tested using three
Bone fracture 6 commercially available sTfR assays for comparison.
Appendicitis 5
Heterogeneous surgical afflictions of only one cause 20 diagnostic testing
Internal medicine Blood counts were measured by assaying one specimen
Myocardial infarction and angina pectoris 12 from every patient for Hb, red blood cells, hematocrit,
Heart failure 28 mean cellular volume (MCV), CH, RDW, and white blood
Circulatory collapse 2 cells on the Advia 120 (Bayer Diagnostics) automated
Thrombosis, thromboembolism 12 hematology analyzer. Reticulocyte measurements in-
High blood pressure 4 cluded percentage of reticulocytes (% retic), mean Hb
Anemia of unknown origin 9 concentration of reticulocytes (CHCMr), and mean cell
Abdominal pain 11
volume (MCVr). Calculated indices were absolute reticu-
Gastrointestinal bleeding 13
locyte count (# retic ⫽ red blood cells ⫻ % retic), Hb
Gastroenteritis 11
content per reticulocyte (CHr ⫽ MCVr ⫻ CHCMr), and
Acute and chronic liver disease 8
total reticulocyte Hb (RetHb ⫽ CHr ⫻ # retic). Plasma Tf
Cholecystitis, cholelithiasis 7
Pancreatitis 3
(reference interval, 2.0 –3.6 g/L) (23 ) was assayed with the
Fever of unknown origin 14
Image immunonephelometer (Beckman), using calibra-
Pneumonia 9 tors assigned against the IFCC plasma protein calibrator
Chronic respiratory disease 13 (BCR CRM 470). Serum iron (reference interval, 7.2–27.7
Hyperthyreosis 2 ␮mol/L) and CRP (reference values ⱕ5 mg/L) (23 ) were
Immediate hypersensitivity reaction 2 measured using the Vitros clinical chemistry analyzer
Drug of abuse 2 (Ortho Diagnostics). Ferritin was determined on a Cobas
Endstage renal failure 5 Core analyzer (Roche Diagnostics); the reference intervals
Heterogeneous diseases of only one cause 11 in our hospital are 15–150 ␮g/L for women and 30 –350
All 596 ␮g/L for men. The sTfR concentration for each specimen
was determined using three commercially available as-
says. Two of the assays (Dade Behring and Roche Diag-
biochemical markers of iron metabolism, we used sam-
nostics) use microagglutination of latex particles coated
ples from 227 patients (100 men and 127 women; mean
with an anti-sTfR monoclonal antibody. The third test
age, 62 ⫾ 17 years for the men and 57 ⫾ 19 years for the
(Nichols Institute) uses a noncompetitive sandwich-type
women) with normal red cell hemoglobinization and no
assay with luminescence-labeled monoclonal antibodies
functional ID.
to sTfR. The Dade Behring reagent was measured on a BN
We used ROC curve analysis to evaluate the ability of ProSpect immunoanalyzer, the Roche reagent on an Hi-
biochemical markers for ID to discriminate patients with tachi 917 analyzer, and the Nichols reagent on an Advan-
functional ID from cases without ID. The patient group tage analyzer. The reference intervals (2.5–97.5 percen-
 Clinical Chemistry 48, No. 7, 2002
tiles) given by the manufacturers were 0.4 –1.8 mg/L
(Dade Behring), 0.7– 4.2 mg/L (Nichols), and 1.9 – 4.4
mg/L for women and 2.2–5.0 mg/L for men (Roche).
The CVs for interday precision near the upper bounds
of the reference intervals for HYPO, CHr, ferritin, and Tf
were 4.8%, 5.6%, 4.2%, and 3.9%, respectively. The inter-
day CV was 3.2% for sTfR (Dade), 3.6% for sTfR (Roche),
and 4.1% for sTfR (Nichols) at means of 1.6, (Dade), 3.3
(Nichols), and 3.6 mg/L (Roche).
TfS was calculated based on the formula: TfS (%) ⫽ Fe
(mg/L) ⫻ 70.9/Tf (g/L) (11 ), where Fe is the iron concen-
tration. The CRI was calculated as the reticulocyte count
(percentage of red cells) multiplied by the hematocrit and
divided by 45.
It is normal for CHr to be greater than the mean, mature
CH in individuals who do not have anemia. Patients with
inverted values (CHr ⬍ CH) showed evidence consistent
with recent development of ID (19 ). The development of an
inverted CHr/CH ratio was used as an helpful indicator of
recently developed functional ID. The ratio of sTfR to log
ferritin (sTfR-F index) was also determined because this
index has been reported to be an excellent marker for
biochemical identification of ID (13, 25 ).
Diagnosis of ␤-thal trait was facilitated by determination
of the ratio of the percentage of microcytic cells to the
percentage of hypochromic red cells. Patients with a ratio
⬎0.9 and with ⬎20% microcytic red cells were suspicious for
␤-thal trait (26 ). Column chromatography with minicol-
umns was used for confirming the identification of HbA2.
statistical analysis
Data were evaluated using standard parametric tests, and
calculations were performed with the MedCalc statistical
software package for biomedical research (27 ).
Results
correlations of tests and comparison of sTfR
methods
In the 596 patients studied, median CH was 26.8 pg [mean
(SD), 26.6 (3.4) pg] and median CHr was 27.9 pg [mean
(SD), 28.0 (4.1) pg]; both were normally distributed. There
was a close correlation between CH and CHr (Fig. 1).
Median HYPO was 4.6% [mean (SD), 10.5% (14.2%)], and
the distribution of values was positively skewed. The
three methods used to determine sTfR were compared in
376 patients and showed a close correlation (Fig. 2). The
sTfR concentration increased in hyperproliferative eryth-
ropoiesis. However, there was no correlation between CRI
and the sTfR concentration (Fig. 3).
hematologic indices in the control group
The control group consisted of 71 patients [30 men and 41
women; mean (SD) age, 55 (19) years]. Frequency distribu-
tions of RDW, CH, and CHr were normal, but those of Hb,
HYPO, and CRI were not (Table 2). The 2.5 percentile for
CHr was 28 pg, whereas the 97.5 percentiles for HYPO and
CRI were 5% and 2.6%, respectively. On the basis of cutoffs
1069
Fig. 1. Relationship between CH and CHr in 596 patients.
Equation for the line: y ⫽ 5.3 ⫹ 0.888x (r2 ⫽ 0.680).
of 28 pg for CHr and 5% for HYPO, a CHr ⬍28 pg and a
HYPO ⬎5% were used as indicators of functional ID.
biochemical markers in the absence of
functional id
Ferritin, Tf, TfS, and sTfR, the standard biochemical
markers of ID, as well as the sTfR-F index were deter-
mined in 211–227 patients with normal red cell hemoglo-
binization. Some of the markers could not be determined
in all of the patients because there was insufficient spec-
imen available. The 2.5 percentile for ferritin differed
significantly between male and female patients (P ⫽
0.0097), but not the 97.5 percentile for sTfR (P ⫽ 0.16 – 0.19,
depending on the assay used; Table 3). In male and female
patients, serum ferritin concentrations indicating ID were
⬍20.8 and ⬍11.2 ␮g/L, respectively.
sTfR concentrations in male and female patients, de-
pending on the sTfR assay, were ⬎1.8, ⬎5.5, and ⬎5.5
mg/L, respectively. Cutoffs for the sTfR-F index were not
gender specific, and sTfR-F index indicated ID at values of
⬎1.5, ⬎3.5, and ⬎3.8 in the Dade, Nichols, and Roche
assays, respectively.
biochemical markers in functional id
After eliminating results for 10 patients with ␤-thal trait,
we performed ROC analyses for ferritin, sTfR, and the
sTfR-F index to evaluate the accuracy of these tests at
identifying functional ID (CHr ⬍28 pg, HYPO ⬎5%) in
patients with and without APR. The group consisted of
154 nonanemic and 432 anemic patients. Of the anemic
patients, 288 had APR, 263 had reduced red cell hemo-
globinization, and 195 had both.
As can be seen from Fig. 4 and Table 4, the biochemical
markers displayed a greater sensitivity, specificity, posi-
tive likelihood ratio, and area under the ROC curve
(AUCROC) in patients without APR. The sTfR-F index was
the best marker for biochemical identification of func-
tional ID. Optimal cutoffs (highest sum of sensitivity and
specificity from ROC curves) for functional ID in anemic
1070 Thomas and Thomas: Markers of Iron Deficiency

Fig. 2. Relationship between different sTfR methods in 376 patients.

(A), Dade Behring vs Nichols sTfR assay: y ⫽ 2.58x ⫺ 0.13 (r2 ⫽ 0.851). (B),
Dade Behring vs Roche sTfR assay: y ⫽ 2.94x ⫺ 0.32 (r2 ⫽ 0.934). (C), Roche
vs Nichols sTfR assay: y ⫽ 0.861x ⫹ 0.25 (r2 ⫽ 0.876).

patients with reduced red cell hemoglobinization are
shown in Table 4 for ferritin, sTfR, and the sTfR-F index
[data for comparison of the markers with CHr ⱕ28 pg and
HYPO ⬎5% are available with the online version of this
article at Clinical Chemistry Online (http://www.clinchem.
org/content/vol48/issue7/]. The cutoffs were similar to
the 2.5 percentile for ferritin and the 97.5 percentiles for
sTfR and the sTfR-F index of the control group with
normal red cell hemoglobinization (Table 3). Women had
lower ferritin cutoff values than men irrespective of APR.
In these patients, based on a HYPO ⬎5%, functional ID
was indicated by a ferritin concentration ⬍97.1 ␮g/L in
men and ⬍22.9 ␮g/L in women; based on a CHr of ⬍28
pg, functional ID was indicated by ferritin concentrations
⬍61.7 and ⬍22.9 ␮g/L in male and female patients,
respectively (Table 4). The sTfR-F index criterion value
decreased from 1.5 in patients who had no APR to 0.8 in
patients with a CRP concentration ⬎5 mg/L.
ch inversion
The CHr/CH ratio was inverted in 22% of 586 patients who
did not have ␤-thal trait (Table 5). Only 3.5% of patients with
normal red cell hemoglobinization had inverted values vs
34% of those with reduced red cell hemoglobinization. In

Table 2. Markers of red cell hemoglobinization in the

control group with normal complete blood cell
count (n ⴝ 71).
95% central range (median)
Hb, g/L
Men 140–174 (145)
Women 123–153 (137)
RDW, % 12.5–14.8 (13.5)
CRI, % 0.5–2.6 (1.3)
Fig. 3. Relationship between sTfR concentration in plasma and the CRI CH, pg 25.7–32.4 (29.8)
in 596 patients. CHr, pg 27.7–35.0 (31.4)
The Dade Behring sTfR assay was used. Equation for the line: y ⫽ 1.108x ⫺ 0.05 HYPO, % 1–5 (2)
(r2 ⫽ 0.017).
 Clinical Chemistry 48, No. 7, 2002 1071

0.5–3.8 (1.6)

0.2–3.7 (1.4)

0.6–3.8 (1.6)
(n ⫽ 211)

(n ⫽ 118)
Roche

(n ⫽ 93)
Table 3. Biochemical markers of iron metabolism in patients with normal hemoglobinization of red cells (CHr >28 pg, HYPO <5%).

0.4–3.5 (1.5)

0.2–3.7 (1.4)

0.5–3.5 (1.5)
sTfR-F index

(n ⫽ 217)

(n ⫽ 121)
Nichols

(n ⫽ 93)
0.2–1.6 (0.6)

0.2–1.6 (0.6)

0.1–1.6 (0.6)
Dade Behring

(n ⫽ 227)

(n ⫽ 127)

(n ⫽ 100)
1.3–5.5 (3.0)

1.0–5.7 (3.1)

1.6–5.4 (3.0)
(n ⫽ 211)

(n ⫽ 118)
Roche

(n ⫽ 93)
95% central range (median)

1.3–5.5 (2.9)

0.6–5.7 (2.9)

1.2–5.4 (2.9)
sTfR, mg/L

(n ⫽ 214)

(n ⫽ 121)
Nichols

(n ⫽ 93)
0.5–2.1 (1.2)

0.3–2.2 (1.3)

0.5–1.8 (1.1)
Dade Behring

(n ⫽ 227)

(n ⫽ 100)

(n ⫽ 127)
20.8–2688.0 (159.5)
11.2–1921.7 (91.5)

Fig. 4. ROC plots showing the ability of ferritin, sTfR, and the sTfR-F
11.2–765.9 (88.9)
Ferritin, ␮g/L

index to indicate functional ID in patients with and without APR.

As references, CHr ⬍28 pg (A) and HYPO ⬎5% (B) were used. The solid lines
(n ⫽ 227)

(n ⫽ 100)

(n ⫽ 127)

represent patients without APR, and the dotted lines represent patients with APR
(CRP ⬎5 mg/L). The lines showing ferritin, sTfR, and sTfR-F index are red, green,
and black, respectively. The ROC plots for female patients are presented. The
AUCs for ROC curves and other test characteristics are listed in Table 4.

this group, 26% of patients without APR displayed CH

6.0–76.4 (18.2)

7.2–81.4 (20.1)

5.7–69.4 (19.3)

inversion; this increased to 38% in patients with APR. These

findings suggest that CH inversion is stimulated not only by
TfS, %

(n ⫽ 217)

(n ⫽ 122)
(n ⫽ 95)

functional ID, but also by APR.

diagnostic plots
In anemic patients, the sTfR-F index was the most accu-
1.2–3.7 (2.3)

1.1–3.4 (2.2)

1.3–4.1 (2.3)

rate marker for biochemical identification of functional

(n ⫽ 217)

(n ⫽ 122)
Tf, g/L

ID. However, as shown from the AUCROC and the posi-

(n ⫽ 95)

tive likelihood ratios in Table 4, the accuracy of the sTfR-F

index for identifying functional ID was greater in patients
with ID alone than in those with combined ID and APR.
To evaluate the performance of the sTfR-F index in
Males and
females

combination with CHr and HYPO for identification of the

Females
Males

different phases of advancing ID in anemic patients with

and without APR, we matched these markers in diagnos-
1072 Thomas and Thomas: Markers of Iron Deficiency

Table 4. Characteristics of biochemical markers of iron metabolism in comparison with CHr <28 pg.a
No infection/inflammation Infection/inflammation
(CRP <5 mg/L) (CRP >5 mg/L)

Marker Men Women Men Women

Ferritin n 62 83 93 109
Criterionb ⱕ21.3 ␮g/L ⱕ13.4 ␮g/L ⱕ61.7 ␮g/L ⱕ22.9 ␮g/L
Sensitivity, % 57.1 74.7 38.7 38.5
95% CIc 28.9–82.2 64.0–83.6 28.8–49.4 29.4–48.3
Specificity, % 90.3 92.8 94.4 92.5
95% CI 80.1–93.6 84.9–97.3 86.2–98.4 84.4–97.2
AUC 0.76 0.89 0.68 0.68
95% CI 0.65–0.85 0.84–0.94 0.60–0.75 0.61–0.74
PLR 5.9 10.3 6.9 5.1
sTfR (Dade Behring)
n 62 83 93 109
Criterionb ⱖ2.0 mg/L ⱖ1.7 mg/L ⱖ1.5 mg/L ⱖ1.8 mg/L
Sensitivity, % 78.6 83.1 67.7 67.9
95% CI 49.2–95.1 73.3–90.5 57.2–77.1 58.3–76.5
Specificity 83.9 80.7 66.2 76.2
95% CI 72.3–92.0 70.6–88.6 54.0–77.0 65.4–85.0
AUC 0.87 0.90 0.72 0.77
95% CI 0.77–0.93 0.84–0.94 0.64–0.78 0.70–0.83
PLR 4.9 4.3 2.0 2.9
sTfR-F index
n 62 83 93 109
Criterionb ⬎1.6 ⬎1.5 ⬎0.80 ⬎0.96
Sensitivity, % 71.4 81.9 61.3 65.1
95% CI 41.9–91.4 71.9–89.5 50.6–71.2 55.4–74.0
Specificity, % 95.2 92.8 76.1 81.3
95% CI 86.5–98.9 84.9–97.3 64.5–85.4 71.0–89.1
AUC 0.88 0.93 0.74 0.78
95% CI 0.79–0.95 0.88–0.97 0.66–0.80 0.72–0.84
PLR 14.8 11.3 2.6 3.5
a
Characteristics in comparison with CHr ⬍28 pg and HYPO ⬎5% are available as online data supplement (http://www.clinchem.org/content/vol48/issue7/).
b
Value corresponding with maximum sensitivity and specificity.
c
CI, confidence interval; PLR, positive likelihood ratio.

tic plots (Fig. 5). The plots, which are distributed into four
quadrants, were generated as follows: For the hemato-
logic indices, cutoff values for functional ID were CHr
⬍28 pg and HYPO ⬎5%, as determined in patients with a
normal, complete blood cell count (Table 2). The sTfR-F
index cutoffs for functional ID, selected from ROC anal-
yses of patients with IDA and combined functional ID/
ACD, were 1.5 and 0.8, respectively. The idea was that
Table 5. CH inversions (CHr < CH) in different patient groups.
Patients,
Group n n
Total without ␤-thal trait 586
Normal hemoglobinization of red cells 227
Reduced hemoglobinization of red cells 359
Reduced hemoglobinization of red cells and
CRP ⱕ5 mg/L 111
CRP ⬎5 mg/L 248
data points in quadrants 1 and 2 would be consistent with
normal red cell hemoglobinization in repleted or recently
depleted iron stores. Data points in quadrants 3 and 4
would include patients with reduced red cell hemoglo-
binization with either depleted (quadrant 3) or repleted
iron stores (quadrant 4).
The data point distribution of 432 anemic patients with
and without APR is shown in Fig. 5. Most data points are
located in quadrants 1 and 3. Quadrant 1 patients had
normocytic and normochromic red cells, but the percent-
age of patients with HYPO ⬎5% was higher (13% vs 24%)
CH in the population with APR. Most patients with normal
inversions
hemoglobinization of red cells, CRP ⱕ5 mg/L, and data
% points in quadrant 2 had either ferritin concentrations
130 22.2 ⬍20 ␮g/L or a CRI ⬎2.6%, resulting from hemolysis or
8 3.5 recently started oral iron therapy. Quadrant 2 patients
122 34.0 with a CRP ⬎5 mg/L mainly suffered from cancer-related
anemia (CRA) with reduced hemoglobinization of red
30 27.0
cells (HYPO ⬎5%). In these patients, erythropoiesis was
92 37.1
monitored after chemotherapy, at a time when erythro-
 Clinical Chemistry 48, No. 7, 2002 1073

Fig. 5. Diagnostic plots for the identification of ID in anemic patients with (left) and without (right) APR.
Data for 288 patients with APR and 154 patients without APR are presented. Data from the Dade Behring sTfR assay are shown. The sTfR-F indices separating patients
in the iron-repleted state from those in the iron-depleted state are 0.8 in patients with APR (left) and 1.5 in those without (right). E, HYPO ⬍5%; F, HYPO ⬎5%; 䡺,
inverted CHr/CH ratio; f, combination of HYPO ⬎5% and inverted CHr/CH ratio.

poiesis recovered and changed from hypoproliferation to
hyperproliferation.
Patients with data points in quadrant 3 had anemia with
biochemically and hematologically identified ID. Among
134 quadrant 3 patients with increased CRP, 102 (76%) had
a CHr ⬍28 pg and a HYPO ⬎5%, indicating that ID had
been present for several months, and 57 (43%) had an
inverted CHr/CH ratio, showing evidence of the recent
development of ID. In quadrant 3 patients with CRP within
the reference interval, 50 of 62 (81%) had a CHr ⬍28 pg and
a HYPO ⬎5%, and 20 (32%) revealed CH inversion.
Fourteen of 144 patients (10%) with CRP within the
reference interval had data points in quadrant 4. In six of
these patients, the CHr was ⬍28 pg because of CH inversion.
Of the nine patients with a combination of CHr ⬍28 pg and
HYPO ⬎5%, seven had the combined state of functional
ID/ACD, and two were suffering from refractory IDA.
Quadrant 4 contained 61 of 288 patients with APR (21%).
Combinations of CHr ⬍28 pg with HYPO ⬎5% and CHr
⬍28 pg with an inverted CHr/CH ratio were identified in 32
(57%) and 29 (48%) quadrant 4 patients, respectively. Both
combinations were typical markers in quadrant 3 patients,
with ID occurring as a result of depletion of storage and
functional iron compounds. Therefore, patients with a CRP
⬎5 mg/L and data points in quadrant 4 who met one of
these criteria (CHr ⬍28 pg and HYPO ⬎5%; CHr ⬍28 pg
and CH inversion) were classified as functionally iron defi-
cient; nevertheless, their storage compounds were repleted.
These patients were typically those with combined func-
tional ID/ACD. In contrast to patients with data points in
quadrant 4, in whom the CH inversion rate was 48%,
patients with normal red cell hemoglobinization and storage
iron compound repletion (data points in quadrant 1) had an
inversion rate of only 4.5%. This further demonstrates that
an inverted CHr/CH ratio is also a marker of functional ID
in patients with repleted iron compounds.
To investigate the selectivity of diagnostic plots for
disease-specific anemias, the distributions of patients with
IDA, CRA, ACD without CRA, and third-trimester preg-
nancy were recorded (Table 6). The disease-specific ane-
mias were selected according to clinical findings and
laboratory criteria. Thus, there were differences in the
sTfR-F index between patients with and without APR.
The patient groups were separated into subgroups based
CRP ⱕ5 mg/L or ⬎5 mg/L. Data points for patients with
IDA were localized in quadrant 3, which is indicative of
storage depletion and functional iron compounds. In
ACD without CRA and in CRA, the majority of the data
points were localized in quadrant 1, indicating normal red
cell hemoglobinization in the iron-repleted state. In-

Table 6. Distribution of disease-specific anemias in the diagnostic plots shown in Fig. 5.

Data points

IDA (n ⴝ 72) ACDa (n ⴝ 102) CRA (n ⴝ 211) Pregnancy (n ⴝ 47)

CRP CRP CRP CRP CRP CRP CRP CRP

<5 mg/L >5 mg/L <5 mg/L >5 mg/L <5 mg/L >5 mg/L <5 mg/L >5 mg/L

Quadrant n % n % n % n % n % n % n % n %
1 23 22.5 19 18.6 53 25.1 48 22.9 1 2.1 1 2.1
2 1 1.0 6 5.9 5 2.4 22 10.4 2 4.2 5 10.6
3 55 76.3 17 23.7 18 17.6 8 3.8 44 20.8 21 44.6 14 29.8
4 7 6.8 30 29.4 4 1.9 27 12.8 3 6.4
a
Excluding CRA patients.
1074 Thomas and Thomas: Markers of Iron Deficiency
creased CRP in these diseases was the cause of the data point
distributions in quadrants 3 and 4, representing decreased
red cell hemoglobinization in the iron-depleted or -repleted
states, respectively.
Approximately 65% of pregnant women displayed de-
pletion of storage and functional iron compounds (quadrant
3). In 7 of 47 (15%) pregnancies, data points were localized in
quadrant 2, indicating the increased sTfR concentration
during pregnancy, which reflects increased erythropoietic
activity in the last trimester (28 ). In the 20 pregnancies with
increased CRP concentrations, there were no cases of ID in
which iron stores had been repleted.
Of the 10 patients with ␤-thal trait, 7 had data points in
quadrant 4. Three pregnant women with ferritin concen-
trations ⬍10 ␮g/L had data points in quadrant 3.
Discussion
In functional ID, the imbalance between iron require-
ments of the erythroid marrow and the actual iron supply
leads to a reduction of red cell hemoglobinization, which
causes hypochromic, microcytic anemia. Determination of
the fraction of individual red cells with deficient hemo-
globinization has been recommended as an alternative
measurement method to the use of standard biochemical
markers, such as serum iron, ferritin, Tf, TfS, and sTfR
(15, 17, 18, 20, 29 –31 ). Functional ID is closely related to
the production of hypochromic red cells, and measure-
ment of red cell hemoglobinization provides a sensitive
method for determining the quantity of circulating iron
incorporated into the red blood cells, which reflects recent
changes in erythropoiesis (16 –18, 20, 29 –31 ). CHr is an
early and sensitive marker of iron-restricted erythropoie-
sis (functional ID), whereas the proportion of hypochro-
mic red cells is a time-averaged marker (17, 21, 32 ). Both
these indices are similar in anemic patients, like glucose
and HbA1c in diabetic patients.
The markers HYPO and CHr have been investigated
primarily by monitoring the erythropoietic function in
hemodialysis patients. A HYPO ⬎10% is consistent with
functional ID, which is an important limiting factor in the
effectiveness of recombinant human erythropoietin
(rHuEPO) therapy (17, 30, 31, 33 ). Further studies will,
however, be necessary to evaluate the role of this test (34 ).
CHr has been shown to be a sensitive and specific
indicator of functional ID in healthy individuals (29 ) and
in patients with end-stage renal failure treated with
rHuEPO (18, 19, 30, 31, 35 ). In healthy individuals admin-
istered rHuEPO, the decrease in CHr was inversely cor-
related to the baseline serum ferritin log value, which
shows that CHr is a useful early indicator of iron-deficient
erythropoiesis (29 ). In hemodialysis patients on rHuEPO
therapy, CHr was a much more stable indicator of func-
tional ID than either serum ferritin or TfS (30, 31 ). Less
use of intravenous iron was required in iron management
using CHr than in management with serum ferritin and
TfS (31 ). In one study, the usefulness of CHr in hemodi-
alysis patients was, however, found to be somewhat
limited (35 ). Studies on healthy individuals and hemodi-
alysis patients on rHuEPO therapy have shown that
HYPO and especially CHr are direct measures of iron
sufficiency (16, 29, 31 ). On the basis of these data, we used
these indices as reference tests for functional ID in anemic
patients with a broad spectrum of diseases. Although tests
for these indices are available only on analyzers from a
single manufacturer, various cutoff values for functional
ID are reported in the literature, ranging from 2% to 10%
for HYPO (16, 35 ) and from 23 to 29 pg for CHr
(3, 15, 18, 19, 31, 35 ). The corresponding values in this
study, using the 97.5 and 2.5 percentiles for the control
group, were 5% for HYPO and 28 pg for CHr. In most
studies, the ability of biochemical markers to identify
functional ID and evaluate the phases of ID from storage
to functional ID was compared with indirect methods.
This was the case in studies in which iron staining of the
bone marrow was used or iron therapy was used as the
gold standard for iron depletion (13, 25, 33, 36 ).
In this study, we used normal red cell hemoglobiniza-
tion (CHr ⱖ28 pg, HYPO ⱕ5%), a direct marker of
functional ID, as the standard. In patients with normal red
cell hemoglobinization, cutoff values for biochemical
markers indicating ID (Table 3) were in close agreement
with the data published for ferritin (13, 36, 37 ) and sTfR
(38 ). Significant gender-specific differences could be seen
for ferritin, but not for sTfR or the sTfR-F index. This was
also observed in studies in which iron staining of the bone
marrow or iron therapy was used as the standard
(13, 25, 36 ). The cutoff values for biochemical markers of
ID in patients with normal red cell hemoglobinization
(shown in Table 3) were confirmed by the criterion values
for functional ID in the ROC curves of patients with
reduced red cell hemoglobinization. However, in patients
with anemia and APR (CRP ⬎5 mg/L), cutoff values for
ferritin that indicated ID were skewed to higher concen-
trations from approximately ⬍20 ␮g/L to ⬍100 ␮g/L in
men and from approximately ⬍10 ␮g/L to ⬍20 ␮g/L in
women (Table 4). The cutoff value for sTfR did not change
significantly in patients with APR, whereas the sTfR-F
index decreased from 1.5 to 0.8. These findings do not
agree with data in the literature in which only one sTfR-F
index value was used to distinguish IDA from ACD and
the combined state of functional ID/ACD (13, 25 ). How-
ever, the use of different sTfR-F indices in the diagnosis of
ID in patients with and without APR is plausible. The
increase of serum ferritin in APR decreases the sTfR-F
index because sTfR is unaffected.
Biochemical markers have high diagnostic accuracy for
ID when iron staining of the bone marrow is used as the
criterion (13, 24 ). In our study, in which the diagnostic
accuracy of the biochemical markers was based on red cell
hemoglobinization, the results were different (Table 4).
The reason might be that investigations of stainable bone
marrow iron mainly depict disturbances of the iron me-
tabolism in which iron stores and iron turnover are
involved. In such a situation, biochemical markers of
 Clinical Chemistry 48, No. 7, 2002
ID could be expected to show a good response. However,
for reduced bone marrow activity attributable to APR-
induced disturbances of iron distribution and distur-
bances of iron utilization of the erythroid precursors,
markers such as ferritin, Tf, and sTfR do not reflect the
balance between iron and erythropoiesis.
The weak correlation between CHr and the biochemi-
cal markers sTfR and sTfR-F index suggest that the classic
iron markers are relatively insensitive in diagnosing func-
tional ID. Of the quadrant 4 patients with increased CRP
and an iron-repleted state (Fig. 5), only 15% had increased
sTfR, although 57% displayed reduced red cell hemoglo-
binization and 48% revealed an inverted CHr/CH ratio,
the indicators of functional ID. The reason for the high
percentage of sTfR concentrations within the reference
interval is the hypoproliferation of erythroid marrow in
APR. Functional ID in the presence of hypoproliferative
erythropoiesis does not cause an adequate increase in
sTfR compared with the extent of functional ID. This
might also be the reason for the inadequately low sTfR-F
index in ACD that is accompanied by APR.
Differentiation of functional ID into IDA, ACD, and
combined ID/ACD is possible only by measuring CHr,
HYPO, and CH inversion. The use of these indices in
combination with diagnostic plots identifies functional
ID. The sTfR-F index reflects the iron store status and
allows differentiation of functional ID in the iron-depleted
and -repleted states. On the basis of the present data, we
recommend the use of diagnostic plots to classify ID,
especially in patients with APR. Plots of the data points
enable iron status to be differentiated into the four cate-
gories, as shown in Fig. 6.
Data points in quadrant 1 suggest normal red cell hemo-
globinization, a condition seen mainly in patients with ACD
and CRA, and in hemodialysis patients in the absence of
functional ID. Data points in quadrant 2 indicate three
possible conditions, according to our results:
Fig. 6. Diagnostic plot indicating the correlation between the biochem-
ically indicated iron supply for erythropoiesis and CHr and HYPO.
Iron supply is depleted in cases where the sTfR-F index (x axis) is ⬎1.5 in
patients with CRP ⱕ5 mg/L and ⬎0.8 in patients with CRP ⬎5 mg/L. sTfR was
measured with the Dade assay. The corresponding sTfR-F indices are 3.5 and
1.9 in the Nichols sTfR assay and 3.8 and 2.0 in the Roche assay. CHr ⬍28 pg
and HYPO ⬎5% indicate functional ID.
1075
• Erythropoiesis has not yet initiated reduced red cell
hemoglobinization, although a reduced iron supply
situation exists (seen mainly in tumor patients and
nonanemic patients with latent ID).
• Erythropoiesis has just started normal hemoglobiniza-
tion (CHr ⬎28 pg, HYPO ⬎5%), a condition observed in
IDA patients shortly after they are started on oral iron
supplement therapy.
• Hyperproliferative erythropoiesis where the CRI can be
⬎2.6%, e.g., in acute hemorrhage, hemolytic anemia,
and in third-trimester pregnancies with increased se-
rum sTfR, but no sign of functional ID (CHr ⱖ28 pg,
HYPO ⱕ5%).
Data points in quadrant 3 suggest a reduced iron
supply for erythropoiesis as being the cause of functional
ID attributable to depleted iron stores (the typical situa-
tion in IDA). Patients with data points in quadrant 4 are
iron-repleted but have functional ID (seen mainly in
anemia accompanying infection or chronic inflammation,
and in APR that accompanies CRA).
Patients with ␤-thal trait and those with the combined
state of ID/ACD are microcytic and hypochromic, and
have data points in quadrant 4. Therefore, as a screening
test to avoid mismatching, patients with data points in
this quadrant must be investigated for ␤-thal trait by
determining the ratio of percentage of microcytic to
percentage of hypochromic red cells.
The diagnostic plot used in this study has several
advantages over other diagnostic diagrams used to detect
advancing ID and the division of anemic patients into
those with and without ACD (13, 25 ). ACD can be sepa-
rated from the combined state of functional ID/ACD. A
combination of hematologic indices and biochemical
markers of iron metabolism in the diagnostic plot can
indicate an early and sensitive erythropoietic response to
changes in the iron supply. In pregnant women with
increased sTfR, hyperproliferative erythropoiesis can be
distinguished from ID.
According to preliminary studies, the practical thera-
peutic implications of the quadrants depicted in Figs. 5
and 6 are as follows:
(a) Patients with data points in quadrants 2 and 3
should be administered oral iron supplements. The excep-
tions to this are pregnant women and patients with
reticulocytosis who have data points in quadrant 2. The
response to treatment of IDA is indicated by the data
point shifting from quadrant 3 to quadrant 2 within 10
days, and from quadrant 2 to quadrant 1 (iron-repleted
state) within 4 – 6 weeks.
(b) Anemic patients with data points in quadrants 1
and 4 can be effectively treated with rHuEPO. In patients
with data points in quadrant 4, the response-limiting
factor is functional ID. Therefore, intravenous iron sup-
plements should preferably be given concurrently with
rHuEPO. In patients with data points in quadrant 1,
rHuEPO administration is generally started without iron
1076 Thomas and Thomas: Markers of Iron Deficiency
supplementation as long as the data points remain in that
quadrant. Where there is an inadequate response, the data
point shifts to quadrant 3 or 4, and iron supplements must
be administered.
In conclusion, biochemical markers of ID are of limited
value in diagnosing functional ID, especially in diseases
with APR. The combination of the hematologic indices
CHr and HYPO with the sTfR-F index in a diagnostic plot
provides an attractive tool for diagnosis and therapeutic
monitoring of functional ID.
References
1. Finc C. Regulation of iron balance in humans. Blood 1994;84:
1697–702.
2. Looker AC, Dallman PR, Carroll MD, Gunter EW, Johnson CL.
Prevalence of iron deficiency in the United States. JAMA 1997;
27:973– 6.
3. Brugnara C, Zurakowsky D, DiCanzio J, Boyd T, Platt O. Reticulo-
cyte hemoglobin content to diagnose iron deficiency in children.
JAMA 1999;281:2225–30.
4. Ferguson BJ, Skikne BS, Simpson KM, Baynes RD, Cook JD.
Serum transferrin receptor distinguishes the anemia of chronic
disease from iron deficiency anemia. J Lab Clin Med 1992;19:
385–90.
5. Means RT Jr, Krantz SB. Progress in understanding the pathogen-
esis of anemia of chronic disease. Blood 1992;80:1639 – 47.
6. Sears D. Anemia of chronic disease. Med Clin North Am 1992;
76:567–79.
7. Spivak JL. The blood in systemic disorders. Lancet 2000;355:
1707–12.
8. Fillet G, Beguin Y, Baldelli L. Model of reticuloendothelial iron
metabolism in humans: abnormal behaviour in idiopathic hemo-
chromatosis and in inflammation. Blood 1989;74:844 –51.
9. Porter DR, Sturrock RD, Capell HA. The use of serum ferritin
estimation in the investigation of anemia in patients with rheuma-
toid arthritis. Clin Exp Rheumatol 1994;12:179 – 82.
10. Baynes RD. Assessment of iron status. Clin Biochem 1996;29:
209 –15.
11. Thomas L. Transferrin saturation. In: Thomas L, ed. Clinical
laboratory diagnostics. Frankfurt: TH-Books, 1998:275–7.
12. Cook JD, Skikne BS, Baynes RD. Serum transferrin receptor. Annu
Rev Med 1993;44:63–74.
13. Punnonen K, Irjala K, Rajarnäki A. Serum transferrin receptor and
its ratio to serum ferritin in the diagnosis of iron deficiency. Blood
1997;89:1052–7.
14. Brugnara C. Reticulocyte cellular indices: a new approach in the
diagnosis of anemias and monitoring of erythropoietic function.
Crit Rev Clin Lab Sci 2000;37:93–130.
15. d’Onofrio G, Chirillo R, Zini G, Caenaro G, Tommasi M, Micciuli G.
Simultaneous measurement of reticulocyte and red blood cell
indexes in healthy subjects and patients with microcytic and
macrocytic anemia. Blood 1995;85:818 –23.
16. Mcdougall IC, Cavill I, Hulme B, Brain B, McGregor E, McKay P, et
al. Detection of functional iron deficiency during erythropoietin
treatment: a new approach. Br Med J 1992;304:225– 6.
17. Mcdougall IC. What is the most appropriate strategy to monitor
functional iron deficiency in the dialysed patient on rhEPO ther-
apy? Merits of percentage hypochromic red cells as a marker of
functional iron deficiency. Nephrol Dial Transplant 1998;13:847–9.
18. Cullen P, Söffker J, Höpfl M, Bremer C, Schlaghecker R, Mehrens
T, et al. Hypochromic red cells and reticulocyte hemoglobin
content as markers of iron-deficient erythropoiesis in patients
undergoing chronic hemodialysis. Nephrol Dial Transplant 1999;
14:659 – 65.
19. Fishbane S, Galgano C, Langley RC Jr, Canfield W, Maesaka JK.
Reticulocyte hemoglobin content in the evaluation of iron status in
hemodialysis patients. Kidney Int 1997;52:217–22.
20. Brugnara C, Zemanovic D, Sorette M, Ballas SK, Platt O. Reticu-
locyte hemoglobin. An integrated parameter for evaluation of
erythropoietic activity. Am J Clin Pathol 1997;108:133– 42.
21. Brugnara C, Laufer Mr, Friedman AJ, Bridges K, Platt O. Reticulo-
cyte hemoglobin content (CHr): early indicator of iron deficiency
and response to therapy. Blood 1994;83:3100 –1.
22. Williams WJ, Nelson DA, Morris MW. Examination of the blood. In:
Williams WJ, Beutler E, Erslev AJ, Lichtman MA, eds. Hematology.
New York: McGraw-Hill, 1990:9 –24.
23. Dati F, Schumann G, Thomas L, Aguzzi F, Baudner S, Bienvenu J,
et al. Consensus of a group of professional societies and diag-
nostic companies on guidelines for interim reference ranges for
14 proteins in serum based on the standardization against the
IFCC/BCR/CAP reference material (CRM 470). Eur J Clin Chem
Clin Biochem 1996;34:517–20.
24. Skikne BS, Flowers CH, Cook JD. Serum transferrin receptor: a
quantitative measure of iron deficiency. Blood 1990;75:1870 – 6.
25. Suominen P, Punnonen K, Rajarnäki A, Irjala K. Serum transferrin
receptor and transferrin receptor-ferritin index identify healthy
subjects with subclinical iron deficits. Blood 1998;92:2934 –9.
26. d’Onofrio G, Zini G, Ricerca BM, Mancini S, Mango G. Automated
measurement of red blood cell microcytosis and hypochromia in
iron deficiency and ␤ thalassemia trait. Arch Pathol Lab Med
1992;116:84 –9.
27. Schoonjans F. MedCalc statistics for biomedical research. Mari-
akerke, Belgium: MedCalc Software, 1998:1–163.
28. Choi JW, Im MW, Pai HW. Serum transferrin receptor concentra-
tions during normal pregnancy. Clin Chem 2000;46:725–7.
29. Brugnara C, Colella GM, Cremins JC, Langley RC Jr, Schneider TJ,
Rutherford CJ, et al. Effects of subcutaneous recombinant human
erythropoietin in normal subjects: development of decreased
reticulocyte hemoglobin content and iron-deficient erythropoiesis.
J Lab Clin Med 1994;123:660 –7.
30. Mittman N, Sreedhara R, Mushnick R, Chattopadhyay J, Zel-
manovic D, Mehdi V, et al. Reticulocyte hemoglobin content
predicts functional iron deficiency in hemodialysis patients receiv-
ing rHuEPO. Am J Kidney Dis 1997;30:912–22.
31. Fishbane S, Shapiro W, Dutka P, Valenzuela OF, Faulbert J. A
randomized trial of iron deficiency testing strategies in hemodial-
ysis patients. Kidney Int 2001;60:2406 –11.
32. Cazzola M, Mercuriali F, Brugnara C. Use of recombinant human
erythropoietin outside the setting of uremia. Blood 1997;89:
4248 – 67.
33. Schaefer RM, Schaefer L. Hypochromic red blood cells and
reticulocytes. Kidney Int 1999;55(Suppl 69):S44 – 8.
34. Fishbane S, Maesaka JK. Iron management in end-stage renal
disease. Am J Kidney Dis 1997;29:319 –33.
35. Bhandari S, Turney JH, Brownjohn AM, Norfolk D. Reticulocyte
indices in patients with end stage renal disease on hemodialysis.
J Nephrol 1998;11:78 – 82.
36. Mast AE, Blinder A, Gronowski AM, Chumley C, Scott GM. Clinical
utility of the soluble transferrin receptor and comparison with
serum ferritin in several populations. Clin Chem 1998;44:45–51.
37. Jacobs A, Worwood M. Ferritin in serum. Clinical and biochemical
implications. N Engl J Med 1975;292:951– 6.
38. Vernet M, Doyen C. Assessment of iron status with a new fully
automated assay for transferrin receptor in human serum. Clin
Chem Lab Med 2000;38:437– 42.