Download Is zinc protoporphyrin an indicator of iron

Nephrol Dial Transplant (1996) 11: 492-497
Nephrology
Dialysis
Transplantation
Original Article
Is zinc protoporphyrin an indicator of iron-deficient erythropoiesis in
maintenance haemodialysis patients?
J. Braun1, M. Hammerschmidt1, M. Schreiber1, R. Heidler1 and W. H. Horl2
'Haemodialysis Unit NUrnberg, Ntirnberg, Germany, and department of Nephrology, University of Vienna, Vienna,
Austria
Abstract
Background. Zinc protoporphyrin (ZPP), a metabolic
intermediate generated in the red blood cell by incorporation of zinc instead of iron, has been suggested to
be a sensitive and specific parameter of absolute iron
deficiency in haemodialysis (HD) patients.
Methods. We studied 62 HD patients, 29-86 years old,
with ZPP levels > 50 umol/mol haeme (normal value
of ZPP <40 umol/mol haeme) assessing the value of
ZPP as a marker of functional iron deficiency at
different cut-off points of ZPP. None of the patients
had apparent inflammatory disease, infectious disease,
or malignancy. ZPP, haemoglobin, iron and ferritin
levels were determined before, and after a 24-week
period of once-weekly i.v. administration of 40 mg
iron, to determine whether ZPP levels return to normal
during adequate iron supplementation (960 mg iron/
patient).
Results. There was no significant change in ZPP levels
after iron supplementation in patients with a ZPP
> 50 umol/mol haeme (96.7+49.8 versus 88.4 + 43.5
umol/mol haeme before and after iron administration respectively, P=n.s.). However, in patients with
a ZPP > 90 umol/mol haeme, there was a significant
reduction in ZPP levels (141.2 + 54.5 versus 108.0 ±
48.8 umol/mol haeme, /><0.001). Serum ferritin
increased significantly in both groups. There was no
correlation between ZPP and serum ferritin at any time
during the study. There was also no correlation
between serum aluminium levels and ZPP and
no significant difference in changes in ZPP in
patients receiving desferrioxamine therapy compared
to those not receiving desferrioxamine therapy. We did
find a significant correlation between moderately elevated total blood lead concentrations and ZPP levels at
the end of the study. The ZPP levels were not significantly different in the range from 50-110 umol/mol
haeme before and after i.v. iron supplementation in
the responders (10% increase of haemoglobin or 20%
decrease of the recombinant human erythropoietin
dose) compared with the non-responders.
Conclusions. Our data indicate that ZPP cannot be
used to predict the erythropoietic response to iron
supplementation. However, ZPP levels may be an
indicator of functional iron deficiency due to blockade
of the reticuloendothelial iron release in haemodialysis
patients.
Key words: anaemia; iron deficiency; haemodialysis;
zinc protoporphyrin; lead
Introduction
Iron deficiency is a major cause of anaemia in maintenance haemodialysis (HD) patients, who may also have
an impaired response to recombinant human erythropoietin (rHuEpo) therapy. To date, measurement of
serum ferritin has been the most powerful test in
diagnosing iron deficiency, but it must be interpreted
differently for patients with inflammatory diseases.
Measuring the percentage of hypochromic red blood
cells in the circulation has already been established as
a means of assessing the adequacy of iron supplied to
the marrow [1]. Measurement of zinc protoporphyrin
(ZPP), a metabolic intermediate that is generated in
the red blood cells of patients with lead toxicity and
iron-deficiency anaemia by incorporation of zinc
instead of iron, has been proposed as a sensitive and
specific screening test for iron deficiency. ZPP was
found to be elevated in HD patients treated with
rHuEpo and whose serum ferritin levels were < 50 ug/1
[2]. Hastka et al. concluded that normal ferritin levels
in HD patients do not exclude iron deficiency, and
therefore ZPP would be useful in deciding who should
receive iron supplementation [2]. Two further studies
tested the sensitivity and specificity of ZPP as an assay
of iron deficiency in HD patients. Fishbane and Lynn
[3] considered HD patients with ZPP > 90 umol/mol
Correspondence and offprint requests to: Priv.-Doz. Dr J. Braun, haeme, or ferritin < 50 ng/ml, to be iron-deficient
KfH Dialysezentrum NOrnberg, Virnsbergerstr. 43, D-90431 (normal value of ZPP <40 umol/mol haeme) and
NUrnberg, Germany.
found ZPP > 90 umol/mol haeme to be the most
1
1996 European Dialysis and Transplant Association-European Renal Association
Zinc protoporphyrin an indicator of iron-deficient erythropoiesis
sensitive (90%) and specific (87%) indicator for iron
deficiency in HD patients. In contrast, Ahmed et al.
[4] found that ZPP was the most specific and the least
sensitive test in 22 HD patients (8 patients were
HFV+), divided into responders (10% increase of
haematocrit over 2 months) and non-responders to
rHuEpo.
We therefore investigated a group of 62 HD patients
with ZPP levels > 50 umol/mol haeme, without apparent inflammatory disease, infectious disease, or malignancy, to assess the value of different cut-off points of
ZPP as a marker of functional iron deficiency. In
addition we examined whether elevated ZPP levels
returned to normal following treatment of haemodialysis patients with i.v. iron.
Subjects and methods
We measured ZPP levels in 148 HD patients without apparent
inflammatory disease, infectious disease, or malignancy.
Exclusion of malignancy or inflammatory disease was based
on lack of signs and symptoms. All patients were negative
for antinuclear antibodies, c-ANCA, p-ANCA and rheumatic
factor and had a normal white blood cell count.
There were 62 HD patients (31 male, 31 female) with ZPP
levels > 50 umol/mol haeme who completed the study. These
patients were 29-86 years of age (60+15 years, m + SD) and
had been on maintenance HD from 14 to 184 months
(59 + 37 months, m±SD) without surgical treatment, gastrointestinal blood loss, or unusual blood loss during the study.
Some of the patients had been on oral or i.v. iron therapy
prior to the start of the study. During the study period, each
patient received an intravenous dose of 40 mg iron III sodium
gluconate complex (3.2 ml ampoule Ferrlecit) for 24 weeks,
once a week after the end of the HD session, resulting in a
total i.v. iron load of 960 mg within 6 months. ZPP, serum
iron (Fe), serum ferritin (Fer) and haemoglobin (Hb) were
measured at the beginning (ZPP1, Fel, Ferl, Hbl), 12 weeks
(ZPP2, Fe2, Fer2, Hb2), and 24 weeks after the start of the
study (ZPP3, Fe3, Fer3, Hb3). Vitamin B12, folic acid, and
aluminium were measured before the start of the study. All
patients included in the study had normal folic acid (normal
range 4-40nmol/l) and normal vitamin B12 levels (normal
range 120-700 pmol/1) and were negative for faecal occult
blood loss, as determined by three consecutive tests. Blood
lead concentrations were measured at the end of the study.
Of the 62 patients, 12 were treated with intravenous
desferrioxamine due to elevated aluminium levels and
enhanced mobilization of aluminium after a 0.5 g desferrioxamine (Desferal) test dose. To avoid elimination of iron with
desferrioxamine, 0.5-2.0 g desferrioxamine was administered
once a week at the end of HD therapy and i.v. iron was
administered following the next HD therapy.
ZPP measurement was performed in a front-face haematofluorometer (AVIV Biomedical Company, Lakewood, NJ)
with washed erythrocytes to remove interference from EDTA
blood [5]. The haematofluorometer measures the ZPP-haeme
ratio as a function of the amount of light emitted by ZPP at
594 nm versus the amount of excitatory light absorbed by
haemoglobin at 420 nm. The instrument was calibrated to
measure ZPP in umol/mol haeme. ZPP in non-anaemic
healthy volunteers from our laboratory and the dialysis unit
was 22.7±9.8 umol/mol haeme (m + SD, « = 25).
Serum ferritin was measured by enzyme-linked immuno-
493
adsorbent assays (normal range in non-renal patients
30-300 ug/1 for men and 25-150 ug/1 for women). Total
blood lead concentrations (normal range < 350 ug/1) and
serum aluminium concentrations (normal range < 10 ug/1)
were measured by use of atomic absorption spectrometry in
the laboratories of Dr Schmidt-Gayk in Heidelberg.
Statistics
We used SPSS software (SPSS Inc., Chicago, Illinois) for
data base management and statistical analysis. The statistical
techniques used were linear regression analysis, Levene test
of variance, and the Mann-Whitney U test as a nonparametric test. P<0.05 was considered statistically
significant.
Results
The mean ZPP level in 148 HD patients without
inflammatory disease or malignancy was 76.40 +
49.74 umol/mol haeme (m + SD) (Figure 1). There
is clearly a right-sided tail, reflecting the generally
higher ZPP levels in these patients.
In the 62 patients with ZPP levels > 50 umol/mol
haeme who completed the study, there was no reduction of ZPP (96.7+49.8 versus 88.4±43.5 umol/mol
haeme), whereas serum ferritin increased significantly
(167.0+138.7 versus 280.2 + 202.4 ug/1, P<0.001)
after 24 weeks of iron supplementation. There was no
significant correlation between ZPP levels and serum
ferritin levels at any time during the study. In 35 of
the 62 patients we found ZPP3<ZPP1 and in 27
patients, ZPP3>ZPP1, whereas serum ferritin
increased significantly in both these groups. Only in
those patients with ZPP levels > 90 umol/mol haeme
did i.v. iron supplementation lead to a significant
reduction of ZPP (141.2 ±54.5 versus 108.0 + 48.8
umol/mol haeme, P<0.001). Table 1 summarizes the
parameters of the patients receiving rHuEpo (« = 36)
compared with those not receiving rHuEpo (n = 26).
We reduced the rHuEpo in those patients whose
haemoglobin increased during iron supplementation to
20.0
60.0
40.0
»0.0
100.0
140.0
100.0
220.0
260.0
300.0
340.O
120.0
160.0
200.0
240.0
290.0
320.0
ZPP(pmd/mol Haeme)
Fig. 1. Distribution of zinc protoporphyrin (ZPP) levels in 148
haemodialysis patients without apparent inflammatory or infectious
disease or malignancy.
J. Braun et al.
494
Table 1. Values ±SD of 36 HD patients on rHuEpo therapy compared with 26 HD patients without rHuEpc> therapy
Non-rHuEpo patients (n = 26)
rHuEpo patients (n = 36)
Hb (g/dl)
Fe (ug/1)
Ferritin (ug/1)
ZPP (u.mol/mol haeme)
Baseline
24 weeks
Baseline
24 weeks
9.8 + 0.7
51.4±18.5
208.0 ±162.4
92.9±41.3
10.0+1.0 (P = n.s.)
64.2 + 21.3 (P<0.05)
339.3±221.1 (.P<0.001)
88.3 ±39.7 (/>=n.s.)
11.7 + 1.2
46.9 + 17.5
109.6 + 67.1
99.7 + 59.0
12.2 + 1.3 (P = n.s.)
53.0+15.5 (P = n.s.)
200.3 + 140.1 (P<0.001)
83.5+42.9 (/> = n.s.)
Significances are calculated for the differences between the parameters at baseline and after 24 weeks of i.v. iron supplementation.
keep haemoglobin nearly constant, at a target value of
10 g/dl. The rHuEpo doses after 24 weeks of iron
supplementation decreased significantly (P< 0.001)
from 4638±2319IU/week (mean 73 IU/kg body
weight/week) to 3166+1889 IU/week (mean 50 IU/kg
body weight/week). There was no reduction of ZPP,
whereas serum ferritin increased significantly.
In the responders (10% increase in haemoglobin or
20% decrease in rHuEpo dose), the initial haemoglobin level was significantly lower (9.8 + 0.9 versus
11.1+ 1.4 g/dl, P < 0.001) and the duration of HD
treatment was significantly shorter than in the nonresponders (69.3 + 38.8 versus 46.8 + 32.6 months,
P<0.05). Stepwise analysis of different ZPP cut-offs,
from 50 to 110 umol/mol haeme, revealed that there
was a significant difference only in the initial haemoglobin levels between responders and non-responders,
but there were no significant differences in the ZPP
levels at the start (Table 2) and at the end of i.v. iron
supplementation (Table 3), which would have been
useful in predicting the erythropoietic response.
However, there was a significant inverse correlation
between the relative decrease in ZPP levels and the
relative increase in serum ferritin levels (r = 0.30, P =
0.017). Of the 62 patients with serum aluminium
concentrations of 34.0 + 33.2 ug/1 (m + SD), 12 were
on desferrioxamine therapy at the beginning of the
study. We found no significant reduction of ZPP either
in the desferrioxamine-treated patients (94.1+26.5
versus 83.1 ±32.6 umol/mol haeme, P=n.s., n = 12) or
in the non-treated patients (95.9 + 52.9 versus
87.3+42.8 umol/mol haeme, P = n.s., n = 50), whereas
the serum iron and serum ferritin levels increased
significantly in both groups. Haemoglobin was kept
constant in the desferrioxamine treated patients and
increased from 10.56 + 1.31 to 11.07 +1.56 g/dl
(P<0.05) in the non-treated patients.
100
200
300
400
Total Blood Lead ((jg/1)
Fig. 2. Correlation between the total blood lead concentration and
ZPP level at the end of the study (r = 0.382, P = 0.002, n = 62, 95%
confidence intervals).
and the reduction of rHuEpo dose were not significantly different between the two groups.
Discussion
It has previously been shown that ZPP is a valid
parameter in assessing iron deficiency in both healthy
blood donors [6] and HD patients [2]. Therefore we
selected 62 HD patients with ZPP levels > 50 umol/mol
haeme to study the effect and time course of weekly
i.v. iron administration on ZPP levels. All patients
were without apparent inflammatory disease, infectious
disease or malignancy as it is well known that ZPP
levels are elevated in these disease states [7]. It was
unfortunate that 12 of the 62 HD patients with a ZPP
level > 50 umol/mol haeme had some evidence of
aluminium overload and required desferrioxamine, as
there is clearly an interaction between aluminium and
There was a correlation between the whole blood iron. However, there were no other differences between
lead concentrations and the ZPP3 levels at the end of the desferrioxamine-treated patients and the nonthe study. Mean total blood lead concentration in the treated HD patients.
62 patients was 190.91 ±57.56 ug/1 (m±SD)
The elimination half-life of ZPP from the blood
(Figure 2).
compartment is closely related to the half-life of erythWe found significantly lower ZPP levels in those rocytes and the production rate of new ZPP in the
patients with total blood lead concentrations bone marrow compartment. The half-life of erythro< 200 ug/1 (« = 40) than in those patients with cytes in HD patients has been found to be shorter
higher lead concentrations (86.2 ±38.1 versus than in non-renal patients [8]. The study period of 6
115.5 ±65.8 umol/1, P<0.05). Haemoglobin, ferritin months, therefore, was sufficient to achieve significant
Zinc protoporphyrin, serum ferritin, serum iron, and haemoglobin levels (a) in responders (b) and non-responders to i.v. iron supplementation at the start of the study
ZPP 1
mol
Non-responders
Fer 1
Responders
Fe 1
Hb 1
Non-responders
Responders
Non-responders
Responders
Non-responders
R
98.4 + 58.5
94.7 ±39.4
199.3 ±169.2
132.6 + 86.4
50.9 ±18.4
46.5±17.9
11.1 + 1.4
9
122.2 + 63.0
113.4±38.4
213.5±193.6
127.0±75.6
50.9 + 18.8
42.1 ±12.0
11.3 + 1.4
9
144.1+69.1
137.9±32.8
219.6 + 233.3
134.3 + 79.3
47.7 + 20.5
44.1 + 11.3
11.2±1.5
9
181.7 + 77.1
146.7 ±29.4
257.4±313.5
141.6 + 76.4
46.1 ±15.5
42.6±11.7
11.6±1.8
9
s. n = 30
vs. n=19
s. n = ll
n=9
n values + SD; *P<0.001; **i><0.01; ***/><0.05 compared to non-responders.
increase in haemoglobin or 20% decrease in recombinant human erythropoietin dose.
Zinc protoporphyrin, serum ferritin, serum iron, and haemoglobin levels (a) in responders (b) and non-responders to i.v. iron supplementation at the end of the study
Fer 3
ZPP 3
ol
s. H =
Non-responders
Responders
Fe3
Hb3
Non-responders
Responders
Non-responders
Responders
Non-responders
R
85.0 + 46.2
92.0 ±40.9
320.5 + 228.9
235.7 + 160.7
59.5+16.4
56.3 ±25.8
10.9±1.5
1
102.2 + 50.7
106.3 ±42.2
346.0 ±238.3
253.9± 173.2
60.0 ±17.2
55.6±23.3
11.1 ± 1.8
1
109.4 + 60.2
106.4 + 33.3
359.3 + 274.7
281.0+198.8
59.3+20.3
61.0±27.7
10.9±2.0
1
127.4 + 74.3
112.7 + 31.0
316.4 + 317.6
245.3+98.7
52.2 + 22.2
60.2 + 30.2
11.7 + 2.1
1
30
VS. M=19
. n=ll
n=9
n values ±SD. None of the differences were statistically significant.
increase in haemoglobin or 20% decrease in recombinant human erythropoietin dose.
496
changes in ZPP levels if the production rate of ZPP
was decreased due to iron supplementation.
Currently we do not know all the factors influencing
ZPP synthesis in HD patients. Patients undergoing
maintenance HD therapy have signs of T-cell activation and increased production of interleukin 1, tumour
necrosis factor, and interleukin 6. Those cytokines
have been incriminated in acute and chronic inflammatory reactions associated with HD [9]. The elevated
ZPP levels might therefore be a consequence of these
inflammatory processes and the presence of uraemic
factors that could potentially inhibit ferrochelatase.
Indeed, several abnormalities of the haeme biosynthesis
pathway have been observed in patients with chronic
renal failure [10].
The observation that whole blood lead levels correlate with ZPP indicates that this parameter must be
considered as a confounding factor when interpreting
the results.
The linear correlation between the moderately elevated total blood lead concentrations found in HD
patients in this study and the ZPP3 levels does not
imply a cause-and-effect relationship, but elevated lead
concentrations may contribute to the disturbed
erythropoiesis.
Steassen et al. [11] stated that renal insufficiency by
itself might be responsible for the increased blood lead
concentrations resulting in increased ZPP levels.
Colleoni et al. [12] found significantly higher mean
blood lead concentrations in HD patients than in
healthy subjects. In contrast, two studies [13,14] found
that the lead burdens in renal failure patients were not
different from those patients with normal renal function and no unusual exposure to lead. The sensitivity
of ZPP as a screening test for lead poisoning is at the
moment a matter of controversy [15].
Iron deficiency seems to be an additional contributor
to increased ZPP levels in those patients with ZPP
levels > 90 umol/mol haeme, whereas in patients with
ZPP levels < 90 umol/mol haeme ZPP levels remain
unchanged during iron supplementation. Our results
would support those of Fishbane and Lynn [3,16],
who considered HD patients with ZPP levels
> 90 umol/mol haeme and serum ferritin levels
<100ng/nl, to be iron deficient. These investigators
found that among patients with ferritin higher than
100 ng/ml but ZPP >90 umol/mol haeme, seven of 10
patients (70%) responded positively to therapy with
i.v. iron (sustained 5% increase of haematocrit or a
decrease of rHuEpo dose of more than 2000 IU per
treatment). In our patients with ZPP > 90 umol/mol
haeme, a positive response was noted in 11 of 24
(45%), but we also found positive reactions in 30 of
62 (48%) patients with ZPP levels > 50 umol/mol
haeme.
We regarded patients whose haemoglobin increased
by 10%, or whose rHuEpo dose was reduced by 20%,
as positive responders to i.v. iron supplementation.
ZPP levels were not significantly different in responders
and non-responders at the beginning of the study; a
finding that was independent of the ZPP cut-off in the
J. Braun et al.
range from 50 to 110 umol/mol haeme. Therefore ZPP
concentration cannot be used to predict responders
and non-responders to iron supplementation and does
not provide additional information that would allow
optimization of iron therapy on an individual level.
Red cell protoporphyrin was studied in 1983 by
Moreb et al. [17] prior to the introduction of rHuEpo.
These authors found that red cell protoporphyrin was
not predictive of iron deficiency. In our study, there
was also no difference between initial ZPP levels and
those measured after i.v. iron supplementation in
rHuEpo-treated and non-treated patients. Fishbane
and Lynn [16] overestimated the sensitivity of ZPP
as an indicator of absolute iron deficiency. They
selected iron-depleted patients (serum ferritin levels
< 100 ng/ml) with ZPP >90 umol/mol haeme, and
therefore the positive predictive value of 83% for
determining ZPP level changes in response to intravenous iron can only be valid for this particular group of
patients.
Elevated ZPP levels in HD patients with adequate
iron stores may indicate a functional iron deficiency
due to a blockade of reticuloendothelial iron release
[18], or a disturbed activity of ferrochelatase, an
enzyme which catalyses the addition of iron to
protoporphyrin.
From our results we conclude that nearly 50% of
the HD patients with ZPP levels > 50 umol/mol haeme
in the presence of adequate iron stores (serum ferritin
> 100 ng/ml) will respond to i.v. iron supplementation.
Thus i.v. iron supplementation may overcome functional iron deficiency in these patients. However, ZPP
does not seem to be a suitable parameter in deciding
when to stop i.v. iron supplementation in HD patients
with functional iron deficiency, as the elimination halflife of ZPP from the blood compartment is too long
to allow one to make this decision, and the production
rate of ZPP may be influenced by unknown factors
which are independent of iron stores.
References
1. Macdougall I, Cavill I. Detection of functional iron deficiency
during erythropoietin treatment: a new approach. Br Med J
1992; 304: 225-228
2. Hastka J, Lassere JJ, Schwarzbeck A, Hehlmann R, Strauch M.
Zinkprotoporphyrin als Alternative zu Ferritin bei Steuerung
der Eisensubstitution erythropoietinbedilrftiger Dialysepatienten. Nieren-Hochdruckkrankheiten 1991; 12: 697-700
3. Fishbane S, Lynn RI. The utility of zinc protoporphyrin (ZPP)
as an assay for iron deficiency in hemodialysis patients J Am
Soc Nephrol 1993; 4: 425
4. Ahmed U, Fadia A, Baskin S, Lasker N. What is the best
laboratory indicator of iron availability in hemodialysis patients.
J Am Soc Nephrol 1993; 4: 423
5. Hastka J, Lassere JJ, Schwarzbeck A, Strauch M, Hehlmann R.
Washing erythrocytes to remove interferents in measurement of
zinc protoporphyrin by front hematofluorometry. Gin Chem
1992; 38: 2184-2189
6. Schifman RB, Rivers SL, Finley PR, Thiers C. RBC zinc
protoporphyrin to screen blood donors for iron deficiency
anemia. JAMA 1982; 248: 2012-2015
7. Hastka J, Lassere JJ, Schwarzbeck A, Strauch M, Hehlmann R.
Zinc protoporphyrin in anemia of chronic disorders. Blood 1993;
81: 1200-1204
Zinc protoporphyrin an indicator of iron-deficient erythropoiesis
8. Eschbach JW. The anemia of chronic renal failure: pathophysiology and the effects of recombinant erythropoietin. Kidney Int
1989; 35: 134-148
9. Descamps-Latscha B, Herbelin A. Long term dialysis and cellular immunity: A critical survey. Kidney Int 1993; 43 [suppl 41]:
S135-S142
10. Fontanellas A, Coronel F, Santos JL et al. Heme biosynthesis
in uremic patients on CAPD or hemodialysis. Kidney Int 1994;
45: 220-223
11. Staessen JA, Lauwerys RR, Buchet JP et al. Impairment of
renal function with increased blood lead concentrations in the
general population. N EnglJ Med 1992; 327: 1515-1516
12. Colleoni N, Arrigo G, Gandini E, Corigliano C, D'Amico G.
Blood lead in hemodialysis patients. Am J Nephrol 1993; 13:
198-202
13. Batuman V, Wedeen RP. Impairment of renal function with
497
14.
15.
16.
17.
18.
increasing blood lead concentrations. N Engl J Med 1992;
327: 1394
Emmerson BT. Lead stores in patients with renal insufficiency.
Nephron 1991; 58: 233-234
Turk S, Schonfeld DJ, Cullen M, Rainey P. Sensitivity of
erythrocyte protoporphyrin as screening test for lead poisoning.
N Engl J Med 1992; 326: 137-138
Fishbane S, Lynn RI. The utility of zinc protoporphyrin for
predicting the need for intravenous iron therapy in hemodialysis
patients. Am J Kidney Dis 1995; 25: 426-432
Moreb J, Popovetzer MM, Friedlander MM, Konijin AM,
Hershko C. Evaluation of iron status in patients on chronic
hemodialysis: relative usefulness of bone marrow hemosiderin,
serum ferritin, transferrin saturation, mean corpuscular volume,
and red cell protoporphyrin. Nephron 1983; 35: 196-200
Krantz SB. Pathogenesis and treatment of the anemia of chronic
disease. Am J Med Sci 1994; 307: 353-359
Received for publication: 22.12.94
Accepted in revised form: 23.10.95