Scand J Haernatol(l975) 14, 385-392
Immunoradiometric Assay for Ferritin in Human Serum M. J. LEYLAND, M.D., P. C. GANGULI, M.D., D. BLOWER,B.SC. (Hons) & I. W. DELAMORE, M.D., PH.D. University Departments of Clinical Haeinatology (Chief, I . W . Delamore) and Surgery (Chief, 1. E . Giflespie), Royal Infirmary, Manchester, England
A sensitive, specific and precise immunoradiometric assay for ferritin has been developed. Ferritin was measured in the serum of 160 hospital controls, 101 females (118 ? 9 pgfl) and 59 males (189 ? 16 ,uug/l). This difference was statistically significant. In 28 patients with untreated iron deficiency anaemia, serum ferritin concentration (6.1 f 0.7 pg/l) was significantly lower than in the controls, but it was within the normal range in 14 cases of polycythaemia Vera treated by repeated phlebotomy. In 4 patients with primary haemachromatosis (2884 rf: 56 pg/l), 25 with secondary iron overload states (5702 k 1235 ,ug/l) and 8 with haemdytic anaemia (1612 & 605 pg/l), serum ferritin levels were markedly elevated. In 14 cases of transfusional siderosis there was a highly significant correlation between serum ferritin concentration and units of blood transfused. A circadian rhythm in serum ferritin concentration was observed in 7 healthy subjects. Key words: ferritin - serum - haemochromatosis - polycythaemia Vera - haemolysis
Accepted for publication October 5, 1974 Correspondence to: Dr. M. J. Leyland, Department of Haematology, University of Liverpool, Ashton Street, Liverpool, England
Ferritin is the intracellular storage form of iron and it is found chiefly in the cytoplasm of the cells of the reticuloendothelial system (Crichton 1971). Previous workers (Reissmann & Dietrich 1956, Beamish et al 1971) using insensitive immunodiffusion techniques, showed that serum ferritin was elevated in iron overload states, Hodgkin’s disease and acute myeloid leukaemia; however, these authors were unable to measure serum ferritin in normal subjects or patients with iron deficiency anaemia. Recently sen-
sitive immunoradiometric assays for femtin has been developed (Addison et a1 1972, Lipschitz et a1 1974, Siimes et a1 1974). We report here the serum ferritin concentration of normal subjects, patients with iron deficiency anaemia, iron overload states, the correlation between serum ferritin concentration and units of blood transfused in patients with transfusional siderosis and the circadian rhythm of serum concentration in heaIthy subjects.
M. J . LEYLAND, P. C. GANGULI, D. BLOWER & I. W. DELAMORE
386
Number
Group
Haemoglobin (g/100 ml) Mean 2 SE
Controls
Men Women Anaemia (sideropenic) * Polycythaemia Vera** Haemolytic anaemia Haemachromotosis Sideroblastic anaemia Transfusional siderosis Aplastic anaemia
*
Untreated.
**
59 101 28 14 8 4 6 7 12
14.920.2 13.420.1 8.6T0.3 15.4k0.4 9.7k 1.2 13.821.0 9.6 k 1.3 8.3 k 0.7 9.1 2 0.7
I
Serum iron (ug/100 ml)
Range
Mean ‘I. SE
12.9-16.9 11.8-17.6 4.6-11.5 13.6-19.1 4.9-14.4 12.6-14.6 4.7-14.0 5.7-11.2 4.7-12.5
99k 6 90+ 5 185 2 36+ 3 160 k 41 2382 8 2302 52 289k 24 192 k 41
I
Transferrin saturation (‘%)
Range
Mean 2 SE
56-142 45-135 O L 35 20- 55 80-315 223-260 85-460 150-345 290-680
31.62 2 30.5k 1.4 4.62 0.5 8.0+ 0.5 47.8 2 10.5 86.22 3.6 68 k 9.0 83.0L 7.0 47.0+ 7.0
1
Range 16-53 12-65 0-10 6-11 22-85 78-94 35-92 44-96 24-90
Treated by repeated phlebotomy.
PATIENTS AND METHODS Patient7 The clinical, haematological and biochemical par ticulars of the patients in this report are shown in Table I. Male and female hospital controls with no evidence of anaemia or iron overload states were studied. Serum ferritin was measured in patients with untreated iron deficiency anaemia, with poblycythaemia Vera treated by repeated phlebotomy, with haemolytic anaemia and with a variety of iron overload states (primary haemachromatosis, sideroblastic anaemia, transfusional siderosis and aplastic anaemia). The correlation between units of blood transfused and serum ferritin concentration was examined in 14 patients with refractory anaemia and secondary transfusional siderosis. The circadian rhythm in serum ferritin concentration was studied in 7 healthy volunteers; for this investigation blood samples were orbtained at 0800, 1200, 1600, 2000, 2400 and 0400 h, and food was taken at 0815, 1215 and 1815 h. Methods The haemoglobin and blood counts were estimated by a Coulter S analyser. The serum iron, total iron binding capacity and liver function tests were measured by an autoanalyser technique. In patients with iron deficiency anaemia or iron overload states, the bone marrow aspirates were stained for iron with Perl’s reagent.
Human ferritin was prepared, its concentration determined and purity confirmed by established methods (Beamish et a1 1971). Six New Zealand White rabbits were immunised three times at monthly intervals and bled after another four weeks; all the animals yielded ferritin antibodies of high avidity and titre. A horse ferritin immunoadsorbent was prepared and used for the separation of bound from free antibody (Miles & Hales 1968). Ferritin antibodies were iodinated with Iodine-125 (Amersham, England) at pH 7.5 by the chloramine-T method (Hunter & Greenwood 1962) and the labelled antibodies were stored bound to immunoadsorbent at -20° C. The assay was performed according to the method of Addison and his colleagues (1972). Serum obtained from patients was stored at -20O C for up to six months or until assayed. The effects of dilution of serum, storage, anticoagulants and bacteriostatic agents on the immunoradiometric assay were tested. The specificity of the ferritin measurement was investigated by examining the immunological cross-reaction between ferritin, serum albumin, immunoglobulins, transferrin and nl-foeto-protein. The precision of the assay was estimated by measuring the ‘within’ and ‘between’ assay variations. Statistical analysis was performed using Student’s ‘t’ test for unpaired data and calculating correlation coefficients (Snedecor & Cochran 1967). The data for circadian rhythm of serum ferritin was expressed as fluctuation around each subject’s mean value for the 24 h period (Acland & Gould 1956).
HUMAN FERRITIN IMMUNORADIOMETRIC ASSAY RESULTS
Immunoradiometric assay
In the absence of ferritin, 70 % of the labelled antibody was bound by the immunoadsorbent and the linear portion of the curve covered the range 0.8 to 14 pgll ferritin. As the maximum concentration of serum incorporated in the assay was 42 %, the sensitivity of this method is 2 p g ferritin per 1 of serum. The precision of the assay at this level was F 4 % (coefficient of variation). The correlation coefficients for the ‘within’ assay and ‘between’ assay variations were r = 0.979 (n = 12, P < 0.001) and r = 0.927 (n = 19, P < 0.001). Dilutions of serum between 1 in 10 and 1 in 200, storage at -20’ C for six months, thawing and freezing six times, haemolysis with up to 3.5 g free haemoglobin per 100 ml, heparin, calcium oxalate or citrate in the usual amounts as anticoagulants, rner-
thiolate (1 in 10,000) and sodium azide (2 5%) did not interfere with the assay, but ethylene diamine tetra-acetic acid (EDTA) did. Ferritin concentrations in the serum and plasma prepared from the same sample of blood were not significantly diflerent. Serum albumin, immunoglobulins A, G and M, transferrin and al-foetoprotein did not cross-react with ferritin. Controls The observed and calculated frequency distribution of the serum ferritin concentration (log scale) of the control subjects is shown in Figure 1. The mean ferritin concentration in males, 189 k 16 pgll (mean k SE) and in females, 118 rt 9 pgll differed significantly (P < 0.005). In this study the range of serum ferritin concentration was 16 to 408 pgll for females and 15 to 462 pgll for males. In these subjects there was
L
ar n E
Figure 1. Frequency distribution of serum ferritin concentration (log. scale) in 160 controls. Histogram: Observed frequency distribution. Curve: Calculated freauencv .~ distribution.
I
387
log s e r u m f e r r i t i n c o n c e n t r a t i o n
388
M. J. LEYLAND, P. C. GANGULI, D. BLOWER & I. W. DELAMORE
n o correlation between ferritin level and the age, haemoglobin, MCV, MCH, serum iron or transferrin concentration.
MCV (r = 0.642, P < 0.01) and MCH (r = 0.560, P < 0.05). Primary haemachromotosis
Iron deficiency anaemia Serum ferritin concentration in these untreated patients was 6.1 31 0.7 pg/1 (Figure 2), this was significantly different from the controls (P 0.001). In this group serum ferritin level was not significantly different between males and females.
Immunoradiometric Assay for Ferritin in Human Serum M. J. LEYLAND, M.D., P. C. GANGULI, M.D., D. BLOWER,B.SC. (Hons) & I. W. DELAMORE, M.D., PH.D. University Departments of Clinical Haeinatology (Chief, I . W . Delamore) and Surgery (Chief, 1. E . Giflespie), Royal Infirmary, Manchester, England
A sensitive, specific and precise immunoradiometric assay for ferritin has been developed. Ferritin was measured in the serum of 160 hospital controls, 101 females (118 ? 9 pgfl) and 59 males (189 ? 16 ,uug/l). This difference was statistically significant. In 28 patients with untreated iron deficiency anaemia, serum ferritin concentration (6.1 f 0.7 pg/l) was significantly lower than in the controls, but it was within the normal range in 14 cases of polycythaemia Vera treated by repeated phlebotomy. In 4 patients with primary haemachromatosis (2884 rf: 56 pg/l), 25 with secondary iron overload states (5702 k 1235 ,ug/l) and 8 with haemdytic anaemia (1612 & 605 pg/l), serum ferritin levels were markedly elevated. In 14 cases of transfusional siderosis there was a highly significant correlation between serum ferritin concentration and units of blood transfused. A circadian rhythm in serum ferritin concentration was observed in 7 healthy subjects. Key words: ferritin - serum - haemochromatosis - polycythaemia Vera - haemolysis
Accepted for publication October 5, 1974 Correspondence to: Dr. M. J. Leyland, Department of Haematology, University of Liverpool, Ashton Street, Liverpool, England
Ferritin is the intracellular storage form of iron and it is found chiefly in the cytoplasm of the cells of the reticuloendothelial system (Crichton 1971). Previous workers (Reissmann & Dietrich 1956, Beamish et al 1971) using insensitive immunodiffusion techniques, showed that serum ferritin was elevated in iron overload states, Hodgkin’s disease and acute myeloid leukaemia; however, these authors were unable to measure serum ferritin in normal subjects or patients with iron deficiency anaemia. Recently sen-
sitive immunoradiometric assays for femtin has been developed (Addison et a1 1972, Lipschitz et a1 1974, Siimes et a1 1974). We report here the serum ferritin concentration of normal subjects, patients with iron deficiency anaemia, iron overload states, the correlation between serum ferritin concentration and units of blood transfused in patients with transfusional siderosis and the circadian rhythm of serum concentration in heaIthy subjects.
M. J . LEYLAND, P. C. GANGULI, D. BLOWER & I. W. DELAMORE
386
Number
Group
Haemoglobin (g/100 ml) Mean 2 SE
Controls
Men Women Anaemia (sideropenic) * Polycythaemia Vera** Haemolytic anaemia Haemachromotosis Sideroblastic anaemia Transfusional siderosis Aplastic anaemia
*
Untreated.
**
59 101 28 14 8 4 6 7 12
14.920.2 13.420.1 8.6T0.3 15.4k0.4 9.7k 1.2 13.821.0 9.6 k 1.3 8.3 k 0.7 9.1 2 0.7
I
Serum iron (ug/100 ml)
Range
Mean ‘I. SE
12.9-16.9 11.8-17.6 4.6-11.5 13.6-19.1 4.9-14.4 12.6-14.6 4.7-14.0 5.7-11.2 4.7-12.5
99k 6 90+ 5 185 2 36+ 3 160 k 41 2382 8 2302 52 289k 24 192 k 41
I
Transferrin saturation (‘%)
Range
Mean 2 SE
56-142 45-135 O L 35 20- 55 80-315 223-260 85-460 150-345 290-680
31.62 2 30.5k 1.4 4.62 0.5 8.0+ 0.5 47.8 2 10.5 86.22 3.6 68 k 9.0 83.0L 7.0 47.0+ 7.0
1
Range 16-53 12-65 0-10 6-11 22-85 78-94 35-92 44-96 24-90
Treated by repeated phlebotomy.
PATIENTS AND METHODS Patient7 The clinical, haematological and biochemical par ticulars of the patients in this report are shown in Table I. Male and female hospital controls with no evidence of anaemia or iron overload states were studied. Serum ferritin was measured in patients with untreated iron deficiency anaemia, with poblycythaemia Vera treated by repeated phlebotomy, with haemolytic anaemia and with a variety of iron overload states (primary haemachromatosis, sideroblastic anaemia, transfusional siderosis and aplastic anaemia). The correlation between units of blood transfused and serum ferritin concentration was examined in 14 patients with refractory anaemia and secondary transfusional siderosis. The circadian rhythm in serum ferritin concentration was studied in 7 healthy volunteers; for this investigation blood samples were orbtained at 0800, 1200, 1600, 2000, 2400 and 0400 h, and food was taken at 0815, 1215 and 1815 h. Methods The haemoglobin and blood counts were estimated by a Coulter S analyser. The serum iron, total iron binding capacity and liver function tests were measured by an autoanalyser technique. In patients with iron deficiency anaemia or iron overload states, the bone marrow aspirates were stained for iron with Perl’s reagent.
Human ferritin was prepared, its concentration determined and purity confirmed by established methods (Beamish et a1 1971). Six New Zealand White rabbits were immunised three times at monthly intervals and bled after another four weeks; all the animals yielded ferritin antibodies of high avidity and titre. A horse ferritin immunoadsorbent was prepared and used for the separation of bound from free antibody (Miles & Hales 1968). Ferritin antibodies were iodinated with Iodine-125 (Amersham, England) at pH 7.5 by the chloramine-T method (Hunter & Greenwood 1962) and the labelled antibodies were stored bound to immunoadsorbent at -20° C. The assay was performed according to the method of Addison and his colleagues (1972). Serum obtained from patients was stored at -20O C for up to six months or until assayed. The effects of dilution of serum, storage, anticoagulants and bacteriostatic agents on the immunoradiometric assay were tested. The specificity of the ferritin measurement was investigated by examining the immunological cross-reaction between ferritin, serum albumin, immunoglobulins, transferrin and nl-foeto-protein. The precision of the assay was estimated by measuring the ‘within’ and ‘between’ assay variations. Statistical analysis was performed using Student’s ‘t’ test for unpaired data and calculating correlation coefficients (Snedecor & Cochran 1967). The data for circadian rhythm of serum ferritin was expressed as fluctuation around each subject’s mean value for the 24 h period (Acland & Gould 1956).
HUMAN FERRITIN IMMUNORADIOMETRIC ASSAY RESULTS
Immunoradiometric assay
In the absence of ferritin, 70 % of the labelled antibody was bound by the immunoadsorbent and the linear portion of the curve covered the range 0.8 to 14 pgll ferritin. As the maximum concentration of serum incorporated in the assay was 42 %, the sensitivity of this method is 2 p g ferritin per 1 of serum. The precision of the assay at this level was F 4 % (coefficient of variation). The correlation coefficients for the ‘within’ assay and ‘between’ assay variations were r = 0.979 (n = 12, P < 0.001) and r = 0.927 (n = 19, P < 0.001). Dilutions of serum between 1 in 10 and 1 in 200, storage at -20’ C for six months, thawing and freezing six times, haemolysis with up to 3.5 g free haemoglobin per 100 ml, heparin, calcium oxalate or citrate in the usual amounts as anticoagulants, rner-
thiolate (1 in 10,000) and sodium azide (2 5%) did not interfere with the assay, but ethylene diamine tetra-acetic acid (EDTA) did. Ferritin concentrations in the serum and plasma prepared from the same sample of blood were not significantly diflerent. Serum albumin, immunoglobulins A, G and M, transferrin and al-foetoprotein did not cross-react with ferritin. Controls The observed and calculated frequency distribution of the serum ferritin concentration (log scale) of the control subjects is shown in Figure 1. The mean ferritin concentration in males, 189 k 16 pgll (mean k SE) and in females, 118 rt 9 pgll differed significantly (P < 0.005). In this study the range of serum ferritin concentration was 16 to 408 pgll for females and 15 to 462 pgll for males. In these subjects there was
L
ar n E
Figure 1. Frequency distribution of serum ferritin concentration (log. scale) in 160 controls. Histogram: Observed frequency distribution. Curve: Calculated freauencv .~ distribution.
I
387
log s e r u m f e r r i t i n c o n c e n t r a t i o n
388
M. J. LEYLAND, P. C. GANGULI, D. BLOWER & I. W. DELAMORE
n o correlation between ferritin level and the age, haemoglobin, MCV, MCH, serum iron or transferrin concentration.
MCV (r = 0.642, P < 0.01) and MCH (r = 0.560, P < 0.05). Primary haemachromotosis
Iron deficiency anaemia Serum ferritin concentration in these untreated patients was 6.1 31 0.7 pg/1 (Figure 2), this was significantly different from the controls (P 0.001). In this group serum ferritin level was not significantly different between males and females.
