Normal Serum Ferritin Levels in a Patient with HEMPAS Syndrome and Iron Overload SHAISTA FARUQUI, M.D., ANDREAS ABRAHAM, M.D., MARION R. BERENFELD, B.A., AND THOMAS G. GABUZDA, M.D.

Serum ferritin levels in a patient with HEMPAS syndrome (hereditary erythroblastic multinuclearity associated with positive acidified serum test) were correlated with body iron stores directly measured on spleen and liver biopsy specimens as well as by quantitative serial phlebotomy. Normal serum ferritin concentrations were found in the presence of a moderate excess in iron stores (approximately 6-12 times normal). They temporarily increased after transfusion and splenectomy with a prompt return to the normal range. As repeated phlebotomies over a period of nine months depleted the excess iron stores, the serum ferritin ultimately decreased to a subnormal concentration. The serum ferritin concentration was not a reliable index of increased body iron stores in this iron overloaded patient, but did reflect their depletion by serial phlebotomy. (Key words: HEMPAS; Iron overload; Hypersplenism; Serum ferritin) Am J Clin Pathol 1982; 78: 97-101

Received August 12, 1981; received revised manuscript and accepted for publication February 3, 1982. Partial support provided by US Public Health Service Grant No. HL 16654. Address reprint requests to Dr. Faruqui: Department of Medicine/ Hematology/Oncology, Earl K. Long Memorial Hospital, 582S Airline Highway, Baton Rouge, Louisiana 70805. 0002-9173/82/0700/0097 $00.75 ©

The decision to perform splenectomy in a patient with HEMPAS syndrome (congenital dyserythropoietic anemia type II) recently provided us with an opportunity to correlate changes in serum ferritin concentration with body iron stores, directly measured on the spleen and on a liver biopsy specimen. The patient also received transfusions preoperatively and then was phlebotomized over the months following splenectomy. These changes in body iron status were related to serum ferritin levels. Report of a Case The patient was ten years old in 1971 when she was first seen by Dr. Frank Oski at Children's Hospital of Philadelphia for unexplained splenomegaly. The past medical history was completely unrevealing. The child was the product of a normal pregnancy and weighed 6 lb 9 oz at birth. Blood studies done during the neonatal period were normal and there was no history of jaundice at that period of time. The family history was equally unrevealing, and no members of the family were known to have had anemia, jaundice, splenomegaly, or gall stones. Physical examination in 1971 revealed a bright, healthy young girl with weight of 28.1 kg. The spleen was palpable 5-6 cm below the left costal margin and was quite firm. The liver was not enlarged. There was no associated lymphadenopathy. The following laboratory studies were done: hemoglobin 12.1 g/dl, hematocrit 33.9%, reticulocyte count 7.4%. Examination of the peripheral blood film revealed many distorted red blood cells, with some that were irregularly contracted; an occasional cell was hypochromic. The MCV was 97 /i3 MCH 32.7 pg, MCHC 34 g/dl. The WBC was 7,600 with normal differential. The total bilirubin was 1.8 mg/dl with a direct fraction of 0.3 mg/dl. Special hematology studies included the following: the measurement of unstable, heat precipitable hemoglobin and other tests for abnormal hemoglobin were entirely normal. The unincubated and incubated osmotic fragility tests were normal. Red blood cell glutathione levels and glutathione stability in the presence of acetylphenylhydrazine were normal. The levels of red blood cell glucose-6-phosphate dehydrogenase, 6-phosphogluconate dehydrogenase, hexokinase, glucose phosphate isomerase, triose phosphate isomerase, phosphoglycerate kinase, and pyruvate kinase were assayed and all found to be elevated, consistent with the reticulocytosis present in the patient. The diagnosis was revealed after bone marrow examination, which showed marked erythroid hyperplasia with some megaloblastoid changes along with a fairly striking tendency for multinuclearity among maAmerican Society of Clinical Pathologists

97

Downloaded from http://ajcp.oxfordjournals.org/ by guest on January 15, 2017

FERRITIN is predominantly found in the cytoplasm of reticuloendothelial cells and hepatic parenchymal cells. While most of the body's ferritin is intracellular, small but clinically significant amounts are found in the serum. The detection and quantitation of this serum ferritin has been made practical through immunoradiometric technology.1,2 The serum ferritin concentration has been reported to correlate closely with body iron stores in normal subjects and in patients with iron deficiency and uncomplicated iron storage disease.3,4 The mean concentration is higher in men than in women, with a range between 12-250 ng/ml. The iron deficient person may have a serum ferritin level as low as 1.0 ng/ml while at the opposite extreme the serum concentration in patients with iron overload may be as high as 10,000 ng/ml. After Wands and associates5 reported that some people with precirrhotic hemochromatosis have normal levels of serum ferritin, other studies observed the same phenomenon.6,7 On the other hand, Halliday and colleagues8 found that serum ferritin was raised in 98% of cases with prehemochromatosis and suggested that it was a useful non-invasive screening test for this condition.

Department of Research, Lankenau Hospital and the Cardeza Foundation, Jefferson Medical College, Philadelphia, Pennsylvania

98

FARUQUI ET AL.

A.J.C.P. • July 1982

Downloaded from http://ajcp.oxfordjournals.org/ by guest on January 15, 2017

FIG. 1 Electron micrograph of a HEMPAS binucleated erythroblast; N = nucleus; M = mitochondria; Mem = membrane. ture red blood cell precursors. Under the electron microscope, double cellular membranes were seen in the majority (90%) of the normoblasts of all stages of maturation. These double membranes were present in uni-, bi-, and multinucleated red blood cell precursors as well as in the mature red blood cells of the peripheral blood (Fig. 1). These ultrastructural abnormalities have been described in HEMPAS erythroblasts by Heimpel, Forteza-Vila, and Queisser.9 The Ham test was weakly positive, and anti-i-agglutination studies were also positive. Based on these findings the diagnosis of HEMPAS syndrome was made. During the following years her hemoglobin concentration remained in the range of 10-12 g/dl until 1974 when she had progressive splenomegaly, along with an increase in the severity of the anemia (Fig. 2). During 1976 the hemoglobin concentration fell to the range of 5.87.2 g/dl while the spleen became yet larger. The 51Cr red blood cell

half life was 12.6 days and the plasma 59Fe half life was 34 minutes with 55% utilization. In February 1977 she developed infectious mononucleosis following which her hemoglobin concentration dropped to 3 g/dl. She was given four units of packed red blood cells and in March 1977 splenectomy and cholecystectomy (because of gall stones) were performed without complications. Postoperatively the hemoglobin level rose to 11-12 g/ dl and has remained in that range since. The platelet count also increased postsplenectomy to 2.5 million/mm3, subsequently stabilizing at about 1.25 million/mm3. Aspirin and dipyrimidole were given because of the marked thrombocytosis. Weekly phlebotomies of 300 ml whole blood were begun in September 1977 (body weight 45 kg). By May 1978 a total of about 9600 ml of whole blood was removed, equivalent to about 3.2 g of iron. At this point her hemoglobin level fell to 10 g/dl.

CASE REPORTS

Vol. 78 • No. 1

WEEKLY PHLEBOTOMY

SPLENECTOMY

HEPATITIS

3000K—

99

Tx(x4)m

15-,

imiiiiiiiiiimiiiiiii

2500K

10-

2000K

1500K

1000K

12/2

1/26

I

11/2

en

I

10/2

-* ~

I

9/14

-J CD

3/24

-J CD

NJ CO CO CO CO ^ \ \ w WOO CD - * - • O-^ v^

3/16

CD

-• N)

5/12

I

I -• O

01/9

0 J

CD

to

CO

^

FIG. 2 Clinical and laboratory data in patients with HEMPAS syndrome.

Materials and Methods The serum ferritin concentration was measured by a solid-phase radioimmunoassay (Fer-Iron, Ramco Laboratories, Inc., Houston, TX). The concentration of serum ferritin was calculated by comparing the unknown with a standard curve. All determinations were done in duplicate, some at two or more dilutions. All sera were diluted appropriately so that the actual concentrations measured were below 10 ng/ml. The liver and spleen iron concentrations were measured in duplicate by atomic absorption spectrophotometry on tissues dried to constant weight, ashed, and dis-

solved in 1 N HC1, with appropriate sample dilution to final iron concentrations in the range of 50-500 jtg/dl. 10 Serum iron and total iron binding capacity (TIBC) were measured by standard technics. Results Table 1 shows serum iron, TIBC and total bilirubin concentration during the course of the patient's disease. In 1976, the serum iron was as high as 146 Mg/dl with 70% saturation, but following the episode of infectious mononucleosis it dropped to 111 /*g/dl with saturation of 38%. The most profound effect was seen after the

Table I. Effect of Splenectomy on Serum Iron Levels in Patient with Hempas Syndrome (Weekly Phlebotomy Was Started in Sept. 1977)

Sept. 1976 Jan. 31, 1977 March 4, 1977 March 7, 1977 March 8, 1977 April 15, 1977 June 10, 1977 Jan. 26, 1978

%

S. ferritin (mg/dl)

Hb level %

Total Bilirubin (mg/dl)

210 226 293

70 81 38

50 55 100

5.8 6.2 5.2

4.2 6.1 2.7

14 62 91 49

700 190 60 40

Serum Iron (mg/dl)

TIBC (mg/dl)

146 186 111 275 236 209 366

Saturation

Splenectomy 12.0 11.0 14.0 12.0

1.4

Downloaded from http://ajcp.oxfordjournals.org/ by guest on January 15, 2017

500K

100

FARUQUI ET AL.

A.J.C.P. • July 1982

Table 2. Tissue Iron Stores in Patient with Hempas Syndrome at Splenectomy (Duplicate Values) Iron Content

Tissue Liver Spleen

Wet Weight (mg)

Dry Weight (mg)

Dry:wet

19.2 13.1 52.4 52.9

4.67 3.19 22.9 25.7

.243 .244 .437 .486

Total Organ Weight (g)

mg Fe/g dry weight

mg Fe/g wet weight

Total Fe (g)

1,200 (est.)

12.2 13.2 3.80 3.47

2.97 3.21 1.66 1.70

3.6 3.9 2.4 2.4

1,440

Normal ranges: Liver-0.195 ± 0.113 mg/g wet weight. Spleen-0.336 ± 0.210 mg/g wet weight. Tipton and Cook: Health Phys 9:103, 1963.

6-12 times the normal value of 0.5-1.0 g, excluding iron deposited in other body sites. Discussion In normal adults the concentration of serum ferritin is directly related to the reticuloendothelial and parenchymal iron stores.4 The serum ferritin level is three times higher in males than in females, a finding consistent with the known sex difference in storage iron. In normal adults, a significant correlation has been demonstrated by Cook and co-workers12 between serum ferritin level and radioiron absorption, a sensitive indirect measure of body stores. Most of the pathologic disturbances in iron balance have a marked effect on serum ferritin level. In iron deficiency, a mean of 5 ng/ml was reported in 21 subjects by Jacob and colleagues.3 Similarly, in patients with iron overload due to either hemochromatosis or excess transfusion, the serum ferritin was invariably elevated.4 There are certain diseases, however, in which the measured value of serum ferritin does not correlate with the reticuloendothelial iron stores. For example, elevated serum ferritin levels have been observed in patients with anemia of chronic disease, acute and chronic liver disease, and in patients with leukemia and Hodgkin's disease.4 Wands and associates5 have demonstrated normal serum ferritin levels in iron-loaded subjects with preclinical familial hemochromatosis, whereas recently Halliday and co-workers8 found that serum ferritin was raised in 98% of cases with prehemochromatosis. In our case, the patient had normal serum ferritin levels in spite of excess iron stores which were over 6-12 times normal as measured directly on liver and spleen tissues. We removed about 3.2 g of iron by repeated phlebotomy, without significant initial change, for three months, followed by a drop in the level of 4 ng/ml five months later, indicating exhaustion of the excess stores. After removing a pool of about 2.5 g of iron present in the spleen, the serum ferritin level did not reflect this abrupt decrease in body iron stores. To the contrary it

Downloaded from http://ajcp.oxfordjournals.org/ by guest on January 15, 2017

splenectomy, when the serum iron fell to 39 Mg/dl with only 14% saturation, following which it rose once again. The serum iron level and TIBC saturation gradually fell during the course of serial phlebotomies. The last serum iron done in May 1978 was 46 /ig/dl and TIBC 275 /xg/dl with 17% saturation, reflecting the depletion of iron stores by the phlebotomies. Figure 3 shows serial serum ferritin levels. In 1976 the serum ferritin levels were in the normal range at 30 ng/ml but soon after four blood transfusions, the level rose to 125 ng/ml and then further increased to 700 ng/ml within two to three days after splenectomy. It dropped back to normal levels within a week and then remained in the range of 90 ng/ml during the initial three months of serial phlebotomies. The serum ferritin levels then progressively decreased to 4 ng/ml by May 1978, suggesting depletion of iron stores at this time. Liver biopsy done at the time of splenectomy showed intact liver architecture with marked hemosiderin deposition in both hepatocytes and Kupfer lining cells. The deposition of iron in hepatic parenchymal cells was classified as Grade III by the method of Scheuer and associates." The spleen showed some hemosiderin deposition but not to the degree seen in the liver. Bone marrow iron stain showed diffusely staining material without apparent excess. Table 2 shows iron stores in liver and spleen measured in duplicate. The mean value for iron content in the liver was 12.7 mg/g dry weight (3.09 mg/g wet weight). This is approximately 15 times the normal liver iron concentration. The total liver iron content based on an estimated total weight of 1,200 g was 3.75 g, which agrees well with the amount calculated by quantitative phlebotomy (3.2 g). The mean value for spleen iron was 3.63 mg/g dry weight (1.68 mg/g wet weight) which is about 5 times the normal spleen iron concentration. The total iron in the spleen was 2.4 g, based on the total spleen weight of 1,440 g. The sum of liver and spleen iron was 6.15 g, a minimum estimate of body iron stores at the time of splenectomy. This represents a moderate increase to

Vol. 78 • No. I

101

CASE REPORTS

in iron content in liver and spleen. Indeed, in our patient, the serum ferritin concentration grossly underestimated iron stores. Whether this observation is relevant to the pathogenesis of the abnormal iron accumulation in HEMPAS syndrome remains to be determined. References 1. Addison GM, Beamish MR, Itales CN: An immunoradiometric assay for ferritin in the serum of normal subjects and patients with iron deficiency and iron overload. J Clin Pathol 1972; 25:326-329 2. Cook JD, Lipschitz DA, Miles LEM, Finch CA: Serum ferritin as a measure of iron stores in normal subjects. Am J Clin Nutr 1974; 27:681-687 3. Erslev AJ, McKenna PJ: Effect of splenectomy on red cell production. Ann Intern Med 1967:990-997 4. Feller ER, Pont A, Wands JR, Carter EA, Foster G, Kourides IA, Isselbacher KJ: Familial hemochromatosis. N Engl J Med 1977; 296:1422-1426 5. Green R, Watson LR, Saab GA, Crosby WH: Normal serum ferritin: a caution. Blood 1977; 50:545-547 6. Halliday JW, Russo AM, Cowllishaw JL, Powell LW: Serum ferritin in diagnosis of haemochromatosis. Lancet 1977; 2:621623 7. Heimpel H, Forteza-Vila J, Queisser W: Morphological aberrations of the erythroblasts in congenital dyserythropoietic anemia type I and II. Presented at the XII International Congress of Hematology, Munich, Germany, August 1970 8. Jacobs A, Worwood M: Ferritin in serum: clinical and biochemical implications. N Engl J Med 1975; 292:951-956 9. Jacobs A, Miller F, Worwood M, Beamish MR, Wardrop CA: Ferritin in the serum of normal subjects and patients with iron deficiency and iron overload. Br Med J 1972; 4:206-208 10. Jacobs A, Worwood M: Serum ferritin. N Engl J Med 1976; 294:900-901 11. Miles LEM, Lipschitz DA, Bieleer CP: Measurement of serum ferritin by a 2-site immunoradiometric assay. Anal Biochem 1974;61:209-224 12. Powell LW, Halliday JW: Serum ferritin in hemochromatosis. N Engl J Med 1976; 294:1185 13. Scheuer PJ, Williams R, Muir AR: Hapatic pathology in relatives of patients with hemochromatosis. J Pathol Bacterid 1962; 84:53-64 14. Wands JR, Rowe JA, Mezey SE, Waterbury LA, Wright JR, Halliday JW, et al: Normal Serum ferritin concentrations in precirrhotic hemochromatosis. N Engl J Med 1976; 294:302205

Downloaded from http://ajcp.oxfordjournals.org/ by guest on January 15, 2017

showed a sharp temporary rise demonstrating the capacity of the serum ferritin level to respond by increasing to above normal concentrations. We have observed the same transient rise in patients with sickle cell anemia and iron overload who underwent transfusion and surgery. The serum iron level behaved in a reciprocal manner, falling to subnormal levels as the serum ferritin level rose postsplenectomy. Profound postsplenectomy hypoferremia has been previously documented by Erslev and associates.13 The failure of the serum ferritin level to abruptly decrease following the removal of a significant proportion of the body iron pool in the spleen suggests that input from the spleen into the serum was not a significant determinant for maintaining the serum ferritin level. The explanation given by Wands and associates15 for normal serum ferritin levels in precirrhotic hemochromatosis is that the reticuloendothelial system is not involved in storing excess iron but only the parenchymal cells. Only when reticuloendothelial overload or parenchymal cell damage develops with cirrhosis, for example, does the serum ferritin become abnormally high in these patients. We might postulate in our case that the state of extreme erythroid hyperplasia and increased iron turnover that characterizes HEMPAS syndrome may prevent a rise in serum ferritin level by keeping reticuloendothelial storage iron within the normal range. Alternately, the abnormal double membrane seen in the erythroblasts of patients with HEMPAS syndrome might block entry in some manner from tissue sites into the serum. It is unlikely that the method underestimated the levels of serum ferritin ("hook effect") as discussed by Crosby and colleagues14 since values were measured at two or more dilutions. Thus, we conclude that serum ferritin concentration may be normal despite an abnormally high serum iron, transferrin saturation and moderate to marked increase