* Objective.-Bone marrow examination is rarely required for the diagnosis of childhood anemia, and its diagnostic utility in this setting is unknown.
Design.-Marrow specimens from 25 children aged 11 days to 12 years were reviewed to determine the cause of unexplained anemia.
Results-These samples comprised only 2% of pediatric marrow examinations. Hematocrits ranged from 0.12 to 0.31 (mean 0.23). Marrow findings included erythroid hypoplasia (12 of 25, 48%) and hyperplasia (11 of 25, 44%), dyserythropoiesis (2 cases), ringed sideroblasts (2 cases),
lymphocytosis (3 cases), and megaloblastic change (1 case). Final diagnoses were transient erythroblastopenia of childhood (15 cases, 60%); iron deficiency and sideroblastic anemia (2 cases each); and congenital dyserythropoietic anemia, anemia of chronic disease, hereditary spherocytosis, and intra-abdominal hemorrhage (1 case each). In two patients, a definitive diagnosis was never made.
Conclusions.-Marrow examination contributed to a specific diagnosis in childhood anemia in 92% of cases; the most common diagnosis in this population was transient erythroblastopenia of childhood. (Arch Pathol Lab Med. 1998;122:638-641 )
Anemia raises a different set of diagnostic possibilities when found in children than when found in adults.' Childhood anemia is often caused by iron deficiency, a hemoglobin abnormality, a red cell membrane disorder, or is self-limited and of unknown etiology (transient erythroblastopenia of childhood, or TEC).'-3 The diagnosis can usually be made from the clinical presentation, complete blood count data, reticulocyte count, examination of the blood smear, and other laboratory studies of the peripheral blood. In some instances, however, the etiology cannot be readily identified, and further studies, including bone marrow examination, become necessary. The utility of bone marrow sampling in this setting is unknown. We studied cases of childhood anemia requiring bone marrow aspiration and biopsy to determine the etiology of anemia and the diagnostic efficacy of bone marrow evaluation. Results were correlated with clinical outcomes in these patients. METHODS
We reviewed pathologic material from bone marrow aspirations and biopsies performed on children younger than 18 years with a clinical diagnosis of anemia between 1984 and 1990 at Rainbow Babies and Children's Hospital, Cleveland, Ohio. Children with a previously diagnosed hematologic disorder were excluded from this study. In all cases complete blood count data with red blood cell indices, Wright's-stained peripheral blood smears, and Wright-Giemsa-stained aspirate smears were available for review. Hematoxylin-eosin-stained clot sections were examined in all cases, and core biopsies were examined in 15. Prussian blue stains for evaluation of iron stores were done in all cases on both the core biopsies and clot sections, as well as on aspirate smears in selected cases to evaluate the possibility of sideroblastic anemia. Flow cytometry was performed on mononuclear bone marrow cells in one case on a Coulter Epics 5 (Coulter Corporation, Hialeah, Hla).4 Antibodies to CD1, CD2, CD3, CD7, surface immunoglobulin, Ia, CD21, and CD10 (Coulter) were used. Terminal deoxynucleotidyl transferase (TdT) staining was done by immunofluorescence (Molecular Genetics Resources, Tampa, Fla) s Medical charts were reviewed to determine clinical history pertinent laboratory data, impact of marrow findings on treatment, and clinical outcome. RESULTS
Twenty-five bone marrow specimens from children with anemia were studied, of a total of 1269 pediatric specimens. Thus, only 2% of pediatric marrow examinations were performed for anemia of unknown etiology. The patients ages ranged from 11 days to 12 years, with a mean of 2 years. Eleven patients were female and 14 were male. The presenting symptoms were pallor in 13 patients (52%) and apparent viral illness in 12 (48%).
Complete blood count findings included hemoglobin levels ranging from 38 to 105 g/L, hematocrits of 0.12 to 0.31, and red cell counts of 1.66 to 5.36 x 1012/L. The corrected reticulocyte count, available in 23 of the 25 patients, ranged from 0 to 0.072 (conventional units 0-7.2%) and was greater than 2% in only 25% of patients. Mean corpuscular volume (MCV) of red blood cells ranged from 46 to 109 fL. The MCV was less than 80 fL in 3 patients, between 80 and 100 fL in 21 patients, and greater than 100 fL in a newborn, which was normal for the patient's age. White blood cell counts ranged from 3.1 to 15.1 x 109/L, and platelet counts ranged from 35.0 to 596.0 x 109/L. Four patients with TEC were neutropenic and had neutrophil counts ranging from 0.37 to 0.76 x 109/L. Four of the patients with TEC had an absolute neutropenia, with neutrophil counts ranging from 0.37 to 0.76 x 109/L. Bone marrow examination and clinicopathologic correlation led to the following final diagnoses: TEC, 15 cases; iron deficiency anemia, 2 cases; sideroblastic anemia, 2 cases; congenital dyserythropoietic anemia, 1 case; hereditary spherocytosis, 1 case; anemia of chronic disease, 1 case; and hemorrhage into a mesenteric cyst, 1 case. In two cases, a definite diagnosis for the cause of the anemia was not made (Table 1).
The most common final clinical diagnosis was TEC, representing 60% of all cases. The age range of these patients was 3 months to 12 years (average 32 months). There were eight males and six females in this group. Bone marrow findings included erythroid hypoplasia with maturation arrest at the pronormoblast stage in 11 specimens and erythroid hyperplasia, consistent with recovery phase, in 4 (Table 2). None of the four patients with erythroid hyperplasia had elevated reticulocyte counts at the time of marrow biopsy. Three cases of TEC displayed marked lymphocytosis (>40% lymphocytes on the aspirate smear), which sometimes obscured the erythroid population, particularly in the aspirate clot and biopsy sections. The erythroid precursors were morphologically normal in all cases of TEC except one, in which mild megaloblastic change was seen. Vitamin B,2 and folate levels were not obtained in this patient. The granulocytic and megakaryocytic cells lines were either adequate or slightly increased. Three cases had megakaryocytic hyperplasia and four had granulocytic hyperplasia. In all cases, the white blood cells and megakaryocytes were morphologically normal.
Iron stores were either present or increased in 11 of 12 evaluable cases of TEC. In one case, iron stores were absent and serum studies showed an iron level of 2 [mol/ L (normal value, 4.5-26.8 *,mol/L), total iron-binding capacity of 78 wmol/L (normal value, 13.4-76.1 l,mol/L), and percent saturation of 3% (normal value, 25% to 45%). This patient had had a recent viral illness and did not respond to oral iron therapy. The anemia may have been multifactorial in this child.
Flow cytometry was performed in one case of TEC with 45% lymphocytes on the aspirate smear. The results showed CD2, 61%; CD3, 53%; CD1, 18%; CD7, 58%; CD21, 46%; and CD10, 54%. Terminal deoxynucleotidyl transferase positivity was noted in 22% of marrow cells. Leukemia was never diagnosed in this patient because the lymphoid cells were not thought to represent blasts, and 2 months after admission the patient's hematocrit was 0.38.
All patients diagnosed with TEC fully recovered from the anemia within 2 months of the initial diagnosis. Five patients were tested for viral titers for either cytomegalovirus or Epstein-Barr virus, all of which were negative. Forty-three percent of the patients received one or two blood transfusions. None received additional therapy.
Marrow findings from a 7-year-old girl and a 1-yearold boy with iron deficiency anemia included erythroid and megakaryocytic hyperplasia; one case also showed mild dyserythropoiesis. Serum iron studies revealed moderate iron deficiency in one patient, who had a serum iron level of 2.5 l,mol/L (conventional units 12 jig/dL), total iron-binding capacity level of 16.7 p,mol/L, saturation of 0.08, and serum ferritin level of 35 jig/L (n =15-300 ,ag/ L). Neither patient had iron stores in the marrow. Although the iron studies in the other patient only showed a low-normal ferritin level, the absent iron stores and marked clinical response to oral iron therapy confirmed the diagnosis.
The diagnosis of congenital sideroblastic anemia was made in a 1-year-old boy and a 2-month-old male infant, both of whom showed erythroid hypoplasia and granulocytic hyperplasia in marrow specimens. No dysplastic or megaloblastic changes were present. Ringed sideroblasts were seen in 40% and 25% of the erythroid precursors on the aspirate smears.
One marrow sample from a 3-month-old male infant who had been anemic since birth showed dysplasia of erythroid precursors without increased numbers of blasts. The anemia was normocytic with an MCV of 91 fL. The aspirate smear showed megaloblastic changes, dyserythropoiesis, and intracellular bridges between erythroid precursors. The findings were most consistent with type I congenital dyserythropoietic anemia.
Erythroid hyperplasia was seen in two females, aged 2 months and 2 years, who were subsequently found to have peripheral destruction or sequestration of red blood cells. The hematocrits of these two patients were 0.19 and 0.16, respectively. Both had an MCV of 81 fL and elevated reticulocyte counts of 7.0 and 4.0 x 109/L. The 2-month-old was diagnosed with hereditary spherocytosis. This patient was massively transfused at birth owing to apparent hemolysis, and spherocytes were not appreciated on the peripheral smear at the time of the marrow aspiration. The second child, the 2-year-old, presented with anemia, and the marrow aspiration was performed to rule out TEC. She subsequently developed an abdominal mass and was found to have bled into a mesenteric cyst.
One patient with acquired immunodeficiency syndrome (AIDS) presented with multiple infections and chronic anemia. A bone marrow specimen was obtained to rule out sideroblastic anemia, as the serum iron level was 0.47 l,mol/L, total iron-binding capacity was 9.13 lmol/L, and iron saturation was 51%. Ferritin levels were markedly increased at 2590 jIg/L. The marrow showed erythroid hypoplasia; no ringed sideroblasts were identified, and dysplasia was not seen. The anemia was felt to be due to chronic disease related to human immunodeficiency virus infection.
In the final two cases, occurring in neonates aged 11 days and 3 weeks, no definite diagnosis could be established. Both marrow specimens showed mild erythroid hyperplasia, and one had megaloblastic change. Anemia had been present since birth in both children. Each child recovered over the course of the first year of life without any specific treatment. COMMENT
Anemia appears to be an uncommon cause for bone marrow examination in children, accounting for only 2% (25/1269) of pediatric marrow specimens at one institution from 1984 through 1990. Examination of the bone marrow specimen aided in establishing the final diagnosis in 92% of cases. In cases with definite findings, such as marked dyserythropoiesis or ringed sideroblasts, study of the bone marrow specimen was sufficient for diagnosis; in other cases the absence of ringed sideroblasts or malignancy was the only useful information derived from bone marrow aspirate and biopsy examination. In all but two cases, the findings in the bone marrow aspirate and biopsy were used in combination with the clinical findings and pertinent laboratory data to arrive at a diagnosis. In two cases, the marrow and clinical findings were nonspecific and no diagnosis could be established. The most frequent diagnosis made in this study was red cell aplasia consistent with TEC.'36' The diagnosis was confirmed by spontaneous recovery of the bone marrow with resolution of the anemia. Transient erythroblastopenia of childhood usually requires no specific treatment; in exceptional cases, transfusion for shock and metabolic acidosis may be required.s Although parvovirus B19 is a well-established cause of red cell aplasia in children with hematologic disorders, its association with TEC has not been supported.9 We found no cases of parvovirus in our series, according to morphologic features. This is not surprising, as we had excluded from the study patients with previously known hematologic disorders, such as hemolytic anemia or acute lymphocytic leukemia, in which anemia due to parvovirus has been described. No evidence of other viral infection was apparent in cases where viral titers were performed.
Clinically, TEC must be distinguished from congenital erythroid hypoplasia (Blackfan-Diamond syndrome), which is usually seen in children less than 1 year of age and is not associated with spontaneous recovery.lo Increased red cell fetal hemoglobin and persistence of i antigen on red blood cells may be helpful in distinguishing congenital hypoplastic anemia from TEC; however, the red blood cells of infants and those in the recovery phase of TEC also contain significant amounts of fetal hemoglobin and i antigen." The age range of patients with TEC was 5 months to 12 years, although almost half of the children diagnosed in our study with TEC were less than 1 year old. This may reflect that clinicians are more likely to biopsy a young child with findings suggestive of red cell aplasia, as TEC is less common in this age group.
Bone marrow findings in TEC are variable. Erythroid hypoplasia is usually found, but erythroid hyperplasia may be seen if a marrow sample is obtained during the recovery phase of the disease. In our study, erythroid hyperplasia was seen in 25% of the TEC specimens. Other marrow elements may also be hyperplastic, as demonstrated in a small number of our cases. The most problematic finding is a marked increase in the number of lymphocytes, including many immature cells. Lymphocyte surface marker studies may show an increase in early B cells with significant CD10 and terminal deoxynucleotidyl transferase positivity, as exemplified in one of our cases.2-'4 These findings may be mistaken for acute lymphocytic leukemia. Clinical recovery within 1 to 2 months confirms the diagnosis of TEC. The distinction of acute leukemia from TEC is usually not difficult, owing to the difference in the clinical presentation and course. Cases not diagnosed as TEC fall into two categories, namely findings consistent with a common disorder for which marrow specimens are rarely obtained, such as iron deficiency, or rare disorders, such as congenital sideroblastic or dyserythropoietic anemias.
Iron deficiency anemia is common in all age groups. The diagnosis is usually established by serum iron studies without bone marrow examination. Erythroid hyperplasia and absent iron stores were noted in two cases of iron deficiency in this series. The presence of visually detectable marrow iron is not expected in young children because iron stores as measured by serum ferritin levels are generally low. Bone marrow studies in our patients with iron deficiency were performed because of marked anemia and a clinical suspicion of malignancy. Both children had abnormal iron studies and responded to oral iron therapy.
Congenital sideroblastic and dyserythropoietic anemias are rare disorders, often requiring repeated blood transfusions.ls l6 Complications arise from iron overload rather than transformation to acute leukemia, also a potential outcome in adults with acquired sideroblastic or dyserythropoietic anemia. Bone marrow samples are essential in making these diagnoses in children. Sideroblastic anemia was part of the differential diagnosis in a patient with AIDS, as the serum ferritin level was markedly increased. Ferritin levels have been shown to parallel disease severity in AIDS patients, and the mechanism is thought to be similar to that of the anemia of chronic disease. Anemia is common in AIDS patients and is considered to be multifactorial in origin.l7
Finally, in two cases unusual clinical presentations led to the performance of a bone marrow aspiration, which in retrospect may have been unnecessary. One patient with hereditary spherocytosis could not be diagnosed because transfusion had obscured the peripheral blood findings. The other had an intra-abdominal hemorrhage contained within a mesenteric cyst, and an abdominal mass was not discovered until after the marrow aspiration had been performed.
Hemorrhage into a mesenteric cyst is an extremely rare complication of a rare lesion. Mesenteric cysts are diagnosed in approximately 1 in 140000 hospital admissions.ls The patients generally present with abdominal pain and distension.l9 Only rare cases have been reported with hemorrhage leading to anemia.20 The diagnosis of a mesenteric cyst is made by ultrasound, and surgical resection is the treatment of choice.19
In this study, bone marrow examination contributed to a specific diagnosis in 92% of cases. The greatest utility of bone marrow examination in this setting is to confirm or exclude a serious underlying blood disorder as a cause of childhood anemia. The leading diagnosis was TEC. Although erythroid hypoplasia was the usual bone marrow abnormality in TEC, erythroid hyperplasia was seen in 25% of TEC cases; thus, the presence of erythroid precursors should not exclude this diagnosis. In addition, TEC may occur in children younger than 1 year of age (50% of patients in this study). Thus, this disorder, along with more grave diseases such as congenital hypoplastic anemia, is an important diagnostic possibility in infants with anemia.
1. Welch )C, Lilleyman JS. Anaemia in children. Br] Hosp Med.1995;53:387390.
2. Dallman PR, Yip R, Johnson C. Prevalence and causes of anemia in the United States 1976-1980. Am J Clin Nutr 1984;39:437-445. 3. Cherrick I, Karayalcin G, Lanzkowsky P. Transient erythroblastopenia of childhood: prospective study of fifty patients. Am J Pediatr Hematol Oncol. 1994; 16:320-324.
4. Coon JS, Landay AL, Weinstein R. Diagnostic flow cytometry. Lab Invest. 1987;57:453-479.
5. Bollum F. Terminal deoxynucleotidyl transferase as a hematopoietic cell marker. Blood. 1979;54:1203-1215.
6. Gerrits GPJM, van Oostrom CG, de Vann GAM, et al. Transient erythroblastopenia of childhood: a review of 22 cases. Eur/ Pediatr. 1984;142:266-270. 7. Bhambhani K, Inoue S, Sarnaik SA. Seasonal clustering of transient erythroblastopenia of childhood. Am I Dis Child. 1988;142:175-177. 8. Chabali R. Transient erythroblastopenia of childhood presenting with shock and metabolic acidosis. Pediatr Emerg Care. 1994;10:278-280. 9. Bhambhani K, Inoue S, Sarnaik SA, et al. Transient erythroblastopenia of childhood not associated with human parvovirus infection. Lancet. 1986;509. Letter.
10. Halperin DS, Freedman MH. Diamond-Blackfan anemia: etiology, pathophysiology, and treatment. Am J Pediatr Hematol Oncol. 1989;11:380-394. 11. Link MP Alter BP. Fetal-like erythropoiesis during recovery from transient erythroblastopenia of childhood (TEC). Pediatr Res. 1981;15:1036-1039. 12. Foot ABM, Potter MN, Ropner JE, et al. Transient erythroblastopenia of childhood with CD10, TdT, and cytoplasmic mu lymphocyte positivity in bone marrow. Am] Clin Pathol. 1990;43:857-859.
13. Davis RE, Longacre TA, Cornbleet PJ. Hematogones in the bone marrow of adults. Am/ Clin Pathol. 1994;102:207-211.
14. Sandhaus LM, Chen TL, Ettinger LJ, et al. Terminal significance of increased proportion of CDl 0-positive cells in nonmalignant bone marrow of children. Am J Pediatr Hematol Oncol. 1993;15:65-70.
15. Castaneda VL, Williams TE, Harper JL, et al. Severe refractory anemia with ringed sideroblasts and bone marrow aplasia in a child. Am J Pediatr Hematol Oncol. 1992;14:70-76.
16. Murakami R, Takumi T, Gouji I, et al. Sideroblastic anemia showing unique response to pyridoxine. Am J Pediatr Hematol Oncol. 1991;13:345-350. 17. Aboulafia DM, Mitsuyasu RT. Hematologic abnormalities in AIDS. Hemato/Oncol Clin North Am, 1991;5:195-214.
18. Chung MA, Brandt ML, Yazbeck S. Mesenteric cysts in children. J Pediatr Surg. 1991;26:1306-1308.
19. Egozi El, Ricketts RR. Mesenteric and omental cysts in children. Am Surg. 1997;63:287-290.
20. Solovei G, Alame A, Elchardus JF, et al. Mesenteric cystic lymphangioma in children: report of a case manifested by anemia. Ann Pediatr (Paris).1990;3: 405-08.
Accepted for publication February 16, 1998. From the Department of Pathology and Laboratory Medicine, Emory University School of Medicine, Atlanta, Ga (Dr Farhi), and the Institute of Pathology, Case Western Reserve University, Cleveland, Ohio (Drs Luebbers and Rosenthal).
Reprint requests to Department of Pathology, F-143, Emory University Hospital, 1364 Clifton Rd, NE, Atlanta, GA 30322 (Dr Farhi).
Copyright College of American Pathologists Jul 1998
Provided by ProQuest Information and Learning Company. All rights Reserved