Stanley L. Schrier M.D., F.A.C.P.1
1Professor of Medicine (Active Emeritus), Division of Hematology, Stanford University School of Medicine
The author has served as a consultant for Tularik, Inc., and Receptron, Inc.
Classification of Production Defects
Red blood cell production defects cause anemia that is marked by a low absolute reticulocyte count. Examination of the peripheral blood count and the bone marrow aids in classifying these disorders. The marrow characteristically shows one of the following:
Production Defects Associated with Apparently Normal Bone Marrow
ANEMIA OF CHRONIC DISEASE
The anemia of chronic disease occurs secondary to neoplastic, infectious, and inflammatory diseases and other chronic illnesses, including liver disorders, congestive heart failure, and diabetes mellitus.1,2 Hematocrit values usually range from 27% to 35%, although 20% of patients have hematocrit values below 25%.2
The anemia of chronic disease usually results from a combination of slightly shortened red blood cell survival, the sequestration of iron in the reticuloendothelial system, and erythropoietin levels that are less than expected for the degree of anemia.1,2 Red blood cells usually have a normal morphologic appearance, although they may occasionally be mildly hypochromic and microcytic. The serum iron and transferrin levels are low, and iron saturation is frequently as low as 15%.1,2 The serum ferritin level is usually normal or elevated.2,3 All these changes can be induced by the inflammatory cytokines (e.g., interleukin-1 [IL-1]; tumor necrosis factor-α; interferons alfa, beta, and gamma; and perhaps transforming growth factor-β).4 Under experimental conditions, these cytokines reduce erythropoietin production, cause hypoferremia, increase serum ferritin levels, impair erythropoiesis, and block release of iron from reticuloendothelial cells.5 Hepcidin, a newly described mediator of iron metabolism, may be the major mediator of the anemia of chronic disease6; hepcidin production is increased up to 100-fold with inflammation. Hepcidin seems to be the long-sought mediator that transmits iron stores to the gut. Hepcidin is secreted when iron stores, primarily in the liver, are increased, and it blocks iron absorption from the gut and causes iron to be trapped in macrophages.6
Mild anemia, with normal or elevated levels of leukocytes and platelets, in a patient with a chronic illness suggests the diagnosis of anemia of chronic disease. This normocytic or hypochromic and microcytic anemia is easily misdiagnosed as iron deficiency anemia, thalassemia trait, or a sideroblastic anemia. If the diagnosis is uncertain after careful examination of the blood smear, the most useful tests for making the diagnosis are measurement of the serum ferritin level and, in rare cases, bone marrow examination that includes an iron stain [seeTable 1] and [see Figure 1]. In some cases, there is more than one cause of the anemia, and thorough examination of the patient may be required to establish the primary cause. For example, a patient who has anemia of chronic disease resulting from carcinoma of the colon may also be iron deficient because of intestinal bleeding. HIV infection produces complex hematologic effects, including Coombs-positive autoimmune hemolytic anemia, but it also causes anemia of chronic disease in the majority of patients with AIDS.7
Table 1 Differential Diagnosis of Hypochromic Anemias
Figure 1. Diagnosis of Anemia
Flowchart shows steps in the diagnosis of anemia caused by production defects. This type of anemia is suggested by a low corrected reticulocyte count or the finding of associated leukocyte or platelet abnormalities on the peripheral blood smear.
Identifying and treating the primary disease is the most important part of managing the anemia of chronic disease. Oral or parenteral iron administration is usually not helpful. Erythropoietin is the standard treatment for patients with anemia of chronic disease. For many patients, administration of pharmacologic doses of erythropoietin corrects the anemia of chronic disease by overriding the defect in erythropoietin production. It is useful to obtain a baseline measurement of the plasma erythropoietin level, because a response to erythropoietin is unlikely in patients whose endogenous levels are above 500 mU/ml. Erythropoietin responses have been reported in patients with rheumatoid arthritis,8 AIDS,9 inflammatory bowel disease,10,11 and cancer.12 To respond optimally, the patient must have adequate available iron stores (i.e., normal or elevated ferritin level or marrow iron stain) [see 5:I Approach to Hematologic Disorders]. Previously, the recommendation was to start the patient on 100 to 150 U/kg subcutaneously three times weekly; however, most physicians give a single subcutaneous dose of 40,000 units of erythropoietin weekly.13
If the hemoglobin level does not rise after 12 weeks, erythropoietin should be discontinued. A longer-acting form of erythropoietin, darbepoietin alfa, can be given subcutaneously at doses of 100 µg weekly or 200 µg every other week.
ANEMIA IN SEVERE RENAL DISEASE
Pathophysiology and Etiology
The predominant cause of anemia in renal disease is a deficiency of erythropoietin production by the diseased kidneys. If underlying inflammatory renal disease is present, there may be a component of anemia of chronic disease.14 Anorexia and poor iron intake, frequent blood sampling, and loss of erythrocytes during hemodialysis may produce iron deficiency. Folic acid deficiency, hypersplenism, and secondary hyperparathyroidism with marrow fibrosis4 may also promote anemia.
Anemia in hemodialysis patients can be caused by aluminum toxicity, as well. This anemia was initially identified in patients who had so-called dialysis dementia. Very high plasma aluminum levels probably result from aluminum contamination of the dialysis fluid or gastrointestinal absorption of the aluminum gels taken to bind dietary phosphates. In vitro experiments have shown that aluminum inhibits the growth of the erythroid precursors colony-forming unit-erythroid (CFU-E) and burst-forming unit-erythroid (BFU-E).15
The blood smear should be examined for erythrocyte fragmentation or echinocytosis to exclude other causes of the anemia. The presence of Heinz bodies suggests that oxidative hemolysis has occurred, perhaps caused by oxidants in the hemodialysis fluid.
Erythropoietin is the standard treatment for anemic patients with renal disease. Erythropoietin therapy can eliminate the transfusion requirement for patients on hemodialysis and in patients with progressive renal disease who do not yet require hemodialysis. Such treatment significantly improves their quality of life.16 Side effects, such as hyperkalemia and hypertension, occur infrequently. It is customary to start therapy with 50 U/kg of erythropoietin three times weekly, either intravenously or subcutaneously, and to increase the dosage as necessary to bring the hemoglobin level to the desired value. Parenteral iron supplementation improves the response. ImFed (a form of iron dextran) can be given intramuscularly or intravenously at doses ranging from 100 to 500 mg, with an anticipated frequency of reaction of 4.7%. Ferrlecit (a form of sodium ferric gluconate) can be infused intravenously (125 mg over 1 hour), with the occasional occurrence of hypotension and rash.12 In a study of patients with anemia caused by aluminum toxicity, treatment with I.V. deferoxamine (30 mg/kg I.V. at the end of each dialysis session) produced substantial improvement.17
ANEMIA SECONDARY TO OTHER CONDITIONS
Excessive alcohol ingestion—either acute or chronic—has profound hematologic effects.18 Ingestion of about 80 g of alcohol (one bottle of wine, six pints of beer, or one-third bottle of whiskey) daily may produce macrocytosis,19 stomatocytosis,20 thrombocytopenia,21vacuolization of proerythroblasts, ringed sideroblasts,20 a sharp drop in serum folic acid levels, and a rise in serum iron levels; it may also impair the reticulocyte response to administered folic acid in a patient known to be folic acid deficient. Acute alcohol ingestion itself does not produce a megaloblastic anemia.18 It has been postulated that alcohol-induced hematologic toxicity is mediated through acetaldehyde, the major metabolite of ethanol, which is far more toxic and reactive than ethanol. The mechanism for these alcohol-induced abnormalities may be the formation of antibodies against acetaldehyde-hemoglobin adducts.20 Megaloblasts, macro-ovalocytes, and hypersegmented polymorphonuclear neutrophils (PMNs) usually appear when concomitant folic acid deficiency is present. Chronic alcohol abuse often results in concomitant folic acid or iron deficiency, severe liver disease, GI bleeding, hypersplenism, and the anemia of chronic disease.
Starvation resulting from anorexia nervosa or protein deficiency can cause anemia and even pancytopenia. Hemolysis may also be present [see Figures 2a, 2b, and 2c]. The bone marrow biopsy is hypocellular, with a characteristic gelatinous background material consisting of acid mucopolysaccharides. The anemia can occur despite normal folic acid and cobalamin (vitamin B12) levels and can be corrected with proper nutrition.
Figure 2a. Peripheral Smear Changes in Severe Liver Disease
The peripheral smear changes seen in severe liver disease or starvation include distinct variation in size and shape of red blood cells; both sharply spiculed cells (spur cells) and scalloped erythrocytes are prominent.
Figure 2b. Leukoerythroblastic Blood Smear
The leukoerythroblastic blood smear indicates marrow replacement with extramedullary hematopoiesis. It is characterized by variation in the size and shape of red blood cells, by the presence of nucleated red blood cells in the peripheral blood, by giant platelets, and by immaturity in the myeloid series.
Figure 2c. Blood Smear in Folic Acid or Cobalamin Deficiency
In folic acid or cobalamin deficiency, the smear is characterized by variation in erythrocyte size and by distinct macrocytosis. Occasionally, fish-tailed erythrocytes are present, along with hypersegmented neutrophils.
Hypothyroidism impairs erythrocyte production. The presence of macrocytosis in a hypothyroid patient suggests concomitant dietary folic acid deficiency or pernicious anemia.
The mild anemia that is associated with severe panhypopituitarism can be corrected by replacement of adrenal, thyroid, and gonadal hormones; the enhancing effect of androgens on the action of erythropoietin is well known.
The hemoglobin levels, red blood cell indices, and leukocyte and platelet counts of healthy older people are similar to those of younger adults; this finding was confirmed in a study of patients who were 84 years of age or older.22 Thus, a workup is required when anemia occurs in such older patients. The evaluation and treatment of anemia in the aged has become increasingly important because the presence of anemia (hemoglobin concentration [Hgb] < 12 g/dl in women and < 13 g/dl in men) is an independent risk factor for decline in quality of life.23
Production Defects Associated with Marrow Aplasia or Replacement
The combination of anemia and neutropenia or thrombocytopenia or the combination of all three of these abnormalities (i.e., pancytopenia) usually indicates that the hematopoietic marrow is damaged. If the marrow cavity is infiltrated but pluripotent stem cells are intact, extramedullary hematopoiesis will often develop in the organs of fetal hematopoiesis (i.e., spleen, liver, and distal bones).
Pancytopenia can be congenital or acquired. The finding of combined cytopenias or of immature cells in the blood (myelocytes, metamyelocytes, and erythroblasts)—that is, a leukoerythroblastic blood smear—suggests extramedullary hematopoiesis [see Figures 2a, 2b, and 2c]. These findings are an indication for bone marrow aspiration and biopsy.
Pancytopenia (i.e., anemia, neutropenia, and thrombocytopenia) and aplastic marrow on biopsy examination [see Figures 3a and 3b] establish a working diagnosis of aplastic anemia. The biopsy specimen must not be taken from a marrow site that has been irradiated. It is essential to determine the severity of aplastic anemia. Severe aplastic anemia (SAA) is defined by (1) marrow of less than 25% normal cellularity or marrow of less than 50% normal cellularity in which fewer than 30% of the cells are hematopoietic, and (2) two out of three abnormal peripheral blood values (absolute reticulocyte count < 40,000/µl, absolute neutrophil count [ANC] < 500 µl, or platelet level < 20,000/µl). These criteria have been criticized as being relatively insensitive. Some investigators prefer to identify a cohort of patients with very severe aplastic anemia (VSAA) as those who had an ANC less than 200/µl.24
Figure 3a. Biopsy of Normal Bone Marrow
Shown is biopsy of normal bone marrow.
Figure 3b. Biopsy of Bone Marrow in Aplastic Anemia
Biopsy of bone marrow from a patient with aplastic anemia showing almost complete aplasia.
Aplastic anemia has a number of causes [see Table 2], although in many cases the exact cause cannot be determined.
Table 2 Causes of Aplastic Anemia
Ionizing irradiation and chemotherapeutic drugs used in the management of malignant and immunologic disorders have the capacity to destroy hematopoietic stem cells. With careful dosing and scheduling, recovery is expected. Certain drugs, such as chloramphenicol, produce marrow aplasia that is not dose dependent. Gold therapy and the inhalation of organic solvent vapors (e.g., benzene or glue) can also cause fatal marrow failure.
In 2% to 10% of hepatitis patients, severe aplasia occurs 2 to 3 months after a seemingly typical case of acute disease, usually in young men. Often, the hepatitis has no obvious cause, and tests for hepatitis A, B, and C are negative.25 There is a high incidence of aplastic anemia after liver transplantation in patients with severe non-A, non-B hepatitis.26
Several lines of evidence support the possibility that immune disorders can lead to aplasia. Marrow aplasia occurs in graft versus host disease (GVHD).27 Immunosuppressive preconditioning improves the chances of successful transplantation of syngeneic marrow into patients with aplastic anemia,28 and immunosuppressive therapy has been used successfully to treat idiopathic aplastic anemia.27,28 The blood of some patients with aplastic anemia appears to contain suppressor T cells that suppress the growth of the committed progenitor cells known as colony-forming unit-granulocyte-macrophage (CFU-GM). The suppressor T cells may act by producing interferon gamma.28 The result of these complex immune mechanisms involving suppressor T cells is a profound decrease in primitive hematopoietic cells as measured by both the long-term culture-initiating cell (LTC-IC) assay and the ability to form secondary colonies from the colonies surviving 5 weeks of marrow culture.29
Aplasia can also be part of a prodrome to hairy-cell leukemia [see 12:XV Chronic Lymphoid Leukemias and Plasma Cell Disorders], acute lymphoblastic leukemia [see 12:XVI Acute Leukemia], or, in rare cases, acute myeloid leukemia; or it can develop in the course of myelodysplasia [see 12:XVI Acute Leukemia].
The patient with aplastic anemia may seek medical attention because of fatigue and shortness of breath. Accompanying thrombocytopenia may cause petechiae, oral blood blisters, gingival bleeding, and hematuria depending on the level of the platelet count. By far the major problem associated with aplastic anemia is the recurrent bacterial infections caused by the profound neutropenia. Sepsis, pneumonia, and urinary tract infections are common among patients with aplastic anemia. Invasive fungal infections may cause death, especially in patients with severe neutropenia.
The diagnosis of aplastic anemia requires a marrow aspirate and biopsy [see Figures 3a and 3b], as well as a thorough history of drug exposures, infections, and especially symptoms suggesting viral illnesses and serologic test results for hepatitis, infectious mononucleosis, HIV, and parvovirus [see Figure 4]. Measurement of red cell CD59 is helpful in the diagnosis of paroxysmal nocturnal hemoglobinuria.
Figure 4. Giant Pronormoblast
Giant pronormoblast, evident on this marrow smear, strongly suggests a diagnosis of parvovirus infection.
It is also important to determine the severity of aplastic anemia [see Aplastic Anemia, Definition, above]. Severe cases are associated with a very low rate of spontaneous remission and a mortality of 70% within 1 year. In contrast, 80% of patients who have milder forms of aplastic anemia survive for 1 year.24
The differential diagnosis of pancytopenia includes chronic lymphocytic leukemia, systemic lupus erythematosus, and congestive splenomegaly. In these diseases, however, the marrow is not aplastic but rather shows hyperplasia of the involved cell lines. Other conditions that cause pancytopenia include hypoplastic myelodysplastic syndrome, acute leukemia, megaloblastosis, and large granular lymphocytic leukemia.30
Treatment of Mild Aplastic Anemia
Treatment of milder forms of aplastic anemia involves removing the offending agent and providing supportive therapy, primarily transfusion therapy, anticipating that the remaining pluripotent stem cells will repopulate the marrow.
Thrombocytopenia is often a major problem associated with aplastic anemia. It should be managed by platelet transfusion as needed to control or prevent bleeding. Usually, a threshold of 10,000 platelets/µl is used for transfusion, but conservative treatment is best, and as few transfusions as possible are given. Extensive platelet replacement may result in allosensitization to platelets and may complicate future allogeneic bone marrow transplantation. Red blood cell transfusions are given as required to control the symptoms and signs of anemia.
Granulocyte colony-stimulating factor (G-CSF) and granulocyte-macrophage colony-stimulating factor (GM-CSF) have been given to patients to raise the absolute neutrophil count and help combat infection. They are usually ineffective when used alone, because of the severe deficiency in precursor cells, which are the target for the actions of G-CSF and GM-CSF.31 It is generally preferable to proceed to definitive treatment: immunosuppressive therapy or preferably allogeneic bone marrow transplantation if a matched sibling donor is available [see5:XI Hematopoietic Cell Transplantation].32
Transplantation from a matched sibling after a preparative regimen of high-dose cyclophosphamide and antithymocyte globulin, together with the use of methotrexate and cyclosporine for GVHD prophylaxis, is a very effective regimen for patients with aplastic anemia. Current results suggest a cure rate greater than 90%.33 Results with mismatched or unrelated matched donors are somewhat worse; therefore, patients with aplastic anemia who are without sibling donors are often given a trial of immunosuppressive therapy before transplantation.
Three forms of immunosuppression have been shown to produce partial remission in aplastic anemia.31,32,34 Antithymocyte globulin (ATG) produced sustained remission in about half of the patients in a randomized trial.32 High-dose corticosteroids improved blood counts in about 40% of treated patients, and cyclosporine was also shown to be beneficial.32 (Androgens such as oxymetholone may have a role in the treatment of severe aplastic anemia but are not given alone.31,34)
Although each of these agents can be used individually or consecutively in the treatment of aplastic anemia, a controlled study suggests that results are better when all three are used simultaneously.31,32 The combination of ATG, a corticosteroid, and cyclosporine resulted in an actuarial survival of 62% at 36 months. The first signs of response occurred at about 4 weeks; the median time to remission was 60 to 82 days.32 In this study, patient outcome was related to the quality of hematologic response. An 11-year follow-up report confirmed the effectiveness of the combination of ATG, corticosteroids, and cyclosporine. The relapse rate was 38%, and clonal or malignant diseases developed in 25% of patients.35
One recommendation, based on the usual availability of horse ATG in the United States,31,32 is to administer horse ATG at a dosage of 40 mg/kg/day in 500 ml of saline for 4 days over a period of 4 to 5 hours through an I.V. line equipped with a microaggregate in-line filter. The toxic side effect of ATG is serum sickness, which can usually be controlled with corticosteroids. Prednisone (60 to 100 mg/day) is given orally in divided doses, or methylprednisolone (40 mg) is added to the infusion bottle, and the dose can be increased to 1 mg/kg/day. Corticosteroid therapy is adjusted to control serum sickness, but it can usually be tapered after 2 weeks and stopped after 30 days. Because ATG can lower platelet counts, platelet transfusions are given as needed to maintain the platelet count at more than 20,000 µl.
Cyclosporine (10 to 12 mg/kg/day) is given orally in two divided doses, with the aim of achieving whole blood trough levels of 500 to 800 ng/ml or a serum level of 100 to 200 ng/ml. After 29 days, the cyclosporine dosage can be tapered for a trough whole blood level of 200 to 500 ng/ml.31,32 The cyclosporine is continued for at least 6 months. Cyclosporine can cause hypertension, renal toxicity, hypomagnesemia, vitiligo, tremors, hypertrichosis, susceptibility to Pneumocystis carinii pneumonia (PCP), and gingival hyperplasia.31,32 In one study, 300 mg of aerosolized pentamidine was given every 4 weeks as PCP prophylaxis.32
In another study, G-CSF (5 µg/kg/day) was given subcutaneously for the first 90 days, along with I.V. methylprednisolone (2 mg/kg/day on days 1 through 5, followed by 1 mg/kg/day on days 6 through 10, and tapered off in 30 days), with good results.36
In contrast to patients who undergo allogeneic bone marrow transplantation, patients who respond to immunosuppressive therapy are not actually cured. Many of these patients continue to have moderate cytopenia37; 20% to 36% experience relapses of aplastic anemia,31,32,37and as many as 20% to 36% eventually develop clonal disorders, such as paroxysmal nocturnal hemoglobinuria, myelodysplastic syndrome, and acute leukemia.31,32 Patients also are at increased risk for the development of solid tumors after treatment of aplastic anemia, but the risk is the same for patients who underwent immunosuppressive therapy as it is for those who underwent allogeneic bone marrow transplantation.38 More than 50% of patients who have relapses of aplastic anemia after initially responding to immunosuppressive therapy may respond to a second course of therapy.31,32 For unresponsive patients, a trial of rabbit ATG may work. The rabbit ATG (3.5 mg/kg/day diluted in saline and infused over 6 to 8 hours for 5 consecutive days)39 is given along with cyclosporine (5 mg/kg/day p.o. on days 1 through 180, then tapered) and G-CSF (5 µg/kg/day on days 1 through 90).
An intriguing report concerns 10 patients with severe aplastic anemia who were treated with high-dose I.V. cyclophosphamide (45 mg/kg/day) for 4 consecutive days.40 Some patients also received cyclosporine. Only one course of I.V. cyclophosphamide was given. Seven of 10 patients had a complete hematologic response, and six were still alive after a median follow-up of 10.8 years (range, 7.3 to 17.8 years). However, a trial comparing high-dose cyclophosphamide with ATG was ended early because of excessive cyclophosphamide-induced morbidity and mortality.41 Therefore, the role of high-dose cyclophosphamide in the treatment of aplastic anemia needs extensive clarification.
Treatment of Severe Aplastic Anemia
The choice of appropriate therapy for patients with SAA is influenced by age and disease severity. The European Group for Blood and Marrow Transplantation reported on the results of immunosuppressive therapy in 810 patients subdivided into three age groups: younger than 49, 50 through 59, and older than 60. The 5-year survival rates for those with SAA were 86%, 72%, and 54%, respectively; for those with VSAA, the comparable rates were 49%, 40%, and 21%.42 Older patients had more bleeding and infections.
Patients younger than 20 years
Allogeneic bone marrow transplantation should be performed in patients younger than 20 years if a matched sibling donor is available. Although there are risks, including chronic GVHD and organ dysfunction caused by the conditioning program,31 50% to 80% of patients may be cured; the incidence of later clonal disorders is very low.34 Allogeneic bone marrow transplantation, along with conditioning programs consisting of cyclophosphamide and ATG, produced an actuarial survival rate of 69% after 15 years.34 Patients younger than 20 years who do not have a matched sibling donor should consider transplantation from a matched unrelated donor. Allogeneic transplantation from a matched unrelated donor initially produced a 2-year survival rate of only 29% because of severe GVHD.31 In a study of 15 patients who received unrelated-donor transplantations, all were reported alive at 2 to 86 months (mean follow-up, 51 months); only one patient developed extensive GVHD, and five (33%) developed moderate to acute GVHD. These results suggest that conditioning regimens that contain ATG and cyclophosphamide are improving the treatment outcomes for unrelated donor transplantation in this patient group.43
Patients between 20 and 45 years of age
Patients between 20 and 45 years of age who are in excellent health and have a fully matched sibling donor may be able to tolerate GVHD and thus benefit from the curative potential of an allogeneic bone marrow transplant. Some experts propose that allogeneic bone marrow transplantation should be considered for patients in this age group,34 particularly because newer conditioning programs seem to be capable of reducing the severity of GVHD.31,44 In a study of 154 patients younger than 46 years who received allogeneic transplantation, the median survival was 29 months, and the probability of overall survival at 5 years was 56%.45
Patients older than 45 years
Previously it was thought that the impact of GVHD was too severe for patients older than 45 years, and it was suggested that these patients receive immunosuppressive therapy.31,34 However, conditioning programs containing ATG and cyclophosphamide seem to be more tolerable, and even heavily pretreated patients as old as 59 years have done well after allogeneic marrow transplantation.46
ACQUIRED PURE RED CELL APLASIA
In adults, pure red cell aplasia (PRCA) is an acquired disorder. The anemia is severe (hematocrit usually less than 20%), reticulocytopenia is profound (often 0%), the absolute reticulocyte count is usually less than 10,000/µl, and marrow erythroid precursors are virtually absent. Marrow myeloid and megakaryocytic elements are preserved, however, and the peripheral platelet and white blood cell counts are also normal.
In PRCA, erythropoiesis is thought to be inhibited primarily by immune mechanisms, including autoantibody-mediated and T cell-mediated suppression of erythroid progenitors, usually at a stage after the CFU-E stage of erythroid differentiation and before formation of proerythroblasts. T cells, particularly of the large granular lymphocyte (T-LGL) class, may be involved in the suppression of erythropoiesis, and in some cases, there is evidence that the suppression is caused by clonal T cells.47 Autoantibody inhibition of erythropoietin has also been described, but it is quite uncommon.48 Two other mechanisms probably cause PRCA: (1) a specific attack on erythroid precursors by the parvovirus B19 (one report indicated that 14% of cases were caused by this virus49) and (2) an underlying hematopoietic clonal abnormality that may be a prodrome to myelodysplastic syndrome.48
PRCA may be caused by a variety of processes, including neoplasia, autoimmune disorders, drugs, and infections [see Table 3].
Table 3 Causes of Acquired Pure Red Cell Aplasia
The association of PRCA with LGL proliferation and leukemia is increasingly being recognized.30 The routine use of T cell receptor gene rearrangement studies in one series showed that nine of 14 patients had a clonal LGL disorder.50 Presumably, these LGL cells directly mediate inhibition of erythropoiesis.49,50 In perhaps as many as 20% of cases, PRCA may be a prodrome to the myelodysplastic syndromes or acute myeloid leukemia.49,51
Erythroblastopenia also occurs in a small percentage of patients with autoimmune hemolytic anemia [see 5:IV Hemoglobinopathies and Hemolytic Anemias] and may be caused by autoantibody attack on maturing normoblasts.
The treatment of HIV infection with zidovudine (AZT) produces, in virtually all patients, an anemia that is usually marked by significant macrocytosis.52 Moderate erythroid hypoplasia is the usual cause of this anemia, which can progress to PRCA.
Parvovirus infection is the cause of the transient aplastic crises that occur in patients who have severe hemolytic disorders. The marrow in patients with such disorders must compensate for the peripheral hemolysis by increasing its production up to sevenfold and thus typically shows an intense erythroid hyperplasia. Although parvovirus can affect all precursor cells, the red cell precursors are the most profoundly affected.49
PRCA can complicate ABO-incompatible allogeneic bone marrow transplantation; the recipient's serum continues to express anti-A or anti-B isohemagglutinins against donor A or B antigen expressed on the surface of erythroid progenitors.51 With PRCA of pregnancy, antibodies against BFU-E usually disappear after delivery, coinciding with clinical remission.53
The patient with PRCA presents with symptoms characteristic of anemia—namely, weakness, fatigue, and shortness of breath. White blood cell and platelet counts are normal morphologically and functionally. A very low reticulocyte count—either a relative reticulocyte value of less than 0.2% or a very low absolute reticulocyte count of less than 10,000 µl—should prompt the physician to order a bone marrow aspirate. In a patient with PRCA, a bone marrow aspirate and biopsy typically show normal myelopoiesis, lymphopoiesis, and megakaryocytopoiesis; erythropoiesis is virtually absent. In the absence of any apparent cause of PRCA, four conditions must be considered: idiopathic PRCA, thymoma, hypoplastic myelodysplastic syndromes (MDS), and LGL leukemia. The workup to diagnose PRCA usually includes computed tomography of the chest to evaluate the possibility of thymoma, immunophenotypic analysis of circulating blood or marrow lymphocytes to identify LGL proliferation, marrow cytogenetics to evaluate the possibility of MDS, and antibody tests for parvovirus.49 A diagnostic hallmark of parvovirus infection is the appearance of giant pronormoblasts in the marrow [see Figure 4]. The distinction between PRCA associated with the myelodysplastic syndromes and acute myeloid leukemia may be difficult to determine at the time of diagnosis unless a typical myelodysplastic cytogenetic abnormality is detected during a bone marrow examination.
Two general principles of management in PRCA are transfusions for symptomatic anemia and cessation of possible offending drugs. No specific therapy is indicated in those forms of PRCA that are self-limited, such as pregnancy, ABO-incompatible bone marrow transplantation, and some cases of parvovirus infection.51,53 Treatment of PRCA depends on the identified cause. If a thymoma is present, it should be removed surgically52; this procedure leads to patient improvement in about one third of such cases.51 When surgery is impossible, one should consider a course of prednisone combined with octreotide, a somatostatin analogue that binds to thymomas and may inhibit the function of thymic immune cells.54
Treatment of other causes of PRCA is based on the supposition that the attack is immune mediated and therefore will respond to immunosuppressive therapy. Treatment can begin with the administration of 60 mg of oral prednisone daily in divided doses; this regimen should be continued for 1 to 3 months.49 If a patient fails to respond, as indicated by a rise in the reticulocyte count, cyclophosphamide or azathioprine should be added at a dosage of 2 to 3 mg/kg/day orally. Patients with marrow cytogenetic abnormalities suggestive of myelodysplastic syndrome respond poorly.49,50 Some patients who are refractory to other forms of therapy have responded well to I.V. IgG (0.4 g/kg/day for 5 days).55
Patients with LGL proliferation as the underlying cause respond well to cyclophosphamide.50,56 Usually, low doses of cyclophosphamide (50 to 100 mg/day p.o.) for 3 to 6 months suffice to produce remission, which is sometimes associated with disappearance of LGL proliferation.50,57 Patients who respond poorly usually respond to oral cyclosporine.50,57 Cyclosporine (12 mg/kg/ day) has been shown to produce responses of approximately 65%, even in patients who did not respond to corticosteroids, plasmapheresis, cyclophosphamide, or azathioprine therapy.51,58
For patients in whom parvovirus infection is the cause of PRCA, I.V. IgG works well; the standard dosage is 0.4 g/kg/day for 5 days.49 For AIDS patients with parvovirus infection and PRCA, I.V. IgG may have to be continued.57 Recovery from the transient crises of parvovirus infection occurs spontaneously in 1 to 2 weeks after onset of the infection.
ATG therapy for patients with refractory PRCA is similar to that for patients with aplastic anemia (40 mg/kg/day I.V. for 4 days).51 Other drugs that have been used in refactory cases are azathioprine (2 to 3 mg/kg/day), antilymphocyte globulin, and anti-CD20 monoclonal antibody.31 In very refractory cases, allogeneic bone marrow transplantation can be effective.59
Production Defects with Marrow Erythroid Hyperplasia and Ineffective Erythropoiesis
Anemia with a low reticulocyte count may occur despite intense marrow erythroid hyperplasia. This paradoxic situation is the hallmark of ineffective erythropoiesis or intramedullary hemolysis. Generalized erythroid impairment may be present, or specific subpopulations of erythroid precursors may be involved. Some of these subpopulations escape death in the marrow, but their progeny are so severely damaged that they are rapidly removed from the circulation, thus giving the picture of peripheral hemolysis. Other signs of ineffective erythropoiesis include jaundice, a very high serum lactic dehydrogenase level, and 75% to 90% saturation of serum iron-binding capacity. The classic ferrokinetic picture shows rapid plasma iron clearance, which indicates intense erythroid precursor activity. The delivery of labeled red blood cells to the peripheral circulation, however, is dramatically reduced, which suggests that the precursors are being destroyed by intramedullary hemolysis.
The differential diagnosis includes megaloblastic anemias, sideroblastic anemias, thalassemia [see 5:IV Hemoglobinopathies and Hemolytic Anemias], myelodysplastic syndromes [see 12:XVI Acute Leukemia], and agnogenic myeloid metaplasia [see 12:XVII Chronic Myelogenous Leukemia and Other Myeloproliferative Disorders].
Megaloblastic anemias are caused by cobalamin or folic acid deficiency, by drugs that interfere with the synthesis of DNA or with the absorption or metabolism of cobalamin, and by genetic disorders that interfere with DNA metabolism or with the absorption or distribution of cobalamin.
Megaloblastic erythropoiesis is characterized by defective DNA synthesis and arrest at the G2 phase, with impaired maturation and a buildup of cells that do not synthesize DNA and that contain anomalous DNA. This condition leads to asynchronous maturation between the nucleus and cytoplasm.60 RNA production and protein synthesis continue; thus, larger cells, or megaloblasts, are produced. Ineffective erythropoiesis results, and there is disagreement about the presence of increased apoptosis.61,62 It is presumed that similar defects in DNA synthesis characterize the mucosal abnormalities of the stomach and tongue. In the granulocytic line, the presence of giant metamyelocytes represents ineffective granulopoiesis.60
The role of folic acid and cobalamin
The interactions between folic acid and cobalamin are critical in the metabolism of single carbon units, mainly methylene and formyl analogues, which have a key role in the synthesis of DNA and purines [see Figure 5, part a].63 There are two major coenzymes of cobalamin, adenosylcobalamin and methylcobalamin. Adenosylcobalamin is the coenzyme for methylmalonyl-coenzyme A mutase, which catalyzes a step in the catabolism of propionic acid [see Figure 5b].63 Methylcobalamin is the coenzyme for methionine synthase, which functions as a methyltransferase in the reaction that converts 5-methyltetrahydrofolate (CH3-THF1) to tetrahydrofolate (THF1) [see Figure 5, part a].63Cobalamin and folic acid [see Figure 6] combine in the methionine synthase reaction [see Figure 5, part a], in which the methyl group of CH3-THF1 is transferred to cobalamin to form methylcobalamin. Methylcobalamin then transfers its methyl group to homocysteine to form methionine. The monoglutamated THF1, which is formed by this reaction, is polyglutamated by the enzyme folylpolyglutamate synthase, and a methylene group is added to it by the serine-glycine methyltransferase to form 5,10-methylene THFn. 5,10-Methylene THFn provides its methylene to convert deoxyuridylate to thymidylate, a key step in DNA synthesis. 5,10-Methylene THFn can also be directly converted to CH3-THF1 by the enzyme 5,10-methylene tetrahydrofolate reductase, thereby making its methyl group available.
Figure 5. DNA Synthesis and Metabolism
(a) Intracellular interdependent cofactor activity of cobalamin and folic acid is essential in DNA synthesis and metabolism.63 (b) Adeno-sylcobalamin is a cofactor in the synthesis of succinyl-coenzyme A from methylmalonyl-coenzyme A.63 (CoA—coenzyme A)
Figure 6. Structure of Folic Acid
Folic acid functions as a coenzyme in single-carbon transfer reactions. It is not physiologically active until it is reduced at positions 5, 6, 7, and 8 to tetrahydrofolate (THF). Single-carbon groups (R) such as methyl analogues and formate are added at either position 5 or position 10, or they may bridge from 5 to 10, as shown. There may be several glutamates attached in sequence (R1), which convert the monoglutamate to the polyglutamate form. Enzymes of the intestinal mucosa split polyglutamates back to monoglutamate, whereas liver enzymes add glutamate to tetrahydrofolate or to other reduced folic acids.
Formyl THFn (also called leucovorin, folinic acid, or citrovorum factor) has an important role in purine synthesis and DNA metabolism. It can be generated by oxidation of 5,10-methylene THFn or directly from THFn by the enzyme formyl THF synthase, with methionine providing the formate group [see Figure 5, part a].63 When cobalamin is deficient, CH3-THF1 cannot transfer its methyl group to cobalamin; therefore, THF1 is not free to be polyglutamated by folylpolyglutamate synthase [see Figure 5, part a]. The polyglutamated form is required for synthesis of either 5,10-methylene THFn or formyl THFn; thus, DNA synthesis and purine synthesis are blocked. This hypothesis, the methylfolate trap hypothesis, is supported by the finding of increased levels of CH3-THF1 in the plasma of cobalamin-deficient patients. An alternative explanation is the formate starvation hypothesis, wherein cobalamin deficiency impairs methionine generation, which therefore cannot provide the methyl groups needed by the enzyme formyl THFn synthase to produce formyl THFn.
Other aspects of folic acid and cobalamin metabolism
Neither folic acid nor cobalamin is produced by humans in adequate amounts; both must be absorbed from food. Cobalamin, in particular, is derived from microbial sources and is ingested in the form of meat or eggs.
Most of the dietary folic acid is in the polyglutamate form and is absorbed at the intestinal mucosa. Absorption of radioactively labeled folic acid approaches 80% of a 200 µg dose.60,63 The serum folic acid level appears to be maintained by folic acid absorbed from food. Enterohepatic circulation of folic acid has been observed in which folic acid passing into the bile and small intestine is quantitatively reabsorbed. In an animal model, ethanol administration blocks the entry of folic acid into the bile. This effect could account, in part, for the sharp fall in the serum folic acid level seen 8 hours after alcohol consumption. A similar fall in serum folic acid level follows phenytoin ingestion. The daily requirement of cobalamin is about 1 µg, and the amount usually provided by the Western diet, which is rich in animal products, is about 5 to 15 µg.64
R proteins are a class of cobalamin-binding glycoproteins found in saliva and gastric juice; they are produced by granulocytes and other tissues. Intrinsic factor (IF) is a 45 kd glycoprotein, secreted by gastric parietal cells, that is highly specific for unaltered cobalamin. The R protein-cobalamin complex does not bind to ileal receptors and thus is not absorbed. In the stomach, cobalamin binds preferentially to R proteins rather than to IF60,63,64; thus, it is the physiologically inactive R protein-cobalamin complex that is discharged into the duodenum. In the duodenum and small intestine, however, the pancreatic proteases along with pepsin degrade the R proteins, freeing cobalamin and allowing it to bind to IF. Thus, gastric atrophy and pancreatic insufficiency contribute to cobalamin malabsorption.63,64 The IF-cobalamin complex, in the presence of Ca2+ and at a pH level greater than 5.4, binds specifically to a limited number of sites on the microvilli of mucosal cells in the terminal portion of the ileum, where absorption takes place [see Figure 7].63,64
Figure 7. Cobalamin Assimilation
Cobalamin assimilation. Dietary cobalamin (Cbl) enters the stomach and binds to R protein. This physiologically inactive complex enters the duodenum. In the small intestine, pancreatic enzymes and pepsin digest the R protein, and Cbl binds to intrinsic factor (IF). The Cbl-IF complex passes through the intestine until it reaches receptors on the microvilli of mucosal cells in the distal ileum. The Cbl is then transferred to transcobalamin II (TC-II), which circulates in the blood until it binds to receptors on cells in the body and is internalized.
In the plasma, most of the cobalamin is bound to the physiologically unimportant R proteins, transcobalamins I and III (TC-I and TC-III), which are about 70% saturated with cobalamin.65 The physiologically important transport protein is transcobalamin II (TC-II), which has considerable specificity for cobalamin and is only 5% to 10% saturated with cobalamin. Receptors for the TC-II-cobalamin complex are present on many cell membranes. TC-II binds about 90% of a newly injected dose of cobalamin; and the complex is rapidly cleared, with a half-life of 6 to 9 minutes.66,67 In persons with congenital TC-II deficiency, which results in severe megaloblastic anemia, both plasma cobalamin transport and cobalamin absorption are impaired. Impaired cobalamin absorption implies that TC-II has a role within the ileal enterocyte, where cobalamin is transferred from IF to TC-II.
The elevation of cobalamin levels seen in patients with chronic granulocytic leukemia or significant granulocytosis is caused by increases in TC-I and, to a lesser extent, TC-III, which are produced in granulocytes.
MEGALOBLASTIC ANEMIA CAUSED BY COBALAMIN DEFICIENCY (PERNICIOUS ANEMIA)
Cobalamin deficiency in pernicious anemia is thought to result from an autoimmune gastritis and an autoimmune attack on gastric intrinsic factor. There are two types of anti-IF antibodies: one of these antibodies blocks attachment of cobalamin to IF, and the other blocks attachment of the IF-cobalamin complex to ileal receptors.66 Clinically, highly specific anti-IF antibodies are found in about 70% of patients with pernicious anemia. A second component of pernicious anemia is chronic atrophic gastritis that leads to a decline in IF production. The chronic atrophic gastritis in pernicious anemia is also associated with an increased risk of intestinal-type gastric cancer and of gastric carcinoid tumors.67 Pernicious anemia occurs in association with other autoimmune disorders. In one study, autoimmune thyroid disorders were observed in 24% of 162 patients with pernicious anemia.68
In addition to macrocytic and megaloblastic anemia, the patient with cobalamin deficiency may present with weakness, lethargy, jaundice, and dementia, as well as atrophy of the lingual papillae and glossitis. Neuropathy is the presenting feature in about 12% of patients with cobalamin (vitamin B12) deficiency without concomitant anemia.69 Patients with severe cobalamin deficiency initially complain of paresthesia. The sense of touch and temperature sensitivity may be minimally impaired. Memory impairment and depression may be prominent.69 The disease may progress, involving the dorsal columns, causing ataxia and weakness. The physical examination reveals a broad-base gait, Romberg sign, slowed reflexes, and loss of sense of position and feeling of vibration (especially when tested with a 256 Hz tuning fork). If the disorder is not detected and treated, the lateral columns become involved, resulting in weakness, spasticity, inability to walk, sustained clonus, hyperreflexia, and Babinski sign. Because the peripheral nerves, as well as the dorsal and lateral columns, are involved, these neurologic manifestations are sometimes termed subacute combined degeneration or subacute combined system disease.
Cobalamin deficiency appears to be the cause of various neuropsychiatric disorders, with such symptoms as paresthesia, ataxia, limb weakness, gait disturbance, memory defects, hallucinations, and personality and mood changes.65 These symptoms, however, cannot be easily accounted for by the type of spinal cord lesions that occur in patients with cobalamin deficiency. Investigators have tried to determine whether a defect in methionine synthesis or an abnormality in propionic acid metabolism accounts for the neuropathy associated with cobalamin deficiency [see Figure 5, part b], but the exact mechanism remains obscure. Accruing evidence supports the impairment of methionine synthase as the cause of the neuropathy.60,69 A recent study measured various metabolites of cobalamin and discovered that only high levels of plasma cysteine were predictive of neurologic dysfunction.70
The evaluation of suspected cobalamin deficiency generally proceeds in two stages: documenting the presence of the vitamin deficiency and determining its cause (e.g., pernicious anemia, malabsorption, dietary lack). The diagnosis can often be established by (1) measurement of the serum cobalamin concentration, (2) evaluation of specific metabolites, and (3) use of the Schilling test to establish malabsorption of cobalamin.
Macrocytosis (mean corpuscular volume [MCV] greater than 100 fl) is a hallmark of cobalamin deficiency, but it may be masked by concurrent disorders, such as iron deficiency. If macrocytocis is not apparent on examination of the peripheral smear, it is easily detected when red blood cell counts are made with an electronic particle counter. The peripheral smear shows macro-ovalocytes, fish-tailed red blood cells, hypersegmented neutrophils, and, occasionally, nucleated red blood cells [see Figures 2a, 2b, and 2c]. The finding that a single polymorphonuclear neutrophil has six lobes or that 5% of PMNs have five lobes constitutes strong evidence of megaloblastic anemia. In severe cases, granulocytopenia and thrombocytopenia are present. Examination of the bone marrow is usually not necessary, but if performed, it reveals megaloblastic erythroid hyperplasia and giant metamyelocytes.60 If severe iron deficiency is concurrent with macrocytosis, the full morphologic expression of megaloblastosis is blocked, although the giant metamyelocytes in the marrow and hypersegmented PMNs in the peripheral blood will still be present.
Plasma cobalamin levels and red blood cell folic acid levels should be measured if the MCV is greater than 100 fl. If performed, a bone marrow aspirate and biopsy typically reveal enormous megaloblastic erythroid hyperplasia with giant metamyelocytes.60 The hypercellularity detected on bone marrow examination can be so dramatic and megaloblasts so immature that clinicians still sometimes make the erroneous diagnosis of leukemia.60,71
The standard approach to determining the cause of proven cobalamin deficiency has traditionally relied on the Schilling test, which is becoming difficult to order. The Schilling test measures the absorption of cobalamin labeled with cobalt-57 (57Co). After 1 µg of radioactively labeled cobalamin is given orally, 1,000 µg of unlabeled cobalamin is given parenterally. The parenteral dose saturates transcobalamins I, II, and III, so that a significant portion of the absorbed material is flushed and excreted in the urine. If the amount of57Co-labeled cobalamin measured in an accurately collected 24-hour urine sample is less than 10% of the dose that was administered orally, cobalamin absorption is poor.
There are increasing numbers of reports of patients with proven pernicious anemia who have low or borderline serum cobalamin levels but normal Schilling test results. As the gastric atrophic lesion of pernicious anemia progresses, the ability to produce acid-pepsin is lost before all IF activity disappears. Thus, the ability to cleave the R protein-cobalamin complex, freeing cobalamin to bind to IF, is impaired. Coexisting infection with Helicobacter pylori may further impair production of acid-pepsin.72 However, there may be sufficient IF to bind the free oral cobalamin administered in the Schilling test and therefore yield a normal value.
Malabsorption of cobalamin can be demonstrated by means of a food Schilling test, which is not available clinically. This test is performed with eggs from chickens that have been injected with radioactive cobalamin73 and indicates whether there is insufficient acid-pepsin to split the cobalamin-enzyme complex and release free cobalamin to be bound by IF. If pernicious anemia is strongly suspected in a patient whose Schilling test result is apparently normal and whose plasma cobalamin is not diagnostically low, other steps should be taken to confirm the diagnosis, including examination of the blood cell morphology, measurement of the anti-IF antibody, or performance of a therapeutic trial with parenterally administered cobalamin. Measurement of the serum levels of homocysteine and methylmalonic acid is increasingly being used, because both are elevated as a consequence of cobalamin deficiency [see Figures 5, parts a and b].74
If the initial Schilling test demonstrates reduced excretion of cobalamin, a second phase of the test may be conducted, aimed at correcting cobalamin absorption caused by pernicious anemia. In this phase of the test, supplementary oral IF is administered and will normalize the cobalamin absorption unless the supplementary IF is not fully active, the patient secretes antibodies to IF, or the patient is taking drugs that interfere with cobalamin absorption. In no case, however, will supplementary cobalamin be effective in patients with intestinal malabsorption. It is important to recognize that prolonged cobalamin deficiency impairs intestinal epithelial cells and thus impairs absorption. Therefore, the second stage of the test should only be performed after several weeks of cobalamin replacement therapy. If the result of the second stage of the Schilling test is abnormally low, this suggests the presence of generalized malabsorption, such as may occur in sprue, pancreatic insufficiency, or blind loop syndromes.
Factors affecting test results
Concurrent α-thalassemia may minimize the macrocytosis of pernicious anemia.75 This possibility should be considered particularly in patients of African descent, among whom there is a high incidence of α-thalassemia (about 30%). Anemia of chronic disease or anemia resulting from blood loss and iron deficiency can also reduce the degree of macrocytosis but will not affect the hypersegmentation of neutrophils. In one study, iron deficiency was discovered in 20% of 121 patients with pernicious anemia76; in another study, 19% of patients with pernicious anemia were not anemic, and 33% did not have macrocytosis.77
Falsely low serum cobalamin levels occur during pregnancy and in folic acid deficiency states.74 In the past, a decline in the serum cobalamin level was usually not considered important unless the value was very low (i.e., < 150 pg/ml). It has become clear, however, that patients with serum cobalamin levels as high as 250 pg/ml and perhaps higher may have cobalamin deficiency.73,77,78 Fortunately, the finding of macro-ovalocytes or hypersegmented PMNs on the peripheral smear remains a sensitive indicator for the presence of cobalamin deficiency.
Determining the underlying cause
After the presence of macrocytosis and a reduced cobalamin level have been identified, the cause of these conditions must be determined. It is important to remember that macrocytosis can be caused by conditions other than pernicious anemia, including folic acid deficiency, liver disease, alcohol abuse, reticulocytosis, and ingestion of drugs such as antimetabolites, alkylating agents, and zidovudine.52,75 Cobalamin deficiency can be caused by inadequate absorption resulting from gastric abnormalities (e.g., pernicious anemia, gastritis) and small bowel disease (e.g., tropical sprue, Crohn disease), and pancreatic insufficiency [see Table 4].
Table 4 Causes of Cobalamin Deficiency
Gastric surgery in which the IF, pepsin, and acid-secreting components are removed often results in cobalamin deficiency (it occurred in 31% of patients in one study79). Patients who have undergone gastric surgery should be regularly screened by measurements of plasma cobalamin or homocysteine levels and supplemented with lifelong cobalamin therapy if the levels are low.79
Pancreatic insufficiency can result in malabsorption of cobalamin if the damaged pancreas does not produce enough trypsin and chymotrypsin for digesting the R protein-cobalamin complex and freeing the vitamin to form the complex with IF [see Figure 7] [see 4:V Diseases of the Pancreas].
There are other causes of cobalamin deficiency. In vegetarians, especially vegans, profound nutritional megaloblastic anemia can develop as a result of very low cobalamin intake. Deficiencies of folic acid and iron have also been observed in vegans.80 A careful patient history should indicate the possibility of inadequate dietary intake of cobalamin. Infants of vegan mothers can become severely cobalamin deficient, particularly when they are breast-fed.65 Cobalamin deficiency is surprisingly common in less well developed countries where people are not strict vegans.65 The incidence is particularly high in pregnant women and in preschool-age children.65
Specific replacement should be started promptly after the diagnosis has been made and serum samples have been taken to determine cobalamin levels. Patients who have a low serum cobalamin level and macrocytic anemia should undergo a trial of parenteral cobalamin therapy. The diagnosis of cobalamin deficiency is confirmed if cobalamin therapy produces a reticulocytosis in 3 to 4 days that is associated with a rise in the hemoglobin level and a fall in the MCV.
If the patient has symptoms of severe anemia, packed red blood cells can be transfused; the transfusion should be administered very slowly to avoid precipitating or aggravating congestive heart failure. This circumstance is one of the few in which a single-unit transfusion may be justified, because it may produce a 25% increase in oxygen-carrying capacity. A large dose of cobalamin should be given because the retention of parenterally administered cobalamin is poor but variable; the vitamin is inexpensive and has no harmful side effects. The reticulocyte response begins in 4 to 6 days, and the granulocyte count, if low, begins to increase at the same time. The hypersegmentation of PMNs disappears after 10 to 14 days, which suggests that in the megaloblastic anemias, granulopoiesis is affected by cobalamin deficiency at two different steps: (1) the lobe number of the PMNs is determined, and (2) granulocytes mature and leave the marrow.55Weekly dosages of 1,000 µg of parenteral cobalamin for 6 weeks should be followed by parenteral dosages of 1,000 µg monthly for life. The standard parenteral preparation is cyanocobalamin. For pancreatic insufficiency, cobalamin can be given parenterally or pancreatic enzymes can be administered orally. Specific therapy must be designed for patients with intestinal forms of malabsorption.
Because a small amount of cobalamin is absorbed even in the absence of IF and because only 1 µg/day is required, oral cobalamin has proved adequate for replacement in patients with pernicious anemia, freeing the patient from monthly injections (2,000 µg/day p.o. is recommended).81
MEGALOBLASTIC ANEMIA CAUSED BY FOLIC ACID DEFICIENCY
The patient with folic acid deficiency has a clinical presentation that is distinct from that of the patient with cobalamin deficiency.82 The patient may abuse alcohol or other drugs and have poor dietary intake of folic acid. Patients with folic acid deficiency are more often malnourished than those with cobalamin deficiency. The gastrointestinal symptoms in folic acid deficiency are similar to those in cobalamin deficiency but may be more severe than those in pernicious anemia. Diarrhea is often present. The hematologic manifestations of folic acid deficiency are the same as those of cobalamin deficiency: severe macrocytic anemia, a low absolute reticulocyte count, and a characteristic blood smear showing macro-ovalocytes, occasional megaloblasts, and hypersegmented neutrophils. Patients with megaloblastic anemia who do not have glossitis, a family history of pernicious anemia, or the neurologic features described for cobalamin deficiency may have folic acid deficiency.
A meticulous dietary history is important because food faddism, poor dietary intake, and alcoholism are the usual causes of severe folic acid deficiency [see Table 5]. Cobalamin and folic acid deficiencies frequently coexist and are not easily distinguished. In evaluating patients for folic acid deficiency, values for the levels of serum folic acid, serum cobalamin, and red blood cell folic acid must be obtained. The red blood cell folic acid level reflects tissue stores83 but may be reduced in patients with severe cobalamin deficiency. In isolated cases, the serum folic acid level of cobalamin-deficient patients is usually normal or elevated. Severe, long-standing cobalamin deficiency leads to anorexia and GI disturbances, which may cause dietary folic acid deficiency. As a result, both serum cobalamin and folic acid levels are low, producing a double-deficiency state.
Table 5 Causes of Folic Acid Deficiency
A serum folic acid level less than 2 ng/ml is consistent with folic acid deficiency, as is a red blood cell folic acid level less than 150 ng/ml. If the test results are inconclusive or if it is necessary to distinguish the megaloblastosis of folic acid deficiency from that of cobalamin deficiency, measurements of the serum methylmalonate and homocysteine levels are helpful. If both metabolite tests are normal (i.e., methylmalonate level of 70 to 270 nmol/L and total homocysteine level of 5 to 14 µmol/L), deficiency of both vitamins is ruled out. If the methylmalonate level is normal but the total homocysteine level is increased, folic acid deficiency is likely and investigation into the underlying cause is appropriate.83
Determining the underlying cause
Folic acid deficiency is most frequently caused by poor dietary intake, but it may also result from inadequate absorption secondary to disease or drug administration [see Table 5]. Ingestion of ethanol by well-nourished individuals does not produce megaloblastosis, but in patients with borderline folic acid stores, ethanol can lower serum folic acid levels and block the reticulocyte response to folic acid administration. Alcohol may block release of folic acid from tissues to the serum.
Megaloblastic anemia occurring as a consequence of drug administration or pregnancy is likely to be caused by folic acid deficiency. Many of the antineoplastic and immunosuppressive agents produce megaloblastosis; these include fluorouracil, hydroxyurea, mercaptopurine, thioguanine, cytarabine, and azathioprine. In pregnant women, the presence of megaloblastosis may not be initially apparent. Because the combination of folic acid and iron deficiency is common, full expression of megaloblastosis is often blocked, and the patient will have a dimorphic anemia rather than the easily identifiable macro-ovalocytosis. Hypersegmentation of PMNs persists.60,83
An abnormality in folate metabolism can be caused by a chromosomal mutation, and women who are homozygous for this defect are thought to be at higher risk for pregnancies affected by neural tube defects. One of the enzymes that regulates homocysteine levels, 5,10-methylenetetrahydrofolate reductase has a genetic variant, C677T. Individuals homozygous for this variant have increased plasma homocysteine levels that are lowered by folate supplementation. About 5% to 10% of the general population are homozygous for this variant. Both pregnant and nonpregnant women who are homozygous for the C677T mutation have significantly lower red blood cell folic acid levels.84 These women may be susceptible to cardiovascular disease and stroke and may bear children with neural tube defects.84,85 It would be advisable to know before pregnancy that a woman is homozygous for this variant, and genetic testing would be helpful if a woman has a family history of this defect.
A number of intestinal disorders cause folic acid deficiency. These include severe pancreatic disease and small bowel disease, including malabsorption, ileal disease, Crohn disease, resection, and bypass [see Table 5]. When there is no apparent cause of cobalamin deficiency, it may be practical to suspect an undiagnosed disease of malabsorption. In one prospective study of patients who had laboratory-defined folate deficiency, 10.9% were positive for celiac disease antibodies and 4.7% had histologically confirmed celiac disease.86
Standard therapy for folic acid deficiency is 1 mg/day orally. The response, manifested by reticulocytosis in 4 to 6 days, loss of megaloblastosis, and the return of normal blood counts, confirms the diagnosis of folic acid deficiency. Neutrophil hypersegmentation disappears only after 10 to 14 days, however.60 Patients with megaloblastosis and severe bone marrow depression secondary to administration of drugs that block dihydrofolate reductase, such as pyrimethamine and methotrexate, may be treated with folinic acid. In the case of toxicity after single large doses of methotrexate, a single equivalent dose of I.M. folinic acid (i.e., milligram for milligram) will suffice. For toxicity after chronic pyrimethamine therapy, 1 to 5 mg of folinic acid daily can be given without blocking the antimalarial effects of pyrimethamine. Megaloblastosis caused by anticonvulsant therapy can be treated with 1 mg of folic acid daily. Supplementation during pregnancy is advised and may also be useful for patients who have severe chronic hemolysis.
In most patients (i.e., those who do not require a large amount of folic acid because of conditions such as hemolysis or pregnancy), a hematologic response occurs after administration of 200 µg of folic acid daily. The increased demand of folic acid during pregnancy requires administration of about 200 to 300 µg/day.87 Furthermore, folic acid supplementation seems to prevent fetal neural tube defects.88 Such neural tube defects may occur in the embryo or very early in gestation—even before the pregnancy is confirmed.89,90 Therefore, it is recommended that women of childbearing age or those who plan to become pregnant receive about 400 µg of folic acid a day. Women who are homozygous for the C677T mutation should also take folic acid supplements. Staple foods such as flour and cereal grains can be fortified with folic acid. Concern has been expressed, however, that folic acid supplementation may mask the megaloblastosis of pernicious anemia, causing the development of severe neuropathy rather than anemia.89
The sideroblastic anemias are a heterogeneous group of disorders characterized by anemia, ringed sideroblasts in the marrow, and ineffective erythropoiesis.91 There are hereditary and acquired forms; the latter are subdivided into benign and malignant variants. A fairly common form is the myelodysplastic syndrome called refractory anemia with ringed sideroblasts. Other than alcohol and drugs (e.g., isoniazid), the secondary causes of these diseases remain largely unknown.
Abnormalities of heme synthesis are probably the most frequent cause of the hereditary sideroblastic anemias. Molecular defects of the enzyme 5-aminolevulinate synthase have been described as the cause of this abnormality.90,92 This enzyme initiates the heme synthetic pathway, and its impairment profoundly affects heme synthesis. In other cases, there are major deletions in mitochondrial DNA. Iron enters erythroid precursors, but because heme synthesis is impaired, the iron cannot be incorporated into heme and accumulates on the cristae of mitochondria.90
The principal feature common to all sideroblastic anemias is a refractory or progressive anemia. However, mild, lifelong anemia may go unnoticed. The diagnosis of sideroblastic anemia is established by reticulocytopenia; the red blood cells on smear are frequently profoundly hypochromic and microcytic, and distorted red blood cells and basophilic stippling may be noted.93,94 Occasionally, Pappenheimer bodies (deposits of iron that stain with the Prussian blue reagent) are present in the red blood cells. There are ringed sideroblasts seen on the marrow aspirate (bone marrow normoblasts with heavy incrustations of nonferritin iron on the mitochondria) [see Figure 8]. Because of ineffective erythropoiesis, there is saturation of serum iron-binding capacity (usually approaching 80%) and elevation of the serum lactate dehydrogenase level. Cytogenetic study of the bone marrow may reveal one of the typical patterns seen in the myelodysplastic syndromes. The sideroblastic anemias can be classified into four groupings: hereditary (probably benign), acquired (probably benign), probably benign, and clonal disorder [see Table 6].
Figure 8. Bone Marrow in Idiopathic Sideroblastic Anemia
Prussian blue stain shows ringed sideroblasts in the bone marrow of a patient who has idiopathic sideroblastic anemia.
Table 6 Sideroblastic Anemias
For prognostic purposes, it is important to decide whether the patient has a benign or malignant form of sideroblastic anemia. It is also important to recognize reversible forms of sideroblastic anemia (e.g., those caused by alcoholism, folic acid deficiency, and drugs such as isoniazid and chloramphenicol) and to discontinue any potentially offending agents.
Indicators of myelodysplasia include granulocytopenia, thrombocytopenia, dysplastic marrow granulopoiesis, bilobed megakaryocytes, and typical cytogenetic abnormalities. In rare cases, patients have a reticulocyte and hemoglobin response to pyridoxine (200 to 600 mg/day), with or without folic acid.95
Figures 1 and 6 Talar Agasyan.
Figure 5 Alan D. Iselin.
Figure 7 Tom Moore.
Editors: Dale, David C.; Federman, Daniel D.