Myelodysplastic syndromes (MDS) are clonal myeloid disorders characterized by progressive peripheral blood cytopenias associated with ineffective myelopoiesis. They are typically considered neoplasms because of frequent genetic aberrations and patient-limited survival with progression to acute myeloid leukemia (AML) or death related to the consequences of bone marrow failure including infection, hemorrhage, and iron overload. A progression to AML has always been recognized among the myeloproliferative disorders (MPD) but occurs only rarely among those with essential thrombocythemia (ET). Yet, the World Health Organization (WHO) has chosen to apply the designation myeloproliferative neoplasms (MPN), for all MPD but has not similarly recommended that all MDS become the myelodysplastic neoplasms (MDN). This apparent dichotomy may reflect the extremely diverse nature of MDS. Moreover, the term MDS is occasionally inappropriately applied to hematologic disorders associated with acquired morphologic myelodysplastic features which may rather represent potentially reversible hematological responses to immune-mediated factors, nutritional deficiency states, and disordered myelopoietic responses to various pharmaceutical, herbal, or other potentially myelotoxic compounds. We emphasize the clinical settings, and the histopathologic features, of such AMD that should trigger a search for a reversible underlying condition that may be nonneoplastic and not MDS.
Despite advances in cytogenetic and flow cytometric analyses, aberrant cellular morphology, as identified in the peripheral blood and bone marrow, remains the defining feature leading to a clinical diagnosis of myelodysplastic syndrome (MDS). Certain laboratory values such as blood cell count and cell volume measurements are accurate and reproducible, and the results are not open to dispute, as is the presence of particular unique and obvious morphologic findings such as the presence of acquired Pelger-Huët granulocytes and tear-drop erythrocytes in the peripheral blood or large numbers of ringed sideroblasts or increased numbers of myeloblasts in the bone marrow. Other observations such as reduced mature myeloid cell cytoplasmic granulation and the presence of dimorphic erythrocyte or dysmorphic megakaryocytic populations are more subtle. However, what constitutes a significant variation from normal in each of the three major cell lines in the bone marrow remains very observer dependent. Unfortunately, we are only occasionally but usefully reminded that not all clear-cut examples of acquired and persistent myelodysplasia represent MDS or neoplasia [
The difficulty with morphology, alone, in establishing a diagnosis of MDS is evident in the evolution of the current World Health Organization [WHO] classification system for MDS with respect to the acquired refractory sideroblastic disorders. Germing and associates suggested that careful morphological review allowed some separation within the initial MDS classification system of those individuals with acquired idiopathic sideroblastic anemia (AISA) who were more likely to have an illness that would terminate in AML from those who might not have a neoplastic or preleukemic condition. They separated 232 individuals with MDS associated with ringed sideroblasts into two groups, one without significant myelodysplastic features among nonerythroid bone marrow cells and the other exhibiting such dyspoiesis among multiple cell lines. The 38% with selective erythroid aberrations and the 62% with a more multilineage dysplasia, respectively, exhibited different clinical courses, frequency of cytogenetic defects, and survival patterns [
Such an arbitrary distinction among those with a sideroblastic MDS was subsequently adopted in the WHO MDS classification as refractory anemia with ringed sideroblasts (RARS) and refractory cytopenia with multi-lineage dysplasia and ringed sideroblasts (RCMD-RS). However, a uniform concordance with this dual classification among experts in the field seemed hopelessly lacking. In Pavia, Italy, experienced hematopathologists classified only 28% of 60 such MDS cases with ringed sideroblasts as RCMD-RS while their colleagues in Dusseldorf, Germany, opined that 76% of their 119 patients with MDS and ringed sideroblasts fell into this category [
Current and suggested future MDS classifications seem to focus primarily on survival statistics or risk for evolution into AML to complement prognostic scoring systems [
Age-related incidence of MDS and frequency estimates of MDS subsets. (1)
The ability to “see” and report myelodysplasia where none is likely to exist was emphasized in a recent blinded study involving inspection of the bone marrow aspirate slides of 120 healthy prospective bone marrow donors with normal blood counts by four allegedly experienced morphologists [
When a clinical hematologist is confronted by a bone marrow study interpreted as MDS, typically, that diagnosis has been made with a little difficulty and is thought likely to be correct [
Clinical and laboratory features where AMD may not represent MDS. Young age (<40). Lack of erythrocyte macrocytosis. Lack of cytogenetic aberrations. Presence of ringed sideroblasts. Amegakaryocytic thrombocytopenia. Multiple vacuoles in erythroid and/or myeloid precursor cells. Absence of increased numbers of myeloid blast forms. Prior use of prolonged antibiotic therapy. A history of herbal and/or unregulated alternative medication use. Evidence of systemic or cutaneous autoimmune conditions preceding myelodysplasia. Human immunodeficiency virus (HIV) infection.
The incidence of MDS increases dramatically with advancing age, and MDS is uncommon among individuals younger than 50 years, where it accounts for only 6-7% of all MDS and much less when the diagnosis is restricted to a
Mean corpuscular red blood cell volume (MCV) measurements above 100 fl occur in less than 2% of the general population but are nearly a universal finding among patients with MDS, where the MCV is often above 106 fl and may be as high as 140 fl, particularly among those individuals with RARS. Thus, to be referred an individual with suspected MDS featuring a sideroblastic anemia associated with a normal MCV, and in the absence of chromosomal aberrations, and where vitamin B-12 and folic acid deficiency have been excluded, certain alternate maladies should immediately come to mind. For example, despite the impressive numbers of ringed sideroblasts, both lead and arsenic poisoning typically are associated with an MCV in the normal range [
As discussed in the previous section, reversible sideroblastic anemias may occur in a number of circumstances including prolonged use of certain antibiotics such as chloramphenicol, linezolid, pyrazinamide, isoniazid, rifampin, and some tetracyclines [
Certainly, examples of MDS without numerical or molecular cytogenetic aberrations occur but are uncommon, particularly in secondary or therapy-related MDS. As cytogenetic analysis has become more sophisticated and accompanied by specific and comprehensive fluorescent
Whether a bone marrow demonstrating 10% or more myeloblasts should still be considered a form of MDS, rather than AML, has always been a very controversial issue [
Many have commented on the association of autoimmune disease preceding or coincident to the development of hematological abnormalities consistent with MDS, and this association has also been well documented by epidemiological studies [
While thrombocytopenia in the setting of MDS is common and may be associated with reduced numbers and dysmyelopoietic-appearing megakaryocytes, acquired amegakaryocytic thrombocytopenia with a cellular bone marrow is unusual in MDS but has been described in association with lupus erythematosis and thymoma and as an immune consequence of several disorders [
Years ago, extensive vacuole formation, particularly in the cytoplasm and even the nucleus of bone marrow erythroid precursors, was a hallmark of chloramphenicol toxicity. Today, a similar process has been noted with toxicity to the antibiotics linozolid and certain tetracyclines. This finding has been also described with arsenic poisoning and copper deficiency where it may affect both myeloid and erythroid cells and is well known as a consequence of alcohol-induced hematological toxicity along with ringed sideroblasts [
A 29-year-old white woman who worked in the home as a computer specialist had been scheduled for her initial visit at a bone marrow transplant center in order to discuss, and then undergo, allogeneic stem cell transplantation for progressive cytopenias consequent to MDS. Features of two prior bone marrow studies allowed multiple hematopathologists and her hematologist to concur with the diagnosis of MDS. However, fearful of the complications of bone marrow transplantation, she had been referred to yet another hematologic consultation by her primary physicians not for a review of the diagnosis but simply to convince her of the need for urgent transplantation. Several months earlier she first began to experience easy bruising and was found to have a hemoglobin concentration of 11.0 gm/dL with a platelet count of 63,000/cu mm. She was thought to have idiopathic thrombocytopenic purpura [ITP], but, despite therapy with corticosteroids and Win-Rho SD immune globulin, she became progressively pancytopenic and had required blood transfusions. Numerous laboratory studies including antinuclear antibody, Coombs’ tests, serum vitamin B-12, folate, serum copper and blood lead levels, parvovirus, hepatitis A, B, and C panels, and HIV titres gave normal results as did CT scans of the chest and abdomen. Her CBC now showed a hemoglobin concentration of 6.6 gm/dL with an MCV of 107 fl. Her total leukocyte count was 2,100/cu mm with a neutropenia, and she had a platelet count of 8,000/cu mm.
Her peripheral blood film showed macrocytosis, tear-drop, and other abnormally shaped erythrocytes. The two bone marrow specimens were similar, with about 70% cellularity and with only rare megakaryocytes. Erythropoiesis was increased in activity and extremely dyspoietic. Iron stains showed numerous ringed sideroblasts. Granulocytopoiesis was reduced in activity with mild dysplastic changes. Blast cells were not increased. Routine cytogenetic studies had twice given normal (
However, a year later, asymptomatic and with a normal CBC, she again became pregnant. Her pancytopenia and macrocytosis soon reappeared within the first trimester, and the cytopenias were more profound than on her first presentation. The platelet count was now 2,000/cu mm with a total leukocyte count of 2,200/cu mm with a neutropenia, and she again required blood transfusions. She declined the suggestion by her high-risk obstetrician of a therapeutic termination of pregnancy. She received low-dose prednisone and cyclosporine and required frequent platelet and red blood cell transfusions throughout the pregnancy but uneventfully came to term with only continuous severe bruising. A healthy female infant with a normal blood count was delivered by C-section. She then received oral mycophenolate mofetil and prednisone and within 4 months achieved a normal blood count, and all medications were discontinued. She later underwent bilateral tubal ligation and remained well with a normal CBC, five years later.
Pregnancy-induced pancytopenia is a rare disorder that may occur with the first pregnancy where it may be mild and spontaneously remit to reappear with subsequent pregnancies, often with increasing severity. In this regard, it is similar to the rare syndrome of circulating inhibitor of factor VIII induced by pregnancy, and the infant is not affected by the process. While the peripheral blood and bone marrow, at first glance, seem typical of MDS, the constellation of teardrop erythrocytes in the peripheral blood, increased bone marrow cellularity with erythroid hyperplasia, amegakaryocytic thrombocytopenia, and ringed sideroblasts in a young woman in the proper clinical setting should suggest the correct diagnosis and not MDS.
In a recent publication calling for all MDS to be renamed as the myelodysplastic neoplasms (MDN), the authors’ comment was “
A 52-year-old white attorney with type II diabetes mellitus developed a macular pruritic rash on his upper arms typical of Sweet syndrome that, over several months, became generalized and excoriated. Various topical preparations, antihistamines, and pulse steroid doses were not helpful. He experienced a 20-pound weight loss and developed intermittent fever. He became anemic and thrombocytopenic. He underwent upper and lower endoscopies which gave normal results as did CT scans of the chest and abdomen. His hemoglobin concentration was 10 gm/dL with an MCV of 105 fl. The total leukocyte count was 7,000/cu mm with a monocytosis and the platelet count 102,000/cu mm. A bone marrow study revealed an 80% cellular marrow with an increase in megakaryocytes and trilineage dysplasia, particularly evident in erythroid elements. Iron stores were normal without ringed sideroblasts. Blast forms were not increased. His bone marrow was interpreted by the hematopathology department as consistent with MDS. Cytogenetic studies gave normal results.
Considering an immune-related AMD, he received a trial of oral prednisone and 6-mercaptopurine, and his hemoglobin improved to 14.0 gm/dL with a normal MCV and a platelet count of 232,000/cu mm. His skin rash faded but reappeared when the immunosuppressive drugs were tapered and discontinued. His platelet count then fell to 44,000/cu mm and his hemoglobin concentration to 8.6 gm/dL. Prednisone was resumed, and mycophenolate mofetil substituted for the 6-mercaptopurine. Now, 3 years later and receiving only 500 mg of mycophenolate mofetil daily, aside from his usual diabetic medications, his skin rash is well controlled, his CBC is normal with a normal MCV, and he is asymptomatic.
Sweet syndrome has been described in MDS, AML, and NHL but often abruptly affects younger individuals with no evident underlying neoplasm and typically responds well to immunosuppressive therapy [
A 56-year-old white man, who was employed as mechanic, and was a heavy smoker, developed severe, generalized, and disabling arthritis, thought by his rheumatologists to represent seronegative rheumatoid arthritis. His blood counts and chemistries initially gave normal results. Over a period of 2 years he received courses of at least five different nonsteroidal anti-inflammatory agents. When this therapy proved ineffective, he received oral methotrexate for several months, along with continued anti-inflammatory medications. He then underwent a bone marrow study because over this two-year period his hemoglobin concentration had declined from 14.3 gm/dL to 12.6 gm/dL and his total leukocyte count from 7,000/cu mm to 3,300/cu mm, with a neutropenia. His platelet count remained in the normal range. The striking morphological abnormality in the bone marrow was extensive vacuolization, particularly among the myeloid cell precursors. There was no increase in blast forms, and cytogenetic studies gave normal (
Despite our current knowledge and growing laboratory expertise, whether a prolonged hematopoietic myelodysplastic reaction associated with cytopenias but without a distinct pattern of clonal cytogenetic aberration or increase in blast forms is actually MDS, as we currently employ this term for risk assessment, may remain uncertain. Such an unusual circumstance is presented by the work of Irons and associates in China, involving individuals with alleged major exposures to pure benzene and who manifest unusual and complex bone marrow histology frequently associated with bone marrow hypoplasia, multilineage dysplasia, atypical eosinophilia, and phagocytic histiocytes, among other features [
In Irons and associates two initial reports, this distinctive hematological toxicity to benzene was described simply with morphologic descriptive terms such as dysplasia and dyserythropoiesis [
Other medical groups in China, where major and prolonged benzene exposures are still possible, have classified their alleged examples of benzene-induced hematotoxicity in a different manner than Irons and associates. In one report the authors separate 41 such patients as representing either aplastic anemia, pancytopenia (
A caveat is that cytogenetic studies in China for MDS/AML have shown a differing frequency of particular aberrations than among Western patients, with an increased frequency of trisomy 8 and a very reduced incidence of the 5q-syndrome as well as a marked increase in t (15; 17) acute promyelocytic leukemia [
Olnes and Sloan have demonstrated that, in selected patients with MDS, a significant number may respond favorably to immunosuppressive therapy and observed a complete remission in 18% of 31 evaluable patients, including one manifesting a small clone with trisomy 13 [
In MDS with hypocellular bone marrow, this circumstance may be difficult to clearly separate from aplastic anemia where favorable responses to drugs such as anti-thymocyte globulin [ATG] are often seen. Nevertheless, we believe that a young individual with MDS, regardless of bone marrow cellularity and normal cytogenetics, particularly when associated with morphologic features such as sideroblastic anemia or amegakaryocytic thrombocytopenia occurring in the clinical setting of autoimmune illness, are with high likelihood of achieving a favorable response to immunosuppressive medications. As emphasized, it seems critical to initiate therapy with immunosuppressive regimens early in the course of MDS in order to favorably impact the long-term result [
Bone marrow flow cytometry (FC) analysis may be useful in an attempt to separate MDS from other causes of persistent cytopenias, including forms of hypoplastic anemia. Its application can identify specific aberrations in both immature and maturing cell compartments among the hematopoietic cell lineages but, as yet, it is not reliable as a single parameter to segregate MDS as a specific diagnosis [
The WHO MDS classification seems to favor a parallel association with risk stratification [
Whether or not all MDS should be considered neoplasms and whether or not Irons and associates’ experience with persistent cytopenias among heavily benzene-exposed individuals should be classified as MDS and as neoplasms are just some of the unanswered questions in this field. There is considerable information that immune dysregulation plays an important role in the onset and progression of MDS both confounding a certain confirmation of the diagnosis and its uniform classification as a malignancy. The most straight-forward evidence of this influence is the favorable therapeutic response seen among certain patients with MDS, as defined by current diagnostic criteria, who are treated with medications with immunosuppressive activity. The frequency of long-term control of the illness with this form of therapy is uncertain, but as many as 30% of all MDS cases exhibit some hematologic improvement when treated with a variety of immunosuppressive agents [
In low-risk forms of MDS, based upon current stratification schemes, among those with bone marrow hypocellularity, increased apoptosis of bone marrow cells is an established pathogenic mechanism resulting in ineffective hematopoiesis. Several compelling lines of evidence suggest that disruption of the differentiation of hematopoietic progenitor cells and their increased apoptosis is immunologically mediated [
The increase in apoptosis observed in low-risk MDS is thought to be a key step in the pathology of this disorder. However, there may be additional modifying immunological events in play that collectively result in the observed morphological dyspoiesis. Dysplasia-associated antigens released by degenerating cells that are processed by antigen-presenting cells may initiate an adaptive immune response that may influence the course of the illness [
TNF-
This pattern of immune dysregulation that may be modified by immunosuppressive therapy is not characteristic of more advanced or high-risk cases of MDS. Here, there are increased numbers of regulatory cells, which dampen any type of favorable response to immunosuppressive therapy. This may, in turn, allow for a damaged or dysplastic clone to escape from immune surveillance and progress to a more aggressive expression of the disease [
Immunosuppressive therapy seems to be most effective when applied early in those with low-risk MDS. Many studies have used the combination of anti-thymocyte globulin (ATG) and cyclosporine A (CsA), because of their well-known benefit in forms of aplastic anemia and in the organ transplant field. However, corticosteroids, alemtuzumab, and newer immunosuppressive agents, such as mycofenolate mofetil and TNF-
While the fog obscuring the certain diagnosis of MDS is slowly lifting, clinicians must remain alert to the possibility of an alternate disorder in certain cases of AMD. Our therapeutic armamentarium continues to improve along with efforts to define in which examples of MDS and in which clinical settings they are most likely to initiate a favorable response. While it is clear that immunosuppressive therapy may be beneficial in a subset of individuals with MDS, the most effective types of immunosuppressive regimens and the duration of such therapy remain to be determined. Moreover, such favorable response does not necessarily imply a causative autoimmune illness or exclude a neoplasm.