Guidelines for the diagnosis and therapy of adult myelodysplastic syndromes

The myelodysplastic syndromes (MDS) represent a heterogeneous group of haematopoietic disorders affecting predominantly elderly individuals (median age 69 years). The overall disease incidence is about 4 per 100 000 population but this rises to > 30 per 100 000 in the over 70 year age group. The pathological processes underlying the haematological abnormalities seen in MDS are: augmented apoptosis, leading to ineffective haematopoiesis and peripheral cytopenias; transformation to acute myeloid leukaemia (AML). The precise relationship between these pathological processes remains uncertain but has important implications for the design of new therapeutic strategies targeting one or the other or indeed both. The guideline group was selected to include UK-based medical experts in the clinical management of MDS and to include a representative from a District General Hospital. The drafting group met on six occasions. Each group member was allocated responsibility for preparation of a selected component of the first draft. Medline/Pubmed was systematically searched from 1982. The Cochrane database was searched but contained no references to MDS. Meeting abstracts were not included in the systematic search strategy. The Chairman synthesized the draft components, which were revised by consensus through meetings 3–6. No recommendations are included for which full consensus was not achieved. The draft guideline was reviewed by the Sounding Board and by the Committee of the British Committee for Standards in Haematology, and comments incorporated where appropriate. Following further helpful peer-review by the British Journal of Haematology, a final revision is now presented. Criteria used to quote Levels and Grades of Evidence are outlined in Table I. A full guideline revision is planned for May 2005. The diagnosis and classification of MDS remain dependent on the morphological examination of blood and bone marrow cells. Diagnostic criteria should ideally distinguish MDS from reactive conditions causing dysplastic haematopoiesis and from other clonal myeloid disorders. The minimum clinical assessment and laboratory investigation required for the definitive diagnosis of cases of suspected MDS is shown in Table II. Practical issues regarding the diagnosis of MDS include the following: Should every patient have a bone marrow aspirate? A bone marrow aspirate is usually necessary to make a confident diagnosis and to provide important prognostic information. This may not, however, be necessary in elderly patients in whom a definitive diagnosis of MDS would not alter management or whose poor general health precludes active treatment. Should every patient have a bone marrow trephine? Bone marrow histology complements the morphological information obtained from a marrow aspirate and hence a trephine biopsy should be performed in all cases of suspected MDS in whom bone marrow examination is indicated. Diagnostic certainty can be improved by features such as the presence of megakaryocyte dysplasia, while disordered marrow architecture such as central clustering of immature myeloid cells (abnormal localization of immature precursors or ALIPs) is an adverse prognostic marker (Tricot et al, 1984). The assessment of cellularity and fibrosis define morphological variants, with the identification of hypocellular MDS of particular therapeutic importance (see below) (Mijovic & Mufti, 1998). Should all patients have cytogenetic analysis? A chromosome abnormality confirms the presence of a clonal disorder aiding the distinction between MDS and reactive causes of dysplasia, and in addition has major prognostic value. Cytogenetic analysis should therefore be performed for all patients in whom a bone marrow examination is indicated. What are the minimal morphological diagnostic criteria for MDS? Minimal diagnostic criteria are not clearly defined in MDS. Difficulties arise because a variety of reactive disorders are associated with dysplastic morphology and mild dysplastic features are frequently seen in the marrow of healthy people with normal blood counts (Bain, 1996; Ramos et al, 1999). The following are recommended to increase the reliability of diagnosing MDS. a) At least 200 marrow cells and 20 megakaryocytes should be evaluated where possible. b) Dysplastic features should be present in > 10% of marrow cells (Kouides & Bennett, 1996). c) Particular attention should be given to the presence of pseudo-pelger neutrophils, ring sideroblasts, micromegakaryocytes and increased blasts as these abnormalities correlate most strongly with the presence of clonal markers in MDS and show least interobserver variation (Kuriyama et al, 1986; Ramos et al, 1999). d) The assessment of neutrophil granularity is critically dependent on optimal staining and it is unwise to base a diagnosis of MDS solely on the presence of neutrophil hypogranularity in the absence of other dysplastic features. It is recognized that the definitive diagnosis of early refractory anaemia may be difficult, for example in patients with a single isolated cytopenia or isolated macrocytosis. If the morphological diagnosis remains uncertain, it is recommended that the patient be reviewed regularly in the haematology clinic with repeat blood count and morphological assessment at appropriate intervals. Overlap syndromes. At least three overlap syndromes are recognized namely: fibrotic MDS (Mijovic & Mufti, 1998), MDS with thrombocytosis (Harris et al, 1999) and hypoplastic MDS (Tuzuner et al, 1994, 1995). Presently, only the latter requires a distinctive therapeutic approach and as such is important to distinguish from aplastic anaemia. The diagnosis of hypocellular MDS is of importance, as preliminary data suggest that the response to immunosuppressive therapy is higher than in cases of MDS with normo- or hypercellular marrows. A trephine may be considered hypocellular if cellularity is < 30% in individuals younger than 60 years or < 20% in those over 60 years of age (Tuzuner et al, 1994, 1995). The diagnosis of myelodysplasia requires the presence of dysplastic features in megakaryocytes and/or myeloid cells or an excess of blasts. Erythroid dysplasia is found in aplastic anaemia and cannot be used alone to distinguish MDS from AA. The presence of an abnormal karyotype strongly favours the diagnosis of MDS but cases of aplastic anaemia with an abnormal karyotype without morphological features of MDS and with a low risk of transformation to MDS or AML have been described. The World Health Organization (WHO) recently published proposals for a new classification of MDS to supersede the French–American–British (FAB) classification (Bennett et al, 1982). The new classification is based on a combination of morphology, karyotype and clinical features (Harris et al, 1999), and potential problems with the adoption of the new classification have recently been summarized (Greenberg et al, 2000). The final classification has now been published (Jaffe et al, 2001) and is in the process of independent validation (Germing et al, 2000a; Nosslinger et al, 2001). Given the limited experience with the use of this new classification to date, this guideline will use the more familiar FAB terminology (Bennett et al, 1982). All patients diagnosed with MDS have a reduced life expectancy compared with age- and sex-matched normal controls, regardless of disease subtype. This difference is particularly marked in younger patients (≤60 years) and those with 'high-risk' disease. (Morel et al, 1996). While the FAB/WHO classification systems are themselves of significant prognostic value, scoring systems employing objective parameters allow improved reproducibility for clinical decision-making but also for clinical trials and biological research in MDS patients. A series of different prognostic scoring systems have been developed since the original Bournemouth score (Mufti et al, 1985). The latest of these is the International Prognostic Scoring System (IPSS) (Greenberg et al, 1997), which has improved prognostic power compared with previous scoring systems, namely the Sanz (Sanz et al, 1989) and Lille (Morel et al, 1993) scores. The IPSS is a multivariate analysis of patient characteristics derived from a pool of 816 MDS patients who had been included in the cohorts used to derive previous scoring systems (Table III). A cautionary note is that the patient cohort analysed by the IPSS group were largely untreated and may therefore include the MDS patients with poorest performance status and therefore worst outlook. The IPSS has also excluded chronic myelomonocytic leukaemia (CMML) patients with a white blood cell count (WBC) > 12 × 109/l, considering this group as myeloproliferative rather than myelodysplastic. The IPSS score is computed from three parameters, namely bone marrow blast percentage, bone marrow cytogenetics and number of lineages with cytopenia, in decreasing order of prognostic power. While this guideline advocates bone marrow cytogenetic analysis for all MDS patients in whom a bone marrow examination is deemed appropriate, we accept that this is not always technically successful. In this rare situation, an alternative scoring system most familiar to the clinician will be required, such as the Sanz (Sanz et al, 1989) or Bournemouth (Mufti et al, 1985) scores. The methodology used to derive the Sanz score is more rigorous and provides greater prognostic discrimination than the Bournemouth score, at least in patients with > 5% blasts. While the use of prognostic scoring systems is advocated, it must be recognized that even within each prognostic group the emergence of distinct clinical entities, which have different prognoses, is inevitable. Examples of these include pure sideroblastic anaemia (Germing et al, 2000b) and perhaps the '5q– syndrome', both of which are associated with an excellent prognosis and a low rate of transformation to acute leukaemia. It is recommended that, where possible, management decisions be based upon the patient's IPSS score. It is important that the IPSS score is calculated during a stable clinical state and not, for example, during a florid infective initial presentation. Management decisions should be taken with the informed involvement of the patient and, to aid this, information booklets are available from the Leukaemia Research Fund and the Myelodysplastic Syndromes Foundation. The guideline will critically review individual therapeutic modalities designed to improve the clinical problems specific to an individual patient with MDS, and will conclude with recommendations for management strategies driven by the patient's IPSS score and the overall clinical picture. It is, however, important to stress that most recommendations are made on the basis of a very limited evidence base for the efficacy of interventional and non-interventional therapy in MDS. Most papers reporting interventional therapy describe relatively small cohorts of patients and only one (small) placebo-controlled trial is available. Criteria for defining therapeutic response are also highly variable between studies. The recent publication of standardized response criteria for therapeutic studies in MDS patients (Cheson et al, 2000) may facilitate the interpretation of how clinically meaningful new therapeutic interventions in MDS really are. Supportive care remains the most important aspect of management for patients with good prognosis MDS and those with poor prognosis disease whose age or performance status precludes them from receiving more intensive forms of therapy. The aim is to reduce morbidity and mortality while at the same time providing an acceptable quality of life. However, grade A and grade B evidence for the effectiveness of supportive care in patients with myelodysplasia is absent and is unlikely to ever be obtainable. For patients with good prognosis MDS, it is often feasible to undertake a period of observation without needing to introduce specific therapy. Where possible, this 'wait and watch' approach to management may also be useful for patients with more advanced MDS, allowing one to appraise the stability of the disease process and to assess the need to introduce treatment. At presentation, up to 80% of cases of MDS will have a haemoglobin concentration < 10 g/dl (Sanz et al, 1989). Anaemia in MDS is usually due to ineffective erythropoiesis but other factors that may accentuate anaemia, e.g. nutritional deficiencies, haemorrhage, haemolysis and infection, should be sought and treated as appropriate. Chronic anaemia is seldom life threatening but can lead to significant morbidity and is therefore important in relation to quality of life issues. Figure 1 outlines a flow chart for the management of the anaemic MDS patient. Guidelines for the management of symptomatic anaemia in MDS patients. Red cell transfusion and iron chelation therapy. The use of red cell transfusion should be considered in any patient with symptoms of anaemia. It is not possible to ascribe a single haemoglobin concentration as being the optimal level below which red cell support should be dispensed and each individual case needs to be considered separately (Murphy et al, 2001). Studies in cancer patients have demonstrated a positive correlation between increases in the haemoglobin level (with recombinant Erythropoietin therapy) and improvements in quality of life (QOL) (Glaspy et al, 1997; Demetri et al, 1998). Recommendations for iron chelation treatment in myelodysplasia are based on limited data (evidence grade B, level III). Iron chelation should be considered once a patient has received 5 g iron (approximately 25 units of red cells) but only in patients for whom long-term transfusion therapy is likely, such as those with pure sideroblastic anaemia or the '5q– syndrome'. Desferrioxamine 20–40 mg/kg should be administered by 12 h subcutaneous infusion 5–7 d per week. Audiometry and ophthalmology review are essential prior to commencement of desferrioxamine. The target ferritin concentration should be < 1000 µg/l; if the ferritin concentration falls below < 2000 µg/l, the dose of desferrioxamine should be reduced and should not exceed 25 mg/kg. Vitamin C 100–200 mg daily should be commenced after 1 month of desferrioxamine therapy. Vitamin C should be taken when the infusion is set up. Repeat audiometry and ophthalmology review should be performed at least annually. The use of twice daily subcutaneous bolus injections of desferrioxamine (Franchini et al, 2000) may be considered where infusions are not tolerated, but the common practice of adding a single dose of desferrioxamine at each transfusion episode has no basis and should be discouraged. At present, deferiprone (L1) cannot be recommended for routine use in this group of patients, given the lack of published data in MDS patients and continuing concerns about both efficacy and safety (Pippard & Weatherall, 2000). Erythropoietin (EPO) +/– granulocyte colony-stimulating factor (G-CSF). Many studies have clearly demonstrated that EPO ± G-CSF can increase haemoglobin concentration and reduce/eliminate red cell transfusion in selected MDS patients, and a summary outline of these is provided in Tables IVA and B. These studies were small cohort studies (< 120 patients in each) and only one (small) placebo-controlled randomized study (of EPO therapy alone) has been reported (Italian Cooperative Study Group, 1998). Given the small size of this placebo-controlled trial of EPO therapy alone (87 patients), the grade of recommendation for EPO therapy alone should be considered as grade A/B (level Ib/IIa). The evidence for efficacy of the combination of EPO + G-CSF therapy is grade B (level IIa/IIb). These studies suggest that patients with refractory anaemia with ringed sideroblasts (RARS) are more likely to respond to the combination of EPO + G-CSF (Hellstrom Lindberg et al, 1997, 1998). Patients with refractory anaemia (RA) or RA with excess blasts (RAEB) may respond well to EPO alone, though a proportion will benefit from the addition of G-CSF (Hellstrom Lindberg, 1995, Hellstrom-Lindberg et al, 1998; Italian Cooperative Study Group, 1998; Mantovani et al, 2000). Overall there is sufficient evidence for the efficacy of EPO ± G-CSF therapy in appropriately selected patients. It is recommended that those patients with RA and RAEB [not eligible for chemotherapy/stem cell transplantation (SCT): see later] who are symptomatic of anaemia, with no/low transfusion requirement (< 2 units/month) and a basal EPO level of less than 200 U/l (measured at the haemoglobin nadir in transfusion-dependent patients) be considered for a trial of EPO alone at a dose of 10 000 units daily for 6 weeks (Hellstrom Lindberg, 1995). For non-responders, consideration should be given to either the addition of daily G-CSF, doubling the dose of EPO, or both for a further 6 weeks. The dose of G-CSF should be escalated weekly from 75 µg, to 150 µg to 300 µg (multiple sampling from single vials kept at 4°C) to maintain the WBC between 6 and 10 × 109/l. In responding patients, once the maximum response has been reached, the G-CSF can be reduced to thrice weekly and the EPO to 5 d then 4 d to 3 d per week at 4 weekly intervals to the lowest dose that retains response. For patients with RARS, symptomatic anaemia, basal EPO levels of < 500 U/l and a transfusion requirement of less than 2 units per month, it is recommended that combined therapy with EPO and G-CSF is used from the outset (Hellstrom Lindberg et al, 1997). Dosing recommendations are as for RA/RAEB with consideration given to dose escalation of EPO at 6 weeks in non-responders for a further 6 weeks. While there are no published data demonstrating that growth factor therapy improves QOL, extrapolation from cancer patients treated with EPO suggests that this is likely to be the case and that maintenance of a stable augmented haemoglobin concentration may be preferable to the cyclical fluctuations of red cell transfusion programmes (Bowen & Hellstrom-Lindberg, 2001). It should also be noted that there are a lack of survival data and of quality pharmaco-economic data to support the use of haematopoietic growth factor therapy. We would, therefore, encourage continuing randomized-controlled trials of EPO ± G-CSF, which address the issues of QOL, survival advantage and pharmaco-economics. Immunosuppression. Two groups have demonstrated the efficacy of horse anti-lymphocyte globulin (ALG) in raising the blood counts in a proportion of patients with MDS, although each cohort reported is small, namely n = 25 (Molldrem et al, 1997) [updated in abstract form to n = 60 (Barrett et al, 1998)] and n = 12 (Killick et al, 1999) patients respectively. ALG seems to be more effective in hypoplastic MDS and in patients with a paroxysmal nocturnal haemoglobinuria (PNH) clone (Dunn et al, 1999), but also induces significant improvements in normo- and hypercellular patients with low-risk MDS (IPSS≤INT-1). Preliminary data suggest that responses, when achieved, are durable and prolonged (Barrett et al, 1998). Most clinically meaningful responses are in the erythroid lineage but bi- and trilineage responses are seen. Similar responses have been observed with cyclosporin A (Jonásova et al, 1998), with higher response rates also in hypocellular MDS. Anecdotal reports of corticosteroid-responsive MDS exist but this therapy cannot at present be recommended. The data support a recommendation for a trial of immunosuppressive therapy with ALG at least for patients with hypoplastic MDS (evidence grade B, level IIb). Other agents. Therapeutic agents with limited therapeutic value that are not routinely recommended are outlined in Table V. Bleeding is a common and potentially serious complication of MDS. Both the degrees of thrombocytopenia and platelet functional abnormalities may contribute to this. The role of platelet transfusions in MDS patients should be based on the Royal College of Physicians of Edinburgh Consensus Conference Statement (Ancliff & Machin, 1998). Antifibrinolytic agents (grade C, level IV) and Danazol (grade B, level IIb) are occasionally useful but cannot be routinely recommended. Prophylactic. There are no published data to support the routine use of antibacterial or antifungal prophylaxis in neutropenic MDS patients. Consideration may be given to the use of prophylactic low-dose G-CSF therapy in severely neutropenic patients to maintain the neutrophil count > 1 × 109/l (grade B, level IIb) (Negrin et al, 1996). Therapeutic. in MDS patients should be treated with as for other patients with There is evidence to routine use of low-dose Many relatively small studies are reported but have demonstrated benefit in of survival or improved quality of life (Table these the perhaps the often has a myeloproliferative and is frequently indicated. is considered the treatment for in to (evidence grade level et al, 1996). For the small number of eligible patients, cell transplantation in long-term survival in It is recommended that all patients eligible for cell transplantation with associated with improved following transplantation include younger disease MDS, < 10% cytogenetics et al, 1996; et al, 1996; et al, 2000). IPSS score is also a of with of and in and & 1998). mortality for is in most although the will reduce this. are that may also survival though cell is in cases et al, 1997). A small group of patients who may benefit from intensive alone can also be IPSS intensive cell transplantation can be recommended for this group whose survival without treatment is 60 years (< 60 years) (Greenberg et al, 1997). IPSS All patients < years should be for for as as possible after as is improved if performed early et al, 1996). If eligible and a is it is recommended that patients < years are (evidence grade B, level IIb) and patients > < years are considered for within clinical trials where available (evidence grade C, level Patients with no but with an should also be considered for (< evidence grade B, level or within clinical trials evidence grade C, level though the from these remains prior to is not recommended for this group (evidence grade B, level IIb). Patients > years or < years and not for should be supportive care and/or considered for growth factor therapy Recommendations for the management of IPSS MDS patients years are outlined in Guidelines for the management of IPSS MDS patients All patients < years should be considered as to for cell transplantation early after In this group of patients, cell transplantation should only be considered for those responding to as the for patients is very poor (evidence grade B, level IIb) et al, 1996; et al, 1997; et al, 1998). For patients < eligible for and responding to recommendations for are outlined in cohort studies are available for the assessment of and but the role of and is to be clearly is for both modalities and preliminary evidence suggests that both will have a role et al, 1997; et al, 1998). Guidelines for in the management of IPSS MDS patients Both patients > years and those < years who are for cell transplantation should be considered for intensive There have been no trials following intensive compared with supportive care alone in MDS. studies suggest that of all MDS patients those with RAEB in transformation marrow and an independent adverse risk factor performance of haematological disorder et al, respond to intensive (evidence grade B, level IIb) et al, 1997). intensive alone is recommended for consideration in these patients. No combination is clearly but most used with any of an and/or The number of in most studies is and one and patients more than this. In all other MDS patients those for whom intensive alone is not intensive should be only if cell transplantation is as Supportive therapy. If patients not any for which ± is recommended should be supportive care or within clinical research Leukaemia Research Fund Meeting were by support from an of during the guideline preparation and are of the for the treatment of anaemia. the and information contained in this guideline are to be and at the time of to the the can accept any responsibility or for any or that may be A randomized study of supportive care has demonstrated a survival advantage and reduced AML transformation for MDS patients randomized to et al, Journal of