Myelodysplastic syndromes

• Myelodysplastic syndromes (MDS) are clonal diseases of haematopoietic stem cells: ineffective blood-cell production, refractory cytopenias and morphological dysplasia, with a risk of progression to acute myeloid leukaemia (AML).
• The stem cell still produces cells, but they are dysplastic and often die in the marrow, so the peripheral blood contains too few red cells, neutrophils and platelets despite a cellular marrow.
• Classification has two axes: how the disease looks under the microscope and genetics (especially WHO 2022 and ICC 2022, which differ in how they handle patients with 10 to 19 percent blasts). Risk is then scored, most often with the Revised International Prognostic Scoring System (IPSS-R) or, when molecular data exist, the Molecular IPSS (IPSS-M).
• For every patient the decision is the same two-part question: is this lower-risk or higher-risk disease, and is the patient fit enough, and anatomically suited, for allogeneic stem-cell transplantation.
• Management therefore splits: lower-risk disease is supported and stimulated (transfusions, erythropoiesis-stimulating agents, lenalidomide in del(5q), luspatercept in ring sideroblasts, iron chelation, occasional immunosuppression); higher-risk disease is treated with a hypomethylating agent and, when possible, sent to allogeneic transplantation.

By the end of this page you should be able to:

1. Define MDS and distinguish it from aplastic anaemia, acute leukaemia, clonal cytopenia of undetermined significance and the nutritional causes of cytopenia.
2. Classify MDS using the WHO 5th edition 2022 categories and explain the main points at which the International Consensus Classification (ICC) 2022 differs.
3. Interpret IPSS-R and IPSS-M and explain what changes when molecular genetics are added.
4. Explain why cytopenias occur (ineffective haematopoiesis and dysplasia) and what progression to AML looks like.
5. Plan lower-risk management, including when to use erythropoiesis-stimulating agents, lenalidomide, luspatercept, iron chelation and thrombopoietin analogues.
6. Plan higher-risk management, including hypomethylating agents, timing of allogeneic stem-cell transplantation and what to do when agents fail.
7. Adapt MDS care when molecular testing or transplant facilities are not available.

Myelodysplastic syndromes are, despite the gentle-sounding name, malignant diseases of the haematopoietic stem cell. The clone still differentiates, but badly: erythroid, granulocytic and megakaryocytic precursors show dysplasia, mature incompletely and die in the marrow rather than emerging into the blood. The marrow is usually normocellular or hypercellular, so the problem is not an empty marrow but a disordered one. The result is the cytopenia triad of anaemia (often the dominant symptom), neutropenia with its infection risk and thrombocytopenia with its bleeding risk. ^[1][2]

This single mechanism explains the two faces of MDS. In its indolent face the clone expands slowly and the patient is cytopenic for months or years, needing transfusions and infection prophylaxis. In its aggressive face the clone accumulates further mutations (especially RUNX1, TP53, ASXL1 and others), blasts rise, and the disease behaves like acute leukaemia. About one quarter of patients go on to develop acute leukaemia in the first sense, and one review notes an approximately 30 percent risk of developing AML within two years in many cases. ^[1][2]

Age is the major risk factor for de novo disease: the median age at diagnosis is over 70 years, and over 70 years in most series, with incidence rising steeply with age. Incidence is around 4 cases per 100 000 persons per year and increases significantly with age. Roughly 10 percent of MDS is therapy-related (after chemotherapy, radiotherapy or radioiodine), while the majority is primary. ^[1][2]

For a first-time reader the sensible order is: confirm the cytopenias and dysplasia, classify the subtype, score the risk, then choose between a supportive lower-risk pathway and a disease-modifying higher-risk pathway. Every step below follows that sequence.

Persistent is the operative word: MDS is not a week-long marrow insult. The cytopenia must have lasted long enough, and no other cause found, before the label is applied. ^[3]

Most MDS starts from clonal haematopoiesis, in which a founding mutation (commonly TET2, DNMT3A or ASXL1) gives a stem cell a growth advantage. In the first phase, called clonal haematopoiesis of indeterminate potential (CHIP), mutations are present at a variant allele frequency of 2 percent or more without cytopenia. When cytopenias appear without morphological dysplasia, the condition is called clonal cytopenia of undetermined significance (CCUS). Further mutations, chromosomal losses and epigenetic change turn CCUS into MDS, and further mutations turn MDS into AML. Older MDS classifications used a category called MDS unclassifiable; the WHO 2022 classification removes this and places some of those patients in CCUS instead. ^[2][3]

CHIP is common in health: prevalence is estimated at about 10 percent in people aged 80 years and older, and affected people progress to MDS at a rate of about 0.5 to 1 percent per year. Germline predisposition exists (for example DDX41 mutations and other inherited bone-marrow-failure syndromes) and is more common in MDS than was previously recognised, particularly in younger adults. ^[1][2]

The WHO 2022 scheme splits MDS into two broad families: subtypes defined by genetic abnormalities and subtypes defined by morphology. ^[1][2][3]

Important consequences of the 2022 scheme: MDS unclassifiable has gone; SF3B1-driven disease has replaced the old "MDS with ring sideroblasts" entity, which means more patients may be eligible for targeted therapy with luspatercept; and biallelic TP53 inactivation is separated out as a particularly aggressive category with median survival of about 10 months in one comparative series. Early work comparing the WHO 2016 and 2022 systems reported median overall survival of 45 months under the new categories generally, about 10 months for MDS-biTP53, 15 months for MDS with fibrosis, 24 months for MDS-IB1 and 26 months for MDS-IB2. ^{[3] [2]}

Dysplasia itself is defined quantitatively. MDS requires dysplasia in 10 percent or more of cells in at least one myeloid lineage. ^[2] In practice pathologists report the percentage of cells showing dysmegakaryopoiesis (micromegakaryocytes, hypolobated forms), dysgranulopoiesis (hypogranulation, pseudo-Pelger-Huet neutrophils) and dyserythropoiesis (ring sideroblasts, nuclear budding) as well as the blast percentage, which is the single most powerful predictor of myeloid neoplasms. One review summarises the same definition: persistent dysplasia in at least 10 percent of cells in one or more lineages assessed on marrow cytology. ^[1]

The International Consensus Classification (ICC) 2022 was proposed by a separate clinical advisory committee that includes many former WHO authors but is no longer affiliated with the WHO. The two systems share most of the patient, but part company on three points that affect patient care and access to treatments. ^[2]

• Blasts between 10 and 19 percent. WHO 2022 keeps "MDS with increased blasts-2" for this group. The ICC renames the same group "MDS/AML" so that trial and therapy pathways (MDS versus AML) can be opened to them. ^[2]
• SF3B1 mutation. WHO 2022 defines MDS-SF3B1 by the mutation itself; the ICC asks for a variant allele frequency of 10 percent or more in the absence of multihit TP53 or RUNX1. ^{[2] [1]}
• TP53. The ICC defines a separate myeloid neoplasm with mutated TP53 that spans both MDS and MDS/AML (two distinct TP53 mutations each at 10 percent variant allele frequency or more, or one mutation plus 17p deletion, variant allele frequency above 50 percent, or copy-neutral loss of heterozygosity at 17p). WHO 2022 recognises biallelic TP53 inactivation within MDS. ^[2]

In everyday clinical practice the practical message is that the names differ at 10 to 19 percent blasts, but both systems agree that these patients behave like very-high-risk disease. ^[2]

• CHIP: mutations at 2 percent variant allele frequency or more, no cytopenia. ^[2]
• CCUS: CHIP plus unexplained cytopenia, no dysplasia. ^[2]
• MDS: cytopenia plus dysplasia or a defining abnormality. ^[2]
• AML: 20 percent blasts or more in most classifications, or certain defining genetic abnormalities (for example NPM1 mutation) regardless of blast count in WHO 2022. ^[3]

This boundary is where students stumble: AML is not defined only by blasts. In WHO 2022 an NPM1 mutation or KMT2A, MECOM or NUP98 rearrangement upgrades even a low-blast marrow to AML. ^[3]

Incidence is approximately 4 cases per 100 000 persons per year, and rises steeply with age. The median age at onset is 70 to 75 years. Because MDS is largely a disease of older people, the "cancer epidemic" problem is a population-ageing problem; true age-standardised incidence changes are modest. ^{[1] [5]}

• About 30 percent of patients progress to AML. Two reviews phrase this differently but are consistent: in one, one quarter of patients go on to develop acute leukaemia; in the German review, genetic alteration causes progressive loss of differentiation with an approximately 30 percent risk of AML within two years. ^[1]
• Roughly 10 percent of MDS is therapy-related (after prior chemotherapy, radiotherapy, radioiodine or benzene exposure). ^[1]
• Occupational risk: long-standing benzene exposure (painters, petrol-pump attendants) counts as an occupational disease in some jurisdictions. ^[1]
• Clonal evolution: re-acquisition or gain of cytogenetic abnormalities during follow-up worsens prognosis. ^[1]

A population caveat that matters for non-epidemiological reasons: MDS is markedly under-diagnosed in low-resource settings because a full blood count and film, let alone marrow cytogenetics, may be missing. The biology does not differ dramatically by region; access does. ^[1]

The mechanism of primary MDS is accumulated DNA damage and mutation in long-lived haematopoietic stem cells, amplified by age. Factors repeatedly identified include:

• Age, which is the dominant risk factor. ^[1][2]
• Prior therapy: alkylating agents, topoisomerase II inhibitors and radiotherapy produce therapy-related MDS, often via TP53 or chromosome 7 lesions. ^[2]
• CHIP/CCUS, which progresses at roughly 0.5 to 1 percent per year, more often in the presence of SRSF2, TP53 or IDH mutations. ^[1][2]
• Benzene and solvents, usually occupational. ^[1]
• Germline predisposition: DDX41, RUNX1 and related inherited bone-marrow-failure syndromes in younger adults or familial cases. ^[1]
• Paroxysmal nocturnal haemoglobinuria (PNH) clones and immune-mediated marrow damage, which overlap with hypoplastic MDS and aplastic anaemia, and where anti-thymocyte globulin can be useful. ^[6]

Smoking and alcohol are inconsistently linked and no strong modifiable risk factor other than avoiding benzene is well established. ^[1]

The entire disease can be read as a sequence, and this sequence is also the justification for the treatment algorithm in the management section. ^[2][1]

1. Founding lesion. A mutation in TET2, DNMT3A, ASXL1 or another "CHIP gene" confers a fitness advantage and the clone expands. ^[2]
2. Second hits and epigenetic disruption. Splicing-factor mutations (SF3B1, SRSF2, U2AF1, ZRSR2) are reported in roughly 40 to 60 percent of MDS, with series-dependent estimates from about 40 to 50 percent in one Hoff 2023 table and about 60 percent in the same review's text; DNA-methylation mutations (TET2, DNMT3A, IDH1/2) in 30 to 40 percent, and chromatin-modification (ASXL1, EZH2) mutations in around 20 percent, alter RNA processing and transcription. Published mutation-band estimates differ between series (Hoff 2023 lists SRSF2 at about 15 percent and TET2 at 20 to 30 percent, whereas the Nachtkamp 2023 table lists SRSF2 at 5 to 10 percent and TET2 at 15 to 25 percent); DNMT3A is reported in about 5 to 10 percent in two agreeing sources, although some tables cite about 15 percent. ^{[2] [1]}
3. Differentiation block. Precursors fail to mature; megakaryocytes, granulocytes and erythroid precursors show dysplasia. ^[1][2]
4. Ineffective haematopoiesis. Dysplastic cells die in the marrow. The marrow is usually normo- or hypercellular despite the low blood counts. ^[1][2]
5. Cytopenias in blood. Erythroid failure causes anaemia; granulocytic failure causes neutropenia and infection risk; megakaryocytic failure causes thrombocytopenia and bleeding. ^[1]
6. Accumulated DNA damage and clonal evolution. Further mutations (RUNX1, TP53, NRAS/KRAS) mark progression; about a quarter to a third of patients progress to AML. ^[1][2]

The point that dysplasia and inefficiency, rather than simple absence of marrow cells, define MDS is what separates it from aplastic anaemia, and it is why marrow cellularity is an important distinguishing test. ^[6][1]

• Peripheral blood may show macrocytosis, pseudo-Pelger-Huet neutrophils, hypogranulation, thrombocytopenia with giant or hypogranular platelets and, in advanced disease, circulating blasts. ^[1]
• Bone marrow cytology shows dysplasia in 10 percent or more of cells in one or more lineages, ring sideroblasts (15 percent or more in the classic subgroup, or substitution by SF3B1 mutation), and a blast count that defines the subtype. ^[1][2]
• Bone marrow histology (trephine) shows age-adjusted cellularity, fibrosis, histiocytes and iron stores; in hypoplastic MDS cellularity under 25 percent (age-adjusted) is by itself a subtype. ^[1][3]
• Cytogenetics are abnormal in about half of de novo patients, and complex karyotypes (three or more abnormalities) are associated with very high risk. ^[2]

Ring sideroblasts are erythroid precursors in which at least five iron-laden (Prussian-blue-positive) granules form a ring around at least a third of the nucleus. The 15 percent threshold is a long-established morphological proxy that the SF3B1 mutation now explains and, in many patients, substitutes for. ^[2][3]

MDS is usually diagnosed during investigation of an incidental abnormal blood count. When patients are symptomatic, the complaints follow the cytopenias. ^[1]

• Anaemia: fatigue, breathlessness on exertion, palpitations, cold intolerance. ^[1]
• Infection: fever, recurrent or unusually severe infections, failure to clear an infection. ^[1]
• Bleeding: bruising, epistaxis, petechiae, gastrointestinal bleeding, menorrhagia. ^[1]
• Constitutional: weight loss, night sweats, bone pain (uncommon but reported). ^[1]
• Transfusion history: number of red-cell units and any iron accumulation; previous ESA trials. ^[1][5]

• Pallor; signs of active infection or bleeding.
• Signs of iron overload in heavily transfused patients (skin pigmentation, hepatomegaly in late cases).
• Splenomegaly is uncommon in uncomplicated MDS and should make you think of a myeloproliferative overlap disorder rather than pure MDS. ^[6]
• Minimal findings in many patients. An unexpectedly normal examination should not reassure you; the blood counts and marrow tell the story.

A 72-year-old retired electrician complains of six weeks of breathlessness on the stairs. His full blood count shows anaemia with neutropenia and thrombocytopenia; the blood film shows macrocytosis and hypogranular neutrophils with pseudo-Pelger-Huet forms. Iron, folate and vitamin B12 levels are normal. Bone marrow is cellular and shows dysplasia in two lineages, with a low blast count; cytogenetics show a deletion of 5q. This is lower-risk MDS with isolated del(5q), the type that responds best to lenalidomide. ^{[1][2][6] [5]}

MDS must be separated from the many other causes of cytopenia. The marrow biopsy with blast percentage, cytogenetics and molecular genetics usually settles it. ^[1][6]

A dangerous exam trap: a high total white count with neutropenia is common in MDS with circulating blasts and dysplastic cells. Look at the differential count and the film, and never assume a high total white excludes marrow failure. ^[1]

The pathway moves from screening through confirmation to characterisation, because treatment follows characterisation rather than the diagnosis alone. ^[1]

• Full blood count with differential and manual film review, looking for dysplasia and blast count.
• Reticulocyte count (often reduced).
• Lactate dehydrogenase (elevated, unfavourable prognostic signal), ferritin, erythropoietin level.
• Blood typing if transfusions are expected.
• Vitamin screen (iron, B12, folate, copper) to exclude reversible causes. ^[1]

• Cytology with iron, peroxidase and esterase staining: extent of dysplasia, blast percentage.
• Histology (trephine) for cellularity, fibrosis and iron.
• Chromosomal analysis of at least 25 metaphases, complemented by fluorescence in situ hybridisation in individual cases, with cytogenetic risk assignment by IPSS-R definitions (very good, good, intermediate, poor, very poor).
• Somatic mutation analysis of the genes important in MDS, including TP53, SF3B1, SRSF2, U2AF1, ASXL1, RUNX1, TET2, DNMT3A, EZH2, BCOR, STAG2, ETV6, NRAS, KRAS, CBL, JAK2, IDH1/2 and others listed in open resources; the more genes covered, the more robust the IPSS-M estimate.
• Not mandatory: immunophenotyping, WT1 expression, HLA typing and CMV status for transplant candidates. ^[1][2]

• Dysplasia at or above 10 percent plus cytopenia establishes MDS in most patients, unless another cause explains everything. ^[2]
• Blast count places the subtype (low blasts, IB1, IB2) and largely sets the risk. ^[1]
• Ring sideroblasts at or above 15 percent identify the SF3B1-driven subgroup even if sequencing is delayed. ^[2][3]
• Cytogenetics place the patient into an IPSS-R category; in IPSS-R, del(5q), -Y and del(20q) are favourable, complex karyotype (especially more than three abnormalities) very unfavourable. ^[2][6]
• Somatic mutations explain the complication profile and refine prognosis in IPSS-M; TP53 multihit status in particular defines a very-high-risk, treatment-refractory subtype. ^[2][3]

A practical diagnostic rule combines three domains. ^[1][2][3]

1. Persistent cytopenia in at least one lineage (neutrophils under 1.8 x 10^9/L, platelets under 100 x 10^9/L, or haemoglobin under standard thresholds depending on sex) that is unexplained.
2. Dysplasia of 10 percent or more of nucleated cells in one or more myeloid lineages, and/or
3. An MDS-defining abnormality (for example del(5q), monosomy 7, complex karyotype, SF3B1 mutation with low blasts), even when dysplasia is equivocal.

Blast thresholds separate MDS from AML. In WHO 2022 MDS-IB2 carries 10 to 19 percent marrow or blood blasts. Most classifications still use 20 percent blasts or more as the AML threshold, but WHO 2022 classifies some patients with fewer blasts (NPM1 mutation, or KMT2A, MECOM or NUP98 rearrangements) as AML regardless of blast percentage. ^[3][2]

Pre-MDS conditions matter in the clinic: CHIP without cytopenia is observed; CCUS with cytopenia and mutations but no dysplasia needs careful follow-up because it can evolve into MDS. ^[2]

The original International Prognostic Scoring System (IPSS) used bone-marrow blasts, karyotype and the number of cytopenias to place patients in four risk groups: low (0), intermediate-1 (0.5 to 1), intermediate-2 (1.5 to 2) and high (2.5 to 3.5). Many clinical trials still use these cut-offs to distinguish lower from higher risk. ^[2]

IPSS-R refined IPSS cytogenetic risk (very good, good, intermediate, poor, very poor), refined blast cut-offs and the degree of cytopenia, and added haemoglobin and platelet detail. Patients fall into five groups with different median survival and AML-evolution times. ^[2][6]

A practical hazard is the intermediate-risk grey zone. IPSS-R intermediate-risk patients can have relatively indolent disease (similar to lower-risk MDS) or relatively aggressive disease. The decision to treat as lower or higher risk should draw on molecular genetics, marrow cellularity, symptoms and cytopenia severity, not on the score alone. ^[2][6]

The simplest practical cut-off used in many guidelines stratifies patients into lower-risk (IPSS-R very-low, low or intermediate at 3.5 or under) versus higher-risk (intermediate at 4.0 or more, high and very high). ^[2][6]

IPSS-R scores rely on cytogenetics and blood counts measured at diagnosis and predict outcomes for patients treated with supportive care. They are validated in adults with de novo disease and, strictly speaking, do not apply to therapy-related MDS (these patients should be treated as very high risk) or to every patient beyond the original cohort. ^[2][6]

IPSS-R deliberately ignored somatic mutations because it was built before mutation testing became routine. IPSS-M molecular (IPSS-M) adds somatic mutation data for 31 genes, combined with the IPSS-R clinical and cytogenetic parameters, to produce a personalised risk score. ^[2][5]

Key IPSS-M facts from open-access reviews: ^[2][5]

• It is a regression model designed for web-based use (mds-risk-model.com) and is most accurate when the full mutation panel is available. ^[2]
• It stratifies patients into six risk categories: very low, low, moderate low, moderate high, high and very high. ^[1][2]
• Median overall survival ranges from about 10.6 years in the very-low-risk group to about 1.0 year in the very-high-risk group. ^[2]
• It restratifies roughly half of patients compared with IPSS-R; about 74 percent of those are up-staged and 26 percent down-staged. ^[2]
• It applies to primary and secondary or therapy-related MDS. ^[2]
• Most mutations worsen prognosis (TP53, RUNX1, ASXL1, EZH2, SRSF2, U2AF1, DNMT3A and others); SF3B1 is the notable exception with a favourable association unless combined with del(5q) or additional aberrations (RUNX1). ^[1][2]

The IPSS-M web calculator tolerates missing data, but incomplete mutation information degrades the estimate. ^[2]

The golden thread is risk adaptation, decided jointly on disease biology (risk group), symptoms, patient fitness and available resources.

Lower-risk disease accounts for about two thirds of patients at diagnosis. The paradigm is not "do nothing" but "act on the specific problem the clone creates": anaemia, severe cytopenias, iron overload and, importantly, quality of life. ^[5][6]

Observation is reasonable for patients who are asymptomatic or mildly cytopenic, with regular blood counts and a plan for when cytopenias worsen. ^[6]

Transfusion support is the backbone: red-cell transfusions for symptomatic anaemia, platelet transfusions for severe thrombocytopenia or bleeding. ^[1][6]

Iron chelation should be considered when serum ferritin exceeds 1000 ng/mL in transfusion-dependent patients, to reduce iron-overload injury to the heart and liver. Data from a randomised chelation trial support a survival benefit, and about 10 to 15 percent of patients show improved blood counts. ^[1][6]

Erythropoiesis-stimulating agents (ESAs) are typical first-line therapy for anaemia in lower-risk MDS. ^[5][6]

• Epoetin alfa: 85 percent of very-low-risk and 68 percent of low-risk patients with moderate transfusion need and endogenous erythropoietin (EPO) levels under 200 ng/mL respond within weeks. ^[1]
• Across broader populations the best-studied predictor is the serum EPO level: over 70 percent response when serum EPO is under 100 U/L, and under 10 percent when serum EPO exceeds 500 U/L, where a trial of ESA is usually not warranted. ^[6]
• A validated model combines EPO with granulocyte colony-stimulating factor (G-CSF) for anaemia of MDS. ^[5][6]

Notably, only around 31.8 percent of ESA responders have EPO levels under 200 mU/mL, and the older predictive model uses EPO under 500 mU/mL, so EPO level is predictive but not a hard cut-off for an individual patient. ^[5]

Lenalidomide is specifically useful in lower-risk MDS with an isolated del(5q) abnormality and no excess blasts, producing transfusion independence in roughly 65 to 70 percent and cytogenetic remission in 30 to 40 percent. In German registry data about two thirds become transfusion-free within four months. Disposition-of-clone analysis (interphase FISH in peripheral blood) allows molecularly guided treatment interruptions. Screening for TP53 mutation before starting lenalidomide is advised, as biallelic TP53 disease can behave more aggressively. ^{[1][6] [2]}

Luspatercept was approved for adult patients with IPSS-R very-low to intermediate-risk MDS with ring sideroblasts (or MDS/MPN-RS-T), who require at least 2 red-cell units per 8 weeks and have failed or are unlikely to respond to ESA (EPO over 200 U/L or EPO over 200 U/L at entry). It is given subcutaneously at a starting dose of 1 mg/kg every 3 weeks, titratable to 1.33 mg/kg and then 1.75 mg/kg after two consecutive 6-week cycles without transfusion independence; it should be discontinued if there is no reduction in red-cell transfusion burden after 3 cycles at 1.75 mg/kg. In the double-blind phase III MEDALIST trial 229 such patients received luspatercept or placebo: 38 percent versus 13 percent achieved 8 weeks or more of transfusion independence in weeks 1 to 24 (p less than 0.001); 28 percent versus 8 percent achieved 12 weeks in the first 24 weeks; with the benefit seen in the 200 to 500 U/L EPO subgroup, where 40 percent had fewer transfusions. Dose is adjusted for haemoglobin 11.5 g/dL or more without transfusions and for haemoglobin rises over 2 g/dL in three weeks. ^{[5] [1]} Newer comparative data from the COMMANDS trial support luspatercept as the preferred first-line agent for previously untreated, transfusion-dependent, lower-risk, ring-sideroblast-positive disease, with response rates of 60.4 percent versus 34.8 percent for epoetin alfa in that setting and longer response duration; imetelstat is a second-line option for transfusion-dependent lower-risk patients after ESA or luspatercept failure, or when the erythropoietin level is high. ^[6]

Thrombopoietin analogues (eltrombopag, romiplostim) may be used case by case for severe thrombocytopenia to avoid platelet transfusions, in line with phase II data. ^[1][6]

Immunosuppressive therapy with anti-thymocyte globulin (plus or minus cyclosporine) can improve haematopoiesis in hypoplastic or suspected immune-mediated MDS, usually in younger patients (under 65 to 70 years) with favourable cytogenetics and a hypocellular marrow. ^[1][6]

Oral hypomethylating agents in a reduced three-day schedule have been studied in lower-risk disease, with better response rates to three-day decitabine than three-day azacitidine in one multicentre adaptive trial; azacitidine or decitabine can be reserved for lower-risk patients with persistent cytopenias or higher-risk behaviour after failure of the above options. ^[6]

The transplant question comes first. For patients with higher-risk disease who are transplant-eligible, life expectancy is improved by allogeneic haematopoietic stem-cell transplantation as soon as possible once a donor is identified. Age up to 75 years is now routine when performance status is good, and myeloablative conditioning is the standard when tolerated, with reduced-intensity conditioning trading lower transplant-related mortality for higher relapse. ^[1][6]

Practical transplant points supported by the cited reviews: ^[1][6]

• Median overall survival after allogeneic transplantation in selected Dusseldorf cohorts was about 96 months, versus 18 months with best supportive care alone in higher-risk disease in the same dataset. ^[1]
• Two-year non-relapse mortality after allogeneic transplantation is around 16 percent. ^[1]
• Adverse genetics (RUNX1, TP53) increase the likelihood of post-transplant relapse (5-year relapse around 61 percent in such groups). ^[1]
• Pre-transplant therapy with a hypomethylating agent or intensive chemotherapy mostly controls disease at best and can select resistant clones; a proportion of patients fail to proceed to transplant because of complications. In stable disease one reasonable approach is observation during the donor search with an early primary transplant. ^[1]
• Once transplanted, minimal residual disease monitoring is essential; when MDS relapses, 29 to 71 percent of patients can be successfully treated with hypomethylating agents plus donor lymphocyte infusions. ^[1]

Hypomethylating agents (HMAs) are the standard of care for patients who are not transplant candidates, or as bridging therapy.

• Azacitidine 75 mg/m^2 per day for 7 consecutive days every 28 days, subcutaneously or intravenously. ^[4][6]
• In the AZA-001 trial (358 patients) azacitidine improved overall survival to 24.5 months versus 15.0 months (log-rank p 0.0001) and delayed AML progression to 17.8 versus 11.5 months versus conventional care, with 45 percent of transfusion-dependent patients becoming transfusion independent. ^[4]
• When the standard 7-day schedule is impractical for outpatients, 5 consecutive days at 75 mg/m^2 can be used for lower-risk disease and a 5-2-2 schedule (5 days, 2-day pause, 2 days) for higher-risk disease; fewer than 5 days or a reduced total dose is not recommended. ^[4]
• Durability planning: 2 to 3 cycles usually show the first response, around 90 percent of responses occur within the first 6 cycles, and a minimum of 4 to 6 cycles is recommended before concluding non-response; the median cycle number to show survival benefit in AZA-001 was 9. Continue while clinical benefit exists or until progression or unacceptable toxicity. ^[4]
• About half of patients respond (complete remission in 10 to 17 percent in pooled CALGB data), and most responders lose response within 2 years. ^{[4][5] [1]}
• Decitabine 20 mg/m^2 per day for 5 days every 28 days is the other commonly used HMA; larger differences in trial design make the two agents hard to compare directly. ^[6]

Azacitidine after HMA failure. There is no approved second-line therapy and median survival falls below 6 months after failure. Options include transplant (best if undertaken before progression to AML), investigational agents (shown in series to give 13.2 months versus 4 months with supportive care), re-challenge with intensive chemotherapy in fit patients, or best supportive care. ^{[4][6] [1]}

Additions to HMAs (venetoclax, magrolimab, sabatolimab, rigosertib, immune checkpoint inhibitors) are being studied in trials; combinations with lenalidomide or histone deacetylase inhibitors have not improved outcomes on the whole, mainly because of added myelosuppression. ^[1][6]

Classical induction chemotherapy as used in AML is no longer recommended for high-risk MDS: remission rates are low, durations short, and long-term results poor. ^[1][6]

• Therapy-related or secondary MDS (after genotoxic exposure) frequently carries TP53 mutations or complex karyotype, behaves like very-high-risk disease even when the IPSS-R omits it, and should be treated as such. ^[2][6]
• del(5q) (isolated) is the main genotype-specific lower-risk target, as above, and routine TP53 screening before lenalidomide is sensible. ^[1][6]
• Hypoplastic MDS has more severe cytopenias but better prognosis than hypercellular MDS of similar risk scores, with distinct immune/genomic features; anti-thymocyte globulin or transplant is the directional therapy. ^[3][1][6]
• SF3B1/MDS-RS now qualifies for luspatercept after ESA failure because WHO 2022 no longer requires large numbers of ring sideroblasts as long as the mutation is present, allowing more patients to receive targeted therapy. ^[3][5]
• Older adults. Fitness for transplant, not calendar age, determines the pathway: patients up to about 75 years undergo transplant routinely when performance status permits. ^[6]

The biology is the same worldwide; the constraints are diagnostic and therapeutic. ^[6][1]

• Diagnosis is often possible with a manual film, a marrow aspirate and trephine, and a modest cytogenetic panel. Full mutation analysis improves scoring but is not a requirement to start sensible supportive care. ^[1][3]
• Where ESA, lenalidomide and luspatercept are unavailable or unaffordable, red-cell transfusion with pragmatic EPO analogue use and iron management remains the standard.
• Azacitidine and decitabine are simply not available in many settings; where access exists, the standard 7-day schedule is preferable and 5-day schedules acceptable. ^[4]
• Allogeneic transplant requires a donor, cell-processing capacity and long-term follow-up infrastructure; where absent, maximising supportive care and access to trials is the realistic plan.
• Iron overload management (ferritin testing and chelation) is frequently under-resourced; transfusion triggers and local thresholds should be agreed explicitly. ^[1][6]

• Infection (neutropenic and otherwise) is the leading cause of death. ^[6]
• Bleeding from severe thrombocytopenia; platelet transfusions and (in selected cases) TPO analogues help. ^[1][6]
• Transfusion-related iron overload: cardiology, hepatic and endocrine sequelae when ferritin rises above 1000 ng/mL. ^[1][6]
• Progression to AML: about a quarter to a third of patients; elevating blasts, new cytogenetic clones and accumulating mutations are the warning signs. ^[1][2]
• Treatment toxicity: azacitidine commonly causes cytopenias, gastrointestinal upset and injection-site reactions in the first 2 cycles, falling thereafter; venetoclax and checkpoint inhibitors are investigated alongside HMAs. ^[4][6]
• Psychosocial and quality-of-life burden, in particular fatigue, which is often the dominant symptom even before the anaemia is severe. ^[6]

Prognosis in MDS spans years to months and is driven by the biology, the risk score and patient fitness.

• Median survival in a broad MDS population is roughly 3 years. ^[1]
• IPSS-R risk groups carry median survivals from 8.8 years (very low) through 5.3 (low), 3.0 (intermediate), 1.5 (high) to 0.8 years (very high), and median times to AML progression of 10.8, 3.2, 1.4 and 0.7 years respectively. ^{[6] [2]}
• IPSS-M stretches these from about 10.6 years (very low risk) to about 1.0 year (very high risk). ^[2]
• Favourable subtypes include isolated del(5q) (about 67 percent 6-year overall survival, 53 percent progression-free, 15 percent 5-year AML progression risk) and SF3B1-mutated low-blast disease (median overall survival 79 months versus 53 months for wild-type; only 7 percent progress to AML after 9 years). ^[2]
• Unfavourable subtypes include biallelic TP53 (about 10 months), MDS with fibrosis (about 15 months) and complex karyotype (median overall survival roughly 1.5 years with three abnormalities, falling to 0.7 years with more than three). ^{[2][3] [1]}
• Therapy-related MDS can be expected to behave like very-high-risk disease even when scoring formulas do not include it. ^[2]
• Luspatercept improved haemoglobin-related endpoints (38 percent transfusion independence at 8 weeks) and azacitidine improved overall survival (24.5 versus 15.0 months in AZA-001). ^[5][4]

Measurable residual disease monitoring after transplant is central to catching relapse at a point when it can be treated. ^[1]

There is no population screening for MDS. What can be done is risk reduction and early detection of predisposition. ^[1][2]

• Avoid unnecessary exposure to alkylating agents and topoisomerase inhibitors in young patients (and use radiotherapy judiciously). ^[2]
• Reduce occupational benzene exposure and include occupational history in older adults' assessments. ^[1]
• Detect and follow CHIP/CCUS so progression is caught when intervention still has room to work. ^[2]
• Test for germline predisposition in young patients (under 50 years), familial cases or multiple neoplasms, because the implications for treatment and surveillance differ. ^[1]
• Early diagnosis matters because supportive care and risk-adapted treatment (including transplant) do change outcomes, and an incidental cytopenia work-up should not stop at "probably iron deficiency". ^[1]

• Children and young adults. Primary MDS is rare in children and adolescents; inherited predisposition should be sought (DDX41, Fanconi, dyskeratosis congenita, germline RUNX1 and related syndromes). Management often runs through transplant. ^[1][2]
• Pregnancy. MDS in pregnancy requires joint haematology, obstetric and transfusion-medicine care; it is rare and the literature is case-based, so decisions are individualised. ^[1]
• Older adults. Comorbidity is common and should drive fitness assessment rather than calendar age alone. ^[1][6]
• Hypoplastic MDS. As above: younger patients, a hypocellular marrow and absence of complex karyotype make immunosuppressive therapy attractive. ^[6]

See also the earlier practical notes. The reality of many low- and middle-income regions is that the limiting factors are diagnostic (marrow, cytogenetics, mutation panels) and therapeutic (access to HMAs, lenalidomide, luspatercept and transplant), not biology. ^[1][6]

• A manual film plus a well-processed marrow aspirate and trephine reaches the diagnosis in many patients; ring sideroblasts and dysplasia percentages are assessable without sequencing. ^[1]
• Where mutation testing is unavailable, risk stratification falls back to IPSS-R which still discriminates.
• Transfusion dependence with iron overload remains the commonest killer of lower-risk patients; ferritin testing and pragmatic chelation policies give population-level benefit. ^[1][6]
• Patient registries and closer links with MDS centres (where these exist) allow access to trials and to transplant pathways. ^[1]

Related MedGrasp topics are named here in plain text because this page was written from primary sources and the other pages are written separately. Topics that link up naturally: acute leukaemias; aplastic anaemia; anaemia (approach and iron deficiency); transfusion reactions (blood product support during marrow failure); non-Hodgkin lymphoma; multiple myeloma; neutropenia; immune thrombocytopenia; sideroblastic and megaloblastic anaemia; sickle cell disease; thalassaemias; haemolytic anaemias; normal haemostasis and thrombosis. In the meantime this page stands alone because myelodysplastic syndromes is the specific disease those pages mention in passing but never describe.