MDS What causes it?

Amongst the most frequent questions asked by patients soon after the diagnosis of MDS are: 

  • “Why me?”
  • “What causes MDS?” 
  • “Could I have done anything different to avoid getting MDS?” and “Can my children get it?” 
The simple answer to these questions is that for the vast majority of patients, we have few clues as to the cause of their MDS. The study of the causes of these diseases is proving difficult for the following reasons: 
  1. MDS is more than one disease
  2. Few comprehensive patient registries exist to accurately determine who gets the different subtypes of MDS (e.g. age/sex distribution)
  3. Determining the length of time to develop MDS is difficult 

More Than One Disease

The diagnostic process involves categorising an individual’s disease into one of five French-American-British (FAB) groups or one of six World Health Organisation (WHO) subgroups (the latter excluding sub-types CMML and the old RAEB-t). These classification systems recognise the differences in the bone marrow appearance and the chromosome abnormalities within the different subgroups of MDS. It does not therefore require a large leap of faith to expect that diseases that look different down a microscope (albeit with certain overlapping similarities), might have different causes. 

Attempts to study the cause of MDS (“epidemiology”) have focussed mainly on case-control studies, consisting of questionnaires requesting information about the work and recreational background of MDS patients, compared with a “control” group of individuals who do not have MDS. Whilst these efforts are commendable, and the best that can be achieved, there are many limitations to such studies. These include “recall bias”, relying on the patient’s memory for accuracy, and size of the patient group studied, which in turn determines the “power” of the study, and hence the confidence that the results are truly accurate and not simply statistical chance. For the purposes of most of these studies, MDS has been considered as one disease, given that the numbers in each subgroup will be small. 

Who Gets MDS?

MDS is a rare disease, whose incidence is 4 per 100,000. The disease becomes more common with increasing age, such that the incidence rises to >30 per 100,000 for people over 70 years of age. Males are more commonly affected than females, although there is some evidence that this is not so for Refractory Anemia with Ring Sideroblasts (RARS) (Dr. U. Germing, personal communication). 

MDS in children is more infrequent still, and has different characteristics to the adult form. Examples of these differences include the spectrum of FAB/WHO types (RARS and 5q- syndrome are almost never seen in childhood), and of chromosome abnormalities (a higher proportion of children have abnormalities of chromosome 7). 

MDS may also evolve from related disorders such as Aplastic Anemia (following immuno-suppression treatment) or Paroxysmal Nocturnal Haemoglobinuria (PNH). 

How Long Does it Take to Develop MDS?

For patients with de novo MDS, the latency time to disease development is unknown. From a biological angle, there will be at least two phases of disease development, namely: 
  1. The time from the first damage in the bone marrow to the appearance of a change in the blood count, then,
  2. the time from the first blood count abnormality to the presentation with clinically relevant disease (usually symptoms of anaemia). Both are impossible to study systematically at present.
For cases of MDS developing after exposure to an agent known or presumed to cause MDS, the latency period varies. This may be from 1–41 years for different radiation exposures[1], 1–10 years for alkylator cytotoxic drugs[2,3] and more difficult to assess (but up to 30 years?) for benzene[4]. 

Established Causative Factors for MDS

  • Cytotoxic drugs Therapy-related MDS and AML (t-MDS/AML) are well-recognised though rare complications following cytotoxic drug therapy for malignant and some non-malignant (mainly autoimmune) diseases. The risk of developing t-MDS/AML following therapeutic radiotherapy remains less clear[5]. Several classes of cytotoxic drugs are implicated, but alkylators most frequently cause an MDS phase. One of the major challenges in the treatment of highly curable diseases such as Hodgkin’s lymphoma is now to reduce the risk of late complications, and newer therapies should prove less likely to produce t-MDS/AML.
  • t-MDS/AML constitute <10% all cases of adult MDS however, and the study of t-MDS/AML as a model for the causes of de novo MDS is problematic. Many cases of t-MDS cannot be easily classified due to bone marrow fibrosis (scarring). RARS and CMML are relatively under-represented, and the chromosome abnormalities in t-MDS/AML differ from those of the de novo diseases.3 Survival of patients with t-MDS is also poorer than for de novo MDS. 

Possible Causative Factors for MDS 

  • Radiation MDS cases are reported in cohorts of people exposed to radiation, for treatment of diseases such as ankylosing spondylitis, or following exposure to the A-bomb in Hiroshima and Nagasaki. Some of these cases occurred up to 40 years after exposure and thus the precise association between the development of MDS and exposure to radiation is not possible to quantify. Similarly, weak associations between radiation exposure and MDS are identified in some (but not all) case-control epidemiology studies. 
  • Benzene Legislation now ensures that exposure to high concentrations of benzene in the workplace or the environment should not occur. Thus, the main sources of exposure to low concentrations of benzene in daily life are tobacco smoke and petrol (cartoon). Tobacco smoking is a weak but consistent risk factor (less than two-fold) for MDS in several case-control studies[6], but also contains many other carcinogens in addition to benzene. 
  • In contrast, exposure to high concentrations of benzene clearly causes bone marrow toxicity, usually aplasia, some of which will progress to MDS and/or AML[7,8]. 
  • Miscellaneous Whilst each individual published case-control study has identified a number of occupations and substances, which may be risk factors for MDS, there is little consistency between these studies. An exhaustive list of these possible risk factors is therefore not helpful, as many will represent statistical chance or very weak relative risks. 
The incidence of MDS increases with increasing age. This has been interpreted in two ways; the disease must result from a progressive accumulation of a lifetime’s exposure to a toxic agent or, the aged bone marrow “stem” cell is easier to damage than its younger counterpart. The process of “ageing” is not well defined, though much blame is heaped upon “free radicals”[9], defence against which deteriorates with age. It remains unclear in what way this may be relevant to diseases of older age such as MDS. 

Is There an Inherited Tendency to Develop MDS? 

The vast majority of MDS patients presenting in adulthood have no relatives with the disease, and no obvious inherited disease with a tendency to MDS. 30% children with MDS have other associated abnormalities, and some of these are part of well-recognised syndromes including Fanconi Anaemia, and Blooms Syndrome.
Families with several cases of MDS are described, but are exceptionally rare. Although still very rare, familial MDS may be more likely in the family of a child with monosomy [7,10]. In the absence of other family members with MDS, it is safe to confirm that the disease is vanishingly unlikely to run down the generations. 

How Can We Study the Cause in the Future? 

A whole new avenue of research into the contribution of genetics to the cause of diseases, involves the study of natural variations in our DNA from person to person. These variations can often lead to changes in the function of cells, and may therefore affect the natural functions of a cell, such as neutralising toxic chemicals that enter the body. To date, no definite natural variation in a gene has been associated with increasing the risk of developing MDS, but this field is only in its infancy. Large numbers of patient samples are needed for such studies, particularly in MDS, where these large numbers are required for each of the FAB/WHO groups. 

This initiative requires the development of large Biobanks, married to high-quality registries of clinical and laboratory information cataloguing the characteristics of each patient’s disease. 


Despite more than a decade of dedicated effort from epidemiologists, clinicians and scientists, the cause of MDS remains largely unknown. It is inevitable that the different subtypes of MDS will have different causes. We must use the little high-quality demographic data (Who gets MDS?) to develop hypotheses, and test these in a combination of clinical and molecular epidemiology studies, which by definition will need to involve very large patient numbers. National and international collaborative efforts are underway to attempt to achieve this. 

David T. Bowen, MD at University of Dundee Medical School, Dundee, Scotland


  1. Moloney WC. Radiogenic leukemia revisited. Blood. 1987;70:905–908.
  2. Pedersen-Bjergaard J, Andersen MK, Christiansen DH, Nerlov C. Genetic pathways in therapy-related myelodysplasia and acute myeloid leukemia. Blood. 2002;99:1909–1912.
  3. Mauritzson N, Albin M, Rylander L et al. Pooled analysis of clinical and cytogenetic features in treatment-related and de novo adult acute myeloid leukemia and myelodysplastic syndromes based on a consecutive series of 761 patients analyzed 1976–1993 and on 5098 unselected cases reported in the literature 1974–2001. Leukemia. 2002;16:2366–2378.
  4. Voytek PE, Thorslund TW. Benzene risk assessment: status of quantifying the leukemogenic risk associated with the low-dose inhalation of benzene. Risk Anal. 1991;11:355–357.
  5. Andersen MK, Johansson B, Larsen SO, Pedersen-Bjergaard J. Chromosomal abnormalities in secondary MDS and AML. Relationship to drugs and radiation with specific emphasis on the balanced rearrangements. Haematologica. 1998;83:483–488.
  6. Bjork J, Albin M, Mauritzson N et al. Smoking and myelodysplastic syndromes. Epidemiology. 2000; 11:285–291.
  7. Aksoy M, Dincol K, Erdem S, Dincol G. Acute leukemia due to chronic exposure to benzene. Am J Med. 1972; 52:160–166.
  8. Rothman N, Smith MT, Hayes RB et al. Benzene poisoning, a risk factor for haematological malignancy, is associated with the NQO1 609C->T mutation and rapid fractional excretion of chlorzoxazone. Cancer Research. 1997;57:2839–2842.
  9. Kirkwood TB, Austad SN. Why do we age? Nature. 2000;408:233–238.
  10. Luna-Fineman S, Shannon KM, Lange BJ. Childhood monosomy 7: epidemiology, biology, and mechanistic implications. Blood. 1995;85:1985–1999. 


  • Epidemiology: The study of frequency and cause of disease.
  • Incidence: The number of new cases presenting per year in a given population.
  • Power (of a study): A calculation used in preparing a study to determine whether the study is capable of answering the question asked. The power is determined by the number of patients in the study and the size of the effect that the researchers expect to see and would consider to be medically important.
  • Case-control study: A widely used epidemiology method for studying cause of disease. A group of patients (“cases”) are compared with a matched group of individuals lacking the disease (usually healthy), “controls”.
  • Relative risk: The numerical chances of developing a disease in a given set of circumstances. In this context the chance of developing MDS if you are exposed to chemical x, or have genetic tendency y. Relative risk of 2=two-fold, 4=four-fold etc.
  • Latency: In this context, the “lag” time from the first event that might start the disease process to the time the disease presents to a doctor.
  • De novo: Disease presenting with no obvious factors known to cause the disease.