Ann Hematol (2006) 85: 514–519 DOI 10.1007/s00277-006-0128-y
ORIGINA L ARTI CLE
Tatjana Pekmezovic . Nada Suvajdzic Vukovic . Darija Kisic . Anita Grgurevic . Andrija Bogdanovic . Mirjana Gotic Milena Bakrac . Nenad Brkic
A case-control study of myelodysplastic syndromes in Belgrade (Serbia Montenegro) Received: 23 December 2005 / Accepted: 6 April 2006 / Published online: 12 May 2006 # Springer-Verlag 2006
Abstract The objective of the study was to investigate factors related to the occurrence of myelodysplatic syndromes (MDS) in the population of Belgrade (Serbia Montenegro). The case-control study was conducted during the period 2000–2003. The study group consisted of 80 newly diagnosed MDS patients and 160 sex- and agematched hospital controls with nonmalignant and noninfectious diseases. The disease categories in the control group were circulatory (51 patients, 32%), gastrointestinal (53 patients, 33%), and ophthalmological (56 patients, 35%) disorders. Conditional univariate and multivariate logistic regression analyses were applied. Multivariate analysis showed the following factors to be significantly related to MDS: exposure to chemicals (OR=10.8, 95%CI 3.2–36.2, p=0.0001), viral upper respiratory tract infections (twice a year or more, OR=5.8, 95%CI 2.5–13.6, p=0.0001), exposure to insecticides, pesticides and herbicides (OR=5.2, 95%CI 1.8–15.1, p=0.003), coffee (OR=5.1, 95%CI 1.9–13.7, p=0.001), and alcohol consumption (OR=2.2, 95%CI 1.1–4.6, p=0.033). The findings support the hypotheses that exposure to chemical agents, pesticides, insecticides, and herbicides, certain lifestyle factors (alcohol and coffee consumption), and frequent viral infections may be involved in the etiology of MDS, but these results should be confirmed by further investigations.
T. Pekmezovic (*) . D. Kisic . A. Grgurevic Institute of Epidemiology, School of Medicine, University of Belgrade, Visegradska 26A, Belgrade 11000, Serbia Montenegro e-mail:
[email protected] Tel.: +381-11-3615768 Fax: +381-11-3615768 N. Suvajdzic Vukovic . A. Bogdanovic . M. Gotic . M. Bakrac . N. Brkic Clinical Centre of Serbia, Institute of Hematology, Dr Koste Todorovica 4, Belgrade 11000, Serbia Montenegro
Keywords Myelodysplastic syndrome . Risk factors . Epidemiology . Case-control study . Belgrade
Introduction Myelodysplastic syndromes (MDS) are heterogeneous blood disorders characterized by ineffective hematopoiesis, leading to one or more peripheral blood cytopenias [11]. Although the occurrence of MDS after exposure to antineoplastic drugs (mainly alkylating agents) and radiotherapy is well-established [10, 19], the etiology of idiopathic (primary) MDS remains largely unknown, and there are many epidemiological studies conducted specifically for MDS [15]. In previous case-control studies, several positive associations between environmental and occupational factors have been observed. Occupational exposure to chemicals, such as petrol and diesel vapors, exhaust gases, metals, solvents, and pesticides, have been established as risk factors for MDS [14, 15, 25]. The association with cigarette smoking has been reported in some studies [2, 25], but not in others [14]. Similarly, heavy alcohol consumption has been associated with a high risk in some studies [9], but not in others [18]. However, a recent case-control study suggests that both exogenous and endogenous factors may play a role in the etiology of MDS [4]. The aim of this study was to determine the risk factors for primary MDS among adults in Belgrade (Serbia Montenegro).
Materials and methods Study population The study group consisted of 80 consecutive MDS patients treated at the Institute of Hematology, Clinical Centre of Serbia, Belgrade, over the period 2000–2003. This is the national reference center for hematology and oncology, and virtually all MDS cases from Belgrade are diagnosed and
515
treated at the Institute. The inclusion criteria were patients (a) with MDS confirmed according to the French-American-British (FAB) classification [1], (b) aged between 20 and 85, (c) with a diagnosis of primary MDS at least 2 years before the study. The control group consisted of 160 patients matched by sex and age (±2 years), attending the Outpatient Departments of Cardiology, Gastroenterology, and Ophthalmology (also in the Clinical Centre of Serbia, Belgrade) for nonmalignant and noninfectious diseases. All subjects in the control group passed a routine hematological checkup before inclusion; patients diagnosed with any blood disorders or liver disease were not chosen as controls because both of these may have been related to exposure to organic solvents and alcohol (considered risk factors for MDS). The disease categories in the control group were circulatory (51 patients, 32%), gastrointestinal (53 patients, 33%), and ophthalmologic (56 patients, 35%) disorders. Consent for inclusion in the study was obtained for all the MDS patients and controls. The research was approved by the Institutional Review Committee for human subject investigations. Data collection All study participants were interviewed during their visits to the Outpatient Departments of the above-mentioned institutions by two physicians (DK, AG) using the same questionnaire. The questionnaire collected demographic information and information regarding personal and family history, lifetime residence, habits, occupational exposures to radiation and chemicals, and diet. This was supplemented and validated by an examination of the medical records. Participants were asked about the number of cigarettes smoked per day and duration of smoking. Questions about alcohol intake comprised of the type of liquor and the amount and frequency of consumption per week. Occupational exposure to radiation, a wide variety of chemical agents, insecticides, pesticides, and herbicides were analyzed by using a qualitative method (exposure to the hazard or not). Subjects were asked how often (per week) they had consumed each one of seven selected food groups or food items. All exposure data referred to a time period before the diagnosis of MDS for the patients and a corresponding period for the controls. Data analysis Odds ratios (OR) and 95% confidence intervals (95%CI) according to the Miettinen procedure were calculated separately for each variable using univariate conditional logistic regression analysis [13]. Variables that were related
to MDS at a significant level of p≤0.10 entered the final model of multivariate conditional logistic regression analysis to evaluate their independent contribution to the overall risk of MDS [12].
Results Eighty patients (41 men and 39 women) with MDS and 160 matched control subjects (82 men and 78 women) were included in the study. Most were aged 65 years and over. Other demographic characteristics of the MDS patients and the controls are presented in Table 1. According to the FAB classification, the refractory anemia (RA) was the most frequent morphological subtype (38/80 patients). The other MDS subtypes were less frequent: five patients with RA with ringed sideroblasts (RARS), 20 with RA with excess of blasts, low-blast-count type (RAEB-1, blasts10%), three patients with refractory anemia with excess of blasts in transformation (RAEB-t, blasts >20%) and two patients with dysplastic form of chronic myelomonocytic leukemia. Univariate logistic regression analysis found that MDS patients lived in rural areas significantly more frequently Table 1 Demographic characteristics of the MDS cases and their controls Variable
Sex Male Female Age (years) ≤45 46–55 56–65 66–75 76–84 Residential area Ruralb Urban Education Elementary schoolb Secondary and high school Occupation Farmerb Housewifeb Manual workerb Clerk and professional
Controls Odds Cases (No.=80) (No.=160) ratio (95%CI)
pa
41 39
82 78
Matched
12 14 15 24 15
25 26 31 51 27
Matched
19 61
19 141
2.3 (1.1–4.7) 0.020
36 44
53 107
1.6 (0.9–2.9) ns
5 13 29 33
12 28 45 75
1.0
1.3 (0.7–2.2) ns
95%CI 95% confidence interval, ns not significant a According to univariate logistic regression analysis b Reference category
516
than did the controls (OR=2.3, 95%CI 1.1–4.7, p=0.020). Those with only elementary school education (8 years of education) had a higher risk of MDS compared to those with secondary and high school education (OR=1.6, 95% CI 0.9–2.9), and MDS patients were more frequently farmers, housewives, and manual workers compared with controls (OR=1.3, 95%CI 0.7–2.2). There were significant differences between MDS patients and controls concerning several variables in the past medical history, birth history, and family history (Table 2). Viral (OR=9.5, p=0.001) and streptococcal (OR=6.5, p=0.001) upper respiratory tract infections (twice or more per year) were significantly associated with an increased risk of MDS. Appendectomy (OR=2.2, p=0.023), other
Table 2 Birth history, personal history, and family characteristics of the MDS patients and the controls Patients Controls Odds 95% (No.=80) (No.=160) ratio CI Birth history Maternal age at participants birth (>26 years)b Paternal age at participant birth (>29 years)b Birth order (4+)b Total number of children in the family (>4)b Personal history Upper respiratory infections—viral ≥2 times per yearb Upper respiratory infections— streptococcal ≥2 times per yearb Allergic diseases Diabetes mellitus Tonsillectomy Appendectomy Other operations Family history Diabetes mellitus Cancer
pa
35
55
1.4
(0.8–2.6)
ns
29
46
1.4
(0.8–2.6)
ns
15 24
20 42
1.6 1.2
(0.7–3.6) (0.6–2.3)
ns ns
33
11
9.5
(4.5–20.3) 0.001
14
5
6.5
(2.3–18.9) 0.001
6 10 10 20 34
13 8 15 21 45
0.9 2.7 1.4 2.2 1.9
(0.3–2.5) (1.0–7.2) (0.6–3.2) (1.1–4.4) (1.1–3.3)
ns 0.044 ns 0.023 0.026
17 17
13 12
3.1 3.3
(1.4–6.7) (1.5–7.4)
0.005 0.003
95%CI 95% confidence interval, ns not significant a According to univariate logistic regression b All variables with lower values were taken as a reference group
surgical procedures (hysterectomy, cholecystectomy, prostatectomy and inguinal hernia surgery) (OR=1.9, p=0.026), and diabetes mellitus (OR=2.7, p=0.044) were also significantly more frequently reported in MDS patients. There was no difference regarding occurrence of allergic diseases (hay fever, allergic rhinitis, conjunctivitis, atopic dermatitis, urticaria, reactions to insect stings, and allergies to food and drugs). Data related to birth history showed no differences between MDS patients and controls (Table 2), although a higher birth order and more than four children in the family were associated with a higher risk of MDS. Investigation of the family history included the presence of chronic diseases (Table 2); there was a significantly higher frequency of different types of cancer (OR=3.3, p=0.003) and diabetes mellitus (OR=3.1, p=0.005) among first degree relatives of MDS patients compared with controls. An analysis of smoking habits of MDS patients and controls are presented in Table 3. We found almost equal number of smokers among patients with MDS and their controls (OR=1.1, 95%CI 0.6–1.9), and there were no differences between the MDS cases and controls regarding the number of cigarettes smoked per day or the duration of smoking. Consumption of alcohol was a significant risk for MDS (OR=3.0, 95%CI 1.7–5.3). This was especially the case for harder liquor (spirits) in comparison with consumption of wine and beer (OR=11.8, 95%CI 3.8– 36.4). Coffee intake was more frequently reported by MDS patients than controls (OR=4.8, 95%CI 2.2–10.8), with a significant positive association with a daily consumption of four cups of coffee or more (OR=7.5, 95%CI 1.9–34.6). An analysis of occupational exposure (Table 4) showed that MDS cases were significantly more often exposed to X-ray radiation (OR=10.6, 95%CI 1.2–92.3, p=0.033), different types of chemicals (listed in Table 4; OR=12.5, 95%CI 4.5–34.5, p=0.001), and insecticides, pesticides, and herbicides (OR=7.3, 95%CI 2.9–18.1, p=0.001) than were the controls; it is important to emphasize that some of the patients were exposed to more than one agent. Dietary habits of MDS patients and their controls are presented in Table 5. The risk of MDS was statistically significant in those consuming eggs seven or more times per week (OR=2.8, 95%CI 1.2–6.3, p=0.005). All variables related to MDS at a significant level of p≤0.10 using univariate logistic regression analysis were included in the model of multivariate logistic regression analysis. According to multivariate analysis, the following factors were significantly related to MDS: an exposure to chemicals (OR=10.8, 95%CI 3.2–36.2, p=0.0001), viral upper respiratory tract infections (twice a year or more; OR=5.8, 95%CI 2.5–13.6, p=0.0001), exposure to insecticides, pesticides, and herbicides (OR=5.2, 95%CI 1.8– 15.1, p=0.001), consumption of coffee (OR=5.1, 95%CI 1.9–13.7, p=0.001), and alcohol (OR=2.2, 95%CI 1.1–4.6, p=0.033).
517 Table 3 Personal habits of the MDS patients and the controls Patients (No.=80) Smoking Duration of smoking (>25 years)b No. cigarettes per day (>20)b Alcohol consumption Type of alcoholic beverages (spirits vs wine and beer) Frequency of consumption (weekly vs rarely) Coffee consumption Duration of coffee consumption (>30 years)b No. cups per day (>4)b
37 19 9 37 26 25 72 30 13
Controls (No.=160) 70 36 14 36 6 20 104 50 3
Odds ratio 1.1 1.0 1.3 3.0 11.8 1.7 4.8 0.8 7.5
pa
95%CI (0.6–1.9) (0.4–2.3) (0.4–3.6) (1.7–5.3) (3.8–36.4) (0.6–4.3) (2.2–10.8) (0.4–1.5) (1.9–34.6)
ns ns ns 0.001 0.001 ns 0.001 ns 0.002
95%CI 95% confidence interval, ns not significant a According to univariate logistic regression b All variables with lower values were taken as a reference group
Discussion This case-control study of patients with MDS has confirmed many statistically significant and possibly important risk factors for the disease. These include some that have been previously documented (such as an exposure to chemical agents, insecticides, pesticides and herbicides, and alcohol consumption) and some new observations (of frequent viral upper respiratory tract infections and coffee consumption). There have been numerous investigations regarding the association between chemical exposure and MDS [10, 14, 15, 18, 25]; however, we have now calculated that subjects exposed to different chemical agents have more than ten times greater risk of developing MDS compared to those not exposed (multivariate analysis). Although an exposure to a wide variety of chemical agents made it impossible for us to investigate their individual roles in MDS, analysis showed that MDS patients had been exposed to different Table 4 Exposure to exogenous agents by the MDS cases and their controls Occupational exposure Exposure to radiation Exposure to chemicalsb Exposure to insecticide, pesticide, and herbicide
Patients Controls Odds 95%CI (No.=80) (No.=160) ratio
pa
5
1
10.6
(1.2–92.3)
0.033
23
5
2.5
(4.5–34.5)
0.001
20
7
7.3
(2.9–18.1)0 0.001
95%CI 95% confidence interval, ns not significant According to univariate logistic regression b Case group petrol and diesel liquids and vapors (six cases), organic solvents (five cases), coal tar (four cases), glues (three cases), textile dyes (three cases), organic acids (two cases), fertilizers (two cases), inorganic solvents (two cases), vinyl chloride (one case), dry cleaning fumes (one case), metal alcohol (one case), explosives (one case), steel wool (one case); Control group glues (two cases), petrol (one case), explosives (one case), polishes and dyes (one case) a
agents such as petrol and diesel liquids/vapors, organic and inorganic solvents, and many others (Table 4). All these agents are documented risk factors for MDS [7–9, 14, 15, 20]. This study also confirmed that exposure to insecticides, pesticides, and herbicides (OR=5.2, 95%CI 1.8– 15.1) was significantly associated with MDS [8, 15, 20]. MDS patients were more frequently farmers, housewives, and manual workers (OR=1.3, 95%CI 0.7–2.2), with lower levels of education (OR=1.6, 95%CI 0.9–2.9), and residing in a rural environment (OR=2.3, 95%CI 1.1– 4.7). Nisse et al. [15] also found an excess of MDS patients amongst agricultural workers and an association between MDS and gardening, contrary to West et al. [25] and Rigolin et al. [20]. Although specific agricultural factors were identified, it is probable that exposure to insecticides, pesticides, herbicides, and fertilizers encountered in Table 5 Consumption frequency of different food group and food items by the MDS patients and the controls Food group and food items Meat—seven times per week Nitrate-treated meat—once a week Meat productsb— once a week Milk—seven times per week Eggs—seven times per week Vegetables—seven times per week Fruit—seven times per week
Patients (No.=80)
Control (No.=160)
Odds ratio
34
59
1.3
12
20
1.6
11
23
1.0
45
94
0.9
18
15
2.8
57
128
0.6
74
152
0.8
95% CI
pa
(0.7– ns 2.3) (0.7– ns 3.6) (0.4– 2.2) (0.5– 1.6) (1.2– 6.3) (0.3– 1.3) (0.3– 2.5)
ns ns 0.005 ns ns
All those whose consumption frequency was lower were taken as a reference group 95%CI 95% confidence interval, ns not significant a According to univariate logistic regression b Ham, sausage, hot dog, meat paste
518
agricultural work is the cause of this increased risk of MDS. MDS patients significantly more frequently reported occupational exposure to X-ray radiation (OR=10.6, 95% CI 1.2–92.3) than controls. In a case-control study of MDS carried out in UK [25], relationships were found between occupational and environmental radiation and myelodysplasia. These findings were also compatible with the wellrecognized association between radiation and leukemia among atom bomb survivors, after radiotherapy for nonmalignant diseases, and after fetal irradiation, discussed by West et al. [25]. There was a significant increased risk of MDS in patients suffering frequent viral upper respiratory tract infections (twice or more per year; OR=5.8, 95%CI 2.5–13.6). Similar results were also obtained by univariate analysis for frequent streptococcal infections. Although relationships between frequently occurring infections and susceptibility for MDS are unclear, there is some evidence that chronic immune stimulation may be important in the etiology of primary MDS in adults [6, 17]. In addition, infectious agents have also been hypothesized to be related to the genesis of leukemia, and some viruses are well-known to induce leukemia [23]. We found that MDS patients suffered significantly more frequently from diabetes mellitus (OR=2.7, p=0.044) and were more likely to have undergone surgical procedures such as appendectomy (OR=2.2, p=0.023), hysterectomy, cholecystectomy, prostatectomy, and inguinal hernia surgery (OR=1.9, p=0.026). There was no difference with regard to the occurrence of allergic diseases when compared to the controls. Such associations have not been previously reported, though Dalamaga et al. [4] showed that autoimmune diseases (but not allergic diseases) tended to increase the risk of MDS, irrespective of the timing of their occurrence (before MDS). Having first degree relatives with hematological malignancy have been previously associated with an increased risk of MDS [15, 25]. Our observations additionally found a significant increased risk of MDS in the presence of any cancer (OR=3.3, p=0.003) or diabetes mellitus (OR=3.1, p=0.005) in the family. Contrary to the literature [25], our MDS patients came from families with a high number of children and had a higher birth order. This study also looked at behavioral risk factors such as smoking and the consumption of alcohol and coffee; both alcohol and coffee consumption were independent risk factors in the multivariate model (alcohol: OR=2.2, 95%CI 1.1–4.6; coffee: OR=5.1, 95%CI 1.9–13.7) Other studies have previously investigated the association with alcohol and have published conflicting results [2, 15, 25]. A casecontrol study undertaken in Greece registered a significant difference (p=0.001) [4]. Ido et al. [9] also found a positive association with a dose–response relationship, whereas Nagata et al. [14] found no significant association with any type of liquor. Alcohol may have a direct toxic affect on the bone marrow [3], and experimental studies have demon-
strated that alcohol exposure may induce chromosomal aberrations in hematopoietic cells [5, 16]. No previous studies have implicated coffee consumption in the etiology of MDS; thus, our positive findings are unique and should be confirmed by a further larger study group. An analysis of smoking habits among the MDS patients and their controls showed that there was an almost equal number of smokers in both groups (OR=1.1, 95%CI 0.6– 1.9), and there were no differences between the cases and controls regarding the number of cigarettes smoked per day and duration of smoking. This is in accordance with the results of Nagata et al. [14], Ido et al. [9], and Pasqualetti et al. [18], and contrary to those of Nisse et al. [15], Bjork et al. [2], and Dalamaga et al. [4], who showed positive and dose-dependent relationships between tobacco smoking and MDS. MDS patients consumed more meat, nitrate-treated meat, and eggs, than did the controls, with frequent egg consumption being a significant risk factor (OR=2.8, p=0.005). This may implicate animal fats as a dietary factor related to MDS, with fruit and vegetable consumption conversely protecting against it as suggested in many hematological malignancies [21, 22, 24]. One can criticize that recall bias may have influenced the results in this case-control study, that the study group was small, and that the use of healthy controls may have been more appropriate. However, due to the rarity of the disease investigated, the sample sizes were as large as possible and our results were comparable to the larger published series, with longer durations of observation and locations [14, 15, 25]. Our study may be also biased with biological heterogeneity of MDS itself. Possible influence of some factors on the type of MDS was not accessible in our study due to rather small numbers of patients in the preleucemic phase (RAEB-1, RAEB-2, RAEB-t; 35/80 patients) of the disease. We believe that all analyzed factors have similar influence on all MDS group; therefore, we have conducted our investigation in MDS group as a whole. In conclusion, the findings obtained in the Belgrade study support the hypotheses that an exposure to chemical agents, pesticides, insecticides, and herbicides, and certain lifestyle factors, such as diet, alcohol and coffee consumption, as well as frequent viral infections, might be involved in the etiology of MDS. Further studies should be undertaken to assess these specific risk factors in more detail. Acknowledgements This work was financially supported by the Ministry of Science and Environment Protection of Republic of Serbia through contract no. 1927/ 2002–2005. We would very much like to thank Dr H.D.E. Atkinson, Imperial College Medical School, Hammersmith Hospital, Du Cane Road, London, UK, for his critical review and assistance in the writing of this paper in English.
519
References 1. Bennett M, Catovsky D, Daniel MT et al (1982) Proposals for the classification of MDS. Br J Haematol 51:189–199 2. Bjork J, Albin M, Mauritzson N, Stromberg U, Johansson B, Hagmar L (2000) Smoking and myelodysplastic syndromes. Epidemiology 11:285–291 3. Casagrande G, Michot F (1989) Alcohol-induced bone marrow damage; status before and after a 4-week period of abstinence from alcohol with or without disulfram. A randomized bone marrow study in alcohol-dependent individuals. Blut 59:231– 236 4. Dalamaga M, Petridou E, Cook FE, Trichopoulos D (2002) Risk factors for myelodysplastic syndromes: a case-control study in Greece. Cancer Causes Control 13:603–608 5. Datta R, Sherman ML, Kufe DW (1990) Regulation of protooncogene and tumor necrosis factor gene expression by ethanol in HL-60 myeloid leukemia cells. Blood 76:298–301 6. Enright H, Miller W (1997) Autoimmune phenomena in patients with myelodysplastic syndromes. Leuk Lymph 24:483–489 7. Farrow A, Jacobs A, West RR (1989) Myelodysplasia, chemical exposure and other environmental factors. Leukemia 3:33–35 8. Goldberg H, Lusk E, Moore J, Nowell PC, Besa EC (1990) Survey of exposure to genotoxic agents in primary myelodysplastic syndrome—correlation with chromosome patterns and data in patients without haematological disease. Cancer Res 50:6876–6881 9. Ido M, Nagata C, Kawakami N et al (1996) A case-control study of myelodysplastic syndromes among Japanese men and women. Leuk Res 20:727–731 10. Levine EG, Bloomfield CD (1992) Leukemia and myelodysplastic syndromes secondary to drugs, radiation and environmental exposure. Semin Oncol 19:47–84 11. List AF, Doll DF (1999) The myelodysplastic syndromes. In: Foerster GR, Lukens LJ, Paraskevas F, Greer JP, Rodgers GM (eds) Wintrobe’s clinical hematology. Williams & Willkins, Baltimore, MD, pp 2320–2333 12. McNeil D (1996) Epidemiological research methods. Wiley, New York
13. Miettinen OS (1974) Simple interval estimations of risk ratio. Am J Epidemiol 100:515–516 14. Nagata C, Shimizu H, Hirashima K et al (1999) Hair dye use and occupational exposure to organic solvents as risk factors for myelodysplastic syndrome. Leuk Res 23:57–62 15. Nisse C, Haguenoer JM, Grandbastein B et al (2001) Occupational and environmental risk factors of the myelodysplastic syndromes in the North of France. Br J Haematol 112:927–935 16. Obe G, Gobel D, Engeln H, Herha J, Natarajan AT (1980) Chromosomal aberrations in peripheral lymphocytes of alcoholics. Mutat Res 73:229–259 17. Okamoto T, Okada M, Mori A (1997) Correlation between immunological abnormalities and prognosis in myelodysplastic syndrome patients. Int J Hematol 66:345–351 18. Pasqualetti P, Casale R, Burmeister LF et al (1991) Occupational risk for hematological malignancies. Am J Hematol 38:147–152 19. Pedersen-Bjergaard J (1992) Radiotherapy and chemotherapy induced myelodysplasia and acute myeloid leukaemia. A review. Leuk Res 16:61–65 20. Rigolin GM, Cuneo A, Roberti MG, Bardi A, Bigoni R, Piva N, Monotto C, Agostini P, De Angeli C, Del Senno L, Spanedda R, Castoldi G (1998) Exposure to myelotoxic agents and myelodysplasia: case-control study and correlation with clinicobiological findings. Br J Haematol 103:189–197 21. Tavani A, Pregnolato A, Negri E et al (1997) Diet and risk of lymphoid neoplasms and soft tissue sarcomas. Nutr Cancer 27:256–260 22. Vlajinac H, Pekmezovic T, Adanja B, Marinkovic J, Kanazir M, Suvajdzic N, Colovic M (2003) Case-control study of multiple myeloma with special reference to diet as risk factor. Neoplasma 50:79–83 23. Vogel VG, Fisher RE (1993) Epidemiology and etiology of leukemia. Curr Opin Oncol 5:26–34 24. Ward MH, Zahm SH, Weisenburger DD et al (1994) Dietary factors and non-Hodgkin’s lymphoma in Nebraska (United States). Cancer Causes Control 5:422–432 25. West RR, Stafford DA, Farrow A, Jacobs A (1995) Occupational and environmental exposures and myelodysplasia: a case-control study. Leuk Res 19:127–139