Search tips
Search criteria 


Logo of amjepidLink to Publisher's site
Am J Epidemiol. 2009 June 15; 169(12): 1492–1499.
Published online 2009 April 24. doi:  10.1093/aje/kwp074
PMCID: PMC2727203

Obesity, Lifestyle Factors, and Risk of Myelodysplastic Syndromes in a Large US Cohort


The etiology of myelodysplastic syndromes (MDS) is not well understood. The authors examined the relations of obesity and lifestyle factors to MDS in a cohort of 471,799 persons aged 50–71 years who were recruited into the National Institutes of Health-AARP Diet and Health Study, a large US prospective study, in 1995–1996. Incident MDS was diagnosed in 193 persons during 2001–2003. A significant positive association was observed between body mass index (BMI; weight (kg)/height (m)2) at baseline and MDS. Compared with persons with a BMI less than 25.0, the hazard ratios for persons with BMIs of 25.0–<30.0 and ≥30.0 were 1.15 (95% confidence interval (CI): 0.81, 1.64) and 2.18 (95% CI: 1.51, 3.17; P for trend < 0.001), respectively. The association was not affected by physical activity, cigarette smoking, or alcohol intake. As reported in previous studies, the risk of MDS was elevated among former smokers (hazard ratio = 1.68, 95% CI: 1.17, 2.41) and current smokers (hazard ratio = 3.17, 95% CI: 2.02, 4.98) as compared with never smokers. Physical activity, alcohol consumption, meat intake, and fruit and vegetable intake did not appear to significantly influence the risk of MDS in this analysis. This prospective investigation of MDS implicates both obesity and smoking as modifiable risk factors.

Keywords: life style, myelodysplastic syndromes, obesity, smoking

Myelodysplastic syndromes (MDS), a group of clonal disorders of the hematopoietic stem cell, are characterized by ineffective hematopoiesis and frequent (~30%) progression to acute myeloid leukemia (1). MDS are most common among the elderly, and the 3-year survival rate is only 35% (2). Data from the Surveillance, Epidemiology, and End Results (SEER) Program indicate that the risk of MDS increases with age, and approximately 86% of MDS cases are diagnosed in persons aged 60 years or more (median age at diagnosis, 76 years) (2). Men have a significantly higher incidence rate than women (4.5 per 100,000 person-years vs. 2.7 per 100,000 person-years). Among racial groups, whites have the highest incidence rate. In 2003, approximately 10,300 incident cases of MDS were diagnosed in the United States (2).

The etiology of MDS is not well understood, although roles for chemotherapy and/or radiation therapy, certain genetic disorders, smoking, and exposure to solvents and pesticides have been indicated (3). Previous studies have all been of the case-control design and therefore may not have been ideal for investigation of lifestyle factors potentially affected by the disease and its prevalent symptoms (e.g., fatigue and possible weight change). Although adiposity has been linked to an increased risk of several cancers, including hematologic malignancies such as leukemia (4), non-Hodgkin lymphoma (5), and multiple myeloma (6, 7), an association with MDS has not previously been reported.

Our aim in the current analysis was to explore the relations of obesity and other lifestyle factors, including smoking, physical activity, alcohol consumption, and diet, to MDS in a prospective cohort study. The National Institutes of Health (NIH)-AARP Diet and Health Study is a collaboration between the NIH and AARP (formerly the American Association of Retired Persons) that involves over 560,000 men and women across the United States. The large size of the cohort and the older age of the participants provided us with a unique opportunity to evaluate the etiology of MDS prospectively.


Study population

The NIH-AARP Diet and Health Study has been previously described in detail (8). Briefly, the study was established in 1995–1996 by recruiting AARP members aged 50–71 years who resided in 1 of 6 US states (California, Florida, Louisiana, New Jersey, North Carolina, and Pennsylvania) or 2 metropolitan areas (Atlanta, Georgia, and Detroit, Michigan). A self-administered questionnaire eliciting information on demographic characteristics, dietary intake, and health-related behaviors was mailed to participants. Out of 617,119 questionnaires returned, a total of 567,169 (92%) subjects satisfactorily completed the questionnaire and agreed to participate in the study (8). The study was approved by the Special Studies Institutional Review Board of the National Cancer Institute.

For the current analysis, we excluded subjects who had duplicate representation (n = 179), moved out of 1 of the 8 study areas before returning the baseline questionnaire (n = 321), died before study entry (n = 261), withdrew (n = 6), had the questionnaire completed by a proxy respondent (n = 15,760), had a history of cancer before study entry (n = 51,205), were identified as having cancer through death reports only (n = 3,890), or had extreme values for energy intake (more than 3 interquartile ranges outside the 75th and 25th percentiles of log-transformed energy intake; n = 4,181). We further excluded participants with missing height or weight data (n = 12,051) or extreme values for BMI (more than 3 interquartile ranges outside the 75th and 25th percentiles of log-transformed BMI; n = 3,872). A total of 3,644 persons with a BMI less than 18.5 were also excluded out of concern about disease-related weight loss, similarly to what was done in a previous analysis of data from the NIH-AARP Diet and Health Study (9). Our final analytic cohort consisted of 471,799 persons (283,604 men and 188,195 women; 83% of the 567,169 subjects who satisfactorily completed the baseline questionnaire).

Ascertainment of cases

In the NIH-AARP Diet and Health Study, vital status was ascertained through linkage of the cohort to the Social Security Administration Death Master File, follow-up searches of the National Death Index Plus for participants who were matched to the Death Master File, cancer registry linkage, questionnaire responses, and responses to other mailings. Follow-up time extended from study baseline (1995–1996) to the date of death (ascertained by the National Death Index or the Death Master File), diagnosis of incident MDS, relocation out of the registry ascertainment area, or December 31, 2003, whichever was earliest. Of all the subjects included in this analysis, 4.4% relocated out of the registry ascertainment area.

In the NIH-AARP cohort, incident cases were identified through linkage with state cancer registry databases. A validation study showed that approximately 90% of all incident cancer cases in the NIH-AARP cohort were identified using linkage to cancer registries (10). Having long been viewed as preleukemic disorders, MDS are now considered malignant due to their clonal nature (International Classification of Diseases for Oncology, Third Edition, behavior code change from “1” (uncertain whether benign or malignant) to “3” (malignant) (11)) and have been reportable to the SEER Program since 2001. Eight International Classification of Diseases for Oncology, Third Edition, codes are used to identify MDS in the SEER Program: 1) 9980—refractory anemia; 2) 9982—refractory anemia with ringed sideroblasts; 3) 9983—refractory anemia with excess blasts (RAEB) (including RAEB under the French-American-British classification (12) and both RAEB-1 and RAEB-2 under the World Health Organization recommendation (13, 14)); 4) 9984—RAEB in transformation (RAEB-t); 5) 9985—refractory cytopenia with multilineage dysplasia; 6) 9986—MDS associated with 5q deletion; 7) 9987—therapy-related MDS; and 8) 9989—MDS not otherwise specified. Linkage with registry files from 2001–2003 resulted in a total of 193 incident cases of MDS (137 men, 56 women). Since the morphologic feature of RAEB-t is considered similar to that of acute myeloid leukemia (14), we conducted analyses with and without the inclusion of RAEB-t cases (n = 4). None of the MDS cases were identified through death reports only; that is, among the 3,890 persons who were excluded from the current analysis because they were identified as having cancer through death reports only, none had MDS.

Data collection and statistical analysis

In this analysis, we were mainly interested in obesity and other lifestyle factors, including cigarette smoking, physical activity, alcohol consumption, and dietary intake (major food groups). We also included demographic variables (i.e., age, sex, race/ethnicity, and education) as potential confounders. From the baseline questionnaire responses on self-reported weight and height, BMI was calculated as weight in kilograms divided by height in meters squared. We grouped BMI into 3 categories—18.5–<25, 25–<30, and ≥30—which are consistent with the definitions of normal weight, overweight, and obesity proposed by the World Health Organization (15). To further assess the role of adiposity, we also grouped BMI into 6 more refined categories (18.5–<25, 25–<27.5, 27.5–<30, 30–<32.5, 32.5–<35, and ≥35). We constructed a categorical smoking variable based on smoking status and usual dose: never smokers, former smokers who had smoked 1 pack/day or less, former smokers who had smoked more than 1 pack/day, current smokers who were smoking 1 pack/day or less, and current smokers who were smoking more than 1 pack/day. Participants were also categorized as never smokers versus ever smokers (those who had smoked 100 or more cigarettes during their entire lives as of baseline), as well as never, former, and current smokers. Additional variables of interest included vigorous physical activity (activity that lasted 20 minutes or more and caused either increases in breathing or heart rate or working up a sweat; ≤3 times per month, 1–2 times per week, or ≥3 times per week), alcohol consumption (persons who never drank vs. those who drank, in tertiles (drinks/day)), total meat intake (g/day), and fruit and vegetable intake, excluding white potatoes (servings/day). Intakes of meat, fruit, and vegetables were adjusted for total energy intake (calories/day) with the density method (16) and then classified into tertiles based on the distributions in the entire cohort.

To maintain a relatively large sample size, we grouped participants who had missing values for smoking, physical activity, or educational level in a separate category rather than excluding them. To evaluate the impact of this analytical strategy, we conducted sensitivity analysis by excluding participants who had missing values for any of these variables and comparing the results with those derived from analyses including participants with missing values.

Multivariable-adjusted hazard ratios and 95% confidence intervals were estimated using Cox proportional hazards regression with follow-up time as the underlying time metric. Using age as the underlying time metric in models led to essentially the same results. We tested for the proportional hazards assumption by plotting Schoenfeld residuals versus time, and the patterns appeared random, suggesting no deviations from the assumption. All multivariate models included adjustment for sex, age at baseline (continuous), race/ethnicity, education, and total energy intake. An additional model included BMI, smoking, and physical activity simultaneously. Tests for trend were performed by modeling categorical variables as ordinal variables. To examine possible effect modification of the BMI associations by sex, smoking, and physical activity, we conducted stratified analyses and compared models with and without interaction terms using likelihood ratio tests. To obtain a graphic illustration of the association between BMI and MDS, we used natural cubic splines, placing knots at the 2nd, 25th, 75th, and 98th percentiles of BMI, centering BMI at 21 and adjusting for covariates. All analyses were performed with SAS, version 9.1 (SAS Institute Inc., Cary, North Carolina). An α level of 0.05 was considered statistically significant, and all tests were 2-sided.


Baseline characteristics of the study participants are shown in Table 1. Among the 471,799 participants, 35% had normal weight (BMI 18.5–<25), whereas 43% were overweight (BMI 25–<30) and 22% were obese (BMI ≥30). Heights were similar across different BMI levels. Obese participants were more likely to be male, less educated, and physically inactive, to not be current smokers, and to have higher energy and meat intakes and lower alcohol consumption.

Table 1.
Characteristics of Participants in the National Institutes of Health-AARP Diet and Health Study by Body Mass Index, United States, 1995–2003

A more than 2-fold increased risk of MDS was observed among obese participants (hazard ratio (HR) = 2.18, 95% confidence interval (CI): 1.51, 3.17), and there was a significant positive trend for the relation between BMI and MDS when the 3 categories when modeled as an ordinal variable (P for trend < 0.001). Classifying BMI into 6 narrower categories or modeling BMI as a continuous variable generated further support for an association between BMI and MDS (Table 2). The hazard ratio for persons whose BMIs were in the range of 25–<27.5 was hardly elevated. Compared with persons who had never smoked, both former smokers (HR = 1.68, 95% CI: 1.17, 2.41) and current smokers (HR = 3.17, 95% CI: 2.02, 4.98) had significantly elevated risks of MDS, with the highest risk being observed among current smokers who smoked more than 1 pack of cigarettes each day (HR = 4.70, 95% CI: 2.68, 8.24). When compared with physically inactive persons (vigorous physical activity ≤3 times per month), participants who reported engaging in vigorous physical activity at least 3 times per week had a significantly reduced risk (HR = 0.68, 95% CI: 0.49, 0.95). Neither alcohol consumption nor fruit and vegetable intake nor total meat intake was associated with MDS (Table 2).

Table 2.
Relation Between Selected Lifestyle Factors and Myelodysplastic Syndromes in the National Institutes of Health-AARP Diet and Health Study, United States, 1995–2003

We further evaluated the association between BMI and MDS by adjusting simultaneously for physical activity and smoking. The estimated hazard ratio for obesity was almost the same and still statistically significant, and the positive trend for the relation between BMI and MDS persisted. Smoking remained a significant risk factor for MDS, and the magnitude of association was also extremely similar. The protective effect for engaging in vigorous physical activity at least 3 times per week was attenuated and became nonsignificant (HR = 0.83, 95% CI: 0.59, 1.16). Analysis of BMI on a continuous basis with natural cubic splines suggested that the increased risk of MDS was mostly apparent among persons who were obese (BMI ≥30) and appeared to plateau around BMIs of 34–35 (Figure 1). A likelihood ratio test comparing the model with BMI as a continuous variable and the spline model yielded a P value of 0.16.

Figure 1.
Hazard ratio for myelodysplastic syndromes (solid line) in relation to body mass index (BMI; weight (kg)/height (m)2), National Institutes of Health-AARP Diet and Health Study, United States, 1995–2003. Relative risks are modeled on a continuous ...

Subgroup analyses in which results were stratified by sex, smoking, and physical activity showed that obesity was associated with MDS in both males and females and in both smokers and nonsmokers, as well as persons with different levels of vigorous physical activity (Table 3). Likelihood ratio tests comparing the models with and without the interaction terms involving these variables all had P values greater than 0.5 (data not shown).

Table 3.
Relation Between Body Mass Index and Risk of Myelodysplastic Syndromes, by Sex, Smoking, and Physical Activity, in the National Institutes of Health-AARP Diet and Health Study, United States, 1995–2003

Excluding cases with RAEB-t (n = 4) had no impact on the results (data not shown). When the analyses were restricted to 167 MDS cases who did not have any other cancer prior to the diagnosis of MDS, the changes in risk estimates were minimal and negligible (data not shown). Analyses excluding participants who had missing values for smoking, physical activity, or educational level generated essentially the same results (data not shown). Inclusion of participants who had been excluded because they had a BMI less than 18.5 (n = 3,644) also had no impact on the results (data not shown).


To our knowledge, this study was the first to prospectively evaluate the relation between various lifestyle factors and MDS. A more than 2-fold increased risk of MDS was observed among persons who were obese, and there was a significant trend for the relation between BMI and the risk of MDS. Subgroup analyses stratified according to sex, smoking, and physical activity did not provide evidence for possible modification of the BMI-MDS relation by those variables. Given that MDS are malignancies with a poor prognosis and a probably increasing incidence in the United States (2) and given that a substantial proportion of the US adult population is obese (17), this finding has important implications.

This study confirmed the previously reported association between cigarette smoking and MDS (1822). Our analysis showed that current smokers had a higher risk than former smokers, and heavy smokers had a higher risk than light smokers. The relation of alcohol consumption to the risk of MDS has been assessed in several studies, with conflicting results. Overall, the consumption of alcoholic beverages did not appear to affect the risk of MDS (2325), which is consistent with our finding. However, in a recent case-control study, Strom et al. (19) found that moderate consumption of wine was associated with a significantly decreased risk of MDS in both sex groups and across all French-American-British subtypes. In addition to assessing the role of alcohol consumption, we also evaluated wine intake specifically but found no significant association (data not shown). Given the relatively small numbers of cases included in the previous studies and in the current analysis, more research is needed to clarify the role of alcohol consumption in general and wine intake specifically. In this analysis, physical activity, total meat intake, and fruit and vegetable intake did not appear to significantly influence the risk of MDS. Since these factors were not evaluated in previous studies of MDS, these null findings need to be interpreted with caution.

The mechanisms by which obesity could affect the pathogenesis of MDS are unclear. One potential causal pathway operates through elevated levels of insulin and insulin-like growth factor 1. A metabolic consequence of obesity is insulin resistance followed by an increase in insulin secretion. Insulin may promote tumorigenesis directly through insulin receptors in (pre)neoplastic target cells or indirectly by increasing levels of bioavailable insulin-like growth factor 1 (26). Insulin-like growth factor 1 is involved in hematopoiesis and is mitogenic for myeloid cells; almost all normal and neoplastic hematopoietic cells express the insulin-like growth factor receptors (27). Among older persons, weight at present and weight from 4 years prior have been reported to have a correlation of 0.92, while weight at present and weight from 28 years prior had a correlation of 0.71 (28). The effect on hematopoiesis of obesity-related disruptions in the insulin-like growth factor 1 axis could persist over many years. In addition, myeloid precursor cells in the bone marrow have leptin receptors, and leptin plays an important role in the modulation of the innate immune response, inflammation, and hematopoiesis (29). Leptin inhibits food intake and stimulates energy expenditure by acting on the appetite centers in the hypothalamic regions of the brain (30). The abnormal level of circulating leptin and/or different sensitivity to leptin among obese persons could interfere with hematopoiesis over an extended period of time.

A major strength of this study was its prospective design, with BMI and other information being recorded at baseline prior to diagnosis—an approach that essentially precludes recall bias. The large size of the cohort generated enough cases, comparable to the number included in previous case-control studies, to make prospective analysis feasible. We missed incident cases arising during the earlier follow-up (1995–2000), since the tumor registries did not start collecting information on incident MDS until 2001. A total of 193 MDS cases were diagnosed during 2001–2003, with a total of 1,284,373 person-years of follow-up. Assuming the same incidence rate, approximately 319 MDS cases might have occurred during 1995–2000. In other words, of the 471,606 non-MDS participants included in this study, 319 (0.068%) might have been misclassified. Such a misclassification, most likely nondifferential with regard to obesity, would have yielded slightly attenuated estimates. If obesity is related to the age of onset, the inability to identify cases during the earlier follow-up might have introduced a differential bias. Given the lag time between the assessment of adiposity (1995–1996) and the ascertainment of MDS (2001–2003), reverse causation is not likely in this study.

This study also had some other limitations. First, the findings are based on AARP members who were aged 50 years or older at baseline; results are not necessarily generalizable to younger persons. We note, however, that MDS occurs predominantly in older people (31). The ages of the 193 MDS cases identified ranged from 57 years to 78 years (mean = 71; median, 72). Second, as with many large epidemiologic studies, BMI in the current analysis was based on self-reported height and weight at the time of enrollment. In general, though, self-reported height and weight are quite accurate, with correlations between measured and self-reported BMIs typically being greater than 0.9 (32). It might have been useful to measure BMI at different ages or evaluate possible changes over time, but BMI at baseline, which was 5–8 years before the diagnosis of MDS, was probably relevant timewise. Third, although the number of MDS cases was sufficient for our primary analyses, risk estimates from some of the subgroup analyses were imprecise, as reflected by the wide confidence intervals presented in Table 3. Fourth, we could not pursue analyses by MDS subtype because approximately half of the cases were categorized as having an unspecified subtype (consistent with SEER data (2)) and the number of cases with identified subtypes was relatively small. Analyses focusing on MDS subtypes may become feasible with additional follow-up and case ascertainment.

Finally, since MDS did not become reportable to cancer registries until 2001 and MDS are more commonly diagnosed and managed outside of hospitals in comparison with other cancers, it is possible that some cases were not reported to cancer registries (3). If there was serious underreporting and the reasons for underreporting were related to the exposures of interest in this analysis, our findings might have been biased. On the other hand, there are multiple observations/factors that appear to alleviate the concern about possible underreporting. Our analysis of SEER data showed that the incidence of MDS in whites is consistent with what has been reported in European countries. MDS incidence based on SEER data remained relatively stable between 2001 and 2005 (the most recent year for which SEER limited-use data were available). If there had been serious underreporting when the disease had just become reportable, we might have seen a larger difference between the 2001 incidence and the 2005 incidence. Third, unlike diagnosis of some other hematologic malignancies such as myeloproliferative neoplasms, the diagnosis of MDS requires bone marrow tests, which makes it harder (though still possible, of course) for cases to be “missed.”

Our analysis suggests that obesity is significantly associated with an increased risk of MDS. Since our study is the first to have observed such an association, the finding needs to be confirmed in other well-designed studies. Further exploration of the mechanisms underlying an obesity-MDS link may provide additional insight into both the biologic effects of adiposity and the pathogenesis of hematopoietic malignancies.


Author affiliations: Division of Chronic Disease Epidemiology, Yale School of Public Health, New Haven, Connecticut (Xiaomei Ma, Susan T. Mayne, Rong Wang); Cancer Research Center of Hawaii, University of Hawaii, Honolulu, Hawaii (Unhee Lim); Nutritional Epidemiology Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, Maryland (Yikyung Park, Patricia Hartge, Arthur Schatzkin); and AARP, Washington, DC (Albert R. Hollenbeck).

This work was supported by grant K07 CA119108 from the National Cancer Institute.

Cancer incidence data from the Atlanta, Georgia, metropolitan area were collected by the Georgia Center for Cancer Statistics, Department of Epidemiology, Rollins School of Public Health, Emory University. Cancer incidence data from California were collected by the Cancer Surveillance Section, California Department of Health Services. Cancer incidence data from the Detroit, Michigan, metropolitan area were collected by the Michigan Cancer Surveillance Program, Community Health Administration, state of Michigan. The Florida cancer incidence data used in this report were collected by the Florida Cancer Data System under contract to the Department of Health. Cancer incidence data from Louisiana were collected by the Louisiana Tumor Registry, Louisiana State University Medical Center in New Orleans. Cancer incidence data from New Jersey were collected by the New Jersey State Cancer Registry, Cancer Epidemiology Services, New Jersey State Department of Health and Senior Services. Cancer incidence data from North Carolina were collected by the North Carolina Central Cancer Registry. Cancer incidence data from Pennsylvania were supplied by the Division of Health Statistics and Research, Pennsylvania Department of Health, Harrisburg, Pennsylvania. Cancer incidence data from Arizona were collected by the Arizona Cancer Registry, Division of Public Health Services, Arizona Department of Health Services. Cancer incidence data from Texas were collected by the Texas Cancer Registry, Cancer Epidemiology and Surveillance Branch, Texas Department of State Health Services. Cancer incidence data from Nevada were collected by the Nevada Central Cancer Registry, Center for Health Data and Research, Bureau of Health Planning and Statistics, State Health Division, Nevada Department of Health and Human Services.

The views expressed herein are solely those of the authors and do not necessarily reflect those of the Florida Department of Health or its contractor. The Pennsylvania Department of Health specifically disclaims responsibility for any analyses, interpretations, or conclusions.

Conflict of interest: none declared.



body mass index
confidence interval
hazard ratio
myelodysplastic syndromes
National Institutes of Health
refractory anemia with excess blasts
Surveillance, Epidemiology, and End Results


1. Corey SJ, Minden MD, Barber DL, et al. Myelodysplastic syndromes: the complexity of stem-cell diseases. Nat Rev Cancer. 2007;7(2):118–129. [PubMed]
2. Ma X, Does M, Raza A, et al. Myelodysplastic syndromes: incidence and survival in the United States. Cancer. 2007;109(8):1536–1542. [PubMed]
3. Rollison DE, Howlader N, Smith MT, et al. Epidemiology of myelodysplastic syndromes and chronic myeloproliferative disorders in the United States, 2001–2004, using data from the NAACCR and SEER programs. Blood. 2008;112(1):45–52. [PubMed]
4. Larsson SC, Wolk A. Overweight and obesity and incidence of leukemia: a meta-analysis of cohort studies. Int J Cancer. 2008;122(6):1418–1421. [PubMed]
5. Larsson SC, Wolk A. Obesity and risk of non-Hodgkin's lymphoma: a meta-analysis. Int J Cancer. 2007;121(7):1564–1570. [PubMed]
6. Reeves GK, Pirie K, Beral V, et al. Cancer incidence and mortality in relation to body mass index in the Million Women Study: cohort study. BMJ. 2007;335(7630):1134. [PMC free article] [PubMed]
7. Pan SY, Johnson KC, Ugnat AM, et al. Association of obesity and cancer risk in Canada. Am J Epidemiol. 2004;159(3):259–268. [PubMed]
8. Schatzkin A, Subar AF, Thompson FE, et al. Design and serendipity in establishing a large cohort with wide dietary intake distributions: The National Institutes of Health-American Association of Retired Persons Diet and Health Study. Am J Epidemiol. 2001;154(12):1119–1125. [PubMed]
9. Adams KF, Leitzmann MF, Albanes D, et al. Body mass and colorectal cancer risk in the NIH-AARP cohort. Am J Epidemiol. 2007;166(1):36–45. [PubMed]
10. Michaud DS, Midthune D, Hermansen S, et al. Comparison of cancer registry case ascertainment with SEER estimates and self-reporting in a subset of the NIH-AARP Diet and Health Study. J Regist Manage. 2005;32:70–75.
11. Fritz A, Percy C, Jack A, et al. International Classification of Diseases for Oncology (ICD-O) Third Edition. Geneva, Switzerland: World Health Organization; 2000.
12. Bennett JM, Catovsky D, Daniel MT, et al. Proposals for the classification of the myelodysplastic syndromes. Br J Haematol. 1982;51(2):189–199. [PubMed]
13. Harris NL, Jaffe ES, Diebold J, et al. World Health Organization classification of neoplastic diseases of the hematopoietic and lymphoid tissues: report of the Clinical Advisory Committee meeting—Airlie House, Virginia, November 1997. J Clin Oncol. 1999;17(12):3835–3849. [PubMed]
14. Vardiman JW, Harris NL, Brunning RD. The World Health Organization (WHO) classification of the myeloid neoplasms. Blood. 2002;100(7):2292–2302. [PubMed]
15. World Health Organization. Physical Status: The Use and Interpretation of Anthropometry. Report of a WHO Expert Committee. (WHO Technical Report Series no. 854) Geneva, Switzerland: World Health Organization; 1995. [PubMed]
16. Willett W. Nutritional Epidemiology. 2nd ed. New York, NY: Oxford University Press; 1998.
17. Ogden CL, Carroll MD, Curtin LR, et al. Prevalence of overweight and obesity in the United States, 1999–2004. JAMA. 2006;295(13):1549–1555. [PubMed]
18. Nisse C, Lorthois C, Dorp V, et al. Exposure to occupational and environmental factors in myelodysplastic syndromes. Preliminary results of a case-control study. Leukemia. 1995;9(4):693–699. [PubMed]
19. Strom SS, Gu Y, Gruschkus SK, et al. Risk factors of myelodysplastic syndromes: a case-control study. Leukemia. 2005;19(11):1912–1918. [PubMed]
20. Pasqualetti P, Festuccia V, Acitelli P, et al. Tobacco smoking and risk of haematological malignancies in adults: a case-control study. Br J Haematol. 1997;97(3):659–662. [PubMed]
21. Nisse C, Haguenoer JM, Grandbastien B, et al. Occupational and environmental risk factors of the myelodysplastic syndromes in the North of France. Br J Haematol. 2001;112(4):927–935. [PubMed]
22. Björk J, Albin M, Mauritzson N, et al. Smoking and myelodysplastic syndromes. Epidemiology. 2000;11(3):285–291. [PubMed]
23. Ido M, Nagata C, Kawakami N, et al. A case-control study of myelodysplastic syndromes among Japanese men and women. Leuk Res. 1996;20(9):727–731. [PubMed]
24. Dalamaga M, Petridou E, Cook FE, et al. Risk factors for myelodysplastic syndromes: a case-control study in Greece. Cancer Causes Control. 2002;13(7):603–608. [PubMed]
25. Miller KB. Myelodysplastic syndromes. In: Wiernik PH, Goldman JM, Dutcher JP, et al., editors. Neoplastic Diseases of the Blood. Cambridge, United Kingdom: Cambridge University Press; 2003. pp. 395–414.
26. Calle EE, Kaaks R. Overweight, obesity and cancer: epidemiological evidence and proposed mechanisms. Nat Rev Cancer. 2004;4(8):579–591. [PubMed]
27. Shimon I, Shpilberg O. The insulin-like growth factor system in regulation of normal and malignant hematopoiesis. Leuk Res. 1995;19(4):233–240. [PubMed]
28. Stevens J, Keil JE, Waid LR, et al. Accuracy of current, 4-year, and 28-year self-reported body weight in an elderly population. Am J Epidemiol. 1990;132(6):1156–1163. [PubMed]
29. Fantuzzi G, Faggioni R. Leptin in the regulation of immunity, inflammation, and hematopoiesis. J Leukoc Biol. 2000;68(4):437–446. [PubMed]
30. Friedman JM, Halaas JL. Leptin and the regulation of body weight in mammals. Nature. 1998;395(6704):763–770. [PubMed]
31. Cheson BD, Bennett JM, Kantarjian H, et al. Report of an international working group to standardize response criteria for myelodysplastic syndromes. Blood. 2000;96(12):3671–3674. [PubMed]
32. McAdams MA, Van Dam RM, Hu FB. Comparison of self-reported and measured BMI as correlates of disease markers in US adults. Obesity (Silver Spring) 2007;15(1):188–196. [PubMed]

Articles from American Journal of Epidemiology are provided here courtesy of Oxford University Press