SEARCH

SEARCH BY CITATION

Keywords:

  • body mass index;
  • children;
  • acute myeloid leukemia;
  • survival;
  • toxicity

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

BACKGROUND:

The effect of body mass index (BMI) on the treatment outcomes of children with acute myeloid leukemia (AML) is unclear and needs further evaluation.

METHODS:

Children with AML (n = 314) who were enrolled in 4 consecutive St. Jude protocols were grouped according to BMI (underweight, <5th percentile; healthy weight, 5th to 85th percentile; and overweight/obese, ≥85th percentile).

RESULTS:

Twenty-five patients (8%) were underweight, 86 patients (27.4%) were overweight/obese, and 203 patients (64.6%) had healthy weight. The 5-year overall survival rate of overweight/obese patients (46.5% ± 7.3%) was lower than the rate of patients with healthy weight (67.1% ± 4.3%; P < .001); underweight patients also tended to have lower survival rates (50.6% ± 10.7%; P = .18). In a multivariable analysis that was adjusted for age, leukocyte count, French-American-British classification, and study protocols, patients with healthy weight had the best survival rate among the 3 groups (P = .01). When BMI was considered as continuous variable, patients with lower or higher BMI percentiles had worse survival (P = .03). There was no difference in the occurrence of induction failure or relapse among BMI groups, although underweight and overweight/obese patients had a significantly higher cumulative incidence of treatment-related mortality, especially because of infection (P = .009).

CONCLUSIONS:

An unhealthy BMI was associated with worse survival and more treatment-related mortality in children with AML. Meticulous supportive care with nutritional support and education, infection prophylaxis, and detailed laboratory and physical examination is required for these patients. These measures, together with pharmacokinetics-guided chemotherapy dosing, may improve outcome. Cancer 2012. © 2012 American Cancer Society.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

The use of comprehensive risk-classification algorithms, improvements in supportive care, and careful application of hematopoietic stem cell transplantation have increased survival rates for children with acute myeloid leukemia (AML) to approximately 70% in recent clinical trials.1, 2 Current risk-classification strategies are based primarily on the analysis of genetic abnormalities of leukemic blasts and the careful monitoring of each patient's early response to induction therapy, both of which are used to adjust therapy.2, 3 However, certain clinical features, such as age and body mass index (BMI), that are not typically used for risk stratification, also are associated with outcome.2, 4 For example, age >10 years was an independent, adverse prognostic feature in our AML02 trial.2 In the Children's Cancer Group (CCG) 2961 trial, underweight and overweight patients with AML were less likely to survive and were more likely to experience treatment-related mortality than middle-weight patients,4 but these results await confirmation. Thus, we examined the effects of BMI on treatment outcome of children with AML in our patient cohort.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

Patients

Patients with AML (n = 314) ages 2 to 20 years and who were enrolled in 4 consecutive St. Jude AML protocols (AML87,5 AML91,6 AML97,7 and AML022) were included in this analysis. Patients with acute promyelocytic leukemia (APL) (enrolled in only AML87), mixed phenotype acute leukemia (enrolled in only AML97 and AML02), and secondary AML were excluded. The chemotherapy dose was not modified on the basis of BMI. The protocols were approved by the institutional review boards at participating hospitals, and written informed consent was obtained from all patients or their guardians or parents.

Body Mass Index Classification

The BMI at diagnosis for each of the 314 eligible patients was calculated by using the macro for SAS (SAS Institute, Inc., Cary, NC) provided by the Centers for Disease Control and Prevention (CDC) (Atlanta, Ga: available at: http://www.cdc.gov/nccdphp/dnpao/growthcharts/resources/sas.htm; [Accessed October 15, 2011]). The patients initially were classified into 4 groups based on CDC weight status and BMI percentiles for children and teens (available at: http://www.cdc.gov/healthyweight/assessing/bmi/childrens_bmi/about_childrens_bmi.html#What; [Accessed October 15, 2011]) as follows: underweight, <5th percentile; healthy weight, 5th to 85th percentile; overweight, 85th to 95th percentile; and obese, ≥95th percentile.

Statistical Analyses

The association of BMI category with other categorical variables was explored by using a Monte-Carlo approximation to the exact chi-square test; this approximation was based on 10,000 permutations. The association of BMI category with continuous variables was explored by using the Kruskal-Wallis test.8

Overall survival (OS) was measured from time of study enrollment to death, and those who remained alive at the last follow-up were censored. Event-free survival (EFS) was calculated from the time of study enrollment to induction failure, withdrawal, relapse, secondary malignancy, or death, and those who remained alive and event-free at the last follow-up also were censored. The time to first occurrence of infection was calculated from the time of study enrollment to the first occurrence of a grade ≥2 infection during chemotherapy. Withdrawal, death, and relapse were treated as competing events. Those without events were censored sequentially at the end of the last course, when off therapy, or when off study.

The Kaplan-Meier method9 was used to estimate the probability of OS and EFS, and standard errors were determined by using the method described by of Peto et al.10 Survival comparisons were performed by using the Mantel-Haenszel log-rank test, and significance was determined by 10,000 permutations.11 The log-rank test was stratified by study protocol for 2-group comparisons. In multivariable analysis, OS and EFS were modeled by using a Cox proportional hazards model with age, leukocyte count, French-American-British (FAB) classification, and weight status as class predictor variables, stratified by study protocol.12 The Gray method was used to estimate and compare the cumulative incidence of infection, induction failure and relapse, and death during complete remission.13 OS and EFS also were modeled in a Cox proportional hazards regression analysis by using BMI percentile as a continuous variable in a spline model that was stratified by study protocol.14 In multivariable analysis, OS and EFS were modeled in a Cox proportional hazards regression analysis with age, leukocyte count, and FAB classification as class predictor variables and BMI percentile as a continuous variable in a spline model, all stratified by study protocol. The degrees of freedom were selected automatically by Akaike Information Criteria. All analyses were performed using SAS software for Windows (version 9.2; SAS Institute Inc) and StatXact Windows version 7.1 (Cytel Corporation, Cambridge, Mass).

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

Patient Demographics

Among the 314 patients, 25 patients (8%) were classified as underweight, 203 (64.6%) were classified as healthy weight, 37 (11.8%) were classified as overweight, and 49 (15.6%) were classified as obese. Because there was no significant difference in all analyses between overweight and obese patients, they were grouped into 1 BMI category: ie, overweight/obese (86 patients; 27.4%) (Table 1). The prevalence of the 3 BMI groups was significantly different among patients of different ages. Underweight children were younger than healthy weight children, who, in turn, were younger than overweight/obese children (P < .001). We also noted that patients enrolled in the most recent protocol had a higher BMI (P = .001).

Table 1. Demographics and Disease Characteristics of Patients in the 3 Body Mass Index Groups
 No. of Patients 
Clinical FeatureUnderweight, N = 25Healthy Weight, N = 203Overweight/Obese, N = 86P
  • Abbreviations: AML, acute myeloid leukemia; FAB, French-American-British; HSCT, hematopoietic stem cell transplantation; Inv, inversion; MDS, myelodysplastic syndrome; WBC, white blood cell.

  • a

    Kruskal-Wallis test: all other P values were calculated with exact chi-square tests.

Age, y    
 Median (range)6.7 (2.1-18.6)9.6 (2.0-19.3)13.2 (2.0-19.9)<.001
 2-9.91510321<.001
 10-201010065 
Sex    
 Male1210745.92
 Female139641 
Race    
 White1614856.66
 Black73623 
 Hispanic152 
 Other1135 
 Not available010 
WBC count, ×109/L    
 Median (range)12.2 (1.7-209)12.8 (0.01-412.2)23.7 0.03-358.1).09a
 <502015756.08
 ≥5054630 
FAB classification    
 M0132.04
 M112717 
 M286420 
 M454919 
 M533019 
 M6011 
 M7694 
 MDS170 
 Not available0134 
Cytogenetics    
 Normal67027.50
 t(8;21)44415 
 Inv(16)41711 
 t(9;11)0137 
 11q2321310 
 Other94415 
 Not available021 
Study protocol: Reference    
 AML87: Arnaout 200057226.001
 AML91: Krance 200167437 
 AML97: Rubnitz 2009744518 
 AML02: Rubnitz 2010279355 
HSCT    
 Yes88327.28
 No1712059 

Associations Between Body Mass Index and Survival

Five-year OS rates differed significantly when the 3 BMI groups were compared (P = .003) (Fig. 1A). In comparisons between healthy weight and either unhealthy BMI categories, overweight/obese patients had a significantly lower 5-year OS rate than healthy weight patients (46.5% ± 7.3% vs 67.1% ± 4.3%, respectively; P < .001); underweight patients also had a lower, but not a statistically significantly different, survival rate (50.6% ± 10.7%; P = .18). Five-year EFS rates also differed significantly among the 3 BMI categories (P = .04) (Fig. 1B): Overweight/obese patients had a significantly lower 5-year EFS rate than healthy weight patients (39.4% ± 7.2% vs 54.6% ± 4.7%; P = .005), and underweight patients had a lower but not significantly inferior EFS rate (40% ± 10.3%; P = .21).

thumbnail image

Figure 1. These charts illustrate (A) overall survival and (B) event-free survival in 314 children with acute myeloid leukemia. Data were analyzed with patients classified into 3 groups of body mass index: underweight, healthy weight, and overweight/obese.

Download figure to PowerPoint

In a multivariable analysis that included BMI, age, leukocyte count, and AML FAB subtype, stratified by study protocol, BMI status retained its significant association with OS (P = .01) but not with EFS (P = .14) (Table 2). In this analysis, overweight/obese patients had a significantly lower OS (P = .004) compared with healthy weight patients, and they tended to have lower EFS (P = .08). OS and EFS did not differ significantly between underweight patients and healthy weight patients (Table 2).

Table 2. Multivariable Cox Regression Model of Survival Stratified by Study Protocol
 Event-Free SurvivalOverall Survival
 HR95% CIPHR95% CIP
Weight Category  .14  .01
  1. Abbreviations: CI, confidence interval; FAB, French-American-British; HR, hazard ratio; WBC, white blood cell.

 Age: 10-20 y vs 2-9.9 y1.691.166-2.451.0061.761.16-2.671.008
 FAB classification  .005  .004
 WBC: ≥50 × 109/L vs <50 × 109/L1.3940.951-2.045.091.2530.815-1.926.30
Overweight/obese vs healthy weight1.4010.963-2.038.081.841.218-2.78.004
 Age: 10-20 y vs 2-9.9 y1.4640.992-2.16.061.6441.053-2.569.03
 FAB classification  .19  .38
 WBC: ≥50 × 109/L vs <50 × 109/L1.3750.933-2.027.111.2210.79-1.888.37
Underweight vs healthy weight1.4270.81-2.514.221.3910.737-2.627.31
 Age: 10-20 y vs 2-9.9 y2.1421.389-3.303<.0012.321.396-3.854.001
 FAB classification  .002  <.001
 WBC: ≥50 × 109/L vs <50 × 109/L2.0451.257-3.329.0041.3480.748-2.431.32

Because BMI criteria for underweight, healthy weight, and overweight/obese differ among studies,4, 15, 16 we evaluated OS and EFS by using a Cox proportional hazards model with BMI percentile as a continuous variable in a spline model. There was a significant, nonlinear association of BMI percentile with OS (P = .008) (Fig. 2A) and with EFS (P = .02) (Fig. 2B). Patients with low or high BMI percentiles had worse OS and EFS. In multivariable analysis, BMI percentile had a significant, nonlinear association with OS (P = .03) (Fig. 2C) but not with EFS (P = .07) (Fig. 2D).

thumbnail image

Figure 2. These charts illustrate the association of body mass index (BMI) percentile with (A,C) overall survival (OS) and (B,D) event-free survival (EFS). (A,B) Data were evaluated in Cox regression analyses with BMI percentile as a continuous variable in a spline model that was stratified according to study protocol. (C,D) In multivariable analyses, OS and EFS were analyzed in a Cox proportional hazards regression model with age, leukocyte count, and French-American-British classification as class predictor variables and with BMI percentile as a continuous variable in a spline model that was stratified according to study protocol. In each graph, a black line indicates a smoothing spline fit, and gray lines indicate ±1 standard error.

Download figure to PowerPoint

Causes of Treatment Failure

The prevalence of remission-induction failure or AML relapse among the 3 BMI groups did not differ significantly: the mean (±standard error) 5-year cumulative incidence of these events was 38% ± 5.4% for overweight/obese patients, 40% ± 10.1% for underweight patients, and 36.4% ± 3.4% for healthy weight patients (P = .88). However, there was a significantly different cumulative incidence of grade ≥2 infections (P = .03) (Fig. 3A) and of grade ≥3 infections (P = .006) (Fig. 3B) among the 3 BMI groups. Both types of infection occurred more frequently during the early treatment phase in overweight/obese patients, whereas grade 4 infections tended to occur more frequently in underweight patients (P = .10) (Fig. 3C).

thumbnail image

Figure 3. These charts illustrate the cumulative incidence of (A) grade 2 or higher infections, (B) grade 3 or higher infections, and (C) grade 4 infections. Data were analyzed with patients classified into 3 groups of body mass index: underweight, healthy weight, and overweight/obese.

Download figure to PowerPoint

Among all 314 patients, 35 died of toxicity, including 29 deaths from infection (13 fungus infections, including 7 Aspergillus and 4 Candida, 6 bacterial, 4 viral, and 6 organisms not identified), 5 from hemorrhage, and 1 from encephalopathy of unknown etiology. Infection was the cause of death for all 14 overweight/obese patients who died and for 5 of 6 underweight patients who died. The cumulative incidence of treatment-related mortality was associated significantly with BMI, with higher mortality rates observed in underweight patients (3-year cumulative incidence, 24% ± 8.8%) and overweight/obese patients (15.2% ± 3.9%) than in healthy weight patients (7.4% ± 1.9%; P = .009) (Fig. 4). When the unhealthy BMI categories were compared individually with the healthy weight category, the cumulative incidence of treatment-related mortality was significantly higher in overweight/obese patients and underweight patients than in healthy weight patients (P = .02 and P = .006, respectively). Although we did not examine the dose modifications because of toxicities or organ dysfunction in individual patients, we did evaluate the duration of each chemotherapy course in each protocol. We did not observe any statistically significant differences among the 3 BMI groups except during consolidation I of AML02: overweight/obese patients (median, 40 days; range, 22-149 days) had longer duration than either healthy weight patients (median, 32.5 days; range, 21-88 days) or underweight patients (median, 33 days; range, 24-54 days; P < .001).

thumbnail image

Figure 4. The cumulative incidence of treatment-related mortality is illustrated in 314 children with acute myeloid leukemia. Data were analyzed with patients classified into 3 groups of body mass index: underweight, healthy weight, and overweight/obese.

Download figure to PowerPoint

We compared the cumulative incidence of treatment-related mortality in patients who were treated on AML02 (2002-2008) with that in patients who were treated on earlier trials (1987-2001). Although the difference was not statistically significant, there was a trend toward lower rates of treatment-related mortality for patients on AML02 compared with those on previous trials (3-year cumulative incidence, 7.9% ± 2.2% vs 13.8% ± 2.8%; P = .07), suggesting that improvement in supportive care could result in improved outcomes. Even in AML02, however, overweight/obese patients still had higher rates of treatment-related mortality than healthy weight patients (3-year cumulative incidence, 13% ± 4.7% vs 5.5% ± 2.4%; P = .12). When obese patients (BMI, ≥95th percentile) were analyzed separately, their treatment-related mortality rate was 16.7% ± 7% (P = .06).

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

The current study demonstrates that overweight/obese patients with AML have significantly worse survival. Excessive treatment-related mortality was observed in unhealthy BMI groups, including both underweight and overweight/obese patients, whereas the rates of refractory leukemia and relapse were nearly identical across weight categories.

In pediatric acute lymphoblastic leukemia (ALL), reports of the effect of BMI on patient outcome are conflicting. A study from the CCG indicated that obese children (BMI, ≥95th percentile) aged ≥10 years had a significantly higher risk of events and relapses.15 In contrast, a study from St. Jude Children's Research Hospital reported no differences in survival, events, or relapse rates among 4 BMI groups (underweight [BMI, ≤10th percentile], normal [BMI, 10th-85th percentile], at risk of overweight, [BMI, 85th-95th percentile], and overweight [BMI, ≥95th percentile]).16 Reports on the impact of BMI in children with AML are in better agreement. A study from the CCG indicated that overweight patients (BMI, ≥95th percentile) and underweight patients (BMI, ≤10th percentile) had worse survival rates, a result primarily caused by the higher rate of infection-related deaths during the first 2 courses of chemotherapy.4 Our current study confirms that overweight patients (BMI, 85th-95th percentile) as well as obese patients (BMI, ≥95th percentile) have worse survival and suggests that their prognosis is unfavorable even in the context of contemporary supportive care. When comparing BMI data, the differences in category criteria must be taken into account. Currently, the CDC categorizes those with a BMI <5th percentile as underweight; however, the CCG AML study used the 10th percentile as the cutoff value for the category.4 Although our data indicated that underweight patients had higher treatment-related mortality, significantly lower survival was not observed. Because our more stringent criteria to define the underweight category may have limited the statistical power of our survival analysis, we evaluated the association between BMI percentile as a continuous variable and OS and EFS by using a Cox regression model, which revealed a significant, nonlinear effect of BMI percentile on OS; patients with lower or higher BMI percentiles had worse OS. We also analyzed our data based on the 10th percentile cutoff, the criterion used in the CCG AML study (data not shown). Although the 5-year OS rate for underweight patients in this cutoff range was significantly lower than that for healthy weight patients in univariate analysis, this difference was not significant in multivariable analysis, perhaps because of the lower number of patients examined in our study.

In contrast to the results from pediatric studies, those from 2 recent adult AML studies indicated that obesity was not associated with worse survival or toxicities.17, 18 Although there is no clear explanation for this discrepancy between pediatric and adult findings, adults tend to have poorer results and higher risk cytogenetic features, which may overcome any association of BMI with outcome. Increasing BMI was strongly associated with APL but not with other subtypes of AML.19 The results from a recently published study of children and adults with APL indicated that overweight/obese patients had a significantly higher relapse rate and incidence of differentiation syndrome.20 These results may suggest a relation between obesity-related metabolic mediators and the pathogenesis of APL.

The prevalence of obesity has increased substantially over the past few decades among children and adolescents in the United States; 16.9% of the surveyed population was categorized as obese, and 31.7% was within overweight ranges of BMI for age.21 Therefore, obesity is a factor frequently encountered in pediatric oncology,22 and there have been significantly increased percentages of obese patients in our recent protocols, especially among older children. Overweight/obese patients can have subtle immunologic abnormalities that may contribute to an increased likelihood of infection-related death.23 For example, adipocytokines, such as adiponectin, can suppress the synthesis of tumor-necrosis factor and interferon-γ and can induce the production of anti-inflammatory cytokines, such as interleukin-10 (IL-10) and IL-1 receptor antagonist.23 Another adipocytokine, leptin, can induce the production of proinflammatory cytokines, such as tumor necrosis factor, IL-6, and IL-12.23 Overweight and obese patients also may have comorbid conditions, such as diabetes and hypertension.22 Furthermore, physical examination of obese patients may be difficult and less effective, especially at identifying abdominal and perirectal pathology, which can affect outcomes.24 These issues necessitate detailed laboratory tests, including metabolic and immunologic profiles, along with nutrition assessment, dietary education, early use of prophylactic antibiotics, careful physical examination, and frequent clinic visits or inpatient care. Underweight patients also had worse treatment-related mortality in our study. Chronic malnutrition in developed countries, including the United States, is observed in approximately 1% of children.25 In addition to impaired physical growth and cognitive functions, the immune response is changed early in the course of significant malnutrition in a child.26 Associated problems include loss of delayed hypersensitivity; fewer T lymphocytes; impaired lymphocyte response; and impaired phagocytosis secondary to decreased complement, cytokines, and immunoglobulins. Therefore, underweight children with AML need immediate nutritional intervention to help prevent or reverse nutrition deficiencies as well as infection prophylaxis with antibiotics and supplemental immunoglobulin.

The data comparing the pharmacokinetics of various drugs in adult or pediatric obese patients with the pharmacokinetics in normal weight controls are conflicting,22 and few studies have examined this issue in underweight patients. Pharmacokinetics can be affected by multiple factors, such as plasma protein binding capacity, water or lipid solubility of the compound, liver metabolism, and the function of the excretion pathway. Hijiya et al16 demonstrated that intracellular levels of thioguanine nucleotides and methotrexate polyglutamates, as well as systemic clearance of methotrexate, teniposide, etoposide, and low-dose cytarabine, did not differ between 4 BMI groups of patients with pediatric ALL. In AML, however, high-dose cytarabine, the mainstay of treatment, is highly associated with alpha-streptococcus bacteremia,27 and similar pharmacokinetics studies have not been performed in obese or underweight patients. Doxorubicin had lower clearance and longer half-life in the most obese patients, resulting in a larger area under the curve and more exposure, although the pharmacokinetics of doxorubicinol, an active metabolite, did not differ among patient groups.28 Prospective pharmacokinetic studies of medications used in AML, especially cytarabine, anthracycline, and etoposide, are needed. In most patients, neither we nor CCG modified the chemotherapy dosing on the basis of BMI or weight in AML studies.4 However, pharmacokinetics-guided dosing may reduce toxicity in those patients with AML who have an unhealthy BMI.

In conclusion, children with an unhealthy BMI (ie, underweight, overweight, and obese) have lower survival rates because of treatment-related mortality. BMI-oriented supportive care should be provided to these patients during treatment for AML.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

We thank Cherise Guess, PhD, for editorial assistance and Soheil Meshinchi, MD; Barbara Degar, MD; and Gladstone Airewele, MD for participating in the AML02 study.

FUNDING SOURCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

This work was supported in part by Cancer Center Support (CORE) grant P30 CA021765-30 from the National Institutes of Health and by the American Lebanese Syrian Associated Charities.

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES
  • 1
    Tsukimoto I, Tawa A, Horibe K, et al. Risk-stratified therapy and the intensive use of cytarabine improves the outcome in childhood acute myeloid leukemia: the AML99 trial from the Japanese Childhood AML Cooperative Study Group. J Clin Oncol. 2009; 27: 4007-4013.
  • 2
    Rubnitz JE, Inaba H, Dahl G, et al. Minimal residual disease-directed therapy for childhood acute myeloid leukaemia: results of the AML02 multicentre trial. Lancet Oncol. 2010; 11: 543-552.
  • 3
    Grimwade D, Hills RK. Independent prognostic factors for AML outcome [serial online]. Hematol Am Soc Hematol Educ Program. 2009: 385-395.
  • 4
    Lange BJ, Gerbing RB, Feusner J, et al. Mortality in overweight and underweight children with acute myeloid leukemia. JAMA. 2005; 293: 203-211.
  • 5
    Arnaout MK, Radomski KM, Srivastava DK, et al. Treatment of childhood acute myelogenous leukemia with an intensive regimen (AML-87) that individualizes etoposide and cytarabine dosages: short- and long-term effects. Leukemia. 2000; 14: 1736-1742.
  • 6
    Krance RA, Hurwitz CA, Head DR, et al. Experience with 2-chlorodeoxyadenosine in previously untreated children with newly diagnosed acute myeloid leukemia and myelodysplastic diseases. J Clin Oncol. 2001; 19: 2804-2811.
  • 7
    Rubnitz JE, Crews KR, Pounds S, et al. Combination of cladribine and cytarabine is effective for childhood acute myeloid leukemia: results of the St Jude AML97 trial. Leukemia. 2009; 23: 1410-1416.
  • 8
    Kruskal WH, Wallis WA. Use of ranks in 1-criterion variance analysis. J Am Stat Assoc. 1952; 47: 583-621.
  • 9
    Kaplan EL, Meier P. Non-parametric estimation for incomplete observations. J Am Stat Assoc. 1958; 53: 457-481.
  • 10
    Peto R, Pike MC, Armitage P, et al. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. analysis and examples. Br J Cancer. 1977; 35: 1-39.
  • 11
    Mantel N. Evaluation of survival data and 2 new rank order statistics arising in its consideration. Cancer Chemother Rep. 1966; 50: 163-170.
  • 12
    Cox D. Regression models and life-tables. J R Stat Soc B. 1972; 34: 187-220.
  • 13
    Gray RJ. A class of K-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat. 1988; 16: 1141-1154.
  • 14
    Gray RJ. Spline-based tests in survival analysis. Biometrics. 1994; 50: 640-652.
  • 15
    Butturini AM, Dorey FJ, Lange BJ, et al. Obesity and outcome in pediatric acute lymphoblastic leukemia. J Clin Oncol. 2007; 25: 2063-2069.
  • 16
    Hijiya N, Panetta JC, Zhou Y, et al. Body mass index does not influence pharmacokinetics or outcome of treatment in children with acute lymphoblastic leukemia. Blood. 2006; 108: 3997-4002.
  • 17
    Medeiros BC, Othus M, Estey EH, Fang M, Appelbaum FR. Impact of body-mass index in the outcome of adult patients with acute myeloid leukemia [published online ahead of print February 7, 1012]. Haematologica. 2012.
  • 18
    Lee HJ, Licht AS, Hyland AJ, et al. Is obesity a prognostic factor for acute myeloid leukemia outcome? Ann Hematol. 2012; 91: 359-365.
  • 19
    Estey E, Thall P, Kantarjian H, Pierce S, Kornblau S, Keating M. Association between increased body mass index and a diagnosis of acute promyelocytic leukemia in patients with acute myeloid leukemia. Leukemia. 1997; 11: 1661-1664.
  • 20
    Breccia M, Mazzarella L, Bagnardi V, et al. Increased BMI correlates with higher risk of disease relapse and differentiation syndrome in patients with acute promyelocytic leukemia treated with the AIDA protocols. Blood. 2012; 119: 49-54.
  • 21
    Ogden CL, Carroll MD, Curtin LR, Lamb MM, Flegal KM. Prevalence of high body mass index in US children and adolescents, 2007-2008. JAMA. 2010; 303: 242-249.
  • 22
    Rogers PC, Meacham LR, Oeffinger KC, Henry DW, Lange BJ. Obesity in pediatric oncology. Pediatr Blood Cancer. 2005; 45: 881-891.
  • 23
    Tilg H, Moschen AR. Adipocytokines: mediators linking adipose tissue, inflammation and immunity. Nat Rev Immunol. 2006; 6: 772-783.
  • 24
    Wiernik PH, Serpick AA. Factors effecting remission and survival in adult acute nonlymphocytic leukemia (ANLL). Medicine (Baltimore). 1970; 49: 505-513.
  • 25
    de Onis M, Blossner M, Borghi E, Frongillo EA, Morris R. Estimates of global prevalence of childhood underweight in 1990 and 2015. JAMA. 2004; 291: 2600-2606.
  • 26
    Chandra RK. Nutrition and the immune system from birth to old age. Eur J Clin Nutr. 2002; 56( suppl 3): S73-S76.
  • 27
    Gamis AS, Howells WB, DeSwarte-Wallace J, Feusner JH, Buckley JD, Woods WG. Alpha hemolytic streptococcal infection during intensive treatment for acute myeloid leukemia: a report from the Children's Cancer Group Study CCG-2891. J Clin Oncol. 2000; 18: 1845-1855.
  • 28
    Rodvold KA, Rushing DA, Tewksbury DA. Doxorubicin clearance in the obese. J Clin Oncol. 1988; 6: 1321-1327.