Background: In adults, obesity has been associated with increased risk of developing acute myeloid leukemia (AML) with recurrent cytogenetic abnormalities and preclinical studies indicate that certain proteins associated with obesity are also involved in the development of subtypes of AML. Studies on the effect of body-mass index (BMI) in childhood AML have suggested differences in cytogenetic abnormalities across weight-groups, but the association has not been well-studied. Further, the effect of BMI on outcome in children with AML is controversial. While previous studies have found that overweight is associated with worse survival, mainly due to increased treatment-related mortality, others have not confirmed this association. We investigated differences in cytogenetic abnormalities, treatment-related mortality, relapse and overall survival in four weight groups in a large multi-institutional cohort of children diagnosed with AML.

Methods: We included patients, age 2-17, diagnosed with de novo AML from the five Nordic countries (2004-2016), Hong Kong (2007-2016), the Netherlands and Belgium (2010-2016) and Canada and USA (1995-2012). The cohort was population-based for the European countries, Hong Kong and Canada from 1995-2004 for all provinces except for Saskatchewan. In USA and Canada from 2005-2012 the cohort was based on cases included for a prospective study at 15 Canadian and 2 American centers. Cases with Down syndrome, acute promyelocytic leukemia, isolated granulocytic sarcoma or missing data on BMI at diagnosis (n=7) were excluded. BMI standard deviation (SD) for age and sex was calculated and categorized as following: underweight: <-2 SD, healthy weight: -2-1 SD, overweight: >1-2 SD and obesity: >2 SD according to the World Health Organization. Treatment-related mortality was classified according to an international consensus-based definition of treatment-related mortality in children with cancer. Logistic regression was used to investigate associations between weight groups and cytogenetic abnormalities. Thirty-six cases had missing data on one or more cytogenetic abnormality and were excluded from the cytogenetic abnormality analyses. Time-to-event analyses and Cox regression was used to investigate associations with treatment-related mortality, relapse and overall survival. Reason for death was unknown in two cases (excluded from treatment-related mortality analyses). For all analyses, the healthy weight group was used as reference. For relapse analyses death was a competing risk and for treatment-related mortality analyses, death from progressive disease was a competing risk. Treatment-related mortality analyses were adjusted for sex, age, and year of diagnosis; relapse and overall survival analyses also for country.

Results: In total 867 patients were included in the study. The median age was 10 (range 2-17) and 53% were male. Patients were treated on 17 different protocols with AAML0531, COG9421, NOPHO-AML 2004, DB AML 01 and NOPHO-DBH AML 2012 accounting for 79%. None of the protocols recommended chemotherapy dose capping for overweight patients. At diagnosis 32 (4%) were underweight, 593 (68%) were healthy weight, 160 (18%) were overweight and 82 (9%) were obese. Boys and older patients were more likely to be underweight or obese. North Americans were less likely to be underweight and more likely to be obese. There were 326 first relapses, 104 treatment-related mortalities and 251 deaths. The frequency of t(8;21) and inv(16) were higher in the obese group and the frequency of MLL rearrangements were higher in the overweight group. The frequency of t(8;21) and inv(16) increased with increasing weight-group (table 1). There was no evidence of difference in the risk of relapse, treatment-related mortality or overall survival based on weight groups (table 2).

Conclusion: This multi-institutional study of 867 pediatric patients with de novo AML found an association between favorable cytogenetic abnormalities and obesity. The study did not confirm previous reports of overweight/obesity being associated with decreased overall survival, risk of relapse or treatment-related mortality. Our results could suggest that overweight is a risk factor for developing specific cytogenetic subgroups of AML in children.


Kaspers: Janssen-Cilag: Research Funding.

Author notes


Asterisk with author names denotes non-ASH members.