Amid the ever-increasing concern about the escalating prevalence of obesity among young people in the general populations of North America and Western Europe, are dire predictions of a future burden of coronary heart disease in adult life1; a prediction that has come true already, as exemplified by a study of more than 1/4 million former Danish school children in their adult years.2 Such concerns relating to obesity are at least as relevant to long-term survivors of cancer in childhood, particularly those who experienced cranial irradiation as part of their treatment for acute lymphoblastic leukemia (ALL).3 This problem is compounded by the poor adherence of adult survivors of ALL to dietary guidelines.4
The metabolic syndrome—a clustering of obesity, hypertension, dyslipidemia, and insulin resistance—is linked strongly to cardiovascular disease and type 2 diabetes. In children with ALL, acquisition of the features of metabolic syndrome seems to be initiated during antileukemic therapy5,6 and a plea has been made to begin nutritional and behavioral interventions before this treatment has been completed.7
Examination of the changes in body mass index (BMI) across the trajectory of antileukemic therapy should be informative in this regard. More than a decade ago, we drew attention to the gain in weight, disproportionate to height, during the treatment of ALL in children.8 Others have reported similar findings since, in both the United Kingdom9 and the United States,10,11 including a study of children treated according to Dana Farber Cancer Institute (DFCI) protocols.12
The current investigation was undertaken to examine in detail the trajectory of changes in BMI from diagnosis through completion of treatment in a single institution that used DFCI treatment regimens. The following hypotheses were tested: the prevalence of overweight/obesity at the time of diagnosis of ALL will be higher in the decade 1995 to 2004 than in the preceding 10 years; overweight/obesity will be more common in patients who received cranial irradiation than in those who did not; the administration of dexamethasone after remission induction is associated with a greater frequency of overweight/obesity than the corresponding administration of prednisone; and high-dose steroid therapy (3-fold higher than the alternate) will cause more patients to be overweight/obese than low-dose therapy.
PATIENTS AND METHODS
All children (n=230) who were diagnosed with ALL at McMaster Children's Hospital between January 1985 and the beginning of December 2004 were eligible for this historical cohort study. The start date was determined by the year in which McMaster Children's Hospital became the first Canadian Institution to join the DFCI Childhood ALL Consortium. The end date allowed all children to complete “first-line” therapy by the end of December 2006. The health records of these children were searched for hospital ID number, sex, age at diagnosis, date of diagnosis, protocol assignment, prognostic risk category of disease, cranial radiation dose, postinduction steroid use (dexamethasone or prednisone), and intensity of steroid therapy (high or low dose). In addition, height and weight records (used to calculate doses of chemotherapy on the basis of body surface area) were extracted at the time of diagnosis and at 6-month intervals thereafter until the end of treatment. Patients were treated successively on protocols 85-01, 87-01, 91-01, 95-01, and 2000-01.
Exclusion criteria were age below 2 years at diagnosis, (validated BMI Z scores not available,13) Down syndrome, enrollment on other protocols, and event-free survival of less than 2 years.
BMI was calculated as weight/height2 (kg/m2) and expressed as standardized Z scores and centiles adjusted for age and sex. Using the BMI centile at diagnosis, each child's nutritional status was categorized as undernourished, normal, at risk of overweight, or overweight according to the respective percentiles (<5th, 5th to 85th, >85th to 95th, and >95th).
To determine the association of steroid therapy and cranial irradiation, repeated-measures linear model was used to model the standardized BMI scores (Z scores). The model was adjusted for baseline BMI score, drug type (prednisone vs. dexamethasone), risk category based on steroid dose (high vs. low), and cranial irradiation (yes vs. no). The covariance structure was assumed to be compound symmetric. Initially, a full model with all possible interactions was investigated. A backward selection method was used with a staying criterion of P=0.05 for all variables except cranial irradiation. Cranial irradiation was forced to remain in the model considering its primary role with respect to our hypotheses.
Data were entered and maintained in a database at the Department of Epidemiology and Biostatistics, University of Texas Health Science Center at San Antonio. All statistical testing was 2-sided with a significance level of 5% and was performed using SAS (version 9.2, SAS Institute, Cary, NC). All graphical presentations were created using R (version 2.9.0).
The study was approved by the Research Ethics Board of Hamilton Health Sciences and McMaster University.
The study population (n=162) consisted of 98 (61%) boys and 64 (39%) girls with an average age (mean±SD) at diagnosis of 77±48 months. There were 93 (57%) standard risk patients with 69 (43%) were categorized as high/very high risk. A total of 29 children were excluded on the basis of age (<2 y), 4 had Down syndrome, 10 were treated on other protocols, and 18 had an event-free survival less than 2 years. Seven patients had missing data.
The distribution of patients (n, %) by treatment protocol was: 85 to 01 (16, 10%); 87 to 01 (30, 18.5%); 91 to 01 (30, 18.5%); 95 to 01 (41, 25%); and 2000 to 01 (45, 28%) (Table 1). Almost 60% (n=97) of the patients had received cranial irradiation, 51 (31%) had been given dexamethasone and 69 (43%) were treated with high-dose steroid therapy.
On the basis of the BMI centiles at diagnosis, the distribution of patients (n, %) by nutritional status were undernourished (17, 10.5%), normal (121, 74.7%), risk for overweight (13, 8.0%), and overweight (11, 6.8%) (Table 1). The Z scores at diagnosis ranged from −3.0 to +3.3 with a mean±SD of −0.2±1.2 and were normally distributed based on the Shapiro-Wilks test (P=0.69) (data not shown). The changes in scores over time, stratified by risk group, according to type of steroid are shown in Figure 1.
At diagnosis, 24 (14.8%) patients had a BMI score above the 85th centile. As shown in Table 2, this group steadily increased as treatment progressed: 6 months (27, 16.7%), 12 months (48, 29.6%), 18 months (60, 37%), and 24 months (64, 39.5%). As hypothesized, of those above the 85th centile, a higher proportion were diagnosed from years 1995 to 2004 than the preceding decade of 1985 to 1994 (54.2% vs. 45.8%), although this difference was not statistically significant (P=0.68). The proportion of patients who received cranial irradiation was approximately 60% in both the at-risk for overweight/overweight group and the others (Table 2). In addition, there was not a significant association (P=0.73) of high/low risk of disease with respect to weight category (Table 2).
As hypothesized, over the course of the treatment the frequency of overweight/obesity was consistently higher in the patients who received dexamethasone versus prednisone (baseline: 24% vs. 11%, 6 mo: 18% vs. 17%, 12 mo: 39% vs. 27%, 18 mo: 55% vs. 31%, and 24 mo: 59% vs. 34%) (Supplement Table 1, Supplemental Digital Content 1, http://links.lww.com/JPHO/A1). The BMI Z scores showed a similar trend of dexamethasone-treated patients being consistently higher than prednisone-treated patients, but the difference was not significant (P=0.25) in a repeated-measures linear model (Table 3). Although there was a trend toward higher BMI Z scores over time for the cranial irradiation group, it did not reach statistical significance (P=0.06). There was a statistically significant difference (P=0.001) in BMI Z scores for those who received high/low risk steroid therapy (Fig. 1), with this significance diminishing over time (6 mo: P<0.001, 12 mo: P<0.001, 18 mo: P=0.006, and 24 mo: P=0.06), as shown in Table 4. However, it should be noted that cranial irradiation was confined largely to patients who had high-risk disease (98.6%), as is true also for high-dose steroid therapy (98.5%). The BMI Z scores (median, interquartile range) were higher at 24 months in patients with high/very high risk disease [median: 0.9, interquartile range: (0, 1.8)] than in those with standard risk disease [0.7, (0.3, 1.3)].
A clinically valuable measure of nutritional status in children with cancer should be simple, accurate, noninvasive, inexpensive, and readily available for repeated use. Although dual energy x-ray absorptiometry scan offers a “gold standard” in the clinical setting,14 it does not fulfill all of these requirements. A recent report15 examined the validity of various anthropometric measures based on a comparison with a criterion method [total body potassium, yielding calculated values for body cell mass index (BCMI) Z scores] in 37 children with cancer. The strongest relationship (P=0.02) was between the BCMI Z score and the BMI Z score, but the authors were of the opinion that this was not biologically significant. It was noted that the normative data for BCMI were based on a small number of subjects and the validity in children of the formula for converting total body potassium to body cell mass, generated in adults, was uncertain. Furthermore, the BCMI Z score classified 48% of these children with cancer in Australia as being undernourished.
Measures of nutritional status that are based on weight can be problematic in children with cancer, leading to underestimation of malnutrition in those with large solid tumors,16 but this does not seem to pose the same challenge in those with leukemia. Reilly et al,17 in a study of more than 1000 children with standard risk ALL at diagnosis in the UK, found that BMI Z scores provided a good measure of protein energy nutrition. These investigators noted that approximately 7% of the children were malnourished (BMI Z scores less than –2.0); (Fig. 2) a number 2 fold greater than expected statistically. A study from St Jude Children's Research Hospital found that more than 16% of children with ALL were malnourished at diagnosis, as defined by a BMI of less than the 10th centile.18 The corresponding proportion in this study was 15.4%.
Reilly and et al9 reported on a serial study of 26 children with ALL. The BMI Z score rose by an average of 0.38 in the first year and by 0.68 from diagnosis to the end of the second year of treatment. It was observed that energy intake was 20% higher on than off-steroid therapy, with no difference between patients receiving prednisone and dexamethasone. A Dutch study of 16 children with ALL on maintenance therapy19 found that energy intake was almost 40% higher on than off-steroid (dexamethasone) therapy and that, while on steroid therapy, the children were less physically active. These investigators noted also that BMI may underestimate the problem of obesity in children with ALL20 because of underestimation of linear growth retardation (the difference between target height, based on parental height, and final height).
The focus of attention, with respect to BMI in children with ALL, is mainly on increasing values.7 Additional justification for this priority comes from the observation that overweight children with ALL, who are 10 years of age or above, had a significantly greater risk of relapse and lower probability of event-free survival when treated on Children's Cancer Group protocols.21
The authors thank Amy Cranston BA, Yavisha Nayiager, Paul Pencharz MD, PhD, and Sabrina Siciliano for their contributions to this study.
1. Bibbins-Domingo K, Coxson P, Pletcher MJ, et al. Adolescent overweight and future adult coronary heart disease. N Engl J Med. 2007;357:2371–2379.
2. Baker JF, Olsen LW, Sorensen TIA. Childhood body mass index and the risk of coronary heart disease in adulthood. N Engl J Med. 2007;357:2329–2337.
3. Garney EG, Liu Q, Sklar CA, et al. Longitudinal changes in obesity and body mass index among adult survivors of childhood acute lymphoblastic leukemia: a report from the Childhood Cancer Survivor Study. J Clin Oncol. 2008;26:4639–4645.
4. Robien K, Ness KK, Klesges LM, et al. Poor adherence to dietary guidelines among adult survivors of childhood acute lymphoblastic leukemia. J Pediatr Hematol Oncol. 2008;30:815–822.
5. Chow EJ, Pihoker C, Hunt K, et al. Obesity and hypertension among children after treatment for acute lymphoblastic leukemia. Cancer. 2007;110:2313–2320.
6. Lowas S, Malempati S, Marks D. Body mass index predicts insulin resistance in survivors of pediatric acute lymphoblastic leukemia. Pediatr Blood Cancer. 2009;53:58–63.
7. Withycombe JS, Post-White JE, Meza JL, et al. Weight patterns in children with higher risk ALL: a report from the Children's Oncology Group (COG) for CCG 1961. Pediatr Blood Cancer. 2009;53:1249–1253.
8. Halton JM, Atkinson SA, Barr RD. Growth and body composition in response to chemotherapy in children with acute lymphoblastic leukemia. Int J Cancer. 1998;11(suppl):81–84.
9. Reilly JJ, Brougham M, Montgomery C, et al. Prevalence of protein-energy malnutrition at diagnosis in children with acute lymphoblastic leukemia. J Clin Endocrinol Metab. 2001;86:3742–3745.
10. Baillargeon J, Langevin AM, Lewis M, et al. Therapy-related changes in body size in Hispanic children with acute lymphoblastic leukemia. Cancer. 2005;103:1725–1729.
11. Bailargeon J, Langevin AM, Lewis M, et al. Demographic correlates of body size changes in children undergoing treatment for acute lymphoblastic leukemia. Pediatr Blood Cancer. 2007;49:793–796.
12. Dalton VK, Rue M, Silverman LB, et al. Height and weight in children treated for acute lymphoblastic leukemia: relationship to CNS treatment. J Clin Oncol. 2003;21:2953–2960.
13. Ogden C, Flegal K, Carroll M, et al. Prevalence and trends in overweight among US children and adolescents 1999 to 2000. JAMA. 2002;288:1728–1731.
14. Sala A, Pencharz P, Barr RD. Children, cancer and nutrition– a dynamic triangle in review. Cancer. 2004;100:677–687.
15. Murphy AJ, White M, Davies PSW. The validity of simple methods to detect poor nutritional status in paediatric oncology patients. Br J Nutr. 2009;101:1388–1392.
16. Smith DE, Stevens MCG, Booth JW. Malnutrition at diagnosis of malignancy in childhood: common but mostly missed. Eur J Pediatr. 1991;150:318–322.
17. Reilly JJ, Weir J, McColl JA, et al. Prevalence of protein-energy malnutrition at diagnosis in children with acute lymphoblastic leukemia. J Pediatr Gastroenterol Nutr. 1999;29:194–197.
18. 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.
19. Jansen H, Postma A, Stolk RP, et al. Acute lymphoblastic leukemia and obesity: increased energy intake or decreased physical activity? Support Care Cancer. 2009;17:103–106.
20. Bongers MEJ, Francken AB, Rouwe C, et al. Reduction in adult height in childhood acute lymphoblastic leukemia survivors after prophylactic cranial irradiation. Pediatr Blood Cancer. 2005;45:139–143.
21. Butturini AM, Dorey FJ, Lange BJ, et al. Obesity and outcomes in pediatric acute lymphoblastic leukemia. J Clin Oncol. 2007;25:2063–2069.
© 2010 Lippincott Williams & Wilkins, Inc.