SEARCH

SEARCH BY CITATION

Keywords:

  • acute lymphoblastic leukemia;
  • body mass index;
  • obesity;
  • cancer risk;
  • Hispanic

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

BACKGROUND

The objective of this study was to examine changes over time in body mass index (BMI) from diagnosis through chemotherapy for pediatric patients with B-precursor acute lymphoblastic leukemia (ALL).

METHODS

The study cohort consisted of 141 white Hispanic pediatric patients who were diagnosed with ALL and were treated at 2 South Texas pediatric oncology centers between 1993 and 2002. Changes in age-standardized and gender-standardized BMI scores were assessed.

RESULTS

The study cohort exhibited a steady increase in age-adjusted and gender-adjusted BMI scores for the first 12 months of therapy, a modest increase in BMI scores during the 18–23 month and 24–29 month periods, followed by a slight decrease in BMI scores at 30 months (end of therapy). A repeated-measures analysis indicated significant effects for time (P = 0.019) and time by baseline BMI category interaction (P = 0.0001) but no significant interaction effect between time and gender (P = 0.65).

CONCLUSIONS

Although it is known that leukemia therapy is associated with prevalent obesity in survivorship, its pattern of development during therapy has not been elucidated. In the current cohort of Hispanic children with ALL, BMI scores were elevated at diagnosis (mean ± standard deviation standardized BMI Z score, 0.33 ± 1.4), then increased, and remained elevated for the entire duration of chemotherapy. Patients who were classified as normal weight exhibited an increase in BMI over time; patients who were classified as overweight at diagnosis exhibited BMI patterns that were relatively stable; and patients who were classified as obese exhibited a very slight decline over time. These findings suggest that the risk for chemotherapy-related weight gain applies predominantly to children who begin ALL therapy within a normal weight range. Cancer 2005. © 2005 American Cancer Society.

Obesity is one of the most commonly occurring late effects among childhood cancer survivors.1–9 A number of etiologic mechanisms reportedly underlie this problem, including exposure to corticosteroids, exposure to cranial irradiation, diminished exercise capacity, and psychological barriers to physical activity.3, 8, 9 Development of obesity during and after childhood cancer treatment places survivors at increased risk for hypertension, hyperlipidemia, Type II diabetes, cardiovascular disease, and persistent obesity into adulthood.5, 10–14 Childhood cancer patients who already are obese at the time of diagnosis represent a particular challenge for the oncologist. Treatment-related weight gain may exacerbate the obese patient's already increased risk for many of the aforementioned chronic diseases. In addition, obesity complicates determination of appropriate chemotherapy dosing. Because the prevalence of childhood and adolescent obesity is particularly high among Hispanic adolescents, this group may experience a risk of treatment-related obesity and its effects. Our study objective was to assess changes in body mass index (BMI) during treatment for acute lymphoblastic leukemia (ALL) among a cohort of predominantly Mexican-American patients.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Study Cohort

The study cohort consisted of a consecutive sample of 141 Hispanic pediatric patients diagnosed with B-precursor ALL in first remission who were treated at CHRISTUS Santa Rosa Children's Hospital (San Antonio, TX) or Driscoll Children's Hospital (Corpus Christi, TX) between January 1, 1993 and December 31, 2002. Because the Centers for Disease Control and Prevention (CDC) recommends the use of BMI growth charts beginning at age 2 years, when an accurate stature can be determined,15 we excluded all children who were age < 2 years at diagnosis. To ensure assessment of BMI changes over time, we restricted our study to patients who were followed for a minimum of 6 months. In addition, to minimize treatment-related variability, we excluded six patients who received cranial irradiation as part of their treatment. Of the original 147 study patients, 141 patients met the entry criteria. The included and excluded study patients exhibited comparable distributions of the demographic study factors. All data were obtained by reviewing existing patient records. Patients were treated primarily according to Pediatric Oncology Group legacy protocols (P9000, P9400, and P9900 series) following the National Cancer Institute definition for standard risk or high risk based on age and leukocyte count at diagnosis.16 Patients received either a three-drug or a four-drug induction regimen that consisted of prednisone, asparaginase, and vincristine with or without daunorubicin. This study was reviewed formally and was approved by the University of Texas Health Science Center–San Antonio Institutional Review Board.

Measurements

BMI was assessed using the following formula: weight (kilograms)/height (meters)2. BMI is considered the best single anthropomorphic measure in both children and adults with respect to independence of height, correlation with body fat, and prediction of mortality.17, 18 Ethnicity was assigned on the basis of parental report according to National Cancer Institute/Children's Clinical Oncology Group conventions

Statistical Methods

Age-standardized and gender-standardized BMI Z-scores were calculated for each study patient using height, weight, gender, and age data based on CDC National Center for Health Statistics growth curves.15, 19 The Z-score indicates the number of standard deviations the measurement is away from the mean for the normal age-gender cohort. Obesity status was defined as a BMI Z-score ≥ the 95th percentile, and overweight status was defined as a BMI Z-score ranging from the 85th percentile to the 95th percentile. Both definitions are based on the recommendations of an expert panel on childhood obesity20 and are in accordance with previous studies.12, 21, 22 We assessed the prevalence of obese, overweight, and normal weight status according to demographic and other study characteristics. To examine BMI changes over time, we employed a general linear model approach using the MIXED procedure in SAS (SAS Institute, Cary, NC). A first-order, autoregressive structure was used to model covariance. Between-group comparisons were made using differences of least squares means.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Table 1 presents the demographic characteristics of the study cohort according to BMI percentile categories. Overall, 29.8% of the study cohort was either overweight or obese (BMI > 85th percentile) at diagnosis, and 17.0% of the cohort was obese (BMI > 95th percentile). Female patients ages 10–18 years had elevated rates of classification as either overweight or obese.

Table 1. Study Cohort Characteristic Percentages by Body Mass Index Percentile Category at Diagnosis
CharacteristicProportion of patients (%)
Normal weight (BMI ≤85th percentile)Overweight (BMI 85–95th percentile)Obese (BMI > 95th percentile)
  1. BMI: body mass index.

Overall (n = 141)70.212.817.0
Gender   
 Male (n = 90)68.617.713.7
 Female (n = 51)71.110.018.9
Age group   
 2–9 yrs (n = 120)75.812.511.7
 10–18 yrs (n = 21)38.114.347.6

Table 2 presents the overall and BMI category-specific mean BMI Z-scores at diagnosis and at 6-month intervals after diagnosis. Average BMI Z-scores increased for the entire cohort (n = 141 patients). When examining these patterns for subgroups defined by their BMI at diagnosis, for the normal BMI subcohort (n = 99 patients), BMI increased substantially between baseline and 12 months, increased moderately between 12 months and 24 months, then showed a slight decrease at 30 months. For the overweight subcohort (n = 23 patients), BMI exhibited no consistent pattern of increase or decrease over time. For the obese subcohort, there was a slight overall decrease over time. Consistent with this finding, a repeated-measures analysis indicated a significant interaction for time × BMI baseline category (P < 0.0001), indicating that the normal group followed a different pattern than the other 2 groups, which either appeared flat or exhibited a slight decrease over time (Fig. 1). A random-effects model that was fitted with separate slopes for BMI baseline category subgroups indicated that, on average, patients in the normal BMI group had growth slopes that differed significantly from 0 (slope = 0.18; P < 0.0001), whereas patients in the overweight group (slope = 0.04; P < 0.39) and in the obese group (slope = − 0.06; P = 0.11) did not. Assessment of the entire study cohort, independent of baseline BMI categories, indicated a significant effect for time (slope = 0.16; P < 0.0001) but no significant interaction between gender and time (P = 0.68), suggesting that neither of these factors influenced standardized BMI score changes over time. The proportion of patients who changed BMI percentile categories over the follow-up period also was assessed. Among the patients who presented initially in the normal BMI range, 16% became overweight, and 6% became obese by the end of the 30-month study period. Among patients who were overweight at baseline, 36% became obese; and, among the patients who were obese at baseline, 13% became overweight, and 6% became normal weight by the end of the study.

Table 2. Mean Body Mass Index Z Scores at Diagnosis and at 6-Month Intervals
IntervalOverallNormal weightOverweightObese
No. of patientsBMI Z score (mean ± SD)No. of patientsBMI Z score (mean ± SD)No. of patientsBMI Z score (mean ± SD)No. of patientsBMI Z score (mean ± SD)
  1. BMI: body mass index; SD: standard deviation.

Baseline1410.33 ± 1.499−0.35 ± 1.1181.37 ± 0.2242.37 ± 0.5
6 mos1410.73 ± 1.2990.32 ± 1.0181.09 ± 1.1242.14 ± 0.9
12 mos1360.91 ± 1.1950.54 ± 0.9181.27 ± 0.9232.15 ± 0.7
18 mos1250.92 ± 1.1900.55 ± 0.9151.47 ± 0.7202.20 ± 0.5
24 mos1070.94 ± 1.0770.60 ± 0.9131.51 ± 0.7172.09 ± 0.6
30 mos990.81 ± 1.1700.43 ± 0.9131.39 ± 0.7152.00 ± 0.7
thumbnail image

Figure 1. Body mass index (BMI) at 6-month intervals during therapy.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Obesity is a common late complication of treatment for childhood ALL1, 3, 4, 6–9 and has been attributed to a number of factors, including corticosteroid exposure, diminished exercise capacity, cranial irradiation,5 and psychological barriers to physical activity.3, 8, 9 To our knowledge, little information concerning relative weight change both during and after chemotherapy has been reported to date. The objective of the current study, therefore, was to describe this phenomenon by assessing the specific time-related pattern of BMI change over the course of chemotherapy.

We observed an elevated prevalence of obesity prior to chemotherapy in our study cohort of Hispanic-Caucasian patients with ALL. After this, patients exhibited a consistent increase in BMI for the first 24 months of therapy. A separate analysis, in which patients were stratified according to their BMI category at diagnosis, indicated that this pattern was seen predominantly in children who had BMI scores in the normal range at diagnosis. Patients who were classified as overweight at baseline exhibited BMI patterns that remained relatively stable, whereas patients who were classified initially as obese exhibited a very slight decrease. However, a significant proportion of patients who were overweight at diagnosis became obese during therapy. These data suggest that the largest increase in BMI occurs during the first 12 months of treatment and that the risk of chemotherapy-associated weight gain occurs in children who are within the normal weight range at diagnosis. These findings are consistent with a number of previous investigations, which reported that children with ALL experience an increase in BMI during chemotherapy that may persist after treatment.2, 3, 23 Because the vast majority of previous ALL studies included substantial numbers of patients who received cranial irradiation, our findings that BMI changes occurred independent of this risk factor are particularly noteworthy. In a recent study, Davies et al.2 similarly reported that a cohort of childhood ALL patients exhibited a substantial increase in standardized BMI scores during chemotherapy with no exposure to cranial irradiation. By contrast, in an earlier study, Sklar et al.24 reported that standardized BMI scores increased significantly in patients who were treated with cranial irradiation but remained unchanged in patients who were treated exclusively with chemotherapy. In view of the reduced role of cranial irradiation in most conventional treatments for childhood ALL, it will be important for future studies to continue assessing the risk of obesity in patients who receive chemotherapy agents without adjunct radiation therapy.

Although previous studies have suggested a number of potential mechanisms underlying the increased rates of obesity among ALL survivors, some controversy remains regarding which of these risk factors is the primary culprit. Some investigations have shown that obesity is associated with exposure to corticosteroids.3, 23, 25 However, conventional responses to steroids, such as increased energy intake, have not been observed in these cancer patients.26 Moreover, excess weight gain consistently occurs long after the completion of ALL therapy, indicating that other factors must contribute to this problem.23 A number of studies have reported that reduced habitual physical activity during and ALL therapy is the primary cause of obesity among children on modern treatment protocols.9, 23, 27, 28 This may be due to a number of factors, including diminished exercise capacity,29 impaired motor function,27 impaired lung function,29 diminished interest in recreational activity,29 and over-protectiveness of the child's primary caregiver.29 Because the current study was restricted to Hispanic patients, it is possible that BMI changes were driven in part by behaviors or responses that are not representative of other ethnic or race groups. In addition, because this cohort was restricted to patients with B-precursor ALL, the current findings are not necessarily generalizable to other subgroups of cancer patients.

Understanding the etiology of obesity in childhood cancer survivors is an important public health concern given the recent epidemic rise in obesity rates in the general U.S. pediatric population. The prevalence of childhood and adolescent overweight and obesity has doubled in the past 2 decades in the U.S.15 and has been particularly high among Hispanics.15 The most recent estimates of obesity prevalence in the U.S., based on the National Health and Nutrition Examination Study (NHANES III) collected in 1999–2000, indicate that, among adolescents (ages 12–19 years), Hispanics had the highest overall prevalence (43.8%) of obesity and obesity-at-risk status combined (BMI ≥ 85%) compared with non-Hispanic whites (26.5%) and non-Hispanic blacks (40.4%). Examining the specific patterns of weight gain in this cohort of predominantly Hispanic patients with ALL represents an important first step in identifying the clinical and behavioral factors that underlie the problem of obesity among ALL survivors. It will be important for future studies to assess specific serum biomarkers associated with weight gain, such as glucose, insulin, lipid profile, and leptin. Learning about how these markers are associated with chemotherapy-related weight gain is crucial for the development of clinical interventions. In addition, it will be important serially to track the extent to which elevated BMI persists after the completion of chemotherapy. Understanding the long-term patterns of BMI changes over time will provide greater insight into the etiology of this important late effect.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES