SEARCH

SEARCH BY CITATION

Keywords:

  • survivor;
  • body mass index;
  • childhood cancer;
  • underweight

Abstract

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

BACKGROUND

The goals of the current study were to determine the distribution of body mass index (BMI) of survivors of common pediatric malignancies and to identify factors associated with abnormal BMI.

METHODS

The Childhood Cancer Survivor Study (CCSS) is a multicenter cohort study of ≥ 5-year survivors of pediatric cancer diagnosed between 1970 and 1986. Self-reported heights and weights were used to calculate BMI for 7195 adult survivors, compared with population-based, age-specific, and gender-specific norms from the 1995 National Health Interview Survey. Underweight was defined as a BMI < 18.5 kg/m2 and obese as BMI ≥ 30 kg/m2.

RESULTS

Survivors of leukemia were more likely to be obese (females: odds ratio [OR] = 1.5; 95% confidence interval [CI], 1.2–1.8; males: OR = 1.2; 95% CI, 1.0–1.5). Survivors more likely to be underweight included female and male survivors of Hodgkin disease (OR = 1.7; 95% CI, 1.3–2.3 and OR = 3.5; 95% CI, 2.3–5.3) and Wilms tumor (OR = 1.8; 95% CI, 1.2–2.8 and OR = 5.5; 95% CI, 3.1–9.7), female survivors of bone carcinoma without amputation (OR = 1.9; 95% CI, 1.2–2.9), and male survivors of leukemia (OR = 2.4; 95% CI, 1.6–3.6), brain tumors (OR = 2.7; 95% CI, 1.6–4.4), non-Hodgkin lymphoma (OR = 3.1; 95% CI, 1.9–5.2), neuroblastoma (OR = 4.9; 95% CI, 2.48–10.0), and soft tissue sarcoma (OR = 3.5; 95% CI, 2.0–6.0). In females, treatment with total body irradiation, alkylating agents, and anthracyclines and in males, treatment with abdominal radiation, younger age at treatment, and treatment with anthracyclines and alkylating agents were associated with being underweight. Underweight survivors were more likely to report adverse health and major medical conditions.

CONCLUSIONS

A significant proportion of childhood survivors of cancer are underweight as adults and the impact of this on the general health of survivors will need to be addressed further. Cancer 2005. © 2005 American Cancer Society.

Among children diagnosed with cancer, 75% can expect to be long- term survivors. Currently, there is an estimated quarter million childhood survivors of cancer residing in the United States, which is nearly 1 in 640 young adults in the general population.1, 2 The health consequences of childhood cancer therapies are beginning to unveil as focus on identification and treatment of late effects becomes standardized in long-term follow-up programs. Alterations in the body mass index (BMI; kg/m2) of survivors of cancer, specifically the presence of obesity, has been reported for survivors of acute lymphoblastic leukemia (ALL) and in some subgroups of pediatric patients with brain tumors.3–12 In addition, the national focus on the epidemic of obesity and consequent obesity-related morbidity and mortality led us to question the BMI distribution of adult survivors of other pediatric malignancies. In the general population, the long-term health consequences of altered BMI have been primarily focused on individuals who are overweight or obese but altered health outcomes for underweight populations are described as well.13, 14

The purpose of the current study was to assess BMI in adult survivors of pediatric cancer across the major oncologic diagnoses using data from the Childhood Cancer Survivor Study (CCSS). Percentages of patients in each BMI category (underweight, normal weight, overweight, and obese) were determined for each type of cancer. Host factors and cancer therapies associated with variations from a normal BMI were identified.

MATERIALS AND METHODS

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

Study Participants

The CCSS, previously described elsewhere,15 is a 26-institution retrospective cohort study tracking the health outcomes of long-term survivors of childhood cancer. Inclusion criteria for the CCSS were limited to individuals who were ≤ 20 years at diagnosis and who were diagnosed between 1970 and 1986. Eligible participants included children who survived ≥ 5 years after diagnosis of leukemia, primary brain tumor, Hodgkin disease (HD), non-Hodgkin lymphoma (NHL), Wilms tumor, neuroblastoma, soft tissue sarcoma, or malignant bone tumor, and who received their primary treatment at one of the collaborating institutions. Protocols, questionnaires, and documents for human subject participation were approved by institutional review boards at the University of Minnesota (the study coordinating center) and at each collaborating institution. Informed consent for study participation was obtained from each eligible participant. Separate consent was obtained to allow release and abstraction of medical and treatment records. Among the 20,602 five-year survivors identified by the collaborating institutions, at the time of this analysis, 14,370 were enrolled and completed a baseline questionnaire, 3197 declined to participate, and 3035 were lost to follow-up and were never offered enrollment. Treatment records were obtained for 13,134 (91%) of the 14,370 participating survivors. The analysis was restricted to the 7195 survivors of cancer with treatment information who were ≥ 20 years and alive at the time of the baseline questionnaire assessment.

Data Collection

Participants completed a baseline survey questionnaire covering demographic characteristics, health habits, and medical conditions. They were asked to report their current height and weight. The baseline survey is available for review at www.cancer.umn.edu/ccss. Baseline surveys were completed primarily in 1995 and 1996 either by mailed questionnaire or by telephone with a trained interviewer. Cancer-related information, including morphology and treatment for the original cancer, was obtained from each treating institution for all eligible participants who returned a signed medical release. The medical record abstraction form can also be viewed at www.cancer.umn.edu/ccss. For comparative analyses, we calculated self-reported age-specific and gender-specific population norms for height, weight, and BMI from the 1995 National Health Interview Survey (NHIS). The NHIS is conducted annually by the National Center for Health Statistics of the Centers for Disease Control and Prevention.16 The 1995 NHIS included interviews with 102,467 persons, including 40,899 who were aged 20–47 years17 to correspond with the age range at baseline questionnaire of the CCSS participants included in this analysis. In addition, BMI distribution comparisons were made in survivor-sibling matched pairs. In the analyses, data were adjusted for gender and age. There were 1870 survivors in our analysis with a sibling match who reported both their height and weight with 868 of them being male (955 male survivors) and 1002 being female (915 female survivors).

Outcome Measures

BMI was calculated by dividing self-reported weight in kilograms by height in meters squared. BMI was classified according to the standards of the National Heart, Lung, and Blood Institute and the World Health Organization as follows: 1) < 18.5 kg/m2 for individuals of underweight, 2) 18.5–24.9 kg/m2 for individuals of normal weight, 3) 25–29.9 kg/m2 for individuals of overweight, and 4) ≥ 30 kg/m2 for obese individuals.18, 19 Because BMIs were not corrected for loss of limb, survivors with an extremity amputation were not included in the analysis. Both underweight status and obesity were evaluated as dichotomous outcomes.

Independent Variables

Patient and disease characteristics that were examined included age at cancer diagnosis, cancer morphology, radiation site, chemotherapy exposure, and current medical and health status. All analyses were stratified by gender and adjusted for age at the time of baseline questionnaire. A radiation physicist evaluated diagrams and photographs taken in the treatment position to determine exposed body regions. If diagrams were not available, a written description from the medical record was used to estimate the exposed regions. Radiation sites considered for these analyses included the brain or head, spine, abdomen, pelvis, chest, entire body, or an unspecified site. Five groupings of chemotherapy previously defined for the current cohort were considered as potential independent risk factors of being underweight (BMI < 18.5 kg/m2) or obese (BMI ≥ 30 kg/m2) based on quantitative information abstracted from the medical records. The chemotherapy categories considered in these analyses were 1) anthracyclines, 2) alkylating agents, 3) anthracyclines and alkylating agents used in combination, 4) other chemotherapeutic agents, or 5) no chemotherapy.

As previously reported by Hudson et al.,20 identification of major medical conditions and characterization of states of adverse health were used to describe the overall health of CCSS participants. Individuals were classified as having a major medical condition if they reported current use of a selected medication (anticonvulsants, cardiovascular medications, or chemotherapy/immune suppressants) or the presence of selected medical conditions (complete deafness, renal failure requiring dialysis, congestive heart failure, myocardial infarction, stroke or cerebrovascular accident, current use of oxygen, cirrhosis, coronary artery bypass surgery, angioplasty, heart transplant, lung transplant, kidney transplant, repeated seizures, convulsions or blackouts, confirmed new malignancy excluding basal cell carcinoma or disease recurrence, amputation, or a joint replacement).20 Individuals who reported one of the following were classified as having adverse health status: fair or poor overall health, needing help with personal cares or routine activities, whose health prevented working, the inability to walk one block or participate in moderate physical activity, continued pain as a result of their cancer, continued medium to extreme anxiety or fear as the result of their cancer, or individuals who scored in the highest 10% of population norms on any scale of the Brief Symptom Inventory (a measure of mental health).

Data Analysis

The frequencies and percentages of each independent and dependent variable were calculated for the study population, both overall and stratified by gender. The frequency and percentage of persons in each category of BMI were calculated and stratified by gender and cancer type. The relative odds of having a BMI < 18.5 kg/m2 or ≥ 30 kg/m2 among survivors of cancer when compared with population norms from the 1995 NHIS were calculated separately for males and females, and matched to age in years. Initial analysis included stratification by 10-year age groups. however, the age-specific odds ratios (OR) in the 10-year age groups did not differ significantly from each other, so the multivariate analyses were adjusted for age. Chi-square tests were used to evaluate differences between survivors of cancer and the NHIS population sample. The relative risk (RR; hazard ratio [HR]) of having a BMI < 18.5 kg/m2 and ≥ 30 kg/m2 among survivors of cancer when compared with their matched sibling pairs were evaluated using conditional logistic regression, adjusting for age and gender.

For each of the dichotomous outcome variables of BMI < 18.5 kg/m2 and ≥ 30 kg/m2 logistic regression analyses were conducted to simultaneously account for the effects of age at diagnosis, race/ethnicity, radiation site, and chemotherapy group. These models were stratified by gender and adjusted for age at baseline questionnaire. Multivariate logistic regression models were also calculated separately for each cancer category, again stratified by gender for both BMI < 18.5 kg/m2 and ≥ 30 kg/m2. Initial multivariate models for these stratified analyses included variables for age at diagnosis, race/ethnicity, radiation site, and type of chemotherapy, and were adjusted for age at baseline questionnaire. Models were reduced using stepwise regression and by comparing nested −2log likelihood to determine the best fit. Only the final models are presented in the results. Data were analyzed with SAS version 8.2 (SAS Institute, Cary, NC).

RESULTS

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

Subject Characteristics

Demographic, cancer, and treatment characteristics of the study participants are shown in Table 1. Individuals who were diagnosed with ALL, HD, or a brain tumor accounted for 56% of the study population. Forty-three percent of the participants were < 10 years at diagnosis, and 63% completed the baseline survey before age 30 years. Anthracyclines with or without alkylating agents were received by 52% of the subjects.

Table 1. Demographic, Cancer, and Treatment Characteristics of the Study Participants
CharacteristicsNo. of females (N = 3456) (%)No. of males (N = 3739) (%)Total no. of patients (N = 7195) (%)
Histology   
 Acute lymphoblastic leukemia828 (24.0)837 (22.4)1665 (23.1)
 Other leukemia163 (4.7)143 (3.8)306 (4.2)
 Brain tumors433 (12.5)507 (13.6)940 (13.1)
 Hodgkin disease689 (19.9)713 (19.1)1402 (19.5)
 Non-Hodgkin lymphoma207 (6.0)452 (12.1)660 (9.2)
 Wilms tumor228 (6.6)184 (4.9)412 (5.7)
 Neuroblastoma132 (3.8)117 (3.1)249 (3.5)
 Soft tissue sarcomas347 (10.0)357 (9.5)704 (9.8)
 Bone malignancies   
  Amputation215 (6.2)228 (6.1)443 (6.2)
  No amputation214 (6.2)201 (5.4)415 (5.8)
Age at diagnosis (yrs)   
 ≤ 4720 (20.8)720 (19.3)1440 (20.0)
 5–9735 (21.3)891 (23.8)1626 (22.6)
 10–141071 (31.0)1138 (30.4)2209 (30.7)
 15–20930 (26.9)990 (26.5)1920 (26.7)
Age at interview (yrs)   
 20–292128 (61.6)2398 (64.1)4526 (62.9)
 30–391191 (34.5)1194 (31.9)2385 (33.1)
 40–47137 (4.0)147 (3.9)284 (3.9)
Race/ethnicity   
 White, non-Hispanic2958 (85.6)3209 (85.8)6167 (85.7)
 Black, non-Hispanic113 (3.3)103 (2.7)216 (3.0)
 Hispanic149 (4.3)113 (3.0)262 (3.6)
 Other, not specified236 (6.8)314 (8.4)550 (7.6)
Radiation treatment site   
 Brain1015 (30.2)1132 (31.0)2147 (30.6)
 Other head207 (6.1)285 (7.8)492 (7.0)
 Spine222 (6.6)274 (7.5)496 (7.1)
 Chest888 (26.4)885 (24.3)1773 (25.3)
 Abdomen711 (21.1)746 (20.5)1457 (20.8)
 Pelvis492 (14.6)644 (17.7)1136 (16.2)
 Total body irradiation49 (1.5)53 (1.4)102 (1.4)
 Unspecified radiation site87 (2.5)79 (2.1)166 (3.2)
 Radiation status unknown5 (0.1)13 (0.1)18 (0.2)
 No radiation948 (27.3)994 (26.5)1942 (27.0)
Chemotherapy included   
 Anthracycline770 (22.3)891 (23.8)1661 (23.1)
 Alkylating agent296 (8.6)298 (8.0)594 (8.3)
 Anthracycline and alkylating agent963 (27.9)1116 (29.8)2079 (28.9)
 Other chemotherapy607 (17.6)587 (15.7)1194 (16.6)
 No chemotherapy820 (23.7)847 (22.6)1667 (23.2)
Adverse health status   
 Yes1340 (35.8)1537 (44.5)2877 (40.0)
 No2146 (57.4)1733 (50.1)3879 (53.9)
 Unknown253 (6.8)186 (5.4)439 (6.1)
Major medical condition   
 Yes858 (25.2)901 (26.1)1760 (24.5)
 No2518 (67.3)2221 (64.3)4739 (65.9)
 Unknown362 (9.7)334 (9.7)696 (9.7)
Current smoker   
 Yes756 (20.2)586 (16.9)1342 (18.6)
 No2975 (79.6)2865 (82.9)5840 (81.2)
 Unknown8 (0.2)5 (0.2)13 (0.2)
Body mass index   
 < 18.5298 (8.6)152 (4.1)450 (6.2)
 18.5–24.92002 (57.9)1828 (48.9)3830 (53.2)
 25–29.9692 (20.0)1299 (34.7)1991 (27.7)
 ≥ 30464 (13.4)460 (12.3)924 (12.8)

Table 2 provides the gender-specific BMI distribution of adult survivors of pediatric cancers in the 4 standard categories: underweight BMI < 18.5 kg/m2, normal weight BMI 18.5–24.9 kg/m2, overweight BMI 25–29.9 kg/m2, or obese BMI ≥ 30 kg/m2. Data are presented for males and females in each of the major cancer categories represented in the CCSS with comparisons made to age-matched and gender-matched groups in the general population. A smaller percentage of survivors of childhood cancer were classified as obese in each category, except for females treated for ALL or brain tumors. For most cancer types, however, survivors were more likely to be underweight than would be expected from population norms. Table 3 shows the ORs and associated 95% confidence intervals (95% CI) of either having a BMI < 18.5 kg/m2 or ≥ 30 kg/m2 relative to same-gender, same-age population norms. Only survivors of ALL were at increased risk of obesity (OR = 1.5; 95% C, 1.2–1.8 for females and OR = 1.2; 95% CI, 1.0–1.5 for males). Survivors found to have an abnormally low BMI (< 18.5 kg/m2) were females who had been treated for HD (OR = 1.7; 95% CI, 1.3–2.3), Wilms tumor (OR = 1.8; 95%CI, 1.2–2.8), and bone malignancies without amputations (OR = 1.9; 95%CI, 1.2–2.9). Male survivors at increased risk for being underweight included those treated for ALL (OR = 2.4; 95% CI, 1.6–3.6), brain tumors (OR = 2.7; 95% CI, 1.6–4.4), HD (OR = 3.5; 95% CI, 2.3–5.3), NHL (OR = 3.1; 95% CI, 1.9–5.2), Wilms tumor (OR = 5.5; 95% CI, 3.1–9.7), neuroblastoma (OR = 4.9; 95% CI, 2.4–10.0), and soft tissue sarcoma (OR = 3.5; 95% CI, 2.0–6.0).

Table 2. Distribution of BMI by Cancer Type
CharacteristicsBMI < 18.5 Underweight No. (%)BMI 18.5–24 Normal weight No. (%)BMI 25–29 Overweight No. (%)BMI ≥ 30 Obese No. (%)
  • BMI: body mass index.

  • a

    Age and gender-specific BMI prevalence data were derived from the 1995 National Health Interview Survey.

Females    
 General populationa1003 (4.7)11851 (55.4)5008 (23.4)3531 (16.5)
 Survivors of cancer    
  Acute lymphoblastic leukemia35 (4.2)454 (54.8)186 (22.5)153 (18.5)
  Other leukemia12 (7.4)87 (53.4)39 (23.9)25 (15.3)
  Brain tumors35 (8.1)216 (49.9)108 (24.9)74 (17.1)
  Hodgkin disease56 (8.1)423 (61.4)123 (17.8)87 (12.6)
  Non-Hodgkin lymphoma18 (8.7)120 (58.0)54 (26.1)15 (7.2)
  Wilms tumor28 (12.3)157 (68.9)24 (10.5)19 (8.3)
  Neuroblastoma14 (10.6)88 (66.7)17 (12.9)13 (9.8)
  Soft tissue sarcomas27 (7.8)216 (62.2)66 (19.0)38 (10.9)
  Bone malignancies    
   Amputation52 (24.2)100 (46.5)40 (18.6)23 (10.7)
   No amputation21 (9.8)141 (65.9)35 (16.4)17 (7.9)
Males    
 General populationa181 (0.9)7729 (39.6)8347 (42.8)3249 (16.7)
 Survivors of cancer    
  Acute lymphoblastic leukemia29 (3.5)367 (43.8)303 (36.2)138 (16.5)
  Other leukemia3 (2.1)71 (49.6)52 (36.4)17 (11.9)
  Brain tumors18 (3.5)229 (45.2)193 (38.1)67 (13.2)
  Hodgkin disease26 (3.6)342 (48.0)253 (35.5)92 (12.9)
  Non-Hodgkin lymphoma17 (3.8)224 (49.6)165 (36.5)46 (10.2)
  Wilms tumor15 (8.1)105 (57.1)49 (26.6)15 (8.1)
  Neuroblastoma9 (7.7)71 (60.7)31 (26.5)6 (5.1)
  Soft tissue sarcomas15 (4.2)193 (54.1)107 (30.0)42 (11.8)
  Bone malignancies    
  Amputation15 (6.6)103 (53.9)76 (33.3)14 (6.1)
  No amputation5 (2.5)103 (51.2)70 (34.8)23 (11.4)
Table 3. Relative Odds of Being Underweight or Obese in Adult Survivors of Childhood Cancer Compared with Population Norms, Adjusted for Age in Years
CharacteristicsBMI < 18.5BMI ≥ 30
OR95% CIP valueOR95% CIP value
  1. BMI: body mass index; OR: odds ratio; 95% CI: 95% confidence interval.

Females      
 Acute lymphoblastic leukemia0.60.4–0.90.0061.51.2–1.80.001
 Other leukemia1.20.6–2.10.581.10.7–1.70.59
 Brain tumors1.30.9–1.90.131.31.0–1.60.06
 Hodgkin disease1.71.3–2.30.0010.80.6–1.00.02
 Non-Hodgkin lymphoma1.60.9–2.50.090.50.3–0.80.004
 Wilms tumor1.81.2–2.80.0030.60.4–1.00.04
 Neuroblastoma1.50.9–2.70.150.80.4–1.30.34
 Soft tissue sarcomas1.40.9–2.10.090.70.5–1.00.05
 Bone malignancies      
  No amputation1.91.2–2.90.0080.50.3–0.80.006
Males      
 Acute lymphoblastic leukemia2.41.6–3.60.0011.21.0–1.50.02
 Other leukemia1.60.5–5.00.450.80.5–1.30.42
 Brain tumors2.71.6–4.40.0010.90.7–1.20.50
 Hodgkin disease3.52.3–5.30.0010.80.5–1.00.06
 Non-Hodgkin lymphoma3.11.9–5.20.0010.70.5–0.90.007
 Wilms tumor5.53.1–9.70.0010.60.3–1.00.04
 Neuroblastoma4.92.4–10.00.0010.40.1–0.80.01
 Soft tissue sarcomas3.52.0–6.00.0010.80.5–1.10.11
 Bone malignancies      
  No amputation2.10.9–5.30.100.70.5–1.10.17

Table 4 shows the ORs and 95% CIs from the multivariate analysis that included age at diagnosis, race/ethnicity, radiation site, and chemotherapy as predictor variables for either being underweight or obese among all survivors. The analysis was adjusted for age at baseline questionnaire and stratified by gender. Total body irradiation (OR = 2.4; 95% CI, 1.2–5.0) and the use of alkylating agents with or without anthracyclines (OR = 1.8; 95% CI, 1.1–2.7 and OR = 1.4; 95% CI, 1.0–2.0, respectively) were associated with an abnormally low BMI in females. Among males, young age at diagnosis (OR = 2.3; 95% CI, 1.2–4.4 at age < 4 years), abdominal radiation (OR = 1.8; 95% CI, 1.1–2.9), and alkylating agents with anthracyclines (OR = 2.2; 95% CI, 1.3–3.6) were associated with an elevated risk of being underweight. Univariate models for age at diagnosis, race/ethnicity, radiation site, and chemotherapy group were conducted separately for lymphoma, Wilms tumor, neuroblastoma, soft tissue sarcoma, and bone malignancies to determine significant predictors in a multivariate model with BMI < 18.5 kg/m2 as the outcome variable. The results of the multivariate models are shown by cancer type in Table 5. Among female survivors of lymphoma, only age 5–9 at diagnosis, when compared with age 15–20 years at diagnosis, increased the odds of being underweight (OR = 2.5; 95% CI, 1.1–5.3). Among male survivors of lymphoma, younger age at diagnosis was associated with a low BMI and the younger the age at diagnosis of lymphoma, the greater the risk of being underweight as an adult. Being underweight as an adult male survivor of lymphoma was also associated with radiation to the abdomen or pelvis (OR = 2.7; 95% CI, 1.3–5.4), and with a combination of anthracyclines and alkylating agents (OR = 2.6; 95% CI, 1.0–6.6). Among survivors of Wilms tumor, females who received chest radiation had a 2.5-fold increased risk of being underweight (95% CI, 1.0–5.8) when compared with females who were not treated with chest radiation. Male survivors of soft tissue sarcoma who received pelvic radiation were 4.6 times (95% CI, 1.4–15.1) more likely to be underweight than those who did not receive pelvic radiation. Female survivors of soft tissue sarcomas who received a combination of anthracyclines and alkylating agents were 6.7 times as likely (95% CI, 1.5–30.4) as those who received no chemotherapy to be underweight. There were no significant predictors of a low BMI among either survivors of neuroblastoma or bone cancer survivors.

Table 4. Multivariable ORs for Being Underweight or Obese among Adult Survivors of Childhood Cancera
CharacteristicsBMI < 18.5 UnderweightBMI ≥ 30 Obese
FemalesMalesFemalesMales
OR95% CIP valueOR95% CIP valueOR95% CIP valueOR95% CIP- value
  • BMI: body mass index; OR: odds ratio; 95% CI: 95% confidence interval.

  • a

    Analysis was adjusted for age at baseline.

Age at diagnosis (yrs)            
 ≤ 40.90.5–1.50.672.31.2–4.40.011.41.0–2.20.081.10.7–1.60.67
 5–91.00.7–1.60.841.50.8–2.80.171.81.3–2.6< 0.0011.20.8–1.60.38
 10–141.30.9–1.90.121.40.8–2.40.261.10.8–1.60.351.00.7–1.30.76
 15–201.0Reference1.0Reference1.0Reference1.0Reference
Race/ethnicity            
 White, non-Hispanic1.0Reference1.0Reference1.0Reference1.0Reference
 Black, non-Hispanic0.60.3–1.30.190.80.3–2.30.742.61.6–4.0< 0.0011.10.6–2.00.77
 Hispanic0.80.4–1.40.451.90.9–3.90.071.10.7–1.90.582.01.2–3.10.004
 Other, not specified1.00.6–1.50.900.80.4–1.50.421.10.7–1.60.580.90.6–1.20.41
Radiation site            
 Brain0.50.4–0.7< 0.0010.90.6–1.40.721.91.5–2.3< 0.0011.31.0–1.60.02
 Abdomen1.10.8–1.60.581.81.1–2.90.010.80.6–1.20.470.80.5–1.10.12
 Pelvis1.20.8–1.90.301.20.7–1.90.470.70.4–1.10.100.90.6–1.30.56
 Total body irradiation2.41.2–5.00.011.80.6–5.40.270.50.1–1.60.231.00.4–2.50.94
Chemotherapy            
 Anthracycline1.20.8–1.70.421.50.9–2.70.121.00.7–1.30.821.10.8–1.40.57
 Alkylating agent1.81.1–2.70.011.50.7–3.20.240.80.6–1.20.380.90.6–1.40.73
 Anthracycline and alkylating agent1.41.0–2.00.032.11.3–3.60.0050.60.5–0.80.0020.80.6–1.10.25
 Other chemotherapy0.70.4–1.20.200.80.4–1.60.640.90.7–1.30.721.00.7–1.40.91
 No chemotherapy1.0Reference1.0Reference1.0Reference1.0Reference
Table 5. Cancer and Gender-Specific Multivariable OR for Being Underweight among Adult Survivors of Childhood Cancera
CharacteristicsBMI < 18.5
FemalesMales
OR95% CIP valueOR95% CIP value
  • BMI: body mass index; OR: odds ratio; 95% CI: 95% confidence interval.

  • a

    Analysis was adjusted for age at baseline questionnaire.

Lymphoma      
 Age at diagnosis (yrs)      
  ≤ 41.20.3–4.80.8212.42.9–52.80.001
  5–92.51.1–5.30.025.61.9–16.40.002
  10–141.00.5–1.80.882.51.0–6.60.05
  15–201.0Reference1.0Reference
 Radiation site      
  Abdomen or pelvis1.00.6–1.80.872.71.3–5.40.005
  Chest0.90.5–1.60.681.00.5–2.00.99
 Chemotherapy      
  Anthracycline1.00.5–1.80.951.10.4–2.80.82
  Alkylating agent0.60.1–5.20.673.30.6–17.70.16
  Anthracycline and alkylating agent1.00.5–2.00.992.61.0–6.60.05
  Other chemotherapy1.00.2–4/90.991.50.1–13.20.74
  No chemotherapy1.0Reference1.0Reference
Wilms tumor      
 Radiation site      
  Pelvis2.40.9–6.30.073.10.8–11.40.09
  Abdomen1.00.3–3.80.980.60.1–2.90.55
  Chest2.51.0–5.80.041.20.3–4.30.83
  No radiation1.0Reference1.0Reference
Neuroblastoma      
 Radiation site      
  Pelvis2.70.6–11.90.190.60.1–6.30.67
  Abdomen1.60.4–6.80.523.30.7–14.20.13
  Chest3.10.9–10.30.074.30.9–20.10.06
  No radiation1.0Reference1.0Reference
Soft tissue sarcomas      
 Radiation site      
  PelvisNot included in final model4.61.4–15.10.01
  AbdomenNot included in final model0.50.1–4.40.50
  ChestNot included in final model3.40.8–13.90.09
  No radiationNot included in final model   
 Chemotherapy      
  Anthracycline3.90.8–19.20.09Not included in final model
  Alkylating agent0.00.0–99.90.98Not included in final model
  Anthracycline and alkylating agent6.71.5–30.40.01Not included in final model
  Other chemotherapy0.00.0–99.90.97Not included in final model
  No chemotherapy1.0ReferenceNot included in final model
Bone malignancies      
 Chemotherapy      
  Anthracycline1.90.5–7.30.340.00.0–99.90.97
  Alkylating agent2.60.9–7.40.080.70.1–4.30.68
  Anthracycline and alkylating agent1.50.6–4.20.400.90.2–4.20.89
  Other chemotherapy4.40.9–22.60.070.00.0–99.90.98
  No chemotherapy1.0Reference1.0Reference

To evaluate the possibility that either underlying disease or cigarette smoking was associated with being underweight, we performed multivariate models adjusted for age at baseline questionnaire to simultaneously account for adverse health status, major medical condition, and smoking status. Among males, the risk of being underweight was increased by 60% (OR = 1.6; 95% CI, 1.0–2.6) when comparing smokers with non-smokers, by 80% (OR = 1.8; 95% CI, 1.1–2.9) when comparing those with a major medical condition with those without a major medical condition, and by 150% (OR = 2.5; 95% CI, 1.6–3.8) when comparing those with adverse health status to those without adverse health status. Among females, only a concomitant major medical condition was associated with being underweight (OR = 1.8; 95% CI, 1.3–2.5). Adverse health status and major medical conditions were not associated with cancer therapies that included alkylating agents or anthracyclines, total body irradiation, or abdominal radiation (range of correlation coefficients = −0.03–0.08). We could not directly assess whether the intensity of these therapies inhibited weight gain during childhood or adolescence, or affected future weight gain. Further study will be needed to better understand these potential relations.

Finally, we studied the influence of familial tendency toward low, normal, overweight, and obese BMI to explain the variation noted in the CCSS subjects. Conditional logistic regression was used to estimate the RR (HR) of being obese or underweight in the matched pairs, adjusting for age and gender. Survivors are less likely (HR = 0.73, P = 0.008) to be obese and more likely (HR = 2.8, P < 0.001) to be underweight than their siblings.

DISCUSSION

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

Utilizing the large database of the CCSS, we are able to describe the distribution of BMI in 7195 adult survivors of major types of childhood cancer. The BMI categories we used were underweight BMI < 18.5 kg/m2, normal weight 18.5–24.9 kg/m2, overweight 25–29.9 kg/m2, and obese ≥ 30 kg/m2. Comparisons were made to BMIs of a large national population-based sample of adults of similar age. With the national prevalence of obesity increasing and the previous reports of obesity in survivors of ALL and brain tumors, we had a priori anticipated increased rates of obesity among adult survivors of other cancers as well. The only survivors in this large multicenter cohort found to be at risk for a BMI ≥ 30 kg/m2 were female patients treated for ALL. The data on subjects with ALL in the CCSS were previously published.4 Obesity among patients with ALL in this cohort was associated with cranial radiotherapy ≥ 20 gray (Gy). When the subgroup of male patients with ALL treated with ≥ 20 Gy was analyzed, they were found to be at increased risk for obesity (OR = 1.86; 95% CI, 1.33–2.57). The subjects with ALL at highest risk for obesity were female patients diagnosed at ≤ 4 years who received ≥ 20 Gy of cranial radiation. The BMI data on survivors of brain tumors in the CCSS cohort have also been previously reported.21 It is noteworthy that subjects with craniopharyngioma were excluded from the cohort. BMI distribution in survivors of brain tumors did not differ from that of population norms. Female patients with brain tumors who were diagnosed at a younger age and treated with cranial irradiation were at risk for obesity. It is important to realize that nearly two-thirds of the CCSS survivors reported in the current study were in their 20s and one-third in their 30s at the time of the questionnaire. Their relatively older age may have an effect on the distribution of their BMI compared with survivors who are assessed closer to the time of diagnosis.

From the remainder of the CCSS cohort, female survivors of HD, Wilms tumor, and bone carcinoma and male survivors of leukemia, brain tumor, HD, NHL, Wilms tumor, neuroblastoma, and soft tissue sarcoma were more likely to be underweight as adults. Survivors of childhood cancer were more likely to be underweight than their siblings, arguing against familial factors as an explanation for our results. The finding of increased rates of underweight survivors directs the focus of follow-up programs to the assessment of adequate and appropriate nutrition in survivors of childhood cancer and counseling for both abnormally high and low BMIs.

The health consequences of obesity have been documented in the adult and pediatric literature and include hypertension, type 2 diabetes, cardiovascular disease, hypercholesterolemia, sleep disorders, and musculoskeletal disorders.22, 23 Increased rates of morbidity and mortality have been reported in patients with abnormally high BMIs.24–27 Similar concerns have been identified in pediatric oncology patients as described in a previous CCSS report which identified the need to monitor comorbid cardiovascular risk factors in obese survivors of childhood ALL.4

In the general population, the clinical implications of being underweight are controversial. All-cause mortality is increased across all adult age categories for those with a BMI < 18.5 kg/m2.13 This increase in mortality risk has generally been attributed to smoking or underlying serious disease, either diagnosed or undiagnosed. In a recent prospective study of more than 1 million adults in the United States, Calle et al.13 reported that leanness (BMI < 18.5 kg/m2) was most strongly associated with an increased risk of death among current or former smokers with a history of disease. However, among 30–64-year-old participants, the risk of death was increased even in those who did not smoke or report an underlying disease. Katzmarzyk et al.14 reported similar findings from the Canadian Fitness Study.

We assessed three factors in our study that may contribute to the likelihood of a survivor being underweight. Of those who were underweight, 55% had moderate to extreme adverse health status, 39% reported a major medical condition, and 22% were current smokers. Nearly one-third (29%) of the underweight survivors did not report any of these 3 conditions or habits. Thus, the majority of underweight survivors have an underlying problem that may be contributing to their leanness that should be identified and treated. Cancer treatment modalities and their association with adverse health, major medical conditions, and being underweight will be studied further in follow-up surveys of the CCSS population. Because of the cross-sectional nature of the current study, we were not able to assess whether being underweight subsequently increased the risk of adverse health status or a major medical condition or vice versa. Continued longitudinal follow-up of the CCSS cohort will provide some insights into these relations. However, prospective study from the time of diagnosis is needed to determine the timing of the outcomes and to explore mechanisms of causation.

These data from the CCSS suggest that both underweight and overweight survivors need to be monitored for ongoing health consequences of abnormal BMI. Emphasis on balanced nutrition, proactive instruction for exercise and avoidance of sedentary lifestyles, the prevention of osteopenia, and optimization of health should begin soon after the diagnosis of childhood cancer, continue through long-term follow up, and extend into adulthood.28

The current study has several methodologic limitations. First, heights and weights for the calculation of BMI were self-reported. Self-reported values have, in general, been reasonably accurate.29, 30 In the current report, height and weight values were self-reported in both the survivors and the national comparison group, thus minimizing the likelihood of differential measurement bias. Second, BMI is recognized as only a screening mechanism to assess adequacy of nutrition and additional studies with more sophisticated measurement may be warranted. More accurate assessments of body composition such as percent body fat by dual-energy X-ray absorptometry would be desirable but are unavailable in the CCSS questionnaire format. In addition, it would be of interest to assess the composition and amount of calories consumed in the various BMI subgroups of the CCSS cohort. Third, it is possible that those who elected not to respond to the questionnaire may have been of better health and may have had a higher, more normal, BMI than the participants in the current study. In addition, the underrepresentation of minorities in the CCSS cohort makes generalization of these data to minority survivor populations tenuous. Finally, the results of our study, which represent the outcome of children treated in the 1970s and early 1980s, may not be applicable to survivors of current, often more aggressive therapies.

In summary, adult survivors of childhood cancer are at risk for abnormal BMI. Survivors should be monitored for variations in BMI, and those with high and low BMI evaluated for associated health consequences related to altered BMI. The American Cancer Society has developed specific guidelines on diet, nutrition, and cancer prevention for the maintenance of health in the general population as well as in survivors of cancer. Recommendations include a diet with five or more servings of fruits and vegetables, whole grain foods instead of processed refined foods, and limited consumption of red meat, foods high in fat content, and alcoholic beverages. In addition, adults and children should engage in 30–60 minutes of moderate to vigorous activity ≥ 5 days a week. For our survivors of childhood cancer, all members of the health team, primary care physicians, and subspecialists should encourage a diet of healthful foods, a physically active lifestyle, and the maintenance of a healthful weight throughout life.31, 32

Acknowledgements

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

The authors acknowledge and thank the participating CCSS investigators and institutions: Arthur Ablin, M.D. (Institutional Principal Investigator), University of California-San Francisco, CA; Robert Castleberry M.D. (Institutional Principal Investigator), Roger Berkow M.D. (Former Institutional Principal Investigator), University of Alabama, Birmingham, AL; John Boice, Sc.D. (CCSS Steering Committee Member), International Epidemiology Institute, Rockville, MD; Norman Breslow, Ph.D. (CCSS Steering Committee Member), University of Washington, Seattle, WA; George R. Buchanan, M.D. (Institutional Principal Investigator), Kevin Oeffinger, M.D. (CCSS Steering Committee Member), UT-Southwestern Medical Center at Dallas, TX; Stella Davies, M.D., Ph.D. (CCSS Steering Committee Member), Cincinnati Children's Hospital Medical Center, Cincinnati, OH; Lisa Diller, M.D. (Institutional Principal Investigator), Holcombe Grier, M.D. (Former Institutional Principal Investigator), Frederick Li, M.D. (CCSS Steering Committee Member), Dana-Farber Cancer Institute, Boston, MA; Zoann Dreyer, M.D. (Institutional Principal Investigator), Texas Children's Center, Houston, TX; Debra Friedman, M.D., M.P.H. (Institutional Principal Investigator), Thomas Pendergrass, M.D. (Former Institutional Principal Investigator), Children's Hospital and Medical Center, Seattle, WA; Daniel M. Green, M.D. (Institutional Principal Investigator; CCSS Steering Committee Member), Roswell Park Cancer Institute, Buffalo, NY; Mark Greenberg, MB, Ch.B. (Institutional Principal Investigator), Hospital for Sick Children, Toronto, ON, Canada; Robert Hayashi, M.D. (Institutional Principal Investigator), Teresa Vietti, M.D. (Former Institutional Principal Investigator), St. Louis Children's Hospital, St. Louis, MO; Melissa Hudson, M.D. (Institutional Principal Investigator, CCSS Steering Committee Member), St. Jude Children's Research Hospital, Memphis, TN; Raymond Hutchinson, M.D. (Institutional Principal Investigator), University of Michigan, Ann Arbor, MI; Neyssa Marina, M.D. (Institutional Principal Investigator), Michael P. Link, M.D. (Former Institutional Principal Investigator), Sarah S. Donaldson, M.D. (CCSS Steering Committee Member), Stanford University School of Medicine, Stanford, CA; Lillian Meacham, M.D. (Institutional Principal Investigator), Emory University, Atlanta, GA; Anna Meadows, M.D. (Institutional Principal Investigator, CCSS Steering Committee Member), Bobbie Bayton (CCSS Steering Committee Member), Children's Hospital of Philadelphia, PA; John Mulvihill, M.D. (CCSS Steering Committee Member), Children's Hospital, Oklahoma City, OK; Brian Greffe, M.D. (Institutional Principal Investigator), Lorrie Odom, M.D. (Former Institutional Principal Investigator), Children's Hospital, Denver, CO; Maura O'Leary, M.D. (Institutional Principal Investigator), Children's Health Care-Minneapolis, MN; Amanda Termuhlen, M.D. (Institutional Principal Investigator), Frederick Ruymann, M.D. (Former Institutional Principal Investigator), Stephen Qualman, M.D. (CCSS Steering Committee Member), Columbus Children's Hospital, Columbus, OH; Gregory Reaman M.D. (Institutional Principal Investigator), Roger Packer M.D. (CCSS Steering Committee Member), Children's National Medical Center, Washington, DC; A. Kim Ritchey, M.D. (Institutional Principal Investigator), Julie Blatt, M.D. (Former Institutional Principal Investigator), Children's Hospital of Pittsburgh, Pittsburgh, PA; Leslie L. Robison, Ph.D. (Institutional Principal Investigator, CCSS Steering Committee Member), Ann Mertens, Ph.D. (CCSS Steering Committee Member), Joseph Neglia, M.D., MPH (CCSS Steering Committee Member), Mark Nesbit, M.D. (CCSS Steering Committee Member), University of Minnesota, Minneapolis, MN; Kathy Ruccione, RN, MPH (Institutional Principal Investigator), Children's Hospital Los Angeles, Los Angeles, CA; Charles Sklar, M.D. (Institutional Principal Investigator, CCSS Steering Committee Member), Memorial Sloan-Kettering Cancer Center, New York, NY; Barry Anderson, M.D. (CCSS Steering Committee Member), Peter Inskip, Sc.D. (CCSS Steering Committee Member), National Cancer Institute, Bethesda, M.D.; Vilmarie Rodriguez, M.D. (Institutional Principal Investigator), W. Anthony Smithson, M.D. (Former Institutional Principal Investigator), Gerald Gilchrist, M.D. (Former Institutional Principal Investigator), Mayo Clinic, Rochester, MN; Louise Strong, M.D. (Institutional Principal Investigator, CCSS Steering Committee Member), Marilyn Stovall, Ph.D. (CCSS Steering Committee Member), U.T.M.D. Anderson Cancer Center, Houston, TX; Terry A. Vik, M.D. (Institutional Principal Investigator), Robert Weetman, M.D. (Former Institutional Principal Investigator), Riley Hospital for Children, Indianapolis, IN; Yutaka Yasui PhD (Institutional Principal Investigator, CCSS Steering Committee Member), John Potter M.D., Ph.D. (Former Institutional Principal Investigator, CCSS Steering Committee Member), Fred Hutchinson Cancer Center, Seattle, WA; Lonnie Zeltzer, M.D. (Institutional Principal Investigator, CCSS Steering Committee Member), University of California-Los Angeles, Los Angeles, CA.

REFERENCES

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