Changes in Body Mass Index in Long-Term Childhood Cancer Survivors

Transcription

1 Changes in Body Mass Index in Long-Term Childhood Cancer Survivors Hanneke M. van Santen, MD, PhD 1,2 ; Ronald B. Geskus, PhD 3 ; Steven Raemaekers, MSc 4 ; A. S. Paul van Trotsenburg, MD, PhD 1 ; Thomas Vulsma, MD, PhD 1 ; Helena J. H. van der Pal, MD, PhD 5,6 ; Hubert N. Caron, MD, PhD 6 ; and Leontien C. M. Kremer, MD, PhD 6 BACKGROUND: Previous studies have reported changes in the body mass index (BMI) with time in childhood cancer survivors (CCSs) during follow-up. The limitations of these studies include that they described only a subgroup of survivors or used questionnaires with self-reported heights and weights. The goal of this study was to examine BMI in a large cohort of long-term CCSs and relate this to the BMI at diagnosis, age, sex, tumor type, treatment, and endocrine defects. METHODS: All patients treated for childhood cancer at the Emma Children s Hospital/Academic Medical Center between 1966 and 1996 who had survived for at least 5 years were eligible for inclusion. For 893 CCSs with a mean follow-up of 14.9 years, the BMI at the late effects outpatient clinic was compared with the BMI for the general Dutch population. RESULTS: For girls, an increased prevalence of obesity was found. Risk factors for developing a high BMI at follow-up were a younger age and a high BMI at diagnosis and treatment with cranial radiotherapy. A significantly increased prevalence of severe underweight was found in all adult subgroups except for females aged 26 to 45 years. An association was found between a low BMI at diagnosis and a low BMI at follow-up. No treatment-related variables could be related to changes in BMI. CONCLUSIONS: The BMI at diagnosis is one of the most important predictors for the BMI at follow-up, and this suggests an important genetic or environmental cause. Adult CCSs are at high risk for developing severe underweight at follow-up. Future studies should focus on the causes and clinical consequences of underweight. Cancer 2015;121: VC 2015 American Cancer Society. KEYWORDS: body mass index, childhood cancer survivor, late effects, obesity, underweight. INTRODUCTION The increasing prevalence of overweight is a development of great concern. Not only in adults but also during childhood, an increase in the average body mass index (BMI) has been observed. 1,2 Obese individuals are at increased risk for developing diabetes mellitus type 2, dyslipidemia, hypertension, and cardiovascular disease. In children, it has been shown that a higher BMI during childhood is associated with an increased risk of coronary heart disease in adulthood. 3 More than 80% of all children treated for childhood cancer can be expected to become long-term survivors. For this reason, the long-term health consequences of the treatments given for pediatric malignancies are of increasing importance. Childhood cancer survivors (CCSs) have been reported to be at increased risk for developing obesity and metabolic syndrome (defined as the presence of 3 of the following risk factors: obesity, impaired fasting glucose, altered lipid profile, and hypertension). 4-6 Suggested risk factors for obesity developing in CCSs are female sex (leukemia survivors), brain tumors (especially after cranial irradiation), and a younger age during treatment. 7 Besides treatment-related factors, the development of obesity also has links to genetic characteristics. For instance, in irradiated female survivors of childhood leukemia (acute lymphoblastic leukemia), support has been found for the role of a leptin receptor gene polymorphism in the development of obesity. 8 For this reason, one of our aims was to relate the BMI at diagnosis to the BMI at follow-up in CCSs. In addition to obesity, underweight may be a problem for CCSs because it was reported by the CCS study that the prevalence of underweight was significantly increased. 9 Suggested risk factors for underweight were treatment with total body irradiation (females) or abdominal irradiation (males), use of alkylating agents, and anthracyclines. 9,10 Corresponding author: Hanneke M. van Santen, MD, PhD, Department of Pediatric Endocrinology, Wilhelmina Children s Hospital, University Medical Center Utrecht, KC , Postbus 85090, 3508 AB Utrecht, the Netherlands; Fax: (011)31-(0) ; h.m.vansanten@umcutrecht.nl 1 Department of Pediatric Endocrinology, Emma Children s Hospital, Academic Medical Center, University of Amsterdam, Amsterdam, the Netherlands; 2 Department of Pediatric Endocrinology, Wilhelmina Children s Hospital, University Medical Center Utrecht, Utrecht, the Netherlands; 3 Department of Clinical Epidemiology, Biostatistics, and Bioinformatics, Academic Medical Center, University of Amsterdam, Amsterdam, the Netherlands; 4 Department of Medical Informatics, Academic Medical Center, University of Amsterdam, Amsterdam, the Netherlands; 5 Department of Medical Oncology, Academic Medical Centre, University of Amsterdam, Amsterdam, the Netherlands; 6 Department of Pediatric Oncology, Emma Children s Hospital, Academic Medical Center, University of Amsterdam, Amsterdam, the Netherlands DOI: /cncr.29614, Received: April 25, 2015; Revised: June 26, 2015; Accepted: July 15, 2015, Published online August 19, 2015 in Wiley Online Library (wileyonlinelibrary.com) Cancer December 1,

2 Most of the studies on BMI after treatment for childhood cancer address only small numbers or a select group of patients or use questionnaires with self-reported heights and weights, or the BMI outcome has not been corrected for age or for the BMI at diagnosis In this study, we evaluated the prevalence of an abnormal BMI (corrected for age) in a large cohort of CCSs and the relation with the BMI at diagnosis and with treatment-related risk factors. MATERIALS AND METHODS Study Population All patients who had been treated for childhood cancer at the Emma Children s Hospital (EKZ)/Academic Medical Center (AMC) in Amsterdam between 1966 and 1996 and who were still alive at the time of data collection were identified by the EKZ/AMC Childhood Cancer Registry. In total, 1362 of 2596 patients survived their primary malignancy for 5 years or more. This cohort has been described previously. 17,18 In 1996, the EKZ/AMC started an outpatient clinic for the assessment of late effects of childhood cancer treatment. In 1055 of the 1362 survivors (77%), height and weight were measured at the first visit to the late effects clinic. Patients with scoliosis (n 5 103) or amputated extremities (n 5 58; 9 of these patients also had scoliosis) were excluded because the BMI could not be accurately calculated. Also, patients with recognized early (n 5 10) or late puberty (n 5 1; this patient also had scoliosis) were excluded because different age cutoff levels should be used for interpreting BMI. In total, 893 patients could be included for the analysis of BMI. Data Collection and Outcome Definition The primary cancer diagnosis and treatment data were retrieved from the EKZ/AMC Childhood Cancer Registry. For this study, we analyzed height, weight, and hormonal status data at the first visit at the outpatient clinic more than 5 years after diagnosis. A high BMI was defined as overweight or obesity. For children (age < 18 years), definitions of overweight and obesity were dependent on age as described in the literature. 1,19-23 Overweight for adult CCSs (age 18 years) was defined as a BMI between 25 and <30 kg/m 2. Obesity was defined as a BMI 30 kg/m A low BMI was defined as underweight or severe underweight. For children (age < 18 years), definitions of underweight were dependent on age as described in the literature ,23 Underweight for adults was defined as a BMI < 20 kg/m 2, and severe underweight was defined as a BMI < 18.5 kg/m 2. 21,25 For the BMI at diagnosis, a chart evaluation was performed to collect the height and weight at the time of the diagnosis. In total, data from 445 survivors could be retrieved. Hypothyroidism was defined as the use of thyroxine at the time of medical follow-up. Growth hormone (GH) deficiency was defined as the use of GH at the time of medical follow-up. Any hormone substitution was defined as the use of thyroxine, GH, corticosteroids, estrogens (not including oral contraceptives), testosterone/androgens, or vasopressin at the time of medical follow-up. Ethics All CCSs gave informed consent for data collection. The EKZ/AMC institutional review board reviewed and approved the collection of data. Statistical Analyses Prevalence of obesity and underweight All survivors were divided into 2 groups according to their age at follow-up (children were younger than 18 years, and adults were 18 years old or older). The prevalence of overweight, obesity, underweight, and severe underweight in the CCSs was compared with the age-specific prevalence in the Dutch population in the period of ,20-23,26 Logistic regression with an offset term was used to relate the observed outcome to the prevalence in a Dutch population of comparable age and sex (SPSS 15.0 and R 3.1.0). 27 The prevalence of adults who were underweight but not severely underweight after treatment for childhood cancer could not be compared with the prevalence in the reference population because of a lack of reference data. P values were based on the Wald test, whereas confidence intervals were based on the likelihood ratio test. Risk factors for abnormal BMI A multivariate analysis was performed for the risk factors for developing an abnormal BMI separately for children and adults. The multivariate logistic models included the BMI at diagnosis, age at treatment, age at follow-up, sex, head-neck irradiation, abdominal irradiation, treatment with chemotherapy, and hormonal deficiency (defined as the presence of hypothyroidism or GH deficiency). RESULTS Patient Characteristics (Table 1) Of the 893 survivors included in the study, 214 were children and 679 were adults at the time of the assessment of BMI during the first visit at the outpatient clinic more 4198 Cancer December 1, 2015

3 BMI in Childhood Cancer Survivors/van Santen et al TABLE 1. Patient Characteristics of Included CCSs (n 5 893) Characteristic CCSs < 18 y (n 5 214) CCSs 18 y (n 5 679) Sex, No. (%) Male 119 (55.6) 378 (55.7) Female 95 (44.4) 301 (44.3) Primary childhood cancer diagnosis, No. (%) Leukemia 84 (39.2) 163 (24) Lymphoma 19 (8.8) 176 (25.9) Kidney/Wilms tumor 44 (20.6) 66 (9.7) Brain/CNS tumor 8 (3.7) 43 (6.3) Soft tissue sarcoma 5 (2.3) 22 (3.2) Neuroblastoma 20 (9.3) 28 (4.1) Gonadal tumor 4 (1.9) 30 (4.4) Ewing sarcoma/pnet/osteosarcoma 2 (0.9) 22 (3.2) Other 28 (13.0) 129 (19.0) Mean age at diagnosis of childhood cancer, No. (%) 0-4 y 164 (76.6) 225 (33) 5-9 y 44 (20.6) 209 (30.8) y 6 (2.8) 196 (28.9) y 0 49 (7.2) Overall treatment category, No. (%) Chemotherapy 201 (93.9) 587 (86.5) Head/neck irradiation 28 (13.1) 223 (32.8) Abdominal irradiation 13 (6.1) 64 (9.4) Total body irradiation 4 (1.9) 13 (1.9) Follow-up time, mean (SD), y 10.3 (2.7) 16.4 (6.3) Abbreviations: CCS, childhood cancer survivor; CNS, central nervous system; PNET, primitive neuroectodermal tumor; SD, standard deviation. than 5 years after diagnosis. The most prominent malignancy was leukemia, and most survivors were 0 to 4 years old at the time of treatment. The mean follow-up time after the first treatment directly after diagnosis was 14.9 years (range, years); the mean age at follow-up was 21.9 years (range, years). The BMI at diagnosis could be found for 445 survivors. To analyze whether a selection bias had occurred with the exclusion of the group of survivors without BMI data at the time of diagnosis, the 2 groups were compared (Table 2). Compared with the 448 CCSs for whom the BMI at the time of diagnosis was missing, the 445 survivors with measurements of the BMI at diagnosis had been treated more frequently with chemotherapy (93% vs 81%) and less frequently with radiotherapy (35% vs 42%) and had been diagnosed more frequently with leukemia (30% vs 20%) and less frequently with a brain tumor (3% vs 9%; Table 2). These differences were not considered to be of clinical significance for introducing a selection bias. Prevalence of a High BMI and Obesity In total, 208 of all CCSs (23.3%) had a high BMI or were obese (Table 3). TABLE 2. Differences in Patient Characteristics Between Childhood Cancer Survivors With and Without a Known Measured BMI at Diagnosis Characteristic BMI at Diagnosis Known (n 5 445), No. (%) BMI at Diagnosis Not Known (n 5 448), No. (%) OR (95% CI) Sex: male 245 (55) 252 (56) 1.05 ( ) No chemotherapy ever 32 (7) 73 (19) 2.5 ( ) No RT ever 290 (65) 259 (58) 0.7 ( ) Brain tumor 12 (3) 39 (9) 0.29 ( ) Lymphoma 98 (22) 97 (22) 1.0 ( ) Leukemia 133 (30) 92 (20) 1.6 ( ) Solid tumor 165 (37) 184 (41) 0.8 ( ) Abbreviations: BMI, body mass index; CI, confidence interval; OR, odds ratio; RT, radiotherapy. Forty of the 214 children (18.7%) had a high BMI. Fifteen were obese (7.0%), and this rate was significantly higher than the rate for the general population in the Netherlands (3.1%). When we analyzed girls and boys separately, girls were found to have a significantly increased prevalence of obesity in comparison with an age- and sex-matched Dutch reference population (8.4% vs 3.4%). For boys, a trend toward more obesity in CCSs was found. Of the 679 adults, 168 (24.7%) had a high BMI, and 37 of these were obese (5.4%). In comparison with the normal population, a nonsignificantly lower prevalence of overweight and obesity was found in the CCSs (Table 3). For the subgroup of adult female CCSs aged 25 to 45 years, a trend toward more obesity was found in comparison with the normal population (near significance with a 95% confidence interval of ). Prevalence of a Low BMI and Severe Underweight The prevalence of a low BMI and severe underweight in children and adults is presented in Table 4. In total, 162 of all CCSs (18%) had a low BMI (underweight or severe underweight). Nineteen of the 214 children (8.9%) had a low BMI, and 3 of these children (1.4%) were severely underweight; neither group was significantly different from the general population. Among the adult survivors, a low BMI (BMI < 20 kg/m 2 ) was seen in 143 (21.1%). Sixty-three adult survivors (9.3%) were severely underweight (BMI < 18.5 kg/m 2 ), and this rate was significantly higher than the rate for the general population in the Netherlands (1.8%; P <.0001). This significantly increased prevalence of severe underweight was found in most age groups: males Cancer December 1,

4 TABLE 3. Prevalence of a High BMI and Obesity in CCSs Versus the General Population (n 5 893) No. CCS Study Cohort, % Dutch Population, % OR P 95% CI CCSs < 18 y 214 High BMI Boys (n 5 119) Girls (n 5 95) Obesity a Boys (n 5 119) Girls (n 5 95) a CCSs 18 y 679 BMI > 25 kg/m 2 Males y (n 5 236) y (n 5 141) >45 y (n 5 1) 0 Females y (n 5 189) y (n 5 112) BMI > 30 kg/m 2 37 Males y (n 5 236) y (n 5 141) >45 y (n 5 1) 0 Females y (n 5 189) y (n 5 112) Abbreviations: BMI, body mass index; CCS, childhood cancer survivor; CI, confidence interval; OR, odds ratio. For reference values of BMI, average values of the Dutch population were used. 1,19-22 A high BMI was defined as overweight or obesity. For children (age < 18 years), definitions of overweight and obesity were dependent on age as described in the literature. 22,23 For the definition of overweight for adults (age 8 years), standard international cutoff points of 25 to 30 kg/m 2 were used. Obesity was defined as a BMI 30 kg/m Significance was analyzed with 2-tailed binomial tests (SPSS 15.0). a Statistical significance was set at P <.05. aged 18 to 25 years (13.6% vs 2.4%), males aged 26 to 45 years (4.3% vs 0.8%), and females aged 18 to 25 years (12.2% vs 4.6%). Tumor Type and BMI The distribution of an abnormal BMI by tumor type is presented in Table 5. Risk Factors for BMI at Follow-Up (Tables 6 and 7). The analysis of risk factors for the BMI at follow-up was performed for the survivors for whom the BMI at diagnosis was known (n 5 445). For the child survivors, no significant risk factors were found. For the adult survivors, a higher BMI at diagnosis was associated with an increased risk of having a high BMI at follow-up (odds ratio 1.75; 95% confidence interval ). In addition, a younger age at diagnosis, an older age at follow-up, and treatment with cranial irradiation were associated with a significantly increased risk for developing a high BMI. A lower BMI at diagnosis and a younger age at follow-up were associated with a low BMI at follow-up. No specific treatment-related risk factors could be identified for developing a low or high BMI after treatment for childhood cancer. DISCUSSION In this large Dutch cohort of CCSs, we found a significantly increased prevalence of obesity only in girls < 18 years old in comparison with the general Dutch population. Children who had a high BMI, were younger at the time of diagnosis, and were treated with cranial radiotherapy had the greatest risk of having a high BMI after treatment for childhood cancer. Perhaps even more importantly, we found an increased prevalence of severe underweight in almost all adult CCSs in comparison with the general Dutch population. A low BMI at diagnosis and a younger age at follow-up were the only significant risk factors for underweight at follow-up. In contrast to earlier studies, we could not identify treatment-related risk factors for developing a low BMI. The unique features of our study are the long-term follow-up, the objective measurements of height and weight instead of self-reported questionnaires, and the inclusion of the BMI at diagnosis in the analysis. Because 4200 Cancer December 1, 2015

5 BMI in Childhood Cancer Survivors/van Santen et al TABLE 4. Prevalence of a Low BMI and Severe Underweight in CCSs Versus the General Population (n 5 893) No. CCS Study Cohort, % Dutch Population, % OR P 95% CI CCSs < 18 y 214 Low BMI Boys (n 5 119) Girls (n 5 95) Severe underweight Boys (n 5 119) Girls (n 5 95) CCSs 18 y 679 BMI < 18.5 kg/m 2 63 Males y (n 5 239) <.001 a y (n 5 145) <.001 a Females y (n 5 189) <.001 a y (n 5 112) Abbreviations: BMI, body mass index; CCS, childhood cancer survivor; CI, confidence interval; OR, odds ratio. A low BMI was defined as underweight or severe underweight. For children (age < 18 years), definitions of underweight were dependent on age as described in the literature. 22,23 Underweight for adults was defined as a BMI < 20 kg/m 2, and severe underweight was defined as a BMI < 18.5 kg/m For reference values of BMI, average values of the Dutch population were used. 1,19-21,23 Significance was analyzed with 2-tailed binomial tests (SPSS 15.0). a Significance was set at P <.05. TABLE 5. Distribution of Abnormal BMIs by Tumor Type Characteristic No. High BMI, Obesity, Low BMI, Severe Underweight, % (No.) a % (No.) a % (No.) a % (No.) a General population of the Netherlands < 18 y CCSs < 18 y Brain tumor 8 25 (2) Acute lymphoblastic leukemia (16) 8 (6) 10.7 (8) 4 (3) Myeloid leukemia (3) 11.1 (1) 0 0 Hodgkin disease Non-Hodgkin lymphoma (4) (2) 0 Rhabdomyosarcoma (1) (3) 0 Nonrhabdomyosarcoma 5 40 (2) 40 (2) 5.3 (1) 0 Ewing sarcoma/pnet (1) 0 Neuroblastoma 20 5 (1) 5 (1) 15 (3) 0 Ovarian/testis carcinoma Nephroblastoma (10) 11.1 (5) 2.3 (1) 0 Hepatoblastoma Other (rare) childhood cancer n 5 4 Langerhans cell histiocytosis Total child survivors (40) 7.0 (15) 8.9 (19) 1.4 (3) General population of the Netherlands 18 y 40% 9.5% Unknown 1.8% CCSs 18 y Brain tumor (19) 11.6 (5) 9.3 (4) 7 (3) Acute lymphoblastic leukemia (34) 4.0 (6) 22.6 (34) 8 (12) Myeloid leukemia (2) 6.7 (1) 40 (6) 20 (3) Hodgkin disease (21) 3.6 (3) 20.2 (17) 3.6 (3) Non-Hodgkin lymphoma (23) 4.3 (4) 19.6 (18) 14.1 (13) Rhabdomyosarcoma (12) 6.5 (4) 19.4 (12) 8.1 (5) Nonrhabdomyosarcoma (7) 9.1 (2) 22.7 (5) 4.5 (1) Ewing sarcoma/pnet (9) 13.6 (3) 4.5 (1) 0 Neuroblastoma (7) 14.3 (4) 28.6 (8) 14.3 (4) Ovarian/testicular carcinoma (3) (8) 16.7 (5) Nephroblastoma (12) 3 (2) 24.2 (16) 9.1 (6) Hepatoblastoma (2) 20 (1) Other type of (rare) childhood cancer n 5 55 Langerhans cell histiocytosis (3) 75 (3) Total adult survivors (168) 5.4 (37) 21.1 (143) 9.3 (63) Abbreviations: BMI, body mass index; CCS, childhood cancer survivor; PNET, primitive neuroectodermal tumor. a Percentage of cancer survivors with an abnormal BMI (high or low) with respect to the tumor type. Cancer December 1,

6 TABLE 6. Multivariate Logistic Regression Model for Risk Factors for Developing a BMI > 25 kg/m 2 in Childhood Cancer Survivors Aged 18 Years or Older (n 5 334) Covariate OR 95% CI P BMI at diagnosis a Age at diagnosis a Age at follow-up a Sex: female vs male Radiotherapy: yes vs no Head/neck b a Abdomen TBI Chemotherapy overall: yes vs no Endocrine state: hypothyroidism or GH deficiency c Abbreviations: BMI, body mass index; CI, confidence interval; GH, growth hormone; OR, odds ratio; TBI, total body irradiation. a Statistical significance was set at P <.05. b Head/neck and abdominal irradiation might include total body irradiation. c Hypothyroidism was defined as using thyroxine at the time of medical follow-up, and a GH deficiency was defined as using GH at the time of medical follow-up. TABLE 7. Multivariate Logistic Regression Model for Risk Factors for Developing a BMI < 20 kg/m 2 in Childhood Cancer Survivors Aged 18 Years or Older (n 5 334) Covariate OR 95% CI P BMI at diagnosis a Age at diagnosis Age at follow-up a Sex: female vs male Radiotherapy: yes/no Head/neck b Abdomen TBI Chemotherapy: yes vs no Endocrine state: hypothyroidism or GH deficiency c Abbreviations: BMI, body mass index; CI, confidence interval; GH, growth hormone; OR, odds ratio; TBI, total body irradiation. a Statistical significance was set at P <.05. b Head/neck and abdominal irradiation might include total body irradiation. c Hypothyroidism was defined as using thyroxine at the time of medical follow-up, and a GH deficiency was defined as using GH at the time of medical follow-up. the reference values for BMI in children are greatly dependent on age and sex, age- and sex-matched reference values were used. Unique to this study is the fact that we also corrected for age and sex in the adult population; this is relevant because BMI changes with time and also after the age of 18 years. 21 A possible drawback of the study is the large number of survivors whose BMI was unknown at the time of diagnosis, although we believe that also for these patients height and weight must have been measured because they are often used for dose calculations for chemotherapy. Table 2 shows that there are some differences between the group with BMI values and the group without BMI values. However, the CCS group for which the BMI was measured was a very heterogeneous and unselected cohort, and this makes the analysis representative and valuable. No relevant relations could be found between the tumor type and abnormal BMIs, partly because the patient numbers became small in these subanalyses. It may be commented that children surviving acute myeloid leukemia, acute lymphoblastic leukemia, and nephroblastoma seem to develop a high BMI more frequently, and adults surviving childhood brain tumors, neuroblastoma, and Ewing sarcoma seem to have this tendency. A great issue of concern is the strikingly high percentage of severely underweight survivors that we found in this study. Among the male adult survivors aged 18 to 25 years, 13.6% were severely underweight. Also, among females aged 18 to 25 years and males aged 26 to 45 years, we found a significantly higher prevalence of severe underweight in comparison with the general Dutch population. Underweight in CCSs has been reported previously, 9-11 and total body irradiation, abdominal radiation, treatment with anthracyclines, and alkylating agents have been identified as possible risk factors. 4,10,11 None of these studies included the BMI at diagnosis in their analyses. In our study, we could not verify these treatment-related risk factors but identified only a low BMI at diagnosis and a younger age at follow-up. The clinical consequences of underweight for CCSs are not clear. Underweight may contribute to fatigue, which has been described after treatment for (childhood) cancer. Also, long-lasting underweight may result in a reduction of the final height because of a lack of catch-up growth after cancer treatment. Obesity is the result of a disturbed energy homeostasis. For survivors of childhood cancer, special attention must be paid to the development of overweight or obesity because some survivors are already at increased risk for cardiovascular disease (eg, after treatment with anthracyclines, mitoxantrone, or cardiac radiotherapy 28 ) or hormonal deficiencies because of their earlier treatment (eg, after cranial radiotherapy or alkylating agents 29 ). For this reason, identifying the groups of survivors of childhood cancer that are at increased risk to become obese is necessary because these risks may be co-associated. In this large prospective cohort of CCSs, next to treatment with cranial radiotherapy, the BMI at diagnosis was independently related to the BMI at follow-up. The explanation for this finding probably lies in an 4202 Cancer December 1, 2015

7 BMI in Childhood Cancer Survivors/van Santen et al individual s genetic susceptibility for gaining weight. 8,30 Next to genetic factors, social and environmental factors may also be involved. For children with cancer as well as healthy children, the environment and parents play essential roles in the development of children s food preferences and energy intake, and this must be a point of attention. In addition, it has been shown that CCSs displayed excessive energy intake and poor dietary habits. 31 Wellintended parental or familial involvement and nourishing of the sick child may lead to unhealthy feeding. This may lead to difficulties for parents in setting the right boundaries for eating habits for their ill child and may suggest that the way to address this issue in CCSs may be different than the way for other healthy children. There are, of course, also many comparable factors found in children with regular obesity that shape children s daily dietary habits and physical activity, such as their religion, community, and environment, which all influence the development of obesity. These educational, physical, and social aspects should all be points of attention for professional support and the follow-up of these children. Cranial irradiation has been found to be strongly related to GH deficiency and, with that, to lower insulinlike growth factor 1 concentrations, higher fasting insulin concentrations, abdominal obesity, and dyslipidemia. 32,33 The finding that overweight did not seem to be related to any peripheral hormonal deficiency may be explained by the fact that all survivors had already received adequate hormone supplementation therapy. The most effective intervention for increasing BMI is primary prevention. This suggests that for a young child with cancer with a high BMI at diagnosis, special attention must be paid to weight, diet, and exercise during and after cancer treatment. For a young child with a low BMI at diagnosis, special attention during treatment and follow-up should be addressed to maintain an optimal BMI. In conclusion, in our study, we did not find adults to have an increased risk for developing a high BMI after treatment for childhood cancer. However, adults were found to have an increased risk for developing underweight in comparison with the normal Dutch population. Furthermore, the BMI at diagnosis was associated with the BMI at follow-up, and this suggested a genetic and environmental component. Future studies regarding BMI in CCSs should be focused on preventive interventions to maintain a normal BMI during and after treatment for childhood cancer. Also, the fact that young women are more frequently underweight after treatment for childhood cancer must be further explored. Etiological and genetic studies of overweight should be performed in the general population because the role of cancer treatment seems to be minimal. New research should be developed that focuses on the causes and clinical consequences of underweight in survivors of childhood cancer. FUNDING SUPPORT No specific funding was disclosed. CONFLICT OF INTEREST DISCLOSURES The authors made no disclosure. REFERENCES 1. van den Hurk K, van Dommelen P, van Buuren S, Verkerk PH, Hirasing RA. Prevalence of overweight and obesity in the Netherlands in 2003, compared to 1980 and Arch Dis Child. 2007; 92: Wang Y, Lobstein T. Worldwide trends in childhood overweight and obesity. Int J Pediatr Obes. 2006;1: Baker JL, Olsen LW, Sorensen TI. Childhood body-mass index and the risk of coronary heart disease in adulthood. N Engl J Med. 2007; 357: Talvensaari KK, Lanning M, Tapanainen P, Knip M. Long-term survivors of childhood cancer have an increased risk of manifesting the metabolic syndrome. J Clin Endocrinol Metab. 1996;81: Srinivasan S, Ogle GD, Garnett SP, Briody JN, Lee JW, Cowell CT. Features of the metabolic syndrome after childhood craniopharyngioma. J Clin Endocrinol Metab. 2004;89: Meacham LR, Chow EJ, Ness KK, et al. Cardiovascular risk factors in adult survivors of pediatric cancer a report from the Childhood Cancer Survivor Study. Cancer Epidemiol Biomarkers Prev. 2010;19: Garmey 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: Ross JA, Oeffinger KC, Davies SM, et al. Genetic variation in the leptin receptor gene and obesity in survivors of childhood acute lymphoblastic leukemia: a report from the Childhood Cancer Survivor Study. J Clin Oncol. 2004;22: Meacham LR, Gurney JG, Mertens AC, et al. Body mass index in long-term adult survivors of childhood cancer: a report of the Childhood Cancer Survivor Study. Cancer. 2005;15: Brouwer CA, Gietema JA, Vonk JM, et al. Body mass index and annual increase of body mass index in long-term childhood cancer survivors; relationship to treatment. Support Care Cancer. 2012;20: Brouwer CA, Gietema JA, Kamps WA, de Vries EG, Postma A. Changes in body composition after childhood cancer treatment: impact on future health status a review. Crit Rev Oncol Hematol. 2007;63: Hoffman KE, Derdak J, Bernstein D, et al. Metabolic syndrome traits in long-term survivors of pediatric sarcoma. Pediatr Blood Cancer. 2008;50: Kourti M, Tragiannidis A, Makedou A, Papageorgiou T, Rousso I, Athanassiadou F. Metabolic syndrome in children and adolescents with acute lymphoblastic leukemia after the completion of chemotherapy. J Pediatr Hematol Oncol. 2005;27: Oeffinger KC. Are survivors of acute lymphoblastic leukemia (ALL) at increased risk of cardiovascular disease? Pediatr Blood Cancer. 2008;50: Taskinen M, Lipsanen-Nyman M, Tiitinen A, Hovi L, Saarinen- Pihkala UM. Insufficient growth hormone secretion is associated with metabolic syndrome after allogeneic stem cell transplantation in childhood. J Pediatr Hematol Oncol. 2007;29: Cancer December 1,

8 16. Trimis G, Moschovi M, Papassotiriou I, Chrousos G, Tzortzatou- Stathopoulou F. Early indicators of dysmetabolic syndrome in young survivors of acute lymphoblastic leukemia in childhood as a target for preventing disease. J Pediatr Hematol Oncol. 2007;29: Geenen MM, Cardous-Ubbink MC, Kremer LC, et al. Medical assessment of adverse health outcomes in long-term survivors of childhood cancer. JAMA. 2007;27: Sieswerda E, Mulder RL, van Dijk IW, et al. The EKZ/AMC childhood cancer survivor cohort: methodology, clinical characteristics, and data availability. J Cancer Surviv. 2013;7: van Buuren S. Body-mass index cut-off values for underweight in Dutch children [in Dutch]. Ned Tijdschr Geneeskd. 2004;148: Health Council of the Netherlands. Overweight and Obesity. The Hague, the Netherlands: Health Council of the Netherlands; Publication 2003/ Central Bureau of Statistics. Permanent Onderzoek Leefsituatie (POLS). Available at: Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition for child overweight and obesity worldwide: international survey. BMJ. 2000;320: Cole TJ, Flegal KM, Nicholls D, Jackson AA. Body mass index cut offs to define thinness in children and adolescents: international survey. BMJ. 2007;335: Central Bureau of Statistics. Aspecten van gezondheid en ziekte in de bevolking. In: Vademecum Gezondheidsstatistiek Nederland. Voorburg/Heerlen, the Netherlands: Ministry of Health, Welfare, and Sport; 1999: Ali SM, Lindstrom M. Socioeconomic, psychosocial, behavioural, and psychological determinants of BMI among young women: differing patterns for underweight and overweight/obesity. Eur J Public Health. 2006;16: Hirasing RA, Fredriks AM, van Buuren S, Verloove-Vanhorick SP, Wit JM. Increased prevalence of overweight and obesity in Dutch children, and the detection of overweight and obesity using international criteria and new reference diagrams. Ned Tijdschr Geneeskd. 2001;145: R Development Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: Foundation for Statistical Computing; Tonorezos ES, Hudson MM, Edgar AB, et al. Screening and management of adverse endocrine outcomes in adult survivors of childhood and adolescent cancer. Lancet Diabetes Endocrinol. 2015;3: Armenian SH, Hudson MM, Mulder RL, et al. Recommendations for cardiomyopathy surveillance for survivors of childhood cancer: a report from the International Late Effects of Childhood Cancer Guideline Harmonization Group. Lancet Oncol. 2015;16: Ross JA. Genetic susceptibility and body mass in childhood cancer survivors. Pediatr Blood Cancer. 2007;48: Cohen J, Wakefield CE, Fleming CA, Gawthorne R, Tapsell LC, Cohn RJ. Dietary intake after treatment in child cancer survivors. Pediatr Blood Cancer. 2012;58: Brennan BM, Rahim A, Mackie EM, Eden OB, Shalet SM. Growth hormone status in adults treated for acute lymphoblastic leukaemia in childhood. Clin Endocrinol. 1998;48: Brennan BM, Rahim A, Blum WF, Adams JA, Eden OB, Shalet SM. Hyperleptinaemia in young adults following cranial irradiation in childhood: growth hormone deficiency or leptin insensitivity? Clin Endocrinol. 1999;50: Cancer December 1, 2015