Risk Factors for the Development of Obesity in Children Surviving ...

0013-7227/03/$15.00/0 Printed in U.S.A.

The Journal of Clinical Endocrinology & Metabolism 88(2):611– 616 Copyright © 2003 by The Endocrine Society doi: 10.1210/jc.2002-021180

Risk Factors for the Development of Obesity in Children Surviving Brain Tumors ROBERT H. LUSTIG, SUSAN R. POST, KLEEBSABAI SRIVANNABOON, SUSAN R. ROSE, ROBERT K. DANISH, GEORGE A. BURGHEN, XIAOPING XIONG, SHENGJIE WU, AND THOMAS E. MERCHANT Departments of Endocrinology (R.H.L., K.S., S.R.R., R.K.D., G.A.B.), Biostatistics (X.X., S.W.), and Radiation Oncology (T.E.M.), St. Jude Children’s Research Hospital, Memphis, Tennessee 38105; and Chicago Medical School (S.R.P.), North Chicago, Illinois 60064-3095 Risk factors were: age at diagnosis (P ⴝ 0.04), radiation dosimetry to the hypothalamus (51–72 Gy, P ⴝ 0.002 even after hypothalamic and thalamic tumor exclusion), and presence of any endocrinopathy (P ⴝ 0.03). In addition, risk factors when compared with BMI slope for the general American pediatric population included: tumor location (hypothalamic, P ⴝ 0.001), tumor histology (craniopharyngioma, P ⴝ 0.009; pilocytic astrocytoma, P ⴝ 0.043; medulloblastoma, P ⴝ 0.039); and extent of surgery (biopsy, P ⴝ 0.03; subtotal resection, P ⴝ 0.018). These results verify hypothalamic damage, either due to tumor, surgery, or radiation, as the primary cause of obesity in survivors of childhood brain tumors. In particular, hypothalamic radiation doses of more than 51 Gy are permissive. These results reiterate the importance of the hypothalamus in energy balance, provide risk assessment criteria for preventative measures before the development of obesity in at-risk patients, and suggest therapeutic strategies to reduce the future development of obesity. (J Clin Endocrinol Metab 88: 611– 616, 2003)

Hypothalamic obesity, a syndrome of intractable weight gain due to hypothalamic damage, is an uncommon but devastating complication for children surviving brain tumors. We undertook a retrospective evaluation of the body mass index (BMI) curves for the St. Jude Children’s Research Hospital brain tumor population diagnosed between 1965 and 1995 after completion of therapy to determine risk factors for the development of obesity. Inclusion criteria were: diagnosis less than 14 yr of age, no spinal cord involvement, ambulatory, no supraphysiologic hydrocortisone therapy (>12 mg/m2䡠d), treatment and follow-up at St. Jude Children’s Research Hospital, and disease-free survival greater than 5 yr (n ⴝ 148). Risk factors examined were age at diagnosis, tumor location, histology, extent of surgery, hydrocephalus requiring ventriculoperitoneal shunting, initial high-dose glucocorticoids, cranial radiation therapy, radiation dosimetry to the hypothalamus, intrathecal chemotherapy, and presence of endocrinopathy. Analyses were performed both between groups within a risk factor and against BMI changes for age in normal children older than 5.5 yr (the age of adiposity rebound).

A

endocrinopathies and appears as a result of damage to the hypothalamus (22, 23). We undertook a systematic evaluation of the rate of increase in body mass index (BMI) of survivors of brain tumors in an attempt to identify risk factors for the development of obesity in this population. We then compared these rates of BMI increase with that of the general American pediatric population, based on the BMI curves developed in 1999 by the Centers for Disease Control (CDC; Atlanta, GA) (24).

S THERAPEUTIC MODALITIES become more effective in curing childhood cancer, and as a greater proportion of these patients are reaching their adult years, a number of late effects have recently been described that have implications for long-term medical well being. These include impaired learning and memory, osteoporosis, and various endocrinopathies (1–9). Another such late effect is obesity. Various reports in survivors of acute lymphoblastic leukemia (ALL) suggest an obesity prevalence of 20 –50% (10 –17). Numerous risk factors, including glucocorticoid treatment, chemotherapy, cranial irradiation, and psychosocial stressors have been implicated in the pathogenesis of obesity in this vulnerable population. Children with brain tumors, particularly craniopharyngioma, are also at extremely high risk for the development of obesity after tumor therapy (18 –21). In these patients, the weight gain is often intractable, and not responsive to diet and exercise interventions. This form, termed “hypothalamic obesity,” is frequently associated with other hypothalamic

Patients and Methods This study was a retrospective analysis of data obtained during the course of patient care and was approved by the St. Jude Children’s Research Hospital (SJCRH) Institutional Review Board before data collection. The records of all children diagnosed and treated for primary brain tumors at SJCRH from 1965 until 1995 (n ⫽ 675) were evaluated to determine heights and weights at each follow-up visit. From these data, the rate of BMI increase was computed, using the formula: BMI ⫽ weight (kg) ⫼ height (m2). Patients were eligible for this analysis if: 1) they were 14 yr of age or younger at the time of tumor diagnosis, 2) treatment and follow-up took place at SJCRH, 3) there was no spinal cord involvement or spinal radiation therapy, 4) they were ambulatory (as lack of ambulation may predispose to positive energy balance), 5) they were not receiving supraphysiologic doses of glucocorticoid (e.g. hydrocortisone, prednisone, or dexamethasone to exceed 12 mg/m2䡠d hydrocortisone equivalent) for greater than 6 months after tumor therapy, 6) they were followed for at least 5 yr after tumor therapy was

Abbreviations: ALL, Acute lymphoblastic leukemia; BMI, body mass index; CDC, Centers for Disease Control; CrXRT, cranial radiation therapy; Gy, Gray; SJCRH, St. Jude Children’s Research Hospital; VMH, ventromedial hypothalamus; V-P, ventriculoperitoneal.

611

612

J Clin Endocrinol Metab, February 2003, 88(2):611– 616

complete, and 7) there were no documented recurrences or second tumors at the time of chart review.

Risk factors assessed The following were recorded from each patient’s chart and were used for BMI stratification: 1) age at diagnosis; 2) tumor histology (astrocytoma, pilocytic astrocytoma, brain stem glioma, craniopharyngioma, ependymoma, medulloblastoma, primitive neuroectodermal tumor, germ cell tumor, pituitary adenoma); 3) tumor location (hypothalamic, thalamic, posterior fossa, lateral ventricle, temporal lobe, other hemispheric); 4) extent of surgery (none, biopsy, subtotal resection, gross total resection; 5) hydrocephalus at diagnosis, as determined by the need for surgical placement of a ventriculoperitoneal (V-P) shunt (Yes/No); 6) steroid use after diagnosis [none, short-term (less than 2 months), longterm (less than 6 months)]; 7) exposure to chemotherapy (Yes/No); and 8) cranial radiation therapy (CrXRT) (Yes/No). In addition, 9) the dose of radiation to the hypothalamus was estimated by a review of simulation and portal films and treatment dosimetry for all patients treated with CrXRT. Lastly, 10) diagnoses of hypothalamic endocrinopathies, as determined by the need for hormone replacement therapy (GH, levothyroxine, hydrocortisone, estrogen-testosterone-leuprolide, DDAVP), were recorded. Most, but not all of these assumptions were confirmed by formal endocrine dynamic testing before treatment.

Statistical analysis The goal of this analysis was to identify and assess those factors that correlated with longitudinal changes in BMI from diagnosis to last follow-up. We used the mixed model (measuring fixed and random effects) for the analysis. In this model, each patient was a cluster with repeated measurements (25, 26). According to descriptive BMI curve plots, we assumed that BMI change was a linear function of time. The mixed model was adjusted by age at diagnosis when it was fitted by each clinical parameter. As a secondary analysis, BMI change in brain tumor survivors was compared with BMI change in the general pediatric population, using calculations based on the BMI curves released in 1999 by the CDC (24). The normal BMI monotone decreases before, and increases after age 5.5 yr. Therefore, we compared BMI change rates in brain tumor survivors and in the general pediatric population over age 5.5 yr. Comparisons were made based on BMI changes adjusted for, and according to the age of each patient, and corrected for the age-dependent BMI velocity of the general population. Results are expressed as the change in BMI (⌬BMI) per month following treatment. We chose to assess this parameter, rather than the change in BMI sd score (z-score) (16), based on two statistical precepts: 1) The BMI curves of normal children are not Gaussian, particularly beyond the ⫹2.5 sd level, where they exhibit an inherent skewness (24), which would artifactually decrease their degree of overweight. Many of our subjects were above this sd level. 2) This study is an assessment of risk for the development of obesity within the pediatric population surviving brain tumors, not a comparison of incidence or prevalence of obesity against the general pediatric population. During the 30-yr period of eligibility for inclusion in this study, and in particular, the 5-yr follow-up period 1995–2000, the mean and median BMI of the general American pediatric population have risen markedly (27, 28). Standard BMI curves for American children before 1999 are not available, and therefore the computation of the BMI z-score for our population against historical peers is fraught with difficulty. Rather, comparison of ⌬BMI in brain tumor survivors against the 1999 BMI curves allow for the assessment of biological differences in our subjects across time, whereas comparing ⌬BMI z-scores against the 1999 norms would likely underestimate the risk for the development of obesity in this population.

Results

Of the 675 patients treated at SJCRH for primary brain tumors between 1965 and 1995, a total of 156 satisfied the inclusion criteria. Table 1 summarizes the clinical and demographic data of this study population. Eight of these were

Lustig et al. • Obesity Risk Factors after Childhood Brain Tumors

missing clinical or demographic data and were excluded from further analysis. Our statistical analysis includes the remaining 148 children. BMI stratification by age at diagnosis

Time (in months) at which measurements were taken was treated as a continuous variable. The rates of BMI increase were computed based on age of diagnosis, ranging from 6 –12 yr. As BMI velocity increases with increasing age, the mean age-dependent BMI velocities calculated from the 1999 CDC BMI charts (24) were subtracted from each of the patientderived BMI curves. Patients diagnosed at a younger age were noted to increase BMI faster than those diagnosed later (Fig. 1). The difference between BMI changes in those diagnosed before or at age 6 yr was significantly greater than at other ages (P ⫽ 0.018). BMI stratification by tumor parameters

The effects of tumor physical characteristics on BMI adjusted by age were estimated. The presenting BMIs at time of diagnosis, and the rates of BMI change over time, although disparate between tumor locations, were not statistically significant. The highest BMI change rate was nearly double for hypothalamic tumors (0.090 ⫾ 0.014) vs. that for lateral ventricular tumors (0.049 ⫾ 0.032). A similar, but significant effect was found when BMI changes were stratified by tumor histology. Presenting BMIs of patients at the time of diagnosis were significantly different between tumor histologies (Fig. 2A; P ⫽ 0.0001). The values ranged between (11.036 ⫹ 0.562 ⫻ median age) in patients with primitive neuroectodermal tumor to (11.036 ⫹ 0.5762 ⫻ median age ⫹ 5.079) in patients with craniopharyngioma. However, the rates of BMI change over time were not significantly different between tumor histologies. BMI stratification by treatment parameters

The effects of various treatment parameters on BMI change (adjusted by age) were estimated over time. The BMI change rate was not significantly different between those who received CrXRT vs. those that did not. However, the radiation dose to hypothalamus did significantly affect the rate of BMI change. Total hypothalamic dosimetry was stratified into low (less than 51 Gy), intermediate (51–55 Gy), and high (greater than 55 Gy) dose groups, respectively. BMI change rate was significantly lower in the low-dose group (0.024 ⫾ 0.016) vs. the other two groups (0.089 ⫾ 0.011, 0.091 ⫾ 0.017) (Fig. 2B; P ⫽ 0.0018), although there was no difference between the intermediate- and high-dose groups. The effect of hypothalamic dosimetry greater than 51 Gy was significant even after subjects with hypothalamic or thalamic tumor location were excluded from the analysis (P ⫽ 0.003). Neither the extent of surgery, the need for V-P shunting, the long-term use of steroid treatment to manage cerebral edema or hydrocephalus, nor the use of chemotherapy, predicted the rate of BMI increase in later years. BMI stratification by residual endocrinopathy

No significant differences were noted in the BMI change rates of those patients who had individual endocrinopathies


J Clin Endocrinol Metab, February 2003, 88(2):611– 616 613

TABLE 1. Clinical characteristics of patients n (%)

Age at diagnosis (years) Mean ⫾ SEM Median

Treatment parameters Extent of surgery Biopsy Gross total resection Subtotal resection None Missing

6.8 ⫾ 3.6 6.96

23 (14.7) 61 (39.1) 42 (26.9) 12 (7.8) 18 (11.5)

Long-term steroid use No Yes Missing

143 (91.6) 11 (7.1) 2 (1.3)

Intrathecal chemotherapy No Yes Missing

152 (97.4) 2 (1.3) 2 (1.3)

n (%)

Supplementary hormone replacement Hydrocortisone No Yes Missing

127 (81.4) 27 (17.3) 2 (1.3)

Sex hormone (estrogen/testosterone/leuprolide) No Yes Missing

135 (86.5) 19 (12.2) 2 (1.3)

DDAVP No Yes Missing

145 (92.9) 9 (5.8) 2 (1.3)

GH No Yes Missing

96 (61.5) 58 (37.2) 2 (1.3)

T No Yes Missing

84 (53.8) 70 (44.9) 2 (1.3)

4

Peripheral chemotherapy No Yes Missing CrXRT No Yes

121 (77.6) 33 (21.1) 2 (1.3) 67 (42.9) 89 (57.1)

Radiation dose to hypothalamus Low (⬍51 Gy) Middle (51 Gy ⱕ 55 Gy) High (⬎55 Gy) V-P shunting No Yes

23 (25.8) 46 (51.7) 20 (22.5) 107 (68.6) 49 (31.4)

Any hormone replacement No Yes Missing

72 (46.1) 82 (52.6) 2 (1.3)

Tumor parameters Tumor location Hemispheric Hypothalamus Temporal lobe Posterior fossa Thalamus Lateral ventricles Tumor type Astrocytoma, NOS Pilocytic astrocytoma Brainstem glioma Craniopharyngioma Ependymoma Medulloblastoma PNET Germinoma Pituitary macroadenoma

(as determined by hormonal therapy), although the trend of BMI change was higher for patients receiving sex hormone/ leuprolide therapy than in those who did not (P ⫽ 0.09). However, when patients were grouped by the presence of any endocrinopathy, BMI change rate over time was significantly increased (Fig. 2C; P ⫽ 0.031).

16 (10.3) 35 (22.5) 15 (9.6) 78 (50.0) 6 (3.8) 6 (3.8) 39 (25.0) 26 (16.7) 14 (9.0) 10 (6.4) 15 (9.6) 40 (25.6) 5 (3.2) 6 (3.8) 1 (0.7)

toma, P ⫽ 0.039); tumor location (hypothalamic, P ⫽ 0.001), extent of surgery (biopsy, P ⫽ 0.038; gross total resection, P ⫽ 0.018), and any hormone replacement therapy (GH, P ⫽ 0.008; T4, P ⫽ 0.001; hydrocortisone, P ⫽ 0.007; sex hormone/ leuprolide, P ⫽ 0.020; DDAVP, P ⫽ 0.016). No effect of V-P shunting, steroid use, or chemotherapy was noted on BMI increase.

BMI change rates of brain tumor patients as compared with normal children

Discussion

When compared with BMI changes for normal children based on the 1999 BMI curves from the CDC (24), BMI change rates for brain tumor survivors were significantly different when stratified for tumor histology (craniopharyngioma, P ⫽ 0.009; pilocytic astrocytoma, P ⫽ 0.043; medulloblas-

Obesity is a devastating late effect in cancer survivors. Obesity is a primary risk factor for morbidity in the general population, including diabetes, dyslipidemia, hypertension, musculoskeletal problems, sleep apnea, impaired well being, depression, and social exile. These are particularly concern-

614


ing in the vulnerable population of brain tumor survivors. It is important to understand the pathogenesis of obesity in this group to treat and prevent further disability and early demise. This retrospective analysis identified several risk factors for the prediction of the development of obesity in survivors of childhood brain tumors. Hypothalamic location, tumor histology consistent with hypothalamic involvement (such as craniopharyngioma), extent of surgery such as biopsy or gross total resection (common in the hypothalamic area, especially with craniopharyngiomas), hypothalamic irradiation exceeding 51 Gy, and the presence of hypothalamic endocrinopathies, were all associated with abnormal posttherapy BMI increase. Thus, it would appear that hypothalamic damage, either due to tumor, surgery, or CrXRT, is a regiospecific and primary risk factor for the development of obesity in this population.

FIG. 1. Rates of BMI increase based on age at tumor diagnosis. When the rate of BMI increase of the general population [based on BMI curves furnished by the CDC, (24)] at each age was subtracted from the BMI velocities of brain tumor subjects, there was a negative age dependence of BMI increase. Those children diagnosed at 6 yr of age gained BMI more rapidly than those diagnosed at 8, 10, or 12 yr of age (P ⫽ 0.0018).


Previous studies have documented increased weight gain in survivors of ALL (10 –17). Studies in ALL survivors suggest that glucocorticoid therapy is an important risk factor (11); however, each ALL patient is treated for long periods of time with high-dose glucocorticoid (29), so risk factor analysis in this population is exceedingly difficult. We specifically excluded patients on long-term high-dose glucocorticoid, and we could not implicate a role for courses of glucocorticoid of shorter duration than 6 months on BMI increase. We found that any form of hypothalamic damage, not just CrXRT, was a risk factor for abnormal BMI increase. The only risk factor unrelated to hypothalamic involvement was a younger age at diagnosis; a finding also seen in some studies of ALL survivors (10, 11). There may be biological reasons for this finding as well, including continued brain growth and myelinization until 4 yr of age (30, 31). CrXRT in younger children leads to significant dysfunction of cognition and attention (7, 32, 33), suggesting increased vulnerability of the developing brain to ionizing radiation, with atrophy, leukomalacia, and lacuna development (34). In this analysis, we assumed that hormonal replacement inferred endocrinopathy; i.e. T4 replacement signified hypothyroidism, etc. This was necessary because the criteria for the diagnosis of various endocrinopathies have been refined over the 30 yr of patient treatment and follow-up over this study and because endocrine evaluation in this cohort was not done prospectively. Such an analysis might include some false positives, such as subjects who were treated based on baseline lab values without formal dynamic testing, and false negatives, such as patients who were not evaluated for GH deficiency because they had already reached an adequate adult height. The most severe weight gain was associated with those patients manifesting diabetes insipidus; probably as a result of surgery for craniopharyngioma (35), the tumor with the fastest BMI velocity. Damage to the hypothalamus has long been associated with obesity. Rats who are subject to bilateral lesions or deafferentation of the ventromedial hypothalamus (VMH) develop a syndrome of hyperphagia, hyperinsulinemia, and weight gain, termed “hypothalamic obesity” (19, 36). This

FIG. 2. Interpatient comparisons of presenting BMI and rate of BMI change. Data were stratified for: A, tumor histology, with craniopharyngiomas exhibiting increased BMI at presentation (P ⫽ 0.0001) and more rapid (although not significant) rate of BMI increase; B, radiation dosimetry to the hypothalamus, with those receiving greater than 51 Gy exhibiting more rapid rate of BMI increase (P ⫽ 0.0018); and C, the presence of residual endocrinopathy, with those who received any hormonal replacement gaining BMI most rapidly (P ⫽ 0.031). Data stratifications based on tumor location, extent of surgery, hydrocephalus, steroid use, or chemotherapy did not reveal statistical differences in presenting BMI or rates of BMI increase.


phenomenon can be suppressed by concomitant pancreatic vagotomy (37–39) to prevent increased muscarinic innervation of the pancreatic ␤-cell, with resultant insulin hypersecretion (40). Children with craniopharyngiomas demonstrate a similar clinical phenomenon, with increased insulin secretion in response to glucose (21, 41– 43). Recent studies by our group have demonstrated insulin hypersecretion in patients with hypothalamic obesity due to brain tumors or cranial irradiation. Octreotide, a long-acting somatostatin analog that binds to the SSTR5 receptor on the ␤-cell, results in inhibition of intracellular calcium influx and attenuation of insulin release, and reversal of the hyperinsulinemia and weight gain in this syndrome (23, 44). The VMH is the site of leptin, ghrelin, neuropepeptide Y-2, and insulin receptors, which transduce peripheral hormonal afferent signals to control efferent sympathetic and vagal modulation, appetite, and energy balance (8, 45– 49). The results of this retrospective analysis provide evidence for hypothalamic damage (probably the VMH) as the primary etiology for obesity in the brain tumor population, and down-play the roles of steroids, hydrocephalus, chemotherapy, and psychological factors. These results clearly recapitulate the phenomenon of hypothalamic obesity in rodents, and strengthen the concept of the hypothalamus as the biological regulator of energy balance in humans (45, 50, 51). These results also provide the radiation oncologist, neurooncologist, and neuroendocrinologist with objective criteria for risk assessment for the future development of obesity in this population, so that close follow-up and early preventive measures can be instituted (44). Clearly, diet, exercise, and medical interventions must be employed early, if the rate of intractable weight gain has any chance for attenuation. Finally, these results reiterate the sensitivity of the hypothalamus to both surgical intervention and/or external beam radiation. Rather than employing gross total or subtotal resection as a primary therapy, recent strategies have been developed to treat craniopharyngiomas more conservatively, using biopsy and focal irradiation (35, 52, 53). Our data suggest that strategies that limit hypothalamic radiation exposure to less than 51 Gy may reduce the future incidence of obesity. This will be particularly important in the treatment of tumors of the posterior fossa and temporal lobes where incidental irradiation occurs because of tumor location and the need to encompass peritumoral normal tissues. Acknowledgments We thank James Boyett and Dana Jones-Wallace for their statistical insights. Received July 29, 2002. Accepted October 30, 2002. Address all correspondence and requests for reprints to: Robert H. Lustig, M.D., Division of Pediatric Endocrinology, Box 0136, University of California San Francisco, 500 Parnassus Avenue, San Francisco, California 94143-0136. E-mail: [email protected]. This work was supported in part by the Cancer Center Support CORE Grant, P30CA12765, and the American Lebanese Syrian Associated Charities. Present address for R.H.L.: Department of Pediatrics, University of California San Francisco, San Francisco, California 94143-0136. Present address for K.S.: Division of Pediatric Oncology, Siriraj Hospital, Mahidol University, Bangkok, 10700 Thailand.

J Clin Endocrinol Metab, February 2003, 88(2):611– 616 615

Present address for S.R.R.: Department of Pediatrics, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio 45229-3039.

References 1. Shalet SM 1993 Radiation and pituitary dysfunction. N Engl J Med 328: 131–133 2. Sklar CA 1997 Growth and neuroendocrine dysfunction following therapy for childhood cancer. Pediatr Clin North Am 44:489 –503 3. Moshang T, Grimberg A 1996 The effects of irradiation and chemotherapy on growth. Endocrinol Metab Clin North Am 25:731–741 4. Boulad F, Sands S, Sklar C 1998 Late complications after bone marrow transplantation in children and adolescents. Curr Prob Pediatr 28:273–297 5. Kaste SC, Jones-Wallace D, Rose SR, Boyett JM, Lustig RH, Rivera GK, Pui CH, Hudson MM 2001 Bone mineral decrements in survivors of childhood acute lymphoblastic leukemia: frequency of occurrence and risk factors for their development. Leukemia 15:728 –734 6. Constine LS, Woolf PD, Cann D, Mick G, McCormick K, Raubertas RF, Rubin P 1993 Hypothalamic-pituitary dysfunction after radiation for brain tumors. N Engl J Med 328:87–94 7. Mulhern RK, Ochs J, Kun LE 1991 Changes in intellect associated with cranial radiation therapy. In: Gutin PH, Leibel SA, Sheline GE, eds. Radiation injury to the nervous system. New York: Raven; 325–340 8. Lustig RH 2001 The neuroendocrinology of childhood obesity. Pediatr Clin North Am 48:909 –930 9. Rose SR, Lustig RH, Pitukcheewanont P, Broome DC, Burghen GA, Li H, Hudson MM, Kun LE, Heideman RL 1999 Hidden central hypothyroidism in survivors of childhood cancer. J Clin Endocrinol Metab 84:4472– 4479 10. Didi M, Didcock E, Davies HA, Oligvy-Stuart AL, Wales JKH, Shalet SM 1995 High incidence of obesity in young adults after treatment of acute lymphoblastic leukemia of childhood. J Pediatr 127:63– 67 11. Sainsbury CPQ, Newcombe RG, Hughes IA 1985 Weight gain and height velocity during prolonged first remission from acute lymphoblastic leukaemia. Arch Dis Child 60:832– 836 12. Zee P, Chen CH 1986 Prevalence of obesity in children after therapy for acute lymphoblastic leukemia. Am J Pediatr Hematol Oncol 8:294 –299 13. Schell MJ, Ochs JJ, Schriock EA, Carter M 1992 A method of predicting adult height and obesity in long-term survivors of childhood acute lymphoblastic leukemia. J Clin Oncol 10:128 –133 14. Odame I, Reilly JJ, Gibson BES, Donaldson MDC 1994 Patterns of obesity in boys and girls after treatment for acute lymphoblastic leukaemia. Arch Dis Child 71:147–149 15. Craig F, Leiper AD, Stanhope R, Brain C, Meller ST, Nussey SS 1999 Sexually dimorphic and radiation dose dependent effect of cranial irradiation on body mass index. Arch Dis Child 81:500 –504 16. Sklar CA, Mertens AC, Walter A, Mitchell D, Nesbit M, O’Leary M, Hutchinson R, Meadows AT, Robison LL 2000 Changes in body mass index and prevalence of overweight in survivors of childhood acute lymphoblastic leukemia: role of cranial irradiation. Med Ped Oncol 35:91–95 17. Nysom K, Holm K, Michaelsen KF, Hertz H, Mu¨ller J, Mølgaard C 1999 Degree of fatness after treatment for acute lymphoblastic leukemia in childhood. J Clin Endocrinol Metab 84:4591– 4596 18. Stahnke N, Grubel G, Lagenstein I, Willig RP 1984 Long-term follow-up of children with craniopharyngioma. Eur J Pediatr 142:179 –185 19. Sorva R 1988 Children with craniopharyngioma: early growth failure and rapid post-operative weight gain. Acta Pediatr Scand 77:587–592 20. Sklar CA 1994 Craniopharyngioma: endocrine sequalae of treatment. Pediatr Neurosurg 21:120 –123 21. Pinto G, Bussieres L, Recasens C, Souberbielle JC, Zerah M, Brauner R 2000 Hormonal factors influencing weight and growth pattern in craniopharyngioma. Horm Res 53:163–169 22. Bray GA, Inoue S, Nishizawa Y 1981 Hypothalamic obesity. Diabetologia 20:366 –377 23. Lustig RH 2002 Hypothalamic obesity: the sixth cranial endocrinopathy. The Endocrinologist 12:210 –217 24. BMI curves 2001 http://www.cdc.gov/growthcharts 25. Rutter CM, Elashoff RM 1994 Analysis of longitudinal data: random coefficient regression modelling. Stat Med 13:1211–1231 26. Little RC, Milliken Ga, Stroup WS, Russell DW 1996 SAS system for mixed models. Cary, NC: SAS Institute 27. Troiano RP, Flegal KM, Kuczmarski RJ, Campbell SM, Johnson CL 1995 Overweight prevalance and trends for children and adolescents. Arch Ped Adol Med 149:1085–1091 28. Styne DM 2001 Childhood and adolescent obesity: prevalence and significance. Pediatr Clin North Am 48:823– 854 29. Gaynon PS, Lustig RH 1995 The use of glucocorticoids in acute lymphoblastic leukemia of childhood: molecular, cellular, and clinical considerations. Am J Pediatr Hematol Oncol 17:1–12 30. Casey BJ, Giedd JN, Thomas KM 2000 Structural and functional brain development and its relation to cognitive development. Biol Psychol 54:241–257 31. Barkovich AJ 2000 Concepts of myelin and myelinization in neuroradiology. Am J Neuroradiol 21:1099 –1109

616


32. Spunberg JJ, Chang CH, Goldman M, Auricchio E, Bell J 1981 Quality of long-term survival following irradiation for intracranial tumors under the age of two. Int J Radiat Oncol Biol Phys 7:727–736 33. Ellenberg L, McComb JG, Siegel SE, Stowe S 1987 Factors affecting intellectual outcome in pediatric brain tumor patients. Neurosurgery 21:638 – 644 34. Steen RG, Koury BSM, Granja CI, Xiong X, Wu S, Glass JO, Mulhern RK, Kun LE, Merchant TE 2001 Effect of ionizing radiation in the human brain: white matter and gray matter T1 in pediatric brain tumor patients treated with conformal radiation therapy. Int J Radiat Oncol Biol Phys 49:79 –91 35. Merchant TE, Kienha EN, Sanford RA, Mulhern RK, Thompson SJ, Wilson MW, Lustig RH, Kun LE 2002 Craniopharyngioma: the St. Jude Children’s Research Hospital experience 1984 –2001. Int J Radiat Oncol Biol Phys 53: 533–542 36. Powley TL, Laughton W 1981 Neural pathways involved in the hypothalamic integration of autonomic responses. Diabetologia 20:378 –387 37. Ionescu E, Rohner-Jeanrenaud F, Berthoud HR, Jeanrenaud B 1983 Increases in plasma insulin levels in response to electrical stimulation of the dorsal motor nucleus of the vagus nerve. Endocrinology 112:904 –910 38. Jeanrenaud B 1985 An hypothesis on the aetiology of obesity: dysfunction of the central nervous system as a primary cause. Diabetologia 28:502–513 39. Bray GA, York DA 1979 Hypothalamic and genetic obesity in experimental animals: an autonomic and endocrine hypothesis. Physiol Rev 59:719 – 809 40. Rohner-Jeanrenaud F, Jeanrenaud B 1985 Involvement of the cholinergic system in insulin and glucagon oversecretion of genetic preobesity. Endocrinology 116:830 – 834 41. Bray GA, Gallagher TF 1975 Manifestations of hypothalamic obesity in man: a comprehensive investigation of eight patients and a review of the literature. Medicine 54:301–333


42. Bray GA 1984 Syndromes of hypothalamic obesity in man. Pediatr Ann 13: 525–536 43. Reeves AG, Plum F 1972 Hyperphagia, rage, and dementia accompanying a ventromedial hypothalamic neoplasm. Arch Neurol 20:616 – 624 44. Lustig RH, Rose SR, Burghen GA, Velasquez-Mieyer P, Broome DC, Smith K, Li H, Hudson MM, Heideman RL, Kun LE 1999 Hypothalamic obesity in children caused by cranial insult: altered glucose and insulin dynamics, and reversal by a somatostatin agonist. J Pediatr 135:162–168 45. Schwartz MW, Woods SC, Porte D, Seeley RJ, Baskin DG 2000 Central nervous system control of food intake. Nature 404:661– 671 46. Baskin DG, Wilcox BJ, Figlewicz DP, Dorsa DM 1988 Insulin and insulin-like growth factors in the CNS. Trends Neurosci 11:107–111 47. Bru¨ ning JC, Gautam D, Burks DJ, Gillette J, Schubert M, Orban PC, Klein R, Krone W, Mu¨ ller-Wieland D, Kahn CR 2000 Role of brain insulin receptor in control of body weight and reproduction. Science 289:2122–2125 48. Schwartz MW, Figlewicz DP, Baskin DG, Woods SC, Porte D 1994 Insulin and the central regulation of energy balance: update 1994. Endocr Rev 2: 109 –113 49. Batterham RL, Cowley MA, Small CJ, Herzog H, Cohen MA, Dakin CL, Wren AM, Brynes AE, Low MJ, Ghatel MA, Cone RD, Bloom SR 2002 Gut hormone PYY3–36 physiologically inhibits food intake. Nature 418:650 – 654 50. Elmquist JK, Ahima RS, Elias CF, Flier JS, Saper CB 1998 Leptin activates distinct projections from the dorsomedial and ventromedial hypothalamic nuclei. Proc Natl Acad Sci USA 95:741–746 51. Elmquist JK, Elias CF, Saper CB 1999 From lesions to leptin: hypothalamic control of food intake and body weight. Neuron 22:221–232 52. Sanford RA 1994 Craniopharyngioma: results of survery of the American Society of Pediatric Neurosurgery. Pediatr Neurosurg 21(Suppl):39 – 43 53. Eisenstat DD 2001 Craniopharyngioma. Curr Treat Opin Neurol 3:77– 88