Introduction

Craniopharyngiomas are benign epithelial tumors arising from embryonic squamous remnants of Rathke’s pouch. They can be found anywhere along the path of the craniopharyngeal tract in the hypothalamic–pituitary regions. Craniopharyngioma is the most common tumor to affect the parasellar region in pediatric populations and they constitute up to 10% of childhood brain tumors [15, 28]. They can occur at any age, but are most common in childhood or adolescence. Despite their benign histological appearance, they often infiltrate surrounding critical parasellar structures and their aggressive behavior, even after successful treatment, may result in significant morbidity especially in endocrine function and satiety regulation [13, 14]. Extreme degrees of obesity have been reported as many as 50–80% of children undergoing extensive surgical treatment of craniopharyngiomas [7, 23, 38]. The mechanism has been postulated to be due to damage to the ventromedial hypothalamus (VMH). This causes insulin hypersecretion due to hyperactivity of parasympathetic nervous system [17, 19]. Despite adequate pituitary hormone replacement including growth hormone (GH) therapy, children with craniopharyngiomas often continue to have disproportionately heavier weight and higher body mass index (BMI) compared with children with other causes of GH deficiency [8].

Childhood obesity predisposes to insulin resistance and type 2 diabetes, dyslipidemia, hepatic steatosis/steatohepatitis, and hypertension, which are the features of the metabolic syndrome [40, 41]. The metabolic syndrome (MS), also called syndrome X, was first described in the 1988 and predisposes individuals to diabetes and cardiovascular disease [30]. Very few studies exist about the metabolic syndrome in children after craniopharyngiomas. Srinivasan et al. first demonstrated that children after craniopharyngioma surgery had more features of the metabolic syndrome compared with controls, but their fasting insulin and insulin sensitivity were not different with those in controls [37]. Recently, Simoneau-Roy et al. reported that children with craniopharygioma had more features of MS, increased insulin secretion, and lower insulin sensitivity compared with BMI-matched controls [34]. The predictors of this finding are not well described, and the insulin dynamics data is conflicting. To obtain more data, we performed an analysis of the changes in height, weight, BMI standard deviation scores (SDS), oral glucose tolerance test, and metabolic syndrome features of children with postsurgical craniopharyngioma and their risk factors compared with controls.

Subjects and methods

Subjects

Twelve subjects (seven males, five females), aged 7.7–18.1 years, were diagnosed with craniopharyngioma at Chulalongkorn University Hospital. They were recruited to take part in this study. Total or near total tumor removal, was performed at diagnosis in all subjects. The median time since surgery was 1.9 years (range, 1.0–11.0 years). Six subjects underwent more than one surgical resection, and four received adjuvant radiotherapy. All patients had tumor involvement of the hypothalamus.

Ten of the 12 subjects were obese (BMI > 95th percentile). All subjects had GH deficiency on previous insulin tolerance tests, but were not treated with GH because of financial problems. Seven subjects had a growth velocity above the 25th percentile for age without GH. T4 was commenced at 50–100 μg/m2/day and hydrocortisone at 5–10 mg/m2/day. This was later titrated according to response. Ten of the 12 subjects were receiving regular hydrocortisone replacement. Eight subjects were pre-pubertal, three were Tanner stage 3, and one was Tanner stage 5. Sex steroid replacement was given at appropriate ages in standard incremental regimens (four subjects). Desmopressin dose and frequency were titrated to control polyuria.

Controls

Twelve subjects were compared with healthy controls that were matched for age, sex, BMI, and pubertal stage. Healthy controls had previously presented to the endocrine clinic with a primary concern of overweight or obesity, and no underlying cause for their excess weight gain was found.

Ethical considerations

Informed consent was obtained from all parents and subjects or controls. The study was approved by the Institutional Review Board of the Faculty of Medicine, Chulalongkorn University. All subjects and controls were given detailed feedback about the results from the study.

Study protocol

At baseline, disease history and clinical features were recorded together with measurements of height, weight, waist, and hip circumferences, and BMI was calculated. The subjects and controls consumed a diet containing at least 250 g of carbohydrates per day for 3 days before the study. They were evaluated at 8 a.m., after a 12-h overnight fast. Blood pressure was measured three times while the subjects were seated, and the last two measurements were averaged for analysis. Physical examination included determination of the stage of puberty according to the criteria of Tanner.

Baseline blood samples were obtained from subjects while they were fasting using an indwelling venous line for measurement of levels of glucose, insulin, lipids, liver enzymes, free thyroxine, IGF-I, and hemoglobin A1C (HbA1C). An oral glucose tolerance test was then performed with the administration of 1.75 g of glucose per kilogram of body weight (maximal dose, 75 g). Blood samples were obtained 0, 30, 60, 90, and 120 min after glucose administration, for glucose and insulin measurements. The homeostasis model assessment (HOMA) index is an index of insulin resistance, calculated as fasting insulin concentration (μU/ml) × fasting glucose concentration (mmol/L)/22.5 [22].

Using values derived from the OGTT, whole body insulin sensitivity (WBISI) was calculated using the Matsuda index [21]. The insulinogenic index (IGI), is an estimate of insulin secretion, defined as the change in insulin from baseline to 30 min, divided by the change in glucose from baseline to 30 min, during OGTT [35]. The area-under-the-curve of insulin (AUCins) during OGTT is an index of insulin secretion, calculated using the trapezoidal rule [31].

Definition

The criteria used to diagnose the metabolic syndrome were modified from those of the National Cholesterol Education Program’s Adult Treatment Panel and the World Health Organization [4]. The subjects in our study were classified as having the metabolic syndrome if they met three or more of the following criteria for age and sex: a waist-to-hip ratio ≥0.90 in boys or ≥0.85 in girls, a triglyceride level above 150 mg/dl, high-density lipoprotein (HDL) cholesterol levels below 40 mg/dl in boys and below 50 mg/dl in girls, systolic or diastolic blood pressure above the 95th percentile, and impaired glucose tolerance. Impaired glucose tolerance was defined as a glucose level greater than 140 mg/dl but less than 200 mg/dl at 2 h. The diagnosis of diabetes mellitus was defined as a glucose level greater than 200 mg/dl at 2 h during an OGTT [2]. The degree of insulin resistance was determined with the use of a HOMA index [22]. The higher scores indicate greater insulin resistance.

In the absence of liver biopsy, presumed nonalcoholic steatohepatitis (NASH) was diagnosed by classical ultrasonographic hepatic appearances together with an elevated serum level of alanine aminotransferase (ALT).

Methods

Anthropometry

Height was measured to the nearest 0.1 cm, and weight was measured to the nearest 0.1 kg using standard techniques. These were expressed as the standard deviation scores for chronological age from the age- and sex-specific reference values used in Thailand (unpublished data). Obesity was defined as a BMI above the 95th percentile. BMI data are presented as SDS from the age- and sex-specific reference values. Waist and hip circumferences were measured to the nearest 0.1 cm using standard techniques. Pubertal status was assessed according to the standards of Tanner and Whitehouse.

Biochemical analysis

Commercial immunoassays were used to measure insulin [Electrochemiluminescence Immunoassay (ECLIA); Diagnostic Products Corporation, Los Angeles, CA, USA], free T4 (ECLIA; Roche Diagnostics, Indianapolis, IN, USA), IGF-I (ELISA; Diagnostic Systems Laboratories, Inc., Webster, TX, USA), and HbA1C (Immunoturbidity; Roche Diagnostics, Indianapolis, IN, USA). Glucose was measured on a Cobas Integra 400 plus (Roche Diagnostics, Indianapolis, IN, USA) using a hexokinase method.

Aspartate aminotransferase, ALT, total cholesterol, high-density lipoprotein cholesterol, and triglycerides (TG) were measured by standard enzymatic methods.

Statistical analysis

Data are expressed as the median and ranges. The Wilcoxon signed-rank test or Mann–Whitney test was used to compare continuous data between groups. Comparisons of proportions were performed using Fisher’s Exact tests. P < 0.05 was considered significant.

Results

Demographics, body weight, height, and BMI

Background characteristics of the study subjects and controls are shown in Table 1. There was no significant difference between subjects and controls for age (14.1 vs. 12.5 years; P = 0.136), and BMI SD score (3.20 vs. 4.61; P = 0.067). Eight of 12 individuals in both groups were in pre-pubertal stage. The craniopharyngioma group was significantly shorter than the control group (height SDS −2.05 vs. 1.04; P = 0.004) and weight SDS for chronological age was significantly lower in subjects than controls (0.89 vs. 3.42; P = 0.004). Nine of 12 subjects were severely obese (BMI SDS above 2.5). All patients exhibited significant weight gain after surgery (BMI SDS before surgery −0.52 vs. after surgery 3.20; P = 0.002; Fig. 1).

Table 1 Background characteristics of the subjects and controls
Full size table
Fig. 1
figure 1

Body mass index (BMI) SDS in subjects with childhood craniopharyngioma before surgery, and at the time of study

Full size image

Metabolic syndrome features

The waist to hip ratio was significantly higher in subjects compared with controls (0.96 vs. 0.89; P = 0.023). All subjects and controls had normal blood pressure. Fasting glucose, fasting insulin, peak insulin, sum of insulin levels, AUCins, the insulinogenic index, whole body insulin sensitivity, and the HOMA index were similar for subject–control pairs. Four of 12 subjects and four of 12 controls had fasting hyperinsulinemia (fasting insulin, >20 μU/ml). The craniopharyngioma group had significantly higher fasting TG (133 vs. 96 mg/dl; P = 0.019) and lower HDL/total cholesterol ratio (0.19 vs. 0.27; P = 0.012). Five of 12 subjects and one of 12 controls met the criteria for metabolic syndrome. One subject and one control had impaired glucose tolerance. Another subject with craniopharyngioma met the criteria for overt diabetes mellitus. Presumed nonalcoholic steatohepatitis was diagnosed in six of 12 subjects and two of 12 controls. The presumed NASH was diagnosed median 2.4 years (range 1.1–8.1 years) after craniopharyngioma surgery (Table 2).

Table 2 Metabolic data comparing subjects and controls
Full size table

Subjects with metabolic syndrome compared with those without metabolic syndrome

There was no significant difference between each group for age, sex, pubertal stage, and anthropometry (height SDS, weight SDS, BMI SDS, and waist to hip ratio). No differences were found in terms of serum IGF-I levels, hypothalamic involvement, irradiation, and progression of residual tumor after incomplete resection (Table 3).

Table 3 Characteristics of subjects with craniopharyngioma with and without metabolic syndrome (MS)
Full size table

The group of patients with metabolic syndrome had significantly higher fasting insulin (35.7 vs. 5.0 μU/ml; P = 0.01), HOMA index (6.35 vs. 1.13; P = 0.007), and HbA1C (6.30 vs. 5.55; P = 0.04), but had significantly decreased insulin sensitivity (median WBISI 1.64 vs. 9.74; P = 0.01). The metabolic syndrome group also had a worse lipid profile, with significantly higher fasting TG (292 vs. 111 mg/dl; P = 0.01) and lower HDL (26 vs. 50 mg/dl; P = 0.03). Alanine aminotransferase levels were significantly higher in subjects with metabolic syndrome compared with those not having metabolic syndrome (106 vs. 26 U/L; P = 0.03).

Discussion

Previous studies demonstrated that obesity is a major sequelae in children with craniopharyngioma, occurring in up to 50–80% of survivors [7, 23, 38]. Comparable with other reports, the rate of obesity was approximately 80% in our study. Most of our patients had normal BMI before surgery and developed rapid postoperative weight gain within a few months. Several factors may play a role in this phenomenon. Lustig et al. postulated that damage to the VMH, which causes insulin hypersecretion due to beta cell dysfunction, promotes the partitioning of ingested energy substrate into adipose tissue, leading to obesity [17, 19]. The weight gain was also driven by hyperphagia which results from the disruption of the mechanisms controlling satiety, hunger, and energy balance that occurs following hypothalamic damage [5, 13]. Possible mechanisms include lack of sensitivity to endogenous leptin [32], vagally mediated hyperinsulinemia, and autonomic imbalance [29]. High levels of the orexigenic gastric hormone ghrelin have not been found in these patients [9]. Muller et al. reported increased daytime sleepiness and reduced nocturnal melatonin levels in patients with childhood craniopharyngioma. This supports the hypothesis that physical activity might be decreased in these patients due to as yet unknown neuroendocrine disorders [26].

The prevalence of metabolic syndrome in childhood craniopharyngioma is not well known since there were only two reports that investigated this [34, 37]. Our findings suggest that the metabolic syndrome is very common (five of 12) among children and adolescents after craniopharyngioma surgery. The craniopharyngioma subjects had significantly higher waist to hip ratio, higher fasting TG, and lower HDL/total cholesterol ratio, compared with age-, sex-, BMI-, and pubertal stage-matched healthy controls. Fasting insulin, peak insulin, sum of insulin levels, AUCins, the insulinogenic index, whole body insulin sensitivity, and the HOMA index were similar for subject–control pairs. Lustig et al. first observed increased insulin and glucose following OGTT in hypothalamic obesity subjects compared with controls [20]. By contrast, Srinivasan et al. reported fasting glucose and insulin levels were similar for subject–control pairs, and found no difference in insulin sensitivity (Si) evaluated by intravenous glucose tolerance test (IVGTT) [37]. However, recent study demonstrated lower Si in craniopharyngioma subjects assessed by IVGTT, but not by OGTT [34]. Several factors that may explain the conflicting results of the insulin dynamics in this population include different methods and assessment of insulin secretion and sensitivity indices, small numbers of patients, the heterogeneity of subjects as well as difficulties in finding adequate controls. Childhood obesity predisposes to the metabolic syndrome and the severity of obesity and the prevalence of the metabolic syndrome were strongly associated [41]. GH deficiency is associated with changes in body composition, characterized by an increase in fat tissue and decreased lean body mass [10, 33]. Despite adequate pituitary hormone replacement including GH therapy, children with craniopharyngiomas continued to have disproportionately heavier weight and higher body mass index compared with children with other causes of GH deficiency [8]. Therefore, despite adequate hormone replacement, children and adolescents after craniopharyngioma surgery are still at risk for having the metabolic syndrome.

The craniopharyngioma patients that met the criteria of metabolic syndrome had significantly higher fasting insulin, HOMA index, and HbA1C, which reflected the greater severity of insulin resistance. One of our patients has been diagnosed with overt type 2 diabetes mellitus. A longitudinal cohort study is required to determine whether there will be an increase in prevalence of diabetes in craniopharyngioma patients compared with the normal population.

Previous studies assessing risk factors of excess weight gain after craniopharyngioma surgery are limited. Some studies have shown a relation between the degree of hypothalamic involvement and the severity of postoperative obesity in children with craniopharyngiomas [7, 24]. In contrast, some authors demonstrated that there was no relation between neuroanatomic extent of the tumor and weight gain [6, 36]. In another study of children surviving brain tumors, the main predictors of weight gain were age at diagnosis, radiation dose, and the presence of any endocrine deficiency [19].

We did not find significant differences in subjects with metabolic syndrome and those without metabolic syndrome in terms of age, sex, pubertal stage, anthropometry, serum IGF-I levels, hypothalamic involvement, irradiation, and progression of residual tumor after incomplete resection. The small numbers of our study might limit any definite conclusions about risk factors for developing the metabolic syndrome in children after craniopharyngioma surgery.

The presumed NASH was diagnosed in six of 12 subjects. Of those that had NASH, they tend to have the more severe degree of insulin resistance. It is well known that NASH is strongly related to obesity, insulin resistance, and other features of metabolic syndrome [3, 27]. Previous reports have also focused on the role of GHD in the pathogenesis of NASH, but this is still unclear [11, 39]. In our series, NASH developed quickly (median 2.4 years) after craniopharyngioma surgery. Similarly, previous study of 21 patients with a diagnosis of either panhypopituitarism or craniopharyngioma showed that NASH developed an average of 6.4 years after the diagnosis. Liver disease in these patients was severe; 60% of those biopsied had cirrhosis, and three of the 21 were either transplanted or died from liver-related causes during follow-up [1].

Treatment of hypothalamic obesity is challenging as this condition is poorly responsive to diet, physical activity, and most pharmacological therapy [16]. Previous study demonstrated insulin hypersecretion in this population [20]. Thus, suppression of insulin secretion with octreotide has been previously used in patients with craniopharyngioma and hypothalamic obesity and demonstrated modest effectiveness in weight reduction in some cases [18, 20]. Bariatric surgery has recently been proposed to be the treatment option in these patients. Either gastric banding or gastric bypass surgery has led to significant weight loss in several patients with history of childhood craniopharyngioma and subsequent morbid obesity [12, 25]. One case report after gastric bypass surgery also demonstrated the correction of fasting hyperinsulinemia, and normalization of postprandial insulin responses [12]. Given the limited available and effective treatment options for hypothalamic obesity, further studies are clearly warranted.

In summary, children after craniopharyngioma surgery are at risk of rapid weight gain and the development of metabolic syndrome that could be associated with a higher risk of atherosclerotic cardiovascular disease compared with BMI-matched healthy controls. Our study provides important implications for monitoring features of the metabolic syndrome, including abdominal obesity, type 2 diabetes mellitus, dyslipidemia, hypertension, and NASH in these patients. Further studies to better understand the mechanism by which rapid weight gain and metabolic syndrome occur in children after craniopharyngioma surgery, are required to design effective treatment and prevention.