Abstract
Context
Children with brain tumors may have pubertal onset at an inappropriately young chronologic age. Hypothalamic-pituitary irradiation ≥18Gy has been found to be a risk factor; age at irradiation is associated with pubertal timing. However, the underlying mechanisms are unknown.
Objective
To determine the impact of body mass index (BMI) and catch-up growth on pubertal timing in females treated for medulloblastoma and other embryonal tumors.
Design, Setting, and Patients
Retrospective cohort analysis of 90 female patients treated for medulloblastoma and other embryonal tumors at Dana-Farber Cancer Institute/Boston Children’s Hospital from 1996 to 2016. Eighteen individuals met inclusion criteria, with a mean ± SD follow-up period of 11.9 ± 3.4 years.
Main Outcome Measures
Multiple linear regression models for age at pubertal onset and bone age discrepancy from chronologic age at pubertal onset assessed the joint influences of age at irradiation, hypothalamic irradiation dose, undernutrition duration, BMI standard deviation score (SDS) at pubertal onset, and catch-up BMI SDS.
Results
The mean ± SD age of pubertal onset was 9.2 ± 1.3 years and hypothalamic radiation dose was 31.9 ± 9.9 Gy. There was a direct relationship between age at irradiation and age at pubertal onset (β = 0.323 ± 0.144 [standard error] year per year; P = 0.04) that was significantly attenuated after adjusting for BMI SDS at pubertal onset (P = 0.5) and catch-up BMI SDS (P = 0.08), suggesting that BMI is a mediator.
Conclusions
Both absolute and catch-up BMI SDS at pubertal onset are significant mediators of pubertal timing and bone age discrepancy in pediatric medulloblastoma and other embryonal tumors, and thus, are targetable risk factors to optimize pubertal timing.
Keywords: puberty, irradiation, undernutrition, medulloblastoma, embryonal tumors
Cranial irradiation greater than or equal to 18Gy to the hypothalamic-pituitary region has been associated with central precocious puberty in childhood cancer survivors (1). The age at irradiation appears to directly influence pubertal timing, as an earlier age at irradiation has been associated with both earlier chronological and bone ages at pubertal onset (2). Medulloblastoma and other embryonal tumors (formally classified as primitive neuroectodermal tumor [PNET]) are the most common malignant brain tumors in the pediatric population accounting for approximately 20% of all childhood brain tumors (3). Current treatment protocols for medulloblastoma and other embryonal tumors include at least 24Gy of cranial irradiation, and studies have reported a 10% to 20% incidence of precocious puberty in survivors of these tumors (4-7). Although early puberty is thought to be due to radiation damage to inhibitory components in the hypothalamus, the factors that may influence this phenomenon are unknown (8).
In the general population, nutritional status and body-mass index have been shown to influence pubertal timing (9). In regard to prenatal growth, both prematurity and small for gestational age birth weight have been associated with earlier pubertal onset (10, 11). In cases of combined prenatal-postnatal undernutrition, studies of children with stunted growth who are adopted from developing countries have reported a period of catch-up growth following adoption. This rapid growth is followed by early pubertal onset, the timing of which appears to be dependent on the degree of growth stunting and the rate of catch-up growth (9).
Similarly, malnutrition is a significant complication of cancer treatment in the pediatric population. Among pediatric cancer patients, those with medulloblastoma have the greatest incidence (up to 94%) and duration of malnutrition (12, 13). We investigated the hypothesis that undernutrition and subsequent weight acceleration influences pubertal timing in girls with medulloblastoma and embryonal tumors. We examined the body-mass index (BMI) standard deviation scores (SDS) at cancer diagnosis, BMI nadir, and onset of puberty to determine the effect of BMI and catch-up weight gain on pubertal timing and bone age. Since age of irradiation has been linked to pubertal timing, we also examined the impact of radiation treatment parameters on chronological age and bone age at pubertal onset.
Materials and Methods
Study participants and eligibility criteria
This single-institution cohort study evaluated all female children diagnosed with medulloblastoma or embryonal tumors not otherwise specified at Boston Children’s Hospital and Dana-Farber Cancer Institute from 1996 to 2016. Institutional Review Board approval was obtained for the collection of these data. Individuals who were deceased prior to pubertal onset or who were lost to follow-up were excluded. In order to assess for precocious puberty, commonly defined as pubertal onset (Tanner stage 2 breasts by palpation) at an age less than 8 years (14), individuals who were 8 years or older at the time of diagnosis were excluded from analysis. All remaining individuals underwent electronic chart review to assess their endocrine evaluations and growth parameters, and individuals with the following characteristics were excluded: 1) incomplete endocrine evaluation, including pubertal status; 2) prepubertal status at the last follow-up visit; 3) diagnosis of primary ovarian insufficiency or hypogonadotropic hypogonadism; and 4) incomplete growth data, including follow-up BMI measurements.
Treatment
All individuals underwent maximally safe resection of the medulloblastoma or embryonal tumor, followed by risk-adapted craniospinal irradiation and chemotherapy. Individuals received either photon, proton, or mixed craniospinal irradiation and one of the following chemotherapy regimens: 1) cisplatin, cyclophosphamide, and vincristine with or without stem-cell rescue; 2) cisplatin, cyclophosphamide, lomustine (CCNU), and vincristine; 3) cisplatin, cyclophosphamide, etoposide, and vincristine with or without intrathecal mafosfamide; 4) cisplatin, cyclophosphamide, carboplatin, etoposide, vincristine, thiotepa, and stem-cell rescue; 5) cisplatin, CCNU, and vincristine; or 6) carboplatin, etoposide, vincristine, and intrathecal mafosfamide. Cranial irradiation doses to the hypothalamic region are reported.
Endocrine evaluation
All individuals underwent endocrine evaluations by a pediatric endocrinologist at least twice-yearly following irradiation and chemotherapy. The onset of puberty was defined by achieving Tanner stage 2 of breast development by palpation by an endocrinologist, and all individuals underwent subsequent biochemical assessment of gonadal function confirming central origin of puberty. All individuals underwent a bone age evaluation at or following pubertal onset. Bone age was estimated independently by a pediatric endocrinologist and a radiologist, according to the standards of Greulich and Pyle (15). In cases in which these bone age evaluations were discrepant, the bone age was evaluated by a second pediatric endocrinologist who was blinded to the individual’s pubertal status, and the older of the 2 pediatric endocrinologists’ bone age evaluations was reported. Bone age discrepancy at pubertal onset was calculated as the difference between bone age and chronological age.
Precocious puberty was defined as Tanner stage 2 breast development at an age less than 8 years. Central activation of the hypothalamic-pituitary-gonadal axis was confirmed by a pubertal morning luteinizing hormone (LH) level or a pubertal stimulated LH level following leuprolide stimulation. Hypothyroidism was defined by one of the following criteria: 1) thyrotropin (thyroid stimulating hormone; TSH) >10 mIU/L with low or normal levels of total thyroxine (T4) or free T4; 2) TSH persistently between 5 and 10 mIU/L with low or normal levels of total T4 or free T4; 3) low total T4 or free T4 levels; or 4) in one individual clinical symptoms of hypothyroidism with low-normal total T4 that had been falling over time. Adrenal insufficiency was defined by a low 8 am cortisol level <5 ug/dL (<138 nmol/L) and clinical symptoms of adrenal insufficiency. The diagnosis of growth hormone (GH) deficiency was based upon a decelerating growth velocity and low insulin-like growth factor 1 (IGF-I) level, with a low IGFBP-3 level (if available). In individuals with a poor growth rate, persistently low IGF-I levels, and normal weight gain, GH deficiency was diagnosed clinically without GH stimulation testing. Other individuals underwent GH provocative testing with 2 tests; either an insulin tolerance test and glucagon stimulation or arginine and glucagon stimulation. A peak GH level <7 ng/mL was considered consistent with a diagnosis of GH deficiency, except for one individual who had a borderline peak value of 7.64 ng/mL, a low IGF-I level, and a clinical picture consistent with GH deficiency.
Anthropometric measures and malnutrition
Weight (in kg) and height (in cm) were obtained in specialty outpatient oncology and/or endocrinology clinics from diagnosis until last follow-up visit. BMI (in kg/m2) was transformed to a standard deviation score (SDS) according to the National Center for Health Statistics of the Center for Disease Control standards for the US population (16). BMI SDS was identified at cancer diagnosis, BMI nadir, and onset of puberty. Catch-up BMI SDS was defined as the difference between BMI SDS at pubertal onset and BMI nadir.
Malnutrition was defined as weight loss >10% from weight at diagnosis, or a BMI or weight-for-length (for the one child less than 2 years at diagnosis) less than −2.0 SDS (13). Recovery from malnutrition occurred when a participant’s weight or BMI exceeded and remained above the malnutrition threshold for a period ≥1 month. For individuals who met the definition for malnutrition as weight loss >10% from weight at diagnosis only, recovery from malnutrition was defined as achieving a weight that exceeded the weight that defined malnutrition (ie, 90% of the weight at diagnosis). For discordant measurements, the highest weight or BMI or weight-for-length in a 7-day period was used. The total duration of malnutrition was calculated and had to exceed 1 month. Any need for enteral supplementation by gastric tube was reviewed.
Statistical analysis
Statistical analyses were conducted with SAS software (version 9.2; SAS Institute, Inc, Cary, NC). Participant characteristics and outcomes are presented as mean ± standard deviation (SD) for variables with normal distributions and median (interquartile range) for variables with nonnormal distributions. All regression coefficients are presented as β ± SE (standard error). Nutritional and treatment parameters and the number of individuals with GH deficiency and hypothyroidism were compared between individuals with an inappropriately early bone age at pubertal onset (<8 years 10 months) and those with an appropriate bone age (≥8 years 10 months). Since GH deficiency can impact growth and pubertal timing, anthropometric measurements, treatment parameters, age at pubertal onset, and malnutrition duration were compared between individuals with and without GH deficiency. A Student independent-sample t-test was used to assess differences between continuous parameters, and a Fisher exact test was used for dichotomous parameters.
Multiple linear regression models for age at puberty and bone age discrepancy at pubertal onset assessed the joint influences of age at irradiation, hypothalamic irradiation dose, malnutrition duration, BMI SDS at pubertal onset, and catch-up BMI SDS. All tests were 2-sided, and a P-value of less than 0.05 was considered statistically significant.
Results
Participant characteristics
Between 1996 and 2016, 90 female individuals were diagnosed with medulloblastoma or embryonal tumor not otherwise specified at Boston Children’s Hospital and Dana-Farber Cancer Institute. Figure 1 details participant eligibility and selection based on age at diagnosis, endocrine evaluation, and growth data. Eighteen individuals met inclusion criteria, 15 with medulloblastoma and 3 with embryonal tumor not otherwise specified. Participant demographics and treatment characteristics are summarized in Table 1.
Figure 1.
Eligibility screening of individuals. n1 denotes number of medulloblastoma individuals and n2 denotes number of embryonal tumor (previously classified as PNET) individuals.
Table 1.
Demographic and Treatment Characteristics
| Characteristic (n = 18) | Mean ± SD, n (%) |
|---|---|
| Age at diagnosis, y) | 5.4 ± 2.1 |
| Follow-up duration, y) | 11.9 ± 3.4 |
| Race and ethnicity | |
| Black, African American | |
| Hispanic | 0 (0) |
| Non-Hispanic | 1 (6) |
| White, Caucasian | |
| Hispanic | 1 (6) |
| Non-Hispanic | 14 (78) |
| Other | |
| Hispanic | 2 (11) |
| Non-Hispanic | 0 (0) |
| Diagnosis | |
| Medulloblastoma | 15 (83) |
| High-risk | 4 (22) |
| Other embryonal tumors | 3 (17) |
| High-risk | 2 (11) |
| Treatment | |
| Age at irradiation | 5.7 ± 1.9 |
| Irradiation type | |
| Photon | 9 (50) |
| Proton | 7 (39) |
| Mixed | 2 (11) |
| Hypothalamic irradiation dose, Gy | 31.9 ± 9.9 |
| Chemotherapy regimen | |
| CDDP + CTX + VCR | 3 (17) |
| CDDP + CTX + VCR + Stem-cell rescue | 2 (11) |
| CDDP + CTX + CCNU + VCR | 2 (11) |
| CDDP + CTX + ETOP + VCR | 2 (11) |
| CDDP + CTX + ETOP + VCR + IT MF | 1 (6) |
| CDDP + CTX + CBDCA + ETOP + VCR + Thiotepa + Stem cell rescue | 1 (6) |
| CDDP + CCNU + VCR | 6 (33) |
| CBDCA + ETOP + VCR + IT MF | 1 (6) |
Abbreviations: CBDCA, carboplatin; CCNU, lomustine; CDDP, cisplatin; CTX, cyclophosphamide; ETOP, etoposide IT, intrathecal; MF, mafosfamide; SD, standard deviation; VCR, vincristine.
Pubertal timing
The mean ± SD age of pubertal onset was 9.2 ± 1.3 years. All individuals had biochemically confirmed central activation of the hypothalamic-pituitary-gonadal axis with a pubertal LH level and an estradiol level (if available) in the pubertal range, within the first year of pubertal onset. The median (interquartile range) baseline LH was 2.1 (1.0-4.6) IU/L.
In the majority of cases (12 of 18), the bone age interpretations of the pediatric endocrinologist and radiologist were concordant. In the remaining 6 cases, the disagreement was less than 1 Greulich and Pyle standard for 5 cases and less than 2 standards for 1 case, and a second blinded pediatric endocrinologist provided an additional bone age evaluation. At pubertal onset, the mean ± SD bone age discrepancy was −0.7 ± 1.4 years, indicating an overall delay in bone age at pubertal onset compared to chronological age. At pubertal onset, half of the individuals (n = 9) had a bone age ≥8 years 10 months, and the remaining half had an inappropriately early bone age <8 years 10 months. There was no significant difference in anthropometric measurements, treatment parameters, age at pubertal onset, or malnutrition duration between the 2 bone age groups, though there was a suggestion that individuals with bone age <8 years 10 months had earlier age at pubertal onset and higher doses of hypothalamic irradiation (Table 2).
Table 2.
Comparison of Individuals with Bone Age < and ≥ 8 Years 10 Months at Pubertal Onset
| Parameter | <8 y 10 mo (n = 9) | ≥ 8 y 10 mo (n = 9) | P |
|---|---|---|---|
| Mean ± SD, n (%) | |||
| ∆ BMI SDS nadir | -2.2 ± 0.9 | -1.7 ± 0.9 | 0.2 |
| ∆ BMI SDS T2 | 2.1 ± 1.3 | 2.4 ± 1.3 | 0.6 |
| BMI SDS T2 | 0.1 ± 1.2 | 0.2 ± 1.5 | 0.8 |
| Hypothalamic RT dose, Gy | 35.4 ± 9.9 | 28.7 ± 9.4 | 0.2 |
| Age at diagnosis, y | 5.3 ± 1.8 | 5.6 ± 2.5 | 0.8 |
| Age at RT, y | 5.5 ± 1.6 | 5.8 ± 2.3 | 0.8 |
| Age at puberty, y | 8.8 ± 1.1 | 9.5 ± 1.4 | 0.3 |
| Malnutrition duration, mo | 4.9 ± 5.3 | 5.0 ± 4.4 | 1 |
| GHD deficiency | 7 (78) | 5 (56) | 0.6 |
| Hypothyroidism | 9 (100) | 6 (67) | 0.2 |
Abbreviations: BMI, body mass index; ∆BMI SDS nadir, difference in BMI SDS between nadir and diagnosis; ∆BMI SDS T2, difference in BMI SDS between T2 and nadir; GHD, GH deficiency; Gy, gray; mo, months; RT, radiation therapy; SD, standard deviation; SDS, standard deviation score; T2, Tanner stage 2/pubertal onset.
Two individuals (11%) were diagnosed with precocious puberty (Tanner 2 stage of breast development at < 8 years), and an additional 6 individuals were diagnosed with “early puberty” based on an overall clinical impression from the physical exam, laboratory evaluation, and bone age. All 6 individuals reached Tanner 2 stage of breast development between 8 and 9 years. For 5 of these individuals, the diagnosis of “early puberty” was based on bone ages that were inappropriately young at the time of pubertal onset (<8 years 10 months). In the remaining individual, the diagnosis was based on the rapid rate of pubertal progression; menarche occurred approximately 9 months after Tanner 2 stage breasts were first noted on examination.
Nutritional outcomes
Anthropometric parameters during treatment, at pubertal onset, and at follow-up are summarized in Table 3. Overall, individuals experienced a drop in BMI SDS during treatment, and 78% (n = 14) met the definition of malnutrition for a mean duration of 4.9 ± 4.7 months. A substantial proportion (n = 8, 44%) required enteral supplementation by gastric tube. Mean BMI SDS normalized to 0.2 ± 1.4 at pubertal onset following a catch-up period (Table 3).
Table 3.
Anthropometric, Endocrinologic, and Nutritional Parameters
| Parameter | Mean ± SD, n (%) |
|---|---|
| Anthropometric parameters | |
| BMI SDS diagnosis | −0.2 ± 1.4 |
| BMI SDS nadir | −2.1 ± 1.5 |
| ∆ BMI SDS nadir | −1.9 ± 0.9 |
| BMI SDS at T2 | 0.2 ± 1.4 |
| ∆ BMI SDS T2 | 2.3 ± 1.3 |
| BMI SDS at adult follow-up | 0.5 ± 1.3 |
| Endocrine evaluation | |
| Age at pubertal onset, y | 9.2 ± 1.3 |
| Bone age at T2, y | 8.9 ± 1.6 |
| ∆ Bone age at T2, y | −0.7 ± 1.4 |
| Precocious puberty | 2 (11) |
| Hypothyroidism | 15 (83) |
| GH deficiency | 12 (67) |
| Adrenal insufficiency | 1 (6) |
| Nutrition | |
| Malnutrition | 14 (78) |
| Duration of malnutrition, mo | 4.9 ± 4.7 |
| G-tube supplementation | 8 (44) |
Abbreviations: BMI, body mass index; ∆BMI nadir, difference in BMI SDS between nadir and diagnosis; ∆BMI T2, difference in BMI SDS between T2 and nadir; ∆ Bone age, difference between chronological age and bone age; mo, months; SD, standard deviation; SDS, standard deviation score; T2, Tanner stage 2/pubertal onset.
Endocrinopathies
Hypothyroidism (most commonly compensated primary) was the most common endocrinopathy, affecting 83% of individuals (n = 15), followed by GH deficiency (n = 12, 67%), precocious puberty (n = 2, 11%), and adrenal insufficiency (n = 1, 6%) (Table 3).
To assess potential impact on pubertal timing, thyroid function and GH deficiency diagnoses and treatment courses were examined in relation to pubertal onset. All individuals with or without a diagnosis of hypothyroidism underwent biochemical evaluation of thyroid function with measurement of a TSH and a total T4 or free T4 level within 6 months of pubertal onset. Mean ± SD TSH, total T4, and free T4 were 3.01 ± 1.69 mIU/L (n = 18), 9.10 ± 2.30 µg/dL (117.1 ± 29.6 nmol/L) (n = 11), and 1.43 ± 0.21 ng/dL (18.4 ± 2.7 pmol/L) (n = 9), respectively. In all but 1 individual, TSH and total T4 or free T4 levels were within normal reference ranges for assay. The remaining individual had a presumed diagnosis of primary hypothyroidism with superimposed TSH deficiency, and the TSH was mildly elevated to 7.19 mIU/L (reference range, 0.3-6.2) with a low total T4 of 4.1 µg/dL (52.8 nmol/L) (reference range, 6.0-12.3 µg/dL [77.2-158.3 nmol/L]) after stopping levothyroxine replacement against medical advice. After resuming levothyroxine treatment, repeat laboratory evaluations demonstrated euthyroid status.
GH deficiency was diagnosed and treated more than 1 year prior to pubertal onset in 3 out of 12 individuals, and 4 were treated within 6 months of pubertal onset. In the remaining 5 individuals, all were treated at least 1 year after pubertal onset. GH replacement therapy was initiated following treatment of hypothyroidism, optimization of nutrition, insurance approval, and discussion with the individual’s primary oncologist and family. In all cases, treatment was started at least 18 months following completion of irradiation and chemotherapy. Although the majority of individuals started GH replacement therapy within 4 months after diagnosis of GH deficiency, GH replacement therapy was initiated more than 1 year after diagnosis in 4 individuals due to ongoing discussions with the family and/or oncologist or concerns regarding weight and nutrition. Individuals with GH deficiency had lower BMI SDS at pubertal onset and a greater discrepancy in bone age at pubertal onset compared with those without GH deficiency (P = 0.01, 0.03, respectively) (Table 4). In addition, 58% of individuals with GH deficiency had an abnormally early bone age at pubertal onset that was <8 years 10 months.
Table 4.
Comparison of Individuals With and Without GH Deficiency
| GHD | No GHD | P | |
|---|---|---|---|
| n = 12 | n = 6 | ||
| Parameter | Mean ± SD, n (%) | ||
| ∆ BMI SDS nadir | −2.2 ± 0.9 | −1.3 ± 0.8 | 0.06 |
| BMI SDS T2 | −0.4 ± 1.1 | 1.3 ± 1.3 | 0.01 |
| Hypothalamic RT dose, Gy | 35.1 ± 8.5 | 24.2 ± 9.7 | 0.04 |
| Age at RT, y | 6.4 ± 1.4 | 4.2 ± 2.2 | 0.02 |
| Age at puberty, y | 9.4 ± 0.9 | 8.6 ± 1.8 | 0.2 |
| ∆ Bone age at puberty | −1.2 ± 1.0 | 0.2 ± 1.7 | 0.03 |
| Bone age T2 <8 y 10 mo | 7 (58) | 2 (33) | 0.6 |
| Malnutrition duration, mo | 6.3 ± 4.4 | 2.2 ± 4.4 | 0.08 |
Abbreviations: BMI, body mass index; ∆BMI SDS nadir, difference in BMI SDS between nadir and diagnosis; ∆BMI SDS T2, difference in BMI SDS between T2 and nadir; ∆ Bone age, difference between bone age and chronological age; GHD, GH deficiency; Gy, gray; mo, months; RT, radiation therapy; SD, standard deviation; SDS, standard deviation score; T2, Tanner stage 2/pubertal onset.
All individuals with precocious puberty were treated at the time of diagnosis with a gonadotropin-releasing hormone (GnRH) agonist. For the 1 individual diagnosed with adrenal insufficiency, steroid replacement was initiated during an acute inpatient hospitalization. As there was a concern for possible iatrogenic central adrenal insufficiency, maintenance steroids were weaned in the outpatient setting and repeat 8 am cortisol levels demonstrated normal levels >10 ug/dL (>276 nmol/L).
Effects of treatment and nutritional parameters on pubertal timing and bone age discrepancy
Younger age at irradiation was correlated with a younger age at pubertal onset in a bivariate regression model (β = 0.323 ± 0.144 year per year; P = 0.04) (Fig. 2a, Table 5). A higher BMI SDS at pubertal onset was also associated with a younger age at pubertal onset (β = −0.603 ± 0.167 year per 1 SDS; P = 0.002) (Table 5). In addition, a younger age at irradiation was associated with a higher BMI SDS (β = −0.482 ± 0.139 SDS per year; P = 0.003) and higher catch-up BMI SDS at pubertal onset (β = −0.320 ± 0.144 SDS per year; P = 0.04). To investigate the relationships between age at irradiation, BMI SDS and catch-up BMI SDS, and age at pubertal onset, multiple regression was performed. The relationship between age at irradiation and age at pubertal onset appeared to be mediated by the BMI SDS at pubertal onset and catch-up weight gain, as the association between age at irradiation and age at pubertal onset was no longer significant after adjusting for BMI SDS at pubertal onset and catch-up BMI SDS in covariate-adjusted regression models (Table 6).
Figure 2.
Relationship of age at irradiation with age at puberty and bone age discrepancy. (a) shows individuals plotted by age at irradiation and age at puberty. (b) shows individuals plotted by age at irradiation and bone age discrepancy (difference between bone age and chronological age). Abbreviation: yr, year.
Table 5.
Influence of Treatment and Nutrition Parameters on Age at Puberty and Bone Age Discrepancy
| Predictor (Parameter) | Regression Coefficient (95% CI); pa | |||
|---|---|---|---|---|
| Age at Pubertyb | ∆ Bone Age at Pubertyb | |||
| β ± SE (95% CI) | P | β ± SE (95% CI) | P | |
| Treatment | ||||
| Age at RT, y | 0.323 ± 0.144 (0.017, 0.628) | 0.04 | −0.357 ± 0.160 (−0.697, −0.017) | 0.04 |
| Hypothalamic RT dose, Gy | 0.055 ± 0.030 (−0.009, 0.119) | 0.09 | −0.036 ± 0.036 (−0.111, 0.040) | 0.3 |
| Nutrition | ||||
| Malnutrition duration, mo | 0.090 ± 0.064 (−0.044, 0.225) | 0.2 | −0.082 ± 0.072 (−0.235, 0.070) | 0.3 |
| BMI SDS diagnosis | −0.220 ± 0.224 (−0.695, 0.255) | 0.3 | −0.036 ± 0.256 (−0.579, 0.507) | 0.9 |
| BMI SDS nadir | −0.254 ± 0.202 (−0.682, 0.174) | 0.2 | 0.177 ± 0.231 (−0.313, 0.667) | 0.5 |
| ∆BMI SDS nadir | −0.182 ± 0.341 (−0.905, 0.542) | 0.6 | 0.548 ± 0.357 (−0.209, 1.305) | 0.1 |
| BMI SDS T2 | −0.603 ± 0.167 (−0.956, −0.249) | 0.002 | 0.693 ± 0.180 (0.313, 1.074) | 0.001 |
| ∆BMI SDS T2 | −0.394 ± 0.230 (−0.882, 0.095) | 0.1 | 0.615 ± 0.232 (0.123, 1.107) | 0.02 |
Abbreviations: BMI, body mass index; ∆BMI SDS nadir, difference in BMI SDS between nadir and diagnosis; ∆BMI SDS T2, difference in BMI SDS between T2 and nadir; ∆ Bone age, difference between bone age and chronological age; CI, confidence interval; Gy, gray; mo, months; RT, radiation therapy; SDS, standard deviation score; SE, standard error; T2, Tanner stage 2/pubertal onset; .
p* tests hypothesis of no association (zero regression coefficient).
a Bivariate regression, unadjusted (indicated predictor only).
b units = year of age or bone-age discrepancy per unit of row variable.
Table 6.
Influence of Age at Irradiation on Age at Puberty and Bone Age Discrepancy, as Adjusted for Other Parameters
| Predictors (Parameters) | Regression Coefficient for Age at Irradiation (95% CI); pa | |||
|---|---|---|---|---|
| Age at Pubertyb | ∆ Bone Age at Pubertyb | |||
| β ± SE (95% CI) | P | β ± SE (95% CI) | P | |
| Age at RT only, y | 0.323 ± 0.144 (0.017, 0.628) | 0.04 | −0.357 ± 0.160 (−0.697, −0.017) | 0.04 |
| Age at RT, RT dose, Gy | 0.379 ± 0.134 (0.092, 0.666) | 0.01 | −0.453 ± 0.157 (−0.789, −0.116) | 0.01 |
| Age at RT, RT dose, Malnut (mo) | 0.353 ± 0.160 (0.006, 0.699) | 0.047 | −0.443 ± 0.189 (−0.850, −0.035) | 0.04 |
| Age at RT, RT dose, Malnut, BMI SDS T2 | 0.147 ± 0.205 (−0.299, 0.594) | 0.5 | −0.157 ± 0.232 (−0.662, 0.348) | 0.5 |
| Age at RT, RT dose, Malnut, ∆BMI SDS T2 | 0.392 ± 0.202 (−0.047, 0.832) | 0.08 | −0.355 ± 0.234 (−0.866, 0.155) | 0.2 |
Abbreviations: BMI, body mass index; ∆BMI SDS T2, difference in BMI SDS between T2 and nadir; ∆Bone age, difference between bone age and chronological age; CI, confidence interval; Gy, gray; Malnut, malnutrition duration in months; mo, months; RT, radiation therapy; SDS, standard deviation score; SE, standard error; T2, Tanner stage 2.
p* tests hypotheses that there are no associations of age at puberty or bone age discrepancy with age at irradiation (zero regression coefficient), after adjustment for other predictors (if any).
a Multiple regression, all predictors mutually adjusted (indicated predictor only).
b units = year of age or bone-age discrepancy per year of age at irradiation.
Greater age at irradiation was associated with greater bone age delay (β = −0.357 ± 0.160 year per year; P = 0.04) (Fig. 2b, Table 5). This association was attenuated after adjusting for BMI SDS at pubertal onset and catch-up BMI SDS, suggesting that the BMI SDS variables also mediate the relationship between age at irradiation and bone age delay (Table 6).
Discussion
We have demonstrated that both absolute and catch-up BMI SDS at pubertal onset are significant mediators of pubertal timing and bone age discrepancy in pediatric patients with medulloblastoma and embryonal tumors. Earlier age at irradiation is associated with a higher BMI SDS and greater catch-up BMI SDS at pubertal onset, which precedes earlier pubertal timing. These associations highlight nutritional status during and following treatment as a critical influence on subsequent growth and pubertal timing in children with medulloblastoma and embryonal tumors.
Prior studies in children with brain tumors and GH deficiency have also shown that an earlier age at irradiation is associated with earlier age at pubertal onset (2). The mechanism of the association is unknown, although a proposed mechanism is damage to inhibitory components on the hypothalamus, leading to increased GnRH signaling that activates the hypothalamic-pituitary-axis (2). However, our finding that BMI mediates the effect of timing of irradiation on pubertal timing suggests a more complex regulation of pubertal timing that intimately involves nutrition and weight. Individuals who were younger at the time of irradiation had more robust catch-up weight gain and absolute BMI SDS by the time of pubertal onset that preceded earlier pubertal timing (Fig. 3). One possible explanation for this observation is that children who were younger at diagnosis had more time to recover from malnutrition and undergo catch-up weight gain prior to pubertal onset. It is also possible that younger children were counseled more aggressively and treated with nutritional supplementation during and following irradiation and chemotherapy, whereas older children may have had more significant side effects of treatment, such as nausea, that impeded feeding tolerance. Data on the management of undernutrition, including counseling, monitoring, and treatment, were not available for analysis. In addition, the age at cranial irradiation exposure on the developing brain may affect the neuroendocrine regulation of metabolism during catch-up growth, such that an earlier age at irradiation may lead to a metabolic state that favors anabolic pathways.
Figure 3.
Mediator role of BMI and catch-up weight gain at puberty. (a) shows the mediator roles on the relationship between age at radiation and age at puberty. (b) shows the mediator roles on the relationship between age at radiation and bone age discrepancy and illustrates possible influences of GH deficiency and yet-to-be identified skeletal maturation factors. Abbreviation: HPG, hypothalamic-pituitary-gonadal.
It is known that higher prepubertal BMI is associated with earlier pubertal onset in healthy children (17, 18). Proposed mechanisms include adiposity as a key permissive signal of central pubertal activation through leptin signaling to the hypothalamus. In addition, physiologic insulin resistance during puberty may be amplified in obese children, which is thought to stimulate excess adrenal androgen production that, in turn, advances pubertal development through peripheral and/or central influences (19). These influences may also advance skeletal maturation, since higher BMI has been highly correlated with more advanced bone age in obese children (20). BMI also appears to influence skeletal maturation and bone age in our sample. Thus, absolute BMI prior to and during pubertal onset appears to have a critical influence on pubertal timing and skeletal development (Fig. 3a).
The role of catch-up weight gain following a period of undernutrition on pubertal timing is less clear. In cases of prenatal and combined prenatal-postnatal undernutrition, catch-up growth in the postnatal period in premature, small for gestational age, and adopted children with stunted growth have been associated with earlier pubertal onset (9-11). Infants with diencephalic syndrome, a rare cause of severe failure to thrive and emaciation associated with central nervous system tumors, have been reported to undergo precocious puberty following tumor treatment and catch-up weight gain (21). In a case series of 11 individuals with diencephalic syndrome, increases in BMI were correlated with higher leptin and free leptin levels, and thus, rapid changes in leptin were proposed to be involved in the development of precocious puberty (21). However, in cases of isolated postnatal undernutrition in later childhood, including due to inflammatory bowel disease, asthma, and eating disorders, growth is often delayed initially and may be followed by a period of catch-up growth that has not been associated with earlier pubertal timing (9). Our finding that catch-up weight gain following treatment for medulloblastoma and embryonal tumors mediates pubertal onset provides direct evidence that catch-up growth in isolated postnatal undernutrition may influence pubertal timing (Fig. 3a). As children with brain tumors often experience significant and prolonged undernutrition, particularly notable in diencephalic syndrome, it is possible that there is a critical window for catch-up growth in early childhood that may not be similarly optimized as in other chronic childhood diseases. In addition, as discussed above, cranial irradiation exposure during childhood brain development may also affect the neuroendocrine regulation of metabolism during catch-up growth (22). Thus, the degree, timing, and neuroendocrine regulation of catch-up growth may be critical components in pubertal regulation.
Another potential influence on the associations between BMI and pubertal timing and bone age discrepancy is GH deficiency. Individuals with GH deficiency had a greater bone age discrepancy, and the majority of these individuals had a bone age that was abnormally early at pubertal onset, less than 8 years 10 months. All individuals with GH deficiency were treated, with most individuals starting GH replacement therapy within months after diagnosis of GH deficiency, although therapy initiation was delayed for up to over 1 year in several cases. Thus, untreated or partially treated GH deficiency during the peri-pubertal period may have contributed to delayed skeletal maturation at pubertal onset. However, patients with GH deficiency had a lower BMI SDS at nadir and later pubertal onset compared to individuals without GH deficiency; this is opposite to the norm where GH deficiency is associated with overweight or obesity (23). Thus, the delayed skeletal maturation is not likely related to GH deficiency alone, but rather multiple factors that include nutrition and BMI.
The finding of pubertal onset at an abnormally early bone age has been previously described in a study of children with brain tumors and GH deficiency (2). In contrast, maturation of the skeleton and the hypothalamic-pituitary-gonadal axis has been shown to be largely synchronous in other conditions that affect skeletal maturation, with a direct temporal correlation in conditions that delay maturation, including constitutional delay of growth and maturation, malnutrition, hypothyroidism, and idiopathic GH deficiency, and in conditions that accelerate maturation, including peripheral and familial precocious puberty (24-26). However, this synchrony between skeletal and pubertal maturation was not observed in healthy boys, suggesting that these processes are regulated in parallel by many intrinsic and extrinsic factors rather than by direct causation (27). Our finding of pubertal initiation in the setting of an immature bone age further supports this parallel relationship and suggests that CNS regulation of the hypothalamic-pituitary-gonadal axis appears to be independent of skeletal maturation (Fig. 3b).
The frequency of endocrinopathies in this cohort appears comparable to recent reports in survivors of medulloblastoma when factors including age at diagnosis, type of irradiation, irradiation dose, and follow-up time are considered. Hypothyroidism is one of the most common endocrinopathies associated with craniospinal irradiation, and our report of 83% is consistent with the previous reports in individuals receiving standard dose, photon irradiation therapy (~60%-80%) (4, 5, 7). GH deficiency occurred in 67% of individuals, which is lower than the >90% frequencies reported in survivors of medulloblastoma with long-term follow-up of >10 years (4, 5), but more comparable to the ~50% to 60% frequency reported in a study with a median follow-up duration of 5.8 to 7 years (7). Similarly, the frequency of precocious puberty (11%) and adrenal insufficiency (6%) are similar to recent studies with modest follow-up durations of <10 years (6, 7), but lower than studies with more extensive follow-up of >10 years (4, 5). Proton irradiation has also been shown to reduce the risk of some radiation-associated endocrinopathies compared to photon irradiation (7). In our cohort, 39% received proton irradiation and 11% received a mixed regimen of both photon and proton irradiation, which may contribute to the lower rates of GH deficiency, precocious puberty, and adrenal insufficiency in our cohort compared to cohorts with photon irradiation (4, 5). However, the rates of these 3 endocrinopathies were not significantly different between individuals who received proton versus photon therapy in a recent study (7).
Strengths of our study include the availability of long-term data on growth, pubertal development, and endocrine evaluation and treatment. Limitations include the small number of individuals at a single institution; however, the number of individuals is comparable to other studies of long-term endocrine outcomes in survivors of medulloblastoma (4, 5, 7). Because the study was retrospective, patients were not systematically evaluated for growth and pubertal development, although they were seen regularly by a pediatric endocrinologist. Exact timing of pubertal onset could not be identified and could have occurred prior to the first visit where Tanner stage 2 breasts were noted. However, our methodology would have overestimated, rather than underestimated the age at pubertal onset. In addition, the bone age determination was based on a combination of evaluations by the individual’s pediatric endocrinologist, a radiologist, and in cases of discrepant evaluations between the endocrinologist and the radiologist, a second blinded pediatric endocrinologist. In the 6 cases of discrepant evaluations, our methodology may have overestimated the bone age. Our sample was predominantly non-Hispanic White children and may not be representative of the broader survivor population. However, non-Hispanic White children on average have later puberty onset than Black children and children of Hispanic ethnicity (28, 29). In addition, data on nutritional management of undernutrition, such as type and duration of enteral tube supplementation, was limited.
Variation in pubertal timing due to cranial irradiation in children with brain tumors can serve as a model for pubertal regulation under environmental stress. Survivors of pediatric brain tumors are at risk of developing multiple complications, including short stature, obesity, adverse bone health, and negative psychosocial functioning, which may be linked to early and/or delayed pubertal timing (30). Thus, optimizing pubertal timing and limiting bone age discrepancy may be key to improving long-term outcomes in survivors of pediatric brain tumors. We have demonstrated that low absolute BMI SDS and catch-up weight gain are associated with an earlier pubertal onset and may be targetable risk factors with intensive nutritional therapy. Our findings further reinforce the relationship between nutrition and pubertal timing and provide direct evidence that catch-up growth in isolated postnatal undernutrition may influence pubertal timing.
Acknowledgments
This work was supported in part by Team Xaverian’s Pan-Mass Challenge and the Stop and Shop Family Pediatric Neuro-Oncology Outcomes Program.
Financial Support: J.Z. was supported by grant 5T32DK007699 from the National Institute of Diabetes and Digestive and Kidney Diseases.
Glossary
Abbreviations
- BMI
body mass index
- CCNU
cisplatin, cyclophosphamide, lomustine
- GH
growth hormone
- GnRH
gonadotropin-releasing hormone
- IGF-I
insulin-like growth factor 1
- IGFBP-3
insulin-like growth factor binding protein 3
- LH
luteinizing hormone
- PNET
primitive neuroectodermal tumor
- SDS
standard deviation score
- T4
thyroxine
- TSH
thyrotropin (thyroid stimulating hormone)
Additional Information
Disclosure Summary: Laurie E. Cohen is co-chair of the Tumor Related Endocrine and Neuroendocrine Disorders Special Interest Group of the Pediatric Endocrine Society, member of the Children’s Oncology Group Long-Term Follow-up Guidelines Late Effects Expert Panel, and former member of the Children’s Oncology Group Endocrine Task Force. Laurie E. Cohen received a lecture honorarium from Novo Nordisk. Jia Zhu, Henry A. Feldman, Christine Chordas, Ari J. Wassner, and Peter Manley have no conflicts of interest.
Data Availability
Restrictions apply to the availability of data generated or analyzed during this study to preserve patient confidentiality or because they were used under license. The corresponding author will on request detail the restrictions and any conditions under which access to some data may be provided.
References
- 1. Sklar CA, Antal Z, Chemaitilly W, et al. Hypothalamic-pituitary and growth disorders in survivors of childhood cancer: an endocrine society clinical practice guideline. J Clin Endocrinol Metab. 2018;103(8):2761-2784. [DOI] [PubMed] [Google Scholar]
- 2. Ogilvy-Stuart AL, Clayton PE, Shalet SM. Cranial irradiation and early puberty. J Clin Endocrinol Metab. 1994;78(6):1282-1286. [DOI] [PubMed] [Google Scholar]
- 3. Fossati P, Ricardi U, Orecchia R. Pediatric medulloblastoma: toxicity of current treatment and potential role of protontherapy. Cancer Treat Rev. 2009;35(1):79-96. [DOI] [PubMed] [Google Scholar]
- 4. Xu W, Janss A, Packer RJ, Phillips P, Goldwein J, Moshang T Jr. Endocrine outcome in children with medulloblastoma treated with 18Gy of craniospinal radiation therapy. Neuro Oncol. 2004;6(2):113-118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Uday S, Murray RD, Picton S, et al. Endocrine sequelae beyond 10 years in survivors of medulloblastoma. Clin Endocrinol (Oxf). 2015;83(5):663-670. [DOI] [PubMed] [Google Scholar]
- 6. DeWire M, Green DM, Sklar CA, et al. Pubertal development and primary ovarian insufficiency in female survivors of embryonal brain tumors following risk-adapted craniospinal irradiation and adjuvant chemotherapy. Pediatr Blood Cancer. 2015;62(2):329-334. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Eaton BR, Esiashvili N, Kim S, et al. Endocrine outcomes with proton and photon radiotherapy for standard risk medulloblastoma. Neuro Oncol. 2016;18(6):881-887. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Gleeson HK, Shalet SM. The impact of cancer therapy on the endocrine system in survivors of childhood brain tumours. Endocr Relat Cancer. 2004;11(4):589-602. [DOI] [PubMed] [Google Scholar]
- 9. Proos L, Gustafsson J. Is early puberty triggered by catch-up growth following undernutrition? Int J Environ Res Public Health. 2012;9(5):1791-1809. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Persson I, Ahlsson F, Ewald U, et al. Influence of perinatal factors on the onset of puberty in boys and girls: implications for interpretation of link with risk of long term diseases. Am J Epidemiol. 1999;150(7):747-755. [DOI] [PubMed] [Google Scholar]
- 11. Wehkalampi K, Hovi P, Dunkel L, et al. Advanced pubertal growth spurt in subjects born preterm: the Helsinki study of very low birth weight adults. J Clin Endocrinol Metab. 2011;96(2):525-533. [DOI] [PubMed] [Google Scholar]
- 12. Bakish J, Hargrave D, Tariq N, Laperriere N, Rutka JT, Bouffet E. Evaluation of dietetic intervention in children with medulloblastoma or supratentorial primitive neuroectodermal tumors. Cancer. 2003;98(5):1014-1020. [DOI] [PubMed] [Google Scholar]
- 13. Zimmermann K, Ammann RA, Kuehni CE, De Geest S, Cignacco E. Malnutrition in pediatric patients with cancer at diagnosis and throughout therapy: A multicenter cohort study. Pediatr Blood Cancer. 2013;60(4):642-649. [DOI] [PubMed] [Google Scholar]
- 14. Lee PA, Kerrigan JR. Precocious puberty. In: Pescovitz OH, Eugster EA, eds. Textbook of Pediatric Endocrinology. Philadelphia: Lippincott Williams & Wilkins; 2004:316-330. [Google Scholar]
- 15. Greulich WW, Pyle SI Radiographic Atlas of Skeletal Development of the Hand and Wrist. Stanford: Stanford University Press; 1959. [Google Scholar]
- 16. Kuczmarski RJ, Ogden CL, Guo SS, et al. 2000 CDC growth charts for the United States: methods and development. Vital Health Stat. 2002; 11(246):1-190. [PubMed] [Google Scholar]
- 17. De Leonibus C, Marcovecchio ML, Chiavaroli V, de Giorgis T, Chiarelli F, Mohn A. Timing of puberty and physical growth in obese children: a longitudinal study in boys and girls. Pediatr Obes. 2014;9(4):292-299. [DOI] [PubMed] [Google Scholar]
- 18. Tomova A, Robeva R, Kumanov P. Influence of the body weight on the onset and progression of puberty in boys. J Pediatr Endocrinol Metab. 2015;28(7-8):859-865. [DOI] [PubMed] [Google Scholar]
- 19. Burt Solorzano CM, McCartney CR. Obesity and the pubertal transition in girls and boys. Reproduction. 2010;140(3):399-410. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. de Groot CJ, van den Berg A, Ballieux BEPB, et al. Determinants of advanced bone age in childhood obesity. Horm Res Paediatr. 2017;87(4):254-263. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Brauner R, Trivin C, Zerah M, et al. Diencephalic syndrome due to hypothalamic tumor: a model of the relationship between weight and puberty onset. J Clin Endocrinol Metab. 2006;91(7):2467-2473. [DOI] [PubMed] [Google Scholar]
- 22. Tena-Sempere M. Neuroendocrinology in 2016: Neuroendocrine control of metabolism and reproduction. Nat Rev Endocrinol. 2017;13(2):67-68. [DOI] [PubMed] [Google Scholar]
- 23. Aguiar-Oliveira MH, Bartke A. Growth hormone deficiency: health and longevity. Endocr Rev. 2019;40(2):575-601. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Leschek EW, Jones J, Barnes KM, Hill SC, Cutler GB Jr. Six-year results of spironolactone and testolactone treatment of familial male-limited precocious puberty with addition of deslorelin after central puberty onset. J Clin Endocrinol Metab. 1999;84(1):175-178. [DOI] [PubMed] [Google Scholar]
- 25. Holland FJ, Kirsch SE, Selby R. Gonadotropin-independent precocious puberty (“testotoxicosis”): influence of maturational status on response to ketoconazole. J Clin Endocrinol Metab. 1987;64(2):328-333. [DOI] [PubMed] [Google Scholar]
- 26. Flor-Cisneros A, Leschek EW, Merke DP, et al. In boys with abnormal developmental tempo, maturation of the skeleton and the hypothalamic-pituitary-gonadal axis remains synchronous. J Clin Endocrinol Metab. 2004;89(1):236-241. [DOI] [PubMed] [Google Scholar]
- 27. Flor-Cisneros A, Roemmich JN, Rogol AD, Baron J. Bone age and onset of puberty in normal boys. Mol Cell Endocrinol. 2006;254-255:202-206. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Sun SS, Schubert CM, Chumlea WC, et al. National estimates of the timing of sexual maturation and racial differences among US children. Pediatrics. 2002;110(5):911-919. [DOI] [PubMed] [Google Scholar]
- 29. Wu T, Mendola P, Buck GM. Ethnic differences in the presence of secondary sex characteristics and menarche among US girls: the Third National Health and Nutrition Examination Survey, 1988-1994. Pediatrics. 2002;110(4):752-757. [DOI] [PubMed] [Google Scholar]
- 30. Zhu J, Chan YM. Adult consequences of self-limited delayed puberty. Pediatrics. 2017;139(6):e20163177. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Restrictions apply to the availability of data generated or analyzed during this study to preserve patient confidentiality or because they were used under license. The corresponding author will on request detail the restrictions and any conditions under which access to some data may be provided.