Hypothalamic–Pituitary and Growth Disorders in Survivors of Childhood Cancer

An Endocrine Society Clinical Practice Guideline

Charles A. Sklar; Zoltan Antal; Wassim Chemaitilly; Laurie E. Cohen; Cecilia Follin; Lillian R. Meacham; M. Hassan Murad


J Clin Endocrinol Metab. 2018;103(8):2761-2784. 

In This Article

GHD in Childhood Cancer Survivors

GHD is characterized by inadequate GH secretion from the pituitary and is defined using different diagnostic tests. GHD can result from damage to the HP area due to tumors, surgery, and/or HP axis RT.[8,45–47] Additionally, researchers have described a few cases of GHD associated with the TKI imatinib,[48,49] and the immune modulator ipilimumab can cause hypophysitis.[50]

GHD is the most common endocrine late effect in childhood cancer survivors treated with CRT.[46,51,52] The prevalence of GHD varies depending on the type of tumor and treatment and is most frequent in survivors of suprasellar tumors and after high-dose HP axis RT.[45–47] Adults with hypopituitarism on conventional hormone therapy that does not include GHT have an increased cardiovascular mortality in comparison with the general population.[53,54]

Clinical Manifestations

GHD that develops in childhood usually affects linear growth.[8,55,56] GHD in the cancer population has similar symptoms as we see in the noncancer population and may be associated (particularly in adults) with reduced lean body mass and increased fat mass, an adverse lipid profile, increased cardiovascular morbidity, impaired bone mineral density, impaired quality of life, and psychosocial problems.[57–61] In a large cohort of 695 survivors of childhood cancer enrolled in the St. Jude Lifetime Cohort study, survivors with untreated GHD were more likely to have an increased weight-to-height ratio, decreased lean muscle mass, low energy expenditure, muscle weakness, and poor exercise tolerance compared with individuals without GHD.[47]

Diagnosis of GHD in Childhood Cancer Survivors

   2.1 We recommend lifelong periodic clinical assessment for GHD in survivors treated for tumors in the region of the HP axis and in those exposed to HP axis RT ≥18 Gy (e.g., various brain tumors, nasopharyngeal carcinoma, acute lymphoblastic leukemia, lymphoma). (1∣⊕⊕⊕O)

    Technical remark: The consensus of the writing committee is to assess height in children every 6 to 12 months.

   2.2 We recommend against relying solely on serum IGF-I levels in childhood cancer survivors exposed to HP axis radiotherapy to make the diagnosis of GHD. (1∣⊕⊕OO)

   2.3 We advise using the same provocative testing to diagnose GHD in childhood cancer survivors as are used for diagnosing GHD in the noncancer population (Table 2). (Ungraded Good Practice Statement)

   2.4 We recommend against the use of GHRH alone or in combination with arginine in childhood cancer survivors to diagnose GHD after HP axis radiation. (1∣⊕⊕OO)

   2.5 We suggest against using spontaneous GH secretion (e.g., 12-hour overnight sampling) as a diagnostic test in determining GHD in childhood cancer survivors. (2∣⊕OOO)

   2.6 We recommend against formal testing to establish a diagnosis of GHD in childhood cancer survivors with three other confirmed anterior pituitary hormone deficits (Table 2). (1∣⊕⊕OO)

   2.7 We recommend retesting adult cancer survivors exposed to HP axis RT and with a diagnosis of isolated GHD in childhood (Table 2). (1∣⊕⊕OO)

Evidence. HP axis radiation is a potent cause of GHD and the risk is directly related to the total dose delivered, the dose per fraction, and the time interval postirradiation.[62–64] HP axis RT in children frequently causes abnormal HP function later in life.[65] HP axis RT initially affects the hypothalamus, which is more sensitive to irradiation than is the anterior pituitary, based on responses to anterior pituitary–releasing hormone stimulation in patients with anterior pituitary hormone deficiencies.[65,66] GHD is usually the first established endocrine sequela of HP axis RT.[51,63] The prevalence varies depending on the population studied, follow-up time, type of stimulation test used, and peak GH cut-off levels.[67]

GHD is a frequent consequence in childhood cancer survivors treated for tumors in the region of the HP[45,68] and in brain tumor and nasopharyngeal carcinoma survivors exposed to HP axis radiotherapy ≥30 Gy.[47,62,69] Furthermore, GHD is also reported in some acute lymphoblastic leukemia and lymphoma survivors exposed to 18 to 24 Gy to the HP axis.[23,70–72] In the even lower doses used for hematopoietic stem cell transplantation, GHD may occur after a single TBI dose of 10 Gy or fractionated doses of 12 Gy;[73] however, with repeated assessments over time, there can be recovery.[40]

In children, auxologic data collected every 6 to 12 months should be considered as the initial screen for GHD. Clinicians should further investigate for GHD when there is either linear growth deceleration or no linear growth acceleration during puberty and when they have ruled out other potential etiologies of growth failure (e.g., undernutrition, spinal compromise, hypothyroidism, hypogonadism). Age-adjusted IGF-I levels measured in an accredited laboratory may be useful in screening for severe GHD but have limited utility when using a cut-off of −2 SD.[52,59,74,75] For example, Weinzimer et al.[74] found a sensitivity of 73% for IGF-I for the prediction of GHD in children with brain tumors. Additionally, clinicians need to interpret IGF-I levels in the context of sex steroid exposure (precocious puberty and hypogonadism). In adults, symptoms of GHD are nonspecific. GH testing should be considered in at-risk individuals with fatigue, increased abdominal fat mass, weight gain, low energy level, or hyperlipidemia. As for children, IGF-I levels may be useful in screening for severe GHD.

Clinicians who suspect GHD in childhood cancer survivors should perform provocative testing unless there are three other pituitary hormone deficiencies.[76] In the general population, GHD is established via stimulation testing using the ITT, glucagon, arginine, levodopa, clonidine, or GHRH-arginine (if available).[77–79] Clinicians should not administer GHRH alone when the damage is primarily hypothalamic, as after radiation therapy, because it may give false-negative results (i.e., normal GH responses despite true GHD). Likewise, the GHRH-arginine test can give a falsely normal GH response.[71,78,80] Both ITT and glucagon testing allow evaluation of the complete hypothalamic–somatotroph axis.[81] Based on a recent meta-analysis, the ability to diagnose GHD by different provocative tests after CRT is similar to the general population, with the ITT being most reliable; however, data are limited.[82]

The same GH cut-off levels to stimuli are used in childhood cancer survivors as in the general population. Comparing studies to assess prevalence or incidence of GHD is problematic due to the use of different GH antibodies, GH standards, and GH assays,[83] as well as the poor reproducibility of these tests.[84] Data in noncancer pediatric populations suggest that clinicians often misdiagnose children as having GHD, especially when using peak GH values of 5 to 10 μg/L.[83]

Older studies suggested that some children have GH neurosecretory dysfunction after cranial radiation, especially after low doses, where there is subnormal spontaneous GH secretion despite normal GH responses to stimulation testing.[85] However, Darzy et al.[86] demonstrated normal physiologic GH secretion in adults who received cranial radiation in childhood, suggesting that this particular entity (radiation-induced GH neurosecretory dysfunction) either does not exist or is very rare. Additionally, due to the poor reproducibility of 12-hour overnight GH sampling and the overlap of responses in normal children and nonchildhood cancer survivors with GHD,[76] as well as the impracticality of this test in clinical practice, we suggest against using spontaneous GH sampling, for example, 12-hour overnight sampling, as a diagnostic test in determining GHD in the childhood cancer survivors. In adults, the peak GH cut-off levels to diagnose GHD range from 3 to 5 μg/L for ITT and 3 μg/L for glucagon.[52,87,88] Obesity, sex-steroid deficiency, and hypothyroidism can blunt GH secretion and yield a false-positive result (e.g., falsely low GH levels);[83] for example, in an obese individual, the cut-off of 1 μg/L is used for glucagon.[89,90]

Treatment of GHD in Childhood Cancer Survivors

   2.8 We recommend offering GHT in childhood cancer survivors with confirmed GHD based on the safety and efficacy demonstrated in that population. (1∣⊕⊕OO)

   2.9 In childhood cancer survivors, we suggest waiting until the patient has been 1 year disease-free following completion of therapy for malignant disease, before initiating GHT. (2∣⊕OOO)

   2.10 In childhood cancer survivors who have chronic stable disease and thus may not ever be "disease-free" (particularly survivors treated for optic pathway tumors), we advise discussing the appropriateness of GHT and its timing with their oncologist. (Ungraded Good Practice Statement)

   2.11 We advise treating GH-deficient childhood cancer survivors with similar GHT regimens as are appropriate for individuals with GHD from the noncancer population (Table 2). (Ungraded Good Practice Statement)

Evidence. GHT is approved for both children and adults with confirmed GHD. GH dosing guidelines in the transition period after growth cessation are not well established. However, as GH secretion and IGF-I levels peak in puberty and decline overtime thereafter,[91,92] the effective GH dose needs to be higher in the transition period after growth cessation than in adulthood.

Childhood cancer survivors with GHD who receive GHT have a significant gain in height as compared with those who are not treated [see accompanying meta-analysis[82]]. However, those childhood cancer survivors with GHD who were also treated with CSI or TBI may have impaired spinal growth and not achieve target height. Higher spine radiation doses and radiation at a younger age are associated with impaired spinal growth (see section 1. "Short Stature/Impaired Linear Growth in Childhood Cancer Survivors"). GHT results in either an improvement or no difference in the risk of T2DM, dyslipidemia, and metabolic syndrome [see accompanying meta-analysis[82]]. The discrepant results between studies may be due to metabolic improvements being offset by the effect of GH on increasing insulin resistance. Likewise, quality of life after GHT is either improved or unchanged.[58,60,93]

Concerns have been raised regarding the long-term safety of GHT in childhood cancer survivors, as GH and the target hormone, IGF-I, have in vitro proliferative effects, and IGF-I also has proangiogenic and antiapoptotic properties.[94] Available data on the safety of GH in childhood cancer survivors are based on observational studies with selection bias and a lack of randomized placebo-controlled studies. Childhood cancer survivors have an increased risk of developing meningioma and glioma due to radiation exposure;[95,96] they also are at risk for GHD and will be potential candidates for GHT.[67] Previous data on GH-treated childhood cancer survivors indicated that GHT might potentially induce a small increase in the relative risk of developing second neoplasms compared with survivors not receiving GHT,[97,98] with research indicating that meningiomas are the most common second neoplasm.[97] However, the elevated risk decreased over time.[99] Although the reason for this decrease is unknown, it has been speculated that GH-treated individuals may have been subjected to earlier and increased surveillance.[100] Recent studies have shown no significant association between GHT and the development of a second neoplasm of the central nervous system (CNS) in childhood cancer survivors.[100–102] In the systematic review and meta-analysis conducted for this guideline, there was no statistically significant difference in the occurrence of secondary tumors in survivors treated with GH compared with those not treated (OR, 1.34; 95% CI, 0.92 to 1.96). Similarly, studies show no significant change in the risk of tumor recurrence in survivors treated with GH, compared with those not treated (overall OR, 0.57; 95% CI, 0.31 to 1.02) [see accompanying meta-analysis[82]]. At a recent workshop, the Growth Hormone Research Society concluded that there are no indications of an increased risk of recurrence of primary cancers after GHT in children, and the association between GHT and risk of second tumors is insufficient to make recommendations against GHT.[103]

However, few data are available to provide recommendations when to initiate GHT after cancer treatment.[103] Traditionally, clinicians start GHT for survivors of malignant tumors at least 1 year after a childhood cancer survivor is disease-free; thus, the safety of GHT prior to that time is not clear. An exception is craniopharyngiomas (which are considered benign tumors); in these cases, GHT has been safely initiated earlier (as early as 0.7 year from diagnosis).[104] Additionally, there are patients who may have stable disease, rather than being disease-free. This is often the case in subjects with optic pathway gliomas (low-grade tumors frequently found in association with neurofibromatosis type 1). In these cases, disease can remain stable for prolonged periods, despite radiation and/or chemotherapy treatments.[105] For these patients, clinicians should discuss whether to initiate GHT and its timing with the patient's oncologist.

It is important to note that GHT in children with GHD who had been treated with CSI and TBI may result in improvement in leg length but not spinal height (see section 1. "Short Stature/Impaired Linear Growth in Childhood Cancer Survivors"). Additionally, GHT may exacerbate an existing scoliosis, a condition not uncommon following either spinal surgery and/or spinal RT. GHT in the noncancer population is associated with an increased incidence of slipped capital femoral epiphysis (SCFE).[76] GHT in the noncancer population also commonly results in a decrease in insulin sensitivity and a compensatory increase in insulin secretion.[76] As childhood cancer survivors are at increased risk for both SCFE[106] and metabolic syndrome[107,108] (particularly after TBI), the potential risk for SCFE and T2DM should be a factor when clinicians consider GHT in survivors. Therefore, when considering GHT, clinicians need to carefully weigh the potential risks against the potential benefits. Similar to the Pediatric Endocrine Society's recommendations for the noncancer population,[76] we recommend that clinicians monitor serum IGF-I concentrations in patients on GHT and ensure they are kept within the normal range for sex, age, and pubertal status.