GROWTH AND GROWTH HORMONE IN CHILDREN WITH NEUROFIBROMATOSIS TYPE 1

Neurofibromatosis type 1 ( NF1), also known as von Recklinghausen disease, is an autosomal dominant, commonly inherited disease that affects one of every 3000 individuals.¹ The gene responsible for this condition has been isolated by positional cloning to chromosomal region 17q11.2. It spans over 350 kb of genomic DNA and encodes neurofibromin, a protein product of 2818 amino acids that is expressed in various tissues.² According to the National Institutes of Health Consensus Development Conference (Bethesda, Maryland, July 13-15, 1987), there are 7 key components of the disease (Table 1), at least 2 of which must be present in order to establish the diagnosis.³

Endocrine disorders have been reported in approximately 1% to 3% of all NF1 patients. Pheochromocytoma is the most common endocrinopathy in adults with NF1, occurring in approximately 1% of patients.⁴ In children with NF1, the most prevalent hormonal disorder is central precocious puberty (CPP), with a frequency of 3% compared to 0.06% in the general pediatric population.^4-6 Delayed puberty has also been described, but its exact incidence has not been reported to date. Short stature (defined as a height that is equal to or more than 2 standard deviations [≥ 2 SD] below the population mean) has long been known to be a feature of NF1, affecting approximately 13% to 24% of prepubertal patients and >40% of adults.^7,8 Although short stature is the most common growth disturbance seen in patients with NF1, tall stature has also infrequently been described as a result of growth hormone (GH) hypersecretion linked to brain tumors.^9-15 It is the primary aim of this paper to summarize current knowledge on growth disturbances and GH secretion in children with NF1.

The known molecular functions of the NF1 gene and its protein, neurofibromin, could account for the short stature phenotype of NF1–affected individuals (Figure 1). Neurofibromin is a major regulator of the Ras pathway, a key signal transduction pathway which transmits mitogenic signals to the nucleus, and is expressed in many different tissues, including the brain. It contains a central domain related to Ras-specific guanosine triphosphatase-activating proteins (Ras-GAPs). It stimulates the intrinsic activity of Ras-GTPase and is involved in control of cellular growth and differentiation through down-regulation of Ras activity.¹⁶ Mutations in the GAP-related domain of the NF1 gene lead to increased levels of activated Ras and, thus, to increased downstream mitogenic signaling.¹⁷ The NF1-conserved Drosophila homologue acts as a negative Ras regulator. Homozygous NF1 Drosophila mutants with 2 different mutations that result in lack of expression of NF1 protein are 20% to 25% smaller than flies of the parental strain, but are otherwise patterned normally. Their growth defect is rescued by expression of an hsNF1 transgene, as well as by increasing cAMP-activated protein kinase A (PKA) expression, implying that both Ras and PKA interact in a pathway that controls overall growth.¹⁸ Activated PKA has also been shown to play a critical role in stimulating proliferation of some cell types¹⁹ and may physiologically contribute to body growth. Based on the Drosophila model, it could be postulated that alterations in these pathways could result in smaller phenotypes in humans with NF1 as well.

Figure 1. Cross talks among signaling pathways linked by GAPs. Neurofibromin (NF1) and p120RasGAP (GAP) control the hydrolysis of RasGTP. RasGTP is activated by growth factors via the exchange factor Sos1 and activates a number of effectors such as RAFkinase and PI3K, which in turn activate and phosphorylate the ERK and PKB kinases. In this way, GAPs can serve as key integrators of distinct signaling pathways. Reprinted with permission Donovan S, Shannon KM, Bollag G. Biochim Biophys Acta. 2002;1602:23–45. Copyright © 2002. Elsevier. All rights reserved.

Short stature associated with NF1 usually affects the skeleton symmetrically.²⁰ The etiology of short stature in patients with NF1 does not correlate with disease severity and is multifactorial, stemming from the disease itself or its complications. These complications may include problems that interfere with normal skeletal development, such as scoliosis^21,22 or deep plexiform neurofibromas, or the use of psychostimulant medications²³ for the treatment of attention deficit disorder,²⁴ which is a frequent behavioral problem in children with NF1. Risk factors for suboptimal growth are listed in Table 2.

Riccardi²⁰ suggested that short stature was an “all-or-none” phenomenon that affected only a subset of NF1 patients. Contrary to this suggestion, the National Neurofibromatosis Foundation International Database (NFDB) cross-sectionally analyzed the distribution of heights in 569 Caucasian North American children²⁵ with NF1 (Figure 2). Of note, the mean height SD score (SDS) among their patients was lower than that of the reference population. Thirteen percent of the NF1 patients fell >2 SD below the reference population mean, compared to only 2% of controls. They concluded that the distributions of stature are shifted and unimodal among NF1 patients. The NFDB provided NF1–specific growth charts (Figure 3). From a clinical standpoint, it is important to realize that deviations from the NF1–specific standards may indicate the additive effect of a specific disease feature, such as an optic glioma.

Clementi et al²⁶ also constructed NF1–specific growth charts in a study of 528 Italian patients with comparable stature centile curves to those of the NFDB. In this study, height velocity was normal during childhood for both sexes, whereas the pubertal growth spurt was slightly reduced in boys, but not in girls. During and post-adolescence, the 50^th centile for NF1 patients overlapped with the 25^th centile for normal subjects, but the 3^rd centile was much lower in NF1 subjects than in normal subjects. There was no association of height impairment to disease severity. Carmi et al⁷ prospectively evaluated parameters of growth, puberty, and final height in 89 children with NF1. Short stature was observed in 25.5% of patients during the prepubertal period, with a significant gradual reduction of relative height for age during puberty. Forty-three percent of patients had short adult height; of these, 58% had short stature attributable to familial NF1. Short adult height was more often attributed to central nervous system (CNS) pathology when the father was the affected parent, less when both parents were affected, and rarely when neither parent was affected. There was also a four-fold higher frequency of CPP among their patients compared to that observed in the general population, but the frequency of short stature remained the same even when patients with CPP were excluded. GH deficiency (GHD) as the cause of short stature was found only after neurosurgery and irradiation in a minority of short patients.

Short stature is a cardinal feature in NF1; however, based on the stature distribution analysis of the NFDB, 24% of NF1 patients reside >2 SD above the reference population mean²⁵ (Figure 2). Carmi et al⁷ reported tall stature in 4 of 89 patients with NF1, all without evidence of abnormalities in the GH axis. GH hypersecretion presenting as gigantism has rarely been described in children with NF1, and has always been associated with the presence of optic pathway gliomas (OPG).^9-15 In some of these patients, elevated prolactin was also observed.^12-14 Treatment of the OPG with surgery, radiation, and/or chemotherapy has resulted in a reduction in growth velocity and improved basal and stimulated GH levels in all cases. Bromocriptine¹¹ and, more recently, the somatostatin analogue, octreotide,¹⁵ have also been successfully used in some tall NF1 patients. The mechanism of excess GH in these patients is not clear. There does not appear to be a direct secretory role of the tumor itself. Infiltration of the somatostatinergic pathways by the tumor leading to loss of somatostatinergic tone and, subsequently, increased GH release and loss of pulsatility, appears to be a possible mechanism in some cases.^9,10

Figure 2. Distribution of sex and age standardized stature. NF1 patient measurements are from the National NF Foundation Database. Unaffected norms are from the National Center for Health Statistics and the Fels Institute. Reprinted with permission Szudek J, Birch P, Friedman JM. J Med Genet. 2000;37:933–938. Copyright ©2000. British Medical Journal. All rights reserved.

Malignancy accounts for the development of significant morbidity and mortality in patients with NF1, including intracranial lesions, particularly suprasellar neoplasms. The first 6 years of life appear to be the period of highest risk for development of symptomatic tumors, the median age of detection being 4.2 years.²⁷ OPGs are the most frequent neoplasms, with an overall 19% incidence on routine magnetic resonance imaging (MRI) of the brain, but with a 7% symptomatic incidence.²⁸ Gliomas of the optic chiasm are reported to cause endocrinological disorders, especially CPP and GHD.

In a study by Habiby et al²⁹ of 219 children diagnosed with NF1, 3% had CPP, all associated with OPG. This association also held true in all CPP patients in the study by Carmi et al.⁷ However, in a study by Cnossen³⁰ of 122 children with NF1, the prevalence of CPP was the same as that previously reported; however, there was no evidence that OPG was a prerequisite for CPP, since only 1 of 3 children with CPP had an OPG at the time of diagnosis. Listernick et al³¹ reported CPP in 5 of 17 children with an OPG and NF1, in contrast to no cases of CPP in a group of children with OPG and no features of NF1. Virdis et al³² reviewed the records of 412 NF1 patients and also concluded that CPP is frequently—but not exclusively—associated with OPG. The above studies support an independent association of CPP and NF1 that cannot be solely attributed to OPG. A distinct feature of NF1-associated CPP is its slower rate of pubertal progression compared to CPP not associated with NF1. Whether treatment with gonadotropin-releasing hormone (GnRH) agonists is mandatory and/or efficacious in improving final height in the NF1 population remains under debate. However, there is general agreement that treatment should be offered in children manifesting signs of CPP at a young age and/or in those with a progressive decline in predicted final height.³³

Figure 3. (A) Stature centiles in males 2-18 years. (B) Stature centiles in females 2-18 years. NF1 patient measurements are from the National NF Foundation Database and are denoted by solid lines. Unaffected norms are from the National Center for Health Statistics and are denoted by dashed lines. Reprinted with permission Szudek J, Birch P, Friedman JM. J Med Genet. 2000;37:933-938. Copyright ©2000. British Medical Journal. All rights reserved.

GHD is an important complication in children with NF1, with the etiology in some patients remaining unclear. In the majority of children with NF1, GHD occurs primarily in those with an intracranial tumor who undergo intracranial surgery and cranial irradiation therapy. Indeed, in the study by Carmi et al,⁷ using clonidine or insulin-induced hypoglycemia as GH secretagogues, all children diagnosed with GHD had a history of cranial surgery or irradiation. In a study by Pierce et al³⁴ of 24 patients with OPG, half of whom had NF1, GHD was found in 15 of 18 patients who were evaluated following treatment with radiotherapy. Huguenin et al³⁵ evaluated the relationship of adult height after cranial radiation for OPG to NF1, CPP, and GHD caused by the tumor itself or its management. Cranial irradiation resulted in GHD in 100% of cases. Reduced adult height resulted when there was GHD and CPP in the presence of NF1. In a retrospective review of the Pfizer International Growth (KIGS) database,³⁶ which is a database monitoring recombinant human GH (rhGH)-treated children, a total of 102 children with NF1 were identified, 43 of whom had an intracranial tumor. Ninety-two percent and 80% of the GH-tested patients with a cranial tumor had peak GH responses below 10 and 5 µg/L, respectively. Eighty-one percent and 56% responded below 10 and 5 µg/L, respectively, in the non-tumor group. The median GH peak response to stimuli (most commonly insulin-induced hypoglycemia or arginine) was significantly lower in the tumor group compared to the non-tumor group (3.0 vs 4.6 µg/L; p<0.001). However, Cnossen et al³⁰ reported a 2.5% prevalence of GHD in children with NF1 without an intracranial mass and before surgical or radiation therapy for OPG, a frequency that is significantly higher than the 0.03% observed in the general pediatric population. An OPG was detected in 1 of 3 children with GHD, suggesting that GHD appears independently of the presence of OPG. In a study by Vassilopoulou-Sellin et al,³⁷ the incidence of GHD was investigated in 19 poorly growing children with NF1 and without other identifiable risk factors for shortness. Seventy-nine percent were diagnosed as having GHD on the basis of a peak GH response <10 µg/L after clonidine stimulation, and 42% had a peak GH level <5 µg/L, indicating a high frequency of profound GHD in this cohort. The causal mechanism of increased frequency of GHD in patients with NF1 remains to be elucidated. It is still plausible that despite the high-resolution capability of current MRI neuroimaging, cerebral abnormalities responsible for GHD are present, but not readily identifiable. Another possible explanation could be that there are abnormalities occurring at the cellular level, implicating the known molecular function of neurofibromin in signal transduction.¹⁷

Deficiencies of other anterior pituitary hormones such as thyroid-stimulating hormone (TSH) and adrenocorticotrophin (ACTH) have also been described in subjects with NF1 as a result of surgery and/or irradiation for intracranial tumors. Unrecognized hypothyroidism can account for poor growth, and unrecognized adrenal insufficiency can have potentially fatal consequences. Carmi⁷ described 3 out of 6 children with NF1 and OPG who required thyroid hormone replacement after surgery and/or cranial irradiation. In the review by Huguenin et al,³⁵ no subject had TSH or ACTH deficiency prior to irradiation. However, 80% were found to be TSH-deficient and 17% were found to be ACTH-deficient after irradiation. Gonadotropin deficiency was variable with delayed or even arrested pubertal development in 43% of the patients, and low gonadotropin responses to GnRH were found in 60% of the patients evaluated.

In a retrospective review of patients with NF1 from the KIGS database,³⁶ the outcome of 102 children treated with rhGH, at a mean dose of 0.18 mg/kg/wk with a mean duration of treatment of 2.7 years, was assessed. These included pre- and post-pubertal patients with and without intracranial tumors. The pretreatment median height SDS was –2.4 and the median height velocity was 4.2 cm/year. The median height velocity increased to 7.1 cm/year during the first year of treatment and remained above the baseline value during the next 2 years. The median height SDS increased from –2.4 to –1.9 in the first year and remained stable thereafter. There was no significant difference in the response to treatment between the tumor and the non-tumor groups, nor between those who had received radiation and those who had not. It is notable that the response to treatment was modest and less than that observed in patients with idiopathic GHD. However, the dose of rhGH given to patients with GHD was lower than that in other studies where an average dose of 0.30 mg/kg/wk was used, and it is likely that the growth velocity would have further declined if the patients had been left untreated. Vassilopoulou-Sellin et al³⁷ reported their experience with rhGH replacement therapy in a cohort, including children with NF1 and GHD without suprasellar lesions. This group of patients increased their annual growth rate (from a pre-treatment average) to 5 cm/year to 9 cm/year the first year, 8.3 cm/year the second year, and 6 cm/year during years 3 to 5 of rhGH therapy.

While therapy with rhGH has been shown to be safe, theoretical concerns remain that rhGH treatment may potentially increase an individual’s risk of developing cancer de novo or increase the risk of recurrence of primary tumors and/or the incidence of second tumors in cancer survivors. Analysis of the KIGS database revealed recurrence of a primary CNS tumor and/or appearance of a second tumor in 5 of 102 rhGH-treated subjects³⁸ with NF1. Unfortunately, MRI neuroimaging was not performed in all patients prior to the start of rhGH treatment and, hence, definitive conclusions on the timing of malignancy presentation and its relation to rhGH therapy cannot be drawn. The natural history of OPG in children with NF1, as reported in previous studies,³⁹ suggests an incidence of tumor recurrence of 11% to 14%. There are also reports of a 30% recurrence rate of OPG after 10 years in NF1 patients under the age of 20 treated with surgery.⁴⁰ The occurrence of second intracranial tumors has also been frequently reported in children with NF1 and OPG. Hochstrasser⁴¹ and Kuenzle⁴² reported second tumors in 21% and 52%, respectively, during 9 years of follow-up. Based on the results of the above studies and clinical observations, there does not appear to be an increased risk of primary tumor recurrence nor development of a second malignancy in children with NF1 treated with rhGH.⁴³ However reassuring the data may be, continuous surveillance for all NF1 individuals treated with rhGH is mandatory.⁴⁴

It is well documented that café-au-lait macule size increases during puberty.⁴⁵ It is also known that neurofibromas increase both in size and in number in pubertal patients. Superficial growth of neurofibromas can lead to underlying segmental hypertrophy, whereas deeper structure invasion of the spine and paraspinal areas can create anatomical problems, the most dangerous of which is spinal cord compression. Whether rhGH treatment can accelerate or augment the growth of these lesions with harmful sequelae remains of concern. Indeed, 13% of the NF1 patients in the KIGS database,³⁶ many of whom were pubertal, had changes in café-au-lait macules and neurofibromas. There are no reports that the increase in disease progression was accelerated secondary to rhGH, although one patient had an increase in the size of a pre-lumbar mass thought to be a neurofibroma. Cnossen et al³⁰ reported no growth of neurofibromas that could be ascribed to rhGH replacement in their patient population. The above results are reassuring; however, until larger-scale observations become available, close monitoring of the growth of neurocutaneous lesions is still warranted in rhGH-treated NF1 patients.

Short stature is a well-recognized manifestation of patients with NF1, although its etiology is not fully understood. Insight as to what represents a normal pattern of growth for individuals with NF1 has been gained through the generation of NF1-specific growth charts using information from the NFDB. It is apparent that most children with NF1 grow normally until puberty. Thereafter, their height velocity is diminished compared to their healthy peers, leading to a final height significantly below their predicted genetic target. Disease-specific features, such as scoliosis and extensive neurofibromas, can further compromise final adult height. Suboptimal growth (using the NF1-specific growth charts) is also a compelling argument to look for disease-related complications such as malignancies, the most common being OPG. These tumors are frequently the cause of CPP which if present, may further compromise final height. It is also important to be aware that the increased incidence of CPP in patients with NF1 cannot solely be attributed to the presence of OPG, as CPP may occur in this setting without any tumor. Treatment of symptomatic OPG with radiation, surgery, or chemotherapy may result in decreased final height by causing damage to the hypothalamic-pituitary region and connections thereof, resulting in one or multiple pituitary hormone deficiencies, most often GH. Recent evidence of an increased incidence of isolated GHD without identifiable risk factors in children with NF1 suggests that screening is mandatory when no plausible alternative explanation accounts for a suboptimal growth velocity. Children with NF1 have a modest, although significant response to GH treatment. Current knowledge suggests that such treatment does not influence the progression of any of the features of NF1, including the incidence of recurrence of primary or the development of secondary intracranial tumors. Hence, it appears that the use of GH is efficacious and safe in children with NF1 and GHD, although continuous vigilance is necessary. The discovery of neurofibromin, with its multiple actions on signal transduction and control of cellular growth, has shed light onto aspects of the molecular biology of the disease. Further analysis and exploration of the NF1 gene action and the effects of its mutations may help to elucidate the cellular pathways leading to the phenotypic features (including growth disorders) of neurofibromatosis.


Home	Search	Current issue	Archive	Site help
Current GGH - Vol. 24 No. 1 View table of contents Download PDF
Click here to subscribe