7.2.4 Investigation of the slowly growing child

Reduced height velocity for age and stage of puberty implies slow growth. Although this occurs independently from actual height, children identified for investigation tend to be the ones who are slowly growing, as well as short. The majority of short slowly growing children do not have a recognized endocrinopathy. The commonest growth disorders are those grouped under the heading ‘idiopathic’, which includes constitutional delay in growth and puberty, a disorder of the tempo of maturation, and familial/genetic short stature. These children present with short stature, an unremarkable phenotype, and a variable extent of growth failure. The challenge to the clinician is to differentiate these children from those who may have a defined nonendocrine pathology (for example, chronic systemic disease (Box 7.2.4.1), bone disorder, or psychosocial problem) and those who may have an abnormality within the growth hormone axis.

The decision to undertake a particular test will be guided by the presentation, history, clinical signs, and growth performance (Fig. 7.2.4.1). The presentation of a child with slow growth implies that preceding growth data are available. This may not always be so: a parent/carer may have become aware that their child’s growth is failing against the growth rate of their peer group. Children presenting with a history of growth failure in conjunction with symptoms or signs suggestive of intracranial pathology (for example, morning headaches, visual field defect) require urgent investigation. Under all other circumstances it is helpful to have an objective measure of growth performance—at least two accurate height measurements taken 6 months apart and ideally over a longer period. Growth is nonlinear, with marked variation in growth both within the year and from year to year. Poor growth performance may reflect a genuine problem or possibly measurement error or a period of relative, but normal reduction in growth rate. However, the persistence of a poor growth rate (below the 25th centile on a velocity chart over 12 months) indicates the need for investigation.

Growth assessment should also be directed to the evaluation of weight change. Unsatisfactory gains in both height and weight are more likely to indicate a systemic, rather than an endocrine cause. It is therefore important to define a child’s ‘weight for height’ using parameters such as body mass index (BMI) or percentage of ideal body weight for height. The former can be assessed against a centile chart (1), whereas the latter may be calculated as follows: (observed weight (kg) ÷ expected weight for height age (kg)) × 100%, where height age is defined as the chronological age at which the child’s height (cm) would fall on the 50th centile. Poor nutritional status would be indicated if BMI falls below the second centile or weight for height was less than 85%.

Investigations will be considered in terms of:

Disorders in any body system may be accompanied by growth failure. The primary condition has usually been diagnosed, but this is not always the case. In particular, occult gastrointestinal, renal, and chronic infective disorders may present initially with poor growth and/or weight gain (Box 7.2.4.1). It is also possible for a systemic disorder to develop in a condition already characterized by growth failure. For instance, inflammatory bowel disease or coeliac disease may develop in an adolescent with Turner’s syndrome.

Screening tests can be helpful in the diagnosis of these disorders. Anaemia with microcytic indices suggests iron deficiency, whereas anaemia with macrocytosis would indicate vitamin B₁₂ or folic acid deficiency, which can be secondary to malabsorption. Fat globules in stool samples indicate malabsorption, while a search for ova, cysts, or parasites may reveal conditions such as chronic giardiasis. Measurement of antiendomyseal antibody titres are particularly helpful in the diagnosis of coeliac disease. Renal disease can be identified by measurement of urea, creatinine, electrolytes, acid–base status, and urine analysis for the presence of infection, glomerular, or tubular dysfunction.

Skeletal dysplasia may present with growth failure. Many skeletal dysplasias have a characteristic appearance that immediately indicates a diagnosis. However, the phenotype of the milder bone dysplasias (for example, hypochondroplasia, MIM 146000) may be unremarkable, particularly in the younger child. Thorough auxological assessment using sitting height versus leg length measurements (the latter derived indirectly from (standing height minus sitting height)) plotted on centile charts may reveal skeletal disproportion. If this is present, a skeletal survey can be undertaken, which must include views of:

Interpretation of these films may require specialist expertise not routinely available in the assessment centre.

Disproportion is not always indicative of bone disorders. The lower limbs are longer in relation to the axial skeleton in disorders of physical maturation, such as constitutional delay in puberty or hypogonadotrophic hypogonadism.

Recognition of the gene mutation responsible for some bone dysplasias also provides an opportunity to test directly for a condition. This can be done for hypochondroplasia, one of the commonest bone disorders presenting to growth clinics. Activating mutations within the fibroblast growth factor receptor 3 gene (FGFR3, MIM 134934) on the short arm of chromosome 4 lead to both achondroplasia (MIM 100800) and hypochondroplasia (2). In addition the short stature homoeobox gene (SHOX) has been implicated in dyschondrosteosis (MIM 127300), a mesomelic skeletal dysplasia with Madelung deformity of the forearm and a SHOX (MIM 312865) deletion can be identified using fluorescence in situ hybridization (FISH) techniques.

Chromosome analysis is an important investigation. This is particularly true in the investigation of the short girl. It provides a definitive test for Turner’s syndrome, which has an approximate incidence of 1 in 2500 female live births. This condition has a very variable phenotype; some girls exhibit few of the classical stigmata. The most consistent features are the short stature and growth failure from early- to mid-childhood, a high-arched palate, hyperconvex shape to the finger and toe-nails, and in the adolescent years amenorrhoea. Turner’s syndrome is associated with a 45, X karyotype, but also with structural abnormalities of the X chromosome, an isochromosome or a ring X (3). In the structural abnormalities, only specific segments of the X chromosome may be lost, conferring short stature, but not gonadal dysgenesis. It is therefore possible to inherit Turner’s syndrome, a fact that should not be overlooked in the assessment of a short mother and daughter.

Chromosome analysis in a short child, who may have dysmorphic features and learning difficulties, may also reveal abnormality such as an unbalanced translocation. This should be followed by analysis of parental chromosomes for evidence of a balanced translocation, which would then have implications for genetic counselling. In addition, the dysmorphic features may suggest a specific genetic disorder (for example, Down’s syndrome—trisomy 21). Genetic conditions can be associated with severe (for example, de Lange’s (MIM 122470) or Seckel’s (MIM 210600) syndromes, premature ageing conditions, Russell–Silver syndrome (MIM 180860) of severe intrauterine growth retardation) or moderate short stature (for example, Noonan’s, (MIM 163950) Williams’ (MIM 194050), or Aarskog’s syndromes (MIM 100050)). Once the genetic syndrome has been recognized, specific molecular genetic tests can confirm the diagnosis in some of these conditions.

Other specific conditions may require confirmation by DNA analysis. One example is Prader–Willi syndrome (MIM 176270), which may present in early childhood, primarily with short stature. It is important to confirm a history of hypotonia and feeding difficulties in early life. There is also likely to be evidence of developmental delay, behaviour disorder, evolving obesity, characteristic facial appearance, and possibly hypogonadism (4). The condition is caused by abnormalities within the imprinted region on the proximal long arm of chromosome 15, with absence of normally active paternal genes. Seventy-five per cent of cases will have a microdeletion of paternal chromosome 15q11–q13, whereas most of the remainder have maternal uniparental disomy for chromosome 15. A minority have a methylation imprinting defect, which inactivates genes on paternal chromosome 15. These causes can all be confirmed by analysis with methylation-sensitive DNA probes.

The presentation of metabolic and storage disorders is usually related to the primary metabolic abnormality. However, these disorders are typically associated with short stature and poor growth, and may present to a growth clinic. Among these are Morquio’s syndrome (mucopolysaccharidosis (MPS) type IVA, MIM 253000; type IVB, MIM 253010), mucolipidosis type III (MIM 309900), juvenile Hunters’ syndrome (MPS type II (MIM 309900)), and connective tissue disorders, such as osteogenesis imperfecta type IV (MIM 166220). Specific metabolic tests may be required to confirm the diagnosis (for example, urine screen for MPS and oligosaccharides, white cell enzyme assays).

Bone age estimation is commonly performed on children with disordered growth after 1 year of age. It is a marker of physical maturation, and can be used to estimate growth potential with prediction of adult height (for further details see Chapter 7.1.1). Diagnostic information is not obtained from the bone age; however, hypothyroidism, growth hormone deficiency, and hypopituitarism can be associated with marked bone age retardation (more than 3 years). In addition, a diagnosis of constitutional delay in growth and puberty would be supported by bone age retardation usually more than 2 years in the absence of any endocrine deficit. Accurate bone age assessment may be difficult or impossible in certain circumstances, such as skeletal dysplasias or in children exposed to long-term immunosuppressive steroid treatment. The main use of bone age estimations is to monitor growth potential over time, particularly if a treatment to modify growth or puberty or both is introduced.

A test of basal thyroid function (total or free thyroxine and thyroid-stimulating hormone (TSH)) is frequently undertaken early in the evaluation of a slowly growing child. Clinical signs of hypothyroidism may be subtle, and the test is easily performed to provide definitive evidence of primary or secondary hypothyroidism. Plasma calcium, phosphate, and alkaline phosphatase are further routine screening tests, and can indicate conditions such as pseudohypoparathyroidism (low calcium, raised phosphate), vitamin D deficiency (raised alkaline phosphatase), or hypophosphataemic (low phosphate) rickets.

However, the most common scenario in the evaluation of growth failure is that the clinician needs an assessment of the growth hormone axis and usually other anterior pituitary function.

The standard method of assessing the integrity of the growth hormone axis is to perform a growth hormone stimulation test (5). Many agents have been used as stimuli (Box 7.2.4.2), acting either through hypothalamic pathways or directly on the somatotroph or both to increase circulating levels of growth hormone. These tests therefore require intravenous cannulation for sample collection before and after administration of the stimulus (usually up to 90–120 min). The insulin stress test has been considered the gold standard for growth hormone stimulation tests. Here, the stimulus is hypoglycaemia. The test has the advantage that the adequacy of cortisol secretion can be assessed at the same time as growth hormone. However, this test has been associated with both morbidity and mortality, principally related to inappropriate correction of symptomatic hypoglycaemia. It is therefore recommended that this test is only carried out in units with considerable experience of paediatric endocrine investigation.

There are many limitations to the interpretation of growth hormone stimulation tests (6–80). Growth hormone secretion falls along a continuum from complete deficiency (growth hormone gene deletion) to growth hormone insufficiency to normality through to hypersecretion (in pituitary gigantism/acromegaly). Defining the point at which mild growth hormone insufficiency becomes normality is not possible. A pragmatic approach to this problem has been to define inadequate growth hormone secretion as a peak growth hormone concentration during a stimulation test not exceeding 7 or 10 µg/l. However, the concentration of growth hormone measured in a blood sample is dependent on the growth hormone assay used; a threefold variation in growth hormone concentration from the same sample has been reported between assays. This is dependent on factors such as the anti-growth hormone antibody used (monoclonal versus polyclonal) and which growth hormone isoforms are being detected, the assay matrix, and the growth hormone standard. It is recommended therefore that each centre regularly undertaking growth hormone testing should define, with their chosen assay, a cut-off level for the diagnosis of growth hormone insufficiency. It has been common practice to undertake more than one growth hormone stimulation test, either separately or sequentially, in an attempt to reduce the number of false positive results. A peak greater than 7 or 10 µg/l in either test would lead to the conclusion that growth hormone secretion was normal. This approach can be modified now that other tests of the integrity of the growth hormone axis are commonly available (see later sections).

The growth hormone response to stimulation tests can also be modified by physiological factors: normal growth hormone responses in the early months of life are higher than in later childhood, while growth hormone levels do increase modestly through childhood and more dramatically in puberty. The use of a single cut-off level may be pragmatic, but it does not accommodate this variation. Nevertheless, the most common time for assessing growth hormone status is in mid-childhood, when growth hormone secretion is relatively stable. The oestrogen-induced rise in growth hormone during puberty has been the impetus for using sex-steroid priming of growth hormone tests in children approaching or in puberty. Priming is usually recommended for pre- and peri-pubertal children with bone age greater than 8 years in girls or 10 years in boys. Sex steroid administration for 3 days prior to the growth hormone test reduces the number of false positive tests (that is, indicating wrongly the presence of growth hormone insufficiency). Other physiological variables influencing growth hormone secretion are body composition and nutritional status. Obesity is associated with reduced growth hormone secretion, but it also occurs in growth hormone deficiency. Poor calorie intake can lead to a state of growth hormone resistance with elevated growth hormone levels. All these factors create difficulties for the interpretation of growth hormone test results. The most important drawback, however, is the relative scarcity of data on growth hormone responses in normal children (9). When such data have been reported, there are, in fact, many normal children who would fulfil the criteria for growth hormone insufficiency. It is important therefore to be cognizant of the limitations of growth hormone stimulation tests.

The difficulties in interpretation of growth hormone stimulation tests have led to the evaluation of other means of assessing growth hormone secretion. Growth hormone is released from the pituitary in a pulsatile manner, under the control of the hypothalamic peptides growth hormone-releasing hormone (GHRH), ghrelin, and somatostatin (SMS), which are, in turn, controlled by multiple cortical neuronal inputs. Measurement of diurnal output of growth hormone can be made by multiple sampling (every 10–20 min) over a 12- or 24-h period to generate a growth hormone profile, from which various parameters can be derived—mean growth hormone peak amplitude, growth hormone area-under-the-curve, maximum growth hormone concentration. Full interpretation requires normative data, which are not readily available. It is proposed that some children with growth failure may have growth hormone neurosecretory dysfunction, such that peak growth hormone levels on stimulation testing are normal, but growth hormone release under physiological circumstances is reduced (10). A growth hormone profile is the only way to make this diagnosis. However, this is a demanding investigation and not routinely carried out.

Another approach to physiological assessment of growth hormone output has been the measurement of the minute amount of intact growth hormone present in urine (11). This growth hormone can be detected by increasing the sensitivity of serum growth hormone assays and in many methodologies by first dialysing the urine. The test has the advantage that it is physiological, noninvasive, and easily repeated. A good correlation between urinary growth hormone and serum growth hormone over the preceding 12 or 24 h has been found. However, the performance of urinary growth hormone tests in the diagnosis of growth hormone deficiency is no better than growth hormone stimulation tests, with the best sensitivity found in those with severe growth hormone deficiency.

GHRH has also been used as a potent stimulus to growth hormone release. When combined with an agent that inhibits somatostatin release (for example, arginine or pyridostigmine), GHRH has been shown consistently to increase growth hormone levels well above 10 µg/l in normal children (9). This would suggest that such a test would be excellent for the diagnosis of growth hormone insufficiency. However, a normal response in this test could occur if the growth hormone insufficiency resulted from hypothalamic, rather than pituitary dysfunction. This is, in fact, a common occurrence in isolated growth hormone deficiency.

Although growth hormone stimulation tests have limitations in the diagnosis of growth hormone deficiency, they do identify accurately those with severe growth hormone deficiency (peak growth hormone level <3 µg/l). Therefore, such information should be available in any slowly growing child on whom a decision to investigate pituitary function has been made. However, it should also be combined with one or more additional measures of the integrity of the growth hormone axis, in particular those peptides regulated by growth hormone, namely insulin-like growth factor 1 (IGF-1) and its principal serum binding proteins IFG binding protein 3 (IGFBP-3) and the acid-labile subunit (ALS).

IGF-1 is critical to both pre- and postnatal growth. Hepatic synthesis of IGF-1 is mainly regulated by growth hormone, such that severe growth hormone deficiency is associated with low circulating levels of IGF-1. Its measurement has now become commonplace in the assessment of the growth hormone axis. IGF-1 is transported through the circulation bound to a 38–42 kDa binding protein, IGFBP-3. This binary complex then forms a ternary complex with a third protein, the 150 kDa ALS. Like IGF-1, IGFBP-3, and ALS are synthesized in the liver, and the process is also, in part, regulated by growth hormone. All three peptides are therefore potential markers of the integrity of the growth hormone axis. However, these peptides can also be affected by nutritional status, liver and renal disease, hypothyroidism, diabetes mellitus, and sex steroids.

There is minimal diurnal variation in these peptides and, therefore, a single blood sample will suffice for measurement. This also facilitates the establishment of comprehensive normal ranges based on age, sex, and pubertal status (12, 13). Assay methodology has been important in serum IGF-1 measurement: it must be separated from the binding proteins before immunoassay. This can be achieved by acidic separation and ethanol precipitation, or more recently by acidic separation then use of insulin-like growth factor-II (IGF-2) to block further binding protein association. The latter eliminates the need for ethanol precipitation, when IGF-1 may be lost with the binding protein, and simplifies the assay procedure. Assays to measure ‘free IGF-1’, the small fraction of IGF-1 (approximately 1%) not bound to IGF binding proteins, are now also available. It might be argued that these assays, in fact, measure IGF-1 that is easily dissociated from binding protein. The performance of free IGF-1 in the diagnosis of growth hormone deficiency does not appear to be greater than that reported for total IGF-1. IGFBP-3 and ALS are present in relatively large amounts in the circulation—mg/l, rather than µg/l for IGF-1. Their measurement by immunoassay is therefore straightforward. To date, most experience using these two peptides as diagnostic tools is limited to measurement of IGFBP-3.

Initial studies indicated that the performance of IGF-1 and IGFBP-3 (sensitivity is number of true positives and specificity is number of true negatives) in the diagnosis of growth hormone deficiency was excellent. However, other reports have confirmed the high specificity, but shown a low sensitivity, particularly for IGFBP-3 (14). This may reflect the fact that, as IGFBP-3 is the most abundant binding protein in the circulation, its concentration reflects the combined concentration of IGF-1 and IGF-2. Although IGF-2 is reduced in growth hormone deficiency, this is a relatively minor effect and its concentration exceeds considerably that of IGF-1. The lowered concentration of IGFBP-3 in growth hormone deficiency then reflects not only very low IGF-1 levels, but also IGF-2 levels. Additionally, IGFBP-3 can be degraded by protease action, generating fragments still detected on immunoassay, but unable to bind IGF-1. In conditions where protease activity may be elevated, a falsely high level of IGFBP-3 would be measured. IGFBP-3 measurement is particularly useful in the diagnosis of GHD in the very young child when IGF-1 levels do not discriminate GHD from normality. The measurement of ALS in the diagnosis of growth hormone deficiency has the potential advantage that, unlike IGFBP-3, it is present at a concentration that exceeds that of the ternary complex. There is therefore free ALS present in the circulation. There is also evidence that growth hormone directly controls hepatic ALS synthesis, while IGFBP-3 may be generated not only by growth hormone, but also IGF-1. ALS levels are, however, reduced to a greater extent by severe GHD than IGFBP-3 levels (15).

Despite these reservations, measurement of IGF-1 and its binding proteins do provide a measure of growth hormone action. Combining their use with a test of growth hormone secretion is now a common approach to the evaluation of growth failure (7). In view of the high specificity of IGF-1 and IGFBP-3, low levels are highly indicative of growth hormone deficiency, while normal values would not necessarily exclude the diagnosis. This infers that peak growth hormone levels during stimulation tests may not always correlate with serum IGF-1. This is particularly relevant to those whose peak growth hormone level falls within the partially deficient range greater than 5 µg/l, but less than 7 or 10 µg/l, when the IGF-1 and IGFBP-3 levels may be normal. A second growth hormone test will be helpful. If low, a diagnosis of growth hormone insufficiency would be appropriate, whereas if the growth hormone test was normal, it could be assumed that the growth hormone–IGF axis was normal. An approach to the interpretation of growth hormone and insulin-like growth factor testing is shown in Fig. 7.2.4.2.

Rare causes of growth failure include congenital and acquired growth hormone insensitivity. The former has been termed the Laron syndrome, characterized by extreme short stature, a phenotype similar to that of severe growth hormone deficiency, elevated growth hormone levels, and very low concentrations of IGF-1, IGFBP-3, and ALS (16). The condition usually arises from a defect in the growth hormone receptor, impairing its expression or action, but also may be caused by defective intracellular growth hormone signalling (e.g. mutations in the signal transducer and activation of signalling molecule STAT-5b). Acquired growth hormone insensitivity occurs in situations, such as malnutrition, chemotherapy treatment, and liver disease.

Growth hormone insensitivity can be confirmed with an ‘IGF generation test’: recombinant growth hormone (0.03 mg/kg per day) is given by subcutaneous injection for 4 days with care taken that nutritional input is adequate. IGF-1 and IGFBP-3 are measured at the start and on day 5. A rise in serum IGF-1 level above 20 µg/l and in IGFBP-3 above 0.4 mg/l would exclude growth hormone insensitivity. Experience with this test is relatively limited. In parallel with all other tests of the growth hormone–IGF axis, it is very difficult to apply cut-off values to what is a continuous variable, namely the degree of growth hormone insensitivity. Nevertheless, it is possible to use this approach to define those with severe growth hormone insensitivity.

One further very rare condition is that of growth hormone bio-inactivity. Here, a mutation within the GH gene introduces a subtle sequence change that reduces or abolishes growth hormone biological activity, but retains its immunological properties, such that it is normally detected on immunoassay (17). This condition can be diagnosed if patient sera are tested in vitro in a growth hormone bioassay. This could utilize a cell line, naturally responsive to growth hormone, such as the IM9 B-lymphocyte, in which growth hormone would normally induce phosphorylation of signalling molecules. Alternatively, a cell line transfected with wild-type growth hormone receptors that would proliferate in response to biologically active growth hormone but not to bio-inactive growth hormone, could be used.

The symptoms and signs of hypothalamic–pituitary disease may be subtle, with the major manifestation being growth failure. If a decision is made to undertake investigation in a slowly growing child, then full assessment of anterior pituitary function should be carried out.

Central nervous system imaging is a mandatory first step, if clinical evaluation of the child with growth failure has suggested that a lesion may be present—persistent early morning headaches, visual disturbance and/or field defect, polyuria and polydipsia, or overt hypopituitarism without explanation. This can be done with high-resolution CT or MRI. Calcification, as detected by CT scan, is highly suggestive of craniopharyngioma, as are lesions with both solid and cystic components. Germinomas may arise in the hypothalamic region, so measurement of tumour markers, such as α-fetoprotein and β-HCG, should be undertaken.

Structural abnormalities within the midline may be associated with growth hormone deficiency and hypopituitarism. The most common diagnosis in this category is septo-optic dysplasia. It is frequently, but not invariably, associated with pituitary dysfunction. If present, however, unusual combinations of dysfunction may occur, e.g. growth hormone, TSH, ACTH, and antidiuretic hormone deficiency, but precocious puberty. In a child with nystagmus, visual problems, poor growth, and possibly other endocrinopathy, MR scanning may be very helpful in delineating the radiological features of septo-optic dysplasia. These are also variable, but may include optic nerve, chiasmatic, and infundibular hypoplasia, and absence of the septum pellucidum. Other conditions, where midline abnormalities associated with growth failure and growth hormone deficiency may be found include cleft lip and palate, solitary central incisor, and holoprosencephaly.

Neuroimaging has also become a useful investigation in the assessment of isolated growth hormone deficiency and hypopituitarism (18, 19). A reduced anterior pituitary height (more than 2 SD below the age-matched population mean), an attenuated or interrupted hypothalamic–pituitary stalk, and/or an ectopically positioned posterior pituitary, defined as a bright spot on MRI, are all associated with pituitary dysfunction (Fig. 7.2.4.3). In a child with isolated growth hormone deficiency, these findings would confirm that the diagnosis is genuine and would also alert the clinician to the possibility of evolving hypopituitarism. It is suggested that all those in whom a diagnosis of growth hormone insufficiency has been made should have imaging of hypothalamic–pituitary structures. There are, however, cases of isolated growth hormone deficiency, where MR abnormalities are not found. This could be a clue that an isolated case of familial growth hormone deficiency, caused by a mutation within the GH gene, may be present.

A search for a specific gene mutation causing or contributing to growth failure and/or pituitary dysfunction (20) should be reserved for those cases where there is a high index of suspicion. Most of the tests are based on polymerase chain reaction (PCR) methodologies with identification of mutations by direct sequencing or altered DNA mobility on gel electrophoresis. DNA for analysis can be readily obtained from any body fluid. For growth-related genes, this would be relevant to a small number of rare conditions (see Chapter 7.2.3).

The investigational approach to the poorly growing child should be targeted on the basis of history, clinical examination and comprehensive evaluation of growth data. Many diagnoses may be apparent on these grounds alone, and only require tests for confirmation (e.g. karyotype in Turner’s syndrome, jejunal biopsy in coeliac disease). However, the majority of children with unremarkable phenotype, but persistent growth failure will need testing of their growth hormone–IGF axis. Interpretation of these test results must be approached with caution. Validation of assay performance and availability of normative data are very important. If pituitary dysfunction is suspected, then neuro-axis imaging is a useful additional investigation. The diagnosis of growth hormone insufficiency and, hence, the decision to embark on potentially lifelong parenteral treatment should be based on all information available to the clinician, including careful and extensive investigation.