| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Christie Hospital National Health Service Trust, Withington, Manchester M20 4BX, U.K.
 |
Abstract |
|---|
 Top Abstract I. IntroductionII. GHD in ChildrenIII. GHD in AdultsReferences |
|---|
|
I. Introduction |
|---|
|
TopAbstractI. IntroductionII. GHD in ChildrenIII. GHD in AdultsReferences |
|---|
The type of underlying pathophysiology differs in childhood-onset compared with adult-onset GHD. In childhood, the commonest etiology is isolated idiopathic GHD (1, 2), a blanket term, which includes some children with distinctive pathophysiology that may be demonstrated radiologically, and others in whom the pathological insult is unknown and the explanation for the GHD ill-understood. In contrast, adult-onset GHD is most frequently due to a pituitary adenoma and/or treatment with surgery or radiotherapy (3, 4); isolated idiopathic GHD acquired in adult life has never been reported. In childhood the differential diagnosis is dominated by alternative causes of poor growth (5, 6, 7, 8, 9), whereas difficulties in the diagnosis of adult-onset GHD (10, 11, 12) exist in the obese and in the elderly. The purpose of this review is to evaluate the usefulness of a variety of clinical and biochemical approaches to the diagnosis of GHD in children and adults. In studies that contain GH measurements in milliunits/liter, the conversion into nanograms/ml has assumed an equivalence of 2 U/mg unless stated otherwise by the authors.
|
II. GHD in Children |
|---|
|
TopAbstractI. IntroductionII. GHD in ChildrenIII. GHD in AdultsReferences |
|---|
); therefore, with rare exception,
the diagnosis of GH deficiency must be made on arbitrary grounds. The
more severe the GHD, the less arbitrary the diagnosis, whereas the
"lesser degrees of GHD" (GH insufficiency) merge into normality.
Until 1985, replacement therapy consisted of pituitary-extract GH
obtained from human cadavers; the supply was limited and thus the
diagnostic strategy was directed initially at detecting severe GHD.
Subsequently, recombinant DNA-derived GH became available, the supply
of which is unlimited. Unfortunately, compared with other endocrine
therapies, GH replacement is expensive. Thus, at the present time in
certain countries, children with all forms of GH insufficiency from
severe to mild are considered for GH replacement, whereas in other
countries, because of economic restraints, only those with severe GHD
have a chance of receiving GH replacement. These observations explain
some of the variation in diagnostic criteria applied to the diagnosis
of GHD in different parts of the world.
|
). Thus,
although the proportion of this cohort of short children lying beneath
the 25th centile for velocity remained constant from year to year, the
identity of the individuals comprising that proportion changed, a
phenomenon that could be largely explained by the random error
associated with height velocity.
|
Typically the GH-deficient child has increased subcutaneous fat especially around the trunk. The face is immature with a prominent forehead and depressed midfacial development; this is related to the lack of GH effect on endochondral growth at the base of the skull, occiput, and the sphenoid bone. Dentition is delayed. In males the phallus may be small, and the average age of pubertal onset is delayed in both boys and girls.
B. Radiology
Bone age is delayed, and the degree of delay is related to the
severity and duration of GHD. Magnetic resonance imaging of the
hypothalamic-pituitary region may reveal a mass lesion such as a
craniopharyngioma or thickening of the pituitary stalk due to an
infiltrative condition such as histiocytosis X. The anterior pituitary
may appear aplastic or hypoplastic, and the posterior pituitary may be
ectopically sited. The more severe the radiological abnormality, the
more likely that additional pituitary hormone deficiencies, other than
GHD, are present (24, 25, 26). Septo-optic dysplasia is hall-marked by
varying degrees of hypoplasia of the optic nerves, chiasm, and
infundibular region of the hypothalamus. The septum pellucidum may be
absent.
C. GH provocative tests
GH secretion is pulsatile and serum concentrations are low during
many hours of the day. Thus provocative tests of GH release, rather
than a single basal GH estimation, were introduced to determine GH
status. The commonly used immunoassays for determination of GH levels
are the RIAs and immunometric assays. The first GH estimations were by
RIAs using polyclonal antibodies. Due to the low specificity of these
assays and the several molecular forms of GH, higher GH levels were
obtained than with the later immunometric assays employing two highly
specific monoclonal antibodies. The variability between assays
increases the need for empiric GH cut-off levels, and the definitions
of normality and abnormality should be established in each laboratory.
In a study in which the GH level was estimated in the same serum
samples in four different laboratories utilizing six different
immunoassays, there was a high degree of agreement on the distribution
of normal and abnormal values despite great variation in the actual
cut-off level between laboratories (27). Thus pediatric
endocrinologists need to know exactly what is being measured in their
own GH assay and to define their own cut-off values. In practice,
however, this is not happening: in a recent audit of practice (28), 251
of 413 US pediatric endocrinologists were surveyed about their current
GH treatment practice. More than 80% used GH stimulation tests to
decide whether to start GH therapy in a child with a growth disorder;
however, 37% did not even know which type of GH assay their laboratory
used and 82% used GH in short poorly growing children regardless of
stimulation test results (28).
The first established pharmacological stimulus introduced for assessment of GH status was insulin hypoglycemia [insulin tolerance test (ITT)]. Advantages of this test include the facts that 1) the ACTH-adrenal axis can be assessed at the same time, 2) it constitutes a powerful stimulus to GH release, meaning that the range between normal and severely GHD is large, and 3) moderate hypoglycemia is sufficient to elicit maximal GH responses (29). The main disadvantages include the lack of normative data in children, a characteristic it shares with many other pharmacological tests, and the unpleasant nature of the test, which in inexperienced hands is frankly dangerous (30). In addition, growing awareness that any normal child might "fail" any single GH provocative test led to the evolving strategy of submitting a child with a growth disorder to two GH provocative tests (31, 32). It should be remembered that the main reason for this strategic development in a child who had "failed" the first GH stimulation test was to exclude normal children ("passed" the second test) from being exposed to a trial of GH therapy rather than adding confirmatory weight to the diagnosis of severe GHD ("failed" the second test).
Thus, a whole series of other pharmacological stimuli were introduced into the GH diagnostic investigational arena including L-dopa (33), arginine (34), glucagon (35), propranolol (33), and clonidine (36). A variety of combinations of these tests has been used (37) and, in some centers, two provocative stimuli are administered sequentially or in combination (33, 38, 39, 40). The latter approach may be time-saving and more economical, but there is no evidence to suggest that the results are more meaningful if the tests are performed in combination rather than individually.
Sex steroid priming with estrogen (41) or androgen (42) administered for a few days before carrying out a GH provocative test is an additional maneuver designed to distinguish between genuine GHD and constitutional delay in growth and puberty (CDGP) (43, 44). At puberty in normal children, there is a marked amplitude-modulated increase in GH secretion (45, 46, 47) directly due to the marked rise in sex steroid concentrations. Teenage children with CDGP often exhibit very low GH secretion while still prepubertal or in early puberty but show a normal increase in GH secretion with progress through puberty. Their GH responses to unprimed provocative stimuli mirror the physiological changes in spontaneous GH secretion. Thus, in a child of peripubertal age with a growth disorder, CDGP is a real possibility in the differential diagnosis. In such a child, priming with sex steroids before carrying out GH provocative tests will delineate severe GHD (minimal GH response) and a significant proportion of those with CDGP (pronounced GH response) providing sufficient data exist to define the degree of change in GH responsiveness to sex steroid priming (40). Nonetheless the gray zone at the interface between the diagnoses of GH insufficiency and CDGP will not be unraveled.
There is a paucity of normative data about the GH response to the
entire range of provocative stimuli (32). Therefore the threshold level
used to define a normal GH response was defined arbitrarily. Initially
it was set at 5 ng/ml, based on the results of studies such as those
performed by Kaplan et al. (48, 49) in which 80 of 91 (88%)
children who were not GH-deficient showed a peak GH response to an
ITT 
5 ng/ml whereas all 53 GH-deficient children had a peak GH
response below 5 ng/ml to the same stimulus. The GH response set-point
was gradually moved to 7 ng/ml (32) as GH testing became more common,
and finally reached 10 ng/ml with the increased availability of
biosynthetic GH. Not only is normality defined arbitrarily, but the
same threshold GH level is used to define normality irrespective of the
pharmacological stimulus applied. In fairness, more normative data have
become available recently (50, 51, 52), but they have not resolved the
question of threshold definition (Fig. 3).
|
The exercise test is safe, simple to perform as an outpatient procedure, and inexpensive (53, 54, 55, 56). The results are highly influenced by the standardization of the exercise test and the degree of work performed by the subject. Unfortunately, an absent GH response to exercise may occur in up to a third of normal prepubertal children (56). However, in a recent very large survey (52) of the reliability of 10 provocative tests to assess GH secretory status carried out in 472 normally growing children, the mean peak GH response to exercise was not significantly different from that seen in response to an ITT or to arginine, clonidine, L-dopa, and glucagon tests. The exercise test, however, is no longer used very often in clinical practice (37).
Many of the criticisms and limitations of provocative GH testing also characterize measurements of spontaneous GH secretion. Such tests typically require blood sampling every 20 min for a minimum of 12–24 h and will require hospitalization. Assays of serum samples (n = 36–72) for GH concentration are expensive. Although continuous serum sampling through a constant blood withdrawal system can be employed, such testing also requires an indwelling catheter and does not permit analysis of GH pulsatility (57, 58).
There have been several reports in children of a poor correlation between spontaneous GH secretion and GH concentrations after provocative tests (59, 60, 61). In some of these studies a group of poorly growing children were described as exhibiting GH neurosecretory dysfunction (GHNSD) (62, 63). These children are characterized by normal GH responses to provocation tests but reduced mean 24-h serum GH concentrations; there may be several aetiologies for the latter combination of GH results including cranial irradiation (64). The exact prevalence of GH neurosecretory dysfunction among children with abnormalities of growth is unknown, and the phenomenon is yet to be described in adult patients. The pathophysiology remains unknown, and very little of scientific or practical value has been gained from the introduction of this expression (GHNSD) into the endocrine literature.
The controversy over whether or not a 24-h spontaneous GH profile is more useful for assessing GH status than provocative GH tests is unresolved. It is clear, however, that while 24-h GH profile group data are highly reproducible, the intraindividual variation is wide, although no greater than that seen with the GH response to a variety of provocative tests (65, 66). This means that the interpretation of a single GH profile is difficult and, in a child being investigated for growth failure, must be interpreted with caution. Although a combination of nighttime spontaneous GH profiles with other indices of GH status, such as insulin-like growth factor I (IGF-I) measurement, demarcated more effectively between normality and short or slowly growing children compared with the GH profile alone (67). From a pragmatic viewpoint, however, very few pediatric centers in the world have the staff, time, or financial resources to carry out 24-h GH profiles, and even fewer centers have normative data for comparative purposes. Therefore, with rare exception, 24-h GH profiles remain a research investigation.
Before 1970, attempts to quantitate GH output in the urine failed because of insensitive assays. Since 1985 the development of ultrasensitive enzyme-linked immunosorbent assay or immunoradiometric assays have enabled detection of urine GH levels as low as 0.4 to 4 pg/ml. This has led to renewed interest in the use of urinary GH estimation in the diagnosis of GHD (68, 69, 70, 71, 72, 73). The test is inexpensive, easy to perform, and noninvasive. Limitations in the interpretation of results relate to large inter- and intraindividual variation (74) and the effect of renal function, which accounts for 52% of the variability in urinary GH measurement. Initial reports were encouraging, but it has since become clear that although urinary GH estimation may match other tests in its ability to confirm severe GHD, it contributes little to the distinction between GH insufficiency and idiopathic short stature (75, 76).
E. IGFs
The IGFs are related GH-dependent peptide factors believed to
mediate many of the anabolic and mitogenic actions of GH (77). The
serum level of the major GH-dependent peptide IGF-I is stable during
the day, due mainly to the complexing of IGF peptides with a family of
IGF-binding proteins (IGFBPs) (78). Thus the potential for assessing GH
status with a single estimation of the circulating IGF-I level proved
attractive and gave rise to the hope that eventually dynamic GH
provocation tests may become unnecessary. Initially IGF assay problems
were caused by interference from the presence of IGFBPs (78, 79). This
problem has been overcome by using a variety of approaches including
acid size exclusion chromatography before the IGF assay, blocking
IGFBP-binding sites with an excess of the nonmeasured IGF peptide
(excess IGF-II for an IGF-I assay) or IGF analogs that do not bind to
IGFBPs as radioligands (80). Additional problems remain, however,
including lack of specificity. A reduced IGF-I level may occur in a
child with malnutrition (81), hypothyroidism (82), hepatic disease, or
diabetes mellitus (83), as well as GHD. The IGF-I level is influenced
markedly by age and pubertal development (84, 85, 86). Particularly low
concentrations of IGF-I occur in normal children younger than 5 yr of
age. Therefore, the use of IGF-I estimation to distinguish between
normal and GH-deficient children is less successful in this age group.
Thus the use of age and puberty-corrected IGF-I values improves the diagnostic use of IGF-I estimation, but overlap with normality remains a problem. Using GH testing as the gold standard to distinguish between 155 children with GHD and 219 with normal short stature, an IGF-I estimation at the fifth percentile level of the normal range provided a sensitivity of 95%, specificity of 60%, and accuracy of 75%, while improved figures of 79%, 95%, and 89%, respectively, were associated with an IGF-I cut-off level of 0.1 centile (87). The IGF-I results discriminated better in those over 8 yr compared with those under 8 yr of age (87). Juul and Skakkebaek (88) also concluded that measurement of IGF-I was useful in the diagnosis of childhood GHD but, unlike the findings reported by Blum (87), more so in younger than older children. They reported an accuracy (predictive value) of 88.8% and 52.3% in children aged less than 10 yr and between 10 and 20 yr, respectively (88). In the latter studies (87, 88) the IGF-I values were corrected for age and sex but not pubertal status.
Despite the promise of these results, multiple studies exist that indicate the serum IGF-I concentration does not correlate perfectly with GH status, as determined by provocative GH testing (76, 89, 90). To improve the diagnostic value of IGF-I measurements, the additional measurement of IGF-II has been recommended (91). Despite the fact that IGF-II is less GH dependent than IGF-I, the normal range for IGF-II levels is relatively high in young children and shows little age-dependence. Thus, although the accuracy of serum IGF-II measurement alone in detecting GHD is moderate, the combined use of IGF-I and IGF-II improves their value as a diagnostic parameter. In a study of 68 GH-deficient children, 197 normal-statured children, and 44 normal short children, 18% of the GH-deficient children had serum IGF-I levels within the normal range, whereas 32% of normal short children exhibited low IGF-I concentrations. Low IGF-II levels were found in 52% of GH-deficient children but also in 35% of normal short children. Utilizing the results from both IGF assays, however, revealed that 4% of GH-deficient children had normal IGF-I and IGF-II levels but only 0.5% of normal children and 11% of normal short children had reduced concentrations of both IGF-I and IGF-II (91). In reality, however, very few pediatric endocrine centers use IGF-II assays in the assessment of children with short stature.
F. IGFBPs
Of the six known IGFBPs, IGFBP-3 is normally the major serum
carrier of IGF peptides (92). IGFBP-3 circulates as part of a ternary
complex consisting of IGFBP-3, an IGF peptide, and an acid-labile
subunit (93, 94). Both acid-labile subunit and IGFBP-3 are GH
dependent. Since IGFBP-3 determinations reflect combined IGF-I and
IGF-II concentrations, age dependency of IGFBP-3 is less striking than
for IGF-I. Similarly the influence of nutritional status on IGFBP-3
levels is less than for IGF-I levels. The IGFBP-3 assays do not require
separation of the IGFBPs from IGF peptides and are technically easier
to perform.
Yet again, the possibility of a single measurement of a GH-dependent peptide with a long half-life as a measure of GH status proved attractive. The close correlation between molar concentrations of IGFBP-3 and the sum of molar concentrations of both IGF-I and IGF-II suggested that IGFBP-3 estimation could reflect GH status more closely than IGF-I estimation alone. Initial (95) and some subsequent (96) reports were highly positive. Blum et al. (95) reported 128 of 132 children with GHD had serum IGFBP-3 levels below the fifth percentile. In contrast, 124 of 130 ’normal’ children had IGFBP-3 concentrations above the fifth percentile. In a more recent evaluation of the usefulness of IGFBP-3 measurement in the diagnosis of GHD, Juul and Skakkebaek (88) reported sensitivity and specificity figures of 60% and 97.9% in children less than 10 yr of age and 56.5% and 78.7%, respectively, in those aged between 10 and 20 yr. These results could not be reproduced in other studies (76, 97, 98, 99, 100). Furthermore, the IGFBP-3 estimation discriminated particularly poorly between GHD and normality in pubertal children and those with radiation-induced GHD (76, 101).
It has been proposed that the measurement of the IGFBP-2 concentration adds to the value of IGF-I and IGFBP-3 assays in the diagnosis of GHD (102). IGFBP-2 values are usually elevated in patients with GHD (102, 103). Despite the fact that in most patients there is agreement between the IGFBP-2/IGF-I ratio, IGFBP-3 measurements, and the results of GH testing, the IGFBP-2/IGF-I ratio was discordant from the GH response in 21% of 80 patients (57 GHD and 23 idiopathic short stature), and IGFBP-3 results were discordant in 18% (102).
The proponents of IGF-I and IGFBP-3 estimations as the best choice of initial investigations in a child with possible GHD point out that in individuals with genetic forms of GH insensitivity (GHI), all have a low IGFBP-3 and most have a low IGF-I concentration (104, 105). By definition, these individuals have elevated GH concentrations associated with a mutation or deletion of the gene for the GH receptor, which renders them insensitive to GH action. Furthermore, despite the universally low IGFBP-3 concentrations and the characteristically severe growth failure, serum IGFBP-3 concentrations still correlate significantly with the height SDS (104, 106).
Thus, low IGF-I and IGFBP-3 concentrations are reliable guides to the diagnosis of severe GHD, providing the investigator considers the alternative possibilities of malnutrition, hypothyroidism, liver disease, and GHI. Discrepancies between IGF-I and IGFBP-3 results compared with the GH responses to dynamic tests exist, however, in those children believed to have lesser degrees of GHD.
The inadequacies of dynamic provocative tests of GH release, outlined previously, may explain the discrepancies in the GH status defined by the GH and IGF axis results. Alternatively, it may be that IGF-I and IGFBP-3 estimations are less useful than the peak GH response to provocative testing in children with GH insufficiency.
G. Newer strategies for assessment of GH status
The effects of the commonly used provocative tests of GH release,
such as arginine, clonidine, and the ITT, are mediated through
hypothalamic mechanisms. The availability of GH releasing hormone
(GHRH) provided a means of assessing the secretory capacity of the
pituitary somatotrope directly. Although in normal subjects, the GHRH
test provokes a greater release of GH than several conventional
provocative stimuli, there is great variability in the GH response to
GHRH probably related to variation in endogenous somatostatin tone
(107, 108).
The use of GHRH in combination with substances that act via inhibition of endogenous somatostatin, such as pyridostigmine (a cholinesterase inhibitor) and arginine, has been explored (109, 110). The suppression of hypothalamic somatostatin release with the latter substances enables GHRH to provoke a much greater GH response, and the GH response to the combined test (GHRH and arginine) shows much reduced inter- and intraindividual variability (111). In a recent comparative study of the reliability of various GH provocative tests, Ghigo et al. (52) studied the GH response to exercise, an ITT, arginine, clonidine, L-dopa, glucagon, pyridostigmine, GHRH plus pyridostigmine, and GHRH plus arginine in 472 children with normal stature or normal short stature. In this normal cohort, the combination of GHRH plus either arginine or pyridostigmine not only provoked a significantly greater GH response than the remainder of the GH provocative stimuli but the minimum individual response in 175 tests was 19 ng/ml! Despite these impressive "stimulated" GH levels in normal children, GHRH alone or in combination with pyridostigmine does not meet the essential requirements for distinguishing between children with GHD and normal subjects. As the authors themselves indicate (52), a normal GH response to these tests cannot exclude GHD due to hypothalamic dysfunction. This is a crucial limitation in that in most children with isolated GHD, the site of abnormal pathophysiology is believed to be hypothalamic rather than pituitary. It is not clear whether the same reservation applies to the combined GHRH and arginine test.
In contrast to the combined approach of a bolus of GHRH plus suppression of endogenous somatostatin tone with pyridostigmine or arginine, there is evidence that somatostatin pretreatment enhances the subsequent GH response to an acute bolus of GHRH 5 h later in normal children (112). The exact mechanism is not clear but possibilities include somatostatin blocking the release of stored intracellular GH without preventing accumulation, chronic somatostatin exposure resulting in somatostatin receptor desensitization, or a central action of somatostatin within the hypothalamus. The results of this new diagnostic approach discriminated well between children with GHD and normal short children, but the numbers were small as only 10 children with normal short stature were studied (112). Six of these children had normal GH responses to standard pharmacological tests and four had borderline GH responses (112).
Maghnie et al. (113) recommended metabolic rather than neuropharmacological pretreatment to enhance the GH response to provocative stimuli. They showed that the stimulated GH response is strongly calorie dependent and that 3 days of a hypocaloric diet can increase the number and amplitude of spontaneous GH pulses and the total GH responses to arginine and to an ITT (113). The exact implications for a child who shows a prediet poor GH response to pharmacological stimuli and a "normal" GH response after the diet are unclear. Therefore, a 3-day hypocaloric diet before GH testing cannot be recommended in the diagnostic assessment of GH status in a child with a growth disorder.
An alternative therapeutic strategy, independent of the GH status of the child, has been introduced in Australia (114). The latter approach has been advocated since 1988 when the guidelines were revised to allow eligibility for GH therapy on auxological criteria alone. Initial entry criteria were height less than the third centile and growth velocity less than 25th centile for bone age. GH testing was continued, however, in most children even though it was not mandatory for the prescription of GH. More than 3100 children have been treated since 1988 including 35% with GHD (peak GH response <10 ng/ml), 12.5% with Turner syndrome, and 52% with other non-GHD forms of short stature. In 1994 the guidelines were revised to restrict the use of GH therapy to subjects with height less than the first centile, and cessation of GH therapy was brought forward to bone age 13.5 yr for girls and 15 yr for boys. New patient accruals have decreased since 1992 from 100/yr to less than 50/yr. Expenditure for GH between 1990–1991 and 1994–1995 has been halved and is in the midrange internationally (114).
The Australian auxology-based approach together with a comprehensive national database has worked well in a single country with a small number of prescribers and a tight audit system. Nonetheless, the identification of children with GHD remains important not only because of awareness of the potential for associated pathology and other hormone deficiencies but also as a predictor of significantly lower GH dose requirements for maximal growth response in children with severe GHD (peak GH response <5 ng/ml) (114). It should be pointed out, however, that the conditions necessary for the Australian system to work successfully would exclude many countries in the world from adopting this strategic approach.
H. Practical approach to the diagnosis of GHD
The usual circumstance in which GH status needs to be assessed is
the child with standing height SDS below -2 and a growth velocity
below the 10–25th centile in whom other causes of poor growth have
been excluded. The major diagnostic difficulty is to distinguish
between idiopathic GHD and idiopathic short stature; the reasons
include 1) lack of age- and puberty-related normative data to define
the threshold for subnormal IGF-I and IGFBP-3 levels and GH responses
to provocative tests, 2) application of a single fixed cut-off level
for GHD independent of the GH provocative test in use, and 3) poor
reproducibility of GH provocative tests or spontaneous GH profiles. The
most important reason, however, is biological in that GHD and normal
short stature do not represent cleanly demarcated entities, a view
supported by the evidence of overlap of all auxological and biochemical
parameters in children with these two diagnostic labels.
GH treatment may improve the short-term growth velocity in any individual short child. Thus, one cannot use the growth response to GH during the first year of GH therapy to confirm the diagnosis of the GH insufficiency. Furthermore, after 3 yr of GH therapy the growth response of children with peak GH responses between 5 and 10 ng/ml is not significantly different from that of children with peak GH responses in excess of 10 ng/ml (115).
The most common approach to the practical assessment of GH status will not utilize 24-h spontaneous GH profiles and will not rely on GHRH tests in combination with arginine or pyridostigmine, in view of the inability to detect GH insufficiency associated with hypothalamic dysfunction reliably with this test.
This leaves IGFs, IGFBPs, dynamic GH provocative tests, and urinary GH estimation. It is the degree of GHD that dictates the extent of the diagnostic confusion; the child with severe GHD is least likely to be confused with the normal short child. Thus, measurement of IGF-I and IGFBP-3 and dynamic tests of GH release all perform best in those children with severe GHD. Urinary GH estimation appears to be a useful diagnostic discriminator only in those children with severe GHD and is singularly unhelpful in those with a lesser degree of GHD.
In practice, most information is gained by the performance of a single IGF-I and IGFBP-3 estimation in combination with a single dynamic GH provocative test. The performance of a second GH provocative test will add little to the diagnosis of severe GHD. In a poorly growing child with normal IGF-I and IGFBP-3 levels but a subnormal peak GH response to a provocative test, a second GH provocative test will help to distinguish between GH insufficiency and normal short stature. If the result of the second GH provocative test is normal, it suggests that the diagnosis of GH insufficiency has been refuted. If both provocative tests produce subnormal GH responses, the diagnosis of GH insufficiency will have been supported. It might be argued that this approach is wasteful because there may be very little difference in the growth response to GH therapy between children with these two sets of biochemical parameters. In the latter child defined as being GH insufficient, however, further biochemical assessment of pituitary function and neuroradiology are required. Furthermore, in this situation we may need to consider biological endpoints other than growth. Compared with the child who shows a normal GH response to at least one provocative test, it is more likely, but unproven, that a child with moderately reduced GH responses to two provocative tests is more prone to failure of acquisition of peak bone mass if left untreated.
A combination of endocrine results, less commonly seen and not yet considered, is that of a low IGF-I level in the presence of normal or exaggerated GH responses to provocative stimuli. The latter biochemistry may be associated with a number of clinical states associated with acquired GH insensitivity such as a chronic systemic disease, i.e., chronic liver disease (116), severe chronic infection (116), and drug therapy, i.e., chemotherapy for a brain tumor (117), but under these circumstances the clinical picture is dominated by the primary disorder and differential diagnosis is not an issue. There are, however, a variety of genetic defects in the GH-IGF-I axis that may result in a short poorly growing child with a low IGF-I level and normal or exaggerated GH responses to provocative stimuli.
The reported genetic defects include a mutation in the GH gene resulting in biological inactivation of GH leaving its immunological properties intact (118), mutations in the GH receptor gene (119, 120), and intrauterine growth retardation and postnatal growth failure associated with deletion of the IGF-I gene (121).
In its classic form, Laron’s syndrome (GH receptor gene mutation) is inherited as an autosomal recessive trait and is characterized by severe postnatal growth failure and low IGF-I and IGFBP-3 levels despite normal or elevated GH secretion (119, 120). The initial cases were all associated with low serum concentrations of GH-binding protein (GHBP), indicating genetic abnormalities in the extracellular domain of the GH receptor. Subsequently, several hundred cases have been identified, some of whom have had normal serum concentrations of GHBP, indicating either defects in the portion of the extracellular domain of the GH receptor molecule that is needed for dimerization or intracellular defects that may affect the signal transduction pathway (122).
Further biochemical investigations to distinguish other genetic defects in the GH-IGF-I axis from that of GH receptor gene mutations are not always helpful in that the GH gene mutation associated with a biologically inactive GH molecule may, like the GH receptor gene mutation, be associated with a low GHBP level and a failure to show an IGF-I response to short-term administration of exogenous GH (118).
The one child thus far described with deletion of the IGF-I gene had normal GHBP and IGFBP-3 levels associated with severe intrauterine growth retardation, sensorineural deafness, and mental retardation (121).
Further investigation of the short slowly growing child with a definitely low IGF-I level and normal or exaggerated GH secretion is likely to consist of screening for either a molecular defect of the GH receptor gene or IGF-I gene, or isoelectric focusing of serum GH.
Increasingly, less classic molecular defects of the GH receptor (123) are being reported, leading to extreme difficulty in determining the contribution of such defects to the growth problem of the child under investigation.
Many areas of the growth/GH/IGF-I field require further research. If one adopts an arbitrary cut-off for the biochemical definition of GH insufficiency, what is the long-term growth response and final height of children below the cut-off, who receive a conventional dose of GH replacement, compared with those children just above the cut-off, who remain untreated? Are there differences in the social or educational performances of these children? Is one group "happier" than the other in the short term, long term, or very long term?
What physiological explanation lies behind the short-term and long-term variation in growth in normal children? Why do children with radiation-induced GHD rarely exhibit a low IGFBP-3 level (76, 101) and more frequently exhibit an impaired GH response to an ITT than to other GH provocative tests (124, 125)? How do children previously treated for a craniopharyngioma grow in the presence of low GH and IGF-I levels (126)? What is the mechanism for suppression of GH secretion in children with obesity (47, 127)? These and many other research questions remain unanswered and will require a combination of the biochemical investigations discussed in this review and new ideas.
In practical terms, in the meantime, we should accept that 100% discrimination between GHD and normality does not exist. The overlap between normality and abnormality is more likely to be reduced, however, if investigators acquire their own age- and puberty-defined normal data for their IGF-I and IGFBP-3 assays and their chosen dynamic GH tests rather than seeking more complicated variations on the theme of conventional testing or new multiple combinations of estimates of GH status.
|
III. GHD in Adults |
|---|
|
TopAbstractI. IntroductionII. GHD in ChildrenIII. GHD in AdultsReferences |
|---|
A. Clinical aspects
GHD in adult life is associated with increased fat mass,
particularly distributed in the truncal region (128, 129), reduced lean
mass (128, 129), osteopenia (130, 131, 132, 133, 134), an adverse lipid profile
(135, 136, 137), glucose intolerance (138), insulin resistance (139),
impaired fibrinolysis (140), altered cardiac structure and function
(141, 142, 143, 144), reduced exercise capacity (145, 146), and reduced quality
of life (147, 148, 149). Despite the impressive list of clinical features,
there is no single symptom or sign that is pathognomonic of GHD in
adult life; this is particularly true of patients with multiple
pituitary hormone deficits in whom T4, glucocorticoids, and
sex steroids may be under- or overreplaced. There is increasing
evidence, however, that the profile of adverse sequelae associated with
GHD in adult life differs in adults with childhood-onset compared with
adult-onset disease (150).
There is no single biological endpoint in an adult suspected of being GH-deficient that offers the same diagnostic usefulness as the growth rate of a child. There is, however, the natural history of evolution of hypopituitarism in patients with a mass lesion of the hypothalamic-pituitary region or in those who have undergone surgery and/or radiotherapy to this region. GH is usually the first of the anterior pituitary hormones to be affected by these various pathological insults, which means that in a patient with multiple pituitary hormone deficits the probability of GHD being present is extremely high (151, 152). Unlike the situation in the pediatric age group in whom pituitary adenomas are rare, this observation is of enormous importance in adults in whom nonfunctioning pituitary adenomas and/or their treatment are the most common causes of GHD (4).
B. GH provocative tests
Over a number of years the ITT has been the gold standard. For the
reasons mentioned earlier, this is partly because the ITT is the only
GH provocative test in which vast experience has been gained by
different investigators. With the synacthen test gaining gradual
acceptance as a reliable substitute for the ITT in the assessment of
the pituitary-adrenal axis in significant numbers of patients with
pituitary disease, an alternative test of GH status will be required in
such patients (153, 154, 155).
There have been few studies of the GH response to different pharmacological stimuli in normal adults (156, 157, 158, 159, 160, 161, 162, 163). Cain et al. (156) assessed the GH status of 20 normal young subjects (14 males) aged 21 to 34 yr by challenging with glucagon, arginine, ITT, and tolbutamide-induced hypoglycemia. They observed a significant GH increment from baseline to peak after each agent and concluded that glucagon was at least as good, if not better than the ITT, at stimulating GH release. Their study, however, reported the mean change in serum GH, from baseline to peak, and mean peak GH response to each agent but did not compare one agent with another. Only 10 subjects had an ITT and although the mean serum GH peak was not significantly different from the response to glucagon, none of the subjects had what Cain et al. (156) considered to be a subnormal GH response to the ITT whereas the "failure rates" (peak GH response or rise in serum GH from baseline to peak of less than 5 ng/ml) after arginine and glucagon were three subjects and one subject, respectively. No information was provided about the sex of the 10 subjects who underwent the ITT; this may be relevant as Hoeck et al. (164) have recently demonstrated that the peak GH response to an ITT is lower in female subjects than in males. This is in contrast to the greater GH response to arginine that is seen in women compared with men (160, 165, 166); thus, there is a gender-related effect on the GH response to certain provocative stimuli that varies in direction depending on the stimulus.
Certain conclusions can be drawn from the few studies that have been carried out. The ITT does appear to be a more effective stimulus for GH release compared with glucagon, arginine, L-dopa, and clonidine. In response to an ITT, all 19 healthy individuals in the Lin and Tucci study (157) achieved a peak GH response of greater than 5 ng/ml and, in the 18 healthy young male students studied by Rahim et al. (161), all but one (10.7 ng/ml) responded with a peak GH concentration in excess of 15 ng/ml. Comparison of the numerical results of these studies, however, is hindered by the fact that the early studies used GH assays that are now obsolete.
Oral clonidine in doses of up to 150 µg/m2 proved singularly unimpressive in stimulating GH release (160, 161). Rahim et al. (161) showed that clonidine (100–200 µg) elicited a GH response no greater than placebo. Intramuscular glucagon (1 mg) stimulated a greater GH response than intravenous arginine (20 g/m2 over 30 min) with 16 of the 18 students achieving a peak GH response in excess of 7.7 ng/ml and the remaining two between 3.8 and 7.7 ng/ml to glucagon. Two of the 18 subjects failed to achieve a peak GH response in excess of 1.9 ng/ml in response to arginine (161).
All the limitations of the GH provocative tests discussed in the
pediatric section of this review apply to the diagnosis of GHD in
adults, but one major difference is in the degree of GHD required to be
present before GH replacement is initiated. In many countries, children
with all grades of GHD ranging from complete absence of GH secretion to
GH insufficiency are offered a trial of GH replacement. Thus far, GH
replacement therapy has only been recommended for adults with severe
GHD. This reduces the scale of the diagnostic problem in defining GHD
in the adult with pituitary disease. Nonetheless, controversy over the
definition remains an issue. In six recent clinical trials of GH
replacement, the diagnostic criteria for GHD varied between a peak GH
response to a provocative test (usually ITT) of less than 0.5 and 5
ng/ml (128, 167, 168, 169, 170, 171). To bring some consistency to the field, a recent
consensus statement recommended that severe GHD be defined as a peak GH
response of less than 3 ng/ml during an ITT (172). The findings of
Hoffman et al. (173) support the choice of this diagnostic
definition; they studied a variety of methods of diagnosing GHD in 23
hypopituitary patients and 35 normal subjects carefully matched for
age, sex, and body mass index (BMI). Compared with their normal counter
parts, the patients with organic hypopituitarism had significantly
lower peak GH responses to an ITT, lower integrated GH concentration
(IGHC) from 20-min sampling over 24 h, and lower IGF-I and IGFBP-3
levels (Fig. 4). There was a complete
segregation of the peak GH response to an ITT between the normal and
GH-deficient subjects irrespective of age or adiposity. No normal
subject had a GH response below 5 ng/ml, and no subject with organic
hypo-pituitarism had a response greater than 3 ng/ml. In contrast
to the ITT, there was a significant overlap of IGHC, IGF-I, and IGFBP-3
values between groups (173).
|
In adult patients with known pituitary disease, not only the presence
but also the severity of GHD is related to the number of additional
pituitary hormone deficits (4, 176). Thus, Toogood et al.
(4) showed that the median peak GH response to an ITT in four
age-matched groups of adult patients with isolated GHD (GHDO) and GHD
plus one (GHD1), two (GHD2), and three (GHD3) additional pituitary
hormone deficits was 3.8, 1.5, 0.8, and 0.7 ng/ml, respectively (Fig. 5). There was a significant downward
trend in the median peak GH responses of the four groups with the
differences between GHDO and GHD1, and GHD1 and GHD2, being highly
significant; however, there was no difference between the GHD2 and GHD3
groups. Ninety one percent of patients in combined groups GHD2 and
GHD3, 55% of GHD1, and 24% in GHDO had a peak GH response less than
1.9 ng/ml. Similar conclusions regarding the relationships between GHD
and additional pituitary hormone deficits have been drawn irrespective
of whether GH status has been assessed by ITT (176), urinary GH
measurement (177), or 24-h spontaneous GH profile (178).
|
Using a newer generation enzyme-linked immunoabsorbent assay to enhance sensitivity (1 ng/liter), the same group reassessed the discriminatory capacity of IGHC (179). Twenty six percent of hypopituitary subjects had IGHC values within the normal range (179). Age stratification improved the separation, but an overlap between normal and GHD remained in those adults aged less than 50 vs. those over 50 yr of age. All samples, including nadirs, from GH-deficient subjects were well within the GH assay detection limit (1 ng/liter) (179). Peak 24-h GH levels in GH-deficient subjects were lower and did not overlap those in the normal subjects (180). Nadir GH concentrations were significantly lower in GH-deficient subjects but the range overlapped that of normal subjects (180). Total daily GH production in GHD was approximately 5% of the production in matched normal subjects (180). This difference resulted from a greater reduction in the pulsatile (by 96%) than in the tonic (by 47%) component, so that the fractional daily contribution by tonic GH release in GH-deficient subjects was markedly greater (180).
Thus, the peak 24-h GH level but not the IGHC or the nadir GH level discriminated satisfactorily between GH-deficient subjects and controls; however, the series contained only 10 GH-deficient patients and 10 controls (180). Furthermore, one could only determine the peak spontaneous GH level by performing a full 24-h GH profile! These physiological studies utilizing highly sensitive GH assays have the potential to advance our understanding of the factors that control tonic and pulsatile GH secretion. Nonetheless, they are unlikely to contribute significantly to the pragmatic need to define GHD biochemically in an individual patient.
There is less information from alternative methods of assessment of physiological GH secretion. Urinary GH excretion showed reasonable separation from normal controls in GH-deficient patients defined by the presence of known pituitary disease and a peak GH response of less than 5 ng/ml to an ITT (177). At a sensitivity of above 90%, the specificity for diagnosing GHD was 79% in adult patients below 40 yr, 67% between 40 and 60 yr, and an unimpressive 36% above the age of 60 yr (177).
D. IGF-I and IGFBP-3
The changes in IGF-I levels throughout life are similar to those
of GH. With the onset of puberty there is a 2 to 3-fold rise in serum
IGF-I concentrations followed by a decline such that average adult
levels are reached by the early twenties (86). There follows a gradual
decline with advancing age (86). Similar to IGF-I but less
age-dependent, serum IGFBP-3 levels rise to a peak during the pubertal
years and then slowly decline in adulthood (95). Thus, in the adult
patient with potential GHD, serum IGF-I and IGFBP-3 measurements can
only be interpreted if decade-based normative data are available.
The usefulness of an IGF-I estimation in the diagnosis of adult GHD is a matter of contention although it has become clear that at least some of the disparity between studies is explained by the timing of the onset of GHD. Hoffman et al. (173) found that 70% of IGF-I and 72% of IGFBP-3 values in adult-onset GHD patients, mean age 45 yr, were within the range of normal subjects even allowing for the effects of age. Holmes (181), in a subsequent study of 65 adults, mean age 35 yr, with GHD defined by a peak GH response of less than 3.8 ng/ml to a provocative test, and associated with other evidence of pituitary disease, found that 70% had an IGF-I SDS below -2. In the latter study, however, the 65 adults consisted of a mixture of childhood-onset and adult-onset GHD (181).
DeBoer et al. (11) focused on childhood-onset GHD and
reevaluated GH status in 89 young adult males who had previously
received GH replacement in childhood. Approximately 93% of the
patients had an IGF-I level below the normal range with a similar
number of subnormal IGFBP-3 levels among the patients (Fig. 6). Attanasio et al. (150)
pursued the same theme in a large study of 74 childhood-onset and 99
adult-onset GHD patients. They concluded that there are profound
differences between these two forms of GHD (150). With group data they
observed that the serum IGF-I levels were below normal in both groups
of GH-deficient patients but were significantly lower in
childhood-onset than adult-onset GHD patients (150). Other authors have
made similar observations (182, 183). In these studies, however, it is
not clear whether sufficient decade-based normative data were available
or whether the severity of GHD was equal in the childhood-onset and
adult-onset groups. Thus, apart from the potential impact of the timing
of onset of GHD, it is possible that the severity of GHD influences the
interpretation of IGF-I results.
|
). In each of
the latter categories, there is a mix of adult-onset and
childhood-onset GHD patients, thereby allowing a comparison of IGF-I
status among groups with tightly defined degrees of GHD. The numbers
are sufficient for statistical comparison in the GHDO, GHD1, and GHD3
groups and, in each category, the median childhood-onset IGF-I SDS is
lower than the median adult-onset IGF-I SDS, reaching significance in
two of the three groups (GHDO, -3.83 vs. -2.69,
P = 0.239; GHDI, -5.75 vs. -2.03,
P = 0.022; GHD3, -6.28 vs. -2.09,
P = 0.0003) (181). These observations suggest that for
a given degree of GHD, the serum IGF-I concentration is lower in adults
with childhood-onset compared with adult-onset disease. The explanation
for this latter observation is unclear; however, it supports the notion
that an IGF-I estimation is more likely to be of diagnostic use in
childhood-onset than in adult-onset GHD.
|
E. Obesity
Obesity is one pathophysiological state that may be difficult to
distinguish from organic GHD in the adult. There is substantial
evidence that morbid obesity is accompanied by suppression of GH
release and that substantial weight loss may restore spontaneous (184, 185) and stimulated GH secretion (184). Veldhuis et al.
(186) observed mean 24-h GH levels in obese men (BMI 41–58
kg/m2) to be 25% that of an age-matched control group (BMI
23–33 kg/m2). Even in subjects of near normal weight,
relative obesity as determined by BMI is negatively correlated with GH
secretion (160, 187, 188) due both to reduced GH production and
increased clearance (186). Each unit increase in BMI, at a given age,
reduced daily GH secretion by 6% (187). Furthermore, in clinically
nonobese healthy adults, relative adiposity in the abdominal region in
particular proved to be a major negative determinant of stimulated GH
secretion (160).
Changes in IGF-I levels in obesity are disputed. There are data showing decreased (187, 189, 190, 191) or normal (186, 192) IGF-I levels in adults and even elevated (193) IGF-I levels in obese children. In the largest population-based study of IGF-I levels in adults, an apparent decline in IGF-I levels attributable to adiposity disappeared when age was accounted for in a multivariate analysis (194).
Exactly how obesity reduces GH secretion has not been clarified; however, there is increasing evidence that FFA play a significant role (195, 196). Recent studies with acipimox, a nicotinic acid analog that blocks lipolysis and is devoid of side effects, indicate that FFA reduction enhanced the GH response to a variety of GH secretagogues, but it did not increase spontaneous GH secretion in obese individuals (197, 198, 199). The possible site of action of FFA is at the pituitary somatotrope.
A recent study in age and sex-matched subjects suggests that there is a much more profound reduction in total GH secretion in a group of individuals with organic GHD compared with a group of obese subjects (200). In the obese individual with pituitary disease, however, and no other pituitary hormone deficit, a reduced GH response to any of the standard provocative tests may reflect organic GHD or obesity itself. In this clinical situation, distinction between the two possibilities is not possible with any degree of certainty at the present time.
F. Elderly
GH secretion in healthy elderly adults is reduced compared with
that in young adults (186, 187, 189, 201, 202, 203, 204). GH secretion declines
by approximately 14% per decade from young adult life (187), and an
early study even suggested that GH secretion may eventually cease in
certain elderly subjects (201). Normal aging is associated with changes
in body composition similar to those seen in patients with GH
deficiency, i.e., a reduction in bone mass, osteoporosis in
severe cases, and a reduction in renal and cardiac function. The
similarity of the changes in adults with GHD and those that are part of
the normal aging process has led some researchers to suggest that aging
may be due to GHD (205). The decline in GH secretion and the
physiological changes seen with aging have been grouped together and
termed the somatopause.
In the clinical setting this raised a number of questions. Can the GH status of elderly patients with organic pituitary disease be distinguished from that of the normal elderly? Do middle-aged adults with GHD need to continue with GH replacement beyond the age of 60 yr?
The prevalence of nonfunctioning pituitary adenomas increases with age (206) and thus pituitary disease is common in the elderly. Toogood et al. (207) have now established that GH secretion is significantly reduced in the elderly with pituitary disease compared with normal controls of similar age. They studied GH secretion in 24 patients (age 61–85 yr) with hypothalamic-pituitary disease and in 24 controls (age 60–87 yr) matched for BMI. The median (range) area under the curve of the 24-h GH profile [<9.6 (<9.6–20) vs. 18.5 (10.7–74.4) ng/ml], the median stimulated peak GH response to arginine [<0.4 (<0.4–7.7) vs. 8.0 (1.6–37.0) ng/ml], and the median serum IGF-I concentration [102 (<14–162) vs. 147 (65–255)ng/ml] were significantly lower in the patients than in the controls (207). For each of the three latter indices of GH status, there was overlap between the two groups. Fifteen of the 24 patients showed no evidence of spontaneous or stimulated GH secretion (using a standard immunoradiometric assay with a sensitivity of 0.4 ng/ml) whereas all controls showed evidence of both.
Under circumstances in which GH secretion is known to be reduced, however, most values during a 24-h GH profile cannot be estimated with conventional RIAs. Thus, a limitation of the original study by Toogood et al. (207) was that 93% and 61% of the GH estimations in the 24-h profiles of patients and controls (72 samples), respectively, fell below the sensitivity of the conventional assay. Subsequently, the latter GH profiles were reexamined using an ultrasensitive chemiluminescent assay (sensitivity 2 ng/liter). GH secretion was detectable throughout the 24-h profile in all patients and controls (178). Furthermore, the pattern of GH secretion remained pulsatile in all but one patient, despite previous pituitary pathology, surgery, and irradiation. GH pulse frequency was unaffected in the patients studied by Toogood et al. (178) and, similar to the findings of Reutens et al. (180) in middle-aged adults with organic GHD, absolute peak GH levels were reduced to a greater extent than absolute nadir GH levels. Nonetheless, overlap between patients’ and controls’ values remained unless the analysis was restricted to those patients with the most severe degree of GHD, i.e., those with two or more additional pituitary hormone deficits in whom the absolute peak GH concentration was distinctly separate from control values (178).
In addition, utilizing the ultrasensitive GH assay on the samples
obtained during the provocative GH test revealed a highly significant
correlation (Fig. 8) between the peak GH
response to arginine and the area under the 24-h GH profile, which was
more pronounced in the patients (r = 0.9) than in the controls
(r = 0.5) (178).
|
Thus, despite the fact that GH secretion in the elderly with pituitary disease may be reduced to about 13% of age-matched controls (178), confirming GH deficiency in an individual patient remains a problem. Twenty four-hour GH profiles and IGF-I and IGFBP-3 measurements are unsatisfactory either because of impracticality or lack of discriminatory power. The excellent correlation seen between the peak GH response to arginine and spontaneous GH secretion supports the view that the arginine stimulation test is a reasonable choice to assess GH status in patients with GHD2 or GHD3 disease, particularly in an age group in whom an ITT may carry an increased risk of morbidity or mortality. In an elderly patient with pituitary disease and either one or no additional pituitary hormone deficits, the distinction between GHD and normality remains a challenge.
G. Childhood onset
Currently, GH replacement in childhood is continued until growth
is completed or there is an agreement between the family and the
endocrinologist that a satisfactory height has been achieved. Over the
last 10 yr there have been a number of studies in which the GH status
of children and young adults, who had received GH replacement during
childhood, was reassessed after completion of growth and puberty (11, 209, 210, 211, 212, 213, 214).
The purpose of the studies was at least 2-fold: first, to establish how many individuals no longer appeared GH-deficient at reevaluation, and second, to determine how many individuals still had severe GHD requiring consideration of GH replacement during adult life. All individuals were documented biochemically to be GH-deficient in childhood but at reassessment GH status was considered to be normal in 20–87% (209, 210, 211, 212, 213, 214, 215). The etiological classification of the childhood diagnosis in the vast majority of the latter subjects was isolated idiopathic GHD. This raises interesting questions regarding the nature of their defect in GH secretion during childhood. The diagnostic threshold for GHD is arbitrarily defined, and the reproducibility of the GH response to provocative testing within individuals is not high. On these grounds alone, it would be anticipated that a percentage of those considered GH-deficient at one point in time might be considered normal at reevaluation. Furthermore, it is likely that in a proportion of these patients, the childhood diagnosis was CDGP and not isolated idiopathic GHD, but the initial GH provocative tests carried out in the "unprimed" state failed to make the distinction. Finally, it remains possible that transient GHD in childhood is a real entity although longitudinally obtained proof is lacking.
In contrast to the findings in the isolated idiopathic GHD population, those young adults diagnosed as having organic GHD in childhood, as a consequence of either a mass lesion, pituitary surgery, or irradiation damage to the hypothalamic-pituitary axis, rarely revert to normal GH status (213).
In addition to influencing the incidence of reversal of GH status, the etiology of the childhood diagnosis of GHD also affects the strategy of how to reevaluate. In our center (213), two provocative tests were carried out in two-thirds of 88 young adults undergoing reassessment of GH status, and one GH provocative test was performed in the remainder. By far, the most commonly used GH provocative tests were the ITT and the arginine stimulation test. Of those undergoing two tests, 55 patients had a peak GH less than 3.5 ng/ml in response to one test, of whom 26 had a peak GH greater than 3.5 ng/ml in response to the second test; similarly, of the 41 patients with a GH peak less than 1.9 ng/ml in response to one test, 26 of them had a GH peak greater than 1.9 ng/ml in response to the second test (213). Although there is a considerable degree of disagreement between the results of the two tests in individuals, the discordance was not explained by weighting associated with a persistently greater or lesser GH response to one particular test (213). Furthermore, discordance between test results is not distributed randomly. Of the 58 patients undergoing two tests at reassessment, 15 had additional anterior pituitary hormone deficits; all of them had a GH peak response below 3.5 ng/ml to both tests. At a diagnostic threshold of 3.5 ng/ml, discordance between the two GH test results only occurred among those patients with a diagnosis of isolated GHD (213).
The results from these studies emphasize that all children who have received GH replacement therapy in childhood should undergo reassessment of GH status in young adult life. The percentage of such patients who merit consideration for GH replacement in adult life will vary depending on the definition of severe GHD in use and the etiological mixture of GH-deficient patient populations from different centers. Patients with isolated GHD should undergo two tests of GH status, but those with additional anterior pituitary hormone deficiencies require only one test at reassessment.
H. Newer strategies for assessment of GH status
Earlier in this review the use of a combination of GHRH with
pyridostigmine as a diagnostic test of childhood GHD was not
recommended (52), the main reason being that in most children with GHD,
the site of abnormal pathophysiology is believed to be hypothalamic
rather than pituitary. In adults the different epidemiology of the
pathophysiology of GHD makes the combined GHRH-pyridostigmine test more
attractive. The most common cause of adult-onset GHD is a pituitary
adenoma or treatment with pituitary surgery and/or irradiation. It is
likely, but not proven, that the majority of patients with a pituitary
mass lesion who are GH-deficient before or after pituitary surgery have
sustained a pituitary insult; in contrast, GHD as a consequence of
irradiation to the hypothalamic-pituitary axis is more likely to be
hypothalamic in origin (216). Thus, provided that the endocrinologist
is comfortable in the knowledge that the patient’s defect is
pituitary-related rather than hypothalamic, the GHRH-pyridostigmine
test can be used to assess GH status.
Anderson et al. (217) have performed a GHRH-pyridostiguine dose response study to determine the optimal dose of pyridostigmine and found it to be 120 mg combined with 1 µg GHRH/kg. They observed a lower reference limit of 21 ng/ml for the peak GH response to the combined test in 40 normal adults, aged between 20 and 60 yr (217); there was no gender difference.
Using the threshold of 21 ng/ml for the peak GH response to the GHRH-pyridostigmine test and an arbitrarily defined level of 10 ng/ml for the ITT, they found consistent classification of normal and subnormal GH responses in 44 of 47 patients with pituitary disease (217).
Perhaps even more promising than GHRH plus pyridostigmine is the combination of GHRH and arginine. Unlike the former combination, the GH response to GHRH and arginine is independent of age; furthermore, the GH response is pronounced and shows much less inter- and intraindividual variability than the more conventional GH provocative tests (104). No overlap was found between GH peak responses in 24 adults with organic GHD and normal subjects (218).
Thus far, the requirement has been only to diagnose severe GHD in adult
life. There is no question, however, that just as in childhood, GH
secretion is a continuum in adult life. Therefore, the concept of GH
insufficiency is a reality. In a current study, GH status was evaluated
in 32 young adults, aged 18 to 33 yr, who previously received cranial
irradiation in childhood as part of their treatment for acute
lymphoblastic leukemia, and 35 age-matched controls (219). Each subject
underwent an ITT and an arginine stimulation test. The median peak GH
responses to both tests were significantly greater in the control group
than in the patient group. Nine patients had a peak GH response to both
tests less than 3.5 ng/ml; 11 patients had a peak GH response less than
7.8 ng/ml to both tests but a peak GH response to one or both tests
greater than 3.5 ng/ml; and the remaining 12 patients had a peak GH
response to one or both tests in excess of 7.8 ng/ml. In contrast, all
the controls bar one (peak GH, 7.4 ng/ml) had a peak GH response
greater than 7.8 ng/ml to one or both tests (219) (Fig. 9).
|
In the future it is likely that the strategy for evaluation of GH status may take into account the degree of GHD, the etiology of the hypothalamic-pituitary disease, and probable site of the d
3rd centile
for height) and their controls (10th to 90th centile for height).
[Reproduced from L. D. Voss et al.: Arch
Dis Child 66:833–837, 1991 (16) with permission from the BMJ
Publishing Group.]
)
and hypopituitary (•) subjects. A, Peak GH response to ITT; B, mean
24-h GH (IGHC); C, IGF-I concentration; D, IGFBP-3 concentration. S,
Assay sensitivity for GH (0.2 ng/ml) and IGF-I (25 ng/ml). [Reproduced
with permission from D. M. Hoffman et al.:
Lancet 343:1064–1068, 1994 (173).]
) GHD or adult-onset (•) GHD stratified into four groups: GHDO,
GHD1, GHD2, and GHD3.
), and 35 age-matched normals
( 
). 1
ng/ml = 2.6 mU/liter. [Reproduced with permission from B.
M. D. Brennan et al.: Clin Endocrinol
(Oxf), in press (219).]