Download GH Values after Clonidine Stimulation Measured

Original Paper
Received: March 1, 2002
Accepted after revision: November 20, 2002
Horm Res 2003;59:229–233
DOI: 10.1159/000070222
GH Values after Clonidine Stimulation
Measured by Immunofluorometric Assay in
Normal Prepubertal Children and GH-Deficient
Patients
Eveline G.P. Silva Natasha Slhessarenko Ivo J.P. Arnhold
Marcelo C. Batista Vivian Estefan Maria G.F. Osorio Suemi Marui
Berenice B. Mendonca
Unidade de Endocrinologia do Desenvolvimento e Laboratório de Hormônios e Genética Molecular LIM /42,
Hospital das Clı́nicas da Faculdade de Medicina da Universidade de São Paulo, Brazil
Key Words
GH peak W Clonidine W Immunofluorometric assay
Abstract
Objective: To establish the cut-off values of GH measured by immunofluorometric assay, a more sensitive
and specific assay, in normal prepubertal children and
compare their values with those of proven GH-deficient
patients. Methods: 30 normal children (20 males) and 26
patients with known causes of GH deficiency were submitted to the clonidine test and their GH values were
compared. A powdered clonidine tablet (0.1 mg/m2) was
given orally and blood samples for GH measurements
were drawn at times –30, 0, 60, 90 and 120 min. Results:
GH peak values presented a wide variation ranging from
1.7 to 25 Ìg/l (mean B SD = 12.87 B 5.8 Ìg/l) in the normal group. The cut-off values for the 5th and 10th percentiles of the distribution curve were 3.3 and 5.5 Ìg/l,
respectively. In the GH deficiency group, maximum GH
levels after clonidine stimulation ranged from ! 0.1 to
2.1 Ìg/l (0.56 B 0.58 Ìg/l). Conclusions: The cut-off values obtained with the immunofluorometric method are
lower than the ones obtained by radioimmunoassay. We
ABC
© 2003 S. Karger AG, Basel
0301–0163/03/0595–0229$19.50/0
Fax + 41 61 306 12 34
E-Mail [email protected]
www.karger.com
Accessible online at:
www.karger.com/hre
suggest a cut-off value of 3.3 Ìg/l (5th percentile) that
ensures 100% of sensitivity along with 93% of specificity
to exclude the diagnosis of GH deficiency when using
this immunofluorometric method.
Copyright © 2003 S. Karger AG, Basel
Introduction
Classically, GH deficiency (GHD) is diagnosed biochemically after at least two stimulation tests, with GH
values !7 or !10 Ìg/l measured by radioimmunoassay
(RIA) using polyclonal antibodies [1–12]. The development of monoclonal antibodies, non-isotopic assays and
the use of more accurate reference standards contributed
to the large variability of the results obtained by different
assays [13, 14]. Therefore, each laboratory must establish
its own cut-off values for each method employed [15,
16].
We measured GH by immunofluorometric assay
(IFMA) after clonidine stimulation test in 30 normal prepubertal children in order to establish cut-off values for a
normal GH response and applied these values in the diagnosis of GHD.
Berenice B. Mendonca, MD
Disciplina de Endocrinologia, Hospital das Clı́nicas da Universidade de São Paulo
Av Dr Eneas de Carvalho Aguiar, 155, 2 andar, Bloco 6, PAMB
São Paulo, SP 05403-900 (Brazil)
Tel. +55 11 3062 6654, Fax +55 11 3083 0626, E-Mail [email protected]
Table 1. Clinical features of normal and
GH-deficient subjects
Groups
n
CA
years
Gender
Height (SD)
WHI, %
Normal prepubertal
30
2.5–10.81
20 M/10 F
GH-deficient
26
1.8–22.5
16 M/10 F
–0.57B0.771
(–1.9 to 0.82)
–4.6B1.42
(–7.7 to –2.0)
96.4B7.72
85U117
100B14
(82U130)
CA = Chronological age; WHI = weight for height index = weight/weight for height age
!100.
1
Mean B SD (range in parentheses).
Subjects and Methods
The present study was approved by the Ethics Committee of Hospital das Clı́nicas da Universidade de São Paulo and the samples
were collected after written consent was obtained from parents or
tutors. The clinical data of all individuals are shown in table 1. The
control group consisted of 30 healthy prepubertal children with normal height (within mean B 2 SD) for chronological age and appropriate weight for height age (WHI: 80–120). All GH-deficient
patients were also at prepubertal Tanner stage. All patients had congenital GHD: 12 had isolated GHD and 14 had multiple hypothalamic-pituitary hormone deficiencies. Eight patients had molecular
defects in the GH, GHRHR or PROP1 genes and the others had an
ectopic posterior pituitary lobe, or pituitary stalk interruption on
MRI. The clonidine stimulation test consisted in the collection of two
basal blood samples and 60, 90 and 120 min after the oral intake of
powdered clonidine tablets (0.1 mg/m2) dissolved in 5 ml of water to
avoid difficulties in clonidine absorption. The test was performed
when thyroid function was normal and was well tolerated. All
patients presented drowsiness 60 min after clonidine intake and
slight fall in blood pressure was also observed. All but 5 GH-deficient
subjects were also submitted to insulin tolerance test (ITT). The ITT
consisted in the collection of two basal blood samples and 15, 30, 45,
60 and 90 min after an intravenous injection of insulin (0.05–0.1U/
kg). All patients presented clinical signs of hypoglycemia confirmed
by a glucose level ! 40 mg/dl.
GH was measured by immunofluorometric kits (Delfia® kit, Wallac, Turku, Finland) based on the direct sandwich technique using
two monoclonal antibodies specific for the 22-kDa form. The assays
were standardized with the international standard 80/505 of the
World Health Organization. The detection limit of this assay was
0.01 Ìg/l and the intra-assay coefficient of variation was lower than
6% in the working range of the standard curve (0.1–36 Ìg/l).
IGF-I was measured in plasma after extraction with ethanol
87.5% and HCl 2 N 12.5%, by RIA, using kits from Nichols Institute
(San Juan Capistrano, Calif., USA) in all individuals. Cross-reaction
with IGF-II and proinsulin was referred to be lower than 2 and 0.1%,
respectively. IGFBP-3 was measured in serum by immunoradiometric assays using kits from the Diagnostic Systems Laboratories Inc.
(DSL, Webster, Tex., USA). No cross-reaction was referred with other binding proteins (IGFBP-1, 2 and 4). The inter-assay variation, as
defined by repeated measurements of serum controls run in different
assays, was ! 10%, for IGF-I levels from 18 to 500 ng/ml and for
IGFBP-3 levels from 0.14 to 12 mg/l.
230
Horm Res 2003;59:229–233
The results were analyzed by the Statistics Program for Windows
(version 4.0, Statsoft Inc. 1993). The normality hypothesis was tested
over the distribution of peak GH values after stimulation, using the
Shapiro-Wilk test. The means of IGF-I and IGFBP-3 and GH peak
values obtained from both groups were compared by Student’s t test,
and p values ! 0.05 were considered statistically significant. GH values ! 10 ng/ml by IRMA were compared to the ones obtained by
IFMA, using Spearman’s correlation test. The cut-off values for GH
by immunofluorometric method were taken at the 2.5th, 5th and
10th percentiles of the distribution curve of GH values after clonidine stimulation in normal subjects. The probabilities of false-positive tests for the classic cut-off values in RIA, 7 or 10 Ìg/l, were also
analyzed on these curves. The cut-off values for basal IGF-I and
IGFBP-3 were taken at –1 SD of the mean values and the analysis of
their diagnostic properties was also performed.
Results
GH values after the clonidine test in the control group
are shown as mean, SD, and minimum-maximum values. We observed a wide variation in GH response to clonidine and the lowest GH response in this group was
1.7 Ìg/l (table 2).
In the GH-deficient group, the highest GH peak value
after clonidine was 2.1 Ìg/l. GH peak value after ITT (n =
21) was not significantly different from GH peak values
after clonidine, 0.32 B 0.34 vs. 0.56 B 0.58 Ìg/l, p = 0.09,
respectively.
The Shapiro-Wilk normality test (W) showed a normal
distribution of GH values after stimulation in the control
group (W = 0.98; p = 0.95). The cut-off values obtained at
the 2.5th, 5th and 10th percentiles of the distribution
curve of the control group of 30 subjects were 1.5, 3.3 and
5.5 Ìg/l, respectively. The characteristics of the clonidine
test for the diagnosis of GHD were calculated with different cut-off values and are presented in table 3. If the current RIA cut-off values of 7 and 10 Ìg/l were used, it
would imply in the diagnosis of GHD in 16.7 and 33.3%
Silva/Slhessarenko/Arnhold/Batista/
Estefan/Osorio/Marui/Mendonca
Table 2. Basal IGF-I, IGFBP-3 and GH peak values after oral clonidine (0.1 mg/m2) stimulation test in normal and
GH-deficient subjects
Groups
Normal prepubertal
GH-deficient
IGF-I, ng/ml
IGFBP-3, mg/l
GH peak, Ìg/l
mean B SD
mean B SD
mean B SD
range
116B68
24.5B14*
3.66B1.06
0.97B0.61*
12.9B5.8
0.56B0.58**
1.7–25
!0.1–2.1
5th centile
10th centile
3.3
5.5
* p ! 0.001; ** p ! 0.05 GH-deficient vs. normal subjects.
Table 3. GH response to clonidine test data
for the diagnosis of GHD by IFMA using
several cut-off levels
GH cut-off level
Ìg/l (centile)
Sensitivity
%
Specificity
%
1.5 (2.5th)
3.3 (5th)
5.5 (10th)
7.01
101
85
100
100
100
100
100
93.3
90
83.3
66.7
1
Table 4. Basal IGF-I and IGFBP-3 levels
for the diagnosis of GHD using –1 SD as
the cut-off levels
FPR
%
FNR
%
Accuracy
%
15
0
0
0
0
93
96
95
91
82
FPR
%
FNR
%
Accuracy
%
6.7
13.3
0
3.8
0
0
0
6.7
10
16.7
33.3
Classical cut-off values.
FPR = False positive rate; FNR = false negative rate.
Test
IGF-I
IGFBP-3
IGF-I + IGFBP-3
Sensitivity
%
96.2
100
100
Specificity
%
93.3
86.7
100
94.6
92.4
100
FPR = False positive rate; FNR = false negative rate.
of the normal individuals, respectively. But if we used the
5th and 10th percentiles of the distribution curve as cutoff values, only 6.7 and 10% of the normal individuals
would be classified as GH-deficient.
Basal IGF-I and IGFBP-3 values are presented in
tables 2 and 4. IGF-1 and IGFBP-3 levels in the GH-deficient group were lower than in the normal one (p ! 0.05).
We established as cut-off value –1 SD of the mean
obtained from the control group for the diagnosis of GHD
(table 4). IGF-I values were low in 2 subjects from the
control group (6.7%) and none had low IGFBP-3, whereas
in the GH-deficient group, 3 subjects presented IGFBP-3
values higher than –1 SD, and 1 presented IGF-I higher
than –1 SD.
GH Values after Clonidine Stimulation
Discussion
Presently, several types of assays for GH measurement
are available: RIA, immunoradiometric assay (IRMA),
enzyme-immunoassay assay (ELISA) chemoluminescence assay and IFMA [17–22]. The presence of GHbinding proteins [23, 24], the heterogeneity of the circulating GH molecules [14, 24, 25] as well as the different
types of antibodies, either monoclonal or polyclonal,
could account for the variability observed in the results
obtained by these methods.
The diagnosis of GHD during childhood and adolescence has been the subject of much controversy [26–28].
To ensure that patients are appropriately identified and
treated, the GH Research Society (GRS) held a workshop
Horm Res 2003;59:229–233
231
on October 17–21 1999, in Eilat, Israel [26]. One of the
conclusions of this workshop was the necessity to establish normative data for GH values using newer monoclonal-based assays and recombinant hGH reference preparations. However, there are limited normative data for
such GH assays and more data in normal children are
desirable within ethical guidelines.
We used an IFMA with monoclonal antibodies specific
for the 22-kDa GH form. This method has several advantages, such as being non-radioactive, requiring only small
amounts of serum, high sensitivity, accuracy and precision. In a preliminary study performed at our laboratory,
Marui et al. [29] compared GH response curves after
stimulation (hypoglycemia and clonidine) by the IRMA
and IFMA methods, in 9 children with short stature (–1.2
to –3.7 SD for height) and/or slow growth velocity. A significant positive correlation between the two methods (r =
0.899, p! 0.001) was observed, although GH peak values
measured by the IFMA method (mean B SD, 9.4 B
4.3 Ìg/l) were lower than the ones for the IRMA method
(17 B 8 Ìg/l). However, for the same IFMA values, there
were different IRMA values, suggesting that the use of the
mathematical correlation (IRMA = 0.575 + 1.58 W IFMA)
would not always represent the actual GH value.
The cut-off level of GH measured by RIA, used in the
diagnosis of GHD, has always been determined arbitrarily as 7 and 10 Ìg/l. If we had used the current GH cut-off
values in our normal individuals analyzed by IFMA, it
would have implied in the diagnosis of GHD in 17 and
33% of them, respectively.
Choosing a cut-off value for a test depends on the
knowledge of the distribution curve obtained from normal individuals and patients. According to the basic principles for the test selection, the ideal test is the one that
does not overlap results from patients and normal individuals [30]. For the majority of the tests, however, there
is some overlapping between these two groups. Statistically, the normal interval comprehends the distribution of
95% of the results and is defined by mean B 2 SD. Therefore, 2.5% of the normal individuals can be found above
or below this interval. Choosing a point that presents high
specificity, but limited sensitivity, is useful when one
wants to confirm a suspected diagnosis, but not as a
screening test to rule out a disease. The most appropriate
cut-off value must present high sensitivity with a reasonable specificity in order to be used as a screening test. It is
difficult to establish a cut-off value with these characteristics for the diagnosis of GHD, since short stature due to
GHD has a spectrum of clinical manifestations that can
vary from partial or transitory deficiency to complete hor-
232
Horm Res 2003;59:229–233
mone deficiency. When analyzing the different cut-off
values obtained from the distribution of GH values in
normal subjects (1.5, 3.3 and 5.5 Ìg/l), and the ones classically described in literature, we observed a better performance or accuracy when the values 3.3 and 5.5 Ìg/l were
used. Our cases with severe GHD were clearly diagnosed
using the cut-off levels at the 5th percentile (3.3 Ìg/l) of
the normal group. However, it is possible that patients
with partial GHD cannot be identified by this cut-off.
Therefore, children with values between 3.3 and 5.5 Ìg/l
should be submitted to a second test, mainly those suspected to have partial GHD.
In our GH-deficient group, 86.7% of the subjects presented IGFBP-3 values below –1 SD and 96% presented
IGF-I below –1 SD. Analyzing together the normal and
GH-deficient subjects, the association of IGF-1 and
IGFBP-3 levels below –1 SD was found only in the GHdeficient group ensuring a 100% of sensitivity along with
100% of specificity for the diagnosis of GHD (table 4).
In a recent study, Attanasio et al. [31] re-evaluated
adult-age GH status in 167 patients with GHD diagnosed
in childhood. They performed a provocative GH stimulation test (cut-off peak level !3 Ìg/l) and measured basal
IGF-I and IGBP-3 (cut-off –2 SDS values using a normal
population as reference). They verified that GH peak
after stimulation and basal IGF-1 levels presented similar
sensitivity (80%) for GHD diagnosis. When IGF-I values
and GH response to the test were discordant, IGFBP-3
levels were not useful to establish a diagnosis.
Thus, the operational characteristics of a test vary
according to the criterion used to define what a normal
value is. Therefore, it is important to know such characteristics, not only in order to promote the adequate interpretation of the results, but also to decide which test
should be used for a given objective.
In conclusion, we determined the normal GH peak values after clonidine stimulation in normal prepubertal
children using an IFMA and suggest the use of a cut-off
value of 3.3 Ìg/l, which ensures 100% of sensitivity along
with 93% of specificity to exclude the diagnosis of GHD
by this method.
Acknowledgements
The authors thank the staff of Laboratório de Hormônios e
Genética Molecular LIM/42, particularly Cassia Regina Mazzi, Marcia Ester Paiva Ferreira and Valeria Samuel Lando for technical
assistance and Sonia Strong for the English review. We also thank the
staff from the Outpatient Clinic of HCFMUSP for performing the
tests and CRIESP Laboratory for performing part of IGFBP-3 measurements.
Silva/Slhessarenko/Arnhold/Batista/
Estefan/Osorio/Marui/Mendonca
References
1 Roth J, Glick SM, Yalow RS, Berson AS: Hypoglycemia: A potent stimulus to secretion of
growth hormone. Science 1963;140:987–988.
2 Kaplan SL, Abrams CAL, Bell JJ, Conte FA,
Grumbach MM: Serum growth hormone response to insulin-induced hypoglycemia in disorders of growth. Soc Pediatr Res 1965;67:
956–959.
3 Stimmler L, Brown GA: Growth hormone secretion provoked by insulin-induced hypoglycaemia in children of short stature. Arch Dis
Child 1967;42:232–238.
4 Kaplan SL, Abrams CAL, Bell JJ, Conte FA,
Grumbach MM: Growth and growth hormone.
I. Changes in serum level of growth hormone
following hypoglycemia in 134 children with
growth retardation. Pediatr Res 1968;2:43–63.
5 Youlton R, Kaplan SL, Grumbach MM:
Growth and growth hormone. IV. Limitations
of the growth hormone response to insulin and
arginine and of the immunoreactive insulin
response to arginine in the assessment of
growth hormone deficiency in children. Pediatrics 1969;43:989–1004.
6 Frasier SD: A review of growth hormone stimulation tests in children. Pediatrics 1974;53:
929–937.
7 Gil-Ad I, Topper E, Laron Z: Oral clonidine as
a growth hormone stimulation test. Lancet
1979;ii:278–180.
8 Ghigo E, Imperiale E, Boffano GM, Mazza E,
Bellone J, Arvat E, Procopio M, Goffi S, Barreca A, Chiabotto P: A new test for the diagnosis
of growth hormone deficiency due to primary
impairment: Combined administration of pyridostigmine and growth hormone-releasing hormone. J Endocrinol Invest 1990;13:307–316.
9 Brook CGD, Hindmarsh PC: Tests for growth
hormone secretion. Arch Dis Child 1991;66:
85–87.
10 Dattani MT, Pringle PJ, Hindmarsh PC, Brook
CGD: What is a normal stimulated growth hormone concentration? J Endocrinol 1992;133:
447–450.
11 Menon PSN, Gupta P, Karmarkar MG: Highand low-dose clonidine tests for the diagnosis
of growth hormone deficiency. Indian Pediatr
1994;31:145–151.
GH Values after Clonidine Stimulation
12 Rosenfeld RG, Albertson-Wikland K, Cassorla
F, Frasier SD, Hasegawa Y, Hintz RL, Lafranchi S, Lippe B, Loriaux L, Melmed S: Diagnostic controversy: The diagnosis of childhood
growth hormone deficiency revisited. J Clin
Endocrinol Metab 1995;80:1532–1540.
13 Kahan I, Papanastasiou-Diamandi A, Ellis G,
Makela SK, McLaurin J, D’Costa M, Diamandis EP: Sensitive time-resolved fluorescence
immunoassay of somatototropin in serum. Clin
Chem 1990;36:503–508.
14 Albertsson-Wikland K, Jansson C, Rosberg S;
Novamo A: Time-resolved immunofluorometric assay of human growth hormone. Clin
Chem 1993;39:1620–1625.
15 Reiter EO, Morris AH, MacGillivray MH, Weber D: Variable estimates of serum growth hormone concentrations by different radioassay
systems. J Clin Endocrinol Metab 1988;66:68–
71.
16 Chevenne D, Beau N, Leger J, Porquet D: Variability of serum human growth hormone levels
in different commercial assays: Specificity of
growth hormone-releasing hormone stimulation. Horm Res 1993;40:168–172.
17 Banfi G, Marinelli M, Casari E, Murone M,
Bonini P: Isotopic and non-isotopic assays for
measuring somatotropin compared: Re-evaluation of cut-off value in provocative tests. Clin
Chem 1991;37:273–276.
18 Chatelain P, Bouillat B, Cohen R, Sassolas G,
Souberbielle JC, Ruitton A Joly MO, Job JC:
Assay of growth hormone levels in human plasma using commercial kits: Analysis of some
factors influencing the results. Acta Paediatr
Scand 1990;370(suppl):56–61.
19 Ghigo E, Bellone J, Aimaretti G, Bellone S,
Loche S, Cappa M, Bartolotta E, Dammacco F,
Camanni F: Reliability of provocative tests to
assess growth hormone secretory status. Study
in 472 normally growing children. J Clin Endocrinol Metab 1996;81:3323–3327.
20 Aimaretti G, Corneli G, Razzore P, Bellone S,
Baffoni C, Arvat E, Camanni F, Ghigo E: Comparison between insulin-induced hypoglycemia
and GH-releasing hormone + arginine as provocative tests for the diagnosis of GH deficiency in adults. J Clin Endocrinol Metab 1998;83:
1615–1618.
21 Strasburger CJ, Dattani MT: New growth hormone assays: Potential benefits. Acta Paediatr
Suppl 1997;423:5–11.
22 Aimaretti G, Baffoni C, DiVito L, Bellone S,
Grottoli S, Maccario M, Arvat E, Camanni F,
Ghigo E: Comparisons among old and new provocative tests of GH secretion in 178 normal
adults. Eur J Endocrinol 2000;142:347–352.
23 Jan T, Shaw MA, Baumann G: Effects of
growth hormone-binding proteins on serum
growth hormone measurements. J Clin Endocrinol Metab 1991;72:387–391.
24 Baumann G: Growth hormone binding proteins and various forms of growth hormone:
Implications for measurements. Acta Paediatr
Scand 1990;370(suppl):72–80.
25 Lewis UJ, Sinha YN, Haro LS: Variant forms
and fragments of human growth hormone in
serum. Acta Paediatr 1994;399(suppl):29–31.
26 Hintz RL: Consensus guidelines for the diagnosis and treatment of GH deficiency in childhood and adolescence: Summary statement of
the GH Research Society. J Clin Endocrinol
Metab 2000;85:3990–3993.
27 Drug and Therapeutics Committee of the Lawson Wilkins Pediatric Endocrine Society:
Guidelines for the use of growth hormone in
children with short stature: A report by the
Drug and Therapeutics Committee of the Lawson Wilkins Pediatric Endocrine Society. J Pediatr 1995;127:857–867.
28 American Academy of Pediatrics: Considerations related to the use of recombinant human
growth hormone in children. Pediatrics 1997;
99:122–129.
29 Marui S, Lemos MM, Cassina C, Osorio GF,
Oliveira SB, Silva EGP, Batista MC, Nicolau
W, Arnhold IJP, Mendonça BB: Comparison
of growth hormone levels measured by immunoradiometric and immunofluorimetric assays
in children with short stature. Pediatr Res
1994;36:676.
30 Griner PF, Mayewski IRJ, Mushlin AI, Greesland P: Selection and interpretation of diagnostic tools and procedures: Principals and applications. Ann Intern Med 1981;94:553–600.
31 Attanasio AF, Howell S, Bates PC, Blum WF,
Frewer P, Quigley C, Shalet SM: Confirmation
of severe GH deficiency after final height in
patients diagnosed as GH deficient during
childhood. Clin Endocrinol (Oxf) 2002;56:
503–507.
Horm Res 2003;59:229–233
233