| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
CLINICAL STUDY |
Division of Endocrinology and Metabolic Diseases, Department of Internal Medicine, University of Turin, Corso Dogliotti 14, 10126 Turin, Italy and 1 Division of Endocrinology, Federico II University of Naples, Naples, Italy
(Correspondence should be addressed to G Aimaretti; Email: gianluca.aimaretti{at}unito.it)
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and methodsResultsDiscussionReferences |
|---|
Design and methods: The GH responses to the GHRH-ARG test were studied in 322 patients with organic hypothalamic-pituitary disease and in 318 control subjects. Patients were subdivided into two groups on the basis of the number of pituitary hormone deficits, except for GH deficiency: (a) patients with total pituitary hormone deficit (TPHD) and (b) patients without or with no more than two pituitary hormone deficits (PHD). Both patients and control subjects were divided into three subgroups according to body mass index (BMI): lean (BMI <25 kg/m2), overweight (BMI 
25 and <30 kg/m2) and obese (BMI 30 kg/m2). TPHD patients were assumed to be GH deficient, whereas PHD patients may include subjects with either normal or impaired GH secretion. The statistical analysis was carried out by the Receiver-Operating Characteristic curve analysis (Medcalc 7.2). The diagnostic cut-off points were calculated for lean, overweight and obese subjects to provide optimal separation of GH-deficient patients and control subjects according to two criteria: (1) a balance between high sensitivity and high specificity; (2) to provide the highest pair of sensitivity/specificity values for GH deficiency.
Results: In the lean population the best pair of values, with highest sensitivity as 98.7% and highest specificity as 83.7%, was found using a peak GH cut-off point of 11.5 µg/l. In the overweight population the best pair of values, 96.7 and 75.5%, respectively, was found using a peak GH cut-off point of 8.0 µg/l. In the obese population the best pair of values, 93.5 and 78.3%, respectively, was found using a peak GH cut-off point of 4.2 µg/l. Applying the above mentioned cut-off points, among PHD patients we found that 80 subjects (72%) were GHD whereas 31 (28%) had normal GH secretion.
Conclusions: In conclusion the GHRH-ARG test is a reliable tool for the diagnosis of adult GH deficiency in lean, overweight and obese patients, provided that specific BMI-related cut-off limits are assumed.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
The clinical features of adult GHD are recognizable but not distinctive, so clinical suspicion must be con-firmed by biochemical tests (3, 12–18). Within an appropriate clinical context, GHD in adults may be shown by a single provocative testing, provided that a reproducible test with clear, normative limits is available (1, 18). Insulin-like growth factor-I (IGF-I) and IGF-binding protein-3 (IGFBP-3), as well as spontaneous GH secretion, are not reliable parameters for the diagnosis of adult GHD (1, 16, 18–20), although the measurement of IGF-I differentiates different degrees of GH deficiency (21).
The insulin tolerance test (ITT) has been indicated as the test of choice (1). A GH peak below 5 µg/l has been reported to be an abnormal response to the ITT (1, 12, 18), whereas severe GHD is defined as a GH response lower than the arbitrary cut-off value of 3 µg/l (1). However, the ITT has been challenged because of its low reproducibility and for its contraindications in several clinical conditions often observed in adult patients with suspected GHD, such as ischaemic heart disease and seizure disorders (1, 16, 22, 23). The GH-releasing hormone-arginine (GHRH-ARG) test represents, at present, one of the best alternatives to the ITT to explore the somatotroph function and for the diagnosis of GHD in non-obese adult patients (1, 16, 18, 24). It is a potent, reproducible and reliable test that shows excellent sensitivity and specificity and is unaffected by gender and aging (1, 16, 18, 25). Moreover, we have recently reported that the GHRH-ARG test can show different degrees of GH deficiency (21). The third and the first centile limits of the peak GH response to the GHRH-ARG test evaluated in a population of 157 normal lean subjects were 16.5 and 9 µg/l, respectively (16). However, in clinical practice the reliability of the GHRH-ARG test, as well as that of ITT, is limited by obesity which is a condition characterized by reduced spontaneous GH secretion and low somatotroph response to all provocative stimuli (26–33). Obesity represents the most important confounding factor for the diagnosis of GHD. In obese subjects the GH response to the GHRH-ARG or insulin tests overlaps frequently with that found in adult patients with GHD (26, 28, 30, 31).
Recently, in a comparative study regarding six different provocative tests for the diagnosis of GHD in adults with hypothalamic-pituitary disease, carefully matched with control subjects, Biller et al. (34) confirmed that the greatest diagnostic accuracy was obtained with the ITT and GHRH-ARG tests. The authors defined the specific cut-off points for each test in order to provide the best pair of sensitivity/specificity values. For the GHRH-ARG test they found that a peak GH response cut-off limit of 4.1 µg/l provided 95% sensitivity and 91% specificity. This cut-off is lower than that we found, but control subjects studied by Biller et al. (34) had a higher body mass index (BMI; 30.3±5.8 kg/m2).
It has to be taken into account that nearly half of the patients with acquired hypothalamic-pituitary disease are overweight or obese (3, 4, 7), indicating the clinical need to define BMI-related cut-off limits of the GH response to provocative tests for the diagnosis of adult GHD. Therefore, the aim of our study was to define appropriate diagnostic cut-off limits of peak GH response to the GHRH-ARG test related to BMI.
|
Subjects and methods |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
All patients were studied during adequate hormonal-replacement therapy, where necessary. Patients treated with dopaminoagonists or with a previous history of acromegaly, active Cushing’s disease, diabetes, malignancy, or renal or hepatic dysfunction were excluded from the study. As a control group we studied 318 normal subjects (147 men and 171 women; age 39.9±0.8 years).
Both patients and control subjects were divided into three subgroups according to BMI: lean (BMI <25 kg/m2), overweight (BMI 25 and <30 kg/m2) and obese (BMI 30 kg/m2; Table 1
). Among the patients, 120 (37.3%) were lean, 120 (37.3%) were overweight and 82 (25.5%) were obese. Among the control subjects, 215 (67.6%), 48 (15.1%) and 55 (17.3%) were lean, overweight and obese, respectively. IGF-I levels in patients and controls are reported in Table 1.
|
Methods
After an overnight fast all subjects underwent a GHRH-ARG test (GHRH1-29; GEREF, Serono, Italy; 1 µg/kg i.v. at 0 min; arginine hydrochloride, 0.5 g/kg i.v. over 30 min from 0 to +30 min, up to a maximum of 30 g). Blood samples for GH evaluation were taken every 15 min from 0 to +90 min.
Serum GH was assayed by immunoradiometric method assay (IRMA) method (HGH-CTK IRMA; Diasorin, Saluggia, Italy). All samples from an individual subject were analysed together. The sensitivity of the method was 0.15 µg/l. The inter- and intra-assay coefficients of variation were 3.5–4.4 and 5.1–7.5%, respectively, at GH levels of 1.98–41.92 and 2.99–42.45 µg/l, respectively.
Serum IGF-I was assayed by RIA method (SM-C-RIA-CT; Pantec, Turin, Italy) after acid-ethanol extraction to avoid interference by binding proteins. The sensitivity of the method was 0.1 µg/l. The inter- and intra-assay coefficients of variation were 5.0–9.5 and 8.8–10.8%, respectively, at IGF-I levels of 79.41–712.3 and 79.6–766.4 µg/l, respectively.
Statistical analysis
Results are expressed as mean±S.E.M. of absolute values. Differences between the groups were tested with the Mann–Whitney U test. Correlations were assessed by linear regression analysis.
The lowest limit of normality for peak GH responses to the GHRH-ARG test was defined as the value that provided the best pair of sensitivity/specificity values based on Receiver-Operating Characteristic (ROC) curve analysis (Medcalc 7.2) (35). For the purpose of ROC analysis, we assumed the patients with TPHD to be GHD, according to other studies and to the statement derived from the Growth Hormone Research Society consensus (1, 34, 36, 37). The PHD group (patients without or with no more than two PHD) included subjects with either normal or impaired GH secretion.
ROC curves are constructed by plotting the sensitivity on the ordinate as a function of the complement of specificity for all the possible cut-off values of the diagnostic test. Each point of the ROC curves represents a sensitivity/specificity pair corresponding to a particular decision threshold. The area under the ROC curve (ROC AUC) represents the probability of correctly distinguishing between affected and non-affected individuals. A perfect diagnostic test has an ROC curve that passes through the upper left-hand corner (area under the curve = 1), where the true-positive fraction is 1.0 or 100% (perfect sensitivity) and the false-positive fraction is 0 (perfect specificity). Tests with an area under the curve of less than 0.5 would not discriminate between affected and non-affected subjects (35).
The diagnostic cut-off points were calculated for the lean, overweight and obese groups. To provide optimal separation of GHD and normal subjects we applied two criteria: (1) to minimize misclassification of control subjects and GHD, balancing between high sensitivity and high specificity; (2) to provide the highest pair of sensitivity and specificity values for GHD.
Positive predictive value (PPV), sensitivity and specificity were calculated using the number of patients with true-positive (TP), true-negative (TN), false-positive (FP) and false-negative (FN) results. Sensitivity (calculated as TP/(TP + FN)) and specificity (calculated as TN/(TN + FP)) were defined as the percentages of TPHD patients who had a peak GH value below and above the specific BMI-related cut-off point, respectively . PPV was defined as the likelihood that a subject with a positive test (peak GH below the specific BMI-related cut-off point) was clinically GH-deficient, based on the presence of TPHD (calculated as TP/(TP + FP)).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
). No differences in IGF-I levels were found among lean, overweight and obese subgroups both in patients and in control subjects.
Peak GH responses to GHRH-ARG test in patients and controls related to BMI are shown in Table 2. The peak always occurred between 30 and 60 min. Peak GH responses were significantly higher in lean than in overweight or obese (P < 0.0001) and in overweight versus obese control subjects (P < 0.0001). On the contrary, there was no difference among lean, overweight and obese TPHD and PHD patients. A significant negative correlation (r = –0.5, P < 0.0001) between peak GH responses and BMI was found in controls but not in patients (Fig. 1).
|
|
; Table 3). For a sensitivity of 95%, a cut-off point of 7.9 µg/l yielded a specificity of 86.0% and for a specificity of 95%, a cut-off point of 2.3 µg/l yielded a sensitivity of 54.5% (Table 3).
|
|
; Table 3). For a sensitivity of 95%, a cut-off point of 7 µg/l yielded a specificity of 76.9% and for a specificity of 95%, a cut-off point of 1.5 µg/l yielded a sensitivity of 52.9% (Table 3).
|
; Table 3). For a sensitivity of 95%, a cut-off point of 4.7 µg/l yielded a specificity of 73.5% and for a specificity of 95%, a cut-off point of 1.0 µg/l yielded a sensitivity of 58.7% (Table 3).
|
. Cut-off points for the sensitivity and the specificity at 95% are also reported. Applying the above-mentioned cut-off points that minimize misclassification of TPHD and controls, among the PHD group we found that 80 subjects (72.0%) were GHD whereas 31 (28%) had normal GH secretion; specifically, 38 out of 43 (88.3%), 31 out of 38 (81.5%) and 11 out of 30 (36.6%) were GHD in lean, overweight and obese patients, respectively.
In the 80 PHD patients who were found to be GHD, IGF-I levels were significantly lower than in the 31 without GHD (106.7±5.5 versus 142.4±7.7 µg/l, P < 0.0001). Moreover, a significant negative correlation (r = –0.6, P < 0.0001) between peak GH responses to the GHRH-ARG test and BMI was found in PHD patients without GHD but not in those with GHD. Finally, PHD patients found to be GHD or not had one or two more pituitary hormone deficits with percentages of 77.4 and 48.3%, respectively.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
The GHRH-ARG and GHRH + GHRP-6 tests represent, at present, the best alternatives to ITT to explore the somatotroph function and to differentiate GHD from normal subjects, provided that appropriate cutoff limits are considered (1, 16–18, 24, 41). In agreement with our previous report (42), as peak GH response always occurred between 30 and 60 min, present results confirm that the test requires only three blood samples for GH evaluation (at 30, 45 and 60 min), avoiding basal determination.
In the diagnostic approach for GH deficiency obesity presents a great limitation since it is characterized by a low response to all provocative stimuli for GH secretion (26–33). In obesity reduction in the half-life of GH (43) as well as a significant decrease in the production and secretion of the hormone have been reported (44). Among potential neuroendocrine causes of the marked impairment in spontaneous and stimulated GH secretion, GHRH hypoactivity has been shown but it is likely that alterations in the modulation of ghrelin, neuropeptide Y and/or leptin could have a role (45). Among metabolic factors, chronic elevation of non-esterified fatty acid levels and hyperinsulinism probably have a key role in reducing GH secretion in obesity (31, 46–48).
In this study clear diagnostic cut-off limits of the GH response to the GHRH-ARG test in lean as well as in overweight and obese populations were provided using the statistical method of ROC analysis (35). For each group the highest pair of values for sensitivity/specificity, the 95% sensitivity and the 95% specificity cutoff points were calculated. High sensitivity cut-points maximize detection of adult GHD, whereas high specificity cut-points minimize misclassification of normal subjects as GH-deficient. In clinical practice the cutoff points derived by ROC analysis in order to provide the best pair of values for sensitivity/specificity should be used commonly. Interestingly in our study these cut-off limits were found to have a sensitivity similar to the ideal value of 95%, specifically 98.7, 96.7 and 93.5% in lean, overweight and obese populations, respectively. In agreement with some clinicians (34), the 95% sensitivity cut-off point might be considered the most appropriate if patients have a clinical context compatible with GHD.
Present results show that the cut-off limits of peak GH response to the GHRH-ARG test obtained by ROC curve analysis with the best pair of sensitivity/specificity values to minimize misclassification of GHD and control subjects were 11.5, 8.0 and 4.2 µg/l for the lean, overweight and obese populations, respectively. The cut-off point of 11.5 µg/l for the lean population is intermediate between 16.5 and 9.0 µg/l, which are the third and first centiles, respectively, of peak GH responses recorded in 157 lean normal subjects (16). The cut-off point of 4.2 µg/l that we found in the obese population is comparable with that (4.1 µg/l) reported by Biller et al. (34) in a population with a BMI of 30.3±5.8 kg/m2, consistent with obesity. The diagnostic reliability of the GHRH-ARG test and of these cut-off limits is confirmed by the good accuracy (87.1, 78.6 and 83.9% in the lean, overweight and obese populations, respectively) and by the high values of ROC AUC (0.943, 0.922 and 0.917 in the lean, overweight and obese populations, respectively).
However, it has to be stressed that the above-mentioned cut-off points referred to the GH assay used in this study. In fact, it is well known that when the same serum sample is tested in different assays, there is a wide variability in the absolute values reported (49, 50). Thus, when using other GH assays the above-mentioned cut-off limits for the peak GH response to the GHRH-ARG test would have to be reconsidered.
The significant inverse correlation that we found between peak GH responses to the GHRH-ARG test and BMI in control subjects was similar to that reported recently by Bonert et al. (33). In patients no correlation was found and this finding may be explained by the presence of GH deficiency that reduced the peak GH response to the test in lean as well as in overweight and obese subjects.
IGF-I levels have been reported previously as high, normal or low in obese subjects (28, 33, 51–53). In our study we have not found significant difference in IGF-I levels as a function of BMI. However, as expected, we have found a significant difference in IGF-I levels between patients and controls and between TPHD and PHD patients.
Applying the cut-off points defined in the present study, among the patients without or with no more than two PHD (the PHD group) we found that 81.5 and 36.6% of overweight and obese patients, respectively, had GHD. This kind of diagnosis was difficult before, because the specific cut-off limits of peak GH response to the GHRH-ARG test in overweight and obese populations had not been calculated. Similarly to the difference in IGF-I levels recorded between patients with TPHD (the TPHD group) and controls, IGF-I levels were significantly lower in PHD patients who resulted to be GHD than in those without GHD. Moreover, a significant negative correlation between peak GH responses to the GHRH-ARG test and BMI was found in PHD patients without GHD, as well as in control subjects. Finally, it has to be noted that there was a higher percentage of other PHD in PHD patients with GHD than in those without GHD. All these findings confirm the reliability of the GHRH-ARG test as diagnostic tool for GH deficiency.
In conclusion, as demonstrated in our previous studies in lean subjects, the GHRH-ARG test is also a reliable diagnostic tool in overweight or obese adult patients suspected of GHD, when appropriate cut-off limits related to BMI are considered. The availability of clear BMI-related cut-off limits is mandatory in common clinical practice in view of the fact that many hypopituitaric patients with suspected GHD are overweight or truly obese.
|
Acknowledgements |
|---|
|
References |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  G. Johannsson Central Adiposity as an Important Confounder in the Diagnosis of Adult Growth Hormone Deficiency J. Clin. Endocrinol. Metab., November 1, 2008; 93(11): 4221 - 4223. [Full Text] [PDF]
|
|
|
|
|
|
H. Makimura, T. Stanley, D. Mun, S. M. You, and S. Grinspoon The Effects of Central Adiposity on Growth Hormone (GH) Response to GH-Releasing Hormone-Arginine Stimulation Testing in Men J. Clin. Endocrinol. Metab., November 1, 2008; 93(11): 4254 - 4260. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. L. Utz, A. Yamamoto, L. Hemphill, and K. K. Miller Growth Hormone Deficiency by Growth Hormone Releasing Hormone-Arginine Testing Criteria Predicts Increased Cardiovascular Risk Markers in Normal Young Overweight and Obese Women J. Clin. Endocrinol. Metab., July 1, 2008; 93(7): 2507 - 2514. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. C. Paulo, M. Cosma, C. Soares-Welch, J. N. Bailey, K. L. Mielke, J. M. Miles, C. Y. Bowers, and J. D. Veldhuis Gonadal Status and Body Mass Index Jointly Determine Growth Hormone (GH)-Releasing Hormone/GH-Releasing Peptide Synergy in Healthy Men J. Clin. Endocrinol. Metab., March 1, 2008; 93(3): 944 - 950. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Corneli, C. Di Somma, F. Prodam, J. Bellone, S. Bellone, V. Gasco, R. Baldelli, S. Rovere, H. J. Schneider, L. Gargantini, et al. Cut-off limits of the GH response to GHRH plus arginine test and IGF-I levels for the diagnosis of GH deficiency in late adolescents and young adults Eur. J. Endocrinol., December 1, 2007; 157(6): 701 - 708. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Klose, T. Watt, J. Brennum, and U. Feldt-Rasmussen Posttraumatic Hypopituitarism Is Associated with an Unfavorable Body Composition and Lipid Profile, and Decreased Quality of Life 12 Months after Injury J. Clin. Endocrinol. Metab., October 1, 2007; 92(10): 3861 - 3868. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. di Iorgi, A. Secco, F. Napoli, C. Tinelli, A. Calcagno, N. Fratangeli, L. Ambrosini, A. Rossi, R. Lorini, and M. Maghnie Deterioration of Growth Hormone (GH) Response and Anterior Pituitary Function in Young Adults with Childhood-Onset GH Deficiency and Ectopic Posterior Pituitary: A Two-Year Prospective Follow-Up Study J. Clin. Endocrinol. Metab., October 1, 2007; 92(10): 3875 - 3884. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Chihara, A. Shimatsu, N. Hizuka, T. Tanaka, Y. Seino, Y. Katofor, and the KP-102 Study Group A simple diagnostic test using GH-releasing peptide-2 in adult GH deficiency Eur. J. Endocrinol., July 1, 2007; 157(1): 19 - 27. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Brabant GH releasing peptide 2 test: the holy grail of testing GH deficiency? Eur. J. Endocrinol., July 1, 2007; 157(1): 29 - 30. [Full Text] [PDF]
|
|
|
|
|
|
L. Manetti, I. Lupi, L. L. Morselli, S. Albertini, M. Cosottini, L. Grasso, M. Genovesi, G. Pinna, S. Mariotti, F. Bogazzi, et al. Prevalence and Functional Significance of Antipituitary Antibodies in Patients with Autoimmune and Non-Autoimmune Thyroid Diseases J. Clin. Endocrinol. Metab., June 1, 2007; 92(6): 2176 - 2181. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Pecori Giraldi, M. Andrioli, L. De Marinis, A. Bianchi, A. Giampietro, M. De Martin, E. Sacco, M. Scacchi, A. Pontecorvi, and F. Cavagnini Significant GH deficiency after long-term cure by surgery in adult patients with Cushing's disease Eur. J. Endocrinol., February 1, 2007; 156(2): 233 - 239. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H J Schneider, B L Herrmann, M Schneider, C Sievers, L Schaaf, and G K Stalla Discrepant results in the diagnosis of GH deficiency with the insulin-tolerance test and the GHRH plus arginine test in patients with traumatic brain injury. Eur. J. Endocrinol., October 1, 2006; 155(4): 553 - 557. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H J Schneider, S Rovere, G Corneli, C G Croce, V Gasco, R Ruda, S Grottoli, G K Stalla, R Soffietti, E Ghigo, et al. Endocrine dysfunction in patients operated on for non-pituitary intracranial tumors. Eur. J. Endocrinol., October 1, 2006; 155(4): 559 - 566. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Bondanelli, M. R. Ambrosio, A. Onofri, A. Bergonzoni, S. Lavezzi, M. C. Zatelli, D. Valle, N. Basaglia, and E. C. degli Uberti Predictive Value of Circulating Insulin-Like Growth Factor I Levels in Ischemic Stroke Outcome J. Clin. Endocrinol. Metab., October 1, 2006; 91(10): 3928 - 3934. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A A van der Klaauw, A M Pereira, S W van Thiel, J W A Smit, E P M Corssmit, N R Biermasz, M Frolich, A Iranmanesh, J D Veldhuis, F Roelfsema, et al. GH deficiency in patients irradiated for acromegaly: significance of GH stimulatory tests in relation to the 24 h GH secretion. Eur. J. Endocrinol., June 1, 2006; 154(6): 851 - 858. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Colao, C. Di Somma, S. Spiezia, F. Rota, R. Pivonello, S. Savastano, and G. Lombardi The Natural History of Partial Growth Hormone Deficiency in Adults: A Prospective Study on the Cardiovascular Risk and Atherosclerosis J. Clin. Endocrinol. Metab., June 1, 2006; 91(6): 2191 - 2200. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. E. Engstrom, P. Burman, C. Holdstock, M. Ohrvall, M. Sundbom, and F A. Karlsson Effects of gastric bypass on the GH/IGF-I axis in severe obesity - and a comparison with GH deficiency Eur. J. Endocrinol., January 1, 2006; 154(1): 53 - 59. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||
