IT IS WIDELY recognized that the diagnosis of GH deficiency

0021-972X/97/$03.00/0 Vol. 82, No. 2 Journal of Clinical Endocrinology and Metabolism Printed in U.S.A. Copyright 1997 by The Endocrine Society Biochemical s in the Diagnosis of Childhood Growth Hormone Deficiency* VALLO TILLMANN, JOHN M. H. BUCKLER, MOHAMMED S. KIBIRIGE, DAVID A. PRICE, STEPHEN M. SHALET, JEREMY K. H. WALES, MICHAEL G. ADDISON, MATHEW S. GILL, ANDY J. WHATMORE, AND PETER E. CLAYTON Royal Manchester Children s Hospital, Christie Hospital (S.M.S.), Manchester; Sheffield Children s Hospital (J.K.W.), Sheffield; General Infirmary (J.M.B.), Leeds; and South Cleveland Hospital (M.S.K.), Middlesbrough, United Kingdom ABSTRACT GH stimulation tests are widely used in the diagnosis of GH deficiency (), although they are associated with a high false positive rate. We have examined, therefore, the performance of other tests of the GH axis [urinary GH excretion, serum insulin-like growth factor I (IGF-I), and IGF-binding protein-3 (IGFBP-3) levels] compared with GH stimulation tests in identifying children defined clinically as GH deficient. Group I comprised 60 children (mean age, 10.3 4.8 yr) whose diagnosis of was based on a medical history indicative of pituitary dysfunction (n 43) or on the typical phenotypic features and appropriate auxological characteristics of isolated (n 17). Group II comprised 110 short children (mean age, 9.8 4 yr) in whom was not suspected, but needed exclusion. The best sensitivity for a single GH test was 85% at a peak GH cut-off level of 10 ng/ml, whereas the best specificity was 92% at 5 IT IS WIDELY recognized that the diagnosis of GH deficiency () in childhood can be difficult. Even in specialist growth centers, 25% of children treated for are found to have a normal peak GH concentration during a stimulation test, when evaluated at the end of growth (1, 2). This can occur despite the application of the classical criteria for at the time of diagnosis and the exclusion of other causes of growth failure. It is no surprise that the issue of the diagnosis of childhood has been revisited (3). Investigators have sought therefore to find other tests of the GH- IGF axis that could be used in diagnosis. Serum insulin-like growth factor I (IGF-I) levels are low in (4), but can overlap levels found in short normal and normal children (5 7). Hence, the specificity of this test in the confirmation of is relatively low (54%) (8). IGFBP-3, the major serum carrier protein for circulating IGFs, is regulated by GH and has been reported to have an excellent sensitivity and specificity in the diagnosis of (8). However, in other studies the performance of IGFBP-3 in this context has been poor (9, 10). Received April 3, 1996. Revision received September 18, 1996. Accepted September 24, 1996. Address all correspondence and requests for reprints to: Dr. Peter E. Clayton, Department of Endocrinology, Royal Manchester Children s Hospital, Manchester, M27 4HA, UK. * The work, M.S.G., and A.J.W. were supported by grants from Serono UK. V.T. is a training fellow supported by the ESPE. ng/ml. The sensitivities of IGF-I, IGFBP-3, and urinary GH, using a cut-off of 2 SD score were poor at 34%, 22%, and 25%, respectively, with specificities of 72%, 92%, and 76% respectively. Only 2 of 21 pubertal children in group I and none of the 27 subjects with radiation-induced had an IGFBP-3 SD score less than 1.5. We devised a scoring system based on the positive predictive value of each test, incorporating data from the GH test and the IGF-I and IGFBP-3 levels. A specificity of 94% could be achieved with a score of 10 or more (maximum 17) (sensitivity 34%). The latter could not be improved above 81% with a score of 5 points or more (specificity, 69%). A high score was, therefore, highly indicative of, but was achieved by few patients. A normal IGFBP-3 level, however, did not exclude, particularly in patients with radiation-induced and those in puberty. A GH test with a peak level more than 10 ng/ml was the most useful single investigation to exclude a diagnosis of. (J Clin Endocrinol Metab 82: 531 535, 1997) It has been recognized that clinical evaluation is the most valuable aspect of assessment in growth disorders (3). Nevertheless, most clinicians will undertake tests of the GH axis. This study was designed, therefore, to compare the performance of GH stimulation tests with that of a single serum IGF-I, IGF-binding protein-3 (IGFBP-3), and urinary GH (ugh) level, using the clinical diagnosis of the growth disorder as the base standard. Patients Subjects and Methods Two hundred and one patients (mean age, 9.9 4.2 yr; 117 boys and 84 girls) were recruited to the study on the basis that a GH test was to be carried out for the evaluation of growth failure (n 187) or reassessment of pituitary function at the end of growth or GH treatment (n 14). A clinical diagnosis of the presence or absence of was made in each case by the physicians responsible for the patient s medical care. All had long-standing experience in the diagnosis and treatment of. Sixty patients were diagnosed clinically as GH deficient (group I; Table 1). In 43 patients this was based in part on medical history [central nervous system (CNS) irradiation, tumors of the hypothalamicpituitary region, and organic hypopituitarism]. In the remaining 17 children, a diagnosis of isolated idiopathic was made on the basis of 1) the presence of more than one typical phenotypic feature (frontal bossing, immature face, midfacial hypoplasia, truncal adiposity, fat dimpling, hypogenitalism in a male, and high-pitched voice), 2) appropriate auxological characteristics (height sd score 2, height velocity below the 10th percentile over 6 months, and a bone age delayed by 2 yr), and 3) the exclusion of other endocrinopathies and chronic disease. A subset of group I (n 19), in which the diagnosis of was supported 531

532 TILLMANN ET AL. JCE&M 1997 Vol 82 No 2 TABLE 1. Clinical diagnoses of subjects in groups I and II Group Diagnosis No. Group I (; n 60) Idiopathic 17 Septo-optic dysplasia 3 Congenital hypopituitarism 2 Tumours of the hypothalamicpituitary 8 area Radiation-induced Cranial tumours outside 12 the hypothalamicpituitary area Malignancy outside the 15 cranium Histiocytosis 3 Group II Constitutional delay in 28 (non-; n 110) growth and puberty Familial short stature 12 Intrauterine growth 2 retardation Clinically defined syndromes 18 Disorders in specific systems 22 (e.g. chronic anaemia) Idiopathic short stature 28 by specific hypothalamic-pituitary pathology, was also identified. This included the children with septooptic dysplasia, congenital hypopituitarism, tumors of the hypothalamic-pituitary area, and 6 with isolated who had abnormal findings (empty sella or stalk lesion) on computed tomography or magnetic resonance imaging scan. Group II comprised 110 short patients, in whom was not suspected by their physicians, but needed to be excluded formally (Table 1). The diagnoses of patients in groups I and II were based on the categories defined in the Kabi Pharmacia International Growth Study (11). In the remaining 31 cases, the physicians were unable to assign a definitive classification of GH status. Data from these children have been excluded from the analysis of test performances. All children had a GH stimulation test (arginine 90; glucagon 65; clonidine 38; insulin 6; exercise 2) carried out in a standard manner. If a child had two GH tests, the results of the first test were used in analysis. Sex steroid priming before GH tests was carried out in three of the five centers in children over 9 yr of age who were prepubertal or in early puberty (n 24). Fifty-four other children in this category were not primed. Serum IGF-I and IGFBP-3 concentrations were measured in a single sample taken before the GH stimulation test, whereas ugh was measured in an overnight urine sample collected on that morning. Height sd score was calculated from the new British standards (12), and height velocity was compared to Tanner standards (13). Body mass index (BMI) was calculated as weight (kilograms)/height (meters 2 ). Bone age was estimated by the method of Greulich and Pyle (14). Assays GH was measured at four different laboratories using two-site immunoradiometric assays with standard commercial kits [Netria (London, UK) and IDS (Tyne and Wear, UK)]. All laboratories participated in the UK National External Quality Assurance Scheme, and each assay conformed to acceptable standards, being calibrated against the International Standard 80/505: the mean bias for the four assays was 2.1% (range, 9.9% to 9.3%). Serum IGF-I, after acid ethanol extraction, and IGFBP-3 were measured by RIA using standard commercial kits (Nichols Institute Diagnostics, Somerville, NJ). The intraassay coefficient of variation ranged from 2.4 3.0%, and the interassay coefficient of variation ranged from 4 5.2% for IGF-I; these values were 3.4 8.0% and 5.3 6.3%, respectively, for IGFBP-3. Results were expressed as the sd score, using the normal ranges, standardized by age and sex, provided by Nichols Institute Diagnostics. ugh excretion in overnight urine was measured by two-site immunoradiometric assay, described previously (15), and the results were expressed as the sd score, using our own standardized normal ranges (15). Statistical analysis We have used clinical assessment of GH status to define groups I () and II (non-). performance is based, therefore, on this classification. Sensitivity was defined as the number of true positive results (below the cut-off point) divided by the total number of results in group I. Specificity was defined as the number of true negative results (above the cut-off point) divided by the total number of results in group II. Efficiency was defined as the number of correct results divided by the total number of tests in both groups. The positive predictive value was defined as the number of positive results in the group divided by the total number of positive results. All values were expressed as a percentage. Student s t test and the Kruskal-Wallis test were used to compare data between groups. Results Patient characteristics The subjects fell into a broad range of diagnostic categories, illustrating the numerous etiologies associated with growth failure, where evaluation of the GH axis may be required (Table 1). Patients with isolated idiopathic could not be distinguished from those in group II by age (mean sd, 7.9 3.4 vs. 9.6 3.8 yr), height sd score ( 2.7 1.0 vs. 2.6 1.1), height velocity (4.2 1.6 vs. 4.3 1.8 cm/yr), body mass index (16.5 3.1 vs. 16.9 4.0), or bone age delay ( 1.5 1.3 vs. 1.8 1.4). performance The efficiency, sensitivity, and specificity of all tests at defined cut-off points are shown in Table 2. GH stimulation tests. The mean peak GH concentration in group I was significantly less than the concentration in group II (Table 3), but there was a broad range of responses in both groups. The mean peak GH concentration in children who had a primed GH test (n 24) did not differ significantly TABLE 2a. Sensitivity, specificity, and efficiency (expressed as number of patients and as percentage of patients) of each test at the given cut-off point for all patients Sensitivity Specificity Efficiency (%) No. % No. % GH test 5 ng/ml 32/60 53 101/110 92 78 7.5 ng/ml 44/60 73 94/110 85 81 10 ng/ml 51/60 85 84/110 76 79 IGF-I 2 SD score 20/58 34 78/109 72 59 1.5 SD score 25/58 43 61/109 56 51 IGFBP-3 2 SD score 13/58 22 100/109 92 68 1.5 SD score 16/58 28 98/109 90 68 ugh 2 SD score 13/51 25 64/84 76 56 1.5 SD score 20/51 38 56/84 67 55 2b. Sensitivity (expressed as percentage) of each test at the given cut-off point for those subjects with a definitive diagnosis of, those undergoing a GH retest, and those with radiation-induced Definitive Sensitivity (%) Retested Radiation-induced GH test 7.5 ng/ml 84 82 63 IGF-I 2 SD score 47 40 28 IGFBP-3 2 SD score 53 10 0 ugh 2 SD score 21 30 32

BIOCHEMICAL TESTS IN DIAGNOSIS OF CHILDHOOD 533 TABLE 3. Results of tests (mean with range) in those classified as (group I), those with a definitive diagnosis of, and those without (group II) Results of tests Group I Definitive Group II Peak GH (ng/ml) 7.2 a 3.9 a 18.5 (0.5 42.2) (0.5 13.0) (0.9 53.6) IGF-I (ng/ml) 164 99.6 b 189 (7.9 604.3) (8.2 417.1) (13.3 745) IGF-I SD score 1.6 2.2 1.2 ( 7.7 to 0.8) ( 7.1 to 0.7) ( 5 to 1.6) IGFBP-3 (mg/l) 2.4 1.7 b 2.7 (0.3 5.0) (0.3 4.3) (0.7 6.6) IGFBP-3 SD score 1.0 b 2.4 b 0.0 ( 9.7 to 1.5) ( 9.7 to 1.4) ( 3.9 to 2.9) ugh (ng) 1.7 1.6 2.2 (0.1 7.7) (0.1 7.7) (0.1 10.1) ugh SD score 1.4 0.9 1.0 ( 6.6 to 1.9) ( 4.2 to 1.9) ( 5.3 to 2.0) a P 0.0001 vs. group II. b P 0.01 vs. group II. from the peak concentration during nonprimed tests in ageand puberty-matched children (n 54; 18.9 14.7 vs. 13.3 11.9 ng/ml). Nine subjects in group I (15%) had a peak GH concentration above 10 ng/ml, of whom six were classified as radiation-induced. Twenty-six children in group II (24%) had a peak GH response below 10 ng/ml. A single GH test with a cut-off value of 7.5 ng/ml proved to be the most efficient of all tests studied (Table 2a). The sensitivity of this test for those in group I with a definitive diagnosis of was still only 84% (Table 2b). In those subjects who were retested at the end of growth, the sensitivity was similar at 82%, whereas in those with radiation-induced, it was lower at 63% (Table 2b). ugh estimations. There was no significant difference between ugh excreted in groups I and II (Table 3). It proved to be a poor test, with an efficiency of 56%, even at a cut-off value of 2 sd score. Serum IGF-I concentration. Values were significantly reduced compared with those for a normal population in both groups, but were not different between the groups. Twenty subjects (34%) in group I and 31 in group II (28%) had an IGF-I sd score below 2, whereas the efficiency in corroborating the clinical diagnosis was very poor and similar to that of ugh (Table 2a). The sensitivity of IGF-I at a cut-off of 2 sd score was particularly poor for children receiving CNS radiation (Table 2b). Serum IGFBP-3 concentration. The IGFBP-3 sd score was significantly lower in group I than that in group II (Table 3). Group II, in fact, had a distribution of IGFBP-3 values very similar to that in a normal population. However, only 2 of 21 pubertal subjects in group I had an IGFBP-3 sd score less than 1.5. Thirteen subjects in group I (22%) and 9 in group II (8%) had an IGFBP-3 sd score below 2, indicative of the poor sensitivity of this test. The latter did improve, but only to 53%, when the test was applied to the group I subset with definitive. The sensitivity of this test was very poor in those undergoing retesting or receiving CNS radiation (Table 2b); none of the 27 subjects with radiation-induced had an IGFBP-3 sd score less than 1.5, although 16 had a GH peak value below 7.5 ng/ml. This test did, however, have a very good specificity (92%; Table 2a). Scoring system To incorporate data from all tests, we devised a scoring system based on the positive predictive value of each test in the group I subset with definitive (Table 4). No score was allocated to an IGF-I sd score above 2 or to any value of ugh because the predictive value fell below 20%. The performance of this system is shown in Table 5. All patients with definitive and 81% of group I achieved at least 5 points with a maximum up to 17 points. Ten of the 11 subjects who scored less than 5 points were in the radiation-induced group. Sixty-nine percent of group II scored less than 5 points, with a maximum of 12 points. The efficiency of this system did not exceed that found for a single GH test. However, the specificity of a score above 10 was 94% and that of a score above 13 was 100%, exceeding that found for a single GH test. If this system is applied to the 31 subjects in whom the physician was unable to allocate a definite classification of GH status, then 9 (29%) scored above 10 points, suggesting that the correct diagnosis was, whereas 10 (32%) scored less than 5, making a diagnosis of unlikely. Discussion The best approach to the investigation of the child with growth failure has been the subject of many studies (reviewed in Refs. 3, 8, and 16 18). In recent years, the debate has widened as alternative tests of the GH axis have become increasingly available (3, 18). A single measurement of serum IGF-I and IGFBP-3, both partly under GH control, provides data on the integrity of the GH transduction pathway, contributing not only to the evaluation of, but also to the determination of GH insensitivity (19). This study was designed to assess the performance of these tests in corroborating the clinical diagnosis of. To evaluate a test, it is necessary to use a gold standard. In the case of, this has proved difficult to define. For instance, if an unsatisfactory growth rate is used as the standard, then relatively poor sensitivity (69%) and specificity TABLE 4. Derivation of the scoring system; this was based on the positive predictive value in those with definitive No. of true positive results Positive predictive value (%) Points GH peak 2.5ng/mL 8/10 80 8 5 ng/ml 13/22 59 6 7.5ng/mL 16/32 50 5 10 ng/ml 18/44 41 4 IGFBP-3 2.5 SDS 7/11 63 6 2.0 SDS 10/19 52 5 1.5 SDS 10/11 48 5 1.0 SDS 13/35 37 4 IGF-I 2.5 SDS 7/26 27 3 2.0 SDS 9/40 22 2 The number of true positive results in this group is expressed as a percentage of the total number of positive results.

534 TILLMANN ET AL. JCE&M 1997 Vol 82 No 2 TABLE 5. Sensitivity, specificity, and efficiency (expressed as number of patients and as percentage of patients) using the scoring system (see Table 4) at the given cut-off score Cut-off Sensitivity Specificity Efficiency (%) No. % No. % 5 points 47/58 81 75/108 69 73 8 points 27/58 47 92/108 85 72 10 points 20/58 34 101/108 94 73 13 points 10/58 17 108/108 100 71 (63%) were achieved with a cut-off peak GH value of 7.5 ng/ml during an insulin tolerance test (20). If the peak GH responses during stimulation tests are used to define the disease, then at least 25% will have normal GH secretion as young adults (1, 2). We have chosen, therefore, a specific approach in this study in which initial clinical evaluation (based on history, examination, and growth parameters) has been used as the gold standard. Our subjects represent a heterogeneous group with disordered growth recruited from a number of growth centers, but including a subset with a very high likelihood of. We believe that this is a realistic clinical setting in which test performance can be properly evaluated. To compare test sensitivity within the subgroups, analysis has also been undertaken separately on the group with definitive, those young people who received a retest at the end of GH treatment, and those treated with CNS radiation to a lesion outside the hypothalamic-pituitary axis. A single peak GH concentration (using a cut-off level of 7.5 ng/ml) during standard stimulation tests proved to be the most efficient (81%) of the tests. Nevertheless, this still resulted in a false positive rate of 15% and a false negative rate of 27%, broadly in agreement with the many previous assessments of GH tests (1, 2, 21, 22). However, the alternative tests, ugh, serum IGF-I, and IGFBP-3 levels, when used alone performed less well. The sensitivity of the GH test was lowest in those receiving CNS radiation (63% at a cut-off of 7.5 ng/ml), but similar in those with a definitive diagnosis of (84%) and those retested at the end of growth (82%). Skinner et al. (23) have shown previously that a sensitivity of 70% and a specificity of 98% could be achieved using a single ugh estimation at a cut-off level of 2 sd. The patients with in that study all had peak GH levels below 4 ng/ml, whereas the short normal group were relatively homogeneous, with 90% having familial short stature and/or delayed growth. If this GH criteria is applied in the present study, then the sensitivity of ugh increases to 75%. A ugh test is, therefore, only helpful in subjects with severe. Serum IGF-I levels were unhelpful in discriminating among our patient groups. This held true across the age range studied despite previous studies suggesting that IGF-I is more informative in older children with (24). We found a similar specificity, but lower sensitivity, for this test than those in other studies (5, 8), with the sensitivity not substantially altered by a change in the cut-off value. This test would certainly not be useful for screening purposes. An IGFBP-3 level has been proposed as an excellent test of GH status (8, 25, 26). We certainly confirmed that this test has a high specificity (92%), and therefore, the finding of a low IGFBP-3 level is highly suggestive of (given that GH insensitivity has been excluded). However, a significant number of children with had normal IGFBP-3 levels; therefore, the sensitivity of the test is poor even in those with a definitive diagnosis of, in agreement with recent studies (10, 27). We also found that children with radiationinduced had consistently normal IGFBP-3 concentrations, as previously noted by Sklar et al. (28). In addition IGFBP-3 levels tended to fall within the normal range in subjects from groups I and II who were in puberty compared to standards for either chronological or bone age. Therefore, IGFBP-3 alone could not be used to exclude the diagnosis of. We went on to explore the use of combined data from the most useful tests, as defined by their positive predictive value in subjects with definitive. Although the efficiency of this system did not exceed that of a single GH test, it was possible to improve specificity and, therefore, confirmation of the diagnosis. In those children not allocated to group I or II, this system has helped to define the GH status of 61%. Half of the remaining 39% had radiation-induced, in which IGFBP-3 levels do not appear to reflect GH status. Although a single GH test remains the most powerful biochemical tool in the evaluation of a child with growth failure, the test remains far from ideal. The combination of this result with a single serum IGF-I and IGFBP-3 measurement can be used very effectively in the confirmation of. If careful clinical assessment and growth rate documented over at least 6 months indicate that biochemical evaluation of the GH-IGF axis should be undertaken, then we recommend the use of a single GH test, followed by serum measurements of IGF-I and IGFBP-3 in those whose peak GH levels fall below 10 ng/ml. References 1. Nicolson A, Toogood AA, Rahim A, Shalet SM. 1996 The prevalence of severe growth hormone deficiency in adults who received growth hormone replacement in childhood. Clin Endocrinol (Oxf). 44:311 316. 2. Clayton PE, Price DA, Shalet SM. 1987 Growth hormone state after completion of treatment with growth hormone. Arch Dis Child. 62:222 226. 3. Rosenfeld RG, Albertsson-Wikland K, Cassorla F, et al. 1995 Diagnostic controversy: the diagnosis of childhood growth hormone deficiency revisited. J Clin Endocrinol Metab. 80:1532 1540. 4. Furlanetto RW, Underwood LE, Van Wyk JJ, D Ercole AJ. 1977 Estimation of somatomedin-c levels in normals and patients with pituitary disease by radioimmunoassay. J Clin Invest. 60:648 657. 5. Rosenfeld RG, Wilson DM, Lee PDK, Hintz RL. 1986 Insulin-like growth factors I and II in evaluation of growth retardation. J Pediatr. 109:428 433. 6. Reiter EO, Lovinger RD. 1981 The use of commercially available somatomedin-c radioimmunoassay in patients with disorders of growth. J Pediatr. 99:720 724. 7. Juul A, Bang P, Hertel NT, Main K, Dalgaard P, Jorgensen K, Müller J, Hall K, Skakkebaek NE. 1994 Serum insulin-like growth factor-i in 1030 healthy

BIOCHEMICAL TESTS IN DIAGNOSIS OF CHILDHOOD 535 children, adolescents, and adults: relation to age, sex, stage of puberty, testicular size, and body mass index. J Clin Endocrinol Metab. 78:744 752. 8. Blum WF, Ranke MB, Keitzmann K, Grauggel E, Zeisel H, Bierich JR. 1990 A specific radioimmunoassay for the growth hormone (GH)-dependent somatomedin-binding protein: its use for diagnosis of GH deficiency. J Clin Endocrinol Metab. 70:1292 1298. 9. Hoffman DM, O Sullivan AJ, Baxter RC, Ho KKY. 1994 Diagnosis of growthhormone deficiency in adults. Lancet. 343:1064 1068. 10. Cianfarani S, Boemi S, Spagnoli A, et al. 1995 Is IGF binding protein-3 assessment helpful for the diagnosis of GH deficiency? Clin Endocrinol (Oxf). 43:43 47. 11. Ranke MB. 1991 The Kabi Pharmacia International Growth Study: aetiology classification list with comments. Acta Paediatr Scand. 379(Suppl):87 92. 12. Freeman JV, Cole TJ, Chinn S, Jones PRM, White EM, Preece MA. 1995 Cross-sectional stature and weight reference curves for the UK, 1990. Arch Dis Child. 73:17 24. 13. Tanner JM, Whitehouse RH, Takaishi M. 1966 Standards from birth to maturity for height, weight, height velocity, and weight velocity: British children. Arch Dis Child. 41:613 635. 14. Greulich WW, Pyle SI. 1959 Radiographic atlas of skeletal development of hand and wrist. Stanford: Stanford University Press. 15. Skinner AM, Price DA, Addison GM, et al. 1992 The influence of age, size, pubertal status and renal factors on urinary growth hormone excretion in normal children and adolescents. Growth Regul. 2:156 160. 16. Oerter KE, Sobel AM, Rose SR, et al. 1992 Combining insulin-like growth factor-i and mean spontaneous nighttime growth hormone levels for the diagnosis of growth hormone deficiency. J Clin Endocrinol Metab. 75:1413 1420. 17. Adan L, Souberbielle JC, Brauner R. 1994 Diagnostic markers of permanent idiopathic growth hormone deficiency. J Clin Endocrinol Metab. 78:353 358. 18. Hindmarsh PC, Swift PGF. 1995 An assessment of growth hormone provocation tests. Arch Dis Child. 72:362 368. 19. Savage MO, Blum WF, Ranke MB, et al. 1993 Clinical features and endocrine status in patients with growth hormone insensitivity (Laron syndrome). J Clin Endocrinol Metab. 77:1465 1471. 20. Dattani MT, Pringle PJ, Hindmarsh PC, Brook CGD. 1992 What is a normal stimulated growth hormone concentration? J Endocrinol. 133:447 450. 21. Cocchi C, Baldasseroni GC, Berherucci P, et al. 1989 Evaluation of GH secretion. J Endocrinol Invest. 12(Suppl 3):9 15. 22. Cacciari E, Tassoni P, Parisi G, et al. 1992 Pitfalls in diagnosing impaired growth hormone (GH) secretion: retesting after replacement therapy of 63 patients defined as GH deficient. J Clin Endocrinol Metab. 74: 1284 1289. 23. Skinner AM, Clayton PE, Price DA, Addison GM, Soo A. 1993 Urinary growth hormone excretion in the assessment of children with disorders of growth. Clin Endocrinol (Oxf). 39:201 206. 24. Moore DC, Ruvalcaba RHA, Smith EK, Kelley VC. 1982 Plasma somatomedin-c as a screening test for growth hormone deficiency in children and adolescents. Horm Res. 16:49 55. 25. Smith WJ, Nam TJ, Underwood LE, Busby WH, Celnicker A, Clemmons DR. 1993 Use of insulin-like growth factor-binding protein-2 (IGFBP-2), IGFBP-3, and IGF-I for assessing growth hormone status in short children. J Clin Endocrinol Metab. 77:1294 1299. 26. Hasegawa Y, Hasegawa T, Aso T, et al. 1994 Clinical utility of insulin-like growth factor binding protein-3 in the evaluation and treatment of short children with suspected growth hormone deficiency. Eur J Endocrinol. 131:27 32. 27. Nunez SB, Municchi G, Barnes KM, Rose SR. 1996 Insulin-like growth factor I (IGF-I) and IGF-binding protein-3 concentrations compared to stimulated and night growth hormone in the evaluation of short children a clinical research center study. J Clin Endocrinol Metab. 81:1927 1932. 28. Sklar C, Sarafoglou K, Whittam E. 1993 Efficacy of insulin-like growth factor binding protein 3 in predicting the growth hormone response to provocative testing in children treated with cranial irradiation. Acta Endocrinol (Copeh). 129:511 515.