Serum free insulin-like growth factor-i in growth hormonedeficient adults before and after growth hormone replacement

European Journal of Endocrinology (1997) 137 132 137 ISSN 0804-4643 Serum free insulin-like growth factor-i in growth hormonedeficient adults before and after growth hormone replacement C Skjærbæk 1, N Vahl 2, J Frystyk 1, T B Hansen 3, J O L Jørgensen 2, C Hagen 3, J S Christiansen 2 and H Ørskov 1 1 Institute of Experimental Clinical Research, Medical Research Laboratories, Aarhus University, Denmark, 2 Medical Department M, Aarhus Kommunehospital, Aarhus, Denmark and 3 Department of Endocrinology M, Odense University Hospital, Odense, Denmark (Correspondence should be addressed to C Skjærbæk, Medical Research Laboratories, Aarhus Kommunehospital, Building 3, 8000 Aarhus C, Denmark) Abstract The objective of the present study was to compare fasting levels of free IGF-I in serum from patients with adult onset growth hormone deficiency (GHD) and from healthy volunteers, and to examine the effect of GH replacement therapy in GHD on serum free IGF-I. Free IGF-I was measured using separation of free IGF-I by ultrafiltration in serum samples from 42 healthy volunteers and 27 patients with GHD, in the latter before and after 1 year of treatment with GH (2 IU/m 2 )(n= 13) or placebo (n = 14). Free IGF-I was significantly decreased in patients with GHD (700 100 ng/l (mean S.E.M.), range 55 2618 ng/l) compared with controls (1010 70 ng/l, range 231 2431 ng/l; P = 0.0016). Total IGF-I was 85 10 mg/l (GHD) and 160 10 mg/l (controls) (P < 0.0001). The ratio of free over total IGF-I was increased in GHD to 0.85 0.08% compared with 0.66 0.05% in controls (P = 0.04). In both GHD and controls, free IGF-I correlated significantly (P < 0.05) with total IGF-I (GHD r = 0.78; controls r = 0.42), IGFBP-1 (GHD r = ¹0.67; controls r = ¹0.46) and the molar ratio of total IGF-I over IGFBP-3 (GHD r = 0.58; controls r = 0.62). After 1 year of GH treatment, free IGF-I was increased to 2780 320 ng/l (P = 0.003) and total IGF-I was increased to 270 30 mg/l (P = 0.006) both of which values were greater than those in healthy volunteers. There were no changes in free or total IGF-I in the placebo-treated group. In conclusion, levels of free IGF-I are decreased in GHD, but measurements of free IGF-I in a single, fasting serum sample do not offer a better separation of patients with GHD from individuals with normal GH status than can be achieved by measurement of total IGF-I. One year of treatment with 2 IU/m 2 GH caused an increase of serum free IGF-I to supraphysiological levels. European Journal of Endocrinology 137 132 137 Introduction Serum levels of total insulin-like growth factor-i (IGF-I) are mainly under the control of growth hormone (GH), and serum levels of IGF-I are decreased in GH deficiency (GHD) (1 4) and increase after GH replacement (5, 6). Measurements of serum total IGF-I are therefore useful in the assessment of GH secretion in GHD (2, 7) and in the evaluation of GH replacement (5, 6). There is, however, a considerable overlap in serum total IGF-I levels between GH-deficient adults and in normal healthy individuals (3, 4). Measurements of serum total IGF-I can therefore be used only as a screening procedure, but cannot replace the more laborious GHprovocative tests in the diagnosis of GHD (3). The vast majority of circulating IGF-I is bound to six specific IGF-binding proteins (IGFBPs -1 to -6). We have recently validated a method for the determination of the free, unbound fraction of circulating IGF-I using ultrafiltration, and found that in normal, fasting adults approximately 0.5-1.0% is in the free form (8). The physiological significance of this unbound moiety remains unclear, but it has been proposed to reflect the bioactivity of IGF-I in serum. Levels of free IGF-I could therefore be expected to be low in GHD and increase after GH replacement, and might add further information, useful in monitoring the clinical effect of GH. To investigate this we have compared levels of free, fasting IGF-I in 27 GH-deficient patients and a control group of 42 healthy volunteers. Furthermore, we have measured free IGF-I after 1 year of treatment with either GH or placebo in the GH-deficient patients. Subjects and methods Subjects and design Twenty-seven patients with adult onset GHD secondary to pituitary pathology were included in the study. None of the patients had been treated with GH before the study. The diagnosis of GHD was confirmed by a GH-stimulation test with insulin. Detailed data on the patients have been provided in a previous publication (9). The patients were 1997 Society of the European Journal of Endocrinology

EUROPEAN JOURNAL OF ENDOCRINOLOGY (1997) 137 Free IGF-I in GH deficiency 133 allocated randomly to groups to receive 1 year of treatment with either placebo (n = 14) or GH (Norditropin, Novo Nordisk A/S, Bagsværd, Denmark) in a dosage of 2 IU/m 2 body surface per day (n = 13). Fasting blood samples were obtained before the start of treatment and after 1 year. For comparison fasting blood samples were obtained from 42 healthy volunteers. Informed consent was obtained from all participants and the study was approved by the Danish National Health authorities and the regional ethics committee. Assays IGFBP-1, IGFBP-3, and insulin were measured by commercial assays: IGFBP-1 by ELISA (Medix Biochemica, Kainainen, Finland), IGFBP-3 by RIA (Diagnostic System Laboratories Inc., Webster, TX, USA) and insulin by RIA (Novo Nordisk A/S). Total IGF-I was determined by an in-house, non-competitive time-resolved immunofluorimetric assay (TR-IFMA) after acid ethanol extraction of serum as described previously (10). Free IGF-I was separated from bound IGF-I by ultrafiltration (8): serum samples were diluted 1 : 11 in Krebs Ringer bicarbonate buffer containing 5% human serum albumin (ph = 7.4), 600 ml of the dilution was applied to a YMT-30 ultrafiltration membrane mounted in an MPS-1 supporting device (both Amicon Division, W R Grace and Co., Beverly, MA, USA) and centrifuged at 300 g at 37 C in triplicate. We have previously demonstrated that dilution of serum from normal individuals and those with GHD before centrifugation can be undertaken without altering the concentration of free IGF-I (8). After appropriate dilution of the filtrate, the concentration of free IGF-I was measured directly in the TR-IFMA. All samples from one individual were run in the same assay. The limit of detection of free IGF-I is 27.5 ng/l and the average intraand interassay coefficients of variation were 14% and 17% respectively. Statistics Data showing normal distribution are expressed as mean S.E.M., otherwise as median and 10 90 percentiles. The Mann Whitney non-parametric test was used to compare groups. Wilcoxon s non-parametric test for paired data was used to compare effect of treatment. Fisher s exact test was used to compare proportions. Correlations were performed on log-transformed data when appropriate. P values less than 0.05 were considered significant. All statistical tests were performed using the SOLO statistical package (BMDP Statistical Software, Los Angeles, CA, USA). Results GH-deficient patients compared with normal volunteers The results are summarized in Table 1. Free IGF-I was decreased in the GH-deficient group compared with the controls, but there was a considerable overlap between the two groups, as shown in Fig. 1A. The same was true for total IGF-I (Fig. 1B) and IGFBP-3. Ten GH-deficient patients (37%) had a serum concentration of free IGF-I below the 5th percentile of the controls (400 ng/l), 19 (70%) had a serum concentration of total IGF-I below this level (94 mg/l) and 19 (70%) had a serum IGFBP-3 below it (2220 mg/l). Serum IGFBP-1 was increased in the GH-deficient group compared with the controls, whereas there was no difference in levels of fasting insulin. In the controls free IGF-I correlated positively with total IGF-I (r = 0.42, P = 0.006) (Fig. 2A and B) and with the molar ratio of IGF-I over IGFBP-3 (r = 0.37, P = 0.0175), and correlated negatively with IGFBP-1 (r = ¹0.46, P = 0.002) (Fig. 2C and D), whereas there were no correlations between free IGF-I and body mass index (BMI), IGFBP-3 or insulin. In the GH-deficient group, free IGF-I was found to be positively correlated to total IGF-I (r = 0.78, P < 0.0001) (Fig. 2B), IGFBP-3 Table 1 Characteristics and results for all participants. Values are mean S.E.M. or median (10 90 percentiles). GHD (n ¼ 27) Controls (n ¼ 42) P Male/female 18/9 20/22 0.08 Age (years) 47 (31 54) 32 (28 52) 0.051 BMI (kg/m 2 ) 27.1 0.7 23.9 0.5 0.0003 Total IGF-I (mg/l) 85 11 157 8 <0.0001 Free IGF-I (ng/l) 700 100 1010 70 0.0016 Free over total IGF-I (%) 0.85 0.08 0.66 0.05 0.038 IGFBP-3 (mg/l) 1903 124 3071 92 <0.0001 IGF-I over IGFBP-3 0.18 (0.07 0.24) 0.20 (0.15 0.29) 0.09 IGFBP-1 (mg/l) 3.6 (1.1 15.4) 2.5 (1.3 6.5) 0.020 Insulin (mu/l) 8 (4 18) 10 (6 16) 0.21

134 C Skjærbæk and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (1997) 137 Figure 1 Individual values of fasting free IGF-I (A), total IGF-I (B), and free over total IGF-I ratio (C) in 27 patients with GHD (W) and 42 normal individuals (X). Horizontal bars represent averages. (r =0.58, P= 0.002) and the molar ratio of total IGF-I over IGFBP-3 (r = 0.59, P = 0.0011), and negatively correlated to IGFBP-1 (r = ¹0.67, P = 0.0001) (Fig. 2D). There were no correlations between free IGF-I and BMI or insulin. Effects of GH replacement There were no differences in BMI, age, insulin, IGFBP-1, IGFBP-3, free or total IGF-I between the GH-treated group and the placebo group before treatment (data not shown). GH treatment caused a significant increase in free IGF-I from 750 125 ng/l to 2780 320 ng/l (P = 0.003; Fig. 3A). Total IGF-I increased from 90 15 mg/l to 275 30 mg/l (P = 0.006; Fig. 3C), IGFBP-3 increased from 1970 167 mg/l to 2790 225 mg/l (P = 0.003) and the molar ratio of total IGF-I over IGFBP-3 increased from 0.17 to 0.44 (P < 0.0001). Total IGF-I increased in all 12 patients receiving GH and free IGF-I increased in 11 of the 12. There was no change in the ratio of free IGF-I over total IGF-I (before: 0.87% (0.27 1.37%); after: 0.89% (0.59 3.21%); P = 0.58) and there was no change in IGFBP-1 (before: 3.5 mg/l (0.9 9.4 mg/l); after: 1.5 mg/l (0.4 4.3 mg/l) P = 0.12). Fasting serum insulin increased from 10 2mU/l to 16 3 mu/l (P = 0.016). After 1 year of GH replacement, both total and free IGF-I levels were significantly (P < 0.0001) increased compared with those in the control group. There was no difference in IGFBP-3 between the control group and the GHtreated group (P = 0.14). In the placebo-treated group, there were no significant changes in free or total IGF-I (Fig. 3B, D), IGFBP-1, IGFBP-3 or insulin (data not shown). Discussion Serum total IGF-I and serum IGFBP-3 are decreased in adult GHD, but the levels of both peptides overlap with those of normal individuals, hence measurements of serum total IGF-I and IGFBP-3 are poor diagnostic tests of adult GHD (3, 4). In this study, we have examined levels of free IGF-I in a group of GH-deficient patients, using separation of free and bound IGF-I by ultrafiltration. As expected, levels of free IGF-I were significantly decreased in the GHdeficient patients. However, the levels in those patients overlapped with levels in the controls, and free IGF-I was actually decreased to a lesser degree than total IGF-I. Measurements of free IGF-I in a single, fasting serum sample therefore did not offer a better separation of adult patients with GHD from healthy individuals than could be achieved by measurement of total IGF-I or IGFBP-3. Fasting serum samples were obtained in this study, to minimize the possible influence of varying levels of insulin and, thereby, IGFBP-1. It is likely that there is a diurnal variation of free IGF-I, but the relation between a single fasting measurement and 24-h levels of free IGF-I is unknown. Hasegawa et al. (11), using a direct IRMA for free IGF-I, found that 94% of children and adolescents with complete GHD and 67% of children with partial GHD had levels of free IGF-I below the 5th percentile of an age-matched normal population sample. In contrast to the findings of our study, there was no difference in the ratio of free IGF-I over total IGF-I between the GHD group and the control group in that study. It was not stated whether fasting serum samples were obtained.

EUROPEAN JOURNAL OF ENDOCRINOLOGY (1997) 137 Free IGF-I in GH deficiency 135 Figure 2 Linear correlation between individual values of fasting free IGF-I and total IGF-I in normal volunteers (A) and patients with GHD (B); and free IGF-I and IGFBP-1 in healthy volunteers (C) and patients with GHD (D). Although there was no statistical difference in the age of the GHD group and the control group in the present study, the latter group was not age-matched to the GHD group, in that the control group contained several individuals in the age range 27 34 years, and this may have contributed to the difference in free IGF-I between the GHD group and the control group (8). The GHD group had a considerably greater BMI than the control group. Although we did not find any correlation between free IGF-I and BMI in this study, we have previously shown that the level of free IGF-I is increased in obese individuals (BMI > 25 kg/m 2 ) compared with that in normal subjects (BMI < 25 kg/m 2 ) (12). However, in that study the obese groups were characterized by hyperinsulinaemia and decreased levels of IGFBP-1 and, using multiple regression analysis, it was shown that IGFBP-1 was the only important determinant for free IGF-I (12). In the present study, the patients with GHD had significantly increased levels of IGFBP-1, and it is therefore difficult to predict which possible influence the difference in BMI may have had on free IGF-I. In contrast to IGFBP-1 concentrations, there were no significant differences in levels of insulin, the principal regulator of IGFBP-1, confirming that IGFBP-1 is increased relative to insulin in GHD (13). The molar ratio of total IGF-I over IGFBP-3 has been proposed to reflect free or easily dissociable IGF-I (14). In this study, we found a correlation between free IGF-I and this molar ratio, both in the GHD patients and in the controls. In the controls, this correlation was weak (r=0.37) and there was no difference in the molar ratio between the two groups, suggesting that this gives only a crude estimate of free IGF-I. We have previously demonstrated significant inverse relationships between serum levels of IGFBP-1 and free IGF-I in human obesity (12), during an oral glucosetolerance test and during short-term fasting (8), in each

136 C Skjærbæk and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (1997) 137 Figure 3 Individual changes in fasting free IGF-I (upper panel) and total IGF-I (lower panel) during 1 year of GH replacement (A and C) or placebo (B and D). case without significant changes in total IGF-I. The observed significant inverse correlations between free IGF-I and IGFBP-1 both in the GH-deficient patients and the controls also found in this study are supportive for the view that IGFBP-1 is an important determinant of free IGF-I, and thereby of IGF-I bioactivity (15 19). In the GHD group treated with GH, levels of free and total IGF-I were significantly increased. Free IGF-I and total IGF-I were increased to the same degree, as there was no change in the ratio of free over total IGF-I. Both free and total IGF-I were increased to levels exceeding those of the control group i.e. to supraphysiological levels. This is in agreement with the view that the GH replacement dose used in this study (2 IU/m 2 per day) and in many other clinical studies to date, is probably too high (6, 20 23). This notion is based on both serum measurements and the frequency of side-effects (21, 23). In this regard, our free IGF-I data emphasize that the GH dose requirement in adults is lower, but whether or not measurements of free IGF-I will provide a useful tool for titration of the optimal GH dose remains to be investigated. In conclusion, we found that free IGF-I is significantly decreased in adult GHD, but measurement of free IGF-I in a single fasting sample is not a better marker for GHD in adults than measurement of total IGF-I or IGFBP-3. After 1 year of replacement therapy, free IGF-I was increased in parallel with total IGF-I. Whether measurements of free IGF-I is a useful tool in the monitoring of GH replacement therapy needs further studies with GH doses in a more physiological range. Acknowledgements We are indebted to Kirsten Nyborg Rasmussen, Susanne Sørensen, Inga Bisgaard, and Joan Hansen for skilled technical assistance. GH and placebo were generously supplied by Novo Nordisk A/S, Bagsværd, Denmark. Aarhus University Research Foundation is thanked for financial support of this study.

EUROPEAN JOURNAL OF ENDOCRINOLOGY (1997) 137 Free IGF-I in GH deficiency 137 References 1 Furlanetto RW, Underwood LE, Van Wyk JJ & D Ercole AJ. Estimation of somatomedin-c levels in normals and patients with pituitary disease by radioimmunoassay. Journal of Clinical Investigation 1977 60 648 657. 2 Bates AS, Evans AJ, Jones P & Clayton RN. Assessment of GH status in adults with GH deficiency using serum growth hormone, serum insulin-like growth factor-i and urinary growth hormone excretion. Clinical Endocrinology 1995 42 425 430. 3 Hoffman DM, O Sullivan AJ, Baxter RC & Ho KK. Diagnosis of growth-hormone deficiency in adults. Lancet 1994 343 1064 1068. 4 Baum HB, Biller BM, Katznelson L, Oppenheim DS, Clemmons DR, Cannistraro KB et al. Assessment of growth hormone (GH) secretion in men with adult-onset GH deficiency compared with that in normal men a clinical research center study. Journal of Clinical Endocrinology and Metabolism 1996 81 84 92. 5 Jørgensen JOL, Flyvbjerg A, Lauritzen T, Alberti KG, Ørskov H & Christiansen JS. Dose response studies with biosynthetic human growth hormone (GH) in GH-deficient patients. Journal of Clinical Endocrinology and Metabolism 1988 67 36 40. 6 Møller J, Jørgensen JOL, Lauersen T, Frystyk J, Næraa RW, Ørskov H etal. Growth hormone dose regimens in adult GH deficiency: effects on biochemical growth markers and metabolic parameters. Clinical Endocrinology 1993 39 403 408. 7 de Boer H, Blok GJ & van der Veen EA. Clinical aspects of growth hormone deficiency in adults. Endocrine Review 1995 16 63 86. 8 Frystyk J, Skjærbæk C, Dinesen B & Ørskov H. Free insulin-like growth factors (IGF-I and IGF-II) in human serum. FEBS Letters 1994 348 185 191. 9 Hansen TB, Vahl N, Jørgensen JOL, Christiansen JS & Hagen C. Whole body and regional soft tissue changes in growth hormone deficient adults after one year of growth hormone treatment: a double-blind, randomized, placebo-controlled study. Clinical Endocrinology 1995 43 689 696. 10 Frystyk J, Dinesen B & Ørskov H. Non-competitive time-resolved immunofluorimetric assay for determination of human insulinlike growth factor I and II. Growth Regulation 1995 5 169 176. 11 Hasegawa Y, Hasegawa T, Takada M & Tsuchiya Y. Plasma free insulin-like growth factor I concentrations in growth hormone deficiency in children and adolescents. European Journal of Endocrinology 1996 134 184 189. 12 Frystyk J, Vestbo E, Skjærbæk C, Mogensen CE & Ørskov H. Free insulin-like growth factors in human obesity. Metabolism 1995 44 (suppl 4) 37 44. 13 Hilding A, Brismar K, Degerblad M, Thoren M & Hall K. Altered relation between circulating levels of insulin-like growth factorbinding protein-1 and insulin in growth hormone-deficient patients and insulin-dependent diabetic patients compared with that in healthy subjects. Journal of Clinical Endocrinology and Metabolism 1995 80 2646 2652. 14 Juul A, Main K, Blum WF, Lindholm J, Ranke MB & Skakkebæk NE. The ratio between serum levels of insulin-like growth factor (IGF)-I and the IGF binding proteins (IGFBP-1, 2 and 3) decreases with age in healthy adults and is increased in acromegalic patients. Clinical Endocrinology 1994 41 85 93. 15 Cox GN, McDermott MJ, Merkel E, Stroh CA, Ko SC, Squires CH et al. Recombinant human insulin-like growth factor (IGF)-binding protein-1 inhibits somatic growth stimulated by IGF-I and growth hormone in hypophysectomized rats. Endocrinology 1994 135 1913 1920. 16 Rajkumar K, Barron D, Lewitt MS & Murphy LJ. Growth retardation and hyperglycemia in insulin-like growth factor binding protein-1 transgenic mice. Endocrinology 1995 136 4029 4034. 17 Lewitt MS, Denyer GS, Cooney GJ & Baxter RC. Insulin-like growth factor-binding protein-1 modulates blood glucose levels. Endocrinology 1991 129 2254 2256. 18 Taylor AM, Dunger DB, Preece MA, Holly JM, Smith CP, Wass JA et al. The growth hormone-independent insulin-like growth factor-i binding protein BP-28 is associated with serum insulin-like growth factor-i inhibitory bioactivity in adolescent insulindependent diabetics. Clinical Endocrinology 1990 32 229 239. 19 Baxter RC. Insulin-like growth factor (IGF) binding proteins: the role of serum IGFBPs in regulating IGF availability. Acta Paediatrica Scandinavica (Suppl) 1991 372 107 114. 20 Wollmann HA, Schonau E, Blum WF, Meyer F, Kruse K & Ranke MB. Dose-dependent responses in insulin-like growth factors, insulin-like growth factor-binding protein-3 and parameters of bone metabolism to growth hormone therapy in young adults with growth hormone deficiency. Hormone Research 1995 43 249 256. 21 de Boer H, Blok GJ, Voerman B, de Vries P, Popp Snijders C & van der Veen E. The optimal growth hormone replacement dose in adults, derived from bioimpedance analysis. Journal of Clinical Endocrinology and Metabolism 1995 80 2069 2076. 22 Shalet SM & Holmes SJ. Growth hormone replacement in adults. British Medical Journal 1994 308 1314 1315. 23 de Boer H, Blok GJ, Popp Snijders C, Stuurman L, Baxter RC & van der Veen E. Monitoring of growth hormone replacement therapy in adults, based on measurement of serum markers. Journal of Clinical Endocrinology and Metabolism 1996 81 1371 1377. Received 6 September 1996 Accepted 28 April 1997