X/00/$03.00/0 Vol. 85, No. 11 The Journal of Clinical Endocrinology & Metabolism Copyright 2000 by The Endocrine Society

Transcription

1 X/00/$03.00/0 Vol. 85, No. 11 The Journal of Clinical Endocrinology & Metabolism Printed in U.S.A. Copyright 2000 by The Endocrine Society The Effect of Four Weeks of Supraphysiological Growth Hormone Administration on the Insulin-Like Growth Factor Axis in Women and Men* R. DALL, S. LONGOBARDI, C. EHRNBORG, N. KEAY, T. ROSÉN, J. O. L. JØRGENSEN, R. C. CUNEO, M. A. BOROUJERDI, A. CITTADINI, R. NAPOLI, J. S. CHRISTIANSEN, B. A. BENGTSSON, L. SACCA, R. C. BAXTER, E. E. BASSET, AND P. H. SÖNKSEN ON BEHALF OF THE GH-2000 STUDY GROUP Department of Medicine M (Endocrinology and Diabetes), Aarhus University Hospital (R.D., J.O.L.J., J.S.C.), DK-8000 Aarhus, Denmark; Department of Clinical Medicine and Cardiovascular Sciences, University Federico II (S.L., A.C., L.S., R.N.), Naples, Italy; Research Center for Endocrinology and Metabolism, Sahlgrenska Hospital (C.E., T.R., B.A.B.), S Gothenburg, Sweden; Metabolic Research Unit, Department of Medicine, and Department of Social and Preventive Medicine, University of Queensland, Princess Alexandra Hospital (R.C.C.), 4102 Brisbane, Queensland, Australia; Department of Endocrinology, St. Thomas s Hospital (N.K., M.A.B., P.H.S.), London, United Kingdom SE1 7EH; Institute of Mathematics and Statistics, University of Kent (E.E.B.), Canterbury, Kent, United Kingdom CT 7NF; and Kolling Institute of Medical Research, University of Sydney, Royal North Shore Hospital (R.C.B.), St. Leonards 2065, New South Wales, Australia ABSTRACT Measurements of serum insulin-like growth factor I (IGF-I) and related markers are routinely used in the diagnosis and treatment of GH deficiency and excess. The validity of these markers for assessment of exogenous GH exposure in healthy adults is, however, unknown. We therefore conducted a double blind, placebo-controlled GH treatment trial in 99 healthy subjects [49 women and 50 men; mean SE age, (women)/ yr (men)]. Blood was collected weekly during a 4-week treatment period (days 1 28), and the subjects were subsequently followed for additional 8 weeks (days 29 84). The treatment arms included: I) 0.1 IU/kg day GH (n 30; GH 0.1), II) 0.2 IU/kg day GH (n 29; GH 0.2), and III) placebo (n 40). At baseline no gender-specific differences existed, except that the acidlabile subunit (ALS) levels were higher in females. Serum insulin-like growth factor I (IGF-I) levels in males receiving GH increased significantly through day 42 with no significant difference between the 2 doses. The absolute IGF-I response was significantly lower in females, and there was a clear dose-response relationship. ALS levels SINCE THE REPORT by Salmon and Daughaday (1) it has become well established that the anabolic actions of GH to a large extent are mediated through hepatic and peripheral synthesis of insulin-like growth factor I (IGF-I) (2). With the Received June 2, Revision received July 31, Accepted August 3, Address all correspondence and requests for reprints to: Rolf Dall, M.D., Aarhus Kommunehospital, Department of Medicine M (Endocrinology and Diabetes), Aarhus University Hospital, DK-8000 Aarhus, Denmark. rd@dadlnet.dk. * This study falls within the framework of the GH-2000 Project, was conducted on behalf of the European Union and the International Olympic Committee by the Universities of Aarhus, Gothenburg, Naples Federica II, and London. This work was supported by grants from the European Union (BIOMED 2, Project BMH4 CT ) and the International Olympic Committee. Additional support and recombinant human GH were provided by Novo Nordisk A/S and Pharmacia & Upjohn, Inc. These authors contributed equally to the study. in males increased through day 30 (P 0.001). In females ALS levels were only modestly increased on day 28 compared with those in the placebo group (P 0.02). IGF-binding protein-3 (IGFBP-3) levels in males increased significantly in the GH 0.1 and the GH 0.2 groups on day 30 (P 0.03), whereas no solid IGFBP-3 increase was detected in females. IGFBP-2 levels decreased insignificantly during GH exposure in both genders. A gender-specific upper normal range for each analyte was arbitrarily defined as 4 SD above the mean level at baseline. On the basis of IGF-I levels alone, GH exposure in the GH 0.2 group was detected in 86% of the males and in 50% of the females on day 21. On day 42 GH exposure was only weakly detectable in males and was not detectable in females. We conclude that 1) males are significantly more responsive than females to exogenous GH; 2) the increase in IGF-I is more robust compared with those in IGFBP-3 and ALS; 3) IGFBP-2 changes very little during GH treatment; and 4) among IGF-related substances, IGF-I is the most specific marker of supraphysiological GH exposure. (J Clin Endocrinol Metab 85: , 2000) development of specific RIAs for IGF-I, it was subsequently revealed that serum IGF-I concentrations are low in GHdeficient children and are distinctly elevated in active acromegaly (3). The dependence of IGF-I on GH status was substantiated by demonstrating that GH administration to GH-deficient children increases serum IGF-I concentrations in a dose-dependent manner (4), and that successful treatment of acromegaly resulted in normalization of circulating IGF-I levels (5). Today, measurement of serum IGF-I is routinely used in the diagnosis and monitoring of both GHdeficient and acromegalic patients. In the circulation most IGF-I circulates as a 150-kDa ternary complex consisting of IGF-I plus a specific binding protein (IGFBP-3) and a nonbinding component, termed acid labile subunit (ALS) (6, 7). Both IGFBP-3 and ALS are partly regulated by GH, as reflected by low and elevated levels in GH deficiency and active acromegaly, respectively (7). 4193

2 4194 DALL ET AL. JCE&M 2000 Vol. 85 No. 11 In contrast to patients with GH disorders, less is known about the regulation of serum IGF-I in healthy adults. Serum IGF-I levels decline gradually with age, which is weakly correlated with the concomitant decline in GH secretion (8), but it has recently been shown that endogenous GH status is a surprisingly weak predictor of serum IGF-I (9). Possible residual determinants of IGF-I in normal adults include gender and sex steroids, body composition, and physical fitness. Even less is known about the impact of exogenous GH exposure on IGF-I and related variables in healthy adults. Such information would seem to be of general interest when considering the use of GH for nonlicensed indications such as catabolic states, sarcopenia, and rejuvenation of age-associated changes in body composition. Moreover, there is increasing evidence to suspect widespread illicit use of GH as a performance-enhancing agent among athletes (10). The aim of the present study was to assess the effects of GH administration in supraphysiological doses on serum concentrations of IGF-I and pertinent related variables and subsequently to evaluate the usefulness of such variables to predict exposure to exogenous GH. To this end a large cohort of healthy, fit, young adults of both sexes were studied before, during, and after GH administration with two doses in a randomized, placebo-controlled, parallel trial. Subjects Subjects and Methods Ninety-nine subjects (50 men and 49 women) were recruited among healthy volunteers in 4 different countries: Denmark (n 31), Italy (n 27), Sweden (n 30), and the United Kingdom (n 11). The inclusion criteria were age between yr, unremarkable medical history, no intake of medications, and regular participation in at least 2 exercise sessions/week for at least 1 yr. None of the subjects was an elite athlete, and participation in organized sport competitions was not allowed during the study period. All women documented a negative pregnancy test before study entry and were confirmed to be using safe contraception; 9 women used oral contraceptives (1 in the placebo group, 3 in the 0.1 IU/kg day GH group, and 5 in the 0.2 IU/kg day group), and the remaining females used barrier methods. Treatment with GH/placebo started randomly with respect to the menstrual cycle. In each country the protocol was approved by the ethical committee system and the national health authorities. Oral and written consents were obtained from each subject in accordance with the principles stated in the Declaration of Helsinki. The baseline characteristics of the subjects are presented in Tables 1 and 2. TABLE 1. Baseline characteristics (mean SE) for females Placebo (n 19) GH 0.1 (n 15) GH 0.2 (n 15) P values a Age (yr) BMI (kg/m 2 ) IGF-I ( g/l) IGFBP-2 ( g/l) IGFBP-3 (mg/l) ALS (nmol/l) IGF-I/IGFBP-3 b The study was a placebo-controlled dose-response study with 4 weeks of GH administration [0.1 IU/kg day (GH 0.1) or 0.2 IU/ kg day (GH 0.2)], followed by an 8-week wash-out period. Blood samples were analyzed at baseline, during treatment (days 21 and 28), and in the washout period (days 30, 33, 42, and 84). a By one-way ANOVA. b To convert mass units to nmol/l for IGF-I ( g/l) and IGFBP-3 (mg/l), multiply by and 23.26, respectively. TABLE 2. Baseline characteristics (mean SE) for males Design The study was performed in a randomized, double blind, placebocontrolled, parallel design involving three arms: I) 0.1 IU/kg day GH (GH 0.1; maximum dose, 9.5 IU/day; n 30; 15 women and 15 men), II) 0.2 IU/kg day GH (GH 0.2; maximum dose, 19 IU/day; n 29; 15 women and 14 men), and III) placebo (PLA; n 40; 19 women and 21 men). GH (Genotropin, Pharmacia & Upjohn, Inc., Stockholm, Sweden; or Norditropin, Novo Nordisk, Copenhagen, Denmark) and placebo were administered as daily sc self-injections in the evening. To minimize side effects, only 50% of the target dose was given during the first week. In case of side effects the dosage was reduced by 50%, and treatment was discontinued if the complaints persisted for more than 1 week. Compliance was monitored by collection of vials and reports about missing injections. At baseline the subjects attended the hospital for 1 day for blood sampling, interview, and physical examination. Blood samples were subsequently collected at the end of each week during the treatment phase (days 7, 14, 21, and 28) and on days 30, 33, 42, and 84 in the wash-out period. Assays All serum samples were stored at 80 C, and all samples from one subject were analyzed in the same run. Serum IGF-I was measured by RIA using a monoclonal antibody after acid-ethanol extraction (Nichols Institute Diagnostics, San Juan Capistrano, CA). IGFBP-2 (11), IGFBP-3 (6), and ALS (7) were assayed using in-house RIAs and polyclonal antibodies (Robert Baxter, Sydney, Australia). The intraassay coefficients of variations (CVs) are given in Table 4. Statistical analyses Placebo (n 21) GH 0.1 (n 15) GH 0.2 (n 14) P values a Age (yr) BMI (kg/m 2 ) IGF-I ( g/l) IGFBP-2 ( g/l) IGFBP-3 (mg/l) ALS (nmol/l) IGF-I/IGFBP-3 b The study was a placebo-controlled dose-response study with 4 weeks of GH administration [0.1 IU/kg day (GH 0.1) or 0.2 IU/ kg day (GH 0.2)], followed by an 8-week wash-out period. Blood samples were analyzed at baseline, during treatment (days 21 and 28), and in the washout period (days 30, 33, 42, and 84). a By one-way ANOVA. b To convert mass units to nmol/l for IGF-I ( g/l) and IGFBP-3 (mg/l), multiply by and 23.26, respectively. Unpaired two-tailed t test was used to test for gender differences at baseline. The variance in the PLA group was tested by ANOVA for repeated measures approached by general linear modeling (GLM repeated measures). The CVs for the variables in the PLA group were calculated as the mean sem of the sd/mean for each subject. Temporal changes in serum levels of each variable were tested by GLM repeated measures with each time point as a with-in subject factor, and randomization and gender as between-factors. If GLM repeated measures revealed significant changes, post-hoc analysis was made by gender-stratified one-way ANOVA, and multiple comparisons were made using the Bonferroni correction. The area under the curve (AUC) during the treatment period (days 1 28) of each variable was estimated according to the trapezoidal rule. Differences in AUC among the three groups were analyzed by one-way ANOVA. Where appropriate, post-hoc analysis was performed by multiple comparisons with the Bonferroni correction. Pearson correlation analyses (r) as well as multiple linear regression analyses were used to relate variables. P 0.05 was considered significant. Results are expressed as the mean se.

3 IGF AXIS DURING GH ADMINISTRATION 4195 Side effects During the study, 8 subjects (3 women and 5 men) of the 30 enrolled in the GH 0.1 group experienced side effects; among these, 1 subject reduced the dose (1 men). In the GH 0.2 group, 20 subjects (7 women and 12 men) experienced side effects; among these, 5 reduced the dose (2 women and 3 men). In the PLA group, 12 subjects (7 women and 5 men) had side effects, of whom 1 subject reduced the dose and 1 subject stopped the treatment on day 14 due to headache and tachycardia. The most frequently reported side effects were attributable to transient fluid retention. In addition, increased sweating and arthralgia were reported. Results The participants in the three study arms were well matched for gender, age, and body composition (Tables 1 and 2). Baseline levels of IGF-related variables were similar in each group, apart from slightly higher serum IGFBP-3 concentrations among females in the PLA group (Tables 1 and 2). Levels of IGF-I, IGFBP-2, and IGFBP-3 and the IGF- I/IGFBP-3 ratio showed no gender differences at baseline, whereas ALS levels were significantly higher in women (P ; Table 3). At baseline, IGF-I correlated positively with IGFBP-3 (r 0.5; P ) and ALS (r 0.6; P ), whereas IGFBP-2 and ALS levels correlated negatively (r 0.3; P 0.002). Gender-stratified multiple linear regression models with IGF-I, IGFBP-2, IGFBP-3, and ALS as dependent variables and age and BMI as independent variables were performed at baseline. IGF-I correlated negatively with age in men [age: 0.5; P ; body mass index (BMI): 0.07; P TABLE 3. Baseline characteristics for females and males (mean SE) Female (n 49) Male (n 50) P values a IGF-I ( g/l) IGFBP-2 ( g/l) IGFBP-3 (mg/l) ALS (nmol/l) IGF-I/IGFBP-3 b The study was a placebo-controlled dose-response study with 4 weeks of GH administration [0.1 IU/kg day (GH 0.1) or 0.2 IU/ kg day (GH 0.2)], followed by an 8-week wash-out period. Blood samples were analyzed at baseline, during treatment (days 21 and 28), and in the washout period (days 30, 33, 42, and 84). a By unpaired two-tailed t test. b To convert mass units to nmol/l for IGF-I ( g/l) and IGFBP-3 (mg/l), multiply by and 23.26, respectively. TABLE 4. Characteristics of the placebo group during the study period Female 0.6; overall: P 0.001; r ], but not in women (P 0.22). For both genders IGFBP-2 levels correlated positively with age and negatively with BMI (women: age: 0.4; P 0.005; BMI: 0.04; P 0.03; overall: P 0.008; r ; males: age: 0.3; P 0.04; BMI: 0.4; P 0.005; overall: P 0.01; r ). ALS levels in females did not correlate significantly with age and BMI (P 0.4), but in males there was a negative correlation with age (age: 0.4; P 0.02; BMI: 0.1; P 0.4; overall: P 0.049; r ). All variables in the PLA group remained stable during the entire study period, as ANOVA revealed no significant time FIG. 1. IGF-I levels (micrograms per L; mean SE). Top, Women; bottom, men. F, PLA group;, GH 0.1 group; Œ, GH 0.2 group. Stability (P) a Intraindividual CV (%) b Stability (P) a Intraindividual CV (%) b Male Intraassay CV (%) IGF-I ( g/l) IGFBP-2 ( g/l) IGFBP-3 (mg/l) ALS (nmol/l) IGF-I/IGFBP-3 c Not applicable The study was a placebo-controlled dose-response study with 4 weeks of GH administration [0.1 IU/kg day (GH 0.1) or 0.2 IU/kg day (GH 0.2)], followed by an 8-week wash-out period. Blood samples were analyzed at baseline, during treatment (days 21 and 28), and in the washout period (days 30, 33, 42, and 84). a P values from ANOVA for repeated measures in the placebo group for temporal changes. b The intraindividual coefficient of variation (CV) estimated as the average ratio (SD/mean) for all subjects in the placebo group during the study period. The intraindividual CV reflects both biological and intraassay CV. Intraassay CVs are shown in the right column. c To convert mass units to nmol/l for IGF-I ( g/l), IGFBP-2 ( g/l), and IGFBP-3 (mg/l), multiply by 0.131, , and 23.26, respectively.

4 4196 DALL ET AL. JCE&M 2000 Vol. 85 No. 11 FIG. 2. IGF-I dot plot (micrograms per L; mean SE). Top, Women; bottom, men. F, All 99 subjects at baseline;, GH 0.1 group; Œ, GH 0.2 group. effect in the PLA group during the study period (Table 4 and Figs. 1 and 3 6). IGFBP-2 levels showed the highest intraindividual CV during the study period (20 25%), whereas the CVs for IGF-I, IGFBP-3, and ALS were between 10 16% (Table 4). The biological CVs, estimated as the intraindividual CVs minus the intraassay CVs, were 6 9% for IGF-I, IGFBP-3, and ALS and approximately 20% for IGFBP-2 (Table 4). In groups receiving GH, serum IGF-I increased significantly with time (P ) and interacted with treatment (P ) and gender (P ). Serum IGF-I levels are depicted separately for each sex due to the pronounced gender difference (Fig. 1). Post-hoc analysis in the female group revealed significantly higher IGF-I levels through day 30 in the GH 0.2 group compared with the PLA group (P ), whereas IGF-I levels in the GH 0.1 group were only different from the PLA group on day 21 (P 0.04). Among females IGF-I levels were significantly higher in the GH 0.2 group compared with the GH 0.1 group through day 30 (day 30; P 0.04; Fig. 1, top panel). The AUC of IGF-I (AUC IGF-I ) in females during the treatment period (days 1 28) was significantly increased in the GH 0.2 group, whereas no significant difference was detected between PLA and the GH 0.1 group [mean se; AUC IGF-I, 9, (Pla) vs. 11,762 1,011 (GH 0.1) vs. 15, (GH 0.2) g/l 28 days; P ]. In men IGF-I levels increased dose independently in the two GH groups and had not returned completely to baseline levels on day 42 (P 0.005; Fig. 1, bottom panel). AUC IGF-I during active treatment in males increased dose independently in the two GH groups compared with that in the PLA group [mean se; AUC IGF-I, 7, (Pla) vs. 17,532 1,062 (GH 0.1) vs. 18, (GH 0.2) g/l 28 days; P ]. Individual IGF-I values are presented as scatterplots for females and males in Fig. 2. The upper 4 sd level of IGF-I measurements at baseline was chosen as an arbitrary cut-off to separate normal from abnormal IGF-I values. Fifty percent of IGF-I measurements among women in the GH 0.2 group were above the cut-off on day 21 compared with 33% on day 28. The corresponding values for women in the GH 0.1 group

5 IGF AXIS DURING GH ADMINISTRATION 4197 FIG. 3. IGFBP-2 levels (micrograms per L; mean SE). Top, Women; bottom, men. F, PLA group;, GH 0.1 group; Œ, GH 0.2 group. were 7% (day 21) and 23% (day 28), respectively. By contrast, in men 86% and 64% of IGF-I levels in the GH 0.2 group were above the cut-off on days 21 and 28, respectively. The corresponding values for males in the GH 0.1 group were 73% (day 21) and 60% (day 28), respectively. The ability to discriminate between basal and GH-stimulated IGF-I values was much poorer with IGFBP-3 and ALS than with IGF-I (data not shown). Serum IGFBP-2 levels interacted significantly with time (P 0.001) and time treatment (P 0.02), but no gender difference was disclosed with ANOVA (Fig. 3). Post-hoc analysis, however, was unable to demonstrate significant differences at any specific time point, indicating that the GHinduced suppression of IGFBP-2 depicted by ANOVA was small. Likewise, AUC IGFBP-2 during active treatment (days 1 28) did not differ between the groups. The changes in serum IGFBP-3 concentrations exhibited a marked gender difference, with a distinct, albeit dose-independent, increase in GH-treated males compared with minimal fluctuations with time in women (Fig. 4). This was reflected by ANOVA, which revealed a significant interaction between time treatment gender (P 0.04). In men, IGFBP-3 levels remained significantly elevated through day 30 with either GH dose (PLA vs. GH 0.1, P 0.03; PLA vs. GH 0.2, P 0.01). AUC IGFBP-3 during active treatment (days FIG. 4. IGFBP-3 levels (milligrams per L; mean SE). Top, Women; bottom, men. F, PLA group;, GH 0.1 group; Œ, GH 0.2 group. 1 28) in women were similar in each group, whereas GH treatment among males induced an increase [mean se; AUC IGFBP-3, (PLA) vs (GH 0.1) vs (GH 0.2) mg/l 28 days; P 0.006]. ALS exhibited significant interaction with time, treatment, and gender (P 0.04; Fig. 5). Again, a dose dependency was recorded in women, where only the GH 0.2 group was significantly different from the PLA group on day 28 (P 0.04). In men, both the GH 0.1 and GH 0.2 groups had elevated levels on day 28 (P ), and ANOVA also revealed a difference among the three groups on day 30 (P 0.001). AUC ALS (nanomoles per L/28 days) in women did not change significantly (P 0.06), whereas in men there were highly significant changes between the two GH groups and the PLA group [ (PLA) vs (GH 0.1) vs (GH 0.2); P ]. The IGF-I/IGFBP-3 ratio showed a treatment effect (P ), but no gender effect (Fig. 6). In women post-hoc analyses revealed a dose dependency, with the GH 0.2 group displaying higher values through day 30 compared with the PLA group (P 0.001) and through day 28 compared with the GH 0.1 group (P 0.001). In men a GH dose-independent increase was detected through day 42 in the two GH groups compared with the PLA group (P 0.008). AUC IGF-I/IGFBP-3 in women was significantly elevated in

6 4198 DALL ET AL. JCE&M 2000 Vol. 85 No. 11 FIG. 5. ALS levels (nanomoles per L; mean SE). Top, Women; bottom, men. F, PLA group;, GH 0.1 group; Œ, GH 0.2 group. the GH 0.2 group compared with both the GH 0.1 and PLA groups [ (PLA) vs (GH 0.1) vs (GH 0.2); P ]. In men AUC IGF-I/IGFBP-3 increased independently of the GH dose compared with placebo [ (PLA) vs (GH 0.1) vs (GH 0.2); P ]. Gender differences in GH responsiveness on day 21 The increments in IGF-I, IGFBP-3, and ALS levels from baseline to day 21 in the combined GH-treated groups (GH 0.1 and GH 0.2) were tested for gender differences by comparing values (mean se): change in ( ) IGF-I, g/l (women) vs g/l (men), P ; IGFBP-3, mg/l (women) vs mg/l (men), P 0.02; ALS, nmol/l (women) vs nmol/l (men), P Discussion The aim of the present study was to assess the impact of GH administration in supraphysiological doses on the circulating levels of IGF-I and related variables in healthy adults of both sexes. Collection of such data is relevant in view of the ongoing interest in the use of exogenous GH for unlicensed indications in general and GH abuse by athletes in particular (12). FIG. 6. IGF-I/IGFBP-3 ratio (mean SE). Top, Women; bottom, men. IGF-I/IGFBP-3 ratios are calculated from molar concentrations. F, PLA group;, GH 0.1 group; Œ, GH 0.2 group. The main findings include 1) IGF-I is the most sensitive marker of GH exposure; 2) the GH-induced increase in IGFBP-3 and ALS is markedly lower compared with that in IGF-I; 3) IGFBP-2 responds very little to GH; and 4) a marked gender difference exists, with men being more responsive to exogenous GH administration. The large sample size and the randomized design, including two different doses, strengthens the validity of the observations, but the study population, on the other hand, covered a narrow age range of young, healthy, and physically fit individuals. Extrapolations to older people or patients should therefore be made with caution. At baseline no gender differences were recorded, apart from higher ALS levels among women. Comparable levels of IGF-I, IGFBP-3, and IGFBP-2 between healthy adult males and females has previously been reported (6, 13), but gender differences in IGF-I levels have been observed in particular age groups (14). Females may exhibit slightly higher IGF-I levels than males at puberty and in early adulthood, whereas the opposite has been reported in middle-aged adults (9, 15). In adult hypopituitary patients with documented GH deficiency IGF-I levels are consistently higher in men than in women (9, 16). Our observation of higher ALS levels in women has been reported recently (17), but is in contrast with

7 IGF AXIS DURING GH ADMINISTRATION 4199 a previous report in which no gender differences were observed in any age group (7). The pattern among males was characterized by high and almost identical increments in IGF-I after the two GH doses. In women, the IGF-I response was lower, but a distinct dose responsiveness was observed. Thus, the lower GH dose group apparently reached the top of the dose-response curve in males, but not in females. It is evident that the present study involved the high end of the dose-response curve, but there is evidence from the literature that a gender difference in GH sensitivity also exists with lower GH doses. The relative GH resistance in women has been described by Ghigo et al. (18), who reported that the minimum exogenous GH dose needed to elicit an IGF-I response in normal subjects is higher in women than in men. Both spontaneous and stimulated GH levels are elevated in women compared with men (16, 19, 20), which has been causally linked to differences in estradiol levels. It is, however, unresolved whether the stimulatory effect of estradiol on GH release involves a central stimulation or a negative feedback linkage to peripheral reduction of IGF-I production. In favor of the latter hypothesis, several studies have shown that administration of exogenous estradiol lowers serum IGF-I levels concomitantly with amplification of endogenous GH release (21). It has, on the other hand, recently been reported that both GH release and serum IGF-I levels increase during the periovulatory phase in normal young women, which coincides with elevated endogenous estradiol levels (22). Alternatively, it could be hypothesized that androgens play a permissive role for GHstimulated IGF-I production. This could explain the suppression of IGF-I production in postmenopausal women during exogenous estradiol administration, which is likely to inhibit endogenous androgen secretion, and it may also account for the low IGF-I levels in hypopituitary females. This theory is supported by Erfurth et al. (23), who found a significant correlation between free testosterone and IGF-I levels. Recently, Span et al. (24) reported that estrogen replacement in GH-deficient women significantly increased GH requirements, and androgen substitution in GH-deficient men increased GH sensitivity. Regardless of the physiological mechanisms the present study demonstrates that women, relative to men, are resistant to GH in terms of IGF-I generation. A similar gender difference has recently been shown regarding the acute lipolytic response to a physiological GH bolus (25). These observations in healthy adults are in accordance with the idea that the GH dose requirements in hypopituitary adults are higher in female patients, as judged by serum IGF-I levels as well as changes in body composition (19). In both sexes GH discontinuation was followed by an abrupt decline in total IGF-I levels. The elevations in IGFBP-3 and ALS were far less pronounced than those in IGF-I, and again, the response was lower in women than in men. Ghigo et al. (18) reported that IGFBP-3 levels were less increased than IGF-I levels after GH exposure. It has previously been shown that the circulating total IGF-I/IGFBP-3 ratio is elevated in active acromegaly (13) and after GH administration in both GH-deficient adults (26) and healthy controls (18). Accordingly, we also recorded a significantly increased IGF-I/IGFBP-3 ratio in the present study, which prevailed even 2 weeks after cessation of GH administration. Serum IGFBP-2 concentrations did not consistently change during GH administration, which contrasts with at least one previous study in which GH administration suppressed circulating IGFBP-2 levels (27). Juul et al. reported decreased IGFBP-2 levels in acromegalic patients (13), and Jørgensen et al. (5) observed suppressed IGFBP-2 levels in active acromegaly, which became normalized after successful surgery, whereas Clemmons et al. (28) reported normal IGFBP-2 levels in active acromegaly and moderately elevated levels in hypopituitary adults. Our study clearly demonstrated that IGF-I is superior to IGFBP-2, IGFBP-3, and ALS regarding the ability to identify exposure to supraphysiological doses of exogenous GH in healthy subjects. Using the upper 4 sd level of IGF-I in the GH-untreated state as an arbitrary cut-off limit, 86% of men were identified while receiving the high GH dose, and the IGF-I level was not completely returned to baseline 14 days after termination of GH administration. As previously mentioned, this percentage was significantly lower in women. As a mean to detect GH abuse in athletes, a single measurement of IGF-I in serum is probably not sufficiently robust, at least not in females. Whether measurements of other GH-dependent growth markers, alone or in combination with IGF-I, will prove more efficacious must await further analysis. Acknowledgments We thank all members of the GH-2000 study group not included among the authors: Marie-Louise Healy, Jake Powrie, and Claire Pentecost (London, UK); Per-Arne Lundberg, Sahlgrenska Hospital (Gothenburg, Sweden); Hans Ørskov, Aarhus University Hospital (Aarhus, Denmark); Kai H. W. Lange and Michael Kjær, Sports Medicine Research Unit (Copenhagen, Denmark); Phil Brown and Mike Kenward, Mathematics Institute, Kent University (Canterbury, UK); Prince Alexander De Merode, International Olympic Committee Medical Commission (Lausanne, Switzerland); Laurent Rivier, Laboratorie Suisse d Analyze du Dopage (Lausanne, Switzerland); Don Catlin, International Olympic Committee Drug Testing Laboratory (Los Angeles, CA); Jennifer Wallace, University of Queensland, Princess Alexandra Hospital (Brisbane, Australia); Linda Fryklund, Pharmacia & Upjohn, Inc. (Stockholm, Sweden); Anne-Marie Kappelgaard, Novo Nordisk A/S (Copenhagen, Denmark); and Christian J. Strasburger, Innenstadt University Hospital (Munich, Germany). The GH-2000 steering committee consists of P. H. Sönksen (Chairman), J. Powrie, B.-Å. Bengtsson, J. S. Christiansen, L. Saccà, and C. Pentecost (Project Manager). The publication committee consists of B.-Å. Bengtsson (Chairman), P. H. Sönksen, J. S. Christiansen, and L. Saccà. Novo Nordisk A/S and Pharmacia & Upjohn, Inc., are thanked for their support and the donation of the study medicine. References 1. Daughaday WH, Hall K, Raben MS, Salmon WDJ, Brande JL, Wyk JJ Somatomedin: proposed designation for sulphation factor. Nature. 235: Jones JI, Clemmons DR Insulin-like growth factors, their binding proteins: biological actions. Endocr Rev. 16: Furlanetto RW, Underwood LE, Van Wyk JJ, D Ercole AJ Estimation of somatomedin-c levels in normals and patients with pituitary disease by radioimmunoassay. J Clin Invest. 60: Copeland KC, Underwood LE, Van Wyk JJ Induction of immunoreactive somatomedin C human serum by growth hormone: dose-response relationships and effect on chromatographic profiles. J Clin Endocrinol Metab. 50: Jørgensen JO, Møller N, Møller J, Weeke J, Blum WF Insulin-like growth factors (IGF)-I, -II and IGF binding protein-1, -2, and -3 in patients with acromegaly before and after adenomectomy. Metabolism. 43: Baxter RC, Martin JL Radioimmunoassay of growth hormone-depen-

8 4200 DALL ET AL. JCE&M 2000 Vol. 85 No. 11 dent insulinlike growth factor binding protein in human plasma. J Clin Invest. 78: Baxter RC Circulating levels and molecular distribution of the acid-labile (alpha) subunit of the high molecular weight insulin-like growth factor-binding protein complex. J Clin Endocrinol Metab. 70: Rudman D, Kutner MH, Rogers CM, Lubin MF, Fleming GA, Bain RP Impaired growth hormone secretion in the adult population: relation to age and adiposity. J Clin Invest. 67: Jørgensen JO, Vahl N, Hansen TB, et al Determinants of serum insulinlike growth factor I in growth hormone deficient adults as compared to healthy subjects. Clin Endocrinol (Oxf). 48: Rickert VI, Pawlak MC, Sheppard V, Jay MS Human growth hormone: a new substance of abuse among adolescents? Clin Pediatr. 31: Ho PJ, Baxter RC Insulin-like growth factor-binding protein-2 in patients with prostate carcinoma and benign prostatic hyperplasia. Clin Endocrinol (Oxf). 46: Wallace JD, Cuneo RC, Baxter R, et al Responses of the growth hormone (GH) and insulin-like growth factor axis to exercise, GH administration, and GH withdrawal in trained adult males: a potential test for GH abuse in sport. J Clin Endocrinol Metab. 84: 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. Clin Endocrinol (Oxf). 41: Juul A, Bang P, Hertel NT, et al Serum insulin-like growth factor-i in 1030 healthy children, adolescents, and adults: relation to age, sex, stage of puberty, testicular size, and body mass index. J Clin Endocrinol Metab. 78: Landin-Wilhelmsen K, Wilhelmsen L, Lappas G, et al Serum insulinlike growth factor I in a random population sample of men and women: relation to age, sex, smoking habits, coffee consumption and physical activity, blood pressure and concentrations of plasma lipids, fibrinogen, parathyroid hormone and osteocalcin. Clin Endocrinol (Oxf). 41: Attanasio AF, Lamberts SW, Matranga AM, et al Adult growth hormone (GH)-deficient patients demonstrate heterogeneity between childhood onset and adult onset before and during human GH treatment. Adult Growth Hormone Deficiency Study Group. J Clin Endocrinol Metab. 82: Juul A, Møller S, Mosfeldt LE, et al The acid-labile subunit of human ternary insulin-like growth factor binding protein complex in serum: hepatosplanchnic release, diurnal variation, circulating concentrations in healthy subjects, and diagnostic use in patients with growth hormone deficiency. J Clin Endocrinol Metab. 83: Ghigo E, Aimaretti G, Maccario M, et al Dose-response study of GH effects on circulating IGF-I and IGFBP-3 levels in healthy young men and women. Am J Physiol. 276:E1009 E Burman P, Johansson AG, Siegbahn A, Vessby B, Karlsson FA Growth hormone (GH)-deficient men are more responsive to GH replacement therapy than women. J Clin Endocrinol Metab. 82: Ho KY, Evans WS, Blizzard RM, et al Effects of sex and age on the 24-hour profile of growth hormone secretion in man: importance of endogenous estradiol concentrations. J Clin Endocrinol Metab. 64: Kelly JJ, Rajkovic IA, O Sullivan AJ, Sernia C, Ho KK Effects of different oral oestrogen formulations on insulin-like growth factor-i, growth hormone and growth hormone binding protein in post-menopausal women. Clin Endocrinol (Oxf). 39: Ovesen P, Vahl N, Fisker S, Veldhuis JD, Christiansen JS, Jørgensen JO Increased pulsatile, but not basal, growth hormone secretion rates and plasma insulin-like growth factor I levels during the periovulatory interval in normal women. J Clin Endocrinol Metab. 83: Erfurth EM, Hagmar LE, Saaf M, Hall K Serum levels of insulin-like growth factor I and insulin-like growth factor-binding protein 1 correlate with serum free testosterone and sex hormone binding globulin levels in healthy young and middle-aged men. Clin Endocrinol (Oxf). 44: Span JP, Pieters GF, Sweep CG, Hermus AR, Smals AG Gender difference in insulin-like growth factor I response to growth hormone (GH) treatment in GH-deficient adults: role of sex hormone replacement. J Clin Endocrinol Metab. 85: Vahl N, Møller N, Lauritzen T, Christiansen JS, Jørgensen JO Metabolic effects and pharmacokinetics of a growth hormone pulse in healthy adults: relation to age, sex, and body composition. J Clin Endocrinol Metab. 82: Skjærbæk C, Vahl N, Frystyk J, et al Serum free insulin-like growth factor-i in growth hormone-deficient adults before and after growth hormone replacement. Eur J Endocrinol. 137: Kicman AT, Miell JP, Teale JD, et al Serum IGF-I and IGF binding proteins 2 and 3 as potential markers of doping with human GH. Clin Endocrinol (Oxf). 47: Clemmons DR, Snyder DK, Busby WHJ Variables controlling the secretion of insulin-like growth factor binding protein-2 in normal human subjects. J Clin Endocrinol Metab. 73: