- Email: [email protected]
Disturbances in growth hormone secretion and action in adolescents with anorexia nervosa
ADOLESCENT MEDICINE
Disturbances in growth hormone secretion and action in adolescents with anorexia nervosa Neville H. G o l d e n , MD, Paula Kreitzer, MD, M a r c S. J a c o b s o n , MD, Fred I. C h a s a l o w , PhD, J a n e t S c h e b e n d a c h , MA, RD, Samuel M. F r e e d m a n , MD, a n d I. Ronald Shenker, MD From the Department of Pediatrics, Divisions of Adolescent Medicine and Pediatric Endocrinology, Schneider Children's Hospital of Long Island Jewish Medical Center, Albert Einstein College of Medicine, New Hyde Park, New York, and the Division of Pediatric Endocrinology, Maimonides Medical Center, Brooklyn, New York Women in whom anorexia nervosa d e v e l o p s during a d o l e s c e n c e have failure of linear growth associated with low levels of insulin-like growth factor I (IGF-I). To investigate the p a t h o p h y s i a l o g y of growth retardation in adolescents with anorexia nervosa, we measured basal growth hormone (GH), growth h o r m o n e binding protein (GHBP), IGF-I, and insulin-like growth factor binding protein-3 (IGFBP-3) in three groups of patients: (I) 28 recently hospitalized female adolescents with anorexia nervosa, (2) 23 of the same patients after partial weight restoration, and (3) 28 healthy control subjects m a t c h e d for age, sex, and pubertal stage. Fasting GH levels in group I did nat differ significantly from those in group 3. In contrast, serum GHBP (p <0.001), IGF-I (p <0.001), and IGFBP-3 (p <0.01) were significantly lower in group t than in group 3. Serum GHBP and IGFBP-3 levels were positively correlated with b o d y mass index. Serum GHBP levels were low in patients in all five pubertal stages and even in those shown to have a d e q u a t e GH secretion. In group 2 (after refeeding) the serum IGF-I concentration increased significantly and GHBP and IGFBP-3 returned to normal. We c o n c l u d e that patients with anorexia nervosa have diminished GH action resulting in decreased secretion of IGF-I. The positive correlation with b o d y mass index and the reversibility with refeeding suggest that these changes are s e c o n d a r y to malnutrition. Altered GH function that occurs during the years of active growth can e x p l a i n the growth retardation seen in anorexia n e r v o s a . (J PEDIATR 1994;125:655-60)
If anorexia nervosa develops during childhood or early adolescence, failure of linear growth can occur, resulting in decreased final adult height,~4 The cause of growth failure in anorexia nervosa has not been elucidated. Decreased levPresented in part at the Society for Pediatric Research, Washington, D.C., May 1993. Submitted for publication Feb. 27, 1994; accepted April 14, 1994. Reprint requests: Neville H. Golden, MD, Divisionof Adolescent Medicinel Room 187, Schneider Children's Hospital, Long Island Jewish Medical Center, New Hyde Park, NY 11042. Copyright © 1994 by Mosby-Year Book, Inc. 0022-3476/94/$3.00 + 0 9/27/56783
els of a growth hormone-dependent peptide, insulin-like growth factor I, have been found in anorexia nervosa, 5 as well as during fasting6, 7 and in other forms of malnutrition.8 In 35% to 50% of patients with anorexia nervosa and low GH GHBP IGF-I IGFBP-3 REE
Growth hormone Growth hormone binding protein Insulin-like growth factor I Insulin-like growth factor binding protein-3 Resting energy expenditure
weight, there are elevated basal levels of growth hormone 9"12 suggestive of GH resistance. However, other patients have inadequate GH secretory responses to provoca655
656
Golden et al.
The Journal of Pediatrics October 1994
T a b l e I. Clinical features: Patients with anorexia nervosa versus control subjects
Age (yr) Tanner stage Height (cm) Weight (kg) BMI (kg/mz) T3 RIA (ng/dl) LH (mIU/rnl) Estradiol (pg/ml)
Patients with anorexia nervosa (n = 28)
Control subjects (n = 28)
p
15.1 ± 2.8 3.9 ± 1.2 157.5 ± 10.2 37.5 ± 7.8 15.0 _+ 2.1 91.7 ± 35.1 3.6 ± 3.1 13.0 ± 8.4
15.0 ± 2.6 3.9 ± 1.3 159.5 ± 10.7 54.2 _+ 13.6 21.1 ± 3.8 124.6 ± 29.3 10.7 ± 14.8 53.7 ± 48.1
NS NS NS <0.001 <0.001 <0.001 <0.05 <0.01
NS, Not significant;BMI, bodymass index;T3, triiodothyronine;RIA, radioimmunoassay;LH, luteinizinghormone.
Table
II.
Serum levels of GHBP, IGF-I, IGFBP-3, and GH: Patients with anorexia nervosa versus control subjects Patients with anorexia nervosa (n = 28)
GHBP (pmol/L) iGF-I (ng/ml) IGFBP-3 (mg/L) GH (ng/ml)
Control subjects (n = 28)
118.5 +_ 49.5 136.9 + 50.4 3.3 _+ 1.8 3.9 + 3.2
292.0 ± 250.0 ± 5.0 + 3.0 ±
145.0 98.9 1.9 4.2
p
<0.001 <0.001 <0.01 NS
¥
Values (exceptp values) are mean _+SD, NS, Not significant.
tive tests, which suggests GH deficiency. 3' 13 Thus different physiologic changes are observed in different individuals. Growth hormone binding protein is a circulating protein, the structure of which is identical to the extracellular domain of the high affinity GH receptor.14 Serum levels of GHBP are low in states of GH resistance, t5, 16 Recently, Counts et al. ~7 found elevated basal GH levels and reduced IGF-1, insulin-like growth factor binding protein-3, and GHBP levels in 12 women with anorexia nervosa. They postulated that down-regulation of the GH receptor Or its postreceptor mechanism Occurred in nutritional deprivation associated with anorexia nervosa. This condition of relative GH resistance would explain the low IGF-I levels despite the elevated basal GH levels noted in their patient population. However, the patients studied were all adults, well past the age of active growth, whereas anorexia nervosais primarily a disease of adolescence that may persist into adulthood. Growth potential will be compromised only before fusion of the epiphyses, and maximal impact will occur when the disease develops before or at the time of the growth spurt. Only if the changes reported by Counts et al. are also present in younger patients with anorexia nervosa would there be an effect on growth potential. The objectives of this study were as follows: 1. To compare the serum levels of GH, IGF-1, IGFBP-3, and GHBP in malnourished adolescent patients with anorexia nervosa with those of a control group matched for age, sex, and Tanner pubertal stage
2. To correlate findings with measures of body composition, resting energy expenditure, and indexes of hypothalamic-pituitary function 3. To evaluate the effects of nutritional rehabilitation on serum GH, IGF-1, IGFBP-3 and GHBP levels METHODS Patients. The study population consisted of 28 female patients, 9 to 21 years of age (mean, 15.1 +_ 218 years) admitted to the Schneider Children's Hospital Eating Disorders Center for nutritional rehabilitation. All patients met the criteria of the Diagnostic and Statistical Manual o f Mental Disorders (third edition, revised) 18 for anorexia nervosa and weighed less than 85% of ideal body weight, according to the National Center for Health Statistics height and weight percentiles. 19 Of the 28 patients, 15 (53.6%) had been dieting for less than ayear. The control subjects were 28 healthy adolescents recruited during health maintenance visits at the adolescent medicine outpatient department; they were free of medical illness and were matched for age, sex, and pubertal stage. Written informed consent was obtained from each subject and from a parent when the subject was a minor. Study protocol. All patients and control patients underwent a physical examination, with special attention given to height, weight, sexual maturity rating, and growth percentiles. Standing height was measured with a stadiometer. Body weight was measured using a digital scale, with subjects undressed and wearing only underwear and a hospital
The Journal of Pediatrics Volume 125, Number 4
gown. Sexual maturity rating was determined by Tanner staging for breast development. 2° In those with anorexia nervosa, anthropometric estimates of lean body mass and adipose stores were obtained by a trained nutritionist using a nonstretchable nylon tape and a pair of Lange skin-fold calipers. Mid-arm muscle circumference (a measure of lean body mass) was derived from measurements of mid-arm circumference and triceps skin fold. Percentage of body fat was calculated from the sum of triceps, biceps, subscapular, and suprailiac crest skin-fold measurements by the method of Durnin and Rahaman.21 Body mass index was calculated as the weight in kilograms divided by the square of the height in meters. Within 72 hours of admission, indirect calorimetry was performed with the Datex DeltaTrac (SensorMedics Corp., Anaheim, Calif.) open-circuit, mixed-chamber collection system. After any preexisting dehydration was corrected and within 72 hours of admission, baseline fasting blood samples were drawn between 7 and 9 AM for GH, IGF-1, IGFBP-3, and GHBP. In 23 of the 28 patients with anorexia nervosa, repeat blood specimens were obtained at the end of hospitalization, usually when a weight within 10% of ideal body weight was achieved. All serum samples were separated and frozen at - 2 0 ° C within 12 hours of collection. Hormone measurements. Insulin-like growth factor I was measured by immunoradiometric assay and IGFBP-3 by radioimmunoassay with kits obtained from Diagnostic Systems Laboratories Inc. (Webster, Tex.). Interassay and intraassay coefficients of variation of both were 12% and 8%, respectively. GH was measured by radioimmunoassay and GHBP by ligand-mediated immunofunctional assay (Nichols Institute, San Juan Capistrano, Calif.). The interassay coefficient of variation for the GHBP assay was < 12% and intraassay coefficient of variation 3%. Estradiol, luteinizing hormone, and triiodothyronine were measured by radioimmunoassay with kits purchased from Diagnostic Products Corp. (Los Angeles, Calif.). Statistics. Serum levels of GH, IGF-1, IGFBP-3, and GHBP in patients with anorexia nervosa and control subjects were compared by means of the Student paired t test. When the Bartlett test of homogeneity of variance showed significant differences in variance between groups, Wilcoxon signed-rank test was used. The Student paired t test was used to assess changes in GH, IGF-1, 1GFBP-3, and GHBP levels in patients with anorexia nervosa before and after refeeding. The Pearson correlation coefficient was used to examine the relationships between levels of IGF-1, IGFBP-3, and GHBP and measures of body composition, REE, and indexes of hypothalamic-pituitary function. Multiple regression analysis was used to ascertain the relative influences of body mass index, percentage body fat, mid-arm muscle circumference, and
Golden et al.
800
i
i
I
657
I
r = 0 . 4 7 , p=O.O01 o o
400
e
e
300
e
|.0 ••
....
100
10
°o
-
•
0
. : .• o
I
I
I
L
12
14
16
18
20
BODY IdA$8 INDEX
10
i
t
t
t
r : 0 . 4 6 , p:O.O01 8
S" •
•
•
°... ;. O0
°
O0
2
0 10
I
I
I
I
12
14
le
18
20
BODY MA88 INDEX Fig. I. Correlation between GHBP, IGFBP-3, and body mass index in anorexia nervosa on admission (dark circles) and after partial weight restoration (clear circles).
REE on IGFBP-3 levels. All data are expressed as mean + SD. RESULTS Patients and control subjects were similar in age, race, Tanner pubertal stage, and height (Table I). All patients with anorexia nervosa had amenorrhea; 12 were premenarchal and 16 had secondary amenorrhea. Patients with anorexia nervosa had lost a mean of 13.0 + 6.0 kg and were significantly underweight compared with control subjects (74.8% ___ 10.7% vs 105.1% _+ 18.5% of ideal body weight; p <0.001). Mean percentage of body fat on admission was
6S8
Golden et al.
The Journal of Pediatrics October 1994
5°°I
I
+4"
~ 3oolo. Q. m 200 ,x.
T T
ioo
A
B
C
A
B
C
T=.
d.
A
B
C
Pig. 9. Serum levels (mean + SD) of GHBP, IGF-I, and IGFBP-3 in low-weightpatients with anorexianervosa(A) (n = 28), patients with anorexia nervosa after nutritional rehabilitation (tt) (n = 23), and age-matched healthy control subjects (C) (n = 28). Symbols: asterisk, A comparedwith C (p <0.01); double asterisks, A compared with C (p <0.001); plus sign, A compared with B (p <0.05); double plus signs, A compared with B (p <0.001); and dark triangle, B compared with C (p < 0.01).
14.5% +_ 4.4%. Mean REE was 930.4 +_ 180.3 kcal/day, significantly lower than the REE predicted for age, height, and weight by the Harris-Benedict equation (1239.6 +
84.1; p <0.001). As expected, levels of triiodothyronine, luteinizinghormone, and estradiol in patients with anorexia nervosa, were significantly lower than in the age-matched control subjects. Serum levels of GHBP, IGF-I, and IGFBP-3 were significantly lower in patients with anorexia nervosa on admission than in control subjects (Table II). A decrease in GHBP levels was present in patients in all five pubertal stages. Even though estradiol levels were lower in those with anorexia nervosa, there was no significant correlation between GHBP and estradiol levels. There were no significant differences in basal GH levels between the two groups. In the patients with anorexia nervosa, 6 (22.2%) of 27 subjects had basal GH levels >7 ng/ml, indicating adequate GH secretion. These 6 patients had had symptoms for 23.8 + 14.1 months versus 18.5 + 16.1 months (p = not significant) and were older (17.0 + 1.3 years vs 14.5 + 2.8 years; p <0.05) than those with basal GH levels <7 ng/ml. The patients with basal GH levels >7 ng/ml had a mean GHBP level of 121.5 + 79.4 pmol/L, which was not significantly different from that in those with basal GH levels <7 ng/ml (118.0 + 44.3 pmol/L). There was a positive correlation between body mass index and both GHBP (r = 0.47; p = 0.001) and IGFBP-3 (r = 0.46; p = 0.001) (Fig. 1). In those with anorexia nervosa, there was no correlation between GHBP levels and amount of weight lost, percentage of body fat, lean body mass, or REE. There was a significant correlation between IGFBP-3 levels and REE (r = 0.50; p <0.05), mid-arm muscle circumference (r = 0.63; p <0.01), and percentage of body fat (r = 0.65; p <0.01). Multiple regression analysis revealed that 50% of the variability of IGFBP-3 was accounted for by body mass index, percentage of body fat, and mid-arm muscle circumference. The REE did not add significantly to this model. After refeeding, patients with anorexia nervosa gained an average of 6.5 kg and increased their percentage of ideal body weight by 14.1%. Mean percentage of ideal body weight at follow-up was 89.8%. After nutritional repletion, mean serum levels of GHBP (262.4 + 87.3 pmol/ L), IGF-I (176.2 +_ 56.7 ng/ml), and IGFBP-3 (4.2 + 1.4 mg/L) increased significantly (Fig. 2). Serum GHBP and IGFBP-3 levels returned to the normal range but serum IGF-I levels remained lower than those in the age-matched control subjects (176.2 _+ 56.7 vs 250.0 + 98.7 ng/ml; p <0.01). DISCUSSION Growth hormone binding protein binds approximately 50% of circulating GH, 22 forming a complex that regulates the amount of free GH available at the cellular receptor. 23 In experimental animals the structure of GHBP is identical to that of the extracellular domain of the high-affinity he-
The Journal o f Pediatrics Volume 125, Number 4
patic cell GH receptor. 24-26In human beings, the structural relation between the hepatic cell receptor and circulating GHBP has not been confirmed, but the two proteins are immunologicaUysimilar.27 Circulating GHBP is thought to be derived from enzymatic cleavage of the liver membrane receptor 25 or alternate splicing of messenger ribonucleic acid coding for the receptor. 26 Low serum levels of GHBP are found in states of GH resistance and reflect decreased numbers of cellular GH receptors. 15' 16 In our study, patients with anorexia nervosa and low weight had reduced GHBP, IGF-I, and IGFBP-3 levels, which reverted to normal after nutritional rehabilitation, consistent with a transient state of diminished GH action. Diminished GH action could be caused by inadequate GH secretion or transient GH resistance or both. The high positive correlations with body mass index, together with the return to normal after refeeding, indicate that the changes noted for GHBP and IGFBP-3 levels are secondary to malnutrition. These findings are in agreement with those of Counts et al. ~7 and Hochberg et al. 28 We found low levels of GHBP at all stages of pubertal development, including those stages associated with peak growth velocity. The occurrence of a state of growth hormone insensitivity during the years of active growth can compromise growth potential. The ability to reverse this state by nutritional rehabilitation emphasizes the need for aggressive intervention in a patient with anorexia nervosa in early adolescence. Diminished GHBP levels may result from decreased numbers of GH receptors, altered production of the protein by the liver, or increased clearance of the circulating protein. Controversy remains as to whether GHBP levels are influenced by GH secretion; some studies have shown normal levels of GHBP in GH deficiency,29, 30 and others, low levels of GHBP in GH deficiency.31 Diminished GH secretion may occur in adolescents at an early stage of the disease and may coexist with GH resistance. Although random unstimulated GH levels <7 ng/ml cannot be taken as proof of GH deficiency, random levels >7 ng/ml (as found in six of our patients) are evidence of adequate GH secretion. The finding of low GHBP levels in the presence of normal or elevated GH levels strongly suggests GH resistance. Mean basal levels of GH in our malnourished patients with anorexia nervosa were not significantly different from those of control subjects, which confirms the results of Nussbaum et al., 13 who found that basal GH levels were normal or low in adolescents with anorexia nervosa who had a relatively short duration of symptoms. Studies of older patients with anorexia nervosa have generally found mean basal GH levels to be higher than in control subjects. In our study, all six patients with basal GH levels >7 ng/ml were 16 years of age or older and had had symptoms for a relatively long period. They did not differ in degree of malnutrition from those with basal GH levels <7 ng/ml. We pos-
Golden et al.
659
tulate that diminished GH secretion may be an early adaptive response to malnutrition in anorexia nervosa and is followed by a state of GH resistance in older patients with more long-standing symptoms. Insulin-like growth factor I is a GH-dependent peptide whose levels would be expected to be low in states of either GH resistance or deficiency. IGF-I mediates much of the growth-promoting effects of GH, including cell proliferation and protein synthesis. The peptide is produced primarily in the liver but also locally by other tissues. IGFBP-3 is regulated both by GH secretion and nutrition.32 Our results demonstrate that IGFBP-3 levels were low in malnourished patients with anorexia nervosa but returned to normal after refeeding. Low IGFBP-3 levels do not help distinguish between GH deficiency and GH resistance, but they do reflect disturbances in the distal GH axis seen in anorexia nervosa. Our results confirm that IGFBP-3 is a good indicator of nutritional status because there was a good correlation between IGFBP-3 levels and not only body mass index but also measures of REE, lean body mass, and percentage of body fat. Levels of GHBP, on the other hand, did not correlate with REE or anthropometric measurements, suggesting that GH resistance in anorexia nervosa may be a long-term protective response to starvation but not necessarily an indicator of current nutritional status. We conclude that young patients with anorexia nervosa in the malnourished state have low levels of GHBP, IGF-I, and IGFBP-3; GHBP levels are low at all five stages of pubertal development and even in those patients found to have adequate GH secretion. Our results support the concept that basal GH levels in young adolescents with anorexia nervosa differ from those of older subjects with this disorder. These findings are consistent with diminished GH action occurring in anorexia nervosa. Fortunately, most patients with anorexia nervosa have final heights in the normal range, though some may be shorter than their genetic potential. The ability to reverse these findings with nutritional rehabilitation emphasizes the need for aggressive treatment to prevent potentially irreversible growth retardation and to maximize height potential. We thank Michelle Kaplan, BS, Raji Dave, MS, and Michael Davis, BS, for their technical assistance; Rosemarie St. Victor, MD, and Stephanie Levine,DO, for help with specimencollection; and Robert Bienkowski,PhD, for his critical review of the manuscript. REFERENCES
1. Root AW, Powers PS. Anorexia nervosa presenting as growth retardation in adolescents. J Adolesc Health Care 1983;4:2530. 2. Nussbaum M, Baird D, Sonnenblick M, et al. Short stature in anorexia ncrvosa patients. J AdolescHealth Care 1985; 6:453-5.
660
Golden et al.
3. Russell GFM. Premenarchal anorexia nervosa and its sequelae. J Psychiatr Res 1985;19:363-9. 4. Huseman C, Johanson A. Growth hormone deficiency in anorexia nervosa. J PEDIATR 1975;87:946-8. 5. Rappaport R, Prevot C, Czernichow P. Somatomedin activity and growth hormone secretion: changes related to body weight in anorexia nervosa. Aeta Paediatr Scand 1980;69:3741. 6. Clemmons DR, Libanski A, Underwood LE, et al. Reduction of plasma immunoreactive somatomedin C during fasting in humans J Clin Endocrinol Metab 1981;53:1247-50. 7. Underwood LE, Clemmons DR, Maes M, D'Ercole A J, Ketelslegers J-M. Regulation of somatomedin-C/insulin-like growth factor I by nutrients. Hormone Res 1986;24:166-76. 8. Grant DB, Hambley J, Becker D, Pimstone BL. Reduced sulphation factor in undernourished children. Arch Dis Child 1973;48:596-600. 9. Maeda K, Kato Y, Yamaguchi N, et al. Growth hormone release following thyrotropin-releasing hormone injection into patients with anorexia nervosa. Acta Endocrinol (Copenh) 1976;81:1-8. 10. Hurd HP, Palumbo P J, Gharib H. Hypothalamic-endocrine dysfunction in anorexia nervosa. Mayo Clin Proc 1977;52: 711-6. 11. Masuda A, Shibasaki T, Hotta M, Suematsu H, Shizume K. Study on the mechanism of abnormal growth hormone (GH) secretion in anorexia nervosa: no evidence of involvement of a low somatomedin-C level in the abnormal GH secretion. J Endocrinol Invest 1988;11:297-302. 12. Casper RL, Davis JM, Pandey GN. The effect of nutritional status and weight change on hypothalamic function tests in anorexia nervosa. In: Vigersky R, ed. Anorexia nervosa. New York: Raven Press, 1977:137. 13. Nussbaum MP, Blethen SL, Chasalow FI, et al. Blunted growth hormone responses to clonidine in adolescent girls with early anorexia nervosa: evidence for an early hypothalamic defect. J Adolesc Health Care 1990;11:145-8. 14. Daughaday WH, Trivedi B. Clinical aspects of GH binding proteins. Acta Endocrinol (Copenh) 1991;124:27-32. 15. Daughaday WH, Trivedi B. Absence of serum growth hormone binding protein in patients with growth hormone receptor deficiency (Laron dwarfism). Proc Natl Acad Sci USA 1987;84:4636-40. 16. Baumann G, Shaw MA, Winter RJ. Absence of the plasma growth hormone-binding protein in Laron-type dwarfism. J Clin Endocrinol Metab 1987;65:814-6. 17. Counts DR, Gwirtsman H, Carlsson LMS, Lesem M, Cutler G. The effects of anorexia nervosa and refeeding on growth hormone--binding protein, the insulin-like growth factors (IGFs), and the IGF-binding proteins. J Clin Endocrinol Metab 1992;75:762-6. 18. American Psychiatric Association. Diagnostic and statistical
The Journal o f Pediatrics October 1994
19.
20. 21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
manual of mental disorders. 3rd ed, revised. Washington, D.C.: The Association, 1987. National Center for Health Statistics. Height and weight of youths 12-17 years, United States. Vital and Health Statistics, series 11, No. 124. Washington, D.C.: U.S. Public Health Service, 1973. Tanner JM. Growth at adolescence. Oxford: Blackwell Scientific Publications, 1962. Durnin JVGA, Rahaman MM. The assessment of the amount of body fat in the human body from measurements of skinfold thickness. Br J Nutr 1967;21:681-9. Baumann G, Amburn KD, Shaw MA. The circulating growth hormone-binding protein complex: a major constituent of plasma growth hormone in man. Endocrinology 1988; 122:97684. Baumann G, Amburn KD, Buchanan TA. The effects of circulating growth hormone-binding protein on metabolic clearance, distribution, and degradation of human growth hormone. J Clin Endocrinol Metab 1987;64:657-60. Smith WC, Kuniyoshi J, Talamantes F. Mouse serum growth hormone (GH) binding protein has GH receptor extracellular and substituted transmembrane domains. Mol EndoerinoI 1989;3:984-90. Spencer SA, Hammonds RG, Henzel W J, et al. Rabbit liver growth hormone receptor and binding protein: purification, characterization and sequence. J Biol Chem 1988;263:7862-3. Baumbach WR, Homer DI, Logan JS. The growth hormone binding protein in rat serum is an alternatively spliced form of the rat growth hormone receptor. Genes Dev 1989;3:11991205. Hocquette J-F, Postel-Vinay M-C, Djiane J, Tar A, Kelley PA. Human liver growth hormone receptor and plasma binding protein: characterization and partial purification. Endocrinology 1990;127:1665-72. Hochberg Z, Hertz P, Colin V, et al. The distal axis of growth hormone (GH) in nutritional disorders: GH-binding protein, insulin-like growth factor-I (IGF-I) and IGF-I receptors in obesity and anorexia nervosa. Metabolism 1992;41:106-12. Hochberg Z, Barkley R J, Even L, Peleg I, Youdim MBH, Amit T. The effects of growth hormone therapy on GH binding protein in GH-deficient children, Acta Endocrinol 1991;125:23-7. Ken KYH, Erefili V, Waters M J, Rajkovic IA. Regulation of growth hormone binding protein in man: comparison of gel chromatography and immunoprecipitation methods. J Clin Endocrinol Metab 1993;76:302-8. Postel-Vinay MC, Tar A, Hocquette J-F, et al. Human plasma growth hormone (GH)-binding proteins are regulated by GH and testosterone. J Clin Endocrinol Metab 1991;73:197-202. Cohen P, Fielder P J, Hasegawa Y, Frisch H, Giudice LC, Rosenfeld RG. Clinical aspects of insulin-like growth factor binding proteins. Aeta Endoerinol (Copenh) 1991 ;124:74-85.
Disturbances in growth hormone secretion and action in adolescents with anorexia nervosa Neville H. G o l d e n , MD, Paula Kreitzer, MD, M a r c S. J a c o b s o n , MD, Fred I. C h a s a l o w , PhD, J a n e t S c h e b e n d a c h , MA, RD, Samuel M. F r e e d m a n , MD, a n d I. Ronald Shenker, MD From the Department of Pediatrics, Divisions of Adolescent Medicine and Pediatric Endocrinology, Schneider Children's Hospital of Long Island Jewish Medical Center, Albert Einstein College of Medicine, New Hyde Park, New York, and the Division of Pediatric Endocrinology, Maimonides Medical Center, Brooklyn, New York Women in whom anorexia nervosa d e v e l o p s during a d o l e s c e n c e have failure of linear growth associated with low levels of insulin-like growth factor I (IGF-I). To investigate the p a t h o p h y s i a l o g y of growth retardation in adolescents with anorexia nervosa, we measured basal growth hormone (GH), growth h o r m o n e binding protein (GHBP), IGF-I, and insulin-like growth factor binding protein-3 (IGFBP-3) in three groups of patients: (I) 28 recently hospitalized female adolescents with anorexia nervosa, (2) 23 of the same patients after partial weight restoration, and (3) 28 healthy control subjects m a t c h e d for age, sex, and pubertal stage. Fasting GH levels in group I did nat differ significantly from those in group 3. In contrast, serum GHBP (p <0.001), IGF-I (p <0.001), and IGFBP-3 (p <0.01) were significantly lower in group t than in group 3. Serum GHBP and IGFBP-3 levels were positively correlated with b o d y mass index. Serum GHBP levels were low in patients in all five pubertal stages and even in those shown to have a d e q u a t e GH secretion. In group 2 (after refeeding) the serum IGF-I concentration increased significantly and GHBP and IGFBP-3 returned to normal. We c o n c l u d e that patients with anorexia nervosa have diminished GH action resulting in decreased secretion of IGF-I. The positive correlation with b o d y mass index and the reversibility with refeeding suggest that these changes are s e c o n d a r y to malnutrition. Altered GH function that occurs during the years of active growth can e x p l a i n the growth retardation seen in anorexia n e r v o s a . (J PEDIATR 1994;125:655-60)
If anorexia nervosa develops during childhood or early adolescence, failure of linear growth can occur, resulting in decreased final adult height,~4 The cause of growth failure in anorexia nervosa has not been elucidated. Decreased levPresented in part at the Society for Pediatric Research, Washington, D.C., May 1993. Submitted for publication Feb. 27, 1994; accepted April 14, 1994. Reprint requests: Neville H. Golden, MD, Divisionof Adolescent Medicinel Room 187, Schneider Children's Hospital, Long Island Jewish Medical Center, New Hyde Park, NY 11042. Copyright © 1994 by Mosby-Year Book, Inc. 0022-3476/94/$3.00 + 0 9/27/56783
els of a growth hormone-dependent peptide, insulin-like growth factor I, have been found in anorexia nervosa, 5 as well as during fasting6, 7 and in other forms of malnutrition.8 In 35% to 50% of patients with anorexia nervosa and low GH GHBP IGF-I IGFBP-3 REE
Growth hormone Growth hormone binding protein Insulin-like growth factor I Insulin-like growth factor binding protein-3 Resting energy expenditure
weight, there are elevated basal levels of growth hormone 9"12 suggestive of GH resistance. However, other patients have inadequate GH secretory responses to provoca655
656
Golden et al.
The Journal of Pediatrics October 1994
T a b l e I. Clinical features: Patients with anorexia nervosa versus control subjects
Age (yr) Tanner stage Height (cm) Weight (kg) BMI (kg/mz) T3 RIA (ng/dl) LH (mIU/rnl) Estradiol (pg/ml)
Patients with anorexia nervosa (n = 28)
Control subjects (n = 28)
p
15.1 ± 2.8 3.9 ± 1.2 157.5 ± 10.2 37.5 ± 7.8 15.0 _+ 2.1 91.7 ± 35.1 3.6 ± 3.1 13.0 ± 8.4
15.0 ± 2.6 3.9 ± 1.3 159.5 ± 10.7 54.2 _+ 13.6 21.1 ± 3.8 124.6 ± 29.3 10.7 ± 14.8 53.7 ± 48.1
NS NS NS <0.001 <0.001 <0.001 <0.05 <0.01
NS, Not significant;BMI, bodymass index;T3, triiodothyronine;RIA, radioimmunoassay;LH, luteinizinghormone.
Table
II.
Serum levels of GHBP, IGF-I, IGFBP-3, and GH: Patients with anorexia nervosa versus control subjects Patients with anorexia nervosa (n = 28)
GHBP (pmol/L) iGF-I (ng/ml) IGFBP-3 (mg/L) GH (ng/ml)
Control subjects (n = 28)
118.5 +_ 49.5 136.9 + 50.4 3.3 _+ 1.8 3.9 + 3.2
292.0 ± 250.0 ± 5.0 + 3.0 ±
145.0 98.9 1.9 4.2
p
<0.001 <0.001 <0.01 NS
¥
Values (exceptp values) are mean _+SD, NS, Not significant.
tive tests, which suggests GH deficiency. 3' 13 Thus different physiologic changes are observed in different individuals. Growth hormone binding protein is a circulating protein, the structure of which is identical to the extracellular domain of the high affinity GH receptor.14 Serum levels of GHBP are low in states of GH resistance, t5, 16 Recently, Counts et al. ~7 found elevated basal GH levels and reduced IGF-1, insulin-like growth factor binding protein-3, and GHBP levels in 12 women with anorexia nervosa. They postulated that down-regulation of the GH receptor Or its postreceptor mechanism Occurred in nutritional deprivation associated with anorexia nervosa. This condition of relative GH resistance would explain the low IGF-I levels despite the elevated basal GH levels noted in their patient population. However, the patients studied were all adults, well past the age of active growth, whereas anorexia nervosais primarily a disease of adolescence that may persist into adulthood. Growth potential will be compromised only before fusion of the epiphyses, and maximal impact will occur when the disease develops before or at the time of the growth spurt. Only if the changes reported by Counts et al. are also present in younger patients with anorexia nervosa would there be an effect on growth potential. The objectives of this study were as follows: 1. To compare the serum levels of GH, IGF-1, IGFBP-3, and GHBP in malnourished adolescent patients with anorexia nervosa with those of a control group matched for age, sex, and Tanner pubertal stage
2. To correlate findings with measures of body composition, resting energy expenditure, and indexes of hypothalamic-pituitary function 3. To evaluate the effects of nutritional rehabilitation on serum GH, IGF-1, IGFBP-3 and GHBP levels METHODS Patients. The study population consisted of 28 female patients, 9 to 21 years of age (mean, 15.1 +_ 218 years) admitted to the Schneider Children's Hospital Eating Disorders Center for nutritional rehabilitation. All patients met the criteria of the Diagnostic and Statistical Manual o f Mental Disorders (third edition, revised) 18 for anorexia nervosa and weighed less than 85% of ideal body weight, according to the National Center for Health Statistics height and weight percentiles. 19 Of the 28 patients, 15 (53.6%) had been dieting for less than ayear. The control subjects were 28 healthy adolescents recruited during health maintenance visits at the adolescent medicine outpatient department; they were free of medical illness and were matched for age, sex, and pubertal stage. Written informed consent was obtained from each subject and from a parent when the subject was a minor. Study protocol. All patients and control patients underwent a physical examination, with special attention given to height, weight, sexual maturity rating, and growth percentiles. Standing height was measured with a stadiometer. Body weight was measured using a digital scale, with subjects undressed and wearing only underwear and a hospital
The Journal of Pediatrics Volume 125, Number 4
gown. Sexual maturity rating was determined by Tanner staging for breast development. 2° In those with anorexia nervosa, anthropometric estimates of lean body mass and adipose stores were obtained by a trained nutritionist using a nonstretchable nylon tape and a pair of Lange skin-fold calipers. Mid-arm muscle circumference (a measure of lean body mass) was derived from measurements of mid-arm circumference and triceps skin fold. Percentage of body fat was calculated from the sum of triceps, biceps, subscapular, and suprailiac crest skin-fold measurements by the method of Durnin and Rahaman.21 Body mass index was calculated as the weight in kilograms divided by the square of the height in meters. Within 72 hours of admission, indirect calorimetry was performed with the Datex DeltaTrac (SensorMedics Corp., Anaheim, Calif.) open-circuit, mixed-chamber collection system. After any preexisting dehydration was corrected and within 72 hours of admission, baseline fasting blood samples were drawn between 7 and 9 AM for GH, IGF-1, IGFBP-3, and GHBP. In 23 of the 28 patients with anorexia nervosa, repeat blood specimens were obtained at the end of hospitalization, usually when a weight within 10% of ideal body weight was achieved. All serum samples were separated and frozen at - 2 0 ° C within 12 hours of collection. Hormone measurements. Insulin-like growth factor I was measured by immunoradiometric assay and IGFBP-3 by radioimmunoassay with kits obtained from Diagnostic Systems Laboratories Inc. (Webster, Tex.). Interassay and intraassay coefficients of variation of both were 12% and 8%, respectively. GH was measured by radioimmunoassay and GHBP by ligand-mediated immunofunctional assay (Nichols Institute, San Juan Capistrano, Calif.). The interassay coefficient of variation for the GHBP assay was < 12% and intraassay coefficient of variation 3%. Estradiol, luteinizing hormone, and triiodothyronine were measured by radioimmunoassay with kits purchased from Diagnostic Products Corp. (Los Angeles, Calif.). Statistics. Serum levels of GH, IGF-1, IGFBP-3, and GHBP in patients with anorexia nervosa and control subjects were compared by means of the Student paired t test. When the Bartlett test of homogeneity of variance showed significant differences in variance between groups, Wilcoxon signed-rank test was used. The Student paired t test was used to assess changes in GH, IGF-1, 1GFBP-3, and GHBP levels in patients with anorexia nervosa before and after refeeding. The Pearson correlation coefficient was used to examine the relationships between levels of IGF-1, IGFBP-3, and GHBP and measures of body composition, REE, and indexes of hypothalamic-pituitary function. Multiple regression analysis was used to ascertain the relative influences of body mass index, percentage body fat, mid-arm muscle circumference, and
Golden et al.
800
i
i
I
657
I
r = 0 . 4 7 , p=O.O01 o o
400
e
e
300
e
|.0 ••
....
100
10
°o
-
•
0
. : .• o
I
I
I
L
12
14
16
18
20
BODY IdA$8 INDEX
10
i
t
t
t
r : 0 . 4 6 , p:O.O01 8
S" •
•
•
°... ;. O0
°
O0
2
0 10
I
I
I
I
12
14
le
18
20
BODY MA88 INDEX Fig. I. Correlation between GHBP, IGFBP-3, and body mass index in anorexia nervosa on admission (dark circles) and after partial weight restoration (clear circles).
REE on IGFBP-3 levels. All data are expressed as mean + SD. RESULTS Patients and control subjects were similar in age, race, Tanner pubertal stage, and height (Table I). All patients with anorexia nervosa had amenorrhea; 12 were premenarchal and 16 had secondary amenorrhea. Patients with anorexia nervosa had lost a mean of 13.0 + 6.0 kg and were significantly underweight compared with control subjects (74.8% ___ 10.7% vs 105.1% _+ 18.5% of ideal body weight; p <0.001). Mean percentage of body fat on admission was
6S8
Golden et al.
The Journal of Pediatrics October 1994
5°°I
I
+4"
~ 3oolo. Q. m 200 ,x.
T T
ioo
A
B
C
A
B
C
T=.
d.
A
B
C
Pig. 9. Serum levels (mean + SD) of GHBP, IGF-I, and IGFBP-3 in low-weightpatients with anorexianervosa(A) (n = 28), patients with anorexia nervosa after nutritional rehabilitation (tt) (n = 23), and age-matched healthy control subjects (C) (n = 28). Symbols: asterisk, A comparedwith C (p <0.01); double asterisks, A compared with C (p <0.001); plus sign, A compared with B (p <0.05); double plus signs, A compared with B (p <0.001); and dark triangle, B compared with C (p < 0.01).
14.5% +_ 4.4%. Mean REE was 930.4 +_ 180.3 kcal/day, significantly lower than the REE predicted for age, height, and weight by the Harris-Benedict equation (1239.6 +
84.1; p <0.001). As expected, levels of triiodothyronine, luteinizinghormone, and estradiol in patients with anorexia nervosa, were significantly lower than in the age-matched control subjects. Serum levels of GHBP, IGF-I, and IGFBP-3 were significantly lower in patients with anorexia nervosa on admission than in control subjects (Table II). A decrease in GHBP levels was present in patients in all five pubertal stages. Even though estradiol levels were lower in those with anorexia nervosa, there was no significant correlation between GHBP and estradiol levels. There were no significant differences in basal GH levels between the two groups. In the patients with anorexia nervosa, 6 (22.2%) of 27 subjects had basal GH levels >7 ng/ml, indicating adequate GH secretion. These 6 patients had had symptoms for 23.8 + 14.1 months versus 18.5 + 16.1 months (p = not significant) and were older (17.0 + 1.3 years vs 14.5 + 2.8 years; p <0.05) than those with basal GH levels <7 ng/ml. The patients with basal GH levels >7 ng/ml had a mean GHBP level of 121.5 + 79.4 pmol/L, which was not significantly different from that in those with basal GH levels <7 ng/ml (118.0 + 44.3 pmol/L). There was a positive correlation between body mass index and both GHBP (r = 0.47; p = 0.001) and IGFBP-3 (r = 0.46; p = 0.001) (Fig. 1). In those with anorexia nervosa, there was no correlation between GHBP levels and amount of weight lost, percentage of body fat, lean body mass, or REE. There was a significant correlation between IGFBP-3 levels and REE (r = 0.50; p <0.05), mid-arm muscle circumference (r = 0.63; p <0.01), and percentage of body fat (r = 0.65; p <0.01). Multiple regression analysis revealed that 50% of the variability of IGFBP-3 was accounted for by body mass index, percentage of body fat, and mid-arm muscle circumference. The REE did not add significantly to this model. After refeeding, patients with anorexia nervosa gained an average of 6.5 kg and increased their percentage of ideal body weight by 14.1%. Mean percentage of ideal body weight at follow-up was 89.8%. After nutritional repletion, mean serum levels of GHBP (262.4 + 87.3 pmol/ L), IGF-I (176.2 +_ 56.7 ng/ml), and IGFBP-3 (4.2 + 1.4 mg/L) increased significantly (Fig. 2). Serum GHBP and IGFBP-3 levels returned to the normal range but serum IGF-I levels remained lower than those in the age-matched control subjects (176.2 _+ 56.7 vs 250.0 + 98.7 ng/ml; p <0.01). DISCUSSION Growth hormone binding protein binds approximately 50% of circulating GH, 22 forming a complex that regulates the amount of free GH available at the cellular receptor. 23 In experimental animals the structure of GHBP is identical to that of the extracellular domain of the high-affinity he-
The Journal o f Pediatrics Volume 125, Number 4
patic cell GH receptor. 24-26In human beings, the structural relation between the hepatic cell receptor and circulating GHBP has not been confirmed, but the two proteins are immunologicaUysimilar.27 Circulating GHBP is thought to be derived from enzymatic cleavage of the liver membrane receptor 25 or alternate splicing of messenger ribonucleic acid coding for the receptor. 26 Low serum levels of GHBP are found in states of GH resistance and reflect decreased numbers of cellular GH receptors. 15' 16 In our study, patients with anorexia nervosa and low weight had reduced GHBP, IGF-I, and IGFBP-3 levels, which reverted to normal after nutritional rehabilitation, consistent with a transient state of diminished GH action. Diminished GH action could be caused by inadequate GH secretion or transient GH resistance or both. The high positive correlations with body mass index, together with the return to normal after refeeding, indicate that the changes noted for GHBP and IGFBP-3 levels are secondary to malnutrition. These findings are in agreement with those of Counts et al. ~7 and Hochberg et al. 28 We found low levels of GHBP at all stages of pubertal development, including those stages associated with peak growth velocity. The occurrence of a state of growth hormone insensitivity during the years of active growth can compromise growth potential. The ability to reverse this state by nutritional rehabilitation emphasizes the need for aggressive intervention in a patient with anorexia nervosa in early adolescence. Diminished GHBP levels may result from decreased numbers of GH receptors, altered production of the protein by the liver, or increased clearance of the circulating protein. Controversy remains as to whether GHBP levels are influenced by GH secretion; some studies have shown normal levels of GHBP in GH deficiency,29, 30 and others, low levels of GHBP in GH deficiency.31 Diminished GH secretion may occur in adolescents at an early stage of the disease and may coexist with GH resistance. Although random unstimulated GH levels <7 ng/ml cannot be taken as proof of GH deficiency, random levels >7 ng/ml (as found in six of our patients) are evidence of adequate GH secretion. The finding of low GHBP levels in the presence of normal or elevated GH levels strongly suggests GH resistance. Mean basal levels of GH in our malnourished patients with anorexia nervosa were not significantly different from those of control subjects, which confirms the results of Nussbaum et al., 13 who found that basal GH levels were normal or low in adolescents with anorexia nervosa who had a relatively short duration of symptoms. Studies of older patients with anorexia nervosa have generally found mean basal GH levels to be higher than in control subjects. In our study, all six patients with basal GH levels >7 ng/ml were 16 years of age or older and had had symptoms for a relatively long period. They did not differ in degree of malnutrition from those with basal GH levels <7 ng/ml. We pos-
Golden et al.
659
tulate that diminished GH secretion may be an early adaptive response to malnutrition in anorexia nervosa and is followed by a state of GH resistance in older patients with more long-standing symptoms. Insulin-like growth factor I is a GH-dependent peptide whose levels would be expected to be low in states of either GH resistance or deficiency. IGF-I mediates much of the growth-promoting effects of GH, including cell proliferation and protein synthesis. The peptide is produced primarily in the liver but also locally by other tissues. IGFBP-3 is regulated both by GH secretion and nutrition.32 Our results demonstrate that IGFBP-3 levels were low in malnourished patients with anorexia nervosa but returned to normal after refeeding. Low IGFBP-3 levels do not help distinguish between GH deficiency and GH resistance, but they do reflect disturbances in the distal GH axis seen in anorexia nervosa. Our results confirm that IGFBP-3 is a good indicator of nutritional status because there was a good correlation between IGFBP-3 levels and not only body mass index but also measures of REE, lean body mass, and percentage of body fat. Levels of GHBP, on the other hand, did not correlate with REE or anthropometric measurements, suggesting that GH resistance in anorexia nervosa may be a long-term protective response to starvation but not necessarily an indicator of current nutritional status. We conclude that young patients with anorexia nervosa in the malnourished state have low levels of GHBP, IGF-I, and IGFBP-3; GHBP levels are low at all five stages of pubertal development and even in those patients found to have adequate GH secretion. Our results support the concept that basal GH levels in young adolescents with anorexia nervosa differ from those of older subjects with this disorder. These findings are consistent with diminished GH action occurring in anorexia nervosa. Fortunately, most patients with anorexia nervosa have final heights in the normal range, though some may be shorter than their genetic potential. The ability to reverse these findings with nutritional rehabilitation emphasizes the need for aggressive treatment to prevent potentially irreversible growth retardation and to maximize height potential. We thank Michelle Kaplan, BS, Raji Dave, MS, and Michael Davis, BS, for their technical assistance; Rosemarie St. Victor, MD, and Stephanie Levine,DO, for help with specimencollection; and Robert Bienkowski,PhD, for his critical review of the manuscript. REFERENCES
1. Root AW, Powers PS. Anorexia nervosa presenting as growth retardation in adolescents. J Adolesc Health Care 1983;4:2530. 2. Nussbaum M, Baird D, Sonnenblick M, et al. Short stature in anorexia ncrvosa patients. J AdolescHealth Care 1985; 6:453-5.
660
Golden et al.
3. Russell GFM. Premenarchal anorexia nervosa and its sequelae. J Psychiatr Res 1985;19:363-9. 4. Huseman C, Johanson A. Growth hormone deficiency in anorexia nervosa. J PEDIATR 1975;87:946-8. 5. Rappaport R, Prevot C, Czernichow P. Somatomedin activity and growth hormone secretion: changes related to body weight in anorexia nervosa. Aeta Paediatr Scand 1980;69:3741. 6. Clemmons DR, Libanski A, Underwood LE, et al. Reduction of plasma immunoreactive somatomedin C during fasting in humans J Clin Endocrinol Metab 1981;53:1247-50. 7. Underwood LE, Clemmons DR, Maes M, D'Ercole A J, Ketelslegers J-M. Regulation of somatomedin-C/insulin-like growth factor I by nutrients. Hormone Res 1986;24:166-76. 8. Grant DB, Hambley J, Becker D, Pimstone BL. Reduced sulphation factor in undernourished children. Arch Dis Child 1973;48:596-600. 9. Maeda K, Kato Y, Yamaguchi N, et al. Growth hormone release following thyrotropin-releasing hormone injection into patients with anorexia nervosa. Acta Endocrinol (Copenh) 1976;81:1-8. 10. Hurd HP, Palumbo P J, Gharib H. Hypothalamic-endocrine dysfunction in anorexia nervosa. Mayo Clin Proc 1977;52: 711-6. 11. Masuda A, Shibasaki T, Hotta M, Suematsu H, Shizume K. Study on the mechanism of abnormal growth hormone (GH) secretion in anorexia nervosa: no evidence of involvement of a low somatomedin-C level in the abnormal GH secretion. J Endocrinol Invest 1988;11:297-302. 12. Casper RL, Davis JM, Pandey GN. The effect of nutritional status and weight change on hypothalamic function tests in anorexia nervosa. In: Vigersky R, ed. Anorexia nervosa. New York: Raven Press, 1977:137. 13. Nussbaum MP, Blethen SL, Chasalow FI, et al. Blunted growth hormone responses to clonidine in adolescent girls with early anorexia nervosa: evidence for an early hypothalamic defect. J Adolesc Health Care 1990;11:145-8. 14. Daughaday WH, Trivedi B. Clinical aspects of GH binding proteins. Acta Endocrinol (Copenh) 1991;124:27-32. 15. Daughaday WH, Trivedi B. Absence of serum growth hormone binding protein in patients with growth hormone receptor deficiency (Laron dwarfism). Proc Natl Acad Sci USA 1987;84:4636-40. 16. Baumann G, Shaw MA, Winter RJ. Absence of the plasma growth hormone-binding protein in Laron-type dwarfism. J Clin Endocrinol Metab 1987;65:814-6. 17. Counts DR, Gwirtsman H, Carlsson LMS, Lesem M, Cutler G. The effects of anorexia nervosa and refeeding on growth hormone--binding protein, the insulin-like growth factors (IGFs), and the IGF-binding proteins. J Clin Endocrinol Metab 1992;75:762-6. 18. American Psychiatric Association. Diagnostic and statistical
The Journal o f Pediatrics October 1994
19.
20. 21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
manual of mental disorders. 3rd ed, revised. Washington, D.C.: The Association, 1987. National Center for Health Statistics. Height and weight of youths 12-17 years, United States. Vital and Health Statistics, series 11, No. 124. Washington, D.C.: U.S. Public Health Service, 1973. Tanner JM. Growth at adolescence. Oxford: Blackwell Scientific Publications, 1962. Durnin JVGA, Rahaman MM. The assessment of the amount of body fat in the human body from measurements of skinfold thickness. Br J Nutr 1967;21:681-9. Baumann G, Amburn KD, Shaw MA. The circulating growth hormone-binding protein complex: a major constituent of plasma growth hormone in man. Endocrinology 1988; 122:97684. Baumann G, Amburn KD, Buchanan TA. The effects of circulating growth hormone-binding protein on metabolic clearance, distribution, and degradation of human growth hormone. J Clin Endocrinol Metab 1987;64:657-60. Smith WC, Kuniyoshi J, Talamantes F. Mouse serum growth hormone (GH) binding protein has GH receptor extracellular and substituted transmembrane domains. Mol EndoerinoI 1989;3:984-90. Spencer SA, Hammonds RG, Henzel W J, et al. Rabbit liver growth hormone receptor and binding protein: purification, characterization and sequence. J Biol Chem 1988;263:7862-3. Baumbach WR, Homer DI, Logan JS. The growth hormone binding protein in rat serum is an alternatively spliced form of the rat growth hormone receptor. Genes Dev 1989;3:11991205. Hocquette J-F, Postel-Vinay M-C, Djiane J, Tar A, Kelley PA. Human liver growth hormone receptor and plasma binding protein: characterization and partial purification. Endocrinology 1990;127:1665-72. Hochberg Z, Hertz P, Colin V, et al. The distal axis of growth hormone (GH) in nutritional disorders: GH-binding protein, insulin-like growth factor-I (IGF-I) and IGF-I receptors in obesity and anorexia nervosa. Metabolism 1992;41:106-12. Hochberg Z, Barkley R J, Even L, Peleg I, Youdim MBH, Amit T. The effects of growth hormone therapy on GH binding protein in GH-deficient children, Acta Endocrinol 1991;125:23-7. Ken KYH, Erefili V, Waters M J, Rajkovic IA. Regulation of growth hormone binding protein in man: comparison of gel chromatography and immunoprecipitation methods. J Clin Endocrinol Metab 1993;76:302-8. Postel-Vinay MC, Tar A, Hocquette J-F, et al. Human plasma growth hormone (GH)-binding proteins are regulated by GH and testosterone. J Clin Endocrinol Metab 1991;73:197-202. Cohen P, Fielder P J, Hasegawa Y, Frisch H, Giudice LC, Rosenfeld RG. Clinical aspects of insulin-like growth factor binding proteins. Aeta Endoerinol (Copenh) 1991 ;124:74-85.