Leptin Levels in Protracted Critical Illness: Effects of Growth Hormone ...

0021-972X/98/$03.00/0 Journal of Clinical Endocrinology and Metabolism Copyright © 1998 by The Endocrine Society

Vol. 83, No. 9 Printed in U.S.A.

Leptin Levels in Protracted Critical Illness: Effects of Growth Hormone-Secretagogues and ThyrotropinReleasing Hormone* GREET VAN DEN BERGHE†, PIETER WOUTERS, LENA CARLSSON, ROBERT C. BAXTER, ROGER BOUILLON, AND CYRIL Y. BOWERS Departments of Intensive Care Medicine (G.V.d.B., P.W.) and Medicine, Division of Endocrinology (R.B.), University Hospital Gasthuisberg, University of Leuven, B-3000 Leuven, Belgium; Research Center for Endocrinology and Metabolism (L.C.), Department of Internal Medicine, Sahlgrenska University Hospital, 5-41345 Go¨teborg, Sweden; Kolling Institute (R.C.B.), University of Sydney, SW2065 St. Leonards, Australia; and Department of Medicine (C.Y.B.), Division of Endocrinology, Tulane University Medical Center, New Orleans, Louisiana 70112-2699 ABSTRACT Prolonged critical illness is characterized by feeding-resistant wasting of protein, whereas reesterification, instead of oxidation of fatty acids, allows fat stores to accrue and associate with a low-activity status of the somatotropic and thyrotropic axis, which seems to be partly of hypothalamic origin. To further unravel this paradoxical metabolic condition, and in search of potential therapeutic strategies, we measured serum concentrations of leptin; studied the relationship with body mass index, insulin, cortisol, thyroid hormones, and somatomedins; and documented the effects of hypothalamic releasing factors, in particular, GH-secretagogues and TRH. Twenty adults, critically ill for several weeks and supported with normocaloric, continuously administered parenteral and/or enteral feeding, were studied for 45 h. They had been randomized to receive one of three combinations of peptide infusions, in random order: TRH (one day) and placebo (other day); TRH 1 GH-releasing peptide (GHRP)-2 and GHRP-2; TRH 1 GHRH 1 GHRP-2 and GHRH 1 GHRP-2. Peptide infusions were started after a 1-mg/kg bolus at 0900 h and infused (1 mg/kgzh) until 0600 h the next morning. Serum concentrations of leptin, insulin, cortisol, T4, T3, insulin-like growth factor (IGF)-I, IGF-binding protein-3 and the acid-labile subunit (ALS) were measured at 0900 h, 2100 h, and 0600 h on each of the 2 study days.

P

ROLONGED critical illness is characterized by feedingresistant wasting of vital protein, whereas fat stores are preserved or even further built up (1, 2). The underlying mechanisms for this apparently inappropriate metabolic response remain largely unknown. Received February 11, 1998. Revision received April 27, 1998. Rerevision received June 8, 1998. Accepted June 12, 1998. Address all correspondence and requests for reprints to: Greet Van den Berghe, M.D., Ph.D., Department of Intensive Care Medicine, University Hospital Gasthuisberg, University of Leuven, B-3000 Leuven, Belgium. E-mail: [email protected]. * This work was supported by research grants from the Fund for Scientific Research, Flanders, Belgium (G.0162.96); the Research Council of the University of Leuven (OT 95/24); The Swedish Medical Research Council (11285, 11502); the Clas Groschinsky’s Foundation, Sweden; and the National Health and Research Council, Australia. Presented at the 80th Annual Meeting of The Endocrine Society, June 24 –27, 1998, New Orleans, Louisiana. † A Clinical Research Investigator of the Fund for Scientific Research, Flanders, Belgium (G.3c05.95N).

Baseline leptin levels (mean 6 SEM: 12.4 6 2.1 mg/L) were independent of body mass index (25 6 1 kg/m2), insulin (18.6 6 2.9 mIU/mL), cortisol (504 6 43 mmol/L), and thyroid hormones (T4: 63 6 5 nmol/L, T3: 0.72 6 0;08 nmol/L) but correlated positively with circulating levels of IGF-I [86 6 6 mg/L, determination coefficient (R2) 5 0.25] and ALS (7.2 6 0.6 mg/L, R2 5 0.32). Infusion of placebo or TRH had no effect on leptin. In contrast, GH-secretagogues elevated leptin levels within 12 h. Infusion of GHRP-2 alone induced a maximal leptin increase of 187% after 24 h, whereas GHRH 1 GHRP-2 elevated leptin by up to 1157% after 36 h. The increase in leptin within 12 h was related (R2 5 0.58) to the substantial rise in insulin. After 45 h, and having reached a plateau, leptin was related to the increased IGF-I (R2 5 0.37). In conclusion, circulating leptin levels during protracted critical illness were linked to the activity state of the GH/IGF-I axis. Stimulating the GH/IGF-I axis with GH-secretagogues increased leptin levels within 12 h. Because leptin may stimulate oxidation of fatty acids, and because GH, IGF-I, and insulin have a protein-sparing effect, GH-secretagogue administration may be expected to result in increased utilization of fat as preferential substrate and to restore protein content in vital tissues and, consequently, has potential as a strategy to reverse the paradoxical metabolic condition of protracted critical illness. (J Clin Endocrinol Metab 83: 3062–3070, 1998)

We have previously shown that the wasting syndrome of protracted critical illness is associated with a uniformly reduced pulsatile secretion of GH, TSH, and PRL related to the low serum concentrations of insulin-like growth factor (IGF)I and thyroid hormones, in the presence of relatively high circulating cortisol levels (3– 6). In a search for novel strategies to reverse the catabolic state of patients treated in intensive care units (ICUs), we previously studied the effects of combined infusion of GH-secretagogues and TRH, and we found that pulsatile GH and TSH secretion could be jointly reactivated and that target organs were readily responsive to the amplified GH and TSH secretion (4 – 6). Leptin is the protein hormone that is expressed in adipocytes and encoded for by the ob gene (7, 8). Circulating levels of leptin vary considerably among normal individuals, with men presenting with lower leptin levels than women and with high age being associated with relatively lower leptin

3062

LEPTIN, GH-SECRETAGOGUES, AND ILLNESS

concentrations for body mass index (BMI) (9). Circulating leptin is known to correlate positively with fat mass and body weight in healthy human volunteers, as well as in subjects with obesity and with conditions of chronic undernutrition (8, 10). Adipocytes have been shown to release leptin in a pulsatile fashion, following a marked circadian rhythmicity with elevated nocturnal values (11, 12). The factors currently known to influence leptin release from adipocytes in man are insulin, GH, IGF-I, thyroid hormones, SRIF, glucocorticoids, cytokines, and b-adrenoreceptor agonists (13–25). Leptin has central effects, playing a role in appetite control (in part, through its effect on neuropeptide Y) and in the regulation of energy expenditure (26 –31). In rodents, a potential role in the neuroendocrine response to starvation has been suggested (32–34), as well as a direct effect on fat metabolism. The latter consists of the ability of leptin to increase intracellular oxidation of fatty acids and to reduce the triglyceride content of adipocytes, hepatocytes, skeletal myocytes, and pancreatic islets (35, 36), hereby counteracting the fat-storing effect of insulin. In view of the metabolic and endocrine alterations present during prolonged critical illness, we here report on leptin serum concentrations in this condition; and we studied the relationship with BMI, insulin, cortisol, thyroid hormones, and somatomedins. In addition, we describe the leptin re-

3063

sponse to the infusion of hypothalamic releasing factors (in particular, GH-secretagogues and TRH) (37–39). Subjects and Methods Patients and concomitant treatment Patients depending on intensive care (including mechanical ventilatory support) for at least 12 days had been eligible for participation in this study. Further inclusion criteria were: 1) a stable condition without dopamine treatment for at least 48 h, because dopamine infusion has been shown to profoundly affect pituitary function in this condition (40, 41); and 2) an expected stay in the ICU for at least another 48 h. Exclusion criteria were: age less than 18 yr; preexisting neurologic, psychiatric, metabolic, or endocrine disease; intracranial lesions; important liver failure [prothrombine time . 2.9 International Normalized Ratio (INR)]; renal failure, requiring dialysis or hemofiltration; and concomitant treatment with glucocorticoids, estrogens, SRIF, thyroid hormones, Ca21-reentry blockers, clonidine, amiodarone, dopamine agonists, or antagonists. A total of 20 patients (7 women, 13 men) were included (Tables 1 and 2). Patient characteristics are described as mean 6 sd. The age was 68 6 13 yr (range, 32– 87 yr). The Apache II score on the ICU-admission day, an indicator of severity of illness (with higher values reflecting a more critical condition) (42), was 14 6 6 (range, 5–28). Patients were critically ill for 25 6 12 (range, 12–59) days at the time of inclusion. BMI was 25 6 5 (range, 17.8 – 40.4) kg/m2. Concomitant treatment included continuously administered total (n 5 14) or partial (n 5 4) parenteral nutrition or full enteral feeding (n 5 2), with normal caloric intake (a mean of 24 nonprotein Cal/kgzday (range, 12–35 Cal/

TABLE 1. Clinical patient characteristics Random

Gender

Age

BMI

Type of illness

Placebo/TRH

M

53

23.9

Placebo/TRH

M

74

24.2

Placebo/TRH

M

60

17.8

Placebo/TRH TRH/Placebo TRH/Placebo

F M M

32 66 69

40.4 25.9 24.2

TRH/Placebo TRH/Placebo GHRP-2/TRH1GHRP-2

F M M

65 48 87

29.4 24.1 22.9

GHRP-2/TRH1GHRP-2 GHRP-2/TRH1GHRP-2

F F

85 76

26.2 23.9

TRH1GHRP-2/GHRP-2 TRH1GHRP-2/GHRP-2

M M

66 80

24.2 26.1

TRH1GHRP-2/GHRP-2

M

76

24.0

G1G/TRH1G1G G1G/TRH1G1G G1G/TRH1G1G TRH1G1G/G1G

F F F M

72 63 77 60

21.4 23.9 31.6 20.6

TRH1G1G/G1G

M

68

22.5

TRH1G1G/G1G

M

81

22.9

Bilobectomy 1 pleural fistula 1 thoracoplasty False aneurysm 1 respiratory insufficiency Complicated esophageal resection with colon graft Varicella pneumonia 1 MOF Lobectomy 1 ARDS Pneumonectomy 1 aspiration pneumonia Complicated spleenectomy 1 MOF Complicated pneumonectomy False aneurysm 1 lung tumor 1 respiratory insufficiency 30% BSA burns Complicated esophageal resection 1 sepsis Complicated CABG Spleen and esophageal resection 1 respiratory insufficiency Mitral valve replacement 1 respiratory insufficiency Post-infarction VSD 1 MOF Necrotising pancreatitis 1 MOF CABG 1 respiratory insufficiency Pneumonectomy 1 respiratory insufficiency Ruptured abdominal aneurysm 1 CABG 1 MOF Complicated esophageal resection 1 sepsis

Total ICU stay

Incl day

Outcome

Feeding

Insulin

5

52

37

Home

TPN

n

19

37

25

Died

TPN

n

7

34

27

Home

PN1EN

n

28 27 12

51 45 33

29 14 18

Home Died Died

TPN TPN PN1EN

n n y

21 11 19

24 116 26

12 59 12

Died Died Died

TPN EN TPN

y n y

11 8

69 108

32 22

Home Home

EN TPN

y y

16 6

36 42

28 15

Home Home

TPN TPN

y y

12

32

13

Home

TPN

y

14 13 12 13

34 181 49 38

21 48 26 24

Died Died Died Died

TPN TPN PN1EN TPN

n y y y

13

35

15

Died

TPN

n

20

39

17

Home

PN1EN

n

Apache II

Randomization group, gender, age (years), BMI (kg/m2), type of illness (MOF, Multiple organ failure; ARDS, adult respiratory distress syndrome; CABG, complicated coronary artery bypass grafts; VSD, ventricular septal defect), Apache II score indicating severity of illness, total ICU stay (days), day of stay in the ICU at the time of inclusion, ultimate outcome, route of continuously administered feeding (TPN, Total parenteral nutrition; PN1EN, parenteral and enteral nutrition; EN, full enteral nutrition), and the need for exogenous insulin infusion in order to maintain blood glucose levels below 12 mmol/L at any time during the course of the study. G1G, GHRH 1 GHRP-2.

3064

JCE & M • 1998 Vol 83 • No 9

VAN DEN BERGHE ET AL.

TABLE 2. Age, Apache II score, caloric intake per 24 h, BMI, and serum concentrations of leptin, insulin, cortisol, IGF-I, IGFBP-3, ALS, and thyroid hormones in 20 prolonged critically ill patients, at the time of inclusion in the study All groups mean 6 SD

Age (y) Apache II Cal/kg/24h BMI (kg/m2) Leptin (mg/L) Insulin (mIU/mL) Cortisol (nmol/L) IGF-I (mg/L) IGFBP-3 (mg/L) ALS (mg/L) T4 (nmol/L) T3 (nmol/L)

Normal mean 6 SD or normal range

Group I mean 6 SD

Group II mean 6 SD

Group III mean 6 SD

P (ANOVA) groups

17.9 6 18.8* 17.0 6 17.5* 276 – 607 100 –300 2.2– 4.6 17.0 –34.0 71–154 1.20 –2.90

58 6 14** 16 6 9 26 6 7 26 6 7 12.0 6 8.0 21.1 6 15.8 438 6 122 91 6 39 2.1 6 0.6 8.5 6 2.7 75 6 27 0.92 6 0.49

78 6 8** 12 6 5 22 6 5 25 6 1 13.5 6 11.5 20.7 6 12.2 578 6 170 86 6 20 1.7 6 0.2 7.0 6 2.1 58 6 15 0.63 6 0.16

70 6 8 14 6 3 23 6 6 24 6 4 11.6 6 10.5 13.0 6 8.7 518 6 282 79 6 17 1.7 6 0.7 5.7 6 2.7 54 6 18 0.55 6 0.15

0.003** 0.5 0.4 0.6 0.9 0.5 0.4 0.7 0.3 0.2 0.2 0.2

68 6 13 14 6 6 24 6 6 25 6 5 12.4 6 9.1 18.6 6 12.9 504 6 194 86 6 28 1.9 6 0.6 7.2 6 2.7 63 6 23 0.72 6 0.36

All parameters were determined in a fed state (normocaloric and balanced parenteral and/or enteral feeding administered over 24 h at a constant rate). The numbers marked by one asterisk refer to 3 h-postprandial mean 6 SD values for insulin and leptin concentration, obtained in a mixed group of healthy men (n 5 13) and women (n 5 8), 18 – 84 yr old, with a BMI of 25 6 5 kg/m2, and measured with the same assay (20). Patients randomized into group I (placebo 6 TRH infusion, n 5 8), group II [GH-releasing peptide (GHRP)-2 6 TRH infusion, n 5 6] and group III (GHRH 1 GHRP-2 6 TRH infusion, n 5 6) were matched at baseline for all parameters except for age (double asterisk, slightly younger patients in group I, compared with group II, as determined with ANOVA and post hoc testing with Fisher’s PLSD). kgzday) and standard composition (0.8 –1.6 g/kg aminoacids/day, 2.8 – 4.3 g/kg glucose per day, 1–1.5 g/kg fat per day (covering 25– 40% of nonprotein calories) (43); inotropic support with exogenous nondopaminergic catecholamines (n 5 7); antibiotics (n 5 16); analgesia and sedation with continuously infused opioids (n 5 17); and/or benzodiazepines (n 5 13). Plasma glucose levels were monitored; insulin was infused when plasma glucose was $ 12 mmol/L (n 5 11) (44). Plasma glucose levels were elevated at the time of study start: 8.3 6 2.1 mmol/L (range, 5.7–14 mmol/L). Human albumin was continuously infused when serum levels were low (mean serum albumin concentration at inclusion was 2.8 6 0.1 g/dL). The mean serum level of triglycerides was 188 6 116 mg/dL (range, 66 –557 mg/dL), and C-reactive protein concentration was elevated (12.5 6 6.2 mg/dL). Continuous hemodynamic monitoring included electrocardiogram (n 5 20), intraarterial blood pressure (n 5 20), central venous pressure (n 5 20), and core and peripheral temperature (n 5 20). During the study period of 45 h, the concomitant ICU therapy, including the nutritional intake, remained virtually unaltered in all patients. The final outcome of these patients was a mean total ICU stay of 54 6 39 (range, 24 –181) days. Eleven patients died in the ICU (55%), a mean 38 6 39 (range, 12–133) days after the study. Nine patients were discharged to the ward and left the hospital subsequently (45%). The study was approved by the Institutional Review Board of the University of Leuven School of Medicine. Informed consent from a first-degree relative was obtained before patient inclusion.

Study design and peptide administration Patients were studied during a total time span of 45 h. They were randomly allocated to one of three study groups: 1) group I (n 5 8) received TRH infusion (1 mg/kg bolus at 0900 h, followed by a 1 mg/kgzh continuous infusion until 0600 h the next morning) vs. placebo during the other day; 2) group II (n 5 6) received TRH 1 GHRP-2 (1 1 1 mg/kg bolus at 0900 h, followed by a 1 1 1 mg/kgzh continuous infusion until 0600 h the next morning) vs. GHRP-2 infusion the other day (1 mg/kg bolus at 0900 h, followed by a 1 mg/kgzh continuous infusion until 0600 h); and 3) group III (n 5 6) received TRH 1 GHRH 1 GHRP-2 infusion (1 1 1 1 1 mg/kg bolus at 0900 h, followed by a 1 1 1 1 1 mg/kgzh continuous infusion until 0600 h) vs. GHRH1GHRP-2 infusion (1 1 1 mg/kg bolus at 0900 h, followed by a 1 1 1 mg/kgzh continuous infusion until 0600 h). Within these three groups, patients were randomized for the order of peptide infusion. This randomized, cross-over design was applied to minimize possible interference by order of peptide administration or by spontaneous recovery. Placebo (NaCl 0.9%), TRH (200 mg/mL NaCl 0.9%; UCB Pharma, Brussels, Belgium), GHRP-2 (50 mg/mL NaCl 0.9%; Kaken Pharmaceu-

tical Co. ltd., Tokyo, Japan), and human GHRH (50 mg/mL NaCl 0.9%; Ferring, Kiel, Germany) infusions were given through a separate lumen of a central venous catheter, inserted for clinical purposes. A PERFUSOR secura FT pump with a 50-ml PERFUSOR syringe (B. Braun, Melsungen, Germany) permitted precise infusions of small volumes at a constant rate. Inadvertent interruption of the infusion or unanticipated bolus injections of the peptides were hereby avoided. During each of the two consecutive study days (at 0900 h, 2100 h, and 0600 h), serum concentrations of leptin, IGF-I, IGF-binding protein (IGFBP)-3, ALS, insulin, cortisol, T4, and T3 were determined. Blood sampling All blood samples were collected through an arterial line, inserted for clinical purposes independently of this study. The Edwards VAMP system (Baxter Healthcare Corporation, Irvine, CA) was used, permitting withdrawal of undiluted blood samples from an indwelling catheter, without undue blood loss. Blood was collected into glass tubes; after clotting and centrifugation, the serum was kept frozen at 220 C until assay.

Assays For determination of leptin, insulin, IGFBP-3, ALS, and cortisol concentrations, all samples were processed in duplicate in the same assay run. For measurement of IGF-I, T4, and T3 levels, all samples of the same patient were processed in duplicate in the same assay run. The serum leptin concentrations were determined by a human leptin RIA (Linco Research, St. Charles, MO). The detection limit was 0.5 mg/L. The intraassay coefficient of variation was 6.3% at a leptin concentration of 15.6 mg/L. The serum insulin levels were determined by a human insulin immunoradiometric assay (Medgenix INS-IRMA, Biocource, Fleurus, Belgium). The detection limit was 1 mIU/mL. The intraassay coefficient of variation was 4.5% at 6.6 mIU/mL and 2.1% at 53 mIU/mL. The plasma concentrations of total IGF-I were measured by RIA, after acid-ethanol extraction. The intraassay coefficient of variation was 10.1% at 95 mg/L and 5.5% at 474 mg/L. The between-assay coefficient of variation was 14.8% at 109 mg/L and 10.1% at 389 mg/L. The detection limit was 10 mg/L. The normal range in healthy adults is 100 –300 mg/L. The serum IGFBP-3 concentrations were measured by RIA, as previously described (45), using antiserum R-100. The intraassay coefficient of variation was 6.2% at 2.5 mg/L and 5.5% at 5.7 mg/L, and the between-assay coefficient of variation was 11.9% at 2.9 mg/L and 14.5% at 6.3 mg/L. Normal ranges are 2.2– 4.6 mg/L. The serum ALS concentrations were assessed by RIA, as described elsewhere (46). The intraassay coefficient of variation was 3.4%, and the between-assay coefficient of variation was 10.5% at 5.3 mg/L and 5.4% at 24 mg/L. Normal ranges are 17–34 mg/L. The serum concentrations of cortisol had been measured by RIA after

LEPTIN, GH-SECRETAGOGUES, AND ILLNESS extraction with dichloromethane. The intraassay coefficient of variation was 3.1% at 417 nmol/L. Normal ranges are 276 – 607 nmol/L at 0800 h; 0 –276 nmol/L at 2000 h; and less than 50 nmol/L at 2400 h, if asleep. The serum concentrations of T4 were measured by RIA using the Tetrabead-125 Diagnostic Kit (Abbott Laboratories, North Chicago, IL). The intraassay coefficient of variation was 4.4% at 44 nmol/L and 2.8% at 94 nmol/L. The between-assay coefficient of variation was 14.6% at 44 nmol/L and 4.3% at 94 nmol/L. Normal values range from 71–154 nmol/L. The serum concentrations of T3 were measured by RIA using the T3 Riabead Kit (Abbott Laboratories). The intraassay coefficient of variation was 4.6% at 0.97 nmol/L and 3.8% at 2.49 nmol/L. The between-assay coefficient of variation was 2.1% at 0.97 nmol/L and 2.8% at 2.49 nmol/L. Normal values range from 1.20 –2.90 nmol/L.

Data analysis At inclusion, groups were compared using factorial ANOVA [post hoc testing using Fisher’s protected least-significant difference (PLSD)], twotailed unpaired t test, Mann-Whitney U-test, and x-square test. Results were analyzed using repeated-measures ANOVA, with post hoc testing using Fisher’s PLSD, when appropriate. Patient characteristics are expressed as mean 6 sd; results are expressed as mean 6 sem, unless indicated otherwise.

Results

Patient characteristics at study inclusion are presented in Tables 1 and 2. The three study groups were matched at baseline, except for age: patients randomized to receive placebo 6 TRH (group I) were found to be slightly younger (mean 6 sd of 58 6 14 yr), compared with those receiving GHRP-2 6 TRH (group II) (mean 6 sd of 78 6 8 yr) (P 5 0.003) but not significantly different from those receiving GHRH 1 GHRP-2 6 TRH (group III) (mean 6 sd of 70 6 8 yr). Patients randomized to one of the groups receiving GHsecretagogues were of the same age. The severity of illness for which patients were treated in the ICU, as indicated by the Apache II score (P 5 0.5, with ANOVA), as well as the nutritional support (P 5 0.4, with ANOVA for the daily caloric intake) were comparable among the three study groups; the apparent difference in ultimate survival did not reach statistical significance (not more survivors in the group receiving GHRP-2 alone, P 5 0.06 with x-square). The baseline serum concentrations of leptin, insulin, cortisol, IGF-I, IGFBP-3, ALS, T4, and T3 measured are delineated in Table 2. There was no difference in these parameters among the three groups. Furthermore, it was impossible to distinguish patients receiving exogenous insulin and patients without insulin treatment, or ultimate survivors and nonsurvivors, by means of these parameters. The normal positive correlation between BMI and leptin was not significantly present during prolonged critical illness [determination coefficient (R2) 5 0.14, P 5 0.1] (Fig. 1). Even after leaving out the highest and outlying BMI of 40.4 kg/m2, the correlation did not reach significance (P 5 0.08). Moreover, there was no significant correlation between inclusion leptin levels and nutritional intake nor with circulating cortisol, T4, T3, insulin, or IGFBP-3; and the absence of correlation was independent of outliers. However, serum leptin levels at inclusion did correlate positively with serum IGF-I (R2 5 0.25, P 5 0.02) and with ALS (R2 5 0.32, P 5 0.009). Leaving out the highest and outlying IGF-I value

3065

strengthened the correlation between IGF-I and leptin (R2 5 0.32, P 5 0.01). The placebo profiles (n 5 8) revealed absence of a diurnal variation in any of these parameters, including leptin (Fig. 2). Effect of TRH

The addition of TRH to the infusion of either placebo, GHRP-2, or GHRH 1 GHRP-2, during one of both study days, which was previously shown to increase circulating levels of T4 and T3 within 24 h (6), did not alter leptin, insulin, cortisol, IGF-I, IGFBP-3, or ALS concentrations (data not shown). Consequently, the infusion of TRH during one of two study days was considered not to play a role in the observations regarding these parameters during the infusion of placebo or GH-secretagogues. Thus, the effects of placebo, GHRP-2, and GHRH 1 GHRP-2 are, from here on, reported over the entire studied time course of 45 h. Effect of placebo, GHRP-2, and GHRH 1 GHRP-2 (Figs. 2 and 3)

Insulin. During placebo infusion, insulin levels remained unaltered over the entire studied episode of 45 h. Within 12 h of infusion of either GHRP-2 or GHRH 1 GHRP-2, insulin levels had increased 312% and 188%, respectively. With GHRP-2, insulin concentration was maximal after 12-h infusion; whereas, with the infusion of GHRH 1 GHRP-2, insulin levels continued to increase, with a maximum of 1372% after 45 h of infusion. These changes in insulin were independent of exogenous insulin infusion. The increases in insulin levels, observed over the studied 45 h during infusion of GHRP-2 alone, were not different from those during GHRH 1 GHRP-2 infusion (P 5 0.9 with ANOVA and Fisher’s PLSD); and both were significantly different, compared with placebo (both P 5 0.004, with ANOVA and Fisher’s PLSD). IGF-I. During placebo infusion, IGF-I levels remained unaltered over the entire studied episode of 45 h. Within 12 h of infusion of either GHRP-2 or GHRH 1 GHRP-2, IGF-I levels had increased 17% and 30%, respectively. During GHRP-2 infusion, IGF-I levels further increased up to a maximum of 176% after 36 h; and during GHRH 1 GHRP-2 infusion, the maximal increase in IGF-I of 1105% was reached after 45 h. These changes in IGF-I concentrations were independent of exogenous insulin infusion. The increases in IGF-I levels observed over the studied 45 h during infusion of GHRP-2 alone were not different from those during GHRH 1 GHRP-2 infusion (P 5 0.1, with ANOVA and Fisher’s PLSD); and both were significantly different, compared with placebo (P 5 0.005 and P , 0.0001, respectively, with ANOVA and Fisher’s PLSD). ALS. During placebo infusion, ALS levels remained unaltered over the entire studied episode of 45 h. Within 12 h of infusion of either GHRP-2 or GHRH 1 GHRP-2, ALS levels had increased 17% and 20%, respectively. With both GHRP-2 and GHRH 1 GHRP-2, ALS-I levels continued to increase progressively up to a maximum

3066


JCE & M • 1998 Vol 83 • No 9

FIG. 1. Leptin, in relation to BMI, and hormonal status in prolonged critically ill patients. The normal positive correlation between BMI and leptin was not significantly present during prolonged critical illness. Moreover, no significant correlation was found between baseline leptin levels and circulating cortisol, T4, T3, insulin, or IGFBP-3. The absence of correlation between leptin and these parameters was independent of outliers. Serum leptin levels at study inclusion did correlate positively with serum IGF-I and ALS concentration. Leaving out the highest and outlying IGF-I value strengthened the correlation with leptin (R2 5 0.32, P 5 0.01).

of 156% and 197%, respectively, after 45 h. These changes in ALS concentrations were independent of exogenous insulin infusion. The increases in ALS levels observed over the studied 45 h during infusion of GHRP-2 alone were not different from those during GHRH 1 GHRP-2 infusion (P 5 0.1, with ANOVA and Fisher’s PLSD); and both were significantly

different, compared with placebo (P 5 0.0002 and P 5 0.007, respectively, with ANOVA and Fisher’s PLSD). IGFBP-3. During placebo infusion, IGFBP-3 levels remained unaltered over the entire studied episode of 45 h. Within 12 h of infusion of either GHRP-2 or GHRH 1 GHRP-2, IGFBP-3 levels had increased 16% and 17%, re-


3067

FIG. 2. Changes over 45 h (‚%) in serum concentrations of leptin, insulin, cortisol, IGF-I, IGFBP-3, and ALS, with infusion of either placebo (triangles), GHRP-2 (1 mg/kgzh, squares), or the combination of GHRH (1 mg/kgzh) and GHRP-2 (1 mg/kgzh) (circles) during prolonged critical illness. The P values indicate the level of significance of the between-group-difference in these changes, as calculated with repeatedmeasures ANOVA.

spectively. With both GHRP-2 and GHRH 1 GHRP-2, IGFBP-3 levels continued to increase progressively up to a maximum of 150% and 165%, respectively, after 45 h. These changes in IGFBP-3 concentrations were independent of exogenous insulin infusion. The increases in IGFBP-3 levels observed over the studied 45 h during infusion of GHRP-2 alone were not different from those during GHRH 1 GHRP-2 infusion (P 5 0.7, with ANOVA and Fisher’s PLSD); and both were significantly different, compared with placebo (P 5 0.03 and P 5 0.01, respectively, with ANOVA and Fisher’s PLSD). Cortisol. The infusion of placebo, GHRP-2, or GHRH 1 GHRP-2 did not alter the moderately elevated serum concentrations of cortisol. Leptin. During placebo infusion, leptin levels remained unaltered over the entire studied episode of 45 h. Within 12 h of infusion of either GHRP-2 or GHRH 1 GHRP-2, leptin levels had increased 72% and 56%, respec-

tively. During GHRP-2 infusion, leptin rose to a maximum of 187% 24 h after initiation of the infusion, and subsequently decreased to 162% at study end. During the infusion of GHRH 1 GHRP-2, leptin levels continued to increase up to a maximum of 1157% after 36 h infusion. These changes in leptin were independent of exogenous infusion of insulin. The increases in leptin levels observed over the studied 45 h during infusion of GHRP-2 alone were different from those during GHRH 1 GHRP-2 infusion (P 5 0.02, with ANOVA and Fisher’s PLSD); and both were significantly different, compared with placebo (P 5 0.005 and P 5 0.006, respectively, with ANOVA and Fisher’s PLSD). The initial change in leptin concentration (after 12 h of infusion of placebo, GHRP-2, or GHRH 1 GHRP-2) correlated positively with change in insulin concentration (R2 5 0.58, P 5 0.0001) (Fig. 3). The change in leptin after 24 h correlated positively with the alterations in insulin (R2 5 0.36, P 5 0.005), IGF-I (R2 5 0.41, P 5 0.002), and ALS (R2 5 0.22, P 5 0.04). The final

3068


JCE & M • 1998 Vol 83 • No 9

FIG. 3. Changes in leptin levels (‚%) during infusion of either placebo, GHRP-2 (1 mg/kgzh) or GHRH 1 GHRP-2 (1 1 1 mg/kgzh) in prolonged critically ill patients. Changes in leptin levels were related to the alterations in serum concentrations of insulin after 12 h, to both insulin and IGF-I after 24 h, whereas after 45 h, only IGF-I was related to leptin.

change in leptin (after 45 h) correlated positively only with the change in IGF-I concentration (R2 5 0.37, P 5 0.004), because the correlation with ALS alterations did not reach significance (R2 5 0.17, P 5 0.07). At all times, leptin was independent of thyroid hormones and cortisol. In the 12 patients randomized to receive GH-secretagogue treatment (GHRP-2 or GHRH 1 GHRP-2), the initial serum leptin levels were independent of BMI, whereas the leptin concentration after 24 h of infusion (R2 5 0.51, P 5 0.009) and also at the end of the treatment (R2 5 0.46, P 5 0.01) correlated positively with BMI. However, it should be noted that the significance of this correlation was highly dependent on the highest BMI in this group. The changes in insulin, IGF-I, ALS, IGFBP-3, and leptin observed in response to GH-secretagogues at any time point during the study were independent of ultimate survival. Discussion

In prolonged critically ill patients supported with intensive care (which includes standardized, balanced, and normocaloric feeding continuously administered over 24 h), leptin concentrations were independent of BMI, insulin, cortisol, thyroid hormones, severity of illness, and ultimate outcome. Leptin levels did correlate positively with low levels of IGF-I and ALS. Infusion of TRH, previously shown to elicit a rise in thyroid hormones, had no effect on circulating leptin levels. In contrast, infusion of GH-secretagogues instantly and robustly elevated leptin concentrations, initially related to the substantial rise in circulating insulin, whereas after 45 h and reaching a plateau, the leptin rise was related to the concomitant increase in IGF-I.

In healthy humans, leptin levels reflect percentage body fat and, accordingly, correlate positively with BMI (9, 21). Within a few weeks and despite feeding, critically ill patients lose total body water and protein, whereas fat stores are preserved or even built up (1, 2), so that BMI is no longer an accurate indicator of body fat and conceivably underestimates the latter after weeks or months of intensive care. Consequently, relatively high leptin levels for a given BMI during critical illness would be anticipated. An activated inflammatory cascade, as present during critical illness, may contribute to leptin increase, mediated by cytokines such as IL-1 and TNF-a (22, 23). In the acute phase of sepsis, the expected rise in leptin was recently confirmed (24). Thus, it was rather surprising to find leptin levels no longer substantially (if at all) elevated in the chronic phase of critical illness. The absence of variability in the leptin levels during the time course of this study is in contrast with the higher leptin concentrations observed during early morning hours in healthy subjects. However, it is in line with loss of diurnal variability of other hypothalamic-pituitary-dependent hormones in critical illness, the cause of which also remains unclear. Together, these findings suggest that an inhibitor of the release of leptin from the adipocytes could be present in the chronic phase of critical illness. A plausible candidate to exert this inhibitory effect is the activated adrenergic system, characterized by high circulating levels of norepinephrine, epinephrine, and dopamine during prolonged critical illness. Activation of the b3 adrenergic receptor on the adipocytes in rodents was found to impair leptin release (47, 48). Recently, infusion of isoproterenol has also been shown to decrease leptin secretion in humans (25). This catecholamine effect could partly explain


3069

the absence of correlation between BMI and leptin during critical illness. Alternatively, because leptin levels at study inclusion were found to correlate only with the low serum concentrations of IGF-I and ALS, the low activity status of the somatotropic axis could play a role. A positive correlation between leptin and IGF-I was previously reported to be present in anorexia nervosa (10). However, unlike in the fed, critically ill patient, the leptin-IGF-I correlation in malnourished anorexia patients was explained by the concomitantly reduced amount of body fat. The situation of the critically ill patient is more one of high age, because a reduced serum IGF-I and relatively lower leptin levels despite an increase in percentage body fat have been documented with aging, a constellation that has been attributed to a reduced activity status of the anterior pituitary (9, 49). Whether leptin plays a causal role in the hypothalamicpituitary dysfunction present in prolonged critical illness, in analogy with what has been described during starvation (32–34), is not clear. The positive correlation with the low IGF-I and ALS levels could be in favor of such a regulatory effect of leptin. However, the finding that leptin was not related in any way to thyroid hormone levels or cortisol does not corroborate a pivotal role of leptin in controlling the neuroendocrine response to critical illness. The observed increase in leptin evoked by GH-secretagogues was striking. A first possible explanation for this increase in leptin is a direct effect of GH-secretagogues on the adipocytes. Although a specific GHRP-receptor has recently been identified on cardiomyocytes1, it is, at present, still unknown whether human adipocytes contain receptors for either GHRH or GHRP. GH deficiency is characterized by high serum leptin levels and a preserved diurnal rhythmicity, whereas during critical illness, the nocturnal rise of GH and leptin release is absent, and both GH and leptin are increased by GH-secretagogues. These findings may corroborate a direct leptin-releasing effect (and eventually a role in the normal nocturnal leptin surge) of GHRH and/or the endogenous ligand for the GHRP-receptor (50), because these hypothalamic releasing factors are abundantly present in pituitary GH deficiency and thought to be less available during prolonged critical illness (51). Alternatively, the leptin rise observed with GH-secretagogues could be evoked, either directly or indirectly, by the amplified GH secretion (6). Addition of GH to cultured mature adipocytes of the rat has no effect on leptin release (52). Nevertheless, the finding of a larger effect on leptin with GHRH 1 GHRP infusion, compared with GHRP alone, in this study may be in favor of such a GH-mediated pathway, because GHRH 1 GHRP has been shown to synergistically increase GH secretion (4, 6). The previously observed transient increase in leptin, within 1 day after initiating GH treatment in GH-deficient adults, corroborates a GH-mediated mechanism underlying our findings (17). The acute leptin-releasing effect of GH in GH deficiency is in contrast with the reduction of high leptin levels documented with GH

treatment at a later stage, after alterations in body composition have occurred (16, 17). In addition, leptin rise obtained with GH-secretagogues in the critically ill patients should perhaps be interpreted within the context of moderate hypercortisolism and increased adrenergic tone. In human volunteers, during high-dose glucocorticoid treatment for 7 days, administration of GH indeed jointly stimulated insulin and leptin release, followed by a rise in oxidation of fatty acids and in energy expenditure (18). Because leptin rise after 12 h of GH-secretagogue infusion in critically ill patients was tightly related to the increase in insulin, an indirect GH effect, mediated by insulin, is suggested. Hyperinsulinism is indeed known to be a potent stimulator of leptin release (14). However, it takes 72 h of hyperinsulinism to increase leptin in healthy volunteers, compared with 12 h or less in the critically ill patients studied here. The increment in leptin seemed to reach a plateau within 45 h. The leptin rise after 45 h of treatment with GH-secretagogues was again, as in the untreated state, positively correlated with IGF-I. It is unclear whether this positive correlation with IGF-I in prolonged critically ill patients suggests a direct effect of IGF-I on leptin release or, alternatively, reflects an increased insulin-sensitivity or an insulin-like effect mediated by rises in free IGF-I. Because, in GH deficient adults, IGF-I administration reduces leptin levels within the short time frame of a few days (17), a direct leptin-releasing effect of IGF-I seems indeed rather unlikely. Increasing leptin levels in prolonged critically ill patients could be of benefit to reverse the paradoxical gain of fat stores and the ongoing protein wasting, despite feeding, in this condition. Indeed, leptin has been shown to exert a direct effect on fat metabolism, stimulating the oxidation of fatty acids and reducing triglyceride content in adipocytes, as well as in the liver, skeletal muscle, and pancreas (35, 36). Consequently, increasing circulating levels of leptin, together with GH, IGF-I, and insulin, is a constellation of endocrine changes that may improve use of fat as preferential substrate, prevent fat accumulation in organs such as the liver, and concomitantly restore the protein content in vital tissues (53, 54). In conclusion, in protracted critical illness, circulating leptin levels were related to the suppressed GH/IGF-I axis. Reactivating the GH/IGF-I axis with GH-secretagogues for 2 days concomitantly increased leptin levels within 12 h. Whether this finding reflects a direct leptin-releasing effect of GH-secretagogues or, rather, points towards a role for GH in the physiological regulation of leptin secretion remains to be determined. Increasing levels of leptin, together with GH, IGF-I, and insulin, may switch substrate use from protein to fat, enabling protein anabolism to restore, which is crucial for the onset of recovery from prolonged critical illness. The data further support future exploration of the therapeutic potential of GH-secretagogues in intensive care-dependent, critically ill patients.

1 Ong H. GHRP receptors: are these binding sites specific to pituitary or cardiac tissue? Oral communication at the 24th International Symposium on GH and Growth Factors in Endocrinology and Metabolism, Antwerp, Belgium, October 3– 4, 1997.

The authors wish to thank the medical and nursing staff of the ICU for their cooperation. Dr. Bengt-Ake Bengtsson is especially acknowledged for actively supporting the collaboration between the Universities

Acknowledgments

3070


of Go¨teborg and Leuven. Dr. Bjo¨rn Carlsson is acknowledged for the valuable discussions, Dr. de Zegher for manuscript review, and Dr. Jaak Billen for providing laboratory support. We thank Sri Meka, Viviane Celis, Danielle Van Sever, Willy Coopmans, Tina Schreurs, Marie-Jeanne Leemput, Christiane Eyletten, Marianne Aerts, and Ulla Karlsson for expert technical assistance.

References 1. Streat SJ, Beddoe AH, Hill GL. 1987 Aggressive nutritional support does not prevent protein loss despite fat gain in septic intensive care patients. J Trauma. 27:262–266. 2. Gamrin L, Esse´n P, Forsberg AM, Hultman E, Wernerman J. 1996 A descriptive study of skeletal muscle metabolism in critically ill patients: free amino acids, energy-rich phosphates, protein, nucleic acids, fat, water, and electrolytes. Crit Care Med. 24:575–583. 3. Vermes I, Bieshuizen A, Hampsink RM, Haanen C. 1995 Dissociation of plasma adrenocorticotropin and cortisol levels in critically ill patients: possible role of endothelin and atrial natriuretic hormone. J Clin Endocrinol Metab. 80:1238 –1242. 4. Van den Berghe G, de Zegher F, Veldhuis JD, et al. 1997 The somatotropic axis in critical illness: effect of continuous GHRH and GHRP-2 infusion. J Clin Endocrinol Metab. 82:590 –599. 5. Van den Berghe G, de Zegher F, Veldhuis JD, et al. 1997 Thyrotropin and prolactin release in prolonged critical illness: dynamics of spontaneous secretion and effects of growth hormone secretagogues. Clin Endocrinol (Oxf). 47:599 – 612. 6. Van den Berghe G, de Zegher F, Baxter RC, et al. 1998 Neuroendocrinology of critical illness: effect of continuous thyrotropin-releasing hormone infusion and its combination with growth hormone-secretagogues. J Clin Endocrinol Metab. 83:309 –319. 7. Zhang Y, Proenca R, Maffei M, Barone M, Leopold L, Friedman JM. 1994 Positional cloning of the mouse obese gene and its human analogue. Nature. 372:425– 432. 8. Considine RV, Sinha MK, Hainman ML, et al. 1996 Serum immunoreactive leptin concentrations in normal weight and obese humans. N Engl J Med. 334:292–295. 9. Ostlund Jr RE, Yang JW, Klein S, Gingerich R. 1996 Relation between plasma leptin concentration and body fat, gender, diet, age, and metabolic covariates. J Clin Endocrinol Metab. 81:3909 –3913. 10. Greenspoon S, Gulicl T, Askari H, et al. 1996 Serum leptin levels in women with anorexia nervosa. J Clin Endocrinol Metab. 81:3861–3863. 11. Sinha MK, Ohannesian JP, Heiman ML, Kriauciunas A, Stephens TW, Magosin S, Marco C, Caro JF. 1996 Nocturnal rise of leptin in lean, obese, and non-insulin-dependent diabetes mellitus subjects. J Clin Invest. 97:1344 –1347. 12. Licinio J, Mantzoros C, Negrao AB, et al. 1997 Human leptin levels are pulsatile and inversely related to pituitary-adrenal function. Nature Med. 3:575–579. 13. Kolaczynski JW, Nyce MR, Considine RV, et al. 1996 Acute and chronic effects of insulin on leptin production in humans: studies in vivo and in vitro. Diabetes. 45:699 –701. 14. Boden G, Chen X, Kolaczynski JW, Polansky M. 1997 Effects of prolonged hyperinsulinemia on serum leptin in normal human subjects. J Clin Invest. 100:1107–1113. 15. Boden G, Chen X, Mozzoli M, Ryan I. 1996 Effect of fasting on serum leptin in normal human subjects. J Clin Endocrinol Metab. 81:3419 –3423. 16. Florkowski CM, Collier GR, Zimmet PZ, Livesey JH, Espiner EA, Donald RA. 1996 Low-dose growth hormone replacement lowers plasma leptin stores without affecting body mass index in adults with growth hormone deficiency. Clin Endocrinol (Oxf). 45:769 –773. 17. Bianda TL, Glatz Y, Boeni-Schnetzler M, Froesch ER, Schmid C. 1997 Effects of growth hormone (GH) and insulin-like growth factor-I on serum leptin in GH-deficient adults. Diabetologica. 40:363–364. 18. Berneis K, Vosmeer S, Keller U. 1996 Effects of glucucorticoids and of growth hormone on serum leptin concentrations in man. Eur J Endocrinol. 135:663– 665. 19. Valcavi R, Zini M, Peino R, Casanueva FF, Dieguez C. 1997 Influence of thyroid status on serum immunoreactive leptin levels. J Clin Endocrinol Metab. 82:1632–1634. 20. Dagogo-Jack S, Selke G, Melson AK, Newcomer JW. 1997 Robust secretory responses to dexamethasone in obese subjects. J Clin Endocrinol Metab. 82:3230 –3233. 21. Korbonits M, Trainer PJ, Little JA, et al. 1997 Leptin levels do not change acutely with food administration in normal or obese subjects, but are negatively correlated with pituitary-adrenal activity. Clin Endocrinol (Oxf). 46:751–757. 22. Janik JE, Curti BD, Considine RV, et al. 1997 Interleukin 1a increases serum leptin concentrations in humans. J Clin Endocrinol Metab. 82:3084 –3086. 23. Zumbach MS, Boehme MWJ, Wahl P, Stremmel W, Ziegler R, Nawroth PP. 1997 Tumor necrosis factor increases serum leptin levels in humans. J Clin Endocrinol Metab. 82:4080 – 4082.

JCE & M • 1998 Vol 83 • No 9

24. Bornstein SR, Licinio J, Tauchnitz R, Engelman L, Negrao AB, Chrousos GP. 1998 Plasma leptin levels are increased in survivors of acute sepsis: associated loss of diurnal rhythm in cortisol and leptin secretion. J Clin Endocrinol Metab. 83:280 –281. 25. Donahoo WT, Jensen DR, Yost TJ, Eckel RH. 1997 Isoproterenol and somatostatin decrease plasma leptin in humans: a noval mechanism regulating leptin secretion. J Clin Endocrinol Metab. 82:4139 – 4134. 26. Barinaga M. 1995 ‘Obese’protein slims mice. Science. 269:475– 476. 27. Pelleymounter MA, Cullen MJ, Baker MB, et al. 1995 Effect of the obese gene on body weight reduction in ob/ob mice. Science. 269:540 –543. 28. Halaas JL, Gaijwala KS, Maffei M, et al. 1995 Weight-reducing effects of the plasma protein encoded by the obese gene. Science. 269:543–546. 29. Romsos DR, Swick AG, Chruncyk BA, Cunningham D, Mistry AM. 1996 Intracerebroventricular recombinant leptin decreases food-intake and increases metabolic rate in ob/ob mice. FASEB J. 10:1287. 30. Mercer JG, Hoggard N, Williams LM, et al. 1996 Coexpression of leptin receptor and preproneuropeptide Y mRNA in arcuate nucleus of mouse hypothalamus. J Neuroendocrinology. 8:733–735. 31. Devos R, Richards JG, Campfield LA, et al. 1996 OB protein binds specifically to the choroid plexus of mice and rats. Proc Natl Acad Sci USA. 93:5668 –5673. 32. Ahima RS, Prabrakan D, Mantzoros C, et al. 1996 Role of leptin in the neuroendocrine response to fasting. Nature. 383:250 –252. 33. Legradi G, Emerson CH, Ahima RS, Flier JS, Lechan RM. 1997 Leptin prevents fasting-induced suppression of prothyrotropin-releasing hormone messenger ribonucleic acid in neurons of the hypothalamic paraventricular nucleus. Endocrinology. 138:2569 –2576. 34. Caro E, Senaris R, Considine RV, Casanueva FF, Dieguez C. 1997 Regulation of in vivo growth hormone secretion by leptin. Endocrinology. 138:2203–2206. 35. Shimabukuro M, Koyama K, Chen G, et al. 1997 Direct antidiabetic effect of leptin through triglyceride depletion of tissues. Proc Natl Acad Sci USA. 94:4637– 4641. 36. Mu¨ller G, Ertl J, Gerl M, Preibisch G. 1997 Leptin impairs metabolic actions of insulin in isolated rat adipocytes. J Biol Chem. 272:10585–10593. 37. Frohman LA, Jansson JO. 1986 Growth hormone-releasing hormone. Endocr Rev. 7:223–253. 38. Jackson IMD. 1982 Thyrotropin releasing hormone. N Engl J Med. 306:145–155. 39. Bowers CY, Momany FA, Reynolds GA, Hong A. 1984 On the in vitro and in vivo activity of a new synthetic hexapeptide that acts on the pituitary to specifically release growth hormone. Endocrinology. 114:1537–1545. 40. Van den Berghe G, de Zegher F, Lauwers P, Veldhuis JD. 1994 Growth hormone secretion in critical illness: effect of dopamine. J Clin Endocrinol Metab. 79:1141–1146. 41. Van den Berghe G, de Zegher F. 1996 Anterior pituitary function during critical illness and dopamine treatment. Crit Care Med.24:1580 –1590. 42. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. 1985 Apache II: a severity of disease classification system. Crit Care Med. 13:818 – 829. 43. Souba WW. 1997 Nutritional support. N Engl J Med. 336:41– 48. 44. Mizock BA. 1995 Alterations in carbohydrate metabolism during stress: a review of the literature. Am J Med. 98:75– 84. 45. Baxter RC, Martin JL. 1986 Radioimmunoassay of growth hormone-dependent insulin-like growth factor binding protein in human plasma. J Clin Invest. 78:1504 –1512. 46. Baxter RC. 1990 Circulating levels and molecular distribution of the acid-labile (a) subunit of the high molecular weight insulin-like growth factor-binding protein complex in normal subjects. J Clin Endocrinol Metab. 70:1347–1353. 47. Gettys TW, Harkness PJ, Watson PM. 1996 The b3-adrenergic receptor inhibits insulin-stimulated leptin secretion from isolated rat adipocytes. Endocrinology. 137:4054 – 4057. 48. Mantzoros CS, Qu D, Frederich RC, et al. 1996 Activation of b3-adrenergic receptors suppresses leptin expression and mediates a leptin-independent inhibition of food intake in mice. Diabetes. 45:909 –914. 49. Vermeulen A. 1987 Nyctohemeral growth hormone profiles in young and aged men: correlation with somatomedin-C-levels. J Clin Endocrinol Metab. 64:884 – 888. 50. Howard AD, Feighner SD, Cully DF, et al. 1996 A receptor in pituitary and hypothalamus that functions in growth hormone release. Science. 273:974 –977. 51. Van den Berghe G, de Zegher F, Bowers CY, et al. 1996 Pituitary responsiveness to growth hormone (GH) releasing hormone, GH-releasing peptide-2 and thyrotropin releasing hormone in critical illness. Clin Endocrinol (Oxf). 45:341–351. 52. Hardie LJ, Guilhot N, Trayhurn P. 1996 Regulation of leptin production in cultured mature white adipocytes. Horm Metab Res. 28:685– 689. 53. Russell-Jones DL, Umpleby M. 1996 Protein anabolic action of insulin, growth hormone and insulin-like growth factor I. Eur J Endocrinol. 135:631– 642. 54. Kupfer SR, Underwood LE, Baxter RC, Clemmons DR. 1993 Enhancement of the anabolic effects of growth hormone and insulin-like growth factor-1 by use of both agents simultaneously. J Clin Invest. 91:391–396.