The effects of growth hormone replacement therapy on overnight metabolic fuels in hypopituitary patients

OBJECTIVE: Hypopituitary adults on conventional replacement have low concentrations of metabolic fuels throughout the night, possibly related to GH deficiency or to decreased cortisol levels overnight. We investigated whether GH replacement corrects the overnight fuel deficiency. DESIGN: We measured circulating levels of metabolic fuels: glucose, non-esterified fatty acids (NEFA), glycerol and 3-hydroxybutyrate (3-OHB) and insulin concentrations over 24 h (from 0730 h to 0700 h) in hypopituitary adults before and after GH treatment in a randomized double-blind placebo-controlled trial of 3 months' duration. PATIENTS: Thirteen hypopituitary patients, 8 women and 5 men, were studied. RESULTS: Six patients (4 women and 2 men) received GH and 7 patients (4 women and 3 men) were allocated to receive placebo. There was no difference in fasting (0730 h), area under the curve (AUC) between 2400 h and 0700 h (overnight) and AUC over 24 h for plasma glucose, 3-OHB, glycerol and insulin concentrations as a result of GH treatment. Fasting and overnight AUC for NEFA were significantly higher on GH treatment ((mean +/- SEM) 243 +/- 29 vs. 446 +/- 90 micromol/l, P = 0.03, 1522 +/- 208 vs. 2167 +/- 123 micromol/l H, P = 0.046, respectively), but AUC over 24 h was not affected significantly. No significant changes in any fuel were seen in the placebo group. The changes in fasting, overnight and 24 h AUC for glucose, 3-OHB, glycerol and insulin levels with GH and with placebo for 3 months were similar. The changes in fasting and overnight AUC for NEFA before and after 3 months were significantly different in the group treated with GH vs. the group treated with placebo (median (lower-upper quartile) 104 (90-276) vs. -89 (-98 to 26) micromol/l, P = 0.002; 633 (263-967) vs. -895 (-1379 to -494) micromol/l h, P = 0.002, respectively), but the changes in 24-h AUC for NEFA were not significant between the two groups. CONCLUSIONS: GH replacement in hypopituitary adults increases fasting and overnight (between 2400 h and 0700 h) non-esterified fatty acid concentrations, consistent with the known lipolytic effect of GH. GH did not influence the concentrations of other metabolic fuels or insulin.     

Effect of glucocorticoid replacement therapy on glucose tolerance and intermediary metabolites in hypopituitary adults

BACKGROUND AND OBJECTIVES Excess Impaired glucose tolerance and diabetes mellitus have been reported in hypopituitary adults on conventional replacement therapy Including glucocorticoids. We investigated the effect of the glucocorticoid component on glucose tolerance and intermediary metabolites in hypopituitary adults. DESIGN A 3-hour 75-g oral glucose tolerance test (OGTT) was performed on two study days, at least one week apart. On one study day, the glucocorticoid replacement morning dose was taken 60 minutes before the OGTT, and on the other it was left until after the OGTT. All other pituitary replacement therapies were kept unchanged on the two study days. PATIENTS Eight hypopituitary adults (3 males and 5 females; aged 46–76 years) on conventional replacement therapy were studied. Their duration of hypopituitarism was mean (range) 15 (5–31) years. Their mean body mass index (BMI) was 28·4 (24·1–35·1) kg/m². Their total daily cortisol dose was 26 (15–30) mg. MEASUREMENTS Plasma glucose, insulin, non-esterified fatty acids (NEFA), glycerol and 3-hydroxybutyrate were measured at 30-minute intervals and plasma cortisol levels were measured hourly. RESULTS Fasting glucose and insulin concentrations were similar on the glucocorticoid day (GD) and the non-glucocorticoid day (NGD) (glucose (mean ± SD) 4·9 ± 0·9 vs 4·4 ± 0·5 mmol/l, insulin (median (range) 5 (1-17) vs 2 (1-15) mU/l, respectively). Post-glucose glycaemia was higher on the GD than on the NGD with a significantly higher glucose area under the curve (AUC) (45·0 ± 82 vs 38·9 ± 11·7 mmol/l h, P < 0·05). Post-glucose insulinaemia was also higher on the GD than on the NGD with significantly higher insulin AUC (270 (47-909) vs 207 (46-687)mU/l h, P < 0·02). Impaired glucose tolerance was found in three patients on the GD, one of whom continued to have impaired glucose tolerance on the NGD. The areas under the curves of NEFA, glycerol and 3-hydroxybutyrate were not significantly different on the two days. On the NGD, plasma cortisol levels were undetectable (<50nmol/l) in all patients and on the GD the median (range) peak was 500 (330–740) nmol/l dropping to 125 (60-330) nmol/l at 180 minutes. The difference in glucose AUC between the two days correlated with the maximal plasma cortisol levels (Spearman's p= 0·83, P < 0·01). CONCLUSIONS Glucocorticoid replacement therapy taken pre-prandially In hypopituitary adults induces mild elevations In circulating glucose and insulin levels even with acceptable plasma cortisol concentrations. Optimal regimens for glucocorticoid replacement require more study.     

Growth hormone treatment in hypopituitary GH deficient adults reduces circulating cortisol levels during hydrocortisone replacement therapy

OBJECTIVES Patients with GH deficiency frequently have multiple hormone deficiencies and require hydrocortisone replacement. We have investigated whether GH treatment alters circulating cortisol levels in hypopituitary patients receiving stable replacement therapy. DESIGN Subjects were studied during 6 or 12 months of S.C. GH at a dose of 0.25 IU/kg/week (0.125 IU/kg/week for the first 4 weeks), and after a wash-out period of at least 2 months off GH (range 2–5 months). PATIENTS Fourteen hypopituitary patients (2F:12M) receiving stable hydrocortisone replacement and thyroxine, gonadal steroids and bromocriptine therapy as required. MEASUREMENTS Serum cortisol values were measured throughout the day over 10.5 hours. Thyroid hormones and cortisol binding globulin (CBG) were measured in the baseline sample. Comparisons of the serum cortisol peak after receiving the first dose of hydrocortisone, the time when the serum cortisol peak was obtained, the area under the curve (AUC) for the cortisol values and the levels of unbound cortisol on and off OH therapy were made. The results are expressed as mean ± SEM. Comparisons were carried out within Individuals, using the Wilcoxon signed rank test. A P-value less than 0.05 was considered statistically significant. RESULTS During OH therapy, there was a significant reduction in the mean cortisol peak (662.2 ± 61.1 vs 848.0 ± 58.6 nmol/l; P= 0.001), and In the AUC for cortisol (185.3 ± 18.3 vs 230 ± 17.9 nmol/l/10.5h; P= 0.03), but there was no significant change in the time of the cortisol peak (55.7 ± 7.6 vs 57.8 ± 4.9 minutes; P= NS). During GH therapy there was a significant reduction In CBG levels (33.64 ± 1.16 vs 40.86 ± 1.34mg/l; P= 0.001); however, no changes were found In the levels of calculated unbound cortisol on and off GH (2.87 ± 0.38 vs 2.90 ± 0.30 nmol/l; P= NS). During OH therapy, there was a significant increase In serum trilodothyronine (T3) (1.88 ± 0.15 vs 1.44 ± 0.11 nmol/l; P= 0.01), and a significant decrease in thyroxine (T4) levels (74.9 ± 11.1 vs 97.6 ± 10.9 nmol/l; P= 0.02) but levels remained within the normal range. No change was observed in serum TSH levels (0.29 ± 0.13 vs 0.83 ± 0.71 mU/I; P = NS). CONCLUSIONS These results suggest that OH therapy In GH deficient adults produces an alteration in the measured serum cortisol profile, with a reduction in concentration of total cortisol in blood after erally administered hydrocortisone. These changes in circulating cortisol probably depend primarily on the fall in CBG levels, as no changes were found in the levels of calculated unbound cortisol on and off GH. Our data show that when measuring circulating cortisol levels, the results should be interpreted with caution in OH deficient patients on GH replacement, and different criteria may have to be applied to the circulating cortlsol profile of these patients. The results highlight the importance of ensuring adequate corticosteroid replacement in patients starting GH therapy.     

The effect of recombinant IGF-I on anterior pituitary function in healthy volunteers

OBJECTIVE Insulin-like growth factor-I is the mediator of many of the actions of GH and is a potent metabolic regulator. Recombinant IGF-I (rhIGF-I) is of potential value in the treatment of syndromes associated with either GH or insulin resistance. This study was designed to assess the effects of subcutaneous (s.c.) rhIGF-I on anterior pituitary function. DESIGN Double-blind, placebo controlled, randomized cross-over study. The interval between investigations was 2 weeks. SUBJECTS Twelve normal volunteers received on one occasion a single S.C. dose of 40 μg/kg rhIGF-I and on the other, placebo. MEASUREMENTS Circulating levels were measured, over 24 hours, of GH, LH, FSH, PRL, TSH, cortisol, ACTH, glucose, IGF-I, IGF-II, Insulin, C-peptide; IGF binding proteins by Western ligand blotting; total IGF bioactivity using FRTL-5 thyroid cells; and glucose by the glucose oxidase method. RESULTS Recombinant IGF-I Increased AUC for plasma IGF-I, measured by radioimmunoassay (rhIGF-I mean 7065 ± SEM 33 vs 3895 ± 204 μg/l, P < 0·0001) and IGF bioactivity (22·5 ± 3·4 vs 14·2 ± 1·8 U/ml, P < 0·001) but plasma IGF-II fell (9308 ± 403 vs 11052 ± 451 μg/l, P < 0·0001). There was no biochemical or clinical evidence of hypoglycaemia and no difference in mean glucose levels. No difference existed in AUC for GH, LH, FSH, ACTH and cortisol between rhIGF-I and placebo; additionally, pulse number and amplitude for GH and LH were unaffected. TSH fell following rhIGF-I (33·0 ± 3·36 vs 42·5 ± 5·98 mU h/l, P= 0·01). Both mean plasma C-peptlde (0·73 ± 0 06 vs 0·91 ± 0±05 nmol/l, P= 0·03), and insulin (10·81 ± 1·02 vs 15·36 ± 1·18 mU/l, P= 0·03) were lower following rhIGF-I. There was no change In IGFBPs. CONCLUSION A single injection of 40 μg/kg of subcutaneous rhIGF-I does not cause hypoglycaemia. IGF bioactivity was increased without inhibition of GH secretion. The only change observed in anterior pituitary function was a fall in plasma TSH.     

Impact of 2 weeks high dose growth hormone treatment on basal and insulin stimulated substrate metabolism in humans

OBJECTIVE Short-term, high dose growth hormone (GH) treatment has been advocated in many catabolic disease states. It is likely that some of the anabolic effects of GH are mediated through activation of lipolysis, but the metabolic impact of therapeutically relevant GH exposure is not known in detail. The present study was accordingly designed to assess the effects of such GH exposure on basal and insulin stimulated intermediary metabolism. DESIGN, PATIENTS AND MEASUREMENTS Six healthy young females were examined following daily injections of GH (12 IU/day) or saline for 2 weeks in a placebo controlled design. Each study consisted of a 3 hour basal period and a 2 hour hyperinsulinaemic euglycaemic clamp. RESULTS GH treatment caused (1) increased levels of IGF-I (382 ± 46 vs 294 ± 22 /μg/l, P < 0 05) and (2) increased basal values of free fatty acids (714 ±40 (GH) vs 634 ±64 (placebo) μmol/l, P<0 05), 3-hydroxybutyrate (118 3 ± 42 8 (GH) vs 57 7 ± 21 6 (placebo) μmol/l, P < 0 05), glycerol (54–3 ±8–2 (GH) vs 41–4 ±8–4 (placebo) μmol/l, P<0 05) and forearm uptake of 3-hydroxybutyrate, together with increments of plasma glucose (5 28 ±0–11 (GH) vs 4 87±0 16 (placebo) mmol/l, P<0 05). Basal forearm uptake of glucose, isotopically determined glucose turnover and serum levels of GH, insulin and C-peptide were unaltered. During the clamp GH treatment was associated with (1) a 40% decrease in the administered amount of glucose (M-value) (P<0 05) and (2) a 70% decrease in forearm glucose uptake (P<0 05). Indirect calorimetry revealed a 15% increase in resting energy expenditure (P<0 05) and a decreased basal respiratory exchange ratio (0 75 (GH) vs 0 80 (placebo), P<0 05), presumably reflecting increased lipid oxidation. CONCLUSIONS Administration of GH in a therapeutic dose for 2 weeks, despite apparently normal daytime levels of major metabolic hormones, induces significant increases in circulating lipid fuel substrates, increased energy expenditure and lipid oxidation, together with insulin resistance. Such effects should be considered when applying GH treatment schedules clinically.     

The effect of growth hormone replacement therapy in hypopituitary adults on calcium and bone metabolism*

OBJECTIVE The importance of growth hormone (GH) for normal skeletal growth in childhood and adolescence is well established but much less is known about its action on the adult skeleton. We therefore wished to investigate the effects of replacement treatment with blosynthetic human GH in hypopituitary adults on aspects of calcium homeostatis, bone metabolism and bone mineral mass. PATIENTS Forty hypopituitary adults (21 females and 19 males; aged 19–67 years). DESIGN A prospective randomized double-blind placebo-controlled trial lasting for 6 months. PROTOCOL Following baseline assessments, GH was given in a daily dose of 0·02–0·05 IU/kg body weight subcutaneously (or a placebo (P)) at bedtime. Patients were reviewed at 1, 3 and 6 months. MEASUREMENTS Plasma calcium, phosphate and total plasma alkaline phosphatase were measured at 0, 1, 3 and 6 months. Serum insulin like growth factor I (IGF-I), osteocalcin, procollagen 1 carboxyterminal peptide (P1CP) and intact parathyroid hormone (PTH) level, 24-hour urinary calcium and creatinine excretion were all measured at 0 and 6 months. Bone mineral density of total body and lumbar spine was also measured by dual energy X-ray absorptiometry at 0 and 6 months in 12 patients on GH and 14 on placebo. RESULTS Thirty-eight patients completed the study (18 on GH, 20 on placebo). Serum IGF-I Increased significantly on GH treatment (mean ± SD) (GH: 276 ± 197 vs P: 88 ± 50 μg/l, P < 0 0001 at 6 months). Plasma calcium increased slightly but significantly in the GH-treated group (2·23 ± 0·11–2·29 ± 0·11 mmol/l, P<0·05). At the end of the study, plasma calcium was however similar on GH and placebo (GH, 2·29 ± 0·11; P, 2·26 ± 0·09 mmol/l). Plasma phosphate increased on GH (GH: 1·02±0·23–1·32±0·19; P: 0·99±0·16–0·96±0·12 mmol/l over the 6 months of treatment, P<0·001). There was no significant change in the urinary calcium excretion on GH therapy. Plasma total alkaline phosphatase, osteocalcin and P1CP were significantly higher on GH than P at 6 months (alkaline phosphatase: GH: 104±32 vs P: 69±32 U/I, P<0·01, osteocalcin: GH: 17·2±8·0 vs P: 5·3±3·2 μg/l, P<0·001 and P1CP: GH: 207 ± 152 vs P: 93±31 μg/l, P<0·01). There was no difference in the intact parathyroid hormone level (GH: 31 ± 14 vs P:31 ± 15 ng/l, NS). No significant change was observed in bone mass after 6 months of GH treatment, either in total body bone mineral content or in the lumbar spine. CONCLUSION In this large study, GH replacement in hypopituitary adults for 6 months increased bone turnover but did not affect bone mineral content. Longer-term studies are required to assess further any effect on bone mass.     

The effects of prolonged growth hormone replacement on bone metabolism and bone mineral density in hypopituitary adults*

OBJECTIVE Short-term GH replacement in hypopituitary adults increases bone turnover; data on the consequences of longer-term GH treatment are limited. We report on the effects of 12–18 months of GH replacement treatment with biosynthetic human GH on bone metabolism and bone mass in hypopituitary adults. DESIGN Patients were studied before and after GH treatment for 12 months (n= 11) and 18 months (n= 27) respectively in an open trial. GH dose was 0·04±0·01 IU/kg daily. MEASUREMENTS Plasma calcium, phosphate and intact PTH concentrations, 24-hour urinary calcium excretion, 3 markers of bone formation (total alkaline phosphatase, osteocalcin and procollagen 1 carboxy terminal peptide (P1CP)) and serum concentration of carboxyterminal cross-linked telopeptide of type 1 collagen (ICTP), as a marker of bone resorption, were measured at 6-month intervals. Lumbar splne and total body bone mineral mass was measured by dual-energy X-ray absorptiometry. RESULTS Small increases were observed in plasma calcium and phosphate concentrations at 12 months of GH therapy but the differences at 18 months were not statistically significant. Serum intact PTH concentration did not change. Plasma total alkaline phosphatase increased significantly on GH from 75 ± 26 to 92 ± 30 (P<0·01) and 85±31 U/I (NS) at 12 and 18 months respectively. Serum osteocalcin increased from 6·5 ± 3·7 to 15·7 ± 6·2 (P<0·0001) and 16·6±5·7 μg/l (P<0·001) at 12 and 18 months respectively and P1CP increased significantly from 106·0 ± 47·3 μ/l to 165·5 ± 95·3 (P< 0·0001) and 177·2 ± 72·2 μg/l (P<0·01) at 12 and 18 months respectively. Plasma ICTP concentration increased also from 3·4 ± 1·8 to 7·3 ± 3·4 (P < 0·0001) and 7·0 ± 2·7 μg/l (P<0·003) at 12 and 18 months of GH therapy respectively. No significant change was observed in total body or lumbar splne bone mass, over the 18 months of GH treatment CONCLUSIONS Replacement therapy with GH in hypopituitary adults for 6–18 months produced a sustained increase in bone turnover (both formation and resorption). Bone mass was maintained but did not increase over the study period.     

The effects of recombinant human insulin-like growth factor I on growth hormone secretion in adolescents with insulin dependent diabetes mellitus

OBJECTIVE It has been proposed that low IGF-I levels and reduced IGF-I bioactivity may lead to elevated GH levels in adolescents with insulin dependent diabetes (IDDM). We have therefore studied the effects of human recombinant insulin-like growth factor I (rhIGF-I) administration on GH levels and GH secretion in adolescents with IDDM. PATIENTS Nine late pubertal adolescents (four male and five female) with IDDM. DESIGN A double-blind placebo controlled study of rhIGF-I administered subcutaneously in a dose of 40 μg/kg body weight at 1800 h. MEASUREMENTS IGF-I and GH concentrations were measured at regular intervals throughout the study. Twenty-two hour GH secretory rates were calculated by deconvolution analysis. Overnight GH profiles were analysed by distribution analysis, and Fourier transformations were performed on both overnight GH concentrations and GH secretory rates. RESULTS Mean IGF-I levels over the 22-hour study period were significantly elevated following rhIGF-I administration (350 ± 26 vs 205 ± 21 μg/l (mean ± SEM), P<0 001). Mean 22-hour GH levels were reduced following rhIGF-I administration (19.4 ± 4.0 compared with 33.6 ± 5.8 mU/l; P= 001). Distribution analysis demonstrated that the reduction in GH levels was due to changes in the proportion of values at both high and low concentrations. Deconvolution analysis also revealed a significant overall reduction in GH secretory rate following IGF-I administration (1.81 ± 0.30 vs 2.98 ± 0.47 mU/min, P= 0.01) which was still apparent during the final 5.5 hours of the study period (1.51 ± 0.30 vs 2.76 ± 0.61 mU/min, P= 002). The dominant periodicity of GH secretory episodes as determined by Fourier transformation was between 120 and 180 minutes after both IGF-I and placebo. CONCLUSIONS In late pubertal adolescents with IDDM the rise in IGF-I levels following rhIGF-l administration in a subcutaneous dose of 40 μg/kg body weight leads to a significant reduction in GH levels and GH secretory rate. The reduction in GH secretion is due to changes in pulse amplitude rather than frequency. A reduction in GH secretion was apparent at the beginning and also towards the end of the 22-hour study period.     

Growth hormone (GH) replacement decreases serum total and LDL-cholesterol in hypopituitary patients on maintenance HMG CoA reductase inhibitor (statin) therapy

Objective Adult onset GH deficiency (GHD) is characterized by abnormalities of serum lipoprotein profiles and GH replacement results in favourable alterations in serum total and low density lipoprotein (LDL)‐cholesterol. Preliminary evidence has indicated that the effect of GH replacement in this respect may be additive to that of HMG CoA reductase inhibitor (statin) therapy. We have examined this possibility during prospective follow‐up of adult onset hypopituitary patients enrolled in KIMS (Pfizer International Metabolic Database), a pharmacoepidemiological study of GH replacement in adult hypopituitary patients. Design Lipoprotein profiles were measured centrally at baseline and after 12 months GH replacement therapy. Patients Sixty‐one hypopituitary patients (30 male, 31 female) on maintenance statin therapy (mean 2·5 ± 2·7 SD years before GH) (statin group – SG) and 1247 (608 male, 639 female) patients not on hypolipidaemic therapy (nonstatin group – NSG) were studied. All patients were naïve or had not received GH replacement during the 6 months prior to study. Patients who developed diabetes mellitus during the first year of GH therapy or in the subsequent year and those with childhood onset GHD were excluded from this analysis. An established diagnosis of diabetes mellitus was present in 18% SG and 4·4% NSG at baseline. Measurements Serum concentrations of total, high density lipoprotein (HDL)‐cholesterol, triglycerides and IGF‐I were measured centrally in all patients and LDL‐cholesterol was estimated using Friedewald's formula. Results The relative frequency of various statin use was simvastatin 52% (15·8 ± 8·1 mg, mean ± SD), atorvastatin 30% (14·4 ± 7·8 mg), pravastatin 9·8% (31·6 mg ± 13·9 mg), lovastatin 6·6% (17·5 ± 5 mg) and fluvastatin 1·6% (40 mg). Baseline serum total and LDL‐cholesterol (mean ± SD) were 5·2 ± 1·4 and 3·1 ± 1·3 mmol/l in SG and 5·8 ± 1·2 and 3·7 ± 1·0 mmol/l in NSG, respectively (P < 0·0001, SG vs. NSG). After 12 months GH replacement (SG: 0·32 ± 0·17 mg/day; NSG: 0·38 ± 0·1 mg/day) serum total and LDL‐cholesterol decreased by a mean (±SD) of 0·48 (± 1·25) mmol/l (P < 0·0004) and 0·53 (± 1·08) mmol/l (P < 0·0001) in SG and by 0·30 (± 0·89) mmol/l (P < 0·0001) and 0·28 (± 0·80) mmol/l (P < 0·0001) in NSG, respectively. There were no significant changes in HDL‐cholesterol or triglycerides in either group (SG vs. NSG: NS). A relationship between LDL‐cholesterol at baseline and the decrease in LDL‐cholesterol after 12 months GH was evident in both groups (SG: R = –0·54, P < 0·001; NSG: R = –0·4, P < 0·001) and a similar relationship for cholesterol was observed. Conclusions These data indicate that GH replacement exerts additional beneficial effects on lipoprotein profiles in patients on maintenance statin therapy, confirming that the effects of these interventions are complementary rather than exclusive.     

Cardiovascular effects of prolonged growth hormone replacement in adults

Abstract. Objectives . To study the cardiovascular effects of human growth hormone (GH) replacement therapy in adults. Intervention . Biosynthetic human GH given in a daily dose of 0.04 ± 0.01 IU kg^?1 for 6–18 months in an open trial. Patients . Thirty-four GH-deficient hypopituitary patients on conventional replacement therapy, aged 19–67 years and with a body mass index of 18.0–410.0 kg/m². Measurements . Resting blood pressure, exercise tolerance, renal function and routine blood counts were assessed every 6 months. Two-dimensional echocardiography and Doppler ultrasound scanning were performed at 0, 6 and 12 months of GH therapy. Results . Exercise time increased significantly on GH from 9.37 ± 2.64min at the start to 10.39 ± 2.86 min (P < 0.001), 10.90 ± 2.48 min (P < 0.001) and 11.11 ± 0.70 min (P < 0.001) at 6, 12 and 18 months respectively. There was no change in the heart rate or in the blood pressure at rest nor at the peak of exercise. No significant changes were observed in measures of cardiac structure (left ventricular mass index, left ventricular posterior wall thickness and interventricular septal thickness), ejection fraction nor in cardiac output. Isovolumic relaxation time, a marker of diastolic function, decreased in 24 patients after 6 months on GH (from 98.6 ± 15.9 to 89.6 ± 15.2 ms; P < 0.03) but it was not different from baseline in the 18 patients who were restudied at 12 months. There was no significant change in the left ventricular filling neither at 6 nor at 12 months. No significant changes were observed in plasma electrolytes, creatinine nor in blood count on GH treatment. Conclusions . Growth hormone replacement therapy in hypopituitary adults for 6–18 months produced sustained increase in exercise tolerance but was not associated with changes in cardiac structure or systolic function.     

An account of the quality of life of patients after treatment for non-functioning pituitary tumours

OBJECTIVE Studies assessing quality of life in GH-deficient adults have shown varying results. This may be due to a number of factors including varying causes of GH deficiency, the use of radiotherapy in treatment and patient selection. We aimed to assess whether anterior pituitary hormone deficiency or external pituitary radiotherapy influenced the quality of life of patients with non-functioning pituitary tumours. PATIENTS We studied 48 patients treated and under follow up for non-functioning pituitary tumours on standard hormone replacement therapy excluding GH. There were 21 females and 27 males with a mean age of 59±12 years. We also studied 42 control patients who had undergone mastoid surgery and were followed at least annually. There were 17 females and 25 males with a mean age of 61±14 years. DESIGN All patients attended a research clinic and completed the Short Form 36 (SF36) and General Well Being Schedule (GWBS) to assess quality of life. Thyroid function tests, IGF1 and IGFBP3 were measured on all patients. Gonadotrophin and cortisol measurements were made on the patients with pituitary disease where appropriate. RESULTS IGF1 and IGFBP3 levels were lower in the pituitary patients compared with controls: 104±98 vs 143±37 μg/l (P<0.0001) and 2.9±0.75 vs 3.3±0.52 mg/l (P<0.004). There were no significant differences in the quality of life scores between the pituitary patients and the control subjects. There was also no difference in quality of life between pituitary patients with two or more hormone deficiencies (n=29) compared with controls. Patients who had received radiotherapy (n=18), when compared with controls, had a decreased mental health score using the SF36 71±21 vs 81±17 (P<0.05) and decreased total GWBS score 70±20 vs 82±17 (P<0.05). Subscore analysis of GWBS showed this to be due to depression and decreased control of emotions. CONCLUSIONS We found that the quality of life of patients treated and under follow up for non-functioning pituitary tumours was similar to that of patients treated by mastoid surgery and under similar follow up. The pituitary patients deficient in two hormones and thus most likely to be GH deficient were also similar to the controls. These results suggest that adding GH replacement in this patient group may not be routinely indicated for improvement in quality of life and needs careful assessment. Patients who had received radiotherapy were more depressed and anxious than controls. Further investigation into the psychological and psychomotor effects of radiotherapy in this group of patients is required.     

The 'dawn phenomenon' in adolescents with insulin dependent diabetes mellitus: possible contribution of insulin-like growth factor binding protein-1

OBJECTIVE Insulin resistance increases during adolescence, and is exaggerated in patients with insulin dependent diabetes mellitus (IDDM). A relative deficiency of insulin-like growth factor-I (IGF-I) may contribute to this increased insulin requirement. Two mechanisms have been proposed: (a) increased GH secretion, caused by failure of IGF feedback control, leading to increased insulin resistance and (b) lack of Insulin-like action of the IGFs which is reinforced by high plasma levels of IGFBP-1, an inhibitor of IGF action. The contribution of these two mechanisms to the ‘dawn phenomenon’ is assessed. DESIGN The two possible mechanisms were studied during the dawn rise of glucose in pubertal adolescent patients with IDDM. Two overnight studies were performed in each subject. Patients remained on the same insulin regimen throughout. SUBJECTS Twenty-two diabetic adolescent subjects, aged (mean ± SEM) 14.0 ± 0.4 years, duration of IDDM 7.9 ± 0.8 years, were recruited. Pubertal status was: group 1 (breast stage 1–2; testicular volume < 4–8ml) 3 male and 4 female, group 2 (breast stage 3; testicular volume 10–12 ml) 0 male 4 female, group 3 (breast stage 4–5; testicular volume 15–25 ml) 4 male and 7 female. Height standard deviation score (mean ± SD) (-0.02 ± 099) and dally insulin dose (50.4 ± 3.1 U/day) did not change between studies. There were no differences in HbA1 (study A 11.26 ± 0.45%, study B 11.09 ± 0.42%). METHODS The subjects were admitted for the two studies 0.3 ± 0.03 years apart. Blood samples were taken via an indwelling cannula every 20 minutes between 1900 and 0700 h. MEASUREMENTS GH was assayed every 20 minutes, IGFBP-1, glucose and free Insulin every hour and lGF-I at 0700h. GH, IGFBP-1, IGF-I and free insulin were measured by radioimmunoassay. IGFBPs were also analysed by Western ligand blotting techniques. GH profiles were analysed by Pulsar and results compared by paired Student's t-test. The relations between the dawn rise In glucose and the changes in IGFBP-1, GH and free insulin were examined by multiple linear regression analysis. RESULTS Serum IGFBP-1 levels rose overnight In the two studies (study A, from 9·1 at 2200 to 59 ± 9μg/l at 0700 h; study B, from 10±1 at 2100 to 64 ± 14μg/l at 0700h) whilst insulin levels fell from 47 ± 5 at 2200 to 16 ± 2mU/l at 0700 h (study A) and from 45·5 at 2000 to 14 ± 2mU/l at 0700 h (study B). Glucose levels fell from 16.0 ± 1.0 to 9.3 ± 0.9 mmol/l at 0400h, and then rose to 11.9 ± 1.1 mmol/l at 0700 h during study A, and from 13.4 ± 1.3 to 10.1 ± 1.1 mmol/l at 0400 h and then rose to 13.5 ± 1.0 mmol/l at 0700 h during study B. There were no differences in GH secretion between studies (mean GH levels (mean ± SD) (study A, 15.7 ± 6.6 mU/I; study B, 16.2 ± 7.1 mU/l; correlation within subjects between studies r= 0.77, P < 0.001), sum of GH peaks (study A, 189.9 ± 903 mU/l; study B, 185.8 ±100.2 mU/l; r= 0.57, P= 0.006)). Mean GH levels varied with pubertal stage (group 1,12.1 ± 1.5 mU/l; group 2, 23.3 ± 2.1 mU/I; group 3, 15.3 ± 1.2mU/I). Serum IGF-l levels were not different (study A, 203 ± 12 μg/l; study B, 218 ± 13 μg/l). REGRESSION ANALYSIS The change In plasma glucose between 0200 and 0700 h In both studies related to free insulin, IGFBP-1 and the sum of the GH levels over the preceding hour (log glucose =7.87 + 5.32 log IGFBP-1 (P= 0.0001) ?5.05 log free insulin (P= 0.0001)-1.44 log GH (P= 0.004); R²= 72%). Mean overnight GH levels did not predict the morning rise In plasma glucose. CONCLUSION The morning rise of IGFBP-1 and plasma glucose appear to be related in this group of subjects with IDDM and this was a consistent finding In the two studies. This relation was additive to the effect of Insulin deficiency. No positive relation was noted between GH secretion and glucose levels. These findings support the hypothesis that the increased GH secretion In IDDM Is a marker of IGF-I deficiency rather than a direct causal factor in the increase In insulin resistance. The IGFs may therefore have a direct role In glucose homeostasis via the ‘free’ fraction of circulating IGFs, the availability of which may be modulated by changes in IGFBP-1 levels.     

Cholinergic control of growth hormone (GH) responses to GH-releasing hormone in insulin dependent diabetics: evidence for attenuated hypothalamic somatostatinergic tone and decreased GH autofeedback

OBJECTIVE We have investigated the effects of cholinergic modulation with pirenzepine and pyridostigmine and of GH pretreatment on the subsequent GH response to a maximal stimulatory dose of GH-releasing hormone (GHRH) in patients with insulin dependent diabetes mellitus (IDDM). We have also investigated the relationship between the differences in metabolic control and other parameters of disease state with the differences in GH responses in IDDM. PATIENTS Thirteen male subjects with IDDM and no clinical evidence of complications were selected based on HbA, levels to provide a wide range of metabolic control. Seven normal subjects were also studied. DESIGN Twelve of the subjects with IDDM and six normal subjects received pirenzepine 200 mg and pyriostigmine 120 mg pretreatment 60 minutes and GH pretreatment 3 hours before an i.v. injection of GHRH (1–44) (80 μg) in random order. All subjects underwent a control study with GHRH alone. MEASUREMENTS Serum GH and plasma glucose were measured at regular intervals throughout the study. Fasting plasma glucose and HbA, were measured before each study to provide measures of metabolic control. RESULTS Subjects with IDDM demonstrated exaggerated GH responses to GHRH compared to normals. Pirenzepine significantly reduced GH responses in both normal and diabetic subjects. However, the GH response to GHRH after pirenzepine was higher in subjects with IDDM (mean GH: IDDM vs normals; 8.1 ± 1.3 vs 2.9 ± 0.7 mU/l, P < 0.05). Pyridostigmine 120 mg significantly augmented the GH response to GHRH in normal subjects. In diabetic subjects, pyridostigmine failed to increase GH response to GHRH compared to GHRH alone (mean GH: pyridostigmine vs control: 75.7 ± 12.6 vs 38.9 ± 5.4 mU/l, P= NS). GH responses to GHRH after pyridostigmine pretreatment in both normal and diabetic subjects did not differ and the GH response to GHRH after pyridostigmine in normal subjects did not differ from the GH response to GHRH alone in diabetic subjects. In normal subjects, GH pretreatment significantly reduced subsequent GH responsiveness to GHRH (Δpeak GH 26.4 ± 5.2 vs 7.7 ± 5.4 mU/I, P< 0.04). In contrast, GH pretreatment did not cause any significant reduction in GH responsiveness to GHRH In diabetics (Δpeak GH 53.6 ± 9.7 vs 33.4 ± 11 mU/I, P= NS). No significant correlation was demonstrated between measures of diabetic control and the responses to GHRH alone or after cholinergic modulation and GH pretreatment. CONCLUSION These data suggest that ambient hypothalamic cholinergic tone in diabetes is high, and of similar degree to the enhanced cholinergic tone in normal subjects pretreated with pyridostigmine. We suggest that in diabetic subjects, the reduced responsiveness to autofeedback may be secondary to the enhanced cholinergic tone demonstrated in these patients. The mechanisms linking the uncontrolled diabetic state to this abnormal neuroregulation of GH remains unknown at present.     

The effect of growth hormone replacement on cortisol metabolism and glucocorticoid sensitivity in hypopituitary adults

OBJECTIVE Growth hormone (GH) replacement therapy in hypopituitary adults is associated with sodium and water retention. The underlying mechanisms are incompletely understood and a possible contribution of altered cortisol metabolism or action has not been evaluated. We have investigated the effect of GH replacement therapy on cortisol metabolism, cortisol binding globulin and In-vitro glucocorticold sensitivity in a group of adult hypopituitary patients. DESIGN AND PATIENTS We studied 19 adult hypopituitary patients (18 adult onset, Y:F, 6:13), who were receiving conventional hydrocortisone (16 patients), thyroxine (14 patients), triiodothyronine (1 patient), sex steroid (9 patients), human chorionic gonadotrophin (1 patient) or desmopressin (6 patients) replacement during a 6-month, double blind controlled trial of GH therapy (active: placebo, 8:11) followed by a 6-month open phase of GH (mean dose: 0·2 IU/kg/week, range 0·051-0·27) and after a 6-week washout phase following discontinuation of GH therapy. MEASUREMENTS Twenty-four-hour urine free cortisol, cortisol metabolites (CoM), ratio 11-hydroxy/ll-oxo CoM (F/E) and ratio 5α/β tetrahydrocortisol were measured at 6 months, 12 months and after the 6 week washout phase. Serum cortisol binding globulin was measured basally, at 6 months, 12 months and after washout. Glucocorticold sensitivity was determined in lymphocyte preparations from 8 patients, during GH therapy and after washout, using an In-vitro technique dependent on dexamethasone suppression of phytohaemagglutinin-stimulated thymidine Incorporation into DNA. Plasma renin activity and aidosterone were measured after 6-12 months GH therapy and after washout. RESULTS After 6 months of GH, In patients on hydrocortisone (n= 9), there were significant decreases in CoM (mean decrement 21%, P < 0·01), F/E (mean decreased from 1·27to 1·0, P= 0·04; reference range0·33-1·29) and 5α/5β tetrahydrocortisol (mean decreased from 0·67 to 0·48, P= 0·01) and a subsequent increase after washout. Patients not on hydrocortisone (n= 2) demonstrated a normal basal F/E falling by 25% on GH therapy but no change in CoM. During 12 months of OH therapy, patients on hydrocortisone (n= 7) demonstrated a further trend to decrement in CoM (P= 0·09) which reversed after washout (P= 0·04). Urine free cortisol tended to fall during OH therapy and Increased significantly following washout after 12 months treatment (P < 0·02). Serum cortisol binding globulin decreased by 20% (P < 0·05) during 12 months GH treatment but remained within the reference range. In-vitro studies demonstrated a trend to reduced giucocorticoid sensitivity on GH therapy; the maximum inhibition of phytohaemaggiutinin by dexamethasone tended to be less on OH therapy (P= 0·052) and was also lower than in 29 normal volunteers (P < 0·05). There were no significant changes in plasma renin but there was a small increment in aidosterone in recumbent patients (P= 0·04) during the open phase of GH therapy in the placebo arm. CONCLUSIONS GH therapy in hypopltuitary adults is associated with an apparent reduction in availability of administered hydrocortisone as measured by urine cortisol metabolites and urine free cortisol. This effect is unlikely to be clinically significant except possibly in ACTH deficient subjects on suboptimal hydrocortisone replacement. The changes in F/E suggest that GH may directly or indirectly modulate the activity of 11β-hydroxysteroid dehydrogenase. The apparent decrease in giucocorticoid sensitivity during GH therapy, demonstrated in vitro, merits further investigation.     

Effects of morning cortisol replacement on glucose and lipid metabolism in GH-treated subjects

OBJECTIVE: Insulin resistance is a frequent consequence of GH replacement therapy but patients on GH replacement therapy often also have replacement of other hormone deficiencies which theoretically could modify the metabolic effects of GH. In particular, cortisol replacement if given in supra physiologic doses immediately before the evaluation of insulin sensitivity could influence insulin sensitivity. The aim of this study was thus to evaluate the effect of morning cortisol replacement given prior to a euglycaemic clamp combined with infusion of [3-(3)H]glucose and indirect calorimetry on glucose and lipid metabolism. METHODS: Ten GH/ACTH-deficient adults received, in a double-blind manner, either cortisol (A) or placebo (B) before the clamp whereas five GH-deficient-ACTH-sufficient adults participated in a control (C) clamp experiment. All subjects received GH replacement therapy. RESULTS: Serum cortisol levels were significantly higher after cortisol than after placebo (324+/-156 vs 132+/-136 mmol/l; P=0.006) and similar to controls (177+/-104 mmol/l). As a measure of the biological effect of cortisol, eosinophil leukocyte counts in peripheral blood decreased (164+/-91 x 10(9)/l vs 216+/-94 x 10(9)/l; P=0.04). Cortisol replacement had no significant effect on insulin-stimulated glucose uptake (11.8+/-1.8 vs 13.2+/-3.9 micromol/kg min), either on glucose oxidation or on glucose storage. There was also no significant effect of cortisol on fasting endogenous glucose production and no effect was seen on serum free fatty acid concentrations. CONCLUSION: Administration of cortisol in the morning before a clamp cannot explain the insulin resistance seen with GH replacement therapy.     

Pharmacokinetics and metabolic effects of growth hormone injected subcutaneously in growth hormone deficient patients: thigh versus abdomen

OBJECTIVE The absorption of insulin following subcutaneous (s.c.) injection is faster in the abdomen than the thigh. We therefore studied the effect of changing the site of injection on the absorption and metabolic effects of human growth hormone. DESIGN AND MEASUREMENTS in a cross-over study human GH (Norditropin) was injected s.c. in the thigh or abdomen in random order. Ultrasonography of the thigh and abdomen was performed in order to evaluate the thickness of the s.c. tissue. After each treatment period (4 weeks), serum profiles of GH, IGF-I, IGF binding proteins 1 and 3 (IGFBP-1 and IGFBP-3), glucose, insulin, non-esterified fatty acids (NEFA), glycerol, 3-hydroxy-butyrate, alanine, lactate and glucagon were measured for 37 hours after GH injection (3 IU/m² at 1900 hour). PATIENTS Nine GH deficient patients (five males, four females). RESULTS The mean (± SEM) thickness of the s.c. tissue (mm) was higher on the abdominal site (9·35 ± 1·38 (thigh), and 22·61 ± 2·19 (abdomen), P<0·001). Mean (±SEM) integrated levels (area under the curves (AUC) divided by time) of GH (mU/l) were identical: 5·54 ± 0·70 (thigh) versus 5·48 ± 0·64 (abdomen) (P = 0·91). AUC (mU/l) for the initial 6 hours were, however, significantly different (14·10 ± 3·76 (thigh) and 19·02 ± 3·18 (abdomen), P = 0·02). Maximal serum concentration (C_max) (mU/l) 23·18 ± 3·86 (thigh) and 29·66 ± 4·78 (abdomen) (P = 0·19) was achieved faster (T_max) following injection in the abdomen. T_max (hours) was 5·89 ± 0·41 (thigh) and 4·26 ± 0·49 (abdomen) (P<0·002). Mean IGF-I levels (μg/l) were unaffected by GH injection sites (355 ± 60 (thigh) and 365 ± 63 (abdomen), P = 0·61). Mean IGFBP-3 levels (μg/l) were significantly different (2100 ± 143 (thigh), and 2350 ± 176 (abdomen), P = 0·05). Mean levels of IGFBP-1, insulin, glucose, lipid intermediates, metabolites and glucagon were not significantly different. CONCLUSIONS Human GH was absorbed faster when injected s.c. In the abdomen as compared with the thigh, despite the thicker s.c. tissue on the abdomen. Apart from higher IGFBP-3 levels after s.c. injections in the abdomen, similar metabolic effects of GH were obtained with the two injection sites.     

Short-term fasting in obesity fails to restore the blunted GH responsiveness to GH-releasing hormone alone or combined with arginine

OBJECTIVE Fasting is known to clearly increase both spontaneous and GHRH-stimulated GH secretion in normal subjects and this effect is likely to be due to hypothalamic mechanism(s). Our aim was to clarify the effect of a 3 or 4-day fast, on the GH response to GHRH alone or combined with arginine, an amino acid probably acting via inhibition of hypothalamic somatostatin release. DESIGN Two tests with GHRH (1 μg/kg i.v.), administered either alone or in combination with arginine (ARG, 0.5 g/kg i.v.) were performed, in a randomized order at least 3 days apart. In obese women the two tests were repeated after a 3 or 4-day fast. PATIENTS Seven obese women (06, aged 17–54 years, BMI 42.4 ± 3.6 kg/m², waist-hip ratio (WHR) 0.85 ± 0.01) and ten healthy women, as control subjects (CS, aged 20–44 years, BMI 23.1 ± 1.1 kg/m², WHR 0.79 ± 0.01) were studied. MEASUREMENTS Serum GH and IGF-I levels were measured by radioimmunoassay. The GH secretory responses were expressed either as absolute values (mU/l) or as areas under the curve (AUC, mU/l/h) calculated by trapezoidal integration. IGF-I concentrations were expressed as absolute values (μg/l) with reference to a pure recombinant IGF-I preparation. Results are expressed as mean ± SEM. RESULTS Basal GH and IGF-l levels in OB were lower than in CS (0.8 ± 0.2 vs 4.8 ± 1.0 mU/I, P< 0.0001 and 120.1 ± 21.4 vs 188.7 ± 13.1 μg/l, P <0.02, respectively). The GHRH-induced GH rise in OB was lower (P <0.00001) than in CS (AUC 340.2 ± 81.0 vs 2125.0 ± 199.6 mU/l/h). ARG increased the GHRH-induced GH rise in both groups, but in OB the GH response to ARG + GHRH (1458.4 ± 439.0 mU/l/h, P < 0.03 vs GHRH alone) remained lower (P <0.0001) than in CS (6396.2 ± 772.2 mU/l/h, P < 0.01 vs GHRH alone). In spite of a reduction in body weight and IGF-I, Insulin and glucose levels, in OB fasting failed to modify both the basal GH levels and the somatotroph responsiveness to GHRH when administered either alone or combined with ARG. An increase in free fatty acids (FFA) was also found after fasting. CONCLUSIONS The results of this study demonstrate that in obesity the somatotroph hyporesponsiveness to GHRH, either alone or combined with arginine, is not improved by short-term fasting. As fasting is considered a CNS mediated stimulus to GH secretion, its ineffectiveness in obesity does not support a hypothalamic pathogenesis and suggests that long standing metabolic alterations, such as hyperinsulinaemia and/or elevated free fatty acids, could play a major role in causing GH insufficiency in obese patients.     

Altered endogenous growth hormone secretory kinetics and diurnal GH-binding protein profiles in adults with chronic liver disease

OBJECTIVE Increased serum GH concentrations and GH responses to a variety of stimuli have been reported in patients with chronic liver disease (CLD). We Investigated the pulsatile pattern of endogenous GH release and GH-binding protein (GHBP) and insulin-like growth factor-I (IGF-I) diurnal profiles in adults with cirrhosis, in comparison with healthy, matched control subjects. DESIGN Case-control, cross-sectional. PATIENTS Seven patients with biopsy proven cirrhosis, and sex, age, height, weight and oestrogen status matched controls. MEASUREMENTS Serum immunoreactive GH concentrations in samples collected at 20-minute Intervals for 24 hours were analysed using a multi-parameter deconvolution method to simultaneously resolve endogenous GH secretory and disappearance rates. Diurnal patterns of GHBP (specific immunoprecipitation method) and serum IGF-I (RIA after acid-ethanol extraction) were assessed. RESULTS The mean dally GH secretion rate in patients with CLD was Increased (210±93 vs 100±55mU/1/day; P= 0.025), and GH disappearance half-time was prolonged (43±10 vs 24±9 min; P= 0.006) compared to controls. Detectable GH secretory bursts were more frequent In patients with CLD (10±1 vs 6±3/day; P= 0.038), but of similar mean mass (21±10 vs 17±8mU/l) compared to controls. In patients with CLD, mean serum GHBP was slightly lower (63±36 vs 71±14% pooled control; P > 0.1). Fasting serum IGF-I concentrations (after size-exclusion HPLC) were lower in the patients with CLD (13±5 vs 21±2 nmol/I; P < 0.0001). Multiple regression analysis showed that GH secretion rate was increased In patients with CLD with higher Child's classifications (R²= 0.86; P= 0.002) and with lower serum IGF-I concentrations measured after HPLC (R²= 0.11; P= 0.044). CONCLUSIONS Adults with chronic liver disease have (1) Increased total daily GH secretion rates, which appear to be influenced by disease severity and diminished serum IGF-I-mediated negative feedback; (2) markedly impaired endogenous GH clearance, possibly reflecting changes In hepatic GH-receptor status; and (3) GHBP levels which do not correlate with GH kinetics or serum IGF-I concentrations.     

Glucose Metabolism and Visceral Fat in GH Deficient Adults: Two Years of GH-Replacement

The aim of this study was to evaluate the effect of 24 months of growth hormone (GH) replacement on glucose metabolism and visceral fat in 17 adults with GH deficiency: 9 men and 8 women; age 40 ± 1.8 yr. [range 20–61] and body mass index 25 ± 0.8 Kg/m². Glucose metabolism was evaluated by a standard oral glucose tolerance test (OGTT), by the homeostatic model assessment (HOMA) insulin resistance index and by the insulin sensitivity index (ISI)-composite derived from the OGTT. Visceral fat was evaluated by CT scan. Twenty-four months of GH replacement induced an increase in the prevalence of abnormal glucose tolerance, with significant progressive increment in 2h-OGTT insulin levels at 3, 12 and 24 months (p = 0.005). Plasma glucose levels and ISI-composite did not alter during the study. HOMA-IR index increased only in the group of patients (n = 8) who had abnormal OGTT at 24 months (p = 0.012). Visceral fat reduced at month 12 and remained decreased until the end of the study (p = 0.009). In conclusion, the present study suggests that adults with GH deficiency after twenty-four months of GH replacement developed abnormal glucose tolerance, probably due to an increase in insulin resistance, associated with higher insulin levels, despite favorable alterations in body composition. Research grant-FAPER. Senior researcher-CNPQ.     

The effect of the deterioration of insulin sensitivity onβ-cell function in growth-hormone-deficient adults following 4-month growth hormone replacement therapy

The purpose of the present study was to evaluate the combined effect of GH treatment on body composition and glucose metabolism, with special focus on β-cell function in adult GHD patients. In a double-blind placebo-controlled design, 24 GHD adults (18M/6F), were randomized to 4 months treatment with biosynthetic GH 2 IU/m²s.c. daily (n=13) or placebo (n=11). At inclusion and 4 months later an oral glucose tolerance test (OGTT), a frequently sampled intravenous glucose tolerance test (FSIGT) and dual-energy X-ray absorptiometry (DXA) whole-body scanning were performed. During the study period, body weight decreased 1.6 kg from 94.0 ± 18.7 to 92.4 ± 19.4 kg (mean ± SD) (P<0.05) in the GH-treated group, but remained unchanged in the placebo group. Fat mass decreased from 32.4 ± 9.6 to 28.1 ± 10.5 kg (P<0.001), whereas lean body mass increased from 58.3 ± 11.5 to 61.0 ± 11.7 kg (P<0.01) in the GH-treated group. Treatment with GH for 4 months resulted in a significant increase in fasting blood glucose (before GH 5.0 ± 0.3 and after 5.4 ± 0.6 mmol/l,P<0.05), fasting plasma insulin (before GH 38.4 ± 30.2 and after 55.3 ± 34.7 pmol/l,P<0.02) and fasting proinsulin (before 8.1 ± 6.7 and after 14.6 ± 16.1 pmol/l,P<0.05). The insulin sensitivity index S_I, estimated by Bergmans Minimal Model, decreased significantly [before GH 1.1 ± 0.7 and after 0.4 ± 0.2 10^–4(min × pmol/l),P<0.003]. The non-insulin-dependent glucose uptake (glucose effectiveness S_Gdid not change (before GH 0.017 ± 0.005 and after 0.015 ± 0.006 min^–1, NS). Insulin secretion was enhanced during GH therapy, but insufficiently to match the changes in S_I, resulting in a higher blood glucose level during an OGTT. Blood glucose at 120 min was 5.5 and 6.3 mmol/l before and after GH treatment, respectively (P= 0.07). One patient developed impaired glucose tolerance.Short-term GH replacement therapy in a dose of about 2 IU/m²daily in GHD adults induces a reduction in insulin sensitivity, despite favourable changes in body composition, and an inadequate enhancement of insulin secretion.