First-trimester fasting glycemia as a predictor of gestational diabetes (GDM) and adverse pregnancy outcomes.

AIMS: Studies to prevent gestational diabetes (GDM) have shown the best results when lifestyle measures have been applied early in pregnancy. We aimed to investigate whether first-trimester fasting plasma glucose (FPG) could predict GDM risk and adverse pregnancy outcomes. METHODS: A retrospective analysis of prospectively collected data from singleton pregnancies who were attended at our hospital between 2008 and 2018 (n = 27,198) was performed. We included patients with a recorded first-trimester FPG and complete pregnancy data (n = 6845). Patients under 18, with pregestational diabetes or reproductive techniques, were excluded. First-trimester FPG was evaluated as a continuous variable and divided into quartiles. GDM was diagnosed by NDDG criteria. The relationship between first- and second-trimester glucose > 92 mg/dL was also investigated. The relationship between FPG and pregnancy outcomes was assessed in 6150 patients who did not have GDM. RESULTS: Maternal age was 34.2 ± 3.9 years, BMI 23.1 ± 3.7 kg/m2 and mean FPG 83.0 ± 7.3 mg/dL. Glucose quartiles were: ≤ 78, 79-83, 84-87 and ≥ 88 mg/dL. First-trimester FPG predicted the risk of GDM (7%, 8%, 10.2% and 16% in each quartile, p < 0.001) and the risk of second-trimester glucose > 92 mg/dL (2.6%, 3.8%, 6.3% and 11.4% in each quartile, p < 0.001). FPG was significantly associated with LGA (8.2%, 9.3%, 10% and 11.7% in each quartile, p = 0.011) but not with other obstetrical outcomes. In a multivariate analysis including age, BMI, tobacco use, number of pregnancies and weight gained during pregnancy, first-trimester FPG was an independent predictor of LGA. CONCLUSIONS: First-trimester FPG is an early marker of GDM and LGA.

Maternal fasting glycemia and adverse pregnancy outcomes in a Mediterranean population.

AIMS: The hyperglycemia and adverse pregnancy outcome study demonstrated a continuous association between fasting plasma glucose (FPG) levels below those diagnostic of diabetes and adverse neonatal outcomes. We aimed to investigate whether the same association was found in a Mediterranean population. METHODS: A retrospective analysis of singleton pregnancies attended at our Hospital between 2008 and 2015 (n = 5203). FPG was evaluated in the second trimester, and it was divided into 7 categories (1 < 75, 2 75-79, 3 80-84, 4 85-89, 5 90-94, 6 95-99 and 7 100-124 mg/dL). Pregnancy outcomes included elective cesarean delivery, gestational hypertensive disorders (GHD), large for gestational age (LGA), small for gestational age (SGA), macrosomia, prematurity, severe prematurity and APGAR at 1 min <7. RESULTS: Maternal age was 33.8 ± 3.8 years, and BMI at first antenatal visit was 22.9 ± 3.5 kg/m2; mean FPG was 79 ± 7 mg/dL. A positive association was observed between FPG and LGA (p < 0.001), GHD (p = 0.004) and prematurity both <37 and <34 weeks of gestation (p = 0.001 and p = 0.004). FPG and SGA were inversely related (p = 0,038). FPG was not significantly related to rate of C-section or APGAR. Adjusted odds ratios associated with 1 standard deviation increase in the fasting plasma glucose (7 mg/dL) were 1.26 (1.15 to 1.37) for LGA, 1.28 (1.09 to 1.49) for GHD and 0.83 (0.74-0.93) for SGA. In a multivariate analysis controlling for confounders, FPG remained associated with LGA. CONCLUSIONS: We found an association between FPG levels, below those diagnostic of gestational diabetes according to our guidelines, and adverse maternal and neonatal outcomes in a Mediterranean population.

Effects of nicotinamide and intravenous insulin therapy in newly diagnosed type 1 diabetes.

OBJECTIVE: To investigate the effect of intravenous insulin therapy combined with nicotinamide in the metabolic control and beta-cell function of newly diagnosed type 1 diabetic subjects in comparison with intensive insulin therapy and nicotinamide alone. RESEARCH DESIGN AND METHODS: A total of 34 newly diagnosed type 1 diabetic patients were included. After the correction of initial metabolic disturbances, subjects were randomly assigned to the following three groups within 72 h after admission: 1) intensive insulin therapy + placebo (C) (n = 12); 2) intensive insulin therapy + nicotinamide, 700 mg three times a day (NIC) (n = 11); and 3) 72-h intravenous insulin followed by intensive insulin therapy + nicotinamide, 700 mg three times a day (NIV) (n = 11). The subjects were monitored for 12 months. GAD, tyrosine phosphatase antibodies, and insulin autoantibodies were measured. C-peptide was measured basally and after 2, 4, 6, 8, and 10 min of 1 mg intravenous glucagon. HbA1c, glucagon, and antibody measurements were determined initially and at 1, 3, 6, 9, and 12 months. RESULTS: HbA1c values declined to normal after treatment was initiated in all groups and remained not significantly different during the follow-up period. We did not find differences between experimental (NIC and NIV) and placebo (C) groups in terms of beta-cell function, considering basal or glucagon-stimulated C-peptide (maximal stimulated C-peptide and area under the curve [AUC] of C-peptide) values during the follow-up period. After pooling data from the NIC and NIV groups (both including nicotinamide) and comparing it with data from the C group, the results remained unchanged. At diagnosis, GAD positivity was observed in 10 of 12, 8 of 11, and 10 of 11 subjects (NS) in the C, NIC, and NIV groups, respectively, and IA2 positivity was observed in 3 of 12, 4 of 11, and 4 of 11 subjects (NS) in the C, NIC, and NIV groups, respectively. Antibody titers displayed a similar behavior in all groups during the follow-up period. CONCLUSIONS: Our pilot study failed to demonstrate that the addition of 72-h intravenous insulin and nicotinamide to conventional intensive insulin therapy produces any beneficial effect in newly diagnosed type 1 diabetic subjects in terms of beta-cell function and metabolic control.

Inflammatory cardiovascular risk markers in women with hypopituitarism.

Patients with hypopituitarism have increased cardiovascular mortality. A high prevalence of conventional cardiovascular risk factors, including obesity, central fat distribution, insulin resistance, and dyslipidemia, have been described in these patients. The inflammatory markers C-reactive protein (CRP) and IL-6 are predictors of cardiovascular events, and high levels of CRP have been reported in men with hypopituitarism and GH deficiency. However, little is known about inflammatory cardiovascular risk markers in women with hypopituitarism. We therefore investigated whether inflammatory and traditional cardiovascular risk markers are elevated in women with hypopituitarism. Fifty-three women with hypopituitarism and 111 healthy control women were included in this cross-sectional study. Morning blood samples were drawn after an overnight fast. Serum was assayed for CRP, IL-6, glucose, insulin, IGF-I, triglycerides, total cholesterol, low density lipoprotein cholesterol, high density lipoprotein (HDL) cholesterol, lipoprotein(a), E2, total testosterone (total T) and free testosterone (free T), and dehydroepiandrosterone sulfate. IL-6 and CRP levels were higher in women with hypopituitarism than in healthy controls (P < 0.0001 for comparison between groups). In a multivariate model, CRP levels depended on hypopituitarism, body mass index (BMI), and estrogen use. There was an interaction between the effect of BMI and hypopituitarism on CRP levels, such that the importance of hypopituitarism in determining CRP levels disappeared at high BMIs. In a similar multivariate model, IL-6 levels depended on hypopituitarism and BMI. Total cholesterol, the total to HDL cholesterol ratio, and triglycerides were higher in hypopituitary patients, but only triglycerides and the total to HDL cholesterol ratio depended on hypopituitarism when controlling for BMI. There was no significant difference in lipoprotein(a) levels between hypopituitary women and control subjects. However, when controlling for estrogen use, lipoprotein(a) levels showed a trend toward being lower in the hypopituitary group (P = 0.075). In patients with hypopituitarism, CRP correlated negatively with IGF-I (r = -0.35; P = 0.010), total T (r = -0.42; P = 0.0020), and free T (r = -0.30; P = 0.031). Similarly, IL-6 correlated negatively with total T (r = -0.39; P = 0.0040) and androstenedione (r = -0.27; P = 0.048) in hypopituitary patients. In conclusion, hypopituitary women have increased levels of IL-6 and CRP, both of which are inflammatory markers of atherosclerosis. GH deficiency and androgen deficiency may contribute to these findings. Chronic inflammation may contribute to the high cardiovascular risk seen in this population.

Withdrawal of long-term physiological growth hormone (GH) administration: differential effects on bone density and body composition in men with adult-onset GH deficiency.

Adults with acquired GH deficiency (GHD) have been shown to have osteopenia associated with a 3-fold increase in fracture risk and exhibit increased body fat and decreased lean mass. Replacement of GH results in decreased fat mass, increased lean mass, and increased bone mineral density (BMD). The possible differential effect of withdrawal of GH replacement on body composition compartments and regional bone mass is not known. We performed a randomized, single blind, placebo-controlled 36-month cross-over study of GH vs. placebo (PL) in adults with GHD and now report the effect of withdrawal of GH on percent body fat, lean mass, and bone density, as measured by dual energy x-ray absorptiometry. Forty men (median age, 51 yr; range, 24-64 yr) with pituitary disease and peak serum GH levels under 5 microg/L in response to two pharmacological stimuli were randomized to GH therapy (starting dose, 10 microg/kg x day, final dose 4 microg/kg x day) vs. PL for 18 months. Replacement was provided in a physiological range by adjusting GH doses according to serum insulin-like growth factor I levels. After discontinuation of GH, body fat increased significantly (mean +/- SEM, 3.18 +/- 0.44%; P = 0.0001) and returned to baseline. Lean mass decreased significantly (mean loss, 2133 +/- 539 g; P = 0.0016), but remained slightly higher (1276 +/- 502 g above baseline; P = 0.0258) than at study initiation. In contrast to the effect on body composition, BMD did not reverse toward pretreatment baseline after discontinuation of GH. Bone density at the hip continued to rise during PL administration, showing a significant increase (0.0014 +/- 0.00042, g/cm2 x month; P = 0.005) between months 18-36. Every bone site except two (radial BMD and total bone mineral content), including those without a significant increase in BMD during the 18 months of GH administration, showed a net increase over the entire 36 months. Therefore, there is a critical differential response of the duration of GH action on different body composition compartments. Physiological GH administration has a persistent effect on bone mass 18 months after discontinuation of GH.

Effects of growth hormone (GH) administration on homocyst(e)ine levels in men with GH deficiency: a randomized controlled trial.

GH deficiency is associated with increased cardiovascular mortality and early manifestations of atherosclerosis. Elevated serum homocyst(e)ine levels have been found to be associated with increased cardiovascular risk. The effect of GH replacement on homocyst(e)ine has not been investigated to date. We evaluated the effect of GH replacement on fasting homocyst(e)inemia in a group of men with adult-onset GH deficiency in a randomized, single blind, placebo-controlled trial. Forty men with adult-onset GH deficiency were randomized to GH or placebo for 18 months, with dose adjustments made according to serum insulin-like growth factor I (IGF-I) levels. Fasting serum homocyst(e)ine, folate, vitamin B12, and total T(3) levels were determined at baseline and 6 and 18 months. Anthropometry, IGF-I levels, insulin, and glucose were measured at 1, 3, 6, 12, and 18 months. Nutritional assessment, body composition, total T(4), thyroid hormone binding index, and free T(4) index were assessed every 6 months. Homocyst(e)ine decreased in the GH-treated group compared with that in the placebo group (net difference, -1.2 +/- 0.6 micromol/L; confidence interval, -2.4, -0.02 micromol/L; P = 0.047). Homocyst(e)ine at baseline was negatively correlated with plasma levels of folate (r = -0.41; P = 0.0087). Total T(3) increased in the GH-treated group vs. that in the placebo group (net difference, 0.17 +/- 0.046 ng/dL; confidence interval, 0.071, 0.26 nmol/L; P = 0.0012). Folate and vitamin B12 levels did not significantly change between groups. Changes in homocyst(e)ine were negatively correlated with changes in IGF-I. For each 1 nmol/L increase in IGF-I, homocyst(e)ine decreased by 0.04 +/- 0.02 micromol/L (P = 0.029). In contrast, changes in homocyst(e)ine did not correlate with changes in folate, vitamin B12, total T(3), C-reactive protein, interleukin-6, or insulin levels. This study shows that GH replacement decreases fasting homocyst(e)ine levels compared with placebo. This may be one of the mechanisms involved in the putative modulation of atherosclerosis and cardiovascular risk by GH replacement.

Androgen deficiency in women with hypopituitarism.

Androgen deficiency in men is associated with severe osteopenia, alterations in body composition including an increase in fat mass, and decreased libido. Little is known about the pathophysiology, metabolic consequences, or gender-specific effects of androgen deficiency in women. Acquired hypopituitarism in women is characterized by central hypogonadism and/or hypoadrenalism and therefore may affect critical sources of androgen production in women. We hypothesized that serum androgen levels would be decreased in women with hypopituitarism. We therefore determined serum androgen levels in 55 women with hypopituitarism and 92 controls. This included 4 subsets of hypopituitary women: 1) women less than 50 yr old not receiving estrogen, 2) women less than 50 yr old receiving estrogen, 3) women more than 50 yr old not receiving estrogen, and 4) women more than 50 yr old receiving estrogen. Premenopausal controls with regular menstrual cycles were studied in the early follicular phase, midcycle, and luteal phase during one cycle. All other subjects were studied 3 times during the month at comparable intervals to mimic these 3 time points of the normal menstrual cycle. Serum testosterone, free testosterone, androstenedione, and dehydroepiandrosterone sulfate levels were markedly reduced in all 4 groups of women with hypopituitarism compared with controls (P < 0.0001). Moreover, serum testosterone, free testosterone, and androstenedione levels were lower in women with central hypoadrenalism and hypogonadism than in subjects with hypoadrenalism or hypogonadism alone (P < 0.025). Mean DHEAS levels were decreased in hypopituitary women with both hypogonadism and hypoadrenalism compared with those in women with hypogonadism alone (P < 0.0001). These data demonstrate hypoandrogenemia in women with hypopituitarism. The physiological consequences of low androgen levels in this population remain to be determined.

Role of thyroid hormones on serum leptin levels.

Leptin is an adipose tissue hormone whose plasma levels reflect energy stores. Although pathological thyroid function is related to changes in energy expenditure and body composition, its possible influence on leptin levels remains to be determined. The objective of the study was to provide new data on the relationship between plasma leptin levels and thyroid function. Sixteen patients with primary autoimmune hypothyroidism, and seventeen patients with primary autoimmune hyperthyroidism were prospectively studied from the time of clinical diagnosis and then every 6-8 weeks until thyroid function was completely restored (plasma tri-iodothyronine, free thyroxine and TSH within normal ranges). Fasting immunoreactive plasma leptin levels and body composition (bioelectrical impedance) were assessed at every visit. Plasma leptin levels were correlated with percentage body fat, as previously described, both at the time of diagnosis (r=0.60, P<0.001) and after normalisation of thyroid function (r=0.63, P< 0.001). There was no correlation between serum leptin and thyroid hormone levels at any time during the study. Plasma leptin levels as well as percentage body fat (BF) did not change significantly from the beginning until the end of the study, either in the hypothyroid (leptin: 14.54+/-2.61 vs 16.92+/-2.61 ng/ml, BF: 25.25+/-2.47 vs 25.90+/-3.22%) or in the hyperthyroid (leptin: 10.69+/-1.81 vs 12.36+/-2.19 ng/ml, BF: 22.01+/-2.31 vs 25.39+/-1.13%) group of patients. In conclusion, these results suggest that thyroid function per se is not a major determinant of plasma leptin levels.

Genetic and clinical characterisation of maturity-onset diabetes of the young in Spanish families.

OBJECTIVE: To investigate the frequencies of the major maturity-onset diabetes of the young (MODY) subtypes in a panel of Spanish families and to assess phenotypic differences in patients with the different subtypes of MODY. METHODS: Forty-eight subjects from twenty families with clinical diagnosis of MODY were studied. They underwent a standardised clinical examination and a 75-g oral glucose tolerance test (OGTT) was performed. Estimations of insulin sensitivity (%S) and insulin secretion capacity (%B) were calculated by the computer-solved homeostasis model assessment (HOMA). Mutations in the coding regions of hepatocyte nuclear factor (HNF)-4alpha/MODY1, glucokinase (GCK/MODY2) and HNF-1alpha/MODY3 genes were investigated by single strand comformation polymorphism and sequencing analysis. RESULTS: Mutations in the GCK and HNF-1alpha genes were observed in 5 (25%) and 7 (35%) families respectively. Novel mutations included R385X, M238fsdelT, V226fsdelTinsAA and S418-7del11 in the GCK gene, and S121fsdelC, V133M, R159Q and V259D in the HNF-1alpha gene. No MODY1 families were found. Subjects which were neither MODY2 nor MODY3 (MODY-X) had a higher fasting glucose than subjects in the other groups. Insulin secretion capacity was similar in the three groups and the insulin sensitivity was decreased in MODY-X subjects. Glucose levels were significantly higher and insulin levels significantly lower, throughout the OGTT, in MODY3 compared with MODY2 subjects. CONCLUSIONS: Mutations in the GCK/MODY2 and HNF-1alpha/MODY3 genes account for the majority of cases in a panel of Spanish MODY families, with MODY3 being the most frequent subtype. The relative frequencies and the clinical characteristics of these MODY subtypes are in agreement with data previously reported in other European populations. MODY-X patients seem to present a heterogeneous clinical profile.

Growth hormone treatment for sustained pain reduction and improvement in quality of life in severe fibromyalgia.

Functional defects in growth hormone (GH) secretion and its efficacy as a complementary treatment have been suggested for fibromyalgia. This study investigated the efficacy and safety of low-dose GH as an add-on therapy in patients with both severe FM and low insulin-like growth factor 1 levels. A total of 120 patients were enrolled in a multicenter, placebo-controlled study for 18 months. They were randomly assigned to receive either 0.006 mg/kg/day of GH subcutaneously (group A, n=60) or placebo (group B, n=60) for 6 months (blind phase). The placebo arm was switched to GH treatment from month 6 to month 12 (open phase), and a follow-up period after GH discontinuation was performed until month 18. Standard treatment for fibromyalgia (selective serotonin re-uptake inhibitors, opioids, and amitriptyline) was maintained throughout the study. Number and intensity of tender points, Fibromyalgia Impact Questionnaire (FIQ) with its subscales, and EuroQol 5 dimensions test (EQ5D) with visual analogue scale (VAS) were assessed at different time points. At the end of the study, 53% of group A patients obtained fewer than 11 positive tender points, vs 33% of group B patients (P<.05). 39.1% vs 22.4% reached more than 50% improvement in VAS (P<.05). Group A patients showed significantly improved FIQ scores (P=.01) compared with group B. Although GH discontinuation worsened all scores in both groups during follow-up, impairment in pain perception was less pronounced in the GH-treated group (P=.05). In this largest and longest placebo-controlled trial performed in FM (NCT00933686), addition of GH to the standard treatment is effective in reducing pain, showing sustained action over time.

High prevalence of growth hormone deficiency in severe fibromyalgia syndromes.

CONTEXT: Fibromyalgia (FM) is characterized by widespread pain and fatigue and is considered a syndrome with different pathogenic mechanisms. Controversial data on GH axis disturbances have been published. Some preliminary trials have shown promising effects of GH therapy on tender points and quality of life in FM. AIM: The aim was to study the patterns of GH secretion/sensitivity in a cohort of severe FM patients. SETTING: The study was conducted in five tertiary hospitals. METHODS: A total of 493 FM women (1990 American College of Rheumatology criteria) recruited from five centers, having more than 16 tender points, Fibromyalgia Impact Questionnaire scores above 75, more than 1 yr of stable medication (serotonin reuptake inhibitors, amitriptyline, and opioids), and body mass index below 35 kg/m(2) underwent baseline IGF-I/GH determinations; an insulin tolerance test (ITT) and a modified IGF-I generation test were performed in those cases showing IGF-I of 150 microg/liter or less. RESULTS: A total of 169 of the 493 patients (34.2%) showed IGF-I of 150 microg/liter or less. Mean peak GH during ITT was 13.3 +/- 9.9 ng/ml in 127 patients in which the test was performed. In 22 of 127 (17.3%), ITT peak GH was 5 microg/ml or less, and in eight of them (6.3%), the peak GH was 3 ng/ml or less. Mean baseline GH (n = 127) was 1.47 +/- 2.58 ng/dl, and eight of 120 (6.8%) showed an insufficient IGF-I response (<50% over baseline) to the IGF-I generation test. CONCLUSION: FM patients show a high prevalence of GH axis dysfunction. A significant number of patients show biochemical patterns of GH deficiency as well as some degree of GH resistance and might be potential candidates for substitution treatment.

A short acute octreotide test for response prediction of long-term treatment with somatostatin analogues in acromegalic patients.

The usefulness of the acute octreotide test in the selection of acromegalic patients for chronic somatostatin depot analogues treatment is controversial. The aim of the present study was to determine its accuracy for chronic response prediction, and the reliability of a short version of the classic 6-hour test. The data from 26 acromegalics (19 women, 7 men, mean age 52.6+/-13.1 years) studied with an acute octreotide test (6 hours sampling for GH measurement after octreotide 100 microg s. c.) were retrospectively analyzed. Eighteen of them followed chronic somatostatin depot analogues treatment for 12 months. GH nadir was always detected at 2 hours (mean decrease 75.9+/-24%). GH levels at 2 hours positively correlated with the other time-points (r(s) 0.97, 0.98, 0.97, 0.96 at 3, 4, 5 and 6 h, respectively; p<0.0001). During chronic treatment with maximal effective dose for 12 months, 61% of the patients achieved IGF1 <3 SD and 22% reached IGF1 <2 SD. GH nadir correlated with IGF1 decrease at 12 months (r(s) 0.76, p<0001). GH nadir of 9.2 ng/ml predicts IGF1 <3 SD with 82% sensitivity and 58% specificity (75% PPV, 67% NPV); for IGF1<2 SD, 75% sensitivity and 58% specificity are obtained for GH nadir 3.6 ng/ml, with 33% PPV and 89% NPV. Acute octreotide test reliably predicts response to long-term treatment; the short, 2-hour version is fully informative for therapeutic decisions in acromegalic patients.

Effects of a growth hormone receptor antagonist on bone markers in acromegaly.

OBJECTIVE: Excess GH secretion, as occurs in acromegaly, is associated with abnormalities in bone turnover markers and bone mineral density (BMD). GH administration in GH deficient patients causes an increase in bone turnover. IGF-I mediates many of the metabolic actions of GH, although GH may have direct effects upon bone. In patients with acromegaly who are treated with a GH receptor antagonist, selective blockade of the GH receptor results in a decrease in circulating IGF-I levels in the majority of cases. We hypothesized that, in acromegaly, antagonism of GH receptors would result in a decrease in serum markers of bone turnover, including serum procollagen I carboxy-terminal propeptide (PICP), osteocalcin and N-telopeptide (NTx). DESIGN AND SUBJECTS: Twenty-seven patients with acromegaly were enrolled as part of a multicentre 12-week trial of a GH receptor antagonist and were randomized to placebo (n = 7) or 10, 15 or 20 mg of pegvisomant (n = 20). MEASUREMENTS: Serum markers of bone turnover were determined at baseline and 12 weeks. RESULTS: Baseline bone turnover markers were above the upper limit of normal in 23%, 19% and 32% of subjects for osteocalcin, PICP and NTx, respectively. During the 12-week placebo-controlled period, there were significant decreases in serum markers of bone formation, osteocalcin (-2.2 +/- 0.44 vs. placebo +0.01 +/- 0.39 nmol/l, P = 0.009) and PICP (-23.6 +/- 9.6 vs. placebo +18.1 +/- 12.8 micro g/l, P = 0.022) and a serum marker of bone resorption, NTx (-4.4 +/- 1.4, placebo +1.0 +/- 0.3 nm, P = 0.024). CONCLUSIONS: Using a specific GH receptor antagonist, we found that normalization of IGF-I is associated with rapid reductions in markers of both bone formation and resorption, and that these processes remain coupled. These data confirm the highly significant effects of GH and IGF-I in modulating bone turnover. The independent contributions of GH and IGF-I to these effects and the long-term effects on BMD in this population remain to be determined.

Effects of growth hormone administration on inflammatory and other cardiovascular risk markers in men with growth hormone deficiency. A randomized, controlled clinical trial.

BACKGROUND: Growth hormone-deficient adults have increased cardiovascular mortality. Growth hormone replacement may affect cardiovascular risk. Inflammation plays an important role in atherosclerosis, and inflammatory markers are predictive of cardiovascular events. OBJECTIVE: To investigate the effect of growth hormone replacement on inflammatory and other cardiovascular risk factors. DESIGN: Randomized, single-blind, placebo-controlled trial. PATIENTS: 40 men with adult-onset growth hormone deficiency. INTERVENTION: Growth hormone or placebo given for 18 months at a dose adjusted for normal serum insulin-like growth factor I level. MEASUREMENTS: Anthropometric, hemoglobin A1c, and central fat values were assessed every 6 months. Levels of glucose, insulin, insulin-like growth factor I, and lipids were measured at 1, 3, 6, 12, and 18 months. C-reactive protein, serum amyloid polypeptide A, inteleukin-6, and lipoprotein(a) levels were determined at baseline and 6 and 18 months. RESULTS: C-reactive protein and inteleukin-6 levels decreased in growth hormone recipients compared with placebo recipients (differences between groups, -1.9 +/- 0.6 mg/L [P = 0.0027] and -1.3 +/- 0.5 ng/L [P = 0.013], respectively). Changes in serum amyloid polypeptide A levels between groups did not reach statistical significance (difference between groups, -2.4 +/- 1.2 mg/L; P = 0.056). Serum cholesterol levels, low-density lipoprotein cholesterol levels, and ratios of total cholesterol to high-density lipoprotein cholesterol decreased in growth hormone recipients in the first 3 months compared with placebo recipients (differences between groups, -0.86 +/- 0.17 mmol/L [-33.2 +/- 6.6 mg/dL] [P < 0.001], -0.63 +/- 0.20 mmol/L [-24.5 +/- 5.9 mg/dL] [P < 0.001], and -0.56 +/- 0.26 [P = 0.040], respectively), but the decrease was not maintained from month 6 to month 18. Lipoprotein(a) levels increased (difference between groups, 22.0 +/- 8.0 mg/L; P = 0.0096). Short-term increases occurred in glucose levels, insulin levels, and insulin-to-glucose ratios (differences between groups, 0.54 +/- 0.16 mmol/L [9.6 +/- 2.8 mg/dL] [P = 0.0018], 37.9 +/- 9.6 pmol/L [P < 0.001], and 6.0 +/- 1.8 [P = 0.0025], respectively), but only the increase in glucose level was maintained over the long term (difference between groups, 0.56 +/- 0.17 mmol/L [10.0 +/- 3.1 mg/dL]; P = 0.0026). Hemoglobin A1c values did not change. Truncal fat-to-total fat ratios decreased (difference between groups, -0.018 +/- 0.007; P = 0.0087). CONCLUSIONS: Long-term growth hormone replacement in men reduces levels of inflammatory cardiovascular risk markers, decreases central fat, and increases lipoprotein(a) and glucose levels without affecting lipid levels.