Leukemia in children treated with growth hormone.

An association between treatment with growth hormone (GH) and the development of leukemia was described in 1988. This perceived association is best explained by the fact that there are more children with GH deficiency (GHD) with risk factors predisposing them to leukemia than in the general population. These factors include previous cancers and their treatment, as well as co-existing conditions such as Down, Bloom and Fanconi syndromes. Examination of large databases of GH-treated individuals shows that GH-treated patients without these risk factors do not have an increased incidence of leukemia.

Somatomedin binding proteins in GH-deficient children with normal plasma somatomedin levels.

The major portion of circulating somatomedin (SM) is specifically carried by a large molecular weight binding protein of 150,000 (150K BP). There also exists a species of unsaturated binding protein in plasma of 40,000 mol wt that binds radiolabeled SM more avidly than the 150K complex. Patients with GH deficiency appear to have higher levels of this smaller BP while exhibiting decreased tracer binding and endogenous SM in the region of the 150K BP when compared to normals. This suggests the GH dependence of the 150K BP. We report studies on four GH-deficient children who had normal growth and normal SM levels after intracranial surgery. All were hyperphagic and obese at the time of study and were on replacement medications with the exception of GH. Endocrine evaluation revealed mild hyperprolactinemia in three patients and high insulin levels in three. SM-C/insulin-like growth factor I (IGF-I) levels by RIA and IGF-II by radioreceptor assay were normal for age. Sephacryl-200 column chromatography on plasma samples from these patients revealed patterns of tracer binding and endogenous SM migration similar to those of normal and GH-treated hypopituitary patients. The chromatograms were different from those usually seen in untreated GH-deficient patients. SM binding in whole serum was also similar to that observed in normal children (38 +/- 7% [125I]IGF-II bound for patients compared to 39 +/- 3% for controls) and less than that observed in GH-deficient children (48 +/- 5%). These data suggest that the characteristics of SM binding in our patients are similar to those of normal subjects. The regulation of the 150K BP is not accomplished by GH in these patients, although the exact mechanism for the maintenance of serum SM is unclear.

Use of a two-site immunoradiometric assay for growth hormone (GH) in identifying children with GH-dependent growth failure.

Children with GH-dependent growth failure have normal GH responses to provocative stimuli, but, nonetheless, show increased growth rates when treated with replacement doses of human GH. We studied the GH secreted by 8 such children using a 2-site immunoradiometric assay for GH (IRMA-GH). To be detected in this assay, a GH molecule must be able to react simultaneously with two different monoclonal antibodies, each specific for a different region of the GH molecule. We compared the GH values observed in the IRMA to those obtained using classical RIA for GH (RIA-GH). The IRMA-GH to RIA-GH ratio was not affected by time of sample collection relative to peak GH serum levels in either normal individuals or children with GH-dependent growth failure. The ratios in serum samples from children with GH-dependent growth failure were significantly lower than those observed in normal individuals [0.35 +/- 0.005 vs. 0.48 +/- 0.02 (+/- SE); P less than 0.001]. All of the children with GH-dependent growth failure had IRMA-GH to RIA-GH ratios less than 0.4, while only 7 of 24 normal individuals had ratios this low (P less than 0.005). We suggest that a comparison of IRMA-GH with RIA-GH may supplement other clinical measurements and aid in identifying children with GH-dependent growth failure.

The role of adrenocorticotropin testing in evaluating girls with premature adrenarche and hirsutism/oligomenorrhea.

To identify biochemical predictors for future development of hirsutism and/or oligomenorrhea (H/O) in girls with premature adrenarche (PA), we performed dexamethasone-suppressed ACTH stimulation tests in girls with PA (n = 46), young women (n = 44) with H/O, and adult women (n = 31). Cortisol, androstenedione, dehydroepiandrosterone, and 17-hydroxyprogesterone were measured. Seven girls with PA (15%) and seven with H/O (16%) had evidence of nonclassical adrenal steroid biosynthetic defects [nonclassical congenital adrenal hyperplasia (NCAH)]. Twenty-five girls with PA (54%) and 28 girls with H/O (64%) had the moderately elevated 17-hydroxyprogesterone response to ACTH that has been reported in obligate heterozygotes for 21-hydroxylase deficiency. There were no clinical features that distinguished the girls with NCAH from the others. ACTH testing is an important tool in distinguishing those girls with PA and H/O who have NCAH. Although we could find no differences in other adrenal steroid hormones that might predict which of the other girls with PA might late develop H/O, black girls comprised a substantially smaller fraction of the population with H/O than of the population with PA (2% vs. 26%; chi 2 = 8.5; P less than 0.005). This observation suggests that PA, in blacks who do not have NCAH, is more likely to be a benign condition/than in other ethnic groups.

Kinetics of the somatomedin C/insulin-like growth factor I: response to exogenous growth hormone (GH) in GH-deficient children.

The somatomedin C/insulin-like growth factor I (SMC/IGF-I) response to human GH (hGH) therapy and the t1/2 of SMC/IGF-I after the cessation of hGH were determined in 15 children with GH deficiency. After 5 injections of hGH (0.1 U/kg), there was a significant increase in total SMC/IGF-I [from 0.27 +/- 0.06 to 1.19 +/- 0.17 U/ml (mean +/- SEM)]. Both the pretreatment SMC/IGF-I and the maximal SMC/IGF-I levels attained were correlated with chronological age and bone age. Body size, as indicated by height and weight, also correlated with pretreatment and maximal SMC/IGF-I levels. For both pretreatment and maximal SMC/IGF-I levels, there was a better correlation of SMC/IGF-I levels with bone age than with chronological age. While the correlation between height and the pretreatment SMC/IGF-I level was stronger, weight was a better predicter of the maximal SMC/IGF-I level. Maximal SMC/IGF-I levels were reached 18.8 +/- 2.9 h after the last hGH injection. The t1/2 for SMC-IGF-I after the attainment of maximal levels was 20.7 +/- 2.3 h, or 39.5 +/- 3.8 h from the last injection of hGH. The t1/2 of SMC/IGF-I determined in this way was longer than previous values reported from studies in the rat. The relatively long t1/2 of SMC/IGF-I which we observed may in part explain the success of present GH treatment regimens which involve every other day injections of hGH.

Insulin-like growth factor-binding protein-2 and insulin: studies in children with type 1 diabetes mellitus and maturity-onset diabetes of the young.

The regulation of circulating insulin-like growth factor-binding protein-2 (IGFBP-2) in humans is not well understood. In vitro and animal data have identified the role of insulin in the regulation of IGFBP-2, but such a relationship has not been established clearly in humans. In the present study, serum IGFBP-2 concentrations were assessed by Western ligand blot and immunoblot analysis in children with newly diagnosed and untreated insulin-dependent diabetes mellitus (IDDM) and maturity-onset diabetes of the young before and at various times after insulin therapy. For comparison, IGFBP-2 levels were also determined in lean and obese age-matched controls. Children with IDDM were grouped according to their serum bicarbonate levels at the time of presentation (group A, > 20; group B, 13-20; group C, < 13 milliequivalents/L). Densitometric analysis demonstrated that before insulin therapy, group A patients had serum IGFBP-2 levels comparable to those in lean controls, and no significant change in IGFBP-2 was observed during insulin therapy. However, group B patients had a 2-fold elevation in IGFBP-2 levels before insulin therapy compared to lean controls. In these patients, IGFBP-2 tended to decrease at 1 week, but was not significantly reduced until 1 month after the initiation of insulin therapy. Group C patients had a 2.5-fold elevation of IGFBP-2 before treatment, which normalized by 1 month after treatment. Children with maturity-onset diabetes of the young, who had insulin levels and body mass indexes greater than IDDM patients and lean controls, had significantly lower IGFBP-2 levels than both lean and obese controls. IGFBP-2 levels tended to decrease further during insulin therapy. These results indicate that long standing alterations in serum insulin concentrations beyond the physiological range have significant influence on serum IGFBP-2 levels in children and confirm earlier findings that serum IGFBP-2 levels are not acutely regulated by insulin.

Early initiation of growth hormone treatment allows age-appropriate estrogen use in Turner's syndrome.

Because estrogen (E) accelerates skeletal maturation it can decrease final height attainable with GH therapy in girls with Turner's syndrome (TS). Nonetheless, as age-appropriate E administration does have psychobehavioral benefits for such patients, we asked whether E treatment in TS could occur without adverse impact on final adult height if GH therapy were started at an earlier age. Near adult height (NAH) was assessed in 344 girls with TS, who had received both GH and E and were followed in the National Cooperative Growth Study database. The groups were divided into quartiles based on age at initiation of GH (2-10, 10-12, 12-14, and 14-18 yr). The longest total and E-free period of GH treatment occurred in the girls who had started treatment in the youngest quartile (mean age, 8.2 +/- 1.5 (SD) yr); they were also exposed to E at the youngest age (12.7 +/- 1.6 yr). Although the girls in the youngest group received E at an earlier age, they had a significantly greater increase (1.8 +/- 0.8) in Lyon height SD score at NAH over Lyon predicted adult height than those in the oldest GH-treated group (0.8 +/- 0.6), which first received E at 15.9 +/- 1.3 yr. Multiple linear regression equations for gain in Lyon height SD score and in height (cm) showed greater increments with a longer period of E-free GH therapy. All four GH age groups had the same NAH, but the youngest quartile was youngest at NAH and likely still having more growth potential. Comparable data were found in 127 TS girls with spontaneous puberty. In conclusion, girls with TS starting GH at an early age have a greater gain in Lyon SD score at NAH compared with those starting later, even though they received E at a younger age. If GH therapy were started early, E treatment could be initiated at a younger, more age-appropriate time without compromising adult height.

17-Hydroxyprogesterone responses to adrenocorticotropin in children with premature adrenarche.

The adrenal secretory response to an iv bolus dose of ACTH was measured in 10 girls (4-8 yr of age), 5 boys (4-9 yr) with premature adrenarche (PA), and 20 normal children. The evening before the ACTH test, each subject took dexamethasone (1 mg at bedtime) to suppress the early morning surge of ACTH. The next morning, 2 serum samples were obtained before the administration of ACTH (Cortrosyn; 0.25 mg), and 2 samples were collected 30 and 45 min after ACTH administration. All samples were assayed for cortisol, dehydroepiandrosterone (DHEA), DHEA sulfate, 17-hydroxyprogesterone (17-OHP), and androstenedione. There was no significant difference in the dexamethasone-suppressed steroid levels between the children with PA and the normal children. After ACTH injection, cortisol, DHEA, 17-OHP, and androstenedione increased significantly. There was no significant change in DHEA sulfate. The mean 17-OHP response to ACTH in girls with PA was significantly higher than that in girls and women whose pubertal development was normal. This response was similar in magnitude to that in a group (n = 5) of obligate heterozygotes for congenital adrenal hyperplasia (CAH). These data suggest that many girls with PA have a mild adrenal steroid secretory defect that resembles the response in adult obligate heterozygotes for CAH due to 21-hydroxylase deficiency. In contrast to the girls, none of the boys with PA had an exaggerated 17-OHP response to ACTH. Thus, these boys had no evidence for an adrenal steroid secretory defect. In summary, although the clinical presentation of boys with PA is similar to that of girls with PA, there is a significant difference in the adrenal steroid secretory response to ACTH. Thus, the biochemical events that cause PA in boys may be different from the corresponding events in girls.

Factors predicting the response to growth hormone (GH) therapy in prepubertal children with GH deficiency.

To identify factors influencing the response to GH therapy, we used a multiple regression model to analyze data from 632 naive prepubertal children with GH deficiency (GHD). There were 523 children with idiopathic and 109 children with organic GHD. They were treated with the same preparation of biosynthetic methionyl GH (somatrem, Protropin) for at least 1 yr. In children with idiopathic GHD, six variables predicted 40% of the response to treatment. They were (listed in relative importance, all P < 0.0001): age, log maximum GH, weight adjusted for height, dosing schedule, dose, and midparental height. Three variables, pretreatment growth rate, log maximum GH, and age, predicted 20% of the GH response in children with organic GHD. When data for all children were analyzed using analysis of covariance, children with idiopathic GHD grew better than those with organic GHD (mean +/- SD, 9.2 +/- 2.4 vs. 8.8 +/- 2.6 cm/yr; P < 0.0001). The children (both organic and idiopathic GHD) who did not respond well to treatment were younger and thinner than those who did. Early diagnosis and initiation of therapy should be beneficial to ultimate height attainment. The best response to GH therapy should be in young children with severe idiopathic GHD who receive daily weight-adjusted doses. The use of GH daily in higher doses would be expected to be most beneficial in older children with acquired and/or less severe GHD or in children who are underweight for height.

Intestinal lymphangiectasia in a patient with autoimmune polyglandular disease type I and steatorrhea.

Steatorrhea is seen in 18-24% of patients with autoimmune polyglandular disease (APD) type 1. The etiology and pathophysiology of the steatorrhea in this disease are unknown. We present a patient with APD type 1 and steatorrhea in whom biopsies revealed intestinal lymphangiectasia. This association has not been previously described. Intestinal lymphangiectasia may explain the steatorrhea in some patients with ADP type 1. As blind intestinal biopsies may miss areas of intestinal lymphangiectasia, endoscopically directed intestinal biopsies should be included in the evaluation of steatorrhea in APD type 1.

Circulating growth hormone isoforms in girls with Turner's syndrome.

To evaluate the presence of different GH isoforms in serum of girls with Turner's syndrome, we measured the serum GH content using RIAs with three different site-specific monoclonal antibodies (MAbs). We compared the results to those obtained with authentic GH and GH isoforms. Compared to pituitary GH (mol wt, 22K daltons) as the standard for all three MAbs, serum from girls with Turner's syndrome did not displace tracer [125I]GH equally with all three MAbs. The relative amounts of GH-immunoreactive material found in Turner's syndrome were different from the amounts observed in normal adults and most children with idiopathic short stature. The presence of GH, other than 22K GH, in serum from girls with Turner's syndrome was confirmed by affinity chromatography. The existence of different isoforms of GH, as shown by different cross-reactivity patterns with different MAbs to GH, may explain the conflicting results reported for GH secretion in girls with Turner's syndrome.

Adrenal steroid secretion in girls with pseudoprecocious puberty due to autonomous ovarian cysts.

To evaluate the role of adrenal steroids in pseudoprecocious puberty due to large ovarian follicular cysts, we studied the serum 17-hydroxyprogesterone response to a combination of dexamethasone suppression followed by iv ACTH administration in two girls and compared the results to those in girls with premature thelarche, normal prepubertal girls, and a girl with true precocious puberty. Although basal serum 17-hydroxyprogesterone levels were normal in all subjects, there was incomplete suppression of 17-hydroxyprogesterone with dexamethasone in the two girls with pseudoprecocious puberty and large ovarian cysts. The 17-hydroxyprogesterone response to ACTH was much greater in these girls (360 and 540 ng/dl) than in the girls with other types of precocious puberty (mean +/- SD, 71 +/- 15 ng/dl) or in normal prepubertal girls (80 +/- 20 ng/dl). The girls with large ovarian cysts had decreased gonadotropin responses to GnRH, which were reversed subsequent to removal of the cyst. Removal of the ovarian cysts also restored the dexamethasone suppressibility of serum 17-hydroxyprogesterone and abolished the progression of pubertal development. However, 17-hydroxyprogesterone responses to ACTH were still elevated (160 and 350 ng/dl). Preoperatively, both girls had increased levels of dehydroepiandrosterone sulfate, 16-hydroxydehydroepiandrosterone sulfate, and another unidentified steroid sulfate. These steroid sulfates were also found in the cyst fluid from the one patient from whom the fluid was obtained. These results suggest that steroid production by the adrenal gland may stimulate the development of small ovarian cysts (which may be present in normal prepubertal girls) into large ovarian cysts capable of causing gonadotropin-independent precocious puberty.

Insulin-like growth factor binding protein-3 proteolysis in children with insulin-dependent diabetes mellitus: a possible role for insulin in the regulation of IGFBP-3 protease activity.

Limited proteolysis of serum insulin-like growth factor (IGF) binding protein (IGFBP)-3 has been described in various conditions and may increase the bioavailability of IGFs. The physiological regulators of serum IGFBP-3 protease activity are unknown. To characterize the relationship between insulin and IGFBP-3 protease activity, we have examined serum IGFBP-3 proteolysis in children with untreated insulin-dependent diabetes mellitus (IDDM) and have followed the effect of insulin therapy on serum IGFBP-3 proteolysis at 1 day, 1 week, and 1 month after the initiation of insulin therapy. Ligand blot analysis of sera from untreated children with IDDM showed that intact IGFBP-3 was 50 +/- 9% of the age-matched control pool. After the initiation of insulin treatment, IGFBP-3 did not change significantly at 1 day after treatment but increased dramatically at 1 week (90 +/- 13%) and 1 month after treatment (102 +/- 13%). In contrast, when measured by immunoradiometric assay (which detects both intact and fragments of IGFBP-3), IGFBP-3 levels were 70% of the control pool before insulin therapy and did not increase significantly until 1 month after treatment. Immunoblot analysis demonstrated that intact IGFBP-3 doublet was diminished to 41 +/- 7% of controls, whereas the major IGFBP-3 fragment (30 kDa) was increased in IDDM sera before insulin therapy. After insulin, intact IGFBP-3 increased and the 30-kDa fragment decreased to values comparable to those observed in controls. In vivo IGFBP-3 proteolysis, which implies preassay exposure of serum IGFBP-3 to proteases, was estimated by immunoblot analysis. IGFBP-3 proteolysis was increased before insulin therapy (160 +/- 9%) and decreased to 81 +/- 9% at 1 week and to 71 +/- 11% at 1 month after insulin treatment. Residual serum IGFBP-3 protease activity was estimated by a 125I-IGFBP-3 degradation assay. Serum IGFBP-3 protease activity increased significantly in untreated diabetics, compared with activity in controls (128 +/- 5% vs. 99 +/- 11%). During insulin therapy, serum IGFBP-3 protease activity decreased gradually to 91 +/- 5% of control values at 1 month. Molecular sizes of the IGFBP-3 proteolytic fragments (30 kDa, 24 kDa, and 19 kDa) and inhibition profile of IGFBP-3 protease were similar in IDDM and pregnancy sera, indicating that similar proteases (cation-dependent serine proteases) were active in both conditions. These results suggest an important role of insulin in the regulation of IGFBP-3 protease activity. Increased IGFBP-3 proteolysis in the sera of children with IDDM may serve to counteract the catabolic state induced by insulin deficiency.

Safety of recombinant deoxyribonucleic acid-derived growth hormone: The National Cooperative Growth Study experience.

The National Cooperative Growth Study has monitored the safety of recombinant human GH (rhGH) since 1985. Data have been collected from more than 19,000 children representing over 47,000 patient-years of rhGH treatment. Children receiving GH for renal disease were more likely to develop problems such as intracranial hypertension than those with GH deficiency (P < 0.01). Children with idiopathic short stature were less likely to develop slipped capital femoral epiphysis than those with GH deficiency or Turner's syndrome (P < 0.01). There was no evidence of an increased recurrence of leukemia or central nervous system tumors. There were 3 new cases of leukemia in children without known risk factors for developing leukemia and 5 cases in children with known risk factors. Growth deceleration associated with high affinity, high capacity antibodies to GH was found in only 2 of 5039 subjects tested (0.04%). Major adverse events in association with rhGH treatment have been rare, and preexisting medical conditions such as renal insufficiency may affect their frequency.

Adult height in growth hormone (GH)-deficient children treated with biosynthetic GH. The Genentech Growth Study Group.

Near-adult height (AH) was determined in 121 children (72 males and 49 females) with GH deficiency (GHD) who were prepubertal when they began treatment with recombinant DNA-derived preparations of human GH. AH as a SD score was -0.7 +/- 1.2 (mean +/- SD), significantly greater than the pretreatment height SD score (-3.1 +/- 1.2), the predicted AH SD score (-2.2 +/- 1.2; Bayley-Pinneau method), and the height SD score at the start of puberty (-1.9 +/- 1.3). In contrast to studies of GH treatment outcome, which used pituitary-derived GH (pit-GH) in lower doses, we found that males did not have a higher AH SD score than females, spontaneous puberty did not diminish AH, and AH was significantly greater than that predicted at the start of GH treatment. In a multiple regression equation, the statistically significant variables (all P < 0.0001) related to AH (r2 = 0.70) were the following: duration of treatment with GH, sex (males were taller than females, as expected for the normal population), age (younger children had a greater AH) and height at the start of GH, and growth rate during first year of GH. For the AH SD score (r2 = 0.47), pretreatment predicted AH, duration of GH, and bone age delay were significant (P < 0.0002) explanatory variables. Bone age delay (chronological age-bone age) had a negative impact on the AH SD score. Target height, etiology of GHD, previous treatment with pituitary GH, and the presence or absence of spontaneous puberty did not significantly improve the prediction of AH. Early diagnosis of GHD and continuous treatment with larger doses of GH to near AH should improve the outcome in children with short stature due to GHD.

Effect of in vitro action of serum proteases or exposure to acid on measurable immunoreactive somatomedin-C in serum.

The proportion of immunoreactive somatomedin-C (IR-Sm-C) in blood that is available for measurement in the RIA is influenced by whether the sample is processed as serum or plasma, by how promptly the sample is chilled and frozen, by whether the reaction is carried out in glass or polystyrene tubes and by whether the incubation mixture contains protamine or heparin. Although protamine buffers and polystyrene tubes increase the availability of purified somatomedin-C (Sm-C), they decrease the detectability of Sm-C in serum. By incubating serum at neutral or acid pH, this IR-Sm-C can be made available for measurement, suggesting that incubation alters the nature of the linkage between Sm-C and its binding proteins or causes a conformational change in the binding protein, resulting in greater exposure of IR-Sm-C. The increment in measurable IR-Sm-C that occurs at neutral pH appears to be due to the action of proteolytic enzymes since it is time, temperature, and pH dependent and is inhibited by a variety of protease inhibitors and chelating agents. The increment which occurs at acid pH is not inhibited by chelating agents or elevated temperature and must be due in part to acid hydrolysis of Sm-C and its binding proteins. However, since the acid-induced increment is optimal within a narrow pH range and falls off sharply below pH 3.5, acid proteases may also be involved. These observations, which provide insight into the nature of the serum proteins with which Sm-C is associated, bear on the interpretation of results of serum somatomedin measurements carried out with different methods. They also may aid in delineating the mechanisms by which the somatomedin contained in the macromolecular complex in serum is made available to tissues.

Plasma somatomedins in children with hyperinsulinism.

Hyperinsulinism has been postulated as a possible stimulus to somatomedin (SM) synthesis by the perfused rat liver and in man. Seven children who presented with severe fasting hypoglycemia and who had documented hyperinsulinism were studied to explore the effect of excessive endogenous insulin secretion on plasma levels of the SMs. Neither the concentration of somatomedin C/insulin-like growth factor-I (SMC/IGF-I) (mean +/- SD, 0.53 +/- 0.37 U/ml) nor that of IGF-II (1.10 +/- 0.58 microgram/ml) was significantly different from values observed in normal children. SMC/IGF-I and IGF-II were inversely correlated. There was a positive correlation between SMC/IGF-I and body weight, and an inverse correlation among IGF-II, body weight, and length. These results suggest that in children with normal GH secretion, hyperinsulinism sufficient to cause severe recurrent hypoglycemia does not result in an increase in either SMC/IGF-I or IGF-II concentrations in plasma.

Plasma somatomedins, endogenous insulin secretion, and growth in transient neonatal diabetes mellitus.

Infants with transient neonatal diabetes mellitus are small for gestational age and fail to thrive postnatally unless insulin is administered. We have measured the concentrations of insulin-related growth factors in an infant girl with this condition to learn if deficiencies in one or more of these factors could be responsible for the impaired growth. Cord blood serum radioimmunoassayable insulin and somatomedin C/insulin-like growth factor I (SMC/IGF-I) were low, but insulin-like growth factor II (IGF-II) measured by a specific radioreceptor assay was normal. Insulin therapy begun on the fourth day of life resulted in a prompt increase in weight and a delayed rise in SMC/IFG-I. No significant changes in IGF-II were observed. After 2.5 months, insulin treatment was discontinued. At that time, endogenous insulin secretion was documented by increased urinary C-peptide. Normal growth and SMC/IFG-I levels persisted. We conclude that growth failure in this condition may be related not only to a lack of insulin but also to SMC/IGF-I deficiency. A deficiency in IGF-II is not involved.

Dehydroepiandrosterone sulfate levels are not suppressible by glucocorticoids before adrenarche.

We measured dehydroepiandrosterone sulfate (DHEA-S) levels in children before and after high dose prednisone therapy. In older children (postadrenarchal), there was a 70% decrease in DHEA-S levels after 1 week. However, even after a month of therapy, DHEA-S was detectable in serum. In contrast, in younger (preadrenarchal) children, the low initial DHEA-S levels were not decreased by prednisone therapy. These findings suggest that there are two distinct regulatory pathways leading to DHEA-S, one of which is independent of ACTH.

Effect of insulin on the insulin-like growth factor system in children with new-onset insulin-dependent diabetes mellitus.

To further characterize the mechanism of impaired growth in children with insulin-dependent diabetes mellitus, we examined the serum components of the insulin-like growth factor (IGF) system in 11 children with new-onset insulin-dependent diabetes mellitus and followed the effect of insulinization on the IGF system longitudinally 1 day, 1 week, and 1 month after starting insulin treatment. Before insulin therapy, serum IGF-I, IGF-II, IGF-binding protein-3 (IGFBP-3), and GH-binding protein (GHBP) levels were significantly decreased, whereas IGFBP-1 and cortisol were significantly increased in diabetic children compared to those in an age-, sex-, and stage of puberty-matched control group. Random serum GH concentrations did not differ significantly. The alterations in the IGF system reversed with insulin therapy in a sequential manner. IGFBP-1 fell rapidly and was comparable to control values within 24 h after insulin treatment. IGF-I rose 1 week after treatment, reaching levels comparable to those in controls and continued to rise through 1 month of treatment. IGF-II, IGFBP-3, and GHBP showed a slower pattern of change, with their levels reaching control values only 1 month after the start of insulin treatment. Improvement in glycemic control, as determined by a change in hemoglobin-A1c, correlated positively with improvement in IGF-I, IGF-II, IGFBP-3, GHBP, and weight gain after 1 month of insulin therapy. These data are consistent with the hypothesis that changes in the IGF system in the insulinopenic state are similar to those during nutritional deprivation and may serve to minimize IGF's anabolic actions. The decreases in IGF-I, IGF-II, and IGFBP-3 may in part be due to a decrease in the GHBP/receptor. However, the observation that an increase in serum IGF-I was observed earlier than an increase in GHBP and without a significant change in serum GH suggests a direct stimulatory effect of insulin on liver IGF-I production or reversal by insulin of some postreceptor defect in GH action independent of GHBP.