Predictors of glycemic control in children with type 1 diabetes: the importance of race.

Diabetes is a common cause of kidney failure and blindness among young adults, particularly of African-American descent. Since glycemic control is a predictor of diabetes complications, we evaluated the impact of multiple factors including a special multidisciplinary management program on glycosylated hemoglobin in children with Type 1 diabetes. Data was collected from pediatric diabetes clinics in New Orleans, LA and Baltimore, MD. In New Orleans, hemoglobin A(1c) was higher in African-American patients 12. 5+/-3.3% (n=71) vs. 10.7+/-2.1% (n=80) in Caucasian children, p<0. 0001. Longer duration of diabetes was also associated with higher hemoglobin A(1c) in both races. The effect of race on hemoglobin A(1c) was independent of the influence of sex, insurance status, body mass index (BMI) z-score, and number of clinic visits. Covariate analysis with mean blood glucose levels indicated that higher hemoglobin A(1c) was attributable to higher mean blood glucose levels in African-American children. From the Baltimore data, a multidisciplinary intervention program led to improved total glycosylated hemoglobin for Caucasian patients but not for African-American children. Poorer glycemic control of African-American children is likely to predispose them to a higher likelihood of developing microvascular complications as they mature. Standard hospital-based multidisciplinary programming for diabetes management may have limited effectiveness in improving glycemic control of African-American children with diabetes. Innovative intervention programs are needed for these high-risk patients.

Can we prevent type 1 diabetes?

Preliminary laboratory and clinical studies suggest that the clinical onset of type 1 diabetes can be delayed or prevented in high-risk individuals. The NIH sponsored Diabetes Prevention Trial Type 1 is currently underway to determine whether prevention is possible using insulin to alter the autoimmune process.

Testosterone effect on growth and growth mediators of the GH-IGF-I axis in the liver and epiphyseal growth plate of juvenile rats.

Several studies have suggested that testosterone may have a direct, GH-independent effect on growth. In order to assess possible mechanism(s) whereby testosterone exerts its growth-promoting effect, we evaluated its effect on growth mediators of the GH-IGF-I axis, in both the liver and the epiphyseal growth plate (EGP). Testosterone was administered to peripubertal rats and the responses of mRNA of GH receptor, IGF-I, IGF-I receptor and IGF-binding proteins-1 and -3 (IGFBP-1 and IGFBP-3) as well as circulating IGF-I were evaluated in two time-related models: over 12 h after a single injection (short-term study) and 10 days after continuous administration (long-term study). Rats in the short-term study were castrated and were killed 1, 4, 6 and 12 h post injection. Rats in the long-term study were divided into two groups: castrated vs castrated and hypophysectomized, in order to assess the effect of testosterone in the presence and absence of GH. mRNA levels were determined by RNase protection assay, and serum IGF-I by RIA. Testosterone enhanced weight gain in the rats treated for 10 days, a change that was similar in the presence or absence of GH. This effect was relatively small, however, by comparison with the total weight gained without testosterone. Testosterone had no effect on hepatic IGF-I mRNA abundance but induced a reduction in circulating IGF-I levels, in both the short- and long-term study. Testosterone had no effect on hepatic GH receptor and IGFBP-3 mRNA levels but resulted in a transient, short-term elevation in IGFBP-1 mRNA levels that was maximal 4 h post injection. In the EGP, neither testosterone administration nor hypophysectomy had any effect on IGF-I and IGF-I receptor mRNA levels. However, testosterone increased GH receptor mRNA abundance after 10 days of continuous administration in hypophysectomized rats only. These data suggest that the effect of testosterone on growth (as assessed by weight gain) is small and is not mediated by changes in hepatic gene expression of IGF-I, IGF-I receptor, IGFBP-1, IGFBP-3 or circulating IGF-I. At the EGP, the testosterone effect on linear growth is not mediated through changes in mRNA abundance of IGF-I and IGF-I receptor. The small but significant elevation of GH receptor mRNA levels in hypophysectomized rats may suggest a testosterone-mediated augmentation of a GH effect at the target organ.

Ketoconazole treatment of gonadotropin independent precocious puberty in girls with McCune-Albright syndrome: a preliminary report.

McCune-Albright syndrome (MAS) in girls is characterized by gonadotropin independent precocious puberty (GIPP). This form of GIPP is resistant to therapy with GnRH analogues. As an alternative treatment, we successfully used ketoconazole 200 mg t.i.d. orally in two girls with MAS, GIPP and advanced bone age Ketoconazole led to rapid control of GIPP with cessation of menses and regression of pubertal signs in both patients. Ketoconazole was temporarily interrupted in one patient due to pruritus but later restarted without problem. After 1 year of therapy both patients have remained free of menses, progression of puberty and other side effects. Repeat sonography on ketoconazole revealed continued presence of ovarian cysts. Our preliminary experience indicates the safety and effectiveness of ketoconazole as a therapy for GIPP with potential advantages over previously used modes of treatment. Longer use of ketoconazole to suppress GIPP is required to determine whether this therapy can prolong linear growth with enhancement of final height.

Effect of age on the response to parathyroid hormone.

Serum phosphate (PO4) levels and the tubular threshold for PO4 corrected for glomerular filtration (TP/GF) are age-dependent, being higher in children than in adults. We evaluated the effect of age on the response to infusion of parathyroid hormone(1-34) (PTH) in healthy children (n = 8) and adults (n = 12). In addition, six patients with pseudohypoparathyroidism (PHP) and two with PTH deficiency (hypoparathyroidism [HP]) were also studied. At baseline, TP/GF in normal subjects was inversely correlated with urinary cyclic adenosine monophosphate corrected for glomerular filtration (UcAMP/GF) (P < .0359). After PTH administration in the controls, UcAMP/GF was inversely correlated with TP/GF (P < .0007) and directly correlated with maximal fractional extraction of PO4 (FEP) (P < .0002). The slope of the regression of TP/GF (P < .0076) and FEP (P < .0034) with UcAMP/GF was steeper in children than in adults. Two HP patients had high PTH-stimulated UcAMP/GF levels, but stimulated TP/GF and FEP were not changed commensurate with levels that would expected from the normative data. In six patients with PHP, PTH-stimulated TP/GF was also correlated with peak UcAMP/GF (r = .96, P < .002). PHP patients could be distinguished from normal controls based on the combination of low peak FEP or high TP/GF together with low peak UcAMP/GF. Thus, in normal subjects, the phosphaturic response to PTH is correlated with the increase in urinary cAMP and is age-dependent, with a greater decrease of TP/GF in children than in adults.

The integrated concentration of cortisone is reduced in obese children.

We evaluated the possibility that there is enhanced conversion of cortisol (F) to cortisone (E) in obese children. IC-E was measured from 15 lean children aged 12.7 +/- 2.2 years, body mass index Z-score (BMI-SD) = -0.35 +/- 0.82, IC-F = 197 +/- 70 nM/l and 9 obese children aged 12.3 +/- 3.2 years, BMI-SD = + 4.7 +/- 2.1, IC-F = 149 +/- 53 nM/l. IC-E was higher in lean children 76 +/- 25 nM/l compared to obese 60 +/- 11 nM/l (p < 0.04). There was no difference in the ratio of IC-E/IC-F between lean 0.40 +/- 0.10 and obese subjects 0.42 +/- 0.09 (p < 0.06). IC-E was directly correlated with IC-F: IC-E = 0.25 x IC-F + 26 (n = 24, r2 = 0.57, p < 0.0001). In a multiple regression model (overall r2 = 0.32, p < 0.02), IC-E was related to BMI-SD inversely (p < 0.0054) and influenced as well by interaction of BMI-SD with sex (p < 0.043), IC-E being lower in boys with increasing body mass. In childhood, obesity is associated with decreased plasma IC-E and IC-F levels, the ratio of IC-E/IC-F is independent of body mass. Reduced IC-E levels in obese children are most likely due to the impact of body mass on IC-F.

Growth hormone deficiency in autoimmune polyglandular syndrome.

We describe a boy with autoimmune adrenal failure and compensated hypothyroidism, associated with isolated growth hormone deficiency (GHD). We suggest an autoimmune mechanism as the underlying etiology for the GHD in this case.

Hypophosphatemia induced in mice by transplantation of a tumor-derived cell line from a patient with oncogenic rickets.

A cultured cell line was derived from a hemangiopericytoma obtained at surgery from a 9 year-old boy with hypophosphatemic rickets. Hypophosphatemia and metabolic bone disease were cured in the patient after tumor removal. Cells from passage 5 were transplanted into 8 week-old athymic female mice. Additional animals received innocula of cells thawed from the stored original tumor tissue, as well as MRC-5 fibroblasts. Serum phosphate levels 3-7 months post-transplantation were lowest in mice which received the cultured cell line (n = 5) 6.1 +/- 0.6 mg/dl (p < 0.05 compared to the other groups), thawed tumor cells (n = 8) 7.2 +/- 0.7 mg/dl, MRC fibroblasts (n = 4) 8.1 +/- 1.0 mg/dl, no transplant (n = 10) 8.7 +/- 1.9 mg/dl. Repeat of the experiment with cultured tumor cells from passage 12 no longer altered phosphate levels. A substance produced and release by the tumor in situ and by tumor-derived cultured cells is capable of producing hypophosphatemia. Experimental manipulation of functional tumor-derived cell lines may help elucidate the factor(s) causing hypophosphatemia in oncogenic osteomalacia/rickets.

Effect of six months of growth hormone therapy, followed by treatment withdrawal in short children with normal quantitative indexes of growth hormone secretion.

Eight short patients with normal 24-hour integrated concentration of growth hormone by continuous withdrawal (IC-GH) received 6 months of GH therapy, followed by 6 months off. GH therapy increased growth rate (+2.9 cm/yr), repeated IC-GH (sixfold), and IGF-1 concentration (twofold). Posttreatment growth reverted to the pretreatment rate. Thus increased growth rate and IGF-1 concentration is associated with supraphysiologic IC-GH after injection.

The relationship of growth rate, plasma growth hormone (GH) concentration, and GH-binding protein.

Growth hormone (GH)-binding protein (GHBP) and GH secretion are potential mediators of linear growth in children. To study the relationship between these variables, we measured GHBP activity, peak stimulated GH (PKGH), and 24-hour integrated GH concentration (ICGH) in 76 children referred for evaluation of growth. Linear growth was expressed as an age- and sex-specific growth rate standard deviation score (GRSD), which was calculated from sequential height measurements in the 6-month period immediately before GH testing. Using multiple regression models, we found that the relationship between GHBP and growth (GRSD) depended on height (height standard deviation [HGTSD] expressed as an age- and sex-specific z score) controlling for ICGH or PKGH. In further analysis of this relationship, we divided the subjects by HGTSD in subsequent analyses. In 19 children of normal stature (HGTSD > -2), GRSD increased with GH concentration (measured both as PKGH and ICGH: P <.013,R2 = .56) but decreased with higher levels of GHBP (P < .005,R2 = .62). In contrast, for 57 subjects with severe short stature (HGTSD < or = -2), GRSD could not be predicted from GHBP, GH secretion, HGTSD, or interaction involving these variables. These data suggest the hypothesis that under normal conditions, GHBP and GH level may be important predictors of growth rate in children.

Urinary cyclic adenosine 3',5'-monophosphate response in McCune-Albright syndrome: clinical evidence for altered renal adenylate cyclase activity.

The recent finding of an activating mutation in the Gs alpha protein, the protein that couples receptors to stimulation of adenylate cyclase, from endocrine and nonendocrine tissues of patients with McCune-Albright syndrome (MAS) suggests that alterations in adenylate cyclase activity may account for the clinical abnormalities in these patients. Many patients with MAS have hypophosphatemia. This may result from the presence of the activating Gs alpha mutation in proximal renal tubules or the elaboration of a phosphaturic factor from fibrous dysplasia. We, therefore, sought to characterize renal cAMP generation and phosphate handling in MAS patients. Intravenous infusion of PTH is a classic clinical test used to evaluate hormonal responsiveness of renal proximal tubule adenylate cyclase and examine PTH-dependent phosphate clearance. We performed PTH infusion in 6 MAS patients, 10 normal subjects, and 6 patients with pseudohypoparathyroidism (PHP). The basal urinary cAMP (UcAMP) level in the MAS group [5.5 +/- 2.6 nmol/dL glomerular filtration (GF)] was elevated (P < 0.05) compared to those in both normal subjects (3.2 +/- 1.2 nmol/dL GF) and patients with PHP (1.9 +/- 0.6 nmol/dL GF). However, PTH-stimulated peak UcAMP (15.0 +/- 7.0 nmol/dL GF) and the peak/basal UcAMP ratio (3.1 +/- 1.7) in MAS were significantly lower than the respective values in normal subjects (30.8 +/- 16.9 nmol/dL GF and 9.3 +/- 2.9; P < 0.05 for both) and were statistically similar to the blunted levels in PHP (respectively, 3.1 +/- 1.5 nmol/dL GF and 2.0 +/- 1.7). By contrast, the PTH-induced phosphaturic response in MAS patients was similar to that in the normal subjects. Our study provides clinical evidence that MAS patients have altered renal adenylate cyclase activity, manifested by an elevated basal UcAMP, but a blunted UcAMP response to PTH stimulation. These observations are presumably due to a mutation in the Gs alpha protein in the renal tubules. Despite the blunted UcAMP excretion, the phosphaturic response to PTH in MAS patients is intact.

Growth failure with plasma GH that is normal by RIA but low by radioreceptor assay: responsiveness to exogenous GH.

We measured the ratio of GH by radioreceptor assay (RRA) using IM-9 cells to its radioimmunoassay (RIA) level in: (a) 25 children and young adults with normal growth (controls); (b) 7 poorly growing children with GH neurosecretory dysfunction (GHND) who had normal stimulated GH but subnormal integrated concentration of GH (IC-GH) who responded to GH therapy; (c) 4 poorly growing children who had both normal stimulated GH and IC-GH but did not respond to GH therapy (GHNR - GH responder), and (d) 7 poorly growing patients with normal stimulated, and IC-GH who responded to GH therapy (GHR - GH responder). The RRA/RIA ratio of the GHND group was 1.3 +/- 0.4 similar to that of the control group - 1.3 +/- 0.3. The mean RRA/RIA ratio of the 7 GHR patients was 0.6 +/- 0.5, significantly lower than the controls and the GHND patients (p < 0.05). The RRA/RIA ratios of the 4 GHNR patients was not different from the controls. We conclude that: (1) GH binding is normal in GHNR and GHND patients and (2) RRA/RIA ratio below the normal range as measured by IM-9 cell GH RRA in short, poorly growing children with normal GH secretion, was a predictor of the response to GH therapy.

Plasma IGFBP-3 and its relationship with quantitative growth hormone secretion in short children.

OBJECTIVE: We assessed the relationship between serum IGFBP-3 levels with IGF-I and quantitative GH secretory status in poorly growing children. DESIGN: We studied the relationship between 24-hour integrated concentration of GH, peak GH to paired sequential stimulation tests, IGF-I and the IGFBP-3 serum levels. PATIENTS: One hundred and two children (82 males, 20 females, age 11.7 +/- 2.7 years) with short stature (height -2.6 +/- 0.7 SDS) were studied. MEASUREMENTS: Quantitative GH secretory status was assessed by the 24-hour integrated GH and by response to arginine and insulin stimulation. GH, IGFBP-3 and IGF-I were measured by radioimmunoassay. To adjust for age and gender, IGFBP-3 levels were converted to SD score. RESULTS: IGFBP-3 SDS was strongly correlated with IGF-I SDS (r = 0.64, P < 0.0001), and weakly with peak GH (r = 0.28, P < 0.0004), but not with the integrated GH concentration (r = 0.07, P < 0.46). IGFBP-3 SDS increased with pubertal maturation (P < 0.0001). There was no difference in mean IGFBP-3 SDS in subgroupings of the patients based on the results of their quantitative GH tests. CONCLUSION: In short children, IGFBP-3 levels increase with puberty, are strongly correlated with IGF-I levels, weakly correlated with peak response to GH stimulation tests, but not correlated with integrated GH. Consequently, diagnostic classifications of patients based on quantitative measurements of GH secretion and IGFBP-3 differ.

Effect of third-party reimbursement on use of services and indexes of management among indigent diabetic patients.

OBJECTIVE: To determine the effect of third-party reimbursement on the use of services and indexes of diabetes-related health management among inner-city diabetic patients. RESEARCH DESIGN AND METHODS: Adult diabetic patients (n = 158; 67% women, 33% men) from an inner-city diabetes clinic were categorized by level of third-party medical coverage: complete reimbursement for all services (full); partial reimbursement (part); and no reimbursement (none). Patients were followed for 13 mo. Use of billable medical services, diabetes clinic visits, emergency room visits, and hospital admissions were recorded. Use of a free, day-time diabetes telephone hot line was also documented. Indexes of diabetes-related health management, HbA1, blood pressure, and weight were compared from the beginning and the end of the study. Diabetes complications were scored and tabulated. RESULTS: Univariate analysis showed that patients with full reimbursement were more likely to use services than patients without reimbursement. When the combined effects of reimbursement status, age, sex, type of diabetes, and diabetes complications on use of services were analyzed together in a multivariate analysis, complications was the best predictor of admissions to the hospital and whether a patient called the hot line. IDDM patients and patients with full reimbursement were most likely to have an emergency room visit. Age was the best predictor of diabetes clinic attendance. No difference was noted in blood pressure or weight among the reimbursement groups at the beginning and end of study. However, the trend was toward (P < 0.05) an increase in HbA1 in the none group. CONCLUSIONS: Among inner-city diabetic patients, multiple factors influence use of medical services. Indigent diabetic patients without third-party reimbursement were observed to have a rise in HbA1. These factors should be taken into consideration when planning strategies to prevent diabetes complications and the most effective allocation of health-care resources.

Short stature: new challenges in growth hormone therapy.

Until recently, the limited supplies of pituitary derived growth hormone (GH) enabled us to treat only those patients who were classical GH deficient. With the unlimited supplies of recombinant GH available, there is no limitation to the number of patients we can treat. It becomes necessary, however, to select those patients who will most benefit from GH therapy. Our preliminary results demonstrate that the short-term growth response to growth hormone is not an all-or-none phenomenon. The lower the growth velocity and the growth hormone reserve, the better the growth response to therapy. On the other hand we do not recommend institution of GH therapy for children with a normal growth rate and a normal GH spontaneous secretion. In children with classical GH deficiency (GHD) and in children with a subnormal spontaneous secretion of GH (NSD) adult height prediction decreases when GH therapy is started at an age older than 12. We have found that GHD and NSD boys differ in their growth pattern. Pubertal maturation and bone age maturation progress more rapidly in NSD patients. Therefore special caution is needed in NSD patients older than 12 years. The older the patient and the longer the treatment period, the faster the pubertal process can advance. Further studies are needed before recommendations for therapy in non-classical GHD patients can be made. Until patients involved in clinical trials reach final height, recommendations for new indications cannot be made.

Cardiovascular reflex abnormalities in children and adolescents with diabetes mellitus.

OBJECTIVE: To assess the usefulness of specific cardiovascular reflex tests in childhood and to estimate the prevalence of cardiovascular reflex abnormalities among children with IDDM. In adults, abnormal cardiovascular reflexes are a frequent complication of diabetes, associated with increased morbidity and mortality. RESEARCH DESIGN AND METHODS: We measured heart-rate responses to deep breathing and standing in ambulatory children with and without IDDM between 6-19 yr of age. A subgroup of the IDDM patients was retested after 1 yr. RESULTS: We found the best techniques for detecting cardiovascular reflex abnormality in children were as follows: to record heart-rate responses to deep breathing either as the change in heart rate corrected for inspiratory heart rate or as the ratio of R-R intervals during expiration and inspiration; and to use the Maximum-minimum ratio for heart-rate responses to standing. HR-DBc was lower in diabetic than nondiabetic children (28.6 +/- 9.2% [n = 248] vs. 33.6 +/- 6.8% [n = 60]; P < 0.0005). Similarly, E:I was lower in children with IDDM than control subjects (1.42 +/- 0.19 [n = 248] vs. 1.52 +/- 0.15 [n = 60]; P < 0.0005). In the IDDM group, 21% of the children had abnormal HR-DBc or E:I responses. HR-STND M/m was lower in children with IDDM than control subjects (1.28 +/- 0.20 [n = 167] vs. 1.38 +/- 0.22 [n = 45]; P < 0.014). Among children with IDDM, 11.4% had abnormal HR-STND M/m responses. Overall, 29% of IDDM children tested abnormal in either HR-DBc or HR-STND M/m; 3% were abnormal in both tests. We found no correlation of HbA1c levels (n = 74) or duration of diabetes with either HR-DB, expiration to inspiration (n = 248), or HR-STND M/m (n = 167). In patients who were reevaluated after 1 yr we found a high correlation of the first and repeat HR-DBc tests (r = 0.47, n = 75, P < 0.0001), E:I (r = 0.53, n = 75, P < 0.0001), and HR-STND M/m (r = .49, n = 37, P < 0.002), but no evidence of an increased number of children with cardiovascular reflex abnormality. CONCLUSIONS: With easily performed HR-DB and HR-STND tests, we detected cardiovascular reflex abnormality in 29% of children with IDDM. We found no correlation of changes in HR-DB and HR-STND with HbA1c or duration of diabetes. These tests provide an objective clinical measurement to monitor autonomic neuropathy in children with diabetes.

The value of urinary growth hormone determination for assessment of growth hormone deficiency and compliance with growth hormone therapy.

OBJECTIVE: To compare the 24-hour integrated concentrations of plasma growth hormone with growth hormone levels in a simultaneously collected sample of urine. SETTING: Pediatric endocrine diagnostic unit. PATIENTS: Forty-six children (41 male and five female) aged 6 to 19 years underwent measurement of integrated concentrations of growth hormone and simultaneous urine collection. MEASUREMENTS AND RESULTS: Integrated concentration of plasma growth hormone was correlated with urinary growth hormone levels from both the 24-hour (r = .67; P < .0001) and 12-hour overnight (r = .52; P < .001) measurements. Peak growth hormone response to paired stimulation was not correlated with either the 24-hour (r = .26; P < .23; n = 28) or 12-hour (r = .16; P < .48; n = 28) urinary growth hormone levels. The mean 24- and 12-hour urinary growth hormone levels for the patients with normal integrated concentrations of growth hormone were significantly higher than those in patient groups having subnormal integrated concentrations of growth hormone (P < .05). However, there was considerable overlap in the 12- and 24-hour urinary growth hormone levels between the patients with normal and those with subnormal integrated concentrations of growth hormone. Only one patient who had subnormal integrated concentrations of growth hormone had a 24-hour urinary growth hormone level higher than 9 ng, and none had a 12-hour urinary growth hormone level higher than 7 ng. The mean 12- and 24-hour urinary growth hormone levels were significantly higher in patients who received growth hormone injection than in those with normal spontaneous integrated concentrations of growth hormone and had no overlap with patients who had subnormal integrated concentrations of growth hormone. CONCLUSIONS: (1) Urinary and integrated concentrations of plasma growth hormone are correlated; (2) patient diagnoses based on integrated plasma growth hormone levels exhibit a high degree of overlap of urinary growth hormone; and (3) urinary growth hormone levels can serve to monitor compliance with growth hormone therapy.

Plasma cortisol levels increase with age in obese subjects.

In order to assess the influence of age, sex, and body mass on plasma cortisol concentrations, we measured the 24-hour Integrated Concentration (IC) of cortisol (F) in 36 obese subjects (16 males, 20 females) aged 5.3-56.4 years, BMI = 35.5 +/- 7.3 kg/m2 and compared with 119 nonobese subjects, body mass indices (BMI) 21.2 +/- 2.7 kg/m2, aged 8.8-66.2 years (55 males, 64 females). Subjects were nondiabetic, normotensive, without history of psychiatric illness, and otherwise in good health. IC studies were performed using a continuous blood withdrawal methodology, and IC-F was assayed from the 24 hour pooled sample by a protein binding method. The effect of age and gender on IC-F was analyzed by multivariate regression. In the nonobese group there was no effect of age or sex on IC-cortisol levels, the mean IC-F = 173.8 +/- 44.1 nmol/L. A statistically significant but weak negative effect of BMI on IC-cortisol (r=-.18, p<0.05) was present. In the obese subjects there was a significant increase in IC-cortisol levels with age IC-F(nmol/L) = 2.76 x age(years) + 85.0 (r2=.36, p<0.0001). IC-cortisol levels tended to be lower in obese males than females when controlled for age (p<0.05). We conclude that in nonobese subjects IC-F levels are independent of age and gender. However, there is a significant increase of IC-cortisol levels with age in obese individuals. The observed increase of IC-cortisol with age may contribute to metabolic complications of obesity.