In Vitro Effects of Pioglitazone on the Expression of Components of Wnt Signaling Pathway and Markers of Bone Mineralization.

Pioglitazone is an insulin-sensitizing thiazolidinedione (TZD) whose use is associated with bone loss. We examined the effects of pioglitazone on components of the Wnt signaling pathway (Wnt1, ß-catenin) and markers of bone mineralization [osteoprotegerin (OPG), bone sialoprotein (BSP), fibroblast growth factor (FGF)23] as well as mineral content in human osteoblast hFOB 1.19 cells. hFOB 1.19 cells were cultured in K12/DMD medium with or without pioglitazone. PPARγ Wnt1, OPG, BSP, or FGF23 mRNA expression was measured using qRT-PCR; ß-catenin, OPG, BSP, or FGF23 using ELISA; and calcium or phosphate content using colorimetry. Treatment with pioglitazone resulted in increased expression of PPARγ mRNA in hFOB 1.19 osteoblasts. Pioglitazone decreased Wnt1 mRNA levels and suppressed components of Wnt signaling pathway as evidenced by a decrease in ß-catenin gene expression and secretion as well as ß-catenin specific activity. The expression and the activity of OPG, BSP, and FGF23 were also reduced by pioglitazone together with total (but not specific) calcium and phosphate content. Pioglitazone affects Wnt1 signaling pathway and mineral matrix regulation components in human osteoblasts.

Thiazolidinediones (TZDs) affect osteoblast viability and biomarkers independently of the TZD effects on aromatase.

Thiazolidinediones (TZDs) are insulin sensitizers used for treatment of diabetes. We have previously reported that TZDs reduce estrogen synthesis by inhibiting aromatase activity in human granulosa cells (HGC). Multiple clinical trials demonstrated that TZDs increase the risk of fractures in postmenopausal women with type 2 diabetes. We studied mouse osteoblasts alone or in a co-culture with HGC to determine whether TZD inhibition of aromatase plays a role in their effects on bone metabolism. Mouse osteoblasts were cultured with and without HGC, and incubated in a medium with or without testosterone, pioglitazone or rosiglitazone. Cell growth, oleic acid uptake, alkaline phosphatase activity, and osteocalcin production were measured. TZDs inhibited estradiol production by up to 84% in HGC/mouse osteoblast co-cultures. TZDs induced mouse osteoblast death and increased oleic acid uptake. TZDs also inhibited alkaline phosphatase activity (58-75%, p<0.046) and osteocalcin production (52-75%, p<0.031). For all the parameters, there were no significant differences between the osteoblast cultures alone and the HCG/osteoblast co-cultures. TZD effects on osteoblast viability, oleic acid uptake, alkaline phosphatase and osteocalcin production are independent of their effects on aromatase.

Continuous insulin infusion is associated with a reduced post-surgical length of stay, but not with the complication rate, in patients with diabetes mellitus undergoing coronary artery bypass graft.

OBJECTIVE: To establish if glucose management with continuous intravenous insulin infusion (CII) in the early post-operative period after coronary artery bypass graft (CABG) surgery is associated with complication rate and length of hospital stay (LOS) in patients with diabetes mellitus (DM). RESEARCH DESIGN AND METHODS: We reviewed the records of 587 patients with DM who underwent CABG from January 1999 until January 2008; 316 patients were placed on CII, while 271 patients were treated with subcutaneous insulin. We examined patient age, glycated hemoglobin (HgbA1c), 24- and 72-h post-operative average capillary blood glucose (CBG), length of stay (LOS), and the rate of complications. RESULTS: There was no difference in HgbA1c between the groups. Mean CBG values at both 24 h and 72 h remained the same in the CII group (167 mg/dl), while in the non-CII group they were 194 mg/dl and 189 mg/dl, respectively (p<0.001 between the groups). Post-surgical median LOS was 6 days in the CII group and 6.5 days in the non-CII group (p=0.003). Complications occurred at similar rate (in 10% and 11% of patients) in the two groups. CONCLUSIONS: CII is associated with a reduced post-surgical LOS in patients with DM who undergo CABG.

Differential roles of MAPK-Erk1/2 and MAPK-p38 in insulin or insulin-like growth factor-I (IGF-I) signaling pathways for progesterone production in human ovarian cells.

Insulin and insulin like-growth factor-I (IGF-I) participate in the regulation of ovarian steroidogenesis. In insulin resistant states ovaries remain sensitive to insulin because insulin can activate alternative signaling pathways, such as phosphatidylinositol-3-kinase (PI-3 kinase) and mitogen-activated protein-kinase (MAPK) pathways, as well as insulin receptors and type 1 IGF receptors. We investigated the roles of MAPK-Erk1/2 and MAPK-p38 in insulin and IGF-I signaling pathways for progesterone production in human ovarian cells. Human ovarian cells were cultured in tissue culture medium in the presence of varying concentrations of insulin or IGF-I, with or without PD98059, a specific MAPK-Erk1/2 inhibitor, with or without SB203580, a specific MAPK-p38 inhibitor or with or without a specific PI-3-kinase inhibitor LY294002. Progesterone concentrations were measured using radioimmunoassay. PD98059 alone stimulated progesterone production in a dose-dependent manner by up to 65% (p<0.001). Similarly, LY294002 alone stimulated progesterone production by 13-18% (p<0.005). However, when used together, PD98059 and LY294002 inhibited progesterone production by 17-20% (p<0.001). SB203580 alone inhibited progesterone production by 20-30% (p<0.001). Insulin or IGF-I alone stimulated progesterone production by 40-60% (p<0.001). In insulin studies, PD98059 had no significant effect on progesterone synthesis while SB203580 abolished insulin-induced progesterone production. Either PD98059 or SB203580 abolished IGF-I-induced progesterone production. Both MAPK-Erk1/2 and MAPK-p38 participate in IGF-I-induced signaling pathways for progesterone production, while insulin-induced progesterone production requires MAPK-p38, but not MAPK-Erk1/2. These studies provide further evidence for divergence of insulin and IGF-I signaling pathways for human ovarian cell steroidogenesis.

Rosiglitazone and pioglitazone inhibit estrogen synthesis in human granulosa cells by interfering with androgen binding to aromatase.

The effects of rosiglitazone or pioglitazone (thiazolidinediones, TZDs) on estrogen production and aromatase activity in human ovarian cells were examined. Human granulosa cells were incubated in the tissue culture medium supplemented with androstenedione or testosterone, with or without insulin, TZDs, or type 1 17ß-hydroxysteroid-dehydrogenase (17ß-HSD) inhibitor. Estrogen concentrations in the conditioned medium, aromatase mRNA and protein expression in the cells and androgen substrate binding to aromatase were measured. With androstenedione as substrate, rosiglitazone or pioglitazone inhibited estrone production by up to 22% (p<0.012) while type 1 17ß-HSD inhibitor enhanced this effect of rosiglitazone or pioglitazone by 37% (p<0.001) and by 67% (p<0.001), respectively. With testosterone as substrate, rosiglitazone or pioglitazone inhibited estradiol production by 32% (p<0.001). With (3)H-testosterone as substrate, rosiglitazone or pioglitazone inhibited the (3)H-tritiated water release by the cultured cells by 45% and 35%, respectively, thus directly demonstrating inhibition of aromatase. Rosiglitazone or pioglitazone, however, had no significant effect on aromatase mRNA or protein expression. Rosiglitazone or pioglitazone inhibited (125)I-androstenedione and (125)I-testosterone binding to aromatase by 38% (p<0.001). It was concluded that rosiglitazone or pioglitazone inhibit estrogen synthesis in human granulosa cells by interfering with androgen binding to aromatase.

Vitamin D regulates steroidogenesis and insulin-like growth factor binding protein-1 (IGFBP-1) production in human ovarian cells.

Vitamin D Receptor (VDR) is expressed in both animal and human ovarian tissue, however, the role of vitamin D in human ovarian steroidogenesis is unknown. Cultured human ovarian cells were incubated in tissue culture medium supplemented with appropriate substrates, with or without 50 pM-150 pM or 50 nM-150 nM of 1,25-(OH)2D3, and in the presence or absence of insulin. Progesterone, testosterone, estrone, estradiol, and IGFBP-1 concentrations in conditioned tissue culture medium were measured. Vitamin D receptor was present in human ovarian cells. 1,25-(OH)2D3 stimulated progesterone production by 13% (p<0.001), estradiol production by 9% (p<0.02), and estrone production by 21% (p<0.002). Insulin and 1,25-(OH)2D3 acted synergistically to increase estradiol production by 60% (p<0.005). 1,25-(OH)2D3 alone stimulated IGFBP-1 production by 24% (p<0.001), however, in the presence of insulin, 1,25-(OH)2D3 enhanced insulin-induced inhibition of IGFBP-1 production by 13% (p<0.009). Vitamin D stimulates ovarian steroidogenesis and IGFBP-1 production in human ovarian cells likely acting via vitamin D receptor. Insulin and vitamin D synergistically stimulate estradiol production. Vitamin D also enhances inhibitory effect of insulin on IGFBP-1 production.

Vitamin D in diabetes mellitus-a new field of knowledge poised for D-velopment.

This commentary reviews the current state of knowledge regarding the role of vitamin D in the pathogenesis of diabetes mellitus. In type 1 diabetes mellitus or in adult onset latent autoimmune diabetes (LADA), vitamin D exhibits immunomodulatory actions, influencing the activity of lymphocytes and interleukins. In type 2 diabetes mellitus vitamin D appears to act through different mechanisms, affecting insulin secretion and insulin sensitivity through its effects on the beta cells, mediators of inflammation and parathyroid hormone. Much work remains to be done in this new field of knowledge before the role of vitamin D in the pathogenesis of diabetes mellitus is completely understood.

Metabolic and hormonal effects of oral DHEA in premenopausal women with HIV infection: a randomized, prospective, placebo-controlled pilot study.

Women with HIV infection use dehydroepiandrosterone (DHEA) because of its potential effects on mood and energy. We examined the effects of DHEA on the hypothalamic-pituitary-adrenal and gonadal axes and on insulin sensitivity. Fifteen HIV-positive women were randomized to receive placebo (6 subjects) or oral DHEA (9 subjects). ACTH-, CRF-, and GnRH-stimulation tests were performed before and after 8 weeks of treatment. DHEA, DHEA-S, dihydrotestosterone, total testosterone, free testosterone, sex hormone-binding globulin, estrone, estradiol, cortisol, insulin, IGF-1, IGFBP-1, IGFBP-3, and adiponectin in plasma or serum were measured. There was a significant increase in DHEA (p<0.004), DHEA-S (p<0.008), total testosterone (p<0.008), dihydrotestosterone (p<0.004), androstenedione (p<0.04), and estrone (p<0.03) from baseline within the DHEA group but not within the placebo group. There was a significant increase in DHEA (p<0.0006), DHEA-S (p<0.032), total testosterone (p<0.01), and dihydrotestosterone (p<0.005) in the DHEA group compared with the placebo group. Oral DHEA produces significant increases in circulating DHEA, DHEA-S, testosterone, DHT, and, possibly, androstenedione and estrone levels in premenopausal women with HIV infection. In the current pilot study these hormone changes did not affect the pituitary or adrenal axis or insulin/IGF indices. Long-term studies with larger groups of patients are needed to confirm these data and to determine their clinical significance.

Reproducibility of a glycemic response to mixed meals in type 2 diabetes mellitus.

OBJECTIVE: To evaluate the inter-patient and intra-patient reproducibility of the glycemic response to a mixed meal in individuals with type 2 diabetes mellitus (DM). SUBJECTS/SETTING: Six individuals with DM were admitted to the General Clinical Research Center for 6 days. INTERVENTION: Subjects consumed 3 different meal plans consisting of 4 meals daily (breakfast, lunch, dinner and snack) on 2 separate occasions. Serum insulin and glucose levels were sampled at 19 time points every day. The glycemic response (GR) to a meal was calculated as the area under the glucose response curve after consumption of a given meal. In addition, the incremental area under the curve (IGR) was calculated assuming a pre-prandial (baseline) glucose value before each meal as zero. RESULTS: Intra-patient correlation coefficients (R) of GR for meals in subjects with DM were quite good, ranging 0.69-0.94. The range of the inter-patient coefficients of variation (CV) for the same meals was 21.5-30.4%. For IGR, the R values ranged from 0.64 to 0.91 for 8 out of 12 meals, confirming good intra-patient reproducibility for these meals. CV for IGR ranged from 31% to 113%. CONCLUSIONS: For patients with DM, the GR of individual meals exhibits excellent intra-patient reproducibility, allowing prediction of the glycemic response to a given meal in an individual subject. However, significant inter-patient variability of the GR precludes its use for the prediction of post-prandial glucose concentrations in groups of patients with diabetes.

Comparison of 1-micro g and 250-micro g corticotropin stimulation tests for the evaluation of adrenal function in patients with acquired immunodeficiency syndrome.

Many patients with acquired immunodeficiency syndrome (AIDS) have symptoms suggestive of adrenal insufficiency, but a normal 250- micro g corticotropin (ACTH) stimulation test. We compared the results of 1- micro g and standard 250- micro g ACTH stimulation tests in patients with AIDS. Each patient was studied on 2 separate days. On day 1, 1 micro g ACTH was given intravenously at 8 am after an overnight fast and serum cortisol levels were measured at baseline, and 30 and 60 minutes after ACTH infusion. On day 2, the procedure was repeated with 250- micro g ACTH. An absolute peak cortisol value of > 18 micro g/dL and an increment of 7 micro g/dL or more from baseline constituted a normal response. Among 31 patients, 16 (52%) had discrepant results: 14 (45%) had subnormal responses to 1 micro g ACTH but normal responses to 250 micro g ACTH (group 1); 2 (6%) had normal responses to 1 micro g but subnormal responses to 250 micro g (group 2) ACTH; 6 patients (19%) had concordant abnormal responses (group 3); and 9 (30%) had concordant normal responses (group 4). Eight patients of group 1 underwent a confirmatory insulin tolerance test (ITT); 4 of these patients had abnormal responses to ITT. Kappa statistic and McNemar's test were used to evaluate the data. A kappa statistic value of 0.095 and a P value less than.003 for the McNemar test indicate only random level of agreement and significant differences in the probability of positive result between the 2 ACTH tests. We conclude that discrepancies between the 1- micro g and the 250- micro g ACTH stimulation tests are common in patients with AIDS, with the likelihood of agreement with the "gold standard" ITT of only 50% for each test in our sample of patients. Larger studies are needed to further evaluate the use of these tests in patients with AIDS.

Use of insulin-sensitizing agents in patients with polycystic ovary syndrome.

OBJECTIVE: To review the subject of polycystic ovary syndrome and the therapeutic use of insulin-sensitizing agents in patients with this endocrinopathy. METHODS: We present background information on this disorder and summarize the pertinent published literature. RESULTS: Polycystic ovary syndrome affects approximately 7.5% of reproductive-age women in the United States. Although specific diagnostic criteria for this condition have not been established, the presence of three major factors-chronic anovulation, hyperandrogenemia, and clinical signs of hyperandrogenism-has been proposed as essential for consideration of the diagnosis. A high ratio of serum luteinizing hormone to follicle-stimulating hormone is found in 60 to 75% of women with this syndrome. Treatment with metformin may yield heterogeneous responses in differing populations with polycystic ovary syndrome, but most studies have shown evidence of restoration of ovulatory cycling. In addition, weight loss and decreases in free and total testosterone levels have been reported. Troglitazone therapy proved somewhat less efficacious than metformin for restoring menstrual cycles and similar to metformin in producing hormonal responses. Because troglitazone is no longer available for clinical use, studies will need to be extended to other thiazolidinediones. Patients treated with another insulin sensitizer, D-chiro-inositol, have demonstrated improved insulin sensitivity, ovulatory rates, and biochemical findings. CONCLUSION: Current evidence suggests that the use of insulin-sensitizing agents in patients with polycystic ovary syndrome not only improves their sensitivity to the effects of insulin on glucose and lipid metabolism but also ameliorates clinical and biochemical manifestations of hyperandrogenism and increases rates of ovulation. Multicenter studies with larger numbers of patients are needed.

Phosphatidyl-inositol-3 kinase-independent insulin action pathway(s) in the human ovary.

Hyperandrogenism observed in women with a variety of insulin-resistant states is thought to be due to a stimulatory effect of insulin on ovarian steroid hormone production. However, it is not known what mechanisms could allow the ovary to remain sensitive to insulin while classical target organs for insulin action (liver, fat, and muscle) exhibit insulin resistance. One hypothesis proposed to explain this paradox suggests that a postbinding divergence of insulin receptor signaling occurs in the ovary and that signaling pathways for steroid hormone synthesis and other ovarian effects of insulin may be distinct from classical glucose signaling pathways. We now report that activation of phosphatidyl-inositol-3 (PI-3) kinase, which is crucial for glucose transport, is not necessary for the insulin-induced stimulation of progesterone production or for the insulin-induced inhibition of insulin-like growth factor binding protein 1 (IGFBP-1) production in cultured human ovarian cells. Human granulosa cells obtained during in vitro fertilization procedures were cultured with 10, 10(2), 10(3), or 10(4) ng/mL insulin with or without preincubation with 100 nM wortmannin, a specific irreversible inhibitor of PI-3 kinase. IGFBP-1 concentration in the conditioned medium was measured using immunoradiometric assay or by Western blot analysis. Progesterone concentration was measured using RIA. Additional studies were carried out in cultures of human ovarian cells prepared from homogenized whole ovarian tissue of a woman with a family history of breast cancer and a mutation of BRCA-1 gene who underwent bilateral oophorectomy. These cells were cultured with 10(3) ng/mL insulin with or without preincubation with 100 nM wortmannin. Two-way ANOVA was used to compare mean values of IGFBP-1 and progesterone according to insulin dose and the use of wortmannin. In cultured granulosa cell medium, progesterone production was stimulated by insulin in a dose-related manner up to 175% of control (P < 0.0001). In tissue culture medium from ovarian cells obtained from a patient with BRCA-gene mutation, concentration of progesterone in the tissue culture medium increased from 2.5 +/- 0.2 ng/mL for control to 5.4 +/- 0.3 ng/mL for cells incubated with insulin (P < 0.001). IGFBP-1 production in tissue culture medium from human granulosa cells was inhibited by insulin to the nadir of 45% of control (P < 0.0001). Preincubation with wortmannin, despite complete inhibition of PI-3 kinase in both cell systems confirmed by Western blot analysis, failed to significantly alter these results. We conclude that inhibition of PI-3 kinase by wortmannin fails to abolish stimulatory effect of insulin on progesterone production or inhibitory effect of insulin on IGFBP-1 production in cultured human ovarian cells. These findings suggest that activation of PI-3 kinase, an enzyme crucial for insulin-stimulated glucose transport, is not necessary for the above effects of insulin in the ovary. These data provide evidence for the presence of PI-3 kinase-independent insulin signaling pathway(s) in human ovarian cells.

Lack of postprandial leptin peaks in patients with type 2 diabetes mellitus.

Leptin production by the adipocyte is acutely stimulated by insulin in vitro. In normal individuals, postprandial insulin peaks are not accompanied by corresponding changes in circulating leptin. Postprandial regulation of leptin in individuals with type 2 diabetes, to our knowledge, has not been previously examined in detail. We examined the effect of meals on circulating leptin levels in six patients with type 2 diabetes who were not treated with insulin and in seven normal individuals. After informed consent was obtained, all subjects were admitted to the General Clinical Research Center for 6 days. They consumed four meals daily (breakfast, lunch, dinner and snack). Eighteen blood samples were drawn between 7.40 a.m. and midnight for the determination of serum leptin, insulin and glucose levels. Postprandial peaks were clearly identifiable for glucose and insulin levels both in normal subjects and in those with type 2 diabetes. However, no postprandial peaks of leptin levels were present. Correlation analysis demonstrated a lack of correlation between leptin levels and the levels of glucose or insulin. We conclude that, in spite of the presence of postprandial insulin peaks, there are no acute changes in circulating leptin levels postprandially in patients with type 2 diabetes who are not on insulin therapy. In this regard, in-vivo regulation of leptin by meals in patients with type 2 diabetes is similar to that in normal individuals.

Synergistic effects of insulin-like growth factor-I and human chorionic gonadotropin in the rat ovary.

Insulin and low doses of lutenizing hormone (LH) activity (human chorionic gonadotropin [hCG]) act synergistically in the rat to produce anovulation, large ovarian cysts, and elevated plasma androstenedione levels. Further, both insulin and insulin-like growth factor-I (IGF-I) affect the ability of gonadotropins to enhance both ovarian theca and granulosa cell function in vitro. The present series of experiments were performed to determine if recombinant human IGF-I (rhIGF-I) can act in a manner similar to insulin when combined with subovulatory doses of hCG in adult normally cycling rats. Fifty-four female Sprague-Dawley rats were randomly assigned to the following treatment groups at the age of 64 days: (A) vehicle alone (controls, phosphate-buffered saline containing 0.09% pig gelatin), (B) twice-daily subcutaneous injections of 0.5 to 3.0 U insulin, (C) twice-daily subcutaneous injections of 1.5 U hCG, (D) both insulin and hCG, (E) twice-daily subcutaneous injections of rhIGF-I (2.5 mg/kg/d), and (F) both hCG and rhIGF-I. After 22 days of treatment, the animals were killed on day 23, trunk blood was collected, and the ovaries were excised for histological study. Eight of 9 control rats and 5 or 6 of 9 rats treated with insulin, hCG, or rhIGF-I alone displayed normal estrus cycles throughout the in vivo treatment period as assessed by daily vaginal smears. In marked contrast, only 1 animal treated with hCG + insulin and 2 animals treated with hCG + rhIGF-I continued to display vaginal smears indicative of normal cycling. Multiple large ovarian follicular cysts were found only in these latter 2 groups (3 of 9 animals in each group). Mean serum testosterone levels were significantly elevated in animals receiving insulin + hCG (0.72 +/- 0.28 v 0.17 +/- 0.03 ng/mL in controls, P = .05). Mean serum androstenedione levels were significantly elevated in animals receiving hCG and animals receiving rhIGF-I + hCG (5.57 +/- 0.99 and 2.39 +/- 0.68 ng/mL, respectively, v0.14 +/- 0.14 ng/mL in controls, P< .01 and P< .05, respectively). We conclude that rhIGF-I and insulin act synergistically with subovulatory doses of hCG to disrupt normal reproductive cycling, elevate serum androgen concentrations, and induce large ovarian cysts in intact adult rats.

Insulin has a biphasic effect on the ability of human chorionic gonadotropin to induce ovarian cysts in the rat.

Hyperinsulinemia enhances the ability of subovulatory doses of human chorionic gonadotropin (hCG) to induce ovarian follicular cysts in the rat. To determine the relative contribution of these hormones to the development of ovarian cysts, adult female rats were treated with either (1) vehicle alone (controls), (2) a high-fat diet (HFD) to control for the effects of weight gain, (3) 1.5 to 6 IU hCG twice daily plus 6 U insulin (Ins)/d, or (4) 1.5 to 9 U Ins/d plus 3 IU hCG twice daily. On day 23 of the in vivo treatments, all groups that received at least 6 U Ins/d displayed increased body weight compared with control and HFD rats (P < or = .05). No control rats and only one HFD rat displayed ovarian cysts on this day. Plasma estrone (E1) and androstenedione (A4) were elevated in HFD rats with noncystic follicles compared with control rats (P < or = .05). Between 64% and 80% of rats on 6 U Ins/d plus twice-daily injections of 1.5 to 6 IU hCG displayed ovarian cysts on day 23. Plasma estradiol (E2) concentrations for these treatment groups were similar to those of control rats. Of the hormonally treated animals, only those that had ovarian cysts in response to twice-daily injections of 4.5 or 6 IU hCG plus 6 U Ins/d displayed elevated plasma A4 and/or testosterone compared with controls. In contrast, plasma E1 concentrations were elevated on day 23 for animals bearing ovarian cysts in response to increasing doses of hCG plus the fixed dose of 6 U Ins/d. Between 70% and 80% of rats treated twice daily with 3 IU hCG plus a daily dose of 1.5 to 6 U Ins displayed ovarian cysts on day 23. In marked contrast, only 25% of rats treated with this dose of hCG plus 9 U Ins/d developed cystic follicles. Of the plasma steroids tested, only E1 and A4 were elevated in these treatment groups compared with controls. However, these increases in plasma steroid concentrations did not correlate with the dose of insulin. We conclude from these data that, although the mechanisms remain to be elucidated, extreme hyperinsulinemia has the paradoxical ability to attenuate the induction of ovarian cysts by hCG in some animals.

Dehydroepiandrosterone sulfate (DHEAS) and testosterone: relation to HIV illness stage and progression over one year.

This study explored associations between serum dehydroepiandrosterone sulfate (DHEAS), free and total testosterone levels, and HIV illness markers, including viral load, and the behavioral problems of fatigue and depressed mood. Subjects were 169 HIV-positive men evaluated at baseline, 6, and 12 months for levels of DHEAS, total and free testosterone, HIV RNA, CD4, HIV symptoms, opportunistic illnesses, fatigue, and depression. Men with AIDS (N = 105), compared with men with less advanced illness, had lower mean levels of DHEAS. Baseline DHEAS was positively correlated with CD4 count, HIV symptom severity, and was inversely correlated with HIV RNA. Baseline DHEAS below the laboratory reference range (96 microg/dl) was associated with history of opportunistic infections and malignancies (adjusted odds ratio [OR], 4.4; 95% confidence interval [CI], 1.9-10.4) and with incidence of these complications or death over 1 year (adjusted OR, 2.6; 95% CI, 1-7.2). Initiating protease inhibitor combination therapy was associated with an increase in DHEAS over 6 months. Free testosterone was inversely correlated with HIV RNA, but there were no other significant associations between testosterone and HIV illness markers. No hormone was related to fatigue or depression. This study confirms that low serum DHEAS is associated with HIV illness markers, including viral load, and carries negative prognostic value. Further, protease inhibitor therapy may result in increased circulating DHEAS.

Hospital length of stay on a diabetes cluster unit.

OBJECTIVE: To determine whether the length of hospital stay for patients with either a primary or a secondary diagnosis of diabetes mellitus can be reduced by use of a diabetes cluster unit. METHODS: For a period of 12 months, we compared the length of hospital stay for 462 patients with diabetes who were admitted to a diabetes cluster unit and for 1,855 patients with diabetes who were admitted to the other medical and surgical units during the same period. Statistical analysis was done by comparing the calculated average length of stay for patients with primary and secondary diagnoses of diabetes admitted to both types of hospital accommodations and comparing these values with the expected lengths of hospital stay based on the New York Group Standard. RESULTS: The mean length of hospital stay for patients who were admitted to the diabetes unit for a primary diagnosis of diabetes was 5.8 days shorter than for those admitted to regular medical and surgical units (7.6 days or 1% below the expected duration versus 13.4 days or 70% longer than the expected duration, respectively; P<0.0002). For a secondary diagnosis of diabetes, patients in the diabetes unit had an average length of stay of 15.5 days (35% longer than the expected duration) in comparison with 17 days (45% longer than expected) for patients on the other medical and surgical units (no significant difference). CONCLUSION: For hospitalization of patients with diabetes, geographic clustering seems to reduce the length of stay for patients with a primary diagnosis of diabetes mellitus.

Twenty-four-hour urinary free cortisol in patients with acquired immunodeficiency syndrome.

Many patients with acquired immune deficiency syndrome (AIDS) have symptoms consistent with adrenal insufficiency, but only a small subset of these patients meet criteria for adrenal insufficiency during a short corticotropin (ACTH) stimulation test. We hypothesized that patients with AIDS and symptoms of adrenal insufficiency who produce normal amounts of cortisol in response to administration of 0.25 mg cosyntropin may nevertheless produce lower amounts of cortisol in a course of 24 hours than comparably sick AIDS patients without symptoms of adrenal insufficiency or comparably sick patients without AIDS. We studied four groups of male patients: AIDS patients with symptoms suggestive of adrenal insufficiency but with a normal response to cosyntropin (group I), AIDS patients without symptoms suggestive of adrenal insufficiency (group II), human immunodeficiency virus (HIV)-negative patients with serious acute or chronic illness (group III), and healthy subjects (group IV). The following variables were examined: age, CD4 cell count, Acute Physiologic and Chronic Health Evaluation (APACHE) score, serum cortisol and plasma ACTH at baseline; serum cortisol at 30 and 60 minutes after intravenous administration of 0.25 mg cosyntropin; and 24-hour urinary free cortisol. The four groups had a similar mean age and baseline plasma ACTH and serum cortisol levels. However, a change in cortisol from baseline to 30 and 60 minutes after administration of cosyntropin was significantly smaller in both groups of AIDS patients than in the sick patients without AIDS and normal subjects. There were also differences noted between the two groups of AIDS patients: both baseline and stimulated levels of cortisol tended to correlate directly with ACTH levels in patients without symptoms of adrenal insufficiency, while this relationship appeared to be inverse in patients with symptoms suggestive of adrenal insufficiency (r = -.57 to -.7, P < .05 to .14). The 24-hour urinary free cortisol levels were similar among all groups, but correlated strongly with baseline and stimulated serum cortisol levels only in patients with AIDS and symptoms of adrenal insufficiency (r = .8 to .9, P < .002 to .015). We conclude that (1) AIDS patients with and without symptoms of adrenal insufficiency may have either normal adrenal function or somewhat suboptimal adrenal reserve as demonstrated by a blunted cortisol response during the short ACTH stimulation test in comparison to HIV-negative comparably sick patients or healthy subjects; and (2) 24-hour urinary free cortisol is not a useful test for detection of subtle abnormalities of adrenal function in patients with AIDS.

Effects of an intervention by a diabetes team in hospitalized patients with diabetes.

OBJECTIVE: Hospitalized patients with diabetes have a prolonged length of stay in the hospital. We conducted a controlled prospective randomized feasibility study of the effects of a diabetes team (a diabetes nurse educator and an endocrinologist) on the length of stay and other outcomes of hospitalization in these patients. RESEARCH DESIGN AND METHODS: A total of 179 hospitalized patients with diabetes were randomly assigned to receive usual care supplemented with (85 patients) or without (94 control patients) a diabetes team intervention. Outcome measures included the length of stay, blood glucose control, and rates of readmission. RESULTS: For the primary diagnosis of diabetes, the median length of stay was 5.5 days (95% CI 4-8 days) for patients who received diabetes team intervention and 7.5 days (5-11 days) for the control patients (NS). For the secondary diagnosis of diabetes, the median length of stay was 10.0 days (8-13 days) in the intervention group and 10.5 days (8-13 days) in the control group (NS). One month after the team intervention was initiated, 75% of patients in the intervention group were in good glycemic control, compared with 46% in the control group. Readmissions at 3 months after discharge included 13 (15%) patients from the intervention group and 30 (32%) patients in the control group (P = 0.01). CONCLUSIONS: Randomized controlled prospective trials of clinical interventions in hospitalized patients with diabetes are feasible. Diabetes team intervention appears to reduce the hospital length of stay and to improve glycemic control. Team intervention significantly reduces the rate of recurrent hospitalization.