| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Departments of Intensive Care Medicine (L.L., S.V.P., P.J.W., G.V.d.B.) and Abdominal Surgery (A.D.), Katholieke Universiteit Leuven, B-3000, Leuven, Belgium; and Medical Department M (Endocrinology and Diabetes) (T.K.H.), Aarhus University Hospital, DK-8000 Aarhus C, Denmark
Address all correspondence and requests for reprints to: Greet Van den Berghe, M.D., Ph.D., Department of Intensive Care Medicine, Katholieke Universiteit Leuven, B-3000, Leuven, Belgium. E-mail: greet.vandenberghe{at}med.kuleuven.be.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Objective: To investigate further the association between insulin dose and adverse outcome, we studied the effect of IIT on circulating insulin levels, markers of insulin sensitivity, and the metabolic and mitogenic insulin signaling molecules in key tissues.
Design: This is a subanalysis of a large randomized, controlled study.
Setting: The study was performed in a university hospital surgical ICU.
Patients: A total of 339 critically ill patients, treated in ICU for at least a week, were included in this subanalysis.
Intervention: Strict normoglycemia with IIT compared with conventional insulin therapy was performed.
Results: Severalfold higher insulin doses than with conventional insulin therapy were required to maintain normoglycemia with IIT. However, serum insulin levels were only transiently higher with IIT, despite the much lower blood glucose levels. IIT normalized the elevated serum C-peptide levels and increased circulating adiponectin levels. The metabolic insulin signal was increased by IIT in muscle, but not in liver. The mitogenic insulin signal in either tissue was not affected by IIT.
Conclusions: Normoglycemia can be maintained in ICU patients without a sustained further elevation of insulinemia. Together with the increased adiponectin levels, this finding suggests that IIT may improve insulin sensitivity. Skeletal muscle, but not liver, revealed an increased metabolic insulin signal. The therapy did not impose mitogenic risk in these tissues.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Insulin actions on metabolism, cell growth, and differentiation are mediated by intracellular signaling pathways. Insulin binding to its receptor can activate two signaling pathways: the so-called "metabolic" pathway, which proceeds through the insulin receptor substrates (IRSs) and depends on the activation of PI3K; and the "mitogenic" pathway, which proceeds through Shc/Grb2 activation, leading to activation of different MAPK isoforms (5). In a rat model of sepsis, abnormalities along the IRS/PI3K pathway have been described (6), whereas in human critical illness, little is known about the defects in insulin signaling.
In two large, prospective, randomized, controlled clinical studies of critically ill patients, we demonstrated that strict maintenance of normoglycemia with insulin infusion for at least a few days reduces morbidity and mortality (7, 8). However, in multivariate analysis and corrected for blood glucose levels, the dose of insulin has been an independent risk factor for mortality (9, 10).
We hypothesized that in critically ill patients, high doses of insulin result in hyperinsulinemia, which in the presence of a responsive mitogenic insulin signaling pathway, may increase the risk of mitogenic complications. This would partly explain the possible link between high insulin dose and adverse outcome, particularly for patients with cancer. To investigate further this possibility, we documented the degree of hyperinsulinemia and C-peptide levels achieved with intensive insulin therapy (IIT). In addition, the impact of IIT on biological markers of insulin action and sensitivity was assessed. Finally, the metabolic and mitogenic insulin signaling pathways in key tissues as liver and skeletal muscle were studied.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
We studied patients who had been included in a large, randomized, controlled clinical trial of IIT in adult, mechanically ventilated patients admitted to a surgical intensive care unit (ICU) (7). In this trial, a total of 1548 patients were randomly assigned to either IIT, in which blood glucose concentrations were kept tightly between 80 and 110 mg/dl, resulting in a mean daily blood glucose of 101 mg/dl, or conventional insulin therapy (CIT), in which insulin was only administered when blood glucose levels exceeded 220 mg/dl, resulting in a mean daily blood glucose of 152 mg/dl. To limit the samples to a reasonable number, we only studied prolonged critically ill patients because IIT was most effective when continued for at least a few days. We analyzed blood samples from all 339 prolonged critically ill patients who were treated in the ICU (Table 1
) for at least 1 wk, that were taken at 0600 h on the day of ICU admission, d 2, d 7, and on the last day of intensive care. Blood samples from 26 overnight fasted healthy volunteers were obtained for comparison [mean age 69 ± 8 yr, mean body mass index (BMI) 26.8 ± 3.5 kg/m2]. After centrifugation, serum was kept frozen at –80 C until analysis. Postmortem biopsy samples of liver (segment IVb) and muscle (right musculus rectus abdominis) were taken within minutes after death (Table 2) from 74 of the 98 patients who died in the ICU. Tissue samples were snap frozen in liquid nitrogen and stored at –80 C until analysis. For gene expression data, good quality RNA was harvested from 66 biopsy samples (44 CIT and 22 IIT patients). For comparison, we collected liver (segment IVb) and muscle (right musculus rectus abdominis) biopsy samples taken from 22 age (mean 69 ± 13 yr) and BMI (mean 25.1 ± 2.6) matched patients during laparotomy for clinical indications at the end of the surgical procedure (e.g. for hemi- or segmentary hepatectomy, or restorative rectal resection in patients with rectal cancer). For the assessment of insulin signaling, we selected 36 patients from whom biopsy samples were available, 18 from each group (IIT vs. CIT), and who were matched for demographic characteristics, reason for admission to the ICU, invasive therapeutic strategies, and cause of death (Table 2). All study protocols were approved by the Institutional Review Board of the Katholieke Universiteit Leuven. Written informed consent was obtained from all healthy volunteers and the patients, or, when the patient was unable to give consent, from the closest family member.
|
|
Serum insulin concentrations were quantified in duplicate by RIA using a guinea pig-derived antibody (kindly provided by Professor R. Bouillon, Katholieke Universiteit Leuven, Leuven, Belgium) and an insulin tracer from BioSource Europe S.A. (Nivelles, Belgium) with a lower detection limit of 5 mIU/liter. Serum C peptide was measured in duplicate by RIA using a rabbit-derived antibody (Milab, Malmö, Sweden) and a C-peptide tracer (BioSource Europe S.A.) with a lower detection limit of 0.03 nmol/liter. Serum IGFBP-1 was analyzed in duplicate by RIA using a rabbit-derived antibody and a human tracer purified from amniotic fluid with a lower detection limit of 0.25 ng/ml (both kindly provided by Professor R. Bouillon). As a marker and mediator of insulin sensitivity, serum adiponectin was determined by a time-resolved immunofluorometric assay based on reagents from R&D Systems, Inc. (Minneapolis, MN) (11).
Gene expression of IGFBP-1 and adiponectin receptors (AdipoRs)
mRNA from muscle and liver biopsies was isolated with RNAeasy Mini columns (QIAGEN, Inc., Valencia, CA). Total RNA (3 µg for liver, 1 µg for muscle) was reverse transcribed in a final volume of 20 µl, as described previously (12). Reactions lacking reverse transcriptase were run to generate controls for assessment of genomic DNA contamination. For hepatic IGFBP-1 expression (NM_000596), forward primer 5'-CAGGAGACATCAGGAGAAGAAATTT-3', reverse primer 5'-TCCCGCCTCTCCATCCAT-3', and FAM-labeled TaqMan probe 5'-TTACCTGCCAAACTGCAACAAGAATGGATT-3' were used. For AdipoR 1 (NM_015999), forward primer 5'-GCCAACCCACCCAAAGCT-3', reverse primer 5'-CCGCACCTCCTCCTCTTCT-3', and FAM-labeled TaqMan probe 5'-AAGAAGAGCAAACATGCCCAGTGCCC-3' were used. For AdipoR 2 (NM_024551), forward primer 5'-TTTGTGGTTGCTGGAGCTTTT-3', reverse primer 5'-CCCCGCCGATCATGAA-3', and FAM-labeled TaqMan probe 5'-TTCACTTCCATGGTGTCTCAAACCTCCA-3' were used. cDNA was quantified in real time with the ABI PRISM 7700 sequence detector (Applied Biosystems, Foster City, CA) as described previously (12). Gene expression data were expressed as a ratio of HPRT (forward primer 5'-TGTAGATTTTATCAGACTGAAGAGCTATTGT-3', reverse primer 5'-AAGGAAAGCAAAGTCTGCATTGTT-3', TaqMan probe 5'-TTTCCAGTTAAAGTTGAGAGATCATCTCCACCAAT-3'). Individual samples with a copy number coefficient of variation greater than 20% were reanalyzed. Data are expressed as a fold increase of the mean of the control patients.
Protein levels of insulin signal transduction molecules in tissue biopsies
Tissue samples were homogenized in a buffer containing 20 mM Tris-HCl (pH 7.6), 10% glycerol, 1% NP-40, 2 µg/ml aprotinin, 5 µg/ml leupeptin, 0.5 µg/ml pepstatin, 10 mM sodium orthovanadate, 34 µg/ml phenylmethylsulfonyl fluoride, 10 mM sodium pyrophosphate, 100 mM sodium fluoride, and 10 mM EDTA. The homogenates were centrifuged for 5 min at 10,000 rpm and 4 C. The protein content in the supernatant was determined with Coomassie Protein Assay Reagent (Pierce, Rockford, IL). Homogenates were separated by denaturating SDS-PAGE and immunoblotted with specific Ab against the insulin receptor (IR) (Upstate Biotechnology, Waltham, MA), phosphoTyr1147-IR, p44/42-MAPK, phosphoThr202/Tyr204-p44/42-MAPK, Akt, phosphoThr308, and Ser473-Akt (both Ab gave the same results), phosphoThr180/Tyr182-p38 MAPK (Cell Signaling Technology, Inc., Danvers, MA), and with a species-specific horseradish peroxidase-conjugated secondary Ab (DakoCytomation, Glostrup, Denmark). In immunoprecipitation experiments, 0.5–1 mg protein was incubated overnight at 4 C with 5 µg anti-IRS-1 or anti-IRS-2 (Upstate Biotechnology), followed by incubation with 50 µl protein A-Sepharose beads. The immunoprecipitates were washed, boiled in sodium dodecyl sulfate sample buffer, separated by denaturating SDS-PAGE, and immunoblotted with an Ab against the 85 kDa subunit of PI3K (Upstate Biotechnology) and with a species-specific horseradish peroxidase-conjugated secondary Ab (DakoCytomation). Immunoprecipitation with 5 µg anti-SHC Ab (Upstate Biotechnology) was followed by a Western blot with a specific anti-Grb2 (Upstate Biotechnology). All blots were analyzed using Image Master Software (Amersham Biosciences, Piscataway, NJ). The amount of phosphorylated p44/42-MAPK in muscle samples was quantified with a commercially available ELISA (Cell Signaling Technology, Inc.). We used 200 µg muscle total protein per 100 µl in the assay, which was developed according to the manufacturer’s instructions. To normalize the data, each data point was divided by the mean value obtained for the CIT patient group.
Statistical analysis
The results obtained from different groups of patients were compared by the 
2 test for comparison of proportions, and ANOVA and post hoc Fisher’s projected least-significant difference for normally distributed continuous data (presented as mean and SEM). We used the unpaired Student’s t test for comparisons of normally distributed data and Mann-Whitney U test for data that were not normally distributed [presented as medians and interquartile range (IQR)]. Statistical significance was considered when two-sided P values were 
0.05. StatView 5.0.1 (SAS Institute Inc., Cary, NC) was used for these statistical analyses.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Table 1
shows the baseline characteristics and outcome of the studied patients with prolonged critical illness (ICU stay of at least 1 wk; randomized for CIT or IIT). Table 2 shows the baseline characteristics and outcome of patients from whom tissue biopsies were available. As a consequence of the better survival with IIT, the patients who died in this treatment group were more severely ill on admission to the ICU, evidenced by a higher acute physiology and chronic health evaluation (APACHE) II score than those in the CIT group, and they died somewhat earlier (Table 2).
Critically ill patients were hyperglycemic on admission (Fig. 1A). According to the study protocol, in all critically ill patients on IIT, normoglycemia was maintained from admission until the last day in the ICU. To reach this normoglycemia, the insulin doses administered to the IIT group were several times higher than in the CIT group (Fig. 1B). On the admission day, nonsurviving patients had higher blood glucose levels compared with surviving patients (147.1 ± 3.0 mg/dl for survivors and 163.3 ± 9.6 mg/dl for nonsurvivors; P = 0.04). Once insulin treatment was started, blood glucose levels were comparable between surviving and nonsurviving patients at all times (Fig. 1A).
|
Critically ill patients were hyperinsulinemic on admission and throughout their ICU stay (Fig. 1C). On the second day of IIT, serum insulin levels were only 54% higher than after CIT. On the seventh and on the last day in the ICU, serum insulin levels in the IIT group were equal to those in the CIT group, in spite of much lower blood glucose levels and severalfold higher amounts of insulin administered. Insulin serum profiles were comparable between surviving and nonsurviving patients (Fig. 1C).
Critically ill patients had elevated levels of C peptide on admission (Fig. 1D). Already on d 2, serum C-peptide concentrations were severalfold higher in the CIT group than in the IIT group. On d 7 and the last day of the ICU stay, IIT reduced C-peptide levels further to values in the healthy reference range. C-peptide profiles were comparable between surviving and nonsurviving patients, although nonsurviving CIT patients had higher C-peptide levels during their ICU stay than surviving CIT patients (Fig. 1D; P < 0.001 for d 2, d 7, and last day).
Elevated serum IGFBP-1 levels at admission decreased thereafter, regardless of the insulin treatment (Fig. 1E). IGFBP-1 serum levels were higher in nonsurviving patients than in surviving patients, already on the day of admission to the ICU throughout day of discharge or day of death (Fig. 1E). Similar to IGFBP-1 serum levels, IGFBP-1 mRNA expression was not affected by IIT [median (IQR) relative mRNA content 3.6 (0.8–5.7) in CIT patients and 3.6 (2–7.8) in IIT patients] but was clearly increased in critically ill patients [3.6 (0.9–6.2)] compared with control values [1.1 (0.6–2.0); P = 0.002].
Adiponectin levels were decreased on admission compared with healthy reference values (Fig. 1F). After 7 d in the ICU, IIT critically ill patients had higher levels of adiponectin than CIT critically ill patients. Also on the last day in the ICU, adiponectin levels were higher in the IIT group than the CIT group. Adiponectin serum profiles were comparable between surviving and nonsurviving patients (Fig. 1F). In those patients in whom we studied insulin signaling (n = 36), adiponectin levels were significantly (P = 0.03) elevated in IIT patients (mean ± SEM 20.2 ± 4.3 µg/ml), as compared with CIT patients (9.4 ± 1.9 µg/ml).
AdipoR expression
In both liver and muscle, expression of the AdipoR 1 was up-regulated in critical illness (Fig. 2). In liver, IIT further increased AdipoR1 expression as compared with CIT. Expression of the AdipoR2 was also up-regulated in critical illness, in both liver and muscle, but unaffected by IIT (Fig. 2).
|
Insulin therapy did not influence Tyr1147-phosphorylated IR levels (Fig. 3A), nor did it change total IR levels (P = 0.4) in muscle [IIT 0.9 (0.5–1.3), CIT 1.2 (0.5–2.9)].
|
) and pAkt (Fig. 3D) but did not alter the association of IRS-2 and PI3K (Fig. 3C). Total IRS-1, IRS-2, and Akt levels in muscle were comparable among prolonged critically ill patients (data not shown).
No difference was found between IIT and CIT for the association of Grb2 and Shc (Fig. 3E) or the phosphorylation of p44/42MAPK (Fig. 3F). Insulin therapy did not alter the amount of phosphorylated p38 MAPK [P > 0.9; IIT 0.9 (IQR 0.7–1.2), CIT 0.9 (IQR 0.7–1.2)]. Total muscle Grb2 and p44/42 MAPK levels were equal among prolonged critically ill patients (data not shown).
There was no correlation between the amount of phosphorylated signaling molecules and time lag between death and freezing of the tissue sample or with the length of insulin therapy (=ICU stay) (Table 2).
Insulin signaling in liver
Insulin therapy did not influence the amount of Tyr1147-phosphorylated IR (Fig. 4A) or total IR levels (P = 0.8) in liver [IIT 1.2 (1.1–1.8), CIT 1.2 (1.0–2.5)].
|
), the association of IRS-2 and PI3K (Fig. 4C), or the amount of pAkt in liver (Fig. 4D). Total IRS-1, IRS-2, and Akt levels in muscle were comparable among prolonged critically ill patients (data not shown).
We found no difference between IIT and CIT for the association of Grb2+Shc (Fig. 4E), or the phosphorylation of p44/42MAPK (Fig. 4F). Insulin therapy did not alter the amount of phosphorylated-p38 MAPK [P = 0.7; CIT 0.84 (IQR 0.72–1.25), IIT 1.13 (IQR 0.76–1.60)]. Total liver Grb2 and p44/42 MAPK levels were equal among prolonged critically ill patients (data not shown).
There was no correlation between the amount of phosphorylated signaling molecules and time lag between death and freezing of the tissue sample or with the length of insulin therapy (=ICU stay) (Table 2).
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Hyperglycemia and elevated insulin levels are long-known features of insulin resistance that is present during illness and after trauma and is associated with severity of illness (13, 14, 15, 16, 17, 18, 19). It was striking to observe that in critically ill patients, who are endogenously hyperinsulinemic, severalfold higher insulin doses than what is done conventionally, continued for at least a few days, evoked sustained normoglycemia without further elevating circulating insulin levels. Identical circulating insulin levels in the face of substantially lower blood glucose levels suggest amelioration of insulin sensitivity with IIT. This may be explained by the prevention of hyperglycemia and/or dyslipidemia, which is known to exert deleterious effects on insulin sensitivity (20, 21, 22, 23, 24, 25). The higher serum adiponectin levels in the IIT group corroborate improved insulin sensitivity. Indeed, adiponectin is a protein hormone produced and secreted exclusively by adipocytes, which increases the tissue response to insulin (26). The underlying molecular mechanism of the insulin-sensitizing effect of adiponectin remains largely unknown, but it was shown recently that adiponectin enhances IRS-1/Akt signaling (27). We indeed observed increased IRS/Akt signaling in the muscle of IIT-treated critically ill patients. However, in the liver, IIT and the concomitantly elevated circulating adiponectin levels did not increase IRS/Akt signaling, and did not alter circulating IGFBP-1 and liver IGFBP-1 mRNA levels, although the major AdipoR isoform was overexpressed. It has been demonstrated previously that the expression levels of the AdipoRs (AdipoR1 and AdipoR2) are associated with insulin secretion and sensitivity (28, 29). This suggested that the metabolic pathway in liver may be resistant or impaired at a level distal to the AdipoR during prolonged critical illness. Alternatively, the lowering of the elevated serum C-peptide levels by IIT, indicating a lowering of endogenous insulin release, may theoretically have counteracted any positive effect of the higher adiponectin levels on the insulin signal in the liver (30).
Elevated C-peptide levels in CIT patients, which were normalized by IIT, indeed suggest reduced endogenous insulin production by the intervention with exogenous insulin. However, a stimulation of insulin clearance by IIT cannot be excluded. From the unaltered serum IGFBP-1 levels, one might conclude that hepatic insulin clearance was not affected by IIT because increased hepatic insulin internalization would have decreased IGFBP-1 expression. However, because muscle and kidney are other major sites of insulin removal (31), IIT may have increased insulin clearance by improving kidney function and inducing more insulin internalization in skeletal muscle of the critically ill (7).
Circulating C-peptide levels were not only dramatically higher in CIT patients compared with IIT patients but were also almost twice as high in nonsurviving CIT patients compared with surviving CIT patients. It has been demonstrated that C peptide accumulates in the vessel wall of diabetics where it colocalizes with macrophages in early arteriosclerotic lesions (32, 33). The protective effect of IIT on the endothelium that we demonstrated previously in the same cohort of patients might have benefited from the lowered circulating C-peptide levels (12).
A higher metabolic insulin signal at the molecular level was documented in muscle, but not in liver. The increased metabolic insulin signal in muscle is in line with our previous finding of increased expression of hexokinase-II and glucose transporter 4 mRNA in muscle (34), and a trend for a higher muscle protein content (35). In further agreement with this observation, an anabolic insulin effect on skeletal muscle was suggested by the observed reduction of ventilator dependency (36). In the liver, metabolic insulin signaling was unaffected by insulin therapy, in line with our previous observation of unaltered phosphoenolpyruvate carboxykinase mRNA and glucokinase mRNA in liver biopsies (34). In addition, hepatic IGFBP-1 gene expression and circulating IGFBP-1 were not affected by insulin therapy.
The metabolic insulin signal transduction pathway in the muscle did, whereas in the liver did not, respond to IIT. Nevertheless, the liver is an organ that is being protected by IIT, with improved function and preserved mitochondrial ultrastructure and function (35, 37). Thus, this constellation of findings fits with the concept of protection of glucose toxicity with IIT in those organ systems that take up glucose independent of insulin, rather than organ protection being directly mediated via an insulin signal. The increased insulin signal in skeletal muscle in turn suggests increased insulin-mediated glucose uptake, which, in view of the large mass of skeletal muscle, may have contributed to the clearing of glucose from the circulation, whereby reducing toxic side effects of high circulating levels of glucose on other cellular systems.
The mitogenic pathway of insulin signaling in the liver and skeletal muscle of prolonged critically ill patients was unaffected by IIT. This is in contrast to what has been reported in cellular models of type 2 diabetes and obesity (38, 39). Montagnani et al. (40) demonstrated in cultured human umbilical vein endothelial cells that blocking the IRS/PI3K pathway with wortmannin in a high-insulin environment causes enhanced activation of the MAPK pathway. Therefore, the absence of such an effect in the liver and muscle of critically ill patients is reassuring because mitogenic complications of hyperinsulinemia, which in ICU patients with cancer could be particularly problematic, are thus likely not brought about by IIT. The association between high insulin doses and adverse outcome of critical illness is, therefore, likely explained by more severe insulin resistance in patients with the highest risk of death and, thus, casual rather than causal. Higher circulating levels of insulin and IGFBP-1 in nonsurvivors as compared with survivors, but both unaffected by insulin therapy, further supports such a casual association.
An important limitation of our study needs to be highlighted. The tissue biopsies of critically ill patients were obtained from nonsurvivors only. This may have caused a bias in the results of the insulin signal transduction analyses induced by postmortem changes. However, there was no correlation between the results of the insulin signal transduction and time delay between death and tissue harvesting. Furthermore, and perhaps more importantly, IIT exerted equal actions in both surviving and nonsurviving patients on blood glucose levels, circulating insulin, adiponectin, and IGFBP-1 levels, which allows to minimize the risk of overinterpretation of the insulin signaling results obtained from tissue analysis from nonsurvivors only.
In conclusion, IIT maintained normal blood glucose levels during prolonged critical illness by only transiently elevating circulating insulin levels, suggesting improved overall insulin sensitivity after a few days of IIT. An increase in circulating adiponectin levels may have played a role. Skeletal muscle, but not liver, revealed an activated metabolic insulin signal transduction pathway. IIT did not impose an increased mitogenic risk in either of these tissues. Therefore, the association between high insulin doses and adverse outcome of critical illness is likely explained by more severe insulin resistance in patients with the highest risk of death and, thus, casual rather than causal.
|
Footnotes |
|---|
First Published Online July 31, 2007
Abbreviations: AdipoR, Adiponectin receptor; APACHE, acute physiology and chronic health evaluation; BMI, body mass index; CIT, conventional insulin therapy; ICU, intensive care unit; IGFBP, IGF binding protein; IIT, intensive insulin therapy; IQR, interquartile range; IR, insulin receptor; IRS, IR substrate.
Received April 11, 2007.
Accepted July 25, 2007.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
. Endocrinology 145:5168–5176This article has been cited by other articles:
 |
|
 |
|
  P. Dandona, A. Chaudhuri, H. Ghanim, and P. Mohanty Insulin as an anti-inflammatory and antiatherogenic modulator. J. Am. Coll. Cardiol., February 3, 2009; 53(5 Suppl): S14 - S20. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
1 wk)
, P 
