| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Division of Endocrinology, Diabetes, and Hypertension (H.V.J., C.R., K.F., G.K.A.) and Cardiovascular Division (R.Y.K., M.D.G.-H.), Department of Medicine, Brigham and Women’s Hospital, Boston, Massachusetts 02115
Address all correspondence and requests for reprints to: Gail K. Adler, M.D., Ph.D., Division of Endocrinology, Diabetes, and Hypertension, Brigham and Women’s Hospital, 221 Longwood Avenue, Boston, Massachusetts 02115. E-mail: gadler{at}partners.org.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Objective: Our objective was to assess whether mineralocorticoid receptor activation contributes to coronary circulatory dysfunction in patients with diabetes who are already receiving ACE inhibitor therapy.
Design and Setting: A randomized, double-blind, crossover study with an intervening washout period of at least 4 wk was conducted with ambulatory patients from the community.
Patients: Patients included 16 subjects (11 men, eight Caucasians; mean age, 53 yr; mean body mass index, 38.0 kg/m2) with diabetes and albuminuria but without clinical cardiovascular disease.
Interventions: ACE inhibitors were switched to enalapril 20 mg daily, and other antihypertensives were discontinued. Amlodipine 5–10 mg daily was added to achieve blood pressures less than 130/80 mm Hg. Subjects then received, in random order, 6 wk of the mineralocorticoid receptor antagonist eplerenone 50 mg (with placebo pill) daily and 6 wk of another diuretic, hydrochlorothiazide 12.5 mg (with potassium 10 mEq) daily.
Main Outcome Measures: Before and after each 6-wk treatment period, we measured coronary circulatory function (adenosine-stimulated myocardial perfusion reserve) and endothelial function (brachial artery reactivity and peripheral arterial tonometry).
Results: The eplerenone and hydrochlorothiazide groups had similar blood pressures, serum potassium, glycemia, and endothelial function. Although pretreatment myocardial perfusion reserve did not differ between groups, myocardial perfusion reserve was significantly higher after eplerenone than after hydrochlorothiazide (median 1.57 vs. 1.30; P = 0.03).
Conclusions: Mineralocorticoid receptor blockade improves coronary circulatory function compared with hydrochlorothiazide in patients with diabetes already receiving ACE inhibitor therapy.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Furthermore, ACE inhibitors do not chronically suppress aldosterone concentrations (8, 9, 10). In patients with heart failure already receiving ACE inhibitor therapy, the addition of an MR antagonist substantially reduces cardiac morbidity and mortality (11, 12). Similarly, combination therapy with an ACE inhibitor and MR antagonist more effectively reduces LV hypertrophy (13) and diabetic albuminuria (14) than ACE inhibitor monotherapy despite similar reductions in blood pressure between the combination therapy and monotherapy groups.
Based on these observations, we hypothesized that MR activation contributes to coronary circulatory dysfunction in patients with diabetes already receiving ACE inhibitor therapy.
Therefore, we performed a randomized, double-blind, crossover, proof-of-concept study comparing the effects of eplerenone, an MR antagonist, vs. another diuretic, hydrochlorothiazide (HCTZ) on vascular function in subjects with diabetes already receiving ACE inhibitor therapy. We used cardiac magnetic resonance imaging (CMR) to assess changes in adenosine-stimulated myocardial perfusion reserve (coronary circulatory function), proteinuria to assess renovascular function, and brachial artery reactivity and peripheral arterial tonometry (PAT) to assess endothelial function in peripheral blood vessels.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
We recruited 21- to 64-yr-old subjects with diabetes. To increase the likelihood that these subjects would have coronary circulatory dysfunction, we required the presence of albuminuria (albumin-to-creatinine ratio 
30 mg/g on two occasions). Exclusion criteria included serum creatinine more than 1.5 mg/dl (>132.6 µmol/liter), clinical evidence of cardiac or cerebrovascular disease, severe peripheral vascular disease, poorly controlled hypertension, contraindications to CMR or adenosine, cigarette smoking, and pregnancy. The Institutional Review Board at Brigham and Women’s Hospital (Boston, MA) approved the study protocol, and all subjects provided written, informed consent. Studies were performed in the General Clinical Research Center.
Study protocol
This randomized, double-blind, crossover study compared the effects of 6 wk of eplerenone 50 mg plus placebo pill daily vs. HCTZ 12.5 mg plus potassium chloride 10 mEq daily on measures of vascular function in subjects with diabetes receiving enalapril 20 mg daily. Subjects not on ACE inhibitor or angiotensin receptor blocker (ARB) therapy at enrollment were started on enalapril 20 mg daily 3 months before the first vascular studies (decreases in aldosterone concentrations after initiation of ACE inhibitor therapy usually reverse within 3 months) (10). All other subjects were switched to enalapril 20 mg daily 4 wk before the first vascular studies. Other antihypertensives were discontinued 2–4 wk before the first vascular studies and amlodipine 5–10 mg daily was initiated for blood pressures higher than 130/80 mm Hg (15). During the run-in period, we adjusted hypoglycemic agents to achieve fasting and preprandial blood glucose to less than 150 mg/dl (<8.3 mmol/liter; approximates hemoglobin A1C of <7%) (16).
We monitored potassium (baseline, d 2, and wk 1, 2, 4, and 6) and blood pressure (baseline and wk 2, 4, and 5.5) during each 6-wk treatment period. Blood pressure was measured using a manual cuff (three readings five min apart) after the subject was seated for 20 min. After the first 6-wk treatment period, subjects entered a washout period of at least 4 wk (median, 34 d; range, 28–95 d due to scheduling conflicts) during which time study medications were discontinued except for enalapril and amlodipine. Subjects then received the alternate treatment for 6 wk.
Vascular studies were performed at the beginning and end of each 6-wk treatment period. For 3 d before vascular testing, subjects consumed an isocaloric diet (50–60% of calories from carbohydrates) containing 200 ± 10 mmol/d sodium, 100 ± 10 mmol/d potassium, 1000 ± 50 mg/d calcium, and 300 ± 50 mg/d magnesium. A 24-h urine sample was collected on each vascular study day for volume, sodium, potassium, protein, and creatinine. Subjects withheld amlodipine for 48 h, avoided caffeine for 24 h, and fasted for at least 8 h before testing. Subjects withheld oral hypoglycemic agents, fast-acting insulin (except for one subject with type 1 diabetes using an insulin pump), long-acting insulin (except for the two other subjects with type 1 diabetes), antioxidants, and multivitamins on the morning of the vascular studies.
On each vascular study day, we obtained two Dynamap blood pressure measurements 10 min apart after subjects were supine for 10 min. An indwelling iv catheter was placed, and subjects lay supine for an additional 45 min before collecting blood for chemistries, hemoglobin A1C, fructosamine, serum aldosterone, and plasma renin activity. Brachial artery reactivity and PAT testing were performed first, followed by CMR. Blood glucose measurements were obtained before ultrasound and CMR testing. No subject required fast-acting insulin for preprocedure hyperglycemia.
Myocardial perfusion reserve
We used CMR to quantify myocardial blood flow (MBF) at rest and during maximal vasodilation with adenosine. We calculated myocardial perfusion reserve by dividing MBF during vasodilation by MBF at rest (17, 18, 19). CMR quantification of MBF correlates well with radioisotope-labeled microspheres in animals (r = 0.99) and intracoronary Doppler measurements in humans (r = 0.84) (17, 18).
We performed CMR examinations with subjects supine in a 1.5-Tesla scanner (General Electric Healthcare, Milwaukee, WI; Signa CV/i) with electrocardiogram gating and an eight-element phase-array surface coil. We used a saturation-prepared fast gradient echo perfusion technique (TR 6.8, TE 2.1, echo-train 4, matrix 128 x 96; FOV = 32 cm) to capture first-pass MBF kinetics. We acquired MBF data during a breath hold of 20–30 sec. During image acquisition, we infused gadolinium-diethylenetriamine pentaacetic acid (DTPA) (Magnevist; Berlex Pharmaceuticals, Wayne, NJ) via a peripheral vein (0.05 mmol/kg at 5 ml/sec followed by a 20-ml saline flush at 5 ml/sec). One image per cardiac cycle was used to maintain adequate temporal resolution of myocardial perfusion imaging at rest and during vasodilation. Myocardial perfusion imaging covering three short-axis locations were acquired after 4 min of adenosine infusion (140 mg/kg·min) and repeated at rest 15 min later in matching locations.
The dose of gadolinium was chosen based on existing literature (20, 21) and from phantom experiments using our current first-pass perfusion sequence demonstrating signal saturation at gadolinium concentrations of at least 4 mmol/liter. In large animal experiments, a peak change of approximately 15 sec–1 in the T1 relaxation rate (
R1, where R1 = 1/T1) of the LV cavity occurred during a first-pass bolus of 0.05 mmol/kg gadolinium (21). With a gadolinium-DTPA r1 relaxivity of 4.45 mmol–1 sec–1, we calculated a peak contrast concentration of 3.37 mmol/liter in the LV blood. In human experiments, this concentration is expected to be slightly lower due to a higher volume of first-pass distribution in the LV and contrast extraction from the pulmonary circulation. Therefore, a 0.05 mmol/kg iv gadolinium-DTPA bolus injection is appropriate for blood flow assessment in the LV cavity.
An investigator blinded to treatment assignment used dedicated software (CineTool, GE Healthcare, version 3.9.1) to analyze CMR data according to validated methodology (19, 22). LV endocardial and epicardial borders were manually mapped as circles. Careful positioning of the tracing at the endocardial border excluded high signal from the blood, which would falsely elevate MBF results. LV myocardium confined between endocardial and epicardial circles was divided into three radial segments by the centerline method. These segments represent the region of myocardium supplied by the major coronary arteries (left anterior descending, left circumflex, and right coronary arteries) according to the American College of Cardiology/American Heart Association myocardial segment model (23). These regions of interest were copied and pasted into all images across different time points to capture MBF in each segment during first-pass transit of the contrast bolus. Image motion due to breathing was manually corrected using landmarks such as endocardial borders to ensure proper alignment of the myocardial segments and the regions of interest. We used a validated Fermi-function deconvolution model (19) to quantify regional and global MBF in milliliters per minute per gram of LV mass. For all experimental data (including the arterial input function curve), the mean value of the baseline signal before appearance of the contrast agent in the LV blood pool was subtracted from all signal vs. time curves. We did not correct for surface coil intensity because the myocardial perfusion reserve calculation canceled out the heterogeneous effect from surface coil placement. The slight delay between myocardial and LV arterial input signal-time curves was adjusted by shifting the LV arterial input function to the right.
Brachial artery reactivity and peripheral arterial tonometry
The brachial artery was imaged in the longitudinal plane with high-resolution ultrasonography (Toshiba Power Vision 8000) equipped with a 7.5-MHz linear array probe using well-validated techniques (24). Pulse wave amplitude was measured with a finger plethysmograph (Itamar-Medical Ltd., Caesarea, Israel), which quantifies pulsatile blood volume changes (25).
Studies were performed in a quiet, temperature-controlled room after subjects rested supine for 10 min. Arterial inflow was occluded to induce reactive hyperemia by inflating a blood pressure cuff placed on the upper arm to at least 50 mm Hg above systolic pressure for 5 min. Images obtained 1 min after cuff release corresponded to the period of maximal endothelium-dependent vasodilation. Ten or more minutes later, rest measures were repeated, and subjects with systolic blood pressure higher than 100 mm Hg and heart rate more than 60 beats/min received 0.3 mg sublingual nitroglycerin. The brachial artery was reimaged 3 min later.
Two independent investigators blinded to subject characteristics analyzed the brachial arterial diameter with edge detection software (Medical Imaging Applications, Coralville, IA). Flow-mediated (endothelium-dependent) vasodilation (FMD) is the percent change in arterial diameter from baseline to reactive hyperemia. Nitroglycerin-mediated (endothelium-independent) vasodilation is the percent change in arterial diameter from second baseline to post-nitroglycerin administration. PAT hyperemia ratio is defined as the ratio of pulse wave amplitude during reactive hyperemia to baseline. PAT data from the study arm were normalized to data obtained from the control arm.
Blood and urine assays
Serum aldosterone was measured with the solid-phase RIA Coat-A-Count procedure (Diagnostic Products Corp., Los Angeles, CA). Plasma renin activity was assayed with the GammaCoat 125I RIA kit (DiaSorin, Stillwater, MN) that measures the generation of angiotensin I. Hemoglobin A1C was measured by HPLC using the Tosoh G7 analyzer (Tosoh Medics Inc., San Francisco, CA). The fructosamine assay was performed on a Roche Modular Analytics P (Hitachi, Japan). Urinary sodium levels were determined with flame photometry with lithium as an internal standard (Nova Biomedical, Waltham, MA). Urinary creatinine levels were assayed with the ACE creatinine reagent (Alfa Wasserman, West Caldwell, NJ). Urine creatinine, protein, and microalbumin were assayed on an Olympus analyzer (Olympus America Inc., Melville, NY).
Statistical analyses
The primary endpoint for our study was the posttreatment analysis of total LV myocardial perfusion reserve. During the design of our protocol, we performed CMR power calculations based on data from a report of myocardial perfusion reserve in patients with microvascular dysfunction without significant coronary lesions (17). These patients had a mean myocardial perfusion reserve of 2.14 (SD 0.57). Assuming a conservative improvement of 25% and 
= 0.05, we determined that 11 subjects would be needed for 80% power.
The power analysis calculations for brachial artery reactivity studies have been extensively reviewed (26). For a crossover trial, 10 subjects are required for 80% power at = 0.05, assuming a conservative improvement of 2.5%.
We describe the continuous data using means with SD or medians with interquartile ranges. Variables collected during the two treatment periods were compared with the Wilcoxon signed rank test (continuous data) and Fisher’s exact test (proportions). All analyses were performed using Stata 8.0 (StataCorp, College Station, TX), and a two-sided P value of 
0.05 was considered significant.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Sixteen subjects completed the study protocol (Table 1
). CMR data were incomplete in four subjects due to technical issues (n = 3) or MRI intolerance (n = 1). Nitroglycerin-mediated vasodilation was not performed in one subject because of low-normal blood pressures. PAT was performed in 11 subjects (PAT was introduced several months after study initiation). Five additional subjects withdrew from the study before the second treatment period and are not included in any of the analyses; three withdrew due to other commitments, and two (both received HCTZ) withdrew due to medical issues (recurrent depression and new-onset atrial fibrillation).
|
Eight subjects received eplerenone as the first treatment. Eleven subjects required amlodipine (10 mg in seven subjects and 5 mg in four subjects) in addition to enalapril for blood pressure control. There were no differences between the eplerenone and HCTZ groups with respect to pretreatment and posttreatment body mass index, serum sodium, serum creatinine, total cholesterol, high-density lipoprotein cholesterol, and urinary sodium, and neither therapy significantly altered any of these characteristics (Table 2). Posttreatment serum triglycerides were lower in the eplerenone group than in the HCTZ group (Table 2).
|
Blood pressures trended downwards over both treatment periods [with eplerenone, systolic decreased from 130 ± 16 to 123 ± 11 mm Hg (P = 0.09) and diastolic from 76 ± 9 to 70 ± 8 mm Hg (P = 0.03); with HCTZ, systolic decreased from 132 ± 22 to 125 ± 19 mm Hg (P = 0.23) and diastolic from 76 ± 11 to 73 ± 9 mm Hg (P = 0.27)]. However, blood pressures measured during eplerenone treatment did not differ from those obtained at corresponding time points during HCTZ treatment (Table 2
).
Serum potassium
Serum potassium concentrations were similar during HCTZ and eplerenone treatment (Table 2). Nine serum potassium measurements exceeded 5.0 mmol/liter (five with eplerenone and four with HCTZ). Two values exceeded 5.5 mmol/liter (5.7 and 5.6 mmol/liter with eplerenone and HCTZ, respectively); both were less than 5.5 mmol/liter on repeat. There was no dose reduction, discontinuation, or unblinding because of hyperkalemia.
Glycemic control
On vascular testing days, we measured hemoglobin A1C, fructosamine, and preprocedure blood glucose to assess chronic, subacute, and acute glycemic control, respectively. None of these measures differed between the eplerenone and HCTZ groups (Table 2).
Proteinuria
There were no differences between the eplerenone and HCTZ groups with respect to pretreatment and posttreatment 24-h proteinuria, and neither therapy significantly altered proteinuria (Table 2).
Vascular function
Adenosine-stimulated myocardial perfusion reserve.
None of the patients had CMR perfusion defects suggestive of occlusive coronary artery disease (CAD). Pretreatment myocardial perfusion reserve was similar before initiation of eplerenone and HCTZ (P = 0.96). After treatment, total LV myocardial perfusion reserve was significantly higher with eplerenone than with HCTZ (median 1.57 vs. 1.30; P = 0.03) (Table 3 and Fig. 1). Nine (75%) subjects had higher myocardial perfusion reserve with eplerenone than with HCTZ (Fig. 2). These results retain statistical significance even after excluding the one potential outlier with a post-eplerenone myocardial perfusion reserve of 4.31 (P = 0.05). When compared with the first study day values, there was a trend (P = 0.09) toward improvement in total LV myocardial perfusion reserve with eplerenone and no change (P = 0.72) with HCTZ. Similar results were obtained when the three coronary territories were evaluated separately (Table 3).
|
|
|
). Furthermore, there were no statistically significant correlations between myocardial perfusion reserve and any of the peripheral measures of vascular function on any of the four study days (P = 0.10–0.89). |
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
This finding is most likely related to study medication because the two treatments resulted in similar blood pressures, serum potassium concentrations, and glycemic control. We chose HCTZ as the active control because this medication is considered inexpensive, first-line antihypertensive therapy for most patients (27). Furthermore, HCTZ has no known direct effects on the structure of small arteries or forearm vasoreactivity (28, 29).
In a cross-sectional study of adults without clinical CAD, the mean myocardial perfusion reserve assessed by CMR was 3.08 in subjects without diabetes and 2.82 in subjects with diabetes (30). The lower myocardial perfusion reserve in our subjects may be related to more severe microvascular disease (our subjects were required to have albuminuria, a marker of microvascular injury). None of our subjects had evidence of occlusive CAD based on clinical history and absence of perfusion defects during stress CMR. Of note, two subjects achieved myocardial perfusion reserve in the nondiabetic range after treatment with eplerenone.
A few studies have evaluated the effects of medications on coronary vasomotor function in patients with diabetes. Neither pravastatin nor pioglitazone improved coronary circulatory function in this patient population (31, 32). In normotensive subjects with diabetes and LV hypertrophy, iv perindoprilat (ACE inhibitor) acutely improved dipyridamole-stimulated myocardial perfusion reserve (6). Our study extends these results. We demonstrated beneficial effects of MR blockade on coronary circulatory function in subjects with diabetes above and beyond potential effects of ACE inhibitor therapy.
Data from human (11, 12, 13, 33, 34) and animal (35) studies support adverse cardiovascular effects of aldosterone and suggest a beneficial effect of MR blockade. In healthy volunteers, a 4-h aldosterone infusion induces endothelial dysfunction (34). In heart failure, combination MR antagonist and ACE inhibitor therapy improves endothelial function within 4 wk; long-term coadministration improves cardiac morbidity and mortality (11, 12, 33). In essential hypertension, combination MR antagonist and ACE inhibitor therapy more effectively reduces LV hypertrophy than monotherapy with either agent alone, effects not fully explained by changes in blood pressure (13).
In contrast to the beneficial effect of eplerenone on myocardial perfusion reserve, we did not detect changes in endothelial function using brachial artery reactivity or PAT despite more than 80% power for detecting a conservative improvement of 2.5% with the former technique (26). Therefore, in patients with type 2 diabetes and albuminuria, endothelial function in the forearm does not appear to reflect endothelial function in the coronary bed. Similar discrepancies between myocardial perfusion reserve and peripheral endothelial function (using different techniques) have been reported in various patient populations (36, 37, 38). In the current study, these disparate results may reflect a preferential effect of MR blockade on the coronary circulation. Alternatively, these findings may be related to differences in the mechanism of vasodilation between CMR and the peripheral tests (adenosine has direct effects on vascular smooth muscle; FMD involves shear stress-mediated production of endothelium-derived nitric oxide).
Our brachial artery reactivity results are not consistent with a report showing impaired endothelial function with spironolactone in patients with type 2 diabetes (7). However, spironolactone was administered for only 4 wk, and there was worsening glycemic control, which may have confounded the results.
We did not detect an effect of our interventions on proteinuria. In contrast, a few studies (14, 39) have shown that an MR antagonist plus ACE inhibitor more effectively reduces proteinuria than either therapy alone in patients with diabetic nephropathy. However, these studies enrolled patients with greater proteinuria, used higher doses of the MR antagonist, or treated patients longer. Perhaps different vascular beds have different sensitivities or response times to MR blockade. Any of these factors in addition to a potential lack of statistical power may explain our proteinuria results.
Although there is a risk of hyperkalemia with combination ACE inhibitor and MR antagonist therapy (40), none of our patients required dose reduction or study discontinuation because of hyperkalemia. The absence of significant hyperkalemia is likely due to the low dose of eplerenone and the discriminating selection of patients with relatively preserved renal function. Epstein et al. (14) also noted no increase in the risk of hyperkalemia with the addition of eplerenone 50 mg daily to enalapril 20 mg daily in patients with type 2 diabetes and good renal function.
Strengths of our study include the randomized, double-blind design with control of confounders (blood pressure, potassium, and glycemia) and the use of each subject as his/her own control. Potential limitations include the short duration of treatment and carryover of medication across treatment periods, although a formal statistical test for carryover was not significant. Because HCTZ has minimal, if any, direct effects on artery structure and function (28, 29), carryover effects from eplerenone into the HCTZ treatment period would be expected to improve the HCTZ results, which would minimize observed differences between groups, and potentially bias our results toward the null hypothesis. We did not assess vascular effects of eplerenone in the absence of background ACE inhibitor therapy or in individuals with clinical evidence of CAD, and the sample size limits conclusions across specific subpopulations. However, there were improvements in myocardial perfusion reserve with eplerenone in men and women, in patients with type 1 and type 2 diabetes, and in those who did and did not use statins. Our study was not designed to assess whether the vascular findings were driven mostly by improvements in vessel function, structure, or both. Finally, our findings could be due to differences in triglyceride concentrations between the two treatments, although this is unlikely because serum triglycerides did not correlate with myocardial perfusion reserve (data not shown).
In conclusion, we have shown that compared with HCTZ, short-term treatment with eplerenone improves coronary circulatory function in patients with diabetes and albuminuria who are already receiving ACE inhibitor therapy. This pilot, proof-of-concept study suggests that the MR is involved in the pathophysiology of impaired myocardial perfusion reserve in diabetes and that MR blockade is a better choice than HCTZ for improving coronary circulatory function in these patients. Future studies should confirm and extend these preliminary findings.
|
Footnotes |
|---|
Disclosure Summary: The authors have nothing to disclose.
First Published Online May 8, 2007
Abbreviations: ACE, Angiotensin converting enzyme; ARB, angiotensin receptor blocker; CAD, coronary artery disease; CMR, cardiac magnetic resonance imaging; DTPA, diethylenetriamine pentaacetic acid; FMD, flow-mediated vasodilation; HCTZ, hydrochlorothiazide; LV, left ventricular; MBF, myocardial blood flow; MR, mineralocorticoid receptor; PAT, peripheral arterial tonometry.
Received February 20, 2007.
Accepted April 27, 2007.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
agonist on myocardial blood flow in type 2 diabetes. Diabetes Care 28:1145–1150This article has been cited by other articles:
 |
|
 |
|
  C. Guo, V. Ricchiuti, B. Q. Lian, T. M. Yao, P. Coutinho, J. R. Romero, J. Li, G. H. Williams, and G. K. Adler Mineralocorticoid Receptor Blockade Reverses Obesity-Related Changes in Expression of Adiponectin, Peroxisome Proliferator-Activated Receptor-{gamma}, and Proinflammatory Adipokines Circulation, April 29, 2008; 117(17): 2253 - 2261. [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 |
