| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Endocrine Care |
Diabetes and Metabolism Research Program, Garvan Institute of Medical Research, and Department of Endocrinology, St. Vincent’s Hospital, Darlinghurst, Sydney, NSW 2010, Australia
Address all correspondence and requests for reprints to: Prof. D. J. Chisholm, Head, Diabetes and Metabolism Research Program, Garvan Institute of Medical Research, 384 Victoria Street, Darlinghurst, Sydney, NSW 2010, Australia. E-mail: . d.chisholm{at}garvan.org.au
Abstract
The increased cardiovascular risk in type 1 diabetes may be related, at least in part, to insulin resistance. The aim of this study was to assess the relationships between insulin sensitivity, abdominal fat, androgens, lipids, and blood pressure in 10 premenopausal women with type 1 diabetes (mean ± SD, hemoglobin A1c 8.1 ± 1.0%) and 10 nondiabetic body mass index-matched controls. Insulin sensitivity (glucose infusion rate during euglycemic-hyperinsulinemic clamp) was significantly less in the type 1 diabetes group than in controls (49.3 ± 14.8 vs. 73.2 ± 21.6 µmol/min·kg fat free mass, respectively, P = 0.01). The two groups were similar with respect to lipids, androgens, energy expenditure, physical activity, blood pressure, and abdominal adiposity (intra-abdominal fat by four-slice computed tomography and central abdominal fat by dual-energy x-ray absorptiometry). There were no relationships between glucose infusion rate, abdominal adiposity, and androgen levels in subjects with type 1 diabetes, in contrast to controls. Our results demonstrate greater insulin resistance in a group of premenopausal women with type 1 diabetes compared with nondiabetic controls, unrelated to abdominal adiposity, lipids, or androgens.
CARDIOVASCULAR DISEASE IS a common cause of morbidity and mortality in type 1 diabetes (1). Relative to the general population, cardiovascular mortality is increased approximately 4-fold, further magnified in patients with albuminuria (1, 2, 3). In women, the relative cardiovascular mortality is at least twice that of men (1) and the premenopausal cardio-protection is lost (4). In type 1 diabetes, the presence of traditional cardiovascular risk factors may not entirely explain this excess cardiovascular risk (4).
Although insulin deficiency is the primary metabolic defect in type 1 diabetes, a number of studies (5, 6, 7, 8, 9, 10, 11, 12, 13, 14) suggest insulin resistance is also a prominent feature and may, at least in part, contribute to the high rates of vascular disease in this group (15, 16). Insulin resistance is, to some extent, genetically inherited (17) and, in nondiabetic and type 2 diabetic populations, is strongly related to abdominal fat, independent of total adiposity (18, 19, 20). In type 1 diabetes, exogenous insulin administration sufficient to achieve adequate portal levels and maintain euglycemia produces relative systemic hyperinsulinemia. It has been proposed that hyperinsulinemia may contribute to abdominal fat deposition in type 1 diabetes (21). Such is possible because insulin has been shown to increase the activity of 11ß-hydroxysteroid dehydrogenase, especially in omental adipocytes (22), favoring a hypercortisolemic milieu and enhanced differentiation of adipose stromal cells to adipocytes, promoting abdominal obesity (22).
It is not known whether abdominal fat is a predictor of insulin resistance, dyslipidemia, and hypertension in individuals with type 1 diabetes. We postulated that systemic hyperinsulinemia may lead to the accumulation of abdominal fat and, therefore, insulin resistance in type 1 diabetes, thereby contributing to the increased cardiovascular risk. In addition, hyperandrogenemia has been shown to be associated with abdominal adiposity, increased insulin response to oral glucose, and reduced levels of high-density lipoprotein (HDL)-cholesterol in premenopausal nondiabetic women (23). The aims of the study, therefore, were to determine, using gold standard measures, whether type 1 diabetes was associated with increased abdominal adiposity in premenopausal women and to assess the relationships between abdominal adiposity, insulin resistance, androgens, and cardiovascular risk factors, specifically lipids and blood pressure (BP).
Subjects and Methods
Recruitment and eligibility criteria
Ten premenopausal women with a 4-yr or more history of type 1 diabetes mellitus were recruited via advertisements in an outpatient diabetes clinic at a teaching hospital and in an educational magazine. The control group consisted of 10 healthy, body mass index (BMI)-matched, premenopausal, nondiabetic women, recruited by advertisements in the same hospital and on recommendation from participants already enrolled. Exclusion criteria included age less than 18 yr, hemoglobin A1c (HbA1c) higher than 9.5%, current smoking, pregnancy, and use of the oral contraceptive. All women were studied in the follicular phase of their menstrual cycle, determined by menstrual history. Details recorded included family history of diabetes and the presence of microalbuminuria; nine of the 10 type 1 subjects were historically free of microalbuminuria. The study protocol was approved by the Research and Ethics Committee at St. Vincent’s Hospital, and informed consent was obtained from each participant.
Baseline assessment, anthropometry, dietary, and physical activity questionnaire
Subjects attended at 0830 h after a 10-h fast. They were advised not to exercise strenuously nor consume alcohol for 24 h before the study. Subjects with type 1 diabetes omitted their usual morning insulin dose. A medical history was obtained on all subjects. BP was measured in the supine position after a 5-min rest using a mercury sphygmomanometer. Weight was measured to the nearest 0.1 kg with the subject in a hospital gown, and height was measured to the nearest 0.01 m by stadiometer with the subject barefoot. BMI was expressed in kilograms per meters2. The average of triplicate measurements of waist (narrowest circumference between the lowest aspect of ribs and anterior superior iliac crests) and hip (widest circumference between the anterior superior iliac crests and greater trochanters) circumferences (to the nearest 0.01 m) were taken. Waist to hip ratio (WHR) was calculated. Triplicate measurements of skin-fold thickness (biceps, triceps, subscapular, suprailiac, and abdominal) were performed by the same experienced operator (John Bull, British Indicators, Bedfordshire, UK) using callipers. Dietary composition and energy intake were measured by a research nutritionist, using a detailed dietary history based on 24-h recall (24). Energy intake and dietary composition were quantified using a standard food composition program (FoodWorks; Xyris Software, Brisbane, Australia). Physical activity (PA) was assessed using a standardized validated questionnaire (Framingham Physical Activity Index) (25), which assessed the time spent performing various activities at home, at work, and in leisure in an average week, to calculate the average weekly energy expenditure (EE) in metabolic units.
Biochemical analysis
Blood samples were collected via a retrograde iv cannula. Venous blood from the sampling hand was "arterialized" throughout the study using a warming device placed over the arm and hand. Baseline fasting blood samples were obtained in all subjects. Whole blood glucose was measured by the Nova 14 Analyzer (Nova Biomedical, Waltham, MA) and the YSI 2300 Stat Plus analyzer (Yellow Springs Instruments, Yellow Springs, OH). Free insulin was extracted using a polyethylene glycol solution and quantified by RIA (Linco Research, Inc., St. Charles, MO). Total and HDL-cholesterol were measured by adaptation of the cholesterol oxidase/peroxidase method; triglycerides (TGs) by the lipase/glycerol kinase method (Sigma, St. Louis, MO); nonesterified fatty acids (NEFAs) by an acyl-CoA oxidase-based calorimetric kit (Wako, Osaka, Japan); and apolipoprotein B (apo B) by Immuno Turbidometric automated analyzer (Hitachi 917; Roche Diagnostics, Indianapolis, IN). Total T was assayed using the ADVIA Centaur System (Bayer Corp., Walpole, MA). SHBG was quantified by an immunometric assay using the IMMULITE Analyzer (Diagnostic Products, Los Angeles, CA). The free androgen index (FAI) was calculated as total T/SHBG x 100.
Euglycemic-hyperinsulinemic clamp
Cannulae were inserted into each antecubital vein; one for infusion of insulin and 25% dextrose and the other for retrograde blood sampling (see above). Neutral insulin (Actrapid HM, Novo Nordisk, Copenhagen, Denmark) was added to 500 ml Hemaccel (Hoechst Marion Roussel, Inc., Auckland, New Zealand) and infused at 50 mU/m2·min for 150 min in control subjects and for 120 min (after a variable period of insulin infusion to lower the blood glucose to 5 mmol/liter) in subjects with diabetes. Similar (mean ± SD) free insulin levels were obtained in subjects with type 1 diabetes and in controls (49.0 ± 12.0 and 56.5 ± 19.0 µU/ml, respectively, 30 min before completion of the insulin infusion; 46.9 ± 13.9 and 49.0 ± 20.3 µU/ml respectively just before completion of the insulin infusion). A variable-rate 25% glucose infusion was adjusted according to 10-min results to maintain blood glucose levels at 5 mmol/liter. The steady state glucose infusion rate (GIR) is an estimate of whole-body insulin sensitivity and was calculated as the mean of the last five recorded rates of glucose infusion (last 25 min of the clamp study).
Calorimetry
Continuous indirect calorimetry (Deltatrac Metabolic Monitor; Datex Instrumentarium, Helsinki, Finland) was performed for two 25-min periods. The first was in the fasted state before commencement of the insulin infusion, following a 15–20 min rest period with the subject in the recumbent position. The second period was the last 25 min of the clamp study. An equilibrium period of 10 min was allowed (data excluded from calculations), after which gas exchange rates were recorded at minutely intervals; EE and respiratory quotient (RQ) (ratio of carbon dioxide production to oxygen consumption) were then calculated (26). Subjects were discouraged from talking, coughing, sneezing, or sleeping for the duration of this period.
Intra-abdominal fat (IAF)
Four-slice abdominal computed tomography (CT) scanning was performed on a General Electric Sytec 3000 scanner (Tokyo, Japan). A computed radiograph was taken to determine the cranio-caudal level of each CT slice. All subjects were examined in the supine position. The uppermost CT slice was taken through the middle of the L2 vertebral body, and the lowermost cut corresponded to the midpoint of L4. The two middle slices were equally spaced between these two set-points. Each CT cut was 10 mm in thickness, obtained at 120 kV (peak) and 450 mAs, with a scanning time of 3 sec during arrested expiration.
The values calculated at each level included IAF area, IAF area expressed as a percentage of the total cross-sectional intra-abdominal area (%IAF), sc abdominal fat (SAF) area, IAF/SAF area, and the maximal sc fat thickness over the rectus sheath. The intra-abdominal cavity was traced manually, being defined as the region enclosed by the inner aspect of the abdominal wall and the anterior margin of the vertebral body. The intra-abdominal area within this boundary was calculated by the computer. The viscera were traced manually, and the IAF area was ascertained by subtracting the summed area of the viscera from the intra-abdominal area. SAF was calculated by the computer and defined as the area enclosed by the outermost aspect of the abdominal muscle wall and the skin surface. All measurements were performed by the same observer who was blinded to the subject category.
Body composition
Body composition was measured by dual-energy x-ray absorptiometry (DXA) (Lunar Corp., Madison, WI). Total body fat (TBF) was expressed in kilograms and as a percentage of body weight (%TBF). The fat content of a central abdominal window, extending from the upper border of L2 to the lower border of L4 and laterally to the inner aspect of the ribs, as described previously (18, 27), was measured. This region has been shown to contain a relatively high proportion of intra-abdominal fat (28). Central abdominal fat (CAF) was expressed in kilograms and as a percentage of the total tissue contained within this window (%CAF).
Statistical analysis
Data are presented as mean ± SD. Statistical analyses were performed using Statview II. Statistical comparisons between controls and subjects with type 1 diabetes were made using ANOVA for normally distributed variables. For non-normally distributed variables (fasting glucose, insulin, total cholesterol, TG, T, FAI, PA, RQ, IAF and IAF/SAF), nonparametric comparisons were made using the Mann-Whitney test. Relationships between normally distributed variables were examined using simple regressions and expressed as Pearson’s correlation coefficients (r). Non-normally distributed variables were log-transformed in regression analyses.
Results
The metabolic and body composition characteristics of subjects with type 1 diabetes and controls are shown in Table 1
. In subjects with type 1 diabetes, mean ± SD HbA1c was 8.1 ± 1.0%, total daily insulin dose was 45 ± 10 U, and duration of diabetes was 23.8 ± 9.9 yr (range, 4.5–34). Subjects with type 1 diabetes were slightly older than controls (38.8 ± 6.5 vs. 31.1 ± 2.8 yr, respectively, P = 0.003), but the groups were matched for BMI and were similar in regard to weight, WHR, PA scores (Table 1), dietary intake, and skin-fold measurements (data not shown).
|
). Similarly, there were no differences in abdominal fat measured by CT (IAF, %IAF, SAF, maximal SAF thickness, IAF/SAF area) (Table 1). Systolic and diastolic BPs were similar in type 1 diabetic subjects and controls (Table 1), as were resting EE and RQ (Table 2). There were no differences in fasting total and HDL-cholesterol, TG, apo B, NEFA, total T, SHBG, or FAI between the groups (Table 2).
|
The GIR was not associated with age, lipids, or BP in either subgroup, nor HbA1c in the type 1 subjects (data not shown). Fasting blood glucose did not predict GIR in either type 1 subjects (r = 0.32, P = 0.37) or controls (r = -0.06, P = 0.86). Although there was an inverse relationship between fasting free insulin and GIR in the total cohort (r = -0.50, P = 0.03), it was not significant when type 1 subjects (r = 0.44, P = 0.20) and controls (R = -0.51, P = 0.13) were examined separately. In type 1 volunteers, GIR was not predicted by IAF area (r = -0.03, P = 0.93) or %CAF (r = 0.17, P = 0.64). In controls, GIR was inversely related to IAF area (r = -0.65, P = 0.04), but the relationship with %CAF did not reach statistical significance (r = -0.56, P = 0.09).
There was a strong correlation between mean IAF area (by CT) and %CAF (by DXA) in both type 1 (r = 0.91, P = 0.0006) and control (r = 0.83, P = 0.003) subjects. In type 1 subjects, there was a significant relationship between IAF and systolic BP (r = 0.77, P = 0.01), which was not evident in controls (r = 0.18, P = 0.61). In the type 1 cohort, neither IAF nor %CAF was predicted by the duration of diabetes (data not shown). Although there was no relationship between IAF area and lipids in type 1 subjects, increasing IAF predicted higher TGs (r = 0.68, P = 0.03) and apo B (r = 0.92, P = 0.0002) and lower HDL-cholesterol levels (r = -0.69, P = 0.03) in control subjects.
In subjects with type 1 diabetes there were no relationships between serum androgens and %TBF, %CAF, or IAF area (data not shown). However, in controls, low plasma SHBG levels and a high FAI predicted increased total and abdominal adiposity (Table 3). In control subjects only, there was a strong inverse relationship between total T and GIR (Table 3). In subjects with type 1 diabetes there were no relationships between plasma androgens and fasting glucose, free insulin levels, lipids (total cholesterol, HDL-cholesterol, TG, NEFA, apo B), or BP (data not shown). In contrast, significant associations were found in controls between T and total cholesterol (r = 0.68, P = 0.03) and NEFAs (r = 0.80, P = 0.006) and between FAI and NEFAs (r = 0.63, P = 0.05) and apo B levels (r = 0.67, P = 0.03).
|
Cardiovascular disease is a significant cause of death in type 1 diabetes, with the strongest predictor being proteinuria (1, 2). Even in the absence of proteinuria, however, cardiovascular risk is still far greater than that in nondiabetic populations (2). This led us to investigate other potential contributors to cardiovascular disease in type 1 diabetes, namely abdominal obesity and insulin resistance.
In the current study, we found that a cohort of 10 premenopausal women with type 1 diabetes were more insulin resistance than 10 BMI-matched premenopausal nondiabetic controls. The finding of a 33% lower GIR in the cohort of type 1 patients is in accordance with the 30–64% reduction reported by others (29, 30). Despite this difference in GIR, there was no significant difference between the groups in any measure of abdominal adiposity (which actually tended to be lower in the diabetic group), rejecting the original hypothesis, but confirming the results of one of the few reports that has used DXA to assess body fat distribution in type 1 diabetes (31). Importantly, there was also no difference between the type 1 and control groups in total T, SHBG, or FAI.
Despite the inverse relationship between IAF and GIR in the control group, there was no relationship between adiposity and GIR in the type 1 diabetic subjects. It is conceivable that failure to detect such relationships was a consequence of small numbers, although the findings in the control arm of the study would favor the validity of the negative results. To our knowledge, the relationship between body fat distribution, using CT or DXA, and insulin sensitivity, measured by the hyperinsulinemic-euglycemic clamp, has not been previously examined in type 1 diabetes. Although an inverse relationship has been demonstrated between GIR and both waist circumference and WHR in a study of 24 volunteers with type 1 diabetes, these relationships were not apparent after controlling BMI (32).
In type 1 diabetic subjects, serum androgens did not predict abdominal adiposity or GIR. In controls, however, there was an inverse relationship between SHBG and total and abdominal adiposity, as found in a recent study of nondiabetic, premenopausal women (23). Consistent with that study, we found that FAI, but not total T, was also significantly related to adiposity in controls (23). An additional finding in our study was the inverse relationship between total T and GIR in the control subjects.
Insulin resistance has previously been shown to be an important predictor of vascular disease in type 1 diabetes in an 18-yr prospective analysis, using the glucose assimilation index (a surrogate estimate of insulin sensitivity) (16). Hypertension was the only other predictor of mortality in that study, although this was not significant after correcting for the glucose assimilation index (16). In our study, although systolic BP and its variance were similar in both type 1 subjects and controls, systolic BP was significantly predicted by IAF in those with type 1 diabetes. It is possible that abdominal adiposity may be a greater contributor to BP regulation in premenopausal women with type 1 diabetes than nondiabetic women, and small differences in the amount of this fat depot may influence cardiovascular risk adversely via its effect on BP.
A number of factors may contribute to insulin resistance in type 1 diabetes. It is possible that the metabolic milieu of type 1 diabetes per se is associated with a defect in peripheral glucose disposal (33), specifically due to reduced glucose transport in skeletal muscle (34). Reduction in insulin binding, and inhibition of binding by insulin antibodies, seems not to play a part (8, 28). Elevated hepatic glucose output due to impaired hepatic insulin sensitivity may be present in poorly controlled patients and contribute to insulin resistance, although this defect is reversed with adequate insulin therapy and improved glycemic control (33). Evidence that poor glycemic control can contribute to insulin resistance has been demonstrated in studies of patients with type 2 diabetes, in which improvements in hyperglycemia are associated with improvements in insulin-mediated glucose uptake (35). In type 1 diabetes, studies of newly diagnosed hyperglycemic patients, both with (9) and without (33) acidosis, demonstrate significant insulin resistance using clamp techniques. Furthermore, an inverse relationship between glycemic control (as determined by HbA1c) and insulin sensitivity (estimated by GIR) has been demonstrated in other studies (8, 28).
The clinical impact of amelioration of insulin resistance in type 1 diabetes (and the most appropriate means to achieve this) is not known. A small study has demonstrated that glucose disposal rates can be improved by 60% and insulin requirements reduced with a controlled physical training program, in insulin pump-treated type 1 diabetic patients (36). In addition, metformin has been shown to improve glucose uptake by 18% in a small study of subjects with type 1 diabetes (37).
Because the number of subjects in this study is small, it is important to exercise care in both positive and negative conclusions. Men have approximately double the amount of intra-abdominal fat compared with premenopausal women, so that a difference in abdominal fat to explain a similar cardiovascular disease risk in type 1 diabetic women to that of men, would have to be large. The results for CAF percentage (which we have previously shown to be the measure most closely associated with insulin sensitivity and lipid abnormalities) indicate that this study had a 97% probability of detecting a 50% increase in type 1 diabetic women. In comparison with previous data (18) a 33% reduction in insulin sensitivity, as seen in this study, if explained by increased abdominal fat, would require an approximate 30% increase in this parameter. The confidence intervals for the diabetic subjects for percentage of CAF would indicate a 95% confidence that this variable in the diabetic subjects was not more than 9% above the mean for controls. Thus, it is reasonable to conclude that premenopausal type 1 diabetic women do not have an increase in abdominal fat likely to explain their impaired insulin sensitivity or increased cardiovascular risk. The results also suggest that the usual relationships between body fat distribution and metabolic risk factors associated with cardiovascular risk are lost in premenopausal type 1 diabetic women, but further study is required to clarify this issue.
In conclusion, we have demonstrated that insulin resistance is a prominent feature in premenopausal women with type 1 diabetes and is not associated with an abdominal pattern of fat distribution or hyperandrogenemia. Our study suggests that the usual relationships between metabolic factors associated with cardiovascular risk and body fat distribution are lost in type 1 diabetes, except for systolic BP. Whether insulin resistance contributes to cardiovascular disease in type 1 diabetes, and whether improvements in insulin sensitivity modify cardiovascular risk in this population, awaits further study. The use of metformin for this purpose, and whether the newer pharmacological insulin- sensitizers, the thiazolidinediones, have a role to play in this respect, has not been examined.
Acknowledgments
We gratefully acknowledge the financial support from the R. T. Hall Trust and the National Health and Medical Research Council of Australia and also acknowledge the assistance of Elizabeth Maclean (Research Nutritionist), Bronwyn Heinrich and Maria Males (Clinic Research Nurses), Lynn Croft (Laboratory Manager), Joanna Edema (Technical Officer), Dr. Judith Freund, Nicole Culton and Melanie Hoy (Department of Nuclear Medicine, St. Vincent’s Hospital/Clinic), and Prof. Bruce Doust and David Joscelyne (St. Vincent’s Private Hospital Medical Imaging/Darlinghurst x-ray).
Footnotes
This work was supported by the National Health and Medical Research Council of Australia and a grant from the R. T. Hall Trust.
Abbreviations: apo B, Apolipoprotein B; BMI, body mass index; BP, blood pressure; CAF, central abdominal fat; CT, computed tomography; DXA, dual-energy x-ray absorptiometry; EE, energy expenditure; FAI, free androgen index; FFM, fat free mass; GIR, glucose infusion rate; HbA1c, hemoglobin A1c; HDL, high-density lipoprotein; IAF, intra-abdominal fat; NEFA, nonesterified fatty acid; PA, physical activity; RQ, respiratory quotient; SAF, sc abdominal fat; TBF, total body fat; TG, tryglyceride; WHR, waist to hip ratio.
Received August 1, 2001.
Accepted November 26, 2001.
References
This article has been cited by other articles:
 |
|
 |
|
  N. A. Sherry, E. B. Tsai, and K. C. Herold Natural History of {beta}-Cell Function in Type 1 Diabetes Diabetes, December 1, 2005; 54(suppl_2): S32 - S39. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. M. Strowig and P. Raskin The Effect of Rosiglitazone on Overweight Subjects With Type 1 Diabetes Diabetes Care, July 1, 2005; 28(7): 1562 - 1567. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Czupryniak, J. Strzelczyk, K. Cypryk, M. Pawlowski, D. Szymanski, A. Lewinski, and J. Loba Gastric Bypass Surgery in Severely Obese Type 1 Diabetic Patients Diabetes Care, October 1, 2004; 27(10): 2561 - 2562. [Full Text] [PDF]
|
|
|
|
|
|
A. A.F. Sima, W. Zhang, Z.-G. Li, Y. Murakawa, and C. R. Pierson Molecular Alterations Underlie Nodal and Paranodal Degeneration in Type 1 Diabetic Neuropathy and Are Prevented by C-Peptide Diabetes, June 1, 2004; 53(6): 1556 - 1563. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Perseghin, G. Lattuada, M. Danna, L. P. Sereni, P. Maffi, F. De Cobelli, A. Battezzati, A. Secchi, A. Del Maschio, and L. Luzi Insulin resistance, intramyocellular lipid content, and plasma adiponectin in patients with type 1 diabetes Am J Physiol Endocrinol Metab, December 1, 2003; 285(6): E1174 - E1181. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. R. Greenfield, K. Samaras, A. B. Jenkins, P. J. Kelly, T. D. Spector, and L. V. Campbell Moderate Alcohol Consumption, Dietary Fat Composition, and Abdominal Obesity in Women: Evidence for Gene-Environment Interaction J. Clin. Endocrinol. Metab., November 1, 2003; 88(11): 5381 - 5386. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. R. Greenfield, K. Samaras, L. V. Campbell, and D. J. Chisholm Type 1 Diabetes Is Not Associated With Increased Central Abdominal Obesity: Response to Sobel Diabetes Care, September 1, 2003; 26(9): 2703 - 2703. [Full Text] [PDF]
|
|
|
|
|
|
B. E. Sobel Is the Fat Still in the Fire?: Response to Greenfield et al. Diabetes Care, September 1, 2003; 26(9): 2703 - 2704. [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 |
