Relationship between insulin sensitivity, obesity, body fat ... - Nature

1 downloads 0 Views 117KB Size Report
OBJECTIVE: To study the associations of obesity (as body mass index (BMI)), of body fat distribution (as waist to hip ratio (WHR)) and of ββ-endorphinaemia ...
International Journal of Obesity (1998) 22, 143±148 ß 1998 Stockton Press All rights reserved 0307±0565/98 $12.00

Relationship between insulin sensitivity, obesity, body fat distribution and b -endorphinaemia in obese women C Percheron1 , C Colette1 , D Mariano-Goulart2 , A Avignon1 , O Capeyron1 , H Boniface1 , N Bressot2 and L Monnier1 1

Department of Metabolism and 2 Department of Nuclear Medicine, Lapeyronie Hospital, Montpellier, France

OBJECTIVE: To study the associations of obesity (as body mass index (BMI)), of body fat distribution (as waist to hip ratio (WHR)) and of b -endorphinaemia (bb-EP-aemia) with fasting insulin and glucose concentrations, with insulin secretion (as ®rst phase insulin response (FPIR)) and with insulin sensitivity (SI) in obese women. DESIGN: a cross-sectional study of insulin sensitivity in obese women. SUBJECTS: 45 obese women (age: 20±70 y, BMI: 27±50). MEASUREMENTS: Frequently sampled intravenous glucose tolerance test (FSIGTT), FPIR, fasting glucose, fasting insulin, BMI, body fat topography (WHR), b -EP-aemia, plasma ACTH. RESULTS: In univariate analysis the following positive associations were observed: fasting glucose with age and WHR, fasting insulin with BMI and WHR, b -EP plasma concentration with WHR; SI was negatively associated with BMI, WHR and b -EP plasma concentrations. This pattern of associations remained unaltered in multivariate analysis including age, BMI and WHR as independent variables. The contribution of b -EP plasma concentrations to SI variability was corroborated by a stepwise multiple regression analysis: 53.8% of SI variation could be explained by BMI (30.7%), by b -EP plasma concentrations (17.2%) and by WHR (5.9%). Finally, women were divided into two groups according to whether they had a peripheral (P-BFD, WHR  0.80, n ˆ 24) or an abdominal (A-BFD, WHR  0.85, n ˆ 16) body fat distribution. After adjustment for age and BMI, SI values were lower while b -EP and ACTH plasma concentrations were higher in the A-BFD compared to the P-BFD group. In this latter group, 54.8% of SI variation was explained by the same variables as in the whole group. In the A-BFD group, higher WHR was associated with lower FPIR. CONCLUSIONS: 1) The major ®nding of this study is that, in non-diabetic obese women (especially those with a PBFD), higher b -EP plasma concentrations are associated with lower insulin sensitivity. This association is independent of both the magnitude of obesity and the pattern of fat distribution, although these two parameters are strong predictors of SI. 2) The major reduction in SI observed in women with A-BFD probably results from the additive effects of obesity, of elevated b -EP plasma concentrations and of metabolic and endocrine alterations in relation with the central pattern of fat distribution. Keywords: obese women; b-endorphinaemia; glucose metabolism; insulin sensitivity

Introduction Obese patients frequently have peripheral hyperinsulinaemia and insulin resistance, the former compensating for the latter in normo glucose tolerant subjects. Impaired glucose tolerance is observed when insulin resistance is associated with impaired ®rst insulin response and overt diabetes can occur when the defect in insulin secretion is more marked. The question of why all obese subjects do not develop diabetes, remains unanswered. Age,1 duration2 and magnitude3 of the obesity, family history of diabetes mellitus4 and abdominal body fat,5±8 are wellknown factors of risk. The latter has been extensively studied in women. Women with abdominal obesity are characterized by several hormonal and metabolic abnormalities includCorrespondence: C Colette, IURC, 75 rue de la Cardonille, 34093 Montpellier, CeÂdex 5, France. Received 20 June 1997; revised 10 September 1997; accepted 25 September 1997

ing moderate to severe hyperinsulinaemia and insulin resistance,9,10 impaired glucose tolerance,10 altered sex hormone metabolism,11 hyperactivity of the hypothalamic pituitary adrenal axis12 and an increased opioid activity,13,14 which could be part of a more complex neuroendocrine alteration. Several studies have provided evidence that endogenous opioids seem to play a role in the pathogenesis of human obesity15,16 since they are involved in the regulation of food intake17,18 and of insulin secretion.19±21 Furthermore, endogenous opioids might play a physiological role in the regulation of glucose homeostasis by the central nervous system.22 The fact that increased opioid activity is observed only in women with abdominal obesity, strongly suggests a role for bendorphins (b-EP) in the development of insulin resistance and/or hyperinsulinaemia. However most of the studies have only focused on the role of b-EP in modulating insulin secretion. The aim of the present study was to evaluate the associations between obesity, fat distribution, b-EP-aemia and insulin sensitivity in obese women.

b-endorphins and insulin sensitivity in obese women C Percheron et al

144

Materials and methods Subjects

Participants in the study were 45 obese women with a body mass index (BMI) (weight/height2, kg/m2)  27, at risk of developing Type 2 diabetes because of family histories of diabetes and/or obesity, associated or not to impaired glucose tolerance. Three out of the 45 patients were hypertensive. Other endocrine and metabolic abnormalities were excluded. Seven women were post menopausal, four of which were taking estrogen-replacement therapy. None of the remaining subjects was taking medication known to affect carbohydrate metabolism. Body fat distribution was de®ned by the waist (W) to hip (H) ratio (WHR), W circumference being obtained as the minimum value between the iliac crest and the lateral costal margin, and H circumference at the maximum value over the buttocks.23 Twentyfour women with WHR values 0.80 were de®ned as having peripheral body fat distribution (P-BFD); 16 women with WHR values 0.85 were de®ned as having abdominal body fat distribution (A-BFD) and ®ve women with WHR values between 0.80 and 0.85 were classi®ed as having a diffused body fat distribution (D-BFD). Protocol

Frequently sampled intravenous glucose tolerance test (FSIGTT). All tests were conducted between 09.00±12.00 h, after an overnight fast and a 30 min period of recumbency. All subjects underwent a FSIGTT, modi®ed by either an administration of tolbutamide (300 mg) at 20 min (n ˆ 14) or an insulin infusion (10 mU.kg71.min71) from 20±25 min (n ˆ 31). Baseline samples for measurement of plasma glucose and plasma insulin were drawn 15 min (715) before initiation of the glucose injection. Additional samples were drawn for plasma glucose and insulin at 710, 75 and 71 min. An i.v. bolus of 50% glucose (0.3 g/kg) was injected at time 0 (over 1 min exactly), followed by a normal saline infusion and subsequent samples were drawn at 2, 3, 4, 5, 6, 8, 10, 12, 14, 16, 18, 20, 22, 23, 24, 25, 27, 28, 30, 35, 40, 45, 50, 55, 60, 70, 80, 90, 100, 120, 140, 160 and 180 min. Blood collection. Blood samples were collected 15 min before the glucose injection for the measurement of plasma b-EP and adrenocorticotropin (ACTH). The timing of blood sampling was strictly respected because of possible diurnal variations of ACTH and b-EP. Assays

Plasma glucose was measured in duplicate by the hexokinase technique using a Beckman Synchron

CX5CE (Beckman Instruments Inc, Diagnostic systems Group, Brea, CA, USA). Plasma insulin and plasma ACTH were measured in duplicate by RIA (Cis bio international, Gif sur Yvette, France). Blood for b-EP was placed in tubes containing ethylene diaminetetraacetic acid (EDTA) and maintained on ice until storage at 780 . b-EP plasma concentrations were measured by radioimmuno assay (RIA) (Incstar, Stillwater, MN, USA) after extraction by means of af®nity chromatography (using a rabbit antibody coupled to sepharose). The lowest detection limit of the method, de®ned as the apparent concentration at three standard deviations from the counts at maximum binding, was 4 pg/ml. Intraassay and interassay coef®cients of variation were 16.3% and 11.6%, at plasma concentrations of 6 and 9 pg/ml respectively. Data analysis

First phase insulin response (FPIR, mU/ml) was estimated from the incremental plasma insulin above basal from 0±10 min. Insulin sensitivity (SI, 1074 min71 mU71 ml) was estimated by the minimal model approach.24 The Marquardt-Lebenberg method was used for non-linear least square estimation of the parameters, and systems of differential equations were solved using a fourth-order Runge-Kutta integration algorithm.25 The values at 0±8 min were zero weighted as suggested by Pacini and Bergman.25

Statistics Data are expressed as mean  s.d. Anthropometric and metabolic variables were log transformed to improve their skewness and kurtosis where necessary. The following statistical tests were made: Pearson correlation, partial correlation analysis and stepwise multiple linear regression analysis. Comparison between subjects with A-BFD and P-BFD were made by using the t-test for unpaired data, after adjusting for other independent variables. A two-sided P value < 0.05 was considered as statistically signi®cant.

Results Whole group

Table 1 shows the general characteristics of the subjects. Univariate analysis. Table 2 shows simple correlation coef®cients between age, anthropometric variables, b-EP-aemia and metabolic variables. Age and WHR (but not BMI) were equally associated with fasting glucose. Both BMI and WHR were positively

b-endorphins and insulin sensitivity in obese women C Percheron et al Table 1 General characteristics (means  s.d.) (n ˆ 45) who participated in the study Parameters (units)

Mean

s.d.

Age (y) Body mass index (BMI) (kg/m72) Waist hip ratio (WHR) Fasting glucose (mM) Fasting insulin (mU ml71) First phase insulin response (FPIR) (mU ml71) Insulin sensitivity (SI) (1074 min71 mU71 ml) b-EP-aemia (pg ml71) ACTH (pg ml71)

40.3 34.4

12.4 5.6

of

women

Median

Range

38.0 34.0

20±70 27±50

0.81 0.07 0.80 5.5 0.7 5.2 10.7 6.0 9.7 353 277 269

0.65±0.93 3.9±7.1 3.0±32.2 56±1498

3.1

1.8

2.6

0.7±7.7

9.9 19.3

4.7 8.9

8.2 17.0

5.0±29.7 7±39

Conversion factors: SI, 1074 min71 mU71 ml, to 1075min71 pM71, divide by 0.7174; insulin, mU ml71 to pM, multiply by 7.174; bEp, pg ml71 to pM, divide by 3.465; ACTH, pg ml71 to pM, multiply by 0.2202. b-EP-aemia ˆ b-endorphinaemia; ACTH ˆ adrenocorticotropin. Table 2 Simple correlation coef®cients between age, anthropometric variables, b-EP-aemia and metabolic variables Age BMI WHR b-EP-aemia Fasting glucose Fasting insulin FPIR SI

70.144 0.191 0.095 0.434* 70.042 70.244 70.084

BMI

WHR

0.214 0.060 0.404* 0.131 0.436* 0.500** 0.370* 0.033 70.016 70.568*** 70.520**

b-EP-aemia

0.178 0.190 0.020 70.458*

*P  0.01, **P < 0.001, ***P < 0.0001 b-EP-aemia ˆ b-endorphinaemia; BMI ˆ body mass index; WHR ˆ waist to hip ratio; FPIR ˆ ®rst phase insulin response; SI ˆ insulin sensitivity

correlated to fasting insulin. Insulin sensitivity was signi®cantly negatively correlated to b-EP-aemia, to BMI and to WHR. b-EP-aemia was positively correlated to WHR. Multivariate analysis. Partial correlations analysis in regression models were used to explore the independent effects of age, BMI and WHR on the metabolic variables and on b-EP-aemia. The pattern of associations observed in univariate analysis remained unaltered in this multivariate analysis (Table 3). Since SI, in univariate analysis, was both correlated to BMI, to WHR and to b-EP-aemia, we evaluated the relative contribution of these variables to SI values after adjusting every one for the others, in regression models. SI remained signi®cantly associated with BMI (r ˆ 70.487, P < 0.0001), with WHR (r ˆ 70.290, P < 0.015), and with b-EP-aemia (r ˆ 70.311, P ˆ 0.008). Table 4 shows the result for stepwise multiple linear regression for fasting glucose, fasting insulin and insulin sensitivity. Variables are shown in the order of entry. The independent variables that could be entered were age, BMI, WHR and b-EP-aemia. Fasting glucose was signi®cantly positively related to WHR and age. Fasting insulin was signi®cantly positively related to BMI and WHR. Insulin sensitivity

Table 3 Independent effects of age, obesity (as BMI), and body fat distribution (as WHR) on fasting glucose, fasting insulin, ®rst phase insulin response (FPIR), insulin sensitivity (SI) and b-EPaemia Independent variables Dependent variables Fasting glucose Fasting insulin SI FPIR b-EP-aemia

a

Age

BMIa

WHRa

0.385* 70.034 70.078 70.252 0.014

0.114 0.434** 70.494*** 70.011 70.025

0.338* 0.283* 70.399** 70.034 0.407**

a Standardized regression coef®cients. *P < 0.05; **P < 0.01; ***P < 0.0001. BMI ˆ body mass index; WHR ˆ waist to hip ratio; b-EPaemia ˆ b-endorphinaemia. ACTH ˆ adrenocorticotropin.

was signi®cantly negatively correlated to BMI (step 1), to b-EP-aemia (step 2) and to WHR (step 3). This multiple regression ®t very well with 53.8% of the variations in insulin sensitivity explained by three variables and con®rmed that b-EP-aemia was an independent predictor of SI. In order to search if in turn, b-EP-aemia could be in¯uenced by carbohydrate metabolism and by the activity of the hypothalamic pituitary adrenal axis, we used a stepwise multiple linear regression analysis with b-EP-aemia as dependent variable. The possible variables were: age, BMI, WHR, ACTH, fasting glucose, fasting insulin, FPIR and SI. SI was the only signi®cant predictor of b-EP-plasma concentrations (r ˆ 0.458, P ˆ 0.0014). Furthermore, b-EPaemia and ACTH plasma concentrations were similar in women with impaired glucose tolerance and in women with normal glucose tolerance (data not shown).

Groups with different patterns of body fat distribution

The characteristics of the two patient groups are given in Table 5. Patients with A-BFD had lower insulin sensitivity, higher b-EP and ACTH plasma concentrations. BMI and fasting insulin showed a tendency to be higher in the A-BFD group (P ˆ 0.08 and 0.07). In univariate analysis, the following correlations were found it the P-BFD group: Fasting glucose with age, fasting insulin with BMI, FPIR with age, SI with BMI. These associations remained statistically signi®cant after adjusting for the independent effect of age BMI and WHR on the metabolic variables. In the A-BFD group, a negative correlation was found between FPIR and WHR (r ˆ 70.52, P ˆ 0.038). After adjusting for age and BMI the statistic signi®cance fell to 0.08. No correlation was found between b-EP-aemia and WHR neither in the A-BFD nor in the P-BFD group. Using the same stepwise regression analysis as for the entire group, we found that in the P-BFD group (Table 6), 54.8% of the variation in insulin sensitivity could be explained by the same variables as in the entire group: BMI (step 1), b-EP-aemia (step 2) and

145

b-endorphins and insulin sensitivity in obese women C Percheron et al

146

Table 4 Stepwise multiple regression for fasting glucose, fasting insulin and insulin sensitivity (SI) Variable Dependent

Independent

Regression coefficient (standardized)

Fasting glucose

WHR age

0.436 0.364

0.0028 0.0077

Fasting insulin

BMI WHR

0.500 0.275

0.0005 0.0402

70.568 70.425 70.290

0.0001 0.0003 0.0154

SI

BMI b-EP-aemia WHR

R2 (adjusted)

P value

change in R2 (%) 17.1 11.4 28.5 23.2 5.8 29.0 30.7 17.2 5.9 53.8

0.285 0.290

0.538

Independent variables are shown in the order of entry. The possible independent variables were age, BMI, WHR and b-EPaemia.

Table 5 Characteristics of subjects with different patterns of body fat distribution (A-BFD or P-BFD) Parameters (units) Age (y) BMI (kg m72) WHR Fasting glucose (mM) Fasting insulin (mU ml71) SI (1074 min71 mU71 ml) FPIR (mU ml71) b-EP (pg ml71 ACTH (pg ml71)

A-BFD n ˆ16

P*

P-BFD n ˆ 24

40.7  12.3 36.6  6.0 0.88  0.03 5.8  0.7 14.3  7.8 2.0  1.1 429  302 13.0  6.2 23  9

ns 0.08 0.0001 ns 0.07 0.05 ns 0.01 0.05

38.8  12.0 33.3  5.1 0.76  0.04 5.3  0.6 8.4  3.1 3.8  1.7 326  282 8.1  2.2 17  8

*P values for the comparison of A-BFD with P-BFD after adjustment for age and BMI (unpaired t-test). A-BFD ˆ abdominal fat distribution; P-BFD ˆ peripheral fat distribution; BMI ˆ body mass index; WHR ˆ waist to hip ratio; SI ˆ insulin sensitivity; FPIR ˆ ®rst phase insulin response; b-EP ˆ b-endorphin; ACTH ˆ adrenocorticotropin; NS ˆ not statistically signi®cant. Table 6 Stepwise multiple regression for fasting glucose, fasting insulin, ®rst phase insulin response (FPIR) and insulin sensitivity (SI) in patients with peripheral body fat distribution (P-BFD, n ˆ 24) Variable Dependent

Independent

Regression coefficient (standardized)

P value

R2 (adjusted)

Fasting glucose Fasting insulin FPIR SI

Age BMI Age BMI b-EP-aemia WHR

0.539 0.738 70.465 70.552 70.489 70.354

0.007 0.0001 0.022 0.005 0.012 0.024

0.258 0.524 0.180

0.548

change in R2 (%) 25.8 52.4 18.0 27.3 16.8 10.7 54.8



Independent variables are shown in the order of entry. The possible independent variables were age, BMI, WHR and b-EP-aemia.

WHR (step 3). Age was the only predictor of both fasting glucose and FPIR. BMI was the only predictor of fasting insulin concentrations.

Discussion The major ®nding of this study is that, in non-diabetic obese women, higher b-EP plasma concentrations are associated with lower insulin sensitivity, independent of the magnitude of obesity and of body fat distribu-

tion. A possible involvement of b-EP-aemia in modulating insulin sensitivity has been suggested in the past by several interventional studies using opioid receptor antagonists.15 In all studies reported, acute or prolonged treatment with such products reduced basal and glucose stimulated insulin release, but did not change plasma glucose tolerance curves. This seemed to suggest either an improvement of tissue sensitivity to insulin or an effect by the antagonist to dampen the increased responsiveness of the obese to the glucoregulatory effects of opioids. The ®rst hypothesis is at variance with recent data of Vettor

b-endorphins and insulin sensitivity in obese women C Percheron et al

et al26 showing that an oral treatment with naltrexone (an opioid receptor antagonist) was unable to modify insulin sensitivity (measured by the minimal model) in nine obese subjects. However, patients of both genders included in this study are a major dif®culty in interpreting these data. Furthermore, data from our study, in agreement with clinical and epidemiological studies,27 show that obesity and body fat distribution are strong factors associated with the reduction in SI. b-EPaemia predicts only about 17% of SI value, leaving little power to demonstrate the effect of naltrexone treatment in a small number of patients of both genders. It remains that, in our correlation study, a cause-and-effect relationship should not be presumed, and that insulin resistance could very well be responsible for elevated b-EP-aemia. Curiously, in women with a P-BFD, the contribution of b-EP-aemia to SI variation was the same as in the entire group, despite a lower b-EP-aemia. This observation reinforces our conclusion that association between b-EP-aemia and SI is independent of body fat distribution in spite of the strong correlations found between Si and WHR and between b-EP-aemia and WHR. Numerous studies have shown that a central pattern of fat is associated with a greater degree of insulin resistance, and with other metabolic and endocrine alterations in relation with the precipitation of noninsulin dependent diabetes mellitus (NIDDM). The background of this syndrome might be an increased sensitivity or hyperactivity of the hypothalamic-pituitary-adrenal (HPA) axis.28 Such an endocrine perturbation seems to be the link between elevated plasma b-EP concentrations and abdominal body fat distribution, since in normal conditions, at the pituitary level, b-EP are a product of proopiomelanocortin (POMC) processing together with ACTH. The increase in ACTH plasma levels as well as those in b-EP-aemia observed in the A-BFD group could be due to the common origin of these peptides from the anterior pituitary cleavage of the same precursor (POMC), suggesting hyperactivity of the HPA axis. Increased cortisol secretion28,29 and increased opioid activity13,14 have been described in women with a central fat distribution. Why associations between SI, BMI, b-EP-aemia and WHR were lost in the A-BFD group remains questionable. A recent prospective study30 demonstrates that greater insulin resistance and reduced insulin secretion precede visceral fat accumulation in non-diabetic Japanese-American men. Our data seem to indicate that obesity and b-EP could play a major role in the ®rst step of this temporal sequence as suggested by the results of the stepwise regression analysis. This hypothesis is supported by the ®nding that hyper b-EP-aemia is present in normal weight relatives of obese people.31 It is well known that in normo-glucose tolerance obese subjects, insulin resistance is ®rst compensated by increased insulin secretion. Impaired glucose tolerance occurs when insulin resistance is associated to

impaired FPIR, which results in a signi®cant increase in plasma glucose levels and in a late insulin hyperresponsiveness. The negative correlation observed between FPIR and WHR in the A-BFD group suggests that WHR could increase gradually as FPIR decreases. Elevated b-EP plasma concentrations in this group could be a response to the decrease in FPIR, since bEP may have role in islet cell function.21 Basal glycaemia is increased in the group and it has been shown that the prevailing glucose concentration is an important determinant of whether b-EP's effect on insulin release will be inhibitory or stimulatory.32,33 Since diabetes is characterized by FPIR disappearance, this observation copes with the actual idea that abdominal visceral fat is a risk factor for the development of Type 2 diabetes. However, the fact that b-EP-aemia was similar in women with impaired glucose tolerance and in women with normal glucose tolerance is not in favour of such a hypothesis. In summary, we have shown an inverse association between SI and b-EP-aemia in 45 obese women with different patterns of body fat distribution. This association was independent of both obesity and body fat distribution although these two parameters were strong predictors of SI. b-EP plasma concentrations were higher in women with A-BFD who had also higher insulin resistance. In these women, the major reduction in SI probably results from additive effects of obesity, of elevated b-EP plasma concentrations and of metabolic and endocrine alterations in relation to the central pattern of BFD. A common feature between A-BFD and hyper b-EP-aemia could be the hyperactivity of the hypothalamo adrenal axis as already suggested by several studies. In this study, it is suggested (but not proved) that in the temporal sequence leading to Type 2 diabetes, b-EP could play a role in modulating ®rst insulin sensitivity and later insulin secretion. References

1 Ohlson LO, Larsson B, Eriksson H, Svardsudd K, Welin L, Tibblin G. Diabetes mellitus in Swedish middle-aged men. The study of men born in 1913 and 1923. Diabetologia 1987; 30: 386±393. 2 Bierman EL, Bagdale JD, Porte D. Obesity and diabetes: the odd couple. Am J Clin Nutr 1968; 21: 1434±1437. 3 Rimm AA, Werner LH, Van Yseloo B, Bernstein RA. Relationship of obesity and disease in 73,532 weight conscious women. Public Health Rep 1975; 90: 44±51. 4 Morris RD, Rimm DL, Hartz AJ, Kalkhoff RK, Rimm AA. Obesity and heredity in the etiology of non insulin dependent diabetes mellitus in 32,662 adult white women. Am J Epidemiol 1989; 130: 112±121. 5 Vague J. The degree of masculine differentiation of obesities: a fact for determining predisposition to diabetes, atherosclerosis, gout and uric calculous disease. Am J Clin Nutr 1956; 4: 20±34. 6 Feldman R, Sender AJ, Sieglaub AB. Difference in diabetic and non diabetic fat distribution patterned by skin fold measurements. Diabetes 1969; 18: 478±486. 7 Kissebah AH, Vydelingum N, Murray R, Evans DJ, Hartz AJ, Kalhoff RK, Adams PW. Relation of body fat distribution to metabolic complications of obesity. J Clin Endocrinol Metab 1982; 54: 254±260. 8 Peiris AN, Struve MF, Mueller RA, Lee MB, Kissebah AH.

147

b-endorphins and insulin sensitivity in obese women C Percheron et al

148

9

10 11

12 13

14

15 16

17 18 19 20

Glucose metabolism in obesity: In¯uence of body fat distribution. J Clin Endocrinol Metab 1988; 67: 760±767. Bonora E, Del Prato S, Bonadonna RC, Gulli G, Solini A, Shank ML, Ghiatas AA, Lancaster JL, Kilcoyne RF, Alyassin AM, De Fronzo RA. Total body fat content and fat topography are associated differently with in vivo glucose metabolism in nonobese and obese non diabetic women. Diabetes 1992; 41: 1151±1159. Kissebah AH, Peiris AN. Biology of regional fat distribution: relationship to non-insulin dependent diabetes mellitus. Diab Metab Res 1989; 5: 83±109. Kirschner MA, Samojlik E, Drejka M, Szmal E, Schneider G, Ertel N. Androgen-Estrogen metabolism in women with upper body versus lower body obesity. J Clin Endocrinol Metab 1990; 70: 473±479. BjoÈrntorp P. Metabolic implications of body fat distribution. Diabetes Care 1991; 14: 1132±1143. Pasquali R, Cantobelli S, Casimirri F, Bortoluzzi L, Boschi S, Capelli M, Melchionda N, Barbara L. The role of the opioid peptides in the development of hyperinsulinemia in obese women with abdominal body fat distribution. Metabolism 1992; 41: 763±767. Pasquali R, Casimirri F, Cantobelli S, Buratti P, Bortoluzzi L, Capelli M, Labate AMM, Barbara L. b-endorphin response to exogenous corticotrophin-releasing hormone in obese women with different patterns of body fat distribution. Int J Obes 1993; 17: 593±596. Giugliano D, Lefebvre PJ. A role for beta-endorphin in the pathogenesis of human obesity? Horm Metab Res 1991; 23: 251±256. Giugliano D, Cozzolino D, Salvatore T, Ceriello A, Torella R, Franchimont P, Lefebvre PJ, D'Onofrio F. Physiological elevations of plasma b-endorphins alter glucose metabolism in obese, but not in normal-weight subjects. Metabolism 1992; 41: 184±190. Morley JE, Levine AS. The role of endogenous opiates as regulators of appetite. Am J Clin Nutr 1982; 35: 757±761. Reid ID. Endogenous opioid peptides and regulation of feeding and drinking. Am J Clin Nutr 1985; 42: 1099±1132. Vettor R, Martini C, Cestaro S, Manno M, Sicolo N, Federspil G. Possible involvement of endogenous opioids in beta-cell hyperresponsiveness in human obesity. Int J Obes 1989; 13: 425±432. Feldman M, Kiser RS, Unger RH, Li CH Beta-endorphin and the endocrine pancreas. Studies in healthy and diabetic human beings. N Engl J Med 1983; 308: 349±353.

21 Reid RL, Yen SSC. b-endorphin stimulates the secretion of insulin and glucagon in humans. J Clin Endocrinol Metab 1981; 52: 592±594. 22 Ipp E, Garberoglio C, Richter H, Rubenstein HA. Naloxone decreases centrally induced hyperglycemia in dogs. Diabetes 1984; 33: 619±621. 23 World Health Organization. Measuring obesityÐclassi®cation and description of anthropometric data. EUR/ICP/NUT 1250612 V, WHO Regional Of®ce for Europe: Copenhagen, 1988. 24 Bergman RN, Ider YZ, Bowden CR, Cobelli C. Quantitative estimation of insulin sensitivity. Am J Physiol 1979; 236(6): E667±E677. 25 Pacini G, Bergman RN. Minimod: a computer program to calculate insulin sensitivity and pancreatic responsivity from the frequently sampled intravenous glucose tolerance test. Comput Methods Programs Biomed 1986; 23: 113± 122. 26 Vettor R, Macor C, Rossi E, De Palo C, Ruggeri A, Cobelli C, Federspil G. Effect of naltrexone treatment on insulin secretion, insulin action and post prandial thermogenesis in obesity. Horm Metab Res 1994; 26: 188±194. 27 Bray GA. Obesity increases risk for diabetes. Int J Obes 1992; 16, (suppl. 4): S13±S17. 28 BjoÈrntorp P. Regional fat distributionÐImplications for Type 2 diabetes. Int J Obes 1992; 16, (Suppl. 4): S19±S27. 29 Pasquali R, Cantobelli S, Casimiiri F, Capelli M, Bortoluzzi L, Flamia R, Labate AMM, Barbara L. The hypothalamic-pituitary-adrenal axis in the obese women with different patterns of body fat distribution. J Clin Endocrinol Metab 1993; 77: 341± 346. 30 Boyko EJ, Leonetti DL, Bergstrom RN, Newell-Morris L, Fujimoto WY. Low insulin secretion and high fasting insulin and C-peptide levels predict increased visceral adiposity. Diabetes 1996; 45: 1010±1015. 31 Cozzolino D, Sessa G, Salvatore T, Sasso FC, Giugliano D, Lefebvre PJ, Torella R. The involvement of the opioid system in human obesity: a study in normal weight relatives of obese people. J Clin Endocrinol Metab 1996; 81: 713±718. 32 Khawaja XZ, Green IC. Dual action of beta-endorphin on insulin release in genetically obese and lean mice. Peptides 1991; 12: 227±233. 33 Brunani A, Pincelli AI, Pasqualinotto L, Tibaldi A, Baldi G, Scacchi M, Fatti LM, Cavagnini F. In¯uence of insulin on bendorphin plasma levels in obese and normal weight subjects. Int J Obes 1996; 20: 710±714.