Association of poorly controlled diabetes with low serum leptin in morbid obesity

Transcription

1 International Journal of Obesity (1997) 21, 556±561 ß 1997 Stockton Press All rights reserved 0307±0565/97 $12.00 Association of poorly controlled diabetes with low serum leptin in morbid obesity K CleÂment 1,2, N Lahlou 1,4, J Ruiz 1,3, J Hager 2, P BougneÁres 4, A Basdevant 1, B Guy-Grand 1 and P Froguel 2 1 Service de Nutrition, HoÃtel-Dieu, place du Parvis Notre Dame, Paris, France; 2 CNRS EP 10, Institut Pasteur de Lille et C.H.U. Lille, 1 rue Calmette, BP 245, Lille Cedex, France; 3 Division d'endocrinologie et du MeÂtabolisme, DeÂpartement de Medecine Interne, CHUV-BH-19, CH-1011 Lausanne, Suisse; and 4 INSERM U342, HoÃpital Saint Vincent de Paul, avenue Denfert Rochereau, Paris, France OBJECTIVES: Leptin may be involved in the regulation of body weight, food intake, and energy expenditure. In view of a possible link between leptin concentrations and diabetes that has been suggested in obese rodents, we investigated the potential relationship between serum leptin concentrations and hyperglycaemia in French patients with morbid obesity. SUBJECTS: Fasting leptin concentrations were measured in 241 morbidly obese patients with various degrees of glucose tolerance in a cross-sectional study. RESULTS: Fasting serum leptin concentrations did not differ between normoglycaemic (NG, ng/ml) and glucose intolerant morbidly obese subjects (IGT, ng/ml) and were slightly lower in those with controlled diabetes ( ng/ml, P when compared to NG subjects). In contrast, leptin concentrations were 30% lower in patients with poorly controlled diabetes ( ng/ml, P vs NG subjects). Leptin concentrations were negatively correlated with fasting glucose in all groups combined (r , P ) and particularly in NIDDM subjects (r , P ). Although leptin concentrations were higher in women than in men, similar signi cant correlation with fasting glucose was found when females were analyzed separately. A positive correlation was found with BMI (r , P ) in all groups. Multivariate analysis revealed that fasting glucose was independently associated with serum leptin concentrations (F , P ). Sex, age, BMI, waist/hip ratio, fasting glucose and insulin, total cholesterol and triglycerides, tested in the model, explained 42% of the leptin variability in this population. CONCLUSIONS: Poorly controlled diabetes was accompanied by a signi cant reduction of serum leptin concentrations in morbidly obese subjects. We suggest that a relative leptin de ciency (lower than expected for the BMI) associated with insulin de ciency in this population might contribute to a vicious cycle maintaining (or even worsening) obesity itself and/or its metabolic complications. Keywords: leptin; hyperglycaemia; morbid obesity; NIDDM Introduction Correspondence: Dr P Froguel, CNRS EP 10, Institut Pasteur de Lille, 1 rue Calmette, BP 245, Lille Cedex, France. Received 25 November 1996; revised 28 February 1997; accepted 12 March 1997 Leptin, the circulating protein of the obese (ob) gene, may play a key role in body weight regulation. Indeed, several studies have suggested that the transduction of the putative hypothalamic leptin signal might be involved in the control food intake and energy balance. 1±4 In ob/ob mice, obesity has been related to the absence of circulating leptin. However, strong positive correlations between blood leptin concentrations (or ob mrna) and BMI and body fat have been reported in both lean and overweight human subjects. 5±8 Since obese subjects have high serum leptin concentrations, some authors have suggested that these subjects might be resistant to the leptin signal, 9 rather than leptin de cient. The morbid form of obesity is a major risk factor for NIDDM and related metabolic disorders like early hyperinsulinaemia, hyperlipidaemia, and insulin resistance. It is not currently known if the presence of morbid obesity related metabolic abnormalities are correlated to serum leptin concentrations. Studies in leptin treated mice have shown that a large part of the improvement of glucose tolerance was independent of reduced food intake or weight loss, 10 suggesting other neuroendocrine effects of leptin. It is now accepted that insulin increases tissue ob mrna. 11±14 Serum leptin measurements in groups of diabetic and normoglycaemic obese subjects have yielded contrasting results 6,15±19 and the potential in uence of severe hyperglycaemia and insulin de ciency on leptin concentrations are unknown. This prompted us to measure fasting leptin concentrations in a large population of 241 morbidly obese subjects and to investigate the possible relationships between the hormonal concentration and the patient's clinical and metabolic characteristics. Subjects Clinical and biological information was obtained from 241 patients (190 females, 51 males) collected from

2 the service de Nutrition at Hopital HoÃtel-Dieu, in Paris. Fifty-seven of females (30%) were post menopausal. This population with morbid obesity as de ned by a Body Mass Index (BMI) 40 kg/m 2 was originally recruited for the study of the genetic determinants of morbid obesity. 20 Clinical and biological characteristics are presented in Table 1. Their glucose tolerance status was evaluated by a 75 g oral glucose tolerance test. We have grouped this population according to categories de ned by the World Health Organization: Normo Glycaemic subjects (NG, N ˆ 125), Impaired Glucose Tolerance subjects (IGT, N ˆ 35) and non insulin dependent diabetes patients (NIDDM, N ˆ 81) (Table 1). Among the patients with NIDDM, 35 were currently being treated with hypoglycaemic agents (6 with biguanides, 13 with sulphonylurea and 16 with both treatments). Five subjects were receiving insulin treatment. However despite treatments, this NIDDM treated group was characterized by a poorly controlled diabetes (NIDDM-pc, mean fasting glucose 11.5 mm). Forty one other NIDDM subjects were quite well controlled with no treatment since their mean fasting glucose was 8.5 mm (NIDDM-c). Subjects from each glycaemic group did not differ in terms of their total caloric intake and their actual (or maximal reached) BMI (Table 2), and none had recently lost weight, therefore there was no clinical evidence that their body composition had recently uctuated. Methods Evaluation of insulin sensitivity Estimations of insulin sensitivity and beta cell function of the morbidly obese subjects were calculated from plasma glucose and insulin values (kit Pharmacia, Del a), after an overnight fast, by homeostatic model assessment (HOMA/CIGMA software). 21 Homa calculation is based on the assumption that the degree of basal hyperglycaemia is determined by a combination of beta cell de ciency and insulin resistance. Homa is a mathematical model of insulin/ glucose interactions which generates an array of fasting plasma insulin and glucose values that would be expected for combinations of various degrees of beta cell de ciency and insulin resistance. From this array one can estimate combined parameters of insulin sensitivity and beta cell function expected from the fasting plasma glucose and insulin concentrations observed in a given patient. Results are expressed as a percent of insulin sensitivity (HomaS) and of beta cell function (HomaB) in a reference caucasian population of young lean subjects. 21 Radioimmunoassay for serum leptin Fasting leptin concentrations were measured in serum samples by radioimmunoassay using reagents supplied by Linco Research Inc (St Louis, MO, USA). The anti-leptin antiserum was raised in rabbit against highly puri ed recombinant human leptin. This antiserum did not crossreact with insulin, insulin-like growth factors 1 and 2, glucagon or interleukin 2 in doses of 10 mg/ml. The tracer was I 125 human leptin. It was stable for at least two months at 4 C, relative binding at the 0 point decreasing only from 55±47%. Recombinant leptin standard or serum samples in duplicate were incubated in phosphate-buffer saline, ph 7.4, containing 0.12% triton X-100 and 1% bovine serum albumin, with anti leptin antiserum and 125 iodinated leptin (approximately cpm in 100 ml) for 21 h at 4 C. The bound fraction was precipitated with a conjugate of polyethylene-glycol and anti rabbit antiserum and centrifuged. The supernatant was decanted and the pellet counted in the gamma counter with a Ria-calc program. The performance of the standard curve was as follows: ED80: ng, ED50: ng, ED20: ng. The intra-assay coef cients of variation were 3.6 and 3.3% at the concentration 1.4 and 14.2 ng/ml, respectively. The detection limit was computed according to Currie (Analyt Chem 1968; 40: 586± 593): the critical limit (95th percentile of the blank value) was 0.1 ng/ml and the qualitative detection limit (concentration different from the blank at 95% signi cance) was 0.3 ng/ml. Recovery of human leptin added to plasma pool was 105 4% (mean s.d.). Serial dilutions of high concentration samples were strictly parallel to the standard curve. Mean fasting concentration ( s.d.) was 3.6 ng 0.7 ng/ml in nonobese men and 8.8 ng/ml in non-obese menstruating women. Statistical analysis Data are expressed as mean standard deviation, unless stated otherwise. The Shapiro±Wilk W test was used to test the guassian distribution of leptin values and other clinical and biological parameters. Skewed variables were logtransformed for analysis of variance. When ANOVA was signi cant, comparisons between pairs were made using Tukey±Kramer HSD test. Qualitative traits were analyzed by contingency table chi-square tests. Univariate and multivariate logistic regression analysis were used to evaluate interactions between serum leptin concentrations and other clinical and biological variables. Nonparametric Spearman correlations were used to analyze the clinical and metabolic data, which were not normally distributed. Statistics were performed with the JMP software (SAS Institute Inc., Cary, NC). Leptin was logarithmically transformed for multivariate analysis. Results The mean leptin concentrations of the 241 morbidly obese subjects were ng/ml (median 55 ng/ 557

3 558 ml). Although mean BMI values were similar in obese women and men (Table 1), serum leptin concentrations were signi cantly higher in females ( ng/ml in females vs ng/ml in males, P < , Table 1). The concentrations of leptin were similar in pre- and post-menopausal women ( vs , ns, respectively.) As expected, average Homa insulin sensitivity was highest in NG subjects, intermediate in IGT subjects and lowest in subjects with overt diabetes (P ˆ 0.002, age adjusted data, Table 2). HomaB was markedly lower in the poorly controlled diabetic group, suggesting that this latter group had a more severe deterioration of their beta cell function than the other groups (Table 2). Despite a similar BMI in NG, IGT, and controlled NIDDM (NIDDM-c) or poorly controlled NIDDM (NIDDM-PC) patients, the four way analysis of variance showed that the mean serum leptin concentrations decreased signi cantly with the worsening of glucose tolerance (61.5 ng/ml 24.0, ng/ml, ng/ml and ng/ml respectively, P ˆ age adjusted data) (Figure 1). The comparisons between each group (NG, IGT, NIDDM-c and NIDDM-pc) mainly showed differences between non-diabetics and poorly controlled NIDDM subjects (Figure 1). Similar results were obtained when the female group was analyzed separately (mean leptin concentrations: ng/ml in NG females and in NIDDMpc females, P ˆ 0.002). In the small group of males, leptin concentrations tended to be lower in the poorly controlled NIDDM subjects than in NG subjects ( ng/ml in NG and ng/ml in NIDDM-pc men, P ˆ 0.09). Among the poorly controlled NIDDM subjects, serum leptin concentrations were not different between those receiving insulin secretagogue drugs, insulin sensitizing drugs or insulin treatment ( ng/ml in subjects treated by biguanides, ng/ml in subjects treated by sulphonylureas, ng/ml in those receiving both oral treatments and in insulin treated subjects, respectively). When we separated all the morbidly obese subjects into quartiles of glucose, we observed that mean leptin concentrations decreased from the lowest to the highest quartile of glucose (P ˆ , BMI and sex adjusted data, Figure 2). Non parametric analyses, by calculating the Spearman coef cient of correlations, were performed. The compared variables are presented in Table 3. These analyses showed that serum leptin concentrations were negatively correlated with fasting glycaemia in the combined groups (P ˆ ) as well as in the female group (P ˆ 0.005) and a trend was found in the small group of males (Table 2). In addition, positive correlations were found with BMI, hip circumferences and HomaS in the combined group. A negative correlation was also found with fasting triglycerides (P ˆ ). The negative correlation of leptin concentrations with fasting glycaemia was principally seen in both controlled and poorly controlled Table 1 Clinical and biological data in morbidly obese subjects Glucose tolerance status Total HDL Fasting leptin (ng/ml) Triglycerides cholesterol cholesterol NIDDM-pc (M%) NIDDM-c (N%) IGT (N%) NG (N%) N Age (y) BMI (kg/m 2 ) w/h (10±156) 40 (16.6) 41 (17.0) 35 (14.4) 125 (52.0) (0.7±1.4) (39.6±87.2) (16±74) All Group (range) *** *** *** (18.5±156) 28 (14.7) 30 (15.8) 26 (13.7) 106 (55.8) *** (0.7±1.4) (39.6±87.2) (16±73) Women (range) (10±80) 12 (23.6) 11 (21.5) 9 (17.7) 19 (37.2) (0.9±1.3) (39.6±73.8) (19±74) Men (range) BMI ˆ Body Mass Index (kg/m 2 ), w/h ˆ waist/hip ratio. ***P < , **P < 0.01 when comparing male vs female groups.

4 Table 2 Clinical and biological data in morbidly obese subjects according to glucose tolerance 559 N Age (y) BMI (kg/m 2 ) Energy intake (kcal) Fasting glucose Fasting insulin (mu/l) Fasting triglycerides HomaS % HomaB % NG IG * * NIDDM-c * * * * * NIDDM-pc * * * * * P value NS NS < < Four-way, age-adjusted, analysis of variance was performed. BMI ˆ Body Mass Index (kg/m 2 ), NG: normal glycaemic subjects, IGT: Glucose intolerant subjects, NIDDM-c: controlled non-insulin dependent diabetic subjects, NIDDM-pc: poorly controlled non-insulin dependent diabetic subjects. * shows mean values that are statistically different to NG. Figure 1 Fasting leptin concentrations and diabetic status in 241 morbidly obese subjects. NG normal glycaemic subjects, IGT: Glucose intolerant subjects, NIDDM-c: controlled non-insulin dependent diabetic subjects, NIDDM-pc: poorly controlled non-insulin dependent diabetic subjects. P ˆ , when doing four ways analysis of variance (age and sex adjusted, ***P < when comparing NG and NIDDM-pc (age and sex adjusted). Figure 2 Leptin concentrations with quartiles of glucose in morbidly obese subjects. 1 lowest quartile (3.8±5.3 mm), 2 ( > 5.3±6.0 mm), 3 ( > 6.0±7.3 mm), 4 highest quartile ( > 7.3± 25.5 mm). Least square means after adjustment for sex and log BMI are presented in the gure, P ˆ NIDDM groups (r ˆ 70.31, P ˆ ), although only a trend of negative correlation was observed in the NG and IGT groups (r ˆ 72, P ˆ 0.1). We performed a multivariate analysis to determine whether the association between fasting glucose and serum leptin concentrations, found in the morbidly obese group, was independent of the other tested factors, particularly gender. Several parameters including sex, age, BMI, waist/hip ratio, fasting glucose and insulin, total cholesterol and triglyceride concentrations, were introduced into a standard least square model of analysis (after logarithmic transformation of the quantitative variables). This analysis showed a signi cant inverse association between fasting glucose and leptin (F ˆ 12.5, P ˆ ), con rming that this parameter was associated with leptin, independently of the other factors. Other independent factors associated with leptin were sex (F ˆ 45.5, P < ), BMI (F ˆ 32.0, P < ) and fasting insulinaemia (F ˆ 8.6, P ˆ 0.004). An inverse association with triglycerides (F ˆ 4.7, Table 3 Correlations between serum leptin levels and several clinical and biological parameters All subjects Female group Men group Parameters r P value r P value r P value Age (y) BMI (kg/m 2 ) Waist (cm) < Hip (cm) 0.32 < Waist/hip ratio < Fasting glucose (mmol/l) Fasting insulin (mu/l) Cholesterol (mmol/l) Triglycerides (mmol/l) HomaS % HomaB % r: Spearman coef cient of correlation.

5 560 Table 4 Logistic regression analysis of variables associated with diabetes Parameters OR 95% CI P value Age (y) ± Sex (F/M) ± BMI (kg/m 2 ) ± Waist/hip ± Leptin < median (ng/ml) ± Insulin (mu/l) ± Ratio; 95% CI ˆ 95% con dence interval. P ˆ 0.03) was also found. In this model, those variables explained 41% of the leptin variability. We also divided the subjects according to their serum leptin concentrations over or under the median value of the cohort (55 ng/ml). Logistic regression analysis showed that leptin concentrations under the median were associated with diabetes mellitus (OR 3.46, P ˆ 0.01) (Table 4). Discussion We have shown that poorly controlled diabetes was accompanied by a signi cant reduction in serum leptin concentrations in females and males from a morbidly obese French population. Consistent with other studies, a gender difference in leptin concentrations was observed, probably due to a higher body fat content of women at any body mass index 7,16 although it is not excluded that gender itself might be associated with leptin independently of body fat mass. 22 A positive correlation between serum leptin concentrations and BMI was also found. 5,8 Since high correlations between leptin concentrations and body fat content have been largely described, 5±8,22 we cannot exclude the possibility that differences in body composition associated with the deterioration of glucose tolerance might contribute to the link of fasting glucose with serum leptin. In rhesus monkeys with obesity, the onset of diabetes was followed by a dramatic decrease of leptin 23 but this effect was associated with weight loss, suggesting reduction of the animals fat content. The assessment of body composition was not possible in these morbidly obese patients but at the higher end of the BMI range, BMI appears to be well correlated with body fat content. 24 Moreover, no clinical evidence of recent changes in body composition or ketosis were reported in our subjects, even in the group of subjects with poorly controlled diabetes, who were weight stable. Reduced leptin concentrations have been previously described in a smaller group of less severe caucasian obese subjects presenting NIDDM compared to obese subjects without NIDDM. 6 In contrast, no signi cant difference in serum leptin concentrations was observed in other non-obese caucasians with or without NIDDM. 15 This might suggest that the effect of diabetes on leptin is limited to obese subjects. Furthermore, no link (or a positive association) between insulin resistance and serum leptin concentrations was seen in diabetic and normoglycaemic populations from different ethnic groups. 15±19 In agreement with these studies, we found no signi cant difference in serum leptin concentrations between NG, IGT or controlled diabetics, despite the worsening of their insulin resistance. The decline of insulin secretion associated with the development of severe hyperglycaemia may lead to an impaired adipocyte secretion of leptin. Indeed, fasting refeeding experiments and insulin-clamp studies in rodents have suggested that insulin increases ob gene expression 11±13 and that the ob protein concentration can be modulated in cell lines by insulin at physiological concentrations. 14 In humans, food consumption and acute administration of insulin were not associated with modi cations of leptin concentrations, 5,8 whereas prolonged exposure to insulin increased serum leptin concentrations. 8,25 This suggests a role for insulin in chronic regulation of plasma leptin concentrations. A decreasing insulin secretion characterizes our population of morbidly obese subjects with various degrees of glucose intolerance. A severe insulin de ciency is likely to be present in the group of NIDDM-pc subjects since their mean fasting glycaemia is uncontrolled, despite treatment. Without taking into account whether they are treated or not for NIDDM, the group of morbidly obese subjects with a mean fasting glucose over 10 mm had signi cantly lower leptin concentrations ( ng/ml) than the subjects with a mean fasting glucose below 10 mm ( ng/ml, P < ) with a similar fasting insulin concentration ( vs respectively), suggesting that lower leptin concentrations were associated with the lack of control of diabetes. Therefore, lower leptin concentrations (than expected according to BMI) in these subjects might be due to insulin de ciency and also to the `starvation' like metabolism that characterizes uncontrolled diabetes. In this regard, we found that leptin was inversely correlated with hyperglycaemia, particularly in the group of morbidly obese diabetics. The possible consequences of the 30% reduction of leptin observed in morbidly obese poorly controlled diabetic subjects are unknown. It has been recently proposed that the regulation of the neuroendocrine system during starvation could, in cooperation with insulin, be the main physiological role for leptin. As leptin-de cient mice show neuroendocrine abnormalities similar to those observed in prolonged starvation circumstances, falling leptin concentration in the blood could be a critical signal that initiates response to starvation. 26 Therefore, we suggest that a relative leptin de ciency (than would be expected for the BMI) in this population of morbidly obese subjects, particularly in those with a decompensated metabolic state, may contribute to a vicious cycle maintaining (or even worsening) the obesity itself and/or its associated metabolic complications.

6 Acknowledgements This work was supported by the French MinisteÁre de la Recherche. We thank the Direction de la Recherche Clinique Assistance Publique-Hopitaux de Paris. We thank the physicians at the HoÃtel-Dieu hospital and particularly Dr Dominique Cassuto. We thank Mme Lefurn (Inserm U 342) for her excellent technical help. We thank Dr J Levy for kindly providing the HOMA/CIGMA software, Dr G Velho for his help in computing the Homa calculations, S Spector for his critical corrections on the manuscript and Pr Fiet for the measurements of serum insulin concentrations. We thank the patients and their families for their cooperation. References 1 Zhang Y, Proenca R, Maffei M, Barone M, Leopold L, Friedman JM. Positional cloning of the mouse obese gene and its human homologue. Nature 1994; 372: 425± Camp eld LA, Smith FJ, Guisez Y, Devos R, Burn P. Recombinant mouse OB protein: evidence for a peripheral signal linking adiposity and central neural networks. Science 1995; 269: 546± Peyllemoutier MA, Cullen MJ, Baker MB, Hecht R, Winters D, Boone T, Collins F. Effects of obese gene product in body weight regulation in ob/ob mice. Science 1995; 269: 543± Halaas JL, Gajiwala KS, Maffei M, Cohen SL, Chait BT, Rabinowitz D, Lallone R, Burley SK, Friedman JM. Weight reducing effects of the plasma protein encoded by the obese gene. Science 1995; 269: 543± Considine RV, Sinha MK, Heiman ML, Kriaciunas A, Stephens TW, Nyce MR, Ohannesian JP, Marco CC, McKee LJ, Bauer TL, Caro JF. Serum immuno reactive-leptin concentrations in normal weight and obese humans. N Engl J Med 1996; 334: 292± Sinha MK, Ohannesian JP, Heiman ML, Kriauciunas A, Stephens TW, Magosin S, Marco C, Caro JF. Nocturnal rise of leptin in lean, obese and non-insulin dependent diabetes mellitus subjects. J Clin Invest 1996; 97: 1344± Maffei M, Halaas J, Ravussin E, Pratley RE, Lee GH, Zhang Y, Fei H, Lallone R, Ranganathan S, Kern PA, Friedman JM. Leptin concentrations in human and rodent: measurement of plasma leptin and ob RNA in obese and weight reduced subjects. Nature Med 1995; 1: 1155± Kolaczynski JW, Nyce MR, Considine RV, Boden G, Nolan JJ, Henry R, Mudaliar SR, Olefsky J, Caro JF. Acute and chronic effect of insulin on leptin production in humans: studies in vivo and in vitro. Diabetes 1996; 45: 699± Caro JF, Sinha MK, Kolaczynski JW, Zhang PL, Considine RV. Leptin: the tale of an obesity gene. Diabetes 1996; 45: 1455± Schwartz MW, Baskin DG, Bukowski TR, Kuijper JL, Foster D, Lasser G, Prunkard DE, Porte D Jr., Woods SC, Seeley RJ, Weigle DS. Speci city of leptin action on elevated blood glucose concentrations and hypothalamic neuropeptide Y gene expression in ob/ob mice. Diabetes 1996; 45: 531± MacDouglas OA, Hwang CS, Fan H, Lane MD. Regulated expression of the obese gene product (leptin) in white adipose tissue and L3T3 adipocytes. Proc Natl Acad Sci USA 1995; 92: 9034± Cusin I, Sainsbury A, Doyle P, Rohner-Jeanrenaud F, Jeanrenaud B. The ob gene and insulin: a relationship leading to clues to the understanding of obesity. Diabetes 1995; 44: 1467± Saladin R, De Vos P, Guerre-Milo M, Leturque A, Girard J, Staels B, Auwerx J. Transient increase in the obese gene expression after food intake or insulin administration. Nature 1995; 377: 527± Leroy P, Dessolin S, Villageois P, Moon BC, Friedman JM, Ailhaud G, Dani C. Expression of ob gene in adipose cells: regulation by insulin. J Biol Chem 1996; 271: 2365± McGregor GP, Desaga JF, Ehlenz K, Fisher A, Heese F, Hegele A, LaÈmmer C, Peiser C, Lang RE. Radioimmunological measurement of leptin in plasma of obese and diabetic human subjects. Endocrinology 1996; 137: 1501± Zimmet P, Hodge A, Nicolson M, Staten M, De Courten M, Moore J, Morawiecki A, Lubina J, Collier G, Alberti G, Dowse G. Serum leptin concentration, obesity and insulin resistance in Western Samoans: cross sectional study. BMJ 1996; 313: 965± Haffner SM, Stern MP, Miettinen H, Wei M, Gingerich RL. Leptin concentration in diabetic and non-diabetic mexicanamericans. Diabetes 1996; 45: 822± Segal KR, Landt M, Klein S. Relationship between insulin sensitivity and plasma leptin concentration in lean and obese men. Diabetes 1996; 45: 988± Dagogo-Jack S, Fanelli C, Paramore D, Brothers J, Landt M. Plasma leptin and insulin relationships in obese and non-obese humans. Diabetes 1996; 45: 695± CleÂment K, Vaisse C, Manning B, Basdevant A, Guy-Grand B, Ruiz J, Silver KD, Shuldiner AR, Froguel Ph, Strosberg AD. Genetic variation in the b3-adrenergic receptor and increased capacity to gain weight in patients with morbid obesity. N Engl J Med, 1995; 333: 352± Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentration in man. Diabetologia 1985; 28: 412± Ostlund RE, Yang JW, Klein S, Gingerich R. Relation between plasma leptin concentration and body fat, gender, diet, age, and metabolic covariates. J Clin Endocrinol Metab 1996; 81: 3909± Hansen BC, Hotta K, Bodkin N, Ortmeyer H, Nicolson M. The leptin-insulin axis: Interactions in the development of obesity and diabetes. Exp Clin Endocrinol Diabetes 1996; 104 (suppl 2): Wellens RJ, Roche AF, Khamis HJ, Jackson AS, Pollock ML, Siervoguel RM. Relationships between the Body Mass Index and body composition. Obes Res 1996; 4: 35± MalmstroÈm R, Taskinen MR, Karonen SL, Yki JaÈrvinen H. Insulin increases plasma leptin concentrations in normal subjects and patients with NIDDM. Diabetologia 1996; 39: 993± Ahima AD, Parabakaran D, Mantzoros C, Qu D, Lowell B, Maratos-Flier E, Flier J. Role of leptin in the neuroendocrine response to fasting. Nature 1996; 382: 250±