Is there a relationship between leptin and insulin sensitivity ... - Nature

International Journal of Obesity (1998) 22, 171±177 ß 1998 Stockton Press All rights reserved 0307±0565/98 $12.00

Is there a relationship between leptin and insulin sensitivity independent of obesity? A population-based study in the Indian Ocean nation of Mauritius PZ Zimmet1, VR Collins1, MP de Courten1, AM Hodge1 , GR Collier2, GK Dowse3, KGMM Alberti4, J Tuomilehto5, F Hemraj6, H Gareeboo6, P Chitson6 and D Fareed 7 on behalf of Mauritius NCD Study Group 1 International Diabetes Institute, Melbourne, Australia; 2 Deakin University, Geelong, Australia; 3 Disease Control Service, Health Department, Western Australia; 4 Department of Medicine, Newcastle-upon-Tyne, UK; 5 Public Health Institute, Helsinki, Finland; 6 Ministry of Health, Port Louis, Mauritius; and 7 WHO Of®ce, Port Louis, Mauritius

OBJECTIVE: It has been shown previously in smaller studies that fasting serum leptin and insulin concentrations are highly correlated, and insulin sensitive men have lower leptin levels than insulin resistant men matched for fat mass. We have examined the association between insulin resistance (assessed by fasting insulin) and leptin after controlling for overall and central adiposity in a population-based cohort. DESIGN: Leptin levels were compared across insulin resistance quartiles within three categories of obesity (tertiles of body mass index (BMI)). Partial correlation coef®cents and multiple linear regression models were used to assess the relationship between leptin and fasting insulin after adjusting for BMI and waist to hip ratio (WHR) or waist circumference. SUBJECTS: Subjects were normoglycemic participants of a 1987 non-communicable diseases survey conducted in the multiethnic population of Mauritius. 1227 men and 1310 women of Asian Indian, Creole and Chinese ethnicity had normal glucose tolerance and fasting serum leptin measurements. RESULTS: Mean serum leptin concentration increased across quartiles of fasting insulin in each BMI group and gender, after controlling for BMI, WHR and age. Furthermore, fasting insulin was a signi®cant determinant of serum leptin concentration, independent of BMI and WHR, in both men and women. Similar results were found if waist circumference replaced BMI and WHR in the model. CONCLUSION: These results suggest that insulin resistance=concentration may contribute to the relatively wide variation in leptin levels seen at similar levels of body mass or alternatively, leptin may play a role in the etiology of insulin resistance. Further studies will be important to determine whether the hyperleptinemia=insulin resistance relationship has a role in the natural history of obesity, Type 2 diabetes mellitus and the other metabolic abnormalities associated with insulin resistance. Keywords: leptin; insulin sensitivity; population-based study; obesity

Introduction Leptin, the circulating product of the recently discovered ob gene, may be a critical factor in the control of energy stores. The effects of leptin suggest regulation of energy balance at the hypothalamic level,1 and leptin receptors have now been demonstrated in the choroid plexus.2 Leptin treatment causes dramatic reductions in food intake and body weight in rodents.3±5 However, only the obese (ob=ob) mouse 3,4 lacks circulating leptin, whereas other animal models of obesity.5±7 and obese humans8±10 show hyperleptinemia. Correspondence: Professor Paul Zimmet, International Diabetes Institute, 260 Kooyong Road, Caul®eld South, Victoria, 3162, Australia. Received 19 March 1997; revised 9 September 1997; accepted 26 September 1997

The effects of short and long term insulin infusions on leptin production have been studied,11 and while the results suggest that ob gene expression and leptin production are indirectly regulated by insulin,12 more detailed studies are needed. Certainly, a strong positive correlation exists between leptin and insulin concentrations,10±14 and insulin sensitive men have lower leptin concentrations than those with insulin resistance, independent of body fat mass.15 This raises important questions about the potential role of leptin in the modulation of insulin sensitivity, or vice versa. The interrelationship of leptin and insulin sensitivity is confounded by the strong association between leptin and obesity6,8±10 and the effect of obesity on insulin sensitivity,16 indicating the need to control for the effect of obesity. In addition, much of human leptin research so far has been based on small clinical studies.8,9,13±15 This study aimed to evaluate the relationship between insulin sensitivity and leptin

Leptin and insulin sensitivity PZ Zimmet et al

172

concentration independent of obesity in data from a large epidemiological study of diabetes and cardiovascular (CVD) disease in the multiethnic population of Mauritius.17 Research design and methods

Subjects. The subjects were drawn from a populationbased non-communicable disease survey performed in Mauritius in 1987 and are described in detail elsewhere.17,18 Mauritius is an island nation located in the southwestern Indian Ocean. The multiethnic population comprises about 70% of Asian Indian origin, 27.9% Creole (mixed with mainly African ancestry) and 2.1% Chinese. The survey protocol was reviewed and approved by the Alfred Healthcare Group Ethics Committee (Melbourne, Australia) and consent of the subjects for leptin measurement was inferred from their voluntary participation. Survey procedure and analyses. All adults aged 25± 74 y were eligible to attend the survey if resident in the population clusters chosen in the representative sample.17 Subjects presented to a survey site between 08:00 and 10:00 h after an overnight fast. All subjects not using oral hypoglycaemic tablets or insulin, underwent an oral glucose tolerance test (75 g dextrose monohydrate). Fasting and 2 h blood samples were collected and plasma glucose was measured on site using a glucose analyser (YSI, Yellow Springs, OH). Glucose tolerance status was classi®ed according to WHO recommendations.19 Body mass index (BMI) was calculated as kg=m2 and waist and hip girth were measured in duplicate to calculate waist to hip ratio (WHR) from the means of measurements as previously described.18 Eligibility for the present study was limited to 1227 men and 1310 women who had normal glucose tolerance19 in 1987, and who also attended a followup survey in 1992. Serum leptin levels were only available for those who participated in both surveys. Leptin measurements on sera frozen since 1987 (720 C), were performed in 1996, using radioimmunoassay kits (Linco, St Louis, MO). Insulin values for the same subjects were measured in 1987 on frozen sera using an in-house assay.20,21 Interassay and intraassay coef®cients of variation were 6% and 4%, respectively. Cross-reactivity with intact proinsulin and 32,33-split proinsulin was 27% and 16%, respectively. There was < 5% difference between results obtained with this assay and a highly speci®c enzymelinked immunosorbent assay. Fasting insulin was used as a marker of insulin sensitivity. It has previously been shown that in non-diabetic subjects fasting insulin is closely correlated with more direct measures of insulin resistance.22 Statistical analysis. All analyses were performed using the Statistical Package for the Social Sciences.23

Spearman rank correlation coef®cients were calculated between all variables using a 2-tailed test for signi®cance. The relationship of serum leptin concentration to insulin resistance was examined within gender-speci®c tertiles of BMI. For men, the cut-off points to de®ne the tertiles were 21.0 and 24.0 kg=m2 and for women they were 21.1 and 25.0 kg=m2. Within each tertile of BMI, gender-speci®c distributions of fasting insulin were divided into quartiles. Mean levels of serum leptin were calculated across quartiles of fasting insulin adjusting for differences in age, BMI and WHR by analysis of covariance. The signi®cance of differences across groups was assessed by the F-test. Log10transformation was used to normalize distributions of leptin, fasting glucose and fasting insulin, and mean values were back-transformed for presentation. The independent linear relationship between log10leptin and log10fasting insulin was assessed using all subjects after adjusting for age, BMI, WHR and log10fasting glucose in multiple linear regression models calculated for men and women separately. Partial correlations of fasting insulin with leptin were calculated as the other variables were sequentially added to the model. These models were also used to calculate the overall R2 and signi®cance of changes in R2 as the independent variables were added. Similar linear regression models using waist circumference instead of BMI and WHR were also calculated. In an alternative analysis, insulin sensitivity computed using the Homeostasis Model Assessment (HOMA) model24 was used instead of fasting insulin.

Results Table 1 shows the mean levels of selected variables in the study subjects. Men were leaner (according to BMI), but had higher mean WHR and waist circumference than women. All subjects had normal glucose tolerance (by selection), although mean fasting glucose levels were higher in men than women. Women had higher fasting insulin concentration and also were more insulin resistant on the basis of HOMA than men. Mean serum leptin levels were about 3-fold higher in women than men (10.2 vs 3.4 ng=ml). To assess associations between the variables used in later multivariate analyses, Table 2 shows Spearman correlation coef®cients between the anthropometric and metabolic variables. Insulin sensitivity calculated from the HOMA model was very strongly correlated with fasting insulin in both men and women. Leptin was signi®cantly correlated with each of the anthropmetric variables, although the relationship with WHR was not as strong as with BMI or waist circumference. Strong correlations were also observed between leptin and the metabolic variables, except fasting glucose. Figure 1 depicts age-, BMI- and WHR-adjusted geometric mean leptin according to quartiles of fast-


173

Table 1 Mean (95% con®dence intervals) levels of selected characteristics of Mauritians in the study Characteristics

Men

n Age (y) BMI (kg=m2) WHR Waist (cm) Fasting plasma glucose (mmo1=1)a Fasting serum insulin (mU=ml)a HOMAS (%)a Serum leptin (ng=ml)a

1227 39.9 (39.2±40.5) 22.7 (22.5±22.9) 0.88 (0.88±0.89) 77.0 (76.6±77.5) 5.17 (5.15±5.20) 4.16 (3.92±4.41) 85.2 (81.1±89.5) 3.36 (3.24±3.47)

Women 1310 40.5 (39.8±41.1) 23.5 (23.3±23.8) 0.80 (0.79±0.80) 72.6 (72.0±73.1) 5.04 (5.02±5.07) 5.44 (5.19±5.69) 69.1 (66.3±72.0) 10.17 (9.78±10.58)

a Geometric mean. BMI body mass index; WHR waist to hip ratio, HOMAS insulin sensitivity from the HOMA model.

ing insulin within categories of obesity. At each level of obesity, and in both men and women, there was a highly signi®cant increase in mean leptin levels with increasing insulin resistance (P < 0.001 for each comparison). It can also be seen from Figure 1, that leptin levels increase across BMI categories (note the difference in the magnitude of the scales of the leptin axes between BMI groups). This analysis was repeated for each ethnic group separately (data not shown) and the relationship between fasting insulin and leptin was consistent for each of Indians, Creoles and Chinese, although differences across fasting insulin were not statistically signi®cant in Chinese, due to small numbers. To assess the linear relationship between leptin and fasting insulin across the entire range of obesity, multiple linear regressions were performed with leptin (log10) as the dependent variable, on men and women separately (Table 3). Partial correlation coef®cients for insulin, adjusting for the other independent variables are presented along with R2 for the models. The partial correlation coef®cient for fasting insulin alone was about 0.5 in both men and women but was

reduced to about half by adjustment for BMI. Further adjustment for age, fasting glucose and WHR did not alter the magnitude of the partial correlation coef®cient. The models shown in Table 3 explained 43% and 53% of the variation in leptin in men and women, respectively. If BMI and WHR were replaced with waist circumference 44% and 48% of this variation was explained. A BMI by fasting insulin interaction term was also tested, but was not signi®cant in either gender. Models using insulin resistance from the HOMA model rather than fasting insulin concentration showed very similar results.

Discussion This study, in a large population-based cohort, con®rms results of clinic-based investigations14,15 and our ®ndings in another population,10 that insulin resistance (as indicated by fasting insulin concentration) is associated with elevated leptin concentrations independent of adiposity and fat distribution. More insulin

Table 2 Spearman correlations of anthropometric and metabolic variables used in analyses BMI

Age

Waist

WHR

Leptin

Fasting insulin

Fasting glucose

Men BMI Age Waist WHR Leptin Fasting insulin Fasting glucose HOMAS

0.024ns 0.866*** 0.594*** 0.633*** 0.561*** 0.102*** 7 0.561***

0.143*** 0.271*** 0.106*** 7 0.064* 0.161*** 0.056*

0.762*** 0.635*** 0.572*** 0.142*** 7 0.573***

0.433*** 0.379*** 0.138*** 7 0.381***

0.540*** 0.051ns 7 0.537***

0.177*** 7 0.999***

7 0.216***

Women BMI Age Waist WHR Leptin Fasting insulin Fasting glucose HOMAS

0.135*** 0.863*** 0.478*** 0.708*** 0.479*** 0.138*** 7 0.499***

0.234*** 0.315*** 0.011ns 7 0.053ns 0.304*** 0.040ns

0.747*** 0.636*** 0.480*** 0.174*** 7 0.487***

0.336*** 0.310*** 0.165*** 7 0.314***

0.506*** 0.068* 7 0.504***

0.179*** 7 0.999***

7 0.219***

* P < 0.05; ** P < 0.01; *** P < 0.001; ns not statistically signi®cant. BMI body mass index; WHR waist to hip ratio, HOMAS insulin sensitivity from the HOMA model.


174

Figure 1 Age-, BMI- and WHR-adjusted mean leptin levels according to quartiles of fasting insulin within obesity groups (BMI tertiles) in Mauritian men and women. ***P < 0.001 across insulin resistance quartiles.

resistant subjects had leptin levels consistently higher than insulin sensitive subjects regardless of gender or level of obesity, and this was true in each ethnic group. Linear regression models con®rmed the associations observed within BMI tertiles. Consistent with earlier reports,8±10 mean leptin concentration was substantially higher in obese than lean Mauritian subjects, and women had leptin levels at least three times higher than men. Moreover, the large variation in leptin at similar levels of BMI suggests that factors besides obesity are important in regulating serum leptin.8±10 Insulin sensitivity appears to be one such factor. Testosterone levels or the ratio of androgen=estrogen have recently been observed to correlate negatively with leptin concentrations in men independent of BMI,25 suggesting another possible factor

contributing to variations in leptin. This observation may also be important in explaining male=female differences in leptin levels which persist after adjusting for body fat in some studies, and do not appear to be explained by levels of female reproductive hormones.26±30 The HOMA model has been suggested as a useful tool for assessing insulin sensitivity in large epidemiological studies, where it is prohibitive, both economically and practically, to undertake more complex assessments of insulin sensitivity.31,32 HOMA estimates of insulin resistance have been shown to correlate well with more direct methods.33 However, when HOMA derived insulin sensitivity is strongly correlated with fasting insulin, as observed in this population, it is not adding any extra information. Haffner and colleagues34,35 have reported similar high correlation coef®cients for HOMA insulin sensitivity and fasting insulin in the Mexico City Diabetes Study and San Antonio Heart Study populations, concluding that the HOMA model may be most useful for estimating b-cell function. Thus it is not surprising that similar results were obtained, whether HOMA insulin sensitivity or fasting insulin were used. To account for the confounding effect of obesity on leptin and insulin sensitivity, Segal et al15 assessed the relationship speci®cally in 18 lean insulin sensitive and resistant men matched closely for body fat mass. Insulin sensitivity was classi®ed using an intravenous glucose tolerance test and minimal model analysis. Leptin was signi®cantly lower in the lean insulin sensitive subjects than in those who were relatively insulin resistant. Our study uses indirect estimates of both insulin resistance and fat mass but given the magnitude of the observed difference, it seems unlikely that this effect would be lost if more direct measures were used. Moreover, the association of insulin sensitivity and leptin, independent of BMI and WHR or waist circumference, was seen over the entire population range of BMI (not just in lean subjects), and in both men and women. There is a possibility that the contribution of adipose tissue to leptin levels was not being fully accounted for by adjusting for BMI and WHR. In fact, there was a trend for increasing waist circumference with insulin resistance, after correcting for BMI (data not shown), although it was not possible to adjust for both BMI and WHR when assessing waist circumference, due to collinearity. However, Segal's study assessing the effect of adipose tissue mass with more sophisticated methods such as densitometry con®rms the independence of the relationship of leptin with measures of insulin resistance.15 A number of other studies have examined the relationship between leptin and insulin resistance or insulin concentration independent of obesity with con¯icting results. Havel et al36 found that leptin was correlated with fasting insulin in women over a range of BMI, but after adjusting for BMI or percent body fat this relationship was no longer signi®cant. In

Leptin and insulin sensitivity PZ Zimmet et al Table 3 Partial correlation coef®cients for log10 fasting insulin with log10 leptin as the dependent variable, after controlling for the other variables which were added to the linear regression model in order of listing. Partial correlations

Stepwise linear regression model

Partial correlation coefficient (PCC)

Significance of PCC

Adjusted R2

Significance of change in R2

Men Fasting insulina BMI Age Fasting glucosea WHR

0.498 0.251 0.264 0.273 0.269

< 0.001 < 0.001 < 0.001 < 0.001 < 0.001

0.25 0.42 0.43 0.43 0.43

< 0.001 < 0.001 < 0.001 0.007 0.106

Women Fasting insulina BMI Age Fasting glucosea WHR

0.507 0.260 0.250 0.256 0.255

< 0.001 < 0.001 < 0.001 < 0.001 < 0.001

0.26 0.52 0.52 0.53 0.53

< 0.001 < 0.001 0.007 0.029 0.759

Variable entered

a

log transformed. BMI body mass index; WHR waist to hip ratio.

African-American women, leptin was not correlated with insulin sensitivity (by intravenous glucose tolerance test) after adjusting for body fat mass by dualenergy X-ray absorbtiometry (DEXA).37 Studies by Larsson et al,38 Kohrt et al39 and Turpeinen et al 40 also found that insulin or insulin sensitivity did not correlate with leptin, independent of obesity, measured by impedance, DEXA or infra-red densitometric assay, respectively, in subjects with normal or impaired glucose tolerance. However, in support of our ®ndings, leptin was independently related to insulin sensitivity (hyperinsulinaemic, euglycaemic clamp) after adjusting for body fat percentage in Danish men and women,30 and in a large epidemiological study, leptin levels were related to fasting insulin independently of BMI.41 None of the positive results discussed so far have indicated the relative strength of the relationships between leptin and insulin concentration or insulin sensitivity. As these two parameters are closely correlated in most populations, it is dif®cult to differentiate between them. Both fasting insulin and insulin sensitivity (hyperinsulinaemic, euglycaemic clamp) were correlated with leptin independently of BMI in middle-aged Finnish men, but no attempt was made to assess which was more important.42 In subjects with Type 2 diabetes mellitus, the association between insulin concentration and resistance is not as close, and in two studies making use of this phenomenon, leptin was found to be more closely correlated with insulin concentration than with insulin resistance. These inconsistent results indicate the need to standardize the approach to investigations into leptin, insulin and obesity. Results from a cross-sectional study such as this do not reveal the direction of the association between leptin and insulin resistance. However, recent studies have shed some light on the possible effects of insulin and=or insulin resistance on leptin. Insulin has been

shown to be a potent regulator of leptin gene expression in cultured human adipocytes and to induce increased leptin production.12 Clinical studies of the effects of insulin on leptin also show mixed results. In some studies, no short-term effect was observed,11,13,14 although Kolaczynski et al11 demonstrated that in the long-term (48±72 h) insulin may regulate ob gene expression and leptin production indirectly, possibly via a trophic effect on adipocytes. More rapid, although not acute, responses of leptin to insulin have also been observed. In normal and Type 2 diabetic subjects, leptin levels were increased by 8.5 h of hyperinsulinemic clamp,43 and leptin levels were increased after only 4 h of supraphysiological hyperinsulinaemia in lean normal subjects.44 However, the lack of acute effects of insulin on leptin implies that post-prandial satiety is not regulated by leptin,44 although insulin may be important in long-term leptin regulation.11,43,44 On the other hand, ®ndings from a study using human liver cells have suggested that leptin when administered in concentrations similar to that found in obesity, antagonizes insulin signaling.45 This led the authors to conclude that secretion of leptin by adipose tissue may be a mechanism whereby adiposity causes insulin resistance. In the Israeli Sand rat (Psammomys obesus), a model of Type 2 diabetes, leptin has been reported to inhibit insulin binding to adipocyte insulin receptors, offering another mechanism whereby hyperleptinaemia could contribute to insulin resistance in obesity.46 At the genetic level, the leptin receptor and acute insulin response loci appear to be closely co-localised on chromosome 1, as demonstrated in Pima Indians.47 It has been suggested therefore, that the leptin receptor gene could play a secondary role in regulation of the acute insulin response (AIR), with high leptin levels down-regulating the receptor gene and diminishing AIR, which may in turn have implications for

175


176

insulinaemia and insulin sensitivity. However, if leptin and insulin are acting as part of a feed-back loop controlling adipose tissue mass, each could modulate the other simultaneously. Direct effects of leptin on insulin sensitivity at the insulin receptor have yet to be examined in humans. Only studies investigating the effect of leptin administration, once leptin is available for human use, on insulin concentrations or insulin sensitivity, will help to clarify these associations. In conclusion, our ®ndings in a large populationbased sample are consistent with a biological link between leptin and insulin. They also suggest that there are determinants of serum leptin concentration beside body fat. Insulin resistance or insulin concentration per se, or via determinants including genetic, biological and environmental factors16,31 may modulate leptin concentration. Alternatively, leptin itself may have an important role in the determination of insulin resistance in the liver,45 or through interactions with the insulin receptor.46 Our ®ndings of a strong association between leptin and a marker for insulin sensitivity independent of obesity suggest an important role for leptin in human metabolism. Further studies on the dynamics of the relationship between leptin, insulin and adipose tissue are required to clarify the underlying importance of leptin. Prospective epidemiological studies will also be required to assess the possible role of leptin in the natural history of obesity and glucose intolerance. Acknowledgements

This study was a collaborative project with invaluable support provided by the Government and Health Department of Mauritius, and the World Health Organization. The work was supported by Grant DK-25446 from the National Institute of Diabetes and Digestive and Kidney Diseases. We are grateful to the whole survey team, especially Mr Ray Spark and Ms Linda Ashworth for technical assistance, and to the British Diabetic Association for laboratory support.

6

7

8

9

10

11

12

13 14 15 16 17

18

References

1 Zhang Y, Proenca R, Maffei M, Barone M, Leopold L, Friedman J. Positional cloning of the mouse obese gene and its human homolgue. Nature 1994; 372: 425±432. 2 Tartaglia L, Dembski M, Weng X, Deng N, Culepper J, Devos R, Richards G, Camp®eld L, Clark F, Deeds J, Muir C, Sanker S, Moriarty A, Moore K, Smutko J, Mays G, Woolf E, Monroe C, Tepper R. Identi®cation and expression cloning of a leptin receptor, OB-R. Cell 1995; 85: 1265±1271. 3 Halaas JL, Gajiwala KS, Maffei M, Cohen SL, Chait BT, Rabinowitz D, Lallone RL, Burley SK, Friedman JM. Weightreducing effects of the plasma protein encoded by the obese gene. Science 1995; 269: 543±546. 4 Pelleymounter M, Cullen M, Baker M, Hecht R, Winters D, Boone T, Collins F. Effects of the obese gene product on body weight regulation in ob=ob mice. Science 1995; 269: 540±543. 5 Cusin I, Rohner-Jeanrenaud F, Stricker-Krongrad A, Jeanrenaud B. The weight-reducing effect of an intracerebro-ventricular bolus injection of leptin in genetically obese fa=fa rats.

19 20 21

22 23 24

Reduced sensitivity compared with lean animals. Diabetes 1996; 45: 1446±1450. Frederich RC, Hamann A, Anderson S, LoÈllman B, Lowell BB, Flier JS. Leptin levels re¯ect body lipid content in mice: Evidence for diet-induced resistance to leptin action. Nature Medicine 1995; 1: 1311±1314. Collier G, Walder K, Lewandowski P, Sanigorski A, Lee S, Zimmet P. Development of obesity and hyperleptinemia in Psammomys obesus. Diabetologia 1996; 39 (Suppl 1): A146. Maffei M, Halaas J, Ravussin E, Prateley RE, Lee GH, Zhang Y, Fei H, Kim S, Lallone R, Ranganathan S, Kern PA, Friedman JM. Leptin levels in human and rodent: Measurement of plasma leptin and ob RNA in obese and weightreduced subjects. Nature Medicine 1995; 1: 1155±1161. Considine RV, Sinha MK, Heiman ML, Kriauciunas A, Stephens TW, Nyce MR, Ohannesian JP, Marco CC, Mckee LJ, Bauer TL, Caro JF. Serum immunoreactive-leptin concentrations in normal-weight and obese humans. N Engl J Med 1996; 334: 292±295. 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±969. Kolaczynski JW, Nyce MR, Considine RV, Boden G, Nolan JJ, Henry R, Mudaliar SR, Olefsky J, Caro JF. Acute and chronic effects of insulin on leptin production in humans. Studies in vivo and in vitro. Diabetes 1996; 45: 699±701. Wabitsch M, Jensen PB, Blum WF, Christoffersen CT, Englaro P, Heinze E, Rascher W, Teller W, Tornqvist H, Hauner H. Insulin and cortisol promote leptin production in cultured human fat cells. Diabetes 1996; 45: 1435±1438. Gabriel M, Jinagouda S, Boyadjian R, Kades W, Ayad M, Saad M. Is leptin the link between obesity and insulin resistance? Diabetes 1996; 45 (Suppl 2): 51A. Dagogo-Jack S, Fanelli C, Paramore D, Brothers J, Landt M. Plasma leptin and insulin relationships in obese and nonobese humans. Diabetes 1996; 45: 695±698. Segal KR, Landt M, Klein S. Relationship between insulin sensitivity and plasma leptin concentration in lean and obese men. Diabetes 1996; 45: 988±991. Kolaczynski JW, Caro JF. Molecular mechanism of insulin resistance in human obesity. Current Opinion in Endocrinology and Diabetes 1996; 3: 36±43. Dowse G, Gareeboo H, Zimmet P, Alberti K, Tuomilehto J, Fareed D, Brissonnette L, Finch C. High prevalence of NIDDM and impaired glucose tolerance in Indian, Creole and Chinese Mauritians. Diabetes 1990; 39: 390±396. Dowse G, Zimmet P, Gareeboo H, Alberti K, Tuomilehto J, Finch C, Chitson P, Tulsidas H. Abdominal obesity and physical inactivity as risk factors for NIDDM and impaired glucose tolerance in Indian, Creole and Chinese Mauritians. Diab Care 1991; 14: 271±282. World Health Organization. Diabetes Mellitus: Report of a WHO Study Group. WHO, Geneva: 1985. Soeldner J, Slone D. Critical variables in radioimmunoassay of serum insulin using the double antibody technique. Diabetes 1965; 14: 771±779. Dowse G, Zimmet P, Alberti K, Brigham L, Carlin J, Tuomilehto J, Knight L, Gareeboo H. Serum insulin distributions and reproducibility of the relationship between 2-hour insulin and plasma glucose levels in Asian Indian, Creole and Chinese Mauritians. Metabolism 1993; 42: 1232±1241. Laakso M. How good a marker is insulin level for insulin resistance? Amer Epidemiol 1993; 137: 959±965. SPSS PC 4.0 for the IBM PC=XT=AT and PS=2 [program]. SPSS Inc., Chicago: 1990. Rudenski A, Matthews D, Levy J, Turner R. Understanding `Insulin Resistance': Both glucose resistance and insulin resistance are required to model human diabetes. Metabolism 1991; 40: 908±917.


25 JockenhoÈvel F, Blum W, Vogel E, Englaro P, MuÈller-Wieland D, Dankwart R, Rascher W, Krone W. Testosterone substitution normalizes elevated serum leptin levels in hypogonadal men. Clin Endocrinol and Metab 1997; 82: 2510±2513. 26 Masuzaki H, Hosoda K, Ogawa Y, Miyawaki T, Hiraoka J, Hanaoka I, Yasuno A, Natsui K, Sugawara A, Arii S, Yoshima Y, Nishi S, Yamori Y, Nakao K. Glucocorticoid and gender regulate plasma leptin levels in humans. Diabetologia 1997; 40 (Suppl. 1): A175. 27 Havel PJ, Kasim-Karak S, Dubuc GR, Muhler W, Phinney SD. Gender differences in plasma leptin concentrations. Nature Medicine 2: 949±950 (Letter). 28 Kennedy A, Gettys T, Watson P, Wallace P, Ganaway E, Pan Q, Garvey W. The metabolic signi®cance of leptin in humans: Gender-based differences in relationship to adiposity, insulin sensitivity, and energy expenditure. J Clin Endocrinol Metab 1997; 82: 1293±1300. 29 Rosenbaum M, Nocolson M, Hirsch J, Heyms®eld S, Gallagher D, Chu F, Leibel R. Effects of gender, body composition, and menopause on plasma concentrations of leptin. J Clin Endocrinol Metab 1996; 81: 3424±3427. 30 Nyholm B, Fisker S, Lund S, Mùller N, Schmitz O. Increased circulating leptin concentrations in insulin-resistant ®rstdegree relatives of patients with non-insulin-dependent diabetes mellitus: relationship to body composition and insulin sensitivity but not to family history of non-insulin-dependent diabetes mellitus. Eur J Endocrinol 1997; 136: 173±179. 31 Dowse G, Qin H, Collins V, Zimmet P, Alberti K, Gareeboo H. Determinants of estimated insulin resistance and B-cell function in Indian, Creole and Chinese Mauritians. Diab Res Clin Prac 1990; 10: 265±279. 32 Multi-Centre S. UK prospective diabetes study. v. Characteristics of newly presenting type 2 diabetic patients: estimated insulin sensitivity and islet B-cell function. Diab Med 1988; 5: 444±448. 33 Matthews D, Hosker J, Rudenski A, Naylor B, Treacher D, Turner R. Homeostasis model assessment: insulin resistance and B-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985; 28: 412±419. 34 Haffner S, Gonzalez C, Miettinen H, Kennedy E, Stern M. A prospective analysis of the HOMA model: The Mexico City Diabetes Study. Diabetes Care 1996; 19: 1138±1141. 35 Haffner S, Miettinen H, Stern M. The homeostatis model in the San Antonio Heart Study. Diabetes Care 1997; 20: 1087± 1092.

36 Havel P, Kasim-Karakas S, Mueller W, Johnson P, Gingerich R, Stern J. Relationship of plasma leptin to plasma insulin and adiposity in normal weight and overweight women: Effects of dietary fat content and sustained weight loss. J Clin Endocrinol Metab 1996; 81: 4406±4413. 37 Dua A, Hennes M, Hoffman R, Maas D, Krakower G, Sonnenberg G, Kissebah A. Leptin: A signi®cant indicator of total body fat but not of visceral fat and insulin insensitivity in African-American women. Diabetes 1996; 45: 1635±1637. 38 Larsson H, ElmstaÊhl S, AhreÂn B. Plasma leptin levels correlate to islet function independently of body fat in postmenopausal women. Diabetes 1996; 45: 1580±1584. 39 Kohrt W, Landt M, Birge Jr S. Serum leptin levels are reduced in response to exercise training, but not hormone replacement therapy, in older women. J Clin Endocrinol Metab 1996; 81: 3980±3985. 40 Turpeinen A, Haffner S, Louheranta A, Niskanen L, Miettinen H, Uusitupa M. Serum leptin in subjects with impaired glucose tolerance in relation to insulin sensitivity and ®rst-phase insulin response. Int J Obes 1997; 21: 284±287. 41 Stolk R, Janssen J, Pols H, Englaro P, Blum W, Attanasio A, Grobbee D, Lamberts S. Insulin and leptin are related independently from body mass index. Diabetologia 1997; 40: A262. 42 Haffner S, Miettinen H, MykkaÈnen L, KarhapaÈaÈ P, Rainwater D, Laakso M. Leptin concentrations and insulin sensitivity in normoglycemic men. Int J Obes 1997; 21: 393±399. 43 MalmstroÈm R, Taskinen M-R, Karonen S-L, Yki-JaÈvinen H. Insulin increases plasma leptin concentrations in normal subjects and patients with NIDDM. Diabetologia 1996; 39: 993± 996. 44 Utriainen T, MalmstroÈm R, MaÈkimattila S, Yki-JaÈrvinen H. Supraphysiological hyperinsuliemia increases plasma leptin concentrations after 4 h in normal subjects. Diabetes 1996; 45: 1364±1366. 45 Cohen B, Novick D, Rubinstein M. Modulation of insulin activities by leptin. Science 1996; 274: 1185±1188. 46 Walder K, Filippis A, Clark S, Zimmet P, Collier G. Leptin inhibits insulin binding in isolated rat adipocytes. Diabetologia 1997; 40:A 176. 47 Norman R, Leibel R, Chung W, Power-Kehoe L, Chua SJ, Knowler W, Thompson D, Bogardus C, Ravussin E. Absence of linkage of obesity and energy metabolism to markers ¯anking homologues of rodent obesity genes in Prima Indians. Diabetes 1996; 45: 1229±1232.

177