Association Between Baseline Plasma Leptin Levels ... - Diabetes Care

Epidemiology/Health Services/Psychosocial Research O R I G I N A L

A R T I C L E

Association Between Baseline Plasma Leptin Levels and Subsequent Development of Diabetes in Japanese Americans MARGUERITE J. MCNEELY, MD EDWARD J. BOYKO, MD DAVID S. WEIGLE, MD JANE B. SHOFER, MS

STEVEN D. CHESSLER, MD DONNA L. LEONNETTI, PHD WILFRED Y. FUJIMOTO, MD

OBJECTIVE — Plasma leptin levels correlate strongly with increased total adipose tissue, a known risk factor for type 2 diabetes, yet the role of leptin in the etiology of diabetes remains unclear. We sought to determine whether leptin is a risk factor for development of diabetes in Japanese Americans. RESEARCH DESIGN AND METHODS — We compared baseline leptin levels in 370 nondiabetic Japanese Americans who remained nondiabetic for 5–6 years of follow-up with those of 40 nondiabetic Japanese Americans who developed diabetes during follow-up. All participants had computed tomography measurements of baseline subcutaneous chest, abdomen, thigh, and intra-abdominal fat, with total fat defined as the sum of all these measurements. RESULTS — The mean age was 51.7 ± 11.7 years for men and 51.9 ± 12.0 years for women. The 23 men who developed diabetes had significantly higher leptin levels than the 212 men who remained nondiabetic (P 0.01). Among men, baseline leptin levels predicted diabetes risk independent of baseline total fat, insulin, insulin resistance, glucose, or age in separate multiple logistic regression models (relative risk adjusted for baseline total fat = 1.80 per SD increase [2.7 ng/ml], 95% CI 1.02–3.17). This association was particularly strong among men in the top decile for intra-abdominal fat. In contrast, the 17 women who developed diabetes had leptin levels similar to those of the 158 women who remained nondiabetic (P = 0.31). CONCLUSIONS — Among Japanese Americans, increased baseline leptin levels are associated with increased risk of developing diabetes in men but not in women. Diabetes Care 22:65–70, 1999

besity is an established risk factor for type 2 diabetes, yet the physiology of obesity is only partially understood. A recent advance was the discovery of leptin, a hormone originally found to be the missing product of the obese(ob) gene in the ob/ob mutant strain of obese mice (1).

O

Due to their gene defect, ob/ob mice have undetectable levels of circulating leptin. When administered exogenous leptin, these mice consume less food, exhibit increased thermogenesis, and lose weight (2). Leptin is produced by adipocytes, and is postulated to act as a satiety factor by

From the Department of Medicine (M.J.M., E.J.B., D.S.W., J.B.S., S.D.C., W.Y.F.), University of Washington School of Medicine; the Seattle Epidemiologic Research and Information Center (E.J.B.), Veterans Affairs Puget Sound Health Care System; and the Department of Anthropology (D.L.L.), University of Washington, Seattle, Washington. Address correspondence and reprint requests to Marguerite J. McNeely, MD, MPH, Division of General Internal Medicine, Box 356429, University of Washington School of Medicine, Seattle, WA 98195-6429. E-mail: [email protected]. Received for publication 3 June 1998 and accepted in revised form 28 September 1998. Abbreviations: ob, obese; RR, relative risk. A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.

DIABETES CARE, VOLUME 22, NUMBER 1, JANUARY 1999

binding to receptors in the hypothalamus (2,3). Leptin levels in humans increase with increasing body fat. Although mutations in the human genes for leptin and the leptin receptor have been reported in individual families (4,5), such mutations do not account for the seemingly paradoxical elevation of leptin in the majority of obese individuals (6,7). The concept of leptin resistance has been postulated as a possible explanation for this observation (8,9). Leptin levels are also higher in women than men, even after adjusting for differences in body fat. Independent of the relationship to body fat, leptin levels are also positively correlated with fasting insulin levels in humans (10–16). While the association between insulin and leptin has not been fully elucidated, several studies indicate that, in the absence of hypoglycemia, prolonged hyperinsulinemia increases leptin levels (17,18). While the acute effects of insulin on leptin production in humans are controversial, there is recent evidence that physiologic levels of insulin stimulate leptin production in a dose-dependent manner within 30–60 min of exposure, as compared with saline infusion (19). In vitro, insulin stimulates leptin secretion and production from rat adipose tissue after only 10 min of exposure (20). Leptin receptors have also been found in pancreatic -cells (21), raising the possibility that leptin may modulate insulin secretion. In ob/ob mice, exogenous leptin lowers plasma insulin levels (22), and in vitro, leptin suppresses insulin release from ob/obmouse (23) and human (22) islet cells. Thus, there is evidence to suggest that leptin may play a role in the pathophysiology of diabetes, possibly by suppressing insulin secretion. Furthermore, since elevated baseline insulin is associated with both diabetes risk and elevated leptin levels, it could confound an association between leptin levels and diabetes. Most cross-sectional studies indicate that leptin levels are similar in people with and without diabetes, after adjusting for 65

Association between leptin and diabetes Table 1—Baseline characteristics of study subjects Follow-up status Baseline characteristic Men Age (years) BMI (kg/m2) Total fat (cm2) Intra-abdominal fat (cm2) Fasting leptin (ng/ml) Fasting glucose (mmol/l) 2-h glucose (mmol/l) Fasting insulin (pmol/l) Insulin resistance Women Age (years) BMI (kg/m2) Total fat (cm2) Intra-abdominal fat (cm2) Fasting leptin (ng/ml) Fasting glucose (mmol/l) 2-h glucose (mmol/l) Fasting insulin (pmol/l) Insulin resistance

Total

Diabetic

Nondiabetic

235 51.7 ± 11.7 (53.6) 25.2 ± 3.1 (25.0) 416.0 ± 171.1 (418.0) 95.9 ± 52.7 (91.3) 4.0 ± 2.7 (3.2) 5.3 ± 0.6 (5.3) 7.2 ± 1.8 (7.1) 76 ± 42 (66) 3.0 ± 1.7 (2.6) 175 51.9 ± 12.0 (53.1) 22.7 ± 3.1 (22.2) 522.0 ± 206.9 (487.5) 61.7 ± 39.0 (51.5) 11.6 ± 7.3 (10.1) 5.0 ± 0.5 (4.9) 7.2 ± 1.7 (7.3) 82 ± 42 (72) 3.1 ± 1.8 (2.6)

23 55.9 ± 10.6 (57.2) 26.8 ± 4.0 (26.9) 506.8 ± 216.0 (482.2) 128.8 ± 66.4 (129.5) 6.0 ± 4.4 (3.9) 5.7 ± 0.4 (5.8) 8.7 ± 1.1 (8.6) 94 ± 52 (72) 4.0 ± 2.1 (3.1) 17 60.3 ± 12.0 (63.3) 23.5 ± 2.6 (22.9) 584.8 ± 163.2 (633.6) 77.4 ± 32.5 (79.5) 12.3 ± 5.4 (10.6) 5.3 ± 0.4 (5.2) 9.3 ± 1.1 (9.3) 89 ± 46 (72) 3.5 ± 1.7 (2.9)

212 51.3 ± 11.8 (53.0) 25.1 ± 2.9 (24.8) 406.1 ± 163.1 (406.4) 92.3 ± 49.9 (89.2) 3.8 ± 2.3 (3.1) 5.3 ± 0.6 (5.3) 7.1 ± 1.7 (7.0) 74 ± 41 (66) 2.9 ± 1.7 (2.5) 158 51.0 ± 11.7 (47.5) 22.6 ± 3.1 (22.1) 515.2 ± 210.4 (477.1) 60.0 ± 39.3 (49.3) 11.6 ± 7.5 (9.9) 4.9 ± 0.5 (4.9) 7.0 ± 1.6 (6.9) 82 ± 42 (72) 3.1 ± 1.8 (2.6)

P value 0.0481 0.0255 0.0192 0.0088 0.0097 0.0002 0.0001 0.16 0.0280 0.0020 0.16 0.07 0.0284 0.31 0.0016 0.0001 0.60 0.19

Data are n or means ± SD (median). P values were determined by Wilcoxon’s rank-sum test, comparing subjects with and without diabetes at follow-up. Insulin resistance is based on the homeostasis model (32): insulin resistance = (fasting glucose)(fasting insulin)/22.5.

body fat (11,13,24–27). However, one study showed decreased leptin levels in obese individuals with poorly controlled diabetes (28). To our knowledge, no studies have examined the relationship between leptin and diabetes prospectively. We sought to determine whether leptin is a risk factor for development of diabetes in Japanese Americans. RESEARCH DESIGN AND METHODS Study subjects The study subjects included second(Nisei) and third-generation (Sansei) Japanese American volunteer enrollees in the Japanese American Community Diabetes Study. Details regarding participant recruitment, selection, and demographic comparison of the Nisei participants to other Japanese Americans in King County, Washington, have been previously reported (29). Exclusion criteria for this study included previous diagnosis of diabetes by a physician, a plasma glucose of 126 mg/dl after a 10-h fast, or a plasma glucose of 200 mg/dl 2 h after a 75-g oral glucose tolerance test at baseline. This study was approved by the University of Washington Institutional Review Board. 66

Measurements Subjects underwent baseline and 5- or 6year follow-up evaluations at the Clinical Research Center, University of Washington. BMI was calculated as weight in kilograms divided by height in meters squared. Plasma glucose was assayed by an automated glucose oxidase method. Diabetic status at follow-up was based on a 10-h fasting plasma glucose level of 126 mg/dl or a plasma glucose of 200 mg/dl 2 h after a 75-g oral glucose tolerance test (30). Subjects who were taking medication for diabetes were considered diabetic at follow-up. Fasting plasma insulin was measured by radioimmunoassay as previously described (31). In addition, we calculated insulin resistance as the product of fasting insulin (microunits per milliliter) and fasting glucose (millimoles per liter) divided by 22.5. This approximation of the homeostasis model assessment of insulin resistance was originally reported by Matthews et al. (32) in a mathematically equivalent form: insulin/(22.5e ln glucose). The model assumes a reference value of 1 based on normal weight nondiabetic control subjects aged 35 years, with larger values representing increased insulin resistance. It correlates well with euglycemic clamp estimates of insulin resistance (32). In several different

ethnic populations, the homeostasis model correlates more closely with Bergman’s model of insulin sensitivity than fasting insulin levels alone (33). A commercial radioimmunoassay kit, using polyclonal antiserum raised against full-length recombinant human leptin (Linco, St. Charles, MO) was used to assay total fasting plasma leptin, in duplicate. The plasma samples were stored at 80°C and thawed just before use. The intra-assay coefficient of variation was 4.8%. The interassay coefficient of variation was 7.2% for a 2.4 ng/ml sample and 5.6% for a 14.8 ng/ml sample. Additional details regarding the leptin measurements have been previously reported (34). Adipose tissue area was measured using computed tomography, since this measure is more closely correlated with leptin levels than BMI (35). Single 10-mm slices were obtained of the thorax (during inspiration, at the nipple level), the abdomen (at the level of the umbilicus), and the right thigh (halfway between the greater trochanter and the superior margin of the patella). Intra-abdominal adipose tissue was measured deep to the transversalis fascia. Total fat area was calculated as the sum of the thorax, abdominal (subcutaneous and intra-abdominal), and twice the


McNeely and Associates

right thigh measurements. Further details regarding computed tomography adipose measurements have been previously reported (36). Statistical analysis Results were mainly analyzed separately for men and women, since leptin levels were disparate in these groups. Some interaction models included both sexes. All P values presented are two-sided. Comparison of baseline measures between groups were made using the Wilcoxon’s rank-sum test, since values were not normally distributed. Spearman’s correlation coefficient was calculated for comparison of continuous variables. A univariate logistic regression model was used to estimate the relative risk of developing diabetes associated with a single standard deviation increase in baseline leptin. Multiple logistic regression models were used to test for the potential confounding factors of age, total fat, intraabdominal fat, fasting glucose, 2-h postload glucose, fasting insulin, and insulin resistance at baseline. Logistic regression was also used to test for the significance of firstorder interaction terms. We limited the number of independent variables in each model, maintaining a ratio of 10 outcome events per independent variable, to avoid overfitting the data. The assessment of interaction was used to determine whether the relationship between diabetic status at follow-up and baseline leptin level differed according to the level of an additional factor in the model, such as baseline total fat or sex. The likelihood ratio test was used to determine the statistical significance of variables and interaction terms in the logistic regression models. All statistics were calculated using Intercooled Stata software, version 5.0 for Windows 95 (Stata, College Station, TX). RESULTS Men A total of 235 men were eligible for this study. The mean age was 51.7 ± 11.7 years (mean ± SD). Men with diabetes at the time of follow-up had higher baseline leptin levels (P = 0.0097) and higher total baseline fat (P = 0.0192) than those who remained nondiabetic (Table 1). In addition, men who developed diabetes during the follow-up period had significantly higher baseline intra-abdominal fat, fasting glucose, 2-h postload glucose, fasting insulin, and insulin resistance than those

Table 2—Correlation between fasting plasma leptin level and other baseline variables Baseline variable

Spearman coefficient

Men Age BMI Total fat Intra-abdominal fat Fasting glucose 2-h glucose Fasting insulin Insulin resistance Women Age BMI Total fat Intra-abdominal fat Fasting glucose 2-h glucose Fasting insulin Insulin resistance

P value

0.0897 0.5509 0.7707 0.6197 0.2495 0.2975 0.4980 0.5348

0.17 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001

0.1395 0.7005 0.7869 0.5783 0.2547 0.2423 0.5295 0.5342

0.07 0.0001 0.0001 0.0001 0.0007 0.0012 0.0001 0.0001

Insulin resistance is based on the homeostasis model (32): insulin resistance = (fasting glucose)(fasting insulin)/22.5.

who did not develop diabetes (Table 1). Baseline leptin levels were significantly correlated with BMI, total fat, intra-abdominal fat, fasting glucose, 2-h postload glucose, fasting insulin, and insulin resistance, but not with age (Table 2). Baseline leptin levels predicted diabetes risk in the univariate logistic regression model (relative risk [RR] = 1.78 per SD [2.7 ng/ml] increase, 95% CI 1.28–2.49; Table 3). This effect was independent of total fat, fasting glucose, fasting insulin, insulin resistance, and age in separate multiple logistic regression models (Table 3). Leptin also remained a significant predictor of future diabetes after controlling for baseline 2-h postload glucose (P = 0.016), although the relative risk was slightly attenuated (Table 3). Leptin was of borderline statistical significance after adjusting for baseline intraabdominal fat (P = 0.05; Table 3). However, the relative risk adjusted for baseline intraabdominal fat should be interpreted with caution, since we also found the interaction between leptin and intra-abdominal fat to be of borderline statistical significance (P = 0.05). The interaction was such that the relationship between diabetes and elevated baseline leptin was stronger among those with large amounts of baseline intra-abdominal fat than those with lower amounts of intra-abdominal fat. In fact, despite small numbers, there was a strong association


between elevated baseline leptin levels and diabetes risk among the 23 men in the top decile of intra-abdominal fat (RR = 3.44 per SD [4.0 ng/ml] increase, 1.08–10.92). The association in this subgroup remained significant after adjusting for intra-abdominal fat (adjusted RR = 5.52 per SD [4.0 ng/ml] increase, 1.08–28.29). Of these 23 men with an intra-abdominal fat area of 165.81 cm2, 8 developed diabetes at follow-up. Among the 212 men below the 90th percentile for intra-abdominal fat, leptin did not predict diabetes risk (RR = 1.16 per SD [2.3 ng/ml] increase in leptin, 95% CI 0.73–1.85), although the number of men who developed diabetes was small (n= 15). Adjusting for intra-abdominal fat did not significantly alter these results. There was no significant first-order interaction between leptin and total fat, fasting glucose, 2-h postload glucose, fasting insulin, insulin resistance, or age. Women A total of 175 women were eligible for this study. The mean age was 51.9 ± 12.0 years (mean ± SD). Women with diabetes at the time of follow-up had similar baseline leptin levels (P = 0.31) and tended to have higher total fat (P = 0.07) compared with those who remained nondiabetic (Table 1). Women who developed diabetes had significantly higher intra-abdominal fat, fasting glucose, and 2-h postload glucose 67

Association between leptin and diabetes

Table 3—Association between baseline leptin level and risk of developing type 2 diabetes al. (15) reported an association between Baseline variables in the logistic regression model Men Leptin Total fat Intra-abdominal fat Fasting glucose 2-h glucose Fasting insulin Insulin resistance Age Women Leptin Total fat Intra-abdominal fat Fasting glucose 2-h glucose Insulin Insulin resistance Age

Relative risk for leptin (95% CI) 1.78 (1.28–2.49) 1.80 (1.02–3.17) 1.50 (1.00–2.24)* 1.75 (1.23–2.51) 1.54 (1.09–2.20) 1.71 (1.15–2.54) 1.62 (1.09–2.39) 1.84 (1.30–2.59) 1.10 (0.68–1.77) 0.66 (0.30–1.48) 0.84 (0.46–1.54) 0.86 (0.52–1.44) 0.88 (0.49–1.58) 1.01 (0.56–1.83) 0.94 (0.52–1.70) 1.06 (0.63–1.80)

Independent variables were entered as continuous values. Relative risk is per SD (2.7 ng/ml for men, 7.3 ng/ml for women) increase in baseline leptin level. Insulin resistance is based on the homeostasis model (32): insulin resistance = (fasting glucose)(fasting insulin)/22.5. *Does not account for interaction between leptin and intra-abdominal fat; see text for details.

compared with those who did not develop diabetes. Interestingly, fasting insulin levels and insulin resistance were similar between the two groups. As in men, baseline leptin levels were significantly correlated with baseline BMI, total fat, intra-abdominal fat, fasting glucose, 2-h postload glucose, fasting insulin, and insulin resistance, but not with age (Table 2). Logistic regression analyses were consistent with the univariate baseline comparisons. Baseline leptin levels did not predict follow-up diabetes status even after controlling for baseline total fat, intraabdominal fat, fasting glucose, 2-h postload glucose, fasting insulin, insulin resistance, or age (Table 3). Unlike in men, leptin levels did not predict diabetes risk among the 18 women in the top decile for intraabdominal fat (RR = 0.93 per SD [7.2 ng/ml] increase, 0.20–4.40), although the numbers were small. Baseline leptin levels also did not predict diabetes risk in a subgroup of 83 postmenopausal women aged 55, 13 of whom had diabetes at followup (unadjusted RR = 0.90 per SD [6.7 ng/ml] increase, 0.48–1.68, P = 0.74). Effect of sex To explore whether the relationship between leptin and diabetes risk varied 68

with sex, we used a multiple logistic regression model to estimate the risk of developing diabetes, including data from the 410 men and women combined. The independent variables in the model included: leptin, total fat, sex, and a leptin-sex interaction term. The coefficient of the firstorder multiplicative interaction term between baseline leptin level and sex was significantly 0 (coefficient = 0.17, P = 0.0295), indicating that the association between leptin and diabetes risk was greater in men than in women. CONCLUSIONS — To our knowledge, this is the first prospective study to demonstrate that elevated leptin is a risk factor for development of type 2 diabetes in men. These findings were independent of baseline total fat, fasting glucose, 2-h postload glucose, fasting insulin, insulin resistance, and age. The relationship was most pronounced among those with higher amounts of intra-abdominal fat. Most cross-sectional studies have shown no difference in leptin levels between diabetic and nondiabetic individuals of similar body weight (11,13,24–27,34). However, there is a single report of reduced leptin levels in obese individuals with poorly controlled diabetes (28). Hanley et

increased leptin levels and glucose intolerance in women, whereas other studies showed no such association (12,37). Our study differed from previous studies of leptin levels in diabetes in several ways. Our sample size was larger than most other studies. We measured body fat using computed tomography, whereas many others used BMI (11,13,25,26), a measure that correlates less closely with leptin levels (35). Considering the borderline lower limit of our adjusted 95% CI for the leptin-diabetes association in men, it is not surprising that previous studies with considerably less statistical power did not show significant results. Another important feature of our study is the prospective design. There appears to be a complex, yet incompletely understood, relationship between insulin and leptin (17–19,22). Results of cross-sectional studies might be affected by variability in disease duration, severity, and treatment among those with diabetes. Furthermore, in a cross-sectional study, it cannot usually be ascertained whether a potential cause (e.g., leptin) preceded the effect of interest (e.g., type 2 diabetes). There is less chance for these types of potential bias in prospective studies. For example, if insulin stimulates leptin production in humans (17–19), then variation in the degree of -cell dysfunction within the diabetic group might obscure differences in leptin levels compared with control subjects. Consistent with this possibility is the recent report that obese patients with poorly controlled diabetes and evidence of reduced -cell function had lower leptin levels than normoglycemic obese control subjects (28). In the same study, obese subjects with less cell dysfunction and controlled diabetes had leptin levels only slightly lower than those of control subjects. If all subjects with diabetes had been combined into a single group in that study, the difference in leptin levels compared with control subjects would have been attenuated. The causal pathway by which baseline leptin levels are associated with increased diabetes risk in men remains unknown. Elevated leptin levels suppress insulin secretion in vitro (22,23), and this represents one potential mechanism. Another possibility is that leptin levels increase diabetes risk through weight gain. Chessler et al. (34) recently reported that elevated baseline leptin levels predicted increased


McNeely and Associates

BMI and adiposity at 5-year follow-up in Japanese Americans. It should be noted that the association between high leptin levels and subsequent weight gain has not been described in other populations. In fact, Pima Indians who gained at least 3.0 kg/year had lower baseline leptin levels than those whose weight remained stable (38). Unlike Japanese Americans, Pima Indians are especially prone to obesity, and it remains to be seen how these findings compare with studies in other populations. Nevertheless, our study design does not allow definitive testing of these hypotheses, since we do not know the time of diabetes onset in relation to changes in insulin secretion or body fat. The mechanism by which intraabdominal fat modifies the effect of elevated leptin on diabetes risk in men is also unclear. Leptin levels are determined by total adiposity rather than regional fat depots, such as intra-abdominal fat (35,39). Increased intra-abdominal fat is a risk factor for diabetes, even after adjusting for BMI and subcutaneous fat (40). One possibility is that men with the predisposition to gain intra-abdominal fat are at an increased risk of doing so when they gain weight, thereby placing them at higher risk for diabetes. Thus elevated leptin levels in such men results in weight gain characterized by increased intra-abdominal fat and may possibly account for the diabetes-leptin association in these men. Again, additional studies will be needed to test this hypothesis, since we cannot determine the time of diabetes onset relative to intraabdominal fat accumulation. In our study, leptin levels predicted risk of diabetes in men but not in women. Unlike in men, there was no statistically significant difference in baseline adiposity, as measured by BMI or total fat, between women who developed diabetes and those who remained nondiabetic. Since leptin levels are thought to reflect overall adiposity, it is not surprising that baseline leptin levels were also similar between these two groups. Although there were fewer women than men in our study, which represents, therefore, less statistical power, our finding of a significant interaction between sex and leptin provides evidence that the association between diabetes risk and leptin is stronger in men than in women. Our study does not address the reason for this difference, however it does not appear to be related to menopausal status. While we combined pre- and postmenopausal

women in our main analysis, adjusting for age did not unmask an underlying effect of leptin on diabetes risk. Limiting the analysis to a subgroup of women aged 55 confirmed the absence of a leptin-diabetes association in postmenopausal women. There are several potential limitations to our study. Computed tomography is probably not as precise a measure of adiposity as underwater weighing. While any such measurement error is likely to be random, even nonsystematic error could theoretically limit our ability to control for baseline adiposity (41). However, it seems unlikely that our findings are explained by measurement error. Computed tomography measures of adiposity correlate closely with leptin levels (35). If the association between leptin levels and diabetes risk in men is due to baseline adiposity, then controlling for adiposity using a reasonable but imperfect measure should result in an attenuated relative risk. In fact, the relative risk adjusted for total fat in men remained unchanged. Likewise, it does not appear that our findings are due to confounding by age, fasting glucose, fasting insulin, or insulin resistance at baseline. Of note, the relative risk of subsequent diabetes was slightly attenuated after adjustment for baseline 2-h postload glucose, but remained statistically significant. This indicates the findings cannot be entirely accounted for by elevated 2-h postload glucose levels at baseline. The attenuation in the relative risk adjusted for baseline intra-abdominal fat (Table 3) was due to interaction between leptin and intra-abdominal fat. The relationship between leptin and diabetes risk was especially prominent among men in the top decile for intra-abdominal fat, even after adjustment for baseline intra-abdominal fat. Thus, our findings are not explained by confounding due to baseline intraabdominal fat. However, it is possible that other, currently unrecognized, confounding variables explain the increased risk of diabetes associated with elevated baseline leptin levels in men. Finally, it is unknown whether these findings are generalizable to other populations. In summary, this study demonstrates that an elevated leptin level is a risk factor for the development of diabetes in Japanese American men, but not in Japanese American women. This finding was independent of baseline total fat, fasting glucose, 2-h postload glucose, fasting insulin, insulin resistance, and age. The magnitude of the association was stronger in men with


high baseline intra-abdominal fat compared with those with lower amounts of intra-abdominal fat. The mechanism by which leptin is associated with diabetes risk remains to be determined. Possible explanations are that elevated baseline leptin is a risk factor for future weight and adiposity gain (34), or that elevated leptin suppresses insulin secretion (22,23). Although the association between sex, body fat, insulin, and leptin appears consistent across many ethnic groups (10,12,14,15), confirmatory studies in other populations would be desirable. Acknowledgments — This work was supported by National Institutes of Health Grants DK-31170 and HL-49293. Facility support was provided by the Clinical Nutrition Research Unit (DK-35816), the Diabetes and Endocrinology Research Center (DK-17047), and the General Clinical Research Center (RR-00037) at the University of Washington. Additional support was provided by the ZymoGenetics Corporation. Preliminary results of this study were presented in abstract form at the 58th Annual Meeting and Scientific Sessions of the American Diabetes Association, Chicago, Illinois, 13–16 June 1998.

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. Nature372:425–432, 1994 2. Halaas JL, Gajiwala KS, Maffei M, Cohen SL, Chait BT, Rabinowitz D, Lallone RL, Burley SK, Friedman JM: Weight-reducing effects of the plasma protein encoded by the obese gene. Science269:543–546, 1995 3. Jacob RJ, Dziura J, Medwick MB, Leone P, Caprio S, During M, Shulman GI, Sherwin RS: The effect of leptin is enhanced by microinjection into the ventromedial hypothalamus. Diabetes46:150–152, 1997 4. Clement K, Vaisse C, Lahlou N, Cabrol S, Pelloux V, Cassuto D, Gourmelen M, Dina C, Chambaz J, Lacorte J-M, Basdevant A, Bougneres P, Lebouc Y, Froguel P, GuyGrand B: A mutation in the human leptin receptor gene causes obesity and pituitary dysfunction. Nature 392:398–401, 1998 5. Montague CT, Farooqi S, Whitehead JP, Soos MA, Rau H, Wareham NJ, Sewter CP, Digby JE, Mohammed SN, Hurst JA, Cheetham CH, Earley AR, Barnett AH, Prins JB, O’Rahilly S: Congenital leptin deficiency is associated with severe early-onset obesity in humans. Nature 387:903–908, 1997 6. Hasstedt SJ, Hoffman M, Leppert MF, Elbein SC: Recessive inheritance of obesity 69

Association between leptin and diabetes

7.

8.

9. 10.

11.

12.

13.

14.

15.

16.

17.

70

in familial non-insulin-dependent diabetes mellitus, and lack of linkage to nine candidate genes. Am J Hum Genet61:668–677, 1997 Considine RV, Considine EL, Williams CJ, Hyde TM, Caro JF: The hypothalamic leptin receptor in humans: identification of incidental sequence polymorphisms and absence of the db/db mouse and fa/fa rat mutations. Diabetes45:992–994, 1996 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 Med334:292–295, 1996 Caro JF, Sinha MK, Kolaczynski JW, Zhang PL, Considine RV: Leptin: the tale of an obesity gene. Diabetes45:1455–1462, 1996 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 313:965–969, 1996 Malmstrom R, Taskinen MR, Karonen SL, Yki-Jarvinen H: Insulin increases plasma leptin concentrations in normal subjects and patients with NIDDM. Diabetologia 39:993–996, 1996 Saad MF, Damani S, Gingerich RL, RiadGabriel MG, Khan A, Boyadjian R, Jinagouda SD, El-Tawil K, Rude RK, Kamdar V: Sexual dimorphism in plasma leptin concentration. J Clin Endocrinol Metab 82:579–584, 1997 Schwartz MW, Prigeon RL, Kahn SE, Nicolson M, Moore J, Morawiecki A, Boyko EJ, Porte D Jr: Evidence that plasma leptin and insulin levels are associated with body adiposity via different mechanisms. Diabetes Care 20:1476–1481, 1997 Widjaja A, Stratton IM, Horn R, Holman RR, Turner R, Brabant G: UKPDS 20: plasma leptin, obesity, and plasma insulin in type 2 diabetic subjects. J Clin Endocrinol Metab82:654–657, 1997 Hanley AJ, Harris SB, Gao XJ, Kwan J, Zinman B: Serum immunoreactive leptin concentrations in a Canadian aboriginal population with high rates of NIDDM. Dia betes Care20:1408–1415, 1997 Mohamed-Ali V, Pinkney JH, Panahloo A, Goodrick S, Coppack SW, Yudkin JS: Relationships between plasma leptin and insulin concentrations, but not insulin resistance, in non-insulin-dependent (type 2) diabetes mellitus. Diabet Med14:376–380, 1997 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. Dia betes45:699–701, 1996

18. Schmitz O, Fisker S, Orskov L, Hove KY, Nyholm B, Moller N: Effects of hyperinsulinaemia and hypoglycaemia on circulating leptin levels in healthy lean males. Diabetes Metab23:80–83, 1997 19. Saad MF, Khan A, Sharma A, Michael R, 31. Riad-Gabriel MG, Boyadjian R, Jinagouda SD, Steil GM, Kamdar V: Physiological insulinemia acutely modulates plasma leptin. Diabetes47:544–549, 1998 20. Barr VA, Malide D, Zarnowski MJ, Taylor SI, 32. Cushman SW: Insulin stimulates both leptin secretion and production by rat white adipose tissue. Endocrinology138:4463–4472, 1997 21. Kieffer TJ, Heller RS, Habener JF: Leptin receptors expressed on pancreatic beta- 33. cells. Biochem Biophys Res Commun 224:522–527, 1996 22. Kulkarni RN, Wang ZL, Wang RM, Hurley JD, Smith DM, Ghatei MA, Withers DJ, Gardiner JV, Bailey CJ, Bloom SR: Leptin rapidly suppresses insulin release from insulinoma 34. cells, rat and human islets and, in vivo, in mice. J Clin Invest100:2729–2736, 1997 23. Kieffer TJ, Heller RS, Leech CA, Holz GG, Habener JF: Leptin suppression of insulin secretion by the activation of ATP-sensitive 35. K channels in pancreatic -cells. Diabetes 46:1087–1093, 1997 24. Kennedy A, Gettys TW, Watson P, Wallace P, Ganaway E, Pan Q, Garvey WT: The metabolic significance of leptin in humans: 36. gender-based differences in relationship to adiposity, insulin sensitivity, and energy expenditure. J Clin Endocrinol Metab 82:1293–1300, 1997 25. McGregor GP, Desaga JF, Ehlenz K, Fischer 37. A, Heese F, Hegele A, Lammer C, Peiser C, Lang RE: Radiommunological measurement of leptin in plasma of obese and diabetic human subjects. Endocrinology 137:1501–1504, 1996 38. 26. Haffner SM, Mykkanen L, Stern MP: Leptin concentrations in women in the San Antonio Heart Study: effect of menopausal status and postmenopausal hormone replacement therapy. Am J Epidemiol146:581–585, 1997 39. 27. de Courten M, Zimmet P, Hodge A, Collins V, Nicolson M, Staten M, Dowse G, Alberti KG: Hyperleptinaemia: the missing link in the metabolic syndrome? Diabet Med 14:200–208, 1997 28. Clement K, Lahlou N, Ruiz J, Hager J, 40. Bougneres P, Basdevant A, Guy-Grand B, Froguel P: Association of poorly controlled diabetes with low serum leptin in morbid obesity. Int J Obes21:556–561, 1997 29. Fujimoto WY, Leonetti DL, Kinyoun JL, Newell-Morris L, Shuman WP, Stolov WC, Wahl PW: Prevalence of diabetes mellitus 41. and impaired glucose tolerance among second-generation Japanese-American men. Diabetes36:721–729, 1987 30. The Expert Committee on the Diagnosis

and Classification of Diabetes Mellitus: Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care 20:1183–1197, 1997 Boyko EJ, Leonetti DL, Bergstrom RW, Newell-Morris L, Fujimoto WY: Visceral adiposity, fasting plasma insulin, and blood pressure in Japanese-Americans. Diabetes Care 18:174–181, 1995 Matthews DR, Hosker JP, Naylor BA, Treacher DF, Turner RC: Homeostasis model assessment: insulin resistance and cell function from fasting plasma glucose and insulin concentrations in man. Dia betologia28:412–419, 1985 Howard G, Bergman R, Wagenknecht LE, Haffner SM, Savage PJ, Saad MF, Laws A, D’agostino RB: Ability of alternative indices of insulin sensitivity to predict cardiovascular risk: comparison with the “minimal model.” Ann Epidemiol8:358–369, 1998 Chessler SD, Fujimoto WY, Shofer JB, Boyko EJ, Weigle DS: Increased plasma leptin levels are associated with fat accumulation in Japanese Americans. Diabetes 47:239–243, 1998 Weigle DS, Ganter SL, Kuijper JL, Leonetti DL, Boyko EJ, Fujimoto WY: Effect of regional fat distribution and Prader-Willi syndrome on plasma leptin levels. J Clin Endocrinol Metab82:566–570, 1997 Shuman WP, Morris LL, Leonetti DL, Wahl PW, Moceri VM, Moss AA, Fujimoto WY: Abnormal body fat distribution detected by computed tomography in diabetic men. Investigative Radiology21:483–487, 1986 Ramachandran A, Snehalatha C, Vijay V, Satyavani K, Latha E, Haffner SM: Plasma leptin in non-diabetic Asian Indians: association with abdominal adiposity. Diabet Med14:937–941, 1997 Ravussin E, Pratley RE, Maffei M, Wang H, Friedman JM, Bennett PH, Bogardus C: Relatively low plasma leptin concentrations precede weight gain in Pima Indians. Nat Med3:238–240, 1997 Dua A, Hennes MI, Hoffmann RG, Maas DL, Krakower GR, Sonnenberg GE, Kissebah AH: Leptin: a significant indicator of total body fat but not of visceral fat and insulin insensitivity in African-American women. Diabetes45:1635–1637, 1996 Bergstrom RW, Newell-Morris LL, Leonetti DL, Shuman WP, Wahl PW, Fujimoto WY: Association of elevated fasting C-peptide level and increased intra-abdominal fat distribution with development of NIDDM in Japanese-American men. Diabetes 39:104– 111, 1990 Rothman KJ, Greenland S: Precision and validity in epidemiologic studies. In Modern Epidemiolog. y2nd ed. Rothman KJ, Greenland S, Eds. Philadelphia, LippincottRaven, 1998, p. 115–134
