Insulin sensitivity and secretion in healthy elderly human subjects with ''abnormal'' glucose tolerance

Glucose tolerance deteriorates dramatically with advancing age. It is not known whether the underlying pathophysiology is different in older subjects. We employed a two step hyperinsulinaemic euglycaemic glucose clamp with [6(14)C] glucose infusion to compare peripheral and hepatic insulin sensitivity in eight elderly (EAGT) with eight young (YAGT) subjects with abnormal (matched) glucose tolerance and nine elderly subjects with normal glucose tolerance (ENGT). There was no difference in basal HGO (EAGT 14.5 +/- 0.9, YAGT 15.3 +/- 1.1 mumol kg-1 min-1). Glucose turnover was similar in both groups at step 1 (EAGT 13.2 +/- 0.8, YAGT 13.4 +/- 0.8 mumol kg-1 min-1) and step 2 (EAGT 25.1 +/- 3.1, YAGT 27.2 +/- 2.7 mumol kg-1 min-1). HGO was lower in the EAGT subjects at step 1 (2.3 +/- 0.4 vs. 4.3 +/- 0.6 mumol kg-1 min-1 P = 0.01). Incremental serum insulin response to oral glucose was comparable (EAGT 66.8 +/- 11.6 YAGT 57.8 +/- 12.2 mU l-1.h). Compared to the ENGT group the EAGT group was insulin resistant with a lower MCR of glucose at step 1 (2.03 +/- 0.28 vs. 3.23 +/- 0.44 ml kg-1 min-1 P = 0.04) and at step 2 (6.18 +/- 0.83 vs. 9.64 +/- 0.38 ml kg-1 min-1 P = 0.004) and had a lower early insulin response (AUC 0-30 min 5.9 +/- 1.1 vs. 9.8 +/- 1.4 mU l-1.h P = 0.04).(ABSTRACT TRUNCATED AT 250 WORDS)     

The effects of the menopause on insulin sensitivity, secretion and elimination in non-obese, healthy women

Abstract. We have carried out intravenous glucose tolerance tests with measurement of plasma glucose, insulin and C-peptide concentrations on 66 premeno-pausal and 92 postmenopausal non-obese Caucasian women. After adjustment for the effects of a number of possible confounding variables, including age and body mass index, there was little difference between pre and postmenopausal women in glucose and insulin concentrations either fasting or in response to intravenous glucose. Mathematical modelling analysis of the resultant plasma concentration profiles was used to obtain measures of insulin sensitivity, secretion and elimination, and non-insulin dependent glucose disposal. We found reciprocal differences in mean insulin sensitivity (increased by 50%) and non-insulin dependent glucose disposal (decreased by 30%). Plasma C-peptide response and pancreatic insulin secretion were markedly lower in the postmenopausal group (- 35% and -50% respectively). However, the rate constant for insulin elimination was also lower in these women. As a result, intravenous glucose tolerance test plasma insulin concentrations were not significantly different between the two groups. We conclude that, despite the occurrence of little or no variation in plasma glucose and insulin concentrations, the menopause is associated with significant changes in insulin metabolism.     

Effects of burn injury on insulin secretion and on sensitivity to insulin in the rat in vivo.

Both insulin resistance and impairment of insulin secretion are know to occur in man after injury. The relative importance of these effects was studied in rats 2 h after a non-lethal 20 percent dorsal scald. No impairment of insulin secretion was found after this injury. Concentrations of both blood glucose and plasma insulin were elevated in scalded rats. Scalded rats responded to intravenous glucose injection (1-0 g/kg) with a further rise in plasma insulin concentration, which remained normal for the prevailing blood glucose concentration. However, marked impairment of glucose tolerance was observed, indicating the presence of insulin resistance. After intravenous insulin injection (1-0 U/kg) the initial rate coefficient for fall of blood glucose concentration was significantly lower (p less than 0-02) in scalded (mean 3-9 percent min.(-1) than in control rats (mean 6-3 percent min.(-1). The minimum in blood glucose concentration after insulin injection was reached at 10 min. in control rats, but not until 60 min. after injection in scalded rats. This difference was due to a delay in compensation for the hypoglycaemia in the scalded rats, since the rate of disappearance of insulin measured by injection of a tracer of 125I-labelled bovine insulin was not decreased after this injury. It was concluded that the impairment of glucose utilization in scalded rats (Heath and Corney, 1973) is due to decreased sensitivity to insulin rather than to suppression of insulin release.     

Decreased insulin sensitivity and muscle enzyme activity in elderly subjects

Skeletal muscle glycogen deposition, and the activation of muscle glycogen synthase and pyruvate dehydrogenase during a hyerinsulinaemic euglycaemic clamp have been measured in six young and six elderly males matched for body mass index, physical activity and diet. Clamp glucose requirement (insulin, 0.1 U kg-1 h-1) was significantly lower in the older subjects (8.0 +/- 0.4 mg kg-1 min-1) than in younger subjects (10.5 +/- 0.6 mg kg-1 min-1, P less than 0.02). Although the older subjects had a 6.5% decrease in lean body mass, clamp glucose requirement expressed per unit of lean body mass was also significantly decreased in the older subjects (10.2 +/- 0.5 vs. 12.4 +/- 0.6 mg kg-1 min-1, P less than 0.05). The increase in muscle glycogen with the clamp was decreased by 33% in the older subjects (elderly: 13.1 +/- 1.3 mg g-1 protein, young: 19.6 +/- 2.2 mg g-1 protein; P less than 0.05), and was strongly correlated with clamp glucose requirement (r = 0.72, P less than 0.01). Glucose-6-phosphate independent glycogen synthase activity increased significantly between fasting and the end of the clamps in both groups (P less than 0.001), but was lower at the end of the clamp in the older subjects (P less than 0.05). Glycogen synthase activity at the end of the clamp correlated with both clamp glucose requirement (r = 0.83, P less than 0.01) and muscle glycogen deposition (r = 0.73, P less than 0.01).(ABSTRACT TRUNCATED AT 250 WORDS)     

Insulin and glucose uptake and insulin sensitivity in patients with peripheral arterial insufficiency

Abstract. There have been previous reports of evidence for increased insulin sensitivity in patients with intermittent claudication as the only symptom of arterial insufficiency. This study was designed to evaluate the role of peripheral tissue with respect to insulin sensitivity in such patients. Intravenous glucose tolerance tests (IVGTT) and intravenous insulin tolerance tests (IVITT) were performed in patients with peripheral arterial insufficiency and in controls. During IVGTT the plasma insulin values were significantly lower in patients with arterial insufficiency. During IVITT the venous concentration of glucose decreased more and the area over the glucose curve was significantly larger, suggesting a higher insulin sensitivity in this group. The low insulin values could not be referred to a larger distribution volume or an increased degradation rate, suggesting decreased pancreatic insulin release in this group. At the time of surgery for arterial insufficiency and for varicose veins in controls, the uptake of insulin and glucose and the release of lactate were measured over the leg before and after a glucose load. The uptake of insulin over the leg correlated positively with the arterial insulin concentration and the uptake of glucose in both groups. The insulin uptake had a tendency to be increased in legs with arterial insufficiency during the glucose challenge. The glucose uptake in the leg did not differ in the basal state, but was 3 times higher in the legs of patients with arterial insufficiency during glucose challenge. The increased glucose uptake in this group could be ascribed to a high insulin sensitivity in the leg, as defined by glucose uptake per unit of insulin taken up.     

糖耐量减低者胰岛素抵抗、胰岛B细胞功能及相关代谢指标分析

目的分析糖耐量减低者的胰岛素抵抗、胰岛B细胞功能及相关代谢指标。方法根据口服葡萄糖耐量试验(OGTT),将243例入选者分为糖耐量减低组(IGT组,108例)及糖耐量正常组(NGT组,135例),测定空腹及服糖后2小时胰岛素(2 hINS)、空腹甘油三酯(TG)、胆固醇(TC)、高密度脂蛋白胆固醇(HDL-C)、低密度脂蛋白胆固醇(LDL-C)、游离脂肪酸(FFA),计算胰岛素抵抗指数(HOMA-IR)、胰岛素分泌指数(HOMA-β)及葡萄糖处置指数(DI)。结果①IGT组HOMA-IR水平显著高于NGT组(0.56±0.25 vs 0.40±0.20,P<0.01);②IGT组DI水平显著低于NGT组(1.66±0.16 vs 2.66±0.21,P<0.01);③IGT组TG、FFA、FINS2、hINS显著高于NGT组(P<0.01),HDL-C显著低于NGT组(P<0.01);两组间TC、LDL-C、HOMA-β差异无统计学意义(P>0.05)。结论糖耐量减低者以胰岛素抵抗及高胰岛素血症为主,脂毒性在这一过程中起重要作用。     

Diurnal triglyceride profiles in healthy normolipidemic male subjects are associated to insulin sensitivity, body composition and diet

BACKGROUND: Elevated fasting and postprandial triglycerides (TG) are established risk factors for Coronary Heart Disease (CHD). Usually, fasting plasma TG are measured, although TG are mainly produced in a postprandial state. Our objective was to investigate diurnal TG profiles using serial capillary TG measurements, in normolipidemic healthy males. MATERIALS AND METHODS: Forty-eight, non-obese, non-smoking males (range: 20-55 years, mean age: 32 +/- 12 years), measured diurnal capillary TG, at six fixed timepoints during the day on three different days and recorded their food intake. Insulin sensitivity was estimated by HOMA. Diurnal capillary TG profiles were calculated as integrated area under the mean capillary TG curve (TGc-AUC). RESULTS: All subjects had normal fasting plasma TG and cholesterol. The average TGc-AUC was 23.6 +/- 6.7 mmol h L-1. Significant correlations with TGc-AUC were: fasting insulin (r = 0. 40, P < 0.005), HOMA (r = 0.32, P < 0.05), relative fat mass (r = 0. 31, P < 0.05), dietary protein-(r = 0.31, P < 0.05) and saturated fat intake (r = 0.30, P < 0.05). Age was not associated to diurnal triglyceridemia. After subdividing the group into quartiles on the base of TGc-AUC, differences were found between the highest (n = 12) and lowest quartile (n = 12) for: fasting capillary TG, fasting insulin, HOMA and systolic blood pressure. Fasting plasma TG and dietary intake were not different. CONCLUSION: Diurnal TG profiles in healthy normolipidemic males are not age-dependent, but are associated to insulin sensitivity, fat mass and diet. Diurnal capillary TG profiles may be a valuable additional tool in estimating a risk profile for CHD since significant differences in diurnal TG are not always reflected by elevated fasting plasma TG.     

Beta-cell function and insulin sensitivity contribute to the shape of plasma glucose curve during an oral glucose tolerance test in non-diabetic individuals

To clarify whether beta-cell function and/or insulin resistance contributes to the shape of plasma glucose curve during an oral glucose tolerance test (OGTT), we investigated 583 Japanese subjects with normal glucose tolerance (NGT, n = 306) or impaired glucose tolerance (IGT, n = 277). Each subject was subdivided into three shapes of plasma glucose curve as follows: monophasic pattern (M type), biphasic pattern (B type) and two peaks (T type). Homeostasis model assessment of insulin resistance, quantitative insulin sensitivity check index and insulinogenic index were assessed by plasma glucose and insulin concentrations obtained at fasting or during an OGTT. There was a greater proportion of M type in the IGT group (M = 80.9%, B = 15.5% and T = 3.6%), whereas the prevalence of B and T types was much higher in the NGT group (M = 66.6%, B = 26.5% and T = 6.9%). There were significant differences in the proportions of shape types between the NGT and IGT groups (p = 0.0006). Among the NGT category, insulin sensitivity was significantly higher in the B type than in the M type, and beta-cell function adjusted for insulin resistance was significantly higher in the B and T types than in the M type. Among the IGT category, no significant differences were seen among the three shape types with respect to insulin sensitivity, but the beta-cell function adjusted for insulin resistance was significantly lower in the M type than in the B and T types. In conclusion, both impaired insulin secretion and insulin resistance may contribute to the underlying mechanisms of the shape of plasma glucose curve in Japanese subjects.     

Capillary permeability in healthy males with different insulin response to glucose

The capillary permeability in human skeletal muscle, expressed as capillary diffusion capacity (CDC) for 51Cr-EDTA, was determined during exercise with a local clearance method in two groups of healthy male subjects: a younger group with a mean age of 30 years and an older one with a mean age of 58 years. The main finding was that, in both age groups, CDC was significantly negatively correlated to the early phase of insulin response to an intravenous glucose load. No correlation was found between CDC and peripheral insulin sensitivity. CDC was significantly higher in the older age group (P less than 0.01) independent of insulin response, blood flow and body weight. These data indicate that subjects with low insulin response to glucose have a higher capillary permeability than high responders, and that muscle capillary permeability increases with age.     

Smoking cessation improves insulin sensitivity in healthy middle-aged men

Cigarette smokers have recently been shown to exhibit insulin resistance, dyslipidaemia and markers of the insulin resistance syndrome (IRS). The aim of this study was to examine the effects of smoking cessation on insulin sensitivity and IRS. Forty male, non-obese healthy smokers participated in this open parallel study with 8 weeks of follow-up. Seventeen subjects were able to stop smoking, while 23 subjects continued to smoke and served as a controls group. Anthropometric and metabolic data were measured. Degree of insulin sensitivity was determined with the euglycaemic hyperinsulinaemic clamp technique. Smoking cessation increased insulin sensitivity and improved the lipoprotein profile in spite of a modest increase in body weight. Initial smoking habits correlated positively with the increase in BMI as well as the improvements in the metabolic variables after smoking cessation. These data support the view that smoking causes insulin resistance and IRS, and also demonstrate that the beneficial metabolic effects of smoking cessation override the effects of an accompanying modest increase in body weight.     

腹型肥胖人群中不同糖代谢水平者胰岛素抵抗及胰岛分泌功能的探讨

目的 探讨腹型肥胖人群中不同糖代谢水平者胰岛素抵抗及胰岛分泌功能的状态。方法 腹型肥胖患者共382例，其中正常糖耐量(NGT)组251例，空腹血糖异常(IFG)组40例，异常糖耐量(IGT)组41例，2型糖尿病(DM组)50例。测腰围、血压、空腹血脂、血糖及血浆胰岛素，应用稳态模式胰岛素抵抗指数(HOMA-IR)作为胰岛素抵抗指标，稳态模式胰岛B细胞功能指数(HBCI)作为胰岛素分泌指标。结果 在腹型肥胖的人群中，不同糖代谢组的HOMA-IR差异具有统计学意义，从NGT→IFG／IGT→DM组HOMA-IR逐渐升高，而HBCI在各组内变化较大，数值分布较分散，与NGT、IFG、IGT组相比，DM组的HBCI明显下降，差别有统计学意义。同时收缩压随着糖代谢的恶化从NGT、IGT IFG到DM组逐步升高，舒张压的变化无明显的规律。结论 腹型肥胖人群中，从NGT经IFG／IGT向2型糖尿病发展的过程中，胰岛素敏感性逐渐下降，β细胞胰岛素分泌功能明显下降是出现DM的主要原因。这一结果与其他种族中的研究结果不完全相同，提示即使是肥胖相关的2型糖尿病，种族、遗传因素仍然在糖尿病发展过程中发挥着重要作用。     

不同糖耐量人群胰岛素抵抗的比较

【目的】研究不同糖耐量人群胰岛素抵抗程度的差异。【方法】将412名门诊患者按OGTT分为4组：糖尿病组（NDM，n=180），空腹血糖受损组（IFG，n=35），糖耐量异常组（IGT，n=46），糖耐量正常组（NGT，n=151）。测定血压、血脂（TG和HDL）、体质指数（BMI）；应用胰岛素抵抗（HOMA-IR）及胰岛素作用指数（IAI）对不同糖耐量人群进行测定。【结果】DM、IFG及IGT组均较NGT组IAI下降，HOMA-IR增高，DM组表现得尤为显著；而IFG组与IGT组比较，亦有显著的IAI下降及HOMA-IR增高（P〈0．05）。【结论】不同糖耐量人群随着糖调节不同程度的受损HOMA-IR和IAI均有增幅变化，这两种指标可较准确的评估胰岛素敏感性。IFG与IGT人群胰岛素抵抗的机制可能有所不同。     

The human insulin analogue insulin lispro

Insulin lispro is a newly developed analogue of human insulin where the positions of the amino acids lysine and proline have been switched at the end of the B chain of the insulin molecule. Insulin lispro with lysine at position B28 and proline at position B29 has a weaker tendency for self-association than human insulin. This leads to three major differences in the pharmacokinetics: the action begins faster, has a higher peak and the duration is shorter than with human insulin. Thus, insulin lispro has a more precise action profile for the mealtime than human regular insulin. Insulin lispro is recommended to be injected within 15 min before the meal in contrast to 30–40 min for human insulin. In clinical trials with insulin lispro, the postprandial rise of blood glucose is smaller, the rate of hypoglycaemia is lower particularly at night-time, the need for snacks is smaller and the patient preference is better than with human insulin. The long-term control as reflected by an improvement in the HbA_]c level is better with insulin lispro than with human regular insulin, provided that an appropriate basal insulin regimen is used to take into account the shorter duration of action. A few patients have been described who have a severe resistance to human insulin but who have been succesfully treated with insulin lispro. Insulin lispro was designed to be used as a mealtime insulin, and it is a step forward in the treatment of diabetic patients using a basal-bolus insulin regimen.     

Hypoglycemic effect of aqueous shallot and garlic extracts in rats with fructose-induced insulin resistance

The present study has been carried out to investigate the effect of aqueous extract of shallot (Allium ascalonicum) and garlic (Allium satium) on the fasting insulin resistance index (FIRI) and intraperitoneal glucose tolerance test (IPGTT) of fructose-induced insulin resistance rats. Male albino Wistar rats were fed either normal or high-fructose diet for a period of eight weeks. Fasting blood glucose level, fasting blood triglyceride level, FIRI, and the area under the glucose tolerance curve were significantly elevated in fructose-fed animals. Fructose-induced insulin resistance rats treated by aqueous shallot or garlic extract (500 mg/kg body weight/day, i.p.) for duration of eight weeks. Control animals only received normal saline (0.9%). The results showed that neither shallot nor garlic extracts significantly altered the FIRI and the IPGTT at the fourth week after treatment. The fasting blood glucose in fructose-induced insulin resistance animals has been significantly decreased in 8-week treated animals by both shallot and garlic extracts. Shallot extract administration, but not garlic extract, for a period of eight weeks can significantly improve the intraperitoneal glucose tolerance and diminish the FIRI. These results indicate that shallot and garlic extracts have a hypoglycemic influence on the fructose-induced insulin resistance animals and aqueous shallot extract is a stronger hypoglycemic agent than the garlic extract.     

Variability of HOMA and QUICKI insulin sensitivity indices

Assessment of insulin sensitivity based on a single measurement of insulin and glucose, is both easy to understand and simple to perform. The tests most often used are HOMA and QUICKI. The aim of this study was to assess the biological variability of estimates of insulin sensitivity using HOMA and QUICKI indices. After a 12-h fast, blood was sampled for insulin and glucose determination. Sampling lasted for 90?min with an intersample interval of 2?min. A total of 56 subjects were included in the study, and in nine subjects sampling was done before and after weight reduction, so total number of analyzed series was 65. To compute the reference value of the insulin sensitivity index, averages of all 46 insulin and glucose samples were used. We also computed point estimates (single value estimates) of the insulin sensitivity index based on the different number of insulin/glucose samples (1–45 consecutive samples). To compute the variability of point estimates a bootstrapping procedure was used using 1000 resamples for each series and for each number of samples used to average insulin and glucose. Using a single insulin/glucose sample HOMA variability was 26.18?±?4.31%, and QUICKI variability was 3.30?±?0.54%. For 10 samples variability was 11.99?±?2.22% and 1.62?±?0.31% respectively. Biological variability of insulin sensitivity indices is significant, and it can be reduced by increasing the number of samples. Oscillations of insulin concentration in plasma are the major cause of variability of insulin sensitivity indices.     

Glucose tolerance in the elderly: the role of insulin and its receptor

Abstract. Oral glucose tolerance tests were performed in young and elderly subjects with minimal risk factors for diabetes mellitus. Compared to the normal glucose tolerance in the young there was a 45% rate of impaired tolerance in the elderly. Fasting insulin levels were significantly lower in the elderly but post-glucose insulin responses in the first hour were similar in young and elderly subjects. Peripheral insulin action was assessed in terms of the ¹²⁵monoiodoinsulin binding to specific insulin receptor sites on circulating lymphocytes in the young, the elderly and a group of age and sex matched obese maturity-onset diabetics. Specific insulin binding was not significantly different in the elderly than in the young but was significantly lower in the diabetics than the young and the elderly. The results suggest that neither defective insulin secretion nor reduced peripheral insulin binding are major causative factors in the reduced glucose tolerance of the elderly.     

Alcohol consumption alters insulin secretion and cardiac autonomic activity

BACKGROUND: Alcohol may have a cardioprotective effect. One possible mechanism is by modifying insulin resistance/secretion. The aims of this study were: (i) to examine the effect of short-term alcohol consumption on the metabolic control of glucose tolerance; (ii) to study the influence of short-term alcohol consumption on cardiac autonomic activity using spectral analysis of heart rate variability. METHODS: Twenty-one healthy subjects, in a randomized crossover design, either received three units of ethanol daily for 1 week or abstained from ethanol. The control of glucose tolerance was assessed using the intravenous glucose tolerance test with minimal modelling. RESULTS: There was no difference in fasting glucose, fasting insulin or insulin sensitivity between the two groups. Alcohol showed a lower insulin first phase insulin response (no alcohol 659.0 +/- 394.1 SD, alcohol 535.2 +/- 309.1) pmol L-1 min-1, P = 0.027). There was no difference in heart rate or blood pressure but a significant difference in the ratio of high to low frequency spectral power of heart rate variability; (no alcohol 4.55 +/- 3.78, alcohol 8.16 +/- 6.77, P = 0.033). This suggests decreased sympathetic and/or increased vagal modulation of heart rate in the alcohol group. CONCLUSION: The finding of no difference in insulin sensitivity between the two groups contrasts with, but does not entirely contradict, the results of previous epidemiological studies--perhaps suggesting that longer term changes such as liver enzyme induction may be important. The difference in insulin secretion questions the validity of previous studies of the influence of alcohol on insulin sensitivity, where insulin levels were used as a surrogate for insulin resistance.     

The glucoregulatory and antilipolytic actions of insulin in abdominal obesity with normal or impaired glucose tolerance: an in vivo and in vitro study

To evaluate the effects of obesity and impaired glucose tolerance on insulin sensitivity, we performed a euglycaemic-hyperinsulinemic clamp at about 350 pmol l-1, combined with 3H-glucose infusion, in 14 obese patients, BMI 36.5 +/- 1.2 and in 12 matched controls, BMI 23.9 +/- 0.4. Six obese patients had normal glucose tolerance (oNGT), and eight had impaired glucose tolerance (oIGT). The ability of insulin to inhibit lipolysis in isolated adipocytes was also studied. Insulin-mediated glucose utilization was more severely impaired in oIGT than in oNGT with respect to the controls (621 +/- 51 vs. 897 +/- 83 vs. 1298 +/- 55 mumol m-2 min-1, P < 0.001). Plasma glycerol was higher in oIGT than in oNGT and in the controls, both fasting (238 +/- 12 vs. 179 +/- 14 vs. 112 +/- 8 mumol l-1, P < 0.001) and during the clamp (175 +/- 21 vs. 120 +/- 12 vs. 36 +/- 6 mumol l-1, P < 0.001). The correlation between glucose utilization and the percent reduction of plasma glycerol during the clamp was significant in the study group as a whole (r = 0.809, P = 0.0001), and in each of the groups separately (oIGT: r = 0.929, P = 0.002; oNGT: r = 0.943, P = 0.036; controls: r = 0.902, P = 0.0001). Inhibition by insulin of noradrenaline-stimulated lipolysis in isolated adipocytes was more severely impaired in oIGT than in oNGT compared with the controls (P < 0.001).(ABSTRACT TRUNCATED AT 250 WORDS)     

Retinol binding protein 4 and insulin resistance in apparently healthy elderly subjects

BackgroundInsulin resistance (IR) increases with advancing age, yet the underlying mechanism is not well established. Although adipocytokine retinol binding protein 4 (RBP4) was recently shown to be linked to IR, their relationship remains controversial and relatively little information exists regarding their roles in the elderly subjects. We investigated the association between RBP4 and IR in obese and nonobese elderly subjects.MethodsA total of 111 (68 nonobese and 43 obese) apparently healthy elderly subjects, aged 75.9 ± 4.8 y were included. IR was determined by homeostasis model assessment (HOMA-IR). Serum RBP4 was measured by enzyme-linked immunosorbent assay.ResultsIn all subjects, RBP4 levels were positively correlated with fasting insulin, HOMA-IR, and triglycerides. However, after subgroup analysis, RBP4 levels were positively correlated with fasting glucose, fasting insulin, and HOMA-IR in the obese group only. In step-wise multiple linear regression analysis, RBP4 was found to be independently associated with triglyceride levels in the nonobese group and independently associated with HOMA-IR in the obese group.ConclusionsThe reason for the differing metabolic role of RBP4 in obese and nonobese elderly subjects remains uncertain, but our findings suggest that RBP4 may be linked to IR and lipid metabolism, at least in the elderly.