Gut microbiota interactions with obesity, insulin resistance and type 2 diabetes: did gut microbiote co-evolve with insulin resistance?

PURPOSE OF REVIEW: The prevalence of obesity, insulin resistance and type 2 diabetes has steadily increased in the last decades. In addition to the genetic and environmental factors, gut microbiota may play an important role in the modulation of intermediary phenotypes leading to metabolic disease. RECENT FINDINGS: Obesity and type 2 diabetes are associated with specific changes in gut microbiota composition. The mechanisms underlying the association of specific gut microbiota and metabolic disease include increasing energy harvest from the diet, changes in host gene expression, energy expenditure and storage, and alterations in gut permeability leading to metabolic endotoxemia, inflammation and insulin resistance. In some studies, the modifications of gut microbiota induced by antibiotics, prebiotics and probiotics led to improved inflammatory activity in parallel to amelioration of insulin sensitivity and decreased adiposity. However, these effects were mainly observed in animal models. Their extrapolation to humans awaits further studies. SUMMARY: The fascinating role of gut microbiota on metabolic disease opens new avenues in the treatment of obesity, insulin resistance and type 2 diabetes. A co-evolutionary clue for microbiota and insulin resistance is suggested.

Enhanced algorithm for glucose estimation using the continuous glucose monitoring system.

BACKGROUND: The CGMS Gold continuous glucose monitor presents a problem of lack of accuracy, especially in the lower range, sometimes leading to missed or false alarms. The new algorithm aims to improve the measurement accuracy and hypoglycemia detection. MATERIAL/METHODS: Twenty-one patients with type 1 diabetes were monitored for 3 days (1 day at the hospital and 2 at home) using the CGMS Gold. For these patients, blood glucose samples were taken every 15 minutes for 2 hours after meals and every half hour otherwise during the first day. A new calibration algorithm was developed and implemented using CGMS Gold intensity readings and capillary glucose. RESULTS: After 1 day, a comparison of results from either the CGMS Gold algorithm and the proposed algorithm, compared with results from blood (2450 points), showed an increase of data in zone A with the proposed algorithm (4.4% in the Clarke error grid analysis (EGA) and 5.0% in the Consensus EGA). After comparing for 3 days, a reduction of 24.7%, p<0.05, in the overall median relative absolute difference (RAD) was also obtained. In the hypoglycemic range, a significant decrease in median RAD was observed (64.4%, p<0.05). Furthermore, the undetected hypoglycemia events in capillary samples by the proposed algorithm were reduced by 59.8% compared to the CGMS Gold algorithm. CONCLUSIONS: The performance as measured with clinical and numerical accuracy criteria illustrates the improved accuracy of the proposed algorithm in comparison with the CGMS Gold algorithm. A significant improvement in hypoglycemia detection was observed.

Alpha defensins 1, 2, and 3: potential roles in dyslipidemia and vascular dysfunction in humans.

OBJECTIVES: Alpha-defensins are natural antibiotics made by neutrophils that have been reported to modulate cholesterol metabolism and vascular function; however, their role in vivo remains largely unknown. We hypothesized that alpha-defensins 1 to 3 (DEFA1-3) are associated with serum lipids and vascular reactivity in humans. METHODS AND RESULTS: One hundred thirteen apparently-healthy White men, participants in a prospective study of cardiovascular risk factors, were assessed for a lipid profile, insulin sensitivity (S(I), frequently-sampled intravenous glucose tolerance test), and non-stressed circulating DEFA1-3 (ELISA). In a subset of 52 subjects, vascular reactivity (high-resolution ultrasound of the brachial artery) was also assessed. Subjects in the highest quartile for plasma DEFA1-3 were found to be leaner and more insulin sensitive, and to have significantly reduced total and LDL-cholesterol, compared with subjects in the lowest quartile for circulating DEFA1-3 (P<0.0001 to P=0.002 for linear trend ANOVA). The associations with serum lipids persisted after adjustment for age, body mass index, insulin sensitivity, and smoking (which was associated with reduced plasma DEFA1-3 concentrations). Finally, endothelium-independent vasodilation increased with increasing circulating DEFA1-3 (P=0.003) and this association was not explained by age, body mass index, serum cholesterol, insulin sensitivity, or smoking. CONCLUSIONS: Circulating DEFA1-3 are associated with serum cholesterol and vascular reactivity in humans. Alpha-defensins may have clinical implications in patients with either hypercholesterolemia or vascular dysfunction.

Urinary interleukin-6 excretion reflects mean 24-hour systolic blood pressure in type 2 diabetes.

AIMS: Research into new risk markers for diabetic kidney disease is required. We aimed to study 24-hour urinary interleukin- 6 excretion (uIL-6) in type 2 diabetic patients in relation to organ damage induced by increased blood pressure. METHODS: 24-hour uIL-6 and albumin excretion and 24-hour blood pressure recording were evaluated in 49 patients with type 2 diabetes and normal renal function. Patients with optimized mean 24-hour systolic blood pressure (SBP), defined as SBP ≤ 130 mmHg, and those with uncontrolled SBP (SBP > 130 mmHg) were compared. Multiple linear regression analysis was performed to study significant contributors to variance in the 24-hour uIL-6 excretion rate. RESULTS: Albumin excretion rate (AER) and uIL-6 were significantly correlated (r=0.63; p<0.0001). Patients with mean 24-hour SBP above 130 mmHg (n=27) had significantly higher mean uIL-6 excretion than those with a mean 24-hour SBP equal to or below 130 mmHg (n=22) (p=0.009). The strength of the association of uIL-6 with diurnal and mean diastolic blood pressure (DBP) was significantly greater than that with AER. Mean SBP (p<0.0001) contributed to 25% of AER variance after body mass index, age, sex, mean SBP, mean DBP, HbA1c and smoking status were accounted for. Mean 24-hour SBP (p=0.005) and smoking (p=0.03) contributed to 15% and 9%, respectively, of uIL-6 variance. CONCLUSIONS: Increased uIL-6, perhaps by reflecting significant tissue damage and remodelling, could be a marker for increased mean SBP in type 2 diabetes.

Autosomal dominant hypercholesterolemia in Catalonia: Correspondence between clinical-biochemical and genetic diagnostics in 967 patients studied in a multicenter clinical setting.

BACKGROUND: Autosomal dominant hypercholesterolemia (ADH) is associated with mutations in the low-density lipoprotein (LDL) receptor (LDLR), apolipoprotein B (APOB), and proprotein convertase subtilisin/kexin 9 (PCSK9) genes, and it is estimated to be greatly underdiagnosed. The most cost-effective strategy for increasing ADH diagnosis is a cascade screening from mutation-positive probands. OBJECTIVE: The objective of this study was to evaluate the results from 2008 to 2016 of ADH genetic analysis performed in our clinical laboratory, serving most lipid units of Catalonia, a Spanish region with approximately 7.5 million inhabitants. METHODS: After the application of the Dutch Lipid Clinic Network (DLCN) clinical diagnostic score for ADH, this information and blood or saliva from 23 different lipid clinic units were investigated in our laboratory. DNA was screened for mutations in LDLR, APOB, and PCSK9, using the DNA-array LIPOchip, the next-generation sequencing SEQPRO LIPO RS platform, and multiplex ligation-dependent probe amplification (MLPA). The Simon Broome Register Group (SBRG) criteria was calculated and analyzed for comparative purposes. RESULTS: A total of 967 unrelated samples were analyzed. From this, 158 pathogenic variants were detected in 356 patients. The main components of the DLCN criteria associated with the presence of mutation were plasma LDL cholesterol (LDLc), age, and the presence of tendinous xanthomata. The contribution of family history to the diagnosis was lower than in other studies. DLCN and SBRG were similarly useful for predicting the presence of mutation. CONCLUSION: In a real clinical practice, multicenter setting in Catalonia, the percentage of positive genetic diagnosis in patients potentially affected by ADH was 38.6%. The DLCN showed a relatively low capacity to predict mutation detection but a higher one for ruling out mutation.

Serum visfatin increases with progressive beta-cell deterioration.

Visfatin has shown to be increased in type 2 diabetes but to be unrelated to insulin sensitivity. We hypothesized that visfatin is associated with insulin secretion in humans. To this aim, a cross-sectional study was conducted in 118 nondiabetic men and 64 (35 men and 29 women) type 2 diabetic patients. Type 1 diabetic patients with long-standing disease (n = 58; 31 men and 27 women) were also studied. In nondiabetic subjects, circulating visfatin (enzyme immunoassay) was independently associated with insulin secretion (acute insulin response to glucose [AIRg] from intravenous glucose tolerance tests) but not with insulin sensitivity (Si) or other metabolic or anthropometric parameters, and AIRg alone explained 8% of visfatin variance (beta = -0.29, P = 0.001). Circulating visfatin was increased in type 2 diabetes (mean 18 [95% CI 16-21] vs. 15 ng/ml [13-17] for type 2 diabetic and nondiabetic subjects, respectively; P = 0.017, adjusted for sex, age, and BMI), although this association was largely attenuated after accounting for HbA1c (A1C). Finally, circulating visfatin was found to be increased in patients with long-standing type 1 diabetes, even after adjusting for A1C values (37 ng/ml [34-40]; P < 0.0001, adjusted for sex, age, BMI, and A1C compared with either type 2 diabetic or nondiabetic subjects). In summary, circulating visfatin is increased with progressive beta-cell deterioration. The study of the regulation and role of visfatin in diabetes merits further consideration.

Sex-specific, independent associations of insulin resistance with erythrocyte sedimentation rate in apparently healthy subjects.

Insulin resistance and erythrocyte sedimentation rate (ESR, a non-specific marker of inflammation) are known risk factors for cardiovascular disease. Although obesity is associated with increased ESR, it is unclear whether insulin resistance is associated with ESR in humans. The relationship between insulin resistance and ESR was studied in a cross-sectional, health-area based study of 140 (89 men and 51 women) apparently healthy Caucasians subjects. ESR, additional inflammatory markers [soluble tumor necrosis alpha receptors 1 and 2 (sTNFR1 and sTNFR2); C-reactive protein (CRP)], and insulin sensitivity (SI, frequently sampled intravenous glucose tolerance test with minimal model analysis) were assessed in all subjects. An interaction with sex was documented in the relationship between ESR and both insulin resistance and obesity (p < 0.05), as log ESR correlated with log SI in men (r=-0.29, p=0.009), but not in women (r=-0.14, p=0.33), and correlated with body mass index (BMI) in women (r=0.49, p=<0.0001), but not in men (r=0.15, p=0.16). On multivariate analyses, these associations proved to be independent of known covariates, such as age, hematocrit, smoking and additional inflammatory markers in both men and women. In a replication study, variables independently associated with ESR were also insulin resistance (homeostasis model assessment) in men and obesity markers (either BMI or fat mass) in women. In conclusion, ESR is independently associated with either insulin resistance or obesity in a sex-specific manner. These findings contribute to explain the known relationship of this inflammatory marker with cardiovascular disease.

Metformin alters the gut microbiome of individuals with treatment-naive type 2 diabetes, contributing to the therapeutic effects of the drug.

Metformin is widely used in the treatment of type 2 diabetes (T2D), but its mechanism of action is poorly defined. Recent evidence implicates the gut microbiota as a site of metformin action. In a double-blind study, we randomized individuals with treatment-naive T2D to placebo or metformin for 4 months and showed that metformin had strong effects on the gut microbiome. These results were verified in a subset of the placebo group that switched to metformin 6 months after the start of the trial. Transfer of fecal samples (obtained before and 4 months after treatment) from metformin-treated donors to germ-free mice showed that glucose tolerance was improved in mice that received metformin-altered microbiota. By directly investigating metformin-microbiota interactions in a gut simulator, we showed that metformin affected pathways with common biological functions in species from two different phyla, and many of the metformin-regulated genes in these species encoded metalloproteins or metal transporters. Our findings provide support for the notion that altered gut microbiota mediates some of metformin's antidiabetic effects.

Dyslipidemia and inflammation: an evolutionary conserved mechanism.

Inflammation leads to changes in lipid metabolism aimed at decreasing the toxicity of a variety of harmful agents and tissue repair by redistributing nutrients to cells involved in host defence. Acute phase response, mediated by cytokines, preserves the host from acute injury. When this inflammation becomes chronic, it might lead to chronic disorders as atherosclerosis and the metabolic syndrome. The activation of the inflammatory cascade will induce a decrease in HDL-cholesterol (HDL-C), with impairment in reverse cholesterol transport, and parallel changes in apolipoproteins, enzymes, anti-oxidant capacity and ATP binding cassette A1-dependent efflux. This decrease in HDL-C and phospholipids could stimulate compensatory changes, as synthesis and accumulation of phospholipid-rich VLDL which binds bacterial products and other toxic substances, resulting in hypertriglyceridemia. The final consequence is an increased accumulation of cholesterol in cells. When the compensatory response (inflammation) is not able to repair injury, it turns into a harmful reaction, and the lipid changes will become chronic, either by repeated or overwhelming stimulus, enhancing the formation of atherosclerotic lesions. Thus, the classical lipid changes associated with the metabolic syndrome (increased triglycerides and decreased HDL-C) may be envisioned as a highly conserved evolutionary response aimed at tissue repair. Under this assumption, the problem is not the response but the persistence of the stimulus.

Obesity changes the human gut mycobiome.

The human intestine is home to a diverse range of bacterial and fungal species, forming an ecological community that contributes to normal physiology and disease susceptibility. Here, the fungal microbiota (mycobiome) in obese and non-obese subjects was characterized using Internal Transcribed Spacer (ITS)-based sequencing. The results demonstrate that obese patients could be discriminated by their specific fungal composition, which also distinguished metabolically "healthy" from "unhealthy" obesity. Clusters according to genus abundance co-segregated with body fatness, fasting triglycerides and HDL-cholesterol. A preliminary link to metabolites such as hexadecanedioic acid, caproic acid and N-acetyl-L-glutamic acid was also found. Mucor racemosus and M. fuscus were the species more represented in non-obese subjects compared to obese counterparts. Interestingly, the decreased relative abundance of the Mucor genus in obese subjects was reversible upon weight loss. Collectively, these findings suggest that manipulation of gut mycobiome communities might be a novel target in the treatment of obesity.

Circulating hepcidin in type 2 diabetes: A multivariate analysis and double blind evaluation of metformin effects.

SCOPE: Very few studies have evaluated serum hepcidin in patients with type 2 diabetes and they have reported conflicting results. In addition, the effect of antidiabetic drugs on circulating hepcidin has not been explored so far. The aims of the study were to evaluate hepcidin concentrations and hepcidin/ferritin ratio in type 2 diabetes subjects and healthy non-diabetic controls and to evaluate the effect of metformin on hepcidin concentrations. METHODS AND RESULTS: Study 1: Cross-sectional multivariate study of 239 non-diabetic individuals and 65 people with type 2 diabetes. The multivariate analysis included covariates of chronic inflammation, BMI, pharmacological treatment, menopausal status and insulin resistance. Study 2: Randomized, double-blinded, placebo-controlled 4-month trial metformin compared to placebo among 36 type 2 diabetic patients. In both groups diet was controlled by maintaining a hypocaloric intake across the trial. Hepcidin levels were significantly lower in patients with type 2 diabetes than in non-diabetic individuals either in crude or adjusted regression models (P<0.05). Hepcidin decreased in both arms of the trial (Placebo, p = 0.004; metformin, p = 0.022). CONCLUSION: Circulating hepcidin was significantly and independently lower in type 2 diabetes. Metformin treatment is not associated with reductions in hepcidin but hypocaloric diet could be involved.

Profiling of circulating microRNAs reveals common microRNAs linked to type 2 diabetes that change with insulin sensitization.

OBJECTIVE: This study sought to identify the profile of circulating microRNAs (miRNAs) in type 2 diabetes (T2D) and its response to changes in insulin sensitivity. RESEARCH DESIGN AND METHODS: The circulating miRNA profile was assessed in a pilot study of 12 men: 6 with normal glucose tolerance (NGT) and 6 T2D patients. The association of 10 circulating miRNAs with T2D was cross-sectionally validated in an extended sample of 45 NGT vs. 48 T2D subjects (65 nonobese and 28 obese men) and longitudinally in 35 T2D patients who were recruited in a randomized, double-blinded, and placebo-controlled 3-month trial of metformin treatment. Circulating miRNAs were also measured in seven healthy volunteers before and after a 6-h hyperinsulinemic-euglycemic clamp and insulin plus intralipid/heparin infusion. RESULTS: Cross-sectional studies disclosed a marked increase of miR-140-5p, miR-142-3p, and miR-222 and decreased miR-423-5p, miR-125b, miR-192, miR-195, miR-130b, miR-532-5p, and miR-126 in T2D patients. Multiple linear regression analyses revealed that miR-140-5p and miR-423-5p contributed independently to explain 49.5% (P < 0.0001) of fasting glucose variance after controlling for confounders. A discriminant function of four miRNAs (miR-140-5p, miR-423-5p, miR-195, and miR-126) was specific for T2D with an accuracy of 89.2% (P < 0.0001). Metformin (but not placebo) led to significant changes in circulating miR-192 (49.5%; P = 0.022), miR-140-5p (-15.8%; P = 0.004), and miR-222 (-47.2%; P = 0.03), in parallel to decreased fasting glucose and HbA1c. Furthermore, while insulin infusion during clamp decreased miR-222 (-62%; P = 0.002), the intralipid/heparin mixture increased circulating miR-222 (163%; P = 0.015) and miR-140-5p (67.5%; P = 0.05). CONCLUSIONS: This study depicts the close association between variations in circulating miRNAs and T2D and their potential relevance in insulin sensitivity.

Real-time glucose estimation algorithm for continuous glucose monitoring using autoregressive models.

BACKGROUND: Continuous glucose monitors (CGMs) present a problem of lack of accuracy, especially in the lower range, sometimes leading to missed or false hypoglycemia. A new algorithm is presented here aimed at improving the measurement accuracy and hypoglycemia detection. Its core is the estimation of blood glucose (BG) in real time (RT) from CGM intensity readings using autoregressive (AR) models. METHODS: Eighteen patients with type 1 diabetes were monitored for three days (one at the hospital and two at home) using the CGMS Gold. For these patients, BG samples were taken every 15 min for 2 h after meals and every half hour otherwise during the first day. The relationship between the current measured by the CGMS Gold and BG was learned by an AR model, allowing its RT estimation. New capillary glucose measurements were used to correct the model BG estimations. RESULTS: A total of 563 paired points were obtained from BG and monitor readings to validate the new algorithm. 98.5% of paired points fell in zones A+B of the Clarke error grid analysis with the proposed algorithm. The overall mean and median relative absolute differences (RADs) were 9.6% and 6.7%. Measurements meeting International Organization for Standardization (ISO) criteria were 88.7%. In the hypoglycemic range, the mean and median RADs were 8.1% and 6.0%, and measurements meeting ISO criteria were 86.7%. The sensitivity and specificity with respect to hypoglycemia detection were 91.5% and 95.0%. CONCLUSIONS: The performance measured with both clinical and numerical accuracy metrics illustrates the improved accuracy of the proposed algorithm compared with values presented in the literature. A significant improvement in hypoglycemia detection was also observed.

Circulating bactericidal/permeability-increasing protein (BPI) is associated with serum lipids and endothelial function.

Bactericidal/permeability-increasing protein (BPI), a major constituent of neutrophils that possesses anti-inflammatory properties, shows a structure similar to some proteins implicated in lipid metabolism. We evaluated circulating BPI as a biomarker of endothelial function and lipid metabolism. Circulating BPI concentrations (ELISA) and serum lipids were measured in 202 Caucasian non-smoking men. In a subgroup of 91 consecutive subjects brachial vascular reactivity (high resolution external ultrasound) was assessed. Plasma BPI concentrations were positively associated with total cholesterol (TC), LDL cholesterol (LDL-C) and HDL cholesterol (HDL-C) (r= 0.203, 0.204 and 0.18; all p<0.05, respectively). In a multiple linear regression analysis, BPI levels were independent contributors to the variance of HDL-C, total cholesterol and LDL-cholesterol after adjusting for age, body mass index and glucose tolerance status. Plasma BPI concentration correlated positively with endothelium-dependent vasodilatation (r=0.277; p<0.05) and HDL-C (r=0.36; p<0.05) in subjects with normal glucose tolerance. In conclusion, circulating BPI could constitute a biomarker of lipid metabolism in subjects with normal glucose tolerance and could help to identify those subjects with preserved endothelial function.

The gene expression of the main lipogenic enzymes is downregulated in visceral adipose tissue of obese subjects.

Contradictory findings regarding the gene expression of the main lipogenic enzymes in human adipose tissue depots have been reported. In this cross-sectional study, we aimed to evaluate the mRNA expression of fatty acid synthase (FAS) and acetyl-CoA carboxilase (ACC) in omental and subcutaneous (SC) fat depots from subjects who varied widely in terms of body fat mass. FAS and ACC gene expression were evaluated by real time-PCR in 188 samples of visceral adipose tissue which were obtained during elective surgical procedures in 119 women and 69 men. Decreased sex-adjusted FAS (-59%) and ACC (-49%) mRNA were found in visceral adipose tissue from obese subjects, with and without diabetes mellitus type 2 (DM-2), compared with lean subjects (both P < 0.0001). FAS mRNA was also decreased (-40%) in fat depots from overweight subjects (P < 0.05). Indeed, FAS mRNA was significantly and positively associated with ACC gene expression (r = 0.316, P < 0.0001) and negatively with BMI (r = -0.274), waist circumference (r = -0.437), systolic blood pressure (r = -0.310), serum glucose (r = -0.277), and fasting triglycerides (r = -0.226), among others (all P < 0.0001). Similar associations were observed for ACC gene expression levels. In a representative subgroup of nonobese (n = 4) and obese women (n = 6), relative FAS gene expression levels significantly correlated (r = 0.657, P = 0.034; n = 10) with FAS protein values. FAS protein levels were also inversely correlated with blood glucose (r = -0.640, P = 0.046) and fasting triglycerides (r = -0.832, P = 0.010). In conclusion, the gene expression of the main lipogenic enzymes is downregulated in visceral adipose tissue from obese subjects.

Subcutaneous fat shows higher thyroid hormone receptor-alpha1 gene expression than omental fat.

The aims of this work were to evaluate thyroid hormone receptor-alpha (TR alpha), TR alpha 1, and TR alpha 2 mRNA gene expression and TR alpha 1:TR alpha 2 ratio, identified as candidate factors for explaining regional differences between human adipose tissue depots. TR alpha, TR alpha 1, and TR alpha 2 mRNA levels, and the gene expressions of arginine-serine-rich, splicing factor 2 (SF2), heterogeneous nuclear ribonucleoprotein H1 (hnRNP H1), heterogeneous nuclear ribonucleoprotein A1 (hnRNP A1), and Spot 14 (S14) were evaluated in 76 paired adipose tissue samples obtained from a population of 38 women who varied widely in terms of obesity and body fat distribution. Gene expression for these factors was also studied in stromal-vascular cells (SVCs) and mature adipocytes (MAs) from eight paired fat depots. TR alpha gene and TR alpha 1 mRNA expression were increased 1.46-fold (P = 0.006) and 1.80-fold (P < 0.0001), respectively, in subcutaneous (SC) vs. visceral fat. These differences in gene expression levels were most significant in the obese group, in which the TR alpha 1:TR alpha 2 ratio was 2.24-fold (P < 0.0001) higher in SC vs. visceral fat. S14 gene expression was also increased by 2.42-fold (P < 0.0001) and correlated significantly with TR alpha and TR alpha 1 gene expression and with the TR alpha 1:TR alpha 2 ratio. In agreement with these findings, hnRNP A1:SF2 ratio was decreased by 1.39-fold (P = 0.001). TR alpha and S14 levels were 2.1-fold (P < 0.0001) and 112.4-fold (P < 0.0001), respectively, higher in MAs than in SVCs from both fat depots. In summary, genes for TR-alpha, their upstream regulators, and downstream effectors were differentially expressed in SC vs. omental (OM) adipose tissue. Our findings suggest that TR alpha1 could contribute to SC adipose tissue expandability in obese subjects.

Using support vector machines to detect therapeutically incorrect measurements by the MiniMed CGMS.

BACKGROUND: Current continuous glucose monitors have limited accuracy mainly in the low range of glucose measurements. This lack of accuracy is a limiting factor in their clinical use and in the development of the so-called artificial pancreas. The ability to detect incorrect readings provided by continuous glucose monitors from raw data and other information supplied by the monitor itself is of utmost clinical importance. In this study, support vector machines (SVMs), a powerful statistical learning technique, were used to detect therapeutically incorrect measurements made by the Medtronic MiniMed CGMS. METHODS: Twenty patients were monitored for three days (first day at the hospital and two days at home) using the MiniMed CGMS. After the third day, the monitor data were downloaded to the physician's computer. During the first 12 hours, the patients stayed in the hospital, and blood samples were taken every 15 minutes for two hours after meals and every 30 minutes otherwise. Plasma glucose measurements were interpolated using a cubic method for time synchronization with simultaneous MiniMed CGMS measurements every five minutes, obtaining a total of 2281 samples. A Gaussian SVM classifier trained on the monitor's electrical signal and glucose estimation was tuned and validated using multiple runs of k-fold cross-validation. The classes considered were Clarke error grid zones A+B and C+D+E. RESULTS: After ten runs of ten-fold cross-validation, an average specificity and sensitivity of 92.74% and 75.49%, respectively, were obtained (see Figure 4). The average correct rate was 91.67%. CONCLUSIONS: Overall, the SVM performed well, in spite of the somewhat low sensitivity. The classifier was able to detect the time intervals when the monitor's glucose profile could not be trusted due to incorrect measurements. As a result, hypoglycemic episodes missed by the monitor were detected.

Serum interleukin-6 correlates with endothelial dysfunction in healthy men independently of insulin sensitivity.

OBJECTIVE: Interleukin (IL)-6 is a proinflammatory cytokine that is implicated in the pathogenesis of atherosclerosis and insulin resistance. Both endothelial dysfunction and insulin resistance are among the earliest abnormalities that can be detected in people at risk for cardiovascular events. We aimed to evaluate whether increased serum IL-6 concentrations associated with endothelial dysfunction are independent of insulin sensitivity in apparently healthy men. RESEARCH DESIGN AND METHODS: Association studies were performed in well-characterized nondiabetic Caucasian men (n = 99) recruited for energy balance studies. Insulin sensitivity (minimal model) and brachial vascular reactivity (high-resolution external ultrasound) were assessed. Circulating IL-6 concentrations were measured by enzyme-linked immunosorbent assay. RESULTS: Serum IL-6 was an independent contributor to the variance of endothelium-dependent vasodilatation after adjusting for age, BMI, smoking status, LDL cholesterol, systolic blood pressure, diastolic blood pressure, and insulin sensitivity (P = 0.001). In fact, circulating IL-6 was negatively associated with endothelium-dependent vasodilatation (r = -0.247, P = 0.014) and insulin sensitivity (r = -0.262, P = 0.011) and correlated positively with age (r = 0.241, P = 0.016), BMI (r = 0.240, P = 0.017), systolic blood pressure (r = 0.299, P = 0.003), diastolic blood pressure (r = 0.295, P = 0.003), and triglycerides (r = 0.212, P = 0.035). No significant associations were observed between endothelium-independent vasodilatation and serum IL-6 concentrations. CONCLUSIONS: Circulating IL-6 is linked to endothelial dysfunction independently of insulin sensitivity in apparently healthy men.