Michael P. Stern

The metabolic syndrome: time for a critical appraisal. Joint statement from the American Diabetes Association and the European Association for the Study of Diabetes.

BackgroundThe term ‘metabolic syndrome’ refers to a clustering of specific cardiovascular disease (CVD) risk factors whose underlying pathophysiology is thought to be related to insulin resistance.MethodsSince the term is widely used in research and clinical practice, we undertook an extensive review of the literature in relation to the syndrome’s definition, underlying pathogenesis, association with cardiovascular disease and to the goals and impact of treatment.DiscussionWhile there is no question that certain CVD risk factors are prone to cluster, we found that the metabolic syndrome has been imprecisely defined, there is a lack of certainty regarding its pathogenesis, and there is considerable doubt regarding its value as a CVD risk marker. Our analysis indicates that too much critically important information is missing to warrant its designation as a ‘syndrome’.ConclusionUntil much-needed research is completed, clinicians should evaluate and treat all CVD risk factors without regard to whether a patient meets the criteria for diagnosis of the ‘metabolic syndrome’.

Prospective Analysis of The Insulin-Resistance Syndrome (Syndrome X)

Many studies have shown that hyperinsulinemia and/or insulin resistance are related to various metabolic and physiological disorders including hypertension, dyslipidemia, and non-insulin-dependent diabetes mellitus. This syndrome has been termed Syndrome X. An important limitation of previous studies has been that they all have been cross sectional, and thus the presence of insulin resistance could be a consequence of the underlying metabolic disorders rather than its cause. We examined the relationship of fasting insulin concentration (as an indicator of insulin resistance) to the incidence of multiple metabolic abnormalities in the 8-yr follow-up of the cohort enrolled in the San Antonio Heart Study, a population-based study of diabetes and cardiovascular disease in Mexican Americans and non-Hispanic whites. In univariate analyses, fasting insulin was related to the incidence of the following conditions: hypertension, decreased high-density lipoprotein cholesterol concentration, increased triglyceride concentration, and non-insulin-dependent diabetes mellitus. Hyperinsulinemia was not related to increased low-density lipoprotein or total cholesterol concentration. In multivariate analyses, after adjustment for obesity and body fat distribution, fasting insulin continued to be significantly related to the incidence of decreased high-density lipoprotein cholesterol and increased triglyceride concentrations and to the incidence of non-insulin-dependent diabetes mellitus. Baseline insulin concentrations were higher in subjects who subsequently developed multiple metabolic disorders. These results were not attributable to differences in baseline obesity and were similar in Mexican Americans and non-Hispanic whites. These results support the existence of a metabolic syndrome and the relationship of that syndrome to multiple metabolic disorders by showing that elevations of insulin concentration precede the development of numerous metabolic disorders.

HYPERINSULINAEMIA : THE KEY FEATURE OF A CARDIOVASCULAR AND METABOLIC SYNDROME

SummaryIn a population-based survey of 2,930 subjects, prevalence rates for obesity, Type 2 (non-insulin-dependent) diabetes mellitus, impaired glucose tolerance, hypertension, hypertriglyceridaemia, and hypercholesterolaemia were 54.3, 9.3, 11.1, 9.8, 10.3 and 9.2%, respectively. The prevalence, however, of each of these conditions in its isolated form (free of the other five) was 29.0% for obesity, 1.3% for Type 2 diabetes, 1.8% for impaired glucose tolerance, 1.5% for hypertension, 1.0% for hypertriglyceridaemia, and 1.7% for hypercholesterolaemia. Two-by-two associations were even rarer. The large differences in prevalence between isolated and mixed forms indicate a major overlap among the six disorders in multiple combinations. In the isolated form, each condition was characterized by hyperinsulinaemia (both fasting and 2 h after oral glucose), suggesting the presence of insulin resistance. In addition, in any isolated condition most of the variables categorising other members of the sextet were still significantly altered in comparison with 1,049 normal subjects. In the whole of the subjects who presented with one or another disorder (1,881 of 2,930 or 64%), marked fasting and post-glucose hyperinsulinaemia was associated with higher body mass index, waist:hip ratio, fasting and post-glucose glycaemia, systolic and diastolic blood pressure, serum triglycerides and total cholesterol levels, and with lower HDL-cholesterol concentrations (all p <0.001). We conclude that (1) insulin sensitivity, glucose tolerance, blood pressure, body fat mass and distribution, and serum lipids are a network of mutually interrelated functions; and (2) an insulin resistance syndrome underlies each and all of the six disorders carrying an increased risk of coronary artery disease.

The Homeostasis Model in the San Antonio Heart Study

OBJECTIVE Both insulin resistance and decreased insulin secretion have been shown to predict the development of NIDDM. However, methods to assess insulin sensitivity and secretion are complicated and expensive to apply in epidemiological studies. The homeostasis model assessment (HOMA) has been suggested as a method to assess insulin resistance and secretion from the fasting glucose and insulin concentrations. However, this method has not been extensively evaluated, particularly in different ethnic groups. RESEARCH DESIGN AND METHODS We applied the HOMA model to cross-sectional analyses of the San Antonio Heart Study (n = 2,465). RESULTS HOMA insulin resistance (IR) was very strongly correlated with fasting insulin (r = 0.98) and HOMA β-cell function (β-cell) was moderately correlated with the 30-min increment in insulin concentration over the 30-min increment in glucose concentration (Δ I30/Δ G30) in an oral glucose tolerance test (OGTT) (r = 0.44). NIDDM was characterized by both high HOMA IR and low HOMA β-cell function. In Mexican-Americans, HOMA IR in NIDDM subjects was 9.5 compared with 2.7 in normal glucose tolerance (NGT) subjects. In contrast, HOMA β-cell function showed only small differences in Mexican-Americans (176 NIDDM; 257 NGT). However, the ΔI30/ΔG30 (pmol/mmol) showed much larger differences (75 NIDDM; 268 NGT). When modeled separately, impaired glucose tolerance (IGT) was characterized by high HOMA IR and high HOMA β-cell function. However, when analyzed in the same regression model, high HOMA IR and low HOMA β-cell function characterized subjects with IGT. These results were similar in both ethnic groups. Mexican-Americans had increased insulin resistance (as judged by both HOMA IR and fasting insulin) and insulin secretion (by HOMA β-cell and ΔI30/ΔG30) relative to non-Hispanic whites. CONCLUSIONS We conclude that HOMA provides a useful model to assess insulin resistance and β-cell function in epidemiological studies in which only fasting samples are available and that, further, it is critical to take into account the degree of insulin resistance in assessing insulin secretion by the HOMA model.

National Cholesterol Education Program Versus World Health Organization Metabolic Syndrome in Relation to All-Cause and Cardiovascular Mortality in the San Antonio Heart Study

Background—To assess the utility of clinical definitions of the metabolic syndrome (MetS) to identify individuals with increased cardiovascular risk, we examined the relation between the MetS, using both the National Cholesterol Education Program (NCEP) and the World Health Organization definitions, and all-cause and cardiovascular mortality in San Antonio Heart Study participants enrolled between 1984 and 1988. Methods and Results—Among 2815 participants, 25 to 64 years of age at enrollment, 509 met both criteria, 197 met NCEP criteria only, and 199 met WHO criteria only. Over an average of 12.7 years, 229 deaths occurred (117 from cardiovascular disease). Moreover, in the primary prevention population of 2372 participants (ie, those without diabetes or cardiovascular disease at baseline), 132 deaths occurred (50 from cardiovascular disease). In the primary prevention population, the only significant association adjusted for age, gender, and ethnic group was between NCEP-MetS and cardiovascular mortality (hazard ratio [HR], 2.01; 95% CI, 1.13–3.57). In the general population, all-cause mortality HRs were 1.47 (95% CI, 1.13–1.92) for NCEP-MetS and 1.27 (95% CI, 0.97–1.66) for WHO-MetS. Furthermore, for cardiovascular mortality, there was evidence that gender modified the predictive ability of the MetS. For women and men, respectively, HRs for NCEP-MetS were 4.65 (95% CI, 2.35–9.21) and 1.82 (95% CI, 1.14–2.91), whereas HRs for WHO-MetS were 2.83 (95% CI, 1.55–5.17) and 1.15 (95% CI, 0.72–1.86). Conclusions—In summary, although both definitions were predictive in the general population, the simpler NCEP definition tended to be more predictive in lower-risk subjects.

A prospective analysis of the HOMA model. The Mexico City Diabetes Study.

OBJECTIVE Both insulin resistance (IR) and decreased insulin secretion have been shown to predict the development of NIDDM. However, methods to assess insulin sensitivity and secretion are complicated and expensive to apply in epidemiological studies. The homeostasis model assessment (HOMA) has been suggested as a method to assess IR and secretion from the fasting glucose and insulin concentrations. RESEARCH DESIGN AND METHODS We applied the HOMA model in the 3.5-year follow-up of the Mexico City Diabetes Study. RESULTS Out of 1,449 subjects, 97 developed diabetes. When modeled separately insulin resistance but not insulin secretion predicted NIDDM. However, when both variables were entered into the same regression model, both increased IR and decreased β-cell function significantly predicted NIDDM. CONCLUSIONS We conclude that the HOMA provides a useful model to assess ²-cell function in epidemiological studies and that it is important to take into account the degree of IR in assessing insulin secretion.

Birthweight and adult health outcomes in a biethnic population in the USA

SummaryRecent data indicate that low-birthweight adults are at a higher risk than their high-birthweight peers of developing ischaemic heart disease or a cluster of conditions known as the IRS, which includes dyslipidaemias, hypertension, unfavourable body fat distribution and NIDDM. Thus far these observations have been limited to Caucasians from the United Kingdom. We extended these observations to a broader segment of the general population by studying the association of birthweight and adult health outcomes in a biethnic population of the United States. We divided a group of 564 young adult Mexican-American and non-Hispanic white men and women participants of the San Antonio Heart Study into tertiles of birthweight and compared metabolic, anthropometric, haemodynamic, and demographic characteristics across these tertile categories. Additionally, we studied birthweight as a predictor of the clustering of diseases associated with the IRS, defined as any two or more of the following conditions: hypertension, NIDDM or impaired glucose tolerance, dyslipidaemia. Normotensive, non-diabetic individuals whose birthweight was in the lowest tertile had significantly higher levels of fasting serum insulin and a more truncal fat deposition pattern than individuals whose birthweight was in the highest tertile, independently of sex, ethnicity, and current socioeconomic status. Also, the odds of expressing the IRS increased 1.72 times (95% confidence interval: 1.16–2.55) for each tertile decrease in birthweight. These findings were independent of sex, ethnicity, and current levels of socioeconomic status or obesity. In conclusion, low birthweight could be a major independent risk factor for the development of adult chronic conditions commonly associated with insulin resistance in the general population.

Role of Insulin in Endogenous Hypertriglyceridemia

Dietary carbohydrate accentuation of endogenous triglyceride production has been studied in 33 patients. A broad and relatively continuous spectrum of steady-state plasma triglyceride concentrations was produced in 31 of the 33 subjects during 3 wk of a high carbohydrate (fat-free) liquid formula diet. Two patients developed plasma triglyceride concentrations in excess of 2000 mg/100 ml, and these were the only patients we have studied in which carbohydrate induction of hypertriglyceridemia seemed to be associated with a defect in endogenous plasma triglyceride removal mechanisms. In the remaining 31 patients the degree of hypertriglyceridemia was highly correlated with the insulin response elicited by the ingestion of the high carbohydrate formula (P < 0.005). No significant correlation existed between fasting plasma triglyceride concentration and either plasma glucose or free fatty acid concentrations after the high carbohydrate diet, nor was the degree of hypertriglyceridemia related to degree of obesity. It is suggested that hypertriglyceridemia in most subjects results from an increase in hepatic triglyceride secretion rate secondary to exaggerated postprandial increases in plasma insulin concentration.

Identification of Persons at High Risk for Type 2 Diabetes Mellitus: Do We Need the Oral Glucose Tolerance Test?

Context Lifestyle and pharmaceutical interventions can prevent overt diabetes in people with impaired glucose tolerance. Oral glucose tolerance testing is the reference standard for identifying impaired glucose tolerance, but it is inconvenient and relatively expensive. Contribution The authors developed multivariable models that use readily available clinical variables to predict the development of diabetes. The models were more accurate than oral glucose tolerance testing alone. Adding results of oral glucose tolerance testing did not substantially improve the models predictions. Cautions More than half the study sample was Mexican American. Validation in other populations and translation for bedside calculation is needed before clinicians can use the model. The Editors In 1979, the National Diabetes Data Group defined an entity called impaired glucose tolerance, which reflects a degree of glucose tolerance that, although abnormal, is considered insufficient to merit a diagnosis of diabetes mellitus (1). This entity, which was later endorsed by the World Health Organization (WHO) (2, 3) and the American Diabetes Association (4), requires a 2-hour oral glucose tolerance test for its detection. It is important to emphasize that impaired glucose tolerance is by itself entirely asymptomatic and unassociated with any functional disability. Indeed, insulin secretion is typically greater in response to a mixed meal than in response to a pure glucose load (5); as a result, most persons with impaired glucose tolerance are rarely, if ever, hyperglycemic in their daily lives (5, 6), except when they undergo diagnostic glucose tolerance tests. Thus, the importance of impaired glucose tolerance resides exclusively in its ability to identify persons at increased risk for future disease. The standard method for identifying persons at high risk for developing diabetes mellitus has been to identify persons with impaired glucose tolerance without regard to other diabetes risk factors. For example, nearly all of the clinical trials on prevention of type 2 diabetes have used impaired glucose tolerance as the principal enrollment criterion. Three of these trialsthe Finnish Diabetes Prevention Study (7), the Diabetes Prevention Program (8), and the Study To Prevent Non-Insulin-Dependent Diabetes Mellitus (STOP-NIDDM) (9)have recently reported positive results. Therefore, a need has arisen to identify persons at high risk for diabetes so that physicians can offer them preventive interventions. Because the 2-hour oral glucose tolerance test, which is necessary for the diagnosis of impaired glucose tolerance, is time consuming, costly, and inconvenient, it becomes relevant to ask whether other, more efficient means exist for identifying persons at high risk for diabetes. A popular method of assessing the predictive discrimination of a test is to use a receiver-operating characteristic (ROC) curve (10) that plots the sensitivity of the test against the corresponding false-positive rate. In the present context, sensitivity refers to the percentage of persons whose initial values were above a given cut point among all persons who later developed diabetes; false-positive rate refers to the percentage of persons whose initial values were above the cut point among persons who nevertheless remained free of diabetes. The area under the ROC curve is a measure of how well a continuous variable can predict the outcome of interest: If the sensitivity increases steeply as the threshold for diagnosis is relaxed, with only a relatively slow accumulation of false-positive results, the area under the ROC curve will be large; conversely, if the sensitivity increases slowly as the threshold for diagnosis is relaxed, with a rapid accumulation of false-positive results, the area under the ROC curve will be correspondingly smaller. The differences in the areas under two curves may be tested to see whether the apparent superiority of one continuous variable over another is statistically significant. We have used this approach to compare the 2-hour glucose value after an oral glucose load to various multivariable models for predicting future diabetes. Methods Participants Our analyses are based on data gathered in the San Antonio Heart Study. The methods of this study have been described elsewhere (11-13). Briefly, households were randomly sampled from three types of neighborhoods: low, middle, and high income. Residents of these households were eligible if they were 25 to 64 years of age and not pregnant. Because the number of nonMexican-American persons residing in the low-income areas was negligible, only Mexican Americans were recruited from these neighborhoods. Stratified random sampling was used to recruit an approximately equal number of Mexican Americans and non-Hispanic whites from the middle-income and high-income neighborhoods. The baseline data were collected in two phases, from 1979 to 1982 and from 1984 to 1988. A total of 5158 participants were enrolled in these two phases, representing a response rate of 65.3% of all eligible participants from the selected households. A follow-up examination was performed 7 to 8 years after the baseline examination on 3682 persons, representing 73.7% of the 4998 surviving study participants. The Institutional Review Board of the University of Texas Health Science Center at San Antonio approved the protocol. All participants gave informed consent. Measurements Fasting plasma glucose levels were measured for all participants; they then drank a standardized 75-g glucose load (Koladex or Orangedex, Custom Laboratories, Baltimore, Maryland). The plasma glucose level was measured again 2 hours later. Although various protocols have been used for oral glucose tolerance testing, the glucose level obtained 2 hours after the administration of the oral glucose load is the only postload value that has been used as the basis for diagnostic categories (1-4); therefore, this was the only postload value that we considered. In line with common usage, we refer to this value as the 2-hour glucose value. Diabetes was diagnosed according to WHO criteria (fasting glucose level 7.0 mmol/L [ 126 mg/dL] or 2-hour glucose level 11.1 mmol/L [ 200 mg/dL]) (3). Persons who reported a history of diabetes diagnosed by a physician and who reported current use of insulin or an oral antidiabetic agent were considered to have diabetes regardless of their plasma glucose levels. Participants were asked to bring to the examination center a list of all prescription medications that they were receiving or the containers in which the medications had been dispensed. Participants who did not adhere to this request were subsequently contacted by telephone to verify their medications. Participants were classified as having diabetes if they met at least one of the above three criteria (fasting glucose value, 2-hour glucose value, or antidiabetic medications), even if all three variables were not recorded. Participants were classified as nondiabetic if all three variables were recorded and none met the criterion for diabetes. Persons with diabetes who were not taking insulin were considered to have type 2 diabetes. Participants with diabetes who used insulin were considered to have type 2 diabetes if they were at least 30 years of age at diabetes onset and if their body mass index (BMI) (weight in kg divided by height in meters squared) was greater than 27.0 kg/m2. Nondiabetic persons were classified as having impaired glucose tolerance if their 2-hour plasma glucose level was 7.8 mmol/L or higher ( 140 mg/dL) but less than 11.1 mmol/L (<200 mg/dL) (1-4). Height; weight; blood pressure; plasma glucose level; serum total, low-density lipoprotein (LDL), and high-density lipoprotein (HDL) cholesterol levels; and serum triglyceride level were measured by using previously reported methods (11, 12). Statistical Analysis Baseline characteristics of the study sample according to sex and ethnicity were adjusted for age by analysis of covariance using SAS software (14). We developed a multiple logistic regression model with incident diabetes as the dependent variable and a panel of baseline characteristics that are ordinarily available in a routine clinical setting as independent variables. We refer to this modelwhich includes (with or without selected interactions, as explained in the Results section) age; sex; ethnicity; fasting and 2-hour glucose levels; systolic and diastolic blood pressures; total, LDL, and HDL cholesterol levels; triglyceride level; body mass index; and parental or sibling history of diabetesas the full model. In addition to testing the statistical significance of each interaction term, we also used likelihood ratio tests to globally compare models with and without interactions. We assessed the importance of the 2-hour glucose value for predicting diabetes by comparing the full model with a nested model that excluded the 2-hour glucose value. We also examined a simplified model based on widely recognized diabetes risk factors, which we call the clinical model. We believe that clinicians would more readily accept this simplified model because it entails fewer variables. The variables used in the clinical model were age, sex, ethnicity, fasting glucose level, systolic blood pressure, HDL cholesterol level, body mass index, and parental or sibling history of diabetes. This model was also examined with and without selected interactions and with and without 2-hour glucose value. We assessed the goodness of fit of all models by using the HosmerLemeshow test (14). We compared the predictive discrimination of the multivariable models to the predictive discrimination of the 2-hour glucose measurement by using ROC curves. The cutoff point defining impaired glucose tolerance represents only one of many possible cutoff points along the 2-hour glucose curve. The ROC curves were calculated for the multivariable models and for 2-hour glucose concentrati

Insulin-resistant Prediabetic Subjects Have More Atherogenic Risk Factors Than Insulin-sensitive Prediabetic Subjects: Implications for Preventing Coronary Heart Disease During the Prediabetic State

BACKGROUND Subjects who convert to type 2 diabetes mellitus have increased cardiovascular risk factors relative to nonconverters. However, it is not known whether these atherogenic changes in the prediabetic state are predominantly due to insulin resistance, decreased insulin secretion, or both. METHODS AND RESULTS We examined this issue in the 7-year follow-up of the San Antonio Heart Study, in which 195 of 1734 subjects converted to type 2 diabetes. At baseline, converters had significantly higher body mass index, waist circumference, triglyceride concentration, and blood pressure and lower HDL cholesterol than nonconverters. Atherogenic changes in converters were markedly attenuated (and no longer significant) after adjustment for the homeostasis model assessment of insulin resistance (HOMA IR, a surrogate for insulin resistance); in contrast, the differences in risk factors between converters and nonconverters increased after adjustment for the ratio of early insulin increment to early glucose increment (DeltaI(30-0)/DeltaG(30-0)) during an oral glucose tolerance test (a surrogate for insulin secretion). We also compared converters who had a predominant insulin resistance (high HOMA IR and high DeltaI(30-0)/DeltaG(30-0)) (n=56) and converters who had a predominant decrease in insulin secretion (low HOMA IR and low DeltaI(30-0)/DeltaG(30-0)) (n=31) with nonconverters (n=1539). Only the converters who were insulin resistant had higher blood pressure and triglyceride levels and lower HDL cholesterol levels than nonconverters. CONCLUSIONS Our data suggest that atherogenic changes in the prediabetic state are mainly seen in insulin-resistant subjects and that strategies to prevent type 2 diabetes might focus on insulin-sensitizing interventions rather than interventions that increase insulin secretion because of potential effects on cardiovascular risk.