The Metabolic Syndrome, Circulating Oxidized LDL, and Risk of Myocardial Infarction in Well-Functioning Elderly People in the Health, Aging, and Body Composition Cohort

The study population was comprised of 3,033 participants in the baseline exam of the Health ABC study. Participants were enrolled between March 1997 and June 1998. The study’s target population was individuals age 70–79 years who did not have lower extremity functional limitation and resided in the Memphis, Tennessee, and Pittsburgh, Pennsylvania, areas. These criteria selected the best functioning 40–60% of the older population. The socioeconomic status patterns reflected the locations from which the samples were drawn: 56.4% of black participants vs. 87.8% of white participants had completed high school, and the median income category was $10,000–25,000 for black participants and $25,000–50,000 for white participants. A participant’s well-functioning was determined by self-report and defined as having the individual having no difficulty in walking a quarter of a mile or going up 10 steps without resting reported on two separate occasions. Exclusion criteria included difficulties with daily living activities, obvious cognitive impairment, an inability to communicate with the interviewer, an intention of moving within 3 years, therapy for cancer within the prior 3 years, and previous participation in a trial involving a lifestyle intervention.

Incident CHD was defined by coronary death or any overnight hospitalization in an acute care hospital for myocardial infarction (MI), angina, coronary angioplasty or artery bypass surgery, or chronic heart failure. Events and diagnoses were adjudicated based on hospital and death records. Participants were asked to report any hospitalizations, and every 6 months they were asked direct questions to elicit information about any event. When an event was reported, hospitalization records were collected and verified by a Health ABC Disease Adjudication Committee. Date and causes of death were taken from the death certificate. Only events that were confirmed by the adjudication committee were included.

Based on cardiac procedures and diagnoses reported to the U.S. Department of Health and Human Services’ Centers for Medicare and Medicaid Services in the 5 years before enrollment, 385 participants had CHD (MI, angina, coronary angioplasty or artery bypass surgery, or chronic heart failure). Based on ATP III criteria, 1,113 had CHD risk equivalents: noncoronary forms of clinical atherosclerotic disease (n = 232), diabetes (n = 577), and/or a 10-year risk for CHD events >20% by Framingham scoring (n = 1,023). Of those with CVD, 60% were treated with antianginal drugs; 42% with antiarrhythmic drugs; 35% with vasodilators; 46% with antiplatelet drugs, including aspirin; 90% with blood pressure−lowering drugs; and 21% with lipid-lowering drugs (10). Since being enrolled in the study (follow-up through June 2002), 418 participants had a CHD event (MI, angina, coronary angioplasty or artery bypass surgery, chronic heart failure, or cardiovascular death). There were 238 events in participants without baseline CHD. Among these, 120 had at least one MI: 106 had one and 14 had more than one based on electrocardiogram and enzyme changes. There were 88 primary and 32 secondary MIs.

During an in-clinic visit, fasting blood specimens were drawn and processed, with plasma and serum stored at −80°C. From these blood samples, total and HDL cholesterol and serum triglycerides were measured using a Vitros 950 analyzer (Johnson & Johnson). The interassay coefficient of variation (CV) was 1.5% for total cholesterol, 2.3% for triglycerides, and 2.3% for HDL cholesterol. LDL cholesterol levels were calculated by the Friedewald equation (15). Levels of oxLDL were measured (2000–2001) blindly at the Center for Experimental Surgery and Anesthesiology. An mAb-4E6−based competition enzyme-linked immunosorbent assay (ELISA) was used for measuring plasma oxLDL levels (5,16,17). The monoclonal antibody mAb-4E6 is directed against a conformational epitope in the apoB-100 moiety of LDL that is generated by substituting aldehydes for at least 60 lysine residues of apolipoprotein B-100. This number of substituted lysines corresponds to the minimal number of substituted lysines required for scavenger-mediated uptake of oxLDL. Substituting aldehydes can be produced by peroxidation of LDL lipids, resulting in the generation of oxLDL. Aldehydes that are released by endothelial cells under oxidative stress or by activated platelets may also induce the oxidative modification of apolipoprotein B-100 in the absence of the peroxidation of LDL lipids. Specificity of the assay is excellent, as the C₅₀ values (i.e., concentrations required to obtain 50% inhibition of antibody binding in the ELISA) are 25 mg/dl for native LDL isolated by ultracentrifugation from the plasma of healthy volunteers and 0.025 mg/dl for oxLDL with at least 60 aldehyde-substituted lysines per apolipoprotein B-100 obtained by copper-ion induced oxidation of the same LDL (16). Therefore, 160 mg/dl LDL in a person’s plasma would contribute <0.2 mg/dl in the oxLDL assay. The interassay CV of oxLDL is 12%.

Measurements of waist circumference were taken with participants standing and were made at the level of widest circumference in the area between the iliac crest and lower rib. Blood pressure was taken three times with participants in the seated position after 5 min of quiet rest. The average of the last two measurements was used for systolic and diastolic blood pressure.

Metabolic syndrome components were defined as detailed in the ATP III report: 1) waist circumference >102 cm in men and >88 cm in women, 2) fasting triglycerides ≥1.70 mmol/l (150 mg/dl), 3) HDL cholesterol <1.03 mmol/l (40 mg/dl) in men and <1.29 mmol/l (50 mg/dl) in women, 4) blood pressure ≥130/85 mmHg, and 5) fasting-glucose ≥6.10 mmol/l (110 mg/dl). Participants with at least three of these components were determined to have the metabolic syndrome.

Before measuring oral glucose tolerance, participants (excluding those taking insulin or hypoglycemic medications) were asked to consume a solution containing 75 g glucose within a 10-min time window after an overnight fast (≥12 h). Glucose and insulin were measured before the glucose load was administered and 2 h after a participant began consuming the drink.

Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS for Windows, release 11; Chicago, IL). The nonparametric Kruskal-Wallis test was used when comparing continuous variables, and a Fisher’s exact test was used for analyzing contingency tables. Logistic regression analysis was used to study the relation of the metabolic syndrome with circulating oxLDL (expressed in mg/dl or as the percent of LDL). The odds ratio (OR) for high oxLDL (>90th percentile of distribution in individuals with no metabolic syndrome components; cutoff value of 1.90 mg/dl or 1.58% of LDL) was determined using individuals without the metabolic syndrome as a reference group. Logistic regression analysis was also used to study the relation of greater waist circumference; higher blood pressure; higher fasting triglycerides, insulin, glucose, and HbA_1c levels; and lower HDL cholesterol with the prevalence of high oxLDL. The OR for high oxLDL for participants in the higher quintiles of waist circumference, systolic blood pressure, triglycerides, insulin, glucose, and HbA_1c and lower quintiles of HDL cholesterol was determined by comparing these individuals with participants in the lowest quintile of waist circumference, systolic blood pressure, triglycerides, insulin, glucose, and HbA_1c and the highest quintile of HDL cholesterol. P < 0.05 was considered statistically significant. The relative risk of total CHD and MI according to the metabolic syndrome, compared with individuals without the metabolic syndrome, and according to oxLDL (per quintile) was determined by logistic regression analysis after adjusting for age, sex, ethnicity, and smoking status. An interaction term was fitted to assess whether the metabolic syndrome determined the oxLDL effect on CHD/MI.

Only 4% of the participants in the Health ABC study had no metabolic syndrome components; 58% of subjects had one or two abnormalities and 38% had three or more abnormalities. Prevalence of the metabolic syndrome as defined by ATP III (three or more components) in the Health ABC study was 38 vs. 42% in a similar-aged population in the NHANES III (14). In all, 13% of the Health ABC participants had CHD compared with 20% in the NHANES III cohort. Prevalence of diabetes was 19 vs. 18% and the prevalence of current smoking was 10 vs. 15%, respectively. The mean systolic blood pressure was 136 vs. 142 mmHg; mean waist circumference, 100 vs. 97 cm; mean HDL cholesterol levels, 1.40 vs. 1.33 mmol/l; and mean triglyceride levels, 1.56 and 1.79 mmol/l, for the two study populations, respectively. Mean BMI was 27 kg/m² in both the Health ABC cohort and the age-matched NHANES III cohort.

Table 1 summarizes the characteristics of the study cohort, contrasting participants without and with the metabolic syndrome. Compared with those without the metabolic syndrome, participants with the syndrome were more likely to be female and have a heavier BMI (weight in kilograms divided by height in meters) (2), larger waist circumference, higher blood pressure, more glucose intolerance and diabetes, lower HDL cholesterol, and higher triglyceride level. Smoking habits were similar for those with and without the metabolic syndrome. OxLDL levels (absolute levels expressed in mg/dl or relative levels expressed in percent of LDL) were higher in those with the metabolic syndrome (Table 1).

Compared with those without the metabolic syndrome, the OR for high oxLDL (>1.90 mg/dl) in participants with the metabolic syndrome was 1.82 (P < 0.001), after adjusting for age, sex, ethnicity, and smoking status (Table 2). No interaction with sex and ethnicity was observed. After further adjusting for LDL cholesterol, the OR was 2.01 (P < 0.001) (Table 2). When oxLDL was expressed as the percent of LDL, the adjusted OR for high oxLDL (>1.58%) was 2.56 (P < 0.001) (Table 2).

Logistic regression analysis was performed to study the relation between the degree of the different metabolic syndrome components and oxLDL. Compared with those in the lowest quintile and adjusting for age, ethnicity, and LDL cholesterol, the prevalence of high oxLDL was higher in higher quintiles of waist circumference, serum triglycerides, and insulin and glucose levels. Compared with the highest quintile of HDL cholesterol, the prevalence of high oxLDL was higher in lower quintiles of HDL cholesterol. No significant association between blood pressure and oxLDL was observed (Table 3). Compared with the lowest quintile of the triglyceride-to-HDL molar ratio, the OR for high oxLDL was 1.57 (95% confidence interval [CI] 1.01–2.44) in the 2nd quintile and increased to 2.03 (1.33–3.09), 2.45 (1.62–3.71), and 3.74 (2.46–5.67) in the 3rd, 4th, and 5th quintiles, respectively. Compared with the lowest quintile, the OR for high oxLDL in the highest quintile of HbA_1c was 1.67 (1.25–2.23), after adjusting for age, sex, ethnicity, lipid levels, smoking status, and glucose and insulin levels.

Compared with participants without the metabolic syndrome and adjusting for age, sex, ethnicity, and smoking status, those with the metabolic syndrome had a 1.69-fold higher risk (95% CI 1.29–2.34) for CHD events. OxLDL did not predict the risk of CHD events (relative risk for participants in highest quintile 1.23; 0.83–1.82). Relative risk ratios for future CHD events for those without CHD at baseline were 1.52 (1.15–2.01) for the metabolic syndrome and 1.45 (0.91–2.34) for oxLDL (highest versus lowest quintile).

Recently, oxLDL was found to be associated with acute coronary events (5,18). We therefore determined the predictive value of oxLDL for new MI events. Table 4 illustrates the relative risk of MI in relation to the metabolic syndrome and oxLDL as determined by logistic regression analysis and adjusted for age, sex, ethnicity, and smoking status. Those with the metabolic syndrome had a 2.0-fold higher risk. The incidence rate for those with the metabolic syndrome was 5.6 vs. 2.9% (P < 0.001) for those without the metabolic syndrome. We also divided the cohort into five groups by levels of oxLDL (Table 4). The incidence of MI was 5.7% in the highest quintile of oxLDL compared with 2.6% (P < 0.01) in the lowest quintile. The risk ratio for participants in the highest quintile was 2.25 (95% CI 1.22–4.15). After adjusting for the metabolic syndrome, the risk ratio for participants in the highest quintile of oxLDL was 1.87 (1.00–3.49). There was no significant interaction between the metabolic syndrome and oxLDL, suggesting that the effect of oxLDL on MI was independent of the metabolic syndrome.

We have shown for the first time in a population cohort that the metabolic syndrome is associated with a higher fraction of oxLDL and thus with higher levels of circulating oxLDL. This association was consistent across sex and ethnicity. Our data further support the importance of identifying individuals with the metabolic syndrome as a high-risk group for developing CHD. Finally, our study identified the oxidation of LDL as a potential mechanism explaining the increased risk for MI among those with the metabolic syndrome.

In elderly individuals, three dominant factors have been identified in the relation between the metabolic syndrome and ischemic heart disease: 1) the central metabolic factor, comprised of obesity, fasting and nonfasting insulin, and dyslipidemia (HDL cholesterol loading negative and triglycerides loading positive), 2) the glucose factor, and 3) the blood pressure factor (19,20). Our study showed a strong association with two out of the three metabolic syndrome factors: the central metabolic syndrome factor, with obesity, dyslipidemia, and insulin resistance, and the glucose factor. The association with these different factors suggests that oxLDL may be generated through several mechanisms.

Recently, an association between small LDL particle size and the oxidation of LDL has been shown in healthy middle-aged men (21). In the elderly population, lipids are significantly associated with ischemic heart disease, largely based on the impact of high triglycerides and low HDL cholesterol and less so for LDL cholesterol (20). Although individuals with the metabolic syndrome often have average levels of LDL, they may have qualitative abnormalities. Small, dense LDL particles (22) have been associated with components of the insulin resistance syndrome, including hypertriglyceridemia, low serum HDL cholesterol levels, and diabetes. Also, small, dense LDL particles, which are particularly prone to oxidation, have been proven to be more atherogenic than larger LDL particles (23–26). In the Health ABC study, the prevalence of small, dense LDLs has not been directly studied. It has, however, been shown that 90% of those with a triglyceride-to-HDL molar ratio >1.33 had small, dense LDL (27). In this study, a triglyceride-to-HDL molar ratio >1.33 was a better predictor of elevated levels of oxLDL than high triglycerides and low HDL cholesterol separately, suggesting an association between the small, dense LDL phenotype and the oxidation of LDL in individuals with the metabolic syndrome.

Previously, an inverse relation between HDL cholesterol and circulating oxLDL has been shown in healthy middle-aged men (21). This inverse relation can be explained by the antioxidant action of HDL components (28). The HDL-associated enzymes paraoxonase and lecithin:cholesterol acyltransferase prevent LDL oxidation and render LDL resistant to oxidation (29,30). Recently, we used mice with both leptin and LDL receptor deficiency to study the relation of insulin resistance and dyslipidemia with atherosclerosis (31). We demonstrated that the high susceptibility of these mice for atherosclerosis was due to increased oxidative stress, as evidenced by the higher levels of antibodies against oxLDL. Those higher levels resulted from decreased antioxidant defense of HDL related to decreased paraoxonase and lecithin:cholesterol acyltransferase activity. It remains to be investigated whether the decreased antioxidant defense of HDL can explain the inverse relation between HDL cholesterol and oxLDL in individuals with the metabolic syndrome.

It has been shown in healthy, nondiabetic volunteers that plasma glucose and insulin levels correlate with a higher susceptibility of ex vivo oxidation of LDL (32). Here we have shown that hyperinsulinemia and impaired glycemic control are associated with increased in vivo LDL oxidation, as reflected by the higher prevalence of high oxLDL. It has been demonstrated that impaired glucose tolerance in elderly individuals is associated with the formation of glycoxidation products of LDL that accumulate in human atherosclerotic plaques (33). Not only can glucose induce LDL oxidation, it also can impair the antioxidant properties of serum albumin (34).

Previously, an inverse correlation has been observed between HbA_1c, our commonly used clinical index of glycemic control, and the lag time for ex vivo LDL oxidation, a proxy for LDL susceptibility to oxidation (35,36). Here we have shown a relation between HbA_1c and in vivo oxLDL using an ELISA that also detects LDL glycoxidation products.

It has been shown that insulin treatment prevents LDL from accelerated in vitro oxidation (37), which suggests a protective effect of insulin against LDL oxidation. Here we demonstrated, however, that hyperinsulinemia is associated with a higher prevalence of having high levels of in vivo oxLDL. Our data are thus in agreement with a synergistic effect of insulin and oxLDL in atherogenesis. Indeed, it has been demonstrated that insulin interacts with cytokines and growth factors in vascular wall cells in concert with oxidized and/or glycated lipoproteins by inducing the expression of monocyte/macrophage specific adhesion molecules such as intracellular adhesion molecule-1 (38).

As in the Healthy American Women Study and the West of Scotland Coronary Prevention Study (25,39), the modified National Cholesterol Education Program metabolic syndrome definition predicted total CHD and MI risk in the Health ABC cohort. OxLDL did not predict the total CHD risk, but did predict MI. These observations, along with previous research into the association of LDL oxidation with atherosclerosis and CHD (8,10,40,41), provide evidence that LDL oxidation is a common basis for the metabolic syndrome and CHD or, more particularly, for inducing atherothrombotic coronary disease. Previous studies in the hypercholesterolemic miniature pig have suggested that the oxidation of LDL occurs in the arterial wall and not in the circulation (42,43). Recently, Tsimikas et al. (18) demonstrated temporal increases of circulating oxLDL in association with acute coronary syndromes. In aggregate, these data support the hypothesis that an increase in oxLDL reflects plaque instability.

This epidemiological study does not give insight into the mechanisms by which the metabolic syndrome and oxLDL might contribute to the development of CHD or plaque instability. Previous studies into the effect of in vitro oxLDL on vascular cell function did not determine the minimal amounts of oxLDL required to alter the redox state of vascular cells. These studies typically use concentrations of oxLDL that are 10-fold higher than those detected in the plasma.

This is the first report of an association between the metabolic syndrome and a high prevalence of high oxLDL in a population-based cohort. This association remained valid after adjusting for LDL cholesterol, sex, and ethnicity. Our data also further support the predictive value of the metabolic syndrome for CHD and suggest that baseline levels of oxLDL add prognostic information concerning future risk for MI.

This work was supported by National Institutes of Health Grants N01-AG-6-2103, N01-AG-6-210, and N01-AG-6-2106; the Interuniversitaire Attractiepolen Programma of the Belgian Federal Government (P05/02); and the Fonds voor Wetenschappelijk Onderzoek-Vlaanderen (G.0088.02).

We thank M. Landeloos for excellent technical assistance.