The Metabolic Syndrome and Antioxidant Concentrations: Findings From the Third National Health and Nutrition Examination Survey

Between 1988 and 1994, a representative sample of the noninstitutionalized civilian U.S. population, selected by using a multistage, stratified sampling design, participated in the Third National Health and Nutrition Examination Survey (NHANES III). Survey participants were interviewed and invited for a clinical examination. For most participants, blood was drawn at the examination clinic, but for some who were unable to attend the examination because of health reasons, a blood sample was obtained during the home interview. People ages ≥60 years and African Americans and Mexican Americans were oversampled. Details about the survey and its methods have been previously published (12,13,14).

According to Adult Treatment Panel III criteria (15), a participant has the metabolic syndrome if he or she has three or more of the following criteria: 1) abdominal obesity: waist circumference >102 cm in men and >88 cm in women; 2) hypertriglyceridemia: ≥150 mg/dl (1.695 mmol/l); 3) low levels of HDL cholesterol: <40 mg/dl (1.036 mmol/l) in men and <50 mg/dl (1.295 mmol/l) in women; 4) high blood pressure: ≥130/85 mmHg; 5) high fasting glucose: ≥110 mg/dl (≥6.1 mmol/l).

The waist circumference was measured at the high point of the iliac crest at minimal respiration to the nearest 0.1 cm. Three readings of systolic and diastolic blood pressure were obtained from participants who attended the mobile examination center; the average of the last two measurements was used. We considered the current use of antihypertensive medication an indication of high blood pressure. Serum concentrations of HDL cholesterol (after precipitation with a heparin-MnCl₂ solution) and triglyceride concentrations (after hydrolysis to glycerol) were measured enzymatically with a Hitachi 704 analyzer (Boehringer Mannheim Diagnostics, Indianapolis, IN). Plasma glucose concentrations were measured using an enzymatic reaction (Cobas Mira Chemistry System; Roche Diagnostic System, Montclair, NJ). Participants who reported using insulin or oral antidiabetic medications were considered to have hyperglycemia.

NHANES III participants attended one of three examination sessions: morning, afternoon, or evening. Those attending the morning sessions were asked to fast for 10–16 h. Those attending the afternoon and evening sessions were asked to fast for at least 6 h.

Concentrations of retinol and retinyl esters, vitamin E, and five carotenoids (α- and β-carotene, β-cryptoxanthin, lutein/zeaxanthin, and lycopene) were measured in serum with an isocratic reverse-phase high-performance liquid chromatography (HPLC) method, with detection at 325, 300, and 450 nm respectively. Selenium concentrations were measured in serum using graphite furnace atomic absorption spectroscopy. Detailed procedures for these assays were published in the NHANES III laboratory manual (14).

The analyses included the following variables: age, sex, race or ethnicity, education, smoking status, serum cotinine concentration, BMI, leisure-time physical activity, total cholesterol concentration, non-HDL cholesterol concentration, insulin concentration, and vitamin or mineral use during the previous 24 h, and intake of fruits and vegetables. We created three levels of smoking status: participants who currently smoked (had smoked 100 cigarettes and were currently smoking), participants who had quit smoking (had smoked 100 cigarettes and were not currently smoking), and participants who had never smoked (had never smoked 100 cigarettes). Serum cotinine concentration was determined using HPLC atmospheric-pressure chemical ionization tandem mass spectrometry. BMI was calculated as weight in kilograms divided by height in meters squared. Respondents were asked about their participation and frequency of participation in the following activities during the previous month: walking, jogging or running, bicycling or using an exercise bicycle, swimming, aerobics or aerobic dancing, other dancing, calisthenics or exercises, gardening or yard work, and lifting weights. In addition, participants could report up to four additional physical activities. We created a physical activity index by summing the products of the frequency of participation by the metabolic equivalent levels for each reported activity. One metabolic equivalent is the energy expenditure of ∼3.5 ml oxygen · kg body wt⁻¹ · min⁻¹ or 1 kcal · kg body wt⁻¹ · h⁻¹. Serum concentrations of total cholesterol were measured enzymatically with a Hitachi 704 analyzer (Boehringer Mannheim Diagnostics). We calculated non-HDL cholesterol concentrations by subtracting the concentration of HDL cholesterol from that of total cholesterol. Plasma insulin was measured by radioimmunoassay using the Pharmacia Insulin RIA Kit (Pharmacia Diagnostics AB, Uppsala, Sweden). Participants were asked about their vitamin or mineral use with the following two questions: “Have you taken any vitamins or minerals during the past 24 h?”and “Have you taken any vitamins or minerals in the past month?” To estimate the number of monthly fruit and vegetable servings, we summed responses to 21 items on the NHANES III food frequency questionnaire. The food frequency questionnaire, with a special emphasis on foods that contribute to calcium, vitamin A, and vitamin C intake, was developed for use in NHANES III and relies in part on data about food consumption reported by African Americans and Mexican Americans who participated in NHANES II and Hispanic HANES. Although the reliability and validity of these questions have not been tested, the questionnaire has been subjected to cognitive testing and was pretested. In addition, we calculated mean dietary intakes for vitamin A, vitamin C, vitamin E, and carotenes from a single 24-h dietary recall.

We limited our analyses to participants who were aged ≥ 20 years and who had fasted ≥8 h. We also excluded women who were pregnant. When age adjustment was performed, the data were adjusted to the 2000 U.S. population ages ≥20 years using the direct method (16). To compare categorical and continuous variables of participants with and without the metabolic syndrome, we used χ² and t tests, respectively. Associations between antioxidant concentrations and metabolic syndrome status were examined using multiple linear regression analysis. For vitamin A, retinyl esters, and vitamin E, we also examined the association between residuals, calculated from regressing concentrations of cholesterol and triglycerides on these vitamins, and metabolic syndrome status. To account for the complex survey design, we used SUDAAN and the medical examination clinic sampling weights to produce our weighted estimates and standard errors (17).

The unadjusted prevalence of the metabolic syndrome was 21.9 ± 0.9% and the age-adjusted prevalence was 23.7 ± 0.8% for the 8,808 participants for whom we were able to establish their status. Participants with the metabolic syndrome were older, more likely to be white, had fewer years of education, were less likely to engage in moderate or vigorous physical activity, and had higher lipid concentrations (total cholesterol, non-HDL cholesterol, and triglycerides) and higher insulin concentrations, and consumed fewer fruits and vegetables (Table 1). The use of vitamin and mineral supplements was similar between the two groups. No significant differences in dietary intake of vitamin C, vitamin E, and carotenes were present among participants with and without the metabolic syndrome. However, the dietary intake of vitamin A was significantly lower among participants with than among those without the metabolic syndrome.

The age-adjusted concentrations of vitamin C, α- and β-carotene, β-cryptoxanthin, and lutein/zeaxanthin were lower among participants with than those without the metabolic syndrome (Table 1). On the other hand, those with the syndrome had higher concentrations of retinol and vitamin E than those without the syndrome. No significant differences in concentrations of serum retinyl esters, lycopene, and selenium existed.

To examine whether metabolic syndrome status was independently associated with the antioxidants, we ran multiple linear regression models that included age, sex, race or ethnicity, education, smoking status, cotinine concentration, physical activity, fruit and vegetable intake, and vitamin or mineral use during the previous 24 h (Table 2). Participants with the syndrome had lower concentrations of retinyl esters, vitamin C, and all carotenoid concentrations, except lycopene, than participants without the syndrome. Concentrations of vitamins A and E were positively associated with metabolic syndrome status.

Because concentrations of fat-soluble vitamins are closely correlated with blood lipid concentrations, we examined several models to take into account concentrations of total cholesterol and triglycerides. After adding concentrations of total cholesterol and triglycerides to the regression models or in models using residuals after adjusting for concentrations of cholesterol and triglycerides, vitamin A was no longer significantly associated with and vitamin E was inversely associated with metabolic syndrome status.

To examine the question of whether vitamin or mineral supplement use affected the association between antioxidant concentrations and metabolic syndrome status, we tested interaction terms between vitamin or mineral use and metabolic syndrome in the multiple linear regression models. Significant interactions were observed for models of retinol, vitamin E, and cryptoxanthin. In all three instances, the magnitude of the regression coefficient was stronger among participants who did not use supplements than among those who did use supplements. In addition, in analyses limited to participants with the metabolic syndrome, those who were using vitamin or mineral supplements had significantly higher concentrations of retinol, retinyl esters, vitamin C, vitamin E, α- and β-carotene, and β-cryptoxanthin than those who did not use such supplements (data not shown).

We repeated the analyses (results in Tables 1 and 2) after eliminating participants with self-reported diabetes and participants with a fasting glucose concentration ≥126 mg/dl. Among 7,980 participants, the unadjusted prevalence of the metabolic syndrome was 18.0 ± 0.7% and the age-adjusted prevalence was 20.1 ± 0.7%. The pattern of the results in these two tables was unchanged after these subanalyses (data not shown).

Of the five criteria for the metabolic syndrome, waist circumference, low levels of HDL cholesterol concentration, and hypertriglyceridemia were most often significantly associated with various antioxidant concentrations (Table 3). Waist circumference was inversely associated with all antioxidant concentrations except retinol and lycopene. HDL cholesterol concentration was inversely associated with retinol, retinyl esters, carotenoids (except α- and β-carotene), and selenium. In contrast, hypertriglyceridemia was more often directly associated with the antioxidants. Hypertension and hyperglycemia were generally inversely associated with some of the antioxidants.

Concentrations of retinol and vitamin E were positively associated with the number of metabolic syndrome criteria (Table 4). In contrast, concentrations of vitamin C and carotenoids (except for lycopene) were inversely associated with the number of metabolic syndrome criteria. Compared with participants with no metabolic syndrome criteria, those with five criteria had 40% lower vitamin C concentrations and 27% lower total carotenoid concentrations. Concentrations of retinyl esters, lycopene, and selenium were not significantly associated with the number of metabolic syndrome criteria after adjustment for covariates.

The metabolic syndrome is highly prevalent in the U.S. population (18). Thus, a large group of people are at increased risk for developing diabetes and cardiovascular disease. Because antioxidants may play a role in the pathophysiology of both of these conditions, our finding that people with the metabolic syndrome have lower concentrations of several antioxidants is of potential concern. We are not aware of previous studies that have examined antioxidant status among people with the metabolic syndrome.

The lower antioxidant concentrations among those with the metabolic syndrome may have resulted from lower intakes of antioxidants, increased use of antioxidants, or both. Our data show that participants with the metabolic syndrome consumed fewer fruits and vegetables than those without this syndrome and that about equal proportions of the two groups reported using vitamin or mineral supplements. Thus, our results suggest that antioxidant intake may be lower among participants with than those without the metabolic syndrome. Lower consumption of fruits and vegetables may have contributed to the reduced concentrations of vitamin C and carotenoids among participants with the metabolic syndrome. The dietary data from NHANES III provided us with limited ability to examine the contribution of diet to the antioxidant patterns we observed.

After adjusting for use of vitamins or minerals and intake of fruits and vegetables, people with the metabolic syndrome still had lower concentrations of several antioxidants, including retinyl esters, vitamin C, vitamin E, and several carotenoids. This finding suggests that increased use of antioxidants probably contributed to the reduced antioxidant concentrations found among participants with the metabolic syndrome. A likely mechanism for this increased use is that high levels of oxidative stress deplete endogenous and exogenous pools of antioxidants. NHANES III did not include measures of oxidative stress and, therefore, we were unable to examine whether people with the metabolic syndrome have higher levels of oxidative stress than people without this syndrome. However, it seems reasonable to postulate this because four of the five components of the metabolic syndrome (obesity, hyperglycemia, hypertension, and hypertriglyceridemia) are characterized by high oxidative stress (7,9,19–21). Additional sources of increased oxidative stress include hypercholesterolemia (22) and inflammation (23,24,25).

We did not find that concentrations of retinol, lycopene, and selenium differed according to metabolic syndrome status. Similar intake or use of these nutrients may have accounted for our findings. In a study of the depletion of various antioxidants in human plasma exposed to cigarette smoke, lycopene was the first and retinol and tocopherol were the last to be depleted (26). The similar concentration of lycopene among participants with and without the syndrome was somewhat unexpected. Whether this was attributable to similar intakes of tomatoes and tomato-based products coupled with similar rates of use between the two groups or to higher lycopene intake coupled with higher use among people with compared with those without the metabolic syndrome is unclear.

Usually, retinol from the liver, the main storage site for retinol, is transported to peripheral tissues by retinol-binding protein. Retinol may be released as a retinyl ester, however, when the ability of the liver to store retinol is exceeded or when liver function is impaired (27). Thus, the higher retinyl ester concentrations among those who did not have the metabolic syndrome may indicate that they consumed larger amounts of vitamin A compared with people who have this syndrome.

Our findings may have implications for people with the metabolic syndrome, health care professionals who care for them, and researchers who study the metabolic syndrome. People with the metabolic syndrome are at increased risk for diabetes and cardiovascular disease, and a role for oxidative stress in the pathophysiology of these conditions has been postulated. Although quenching of free radical species is one of the principal mechanisms of action of antioxidants, other mechanisms that affect the pathophysiology of diabetes and cardiovascular disease may be operating as well (28). The effects of vitamins C and E have received a great deal of interest. Through effects on oxidation of LDL cholesterol concentration, leukocyte adhesion, and endothelial function, vitamins C and E may slow atherosclerosis (29). For example, vitamin C and E intakes were positively associated with paraoxonase activity (30). Although the effects of carotenoids on atherosclerosis may be similar to those of vitamins C and E (i.e., a free radical species inhibition effect), lycopene may also have specific effects on cholesterol synthesis (8).

Much remains to be learned about the role of antioxidants in the prevention of diabetes. Results from prospective studies have suggested that the intake of fruits and vegetables is inversely related to the risk of developing diabetes (31,32,33,34,35). In addition, the intake of vitamin or mineral supplements may be associated with a reduced risk for diabetes (36). Studies of circulating concentrations of antioxidants have provided mixed evidence of a benefit of antioxidants on the risk of diabetes (37,38). In a trial of β-carotene supplementation, diabetes incidence was not reduced among the experimental group compared with the control group (39).

Oxidation and attendant free radical damage may contribute to the pathogenesis of diabetes through impairment of insulin-mediated phosphatidylinositol 3-kinase and the resultant impairment of GLUT4 translocation (40). In addition, β-cells, which are characterized by low concentrations of radical scavengers (41), are thought to be highly susceptible to damage from reactive oxygen species (42). Increased free radical production has been shown to correlate inversely with insulin action (43). Although several studies have shown inverse associations between measures of insulin resistance and antioxidant concentrations (44) and improvements in insulin action with antioxidant supplementation (45), not all studies have shown such associations (46). Of note is that people with the metabolic syndrome have a high prevalence of insulin resistance, which is associated with an increased risk of developing diabetes.

Our results may have some clinical implications. Although measurement of antioxidants is not routinely performed in clinical practice, reviewing the intake of foods rich in antioxidants (particularly fruits and vegetables) and the use of antioxidant supplements among patients with the metabolic syndrome may be instructive. As a group, people with the metabolic syndrome should be encouraged to consume adequate foods rich in antioxidants. If their dietary intake of vitamins A, C, and E fails to meet the recommended daily allowance, health care professionals should encourage people with the metabolic syndrome to increase their intake of these vitamins, preferably through the consumption of healthy food sources rich in these vitamins or otherwise through the use of appropriate vitamin supplements. Participants with the metabolic syndrome who used vitamin or mineral supplements had higher concentrations of several antioxidants than those who did not use such supplements in our analyses.

Our results point to several research opportunities. First, our findings need to be confirmed by other studies. Second, studies of other endogenous and exogenous antioxidants are of interest in the setting of the metabolic syndrome. Third, establishing whether people with the metabolic syndrome have higher levels of oxidative stress needs to be examined. Fourth, studies of antioxidant intake with better dietary instruments would be helpful in establishing how much of the antioxidant deficit among people with the metabolic syndrome may be attributable to inadequate intake of antioxidants. Fifth, studies of the effects of antioxidant supplements on the prevention or delay of chronic diseases among people with the metabolic syndrome have not been performed. Because our results show that these people have low concentrations of several antioxidants, they may be an interesting group to study the effects of antioxidant supplementation or dietary modification to enhance the intake of antioxidants.