SEARCH

SEARCH BY CITATION

Keywords:

  • alcohol;
  • metabolic syndrome;
  • epidemiology;
  • wine

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Objective: To examine the association between total and beverage-specific alcohol consumption and the prevalence odds of metabolic syndrome (MS).

Research Methods and Procedures: Using a cross-sectional design, we studied 4510 white participants of the National Heart, Lung, and Blood Institute Family Heart Study. We used generalized estimating equations adjusting for age, education, risk group, smoking, physical activity, diabetes mellitus, coronary heart disease, energy intake, energy from fat, fruits, and vegetables, dietary cholesterol, dietary fiber, and use of multivitamins to estimate the prevalence odds of MS by alcohol intake.

Results: Compared with never-drinkers, multivariate odds ratios (95% confidence interval) for MS were 1.12 (0.85 to 1.49), 0.68 (0.36 to 1.28), 0.72 (0.50 to 1.03), 0.66 (0.44 to 0.99), and 0.80 (0.55 to 1.16) among men who were former drinkers and who were current drinkers of 0.1 to 2.5, 2.6 to 12.0, 12.1 to 24.0, and >24.0 g/d of alcohol, respectively (p for linear trend 0.018). Corresponding values for women were 0.86 (0.69 to 1.09), 0.80 (0.43 to 1.34), 0.47 (0.33 to 0.66), 0.47 (0.30 to 0.74), and 0.39 (0.21 to 0.74), respectively (p for trend < 0.0001). The reduced prevalence odds of MS was observed across all beverage types: compared with never-drinkers, multivariate adjusted odds ratios (95% confidence interval) of MS were 0.32 (0.14 to 0.73), 0.42 (0.23 to 0.77), 0.57 (0.30 to 1.09), and 0.56 (0.36 to 0.88) for subjects who consumed >7 drinks/wk of wine only, beer only, spirits only, and more than one type of beverage, respectively.

Discussion: Our data indicate that alcohol consumption is associated with a lower prevalence of MS irrespective of the type of beverage consumed. Prospective studies are needed to confirm these findings and to assess the influence of drinking patterns on the alcohol-MS association.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Multiple metabolic syndrome (MS)1 affects ∼20% of the U.S. population (1, 2). Features of MS—obesity, high blood pressure, hypertriglyceridemia, low concentration of high-density lipoprotein (HDL)-cholesterol, and hyperglycemia—are established risk factors for cardiovascular disease and diabetes. The relation of alcohol consumption to these features is complex: although alcohol consumption has been associated with increased risk of hypertension (3), moderate alcohol consumption is associated with a lower risk of diabetes mellitus (4, 5, 6), perhaps through improved insulin sensitivity (7, 8). In addition, alcohol consumption is associated with higher HDL-cholesterol (9, 10, 11) and higher triglycerides (TGs) (12). Limited data are available on the relation of total alcohol and MS. Further, it is not established whether the type of beverage consumed influences the alcohol-MS relation.

We used data collected on 4510 white participants of the National Heart, Lung, and Blood Institute (NHLBI) Family Heart Study to assess the association between total alcohol consumption and the prevalence of MS in men and women. In addition, we also evaluated whether the type of beverage consumed (beer, wine, or spirits) influences the alcohol-MS association.

Research Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Study Population

The NHLBI Family Heart Study is a multicenter, population-based study designed to identify and evaluate genetic and nongenetic determinants of coronary heart disease (CHD), preclinical atherosclerosis, and cardiovascular risk factors. Families in the study had been chosen either at random (referred to as random group) or based on a higher-than-expected risk of CHD (referred to as high-risk group) from previously established population-based cohort studies: The Framingham Heart Study in Framingham, MA; the Atherosclerosis Risk in Communities Study cohorts in North Carolina and Minnesota; and the Utah Health Family Tree Study in Salt Lake City, UT. The high-risk group was defined based on a family risk score, which compares the family's age and sex-specific incidence of CHD to that expected in the general population (13). A detailed description of the methods and design of the study has been published (14). Briefly, between 1993 and 1995, groups of individuals participating in each of the parent studies were selected at random and invited to furnish an updated family health history that contained information on their parents, children, and siblings. Of 4679 individuals contacted, responses were obtained from 3150 (67%); their family members were then contacted, and self-reported health data were obtained from a total of 22, 908 individuals (86% of those contacted). From the families responding to the health questionnaire, 588 families were chosen at random, and 566 families were selected based on higher-than-expected risk of CHD. We obtained written consent from each participant, and the study protocol was reviewed and approved by the Institutional Review Board at each of the participating institutions.

Blood Collection and Assays

All participants were asked to fast for 12 hours before arrival at the study center. For blood samples for lipids, glucose, and insulin, vacutainers without additives were used. Blood samples were then spun at 3000g for 10 minutes at 4 °C. Sera were stored at −70 °C until shipped periodically to a central laboratory at the Fairview-University Medical Center (Minneapolis, MN) for processing. Serum insulin was measured by a radioimmunoassay method (Diagnostic Products Corporation, Los Angeles, CA). Low-density lipoprotein (LDL)-cholesterol was estimated using the method of Friedewald (15) except for subjects with TGs above 4.5 mM, whose low-density lipoprotein was measured by ultracentrifugation.

Alcohol Assessment

Alcohol consumption data were collected during the clinic interviews by asking whether the person “ever consumed alcoholic beverages” or “presently drinks alcoholic beverages at all”; if the latter, subjects were asked specifically about the number of drinks per week of each type of beverage (beer, wine, spirits). In addition, subjects were asked: “Was there a time in your life when you drank five or more drinks of any kind of alcoholic beverage almost every day?” and whether each type of beverage (wine, beer, spirits) is “normally consumed with meals.” For purposes of the study, a “drink” was defined as a 0.36-liter bottle or can of beer containing 12.6 grams of alcohol, a 0.12-liter glass of wine containing 13.2 grams of alcohol, or a 0.038-liter shot of 80-proof spirits containing 15 grams of alcohol. Total alcohol was computed as the sum of the three beverage-specific alcohol contents. The reported alcohol intake was highly correlated with serum γ-glutamyltranspeptidase (an indicator of the effects of alcohol on the liver) measured on a subsample of subjects (r = 0.51, p < 0.0001 among drinkers). Detailed description of alcohol assessment in the NHLBI Family Heart Study has been published previously (16, 17, 18).

MS Definition

We used the National Cholesterol Education Program Adult Treatment Program guidelines (19) to define MS. Specifically, a subject was considered to have MS if he/she met at least three of the following criteria: abdominal waist circumference >1.02 m for men and >0.88 m for women, serum TGs of at least 1.69 mM, serum HDL-cholesterol < 1.03 mM for men or <1.29 mM for women, average systolic blood pressure of at least 130 mm Hg or diastolic blood pressure of at least 85 mm Hg or current use of antihypertensive medication, and fasting serum glucose of at least 6.1 mM or current use of hypoglycemic agents.

Other Variables

Dietary information was collected through a modified semiquantitative food frequency questionnaire (20). From the food frequency questionnaire, intake of specific nutrients was computed by multiplying the frequency of consumption of an item by its nutrient content. Composition values for nutrients were obtained from the Harvard University Food Composition Database derived from the U.S. Department of Agriculture sources and manufacturer information (21). Information on cigarette smoking was obtained by interview during the clinic visit. Oral contraceptive use and hormone replacement therapy were assessed using a reproductive history questionnaire and medication inventory. Frequency and duration of strenuous, moderate, and light physical activity during the previous year were estimated from a physical exercise questionnaire. Anthropometric data were collected with subjects wearing scrub suits. A balance scale was used to measure body weight, and height was measured using a wall-mounted vertical ruler. Cardiovascular disease was assessed from the medical history and a 12-lead electrocardiogram. Prevalent CHD was defined as a self-reported history of myocardial infarction, percutaneous transluminal coronary angioplasty, or coronary artery bypass graft, or the presence of Q-waves (Minnesota codes 1.1 to 1.2) on the resting 12-lead electrocardiogram. Detailed information on these covariates has been published (14, 18, 22).

Statistical Analysis

Because men consumed more alcohol than women, we initially conducted sex-specific analyses. We created the following categories of alcohol: never-drinkers, former drinkers, and current drinkers of 0.1 to 2.5, 2.6 to 12.0, 12.1 to 24.0, and >24.0 g/d of alcohol. Because subjects in this study were not independent (familial clustering), we used generalized estimating equations to calculate the prevalence odds ratios (ORs) for MS. This technique corrects the variance of the point estimate in the presence of familial clustering. We used never-drinkers as the reference category and made adjustment for age, age squared, education, risk group (random group or high-risk group), smoking, physical activity, history of CHD, energy intake, percentage of energy intake from fat, intake of fruit and vegetables, dietary cholesterol, dietary fiber, and use of multivitamins. Results from parsimonious models are presented. To test for trend, we computed median alcohol intake for each category of alcohol intake and created a new variable to which median alcohol values were assigned. This new variable was used as a continuous variable in testing for a linear trend. We also conducted sensitivity analyses to evaluate whether our results were influenced by the exclusion of 603 subjects with CHD and those currently receiving treatment for hypertension and diabetes mellitus, the number of MS components (three, four, and five), the exclusion of 493 subjects who reported that they have ever consumed at least five drinks per day, and whether or not alcohol was consumed with meals. We also evaluated the association between alcohol consumption and individual components of MS.

To assess beverage-specific effects, we excluded former drinkers and divided subjects into categories of never-drinkers and subjects who consumed: beer only, wine only, spirits only, and more than one type of alcoholic beverage. For each type of beverage, we created indicator variables for alcohol categories of 0.1 to 7 and >7 drinks/wk to assess for a dose-response relation. Using never-drinkers as the referent category, we fitted a generalized linear model with adjustments, as above. In addition, we controlled for total alcohol consumption in the multivariate model. All analyses were completed on PC SAS (windows version 5.1, release 8.02; SAS Institute Inc., Cary, NC).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Of the 4510 white subjects included in the analysis, 2077 were men, and 2433 were women. The mean age was 51.6 ± 13.7 years (range 25 to 91) for men and 52.4 years (range 25 to 93) for women. The range of alcohol intake was 0 to 124 g/d for men and 0 to 69 g/d for women.

Tables 1 and 2 show the gender-specific baseline characteristics according to categories of alcohol consumption. Subjects with higher alcohol consumption tended to be younger and leaner and consumed less dietary fiber, smoked more cigarettes, and had higher HDL-cholesterol and lower insulin concentration than never-drinkers.

Table 1.  Characteristics of participants (men) according to alcohol consumption in the NHLBI Family Heart Study*
 Alcohol (g/d)
Characteristics0 (Never drinkers) (n = 359)0 (Former drinkers) (n = 822)0.1 to 2.5 (n = 58)2.6 to 12.0 (n = 313)12.1 to 24.0 (n = 238)>24.0 (n = 287)
  • *

    F&V, fruits and vegetables; HS, high school education; continuous variables are shown as mean ± SD.

Age (years)53.8 ± 13.653.0 ± 13.752.8 ± 13.049.9 ± 14.948.0 ± 14.449.7 ± 13.7
BMI (kg/m2)27.8 ± 4.628.3 ± 4.927.2 ± 3.927.3 ± 4.027.1 ± 4.027.2 ± 3.9
F&V (servings/d)3.59 ± 1.883.22 ± 1.742.90 ± 1.553.18 ± 1.742.95 ± 1.572.67 ± 1.46
Dietary fiber (g/d)19.6 ± 8.917.9 ± 8.118.1 ± 7.818.4 ± 9.816.2 ± 8.415.5 ± 7.7
Dietary cholesterol (g/d)0.26 ± 0.150.26 ± 0.140.25 ± 0.120.26 ± 0.120.26 ± 0.120.28 ± 0.16
Energy intake (kJ)7808 ± 25727729 ± 26647953 ± 30317887 ± 27178129 ± 27648832 ± 2798
Percent energy from      
 Saturated fat11.5 ± 3.111.9 ± 3.011.8 ± 2.911.1 ± 3.011.0 ± 3.010.4 ± 2.9
 Monounsaturated fat12.3 ± 3.112.9 ± 3.313.0 ± 3.312.2 ± 3.212.0 ± 2.911.3 ± 2.9
 Polyunsaturated fat4.67 ± 1.344.75 ± 1.374.64 ± 1.314.66 ± 1.414.38 ± 1.244.00 ± 1.24
 Protein18.3 ± 3.718.4 ± 4.117.7 ± 3.718.1 ± 3.717.1 ± 3.316.5 ± 4.2
 Carbohydrate52.5 ± 9.551.1 ± 9.751.4 ± 9.250.0 ± 9.147.6 ± 9.042.5 ± 9.3
Exercise (min/d)29.7 ± 34.734.1 ± 42.134.8 ± 31.641.0 ± 42.544.3 ± 49.037.3 ± 41.8
LDL-cholesterol (mM)3.18 ± 0.823.30 ± 0.873.31 ± 0.953.24 ± 0.843.36 ± 0.873.35 ± 0.93
HDL-cholesterol (mM)1.04 ± 0.251.04 ± 0.241.12 ± 0.271.16 ± 0.291.23 ± 0.291.27 ± 0.34
Triglycerides (mM)1.91 ± 1.401.82 ± 1.181.69 ± 0.831.62 ± 1.201.66 ± 1.022.03 ± 1.88
Glucose (mM)5.67 ± 1.785.92 ± 2.205.67 ± 1.725.61 ± 1.725.58 ± 1.365.69 ± 1.49
Insulin (pM)95.8 ± 153.8101.5 ± 139.5137.4 ± 364.773.3 ± 45.179.8 ± 146.279.0 ± 57.6
Coronary heart disease (%)15.421.722.415.713.015.3
Diabetes mellitus (%)15.019.815.511.816.817.8
Current smoking (%)3.114.613.814.126.730.0
HS graduate or less (%)20.333.222.431.627.334.5
Risk group (% random)40.755.844.851.145.044.6
Use of multivitamins (%)35.438.336.233.135.437.6
Table 2.  Characteristics of participants (women) according to alcohol consumption in the NHLBI Family Heart Study*
 Alcohol (g/d)
Characteristics0 (Never drinkers) (n = 659)0 (Former drinkers) (n = 1084)0.1 to 2.5 (n = 87)2.6 to 12.0 (n = 364)12.1 to 24.0 (n = 159)>24.0 (n = 80)
  • *

    F&V, fruits and vegetables; HS, high school education; continuous variables are shown as mean ± SD.

Age (years)55.8 ± 13.451.5 ± 13.650.4 ± 15.149.0 ± 13.751.6 ± 12.853.9 ± 11.3
BMI (kg/m2)28.0 ± 6.227.7 ± 6.426.2 ± 6.125.6 ± 5.125.8 ± 5.425.8 ± 4.7
F&V (servings/d)3.83 ± 2.023.43 ± 1.753.71 ± 1.823.40 ± 1.593.52 ± 1.652.99 ± 1.49
Dietary fiber (g/d)19.0 ± 8.917.4 ± 7.816.7 ± 6.917.4 ± 7.317.4 ± 8.514.3 ± 7.2
Dietary cholesterol (g/d)0.22 ± 0.120.22 ± 0.100.19 ± 0.080.22 ± 0.100.23 ± 0.110.21 ± 0.09
Energy intake (kJ)6820 ± 22946700 ± 23036273 ± 21646772 ± 23167017 ± 23196642 ± 1996
Percent energy from      
 Saturated fat11.1 ± 3.211.3 ± 3.310.7 ± 3.110.8 ± 3.111.0 ± 2.910.0 ± 2.7
 Monounsaturated fat11.5 ± 3.211.8 ± 3.311.2 ± 3.411.3 ± 3.111.0 ± 3.010.6 ± 2.8
 Polyunsaturated fat4.43 ± 1.384.50 ± 1.394.40 ± 1.284.48 ± 1.474.20 ± 1.283.94 ± 1.28
 Protein18.7 ± 4.018.9 ± 4.218.9 ± 3.619.2 ± 3.818.7 ± 3.517.2 ± 4.0
 Carbohydrate53.9 ± 9.552.9 ± 10.153.4 ± 10.250.5 ± 8.947.3 ± 8.840.8 ± 9.6
Exercise (min/d)21.4 ± 24.325.4 ± 38.626.9 ± 34.024.9 ± 26.127.3 ± 27.420.3 ± 21.5
LDL-cholesterol (mM)3.20 ± 0.973.20 ± 0.923.08 ± 0.973.14 ± 0.933.24 ± 0.953.14 ± 0.92
HDL-cholesterol (mM)1.39 ± 0.371.41 ± 0.381.46 ± 0.391.55 ± 0.411.62 ± 0.471.78 ± 0.54
Triglycerides (mM)1.72 ± 1.121.61 ± 0.961.81 ± 2.251.33 ± 0.811.48 ± 1.331.68 ± 1.52
Glucose (mM)5.47 ± 1.725.42 ± 1.595.13 ± 0.935.15 ± 0.815.51 ± 2.145.38 ± 1.18
Insulin (pM)85.0 ± 79.682.8 ± 110.477.6 ± 80.859.5 ± 37.066.9 ± 88.070.3 ± 115.1
Coronary heart disease (%)6.45.34.62.83.88.8
Diabetes mellitus (%)15.811.96.99.110.16.3
Current smoking (%)3.014.09.218.130.840.0
HS graduate or less (%)40.942.427.632.035.936.3
Risk group (% random)47.855.450.647.538.442.5
Use of multivitamins (%)45.648.862.150.051.051.9

MS prevalence in our population was 28.5% in the random sample and 33.0% in the high-risk sample. Because we observed an inverse association between alcohol and MS in both the random and high-risk groups (data not shown), we combined both groups for subsequent analyses. In a multivariate regression accounting for familial correlation, age, age squared, education, risk group (random group vs. high-risk group), smoking, physical activity, history of CHD, energy intake, percentage of energy intake from fat, intake of fruits and vegetables, dietary cholesterol, dietary fiber, and use of multivitamins, alcohol consumption was inversely associated with the prevalence odds of MS in both men and women (p for linear trend 0.018 for men and <0.0001 for women, Table 3). When both sexes were combined, similar findings were observed. Multivariate adjusted prevalence ORs [95% confidence interval (CI)] were 1.0 (reference), 0.94 (0.79 to 1.13), 0.70 (0.47 to 1.06), 0.57 (0.45 to 0.72), 0.56 (0.42 to 0.75), and 0.65 (0.481 to 0.87) for never-drinkers and former drinkers, and current drinkers of 0.1 to 2.5, 2.6 to 12.0, 12.1 to 24.0, and >24.0 g/d, respectively (p for linear trend <0.0001). Exclusion of subjects who reported alcohol consumption of five or more alcoholic drinks almost every day did not substantially change these results (data not shown). In addition, when current drinkers were classified as those consuming alcohol with or without meals, the inverse association between alcohol and MS was observed in both groups (data not shown). There was no statistically significant interaction between education or income and alcohol (data not presented).

Table 3.  Prevalence odds ratio (95% CI) of multiple MS according to alcohol consumption among participants in the NHLBI Family Heart Study
Alcohol categoriesnCasesCrude modelModel 1*Model 2
  • *

    Model 1 adjusted for age, age squared, education, current smoking, physical activity, and CHD using generalized estimating equations.

  • Model 2 adjusted for variables in model 1 plus energy intake, risk group (random versus high risk for CHD), dietary cholesterol, use of multivitamins, dietary fiber, and percent energy from polyunsaturated fatty acids, monounsaturated fatty acids, and saturated fatty acids using generalized estimating equations.

Men     
 Never drinkers3591241.01.01.0
 Former drinkers8223111.13 (0.87 to 1.48)1.14 (0.87 to 1.51)1.12 (0.85 to 1.49)
 Current drinkers of     
  0.1 to 2.5 g/d58160.69 (0.37 to 1.27)0.70 (0.37 to 1.32)0.68 (0.36 to 1.28)
  2.6 to 12.0 g/d313750.62 (0.44 to 0.86)0.71 (0.50 to 1.01)0.72 (0.50 to 1.03)
  12.1 to 24.0 g/d238520.49 (0.33 to 0.72)0.66 (0.44 to 0.99)0.66 (0.44 to 0.99)
  >24.0 g/d287790.70 (0.50 to 0.99)0.82 (0.57 to 1.17)0.80 (0.55 to 1.16)
 p for linear trend  0.00010.0140.018
Women     
 Never drinkers6592491.01.01.0
 Former drinkers10843460.73 (0.60 to 0.90)0.86 (0.69 to 1.08)0.86 (0.69 to 1.09)
 Current drinkers of     
  0.1 to 2.5 g/d87230.59 (0.36 to 0.95)0.76 (0.46 to 1.28)0.80 (0.43 to 1.34)
  2.6 to 12.0 g/d364660.36 (0.26 to 0.50)0.47 (0.34 to 0.66)0.47 (0.33 to 0.66)
  12.1 to 24.0 g/d159350.43 (0.28 to 0.66)0.49 (0.31 to 0.77)0.47 (0.30 to 0.74)
  >24.0 g/d80170.41 (0.23 to 0.75)0.39 (0.21 to 0.73)0.39 (0.21 to 0.74)
 p for linear trend  <0.0001<0.0001<0.0001

When cases of MS were classified according to the number of individual components, we observed an inverse association between alcohol consumption and MS when three, four, and five components were present. Of the 1393 cases of MS, 842 had three components, 382 had four components, and 169 had five components. From the lowest to the highest category of alcohol intake, multivariate adjusted ORs (95% CI) for subjects with three components were 1.0 (reference), 0.92 (0.74 to 1.13), 0.78 (0.48 to 1.24), 0.59 (0.44 to 0.77), 0.53 (0.37 to 0.74), and 0.70 (0.50 to 0.97) (p for linear trend 0.0031); for subjects with four components, ORs were 1.0, 1.02 (0.78 to 1.35), 0.66 (0.33 to 1.33), 0.49 (0.33 to 0.74), 0.65 (0.41 to 1.04), and 0.59 (0.36 to 0.95) (p for linear trend 0.0032); and for subjects with five components, they were 1.0, 0.84 (0.58 to 1.22), 0.41 (0.12 to 1.37), 0.36 (0.19 to 0.69), 0.44 (0.22 to 0.91), and 0.29 (0.12 to 0.68) (p for linear trend 0.0022).

As expected, alcohol consumption was associated with HDL-cholesterol in a linear fashion in both men and women (Table 4). In addition, alcohol consumption was associated with prevalent hypertension in a dose-response manner and with obesity, diabetes mellitus, and elevated TGs in a nonlinear manner among women (Table 4). In men, alcohol consumption was positively associated with prevalent hypertension, and a suggestive U-shaped curve was observed with obesity, diabetes mellitus, and high TGs (Table 4).

Table 4.  Prevalence OR (95% CI) of multiple MS according to alcohol consumption and individual components of the MS among participants in the NHLBI Family Heart Study*
 Components of MS
Alcohol groupHypertensionObesityDiabetes mellitusHigh TGLow HDL
  • *

    Adjusted for age, age squared, sex, center, risk group (random versus high risk for CHD), current smoking, education, physical activity, and fruit and vegetable intake.

Men     
 Never drinkers1.01.01.01.01.0
 Former drinkers1.15 (0.83 to 1.60)1.19 (0.90 to 1.59)1.23 (0.84 to 1.79)1.14 (0.86 to 1.51)0.92 (0.70 to 1.20)
 Current drinkers of     
  0.1 to 2.5 g/d0.81 (0.45 to 1.46)0.87 (0.47 to 1.59)0.89 (0.38 to 2.05)1.33 (0.74 to 2.40)0.56 (0.31 to 1.01)
  2.6 to 12.0 g/d0.80 (0.53 to 1.21)0.93 (0.64 to 1.35)0.72 (0.43 to 1.19)0.83 (0.57 to 1.19)0.59 (0.41 to 0.84)
  12.1 to 24.0 g/d1.17 (0.75 to 1.81)0.80 (0.54 to 1.19)1.36 (0.80 to 2.32)0.86 (0.58 to 1.28)0.25 (0.16 to 0.38)
  >24.0 g/d1.85 (1.21 to 2.84)0.97 (0.66 to 1.43)1.29 (0.78 to 2.11)1.18 (0.81 to 1.71)0.28 (0.19 to 0.41)
 p for linear trend0.00070.160.410.87<0.0001
Women     
 Never drinkers1.01.01.01.01.0
 Former drinkers0.88 (0.67 to 1.15)1.09 (0.85 to 1.41)0.82 (0.58 to 1.14)0.84 (0.66 to 1.07)0.90 (0.72 to 1.13)
 Current drinkers of     
  0.1 to 2.5 g/d0.77 (0.42 to 1.39)0.88 (0.52 to 1.48)0.52 (0.22 to 1.21)0.85 (0.50 to 1.46)0.86 (0.53 to 1.40)
  2.6 to 12.0 g/d0.74 (0.52 to 1.08)0.62 (0.44 to 0.87)0.73 (0.45 to 1.17)0.57 (0.41 to 0.79)0.51 (0.37 to 0.70)
  12.1 to 24.0 g/d1.21 (0.78 to 1.86)0.67 (0.45 to 0.99)0.68 (0.37 to 1.27)0.49 (0.32 to 0.74)0.42 (0.27 to 0.64)
  >24.0 g/d1.41 (0.80 to 2.48)1.01 (0.60 to 1.74)0.33 (0.11 to 0.96)0.61 (0.35 to 1.06)0.24 (0.12 to 0.47)
 p for linear trend0.050.040.040.007<0.0001

We evaluated the sensitivity of the results with exclusion of subjects with a history of CHD and subjects reporting current treatment for diabetes mellitus or hypertension and found similar findings. From the lowest to the highest category of alcohol intake, adjusted ORs with these exclusions were 1. 0 (reference), 1.16 (0.88 to 1.53), 0.72 (0.39 to 1.34), 0.73 (0.51 to 1.04), 0.67 (0.44 to 1.00), and 0.80 (0.55 to 1.16), respectively, for men (p for linear trend 0.012) and 1.0, 0.86 (0.69 to 1.08), 0.79 (0.47 to 1.33), 0.46 (0.33 to 0.65), 0.46 (0.30 to 0.73), and 0.41 (0.22 to 0.77), respectively, for women (p for trend <0.0001).

The frequency of nondrinkers, drinkers of wine only, beer only, spirits only, and more than one type of beverage was 28.6%, 6.3%, 23.1%, 9.0%, and 33.0%, respectively, among men and 48.9%, 17.1%, 8.3%, 9.7%, and 16.1%, respectively, among women. A lower prevalence odds of MS was observed across all beverage types in a dose-response fashion (Table 5). Compared with never-drinkers, the prevalence odds of MS were 68%, 58%, 43%, and 44% lower among subjects who consumed >7 drinks/wk of wine only, beer only, spirits only, and at least two beverage types, respectively, in a multivariate model that also controlled for total alcohol (Table 5). As expected, subjects who consumed only wine had different characteristics than subjects in other categories (Table 6). For example, the average alcohol consumption was about one-half a drink per day for wine drinkers compared with approximately one drink per day for subjects who consumed beer or spirits only. In addition, compared with drinkers of other beverage types, wine consumers had lower energy intake and lower prevalence of current smoking (Table 6).

Table 5.  Prevalence OR (95% CI) of multiple MS according to type of alcoholic beverage among participants in the NHLBI Family Heart Study*
Beverage typesnCasesCrude modelModel 1Model 2Model 3§
  • *

    Former drinkers were excluded from these analyses.

  • Model 1 adjusted for age, age squared, sex, risk group (random versus high risk for CHD), current smoking, education, income, physical activity, other types of beverage, energy intake, and history of diabetes mellitus and coronary heart disease using generalized estimating equations.

  • Model 2 adjusted for variables in model 1 plus fruits and vegetables, dietary cholesterol, dietary fiber, use of multivitamins, and percentage of energy from polyunsaturated fatty acids, monounsaturated fatty acids, and saturated fatty acids using generalized estimating equations.

  • §

    Model 3 adjusted for total alcohol consumption and all variables in model 2.

None10183731.01.01.01.0
Wine only      
 1 to 7 drinks/wk263500.40 (0.29 to 0.55)0.44 (0.32 to 0.62)0.44 (0.31 to 0.62)0.43 (0.32 to 0.65)
 >7 drinks/wk4680.33 (0.15 to 0.75)0.35 (0.15 to 0.79)0.35 (0.16 to 0.79)0.32 (0.14 to 0.73)
Beer only      
 1 to 7 drinks/wk245450.38 (0.27 to 0.55)0.58 (0.39 to 0.87)0.60 (0.40 to 0.89)0.58 (0.39 to 0.88)
 >7 drinks/wk157270.33 (0.21 to 0.52)0.46 (0.29 to 0.75)0.48 (0.29 to 0.79)0.42 (0.23 to 0.77)
Spirits only      
 1 to 7 drinks/wk164480.66 (0.45 to 0.97)0.70 (0.48 to 1.04)0.72 (0.48 to 1.07)0.69 (0.46 to 1.04)
 >7 drinks/wk80260.76 (0.45 to 1.27)0.69 (0.40 to 1.19)0.65 (0.37 to 1.14)0.57 (0.30 to 1.09)
Any combination      
 1 to 7 drinks/wk246640.57 (0.42 to 0.79)0.68 (0.48 to 0.96)0.68 (0.48 to 0.96)0.65 (0.45 to 0.94)
 >7 drinks/wk385950.54 (0.41 to 0.71)0.64 (0.47 to 0.87)0.64 (0.47 to 0.89)0.56 (0.36 to 0.88)
Table 6.  Characteristics of participants according to type of alcohol in the NHLBI Family Heart Study*
 Type of alcoholic beverage consumed
 Never drinkers (n = 1018)Wine only (n = 309)Beer only (n = 402)Spirits only (n = 244)Mixed type (n = 631)
  • *

    F&V, fruits and vegetables; HS, high school education; continuous variables are shown as mean ± SD.

Age (years)55.1 ± 13.452.7 ± 12.944.3 ± 13.755.5 ± 13.749.9 ± 13.5
Alcohol intake (g/d)07.8 ± 8.915.6 ± 16.415.1 ± 13.522.5 ± 16.9
BMI (kg/m2)28.0 ± 5.725.3 ± 4.726.3 ± 4.227.1 ± 4.927.1 ± 4.6
F&V (servings/d)3.75 ± 1.983.56 ± 1.572.85 ± 1.623.11 ± 1.653.15 ± 1.63
Dietary fiber (g/d)19.2 ± 8.917.9 ± 8.316.5 ± 8.416.2 ± 8.116.9 ± 8.2
Dietary cholesterol (g/d)0.23 ± 0.130.20 ± 0.090.26 ± 0.130.25 ± 0.110.25 ± 0.13
Energy intake (kJ)7171 ± 24416540 ± 22728220 ± 28267307 ± 25577837 ± 2678
Percent energy from     
 Saturated fat11.2 ± 3.210.3 ± 3.211.1 ± 3.111.5 ± 3.110.7 ± 2.7
 Monounsaturated fat11.8 ± 3.210.8 ± 3.311.8 ± 3.112.3 ± 3.111.5 ± 2.9
 Polyunsaturated fat4.52 ± 1.374.31 ± 1.374.30 ± 1.334.51 ± 1.354.36 ± 1.37
 Protein18.6 ± 3.918.8 ± 3.917.2 ± 3.818.3 ± 3.617.9 ± 4.0
 Carbohydrate53.4 ± 9.551.7 ± 10.548.6 ± 9.446.7 ± 9.946.1 ± 9.1
Exercise (min/d)24.3 ± 28.632.1 ± 37.634.6 ± 37.631.2 ± 40.935.0 ± 37.4
LDL-cholesterol (mM)3.20 ± 0.923.19 ± 0.973.21 ± 0.913.31 ± 0.903.26 ± 0.88
HDL-cholesterol (mM)1.27 ± 0.371.51 ± 0.451.30 ± 0.391.36 ± 0.411.36 ± 0.41
Triglycerides (mM)1.79 ± 1.231.47 ± 1.351.59 ± 1.271.78 ± 1.501.69 ± 1.39
Glucose (mM)5.54 ± 1.745.27 ± 1.295.39 ± 1.195.61 ± 1.705.57 ± 1.59
Insulin (pM)88.8 ± 111.665.3 ± 56.565.9 ± 40.797.0 ± 228.475.5 ± 79.9
Coronary heart disease (%)9.59.47.212.711.9
Diabetes mellitus (%)15.58.110.716.014.4
Current smoking (%)3.09.728.628.321.0
HS graduate or less (%)33.823.335.738.930.1
Risk group (% random)54.852.841.347.545.7
Use of multivitamin (%)42.053.330.047.743.5

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

In the present study, we demonstrated that moderate alcohol consumption was associated with lower prevalence odds of MS among men and women, with a U-shaped relation in men and a dose-response relation in women. The lowest prevalence odds were observed among men who consumed ∼1 to 2 drinks/d and among women whose alcohol intake was >2 drinks/d. We found evidence that the alcohol-MS association was strongest among cases who had all five components. As expected, alcohol consumption was associated with prevalent hypertension and higher HDL-cholesterol in both men and women. A U-shaped relation was observed among alcohol and TGs and diabetes mellitus, mainly among women.

Although the association between alcohol consumption and individual components of the MS is well documented (7, 8, 23, 24, 25), limited data are available on the alcohol-MS relation. In a cross-sectional study of 793 men and 315 women, the frequency of alcohol consumption was significantly lower among subjects with MS than among those without MS (33.2% and 42.1%) (26). Unfortunately, that study did not assess dose-response relation of alcohol to MS; in addition, only crude results were presented, and sex-specific data were not shown. In another cross-sectional study of 391 men, Goude and colleagues (8) reported that men with MS consumed, on average, 15.5 g/d of alcohol, whereas the average alcohol intake in men without any risk factor for MS was 11.4 g/d (p = 0.4). Among 6805 women in Sweden, in comparison with nondrinkers, alcohol consumption of <12 and 12 to 23 g/d was associated with lower prevalence odds of one or more features of the MS [OR (95% CI): 0.71 (0.63 to 0.81) and 0.78 (0.65 to 0.93), respectively, adjusting for age, bone density, perimenopausal status, family history of hypertension, and exercise] (24). In that study, the OR for women consuming at least two drinks per day was not presented. Data from the Korean National Health and Nutrition Examination Survey (27) are consistent with our findings on total alcohol and MS in that Korean men and women consuming 1 to 14.9 g/d of alcohol had ∼20% lower prevalence of MS than nondrinkers. Unfortunately, the Korean study (27) did not evaluate the effects of beverage types on MS.

In our study, a lower prevalence of MS was observed with wine, beer, and spirits consumption in a dose-response manner. In a multivariate model that also controlled for total alcohol intake, consumption of >7 drinks/wk of wine, beer, spirits, and any combination was associated with 68%, 58%, 43%, and 44% lower prevalence odds of MS, respectively. The slightly greater risk reductions associated with beer and wine consumption may suggest that other components in wine and beer (other than ethanol) and/or behavior associated with wine and beer consumption may also reduce the risk of MS. Rosell et al. (28) found that wine consumption was associated with lower prevalence odds of MS compared with subjects consuming 1 to 10 g/d of any type of alcohol [OR (95% CI) = 0.60 (0.40 to 0.91)] in women but did not find a significant difference in prevalence odds of MS across beverage types in men. In the same study (28), spirits consumption of 11 to 30 g/d was suggestive of increased prevalence odds of MS, with ORs of 1.7 (0.8 to 3.3) for men and 2.0 (0.8 to 4.9) for women. These results were adjusted for smoking, education, immigration, employment, and intake of vegetables. Dorn et al. (29) reported that wine, but not beer and spirits, consumption was associated with lower abdominal height (a measure of central obesity); beer had no effect, but spirits consumption was associated with an increased abdominal height among women.

The divergence of these findings with our results merits some comments. In the paper of Rosell et al. (28), the “low alcohol group” referred to the daily consumption of 1 to 10 g/d of any type of alcohol, and moderate consumption referred to the intake of 11 to 30 g/d of any type of alcohol. Subjects in the moderate category were subdivided into wine, beer, spirits, and mixed group. We cannot compare these ORs with our results because of several reasons: 1) Although in our analyses, beverage-specific categories were limited to subjects consuming only one type of beverage (except for mixed group), Rosell et al. (28) used a different method for beverage-specific classification. For example, subjects with >10 grams of alcohol from wine and <10 grams from beer and <10 grams from spirits were classified as wine drinkers, and the reference group might include wine drinkers only and beer drinkers only whose total alcohol intake was up to 10 g/d. We do not believe this classification is optimal in assessing the effects of wine because subjects consuming small amounts of beer and spirits would be included in the wine group. Because that study found increased prevalence odds of MS among spirits drinkers, the true effects of wine might have been attenuated. 2) A single estimate of OR among consumers of 11 to 30 g/d of alcohol is not available for comparison. 3) Our study did not have enough subjects consuming >30 g/d for contrast. In the paper by Dorn et al., beverage-specific analyses assessed ethanol effects from wine, beer, or spirits using nondrinkers of each type of beverage as reference. Thus, the goal in that paper was not to compare the effects of nonethanol components contained in each type of beverage. As shown in Table 6, and demonstrated by others (30, 31), wine consumption is associated with a healthier lifestyle than the consumption of beer and spirits. For example, wine consumers have lower BMI, a better lipid profile, and higher educational attainment, eat less saturated fat, and are less likely to smoke cigarettes than consumers of other types of beverage. For these reasons alone, residual confounding is more likely to explain effect measures that use nonwine drinkers as reference and, thus, may exaggerate the true association. We recognize that the issue of comparing beverage-specific effects on outcome is complex in observational studies because beverage preference is associated with other habits that, if uncontrolled, can bias the study results.

Our study has some limitations. Causal inferences from our data are not possible due to the cross-sectional design. Because alcohol assessment was through self-report, it is possible that underestimation of alcohol consumption might have introduced bias into our findings. Our data showed a slightly lower prevalence of MS among light drinkers (0.1 to 2.5 g/d), suggesting that, perhaps, subjects in these categories had underreported their alcohol intake. It is also possible that residual confounding by other factors related to light consumption of alcohol might explain the effects seen among very light drinkers. However, underreporting of alcohol is more likely to occur among heavy drinkers, and if this were the case, then the observed inverse association between alcohol and MS among women might have been attenuated by underreporting. However, the fact that 1) very few subjects consumed >2 drinks/d; 2) we observed a strong correlation between reported alcohol and serum γ-glutamyltranspeptidase; and 3) alcohol was positively associated with hypertension and HDL-cholesterol, as expected, suggest that our estimation of alcohol was reasonably valid. Some subjects may have changed their drinking habits after developing one or more features of MS. Nevertheless, we did have information to identify former drinkers and separate them from life-long abstainers. We did not have detailed data on drinking patterns to assess the influence of alcohol consumption patterns on our findings. Also, our data were limited to white participants because we did not have adequate numbers of subjects in other ethnic groups for stable estimates. Contrary to most previous studies, we had a large sample size with data on both men and women and detailed information on lifestyle, diet, and metabolic factors to minimize residual confounding.

Mechanisms by which alcohol consumption might favorably influence the risk of MS have been proposed. Moderate alcohol has been found to improve insulin sensitivity (7, 8, 23, 32), raise HDL-cholesterol (10, 11), and lower TGs (12, 33). The observed associations between alcohol and HDL-cholesterol, TGs, serum insulin, and serum glucose are consistent with those of previous reports. Also consistent with previous studies (34, 35) was our finding that alcohol consumption was associated with a higher prevalence of hypertension.

In conclusion, alcohol consumption is associated with lower prevalence odds of MS in men and women, and this association is observed among subjects consuming all three types of alcoholic beverage. Prospective studies are needed to confirm these findings and to assess the influence of drinking patterns on the alcohol-MS association.

Acknowledgment

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

This study was partially supported by Grant 05 K01 HL70444-02 and Contracts NO1-HC-25104, NO1-HC-25105, NO1-HC-25106, NO1-HC-25107, NO1-HC-25108, and NO1-HC-25109 from NHLBI (Bethesda, MD). This paper is presented on behalf of the Investigators of the NHLBI Family Heart Study. Participating Institutions and Principal Staff of the study are as follows: Forsyth County/University of North Carolina, NC, Wake Forest University, Winston-Salem, NC, Gerardo Heiss, Stephen Rich, Greg Evans, James Pankow, H.A. Tyroler, Jeannette T. Bensen, Catherine Paton, Delilah Posey, and Amy Haire; University of Minnesota Field Center, St. Paul, MN, Donna K. Arnett, Aaron R. Folsom, Larry Atwood, James Peacock, and Greg Feitl; Boston University/Framingham Field Center, MA, R. Curtis Ellison, Richard H. Myers, Yuqing Zhang, Andrew G. Bostom, Luc Djoussé, Jemma B. Wilk, and Greta Lee Splansky; University of Utah Field Center, Salt Lake City, UT, Steven C. Hunt, Roger R. Williams (deceased), Paul N. Hopkins, Jan Skuppin, and Hilary Coon; Coordinating Center, Washington University in St. Louis, St. Louis, MO, Michael A. Province, D.C. Rao, Ingrid B. Borecki, Yuling Hong, Mary Feitosa, Jeanne Cashman, and Avril Adelman; Central Biochemistry Laboratory, University of Minnesota, St. Paul, MN, John H. Eckfeldt, Greg Rynders, Catherine Leiendecker-Foster, and Michael Y. Tsai; Central Molecular Laboratory, University of Utah, Salt Lake City, UT, Mark F. Leppert, Jean-Marc Lalouel, Tena Varvil, and Lisa Baird; and NHLBI Project Office, Bethesda, MD, Phylliss Sholinsky, Millicent Higgins (retired), Jacob Keller (retired), Sarah Knox, and Lorraine Silsbee.

Footnotes
  • 1

    Nonstandard abbreviations: MS, metabolic syndrome; HDL, high-density lipoprotein; TG, triglyceride; NHLBI, National Heart, Lung, and Blood Institute; CHD, coronary heart disease; LDL, low-density lipoprotein; OR, odds ratio; CI, confidence interval.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References
  • 1
    Park, Y. W., Zhu, S., Palaniappan, L., Heshka, S., Carnethon, M. R., Heymsfield, S. B. (2003) The metabolic syndrome: prevalence and associated risk factor findings in the US population from the Third National Health and Nutrition Examination Survey, 1988-1994. Arch Intern Med. 163: 427436.
  • 2
    Ford, E. S., Giles, W. H., Dietz, W. H. (2002) Prevalence of the metabolic syndrome among US adults: findings from the third National Health and Nutrition Examination Survey. JAMA. 287: 356359.
  • 3
    Nakanishi, N., Makino, K., Nishina, K., Suzuki, K., Tatara, K. (2002) Relationship of light to moderate alcohol consumption and risk of hypertension in Japanese male office workers. Alcohol Clin Exp Res. 26: 988994.
  • 4
    Wannamethee, S. G., Camargo, C. A., Jr, Manson, J. E., Willett, W. C., Rimm E. B. (2003) Alcohol drinking patterns and risk of type 2 diabetes mellitus among younger women. Arch Intern Med. 163: 13291336.
  • 5
    Conigrave, K. M., Hu, B. F., Camargo, C. A., Jr, Stampfer, M. J., Willett, W. C., Rimm E. B. (2001) A prospective study of drinking patterns in relation to risk of type 2 diabetes among men. Diabetes. 50: 23902395.
  • 6
    Ajani, U. A., Hennekens, C. H., Spelsberg, A., Manson, J. E. (2000) Alcohol consumption and risk of type 2 diabetes mellitus among US male physicians. Arch Intern Med. 160: 10251030.
  • 7
    Bell, R. A., Mayer-Davis, E. J., Martin, M. A., D'Agostino, R. B., Jr, Haffner S. M. (2000) Associations between alcohol consumption and insulin sensitivity and cardiovascular disease risk factors: the Insulin Resistance and Atherosclerosis Study. Diabetes Care. 23: 16301636.
  • 8
    Goude, D., Fagerberg, B., Hulthe, J. (2002) Alcohol consumption, the metabolic syndrome and insulin resistance in 58-year-old clinically healthy men (AIR study). Clin Sci. 102: 345352.
  • 9
    Hulley, S. B., Gordon, S. (1981) Alcohol and high-density lipoprotein cholesterol: causal inference from diverse study designs. Circulation. 64: 5763.
  • 10
    van Tol, A., van der Gaag, M. S., Scheek, L. M., Van Gent, T., Hendriks, H. F. (1998) Changes in postprandial lipoproteins of low and high density caused by moderate alcohol consumption with dinner. Atherosclerosis. 141(Suppl 1): S101S103.
  • 11
    Hendriks, H. F., Veenstra, J., van Tol, A., Groener, J. E., Schaafsma, G. (1998) Moderate doses of alcoholic beverages with dinner and postprandial high density lipoprotein composition. Alcohol Alcohol. 33: 403410.
  • 12
    Kato, I., Kiyohara, Y., Kubo, M., et al (2003) Insulin-mediated effects of alcohol intake on serum lipid levels in a general population: the Hisayama Study. J Clin Epidemiol. 56: 196204.
  • 13
    Hunt, S. C., Williams, R. R., Barlow, G. K. (1986) A comparison of positive family history definitions for defining risk of future disease. J Chron Dis. 39: 809821. [Published erratum inJ Chron Dis. 1987;40:369.].
  • 14
    Higgins, M., Province, M., Heiss, G., et al (1996) NHLBI Family Heart Study: objectives and design. Am J Epidemiol. 143: 12191228.
  • 15
    Friedewald, W. T., Levy, R. I., Fredrickson, D. S. (1972) Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem. 18: 499502.
  • 16
    Ellison, R. C., Myers, R. H., Zhang, Y., et al (1999) Effects of similarities in lifestyle habits on familial aggregation of high density lipoprotein and low density lipoprotein cholesterol: the NHLBI Family Heart Study. Am J Epidemiol. 150: 910918.
  • 17
    Djousse, L., Pankow, J. S., Arnett, D. K., et al (2000) Alcohol consumption and plasminogen activator inhibitor type 1: the National Heart, Lung, and Blood Institute Family Heart Study. Am Heart J. 139: 704709.
  • 18
    Ellison, R. C., Zhang, Y., Qureshi, M. M., Knox, S., Arnett, D. K., Province, M. A. (2004) Lifestyle determinants of high-density lipoprotein cholesterol: the National Heart, Lung, and Blood Institute Family Heart Study. Am Heart J. 147: 529535.
  • 19
    Expert Panel on Detection EaToHBCiA. (2001) Executive Summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA. 285: 24862497.
  • 20
    Willett, W. C., Sampson, L., Stampfer, M. J., et al (1985) Reproducibility and validity of a semiquantitative food frequency questionnaire. Am J Epidemiol. 122: 5165.
  • 21
    U.S. Department of Agriculture. (1989) Composition of foods: raw, processed, and prepared, 1963-1988. Agriculture Handbook No. 8. U.S. Government Printing Office: Washington, DC.
  • 22
    Djousse, L., Pankow, J. S., Eckfeldt, J. H., et al (2001) Relation between dietary linolenic acid and coronary artery disease in the National Heart, Lung, and Blood Institute Family Heart Study. Am J Clin Nutr. 74: 612619.
  • 23
    Kiechl, S., Willeit, J., Poewe, W., et al (1996) Insulin sensitivity and regular alcohol consumption: large, prospective, cross sectional population study (Bruneck study). BMJ. 313: 10401044.
  • 24
    Lidfeldt, J., Nyberg, P., Nerbrand, C., Samsioe, G., Schersten, B., Agardh, C. D. (2003) Socio-demographic and psychosocial factors are associated with features of the metabolic syndrome: the Women's Health in the Lund Area (WHILA) study. Diabetes Obes Metab. 5: 106112.
  • 25
    Dixon, J. B., Dixon, M. E., O'Brien, P. E. (2002) Alcohol consumption in the severely obese: relationship with the metabolic syndrome. Obes Res. 10: 245252.
  • 26
    Solymoss, B. C., Bourassa, M. G., Lesperance, J., et al (2003) Incidence and clinical characteristics of the metabolic syndrome in patients with coronary artery disease. Coron Artery Dis. 14: 207212.
  • 27
    Yoon, Y. S., Oh, S. W., Baik, H. W., Park, H. S., Kim, W. Y. (2004) Alcohol consumption and the metabolic syndrome in Korean adults: the 1998 Korean National Health and Nutrition Examination Survey. Am J Clin Nutr. 80: 217224.
  • 28
    Rosell, M., de Faire, U., Hellenius, M. L. (2003) Low prevalence of the metabolic syndrome in wine drinkers: is it the alcohol beverage or the lifestyle? Eur J Clin Nutr. 57: 227234.
  • 29
    Dorn, J. M., Hovey, K., Muti, P., et al (2003) Alcohol drinking patterns differentially affect central adiposity as measured by abdominal height in women and men. J Nutr. 133: 26552662.
  • 30
    Klatsky, A. L., Armstrong, M. A., Kipp, H. (1990) Correlates of alcoholic beverage preference: traits of persons who choose wine, liquor or beer. Br J Addict. 85: 12791289.
  • 31
    Barefoot, J. C., Gronbaek, M., Feaganes, J. R., McPherson, R. S., Williams, R. B., Siegler, I. C. (2002) Alcoholic beverage preference, diet, and health habits in the UNC Alumni Heart Study. Am J Clin Nutr. 76: 466472.
  • 32
    Davies, M. J., Baer, D. J., Judd, J. T., Brown, E. D., Campbell, W. S., Taylor, P. R. (2002) Effects of moderate alcohol intake on fasting insulin and glucose concentrations and insulin sensitivity in postmenopausal women: a randomized controlled trial. JAMA. 287: 25592562.
  • 33
    Mukamal, K. J. (2003) Alcohol use and prognosis in patients with coronary heart disease. Prev Cardiol. 6: 9398.
  • 34
    Xin, X., He, J., Frontini, M. G., Ogden, L. G., Motsamai, O. I., Whelton, P. K. (2001) Effects of alcohol reduction on blood pressure: a meta-analysis of randomized controlled trials. Hypertension. 38: 11121117.
  • 35
    Ohmori, S., Kiyohara, Y., Kato, I., et al (2002) Alcohol intake and future incidence of hypertension in a general Japanese population: the Hisayama study. Alcohol Clin Exp Res. 26: 10101016.