ORIGINAL COMMUNICATION Low prevalence of the metabolic syndrome in wine drinkers is it the alcohol beverage or the lifestyle?

Transcription

1 ORIGINAL COMMUNICATION Low prevalence of the metabolic syndrome in wine drinkers is it the alcohol beverage or the lifestyle? M Rosell 1, U de Faire 1,2 and M-L Hellénius 2 (2003) 57, ß 2003 Nature Publishing Group All rights reserved /03 $ Division of Cardiovascular Epidemiology, Institute of Environmental Medicine, Karolinska Institute, Stockholm, Sweden; and 2 Cardiovascular Laboratory, Department of Cardiology, Karolinska Hospital, Stockholm, Sweden Objective: To study how the intake of alcohol and the choice of wine, beer, and spirits is related to lifestyle factors and the metabolic syndrome in 60-y-old men and women. Design: Cross-sectional population based study. Setting: Stockholm County, Sweden. Subjects: Sixty-year-old men and women (n ¼ 4232). Results: intake of wine (10 30 g=day) was associated with a lifestyle characterized by being married, having a university education, being employed, being Swedish-born, having a good quality of life according to economy, leisure time and health, compared with a group with low alcohol intake. The opposite characteristics were seen among the non-drinkers. Drinkers of spirits were more often smokers and also reported higher intake of sausage and fried potatoes compared with a group with low alcohol intake. In women, the metabolic syndrome was significantly more common in non-drinkers (20%), P < 0.05, and less common among wine drinkers (8%), P < 0.01, compared with a group with low alcohol intake. After adjustments, a significant lower odds ratio for the metabolic syndrome were seen in wine drinkers in women (OR ¼ 0.60, P < 0.05). Conclusions: Compared with low alcohol drinkers, moderate wine drinkers exhibited a more favorable pattern according to both lifestyle factors and metabolic parameters. The close link between alcohol drinking behaviour and lifestyle habits illustrate the complex relationship between alcohol and health. Sponsorship: Swedish Heart-Lung Foundation, the Stockholm City County Council, the Swedish Society for Medicine, and the Swedish Medical Research Council. (2003) 57, doi: =sj.ejcn Keywords: alcohol intake; life style; the metabolic syndrome Introduction Epidemiological studies have shown an association between light to moderate alcohol consumption and reduced risk of cardiovascular disease (Fagrell et al, 1999; Rimm et al, 1996). Although some studies have indicated that wine is more protective (Gronbaek et al, 1995; Klatsky et al, 1997; *Correspondence: M Rosell, Division of Cardiovascular Epidemiology, Institute of Environmental Medicine, Karolinska Institutet, Box 210, S Stockholm, Sweden. magdalena.rosell@imm.ki.se Guarantors: M Rosell and M-L Hellénius. Contributors: MR contributed to the data analyses, the interpretation and the writing of the paper. MLH and UdF designed the study and contributed to the interpretation of the analyses. Received 8 November 2001; revised 29 May 2002; accepted 4 June 2002 Wannamethee & Shapes, 1999), the influence of different alcoholic beverages is unclear (Fagrell et al, 1999; Rimm et al, 1996). Various plausible mechanisms have been discussed regarding the beneficial role of moderate alcohol intake on cardiovascular disease. The most important effect of alcohol is its ability to increase HDL-cholesterol concentration, but also Lp(a) and haemostatic factors are favourably affected by alcohol (Gaziano et al, 1993; Rimm et al, 1999). Red wine may be particularly beneficial due to its high content of phenolic compounds (Serafini et al, 1998; Caccette et al, 2000). These substances have antioxidative properties and have been shown to reduce the susceptibility of low-density lipoprotein oxidation (Nigdikar et al, 1998). Favourable association of alcohol intake and cardiovascular disease, however, has not been reported in all population groups depending on several factors such as age, sex and the

2 228 prevalence of cardiovascular risk factors such as smoking and LDL-cholesterol concentration (Hein et al, 1996; Thun et al, 1997; Shaper et al, 1988). There are also reports of populations with low or no alcohol intake where the cardiovascular mortality is lower than the general population, which illustrates the significant role of other lifestyle factors (Fonnebo et al, 1992; Kjaerheim et al, 1993). For many people, alcohol intake is an important part of their social life and is often used in such contexts. Therefore, in epidemiological studies it has been difficult to estimate the effect of alcohol exposure per se due to difficulties in estimation and adjustments for such influences. Smoking, physical activity, education and economy are related to alcohol intake and it has been shown that wine drinkers consume more fruit and vegetables than beer and spirits drinkers do (Tjonneland et al, 1999). During recent years, an increasing interest has been directed towards the mechanisms behind the metabolic syndrome and its role in cardiovascular disease and type 2 diabetes. The metabolic syndrome describes a state of clustered metabolic disturbances such as abdominal obesity, insulin resistance, raised triglyceride concentration, low HDL-cholesterol concentration and elevated blood pressure (DeFronzo & Ferrannini, 1991). Alcohol affects the separate components in the metabolic syndrome in different ways with particularly favorable effects on HDL-cholesterol, but triglycerides may be negatively affected as well (Rimm et al, 1999; Temelkova-Kurktschiev et al, 2001). Inverse associations between alcohol and insulin resistance have been noted (Lazarus et al, 1997; Goude et al, 2002; Kiechl et al, 1996). The relationship, however, was attenuated when controlling for BMI and waist circumference (Bell et al, 2000). The effects of alcohol on weight are unclear ( Jéquier, 1999), but findings of positive associations between alcohol and waist:hip ratio, independent of BMI, have been noted (Dallongeville et al, 1998; Sakurai et al, 1997). To get further insight into the associations between alcohol and the metabolic syndrome, we have investigated how alcohol intake and the choice of wine, beer or spirits are related to lifestyle factors and to the metabolic syndrome. Methods A study of 60-y-old men and women in Stockholm County From August 1997 to March 1999, every third man and woman living in Stockholm County who was born between 1 July 1937 and 31 June 1938 was invited to participate in a thorough health screening study. The participants underwent a physical examination that included measuring the participant s height, weight, waist and hip. Systolic and diastolic blood pressures were measured after 5 min of rest, and the mean values of two measurements were calculated. Fasting blood samples were taken and a comprehensive questionnaire was completed. The study was approved by the ethical committee at the Karolinska Institute. Medical history History of cardiovascular disease, hospitalizations, drugs use and occurrence of cardiovascular symptoms was asked for in the questionnaire. Alcohol intake The questionnaire included five questions concerning the intake of beer (light, medium and strong), wine and spirits. For example, for the question How much strong beer do you usually drink? the following alternatives were provided: More than 1 bottle=day, 4 6 bottles=week, 2 3 bottles=week, 1 or less bottle=week, and None at all (1 bottle ¼ 0.45 l). An average daily intake of alcohol was calculated and three categories of drinking groups were defined: no alcohol intake (0 g alcohol=day), low intake (1 10 g=day), moderate intake (> g=day), and high intake (> 30 g=day). alcohol intake is equivalent to approximately two glasses of wine per day. The moderate group was divided into four groups according to the dominant type of beverage consumed: a wine group (> 10 g alcohol from wine, < 10 g from beer and < 10 g from spirits); a beer group (> 10 g alcohol from beer, < 10 g from wine and < 10 g from spirits); a spirits group (>10 g alcohol from spirits, < 10 g from wine and < 10 g from spirits); and a mixed group that comprised people whose alcoholic consumption did not fit the criteria for the other groups. The high alcohol intake group was not in focus in this study and the remaining groups will be referred to as no, low, wine, beer, spirits and mixed alcohol group. Lifestyle factors The education level was defined as (1) university, (2) 12 y school, and (3) lower or no education. Civil status was defined as married=living together (yes=no). An employment variable was defined according to working status (yes, full or part-time=no). An immigration variable was used (yes=no). The subjects were asked to rate their wellbeing on a seven-degree scale ranging from very bad to excellent, could not be better. A score of 1 5 was defined as bad and 5 7as good (Tibblin et al, 1990). The items included health, economy and leisure time. Physical activity in leisure time the past year was asked for and grouped into the following categories: (1) inactive; (2) light activity at least 2 h=week; (3) moderate activity 1 2 times=week; and (4) intensive activity 3 times=week. Categories 1 and 2 were classified as inactive and categories 3 and 4 were classified as active. Current smoking was coded as yes=no. Dietary intake was measured using a food frequency questionnaire that included 17 questions, each with four answer alternatives. The questions used here concerned the intake of fruit, vegetables, fish, fried potatoes and sausage=bacon.

3 Symptoms of illness The questionnaire included a list of symptoms of illnesses where the subjects were asked to indicate whether they had experienced any of these symptoms during the previous 3 months (yes=no; Tibblin et al, 1990). Muscle-skeletal symptom was defined when at least two of the three following items were indicated: pain in legs, pain in joints, and pain in back. Heart=lung symptom, which is defined when breathlessness and=or chest pain was indicated. Blood analyses All blood samples were analysed on-line during the study. Serum glucose was measured with an enzymatic colorimetric test (Bayer Diagnostics, Tarrytown, NY, USA). Serum insulin levels were determined using the ELISA technique (Boeringer Mannheim Gmbh, Diagnostica, Germany). Cholesterol and triglycerides in serum were analyzed using enzymatic methods (Bayer Diagnostics, Tarrytown, NY, USA; Allain et al, 1974; Fossati & Prencipe, 1982). HDL-cholesterol in serum was measured enzymatically after isolation of LDL and VLDL (Boehringer Mannheim Gmbh, Germany) and LDLcholesterol was estimated using Friedewald s method (Friedewald et al, 1972). Lp(a) in serum was analyzed with an immunoassay method (DiaSys Diagnostica Systems, Holzheim, Germany). Plasma fibrinogen was measured with a functional spectrophotometric test (Boehringer Mannheim, Germany; Hemker et al, 1979; Becker et al, 1984). Serum urate was measured using an enzymatic method (Bayer Diagnostics, Tarrytown, NY, USA; Trinder, 1969; Fossati et al, 1980). Gamma glutamyltransferase in serum was determined using an enzymatic colorimetric test (Bayer Diagnostics, Tarrytown, NY, USA). The metabolic syndrome The metabolic syndrome was classified using the criteria proposed by the European Group for the Study of Insulin Resistance (EGIR; Balkau et al, 1999). The metabolic syndrome was accordingly defined by the presence of fasting hyperinsulinemia (the highest 25%), and at least two of the following abnormalities: hyperglyceridemia (plasma glucose 6.1), hypertension (systolic=diastolic blood pressures 140=90 mmhg, or treated for hypertension), dyslipoproteinemia (triglycerides > 2mmol=l or HDL-cholesterol < 1.0 mmol=l, or treated for dyslipidemia), and central obesity (waist circumference 94 and 80 cm in men and women respectively). Statistics To compare the prevalence of lifestyle factors and the metabolic syndrome among the different alcohol groups, the Mantel Haenzel chi-square test was used. To compare the means of the biological variables among the different alcohol groups, the two-sample t-test was used. If the distribution of the biological variable was skewed in at least one group, the Wilcoxon rank-sum test was used. Odds ratios for the metabolic syndrome in the different alcohol groups were calculated with logistic regression. The Low alcohol intake group was used as reference in all analyses and a P-value of < 0.05 was considered as statistically significant. All statistical analyses were performed using SAS System for Windows 8.01 software. Results Seventy-eight percent participated in the study, 2039 men and 2193 women. The response rate among immigrants was 68% (n ¼ 477). The main reasons for not participating were the following: not interested (27%), unknown reason (27%), contact with physician (20%), and disease (11%). Sixty-three men and 59 women did not complete the questionnaire and were therefore excluded. People diagnosed with cardiovascular diseases (angina pectoris, stroke, coronary heart disease), pharmacological treated diabetes or undergoing treatment for cancer were excluded (246 men and 169 women). This left 1730 men and 1974 women for the study. Alcohol intake Approximately 50% of the men reported a moderate alcohol intake and approximately 50% of the women reported a low alcohol intake (Table 1). The men were more evenly distributed with respect to the consumption of wine, beer and mixed groups compared with the women where a majority preferred wine. In men, the medians of average alcohol intake (g=day) in the groups were as follows: 0 (no), 6.2 (low), 18.7 (wine), 19.2 (beer), 18.7 (spirits), 14.7 (mixed). In women, the medians of average alcohol intake (g=day) in the groups were as follows: 0 (no), 4.2 (low), 16.3 (wine), 16.2 (beer), 18.8 (spirits), 14.8 (mixed). Lifestyle characteristics The wine group and the mixed group were more often married=lived together, university educated and currently employed compared with the low alcohol group, as shown in Table 2. The proportion of immigrants was lowest in the wine group and highest in the non-drinking group, particularly among women. The wine and the mixed group reported Table 1 Reported alcohol intake in 60-y-old men and women in Stockholm Country, number and percentage Men (n ¼ 1730) Women (n ¼ 1974) No intake (0 g=day) 89 (5%) 189 (10%) Low intake (1 10 g=day) 383 (22%) 1059 (54%) intake (>10 30 g=day) 879 (51%) 652 (33%) Wine 238 (27%) 401 (62%) Beer 191 (22%) 37 (6%) Spirits 58 (7%) 32 (5%) Mixed 392 (45%) 182 (28%) High intake (>30g=day) 379 (22%) 74 (4%) 229

4 230 Table 2 Life style characteristics in relation to alcohol intake in 60-y-old men and women. Prevalence (%) in different alcohol groups: no, low and moderate intake, divided into a wine, beer, spirits and a mixed group Men Women Alcohol group No Low Wine Beer Spirits Mixed No Low Wine Beer Spirits Mixed Married=living together 67* 78 88** * 58** 69 75* ** Higher education *** 18** 14** 30 17* 25 36*** * Being employed 59* 71 82** *** Immigrants *** *** 20 8*** Quality of life Good economy ** ** * Good leisure time * *** 72*** Good health 62** 78 87** * 56*** 76 81* Smokers *** * 24 47*** 17 Ex-smokers *** 44* 35 47*** 20** 29 40*** * Physical active * 46** Symptoms of illness Muscle skeletal *** 37 30* Lung heart symptoms 32* * 20 43** 31 23** Chi-square tests were performed using the low alcohol group as reference. *P < 0.05; **P < 0.01; ***P < more often a good well-being with respect to economy, leisure time and health compared with the low alcohol group. The non-drinkers had lower rates of these variables. Smoking was most prevalent among the spirits drinkers in both men and women. Ex-smokers were most common in the wine and the mixed group. Forty-nine percent of the non-drinking woman reported muscle skeletal symptoms. The corresponding rate among wine drinkers was 30%. The prevalence of heart=lung symptoms was highest in the non-drinking group (men and women) and in the spirits group (men only). Dietary habits Daily intake of fruit and vegetables was less frequent in the beer and spirits group compared with the low alcohol group (Table 3). The wine group had a higher intake of fish and vegetables. In men, weekly intake of sausage= bacon and fried potatoes was more common in the spirits drinkers. Biological variables The wine group had a more favourable metabolic profile with significantly lower levels of fasting insulin, higher HDL-cholesterol, and lower P-fibrinogen concentrations compared with the low alcohol group (Table 4). The mixed group looked similar to the wine group in these respects. The spirits group showed a more unfavourable metabolic profile with higher waist-to-hip ratio and insulin levels in both men and women. The non-drinkers had a higher waist-to-hip ratio and in women there were higher insulin and triglyceride levels. Table 3 Dietary habits in relation to alcohol intake in 60-y-old men and women. Prevalence (%) of food intake in different alcohol groups: no, low and moderate intake, divided into a wine, beer, spirits and a mixed group Men Women Alcohol group No Low Wine Beer Spirits Mixed No Low Wine Beer Spirits Mixed Daily intake of fruit *** * 44*** 74 Daily intake of vegetables * *** 71 Non-oily fish 3 times=week *** ** 70 38*** 71 Sausage, bacon > 1 times=week * 48 76*** ** 59*** 28 20** Fried potatoes > 1 times=week * 47* 40* * 13 24* Chi-square tests were performed using the low alcohol group as reference. *P < 0.05; **P < 0.01; ***P <

5 Table 4 Metabolic parametern in relation to alcohol intake in 60-y-old men and women presented as means (or medians when skewed distribution) in different alcohol groups: no, low and moderate intake, divided into a wine, beer, spirits and a mixed group 231 Men Women Alcohol group No Low Wine Beer Spirits Mixed No Low Wine Beer Spirits Mixed Waist-to-hip ratio 0.96* ** *** ** 0.81 Syst blood pressure (mmhg) Diast blood pressure (mmhg) Fasting S-glucose (mmol=l) * Fasting S-insulin (mu=l) * * ** ** * 8.0 S-Cholesterol (mmol=l) ** 5.8* ** S-HDL-cholesterol (mmol=l) 1.15* *** 1.29** *** 1.46*** *** *** S-LDL-cholesterol (mmol=l) * * * S-Triglycerides (mmol=l) * ** S-Lp(a) (g=l) P-Fibrinogen (g=l) 3.1* * 2.7* *** *** *** Uric acid (mmol=l) ** 340* g-gt (mkat=l) * 0.53** 0.63*** 0.53*** ** *** 0.42** t-tests were performed using the low alcohol group as reference. If the distribution was skewed in at least one alcohol group, the Wilcoxon rank-sum test was used for that variable. *P < 0.05; **P < 0.01; ***P < The metabolic syndrome Thirteen percent of the women and 19% of the men fulfilled the criteria for the metabolic syndrome. No significant differences in the metabolic syndrome were seen in the different alcohol groups among the men. In women, the non-drinkers had higher prevalence (20%) and the wine group had lower prevalence (8%) of the metabolic syndrome compared with the low alcohol group, as illustrated in Figure 1. When adjustments were made for smoking, education, immigration, employment, physical activity and the intake of vegetables, a significant lower odds ratio was seen in the wine group in women (Table 5). Figure 1 Prevalence (%) of the metabolic syndrome in different alcohol groups in 60-y-old men and women. The alcohol groups were: none, low and moderate intake, which was divided into a wine, beer, spirits and a mixed group. Chi-square tests were performed using the low alcohol group as reference. *P < 0.05; **P < 0.01.

6 232 Table 5 Odds ratios (OR) with 95% confidence interval for the metabolic syndrome in the different alcohol groups: no, low and moderate intake, which was divided into a wine, beer, spirits and a mixed group Alcohol group No Low Wine Beer Spirits Mixed Men n Metabolic syndrome, n Crude OR 0.85 ( ) ( ) 0.87 ( ) 1.46 ( ) 0.89 ( ) Adjusted a OR 0.79 ( ) ( ) 0.88 ( ) 1.67 ( ) 0.95 ( ) Women n Metabolic syndrome, n Crude OR 1.56 ( ) ( ) 0.97 ( ) 2.07 ( ) 0.77 ( ) Adjusted a OR 1.23 ( ) ( ) 0.95 ( ) 2.01 ( ) 0.84 ( ) The metabolic syndrome was defined when three of six metabolic criteria where fulfilled. Odds ratios for the metabolic syndrome was calculated with logistic regression. a Adjusted for smoking, education, immigration, employment, physical activity, and intake of vegetables. Discussion In this population-based study, every third 60-y-old man and woman in Stockholm County was invited and 78% participated. From self-reported alcohol intake, three groups were identified, no, low and light to moderate groups, which in turn were divided into a wine, beer, spirits and mixed groups. In a validation study by Gronbaek & Heitmann (1996), a simple questionnaire for alcohol intake was compared with a diet history method. The mean frequency of beer, wine and spirits intake (drinks=day) did not differ. The correlation between the two methods ranged from 0.59 (spirits intake in women) to 0.87 (beer intake in men). The classification of the alcohol groups may be more or less crude. The non-drinking group probably consists of former drinkers, life-long non-drinkers, and perhaps under-reporters of occasional drinkers. It has been found that that former drinkers are more likely to be heavy smokers, depressed, unemployed and in poorer health than long term non-drinkers (Middleton Fillmore et al, 1998). Nondrinkers who abstain from alcohol for principle reasons (religious or ideological) have on the other hand been shown to have lower mortality rates compared with the general population (Fonnebo, 1992; Kjaerheim et al, 1993). However, this group most likely represented a minor part of our nondrinking group (Andréasson, 1998). Our results showed that light to moderate intake of alcohol was associated with a social background characterized by being married, university educated, employed, and Swedish born. Importantly, these associations were seen only in the wine and the mixed group, and not in the beer and spirits groups. In non-drinking women, a high proportion of non-married, immigrants and unemployed were seen. Smokers were most prevalent in the spirits group in both men and women. Exsmokers were most common among the wine drinkers, which could be an indicator of health consciousness in this group. These lifestyle patterns correspond to what other studies have found (Klatsky et al, 1990; Truelsen et al, 1998; Wannamethee & Shaper, 1999). The variables regarding quality of life (economy, leisure time and health) further exhibited the differences in the characteristics. The self-reported health variable was most prominent, where 87 and 81% of wine-drinking men and women, respectively, reported good health. The corresponding figures in non-drinkers were 62 and 56%, respectively. Similar results have been found in studies from Finland and Denmark where good subjective health was significantly more prevalent among wine drinkers compared with non-wine drinkers after adjustments for covariates (Poikolainen & Vartianen, 1999; Gronback, et al, 1999). The interpretation of this variable is not clear; self-reported health has however turned out to be an independent predictor of mortality (Idler & Benyamini, 1997). The high prevalence of muscle skeletal symptoms in non-drinking women was noteworthy, 49% compared with 30% in the wine group. The dietary habits tended to be healthier in the wine and mixed groups and less healthy among the beer and spirits drinkers. This agrees with the results from the Danish Diet, Cancer and Health Study (Tjonneland et al, 1999). Altogether, this study showed that alcohol intake, and the choice of wine, beer and spirits, were related to a number of lifestyle variables in different ways, indicating that alcohol intake is an integrated part of the ordinary way of living. Regarding the metabolic syndrome, a more favorable metabolic profile was seen in the wine and the mixed group, with lower fasting insulin, higher HDL-cholesterol, and lower fibrinogen concentrations. The waist-to-hip ratio was highest in non-drinkers and spirits drinkers, which probably explains much of the unfavorable metabolic profile in these groups. When the biological parameters were combined in the metabolic syndrome, a higher prevalence was seen in non-drinking women compared with the low alcohol group. A lower prevalence was seen in wine-

7 drinking women. After adjustments, a significant lower odds ratio was seen in the wine group in women. The analyses are based on cross-sectional data, and the odds ratios should therefore be interpreted with caution. The main purpose of these analyses was not to estimate a relative risk, but rather to see how the estimates could change when adjustments were made for some confounding factors. If the favorable metabolic pattern in wine drinkers is due to the intake of alcohol, the question remains: why did the other groups of moderate alcohol intake not show the same metabolic pattern? Differences in drinking frequency could be one explanation (Britton & McKee, 2000). The drinking frequency was not known in this study, and binge drinking might have been more common in the spirit group than among those who preferred wine. Changes in alcohol habits during one s life could also play a role. An increased alcohol intake was more often reported among the wine-drinkers compared with the other groups (data not shown). This corresponds with food consumption data indicating that wine consumption is increasing in Sweden, while the intake of spirits is decreasing (Statistics Sweden, 1997). Another possible explanation is miss-reporting. The serum concentration of g-gt was significantly raised in the spirit group compared with the wine group in both men and women (P < 0.05). Since the concentration of g-gt is related to alcohol consumption, this might indicate that underreporting was present to a greater extent in this group. Moreover, wine-specific components such as flavonoids may also contribute the diverging results. However, factors such as diet, smoking, physical activity, economy, social and mental well-being may also contribute to the metabolic effects seen in the different alcohol groups. Findings from a recently published study from Denmark, where wine drinking was associated with higher IQ, further illustrate the role of alcohol intake as an indicator of social and psychosocial characteristics (Mortensen et al, 2001). It is probably difficult to obtain full adjustments for the confounding influences of all such potential factors that may obscure the estimates of the effects of alcohol on health. In conclusion, compared with low alcohol drinkers, moderate wine drinkers exhibited a more favorable pattern according to both lifestyle factors and metabolic parameters. The close link between alcohol drinking behavior and lifestyle factors contribute to the difficulties inherited in delineating the true effects of alcohol beverage on health. References Allain CC, Poon LS, Chan CSG, Richmond W & Fu PC (1974): Enzymatic determination of total serum cholesterol. Clin. Chem. 20, Andréasson S (1998): Alcohol and J-shaped curves. Alcohol Clin. Exp. Res. 22, 359S 364. Balkau B & Charles MA for the European Group for the Study of Insulin Resistance (1999): Comment on the provisional report from the WHO consultation. Diabet. Med. 16, Becker U, Bartl K & Wahlfeld AW (1984): A functional photometric assay for plasma fibrinogen. Thromb. Res. 35, Bell RA, Mayer-Davis EJ, Martin MA, D Agostino RB & Haffner SM (2000): Associations between alcohol consumption and insulin sensitivity and cardiovascular disease risk factors. Diabet. Care 23, Britton A & McKee M (2000): The relation between alcohol and cardiovascular disease in Eastern Europe: explaining the paradox. J. Epidemiol. Community Hlth 54, Caccette RA-A, Croft KD, Beilin LJ & Puddey IB (2000): Ingestion of red wine significantly increases plasma phenolic acid concentrations but does not acutely affect ex vivo lipoprotein oxidizability. Am. J. Clin. Nutr. 71, Dallongeville J, Marécaux N, Ducimetiére P, Ferriéres J, Arveiler D, Bingham A, Ruidavets JB, Simon C & Amouyel P (1998): Influence of alcohol consumption and various beverages on waist girth and waist-to-hip ratio in a sample of French men and women. Int. J. Obes. Relat. Metab. Disord. 22, DeFronzo RA & Ferrannini E (1991): Insulin resistance. A multifaceted syndrome responsible for NIDDM, obesity, hypertension, dyslipidemia, and atherosclerotic cardiovascular disease. Diabet. Care 14, Fagrell B, de Faire U, Bondy S, Criqui M, Gaziano M, Gronbaek M, Jackson R, Klatsky A, Salonen J & Shaper AG (1999): The effects of light to moderate drinking on cardiovascular diseases. J. Intern. Med. 246, Fonnebo V (1992): Mortality in Norwegian Seventh-day Adventists J. Clin. Epidemiol. 45, Fossati P & Prencipe L (1982): Serum triglycerides determined colorimetrically with an enzyme that produces hydrogen peroxide. Clin. Chem. 28, Fossati P, Prencipe L & Berti G (1980): Use of 3,5-dichloro-2- hydroxybenzenesulfonic acid=4-amino-phenazone chromogenic system in direcy enzymatic assay of uric acid in serum and urin. Clin. Chem. 26, Friedewald WT, Levy RI & Fredrickson DS (1972): Estimation of the concentration of low-density lipoprotein cholesterol n plasma, without use of the preparative ultracentrifuge. Clin. Chem. 18, Gaziano JM, Buring JE, Breslow JL, Goldhaber SZ, Rosner B, VanDenburgh M, Willet W & Hennekens CH (1993): alcohol intake, increased levels of high-density lipoprotein and its subfractions, and decreased risk of myocardial infarction. New Engl. J. Med. 329, 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, Gronbaek M & Heitmann BL (1996): Validity of self-reported intakes of wine, beer and spirits in population studies. Eur. J. Clin. Nutr. 50, Gronbaek M, Deis A, Sorensen TIA, Becker U, Schnohr P & Jensen G (1995): Mortality associated with moderate intakes of wine, beer, or spirits. Br. Med. J. 310, Gronbaek M, Mortensen EL, Mygind K, Andersen AT, Becker U, Gluud C & Sorensen TIA (1999): Beer, wine, spirits and subjective health. J. Epidemiol. Community Hlth 53, Hein HO, Suadicani P & Gyntelberg F (1996): Alcohol consumption, serum low density lipoprotein cholesterol concentration, and risk of ischaemic heart disease: six year follow up in the Copanhagen male study. Br. Med. J. 312, Hemker HC, Kop J & Willems GM (1979): Kinetic aspects of the interaction of blood clotting enzymes VIII. The relation between clotting time and clotting velocity. Thromb. Haemostas. 41, Idler EL & Benyamini Y (1997): Self-rated health and mortality: a review of twenty-seven community studies. J. Hlth Soc. Behav. 38, Jéquier E (1999): Alcohol intake and body weight: a paradox. Am. J. Clin. Nutr. 69, Kiechl S, Willeit J, Poewe W, Egger G, Oberhollenzer F, Muggeo M & Bonora E (1996): Insulin sensitivity and regular alcohol consumption: large, prospective, cross sectional population study (Bruneck study). Br. Med. J. 313, Kjaerheim K, Andersen A & Helseth A (1993): Alcohol abstainers: a low-risk group for cancer a cohort study of Norwegian teetotalers. Cancer Epidemiol. Biomarkers Prev. 2,

8 234 Klatsky AL, Armstrong MA & Kipp H (1990): Correlates of alcoholic beverage preference: traits of persons who choose wine, liquor or beer. Br. J. Addict. 85, Klatsky AL, Armstrong MA & Friedman GD (1997): Red wine, white wine, liquor, beer, and risk for coronary artery disease hospitalization. Am. J. Cardiol. 80, Lazarus R, Sparrow D & Weiss ST (1997): Alcohol intake and insulin levels. The Normative aging study. Am. J. Epidemiol. 145, Middleton Fillmore K, Golding JM, Graves KL, Kniep S, Leino EV, Romelsjö A, Shoemaker C, Ager CR, Allebeck P & Ferrer HP (1998): Alcohol consumption and mortality. I. Characteristics of drinking groups. Addiction 93, Mortensen EL, Jensen HH, Sanders SA & Reinisch JM (2001): Better psychological functioning and higher social status may largely explain the apparent health benefits of wine. Arch. Intern. Med. 161, Nigdikar SV, Williams NR, Griffin BA & Howard AN (1998): Consumption of red wine polyphenols reduces the susceptibility of low-density lipoproteins to oxidation in vivo. Am. J. Clin. Nutr. 68, Poikolainen K & Vartianen E (1999): Wine and good subjective health. Am. J. Epidemiol. 150, Rimm EB, Klatsky A, Grobbee D & Stampfer MJ (1996): Review of moderate alcohol consumption and reduced risk of coronary heart disease: is the effect due to wine, beer, or spirits? Br. Med. J. 312, Rimm EB, Williams P, Fosher K, Criqui M & Stampfer MJ (1999): alcohol intake and lower risk of coronary heart disease: meta-analyses of effects on lipids and heamostatic factors. Br. Med. J. 319, Sakurai Y, Umeda T, Shinchi K, Honjo S, Wakabayashi K, Todoroki I, Nishikawa H, Ogawa S & Katsurada M (1997): Relation of total and beverage-specific alcohol intake to body mass index and waist-tohip ratio: A study of self-defence officials in Japan. Eur. J. Epidemiol. 13, Serafini M, Maiani G & Ferro-Lussi A (1998): Alcohol-free red wine enhances plasma antioxidant capacity in humans. J. Nutr. 128, Shaper AG, Wannamethee SG & Walker M (1988): Alcohol and mortality in British men: explaining the U-shaped curve. Lancet 2, Statistics Sweden (1997): När mat kommer på tal (in Swedish). Temelkova-Kurktschiev T, Henkel E, Koehler C & Hanefeld M (2001): Light-to-moderate alcohol consumption, risk factor profile and early atherosclerosis: the RIAD Study. Atherosclerosis 156, Thun MJ, Peto R, Lopez AD, Monaco JH, Henley J, Heath CW & Oll R (1997): Alcohol consumption and mortality among middle-aged and elderly U.S adults. New Engl. J. Med. 337, Tibblin G, Cato K & Svärdsudd K (1990): Göteborg Quality of Life Study of men born in 1913 and 1923 age, sex, job satisfaction and cardiovascular diseases. Scand. J. Prim. Health Care 1(Suppl 1), Tjonneland A, Gronbaek M, Stripp C & Overvad K (1999): Wine intake and diet in a random sample of Danish men and women. Am. J. Clin. Nutr. 69, Trinder P (1969): Determination of blood glucose using an oxidaseperoxidase system with a non-carcinogenic chromogen. Clin. Pathol. 22, Truelsen T, Gronbaek M, Schnohr P & Boysen G (1998): Intake of beer, wine, and spirits and risk of stroke. The Copenhagen City Heart Study. Stroke 29, Wannamethee SG & Shaper AG (1999): Type of alcohol drink and risk of major coronary heart disease events and all-cause mortality. Am. J. Publ. Hlth 89,