Alcohol and adiposity: effects of quantity and type of drink and time relation with meals

Transcription

1 (2005) 29, & 2005 Nature Publishing Group All rights reserved /05 $30.00 PAPER : effects of quantity and type of drink and time relation with meals SG Wannamethee 1 *, AG Shaper 1 and PH Whincup 2 1 Department of Primary Care and Population Sciences, Royal Free and University College Medical School, Rowland Hill St, London, UK; and 2 Department of Public Health Sciences, St George s Medical School Hospital, London, UK BACKGROUND: It has been suggested that the effects of alcohol on body weight and fat distribution may be influenced by the quantity and type of drink and may differ according to whether the alcohol is consumed with meals or not. OBJECTIVE: We have examined the cross-sectional association between alcohol intake, patterns of drinking and adiposity (body mass index (BMI), waist-to-hip ratio (WHR), waist circumference (WC) and percentage body fat (%BF)). METHODS: We studied 3327 men aged y with no history of myocardial infarction, stroke or diabetes drawn from general practices in 24 British towns. RESULTS: BMI, WHR, WC and %BF increased significantly with increasing alcohol intake even after adjustment for potential confounders (all Po0.0001), although the effects were stronger for WC and WHR (measures of central adiposity). Men who consumed Z21 units/week showed higher levels of central adiposity (WHR, WC) and general adiposity (BMI, %BF) than nondrinkers and lighter drinkers, irrespective of the predominant type of drink consumed (wine, beer, spirits or mixed). The positive association was most clearly seen in beer and spirit drinkers; positive but weaker associations were seen for wine. Among drinkers, a positive association was seen between alcohol intake and the adiposity variables irrespective of whether the alcohol was drunk with or separately from meals. CONCLUSION: Higher alcohol consumption (Z21 units/week) is positively associated with general and to a greater extent with central adiposity, irrespective of the type of drink and whether the alcohol is drunk with meals or not. (2005) 29, doi: /sj.ijo ; published online 2 August 2005 Keywords: alcohol intake; type of drink; BMI; abdominal fat; % body fat Introduction Increased body weight and in particular, abdominal obesity is associated with increased cardiovascular disease risk. 1 3 Alcohol ingestion suppresses the oxidation of fat, favouring fat storage, and can serve as a precursor for fat synthesis. 4 A number of epidemiological studies in men have reported positive associations between alcohol and body weight and abdominal obesity as measured by the waist-to-hip ratio (WHR) but some have reported no association Several factors have been proposed, which may explain the inconsistencies between studies, including the suggestion that the effect of alcohol on adiposity is influenced by type of drink. 4 There is a widespread belief that drinking beer promotes abdominal fat ( beer belly ) 6 and that wine has little or no such effect. 9,14 The few studies that have examined the influence of type of beverage on fat distribution in men have been inconsistent, some showing a positive effect of wine on abdominal fat 12,15 and others observing no effect or even an inverse effect with wine. 9,14 It has also been postulated that the effects of alcohol on body weight and fat distribution may differ according to whether the alcohol is consumed with meals or not 7 although data are limited. This paper examines the association between alcohol intake, patterns of alcohol intake and adiposity (body mass index (BMI), WHR, waist circumference (WC) and percentage (%) body fat (%BF)) in a large cross-sectional study of men aged y, with particular focus on the effects of quantity and type of drink and on the effects of drinking with or separately from meals. *Correspondence: Dr SG Wannamethee, Department of Primary Care and Population Sciences, Royal Free and University College Medical School, Rowland Hill St, London NW3 2PF, UK. goya@pcps.ucl.ac.uk Received 18 November 2003; revised 4 August 2004; accepted 12 August 2004; published online 2 August 2005 Subjects and methods The British Regional Heart Study is a prospective study of cardiovascular disease involving 7735 men aged y, selected at random from the age sex registers of one general practice in each of 24 British towns and screened between

2 1978 and The practices were selected to reflect the socioeconomic distribution of men in each town. All men provided informed written consent to the investigation, which was carried out in accordance with the Declaration of Helsinki. Postal questionnaires measuring medical history and lifestyle changes were sent to all survivors 5 y after screening ( ; Q5), again in 1992 (Q92) and in 1996 (Q96). In , all surviving men, now aged y, were invited for a 20th year follow-up examination. The men completed a questionnaire (Q20) providing information on their medical history, smoking and drinking habits, physical activity and occupation, then attended for physical examination, and provided a fasting blood sample. Of the 5565 surviving subjects, 4252 (76.4%) attended for examination. As a result of the strong influence that the diagnosis of CHD, stroke or diabetes is known to have on drinking patterns, 17 and because of the strong association between BMI and cardiovascular disease and diabetes, we have excluded from the analyses 869 men with recall of a doctor diagnosis of CHD, stroke or diabetes leaving 3383 men for analysis. Alcohol intake Standard questionnaire (Q20). The men were asked about current frequency of drinking (daily, most days, weekend only, occasional, special occasions only or none) and were asked to estimate the number of alcoholic drinks during an average week. 1 UK unit of alcohol (drink) was defined to the men as half a pint of beer, a single measure of spirits or a glass of wine (approximately 10 g alcohol). Since our earlier reports on these men had shown that BMI is increased at levels of 3 or more drinks a day (Z21 units/week), the men were classified into five groups according to their reported weekly intake: none or o1/week, 1 6/week, 7 20/week, 21 34/week and Z35 drinks/week. Heavy drinking refers to those drinking Z35 units/week (five or more drinks daily). The men were asked to indicate which types of drink they usually took that is, the predominant drinks: (1) beer (N ¼ 1071; 42%); (2) wine (N ¼ 306; 12%); (3) spirit (N ¼ 201; 8%); (4) mixed wine and sherry, beer and spirits (N ¼ 937; 37%) or (5) low alcohol drinks (N ¼ 28; 1%). These groups were mutually exclusive. They were also asked whether the alcohol they drank was usually taken (1) before meals, (2) with meals, (3) after meals or (4) separately from meals. In the analyses, groups (1) and (3) were combined and referred to as before/after meals. A total of 54 men did not provide information on alcohol intake and were excluded. A further 20 men did not provide details on type of drink and 55 men did not provide data on drinking with meals and were excluded from analyses involving type of drink and the influence of meals. Dietary questionnaire (Q20-7). Information on alcohol intake was also obtained from a 7-day recall diet questionnaire, which was posted to the men for completion prior to attending screening. The men were asked to report the number of drinks of wine, beer and spirits consumed for each day of the week in the previous 7 days as part of a dietary assessment. Intakes of beer, wine and spirits were categorised quantitatively for the individual types of beverages (none/ o1 week, 1 6 drinks/week, 7 20 drinks/week and Z21 drinks/week). The correlation between quantity of alcohol obtained from the standard questionnaire (Q20) and the 7- day recall dietary questionnaire (Q20-7) was There was good agreement between the response to the Q20 question on the type of alcohol usually drunk and the 7-day recall of specific beverages drunk (Q20-7). In men who stated they usually drank beer, beer accounted for 79% of total alcohol intake based on 7-day recall. Wine and spirit accounted for 77% of total alcohol intake in those who on questionnaire (Q20) usually drunk wine and spirit, respectively. Among those who usually drunk mixed drinks, beer accounted for 37%, spirit 26% and wine 37% of their total intake. Since information on the time at which alcohol is usually taken in relation to the meal was asked on the standard questionnaire (Q20), we have in the main used reported alcohol intake from the standard questionnaire (Q20) to assess the relationship between alcohol and adiposity. Since the standard questionnaire only had questions on beverage usually drunk rather than the precise quantity for the specific type of drink, we have presented data using two sources of information in assessing the effects of type of drink: (i) beverage usually drunk, that is, predominant type of drink obtained from the standard questionnaire (Q20) (Figure 1) and (ii) data on type of drink from the dietary questionnaire (Q20-7) (Figure 2). Data from the dietary questionnaire provides the exact number of drinks reported for each specific beverage and allows quantification of the specific beverage type. This allows us to examine the specific effects of wine, beer and spirits on adiposity in addition to using information from Q20 on the beverage usually drunk. Dietary intake Dietary intake was obtained from a detailed 7-day recall food frequency questionnaire (FFQ) that was developed for use in the Scottish Heart Health Study 18 and later for the World Health Organization MONICA survey and was validated against weighed intake in both a Welsh 19 and a Scottish population. 20 Study participants were asked to recall their usual intake during the past 7 days by reporting amounts and frequency of food consumed, which included 86 different food and drinks. Nutrient intakes were calculated by using a validated program that multiplied the food frequency by the standard portion sizes for each food and by the nutrient composition of the foods obtained from the UK food composition tables. 21 Potential confounders Details of classification methods for smoking status, physical activity, BMI and social class have been described. 16,22 From 1437

3 1438 the combined information at initial screening (Q1; ) and follow-up questionnaires, the men were classified into five smoking groups: (i) those who had never smoked, (ii) ex-smokers since screening, (iii) smokers at baseline who gave up between screening and 1996 (Q96), (iv) smokers at screening and Q96 who gave up after 1996, and (v) current cigarette smokers at Q20. A physical activity score was derived for each man at Q20 and the men were grouped into six broad categories: inactive, occasional, light, moderate, moderately vigorous and vigorous. 22 The longest-held occupation of each man was recorded and coded in accordance with the Registrar General s occupational classification into six social class groups: I, II, III nonmanual, III manual, IV and V. The Armed Forces formed a separate group. Other pre-existing cardiovascular disease The men were also asked about a doctor diagnosis of heart failure, other heart trouble, aortic aneurysm, claudication, deep vein thrombosis or pulmonary embolism. Other preexisting CVD included men with a history of any of the above conditions. Statistical methods Analysis of covariance was used to obtain adjusted means by alcohol categories using the PROC GLM procedures of SAS version 6.12;. SAS Institute Inc., Cary, NC, USA. In the adjustment smoking (five groups), social class (seven groups) and current physical activity (six groups) were fitted as categorical variables. Age, total fat intake and total nonalcohol calories were fitted as continuous variables. We conducted tests of linear trend across increasing categories of alcohol consumption by assigning the median for the alcohol categories and treating the categories a continuous variable. Standardised regression coefficients (regression coefficient/s.d. of the adiposity measure) were calculated to compare the strength of association between alcohol and the adiposity measures. We assessed whether the trends varied by whether alcohol was taken with meals or by predominant type of drink by including cross product terms for alcohol consumption (continuous) and type of drink/meals in the multivariate model. Since BMI and WC were highly correlated (r ¼ 0.87), we have used WHR (correlation with BMI ¼ 0.54) when adjusting for central adiposity in the multivariate analysis. Anthropometric measurements The anthropometric measurements at re-examination included height, weight, waist and hip circumference and %BF. Subjects were measured in light clothing. Height and weight were both measured standing. Height was measured with a Harpenden stadiometer to the least complete 0.1 cm and weight with a Soehnle digital electronic scale to the last complete 0.1 kg. BMI (weight/height 2 in kg/m 2 ) was calculated for each man. BMI was not available in 10 men. Waist and hip circumferences were measured in duplicate with an insertion tape (CMS Ltd London); hip circumference was measured at the point of maximum circumference over the buttocks. The waist measurement was taken from the midpoint between the iliac crest and the lower ribs measured at the sides. WHR was calculated as WC divided by hip circumference (cm). Adjustments were made for the observer variation in WHR measurements. WC was not available in 16 men and WHR was missing in 19 men. %BF was estimated using bioelectric impedance method measured with the Bodystat 500 apparatus and applying the equation of Deurenberg et al. 23 Fat-free mass was calculated as 6710 height in metres squared/resistance in ohms þ 7. %BF was calculated as [body weight fat free mass]/body weight. Data on %BF were missing in 108 men. Of the 3329 men with data on alcohol, 3327 men had at least one anthropometric measurement. WHR and WC were used as measures of abdominal (central) adiposity. BMI and %BF were used as measures of general adiposity. High WC was defined as WC 4102 cm. 24 Obesity was defined as BMI Z30 kg/m 2. Results In the 3327 men free of CVD or diabetes with data on alcohol intake and at least one adiposity measurement (BMI, WHR, WC, %BF), the mean (s.d.) BMI was 26.8 (3.6) kg/m 2, the mean WC was 96.8 (10.2), the mean WHR was (0.06) and the mean %BF was 34.8 (8.1). Mean total calorie intake increased significantly with increasing alcohol intake (Table 1). There was little difference in intake of nonalcohol calories between the alcohol groups. Nondrinkers and heavy drinkers had higher total dietary fat intake than other alcohol intake groups. Alcohol intake was significantly and positively associated with age-adjusted mean BMI, WHR, WC and %BF (Table 1). Adjustment for confounders including smoking, physical activity, other pre-existing CVD, social class, dietary total fat and total nonalcohol calories strengthened the associations with both BMI and WHR. Prevalence of obesity (BMI Z30 kg/m 2 ) increased slightly (nonsignificantly) with increasing alcohol intake but the percentage of men with high WC (4102 cm) increased strongly and significantly with increasing alcohol intake. Comparisons of standardised regression coefficients (see Subjects and methods) showed the effects of alcohol on adiposity to be greater for abdominal adiposity (WHR and WC) than for general adiposity (BMI and %BF) (Table 1). Further adjustment for WHR attenuated the positive relationship between alcohol and BMI, although the trend remained significant (P ¼ 0.02). The positive association between alcohol and WHR remained after adjustment for BMI (P ¼ 0.002).

4 Table 1 Alcohol intake (Q20) and mean total daily calories, dietary total fat intake and adjusted mean BMI, waist to-hip ratio (WHR), waist circumference and percentage body fat 1439 Alcohol intake (drinks/week) None/o1 week 1 6/week 7 20/week 21 34/week 35+/week P-value trend SRE (s.e.) N Total calories (kcal) a Total nonalcohol calories (kcal) NS Dietary total fat (g/day) BMI (kg/m 2 ) Mean b Adjusted mean c (0.001) % obese (430 kg/m 2 ) WHR Mean b Adjusted mean c (0.001) Waist circumference (cm) Mean b Adjusted mean c (0.001) % 4102 cm % body fat Mean b Adjusted mean c (0.0003) a Total calories and nonalcohol calories was obtained from dietary recall. b Age-adjusted. c Adjusted for age, social class, smoking, physical activity, other preexisting CVD disease, total fat and total nonalcohol calories. SRE (s.e.) ¼ standardised regression coefficient (standard error). The relationships between total alcohol intake and the adiposity variables were very similar when the 7-day recall of alcohol intake was used. Quantity and type of drink and relationships of alcohol intake to meals Light drinkers were far more likely than others to drink with their meals (Table 2). Heavier drinkers were less likely to drink with meals. Men who reported drinking wine (predominantly or mixed) were far more likely to drink with meals than those who usually drink only beer or spirits. Spirit drinkers were least likely to drink with meals. Table 2 meals Alcohol intake (drinks/week) Alcohol quantity and type of beverage (Q20) and relation with Before/after meal (N ¼ 563) With meal (N ¼ 1227) Separate from meals (N ¼ 563) % 56% 30% % 46% 30% % 43% 31% % 32% 39% Type of drink a Beer 30% 30% 40% Wine 9% 83% 8% Spirits 35% 22% 43% Mixed 13% 67% 20% a In all, 28 men who consumed low alcoholic drinks were omitted. Type of drink and time of meal Wine drinkers had significantly lower average consumption, lower total nonalcohol calorie intake and dietary total fat intake than each of the other drinking groups (Po0.05) (Table 3). Beer drinkers had the highest dietary total fat intake. Men who drank with meals tended to have lower total calorie and dietary total fat intake than men who drank separately or before or after meals. We examined the differences in mean adiposity by types of drink usually drunk and by time of alcohol consumption with meals in regular drinkers (41/week) using beer drinkers and those who drunk with meals as the comparison groups. Wine drinkers showed significantly lower BMI but not WHR, WC or %BF compared with beer drinkers after adjustment for confounders. Those who usually drank with meals had significantly lower mean levels of BMI and WHR than those who drank separately from meals and lower levels than those who usually drank before/after meals. Since wine drinkers were more likely to drink with meals and were more likely to be lighter drinkers, adjustment for relationship with meals attenuated the difference seen for BMI between wine and beer drinkers and the difference was of marginal significance. Further adjustment for amount and type of drink made minor differences to the relationships seen between patterns of drinking in relation to meals and the adiposity variables.

5 1440 Table 3 Type of beverage and time of meal (Q20) and adjusted a mean BMI, WHR, WC and % body fat in men drinking at least 1 unit/week Usual consumption of alcohol in relation to meals N Beer 1037 Wine 303 Spirit 198 Mix 931 Before/after 563 Separate 716 With meal 1227 Average no. drinks/week (s.e.) 12.4 (0.42) 9.6 (0.45) 12.0 (0.96) 12.3 (0.36) 13.5 (0.50) 13.2 (0.51) 10.4 (0.28) Total nonalcohol calories (kcal) (s.e.) 2126 (17.9) 1928 (26.1) 1990 (37.5) 2033 (16.1) 2070 (23.5) 2097 (21.5) 2024 (14.1) Dietary Total Fat (g/day) (s.e.) 75.5 (0.9) 66.4 (1.28) 70.1 (1.81) 72.4 (0.85) 74.4 (1.16) 74.9 (0.71) 70.7 (1.09) Adiposity measures Mean BMI (kg/m 2 ) a 27.0 (0.06) 26.6 (0.11)* 26.7 (0.16) 26.8 (0.07) 26.8 (0.10) 27.0 (0.09)* 26.7 (0.07) (s.e.) # * 26.7 Mean WHR a (s.e.) b (0.002) (0.003) (0.003) (0.002) (0.003) (0.002)* (0.002) # Mean WC (cm) a (s.e.) b 97.1 (0.24) 96.5 (0.42) 96.3 (0.53) 96.8 (0.24) 96.9 (0.33) 97.1 (0.29) 96.6 (0.23) Mean % body fat a (s.e.) b 35.0 (0.24) 34.5 (0.43) 35.5 (0.54) 34.7 (0.25) 34.4 (0.33) 35.2 (0.30) 34.8 (0.23) s.e. ¼ standard error. a Adjusted for age, social class, smoking, physical activity, other pre-existing CVD disease, total fat, total nonalcohol calories and total usual alcohol intake/week. b Adjusted in addition for type of beverage /time of meal. Comparisons with beer drinking/comparisons with with meal *Po0.05; # 0.05o Po0.06. Amount of alcohol, predominant/usual type of drink and adiposity When examined by predominant type of drink, for example, beer, wine, spirits and mixed, the positive association between total alcohol intake and BMI was seen for all types of drink although it was less apparent for wine drinking (Figure 1). As a result of the small number of men who drank predominantly wine or spirits, in particular those whose total intake was 421 units (38 and 30 men, respectively), we are primarily interested in the pattern of relationship rather than the statistical significance of the trend within each preference group. A positive association was seen between weekly alcohol intake and WHR, WC and %BF for all types of drink. All moderate/heavy drinkers (Z21 drinks/day) irrespective of type of drink showed higher adiposity levels than those consuming o1 drink/week. There were no heavy wine drinkers (Z35 drinks/week) in the study but moderate wine drinkers (21 34 drinks/week) showed higher levels of adiposity (BMI, WHR, WC, body fat) than nondrinkers, whereas lighter wine drinkers (7 20/drinks week) showed lower adiposity levels than nondrinkers, although these differences were not statistically significant (Figure 1). Formal tests for interaction to see whether the relationships between alcohol and measures of adiposity differed by type of drink were not significant (P ¼ 0.80 for BMI, 0.53 for WHR, 0.65 for WC and 0.47 for %BF). Specific types of drink We have also examined the effects of type of drink using alcohol data from the 7-day dietary assessment (Q20-7), which enabled us to determine the quantity drunk of each specific beverage (Figure 2). In all, 57% (N ¼ 1906) of the men reported drinking beer, 43% (N ¼ 1445) wine and 41% (N ¼ 1362) spirits. As the categories were not mutually exclusive, the relationship between weekly amount of wine, beer and spirits was examined in relation to the adiposity measures adjusting for potential confounders, time of meal and each of the other type of drink. The patterns were similar to the findings in Figure 1 based on the predominant type of drink. Those drinking Z21 drinks/week, irrespective of type of drink, showed higher central adiposity (WHR and WC) and %BF than lighter drinkers, although the effect was most marked for beer drinking. The number of men drinking more than 21 drinks/week of wine or spirits was small (66 and 67 men, respectively) and the trends were only significant for beer drinkers (Po0.001) except for WHR, which was also significant in spirit drinkers. For BMI little association was seen with wine. Amount of alcohol, time of meal and adiposity We examined the relationship between total alcohol intake and adiposity in relation to meals adjusting for potential confounders and predominant type of drink (Figure 3). Among drinkers, the positive association between usual alcohol intake and adiposity was seen irrespective of time relation to meals, although the effect was more marked in those who did not usually drink with their meals (test for trend all Po0.05 in men who drank after/before and separate combined). In men who drank with meals the test for trends were P ¼ 0.13 for BMI, P ¼ 0.04 for WHR, P ¼ 0.11 for WC and P ¼ for %BF. However, formal tests for interaction to see whether the effects of alcohol differed between drinking with or separately (separate and before/after groups combined) from meals was not statistically significant for any of the adiposity variables BMI (P ¼ 0.44), WHR (P ¼ 0.28), WC (P ¼ 0.78) and %BF (P ¼ 0.97).

6 BMI (kg/m 2 ) Waist circumference (cm) Beer Wine Spirit Mix Nondrinker Nondrinker Beer Wine Spirit Mix 0 1-6/w 7-20/w 21+/w 0 1-6/w 7-20/w 21+/w Waist-to-hip ratio % body fat Nondrinker Nondrinker Beer Wine Spirit Mix Beer Wine Spirit Mix Figure 1 Weekly alcohol intake of predominant/usual type of drink (Q20) and adjusted mean BMI, WHR, WC and % BF. Adjusted for age, social class, smoking, physical activity, other pre-existing CVD, dietary total fat intake and time relation with meal BMI (kg/m 2 ) P< Beer Wine Spirit 0 1-6/w 7-20/w 21+/w Waist-to-hip ratio P< P=0.04 Beer Wine Spirit Waist circumference (cm) P< /w 7-20/w 21+/w % body fat 33.5 Beer Wine Spirit Beer Wine Spirit P< Figure 2 Weekly alcohol intake of specific drinks (beer, wine, spirits, mixed) and adjusted mean BMI, WHR, WC and %BF. Adjusted for age, social class, smoking, physical activity, other pre-existing CVD, dietary total fat intake, time taken with meal and other types of alcohol. P=0.06 Discussion Most studies report that alcohol energy is usually added to total food energy as seen in this study. 7,25 We have observed a significant positive association between alcohol intake and general adiposity (BMI and %BF) and central adiposity (WHR and WC). The positive relationship between alcohol and BMI

7 1442 BMI (kg/m 2 ) /w 7-20/w 21+/W Alcohol intake/week Waist-to-hip ratio /w 7-20/w 21+/w Alcohol intake/week Waist circumference (cm) % body fat /w 7-20/w 21+/w 1-6/w 7-20/w 21+/W Alcohol intake/week Alcohol intake/week Before/after With Separate Figure 3 Total alcohol intake and adjusted mean BMI, WHR, WC and %BF in relation to time taken with meals. Adjusted for age, social class, smoking, physical activity, other pre-existing CVD, dietary total fat intake and predominant type of drink. Test for trend all Po0.05 in men who drank without meals (after/before and separate combined). Test for trend in men who drank with meals BMI (P ¼ 0.13), WHR (P ¼ 0.04), WC (P ¼ 0.11), %BF (P ¼ 0.001). in these older men is consistent with our previous report carried out when these men were in middle age. 13,26 The present study extends the findings and examines the patterns of drinking as well as other important measures of adiposity including %BF, WC and WHR. The relationships between alcohol and general adiposity and central adiposity were seen for all types of drink. It was most apparent for beer drinking and was independent of meals, although there appeared to be a stronger effect of alcohol on body weight in men who drank separately from meals. However, we could not confirm a significant difference in the alcohol adiposity relationships in relation to time of meal. Alcohol and body fat distribution Evidence from a number of studies suggest that in drinkers, fat is preferentially deposited in the abdominal area. 7. In contrast to the cross-sectional relationship between alcohol and body weight, which has been found to be almost equally positive or nonexistent in men and negative in women, 5 8 more recent studies tend to report positive associations between alcohol and body fat distribution. In the French MONICA study, alcohol consumption was positively associated with WHR independently of BMI in both men and women. 12 Positive associations have also been reported in studies from Switzerland, 10,27 US men and women, 9,15,28 Italian women 29 and in Japanese men. 11 Some studies have observed stronger association between alcohol and central adiposity as measured by the WHR than with BMI. 11,12 We have observed a positive relationship with both central and general adiposity but the effects as measured by the standardised regression coefficients were greater for WC and WHR than for BMI and %BF, and the increase in percentage of men with large WC was more marked than the increase in rates of obesity as measured by BMI. The relationship with BMI was attenuated after adjustment for the WHR, although it remained significant, suggesting that alcohol is more associated with abdominal obesity than with general obesity. Type of alcohol and adiposity It has been suggested that the type of alcohol consumed might explain the discrepant results in studies of alcohol and adiposity. The common belief is that drinking beer promotes abdominal fat distribution and that wine in contrast has no

8 effect and may even have beneficial effects on metabolism. 6,9,30 Some studies have reported differing effect of type of beverage on body weight and fat distribution 9,11,25,31 while others have not. 12,13,15 In a large US study of men and women, beer drinking (6 drinks/week) was positively associated with increased WHR compared to nondrinkers, whereas wine drinking (6 drinks/week) was inversely related to WHR. The findings were regarded as supporting the popular concept of the beer belly. 9 By contrast, in another US Study (CARDIA) beer, wine and liquor were all positively associated with WHR in white men. 15 In the French MONICA study where wine was the main source of alcohol (67% of intake), alcohol consumption was positively associated with WHR in men and women. 12 Our findings of a positive relationship between alcohol intake and adiposity measures irrespective of type of drink support the suggestion that alcohol per se rather than any alcoholic beverage consumption is associated with increased abdominal fat deposition. 12 Prospective studies on the effects of alcohol on weight gain and central adiposity by type of drink are sparse. The Copenhagen City Heart Study reported high consumption of beer and spirits to be associated with increased WC whereas moderate to high wine consumption was associated with lower WHR. 14 The differences in findings between studies may be associated with unrecorded differences in lifestyle or differences in nutritional characteristics between wine, spirit and beer drinkers or to the level of drinking in the population. It appears that central and general adiposity is increased at 3 or more drinks a day. In many studies including our study, the number of heavier wine drinkers is very small (Z21 units/ week) and this may explain the lack of positive effect in wine drinkers. There has been suggestion that wine drinkers may take their alcohol more frequently with meals than other drinkers and consume it more slowly which in turn may have lesser effect on adiposity. 7 Regular alcohol use at meals may increase total energy expenditure by potentiating normal-dietary-induced thermogenesis. 7 Data on patterns of drinking in relation to meals are limited. We observed that wine drinkers were more likely to drink with meals than other drinkers but total alcohol intake (Z21 drinks/week) is positively associated with adiposity irrespective of whether the alcohol is usually drunk with meals or separately and is thus unlikely to explain the differences in finding between studies. Wine drinkers in this study tended to have lower total fat intake than other drinkers which is in consistent with other studies. 32 The lack of heavy wine drinkers and the multiple healthy lifestyle characteristics associated with light to moderate wine drinking 32,33 is more likely to explain why many studies have shown no association or even inverse associations for wine. Mechanisms The mechanisms involving alcohol and abdominal fat deposition are not clearly established but endocrine changes reflected by various hormonal changes including increased cortisol secretion appear to be involved. 34,35 These hormones are involved to a certain extent in the regulation of energy balance and affect fat-tissue enzymatic activities, which may promote abdominal fat deposition. 7 Limitations Although the study is a prospective cohort investigation, the present findings are based on cross-sectional associations and it is not possible to infer cause and effect from the present analyses. Data on WC, WHR and %BF was only available at re-examination. However, previous prospective data from this cohort have shown that moderate/heavy drinkers (Z3 drinks/day) have significant increase in 5-y weight gain as measured by BMI compared to none/ occasional drinkers, suggesting that heavier drinking predisposes to increased adiposity. Although the cross-sectional data from the present study are based on 77% of survivors, the mean BMI as measured at initial screening of those who attended re-examination and nonattenders were almost identical, although heavy drinkers (46 drinks/day) and nondrinkers were less likely to attend. 36 Although the prevalence of nondrinkers and heavy drinkers may be lower this should not greatly affect the alcohol adiposity relationships. The positive association between alcohol and adiposity in this study is unlikely to be due to selection bias. Conclusion In this cross-sectional study, higher total alcohol intake (Z21 units/week) is associated with increased adiposity and the effects appear to be greater for abdominal adiposity (WC) than for general adiposity (BMI). The positive association between alcohol intake and adiposity is seen irrespective of type of drink or time in relation to meals although the greatest effect appeared to be seen with beer drinking (Z3 drinks/day). There is no evidence that wine protects men from abdominal fat deposition. Acknowledgements The British Regional Heart Study is a British Heart Foundation Research Group and receives support from the Department of Health (England). We are grateful to Dr Caroline Bolton-Smith for making available the food frequency questionnaire and analysis package. The views expressed in this publication are those of the authors and not necessarily those of the Department of Health (England). References 1 Rimm EB, Stampfer MJ, Giovannucci E, Ascherio A, Spiegelman D, Colditz GA, Willett WC. Body size and fat distribution as predictors of coronary heart disease among middle-aged men and older US men. Am J Epidemiol 1995; 141:

9 Folsom AR, Stevens J, Schreiner PJ, McGovern PG. Body mass index, waist/hip ratio, and coronary heart disease incidence in African Americans and whites. Atherosclerosis Risk in Communities Study Investigators. Am J Epidemiol 1998; 148: Lakka HM, Lakka TA, Tuomilehto J, Salonen JT. Abdominal obesity is associated with increased risk of acute coronary events in men. Eur Heart J 2002; 23: Suter PM, Schutz Y, Jequier E. The effect of ethanol on fat storage in healthy subjects. N England J Med 1992; 326: Hellerstedt WL, Jeffery RW, Murray DM. The association between alcohol intake and the general population. Reviews and commentary. Am J Epidemiol 1990; 132: McDonald I, Debry G, Westerterp K. Alcohol and overweight. In: Verschuren PM (ed). Health issues related to alcohol consumption. ILSI Europ: Brussels; pp Suter PM, Hasler E, Vetter W. Effects of alcohol on energy metabolism and body weight regulation: is alcohol a risk factor for obesity? Nutr Rev 1997; 55: Westerterp KR, Prentice AM, Jequier E. Alcohol and body weight. In: McDonald I (ed). Health issues related to alcohol consumption 2nd edn. ILSI Europe, Brussels; pp Duncan BB, Chambless LE, Schmidt MI, Folsom AR, Szklo M, Crouse III JR, Carpenter MA. Association of the waist-to-hip ratio is different with wine than with beer or hard liquor consumption. Atherosclerosis Risk in Communities Study Investigators. Am J Epidemiol 1995; 142: Suter PM, Maire R, Vetter W. Alcohol consumption: a risk factor for abdominal fat accumulation in men. Addiction Biol 1997; 2: Sakurai Y, Umeda T, Shinchi K, Honjo S, Wakabayashi K, Todoroki I, Nishikawa H, Ogawa S, Kitsurada M. Relation of total and beverage-specific alcohol intake to body mass index and waist-tohip ratio: a study of self-defense officials in Japan. Eur J Epidemiol 1997; 13: Dallongeville J, Marecaux N, Ducimetiere P, Ferrieres J, Arveiler D, Bingham A, Ruidavets JB, Simon C, Amouyel P. 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 1998; 22: Wannamethee SG, Shaper AG. Alcohol, body weight and weight gain in middle-aged men. Am J Clin Nutr 2003; 77: Vadstrup ES, Petersen L, Sorensen TIA, Gronbaek M. Waist circumference in relation to history of amount and typr of alcohol: results from the Copenhagen City heart Study. Int J Obes Realt Metab Disord 2003; 27: Slattery ML, McDonald A, Bild DE, Caan BJ, Hilner JE, Jacobs DR, Liu K. Associations of body fat and its distribution with dietary intake, physical activity, alcohol, and smoking in blacks and whites. Am J Clin Nutr 1992; 55: Shaper AG, Pocock SJ, Walker M, Cohen NM, Wale CJ, Thomson AG. British Regional Heart Study: cardiovascular risk factors in middle-aged men in 24 towns. BMJ 1981; 282: Shaper AG, Wannamethee SG. The J-shaped curve and changes in drinking habit. Novartis Foundation Symp 1998; 216: Bolton-Smith C, Casey CE, Gey GF, Smith WCS, Tunstall Pedoe H. Anti-oxidant vitamin intakes assessed using a food frequency questionnaire: correlation with biochemical status in smokers and non-smokers. Br J Nutr 1991; 65: Yarnell JWG, Fehily AM, Milbank JE, Kubicki AJ, Eastham R, Hayes TM. A short dietary questionnaire for use in an epidemiological survey: comparisons with weighed dietary record. Hum Nutr Appl Nutr 1983; 37: Bolton-Smith C, Milne AC. Food frequency v weighed intake data in Scottish men. Proc Nutr Soc 1991; 50: 35A. 21 Holland B, Welch AA, Unwin ID, Buss DH, Paul AA, Southgate DAT. McCance and Widdowson s the composition of foods 5th edn. Royal Society of Chemistry and MAFF: London; Shaper AG, Wannamethee G, Weatherall R. Physical activity and ischaemic heart disease in middle-aged British men. Br Heart J 1991; 66: Deurenberg P, van der Kooij K, Evers P, Hulshof T. Assessment of body composition by bioelectrical impedance in a population aged 460 years. Am J Clin Nutr 1990; 51: National Institutes of Health. National Heart Lung and Blood Institute. Clinical guidelines on the identification, evaluation and treatment of overweight and obesity in adults: the evidence report. Obes Res 1998; 6 (suppl 2): S51 S Colditz GA, Giovannucci E, Rimm EB, Stampfer MJ, Rosner B, Speizer FE, Gordis E, Willett WC. Alcohol in relation to diet and obesity in women and men. Am J Clin Nutr 1991; 54: Wannamethee G, Shaper AG. Blood lipids: the relationship with alcohol intake, smoking and body weight. J Epidemiol Comm Health 1992; 46: Suter PM, Maire R, Vetter W. Is an increased waist:hip ratio the cause of alcohol-induced hypertension: The AIR94 study. J Hypertens 1995; 13: Laws A, Terry RB, Barrett-Connor E. Behavioural covariates of waist-to-hip ratio in Rancho Bernardo. Am J Public Health 1990; 80: Armellini F, Zamboni M, Frigo L, Mandragona R, Robbi R, Micciolo R, Bosello O. Alcohol consumption, smoking habits and body fat distribution in Italian men and women aged years. Eur J Clin Nutr 1993; 47: Gronbaek M, Deiss A, Sorensen TIA, Becker U, Schnohr P, Jensen G. Mortality associated with moderate intakes of wine beer or spirits. BMJ 1995; 310: Gutierrez-Fisac JL, Rodriguez-Artalejo F, Rodriguez-Blas C, del Rey-Calero J. Alcohol consumption and obesity in the adult population of Spain. J Epidemiol Community Health 1995; 49: Barefoot JC, Gronbaek M, Feagenes JR, Mcpherson RS, Williams RB, Siegler IC. Alcoholic beverage preference, diet and health habits in the UNC Alumni Heart Study. Am J Clin Nutr 2002; 76: Wannamethee SG, Shaper AG. Type of alcoholic drink and risk of major coronary heart disease events and all-cause mortality. Am J Public Health. 1999; 89: Zakhari S. Alcohol and the endocrine system. Research Monograph No. 23, National Institutes of Health, National Institute on Alcohol Abuse and Alcoholism (NIH-NIAAA), Bethesda, MD; pp Kissebah A, Krakower GR. Regional adiposity and morbidity. Physiol Rev 1994; 74: Thomas MC, Walker M, Lennon LT, Thomson AG, Lampe FC, Shaper AG, Whincup PH. Non-attendance at re-examination 20 years after screening in the British Regional Heart Study. JPub Health Med 2002; 24: