|Home | About | Journals | Submit | Contact Us | Français|
Aim: The aim of this study was to demonstrate a methodology for estimating detailed energy intake from alcoholic beverages. Methods: Participants were 315 monthly drinkers who completed a drink-measuring exercise. Energy intake from alcohol and non-alcohol ingredients was calculated for all beverages consumed. Results: Measured alcoholic beverages had on average 140 kilocalories, with 26% of the energy coming from non-alcohol ingredients. The average monthly kilocalorie intake, from all alcoholic beverage types, was 6423 kilocalories. Self-measured wine and spirits drinks contained more energy than reference standards for size and ethanol concentration. Conclusions: Amount and sources of kilocalories differ by drink type, gender, age, education and BMI. Researchers and consumers should be aware of this variation and its sources.
Obesity and being overweight are important public health problems, significant in both scale and consequence. In the United States, data from the 2003–2004 National Health and Nutrition Examination Survey signify a nearly one-third increase in the age-adjusted prevalence of obesity since the 1988–1994 collection period (Ogden et al., 2006). The upward trend in obesity in the United States, however, has not been associated with a similar rise in alcohol consumption that peaked in 1981 at 10.94 L of ethanol per capita aged 15 and older (Kerr et al., 2006) and has since declined with consumption dropping to 8.18 L per capita by 1993, rising slightly to 8.23 L by 2002.
Despite the non-correspondence of trends, there are several reasons why alcohol consumption merits careful investigation in the etiology of obesity. First, alcohol is an energy-dense, yet nutritionally poor, food source with an energy content of 7.1 kilocalories per gram and as a food group is the ninth largest contributor to daily energy intake in the US diet (Block, 2004). Second, total energy can vary considerably by alcoholic beverage type. While a 12 ounce (355 mL) bottle of light beer contains ~103 kilocalories and 11 g of ethanol, a 4.5 ounce (133 mL) Pina Colada contains considerably more ethanol and energy with 245 kilocalories and 14 g of ethanol (US Department of Agriculture, 2005). The difference between these two beverages also emphasizes that non-alcohol ingredients can contribute a sizable number of kilocalories to some drinks. And third, quantification of energy and sources of energy intake from alcoholic beverages to date has been limited by imprecise kilocalorie and alcoholic beverage intake estimation with studies typically employing relatively simplistic food frequency or quantity-frequency measures. In these studies, a U-shaped curve typically describes the relationship between volume of alcohol consumed and obesity (Arif and Roher, 2006), although results have not been consistent (Männistö et al., 1997; Kleiner et al., 2004). Recent studies have begun to consider alcohol intake patterns, finding lower body mass index among frequent low quantity drinkers compared to those who drink larger amounts less frequently (Breslow and Smothers, 2005).
This study explores the contribution to energy intake from alcoholic beverages based on detailed measures of drink size and contents including beaker measures of subjects’ usual drinks in their own glassware. Kilocalorie and alcohol contents of drink ingredients were identified by specific brand for 98% of beer drinks, 90% of spirits drinks and 68% of wine drinks, and were identified by type in nearly all other cases.
Participants (N = 315) were drawn from a sub-sample of 751 monthly drinkers in the 2000 National Alcohol Survey study (Midanik and Greenfield, 2003) who agreed to be re-contacted and, 3 years later, agreed to participate (43%) in a comprehensive study of drink sizes. A detailed description of this sample has been described previously (Kerr et al., 2005). Human subjects’ approval for this study was granted by the Institutional Review Board of the Public Health Institute.
Participant characteristics collected in the original 2000 National Alcohol Survey include self-reported weight and height and educational attainment. Gender, age and ethnicity were collected in 2000 and 2003, with the latter collection period used for these analyses. Educational attainment was categorized as having completed high school or less, having attended some years of college and having graduated college or more. The age of participants at the time of the follow-up survey was divided into three age groups; 18–34 years, 35–54 years, and 55 years and older. Body mass index was calculated from self-reported height and weight as
World Health Organization categories were used to create the cut points for categorizing BMI into the normal or underweight (BMI ≤ 24.9), overweight (BMI ≥ 25 and ≤29.9) and obese (BMI ≥ 30) groups used in all analyses (World Health Organization, 1995).
In earlier work, the authors reported that respondents listed their usual brands and types of alcohol and used a provided measuring beaker to report the amount of each drink ingredient during an exercise in which they poured these drinks into their own glassware (Kerr et al., 2004). Drinks and individual ingredients were reported in fluid ounces. In the current study, we built on this detailed alcohol intake assessment to assign specific kilocalorie content to respondents’ wine, beer and spirits drinks. Kilocalories attributed to the ethanol portion of each beverage were calculated directly from the percentage alcohol by volume (%ABV) as
where ‘drink size’ is equal to the amount of wine, beer or spirits poured.
To calculate total kilocalories, beer data were collected from the producer or one of several online beer aficionado websites that contain databases listing beer brands, kilocalories and %ABV. For total wine kilocalories, each wine in the sample was assigned the one of 30 wine types or varietals in the USDA database (US Department of Agriculture, 2005) that most closely fit the respondent's given wine brand or type. For each of the USDA wine types, a ratio of alcohol kilocalories to total kilocalories was calculated and then applied to the respondent-specific %ABV of the wines in our sample to get a more accurate total kilocalorie estimate.
Kilocalorie estimates for spirits-based drinks were more complicated due to the number of different ways drinks can be prepared and to the variety of possible ingredients among various types of liquors. (A detailed description of these methods is available from the authors.) In brief, alcohol kilocalories were calculated as described above. To calculate non-alcohol kilocalories, mixer size was multiplied by kilocalories per ounce for each mixer ingredient, obtained from the USDA Nutrient Database. Additionally, certain liquors contain a significant number of kilocalories in addition to their alcohol kilocalories. These ‘extra’ kilocalories were calculated by difference (total kilocalories − alcohol kilocalories.) Total kilocalorie content was obtained from an online spirits aficionado website with brand-level nutritional information. Spirits drinks have two potential sources of non-alcohol kilocalories, mixer and non-alcohol liquor ingredients. In the analysis, both sources were combined and described as non-alcohol ingredients or kilocalories.
Alcohol and non-alcohol kilocalories per drink were multiplied by each respondent's beverage-specific usual quantity to obtain the number of kilocalories per occasion, and by their monthly volume to obtain monthly kilocalorie intake. Monthly volume of each beverage was obtained from beverage-specific graduated frequency series where respondents were asked to report the frequency of consuming each beverage type and the proportion of drinking days on which they consumed one to two, three to four, and five or more drinks. For those reporting having five or more drinks, the average number of drinks on these days was calculated from a respondent's report of how frequently they consumed 5–7, 8–11, and 12 or more drinks of any type of alcohol. A drink average was calculated as each respondent's average number of kilocalories, weighted by the relative proportion of wine, beer and/or spirits monthly volume. Respondents were classified as ‘Heavy’ drinkers if their reported monthly volume was 80 drinks or more, representing the top 10% of drinkers in the 2000 National Alcohol Survey. They were classified as ‘Light’ drinkers if their monthly volume was ≤30 drinks and ‘Moderate’ drinkers if their monthly volume fell between 30 and 80 drinks. The alcohol consumption distribution is highly concentrated such that the top 10% of drinkers in the 2000 NAS reported drinking 55.3% of the total amount of alcohol (Kerr and Greenfield, 2007) and therefore represent a key subgroup who are expected to consume larger and more concentrated drinks.
In this sample of self-poured and measured beverages, respondents reported consuming a variety of drink types and sizes, both factors affecting the number of kilocalories consumed in a drink. Wine varietals, beer types and cocktails can have sizably different nutrient profiles and brand differences in %ABV or in non-alcohol ingredients can be considerable. As a result, the total energy consumed in a drink bears a wide distribution and differs by drink type (Fig. (Fig.11.)
Spirits-based drinks are found to have a higher average number of total kilocalories per drink than wine and beer (Table (Table1),1), containing 211 kilocalories, 64 (30%) of which are from non-alcohol ingredients. When respondent details relating to drink size and composition are applied to the usual number of drinks consumed during a drinking occasion, the mean total kilocalories from spirits drinks is again higher than wine and beer drinks (Table (Table1.)1.) When drink size and content are applied to average monthly volume of wine, beer or spirits drinks, beer, not spirits drinks, contribute more total monthly kilocalories and 28% of the nearly 3457 total kilocalories come from the non-alcohol ingredients in beer.
Kilocalories consumed per drink, occasion and month differed by drink type according to gender, age, educational attainment and body mass index category (Table (Table1.)1.) The largest gender-related difference in energy intake is related to monthly volume, primarily from beer consumption. Men consume almost five times more beer kilocalories per month than women. Women's individual drinks, on the other hand, contain more total energy with a higher percentage of kilocalories coming from mixer or added ingredients. While this trend was consistent across all beverages, this gender difference reached statistical significance only for spirits drinks; 41% of the 238 total kilocalories in the average spirits drink for women come from non-alcohol ingredients, more than twice as many as men's drinks [97.5 kilocalories (95% confidence interval: 56, 139) versus 43 kilocalories (95% confidence interval: 6, 31); P = 0.01]. Many respondents drink more than one type of beverage, and when all beverage types are combined, men consume significantly more kilocalories each month than women do: 8210 (95% confidence interval: 6403, 10,018) versus 3790 (95% confidence interval: 2498, 5081; P < 0.01) kilocalories.
Obese respondents consume drinks with higher caloric content than normal weight individuals, except for beer drinks (Table (Table1),1), and consume more kilocalories from non-alcohol ingredients in spirits drinks compared to normal weight respondents, obese: 102 kilocalories (95% confidence interval: 51, 154) versus normal weight: 40 kilocalories (95% confidence interval: 27, 52); P = 0.02. For the drink average, this difference is marginally significant: obese: 51 kilocalories (95% confidence interval: 31, 70) versus normal weight: 30 kilocalories (95% confidence interval: 24, 37), P = 0.055 but suggestive of a similar trend. Although kilocalorie content per average drink is higher for obese respondents than for normal weight respondents, aggregate total kilocalories per month is highest for normal weight respondents; overweight and obese respondents consumed ~2500 kilocalories less per month (not statistically significant). This pattern differs slightly between women and men in the total number of kilocalories consumed and the percentage of alcohol and non-alcohol kilocalorie sources (Fig. (Fig.2).2). Normal or underweight men (n = 57) consume more energy than overweight men (n = 72): 11,228 kilocalories (95% confidence interval: 7154, 15,202) versus 5785 kilocalories (95% confidence interval: 3870, 7700), P = 0.02, and obese women (n = 34) consume less energy than other women, with a higher percentage of kilocalories coming from non-alcohol sources (not statistically significant).
Heavy drinkers (n = 44) predictably consume more energy per month from alcohol than light (n=177) or moderate drinkers (n = 94) [light: 1623 kilocalories (95% confidence interval: 1332, 1915), moderate: 7256 (95% confidence interval: 6341, 8171), heavy: 22,482 (95% confidence interval: 16,791, 28,173), P < 0.01 when compared to light and P < 0.01 when compared to moderate drinkers]; most of these kilocalories come from alcohol, not other ingredients. Interestingly, this pattern is also true of individual drinks (Fig. (Fig.3).3). Heavy drinkers consume a lower proportion of other kilocalories to total kilocalories compared with other drinkers, while total kilocalories contained in the average drink were nearly the same for all drinkers. Heavy drinkers consume 19.9% (95% confidence interval: 16.6, 23.3) and light drinkers consume 25.4% (95% confidence interval: 22.8, 27.9) of 138 total kilocalories (both) from non-alcohol ingredients (P = 0.01).
Body mass index is lower for heavy drinkers compared to both light drinkers and moderate drinkers [light: 27.2 (95% confidence interval: 25.7, 28.7), moderate: 27.5 (95% confidence interval: 25.2, 29.8), heavy: 22.9 (95% confidence interval: 19.8, 26.1), P = 0.02 compared to light drinkers and to moderate drinkers] for women. A similar trend is seen for men but does not reach statistical significance. Results should be interpreted with caution, however, due to the small number of women (n = 4) in the heavy drinker category.
The number of kilocalories consumed per drink and the total number of kilocalories consumed over a month varied also by beverage sub-type (Fig. (Fig.4).4). Light beer has the lowest kilocalorie profile on a per drink basis but light beer drinkers drink more per occasion than wine drinkers or straight liquor drinkers do and a higher percentage of their kilocalories come from the non-alcohol ingredients in beers. Although cocktails with mixer ingredients have the most kilocalories per drink, individuals who report of drinking mixed drinks do not have the greatest total monthly energy intake compared to other drinkers. Spirits drinks made without mixers (either straight, on the rocks or mixed with other spirits) and regular beers contribute the most kilocalories per month while wines, both red and white, contribute the fewest.
In this US national sample of regular drinkers, the average total energy per drink was found to be 140 kilocalories, 103 from alcohol and the remaining 37 from non-alcohol ingredients. As illustrated in Fig. Fig.1,1, this average summarizes a wide variety of individual drinks including as few as 20 kilocalories for a very small drink to over 500 kilocalories for drinks with larger amounts of alcohol and kilocalories from other sources. This variability highlights the importance of obtaining specific information in alcohol surveys from respondents about each type of drink they consume such as the brand, volume and any other ingredients included in a drink. Obtaining this information may be most important for spirits drinks that vary the most in alcohol content and additional kilocalories. Beer and wine kilocalorie variation, while less complex, also fluctuates between brands and with %ABV and pour size. Beer drinks are the most standardized since they are typically purchased in single-serve containers while wine and spirits are poured from large bottles allowing more individual choice and variation.
The heavy drinkers in our sample were found to have the fewest kilocalories from non-alcohol sources but at the same time consumed the most kilocalories per month from alcoholic beverages, an average of 22,482 kilocalories. On a daily basis, this would amount to 749 kilocalories, 37.5% of a reference 2000 kilocalorie daily energy intake. Specific US guidelines for daily energy intake depend on age, activity level and gender. If alcohol beverage kilocalories are consumed in excess of daily energy needs, the light, moderate and heavy drinkers in this study would theoretically gain an excess 0.5, 2.1 and 6.4 pounds of body weight per month, respectively. The message that alcohol and kilocalories are inextricably linked, with one ounce (29.6 mL) of alcohol containing 166 kilocalories, may help focus attention on limiting alcohol intake when one is trying to limit kilocalories. In the United States, alcohol concentration labeling is required only for spirits and kilocalorie or other nutritional information is rarely provided. This information would help consumers to make informed decisions regarding alcoholic beverage intake and would be useful to researchers trying to quantify alcohol and kilocalorie intake (Stockwell and Single, 1997).
This study revealed a number of important sources of variation in the kilocalorie content of drinks and overall kilocalorie intake from alcoholic beverages. Men consume more kilocalories from beer and from all alcoholic beverages over a month due to their much larger volume of drinks. Individuals 55 years of age and older are found to consume fewer kilocalories per drink overall and for each beverage type with significant differences seen for wine and the drink average. This may be due to a reduced preference for sweet tastes with increasing age. Heavy drinkers were found to have about the same number of kilocalories per drink as other drinkers but to have more kilocalories coming from alcohol rather than other ingredients. Heavy drinkers may choose stronger drinks because of their reduced sensitivity to the taste of alcohol. Wine and spirits drinks’ kilocalories are found to be related to body mass index with obese individuals consuming more caloric wine and spirits drinks. This difference does not pertain to monthly kilocalorie intake where normal or underweight individuals actually have the highest intake. This finding seems to fit the expectation that obese individuals would have larger portion sizes and choose sweeter spirits drinks while also confirming a common finding in the literature that higher alcohol intake does not predict obesity or higher body mass index (Männistö et al., 1997; Kleiner et al., 2004).
Frequently in the United States, a standard drink ethanol content of 14 g is used by researchers to estimate total alcohol consumption and is equivalent to 99.5 kilocalories. The number of alcohol kilocalories in a respondent's average drink (that is the average across all drinks) found in this study is similar (103 kilocalories), but beverage-specific averages range from 90 kilocalories for beer, 110 for wine, and 146 for spirits drinks. Alternatively, researchers may use beverage-specific standards to estimate total ethanol or energy intake using a nutrient database (US Department of Agriculture, 2005). The USDA Nutrient Database values for wine, beer and liquor are 124 kilocalories and 15.4 g ethanol for a 5-ounce (148 mL) table wine, 153 kilocalories and 13.9g ethanol for a 12-ounce (355 mL) regular beer, and 97 kilocalories and 14 g ethanol for a 1.5-ounce (44 mL) 40% ABV liquor drink. Compared with the total kilocalories in respondent-measured drinks in this study, the USDA nutrient values underestimate wine, overestimate beer and underestimate spirits drinks by half, in part due to a difference in drink size. Average drink sizes in the current study are 18% and 47% higher than standard wine and spirits drink sizes, respectively. It is important to note that the standard for a spirits drink represents a shot glass of 40% ABV alcohol. There is no standard for a mixed drink; therefore, the standard spirits drink fails to include any kilocalories attributable to mixer ingredients. Regardless of the lack of comparison for a mixed drink, participants consumed considerably more alcohol kilocalories than the standard spirits drink whether they reported drinking a mixed drink (143 kilocalories) or a straight drink (159 kilocalories).
This study relied upon self-reported weight and height measures, which may result in an underestimation of overweight and obesity than if measured weights and heights were available (Rowland, 1990; Kuczmarski et al., 2001). Studies have suggested a greater influence from self-reported height on measured versus self-reported body mass index than did self-reported weight and this difference was notable for the older age groups (60+ years). In the current study, 13% of participants were 60 years or older. This study population does not include sizeable numbers of ethnic sub-groups other than whites and our ability to describe drink kilocalories and to make comparisons among sub-groups is limited. Only the respondent's usual drink of each beverage type was measured and other drinks may differ in kilocalorie content. Some drinkers may ‘top up’ their beer, wine or spirits drinks, which reduces the usefulness of the drink concept for measurement of kilocalorie intake. Due to sample size limitations, the current study did not consider several factors known to influence alcohol consumption and body mass index, such as depression or mental health status, physical activity, dietary intake and smoking, among others. The results of this study therefore are limited to describing energy intake and sources of energy from alcoholic beverages according to basic participant characteristics. Nevertheless, this description adds important information to the fields of alcohol and nutritional epidemiology in providing improved measures of realistic drink content. It will be valuable to repeat this work with a larger sample of respondents so that questions of alcohol consumption related to energy balance and body weight may be explored in greater detail.
This research was supported by grant P30-AA05595 to the Alcohol Research Group, Public Health Institute, from the National Institute on Alcohol Abuse and Alcoholism (NIAAA).