Public Health Nutrition: 22(10), 1727–1734 doi:10.1017/S1368980019000442 How full is your glass? Portion sizes of wine, fortified wine and straight spirits at home in the Netherlands Myrthe FA de Beukelaar1, Marion L Janse2, Aafje Sierksma2, Edith JM Feskens1 and Jeanne HM de Vries1,* 1Division of Human Nutrition and Health, Wageningen University and Research, PO Box 17, 6700 AA Wageningen, The Netherlands: 2The Dutch Beer Institute, Wageningen, The Netherlands Submitted 7 May 2018: Final revision received 23 November 2018: Accepted 8 January 2019: First published online 26 March 2019 Abstract Objective: Alcohol consumption may be wrongly estimated because of inaccurate information on actual portion sizes. We compared portion sizes of wine, fortified wine and straight spirits poured at home with the Dutch standard drink sizes. Design: Participants measured portion sizes of wine, fortified wine and straight spirits at home up to a maximum of three times and reported these via an online survey. Average portion sizes (in millilitres) were compared with the Dutch standard drink sizes. Portion sizes were compared between subgroups of gender, age, BMI and level of education, and for different glass types. Setting: Wageningen and surroundings, the Netherlands. Participants: Adults (N 201) living in the Netherlands and consuming wine and/or straight spirits at home at least once per week. Results: Participants poured on average 129·4 ml white wine and 131·7 ml red fi fi Keywords wine, which is signi cantly more than the standard of 100 ml. For forti ed wine, Portion size · fi the average poured amount was 94 0 ml, signi cantly more than the standard of Alcohol 50 ml; also for straight spirits the poured amount was significantly more than the Wine standard (47·0 v. 35 ml). Spirits Conclusions: Participants’ portion sizes of wine, fortified wine and straight spirits Home poured at home were on average larger than the Dutch standard drink sizes. This Standard drink suggests that at-home alcohol consumption in the Netherlands is underestimated. The Netherlands The alcohol guideline of the Dutch Health Council states: sizes are used for dietary monitoring and to make ‘do not drink alcohol, or at least no more than one glass recommendations about alcohol consumption in relation per day’(1). In 2015, it was estimated that on average 45 % to health. Furthermore, most surveys assessing alcohol of the total Dutch population aged 19–79 years adhered to levels, such as the Dutch National Food Consumption this guideline; approximately 36 % of the men and 53 % of Survey, rely on the assumption that respondents use these the women drank no more than one glass of alcohol standard drink sizes. However, knowledge of the term per day on average(2). ‘standard drink’ by respondents is shown to be poor in – Accurate estimates of energy and nutrient intakes of self-reports of alcohol consumption(5 7). As a result, the individuals and populations rely on information obtained amount of self-reported drinks (given in standard servings) about dietary intake, food composition and portion size(3). might not match the actual amount of alcohol consumed, Portion sizes of alcoholic beverages are often estimated in leading to misreporting of alcohol consumption in the terms of standard drinks. In the Netherlands, one standard Netherlands. drink of alcohol corresponds to 10 g (12 ml) alcohol, To date, most of the research on the practice of pouring which is approximately the amount of alcohol in 250 ml alcoholic beverages has focused on drink size variation in beer (5 % alcohol), 100 ml wine (12 % alcohol), 50 ml for- relation to pouring on-premises. Across these studies, tified wine (15 % alcohol) and 35 ml straight spirits (35 % drinks typically contained greater volumes of alcohol than – alcohol)(1). Noteworthy is that The Netherlands Nutrition one standard drink(8 11), even when alcoholic beverages Centre recently published different drink sizes based on were poured by professional bartenders(10). This effect consumer practices: 150 ml for wine, 75 ml for fortified was demonstrated in wine and in (mixed) straight spirits, wine and 50 ml for straight spirits(4). The standard drink whereas volumes of alcohol in beer and shots (unmixed *Corresponding author: Email [email protected] © The Authors 2019 Downloaded from https://www.cambridge.org/core. 29 Sep 2021 at 09:11:08, subject to the Cambridge Core terms of use. 1728 MFA de Beukelaar et al. spirits) were relatively similar to standard drink volumes. supermarket and at a local public event, both in Wagen- Fewer studies have investigated drink size variation in ingen. Inclusion criteria were a minimum age of 18 years relation to self-serving off-premises. In a study in which old, living in the Netherlands, wine and/or straight spirits the mean alcohol content of beer, wine and spirits was consumption at home at least once per week, and willing estimated in a national sample of US drinkers, a weighted to measure wine and/or straight spirits poured at home on mean alcohol content of 15·6 g overall, 13·1 g for beer, several days. No ethical approval was required for the 15·4 g for wine and 20·8 g for spirits was revealed(12), present study. The Medical Ethics Committee of Wagen- suggesting that the US alcohol drink standard (14 g alcohol ingen University concluded that this research does not fall per drink) underestimates the average alcohol content in within the remit of the ‘Medical Research Involving glasses of wine and spirits poured at home. In Europe, a Human Subjects Act’, because participants are not subject study in a Scottish population showed a considerable to procedures or required to follow rules of behaviour. In variation in the amount of wine or spirits that was poured. addition, it does not concern medical scientific research. On average, the amount poured equalled approximately All participants declared to have read and understood the two UK standard units instead of one (8 g per drink)(13).In goal of the study. 65- to 74-year-olds from Western Australia, larger volumes of wine and spirits were poured in comparison to a stan- Procedure dard drink (10 g); 32 % more for men and 16 % more for Participants of the study were asked to measure the por- women(14). A study conducted in 1994 among Dutch tion size (in millilitres) of wine, fortified wine and straight participants found that drink sizes varied off-premises and spirits in glasses they normally use at home. They received that they were on average larger than a standard drink(15). instructions by email, supported by a short video on how The deviation from the presumed standard (10 g per drink) to measure the content of a glass. After pouring their drink was highest for spirits (+26 %), followed by fortified wine into their usual glass, participants used a measuring cup (+14 %) and least for wine (+4 %)(15). Thus, errors in drink for the measurements, which was provided by the size likely contribute to the under-reporting of alcohol researchers. The total duration of the study was different consumption. This is specifically relevant for beverages for each participant, as they could choose when to mea- that do not come in drink-size containers(16), since not all sure their alcoholic beverages within a period of two individuals use the same type of glassware to pour alco- months (August and September 2017). Participants were holic beverages at home. In fact, elongation of glasses has requested not to change their habits of alcoholic beverage been shown to influence how much alcohol people pour: pouring and drinking. tall, slender glasses lessen the tendency and short, wide Participants filled out an online survey created in glasses increase the tendency to overpour(10,17). LimeSurvey(19) asking for the following information: the The last study in the Netherlands focusing on portion date of measurement, type of beverage, portion size in sizes of alcoholic beverages off-premises was more than millilitres, alcohol content of the beverage indicated on the 20 years ago(15). The aim of the present study was to bottle (volume percentage) and the type of glass (relevant provide an update of the portion sizes of wine and straight number or letter) in which they had poured their drink spirits poured at home and compare them with the Dutch (see Fig. 1). Participants placed wine in one of the fol- standard drink sizes and the drink sizes based on con- lowing categories: white wine, red wine, rosé, sparkling sumer practices. In addition, the study evaluated whether wine, dessert wine, port, sherry and other. The type of gender, age, BMI and level of education, and different straight spirits was identified with an open question. glass types have an association with poured portion sizes. Measurements of the same types of beverages were repe- Beer and pre-mixed drinks were excluded because, at ated up to a maximum of three times, each measurement home, consumers mostly use standard units such as bot- made on a different day. In addition, gender, age, height tles and cans of which the contents are known. and weight (in order to calculate BMI) and highest level of education completed (elementary, secondary, higher pro- fessional or scientific education) were asked to compare Methods portion sizes between specific subgroups. Study design and participants The present study was a cross-sectional observational Data analysis study, designed to obtain information on the portion sizes All data were analysed in the statistical software package of wine, fortified wine and straight spirits poured at home IBM SPSS Statistics for Apple Mac version 24. The average in a sample of adults living in the Netherlands. Between portion sizes (in millilitres) of white wine, red wine, rosé, July and September 2017, adult men and women were other wine (sparkling wine and dessert wine), fortified recruited: (i) from the ‘EetMeetWeet’ research panel(18); (ii) wine (port and sherry) and straight spirits poured at home via a call on social media (Facebook, LinkedIn); and (iii) were calculated per participant.