The association between alcohol intake and obesity in a sample of the Irish adult population, a cross-sectional study

The population in this study comprised those aged 25 years and older who had participated in Wave three of the Healthy Ireland Survey 2017. This survey was part of the Healthy Ireland Framework 2013–2025 that aims to improve overall wellbeing and reduce inequality within the population in Ireland [21]. Access to the anonymized, secondary data was requested from and permitted by the Irish Social Science Data Archive (ISSDA).

The sample of the parent study was selected using a two-stage probability-based methodology with the aid of An Post Ordnance Survey Ireland geographic data [21]. The sample was determined based on electronic division clusters (each containing less than 500 addresses) to ensure coverage of a wide geographical area. Stratification by region was performed. A total of 686 clusters were selected. The addresses of those who took part in previous waves of the Healthy Ireland Survey were excluded from the third wave. Within each cluster, 20 addresses were systematically selected, with a random start address and fixed interval skips. One household member was selected randomly by interviewers to take part in this survey. The survey was conducted face-to-face by trained interviewers. Verbal consent was obtained from participants aged 18 years and older, and the parents/guardians of participants under 18 years of age provided written consent before participation in the survey. All interviews were performed at participants’ homes using computer-assisted personal interviewing (CAPI). Physical measurements (weight (kg), height (cm), and waist circumference (cm)) were measured by trained staff. Out of 12389 eligible addresses preselected to take part in the survey, 7487 households aged 15 and older completed the interview, for an overall response rate of 60.4%. A total of 5868 (78%) participants completed the physical measurement examination [21]. For the purpose of this study only those aged 25 and older were included in this study.

This was an observational, analytical, cross-sectional study where both descriptive and analytical data were presented.

This study was based on secondary data from wave three of the Healthy Ireland Survey 2017. The parent survey was approved by the Research Ethics Committee of the Royal College of Physicians of Ireland (RCPI) on September 18, 2014 [19]. For the current study, a request for exemption from full ethical approval was obtained from the Research Ethics Committee of the School of Public Health, Physiotherapy, and Sport Science at University College Dublin on 25/2/2021.

Out of 134 variables presented in the original Healthy Ireland Survey 2017 dataset, twenty-three variables were included in this study. These data comprised sociodemographic data, health-related data, alcohol-related data, and physical measurement data.

Independent variables included alcohol-related variables, which were obtained via face-to-face interview. In this study, participants were classified into two categories: “drinkers” and “non-drinkers.” Drinkers were defined as those who responded “Yes” to the question “Have you ever consumed an alcoholic beverage in your lifetime?” while non-drinkers were defined as those who responded, “I have never had a drink” or only “I drank a few sips of an alcoholic beverage in the past.” For the alcohol frequency question, "How frequently did you consume alcohol in the last 12 months?" and for those who reported drinking alcohol, alcohol frequency was categorized as "less than three drinks per week" and "three or more drinks per week." Binge drinking was defined as the consumption of six or more standard units per occasion for both sexes. According to the response to the question “During the last 12 months, how often did you consume the equivalent of six standard drinks on one occasion?”, and among those who reported consuming alcohol, binge drinking was categorized as “less than one occasion of binge drinking per week” and “one or more occasions of binge drinking per week”.

The Alcohol Use Disorder Identification Test-Consumption, AUDIT-C, tool was used to screen for harmful alcohol consumption [20]. Three questions included in the AUDIT-C questionnaire were as follows: “How often did you have a drink containing alcohol in the past year?”, “How many drinks containing alcohol did you have on a typical day when you were drinking in the past year?”, and “How often did you have six or more drinks on one occasion in the past year?” The results of these three questions were given scores of 0–2, 3–4, or 5 or higher. A score of 5 or higher was considered harmful drinking [14]. For the regression analysis, binary variables of harmful alcohol consumption (score < 5 and score ≥ 5) were created.

The dependent variables of interest in this study were WC and BMI, which were measured by trained staff as per standard protocol. Waist circumference was considered a continuous variable and was measured in centimetres. BMI is expressed as weight (kg) divided by height squared (m²). BMI was further categorized into three categories: normal weight (< 25. 0 kg/m²), overweight (25.0–29.9 kg/m²) and obesity (≥ 30.0 kg/m²). For the multivariable logistic regression analysis, a binary variable of BMI (< 25.0 kg/m² and ≥ 25.0 kg/m²) was created. The WC was used as a continuous variable in the multivariable linear regression analysis.

Sociodemographic data (including age, sex, marital status, level of education, employment status, urban or rural residency, having a full medical card, and having a private medical card), and health-related data (including general health condition, long-term illness, smoking status, active transportation, and frequency of fruit consumption) were examined to analyse their associations with obesity indicators. For regression analysis, binary variables were created for sociodemographic and health related variables. Answers that were recorded as “don’t know” or data missing from the original dataset were treated as missing variables and thus excluded from the analysis.

For descriptive analysis, the categorical variables were presented as numbers and percentages (n/%). Anthropometric measurements (BMI, WC) were presented as means (standard deviations, SDs) and medians (range). BMI is further presented in categories and thus expressed in numbers and percentages.

Univariate analyses were performed to examine the associations between sociodemographic data, health-related data, alcohol-related data, and obesity indicators (BMI, WC). The means of continuous variables were compared using the independent Student’s t test. The assumption of normality, homogeneity of variance, and absence of outliers were satisfied. The difference in percentages between groups was compared using the Pearson chi-square (χ²) test. Variables with a p value less than 0.05 in the univariate analysis were retained and included in the multivariable analysis.

Multivariable linear regression analysis was performed to analyse the association between harmful alcohol consumption (AUDIT-C score ≥ 5) and WC while controlling for possible confounders. Four different Models were constructed in this regression analysis; with confounding variables (sociodemographic and health-related variables) with significant p values (p < 0.05) were retained and included in the subsequent Model. Model 1 included all variables, while Model 2 was adjusted for sociodemographic data. Model 3 was adjusted for the statistically significant variables in Model 2, along with health-related data (general health, long-term medical illness, smoking status, active travel, and fruit consumption). The fully adjusted Model (Model 4) included the statistically significant variables in Model 3, along with alcohol-related variables (alcohol frequency and binge drinking). The results are presented as linear regression coefficients (β) and 95% confidence intervals (95% CIs). Similarly, a multivariable logistic regression analysis was performed to assess the association between harmful alcohol consumption and BMI using the different Models described earlier. The results are presented as adjusted odds ratios (ORs) and 95% CIs. The Windows-based statistical package (SPSS version 24) was used to perform the analysis. Two-tailed tests were used, and a p value of < 0. 05 was considered statistically significant.

The current study showed positive associations between harmful alcohol consumption (AUDIT-C score ≥ 5) and obesity indicators (both BMI and WC) after controlling for sociodemographic and health related variables. After controlling for binge drinking and alcohol consumption frequency, harmful alcohol consumption continued to be significantly associated with WC and overweight/obesity (BMI ≥ 25 kg/m²). Several studies on obesity looked at either the frequency [22] or quantity [9, 11, 12, 17, 18] of alcohol consumed, or both [10, 13], and revealed mixed results. No other study has employed the AUDIT-C questionnaire to investigate the association between alcohol consumption and obesity. Using the AUDIT-C can help eliminate the inconsistencies in the evaluation of the association between alcohol use and obesity.

The exact underlying mechanism through which alcohol consumption is associated with obesity is not fully understood. According to the existing literature, alcohol consumption is the second-highest source of energy, providing almost 29 kJ (7 kcal) of energy [23]. It has been found to have an additive effect on energy obtained from other non-alcoholic sources, which can lead to weight gain [24]. Some studies have shown that alcohol acts by increasing hunger [25‐27], however, other studies have shown no significant association [28, 29]. Additionally, alcohol is a suppressant of fat oxidation, favours lipid storage, and acts as a precursor for fat synthesis [30, 31]. Furthermore, alcohol intake enhances cortisol secretion, which subsequently affects the distribution pattern of fat in the body [32]. The association between types of alcoholic beverages and obesity is controversial, with some studies showing a positive association [9, 15] and others showing an inverse association [9]. Certain types of alcoholic beverages, such as beer, have higher amounts of carbohydrates per unit of alcohol than others and thereby contribute to an increased risk of obesity [15]. Binge drinking has a catabolic effect on muscle tissue, leading to fat deposition in visceral organs and muscles [33], which further leads to the development of various noncommunicable diseases, including obesity and diabetes mellitus.

The current study showed an inverse association between frequency of alcohol consumption and overweight/obesity (BMI ≥ 25.0 kg/m²); with frequent alcohol consumers (≥ 3 times a week) being 49% less likely to be affected by overweight/obesity. However, no significant association was found between frequency of alcohol consumption and WC. Previous studies examined the association between alcohol frequency and obesity, and the results were discordant [6, 10, 11, 13, 22]. Some studies found inverse associations between the frequency of alcohol consumption (using a one-year recall period) and obesity indicators (BMI and WC) in both sexes [10, 11, 22]. A study in Scotland and England, on the other hand, found a bell-shaped association between the frequency of alcohol consumption (using a 7-day recall period) and obesity indicators (BMI, WHR), with no difference in risk observed between those who never drank and those who drank the most frequently [6]. Another study, which used a 24-hour recall period, on Korean adults aged 40 to 60, found no significant association between the frequency of alcohol consumption and obesity (WC) [13]. However, this lack of an association could be related to the lower prevalence of abdominal obesity in the study population, < 8% in both sexes; thus, an exact association might be difficult to obtain [13].

Several factors could explain the inverse association between alcohol frequency and obesity. First, alcohol intake might affect macronutrient absorption, leading to reduced energy intake [34]. Furthermore, alcohol consumption may stimulate thermogenesis by activating the ethanol oxidizing system, which can result in weight loss [34]. This effect may be balanced by the various metabolic changes caused by alcohol consumption as well as the additional energy obtained from alcohol consumption [10]. As a result, alcohol consumption frequency may not be directly associated with body weight or fat accumulation; rather, high alcohol consumption above the sex threshold may have an effect on abdominal obesity. Overall, based on existing research, encouraging alcohol use to lower the risk of obesity is not recommended, given that the precise mechanism is not fully understood.

Nearly one-third (32.1%) of the study population engaged in binge drinking. Univariate analysis revealed a significant association between binge drinking and obesity indicators (both BMI and WC). After controlling for sociodemographic, health-related, and other alcohol-related variables, the results revealed that frequent binge drinking (≥ 1 time per week) was significantly associated with a larger mean WC than less frequent binge drinking. However, no significant association was found between binge drinking and overweight/ obesity. Several studies have found a positive association between binge drinking and obesity, with individuals engaged in binge drinking being more likely to have a larger waist circumference [13] or suffer from obesity (BMI) [34]. However, a study conducted in the US using self-reported data on adults aged 18 with low income status did not find significant association, possibly due to the small sample size [22]. The reason for the positive association between binge drinking and obesity could be attributed to the presence of other impulsive behaviours, e.g., binge eating or abnormal eating patterns, which could confound the association between binge drinking and obesity [4].

Several factors could have contributed to the inconsistent results of the alcohol-obesity association. First, the underlying cause of obesity is complex, with several factors contributing to its development [1]; however, not all of these characteristics have been researched or controlled for when assessing the association between alcohol consumption and obesity, which may bias the estimated association. The current study controlled for sociodemographic and health-related variables that could potentially confound the association between alcohol consumption and obesity. However, it did not investigate individual consumption patterns or the cultural dynamics that drive alcohol consumption, or the forms of legislation that could influence alcohol consumption. Second, epidemiological studies have examined the association between alcohol consumption and obesity using different instruments [9, 11, 12, 17, 18, 22], making comparisons among studies difficult as no single well-established method was used to measure alcohol consumption. The window period used to recall alcohol intake varied between studies, ranging from a 24-hour recall period [13, 18] to a 7-day recall period [6, 9], to a 12-month recall period [10]. However, a short-term recall period does not account for the usual drinking patterns of individuals, the context in which individuals drink, or the influence of different seasons on alcohol consumption. The current study used a 12-month recall period to examine the association between alcohol consumption and obesity, which is considered more accurate in measuring the usual consumption trends in individuals. Third, the baseline prevalence of both alcohol consumption and obesity in a study population can also contribute to inconsistencies in the results obtained when analysing this association. Fourth, the types of alcoholic beverages might also affect the life habits of individuals, including eating patterns and physical activity, subsequently affecting body weight. Those who consume beer tend to have worse dietary habits than those who consume other beverages [35]. A US study showed that individuals who drank beer tended to eat fewer fruits, vegetables, and grains than those who consumed wine [35] and were more likely to eat ready-made food [36]. Beer consumption was found to be positively associated with smoking, and the interaction of these factors might alter the effect of smoking on body weight [37]. A study showed that current smokers who were heavy drinkers (> 60 g/day) had a lower BMI than never smokers/former smokers. On the other hand, WC was largest among smokers with heavy lifetime alcohol consumption [37]. Last, most of the studies used a self-report questionnaire to collect data; however, self-report questionnaires are prone to several types of bias (nonresponse bias, recall bias, and social desirability bias), and the results may be affected by intentional and unintentional misinterpretations of the questions being asked. One study found that self-reported alcohol intake accounted for only 40–60% of total sales, clearly underestimating the real amount of alcohol consumed [37]. Specifically, heavy drinkers tend to underreport their alcohol intake [37]. The period for recall can also affect the accuracy of self-reported data. However, the data used in the current study were collected during face-to-face interviews that were conducted by trained staff with the aid of CAPI to minimize errors that might occur during the interview. However, a 12-month recall period can lead to recall bias, which cannot be avoided in studies with such a design.

As both alcohol drinking and obesity are prevalent in Ireland, this study has several implications. First, given that obesity is multifactorial in origin, existing national and international guidelines for obesity management have not addressed the possible association between alcohol consumption and obesity; rather, they have given general behavioural recommendations on alcohol consumption [38, 39]. This is partly because of the conflicting evidence regarding the association between alcohol consumption (both quantity and frequency) and obesity. More elucidation on the association between the consumption of alcohol and obesity is advocated in both national and international guidelines. Second, Alcohol consumption is a significant public health problem in Ireland, with one-third of participants engaging in binge drinking and more than half engaging in harmful alcohol consumption. Despite current legislation to control alcohol intake, the prevalence of harmful alcohol consumption is still high, partly due to the normative view of alcohol consumption in Ireland. To mitigate the burden of alcohol use, a multisectoral approach is recommended, including raising awareness of the negative psychological and physical effects, improving the capacity to recognize and treat harmful alcohol behaviours, and raising public awareness of the possible association between alcohol consumption and obesity. Furthermore, the Alcohol Act, which was introduced in Ireland in 2018, included policies that can mitigate and further denormalize alcohol consumption, such as minimum unit pricing, structural separation, health labelling on alcohol-containing products, restrictions on alcohol advertising and marketing, regulation of alcohol sponsorship, and restrictions on certain promotional activities [40]. Third, a more consistent methodology for examining the potential association between alcohol consumption and obesity is required to reduce inconsistency in the results obtained. A prospective study design is better for examining temporal sequences and reverse causation than a cross-sectional design, where such relationships cannot be revealed. The baseline prevalence of obesity and alcohol consumption in the study population is important to consider, as it may alter the overall association. The measurement of exposure should be standardised to improve generalizability and consistency of results. The current study was unique in that it assessed the relationship between alcohol use and obesity using the AUDIT-C questionnaire, which can eliminate inconsistencies in the results obtained. Personal interviews are also preferable to self-report questionnaires because they provide more accurate, complete, and high-quality data. The alcohol intake recall period should be long enough to accurately predict the pattern of alcohol consumption and account for periods where alcohol intake is high. Last, Investment in research on alcohol consumption and obesity is recommended, particularly in Ireland, where both are considered public health problems. This will provide a better understanding of the association between alcohol consumption and obesity, as well as refine future recommendations for alcohol consumption in a community with a high prevalence of obesity. It is also recommended to tailor alcohol consumption guidelines based on an individual's medical history.

This is the first study to examine the association between alcohol consumption and obesity in a nationally representative sample in Ireland. The overall response rate for the 2017 Healthy Ireland Survey was significantly high (above 60%), making the data more generalizable to the entire Irish population. Data collection, including anthropometric measurements, was done by professional personnel; therefore, the risks of misinterpretation or missing data were minimised to some extent [41]. The study used the AUDIT-C score in the analysis of the association between obesity and alcohol consumption, making the results more generalizable as the AUDIT-C is a standardized tool to measure harmful alcohol consumption [16]. Last, this study controlled for a variety of sociodemographic and health-related variables that could confound the association between alcohol consumption and obesity.

Nonetheless, there are some limitations to this study. The major limitation is the cross-sectional design, which precludes the inference of a causal relationship between harmful alcohol consumption and obesity. Additionally, this study was prone to recall bias and social desirability bias, which tend to be more prevalent in face-to-face interviews and in longer recall periods [42]. The study did not look at the type or quantity of alcohol consumed, nor did it account for other potential confounders like sleep patterns, mental health, total energy intake, or type of occupation, as these variables were not collected in the original survey, all of which can confound the results [43]. Furthermore, overweight and obesity were prevalent in nearly two-thirds of the study population, increasing the likelihood of finding an association between alcohol consumption and obesity. The current study did not stratify participants by sex to identify variations in the association between harmful alcohol use and obesity between the strata, as different sexes have different patterns of alcohol consumption [44]; and thus need to be considered in future research. Lastly, underreporting of the amount of alcohol intake is problematic in both self-reporting and face-to-face interviews, especially among heavy drinkers, which might cause significant bias [45].