Critique 231: Association of alcohol consumption with obesity: Does moderate drinking increase or decrease the risk of obesity? 20 August 2019

White GE, Mair C, Richardson GA, Courcoulas AP, King WC.  Alcohol Use Among U.S. Adults by Weight Status and Weight Loss Attempt: NHANES, 2011−2016.  Am J Prev Med 2019;57:220−230.

 Authors’ Abstract

Introduction: Past research examining the relationship between alcohol use and weight status has not differentiated among classes of obesity. There is limited research investigating whether adults trying to lose weight consume less alcohol.

Methods: In 2018−2019, the authors analyzed 2011−2016 National Health and Nutrition Examination Survey data for nonpregnant adults aged ≥20 years with BMI ≥18.5 kg/m2. Multinomial and binomial logistic regression and linear regression were used to test associations between (1) past-year alcohol use and current weight status, differentiating among Class 1, 2, and 3 obesity and (2) past-year weight loss attempt and alcohol use, controlling for potential confounders.  Analyses were stratified by sex.

Results: Male current drinkers versus nondrinkers had lower odds of Class 3 obesity versus healthy weight (AOR=0.62, 95% CI=0.42, 0.92); female current drinkers versus nondrinkers had lower odds of Class 1 (AOR=0.67, 95% CI=0.50, 0.90), Class 2 (AOR=0.62, 95% CI=0.46, 0.83), and Class 3 (AOR=0.66, 95% CI=0.49, 0.89) obesity versus healthy weight. Among current drinkers, less frequent alcohol use was associated with higher odds of Class 1−3 obesity versus healthy weight in both sexes (p<0.05), whereas higher continued volume (heavier drinking) was associated with higher odds of Class 1−3 obesity versus healthy weight in females (p=0.049). Females reporting a weight loss attempt had higher odds of current drinking and more frequent heavy drinking.

Conclusions: Lower frequency of alcohol use (both sexes) and higher continued volume (female adults only) are associated with higher odds of higher weight status. Female adults trying to lose weight drink more, despite guidelines suggesting reducing caloric intake for weight control.

Forum Comments

While all alcoholic beverages contain calories that are absorbed, the association between alcohol intake and body weight and the risk of obesity is unclear.  Further, it has been shown that consumption of alcohol may increase hunger, so it would be expected that consumers of alcohol might have higher levels of obesity from several mechanisms.  However, limited clinical trials have not confirmed a positive association between alcohol intake and weight, and some studies show that regular moderate drinkers are less likely to be obese than are abstainers.

The present study uses data from National Health and Nutrition surveys between 2011 and 2016 to evaluate the association between reported amount and frequency of alcohol consumption and weight status.

Comments from individual Forum members:  Forum member Skovenborg provided an overview of the paper: “While the results of these analyses are not new, the paper adds data that underscore the importance of drinking pattern regarding the association of alcohol and weight.  The strength of the paper is a detailed analysis of a very large amount of data from the NHANES study representing an estimated 202 million US adults.  The authors report that 78.1% of males and 70.1% of females were current drinkers, a high percentage for the US population; however, one or more alcoholic drink in the past 12 months was enough to be classified as a current drinker.

“The prevalence of obesity of any degree among the participants was frightening: about 36% of US males and 41% of US females – and the high obesity prevalence was present in spite of the fact that 4 in 10 males and 6 in 10 females attempted to lose weight in the past year.  Specifically, in the present analyses, 22.8% of males and 20.1 % of females had Class 1 obesity (BMI 30-<35 kg/m²), 8.0% of males and 11.2% of females had Class 2 obesity (BMI 35-<40 kg/m²), and 5.2% of males and 9.4% of females had Class 3 obesity (BMI ≥ 40 kg/m²).

“As for the relation between alcohol consumption and obesity, the authors found that male current drinkers versus nondrinkers had lower odds of Class 3 obesity versus healthy weight subjects (AOR=0.62, 95% CI=0.42, 0.92); female current drinkers versus nondrinkers had lower odds of Class 1 (AOR=0.67, 95% CI=0.50, 0.90), Class 2 (AOR=0.62, 95% CI=0.46, 0.83), and Class 3 (AOR=0.66, 95% CI=0.49, 0.89) obesity versus healthy weight subjects.  Among current drinkers, less frequent alcohol use was associated with higher odds of Class 1−3 obesity versus healthy weight in both sexes (p<0.05).  The observation of an inverse association between a regular, moderate alcohol consumption and obesity is not new; however, the results of this nationally representative study are the first to demonstrate that the magnitude of the inverse association is stronger with higher levels of weight status.”

Forum member Waterhouse questioned: “Is it possible that the observed benefit of alcohol consumption observed here was limited by including very occasional (1 or more drink per year) consumers in the category of ‘drinkers’?”  Others agreed, commenting that including persons reporting very rare consumption of alcohol might be better included in a ”non-drinker” category rather than being mixed with regular or heavy drinkers.  Noted reviewer Mattivi: “I share the same concern about potential bias deriving from the inclusion of  people with a  self-reported  irrelevant number of drinks per year into the category of ‘drinkers.’”

Forum member Parente noted: “Subjects’ alcohol consumption was not stratified or investigated according to beverage type.”  Others agreed that this may have weakened the relation between moderate drinking and lower obesity, as many previous studies have shown beneficial health effects of wine consumption, especially in comparison with the consumption of spirits.

Reviewer Skovenborg also warned about readers relying on the references quoted in this paper:  “For example, their reference 4 is an older study not accessible in most databases; reference 5 is a study of 152 non-diabetic Pima Indians concluding that family membership is the principal determinant of the ratio of fat to carbohydrate oxidation – showing that you should be careful in your choice of parents but these data have little relevance to the relationship between alcohol use and weight status; reference 6 found a statistically significant yet quantitatively immaterial link between increasing alcohol use and weight gain among men, however, alcohol use was not a risk factor for weight gain among women and older adults – the conclusions are hardly sound evidence for the statement ‘alcohol may lead to weight gain and obesity;’ reference 7 found that alcohol consumption increased the incentive salience of smoking cues – a finding that has no relevance to the statement that ‘alcohol may prompt people to eat more and exercise less;’ and reference 8 found that alcohol intoxication is not a good thing for people performing memory tasks – a finding that has no relevance to the statement that ‘alcohol may prompt people to eat more and exercise less.’”  Reviewer Waterhouse agreed with Skovenborg’s remarks about the references listed for the paper: “It appears that the conclusions of the authors may have been unduly influenced by some misinterpreted data from the references they supposedly reviewed.”

Despite a number of weaknesses from the authors’ classification of “drinkers” and failure to include the type of beverage consumed, most Forum members considered this to be generally a well-conducted study.  It supports previous research showing how the regular moderate use of alcohol may lower, rather than increase, the risk of severe obesity.  Further, the decrease in risk was for more frequent drinking, and tended to be lost when there was infrequent or heavy drinking.  The analyses were based on more than 14,000 subjects representative of the US population, the analyses were appropriate, and the authors had the ability to adjust for a large number of potential confounders (including age, race, education, employment, income, marital status, smoking status, depression symptoms, high cholesterol, hypertension, coronary disease, stroke, diabetes COPD, physical activity and weight loss attempt).

Most of the comparisons reported in this paper excluded never drinkers, with the authors stating that they may have other lifestyle characteristics for which it was not possible to adequately adjust for in their analyses.  Thus, it was limited to subjects who reported at least 12 alcoholic drinks during their lives, and most comparisons were between “current drinkers,” those reporting the consumption of one or more alcohol drinks in the past year, and those who reported that they had not had any alcoholic beverage in the past year.

Importance of including the level of obesity as an outcome:  Forum reviewers agreed with the authors that this study is important as it separates subjects into classes of obesity, and the potentially protective effects of regular moderate drinking are significant for greater degrees of obesity, where adverse health effects are greater.  However, Ellison had questions about the concluding sentence of the paper: “Because even modest weight loss can have profound health benefits at the population level, the effectiveness of limiting alcohol use on reducing caloric intake and achieving weight control should be investigated.”  He was not sure that it was based on what the present study demonstrated.

Forum member Finkel wrote: “I agree that obesity is a massive worldwide health problem, counterintuitively most prominently where nutrition is poorest.  The main conclusions of this paper on obesity and alcohol are very familiar.  I can recall a number of other published studies and discussions at meetings from many years back bearing the same or similar messages.  I also recall widespread puzzlement about the preservation of mass and energy that seemed to operate within the bounds of the interrelationship of drinking, particularly of wine, and weight gain or loss, especially among women.  To my knowledge, we have not come to enlightenment lately.  In any case, I have no quarrel with the methods or discussion of this paper, though the authors’ suggestion at the end of the Conclusion section, as queried by Ellison, seems more a gratuitous nod to reflexive alcohol naysayers than reasonably to the results of the study.”

Caution needed in interpreting results:  Reviewer van Velden raised a note of caution in the interpretation of the results of this analysis.  “While I agree with the comments of others, I would like to point out that obesity is closely related to increased carbohydrate, especially refined carbs (sugar).  Excess carbohydrates are deposited in the abdominal fat.

“Moderate alcohol consumption is often related to many components of a healthy lifestyle, including diet, exercise, effective stress management, and weight management.  The sugar in grapes are converted into ethanol during the fermentation process, and alcohol follows a different metabolic pathway than sugars, and is also related to the genetic profile of the individual.  A problem in interpretation may occur since moderate wine drinkers often limit their intake of refined carbohydrates.  Abstainers often have a genetic profile that cannot metabolize alcohol effectively.  Thus, there are many confounders related to alcohol consumption that must be taken into consideration before reliable conclusions can be made on the effect of alcohol and weight gain/loss.”  Other Forum reviewers agreed that adjusting for potential confounders remains a problem in all observational studies.  Also, they noted that in evaluating long-term alcohol intake it is necessary to take into account changes in alcohol consumption over time, as discussed below.

Subjects attempting to lose weight:  The authors also related alcohol consumption to obesity according to whether or not the subject had reported that he or she had been attempting to lose weight.  Their findings indicate that such associations tended to not be significant after adjustments were made for potential confounders.

Need for prospective studies with repeated assessments of alcohol to judge effects on obesity:  The authors also ask for prospective studies of the alcohol-weight association (as “such studies may better characterize individuals who are more likely to gain weight because of alcohol use”), but they do not describe in depth the results from studies that have previously reported such associations.  Overall, there are differences in effect in the few papers based on changes in alcohol and effects on body weight: Mozaffarian et al show a slight increase in weight with increases in total alcohol intake, but Wannamethee et al showed consistent decreases in weight to be associated with moderate drinking.  The latter study noted increases in obesity for heavy drinkers, with some differences according to type of beverage.  Also, Wang et al found that compared with nondrinkers, women who consumed a light-to-moderate amount of alcohol gained less weight over 12.9 years of follow up.  MacInnis et al found that participants were less likely to have elevated weight or waist circumference in a follow-up exam if they consumed low to moderate amounts of alcohol between examinations.  Traversy & Chaput provide a good overview and discuss potential explanations for how alcohol intake can result in either an increase or a decrease in obesity over time depending on amount of alcohol, frequency of drinking, type of beverage, and changes in intake.

An interesting study by French et al showed that the effects of changes in alcohol intake tend to be different for men and women (no change or an inverse associations between alcohol and weight for women).  However, they found that increases in the amount of alcohol are more often associated with an increase in obesity for both men and women.  They found that the type of beverage also has an influence, with wine consumption being less associated with increasing weight than seen for other beverages.  O’Donovan et al reported that over time obesity was lower among moderate drinkers, but was increased among heavy drinkers.   French et al concluded that results in studies incorporating repeated alcohol assessments are much smaller in magnitude and sometimes different in sign compared to the benchmark pooled cross-sectional estimates.  These findings emphasize the importance of large-scale prospective studies that are able to judge the effects on obesity of changes in alcohol intake over time.

References from Forum critique

French MT, Norton EC, Fang H, Maclean JC.  Alcohol consumption and body weight.  Health Econ 2010;19:814–832.  doi: 10.1002/hec.1521

MacInnis RJ, Hodge AM, Dixon HG, et al.  Predictors of increased body weight and waist circumference for middle-aged adults.  Public Health Nutrition 2013;17:1087-1097.

Mozaffarian D, Hao T, Rimm EB, Willett WC, Hu FB Changes in Diet and Lifestyle and Long-Term Weight Gain in Women and Men.  N Engl J Med 2011;364:2392–2404. doi: 10.1056/NEJMoa1014296.

O’Donovan G, Stamatakis E, Hamer M.  Associations between alcohol and obesity in more than 100,000 adults in England and Scotland.  Br J Nutrition 2018;119:222-227.

Traversy G, Chaput J-P.  Alcohol consumption and obesity: An update.  Curr Obes Rep 2015;4:122-130.

Wang L, Lee I-M, Manson JE, Buring JE, Sesso HD.  Alcohol consumption, weight gain, and risk of becoming overweight in middle-aged and older women.  Arch Intern Med 2010;170:453-461.

Wannamethee SG, Field AE, Colditz GA, Rimm EB.  Alcohol intake and 8-year weitht gain in women: A prosptective study.  Obesity research 2004;12:1386-1396.

Forum Summary

While alcoholic beverages contain calories that are absorbed, the association between alcohol intake and body weight and obesity is unclear.  Limited clinical trials have not confirmed a positive association between alcohol intake and weight, and some epidemiologic studies, including prospective cohort studies, show that regular moderate drinkers are less likely to be obese than are abstainers.

The present study uses data from National Health and Nutrition surveys between 2011 and 2016 to evaluate the association between reported amount and frequency of alcohol consumption and weight status.  An estimated 78.1 % of males and 70.1% of females (who reported at least one drink in the past year) were classified as “drinkers.”   A total of about 36% of US males and 41% of US females were described as “obese” (defined as BMI ≥ 30 kg/m²).  The analyses included the level of obesity (Class 1, 2, or 3) as outcomes (with Class II defined as BMI 35-<40 kg/m² and Class III as BMI ≥ 40 kg/m²) making it possible to judge the effects of alcohol on varying levels of obesity.

The present study reports that male current drinkers versus nondrinkers had 38% lower odds of Class 3 obesity versus subjects with a healthy weight; female current drinkers versus nondrinkers had more than 30% lower odds of Class I, Class II, or Class III obesity   In both sexes, more frequent alcohol consumption was associated with less obesity, while heavy alcohol consumption was associated with increased risk of obesity.  Females who reported that they were trying to lose weight consumed more than others, but there was not a strong relation between reported attempts to lose weight and outcomes.

Forum reviewers considered this to be a well-done analysis, with adjustments for most common potential confounding variables.  It provides some support for previous findings of an inverse association with obesity for light-to-moderate alcohol consumption but an increase in obesity for heavy consumption.  The authors state that this is the first study to be able to provide data showing that the magnitude of the inverse association is stronger for higher levels of weight status (Classes 1 to 3).

Forum members noted, however, that this study was based on assessments of different subjects in each wave of the national surveys, rather than on repeated assessments among the same individuals over time.  It is clear that repeated recordings of alcohol intake in prospective cohort studies make it possible to obtain a reliable estimate of any changes in alcohol intake that may occur, rather than having to base results on a single measure, such as only a baseline value.  Further, members caution that there are many genetic and lifestyle variables associated with the development of obesity, and residual confounding is always a potential problem in such analyses.  Thus, while the present paper provides data indicating that more severe obesity may result from higher intake of alcohol, and that light-to-moderate intake may actually lead to a decrease in the risk of obesity, data from cohort studies with repeated assessments of alcohol and body size may provide more precise estimates of effect.

         *                 *                  *                *                *                 *               *

Comments on this critique by the International Scientific Forum on Alcohol Research were provided by the following members:

Andrew L. Waterhouse, PhD, Department of Viticulture and Enology, University of California, Davis, USA

David Van Velden, MD, Dept. of Pathology, Stellenbosch University, Stellenbosch, South Africa

Fulvio Ursini, MD, Dept. of Biological Chemistry, University of Padova, Padova, Italy

Pierre-Louis Teissedre, PhD, Faculty of Oenology–ISVV, University Victor Segalen Bordeaux 2, Bordeaux, France

Arne Svilaas, MD, PhD, general practice and lipidology, Oslo University Hospital, Oslo, Norway

Creina Stockley, PhD, MSc Clinical Pharmacology, MBA; Adjunct Senior Lecturer at the University of Adelaide, Australia

Erik Skovenborg, MD, specialized in family medicine, member of the Scandinavian Medical Alcohol Board, Aarhus, Denmark

Matilda Parente, MD, consultant in molecular pathology/genetics and emerging technologies, San Diego, CA, USA.

Fulvio Mattivi, MSc, CAFE – Center Agriculture Food Environment, University of Trento, via E. Mach 1,  San Michele all’Adige, Italy

Ulrich Keil, MD, PhD, Professor Emeritus, Institute of Epidemiology & Social Medicine, University of Muenster, Germany

Harvey Finkel, MD, Hematology/Oncology, Retired (Formerly, Clinical Professor of Medicine, Boston University Medical Center, Boston, MA, USA)

R. Curtis Ellison, MD, Professor of Medicine, Section of Preventive Medicine & Epidemiology, Boston University School of Medicine, Boston, MA, USA

Giovanni de Gaetano, MD, PhD, Department of Epidemiology and Prevention, IRCCS Istituto Neurologico Mediterraneo NEUROMED, Pozzilli, Italy