Critique 244: Is “ideal cardiovascular health” more common among consumers of wine than among drinkers of other beverages? — 18 December 2020

Ogunmoroti O, Osibogun O, McClelland RL, Lazo M, Mathews L, Okunrintemi V, Oni ET, Burke GL, Michos ED.  Alcohol type and ideal cardiovascular health among adults of the Multi-Ethnic Study of Atherosclerosis.  Drug and Alcohol Dependence 2020: Pre-publication release

Authors’ Abstract

Background: Light to moderate alcohol consumption is associated with favorable cardiovascular health (CVH). However, the association between alcohol type and ideal CVH has not been well-established. We examined the relationship between alcohol type and ideal CVH as measured by the American Heart Association’s seven CVH metrics.

Methods: We analyzed data from 6,389 men and women aged 45-84 years from a multi-ethnic cohort free of CVD. Alcohol type (wine, beer and liquor) was categorized as never, former, 0 but drink other alcohol types, > 0 but < 1 drink/day, 1-2 drinks/day and > 2 drinks/day. A CVH score ranging from 0-14 points was created from the seven CVH metrics (Inadequate score, 08; average, 9-10; optimal, 11-14). We used multinomial logistic regression to examine the association between alcohol type and CVH, adjusting for age, sex, race/ethnicity, education, income, health insurance, field site and total calorie intake.

Results: Mean (SD) age of participants was 62 (10) years. 53% were women. Participants who consumed 1-2 drinks/day of wine had higher odds of optimal CVH scores compared to those who never drank wine [adjusted prevalence odds ratio (POR) 1.64 (1.12-2.40)]. In comparison to participants who never drank beer, those who consumed > 2 drinks/day of beer had lower odds of optimal CVH scores [0.31 (0.14-0.69)]. Additionally, those who consumed > 2 drinks/day of liquor had lower odds of optimal scores compared to those who never drank liquor [0.32 (0.16-0.65)].

Conclusion: Moderate consumption of wine was associated with favorable CVH. However, consumption of beer or liquor was associated with poorer CVH.

Forum Comments

Forum member Djoussé noted: “I have a few methodologic comments regarding this paper:  first, the authors classified each type of alcoholic beverage as never; former; 0 but drink other alcohol types; >0 but <1 drinks/day; 1-2 drinks/day; and >2 drinks/day. Hence, it is possible that for wine, a participant could report 1-2 drinks of wine per day and also consume some amounts of liquor and/or beer. This applies to the last 3 categories assuming never and former categories are set to zero. Such exposure misclassification could have obscured the true association under investigation.  It is reasonable to estimate that only a few people will consume only a single type of alcoholic beverage, especially when frequency is >1 drink/day.  The multivariable model does not seem to control for beer and spirits when assessing wine and vice versa.

“Further, the interpretation of proportional odds is misleading. With 3 outcomes (inadequate, average, and optimal), the reported ORs are relative to “inadequate CVH” and such omission renders the message inaccurate. For example, in the abstract, the text reads ‘Participants who consumed 1− 2 drinks/day of wine had higher odds of optimal CVH scores compared to those who never drank wine [adjusted prevalence odds ratio (POR) 1.64 (1.12− 2.40)].’ This sentence ignores the fact that the outcome is polytomous (and not dichotomous). The 64% higher odds of having optimal CVH is relative to having inadequate CVH score when comparing wine drinkers of 1-2 drinks/day to those who never consumed wine. Throughout most of the text, the line is blurred between interpretation of polytomous and dichotomous outcome.”

Forum member Skovenborg wrote:  “I agree with Professor Djoussé that the study has a possible (and likely) problem of exposure misclassification and a complicated structure of reporting the results due to the facts that many outcomes are polytomous and not dichotomous.  The LS7 metrics consists of a mixture of ideal health behaviors (non-smoking, BMI, physical activity and diet) and ideal health factors (total cholesterol, blood pressure and fasting blood glucose). Further, it seems to me that BMI should be a health factor and not a lifestyle choice.

“I believe that the lifestyle factors physical activity and diet should be considered as confounders associated with beverage choice and not an effect of beverage choice.  The influence of alcoholic beverages on the health factors is mostly a result of the amount of alcohol intake and not a result of intake of the specific beverages.  As for the study upon which this paper is based (the Multi-Ethnic Study of Atherosclerosis), it has been reported that while more than one-half of the cohort reported some alcohol consumption, 78% of such drinkers averaged < 1 drink/day.  Having a small percentage of heavy drinkers would make it difficult to observe a J-shaped curve relation between alcohol intake and disease markers.”

The study does not demonstrate, per se, that wine consumption reduces the risk of cardiovascular disease:  Reviewer Ellison noted that “This paper shows that, in comparison with consumption of beer or liquor, wine consumption is associated with a more favorable cardiovascular health index, CVH, judged from a score based on smoking, physical activity, BMI, diet, total cholesterol, blood pressure, and blood glucose.  It has been shown that people with more healthy lifestyle behaviors tend to have less cardiovascular disease (CVD) and lower mortality rates.  However, this study does not measure effects on the risk of disease occurrence or mortality, but just on the CVH index.  Thus, we cannot tell specifically from this study whether the higher health indices among moderate wine drinkers is the result of the wine that they drink (even though this might be expected to some extent since, for example, moderate wine intake is known to favorably affect BMI, blood lipids, and blood glucose). However, we must also consider that the higher health indices among moderate wine drinkers may be because people with healthier lifestyles in general (who tend to be better educated, with higher income, smoke less, have a moderate diet, etc.) choose to consume wine in moderation rather than to drink beer or liquor, and are less likely to binge drink or drink larger amounts of any alcoholic beverage.  While the authors adjusted for some of these factors, residual confounding is always a possibility.”

Forum member Andrzej Pajak wrote: “The relation described is between consumption of alcohol beverages and behavioral traits or traits which are strongly determined by lifestyle.  Thus, the main finding of this study could be stated as ‘Moderate wine drinkers behave better that beer and spirits drinkers’. This is in agreement with previous studies mentioned in the paper.  I agree that the authors were able to adjust for important possible confounders, but residual confounding can be expected.  The paper does not deal with the problem of to what extent can the relation between alcoholic beverages and such traits as BMI, cholesterol, glucose, or blood pressure be explained by non-behavioral factors.  The authors’ conclusion seems fair, unless CVH is misinterpreted as CVD.  Results presented do not allow for any firm conclusion about casual relation between wine consumption and health.”

Wine as part of a healthy diet:  Reviewer de Gaetano wrote: “Wine, and more rarely beer, is an essential component of the Mediterranean diet. In our Moli-sani study, removing wine consumption from the score of adherence to a Mediterranean diet was associated with about 15% reduction of the diet protection against total and cardiovascular (CV) mortality.  Several studies, including ours, have observed that the relationship between increasing amounts of wine consumption and CV outcomes may be described as a J curve, but often tends to appear as an L-shaped curve, suggesting that even larger doses than the moderate ones are still associated with reduced CV risk (Bonaccio et al; Xi et al). This is less frequently observed with beer but not with spirits. In any case, wine has been consistently shown to be associated with reduced CV risk, independent of healthy lifestyles.”

Reviewer Lanzmann-Petithory noted: “I am very happy with the approach of this paper and its recommendations and conclusion.  The authors are very aware of the ‘French Paradox’ and the extensive work over many decades of Professor Serge Renaud, who was the first scientist to identify many relevant mechanisms for lower cardiovascular risk among wine drinkers (as summarized well by Simini).  Despite the defaults mentioned by other reviewers, the correlations are very significant. I think that a lot of wine drinkers, including me, do not drink beer and spirits, so the results are consistent with my beliefs and practices.”

Importance of also evaluating genetic patterns:  Forum member Van Velden commented: “The health benefits of moderate red wine consumption are well known.  However, lifestyle factors and genetic markers for CVD, as well as oxidative stress, also play a role.  From our own research results at the University of Stellenbosch in this regard, we have found a number of findings:

(1)  The health benefits of red wine, as well as brandy, were evident by a significant increase in HDL cholesterol levels. Moderate alcohol consumption was shown to protect against cardiovascular disease regardless of whether brandy or wine was consumed, with additional benefits from wine ascribed to the presence of polyphenolic antioxidants.

(2)  A beneficial effect was seen of the non-alcoholic fraction of red wine, in particular, with greater protective effects on cardiovascular risk factors than other alcoholic beverages.

(3)  The genetic profile influences the potential health benefits and detrimental effects of moderate alcohol intake. Although the antioxidant effects of red wine are well-documented, it seems clear that alcohol intake may be contra-indicated in individuals with certain genetic alterations that occur relatively frequently in the general population.

“Thus, we conclude that early risk detection through the identification of specific genetic markers that correlate with known biochemical pathological markers for cardiovascular disease may make it possible to develop an individually tailored preventative medicine programme.  By performing a genetic screen of clinically relevant low-penetrance mutations, researchers have been able to identify individuals who are likely or unlikely to derive cardiovascular benefit from alcohol consumption. Since genetic risk factors influence the effect of alcohol on biochemical markers of cardiovascular risk, safe limits for alcohol consumption may in the future be based partly on the genetic profile or knowledge of the importance of genetic variation in this context.”

References from Forum critique:

Bonaccio M, Di Castelnuovo A,  Costanzo S, Persichillo M, De Curtis A, Donati MB, de Gaetano G, Iacoviello L, MOLI-SANI study Investigators.  Adherence to the traditional Mediterranean diet and mortality in subjects with diabetes. Prospective results from the MOLI-SANI study.  Eur J Prev Cardiol 2016;23:400-407. doi: 10.1177/2047487315569409

Simini B.  Serge Renaud: from French paradox to Cretan miracle.  Lancet 2000;355:48. doi: 10.1016/S0140-6736(05)71990-5.

Xi B, Veeranki SP, Zhao M, Ma C, Yan Y, Mi J.  Relationship of Alcohol Consumption to All-Cause, Cardiovascular, and Cancer-Related Mortality in U.S. Adults.  J Am Coll Cardiol 2017;70:913-922.  doi: 10.1016/j.jacc.2017.06.054.

Forum Summary

While a beneficial effect on cardiovascular disease of moderate drinking has been shown repeatedly in epidemiologic studies, the specific effects of different types of alcohol have not been well established.  The authors examined the relation between alcohol type and an index of ideal cardiovascular health (CVH) as measured by the American Heart Association’s seven CVH metrics.  Data from more than six thousand men and women, participants in the Multi-Ethnic Study of Atherosclerosis, were used to relate the association of the consumption of wine, beer, and liquor to the CVH score.

The authors report that participants who consumed 1-2 drinks/day of wine had higher odds of optimal CVH scores compared to those who never drank wine [adjusted prevalence odds ratio (POR) 1.64 (1.12-2.40)]. However, in comparison with participants who never drank beer or liquor, the odds of optimal CVH scores were lower for consumers of these beverages.

Forum members had some questions about the methodology for judging intake of the different beverages, as well as the statistical approach used in these analyses.  However, they consider that the results (more favorable associations with wine consumption than from that of other beverages on indices of cardiovascular health) are in line with most previous studies.

It is important to realize that this study did not measure effects on the risk of disease occurrence or mortality, but just on the CVH index, a measure of overall cardiovascular health.  Thus, we cannot tell specifically from this study whether the higher health indices among moderate wine drinkers was the result of the wine that they drink (even though this might be expected to some extent since, for example, moderate wine intake is known to favorably affect BMI, blood lipids, and blood glucose) or that their results relate to the fact that people with more moderate lifestyles and higher education, income, etc., tend to prefer moderate wine consumption and are less likely to be binge or heavy drinkers.  It may be that wine drinkers make better decisions on all lifestyle behaviors than consumers of other beverages.


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

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

Luc Djoussé, MD, DSc, Dept. of Medicine, Division of Aging, Brigham & Women’s Hospital and Harvard Medical School, Boston, MA, USA

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

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

Dominique Lanzmann-Petithory, MD, PhD, Nutrition Geriatrics, Hôpital Emile Roux, APHP Paris, Limeil-Brévannes, France

Professor Andrzej Pająk, Epidemiology and Population Studies, Jagiellonian University Medical College, Kraków, Poland

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

Creina Stockley, PhD, MSc Clinical Pharmacology, MBA.  Principal, Stockley Health and Regulatory Solutions; Adjunct Senior Lecturer, The University of Adelaide, Australia.

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

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

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