Critique 243: Does the “quality” of the diet modify the relation of alcohol intake to the risk of hypertension? – 19 October 2020

Jiang W, Meng X, Hou W, Wu X, Wang W, Wang M, Chu X, Wang P, Sun C, Han T, Li Y.   Impact of overall diet quality on association between alcohol consumption and risk of hypertension: evidence from two national surveys with multiple ethnics.  Eur J Clin Nutr 2020; doi.org/10.1038/s41430-020-00708-1.

Authors’ Abstract

Background/objectives Alcohol is commonly consumed around mealtimes. This study hypothesized that the association between alcohol and hypertension was influenced by overall diet quality. This study aims to test the hypothesis that overall diet quality influenced associations between alcohol and risk of hypertension across different ethnicities.

Subjects/methods Using nationally representative data from National Health and Nutrition Examination Survey (NHANES 03–12), China Health Nutrition Survey (CHNS), and an independent population-based study, 43,914 adults were included.  Subgroup analysis included 6,984 adults from CHNS with a 14-year follow-up. Light alcohol consumption was defined as < 7 standard drinks/week, moderate as 7–21 drinks/week, and heavy as >21 drinks/week.  Alternative healthy eating index and diet balance index were calculated as indicators of diet quality.

Results There were 3,968 hypertensives in Caucasians (N = 11,325), 1,976 in Africans (N = 5010), 1,907 in Hispanics (N =7,274) and 5,267 (N = 20,305) in Chinese. In context of high diet quality, light alcohol consumption was significantly associated with decreased risk of hypertension in Caucasians, and the risk of hypertension was not significantly increased with increasing in alcohol consumption in Caucasians, Chinese, and Hispanics (all P for trend > 0.05). On the contrary, in context of low diet quality, the risk of hypertension was significantly increased with increasing in alcohol consumption in Caucasians (P for trend = 0.005), Chinese (P for trend = 0.001) and Hispanics (P for trend = 0.022). Associations between alcohol consumption and risk of hypertension significantly varied by diet-quality scores in Caucasians, Hispanics, and Chinese (all P for interaction < 0.01) showing gradually changing from nonsignificant increasing trend to linear association.

Conclusions This study firstly demonstrated that overall diet quality influenced associations between alcohol and risk of hypertension across different ethnicities, emphasizing that when examining health effects of alcohol on blood pressure, diet quality should be considered.

Forum Comments

Overview of study:  Forum member Puddey provided an excellent overview of this study.  “Recent meta-analysis of 36 intervention trials of at least 7 days and up to 2-years duration indicate a causal relationship between the consumption of alcohol and blood pressure (BP) with a dose-related increase in BP evident in those who drank >2 drinks/day (>24-g alcohol/day)(Roerecke, et al). Such causal inference is also strongly supported by at least 6 separate Mendelian randomisation studies (Puddey, et al).  The contention in this paper that a high quality diet can prevent alcohol-related hypertension, even at higher levels of intake, therefore needs to be very carefully examined.

“In their initial cross-sectional analysis of the NHANES data the authors state that compared to abstainers there was a decreased risk of hypertension with light alcohol consumption (0.5–7 drinks per week) in Caucasians and Hispanics. However, NHANES was a cross-sectional study — risk was not ascertained — odds ratios for hypertension were assessed. Further, the decrease in odds of hypertension was no longer significant when dietary covariates and diet quality indices were added to the model. On the other hand an increase in the odds of hypertension in Caucasians and Hispanics with heavy alcohol use (>21 drinks per week) remained significant after adjustment for all covariates. There was surprisingly, in contrast to many previous reports, no association of alcohol use with hypertension in African Americans. In the 2 pooled Chinese populations (The China Health and Nutrition Survey and the Harbin Cohort Study on Diet, Nutrition, and Chronic Noncommunicable Disease) the authors did not see a decrease in odds of hypertension with light alcohol intake but found a significant increase in those drinking >14 drinks per week, even after consideration of all covariates including dietary covariates and diet quality indices. The Chinese populations appeared to be consuming alcohol with a much lower prevalence (~30%) when compared to the NHANES data (73 to 88%) and at lower amounts (given there is no >21 drinks per week category reported or analysed).

“After dichotomising the population into either high or low healthy eating index categories, the authors now find an increasing linear trend in the odds of hypertension with alcohol use in those Caucasians and Hispanics with a low healthy eating index but no increasing trend in those with a high healthy eating index. However, higher alcohol intakes are generally associated with poorer diets (Parekh, et al). In this regard, the NHANES study has previously reported that with increasing quantity of alcohol consumption there is a decrease in the healthy eating index, but with increasing frequency of alcohol use there is an increase in the healthy eating index (Breslow, et al). It is possible therefore that the failure to have included both quantity and frequency measures in the logistic regression models has confounded the outcome in the current study: the high frequency, low volume drinkers falling into the high healthy eating index category and the low frequency, high volume drinkers falling into the low healthy eating index category. Which healthy index category you fall into is a product of your pattern of drinking, the latter a more likely determinant of whether alcohol ultimately increases the risk of alcohol-related hypertension than the healthy eating index itself, with increasing recent evidence for binge drinking as an added risk factor for alcohol-related hypertension (Piano, et al; Hayibor, et al).

“Moreover, as well as this influence of pattern of drinking, alcohol intake has dose-related influences on intake of specific dietary components, especially carbohydrate intake (Cummings, et al). Alcohol intake and binge drinking are related to tendency to skip breakfast and lunch (Patte, et al). The type of alcohol consumed is also linked to dietary quality and components (Gronbaek, et al; Scholz, et al) as well as pattern of drinking (Gronbaek, et al). Given that none of these factors have been considered, it is difficult to reconcile the authors contention that residual confounding or unaccounted for confounding ‘may have been expected to only affect the results slightly’. Until these additional confounding factors are considered I think we should be sceptical that a high quality diet mitigates against alcohol-related hypertension.”

Comments from other Forum members:  Reviewer Skovenborg added: “In complete agreement with Professor Puddey’s comment and questions, I just want to add some further points:

1. The study has no information about beverage preference. In some populations wine drinkers have a healthier Mediterranean type of diet (Johansen et al).
2. The study found a different association of alcohol use and risk of hypertension between Caucasians, African-Americans, and Chinese. The different associations raise questions regarding the possible reasons and also questions the external validity of the study results.
3. In spite of the weaknesses, the possible influence of diet quality in the context of alcohol consumption and risk of hypertension is an interesting hypothesis that should inspire not only further observational studies but also feasible short-time experimental studies.

Reviewer Ellison noted: “I have conflicting impressions of this paper.  The methods seem appropriate, and the analyses are apparently well done.  Further, the results are presented separately for the USA data (from NHANES) and data from China, and there are no attempts to combine the data into a single analysis.  My main worry is whether or not it is the diet solely that is the reason that alcohol consumption was found to be associated with lower, higher, or no effect depending on the type of diet.  As noted above by other Forum members, I am concerned that there may be serious confounding by other factors, especially socio-economic status (SES) (assuming that poorer, less educated people with lower-level jobs and lower income tend to have poorer quality diets).  The authors did adjust for whether or not the subject completed high school, but apparently had no other data on SES.

“There are also worries that while their original data often included information on the type of beverage consumed, no results are shown indicating whether or not there were differences if it was beer, wine, or spirits. And there was insufficient data on the pattern of drinking.  The authors begin their abstract saying that ‘Alcohol is commonly consumed around mealtimes,’ but apparently do not have any data on whether the beverage was consumed with or without food for the present analyses.”

Reviewer van Velden wrote: “It is clear that lifestyle factors have a profound influence on hypertension and overall health.  Exercise, low salt intake, weight management and stress management all play a role in high blood pressure with all the consequences such as stroke, IHD, deep vein thrombosis and other chronic and degenerative diseases such as osteoporosis and arthritis.  The typical Western diet, high in animal products (meat), is acidogenic and has an influence on homeostasis causing calcium loss to maintain the alkaline pH of the body.”

Forum member de Gaetano, after discussions with his colleague Marialaura Bonaccio, wrote:  “It is well established from population studies that diet quality follows a socioeconomic gradient, with more deprived individuals reporting lower diet quality as compared to their wealthy counterpart (Bonaccio et al); these authors also noted that income (not assessed in the current paper) was associated with dietary patterns in all categories of education.  In view of this, a major bias in the analyses possibly emerges: although models have been adjusted by the authors for educational level, it is likely that those participants with poorer diet quality are those from the bottom SES strata, which in turn are at higher health risk, including for hypertension (Leng et al).   Analyses would have benefited from a stratification by additional SES indicators, after verifying that diet quality actually exhibited a SES gradient also in these cohorts, as likely it did.”

Forum member Finkel commented: “The issues this paper addresses are important, and have been previously conflictingly reported. I agree that the methodology appears sound. The paper is written in a straightforward manner, and without any attempt to editorialize. All refreshing.  I am not entirely comfortable with the designations of “quality” for diets, although, taking the case of blood pressure for example, extremes of salt intake certainly would be operative. I agree with the concerns that SES confounders may lay in ambush. Some might judge that diets considered of high quality are also those whose savor has been removed. Further study of a number of facets of eating and drinking in particular, and lifestyle in general, may be revealing.”

References from Forum critique

Bonaccio M, Bonanni AE, Di Castelnuovo A, De Lucia F, Donati MB, de Gaetano G, Iacoviello L; Moli-sani Project Investigators.  Low income is associated with poor adherence to a Mediterranean diet and a higher prevalence of obesity: cross-sectional results from the Moli-sani study.  BMJ Open 2012;2:e001685. doi: 10.1136/bmjopen-2012-001685. PMID: 23166131

Breslow RA, Guenther PM, Smothers BA. Alcohol drinking patterns and diet quality: the 1999-2000 National Health and Nutrition Examination Survey. Am J Epidemiol 2006;163:359-366.

Cummings JR, Gearhardt AN, Ray LA, Choi AK, Tomiyama AJ.  Experimental and observational studies on alcohol use and dietary intake: a systematic review. Obes Rev 2020;21:e12950.

Grønbæk M, Tjonneland A, Johansen D, Stripp C, Overvad K. Type of alcohol and drinking pattern in 56,970 Danish men and women. Eur J Clin Nutr 2000;54:174-176.

Hayibor LA, Zhang J, Duncan A.  Association of binge drinking in adolescence and early adulthood with high blood pressure: findings from the National Longitudinal Study of Adolescent to Adult Health (1994-2008). J Epidemiol Community Health 2019;73:652-659.

Johansen D, Friis K, Skovenborg E, Grønbæk M.  Food buying habits of people who buy wine or beer: Cross sectional study. BMJ 2006;332;519-522.

Leng B, Jin Y, Li G, Chen L, Jin N.  Socioeconomic status and hypertension: a meta-analysis.   J Hypertens 2015;33:221-229. doi: 10.1097/HJH.0000000000000428.

Parekh N, Lin Y, Chan M, Juul F, Makarem N.  Longitudinal dimensions of alcohol consumption and dietary intake in the Framingham Heart Study Offspring Cohort (1971-2008). British Journal of Nutrition 2020;1-10. doi:10.1017/S0007114520002676.

Patte KA, Leatherdale ST. A cross-sectional analysis examining the association between dieting behaviours and alcohol use among secondary school students in the COMPASS study. J Public Health (Oxf) 2017;39:321-329.

Piano MR, Burke L, Kang M, Phillips SA.  Effects of repeated binge drinking on blood pressure levels and other cardiovascular health metrics in young adults: National Health and Nutrition Examination Survey, 2011-2014. J Am Heart Assoc 2018;7:e008733.

Puddey IB, Mori TA, Barden AE, Beilin LJ.  Alcohol and hypertension – New insights and lingering controversies. Curr Hypertens Rep 2019;21:79.

Roerecke M, Kaczorowski J, Tobe SW, Gmel G, Hasan OSM, Rehm J.  The effect of a reduction in alcohol consumption on blood pressure: a systematic review and meta-analysis. Lancet Public Health 2017;2:e108-e120.

Scholz A, Navarrete-Munoz EM, Garcia dlH, Gimenez-Monzo D, Gonzalez-Palacios S, Valera-Gran D, et al.  Alcohol consumption and Mediterranean Diet adherence among health science students in Spain: the DiSA-UMH Study. Gac Sanit 2016;30:126-132.

Forum Summary

The authors of this paper test the hypothesis that overall diet quality influences associations between alcohol consumption and the risk of hypertension, based on data from a study in the USA (NHANES) and two studies in China.  In general, their designation of a “high-quality” diet for Chinese was one with greater amounts of fruit, vegetables, and beans, while for Caucasians and Hispanics it was one with more whole grains, fruit, vegetables and nuts; “low-quality” diets were mainly those with higher intakes of red meat products.  For comparisons with abstainers, subjects reporting < 7 drinks/week for total alcohol were categorized as light and those reporting 7-21 drinks/week categorized as moderate drinkers, with those reporting more referred to as heavy drinkers.

The paper concludes that light alcohol consumption among subjects with a high-quality diet was significantly associated with a decreased risk of hypertension in Caucasians, and the risk of hypertension was not significantly increased with increasing alcohol consumption in Caucasians, Chinese, and Hispanics.  On the contrary, in context of a diet calculated to be low quality the risk of hypertension was significantly increased in a linear fashion with increasing alcohol consumption in all groups except African Americans.

Forum reviewers considered this to be a generally well-done and interesting study, but had concern that the authors may not have adjusted adequately for a factor strongly related to diet quality: socio-economic status.  The authors included in their analyses whether or not subjects had a high school diploma, but no adjustments were made for occupation, income, or other important indices of deprivation, which may be important since lower SES subjects have consistently been shown to have poorer diets.  Thus, we are unsure whether or not it is the diet solely that is the reason that alcohol consumption was found to be associated with lower, higher, or no effect depending on the type of diet.  Further, no results are shown indicating whether or not there were differences if the beverage consumed was beer, wine, or spirits, and there was insufficient data on the pattern of drinking.  The authors begin their abstract saying that ‘Alcohol is commonly consumed around mealtimes,’ but apparently do not have any data on whether the beverage was consumed with or without food for the present analyses.”  All of these factors may modify how alcohol affects the risk of developing hypertension.

Finally, some Forum members were uneasy with the term “quality of diet”, based on the inclusion or exclusion of certain foods, which may be very important to taste.  Said one member, we trust that diets recommended as being of “high quality” are not those whose savor has been removed.

*                 *                     *                   *                   *                  *                 *

This critique by the International Scientific Forum on Alcohol Research has been based on comments by the following members:          

Ian Puddey, MD, Emeritus Professor, Faculty of Health & Medical Sciences, University of Western Australia, Nedlands, Australia

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

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

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

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

Giovanni de Gaetano, MD, PhD, Department of Epidemiology and Prevention, IRCCS Istituto Neurologico Mediterraneo NEUROMED, Pozzilli, Italy (with input from colleague Marialaura Bonaccio of the Fondazione di Ricerca e Cura Giovanni Paolo II, and Epicomed Research, Campobasso, Italy)

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

Ramon Estruch, MD, PhD, Hospital Clinic, IDIBAPS, Associate Professor of Medicine, University of Barcelona, Spain

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

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