Relationship of Alcohol Drinking Pattern to Risk of Hypertension
A Population-Based Study
Epidemiological studies have demonstrated a positive relationship between heavy alcohol use and hypertension, but few studies have directly addressed the role of drinking pattern. This study was designed to investigate the association of current alcohol consumption and aspects of drinking pattern with hypertension risk in a sample of 2609 white men and women from western New York, aged 35 to 80 years, and free from other cardiovascular diseases. Hypertension was defined by systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg or use of antihypertensive medication. Odds ratios (95% confidence intervals) were computed after adjustment for several covariates. Compared with lifetime abstainers, participants reporting drinking on a daily basis (1.75 [1.13 to 2.72]) or mostly without food (1.64 [1.08 to 2.51]) exhibited significantly higher risk of hypertension. When analyses were restricted to current drinkers, daily drinkers and participants consuming alcohol without food exhibited a significantly higher risk of hypertension compared with those drinking less than weekly (1.65 [1.18 to 2.30]) and those drinking mostly with food (1.49 [1.10 to 2.00]), respectively. After additional adjustment for the amount of alcohol consumed in the past 30 days, the results were follows: 0.90 (0.58 to 1.41) for daily drinkers and 1.41 (1.04 to 1.91) for drinkers without food. For predominant beverage preference, no consistent association with hypertension risk was found across the various types of beverages considered (beer, wine, and liquor). In conclusion, drinking outside meals appears to have a significant effect on hypertension risk independent of the amount of alcohol consumed.
The relationship between heavy alcohol consumption and blood pressure elevation is well documented.1,2 In the majority of studies, the assessment of alcohol has been focused primarily on average quantity of alcohol consumed during a period of time. The many different and complex components of drinking, such as the frequency and setting of consumption, have not been sufficiently addressed. It has been suggested that the way in which alcohol is consumed may have important implications for health and, in particular, for cardiovascular disease and cardiovascular risk factors.3–6 Few studies have specifically evaluated the possible effect of pattern of alcohol consumption on blood pressure elevation;7–9 the majority of these studies have given attention to the role of drinking frequency, however, providing conflicting results. Only a study on a large sample of Italian men and women has examined the association between drinking pattern in relation to food consumption and hypertension risk, reporting a higher prevalence of hypertension in individuals consuming wine outside meals compared with drinkers of wine with meals.7
This study was thus conducted to investigate the relationship between pattern of alcohol use and hypertension risk in a general-population sample of men and women. In particular, in addition to the amount of alcohol consumed, our study focused on frequency of drinking and in relation to food consumption and beverage preference (beer, wine, and liquor).
The present study is based on data obtained from a randomly chosen sample of residents of Erie and Niagara counties in New York. The overall sample, identified as the “Western New York Health Study,” was collected between September 1995 and May 2001 as part of a series of studies specifically addressed to examine the complex issue of alcohol drinking pattern and chronic disease risk. The details of the overall study design, participant enrollment, and methodology have been described previously.5 The study protocol was approved by the University at Buffalo institutional review board.
Of the 6837 potential participants identified, contacted, and assessed eligible for our study, a total of 4065 (59.5%) agreed to participate. The exclusion criteria applied to the present analyses were race other from white (n=381), a self-reported history of prevalent coronary heart disease (previous myocardial infarction, coronary artery bypass graft surgery, angioplasty, or diagnosed angina pectoris; n=502), a self-reported history of diabetes (n=221), missing blood pressure measurements (n=63), missing information on drinking habits (n=29), or missing data for various covariates (n=260). The remaining 2609 participants, aged 35 to 80 years, are included in this analysis.
Blood pressures were measured by a trained interviewer 3 times on each individual in a sitting position with a standard mercury sphygmomanometer according to a standardized protocol.10 The mean of the second and third systolic and diastolic blood pressure measurements were used. Hypertension was defined by systolic blood pressure ≥140 mm Hg, or diastolic blood pressure ≥90 mm Hg, or use of medication for hypertension.11
Information about alcohol intake was obtained with a computer assisted in-person interview.12,13
For the present study, the 30 days before the interview date were used as a measure of participants’ most current alcohol consumption. Quantity–frequency questions on alcohol drinking were asked for Fridays, Saturdays, Sundays, and weekdays. The interview also addressed the proportion of drinking that took place with a meal, while snacking, or with no food. The responses to these mutually exclusive questions had to sum to 100%. Responses to the above questions were used to compute the following variables for the present analyses: (1) lifetime abstainers: participants who reported consumption of <12 drinks during their lifetime or in any 1-year period; (2) noncurrent drinkers: participants who reported ≥12 drinks during their lifetime or in any 1-year period but did not consume an alcoholic beverage at least once in the past 30 days; and (3) current drinkers: those who did consume at least 1 alcoholic beverage in the past 30 days. Average volume of alcohol consumed in the 30 days before the interview in drinks per day was calculated. A drink was defined as a 355-mL beer, 148-mL wine, and 37-mL liquor. The number of drinks consumed per drinking day in the past 30 days was used as a measure of drinking intensity. For pattern of drinking in relation to drinking frequency during the week, the following categories were assessed: (1) less than weekly drinkers: current drinkers who reported consuming alcoholic beverages less than once per week; (2) weekly nondaily drinkers: current drinkers who reported drinking weekly but less than daily; (3) weekend only drinkers: current drinkers who reported consuming alcoholic beverages only on weekends (Friday, Saturday, or Sunday) and never or hardly ever on weekdays; and (4) weekly daily drinkers: current drinkers who reported drinking on a daily basis.
For pattern of drinking in relation to food consumption, we computed the following categories: (1) drinkers of alcohol with food: current drinkers who reported consuming their alcoholic beverages with meals or snacks ≥75% of the time; (2) drinkers of alcohol with no food: current drinkers who reported consuming their alcoholic beverages without food ≥75% of the time; and (3) drinkers with a mixed pattern: current drinkers who did not match any of the above categories.
Finally, predominant beverage consumption was defined as drinkers of beer, wine, or liquor (consumption of that beverage ≥75% of all drinking occasions) and mixed beverage consumption (no single beverage consumed ≥75% of all drinking occasions). Because wine coolers were rarely consumed, they were included in the wine category.
For the last 2 patterns, the 75% cut point was used because it ensured reasonable sample sizes in all categories. Results, however, were consistent across a wide range of cut points.
All analyses were conducted using the Statistical Package for Social Sciences (11.5; SPSS). Tests for interaction among all drinking patterns and gender, as well as among women between drinking patterns and menopausal status, were not significant for hypertensive status. Therefore, all analyses were conducted combining men and women. Among current drinkers, categorical variables were created for average amount of alcohol consumed in the 30 days before the interview (drinks per day). Unadjusted and adjusted prevalence of hypertension was calculated for the different aspects of alcohol drinking using the general linear model procedure (analysis of covariance) for multiple comparisons among the various categories of alcohol variables.
Multiple logistic regression was used to calculate odds ratios and 95% confidence intervals, with hypertensive status as the dependent variable. Various alcohol drinking categories served as independent variables, and a number of potential confounding factors were taken into consideration (age, education, gender, smoking status, body mass index, abdominal height, and physical activity). Two sets of analyses were performed. In the first set, lifetime abstainers were the reference category. In the second set, only those participants who were current drinkers were included, and the reference categories varied according to the specific drinking pattern analysis. The second analysis enabled adjustment for average amount of alcohol consumed.
For drinking status, drinkers, noncurrent and current, had a higher adjusted prevalence of hypertension than lifetime abstainers but not at a significant extent. For categories of average alcohol intake in the previous 30 days, heavier drinkers (≥2 drinks per day) had significantly higher unadjusted and adjusted prevalence of hypertension compared with lifetime abstainers or other drinking categories. For frequency of drinking, daily drinkers exhibited a significantly higher adjusted prevalence of hypertension compared with lifetime abstainers and less than weekly drinkers, whereas for drinking pattern in relation to food consumption, participants drinking mostly without food exhibited a significantly higher adjusted prevalence of hypertension compared with either lifetime abstainers or those drinking mostly with food. For predominant beverage consumption, liquor drinkers on average showed a higher adjusted prevalence of hypertension (with borderline significance) than lifetime abstainers (P=0.054) and wine drinkers (P=0.071).
These findings were confirmed when analyses were restricted to current drinkers only, whereas, after further adjustment for alcohol, only participants who drank outside mealtimes still reported a significantly higher prevalence of hypertension compared with participants drinking mostly with food.
Adjusted odds ratios of prevalent hypertension by various drinking categories are reported in Table 3. In the first set of logistic regression analyses, lifetime abstainers were used as reference category. The models included the following covariates: age, education, gender, smoking status, body mass index, abdominal height, and physical activity. Noncurrent drinkers and current drinkers exhibited higher odds ratios compared with lifetime abstainers but not at a significant extent. For categories of average alcohol intake in the past 30 days, only participants in the highest category of consumption (≥2 drinks per day) showed a significantly higher risk of hypertension compared with lifetime abstainers (2.31 [1.47 to 3.62]). With regard to drinking patterns, participants who reported drinking either on a daily basis (1.75 [1.13 to 2.72]) or mostly without food (1.64 [1.08 to 2.51]) exhibited significantly higher risk of hypertension compared with lifetime abstainers, whereas liquor drinkers (1.50 [0.98 to 2.31]) showed a borderline significant (P=0.062) higher risk of hypertension compared with lifetime abstainers.
When analyses were restricted to current drinkers only, participants in the top category of alcohol consumption still reported a significantly higher odds ratio compared with participants in the lowest category (2.07 [1.50 to 2.86]). For drinking patterns, daily drinkers and participants consuming alcohol without food again exhibited a significantly higher risk of hypertension compared with those drinking less than weekly (1.65 [1.18 to 2.30]) and those drinking mostly with food (1.49 [1.10 to 2.00]), respectively.
When these analyses were conducted after adjustment for average amount of alcohol consumed in the past 30 days, the results were follows: 0.90 (0.58 to 1.41) for daily drinkers, 1.41 (1.04 to 1.91) for drinkers without food, and 1.27 (0.90 to 1.80) for liquor drinkers.
Finally, after further subanalysis within categories of drinking volume (Table 4), drinking without food was associated with a strong significant (P=0.028) increased risk of hypertension (1.45 [1.04 to 2.02]) even in individuals with light to moderate alcohol intake (<2 drinks per day).
To ensure that our findings were not influenced by potential changes in drinking habits subsequent to the initiation of antihypertensive drug therapy, all analyses were repeated after the exclusion of participants on medication for hypertension. Results were consistent with those reported in analyses including participants on medication for hypertension (data not shown).
In addition, when analyses were repeated using drinking intensity (drinks per drinking day) instead of volume as a covariate, and also after exclusion of participants who reported consumption of alcohol in the 24 hours before the examination, findings were not substantially affected (data not shown).
The findings of this population-based study confirm that high levels of alcohol consumption (≥2 drinks per day) are associated with an increased risk of hypertension. We also found that some aspects of drinking pattern may affect the risk of hypertension. In particular, those drinking mostly outside mealtimes reported a significant increase in the risk of hypertension compared with either lifetime abstainers or those drinking mostly with food. This association between pattern of drinking in relation to food consumption and hypertension was independent of the effects of the average amount of alcohol consumed and was present even in individuals with light to moderate alcohol intake. The latter is a novel finding, with potentially important clinical implications. In fact, it points out that drinking without food may counteract any benefit associated with moderate alcohol consumption (<2 drinks per day) on the cardiovascular system.
For pattern of drinking in relation to frequency, drinkers on a daily basis showed a significantly higher risk of hypertension than lifetime abstainers and less than weekly drinkers. However, this association disappeared after accounting for the effects of the average amount of alcohol consumed in the previous 30 days. This may suggest that the average volume of alcohol consumed plays a much more important role in the relationship between alcohol consumption and hypertension risk than the frequency of occasion of use does. We did not find any consistent association between predominant beverage preference and hypertension risk.
To our knowledge, the relationship between drinking pattern in relation to food consumption and hypertension risk has been addressed only in Italian data. Trevisan et al, in a large sample of adult men and women from the Italian Nine Communities Study,7 found that drinkers of wine with and without meals experienced significantly higher systolic blood pressure levels and a higher prevalence of hypertension than other drinkers, even after adjustment for differences in alcohol consumption among the various drinking pattern categories. This association was similar in both sexes. In a more recent follow-up of this cohort, it was shown that drinking outside meals was associated with a higher risk of death from all cause and noncardiovascular disease compared with drinking with meals.3 Further, a very recent study by the same principal author, using a case-control design, revealed an increased risk of myocardial infarction in men reporting consuming most of their alcoholic beverages outside mealtimes compared with those reporting consuming alcohol with meals or snacks.6 Finally, Foppa et al found in a controlled randomized trial that moderate consumption of wine (2 glasses of red wine) with the noon meal produced a reduction of the postprandial blood pressure levels in centrally obese, hypertensive individuals.14 The results of our study using a population-based sample in the United States confirm those reported in Italian data from a population characterized by a substantially different drinking culture. We found that participants who drank mostly outside mealtimes had a significantly higher risk of hypertension compared with lifetime abstainers and those drinking mostly with meals or snacks. This association was independent of the effects of the average amount of alcohol consumed and of drinking intensity, as measured by number of drinks per drinking day, and was present even in individuals with light to moderate alcohol intake (<2 drinks per day). We did not detect gender differences in the effect of pattern on hypertension risk, consistent with the previous report on Italian data.7
Several plausible physiological mechanisms may be put forward to explain the observed association between drinking pattern in relation to food consumption and hypertension risk. Drinking with meals has been shown to exert beneficial effects on fibrinolysis15 and lipids.16 It has also been reported that consumption of red wine with meals reduces the susceptibility of human plasma and LDL to lipid peroxidation.17 A potential explanation may reside as well in the effect of food on absorption and intragastric metabolism of ethanol leading to a slower increase and lower peak of blood alcohol concentration.18 Another physiological mechanism may be that food intake produces increased alcohol elimination rates.19 Finally, drinking alcohol with or without food may just represent a marker of lifestyle profile and health-related behaviors, which may itself be the link to protective and detrimental health effects, such as different dietary habits.20,21 Unfortunately, in our study, we did not assess current dietary habits.
During the last decade, frequency of alcohol consumption has been the focus of several investigations aimed at assessing the alcohol–blood pressure relationship.2,4,8,9 However, to date, studies have provided contrasting findings.
In our study, participants who drank on a daily basis had a significantly higher risk of hypertension compared with lifetime abstainers and less than weekly drinkers. However, when differences in the average amount of alcohol consumed across the various drinking pattern categories were taken into account, the relationship was no longer significant. It is possible that by taking into account average volume, we masked much of the “frequency effect” because most daily drinkers tend not to be light drinkers, and most light drinkers do not drink every day,22 as supported as well by our data. In our sample, drinking frequency was somewhat correlated to amount of drinking (Pearson correlation coefficient 0.39). The ideal approach to address this issue would be to analyze the relationship between drinking frequency and hypertension within homogeneous categories of drinkers with regard to amount. Unfortunately, our sample size precluded us from carrying out meaningful comparisons within these subsets of drinkers.
Our study is not without limitations. First of all, the cross-sectional design does not allow us to make any conclusive statement about the temporality of the observed associations. Furthermore, the suboptimal participation rate (59.5%) may leave the possibility for selection bias and restrict the generalization of our findings to the general public. The strength of this study consists in the very detailed information elicited on the many different and complex components of alcohol consumption in terms of amount and pattern, as well as the examination of several covariates known to be related to blood pressure elevation. Nevertheless, we cannot rule out the presence of additional unknown confounding variables (eg, diet and other lifestyle habits) that we were unable to control for in our analyses and the potential of residual confounding that, in the absence of a known physiological link, may have contributed to our findings concerning pattern of drinking.
Our study adds new and important information to the current body of evidence about the role of drinking pattern on cardiovascular health, and specifically that certain aspects of drinking pattern (ie, drinking outside mealtimes) are associated with increased risk of hypertension. Longitudinal studies are necessary to confirm the temporality of this association. However, these findings may have clinical and public health implications from the perspective of recommendations given to patients and the population at large with regard to drinking habits and lifestyle modifications necessary for good health. Our study also points out the need to incorporate more sophisticated approaches to the assessment of alcohol use in the epidemiological research focusing on the complex issue of alcohol and health.
This study was supported in part by grant 5 P50 AA09802 from the National Institute on Alcohol Abuse and Alcoholism.
- Received June 23, 2004.
- Revision received July 9, 2004.
- Accepted September 21, 2004.
Marmot MG, Elliott P, Shipley MJ, Dyer AR, Ueshima H, Beevers DG, Stamler R, Kesteloot H, Rose G, Stamler J. Alcohol and blood pressure: the INTERSALT study. BMJ. 1994; 308: 1263–1267.
Murray RP, Connett JE, Tyas SL, Bond R, Ekuma O, Silversides CK, Barnes GE. Alcohol volume, drinking pattern, and cardiovascular disease morbidity and mortality: is there a U-shaped function? Am J Epidemiol. 2002; 155: 242–248.
Dorn JM, Hovey K, Muti P, Freudenheim JL, Russell M, Nochajski TH, Trevisan M. Alcohol drinking patterns differentially affect central adiposity in women and men. J Nutr. 2003; 133: 2655–2662.
Trevisan M, Krogh V, Farinaro E, Panico S, Mancini M. Alcohol consumption, drinking pattern and blood pressure: analysis of data from the Italian National Research Council Study. Int J Epidemiol. 1987; 16: 520–527.
Meinert CL, ed. Manual of operations for “INTERSALT,” an international cooperative study on the relation of sodium and potassium to blood pressure. In: Controlled Clinical Trials: Design, Methods and Analysis. 1988; 75–125.
Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL Jr, Jones DW, Materson BJ, Oparil S, Wright JT Jr, Roccella EJ; National Heart, Lung, and Blood Institute Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure; National High Blood Pressure Education Program Coordinating Committee. The seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: the JNC 7 report. J Am Med Assoc. 2003; 289: 2560–2572.
Hutchnison H, ed. Sawtooth: Ci3 System User Manual. Version I.I. Evanston, IL: Sawtooth Software; 1995.
Russell M, Marshall JR, Trevisan M, Freudenheim JL, Chan AWK, Markovic N, Vana JE, Priore RL. Test-retest reliability of the cognitive lifetime drinking history. Am J Epidemiol. 1997; 146: 975–981.
Hendriks HF, Veenstra J, Velthuis-te Wierik EJ, Schaafsma G, Kluft C. Effect of moderate dose of alcohol with evening meal on fibrinolytic factors. BMJ. 1994; 308: 1003–1006.
Marques-Vidal P, Montaye M, Haas B, Bingham A, Evans A, Juhan-Vague I, Ferrieres J, Luc G, Amouyel P, Arveiler D, Yarnell J, Ruidavets JB, Scarabin PY, Ducimetiere P. Relationships between alcoholic beverages and cardiovascular risk factor levels in middle-aged men, the PRIME study. Atherosclerosis. 2001; 157: 431–440.
Kesse E, Clavel-Chapelon F, Slimani N, van Liere M. Do eating habits differ according to alcohol consumption? Results of a study of the French cohort of the European Prospective Investigation into Cancer and Nutrition (E3N-EPIC). Am J Clin Nutr. 2001; 74: 322–327.