Effect of beer drinking on risk of myocardial infarction: population based case-control study
Many studies have shown an inverse association between
alcohol consumption and coronary heart disease, with a possible
flattening at higher consumption levels. 1Participants, methods, and results
We conducted a population based case-control study in five Czech
districts. All men aged 25-64 who had a first non-fatal myocardial
infarction that fulfilled the World Health Organization's MONICA
(monitoring trends and determinants in cardiovascular disease) criteria
of definite or probable infarction2 over 18 months were
considered eligible. All cases agreed to participate in the study. An
age stratified random sample of the population (response rate 77%)
served as controls. Data on cases and controls were collected by
identical protocols (details are available elsewhere3).
Participants reported the frequency of drinking any alcohol (never;
less than once a month; once or twice a month; several times a week;
almost daily or daily; and twice a day or more often). They also
reported how much wine, spirits, and beer they consumed during a
typical week. The average consumption of pure alcohol was 148 g a week,
87% of which was consumed as beer. The analyses were restricted to
non-drinkers and "exclusive" beer drinkers (men who typically do
not drink wine or spirits). Participants were categorised into four
groups according to their average weekly intake of beer: <0.5 l
(about 18 g of alcohol), including non-drinkers; 0.5-3.9 l (18-144
g of alcohol); 4-8.9 l (145-324 g of alcohol); and >=9 l (325 g
of alcohol).
|
| Numbers of cases and controls (non-drinkers or only beer drinkers) and odds ratios (95% confidence intervals) of non-fatal myocardial infarction for drinking frequency and average weekly beer consumption |
|
|
| |
No of cases/controls |
Odds ratio adjusted for age and district |
Fully adjusted odds ratio* |
|
|
| Frequency of drinking |
|
|
| Never | 30/63 | 1.0 | 1.0
|
|
|
| Less than once a month | 23/48 | 0.90 (0.43
to 1.87) | 1.14 (0.52 to 2.51) |
|
|
| Less than once a
week | 26/81 | 0.65 (0.32 to
1.29) | 0.62 (0.29 to 1.33) |
|
|
| Several times a
week | 68/276 | 0.56 (0.32 to
0.98) | 0.60 (0.32 to 1.12) |
|
|
| Almost daily or
daily | 37/234 | 0.37 (0.20 to
0.68) | 0.38 (0.19 to 0.74) |
|
|
| Twice a day or
more | 15/31 | 1.04 (0.45 to
2.37) | 0.99 (0.41 to 2.38) |
|
|
| Average weekly beer consumption
|
|
|
| <0.5 l | 77/181 | 1.0 | 1.0
|
|
|
| 0.5 to 3.9 l | 88/325 | 0.68 (0.46 to
1.00) | 0.65 (0.42 to 1.00) |
|
|
| 4 to
8.9 l | 24/178 | 0.38 (0.22 to
0.65) | 0.34 (0.19 to 0.61) |
|
|
| >=9 l
| 13/51 | 0.65 (0.32 to
1.33) | 0.54 (0.25 to 1.14) |
|
|
| *Adjusted for age, district, education, smoking, waist to hip ratio, and personal history of diabetes and high cholesterol concentration. |
The lowest risk was found among men who drank almost daily or daily
(adjusted odds ratio 0.38, 95% confidence interval 0.19 to 0.75) and
among men who drank 4-8.9 l of beer a week (0.34, 0.19 to 0.61)
(table). When beer intake was analysed in narrower categories, the
lowest risk was found for weekly consumption of 5-6 l, but because of
the small numbers of subjects in each category the confidence intervals
were wide (data not shown). The results did not change when men with a
history of heart disease, stroke, diabetes, or cancer were
excluded.
Comment
In this study of beer drinkers, the lowest risk of
myocardial infarction was found among men who drank almost daily or
daily and who drank 4-9 l of beer a week. There was a suggestion that
the protective effect was lost in men who drank twice a day or
more. This is similar to results of studies of other
beverages.
It is unlikely that our results are due to bias or confounding. This
was a population based study with highly complete recruitment of
incident cases through a myocardial infarction register in a well
defined population and with good response rate in controls randomly
selected from the population register. 3 Questions on
average consumption usually lead to underestimation of the real intake,
but the ranking of subjects in terms of long term average intake is
reasonably reliable. 4 Restricting the analysis to
exclusive beer drinkers eliminated potential confounding by other
beverages. It is unlikely that cases and controls answered questions
differently; a cohort study in Bavaria, another beer drinking region,
produced similar findings. 5 These results support the view
that the protective effect of alcohol intake is due to ethanol rather
than to specific substances present in different types of
beverages. 1
Critical appraisal: points to note
What does this paper tell you?
Very good news indeed, on the face of it. We learn that the
daily drinking of beer and the consumption of 4 to 9 litres of beer per
week is associated with a lower risk of heart attack than abstinence or
lower levels of consumption. Higher levels than this, though, and the
risk begins to climb again. As ever with case control studies, do not
be tempted to attribute causes and effects. The study cannot say that
it is beer drinking that reduces the risk of MI or that a reduced risk
of MI causes increased beer drinking or that other (confounding)
factors cause both.
Should you believe it?
No, not on the strength of this one study. Firstly, if you look
back at the methods paper,1 only people with
non-fatal infarcts were considered (obviously, since people are asked
about their beer consumption after their MI). Since a large proportion
of MIs are fatal, the conclusion that moderate beer drinking is
associated with a lower risk of MI is misleading-what if there is a
high rate of fatal MI in moderate beer drinkers? The study would never
detect this.
So the conclusions must be limited to non-fatal MI only. Now they are
more believable, but we must beware of some difficulties. The study
works by taking people with the condition (cases) and people without
(controls). It then looks back (retrospective analysis) to find
consistent differences in drinking habits between the two groups.
However, the study must try to remove other differences between the
groups that may mask the effect of the drinking, such as age, body
weight, smoking, etc. The best way to do this is to select the controls
at the outset to match the cases (at least on average). This is usually
very difficult, so statistical techniques are used to pull apart the
differences during analysis (hence the "adjusted" figures in this
paper). The effectiveness of these techniques is debatable, and
confounding (masking of the true associations) can still occur. In this
study, the controls were on average much younger than the cases. It
might be that younger people drink more than older people (I don't
know this, but it's a possibility). If this were the case, it would
appear that drinking more was associated with lower risk of MI, but in
actual fact the association may be between age and MI risk. The study
says that it has adjusted for this, and we must take it on trust it has
done so correctly.
Another problem is that there may be some systematic difference in the
way people who have had heart attacks report their beer intake,
compared with those who have not. The effect could work both
ways-either group may consistently underreport or exaggerate its
intake, leading to a false association.
So, all in all, this study cannot be seen as conclusive. It represents
very limited evidence that moderate beer consumption may be associated
with a reduced rate of non-fatal MI. But for the deadly heart attack,
it leaves us totally in the dark.
Vivek Muthu BMJ Clinical Evidence
Bobak M, Skodova Z, Hertzmann C, Marmot M. Own education, current conditions, parental material circumstances and risk of myocardial infarction in a former communist country. J Epidemiol Community Health 2000;54:91-6.
We thank local cardiologists in the participating
districts.
Contributors: MB, ZS, and MM jointly
designed the case-control extension of the Czech MONICA project. MB
analysed the data and drafted the paper. ZS coordinated the data
collection and participated in the interpretation of the data and
writing of the paper. MM initiated the project and participated in data
interpretation and writing of the paper. MB will act as guarantor.
Funding: The study was funded by a grant from the
Wellcome Trust and by the Czech Ministry of Health. MB was supported by
the Wellcome Trust fellowship in clinical epidemiology. MM is supported
by an MRC research professorship.
Competing interests: None declared.
Martin Bobak, senior lecturer
Email: martinb@public-health.ucl.ac.uk
Michael Marmot, professor, International Centre for Health and Society, Department of Epidemiology and Public Health, University College London, London WC1E 6BT
Zdenka Skodova senior researcher, Department of Preventive Cardiology, Institute of Clinical and Experimental Medicine, 140 00 Prague, Czech Republic
studentBMJ 2000;08:217-258 July ISSN 0966-6494
- Rimm EB, Klatsky A, Grobbee D, Stampfer MJ. Review of moderate alcohol consumption and reduced risk of coronary heart disease: is the effect due to beer, wine, or spirits? BMJ 1996;312:731-6.
- World Health Organization. Multinational monitoring of trends and determinants of cardiovascular diseases-"MONICA project." Manual of operations. Version 1.1. CDV/MNC. December 1986. Geneva: WHO, 1987.
- Bobak M, Skodova Z, Hertzman C, Marmot M. Own education, current conditions, parental material circumstances and risk of myocardial infarction in a former communist country. J Epidemiol Community Health 2000;54:91-6.
- Rehm J, Greenfield TK, Walsh G, Xie X, Robson L, Single E. Assessment methods for alcohol consumption, prevalence of high risk drinking and harm: a sensitivity analysis. Int J Epidemiol 1999;28:219-24.
- Keil U, Chambless LE, Doering A, Filipiak B, Stieber J. The relation of alcohol intake to coronary heart disease and all-cause mortality in a beer-drinking population. Epidemiology 1997;8:150-6.
