Alcohol and cardiovascular disease: still a research priority?

Alternative explanations for the protective effect of moderate alcohol intake, relying on ever more tenuous confounding effects, have been discredited one by one. The apparently protective effect of moderate alcohol consumption has so far survived the use of controls for sociodemographic status,³ for the "sick quitter hypothesis"⁴ (ie, the suggestion that many abstainers have stopped drinking because of serious illness), for the amount of cholesterol in the diet,⁵ and even for the degree of social isolation.⁶ As evidenced in the study by Simons et al,² the protective effect is fairly specific to cardiovascular disease and does not operate for other major causes of death in older people such as cancer. While a handful of recent studies have failed to find a protective effect for moderate drinking,^7,8 these are still heavily outnumbered by those with positive findings.¹ In fact, the range of different countries and cultures in which the phenomenon has been documented is also testimony to its robustness, even if different levels of consumption appear to provide the benefit in different drinking cultures.¹

So, does medical science need further research on this topic? My contention is that, while the basic protective properties of moderate alcohol consumption appear to have been identified, the precision of the measurement of drinking levels and patterns in these studies needs to be sharpened if we are to have a firmer basis for advising people how to drink to avoid ill-health. The recent major systematic review commissioned by the National Health and Medical Research Council to underpin revisions to Australia's national guidelines on low risk drinking found that nearly all epidemiological studies in this area only attempt to measure one of the important dimensions of alcohol consumption: total volume of drinking, usually expressed as average intake per day.¹

Despite recent evidence that pattern of drinking plays a role independent of volume,⁹ large-scale epidemiological studies rarely include simple items in their questionnaires tapping this dimension, such as frequency of drinking five or more drinks in one day, or maximum amount consumed on one day. Simons et al make a rare contribution to our knowledge by providing an analysis of mortality risk, based not only on average volume of alcohol across all days, but also on usual amount consumed on a drinking day. Clearly, these can be very different measures (eg, seven drinks in a day once a week versus one drink every day of the week).² While the power of the analyses in Simons et al is limited by sample size, usual consumption of five or more drinks for men and of three or four drinks in a day for women was not associated with a significantly reduced risk of death.²

It should be noted, however, that questions regarding "usual" consumption tend to suffer from a bias towards low-consumption occasions and against less frequent occasions of high intake.¹⁰ Forthcoming World Health Organization guidelines on measurement of alcohol consumption advise that a superior method is the "graduated quantity frequency", in which respondents are asked how often they drink at each of different levels of consumption, starting with the highest (eg, "How often do you drink 20 drinks on one day?").¹⁰ The same guidelines identify another problem bedevilling attempts to convert the results of alcohol studies into precise advice for drinkers: assumptions about the alcohol content of drinks reported vary between studies and are usually not empirically based. A number of studies from different countries have attempted to document usual-serve sizes employed by random samples of drinkers and found these to vary significantly from those usually assumed by researchers.^11,12 One study found that the bias created towards under-reporting of consumption was massive in one particular population subgroup: Afro-American women.¹³

One striking consequence of a failure to measure pattern of alcohol consumption adequately can be the false identification of special benefits from one type of alcoholic beverage over another. Wine is often reported as being most associated with benefits, but being a wine drinker as opposed to a beer and spirits drinker is a marker for many other things, including a tendency towards a more consistent pattern of daily drinking rather than occasional "bingeing".¹⁴ A well known Danish study,³ often cited as evidence for a greater benefit of wine over other drinks, measured alcohol consumption by only asking about how much people usually drank if they drank every day. Because this is a less frequent pattern for heavy drinkers of beer and spirits, many of the latter will have been falsely categorised as light or moderate drinkers, thus significantly biasing against finding protective effects for these beverages.

The two fields of epidemiology and alcohol studies have much to learn from each other. If epidemiological studies of risk factors for heart disease do not improve their measurement of patterns and levels of alcohol use, it will remain very hard to give precise advice to drinkers who wish to minimise harms and maximise benefits of alcohol consumption. In reality, despite hundreds of studies into the protective effects of alcohol in relation to heart disease, research into this area has only just begun.

Timothy R Stockwell
Director
National Centre for Research into the Prevention of Drug Abuse
Curtin University of Technology, Perth, WA

1. Single E, Ashley MJ, Bondy S, et al. Evidence regarding the level of alcohol consumption considered to be low-risk for men and women. Final report. Canberra: National Health and Medical Research Council, 2000. URL: <http://www.nhmrc.health.gov.au/advice/alc-comp.htm> (accessed 29 June 2000).
2. Simons LA, McCallumJ, Friedlander Y, et al. Moderate alcohol intake is associated with survival in the elderly: the Dubbo Study. Med J Aust 2000; 173: 121-124.
3. Gronbaek M, Deis A, Sorensen TIA, et al. Mortality associated with moderate intakes of wine, beer or spirits. BMJ 1995; 310: 1165-1169.
4. Rehm J, Sempos CT. Alcohol consumption and all-cause mortality: questions about causality, confounding and methodology. Addiction 1995; 90: 493-498.
5. Rehm J, Sempos CT. Alcohol consumption and all-cause mortality. Addiction 1995; 90: 471-480.
6. Murray RP, Rehm J, Shaten J, Connett JE. Does social integration confound the relation between alcohol consumption and mortality in the Multiple Risk Factor Intervention Trial (MRFIT)? J Stud Alcohol 1999; 60: 740-745.
7. Leino EV, Romelsjo A, Shoemaker C, et al. Alcohol consumption and mortality. II. Studies of male populations. Addiction 1998; 93: 205-218.
8. Hart CL, Smith GD, Hole DJ, Hawthorne VM. Alcohol consumption and mortality from all causes, coronary heart disease, and stroke: results from a prospective cohort study of Scottish men with 21 years follow up. BMJ 1999; 318: 1725-1729.
9. Rehm J, Ashley KJ, Room R, et al. On the emerging paradigm of drinking patterns and their social and health consequences. Addiction 1996; 91: 1615-1621.
10. World Health Organization (Substance Abuse Department). International guidelines for monitoring alcohol consumption and harm. Geneva: WHO. In press.
11. Stockwell T. Information provided in Australia about the size of "standard drinks". Med J Aust 1992; 156, 295.
12. Lemmens P. The alcohol content of self-report "standard drinks". Addiction 1994; 89: 593-602.
13. Kaskutas L, Graves K. An alternative to standard drinks as a measure of alcohol consumption. Paper presented at International Conference on the Measurement of Drinking Patterns, Alcohol Problems and the Connection; 2000 April 2-7; University of Stockholm, Sweden. In press.
14. Doll R. One for the heart. BMJ 1997; 315: 1664-1668.

Make a comment