IJE Advance Access originally published online on January 13, 2005
International Journal of Epidemiology 2005 34(1):205-206; doi:10.1093/ije/dyi003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
IJE vol.34 no.1 © International Epidemiological Association 2005; all rights reserved.
Commentary |
Commentary: Alcohol and mortality: diminishing returns for benefits of alcohol
Department of Primary Care and Population Science, Royal Free and University College Medical School, London NW3 2PF, UK. E-mail: goya{at}pcps.ucl.ac.uk
While no one doubts that heavy consumption of alcohol is harmful the concept that ‘a little bit of what you fancy does you good’ has caused much controversy in the medical world. Ever since Pearl1 in 1926 first showed a shallow U-shaped mortality curve in relation to alcohol consumption, it has been a common finding in epidemiological studies that non-drinkers have a higher mortality for all causes and in particular for coronary heart disease (CHD) than light or moderate drinkers.2–4 This has been interpreted as alcohol being protective against heart attacks and that there may even be an overall protection against all cause mortality. It has been customary to use non-drinkers as the baseline group against which the effects of alcohol on health and disease have been determined. An alternative explanation (the ‘reverse causation’ hypothesis) is that people who develop illnesses, often requiring medication, tend to change their drinking habits towards occasional or non-drinking status.5 Thus, non-drinkers are a heterogenous group comprising both lifelong teetotallers and ex-drinkers, the latter tending to have a high burden of disease and high mortality.4 It must be emphasized that these ex-drinkers need not necessarily have been heavy drinkers.
Although attention has been drawn for more than 20 years to the possibility that the increased risk in non-drinkers may be due to the increased risk in ex-drinkers, it is only in the last decade that attempts have been made to separate ex-drinkers from lifelong teetotallers. Even after taking the characteristics of non-drinkers and the presence of pre-existing disease into account, or separating out ex-drinkers, most studies have shown that light to moderate drinkers have lower CHD risk than non-drinkers.2,4 There is now a consensus that light to moderate drinking is associated with reduced risk of cardiovascular disease and a meta-analysis has indicated that the risk reduction may be as much as 25% in subjects drinking 20–30 g/day compared with non-drinkers.2 The precise mechanisms by which alcohol consumption may reduce the risk of CHD remain uncertain. Increasing of HDL-cholesterol (HDL-C) is the best documented mechanism; other possible mechanisms proposed include the effects of alcohol on insulin sensitivity, haemostasis and inflammation.6
Thus, while there is considerable evidence that alcohol at some level has a beneficial effect on CHD, the benefit of alcohol on overall cardiovascular disease mortality or on all cause mortality is less convincing.3,7,8 Findings from the earlier British Doctor's Study report using 13 year follow-up were based on comparisons between drinkers and all non-drinkers, and this paper is often cited as strong evidence to support the protective effect of alcohol on CHD and all-cause mortality.9 In the latest report from The British Doctor's Study published in this issue of the International Journal of Epidemiology, Doll and his colleagues have re-visited their data and examined the relationship between alcohol and CHD and all cause mortality using 23-year follow-up data, in an attempt to address the issue of reverse causality i.e. illness/medication leading to non-drinking status.10 The study included over 12 000 male British doctors aged 48–78 years in 1978 followed-up for 23 years, in whom there were 7000 deaths. In the first analyses they compared current drinkers with all non-drinkers (lifelong teetotallers and ex-drinkers combined) and showed current drinking to be associated with a relative risk of 0.71 (95% CI 0.63–0.81) for CHD mortality, 0.92 (95% CI 0.80–1.07) for cancer mortality, 0.78 (95% CI 0.63–0.97) for respiratory disease, 1.35 (95% CI 0.91–2.00) for alcohol augmentable causes, and 0.81 (95% CI 0.76–0.87) for all cause mortality after adjustment for age, smoking, and follow-up duration. To allow for the possibility of reverse causality, they combined recent ex-drinkers (some of whom may have stopped recently because of illness) with current drinkers, and combined never drinkers with the long-term ex-drinkers. These combined groupings made a minor difference to the relative risk seen for CHD (RR = 0.72) and reduced the risk for respiratory causes further (RR = 0.69), but they increased the relative risks seen for cancer and all cause mortality. The relative risk for cancer mortality attenuated to 1.04 (95% CI 0.82–1.32) and for all cause mortality, the reduction in relative risk associated with current drinking compared with never or long term ex-drinkers, although still statistically significant, was only 12% (RR = 0.88, 95% CI 0.79–0.98). The study illustrates that the ‘protective’ effects of alcohol diminish when reverse causality is taken into account. The relative risk of 0.88 for all cause mortality is consistent with the effect seen in many other cohort studies which have observed little or no significant benefit for all cause mortality.7,8 The authors conclude that some of the apparently protective effects of alcohol against disease is artefactual but that some of it is real. The range of confounding factors considered in their analyses is very limited with adjustment made only for age, smoking, and follow-up duration. The characteristics of the various alcohol intake groups, which are likely to be very different, are not presented and the magnitude of reduction for all cause mortality, although statistically significant, is small and may well be due to other confounding factors not measured in the study or other untoward/disadvantageous characteristics of lifelong teetotallers which might increase their risk. The reasons for being a lifelong teetotaller may be determined by family history, genetic, environmental, social, religious, or even health conditions, factors that may influence mortality.
From this and from previous studies there is strong evidence that alcohol confers some protection against CHD, but the degree of protection claimed may still be exaggerated by comparisons with a heterogeneous group and by the limited adjustment procedures used to take into account the differing characteristics of the various alcohol intake groups. For all cause mortality there is no general agreement regarding benefit, and the findings from the British Doctor's Study provide further evidence that the benefit of alcohol on all cause mortality in older men is small and this lower risk may well be attributed to residual confounding.
 |
References |
|---|
 Top References |
|---|
1 Pearl R. Alcohol and longevity. New York: Alfred A Knopf, 1926.
2 Corrao G, Rubbiati L, Bagnardi V, Zambon A, Poikolainen K. Alcohol and coronary heart disease: a meta-analysis. Addiction 2000;95:1505–23.[CrossRef][Web of Science][Medline]
3 Shaper AG. Alcohol and mortality: a review of prospective studies. Br J Addict 1990;85:837–47.[CrossRef][Web of Science][Medline]
4 Fagrell B, de Faire U, Bondy S et al. The effects of light to moderate drinking on cardiovascular diseases. J Intern Med 1999;246:331–40.[CrossRef][Web of Science][Medline]
5 Shaper AG, Wannamethee G, Walker M. Alcohol and mortality: explaining the U-shaped curve. Lancet 1988;ii:1268–73.
6 Rimm EB, Williams P, Fosher K, Criqui M, Stampfer MJ. A biologic basis for moderate alcohol consumption and lower coronary heart disease risk: a meta-analysis of effects on lipids and haemostatic factors. BMJ 1999;319:1523–28.
7 Wannamethee SG, Shaper AG. Lifelong teetotallers, ex-drinkers and drinkers: mortality and the incidence of major coronary heart disease events in middle-aged British men. Int J Epidemiol 1997;26:523–31.
8 Strandberg AY, Strandberg TE, Salomaa VV, Pitkala K, Miettienen TA. Alcohol consumption, 29-y total mortality, and quality of life in men in old age. Am J Clin Nutr 2004;80:1366–71.
9 Doll R, Peto R, Hall E, Wheatley K, Gray R. Mortality in relation to consumption of alcohol: 13 years' observations on male British doctors. BMJ 1994;309:911–18.
10 Doll R, Peto R, Boreham J, Sutherland I. Mortality in relation to alcohol consumption: prospective study among male British doctors. Int J Epidemiol 2005;34:199–204.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||