Conditions and diseases other than cardiovascular associated with moderate alcohol consumption.
The relationship between alcohol consumption and health effects
The dose--response relationship for alcohol consumption and some conditions(1) appears to be linear, while for other conditions, such as cirrhosis, esophageal cancer and acute pancreatitis, the relationship is considered to be exponential. A "U"-shaped relationship has been proposed for ischemic heart disease and is now supported by almost 50 studies. It would be reasonable to assume that the dose--response curves for all other conditions associated with chronic excessive alcohol consumption apart from cardiovascular disease and cholelithiasis are either linear or exponential until additional information emerges to challenge this assumption. The dose--response relationships for conditions associated with acute intoxication are likely to be more variable and complex.
If the dose--response relationship is linear, the incidence and severity of negative consequences are proportional to the quantity of alcohol consumed at all levels of consumption. The effects of moderate levels of consumption can be estimated by extrapolating from the known effects of much higher doses. If the relationship is exponential, however, the incidence of problems at lower doses will be very much less than that observed at higher doses.
Contributions to alcohol-related morbidity and mortality
Cerebrovascular disease, road injuries, liver disease, suicide, and cancers of the head and neck (including esophagus), colon (males) and breast (females) are the major contributors to alcohol-related mortality.(2) These conditions account for most of the years of lost life expectancy due to alcohol, although alcohol dependence, falls, drowning, and assault are other important contributors.(2) Alcoholic psychosis and dependence, falls, road injuries, assault and stroke are the major contributors to alcohol-caused hospital morbidity.(2) Special attention should be paid to these conditions when attempting to determine the overall contribution of alcohol consumption (including moderate consumption) to mortality and morbidity.
A major study of alcohol- and drug-related mortality and morbidity in Australia was recently carried out by Holman and colleagues(2) to determine the etiological fraction for conditions considered to be associated with consumption of alcohol, tobacco or illicit drugs. Meta-analysis of published studies of the benefits or adverse health effects was carried out according to a standardized methodology. This paper draws heavily on this analysis,(2) as it is comprehensive, relatively current, carefully conducted, follows a well described methodology, and has the advantage that the material was reviewed by the same panel of authors. However, this analysis has drawn some criticism,(3) which the authors rebutted vigorously.(4)
Caveats to the meta-analysis of morbidity and mortality
Several conclusions emerge from this review. First, the available data do not permit determination of the dose--response relationship for many conditions that contribute substantially to alcohol-related morbidity and mortality. Second, the studies reviewed were conducted by a veritable army of investigators using a wide variety of methodologies, including, most importantly, measurement of alcohol consumption. The reliability and validity of this variety of methods of measurement of alcohol consumption are unknown. Although a common diagnostic definition would have been used in these studies, cross-national comparisons of medical conditions are rarely straightforward. Third, acute and chronic effects of alcohol are considered together, although their characteristics are quite dissimilar. Most of these studies involve chronic diseases (such as cancers) and consumption of alcohol over a long period. However, some conditions, such as road injuries, are more usually associated with acute intoxication from alcohol. The relationship between chronic excessive alcohol consumption and those conditions associated with acute intoxication is further complicated by the effects of tolerance to alcohol. For example, the impairment of psychomotor skills required to safely operate a motor vehicle for an individual used to consuming prodigious quantities of alcohol on a daily basis is clearly less than for an individual who consumes large quantities of alcohol infrequently. Assessment of the acute effects of alcohol is more reliable in the studies that have measured alcohol concentrations. Yet even in these studies, allowance for the possible effects of tolerance to alcohol is difficult to make. One of the important effects that lie hidden in these data is that of age sensitivity. If searched for, it emerges. For example, the etiological fraction for road injuries was calculated(2) to be 0.36 for persons age 15--24 and 0.31 for persons age 25--49 but only 0.16 for persons over 50.
[TABULAR DATA OMITTED]
Modest effects of alcohol on common conditions are likely to assume greater public health significance than more major effects on less common conditions. The etiological fraction for breast cancer--the commonest cause of malignant death in females--has been calculated at only 0.19. But because breast cancer is so common in females, alcohol-related breast cancers become the second commonest adverse health effect related to alcohol in women. The difference in adverse health consequences for moderate alcohol consumption will be significantly greater in public health terms for common conditions like breast cancer if the dose--response relationship is linear rather than exponential. Holman et al.(2) used their estimation of etiological fractions (based on this meta-analysis) to calculate the number of deaths, lost life expectancy and bed days for males and females in Australia in 1988.
[TABULAR DATA OMITTED]
Our understanding of many of these conditions is insufficient to assess the contribution of moderate alcohol consumption to these estimates of morbidity and mortality. Conditions associated with acute intoxication will be well accounted for by the prevention paradox,(5) because these adverse consequences can be largely attributed to individuals drinking immoderately on a few occasions although their consumption appears moderate when averaged out over a period.
There can be little doubt that our knowledge of the health effects of immoderate alcohol consumption far exceeds our understanding of the impact of moderate alcohol consumption. For most conditions other than cardiovascular disease and cholelithiasis, moderate alcohol consumption appears to have intermediate risks compared with the risks of abstinence and of excessive drinking.
Most of the conditions that make major contributions to alcohol-related mortality are associated with chronic excessive consumption, including cancers and cirrhosis. Some conditions are related to both acute and chronic patterns of drinking. For example, acute alcohol consumption has been associated with hemorrhagic stroke, while chronic excessive consumption has been linked to ischemic stroke. Acute alcohol consumption is associated with several other major contributors to mortality, including various kinds of trauma such as road injuries, falls, fires and drowning. Suicides and deaths attributed to alcohol dependence are associated with both acute and chronic patterns of immoderate drinking.
The relative contributions of acute and chronic excessive consumption to mortality and morbidity will vary between countries and be strongly influenced by prevailing drinking patterns. In countries where episodic patterns of drinking are predominant, the effects of the prevention paradox will be more apparent, with moderate drinkers accounting for a substantial proportion of the observed alcohol-related mortality and morbidity. In countries where the pattern of drinking is continuous consumption with minimal intoxication, linear and exponential dose--response relationships will result in the heaviest drinkers accounting for a more substantial proportion of alcohol-related problems. A further complication is the extent to which persons who consume alcohol in binges are counted as moderate drinkers because their consumption averages out below levels considered hazardous.
Although the literature on alcohol consumption and health effects generally assumes that individual drinkers are entrenched members of drinking categories, numerous longitudinal studies of alcohol consumption as well as clinical practice reminds us that drinking patterns in individuals fluctuate considerably, further complicating the interpretation of research data.
(1.)Skog O-J (1992). Epidemiological and biostatistical aspects of alcohol use, alcoholism, and their complications. Windows on Science, Patricia Erickson and Harold Kalant (eds). Addiction Research Foundation, Toronto.
(2.)Holman CDJ, Armstrong BK, Arias LN, Martin CA, Hatton WM, Hayward LD, Salmon MA, Shean RE, Waddell VP (1988). The quantification of drug caused morbidity and mortality in Australia. Commonwealth Department of Community Services and Health, Canberra.
(3.)MacMahon B (1992). The quantification of drug caused morbidity and mortality in Australia: A critique. Med J Aust 157: 557-560.
(4.)Holman CDJ, Armstrong BK (1992). The quantification of drug caused morbidity and mortality in Australia: The authors respond. Med J Aust 157: 560-561.
(5.)Kreitman N (1986). Alcohol consumption and the prevention paradox. Br J Addict, 81: 353-363.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Moderate Drinking and Health: The Scientific Evidence|
|Publication:||Contemporary Drug Problems|
|Date:||Mar 22, 1994|
|Previous Article:||The other health benefits of moderate alcohol intake.|
|Next Article:||Moderate drinking and traffic crashes: a case for health or for safety?|