Light to moderate alcohol consumption reduced mortality in women
ACP J Club. 1995 Nov-Dec;123:80. doi:10.7326/ACPJC-1995-123-3-080
Fuchs CS, Stampfer MJ, Colditz GA, et al. Alcohol consumption and mortality among women. N Engl J Med. 1995 May 11;332:1245-50.
To determine the association between alcohol consumption and the risk for death in women.
Cohort analytic study of women followed for 12 years (Nurses' Health Study).
Population-based study in the United States.
121 700 women (nurses) aged 30 to 55 years at baseline (1976) were surveyed every 2 years. Exclusion criteria were a history of cancer, ischemic heart disease, or stroke; incomplete data; or abstinence in 1980 but previous consumption of alcohol. 85 709 women were studied.
Assessment of risk factors
Food questionnaires completed in 1980, 1984, and 1986 were used to collect data on alcohol consumption (beer [13.2 g/bottle], wine [10.8 g/glass], and spirits [15.1 g/drink]).
Main outcome measure
Death (overall and from cardiovascular disease [CVD], coronary heart disease, stroke, cancer, breast cancer, cirrhosis, and injuries).
2658 women died during follow-up. Multivariate relative risks (RRs) were adjusted for age, smoking, body mass index, aspirin use, exercise, hypercholesterolemia, diabetes, hypertension, premature myocardial infarction in a parent, contraceptive use, menopausal status, hormone replacement therapy, and fiber and fat intake. When compared with no drinking in a multivariate analysis, light to moderate drinking (1.5 to 4.9 g/d, 5.0 to 14.9 g/d, and 15.0 to 29.9 g/d) was associated with decreased risks for death from all causes (RR 0.83, 95% CI 0.74 to 0.93; RR 0.88, CI 0.79 to 0.99; and RR 0.89 CI 0.77 to 1.02 [trend only]), death from CVD (RR 0.57, CI 0.43 to 0.76; RR 0.73, CI 0.56 to 0.95; and RR 0.66, CI 0.47 to 0.93), and death from coronary heart disease (RR 0.51, CI 0.36 to 0.73; RR 0.64, CI 0.46 to 0.89; and RR 0.65, CI 0.43 to 0.99). Women with the highest alcohol consumption (≥ 30 g/d) had increased risk for death from all causes (RR 1.19, CI 1.02 to 1.38), from non-CVD (RR 1.33, CI 1.12 to 1.57), from cancer (RR 1.28, CI 1.04 to 1.58), from breast cancer (RR 1.67, CI 1.10 to 2.53), and from cirrhosis (RR 2.55, CI 1.06 to 6.11). The highest benefits from alcohol consumption were seen in women older than 50 years and in women with risk factors for coronary heart disease.
Women who consumed light to moderate amounts of alcohol had reduced risks for death compared with women who consumed no alcohol; women who consumed large amounts of alcohol had increased risks for death.
Sources of funding: National Institutes of Health and the American Cancer Society.
For article reprint: Dr. C.S. Fuchs, Dana-Farber Cancer Institute, 44 Binney Street, Boston, MA 02115, USA. FAX 617-632- 5424.
Many studies done during the past 15 years have shown that light to moderate alcohol intake decreases mortality from ischemic heart disease and CVD. Even when former heavy drinkers are excluded and only persons who do not drink are considered, the mortality rate for nondrinkers is higher than that for light to moderate drinkers.
The studies by Fuchs and Grønbaek and their colleagues highlight 2 remaining controversies: 1) Can the benefit primarily observed in middle-aged men be extrapolated to women, particularly in light of findings that moderate to heavy alcohol consumption increases the risk for breast cancer (1)? and 2) Does the type of alcohol consumed make a difference?
Fuchs and colleagues did find that light to moderate alcohol consumption reduced death from CVD and overall mortality. This analysis, with its large sample size and wide geographic and ethnic distribution, should be generalizable to most women in the United States. On the basis of their results, we conclude that the health care provider and the woman need to weigh the individual risks for coronary heart disease, breast cancer, and alcoholism to determine whether moderate alcohol consumption is beneficial. A woman at high risk for breast cancer and low risk for coronary heart disease may well decide that drinking is not safe for her. In most cases, however, light to moderate consumption of alcohol should be safe, even beneficial, and should result in a 10% to 20% decrease in overall mortality rates.
A more difficult question to answer concerns the extent to which the type of alcohol consumed makes a difference. Grønbaek and colleagues found that, in a population-based sample in Denmark, drinking wine was protective but the heavy consumption of spirits had an adverse effect. No evidence was found for a U-shaped curve with wine because heavy drinkers received the same benefit as moderate drinkers.
Unfortunately, 2 problems make the results of the study by Grønbaek and colleagues difficult to accept. First, the prevalence of recovering alcoholics—persons with a history of drinking who are currently sober—was unbelievably low (17 of 13 285, or 0.01%), which suggests a strong reporting bias. The article by Fuchs and colleagues supports previous studies showing that former heavy drinkers have an increased mortality rate; therefore, including these former drinkers would increase the RR for nondrinkers. Second, and more important, Grønbaek and colleagues controlled for a very limited number of potential confounders. They did not adjust for physical activity level; aspirin use; or the presence of hypertension, hyperlipidemia, or diabetes mellitus. These factors—all of which are strongly associated with death from CVD and total mortality—could also be associated with the type of alcohol consumed. This is a serious concern because the observed findings could be caused by differences in these factors rather than the type of alcohol consumed. Many studies have looked at this important issue, but the answer is still unclear. This study adds to the data favoring wine, but, unfortunately, its limitations prevent it from being definitive.
Scott E. Sherman, MD, MPH
Donald S. Chang, MDVA/UCLA/RAND Center for the Study of Healthcare Provider BehaviorSepulveda, California, USA