摘要：Objectives. To review critical contributions from the Nurses’ Health Study (NHS) on alcohol consumption and health outcomes. Methods. We performed a narrative review of NHS (1980–2012) and NHS II (1989–2011) publications. Results. Using detailed information on self-reported alcohol drinking patterns obtained approximately every 4 years combined with extensive information on diet, lifestyle habits, and physician-diagnosed health conditions, NHS investigators have prospectively examined the risks and benefits associated with alcohol consumption. Moderate intake, defined as up to 1 drink a day, is associated with a lower risk of hypertension, myocardial infarction, stroke, sudden cardiac death, gallstones, cognitive decline, and all-cause mortality. However, even moderate intake places women at higher risk for breast cancer and bone fractures, and higher intake increases risk for colon polyps and colon cancer. Conclusions. Regular alcohol intake has both risks and benefits. In analyses using repeated assessments of alcohol over time and deaths from all causes, women with low to moderate intake and regular frequency (> 3 days/week) had the lowest risk of mortality compared with abstainers and women who consumed substantially more than 1 drink per day. Evaluating the relationship between alcohol intake and long-term health outcomes in observational studies is challenging owing to methodological complexities including nonlinear dose–response associations, the accuracy of the self-reported measurement of alcohol over time, changes in intake owing to health conditions, pharmaceutical contraindications, and age-related reduction in consumption. The Nurses’ Health Study (NHS) has been critical in contributing substantially to our understanding of the long-term risks and benefits of light, moderate, and heavy consumption of alcohol in relation to hypertension; diabetes; cardiovascular disease (CVD); colon polyps; cancer of the breast, colon, and ovaries; and several other conditions. All of our participants are registered nurses who have volunteered for this health study, so the vast majority were moderate drinkers or did not drink at all, and few participants were heavy drinkers. Therefore, the main focus of our work has been on moderate alcohol consumption. All this work depends on the accuracy of self-reported alcohol consumption from a food frequency questionnaire and several additional questions on drinking patterns, including frequency (days/week), binge drinking, and consuming alcohol with meals. Also, through incorporation of the specific findings of light to moderate alcohol consumption into an overall healthy lifestyle pattern, we have aimed to quantify the relative importance of alcohol with respect to other known risk or preventive factors for chronic disease. In this review, we have summarized some of the key findings on alcohol consumption and health outcomes that have been examined using the rich data on NHS participants. For most studies, we used Cox proportional hazards models or conditional logistic regression with risk set sampling to calculate an estimate of the incidence rate ratio (IRR). Two of the studies cited 1,2 reported odds ratios (ORs) estimated from logistic regression models.
