QIMR Berghofer Medical Research Institute, Brisbane, QLD, Australia.
Department of Psychiatry, Washington University School of Medicine, St Louis, MO, USA.
Addiction. 2018 Jan;113(1):158-166. doi: 10.1111/add.14008. Epub 2017 Sep 18.
Both high alcohol intake and alcohol dependence increase mortality, and both are associated with smoking. We aimed to compare the associations of quantity of alcohol, number of alcohol-related symptoms and smoking history with all-cause mortality, and to assess symptom count and smoking history as confounders or mediators of the effects of high alcohol intake.
Survival was analysed by Cox regression with sex, body mass index, alcohol intake (overall and by beverage), maximum drinks on any day, alcohol symptom count and smoking status as potential predictors of age at death.
Australia.
Participants were apparently healthy volunteers consisting of 33 593 Australian adult twins and their relatives who completed questionnaires or interviews between 1979 and 2005.
Data on alcohol use, smoking and occurrence of symptoms related to alcohol use disorders and death records from the Australian National Death Index.
A total of 3764 participants were matched with deaths occurring within Australia up to July 2014. Individually, alcohol intake [hazard ratio (HR) = 1.0082, 95% confidence interval (CI) = 1.0063-1.0102, per drink per week], beer intake (HR = 1.0159, 95% CI = 1.0123-1.0195, per drink per week), life-time maximum number of drinks in 1 day (HR = 1.0176, 95% CI = 1.0130-1.0221, per drink), symptom count (HR = 1.0867, 95% CI = 1.0633-1.1106, per symptom) and smoking status (HR = 2.82, 95% CI = 2.52-3.16 for smokers of 10+ cigarettes/day versus never-smokers) were each significant predictors of all-cause mortality. After adjustment for the independently significant predictors alcohol symptom count and smoking status, alcohol intake was no longer significant (adjusted HR = 1.0012 per drink per week, 95% CI = 0.9979-1.0145).
Number of symptoms related to high alcohol intake and tobacco smoking appear to account for the positive association between alcohol consumption and premature mortality.
大量饮酒和酒精依赖都会增加死亡率，两者都与吸烟有关。我们旨在比较酒精摄入量、酒精相关症状数量和吸烟史与全因死亡率的关系，并评估症状计数和吸烟史是否为高酒精摄入影响的混杂因素或中介因素。
使用 Cox 回归分析，以性别、体重指数、酒精摄入量（总体和按饮料）、任何一天的最大饮酒量、酒精症状计数和吸烟状况作为死亡年龄的潜在预测因子，分析生存情况。
澳大利亚。
参与者为澳大利亚成年双胞胎及其亲属，他们在 1979 年至 2005 年间完成了问卷调查或访谈，这些参与者被认为是健康的志愿者。
使用澳大利亚国家死亡索引记录的酒精使用、吸烟和与酒精使用障碍相关的症状发生的数据以及死亡记录。
共有 3764 名参与者与截至 2014 年 7 月在澳大利亚发生的死亡情况相匹配。单独来看，酒精摄入量[风险比（HR）=1.0082，95%置信区间（CI）=1.0063-1.0102，每周每喝一杯]、啤酒摄入量（HR=1.0159，95%CI=1.0123-1.0195，每周每喝一杯）、终生每天最大饮酒量（HR=1.0176，95%CI=1.0130-1.0221，每喝一杯）、症状数（HR=1.0867，95%CI=1.0633-1.1106，每增加一个症状）和吸烟状况（HR=2.82，95%CI=2.52-3.16，每天吸烟 10 支以上的吸烟者与从不吸烟者相比）均为全因死亡率的显著预测因子。在调整了独立显著预测因子酒精症状数和吸烟状况后，酒精摄入量不再显著（调整后的 HR=每周每喝一杯 1.0012，95%CI=0.9979-1.0145）。
与大量饮酒相关的症状数量和烟草吸烟似乎解释了饮酒与过早死亡之间的正相关关系。
