- Email: [email protected]
Alcohol Consumption and Mortality among Middle-aged and Elderly Japanese Men and Women
Alcohol Consumption and Mortality among Middle-aged and Elderly Japanese Men and Women YINGSONG LIN, MD, PHD, SHOGO KIKUCHI, MD, PHD, AKIKO TAMAKOSHI, MD, PHD, KENJI WAKAI, MD, PHD, TAKASHI KAWAMURA, MD, PHD, HIROYASU ISO, MD, PHD, ITSURO OGIMOTO, MD, PHD, KIYOKO YAGYU, BA, YUKI OBATA, MPHARM, AND TERUO ISHIBASHI, MD, FOR THE JACC STUDY GROUP*
PURPOSE: We conducted a prospective cohort study to examine the association between alcohol intake and the risk of all-cause mortality among middle-aged and elderly Japanese men and women. METHODS: At baseline (1988–1990), a total of 110,792 Japanese men and women aged 40 to 79 years were asked to complete a questionnaire that included information on alcohol intake, and were followed up for all-cause mortality through December 31, 1999. Relative risks (95% confidence interval) were calculated using Cox proportional-hazards models. RESULTS: The risk of all-cause mortality was lowest among current drinkers with an alcohol intake of 0.1 to 22.9 g/d (RR, 0.80; 95% CI, 0.72–0.88 for men; and RR, 0.88; 95% CI, 0.77–1.00 for women). Excessive mortality associated with heavy drinking (> 69 g/d) was observed for cancer, cardiovascular disease and injuries and other external causes in men, while significantly reduced mortality with light drinking was seen for cancer in men and CVD in women. For men, the benefit associated with light alcohol consumption (! 23 g/d) was more apparent among nonsmokers than among smokers. CONCLUSION: Our prospective data show a 12% to 20% decreased risk of all-cause mortality in both Japanese men and women who consumed less than 23 g/d of alcohol (approximately 2 drinks), although heavy drinking increased that risk. Ann Epidemiol 2005;15:590–597. Ó 2005 Elsevier Inc. All rights reserved. KEY WORDS:
Alcohol Consumption, All-cause Mortality, Japan, Cardiovascular Disease, Cancer.
INTRODUCTION Epidemiological studies in Western industrialized countries have found that light-to-moderate alcohol consumption is associated with decreased total mortality, most of which were due to a reduction in cardiovascular deaths (1–11). In
From the Department of Public Health, Aichi Medical University School of Medicine, Aichi, Japan (Y.L., S.K., K.Y., Y.O.); Department of Preventive Medicine/Biostatistics and Medical Decision Making, Nagoya University Graduate School of Medicine, Nagoya, Japan (A.T.); Division of Epidemiology and Prevention, Aichi Cancer Center Research Institute, Nagoya, Japan (K.W.); Kyoto University Center for Student Health, Kyoto, Japan (T.K.); Department of Public Health Medicine, Institute of Community Health, University of Tsukuba, Tsukuba, Japan (H.I); Department of Public Health, Kurume University School of Medicine, Kyushu, Japan (I.O.); and Asama General Hospital, Nagano, Japan (T.I.). Address correspondence to: Shogo Kikuchi, M.D., Ph.D., Department of Public Health, Aichi Medical University School of Medicine, 21 Karimata, Yazako, Nagakute-cho, Aichi-gun, Aichi 480-1195, Japan. Tel.: C81-56162-3311; Fax: 81-561-62-5270. E-mail: [email protected] *Japan Collaborative Cohort Study for Evaluation of Cancer Risk sponsored by Monbukagakusho. See acknowledgments for members of the JACC Study. The JACC Study has been supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan (Nos. 61010076, 62010074, 63010074, 1010068, 2151065, 3151064, 4151063, 5151069, 6279102 and 11181101). Received July 14, 2004; accepted October 19, 2004. Ó 2005 Elsevier Inc. All rights reserved. 360 Park Avenue South, New York, NY 10010
Japan, the health effects of alcohol consumption have not been clearly defined. Although several prospective studies have been conducted to examine the relation between alcohol consumption and all-cause mortality (12–16), the findings on the protective effect of light-to-moderate alcohol consumption have been controversial, and only one study has reported the corresponding relation for Japanese women (16). A large cohort study involving 490,000 US men and women indicated that the benefit from moderate alcohol consumption depends in part on age and background cardiovascular risk (2). Given the differences in baseline characteristics among subjects, types of alcohol beverages usually consumed, or subjects’ drinking patterns, the balance of risks and benefits of alcohol consumption is likely to differ among age groups and populations. Few studies, however, have addressed this issue in Japanese populations. The main purpose of our study is to examine prospectively the relation between alcohol intake and risk of death from all causes (CVD, cancer, and injuries and other external causes) in a large cohort of Japanese men and women. In addition, we also aim to determine how age and cigarette smoking modify the relation of alcohol consumption to total mortality in our cohort. 1047-2797/05/$–see front matter doi:10.1016/j.annepidem.2004.10.010
AEP Vol. 15, No. 8 September 2005: 590–597
Selected Abbreviations and Acronyms CVD Z cardiovascular disease BMI Z body mass index
METHODS Study Cohort The Japan Collaborative Cohort Study for Evaluation of Cancer Risk (the JACC study) is a large prospective cohort study sponsored by the Japanese Ministry of Education, Culture, Sports, Science and Technology (Monbukagakusho). The details of the JACC study have been described elsewhere (17). Briefly, it was initiated between 1988 and 1990, when 127,500 apparently healthy inhabitants from 45 areas throughout Japan were enrolled as a baseline cohort. Among this cohort population, 46,465 men and 64,327 women (110,792 in total) aged 40 to 79 years were followed up until December 31, 1999. The average age of the participants at enrollment was 58 years. Subjects who reported a history of cancer, stroke, or myocardial infarction were excluded. Informed consent in the majority of the areas was obtained by having subjects sign the cover page of the questionnaire. Our study was approved by the Ethical Board of the Nagoya University School of Medicine.
Lin et al. ALCOHOL CONSUMPTION AND MORTALITY
591
times/week, and less than once/week. The amount of daily ethanol intake was calculated as follows: the average amount of alcohol use per occasion per week multiplied by the number of occasions per week and divided by seven. In the present study, we define light drinkers as people who drank less than 23 g/d of alcohol, which approximates 2 US standard drinks. One drink is assumed to contain, on average, 12g of alcohol. We define moderate drinkers as people who drank less than 46 g/d of alcohol (! 4 drinks) of alcohol, and heavy drinkers as those who drank 46 or more g/d of alcohol (> 4 drinks). Follow-up and Identification of Deaths Our primary end point was death from any cause. Follow-up surveys were conducted annually to verify the vital status of the participants, using the resident-registry data from the municipalities. These data were collected at the central office of the study group. For deceased subjects, the underlying cause was recorded from death certificates and coded according to the International Classification of Diseases, 10th Revision. The following subgroups were analyzed: subjects who died from cancer (C00–C97), CVD (I00– I99), and injuries and external causes (V01–Y98). By December 31, 1999, 11.0% of the participants had died, 85.5% were alive, and 3.5% were lost to follow-up because they moved out of the study area.
Data Collection
Statistical Analysis
At the baseline survey, we used a self-administered questionnaire to collect information on demographic characteristics, medical history of selected diseases such as diabetes mellitus and hypertension, tobacco and alcohol use, coffee consumption, participation in sports or exercise, female reproductive characteristics, and dietary habits. In the section on alcohol consumption, we first asked the subjects to choose their drinking status from three precoded response categories: nondrinkers, ex-drinkers, or current drinkers. We defined these categories as follows: nondrinkers were people who reported no alcohol drinking in the past year and ‘‘never or almost never’’ drinking; exdrinkers were people who drank at some time in their lives but not in the past year preceding the baseline survey; current drinkers were people who reported drinking in the past year. Ex-drinkers or current drinkers were then asked to report the age at which they started drinking, frequency of consumption in the past year (in the year before quitting drinking for ex-drinkers), type of beverages (Japanese sake, Japanese spirits [shochu], beer, whiskey, and wine), and the average amount consumed on one occasion (the unit was converted to gou, with 1 gou equivalent to 23 g of ethanol). The frequency of alcohol consumption was classified into four categories: almost every day, 3 to 4 times/week, 1 to 2
Separate analyses were performed for men and women. Men were grouped into the following six categories based on their daily amount of alcohol intake reported at baseline: nondrinkers, ex-drinkers, and current drinkers (whose intake was classified as 0.1–22.9, 23.0–45.9, 46.0–68.9, or 69.0 or more g/d). Because of the very small number of women in the category of 69.0 g/d or more, such women were combined with those consuming 46.0 to 68.9 g/d, resulting in five categories. Age-adjusted and multivariateadjusted relative risks (95% confidence interval) for death from all causes (cancer, CVD, and injuries and other external causes) were calculated using Cox proportionalhazards models (18). Nondrinkers served as the reference group. The multivariate analyses were adjusted for age (as a continuous variable), body mass index (!18, 18.0–21.9, 22.0–25.9, > 26.0 kg/m2), education levels more than high school and less than high school, history of diabetes (yes/ no), history of hypertension (yes/no), cigarette smoking (never, former, current smokers with 0–19, 20–39, and > 40 cigarettes/d), and participation in sports or exercise (!1 and > 1 hour/week). Tests for linear trend were conducted by treating the median value in each category of alcohol consumption as a continuous variable, with ex-drinkers being excluded in the analyses.
592
Lin et al. ALCOHOL CONSUMPTION AND MORTALITY
AEP Vol. 15, No. 8 September 2005: 590–597
We also conducted analyses stratified by age (40–59 vs. 60–79 years) and smoking status (never vs. ever) to determine their effect on the association between alcohol consumption and all-cause mortality. For these analyses, a cross-product term for each alcohol consumption level and the above variables was included in separate models. All p-values were 2-sided, with values of 0.05 or less indicating statistical significance. All analyses were performed with SAS Version 6.12 (SAS Institute Inc., Cary, NC). RESULTS Among the 110,792 cohort participants, we excluded the 8168 (7.4%) subjects who had missing information on their drinking status, and 5192 (4.7%) subjects who reported a history of cancer, stroke, or myocardial infarction. This left 97,432 (42,072 men and 55,360 women) subjects for the present analysis. During a mean follow-up of 10 years, a total of 9589 subjects (5902 men and 3687 women) died from all causes. Among 5902 deaths in men, 2418 (41.0%) died from cancer, 1584 (26.8%) from CVD, and 185 (3.1%) from injuries and other external causes. Among 3687 deaths in women, 1363 (37.0%) died from cancer, 1191 (32.3%) from CVD, and 93 (2.5%) from injuries and other external causes. Table 1 presents baseline characteristics according to the categories of alcohol consumption. Nearly 76% of the men and 25% of the women were current drinkers. Compared with nondrinkers, current drinkers tended to be younger, more likely to smoke, and have higher education levels.
Among both men and women drinkers, the percentage of current smokers increased consistently with the increasing amount of daily alcohol intake. The association of alcohol consumption with risk of death from all causes (cancer, CVD, and injuries and other external causes) is shown in Table 2 for men and Table 3 for women. In both men and women, after adjustments for age, body mass index, education, history of diabetes and hypertension, cigarette smoking, and exercise, ex-drinkers had an increased risk of all-cause mortality compared with nondrinkers. Similar results were obtained for cancer, CVD in men (Table 2). The risk of all-cause mortality was lowest in the category of current drinkers with alcohol consumption less than 23.0 g/d (RR, 0.80; 95% CI, 0.72–0.88 for men; and RR, 0.88; 95% CI, 0.77–1.00 for women). A statistically significant increase in risk of all-cause mortality was observed for men who consumed 69.0 or more g/d of alcohol (RR, 1.32; 95% CI, 1.18–1.48), but not for women who consumed 46.0 or more g/d of alcohol (RR, 1.22; 95% CI, 0.76–1.95). A high level of alcohol consumption (> 69g/d) significantly increased the risk of death from injuries and other external causes in men. Excessive mortality due to heavy drinking was also observed for cancer, CVD, and injuries or other external causes in men. On the other hand, reduced mortality among light drinkers (! 23 g/d) was seen for cancer in men and CVD in women. In subgroup analyses, we examined alcohol consumption and the risk of all-cause mortality according to age and smoking status (Tables 4 and 5). Risk reduction in all-cause mortality associated with light-to-moderate amount of
TABLE 1. Baseline charactersistics of the study population by categories of alcohol drinking in the JACC study Nondrinkers
Men No. of individuals Age (years) Body mass index (kg/m2)y More than high school education (%) History of hypertension (%) History of diabetes (%) Current smokers (%) Exercise > 5 hours per week (%) Women No. of individuals Age (years) Body mass index (kg/m2)y History of hypertension (%) History of diabetes (%) More than high school education (%) Current smokers (%) Exercise > 5 hours per week (%)
7839 59.0G10.6 22.5G2.9 11.9 14.1 5.1 49.7 5.9 40895 57.8G10.0 22.9G3.1 21.0 3.5 7.6 3.5 3.5
Ex-drinkers
2386 62.4G9.5 22.2G3.0 12.8 24.5 11.8 45.9 6.7 884 58.2G10.2 23.0G3.4 24.8 7.8 7.2 22.8 4.8
Plus-minus values are mean G stardard deviation. y Calculated as weight in kilograms divided by the square of height in meters.
Current drinkers (alcohol intake: g/day) 0.1–22.9
23.0–45.9
46.0–68.9
> 69.0
6140 56.3G10.3 22.7G2.7 18.1 15.8 5.6 46.9 6.2
8072 57.2G10.1 22.6G2.7 14.6 19.6 5.7 52.6 7.4
7085 55.5G9.3 22.8G2.6 12.8 21.0 4.4 61.7 6.4
3337 53.9G9.0 22.9G2.8 10.6 19.1 5.2 69.3 5.8
5860 55.0G9.6 22.9G2.9 16.9 2.4 10.8 7.6 5.4
1020 54.9G9.9 23.0G3.1 21.8 2.1 7.7 22.2 5.9
335 53.0G9.3 23.4G3.4 23.3 2.4 7.5 41.6 4.5
– – –
– – –
AEP Vol. 15, No. 8 September 2005: 590–597
Lin et al. ALCOHOL CONSUMPTION AND MORTALITY
593
TABLE 2. Relative risks of death according to cause and alcohol intake among men in the JACC study Nondrinkers
Ex-drinkers
Current drinkers (alcohol intake: g/day) 0.1–22.9
Person-years 75352 Death from all causes No. of deaths 1281 Age-adjusted RR 1.00 (95% CI) Multivariate RR 1.00 (95% CI) Death from cancer No. of deaths 498 Age-adjusted RR 1.00 (95% CI) Multivariate RR 1.00 (95% CI) Death from cardiovascular disease No. of deaths 329 Age-adjusted RR 1.00 (95% CI) Multivariate RR 1.00 (95% CI) Death from injuries and external causes No. of deaths 30 Age-adjusted RR 1.00 (95% CI) Multivariate RR 1.00 (95% CI)
23.0–45.9
46.0–68.9
> 69.0
P-value for trend
20847
59708
77757
68622
32076
719 1.70 (1.55–1.86)
616 0.78 (0.71–0.86)
998 0.91 (0.83–0.99)
765 0.98 (0.89–1.07)
440 1.43 (1.28–1.60)
!0.001
1.58 (1.44–1.74)
0.80 (0.72–0.88)
0.90 (0.82–0.98)
0.95 (0.86–1.04)
1.32 (1.18–1.48)
!0.001
253 1.56 (1.34–1.81)
251 0.79 (0.68–0.92)
422 0.96 (0.85–1.10)
351 1.09 (0.95–1.25)
185 1.43 (1.20–1.70)
!0.001
1.50 (1.29–1.75)
0.82 (0.70–0.95)
0.96 (0.84–1.10)
1.05 (0.91–1.20)
1.31 (1.10–1.56)
0.001
216 1.95 (1.65–2.32)
171 0.86 (0.72–1.04)
256 0.92 (0.78–1.09)
224 1.18 (0.99–1.40)
105 1.45 (1.16–1.81)
!0.001
1.79 (1.51–2.14)
0.86 (0.73–1.06)
0.89 (0.75–1.05)
1.09 (0.92–1.30)
1.28 (1.02–1.61)
0.01
17 1.70 (0.94–3.08)
28 1.35 (0.81–2.27)
35 1.23 (0.76–2.01)
25 1.11 (0.65–1.90)
18 1.88 (1.04–3.40)
0.17
1.69 (0.92–3.08)
1.52 (0.90–2.58)
1.37 (0.83–2.26)
1.23 (0.71–2.13)
1.99 (1.09–3.64)
0.11
RR, relative risk; CI, confidence interval. Multivariate RR: adjusted for age, body mass index, education, cigarette smoking, exercise, and history of diabetes and hypertension. P for trend among current drinkers.
drinking appeared to be greater in men and women aged 60 to 79 years than in those aged 40 to 59 years. However, no significant effect modification was observed. The relation between light-to-moderate amount of alcohol consumption (! 46 g/d) and the risk of all-cause mortality was significantly modified by smoking status in men (Table 4), but such effect modification was not noted in women (Table 5). Among men, the decrease in the risk for all-cause mortality was greater in never smokers than in ever smokers (Table 4). For both men and women, the estimates remained unchanged after excluding those who had died within the first 2 years of follow-up. Additional adjustment for consumption of green leafy vegetables and female reproductive factors did not materially alter the results. DISCUSSION In this large prospective study, we found that current drinkers who consumed less than 23.0 g/d of alcohol (approximately 2 drinks) had a 12% to 20% significantly decreased risk of all-cause mortality in men and women, while heavier drinking (> 69.0 g/d) was associated with a 30% increased risk in men, in relation to nondrinkers. Our findings were consistent with studies in Western industri-
alized countries and another cohort study in Japan (12), although the definition of ‘‘moderate drinking’’ and the nadir at which the lowest mortality was observed may differ across studies. In our study, deaths from cancer, CVD, and injuries or other external causes accounted for 39.4%, 28.9%, and 2.9% of total mortality, respectively. Among men, light-tomoderate alcohol consumption was significantly associated with a decreased risk of cancer. This finding was consistent with another cohort study on Japanese men, in which moderate alcohol consumption (1–149 g/week) was associated with a 50% decreased risk of cancer mortality (12). Similarly, a prospective study conducted in Shanghai, China showed a 15% reduction in risk of death from any type of cancer for light to moderate male drinkers (19). Studies from Western countries showed the similar results (2, 20). In a large US cohort study, consuming 1 drink/day was associated with a 20% decreased risk (statistically insignificant) for alcohol-related cancers in women, but such an association was not observed in men (2). An American Cancer Society study of middle-aged men found that mortality from cancer was significantly lower among those who consumed up to 1 drink/day, as compared with abstainers (20). It seems that any cancer-related benefits conferred occur only at the
594
Lin et al. ALCOHOL CONSUMPTION AND MORTALITY
AEP Vol. 15, No. 8 September 2005: 590–597
TABLE 3. Relative risks of death according to cause and alcohol intake among women in the JACC study Nondrinkers
Ex-drinkers
Current drinkers (alcohol intake: g/day) 0.1–22.9
Person-years Death from all causes No. of deaths Age-adjusted RR (95% CI) Multivariate RR (95% CI) Death from cancer No. of deaths Age-adjusted RR (95% CI) Multivariate RR (95% CI) Death from cardiovascular disease No. of deaths Age-adjusted RR (95% CI) Multivariate RR (95% CI) Death from injuries and external causes No. of deaths Age-adjusted RR (95% CI) Multivariate RR (95% CI)
23.0–45.9
P-value for trend
> 46.0
406599
8448
57611
9948
3211
2912 1.00 1.00
92 1.49 (1.21–1.83) 1.27 (1.03–1.57)
252 0.84 (0.74–0.95) 0.88 (0.77–1.00)
52 0.98 (0.74–1.29) 0.93 (0.70–1.22)
18 1.39 (0.88–2.21) 1.22 (0.76–1.95)
0.52 0.47
1054 1.00 1.00
30 1.35 (0.94–1.94) 1.21 (0.83–1.74)
119 1.00 (0.83–1.21) 1.03 (0.85–1.25)
26 1.26 (0.86–1.86) 1.20 (0.81–1.77)
6 1.09 (0.49–2.43) 1.04 (0.46–2.33)
0.42 0.53
975 1.00 1.00
20 0.96 (0.61–1.49) 0.79 (0.50–1.23)
62 0.66 (0.51–0.85) 0.69 (0.53–0.90)
12 0.71 (0.40–1.25) 0.65 (0.37–1.15)
6 1.62 (0.72–3.61) 1.22 (0.54–2.76)
0.09 0.06
67 1.00 1.00
2 1.39 (0.34–5.69) 1.69 (0.41–7.02)
13 1.60 (0.88–2.91) 2.11 (1.14–3.89)
2 1.42 (0.36–5.79) 2.01 (0.49–8.32)
– –
0.50 0.17
RR, relative risk; CI, confidence interval. Multivariate RR: adjusted for age, body mass index, education, cigarette smoking, exercise, and history of diabetes and hypertension, and P for trend among current drinkers.
lower levels of drinking. Among all cancer combined, our cohort study showed a decreased risk for men who were lightto-moderate drinkers at the lung and rectal sites (data not shown). A case–control study in Japan also found a U-shaped relationship between alcohol intake and gastric cancer risk (21). Given these findings and the fact that lung cancer and gastric cancer are the most common cancers in Japan, the reduction in risk of all cancers combined may be, in part, due to the decreased risk for lung and gastric cancer at lower levels of drinking in Japanese men.
We consider that the reduction in the risk for all-cause mortality associated with light-to-moderate alcohol consumption in this cohort population may be related to the decreased risk of deaths from CVD, since approximately 30% of deaths occurring during follow-up were attributable to CVD. In particular, we found a 31% decreased risk of morality from CVD among women who consumed 0.1 to 22.9 g/d of alcohol. Several mechanisms have been proposed to explain the protective effect of light-to-moderate alcohol consumption on CVD. The main effect is due to an increase
TABLE 4. Relative risks of death from all causes according to age, smoking, and alcohol intake among men in the JACC study Nondrinkers
Ages 40–59 years No. of deaths Multivariate RR (95% CI) Ages 60–79 years No. of deaths Multivariate RR (95% CI) Nonsmokers No. of deaths Multivariate RR (95% CI) Ever smokers No. of deaths Multivariate RR (95% CI) Excluding deaths within the first 2 years of follow-up No. of deaths Multivariate RR (95% CI)
Ex-drinkers
Current drinkers (alcohol intake: g/day) 0.1–22.9
23.0–45.9
46.0–68.9
> 69.0
190 1.00
137 2.77 (2.21–3.46)
150 0.98 (0.79–1.22)
204 1.02 (0.83–1.25)
240 1.14 (0.94–1.39)
176 1.57 (1.27–1.94)
1091 1.00
582 1.43 (1.29–1.58)
466 0.77 (0.69–0.86)
794 0.88 (0.80–0.97)
525 0.92 (0.83–1.02)
264 1.29 (1.12–1.48)
308 1.00
93 1.61 (1.27–2.04)
78 0.50 (0.39–0.65)
131 0.86 (0.70–1.07)
81 0.97 (0.75–1.25)
47 1.52 (1.11–2.07)
911 1.00
584 1.56 (1.41–1.73)
509 0.85 (0.77–0.95)
829 0.90 (0.82–0.99)
656 0.96 (0.86–1.06)
372 1.34 (1.18–1.51)
1117 1.00
613 1.58 (1.43–1.75)
547 0.82 (0.74–0.91)
903 0.93 (0.85–1.02)
708 1.00 (0.91–1.11)
398 1.38 (1.22–1.55)
RR, relative risk; CI, confidence interval. Multivariate RR: adjusted for age, body mass index, education, cigarette smoking, exercise, and history of diabetes and hypertension.
AEP Vol. 15, No. 8 September 2005: 590–597
Lin et al. ALCOHOL CONSUMPTION AND MORTALITY
595
TABLE 5. Relative risks of death from all causes according to age, smoking, and alcohol intake among women in the JACC Study Nondrinkers
Ages 40–59 years No. of deaths Multivariate RR (95% CI) Ages 60–79 years No. of deaths Multivariate RR (95% CI) Nonsmokers No. of deaths Multivariate RR (95% CI) Ever smokers No. of deaths Multivariate RR (95% CI) Excluding deaths within the first 2 years of follow-up No. of deaths Multivariate RR (95% CI)
Ex-drinkers
Current drinkers (alcohol intake: g/day) 0.1–22.9
23.0–45.9
> 46.0
551 1.00
20 1.36 (0.86–2.15)
77 0.89 (0.70–1.13)
18 1.10 (0.68–1.77)
8 1.40 (0.68–2.87)
2748 1.00
100 1.37 (1.12–1.68)
204 0.82 (0.71–0.95)
40 0.76 (0.56–1.04)
13 1.18 (0.68–2.04)
2823 1.00
58 1.37 (1.05–1.77)
215 0.86 (0.75–1.00)
22 0.59 (0.39–0.90)
10 1.22 (0.66–2.27)
246 1.00
46 1.47 (1.07–2.02)
44 0.77 (0.55–1.07)
23 0.92 (0.60–1.42)
11 1.54 (0.83–2.85)
2979 1.00
104 1.30 (1.07–1.59)
259 0.86 (0.75–0.98)
55 0.89 (0.68–1.17)
19 1.25 (0.79–1.98)
RR, relative risk; CI, confidence interval. Multivariate RR: adjusted for age, body mass index, education, cigarette smoking, exercise, and history of diabetes and hypertension.
in high-density lipoprotein (HDL) cholesterol levels (22), which are inversely related to arteriosclerosis at many sites. Other factors such as insulin sensitivity, platelet aggregation, endothelial function, and inflammation may also be beneficially affected by moderate alcohol consumption (23, 24). In light of these findings, our study suggests that the relationship of moderate alcohol consumption to all-cause mortality in population studies is characterized by the lower CVD risk for moderate drinkers. Our results indicated that the benefit associated with light-to -moderate alcohol consumption may be greater among those aged 60 years or more in both men and women, and among men who were nonsmokers. Two prospective studies have reported a similar modifying effect for smoking among men (12, 19). One possible reason for the lack of effect modification in women is due to the small number of current smokers among them. In contrast, such a effect modification by smoking was not observed in studies from Western countries (25, 26). The strengths of the present study include its large size, prospective design, and the ability to control for cigarette smoking. It is unlikely that our results are biased by a preexisting disease, since the association remained unchanged after excluding subjects who died during the first 2 years of follow-up. Furthermore, we were able to separate exdrinkers from nondrinkers. As shown in our study, exdrinkers had a significantly higher risk of mortality than either nondrinkers or current drinkers. Since they may have given alcohol up because of ill health, the findings in many previous studies may have been biased because ex-drinkers were included among nondrinkers.
Our study has several limitations. First, our results relied exclusively on self-reported alcohol consumption at baseline, and that information was not updated during follow-up. While some misclassification may occur among drinking groups and subjects may change their drinking habits over time, a recent prospective study suggested that among men with an initially low alcohol consumption (< 1 drink), a subsequent moderate increase may lower their risk of CVD (27). Second, because the level of alcohol consumption has been shown to be a marker for several lifestyle factors that strongly influenced death (28), concerns were raised that the low risk of total mortality among light-to-moderate drinkers may be attributable to these favorable lifestyle factors rather than to the role of alcohol consumption itself. We found that light-to-moderate drinkers tended to be younger, better educated, and more physically active compared with nondrinkers. However, our results remained unchanged when we controlled for these confounding factors or conducted stratified analyses. Although some residual confounding may exist, it is unlikely that it has introduced a substantial bias. Third, there has been increasing interest in the association between patterns of drinking and their health consequences. Recent studies have shown that binge drinking is associated with an increased risk of death from any cause, including cardiovascular and external deaths (6, 29–32). However, we could not estimate drinking patterns such as binge drinking or drinking with meals. Although the questionnaire asked about the average amount consumed on one occasion, the participants may have consumed different amounts of alcohol on particular occasions. Our analysis was
596
Lin et al. ALCOHOL CONSUMPTION AND MORTALITY
also limited by the small number of response options for frequency of drinking, which may result in an underestimation of true levels of alcohol intake for subjects in the upper ranges of frequencies. Findings from our prospective study suggest important public health implications for middle-aged and elderly Japanese. Although making public health recommendations for alcohol use is complex in light of the known risks and benefits, those who drink and continue to do so in light-tomoderate amounts, namely, less than 23.0 g/d, will benefit from alcohol use, given the observed low risk of all-cause mortality, cancer, and CVD. On the other hand, heavy drinking is associated with an increased risk of mortality. Thus, the balance between the risks and benefits of alcohol consumption together with factors such as age, genetic predisposition, personality, lifestyle characteristics, and underlying cardiovascular risk factors must be taken into consideration before making individual decisions about appropriate alcohol use. In summary, our prospective data showed a 12% to 20% decreased risk of all-cause mortality for both Japanese men and women who consumed 0.1 to 22.9g/d of alcohol (approximately 2 drinks). The benefit associated with light-to-moderate alcohol consumption is likely to be apparent among subjects aged 60 years or older in both men and women, but only among nonsmokers in men.
The authors thank Dr. Kunio Aoki, Professor Emeritus, Nagoya University School of Medicine and the former chairman of the JACC Study Group, and Dr. Haruo Sugano, the former Director of the Cancer Institute of the Japanese Foundation for Cancer Research, for their great contribution to the initiation of the JACC study. The present members of the JACC Study and their affiliations are as follows: Dr. Akiko Tamakoshi (present chairman of the study group), Nagoya University Graduate School of Medicine; Dr. Mitsuru Mori, Sapporo Medical University School of Medicine; Dr. Yutaka Motohashi, Akita University School of Medicine; Dr. Ichiro Tsuji, Tohoku University Graduate School of Medicine; Dr. Yosikazu Nakamura, Jichi Medical School; Dr. Hiroyasu Iso, Institute of Community Medicine, University of Tsukuba; Dr. Haruo Mikami, Chiba Cancer Center; Dr. Yutaka Inaba, Juntendo University School of Medicine; Dr. Yoshiharu Hoshiyama, University of Human Arts and Sciences Graduate School; Dr. Hiroshi Suzuki, Niigata University Graduate School of Medical and Dental Sciences; Dr. Hiroyuki Shimizu, Gifu University School of Medicine; Dr. Hideaki Toyoshima, Nagoya University Graduate School of Medicine; Dr. Shinkan Tokudome, Nagoya City University Graduate School of Medicine; Dr. Yoshinori Ito, Fujita Health University School of Health Sciences; Dr. Shuji Hashimoto, Fujita Health University School of Medicine; Dr. Shogo Kikuchi, Aichi Medical University School of Medicine; Dr. Kenji Wakai, Aichi Cancer Center Research Institute; Dr. Akio Koizumi, Graduate School of Medicine and Faculty of Medicine, Kyoto University; Dr. Takashi Kawamura, Kyoto University Center for Student Health; Dr. Yoshiyuki Watanabe and Dr. Tsuneharu Miki, Kyoto Prefectural University of Medicine Graduate School of Medical Science; Dr. Chigusa Date, Faculty of Human Environmental Sciences, Mukogawa Women’s University ; Dr. Kiyomi Sakata, Wakayama Medical University; Dr. Takayuki Nose, Tottori University Faculty of Medicine; Dr. Norihiko
AEP Vol. 15, No. 8 September 2005: 590–597
Hayakawa, Research Institute for Radiation Biology and Medicine, Hiroshima University; Dr. Takesumi Yoshimura, Institute of Industrial Ecological Sciences, University of Occupational and Environmental Health, Japan; Dr. Akira Shibata, Kurume University School of Medicine; Dr. Naoyuki Okamoto, Kanagawa Cancer Center; Dr. Hideo Shio, Moriyama Municipal Hospital; Dr. Yoshiyuki Ohno (former chairman of the study group), Asahi Rosai Hospital; Dr. Tomoyuki Kitagawa, Cancer Institute of the Japanese Foundation for Cancer Research; Dr. Toshio Kuroki, Gifu University; and Dr. Kazuo Tajima, Aichi Cancer Center Research Institute. The past investigators of the study group are listed in reference 17 except for the following eight members (affiliations are those at the time they participated in the study): Dr. Takashi Shimamoto, Institute of Community Medicine, University of Tsukuba; Dr. Heizo Tanaka, Medical Research Institute, Tokyo Medical and Dental University; Dr. Shigeru Hisamichi, Tohoku University Graduate School of Medicine; Dr. Masahiro Nakao, Kyoto Prefectural University of Medicine; Dr. Takaichiro Suzuki, Research Institute, Osaka Medical Center for Cancer and Cardiovascular Diseases; Dr. Tsutomu Hashimoto, Wakayama Medical University; Dr. Teruo Ishibashi, Asama General Hospital; and Dr. Katsuhiro Fukuda, Kurume University School of Medicine.
REFERENCES 1. Mukamal KJ, Conigrave KM, Mittleman M, Camargo CA Jr, Stampfer MJ, Willett WC, et al. Roles of drinking pattern and type of alcohol consumed in coronary heart disease in men. N Engl J Med. 2003;348:109–118. 2. Thun MJ, Peto R, Lopez AD, Monaco JH, Henley SJ, Heath CW Jr, et al. Alcohol consumption and mortality among middle-aged and elderly US adults. N Engl J Med. 1997;337:1705–1714. 3. Rehm J, Greenfield TK, Rogers JD. Average volume of alcohol consumption, patterns of drinking, and all-cause mortality: Results from the US National Alcohol Survey. Am J Epidemiol. 2001;153:64–71. 4. Klatsky AL, Friedman GD, Armstrong MA, Kipp H. Wine, liquor, beer, and mortality. Am J Epidemiol. 2003;158:585–595. 5. Hart CL, Smith GD, Hole DJ, Hawthorne VM. Alcohol consumption and mortality from all causes, coronary heart diseases, and stroke: Results from a prospective cohort study of Scottish men with 21 years of follow up. BMJ. 1999;318:1725–1729. 6. Murray RP, Connett JE, Tyas SL, Bond R, Ekuma O, Silversides CK, et al. Alcohol volume, drinking pattern, and cardiovascular disease morbidity and mortality: Is there a U-shaped function? Am J Epidemiol. 2002;155:242–248. 7. Liao Y, McGee DL, Cao G, Cooper RS. Alcohol intake and mortality: Findings from the National Health Interview Surveys (1988 and 1990). Am J Epidemiol. 2000;151:651–659. 8. Valmadrid CT, Klein R, Moss SE, Klein BE, Cruickshanks KJ. Alcohol intake and the risk of coronary heart disease mortality in persons with older-onset diabetes mellitus. JAMA. 1999;282:239–246. 9. Berger K, Ajani UA, Kase CS, Gaziano JM, Burning JE, Glynn RJ, et al. Light-to-moderate alcohol consumption and the risk of stroke among US male physicians. N Engl J Med. 1999;341:1557–1564. 10. Jackson VA, Sesso HD, Burning JE, Gaziano JM. Alcohol consumption and mortality in men with preexisting cerebrovascular disease. Arch Intern Med. 2003;163:1189–1193. 11. Fuchs CS, Stampfer MJ, Colditz GA, Giovannucci EL, Manson JE, Kawachi I, et al. Alcohol consumption and mortality among women. N Engl J Med. 1995;332:1245–1250. 12. Tsugane S, Fahey MT, Sasaki S, Baba S, for the JPHC Study Group. Alcohol consumption and all-cause and cancer mortality among middleaged Japanese men: Seven-year follow-up of the JPHC Study Cohort I. Am J Epidemiol. 1999;150:1201–1207.
AEP Vol. 15, No. 8 September 2005: 590–597
Lin et al. ALCOHOL CONSUMPTION AND MORTALITY
597
13. Miyazaki M, Une H. Japanese alcoholic beverage and all cause mortality in Japanese adult men. J Epidemiol. 2001;11:219–223.
23. Albert MA, Glynn RJ, Ridker PM. Alcohol consumption and plasma concentration of C-reactive protein. Circulation. 2003;107:443–447.
14. Tsubono Y, Yamada S, Nishino Y, Tsuji I, Hisamichi S. Choice of comparison group in assessing the health effects of moderate alcohol consumption. JAMA. 2001;286:1177–1178.
24. Imhof A, Froehlich M, Brenner H, Boeing H, Pepys MB, Koenig W. Effect of alcohol consumption on systemic markers of inflammation. Lancet. 2001;357:763–767.
15. Kono S, Ikeda M, Tokudome S, Nishizumi M, Kuratsune M. Alcohol and mortality: A cohort study of male Japanese physicians. Int J Epidemiol. 1986;15:527–532.
25. Gronbaek M, Deis A, Sorensen TI, Becker U, Borch-Johnsen K, Muller C, et al. Influence of sex, age, body mass index, and smoking on alcohol intake and mortality. BMJ. 1994;308:302–306.
16. Nakaya N, Kurashima K, Yamaguchi J, Ohkubo T, Nishino Y, Tsubono Y, et al. Alcohol consumption and mortality in Japan: The Miyagi Cohort Study. J Epidemiol. 2004;14:S18–S25.
26. Camargo CA Jr, Hennekens CH, Gaziano JM, Glynn RJ, Manson JE, Stampfer MJ. Prospective study of moderate alcohol consumption and mortality in US male physicians. Arch Intern Med. 1997;157:79–85.
17. Ohno Y, Tamakoshi A. Japan Collaborative Study for Evaluation of Cancer Risk Sponsored by Monbusho (JACC study). J Epidemiol. 2001;11:144–150.
27. Sesso HD, Stampfer MJ, Rosner B, Hennekens CH, Manson JE, Gaziano JM. Seven-year changes in alcohol consumption and subsequent risk of cardiovascular disease in men. Arch Intern Med. 2002;16:2605–2612.
18. Cox D. Regression models and life-tables. J R Stat Soc. 1972;34:187– 220.
28. Lantz PM, House JS, Lepkowski JM, Williams DR, Mero RP, Chen J. Socioeconomic factors, health behaviors, and mortality: Results from a nationally representative prospective study of US adults. JAMA. 1998;279:1703–1708.
19. Yuan JM, Ross RK, Gao YT, Henderson BE, Yu MC. Follow-up study of moderate alcohol intake and mortality among middle aged men in Shanghai, China. BMJ. 1997;314:18–23. 20. Ashley MJ, Ferrence R, Room R, Rankin J, Single E. Moderate drinking and health: Report of an international symposium. CMAJ. 1994;151:809– 828. 21. Kikuchi S, Nakajima T, Kobayashi O, Yamazaki T, Kikuchi M, Mori K, et al. U-shaped effect of drinking and linear effect of smoking on risk for stomach cancer in Japan. Jpn J Cancer Res. 2002;93:953–959. 22. Rimm EB, Williams P, Fosher K, Criqui M, Stampfer MJ. Moderate alcohol intake and lower risk of coronary heart disease: Meta-analysis of effects on lipids and haemostatic factors. BMJ. 1999;319:1523–1528.
29. Chenet L, Britton A. Weekend binge drinking may be linked to Monday peaks in cardiovascular deaths. BMJ. 2001;322:998. 30. Hansagi H, Romelsjo A, Gerhardsson de Verdier M, Andreasson S, Leifman A. Alcohol consumption and stroke mortality: 20-year follow-up of 15,077 men and women. Stroke. 1995:1768–1773. 31. Kauhanen J, Kaplan GA, Goldberg DE, Salonen JT. Beer binge and mortality: Results from the Kuopio ischaemic heart prospective population-based study. BMJ. 1997;315:846–851. 32. Li G, Smith GS, Baker SP. Drinking behavior in relation to cause of death among US adults. Am J Public Health. 1994;84:1402–1406.
PURPOSE: We conducted a prospective cohort study to examine the association between alcohol intake and the risk of all-cause mortality among middle-aged and elderly Japanese men and women. METHODS: At baseline (1988–1990), a total of 110,792 Japanese men and women aged 40 to 79 years were asked to complete a questionnaire that included information on alcohol intake, and were followed up for all-cause mortality through December 31, 1999. Relative risks (95% confidence interval) were calculated using Cox proportional-hazards models. RESULTS: The risk of all-cause mortality was lowest among current drinkers with an alcohol intake of 0.1 to 22.9 g/d (RR, 0.80; 95% CI, 0.72–0.88 for men; and RR, 0.88; 95% CI, 0.77–1.00 for women). Excessive mortality associated with heavy drinking (> 69 g/d) was observed for cancer, cardiovascular disease and injuries and other external causes in men, while significantly reduced mortality with light drinking was seen for cancer in men and CVD in women. For men, the benefit associated with light alcohol consumption (! 23 g/d) was more apparent among nonsmokers than among smokers. CONCLUSION: Our prospective data show a 12% to 20% decreased risk of all-cause mortality in both Japanese men and women who consumed less than 23 g/d of alcohol (approximately 2 drinks), although heavy drinking increased that risk. Ann Epidemiol 2005;15:590–597. Ó 2005 Elsevier Inc. All rights reserved. KEY WORDS:
Alcohol Consumption, All-cause Mortality, Japan, Cardiovascular Disease, Cancer.
INTRODUCTION Epidemiological studies in Western industrialized countries have found that light-to-moderate alcohol consumption is associated with decreased total mortality, most of which were due to a reduction in cardiovascular deaths (1–11). In
From the Department of Public Health, Aichi Medical University School of Medicine, Aichi, Japan (Y.L., S.K., K.Y., Y.O.); Department of Preventive Medicine/Biostatistics and Medical Decision Making, Nagoya University Graduate School of Medicine, Nagoya, Japan (A.T.); Division of Epidemiology and Prevention, Aichi Cancer Center Research Institute, Nagoya, Japan (K.W.); Kyoto University Center for Student Health, Kyoto, Japan (T.K.); Department of Public Health Medicine, Institute of Community Health, University of Tsukuba, Tsukuba, Japan (H.I); Department of Public Health, Kurume University School of Medicine, Kyushu, Japan (I.O.); and Asama General Hospital, Nagano, Japan (T.I.). Address correspondence to: Shogo Kikuchi, M.D., Ph.D., Department of Public Health, Aichi Medical University School of Medicine, 21 Karimata, Yazako, Nagakute-cho, Aichi-gun, Aichi 480-1195, Japan. Tel.: C81-56162-3311; Fax: 81-561-62-5270. E-mail: [email protected] *Japan Collaborative Cohort Study for Evaluation of Cancer Risk sponsored by Monbukagakusho. See acknowledgments for members of the JACC Study. The JACC Study has been supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan (Nos. 61010076, 62010074, 63010074, 1010068, 2151065, 3151064, 4151063, 5151069, 6279102 and 11181101). Received July 14, 2004; accepted October 19, 2004. Ó 2005 Elsevier Inc. All rights reserved. 360 Park Avenue South, New York, NY 10010
Japan, the health effects of alcohol consumption have not been clearly defined. Although several prospective studies have been conducted to examine the relation between alcohol consumption and all-cause mortality (12–16), the findings on the protective effect of light-to-moderate alcohol consumption have been controversial, and only one study has reported the corresponding relation for Japanese women (16). A large cohort study involving 490,000 US men and women indicated that the benefit from moderate alcohol consumption depends in part on age and background cardiovascular risk (2). Given the differences in baseline characteristics among subjects, types of alcohol beverages usually consumed, or subjects’ drinking patterns, the balance of risks and benefits of alcohol consumption is likely to differ among age groups and populations. Few studies, however, have addressed this issue in Japanese populations. The main purpose of our study is to examine prospectively the relation between alcohol intake and risk of death from all causes (CVD, cancer, and injuries and other external causes) in a large cohort of Japanese men and women. In addition, we also aim to determine how age and cigarette smoking modify the relation of alcohol consumption to total mortality in our cohort. 1047-2797/05/$–see front matter doi:10.1016/j.annepidem.2004.10.010
AEP Vol. 15, No. 8 September 2005: 590–597
Selected Abbreviations and Acronyms CVD Z cardiovascular disease BMI Z body mass index
METHODS Study Cohort The Japan Collaborative Cohort Study for Evaluation of Cancer Risk (the JACC study) is a large prospective cohort study sponsored by the Japanese Ministry of Education, Culture, Sports, Science and Technology (Monbukagakusho). The details of the JACC study have been described elsewhere (17). Briefly, it was initiated between 1988 and 1990, when 127,500 apparently healthy inhabitants from 45 areas throughout Japan were enrolled as a baseline cohort. Among this cohort population, 46,465 men and 64,327 women (110,792 in total) aged 40 to 79 years were followed up until December 31, 1999. The average age of the participants at enrollment was 58 years. Subjects who reported a history of cancer, stroke, or myocardial infarction were excluded. Informed consent in the majority of the areas was obtained by having subjects sign the cover page of the questionnaire. Our study was approved by the Ethical Board of the Nagoya University School of Medicine.
Lin et al. ALCOHOL CONSUMPTION AND MORTALITY
591
times/week, and less than once/week. The amount of daily ethanol intake was calculated as follows: the average amount of alcohol use per occasion per week multiplied by the number of occasions per week and divided by seven. In the present study, we define light drinkers as people who drank less than 23 g/d of alcohol, which approximates 2 US standard drinks. One drink is assumed to contain, on average, 12g of alcohol. We define moderate drinkers as people who drank less than 46 g/d of alcohol (! 4 drinks) of alcohol, and heavy drinkers as those who drank 46 or more g/d of alcohol (> 4 drinks). Follow-up and Identification of Deaths Our primary end point was death from any cause. Follow-up surveys were conducted annually to verify the vital status of the participants, using the resident-registry data from the municipalities. These data were collected at the central office of the study group. For deceased subjects, the underlying cause was recorded from death certificates and coded according to the International Classification of Diseases, 10th Revision. The following subgroups were analyzed: subjects who died from cancer (C00–C97), CVD (I00– I99), and injuries and external causes (V01–Y98). By December 31, 1999, 11.0% of the participants had died, 85.5% were alive, and 3.5% were lost to follow-up because they moved out of the study area.
Data Collection
Statistical Analysis
At the baseline survey, we used a self-administered questionnaire to collect information on demographic characteristics, medical history of selected diseases such as diabetes mellitus and hypertension, tobacco and alcohol use, coffee consumption, participation in sports or exercise, female reproductive characteristics, and dietary habits. In the section on alcohol consumption, we first asked the subjects to choose their drinking status from three precoded response categories: nondrinkers, ex-drinkers, or current drinkers. We defined these categories as follows: nondrinkers were people who reported no alcohol drinking in the past year and ‘‘never or almost never’’ drinking; exdrinkers were people who drank at some time in their lives but not in the past year preceding the baseline survey; current drinkers were people who reported drinking in the past year. Ex-drinkers or current drinkers were then asked to report the age at which they started drinking, frequency of consumption in the past year (in the year before quitting drinking for ex-drinkers), type of beverages (Japanese sake, Japanese spirits [shochu], beer, whiskey, and wine), and the average amount consumed on one occasion (the unit was converted to gou, with 1 gou equivalent to 23 g of ethanol). The frequency of alcohol consumption was classified into four categories: almost every day, 3 to 4 times/week, 1 to 2
Separate analyses were performed for men and women. Men were grouped into the following six categories based on their daily amount of alcohol intake reported at baseline: nondrinkers, ex-drinkers, and current drinkers (whose intake was classified as 0.1–22.9, 23.0–45.9, 46.0–68.9, or 69.0 or more g/d). Because of the very small number of women in the category of 69.0 g/d or more, such women were combined with those consuming 46.0 to 68.9 g/d, resulting in five categories. Age-adjusted and multivariateadjusted relative risks (95% confidence interval) for death from all causes (cancer, CVD, and injuries and other external causes) were calculated using Cox proportionalhazards models (18). Nondrinkers served as the reference group. The multivariate analyses were adjusted for age (as a continuous variable), body mass index (!18, 18.0–21.9, 22.0–25.9, > 26.0 kg/m2), education levels more than high school and less than high school, history of diabetes (yes/ no), history of hypertension (yes/no), cigarette smoking (never, former, current smokers with 0–19, 20–39, and > 40 cigarettes/d), and participation in sports or exercise (!1 and > 1 hour/week). Tests for linear trend were conducted by treating the median value in each category of alcohol consumption as a continuous variable, with ex-drinkers being excluded in the analyses.
592
Lin et al. ALCOHOL CONSUMPTION AND MORTALITY
AEP Vol. 15, No. 8 September 2005: 590–597
We also conducted analyses stratified by age (40–59 vs. 60–79 years) and smoking status (never vs. ever) to determine their effect on the association between alcohol consumption and all-cause mortality. For these analyses, a cross-product term for each alcohol consumption level and the above variables was included in separate models. All p-values were 2-sided, with values of 0.05 or less indicating statistical significance. All analyses were performed with SAS Version 6.12 (SAS Institute Inc., Cary, NC). RESULTS Among the 110,792 cohort participants, we excluded the 8168 (7.4%) subjects who had missing information on their drinking status, and 5192 (4.7%) subjects who reported a history of cancer, stroke, or myocardial infarction. This left 97,432 (42,072 men and 55,360 women) subjects for the present analysis. During a mean follow-up of 10 years, a total of 9589 subjects (5902 men and 3687 women) died from all causes. Among 5902 deaths in men, 2418 (41.0%) died from cancer, 1584 (26.8%) from CVD, and 185 (3.1%) from injuries and other external causes. Among 3687 deaths in women, 1363 (37.0%) died from cancer, 1191 (32.3%) from CVD, and 93 (2.5%) from injuries and other external causes. Table 1 presents baseline characteristics according to the categories of alcohol consumption. Nearly 76% of the men and 25% of the women were current drinkers. Compared with nondrinkers, current drinkers tended to be younger, more likely to smoke, and have higher education levels.
Among both men and women drinkers, the percentage of current smokers increased consistently with the increasing amount of daily alcohol intake. The association of alcohol consumption with risk of death from all causes (cancer, CVD, and injuries and other external causes) is shown in Table 2 for men and Table 3 for women. In both men and women, after adjustments for age, body mass index, education, history of diabetes and hypertension, cigarette smoking, and exercise, ex-drinkers had an increased risk of all-cause mortality compared with nondrinkers. Similar results were obtained for cancer, CVD in men (Table 2). The risk of all-cause mortality was lowest in the category of current drinkers with alcohol consumption less than 23.0 g/d (RR, 0.80; 95% CI, 0.72–0.88 for men; and RR, 0.88; 95% CI, 0.77–1.00 for women). A statistically significant increase in risk of all-cause mortality was observed for men who consumed 69.0 or more g/d of alcohol (RR, 1.32; 95% CI, 1.18–1.48), but not for women who consumed 46.0 or more g/d of alcohol (RR, 1.22; 95% CI, 0.76–1.95). A high level of alcohol consumption (> 69g/d) significantly increased the risk of death from injuries and other external causes in men. Excessive mortality due to heavy drinking was also observed for cancer, CVD, and injuries or other external causes in men. On the other hand, reduced mortality among light drinkers (! 23 g/d) was seen for cancer in men and CVD in women. In subgroup analyses, we examined alcohol consumption and the risk of all-cause mortality according to age and smoking status (Tables 4 and 5). Risk reduction in all-cause mortality associated with light-to-moderate amount of
TABLE 1. Baseline charactersistics of the study population by categories of alcohol drinking in the JACC study Nondrinkers
Men No. of individuals Age (years) Body mass index (kg/m2)y More than high school education (%) History of hypertension (%) History of diabetes (%) Current smokers (%) Exercise > 5 hours per week (%) Women No. of individuals Age (years) Body mass index (kg/m2)y History of hypertension (%) History of diabetes (%) More than high school education (%) Current smokers (%) Exercise > 5 hours per week (%)
7839 59.0G10.6 22.5G2.9 11.9 14.1 5.1 49.7 5.9 40895 57.8G10.0 22.9G3.1 21.0 3.5 7.6 3.5 3.5
Ex-drinkers
2386 62.4G9.5 22.2G3.0 12.8 24.5 11.8 45.9 6.7 884 58.2G10.2 23.0G3.4 24.8 7.8 7.2 22.8 4.8
Plus-minus values are mean G stardard deviation. y Calculated as weight in kilograms divided by the square of height in meters.
Current drinkers (alcohol intake: g/day) 0.1–22.9
23.0–45.9
46.0–68.9
> 69.0
6140 56.3G10.3 22.7G2.7 18.1 15.8 5.6 46.9 6.2
8072 57.2G10.1 22.6G2.7 14.6 19.6 5.7 52.6 7.4
7085 55.5G9.3 22.8G2.6 12.8 21.0 4.4 61.7 6.4
3337 53.9G9.0 22.9G2.8 10.6 19.1 5.2 69.3 5.8
5860 55.0G9.6 22.9G2.9 16.9 2.4 10.8 7.6 5.4
1020 54.9G9.9 23.0G3.1 21.8 2.1 7.7 22.2 5.9
335 53.0G9.3 23.4G3.4 23.3 2.4 7.5 41.6 4.5
– – –
– – –
AEP Vol. 15, No. 8 September 2005: 590–597
Lin et al. ALCOHOL CONSUMPTION AND MORTALITY
593
TABLE 2. Relative risks of death according to cause and alcohol intake among men in the JACC study Nondrinkers
Ex-drinkers
Current drinkers (alcohol intake: g/day) 0.1–22.9
Person-years 75352 Death from all causes No. of deaths 1281 Age-adjusted RR 1.00 (95% CI) Multivariate RR 1.00 (95% CI) Death from cancer No. of deaths 498 Age-adjusted RR 1.00 (95% CI) Multivariate RR 1.00 (95% CI) Death from cardiovascular disease No. of deaths 329 Age-adjusted RR 1.00 (95% CI) Multivariate RR 1.00 (95% CI) Death from injuries and external causes No. of deaths 30 Age-adjusted RR 1.00 (95% CI) Multivariate RR 1.00 (95% CI)
23.0–45.9
46.0–68.9
> 69.0
P-value for trend
20847
59708
77757
68622
32076
719 1.70 (1.55–1.86)
616 0.78 (0.71–0.86)
998 0.91 (0.83–0.99)
765 0.98 (0.89–1.07)
440 1.43 (1.28–1.60)
!0.001
1.58 (1.44–1.74)
0.80 (0.72–0.88)
0.90 (0.82–0.98)
0.95 (0.86–1.04)
1.32 (1.18–1.48)
!0.001
253 1.56 (1.34–1.81)
251 0.79 (0.68–0.92)
422 0.96 (0.85–1.10)
351 1.09 (0.95–1.25)
185 1.43 (1.20–1.70)
!0.001
1.50 (1.29–1.75)
0.82 (0.70–0.95)
0.96 (0.84–1.10)
1.05 (0.91–1.20)
1.31 (1.10–1.56)
0.001
216 1.95 (1.65–2.32)
171 0.86 (0.72–1.04)
256 0.92 (0.78–1.09)
224 1.18 (0.99–1.40)
105 1.45 (1.16–1.81)
!0.001
1.79 (1.51–2.14)
0.86 (0.73–1.06)
0.89 (0.75–1.05)
1.09 (0.92–1.30)
1.28 (1.02–1.61)
0.01
17 1.70 (0.94–3.08)
28 1.35 (0.81–2.27)
35 1.23 (0.76–2.01)
25 1.11 (0.65–1.90)
18 1.88 (1.04–3.40)
0.17
1.69 (0.92–3.08)
1.52 (0.90–2.58)
1.37 (0.83–2.26)
1.23 (0.71–2.13)
1.99 (1.09–3.64)
0.11
RR, relative risk; CI, confidence interval. Multivariate RR: adjusted for age, body mass index, education, cigarette smoking, exercise, and history of diabetes and hypertension. P for trend among current drinkers.
drinking appeared to be greater in men and women aged 60 to 79 years than in those aged 40 to 59 years. However, no significant effect modification was observed. The relation between light-to-moderate amount of alcohol consumption (! 46 g/d) and the risk of all-cause mortality was significantly modified by smoking status in men (Table 4), but such effect modification was not noted in women (Table 5). Among men, the decrease in the risk for all-cause mortality was greater in never smokers than in ever smokers (Table 4). For both men and women, the estimates remained unchanged after excluding those who had died within the first 2 years of follow-up. Additional adjustment for consumption of green leafy vegetables and female reproductive factors did not materially alter the results. DISCUSSION In this large prospective study, we found that current drinkers who consumed less than 23.0 g/d of alcohol (approximately 2 drinks) had a 12% to 20% significantly decreased risk of all-cause mortality in men and women, while heavier drinking (> 69.0 g/d) was associated with a 30% increased risk in men, in relation to nondrinkers. Our findings were consistent with studies in Western industri-
alized countries and another cohort study in Japan (12), although the definition of ‘‘moderate drinking’’ and the nadir at which the lowest mortality was observed may differ across studies. In our study, deaths from cancer, CVD, and injuries or other external causes accounted for 39.4%, 28.9%, and 2.9% of total mortality, respectively. Among men, light-tomoderate alcohol consumption was significantly associated with a decreased risk of cancer. This finding was consistent with another cohort study on Japanese men, in which moderate alcohol consumption (1–149 g/week) was associated with a 50% decreased risk of cancer mortality (12). Similarly, a prospective study conducted in Shanghai, China showed a 15% reduction in risk of death from any type of cancer for light to moderate male drinkers (19). Studies from Western countries showed the similar results (2, 20). In a large US cohort study, consuming 1 drink/day was associated with a 20% decreased risk (statistically insignificant) for alcohol-related cancers in women, but such an association was not observed in men (2). An American Cancer Society study of middle-aged men found that mortality from cancer was significantly lower among those who consumed up to 1 drink/day, as compared with abstainers (20). It seems that any cancer-related benefits conferred occur only at the
594
Lin et al. ALCOHOL CONSUMPTION AND MORTALITY
AEP Vol. 15, No. 8 September 2005: 590–597
TABLE 3. Relative risks of death according to cause and alcohol intake among women in the JACC study Nondrinkers
Ex-drinkers
Current drinkers (alcohol intake: g/day) 0.1–22.9
Person-years Death from all causes No. of deaths Age-adjusted RR (95% CI) Multivariate RR (95% CI) Death from cancer No. of deaths Age-adjusted RR (95% CI) Multivariate RR (95% CI) Death from cardiovascular disease No. of deaths Age-adjusted RR (95% CI) Multivariate RR (95% CI) Death from injuries and external causes No. of deaths Age-adjusted RR (95% CI) Multivariate RR (95% CI)
23.0–45.9
P-value for trend
> 46.0
406599
8448
57611
9948
3211
2912 1.00 1.00
92 1.49 (1.21–1.83) 1.27 (1.03–1.57)
252 0.84 (0.74–0.95) 0.88 (0.77–1.00)
52 0.98 (0.74–1.29) 0.93 (0.70–1.22)
18 1.39 (0.88–2.21) 1.22 (0.76–1.95)
0.52 0.47
1054 1.00 1.00
30 1.35 (0.94–1.94) 1.21 (0.83–1.74)
119 1.00 (0.83–1.21) 1.03 (0.85–1.25)
26 1.26 (0.86–1.86) 1.20 (0.81–1.77)
6 1.09 (0.49–2.43) 1.04 (0.46–2.33)
0.42 0.53
975 1.00 1.00
20 0.96 (0.61–1.49) 0.79 (0.50–1.23)
62 0.66 (0.51–0.85) 0.69 (0.53–0.90)
12 0.71 (0.40–1.25) 0.65 (0.37–1.15)
6 1.62 (0.72–3.61) 1.22 (0.54–2.76)
0.09 0.06
67 1.00 1.00
2 1.39 (0.34–5.69) 1.69 (0.41–7.02)
13 1.60 (0.88–2.91) 2.11 (1.14–3.89)
2 1.42 (0.36–5.79) 2.01 (0.49–8.32)
– –
0.50 0.17
RR, relative risk; CI, confidence interval. Multivariate RR: adjusted for age, body mass index, education, cigarette smoking, exercise, and history of diabetes and hypertension, and P for trend among current drinkers.
lower levels of drinking. Among all cancer combined, our cohort study showed a decreased risk for men who were lightto-moderate drinkers at the lung and rectal sites (data not shown). A case–control study in Japan also found a U-shaped relationship between alcohol intake and gastric cancer risk (21). Given these findings and the fact that lung cancer and gastric cancer are the most common cancers in Japan, the reduction in risk of all cancers combined may be, in part, due to the decreased risk for lung and gastric cancer at lower levels of drinking in Japanese men.
We consider that the reduction in the risk for all-cause mortality associated with light-to-moderate alcohol consumption in this cohort population may be related to the decreased risk of deaths from CVD, since approximately 30% of deaths occurring during follow-up were attributable to CVD. In particular, we found a 31% decreased risk of morality from CVD among women who consumed 0.1 to 22.9 g/d of alcohol. Several mechanisms have been proposed to explain the protective effect of light-to-moderate alcohol consumption on CVD. The main effect is due to an increase
TABLE 4. Relative risks of death from all causes according to age, smoking, and alcohol intake among men in the JACC study Nondrinkers
Ages 40–59 years No. of deaths Multivariate RR (95% CI) Ages 60–79 years No. of deaths Multivariate RR (95% CI) Nonsmokers No. of deaths Multivariate RR (95% CI) Ever smokers No. of deaths Multivariate RR (95% CI) Excluding deaths within the first 2 years of follow-up No. of deaths Multivariate RR (95% CI)
Ex-drinkers
Current drinkers (alcohol intake: g/day) 0.1–22.9
23.0–45.9
46.0–68.9
> 69.0
190 1.00
137 2.77 (2.21–3.46)
150 0.98 (0.79–1.22)
204 1.02 (0.83–1.25)
240 1.14 (0.94–1.39)
176 1.57 (1.27–1.94)
1091 1.00
582 1.43 (1.29–1.58)
466 0.77 (0.69–0.86)
794 0.88 (0.80–0.97)
525 0.92 (0.83–1.02)
264 1.29 (1.12–1.48)
308 1.00
93 1.61 (1.27–2.04)
78 0.50 (0.39–0.65)
131 0.86 (0.70–1.07)
81 0.97 (0.75–1.25)
47 1.52 (1.11–2.07)
911 1.00
584 1.56 (1.41–1.73)
509 0.85 (0.77–0.95)
829 0.90 (0.82–0.99)
656 0.96 (0.86–1.06)
372 1.34 (1.18–1.51)
1117 1.00
613 1.58 (1.43–1.75)
547 0.82 (0.74–0.91)
903 0.93 (0.85–1.02)
708 1.00 (0.91–1.11)
398 1.38 (1.22–1.55)
RR, relative risk; CI, confidence interval. Multivariate RR: adjusted for age, body mass index, education, cigarette smoking, exercise, and history of diabetes and hypertension.
AEP Vol. 15, No. 8 September 2005: 590–597
Lin et al. ALCOHOL CONSUMPTION AND MORTALITY
595
TABLE 5. Relative risks of death from all causes according to age, smoking, and alcohol intake among women in the JACC Study Nondrinkers
Ages 40–59 years No. of deaths Multivariate RR (95% CI) Ages 60–79 years No. of deaths Multivariate RR (95% CI) Nonsmokers No. of deaths Multivariate RR (95% CI) Ever smokers No. of deaths Multivariate RR (95% CI) Excluding deaths within the first 2 years of follow-up No. of deaths Multivariate RR (95% CI)
Ex-drinkers
Current drinkers (alcohol intake: g/day) 0.1–22.9
23.0–45.9
> 46.0
551 1.00
20 1.36 (0.86–2.15)
77 0.89 (0.70–1.13)
18 1.10 (0.68–1.77)
8 1.40 (0.68–2.87)
2748 1.00
100 1.37 (1.12–1.68)
204 0.82 (0.71–0.95)
40 0.76 (0.56–1.04)
13 1.18 (0.68–2.04)
2823 1.00
58 1.37 (1.05–1.77)
215 0.86 (0.75–1.00)
22 0.59 (0.39–0.90)
10 1.22 (0.66–2.27)
246 1.00
46 1.47 (1.07–2.02)
44 0.77 (0.55–1.07)
23 0.92 (0.60–1.42)
11 1.54 (0.83–2.85)
2979 1.00
104 1.30 (1.07–1.59)
259 0.86 (0.75–0.98)
55 0.89 (0.68–1.17)
19 1.25 (0.79–1.98)
RR, relative risk; CI, confidence interval. Multivariate RR: adjusted for age, body mass index, education, cigarette smoking, exercise, and history of diabetes and hypertension.
in high-density lipoprotein (HDL) cholesterol levels (22), which are inversely related to arteriosclerosis at many sites. Other factors such as insulin sensitivity, platelet aggregation, endothelial function, and inflammation may also be beneficially affected by moderate alcohol consumption (23, 24). In light of these findings, our study suggests that the relationship of moderate alcohol consumption to all-cause mortality in population studies is characterized by the lower CVD risk for moderate drinkers. Our results indicated that the benefit associated with light-to -moderate alcohol consumption may be greater among those aged 60 years or more in both men and women, and among men who were nonsmokers. Two prospective studies have reported a similar modifying effect for smoking among men (12, 19). One possible reason for the lack of effect modification in women is due to the small number of current smokers among them. In contrast, such a effect modification by smoking was not observed in studies from Western countries (25, 26). The strengths of the present study include its large size, prospective design, and the ability to control for cigarette smoking. It is unlikely that our results are biased by a preexisting disease, since the association remained unchanged after excluding subjects who died during the first 2 years of follow-up. Furthermore, we were able to separate exdrinkers from nondrinkers. As shown in our study, exdrinkers had a significantly higher risk of mortality than either nondrinkers or current drinkers. Since they may have given alcohol up because of ill health, the findings in many previous studies may have been biased because ex-drinkers were included among nondrinkers.
Our study has several limitations. First, our results relied exclusively on self-reported alcohol consumption at baseline, and that information was not updated during follow-up. While some misclassification may occur among drinking groups and subjects may change their drinking habits over time, a recent prospective study suggested that among men with an initially low alcohol consumption (< 1 drink), a subsequent moderate increase may lower their risk of CVD (27). Second, because the level of alcohol consumption has been shown to be a marker for several lifestyle factors that strongly influenced death (28), concerns were raised that the low risk of total mortality among light-to-moderate drinkers may be attributable to these favorable lifestyle factors rather than to the role of alcohol consumption itself. We found that light-to-moderate drinkers tended to be younger, better educated, and more physically active compared with nondrinkers. However, our results remained unchanged when we controlled for these confounding factors or conducted stratified analyses. Although some residual confounding may exist, it is unlikely that it has introduced a substantial bias. Third, there has been increasing interest in the association between patterns of drinking and their health consequences. Recent studies have shown that binge drinking is associated with an increased risk of death from any cause, including cardiovascular and external deaths (6, 29–32). However, we could not estimate drinking patterns such as binge drinking or drinking with meals. Although the questionnaire asked about the average amount consumed on one occasion, the participants may have consumed different amounts of alcohol on particular occasions. Our analysis was
596
Lin et al. ALCOHOL CONSUMPTION AND MORTALITY
also limited by the small number of response options for frequency of drinking, which may result in an underestimation of true levels of alcohol intake for subjects in the upper ranges of frequencies. Findings from our prospective study suggest important public health implications for middle-aged and elderly Japanese. Although making public health recommendations for alcohol use is complex in light of the known risks and benefits, those who drink and continue to do so in light-tomoderate amounts, namely, less than 23.0 g/d, will benefit from alcohol use, given the observed low risk of all-cause mortality, cancer, and CVD. On the other hand, heavy drinking is associated with an increased risk of mortality. Thus, the balance between the risks and benefits of alcohol consumption together with factors such as age, genetic predisposition, personality, lifestyle characteristics, and underlying cardiovascular risk factors must be taken into consideration before making individual decisions about appropriate alcohol use. In summary, our prospective data showed a 12% to 20% decreased risk of all-cause mortality for both Japanese men and women who consumed 0.1 to 22.9g/d of alcohol (approximately 2 drinks). The benefit associated with light-to-moderate alcohol consumption is likely to be apparent among subjects aged 60 years or older in both men and women, but only among nonsmokers in men.
The authors thank Dr. Kunio Aoki, Professor Emeritus, Nagoya University School of Medicine and the former chairman of the JACC Study Group, and Dr. Haruo Sugano, the former Director of the Cancer Institute of the Japanese Foundation for Cancer Research, for their great contribution to the initiation of the JACC study. The present members of the JACC Study and their affiliations are as follows: Dr. Akiko Tamakoshi (present chairman of the study group), Nagoya University Graduate School of Medicine; Dr. Mitsuru Mori, Sapporo Medical University School of Medicine; Dr. Yutaka Motohashi, Akita University School of Medicine; Dr. Ichiro Tsuji, Tohoku University Graduate School of Medicine; Dr. Yosikazu Nakamura, Jichi Medical School; Dr. Hiroyasu Iso, Institute of Community Medicine, University of Tsukuba; Dr. Haruo Mikami, Chiba Cancer Center; Dr. Yutaka Inaba, Juntendo University School of Medicine; Dr. Yoshiharu Hoshiyama, University of Human Arts and Sciences Graduate School; Dr. Hiroshi Suzuki, Niigata University Graduate School of Medical and Dental Sciences; Dr. Hiroyuki Shimizu, Gifu University School of Medicine; Dr. Hideaki Toyoshima, Nagoya University Graduate School of Medicine; Dr. Shinkan Tokudome, Nagoya City University Graduate School of Medicine; Dr. Yoshinori Ito, Fujita Health University School of Health Sciences; Dr. Shuji Hashimoto, Fujita Health University School of Medicine; Dr. Shogo Kikuchi, Aichi Medical University School of Medicine; Dr. Kenji Wakai, Aichi Cancer Center Research Institute; Dr. Akio Koizumi, Graduate School of Medicine and Faculty of Medicine, Kyoto University; Dr. Takashi Kawamura, Kyoto University Center for Student Health; Dr. Yoshiyuki Watanabe and Dr. Tsuneharu Miki, Kyoto Prefectural University of Medicine Graduate School of Medical Science; Dr. Chigusa Date, Faculty of Human Environmental Sciences, Mukogawa Women’s University ; Dr. Kiyomi Sakata, Wakayama Medical University; Dr. Takayuki Nose, Tottori University Faculty of Medicine; Dr. Norihiko
AEP Vol. 15, No. 8 September 2005: 590–597
Hayakawa, Research Institute for Radiation Biology and Medicine, Hiroshima University; Dr. Takesumi Yoshimura, Institute of Industrial Ecological Sciences, University of Occupational and Environmental Health, Japan; Dr. Akira Shibata, Kurume University School of Medicine; Dr. Naoyuki Okamoto, Kanagawa Cancer Center; Dr. Hideo Shio, Moriyama Municipal Hospital; Dr. Yoshiyuki Ohno (former chairman of the study group), Asahi Rosai Hospital; Dr. Tomoyuki Kitagawa, Cancer Institute of the Japanese Foundation for Cancer Research; Dr. Toshio Kuroki, Gifu University; and Dr. Kazuo Tajima, Aichi Cancer Center Research Institute. The past investigators of the study group are listed in reference 17 except for the following eight members (affiliations are those at the time they participated in the study): Dr. Takashi Shimamoto, Institute of Community Medicine, University of Tsukuba; Dr. Heizo Tanaka, Medical Research Institute, Tokyo Medical and Dental University; Dr. Shigeru Hisamichi, Tohoku University Graduate School of Medicine; Dr. Masahiro Nakao, Kyoto Prefectural University of Medicine; Dr. Takaichiro Suzuki, Research Institute, Osaka Medical Center for Cancer and Cardiovascular Diseases; Dr. Tsutomu Hashimoto, Wakayama Medical University; Dr. Teruo Ishibashi, Asama General Hospital; and Dr. Katsuhiro Fukuda, Kurume University School of Medicine.
REFERENCES 1. Mukamal KJ, Conigrave KM, Mittleman M, Camargo CA Jr, Stampfer MJ, Willett WC, et al. Roles of drinking pattern and type of alcohol consumed in coronary heart disease in men. N Engl J Med. 2003;348:109–118. 2. Thun MJ, Peto R, Lopez AD, Monaco JH, Henley SJ, Heath CW Jr, et al. Alcohol consumption and mortality among middle-aged and elderly US adults. N Engl J Med. 1997;337:1705–1714. 3. Rehm J, Greenfield TK, Rogers JD. Average volume of alcohol consumption, patterns of drinking, and all-cause mortality: Results from the US National Alcohol Survey. Am J Epidemiol. 2001;153:64–71. 4. Klatsky AL, Friedman GD, Armstrong MA, Kipp H. Wine, liquor, beer, and mortality. Am J Epidemiol. 2003;158:585–595. 5. Hart CL, Smith GD, Hole DJ, Hawthorne VM. Alcohol consumption and mortality from all causes, coronary heart diseases, and stroke: Results from a prospective cohort study of Scottish men with 21 years of follow up. BMJ. 1999;318:1725–1729. 6. Murray RP, Connett JE, Tyas SL, Bond R, Ekuma O, Silversides CK, et al. Alcohol volume, drinking pattern, and cardiovascular disease morbidity and mortality: Is there a U-shaped function? Am J Epidemiol. 2002;155:242–248. 7. Liao Y, McGee DL, Cao G, Cooper RS. Alcohol intake and mortality: Findings from the National Health Interview Surveys (1988 and 1990). Am J Epidemiol. 2000;151:651–659. 8. Valmadrid CT, Klein R, Moss SE, Klein BE, Cruickshanks KJ. Alcohol intake and the risk of coronary heart disease mortality in persons with older-onset diabetes mellitus. JAMA. 1999;282:239–246. 9. Berger K, Ajani UA, Kase CS, Gaziano JM, Burning JE, Glynn RJ, et al. Light-to-moderate alcohol consumption and the risk of stroke among US male physicians. N Engl J Med. 1999;341:1557–1564. 10. Jackson VA, Sesso HD, Burning JE, Gaziano JM. Alcohol consumption and mortality in men with preexisting cerebrovascular disease. Arch Intern Med. 2003;163:1189–1193. 11. Fuchs CS, Stampfer MJ, Colditz GA, Giovannucci EL, Manson JE, Kawachi I, et al. Alcohol consumption and mortality among women. N Engl J Med. 1995;332:1245–1250. 12. Tsugane S, Fahey MT, Sasaki S, Baba S, for the JPHC Study Group. Alcohol consumption and all-cause and cancer mortality among middleaged Japanese men: Seven-year follow-up of the JPHC Study Cohort I. Am J Epidemiol. 1999;150:1201–1207.
AEP Vol. 15, No. 8 September 2005: 590–597
Lin et al. ALCOHOL CONSUMPTION AND MORTALITY
597
13. Miyazaki M, Une H. Japanese alcoholic beverage and all cause mortality in Japanese adult men. J Epidemiol. 2001;11:219–223.
23. Albert MA, Glynn RJ, Ridker PM. Alcohol consumption and plasma concentration of C-reactive protein. Circulation. 2003;107:443–447.
14. Tsubono Y, Yamada S, Nishino Y, Tsuji I, Hisamichi S. Choice of comparison group in assessing the health effects of moderate alcohol consumption. JAMA. 2001;286:1177–1178.
24. Imhof A, Froehlich M, Brenner H, Boeing H, Pepys MB, Koenig W. Effect of alcohol consumption on systemic markers of inflammation. Lancet. 2001;357:763–767.
15. Kono S, Ikeda M, Tokudome S, Nishizumi M, Kuratsune M. Alcohol and mortality: A cohort study of male Japanese physicians. Int J Epidemiol. 1986;15:527–532.
25. Gronbaek M, Deis A, Sorensen TI, Becker U, Borch-Johnsen K, Muller C, et al. Influence of sex, age, body mass index, and smoking on alcohol intake and mortality. BMJ. 1994;308:302–306.
16. Nakaya N, Kurashima K, Yamaguchi J, Ohkubo T, Nishino Y, Tsubono Y, et al. Alcohol consumption and mortality in Japan: The Miyagi Cohort Study. J Epidemiol. 2004;14:S18–S25.
26. Camargo CA Jr, Hennekens CH, Gaziano JM, Glynn RJ, Manson JE, Stampfer MJ. Prospective study of moderate alcohol consumption and mortality in US male physicians. Arch Intern Med. 1997;157:79–85.
17. Ohno Y, Tamakoshi A. Japan Collaborative Study for Evaluation of Cancer Risk Sponsored by Monbusho (JACC study). J Epidemiol. 2001;11:144–150.
27. Sesso HD, Stampfer MJ, Rosner B, Hennekens CH, Manson JE, Gaziano JM. Seven-year changes in alcohol consumption and subsequent risk of cardiovascular disease in men. Arch Intern Med. 2002;16:2605–2612.
18. Cox D. Regression models and life-tables. J R Stat Soc. 1972;34:187– 220.
28. Lantz PM, House JS, Lepkowski JM, Williams DR, Mero RP, Chen J. Socioeconomic factors, health behaviors, and mortality: Results from a nationally representative prospective study of US adults. JAMA. 1998;279:1703–1708.
19. Yuan JM, Ross RK, Gao YT, Henderson BE, Yu MC. Follow-up study of moderate alcohol intake and mortality among middle aged men in Shanghai, China. BMJ. 1997;314:18–23. 20. Ashley MJ, Ferrence R, Room R, Rankin J, Single E. Moderate drinking and health: Report of an international symposium. CMAJ. 1994;151:809– 828. 21. Kikuchi S, Nakajima T, Kobayashi O, Yamazaki T, Kikuchi M, Mori K, et al. U-shaped effect of drinking and linear effect of smoking on risk for stomach cancer in Japan. Jpn J Cancer Res. 2002;93:953–959. 22. Rimm EB, Williams P, Fosher K, Criqui M, Stampfer MJ. Moderate alcohol intake and lower risk of coronary heart disease: Meta-analysis of effects on lipids and haemostatic factors. BMJ. 1999;319:1523–1528.
29. Chenet L, Britton A. Weekend binge drinking may be linked to Monday peaks in cardiovascular deaths. BMJ. 2001;322:998. 30. Hansagi H, Romelsjo A, Gerhardsson de Verdier M, Andreasson S, Leifman A. Alcohol consumption and stroke mortality: 20-year follow-up of 15,077 men and women. Stroke. 1995:1768–1773. 31. Kauhanen J, Kaplan GA, Goldberg DE, Salonen JT. Beer binge and mortality: Results from the Kuopio ischaemic heart prospective population-based study. BMJ. 1997;315:846–851. 32. Li G, Smith GS, Baker SP. Drinking behavior in relation to cause of death among US adults. Am J Public Health. 1994;84:1402–1406.