Abstract

Background Regular alcohol drinking contributes both favourably and adversely to health in the Western populations, but its effects on overall and cause-specific mortality in China are still poorly understood.

Methods A nationally representative prospective cohort study included 220 000 men aged 40–79 years from 45 areas in China in 1990–91, and >40 000 deaths occurred during 15 years of follow-up. Cox regression was used to relate alcohol drinking to overall and cause-specific mortality, adjusting for age, area, smoking and education.

Results Overall, 33% of the participants reported drinking alcohol regularly at baseline, consuming mainly distilled spirits, with an estimated mean amount consumed of 372 g/week (46.5 units per week). After excluding all men with prior disease at baseline and the first 3 years of follow-up, there was a 5% [95% confidence interval (CI) 2–8] excess risk of overall mortality among regular drinkers. Compared with non-drinkers, the adjusted hazard ratios among men who drank <140, 140–279, 280–419, 420–699 and ≥700 g/week were 0.97, 1.00, 1.02, 1.12 and 1.27, respectively (P < 0.0001 for trend). The strength of the relationship appeared to be greater in smokers than in non-smokers. There was a strong positive association of alcohol drinking with mortality from stroke, oesophageal cancer, liver cirrhosis or accidental causes, a weak J-shaped association with mortality from ischaemic heart disease, stomach cancer and lung cancer and no apparent relationship with respiratory disease mortality.

Conclusion Among Chinese men aged 40–79 years, regular alcohol drinking was associated with a small but definite excess risk of overall mortality, especially among smokers.

Introduction

Alcohol is a commonly used beverage in many populations and contributes both favourably and adversely to disease mortality and morbidity. Studies in the Western populations have shown that light-to-moderate alcohol consumption is associated with decreased mortality from cardiovascular diseases (CVD)16 and certain types of cancer (e.g. thyroid cancer and renal cell carcinoma).7,8 On the other hand, alcohol drinking can increase the risks of liver cirrhosis, accidents and many other types of cancer (e.g. oral cavity, pharynx, oesophagus, larynx, liver and female breast cancer).7

Although the association of alcohol drinking with particular diseases may be generalizable from one population to the other, the effect on overall mortality may not be, because of different mixtures of specific diseases. Moreover, even for a specific disease, the effect may also differ somewhat between different populations, depending on the types of alcohol used, the style of drinking and background exposure to other risk factors.6,9,10

Alcohol has been used widely in China for centuries. Over the past three decades, there has been a >4-fold increase in per capita alcohol consumption as a result of rapid economic development and changes in lifestyle.11 Despite this recent increase, alcohol drinking remains predominately a male phenomenon, and the main type of alcoholic beverage used in China is distilled spirits,11 which differs importantly from that in most Western populations. Several prospective studies have investigated the effects of alcohol in Chinese12–18 but each has its limitations such as insufficient numbers,14,18 being confined to a single city1517 or outcome monitoring limited to a few diseases. Consequently, there is still substantial uncertainty about the effects of alcohol drinking on mortality in Chinese adults. We report here a nationally representative prospective study of alcohol and mortality among 220 000 middle-aged men in China, with over 40 000 deaths during 15 years of follow-up.

Materials and Methods

Detailed information about study design, survey methods and participants has been published previously.19,20 Briefly, the original study cohort included 225 721 men recruited from 45 areas (23 urban and 22 rural) randomly selected from China’s national Disease Surveillance Points (DSPs), which was established in 145 areas to provide a nationally representative sample of mortality statistics for the entire country.21 A typical DSP area covers 50 000–100 000 residents in four to eight geographically defined administrative units.22 During 1990–91, all men aged >40 years from two to three randomly selected units within each of 45 areas were invited to participate in a survey. About 80% of the invitees attended the survey clinics and were interviewed by trained health workers. A standardized questionnaire was used to collect information on education, occupation, tobacco, dietary patterns, exposure to indoor air pollution and self-reported health status (good, fair and poor) and prior diseases that were diagnosed by doctors [including any cancer, stroke, heart disease, chronic obstructive pulmonary disease (COPD), asthma, tuberculosis (TB), peptic ulcer, diabetes, hypertension, kidney disease, cirrhosis and chronic hepatitis]. Physical measurements included blood pressure, height, weight and peak expiratory flow rate. The information on alcohol drinking included whether the participant had drunk alcohol regularly (i.e. drank at least once a week on a regular basis) during the past 12 months, and if so, the age at which drinking began, the type (beer, wine or spirits) and the amount of each type consumed on a typical drinking week. The total amount consumed was calculated as grams (g) of pure alcohol, based on the beverage type and amount drunk, assuming the following alcohol content by volume (v/v): beer 4%, rice wine 15% and spirits 53%.

The vital status of participants was monitored regularly by DSP staff, with additional active confirmation obtained annually.22 Causes of death were sought chiefly from official death certificates. The underlying cause of death was coded centrally by the DSP staff in Beijing, using the ninth revision of the International Classification of Diseases-9. In the few deaths without any recent medical attention, standard procedures were used to determine the probable cause from symptoms described by family members.22

The present report is based on follow-up to 1 January 2006 among 218 189 men who were aged 40–79 years at baseline and had complete information on alcohol drinking. Because it is often difficult to reliably assign an underlying cause of death at older ages, all analyses were restricted to deaths occurring at ages 40–79 years, with censoring when men reached 80 years of age (or moved away from study areas). Cox proportional hazards models were used to calculate hazard ratios (HRs). All analyses were stratified by individual area and 5-year age group, and adjusted simultaneously for smoking and education. The 95% confidence interval (CI) for each log HR was estimated using the ‘floating absolute risk’ method,23 which was done by adding ±1.96 times the floated standard error of the log risk to each log HR. This method facilitates different comparisons and tests for trend between different categories, rather than just pair-wise comparisons between one arbitrarily chosen reference group and other categories. Trend tests used the mean value for each alcohol consumption category as a continuous variable. The analyses were done in SAS version 9.2.

Results

Of the 218 189 participants, 33% described themselves as regular alcohol drinkers. Among regular drinkers, the estimated mean amount consumed was 372 g/week (i.e. 46.5 units/week), higher in rural than in urban men (408 vs 276 g/week). Compared with non-drinkers, regular drinkers were slightly younger, more likely to smoke and to drink tea, had higher blood pressure and were less likely to report pre-existing diseases and poor health status (Table 1). Among regular drinkers, those who drank more were more likely to live in rural areas, were less well educated, smoked more and had higher blood pressure and longer duration of alcohol drinking (Table 1). Of the regular drinkers, 93% reported drinking strong spirits only (97% in rural and 83% in urban), with the remainder drinking beer, rice wine or mixed types of beverage.

Table 1

Baseline characteristics of participants by alcohol drinking, among men aged 40–79 years at baseline in 1990–91

    Amount drunk (g/week)
 
Characteristicsa Overall All-drinkers Non-drinkers <140 140–279 280–419 420–699 ≥700 
No. of participants 218 189 72 866 145 323 14 208 19 391 18 681 10 870 9716 
Amount drunk (g/week)  371.9  89.5 201.3 359.9 543.4 956.9 
Year of drinking  29.1  28.0 28.5 29.4 29.9 30.6 
Urban locality (%)a 27.3 27.5 27.2 39.3 32.9 24.4 19.4 14.1 
Age (years)a 54.3 53.8 54.5 53.6 54.0 53.9 53.8 53.3 
BMI (kg/m221.7 21.8 21.7 21.9 21.8 21.8 21.9 21.9 
Systolic blood pressure (mm Hg) 124.0 124.9 123.6 123.4 124.4 125.1 125.9 126.6 
Diastolic blood pressure (mm Hg) 78.7 79.2 78.5 78.6 78.8 79.3 79.8 80.4 
Expiratory peak flow (l/min) 395.6 398.7 394.0 400.5 396.0 396.9 403.2 400.3 
Education ≤6 years (%) 66.9 67.3 66.6 63.8 67.4 68.3 67.5 70.8 
Ever regular smoking (%) 73.3 86.5 66.7 82.4 86.5 87.9 88.3 88.5 
Regular tea drinking (%) 42.2 54.9 35.8 56.2 54.0 53.7 55.2 56.9 
Pre-existing disease (%) 16.0 14.9 16.6 15.6 14.0 14.2 15.2 16.8 
Self-reported poor health status (%) 7.3 5.9 8.0 6.6 5.8 5.6 5.7 5.9 
    Amount drunk (g/week)
 
Characteristicsa Overall All-drinkers Non-drinkers <140 140–279 280–419 420–699 ≥700 
No. of participants 218 189 72 866 145 323 14 208 19 391 18 681 10 870 9716 
Amount drunk (g/week)  371.9  89.5 201.3 359.9 543.4 956.9 
Year of drinking  29.1  28.0 28.5 29.4 29.9 30.6 
Urban locality (%)a 27.3 27.5 27.2 39.3 32.9 24.4 19.4 14.1 
Age (years)a 54.3 53.8 54.5 53.6 54.0 53.9 53.8 53.3 
BMI (kg/m221.7 21.8 21.7 21.9 21.8 21.8 21.9 21.9 
Systolic blood pressure (mm Hg) 124.0 124.9 123.6 123.4 124.4 125.1 125.9 126.6 
Diastolic blood pressure (mm Hg) 78.7 79.2 78.5 78.6 78.8 79.3 79.8 80.4 
Expiratory peak flow (l/min) 395.6 398.7 394.0 400.5 396.0 396.9 403.2 400.3 
Education ≤6 years (%) 66.9 67.3 66.6 63.8 67.4 68.3 67.5 70.8 
Ever regular smoking (%) 73.3 86.5 66.7 82.4 86.5 87.9 88.3 88.5 
Regular tea drinking (%) 42.2 54.9 35.8 56.2 54.0 53.7 55.2 56.9 
Pre-existing disease (%) 16.0 14.9 16.6 15.6 14.0 14.2 15.2 16.8 
Self-reported poor health status (%) 7.3 5.9 8.0 6.6 5.8 5.6 5.7 5.9 

aExcept for percentage of urban locality and mean age, all other variables were adjusted for individual area and age (5-year group) by direct standardization to the whole study population.

During ∼15 years of follow-up, 40 925 men died before 80 years of age, including 14 347 (35.1%) deaths from CVD, 9162 (22.4%) from cancer, 10 100 (24.7%) from respiratory disease and 7316 (17.9%) from other causes. A further 34 165 men (∼1% per annum) were lost to follow-up, mainly because of the demolition of whole neighbourhoods for redevelopment. There was no significant difference in the main baseline characteristics (e.g. age, blood pressure, body mass index and prevalence of smoking or drinking) between those who were lost to follow-up and those who were not.19

Total mortality had a J-shaped association with alcohol consumption both among men with any prior disease and among men without it. However, the HR between non-drinkers and light drinkers was much greater among men with prior disease (Figure 1). This extra risk among non-drinkers may well reflect disease-induced alcohol cessation. Since no separate information was available about ex-drinkers in the study, and to help limit any possible effects of reverse causality, subsequent analyses excluded all men with prior disease, as well as all deaths occurring within the first 3 years of follow-up.

Figure 1

HRs for overall mortality vs baseline alcohol drinking among men aged 40–79 years with (61 365 men, white squares) and without (156 824 men, black squares) prior disease at baseline survey in 1990–91. Analyses were stratified by area and age, and adjusted simultaneously for education and tobacco. The HRs are plotted on a floating absolute scale. Each square has an area inversely proportional to the standard error of the log risk (SE). Vertical lines indicate the corresponding 95% confidence intervals (HR × exp ± 1.96 SE). Numbers above confidence intervals are the hazard ratios and those below confidence intervals are of deaths

Figure 1

HRs for overall mortality vs baseline alcohol drinking among men aged 40–79 years with (61 365 men, white squares) and without (156 824 men, black squares) prior disease at baseline survey in 1990–91. Analyses were stratified by area and age, and adjusted simultaneously for education and tobacco. The HRs are plotted on a floating absolute scale. Each square has an area inversely proportional to the standard error of the log risk (SE). Vertical lines indicate the corresponding 95% confidence intervals (HR × exp ± 1.96 SE). Numbers above confidence intervals are the hazard ratios and those below confidence intervals are of deaths

Figure 2 and Table 2 show the associations of alcohol drinking with total and cause-specific mortality with these exclusions (based on 20 977 deaths). All-cause mortality showed a shallow J-shaped association with alcohol drinking, with the adjusted HRs of 0.97, 1.00, 1.02, 1.12 and 1.27 for those who drank <140, 140–279, 280–419, 420–699 and ≥700 g/week, respectively, compared with non-drinkers (P < 0.0001 for trend among drinkers). Alcohol drinking was hardly related to respiratory diseases mortality but positively related to all CVD, all cancer and mortality from all other non-CVD, non-cancer and non-respiratory causes. For specific diseases, men who drank 140–280 g/week (i.e. 17.5-35.0 units/week) had the lowest risk of ischaemic heart disease (IHD) (HR 0.91, 95% CI 0.75–1.11). Stroke mortality showed a positive association with alcohol drinking, with the adjusted HRs of 0.99, 1.12, 1.14, 1.21 and 1.55, respectively, for men who drank <140, 140–279, 280–419, 420–699 and ≥700 g/week (P < 0.0001 for trend among drinkers) (Table 2 and Figure 3). Alcohol drinking also had a positive association with liver cirrhosis and accidental deaths (e.g. violence, suicide and trauma), a J-shaped association with stomach cancer and lung cancer, but little association with COPD. Since spirits were almost the exclusive type of alcohol used in this male population, it was not possible to examine the effects of other types of alcohol separately.

Figure 2

HRs for death from major diseases vs baseline alcohol drinking among men aged 40–79 years without prior diseases at baseline, excluding the first 3 years of follow-up. (a) All cancer, (b) all CVD, (c) all respiratory disease and (d) others. Conventions as in Figure 1, and non-drinkers (white squares) were used as the reference category

Figure 2

HRs for death from major diseases vs baseline alcohol drinking among men aged 40–79 years without prior diseases at baseline, excluding the first 3 years of follow-up. (a) All cancer, (b) all CVD, (c) all respiratory disease and (d) others. Conventions as in Figure 1, and non-drinkers (white squares) were used as the reference category

Table 2

Numbers of deaths and standardized hazard ratios for cause-specific deaths by alcohol drinking, among men aged 40–79 years without prior diseases at baseline, excluding the first 3 years of follow-up

   Amount drunk (g/week)
 
 
Causes of death All-drinkers Non-drinkers <140 140–279 280–419 420–699 ≥700 P-value for trends* 
Cardiovascular disease         
    IHD         
        No. of deaths 435 1002 99 101 112 57 66  
        HR (95% CI) 1.08 (0.96–1.22) 1.00 (0.93–1.08) 0.99 (0.81–1.21) 0.91 (0.75–1.11) 1.17 (0.97–1.41) 1.12 (0.86–1.45) 1.60 (1.23–2.07) <0.001 
    Stroke         
        No. of deaths 1642 3002 274 400 422 260 286  
        HR (95% CI) 1.16 (1.08–1.24) 1.00 (0.96–1.05) 0.99 (0.88–1.12) 1.12 (1.01–1.24) 1.14 (1.03–1.26) 1.21 (1.07–1.37) 1.55 (1.37–1.75) <0.0001 
    Other CVD         
        No. of deaths 523 979 99 134 127 69 94  
        HR (95% CI) 1.03 (0.92–1.16) 1.00 (0.93–1.07) 1.02 (0.84–1.25) 1.01 (0.85–1.20) 0.99 (0.83–1.18) 0.96 (0.75–1.22) 1.26 (1.01–1.57) <0.0001 
    All CVD         
        No. of deaths 2600 4983 472 635 661 386 446  
        HR (95% CI) 1.12 (1.06–1.18) 1.00 (0.97–1.03) 1.00 (0.91–1.09) 1.06 (0.98–1.14) 1.11 (1.03–1.20) 1.14 (1.03–1.26) 1.48 (1.34–1.64) <0.0001 
Cancer         
    Liver         
        No. of deaths 412 703 62 108 103 79 60  
        HR (95% CI) 1.12 (0.98–1.28) 1.00 (0.91–1.09) 0.98 (0.76–1.25) 1.13 (0.94–1.37) 1.02 (0.84–1.25) 1.38 (1.11–1.73) 1.21 (0.92–1.09) 0.24 
        Stomach         
        No. of deaths 323 814 67 82 81 47 46  
        HR (95% CI) 0.96 (0.84–1.10) 1.00 (0.92–1.09) 0.87 (0.68–1.11) 0.95 (0.77–1.19) 0.96 (0.77–1.20) 1.00 (0.75–1.34) 1.15 (0.85–1.57) 0.02 
    Lung         
        No. of deaths 404 678 60 94 105 77 68  
        HR (95% CI) 1.03 (0.90–1.18) 1.00 (0.92–1.09) 0.86 (0.66–1.11) 0.92 (0.75–1.13) 1.02 (0.84–1.24) 1.30 (1.03–1.63) 1.25 (0.97–1.62) 0.04 
    Oesophageal         
        No. of deaths 242 606 62 62 52 36 30  
        HR (95% CI) 1.25 (1.07–1.47) 1.00 (0.91–1.10) 1.19 (0.93–1.53) 1.28 (1.00–1.65) 1.11 (0.84–1.46) 1.38 (0.99–1.94) 1.63 (1.12–2.39) 0.2 
    Colorectal         
        No. of deaths 63 130 18 11 16 10  
        HR (95% CI) 0.90 (0.65–1.24) 1.00 (0.82–1.22) 1.26 (0.79–2.02) 0.60 (0.33–1.08) 0.88 (0.53–1.45) 0.77 (0.38–1.55) 1.06 (0.55–2.06) 0.5 
    Other sites         
        No. of deaths 208 397 44 50 43 33 41  
        HR (95% CI) 1.05 (0.88–1.26) 1.00 (0.89–1.12) 1.15 (0.85–1.55) 0.96 (0.72–1.27) 0.85 (0.63–1.15) 1.10 (0.77–1.56) 1.39 (1.00–1.94) 0.6 
    All cancers         
        No. of deaths 1655 3328 313 407 400 280 255  
        HR (95% CI) 1.06 (1.00–1.13) 1.00 (0.96–1.04) 0.99 (0.89–1.11) 1.01 (0.92–1.12) 0.99 (0.90–1.10) 1.22 (1.08–1.38) 1.28 (1.12–1.45) <0.001 
Respiratory disease         
    COPD         
        No. of deaths 1298 2589 242 386 343 175 152  
        HR (95% CI) 0.92 (0.86–0.99) 1.00 (0.96–1.04) 0.93 (0.81–1.05) 0.96 (0.86–1.06) 0.94 (0.84–1.05) 0.90 (0.77–1.05) 0.86 (0.73–1.01) 0.93 
    TB         
        No. of deaths 85 156 11 18 20 15 21  
        HR (95% CI) 0.89 (0.67–1.17) 1.00 (0.84–1.19) 0.80 (0.44–1.46) 0.71 (0.45–1.14) 0.75 (0.48–1.18) 0.99 (0.59–1.65) 1.45 (0.91–2.30) 0.03 
    All respiratory disease         
        No. of deaths 1545 2965 275 434 405 227 204  
        HR (95% CI) 0.93 (0.87–0.99) 1.00 (0.96–1.04) 0.93 (0.82–1.05) 0.92 (0.84–1.01) 0.92 (0.83–1.02) 0.96 (0.84–1.09) 0.95 (0.82–1.10) 0.11 
Other diseases         
    Cirrhosis         
        No. of deaths 163 255 27 30 44 27 35  
        HR (95% CI) 1.15 (0.93–1.42) 1.00 (0.87–1.16) 1.12 (0.76–1.64) 0.84 (0.59–1.21) 1.16 (0.85–1.56) 1.22 (0.83–1.79) 1.74 (1.22–2.48) <0.01 
Violence, suicide and other trauma         
    No. of deaths 543 863 68 120 155 98 102  
    HR (95% CI) 1.20 (1.07–1.34) 1.00 (0.92–1.08) 0.90 (0.71–1.15) 1.10 (0.92–1.32) 1.35 (1.15–1.58) 1.36 (1.11–1.67) 1.34 (1.08–1.65) 0.04 
All causes         
    No. of deaths 7237 13740 1273 1838 1833 1137 1156  
    Mortality ratea 1116 1046 1040 1589 1152 1167 4064  
    HR (95% CI) 1.05 (1.02–1.08) 1.00 (0.98–1.02) 0.97 (0.91–1.02) 1.00 (0.96–1.05) 1.02 (0.97–1.07) 1.12 (1.06–1.19) 1.27 (1.19–1.35) <0.0001 
   Amount drunk (g/week)
 
 
Causes of death All-drinkers Non-drinkers <140 140–279 280–419 420–699 ≥700 P-value for trends* 
Cardiovascular disease         
    IHD         
        No. of deaths 435 1002 99 101 112 57 66  
        HR (95% CI) 1.08 (0.96–1.22) 1.00 (0.93–1.08) 0.99 (0.81–1.21) 0.91 (0.75–1.11) 1.17 (0.97–1.41) 1.12 (0.86–1.45) 1.60 (1.23–2.07) <0.001 
    Stroke         
        No. of deaths 1642 3002 274 400 422 260 286  
        HR (95% CI) 1.16 (1.08–1.24) 1.00 (0.96–1.05) 0.99 (0.88–1.12) 1.12 (1.01–1.24) 1.14 (1.03–1.26) 1.21 (1.07–1.37) 1.55 (1.37–1.75) <0.0001 
    Other CVD         
        No. of deaths 523 979 99 134 127 69 94  
        HR (95% CI) 1.03 (0.92–1.16) 1.00 (0.93–1.07) 1.02 (0.84–1.25) 1.01 (0.85–1.20) 0.99 (0.83–1.18) 0.96 (0.75–1.22) 1.26 (1.01–1.57) <0.0001 
    All CVD         
        No. of deaths 2600 4983 472 635 661 386 446  
        HR (95% CI) 1.12 (1.06–1.18) 1.00 (0.97–1.03) 1.00 (0.91–1.09) 1.06 (0.98–1.14) 1.11 (1.03–1.20) 1.14 (1.03–1.26) 1.48 (1.34–1.64) <0.0001 
Cancer         
    Liver         
        No. of deaths 412 703 62 108 103 79 60  
        HR (95% CI) 1.12 (0.98–1.28) 1.00 (0.91–1.09) 0.98 (0.76–1.25) 1.13 (0.94–1.37) 1.02 (0.84–1.25) 1.38 (1.11–1.73) 1.21 (0.92–1.09) 0.24 
        Stomach         
        No. of deaths 323 814 67 82 81 47 46  
        HR (95% CI) 0.96 (0.84–1.10) 1.00 (0.92–1.09) 0.87 (0.68–1.11) 0.95 (0.77–1.19) 0.96 (0.77–1.20) 1.00 (0.75–1.34) 1.15 (0.85–1.57) 0.02 
    Lung         
        No. of deaths 404 678 60 94 105 77 68  
        HR (95% CI) 1.03 (0.90–1.18) 1.00 (0.92–1.09) 0.86 (0.66–1.11) 0.92 (0.75–1.13) 1.02 (0.84–1.24) 1.30 (1.03–1.63) 1.25 (0.97–1.62) 0.04 
    Oesophageal         
        No. of deaths 242 606 62 62 52 36 30  
        HR (95% CI) 1.25 (1.07–1.47) 1.00 (0.91–1.10) 1.19 (0.93–1.53) 1.28 (1.00–1.65) 1.11 (0.84–1.46) 1.38 (0.99–1.94) 1.63 (1.12–2.39) 0.2 
    Colorectal         
        No. of deaths 63 130 18 11 16 10  
        HR (95% CI) 0.90 (0.65–1.24) 1.00 (0.82–1.22) 1.26 (0.79–2.02) 0.60 (0.33–1.08) 0.88 (0.53–1.45) 0.77 (0.38–1.55) 1.06 (0.55–2.06) 0.5 
    Other sites         
        No. of deaths 208 397 44 50 43 33 41  
        HR (95% CI) 1.05 (0.88–1.26) 1.00 (0.89–1.12) 1.15 (0.85–1.55) 0.96 (0.72–1.27) 0.85 (0.63–1.15) 1.10 (0.77–1.56) 1.39 (1.00–1.94) 0.6 
    All cancers         
        No. of deaths 1655 3328 313 407 400 280 255  
        HR (95% CI) 1.06 (1.00–1.13) 1.00 (0.96–1.04) 0.99 (0.89–1.11) 1.01 (0.92–1.12) 0.99 (0.90–1.10) 1.22 (1.08–1.38) 1.28 (1.12–1.45) <0.001 
Respiratory disease         
    COPD         
        No. of deaths 1298 2589 242 386 343 175 152  
        HR (95% CI) 0.92 (0.86–0.99) 1.00 (0.96–1.04) 0.93 (0.81–1.05) 0.96 (0.86–1.06) 0.94 (0.84–1.05) 0.90 (0.77–1.05) 0.86 (0.73–1.01) 0.93 
    TB         
        No. of deaths 85 156 11 18 20 15 21  
        HR (95% CI) 0.89 (0.67–1.17) 1.00 (0.84–1.19) 0.80 (0.44–1.46) 0.71 (0.45–1.14) 0.75 (0.48–1.18) 0.99 (0.59–1.65) 1.45 (0.91–2.30) 0.03 
    All respiratory disease         
        No. of deaths 1545 2965 275 434 405 227 204  
        HR (95% CI) 0.93 (0.87–0.99) 1.00 (0.96–1.04) 0.93 (0.82–1.05) 0.92 (0.84–1.01) 0.92 (0.83–1.02) 0.96 (0.84–1.09) 0.95 (0.82–1.10) 0.11 
Other diseases         
    Cirrhosis         
        No. of deaths 163 255 27 30 44 27 35  
        HR (95% CI) 1.15 (0.93–1.42) 1.00 (0.87–1.16) 1.12 (0.76–1.64) 0.84 (0.59–1.21) 1.16 (0.85–1.56) 1.22 (0.83–1.79) 1.74 (1.22–2.48) <0.01 
Violence, suicide and other trauma         
    No. of deaths 543 863 68 120 155 98 102  
    HR (95% CI) 1.20 (1.07–1.34) 1.00 (0.92–1.08) 0.90 (0.71–1.15) 1.10 (0.92–1.32) 1.35 (1.15–1.58) 1.36 (1.11–1.67) 1.34 (1.08–1.65) 0.04 
All causes         
    No. of deaths 7237 13740 1273 1838 1833 1137 1156  
    Mortality ratea 1116 1046 1040 1589 1152 1167 4064  
    HR (95% CI) 1.05 (1.02–1.08) 1.00 (0.98–1.02) 0.97 (0.91–1.02) 1.00 (0.96–1.05) 1.02 (0.97–1.07) 1.12 (1.06–1.19) 1.27 (1.19–1.35) <0.0001 

aMortality rates are standardized to the geographical area and the 5-year age group structures in the study population aged 40–79 years.

*P-value for trends among alcohol drinkers.

Table 3 shows HRs of total mortality by alcohol drinking subdivided by smoking status, with additional data shown in Figure 4 for other baseline variables. Among current smokers, there appeared to be a dose–response relation of amount drunk with overall mortality, with greater excess risk among heavy smokers or those who began to smoke at a younger age (P < 0.0001 for interaction across each and overall smoking categories). Among non-smokers, no apparent dose–response relationship was seen between alcohol drinking and overall mortality (Table 3), but this heterogeneity in effects of alcohol on total mortality by smoking status was not significant (P = 0.07, Figure 4). With respect to other baseline variables, the effect of alcohol drinking on overall mortality did not seem to be modified significantly by age, locality, education, blood pressure, lung function or baseline health status (Figure 4).

Figure 3

HRs for cause-specific mortality vs baseline alcohol drinking among men aged 40–79 years without prior diseases at baseline, excluding the first 3 years of follow-up. (a) IHD and (b) stroke. Conventions as in Figures 1 and 2

Figure 3

HRs for cause-specific mortality vs baseline alcohol drinking among men aged 40–79 years without prior diseases at baseline, excluding the first 3 years of follow-up. (a) IHD and (b) stroke. Conventions as in Figures 1 and 2

Figure 4

HR of overall mortality for regular drinkers vs non-drinkers by baseline variables among men aged 40–79 years without prior disease at baseline, excluding the first 3 years of follow-up. Each closed square represents a HR with area inversely proportional to the variance of log HR. The dotted vertical line indicates the overall HR; the open diamond indicates summary result and its 95% CI

Figure 4

HR of overall mortality for regular drinkers vs non-drinkers by baseline variables among men aged 40–79 years without prior disease at baseline, excluding the first 3 years of follow-up. Each closed square represents a HR with area inversely proportional to the variance of log HR. The dotted vertical line indicates the overall HR; the open diamond indicates summary result and its 95% CI

Table 3

Standardized hazard ratios for total mortality by alcohol drinking and smoking status, among men aged 40–79 years without prior diseases at baseline excluding the first 3 years of follow-up

    Alcohol drinking category
 
 Non-drinker
 
<280 g/week
 
280–700 g/week
 
≥700 g/week
 
Smoking category No. of deaths HR (95% CI)b No. of deaths HR (95% CI)b No. of deaths HR (95% CI)b No. of deaths HR (95% CI)b 
Never smokers 4132 1.00 (0.97–1.03)a 372 0.94 (0.85–1.04) 352 0.97 (0.87–1.08) 163 1.08 (0.92–1.27) 
Ex-smokers 486 1.14 (1.04–1.25) 103 0.90 (0.74–1.10) 79 0.93 (0.74–1.16) 46 1.51 (1.12–2.02) 
Current smokers 9122 1.12 (1.10–1.15) 2636 1.14 (1.09–1.18) 2539 1.21 (1.17–1.26) 947 1.46 (1.37–1.56) 
    Amount smoked (g/day)         
        <20 4683 1.11 (1.08–1.15) 1310 1.08 (1.02–1.14) 842 1.19 (1.11–1.27) 261 1.48 (1.31–1.68) 
        20–30 2973 1.12 (1.08–1.16) 863 1.14 (1.06–1.22) 900 1.24 (1.16–1.32) 330 1.42 (1.27–1.58) 
        ≥30 1456 1.15 (1.09–1.21) 461 1.29 (1.18–1.42) 797 1.23 (1.14–1.32) 355 1.51 (1.36–1.68) 
    Age began smoking (years)         
        ≥25 2735 1.05 (1.01–1.09) 771 1.08 (1.00–1.16) 532 1.12 (1.03–1.22) 154 1.31 (1.12–1.53) 
        20–25 3843 1.15 (1.11–1.19) 1138 1.14 (1.08–1.21) 1105 1.26 (1.19–1.34) 361 1.39 (1.25–1.55) 
        <20 2536 1.16 (1.12–1.21) 727 1.18 (1.10–1.27) 902 1.23 (1.15–1.32) 432 1.61 (1.47–1.78) 
    Alcohol drinking category
 
 Non-drinker
 
<280 g/week
 
280–700 g/week
 
≥700 g/week
 
Smoking category No. of deaths HR (95% CI)b No. of deaths HR (95% CI)b No. of deaths HR (95% CI)b No. of deaths HR (95% CI)b 
Never smokers 4132 1.00 (0.97–1.03)a 372 0.94 (0.85–1.04) 352 0.97 (0.87–1.08) 163 1.08 (0.92–1.27) 
Ex-smokers 486 1.14 (1.04–1.25) 103 0.90 (0.74–1.10) 79 0.93 (0.74–1.16) 46 1.51 (1.12–2.02) 
Current smokers 9122 1.12 (1.10–1.15) 2636 1.14 (1.09–1.18) 2539 1.21 (1.17–1.26) 947 1.46 (1.37–1.56) 
    Amount smoked (g/day)         
        <20 4683 1.11 (1.08–1.15) 1310 1.08 (1.02–1.14) 842 1.19 (1.11–1.27) 261 1.48 (1.31–1.68) 
        20–30 2973 1.12 (1.08–1.16) 863 1.14 (1.06–1.22) 900 1.24 (1.16–1.32) 330 1.42 (1.27–1.58) 
        ≥30 1456 1.15 (1.09–1.21) 461 1.29 (1.18–1.42) 797 1.23 (1.14–1.32) 355 1.51 (1.36–1.68) 
    Age began smoking (years)         
        ≥25 2735 1.05 (1.01–1.09) 771 1.08 (1.00–1.16) 532 1.12 (1.03–1.22) 154 1.31 (1.12–1.53) 
        20–25 3843 1.15 (1.11–1.19) 1138 1.14 (1.08–1.21) 1105 1.26 (1.19–1.34) 361 1.39 (1.25–1.55) 
        <20 2536 1.16 (1.12–1.21) 727 1.18 (1.10–1.27) 902 1.23 (1.15–1.32) 432 1.61 (1.47–1.78) 

aNon-drinkers and never smokers were used as the reference group.

bHRs were adjusted for education and stratified by age and area. The 95% CI was estimated using the ‘floating absolute risk’ method. Participants with missing values were excluded from the analyses.

Discussion

This is one of the largest prospective studies of the relationship between alcohol and total and cause-specific mortality in China. In this nationally representative male population, about one-third reported drinking alcohol regularly, which involved almost exclusively strong spirits. Overall, among men who had no history of prior disease and did not die during the first 3 years of follow-up, there was a 5% excess risk of total mortality among regular drinkers and the association was largely independent of age, area, education and smoking status.

Alcohol drinking and total mortality

Most previous studies in Western populations have reported a U- or J-shaped relation between alcohol consumption and total mortality.1,2,5,6,2426 In a cohort study of male British doctors involving 7000 deaths, the relative risks of total mortality among men who drank 1–7, 8–14, 15–28 and ≥29 units/week were 0.79, 0.75, 0.8 and 0.93, compared with non-drinkers.1 In a large prospective study among 490 000 US adults (including 238 000 men with 29 000 deaths), the overall death rates were lowest in those reporting one drink per day (i.e. <100 g/week), after which the total mortality among drinkers gradually increased.6 Since many Asians cannot metabolize alcohol effectively, it was hypothesized that the effects of alcohol on health in Asians may differ importantly from those in Caucasians.27 However, there does not seem to be strong supportive evidence for this from the studies in Asian populations.16,17,28,29 A cohort study of 18 244 Chinese men aged 45–64 years in Shanghai, among whom 41% were current drinkers, showed that moderate drinkers had a reduced risk of total mortality (HRs: 0.80, 0.82, 0.84, 0.84, 1.03 and 1.30 for those who drank ≤70, 80–140, 150–210, 220–280, 290–420 and ≥420 g/week, respectively, compared with non-drinkers). The results persisted after exclusion of ex-drinkers, people with prior diseases and the first year of follow-up.17 Similarly, in a large Japanese cohort study of 98 000 individuals (including 42 000 men, of whom 76% were current drinkers) with 10 000 deaths, current male drinkers who consumed <23 g/day of alcohol had a 20% lower risk of total mortality than non-drinkers, whereas heavier drinking (>69 g/day) was associated with a 30% increased risk.29

In the present study, although the size of excess risk associated with heavy drinking appeared to be generally consistent with that reported in other Asian cohort studies, the reduced risk among men who drank lightly or moderately (e.g. <280 g/week) seemed less pronounced. The reasons for the discrepancy are not entirely clear, but it may be due to a number of factors such as differences in the types of alcohol used between different study populations (e.g. 48% of alcohol use involved spirits in the Shanghai cohort vs 93% in the present study), composition of causes of death (e.g. 27% of deaths from CVD in the Japanese cohort vs 35% in our study) and completeness which reverse causality was dealt with. There is evidence from studies in the West that for a given quantity of ethanol, spirits may be associated with greater hazards for certain conditions such as cancer and IHD compared with other types of alcoholic beverage.26 If this were also true in the Chinese population, then it could help to explain in part the lack of any apparent beneficial effects of light-to-moderate drinking on overall mortality in the present study.17,29

In the present study, the mean amount of alcohol consumed per drinker (372 g/week) appeared to be somewhat higher than that reported in some other studies in China,17 and this disparity could be accounted for mainly by differences in the types of alcohol used between various parts of China and at different time periods. Indeed, when restricting analysis to urban men, the level of consumption in the present study (276 g/week) was generally compatible with that reported in other studies of similar populations (e.g. 217 g/week in the Shanghai cohort).17 Likewise, in a recent study of over 200 000 adult men recruited from 10 diverse regions across China in 2004–08, 70% of regular drinkers reported drinking spirits (as opposed to 93% in the present study), with about one-third of them drinking weak spirits (i.e. <40% v/v) rather than the strong spirits (i.e. 53% v/v) seen in previous decades. Nevertheless, in that study the mean consumption per drinker was similar to that in the present study, being 238 g/week for urban and 333 g/week for rural men.30 As Chinese drinkers gradually move away from strong spirits to weak spirits or other non-spirits beverages, the mean levels of consumption per drinker and the associated health effects may change as well.

Alcohol drinking and cause-specific mortality

A lower risk of CVD, especially IHD, mortality associated with light-to-moderate alcohol drinking, has been reported in many,6,26,31 although not all,10 cohort studies in the Western populations. A few Asian studies have also reported an inverse association between light-to-moderate alcohol drinking and IHD, whereas no such relationship was seen for stroke.12,1517,32 In a study of 64 000 Chinese men aged >40 years, with 976 (588 fatal) incident cases of IHD and 3434 (1848 fatal) incident strokes, alcohol drinking was associated with a reduced risk of IHD, with HRs for IHD mortality being 1.20, 0.59 and 0.53 for those who drank 1–6, 7–34 and ≥35 drinks/week, respectively, compared with non-drinkers (the corresponding HRs for IHD incidence were 0.99, 0.67 and 0.58, respectively).12 In contrast, heavy drinking was associated with increased risk of stroke, with HRs among participants consuming 1–6, 7–20, 21–34 and ≥35 drinks/week being 0.93, 0.98, 1.15 and 1.30, respectively, for stroke mortality and 0.92, 1.02, 1.22 and 1.22, respectively, for stroke incidence.13 A cohort study of 18 000 men in Shanghai with 104 IHD deaths and 269 stroke deaths also found that light-to-moderate drinking (<280 g/week) was associated with a 36% lower risk of IHD mortality, whereas for stroke, no such association was seen with light-to-moderate drinking but a significant 70% excess risk of stroke death among heavy drinkers (i.e. >280 g/week).17 However, a cohort study in Hong Kong of 19 000 men aged >65 years with 188 IHD deaths showed that light-to-moderate alcohol use was not associated with lower risk of IHD mortality.33 In the present cohort, although there was a slightly J-shaped association between alcohol drinking and IHD mortality, the confidence intervals were too big to draw any conclusion. For stroke, the present study with more cases than all previous Chinese studies combined provides further supportive evidence that the risk was positively associated with alcohol drinking, with heavy drinkers (≥700 g/week) having about a 60% excess mortality.

Alcohol has been classified by the International Agency for Research on Cancer as a Group 1 carcinogen, i.e. carcinogenic to humans, with particularly strong adverse effects on cancers of the oral cavity, pharynx, larynx, oesophagus, colorectum and liver.34,35 However, most evidence is based on studies in the Western populations, in which the main types of cancer and their underlying causes differ importantly from those in Chinese. As in a few other studies in China,1517 we did not find any excess risk of cancer deaths among light-to-moderate drinkers, but among heavy drinkers there was a significantly increased risk of cancer mortality, especially for oesophageal cancer. Although alcohol drinking is an independent risk factor for liver cancer in most Western populations,35 the main cause of the high incidence of liver cancer in China is chronic infection with hepatitis B virus. In our study, there was only a weak positive association of alcohol drinking with liver cancer mortality. For liver cirrhosis, however, there was a strong positive association with alcohol drinking, confirming the finding of a few other studies in Chinese and Asian populations.15,17 In our study, we found no apparent relationship between alcohol drinking and COPD mortality, which accounts for a large proportion of deaths. In contrast, a number of studies in Western and other Chinese populations have reported a U-shaped relationship between alcohol and COPD mortality.15,36 These studies typically involved small numbers of COPD deaths and generally lacked proper exclusion of people with prior respiratory disease or complete adjustment for other potential confounding factors such as smoking. So, the U-shaped relationship reported in these previous studies may well be an artefact. Consistent with several other studies,6,8,16, 17,28,29,37,38 the present study also found an interaction between smoking and alcohol drinking on overall mortality, with heavy drinkers and heavy smokers or those who began to smoke before 20 years of age having the highest overall mortality, again highlighting the complexity of disease causation and synergistic effects of multivariate exposures for common chronic diseases.

Study strengths and limitations

The present prospective study included a large sample size and a structured questionnaire, so it is unlikely that the study findings could be due to chance or recall bias. Although we have tried to control rigorously for all major confounders, residual confounding may still exist.28 It is certain that an unknown proportion of non-drinkers were abstainers due to ill health, which could account for at least some of the excess risk in non-drinkers.39 Although we were not able to separate ex-drinkers from non-drinkers, we have excluded from the main analyses participants with any prior diseases and the first 3 years of follow-up, thus we believe the effects of reverse causality have largely been controlled for. It has been suggested that different types of alcoholic beverage may contribute differently to disease risk, with wine drinking reportedly having more beneficial effects on IHD and certain cancers compared with other types of alcohol,5,26,40,41 but this question cannot be examined effectively in the present study because distilled spirits were the type of alcohol that was most exclusively used. Another limitation of the study is that only information on average alcohol consumption was collected at baseline, so we could not assess the effect of different drinking patterns (such as binge drinking or drinking with meals) on mortality. As in many other prospective studies, we only assessed alcohol exposure at baseline, and it is likely that alcohol intake patterns have changed over time, leading to misclassification for some individuals, which will tend to dilute rather than augment any real association.

Summary

Globally, alcohol is responsible for ∼4% of all deaths and 5% of the global burden of disease, but its real hazards could still be underestimated, especially since some health benefits associated with light-to-moderate alcohol drinking may well not be causal.42 In the present study of adult Chinese men, regular alcohol drinking is associated with a 5% excess risk of overall mortality, with significant excess mortality among heavy drinkers from stroke, IHD, liver cirrhosis, injury and certain types of cancer. The present study does not provide direct evidence about all aspects of alcohol drinking, nor does it provide direct evidence about the effects of alcohol in women or on non-fatal morbidity. As China continues to develop economically, the prevalence of alcohol drinking is likely to continue to increase, as it has since the 1980s,11 and there may be further changes in the types of alcohol consumed and the ways in which people drink. These, together with other changes in the population (e.g. smoking and other lifestyle factors), may importantly affect the future health effects of alcohol drinking in the population.

Funding

The Chinese Ministry of Health; UK Medical Research Council, British Heart Foundation, Cancer Research UK; the World Bank loan to China; the Canadian International Development Research Centre.

Acknowledgements

We thank the study participants and the China DSP staff in 45 areas. We also acknowledge support received from the BHF Centre for Research Excellence, Oxford.

Conflict of interest: None declared.

KEY MESSAGES

  • Among 220 000 Chinese men aged 40–79 years in 1990–91, 33% of the participants reported drinking alcohol regularly at baseline, consuming mainly distilled spirits.

  • After excluding all men with prior disease at baseline and the first 3 years of follow-up, there was a 5% (95% CI 2–8) excess risk of overall mortality among regular drinkers.

  • Compared with non-drinkers, the adjusted HRs for overall mortality were 0.97, 1.00, 1.02, 1.12 and 1.27 (P < 0.0001 for trend) for men drinking <140, 140–279, 280–419, 420–699 and ≥700 g/week, respectively. The strength of the relationship appeared to be greater in smokers than in non-smokers.

References

1
Doll
R
Peto
R
Boreham
J
Sutherland
I
Mortality in relation to alcohol consumption: a prospective study among male British doctors
Int J Epidemiol
 , 
2005
, vol. 
34
 (pg. 
199
-
204
)
2
Gaziano
JM
Gaziano
TA
Glynn
RJ
, et al.  . 
Light-to-moderate alcohol consumption and mortality in the physicians' health study enrollment cohort
J Am Coll Cardiol
 , 
2000
, vol. 
35
 (pg. 
96
-
105
)
3
Gordon
T
Kannel
WB
Drinking and mortality. The Framingham Study
Am J Epidemiol
 , 
1984
, vol. 
120
 (pg. 
97
-
107
)
4
Kozararevic
D
McGee
D
Vojvodic
N
, et al.  . 
Frequency of alcohol consumption and morbidity and mortality: the Yugoslavia Cardiovascular Disease Study
Lancet
 , 
1980
, vol. 
1
 (pg. 
613
-
16
)
5
Maskarinec
G
Meng
L
Kolonel
LN
Alcohol intake, body weight, and mortality in a multiethnic prospective cohort
Epidemiology
 , 
1998
, vol. 
9
 (pg. 
654
-
61
)
6
Thun
MJ
Peto
R
Lopez
AD
, et al.  . 
Alcohol consumption and mortality among middle-aged and elderly U.S. adults
N Engl J Med
 , 
1997
, vol. 
337
 (pg. 
1705
-
14
)
7
Baan
R
Straif
K
Grosse
Y
, et al.  . 
Carcinogenicity of alcoholic beverages
Lancet Oncol
 , 
2007
, vol. 
8
 (pg. 
292
-
93
)
8
Allen
NE
Beral
V
Casabonne
D
, et al.  . 
Moderate alcohol intake and cancer incidence in women
J Natl Cancer Inst
 , 
2009
, vol. 
101
 (pg. 
296
-
305
)
9
Zaridze
D
Brennan
P
Boreham
J
, et al.  . 
Alcohol and cause-specific mortality in Russia: a retrospective case-control study of 48,557 adult deaths
Lancet
 , 
2009
, vol. 
373
 (pg. 
2201
-
14
)
10
Hart
CL
Smith
GD
Hole
DJ
Hawthorne
VM
Alcohol consumption and mortality from all causes, coronary heart disease, and stroke: results from a prospective cohort study of Scottish men with 21 years of follow up
BMJ
 , 
1999
, vol. 
318
 (pg. 
1725
-
29
)
11
Hao
W
Su
Z
Liu
B
, et al.  . 
Drinking and drinking patterns and health status in the general population of five areas of China
Alcohol Alcohol
 , 
2004
, vol. 
39
 (pg. 
43
-
52
)
12
Bazzano
LA
Gu
D
Reynolds
K
, et al.  . 
Alcohol consumption and risk of coronary heart disease among Chinese men
Int J Cardiol
 , 
2009
, vol. 
135
 (pg. 
78
-
85
)
13
Bazzano
LA
Gu
D
Reynolds
K
, et al.  . 
Alcohol consumption and risk for stroke among Chinese men
Ann Neurol
 , 
2007
, vol. 
62
 (pg. 
569
-
78
)
14
Chen
K
Jiang
Q
Ma
X
, et al.  . 
Alcohol drinking and colorectal cancer: a population-based prospective cohort study in China
Eur J Epidemiol
 , 
2005
, vol. 
20
 (pg. 
149
-
54
)
15
Wang
J
Gao
YT
Wang
XL
Liu
EJ
Zhang
YL
Yuan
JM
Prospective male cohort study on alcohol consumption and mortality in Shanghai
Chin J Public Health
 , 
2005
, vol. 
21
 (pg. 
299
-
302
)
16
Xu
WH
Zhang
XL
Gao
YT
, et al.  . 
Joint effect of cigarette smoking and alcohol consumption on mortality
Prev Med
 , 
2007
, vol. 
45
 (pg. 
313
-
19
)
17
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
, vol. 
314
 (pg. 
18
-
23
)
18
Zhang
LF
Zhao
LC
Zhou
BF
Yang
J
Li
Y
Wu
YF
[Alcohol consumption and incidence of ischemic stroke in male Chinese]
Zhonghua Liu Xing Bing Xue Za Zhi
 , 
2004
, vol. 
25
 (pg. 
954
-
57
)
19
Chen
Z
Yang
G
Zhou
M
, et al.  . 
Body mass index and mortality from ischaemic heart disease in a lean population: 10 year prospective study of 220,000 adult men
Int J Epidemiol
 , 
2006
, vol. 
35
 (pg. 
141
-
50
)
20
Niu
SR
Yang
GH
Chen
ZM
, et al.  . 
Emerging tobacco hazards in China: 2. Early mortality results from a prospective study
BMJ
 , 
1998
, vol. 
317
 (pg. 
1423
-
24
)
21
Medicine CAoP
Annual Report on Chinese Diseases Surveillance Points
1991
Beijing
Hua Xia Press
22
Yang
GH
Murray
CJL
Zhang
Z
Exploring Adult Mortality in China: Levels Patterns and Causes
 , 
1991
Beijing
Hua Xia Press
23
Easton
DF
Peto
J
Babiker
AG
Floating absolute risk: an alternative to relative risk in survival and case-control analysis avoiding an arbitrary reference group
Stat Med
 , 
1991
, vol. 
10
 (pg. 
1025
-
35
)
24
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
, vol. 
153
 (pg. 
64
-
71
)
25
Breslow
RA
Graubard
BI
Prospective study of alcohol consumption in the United States: quantity, frequency, and cause-specific mortality
Alcohol Clin Exp Res
 , 
2008
, vol. 
32
 (pg. 
513
-
21
)
26
Gronbaek
M
Becker
U
Johansen
D
, et al.  . 
Type of alcohol consumed and mortality from all causes, coronary heart disease, and cancer
Ann Intern Med
 , 
2000
, vol. 
133
 (pg. 
411
-
19
)
27
Osier
MV
Pakstis
AJ
Soodyall
H
, et al.  . 
A global perspective on genetic variation at the adh genes reveals unusual patterns of linkage disequilibrium and diversity
Am J Hum Genet
 , 
2002
, vol. 
71
 (pg. 
84
-
99
)
28
Tsugane
S
Fahey
MT
Sasaki
S
Baba
S
Alcohol consumption and all-cause and cancer mortality among middle-aged Japanese men: seven-year follow-up of the JPHC Study Cohort I
Am J Epidemiol
 , 
1999
, vol. 
150
 (pg. 
1201
-
07
)
29
Lin
Y
Kikuchi
S
Tamakoshi
A
, et al.  . 
Alcohol consumption and mortality among middle-aged and elderly Japanese men and women
Ann Epidemiol
 , 
2005
, vol. 
15
 (pg. 
590
-
97
)
30
Chen
Z
Lee
L
Chen
J
, et al.  . 
Cohort profile: the Kadoorie Study of Chronic Disease in China (KSCDC)
Int J Epidemiol
 , 
2005
, vol. 
34
 (pg. 
1243
-
49
)
31
Doll
R
Peto
R
Hall
E
Wheatley
K
Gray
R
Mortality in relation to consumption of alcohol: 13 years' observations on male British doctors
BMJ
 , 
1994
, vol. 
309
 (pg. 
911
-
18
)
32
Ikehara
S
Iso
H
Toyoshima
H
, et al.  . 
Alcohol consumption and mortality from stroke and coronary heart disease among Japanese men and women: the Japan Collaborative Cohort Study
Stroke
 , 
2008
, vol. 
39
 (pg. 
2936
-
42
)
33
Schooling
CM
Sun
W
Ho
SY
, et al.  . 
Moderate alcohol use and mortality from ischaemic heart disease: a prospective study in older Chinese people
PLoS One
 , 
2008
, vol. 
3
 pg. 
e2370
 
34
 
IARC. Alcohol drinking. In: IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Vol. 44. Lyon: IARC Press; 1988
35
 
IARC. Alcohol consumption and ethyl carbamate. In: IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Vol. 96. Lyon: IARC Press; 2010
36
Tabak
C
Smit
HA
Rasanen
L
, et al.  . 
Alcohol consumption in relation to 20-year COPD mortality and pulmonary function in middle-aged men from three European countries
Epidemiology
 , 
2001
, vol. 
12
 (pg. 
239
-
45
)
37
Inoue
M
Tsugane
S
Impact of alcohol drinking on total cancer risk: data from a large-scale population-based cohort study in Japan
Br J Cancer
 , 
2005
, vol. 
92
 (pg. 
182
-
87
)
38
Thun
MJ
Hannan
LM
DeLancey
JO
Alcohol consumption not associated with lung cancer mortality in lifelong nonsmokers
Cancer Epidemiol Biomarkers Prev
 , 
2009
, vol. 
18
 (pg. 
2269
-
72
)
39
Fillmore
KM
Stockwell
T
Chikritzhs
T
Bostrom
A
Kerr
W
Moderate alcohol use and reduced mortality risk: systematic error in prospective studies and new hypotheses
Ann Epidemiol
 , 
2007
, vol. 
17
 (pg. 
S16
-
23
)
40
Klatsky
AL
Friedman
GD
Armstrong
MA
Kipp
H
Wine, liquor, beer, and mortality
Am J Epidemiol
 , 
2003
, vol. 
158
 (pg. 
585
-
95
)
41
Renaud
SC
Gueguen
R
Siest
G
Salamon
R
Wine, beer, and mortality in middle-aged men from eastern France
Arch Intern Med
 , 
1999
, vol. 
159
 (pg. 
1865
-
70
)
42
Rehm
J
Mathers
C
Popova
S
Thavorncharoensap
M
Teerawattananon
Y
Patra
J
Global burden of disease and injury and economic cost attributable to alcohol use and alcohol-use disorders
Lancet
 , 
2009
, vol. 
373
 (pg. 
2223
-
33
)