Alcohol intake and cause-specific mortality: conventional and genetic evidence in a prospective---

Millwood IY, Im PK, Bennett D, Hariri P, Yang L, Du H, Kartsonaki C, Lin K, Yu C, Chen Y, Sun D, Zhang N, Avery D, Schmidt D, Pei P, Chen J, Clarke R, Lv J, Peto R, Walters RG, Li L, Chen Z; China Kadoorie Biobank Collaborative Group. Alcohol intake and cause-specific mortality: conventional and genetic evidence in a prospective cohort study of 512 000 adults in China. Lancet Public Health. 2023 Dec;8(12):e956-e967. doi: 10.1016/S2468-2667(23)00217-7. Epub 2023 Nov 21. PMID: 38000378.

Background Genetic variants that affect alcohol use in East Asian populations could help assess the causal effects of

alcohol consumption on cause-specific mortality. We aimed to investigate the associations between alcohol intake and

cause-specific mortality using conventional and genetic epidemiological methods among more than 512 000 adults in

China.

Methods The prospective China Kadoorie Biobank cohort study enrolled 512724 adults (210205 men and

302 519 women) aged 30–79 years, during 2004–08. Residents with no major disabilities from ten diverse urban and

rural areas of China were invited to participate, and alcohol use was self-reported. During 12 years of follow-up,

56 550 deaths were recorded through linkage to death registries, including 23457 deaths among 168050 participants

genotyped for ALDH2 -rs671 and ADH1B -rs1229984. Adjusted hazard ratios (HRs) for cause-specific mortality by self

reported and genotype-predicted alcohol intake were estimated using Cox regression.

Findings 33% of men drank alcohol most weeks. In conventional observational analyses, ex-drinkers, non-drinkers,

and heavy drinkers had higher risks of death from most major causes than moderate drinkers. Among current

drinkers, each 100 g/week higher alcohol intake was associated with higher mortality risks from cancers (HR 1·18

[95% CI 1·14−1·22]), cardiovascular disease (CVD; HR 1·19 [1·15−1·24]), liver diseases (HR 1·51 [1·27−1·78]), non

medical causes (HR 1·15 [1·08−1·23]), and all causes (HR 1·18 [1·15−1·20]). In men, ALDH2 -rs671 and ADH1B

rs1229984 genotypes predicted 60-fold differences in mean alcohol intake (4 g/week in the lowest group vs 255 g/week

in the highest). Genotype-predicted alcohol intake was uniformly and positively associated with risks of death from all

causes (n=12 939; HR 1·07 [95% CI 1·05−1·10]) and from pre-defined alcohol-related cancers (n=1274; 1·12

[1·04−1·21]), liver diseases (n=110; 1·31 [1·02−1·69]), and CVD (n=6109; 1·15 [1·10−1·19]), chiefly due to stroke

(n=3285; 1·18 [1·12–1·24]) rather than ischaemic heart disease (n=2363; 1·06 [0·99–1·14]). Results were largely

consistent using a polygenic score to predict alcohol intake, with higher intakes associated with higher risks of death

from alcohol-related cancers, CVD, and all causes. Approximately 2% of women were current drinkers, and although

power was low to assess observational associations of alcohol with mortality, the genetic evidence suggested that the

excess risks in men were due to alcohol, not pleiotropy.

Interpretation Higher alcohol intake increased the risks of death overall and from major diseases for men in China. There was no genetic evidence of protection from moderate drinking for all-cause and cause-specific mortality, including CVD.

Research in context

Evidence before this study

Moderate alcohol intake has been associated with lower risks of

mortality overall and from specific diseases—in particular,

ischaemic heart disease. However, these associations might be

largely non-causal, as conventional observational studies of

alcohol use are susceptible to bias from reverse causation and

residual confounding. Genetic evidence from Mendelian

randomisation studies, in particular using the ALDH2 -rs671 and

ADH1B -rs1229984 variants, which strongly affect alcohol

intake and are common in East Asian populations, can help

assess the causal relevance of alcohol intake for cause-specific

mortality.

We searched PubMed from database inception to Feb 25, 2023

using the following search terms (title or abstract) for articles

published in English: ([Alcohol AND Mendelian) or (ALDH2 or

ADH1B or rs671 or rs1229984 or aldehyde dehydrogenase or

alcohol dehydrogenase]) AND (mortality or death or fatal), and

reviewed bibliographies within the identified publications.

Two previous Mendelian randomisation studies of alcohol and

mortality in populations with European ancestry, and one in

Chinese men, reported that higher alcohol intake was

associated with higher risks of all-cause mortality. However,

these studies did not assess causal relevance across a wide range

of alcohol intakes and did not evaluate effects on cause-specific

mortality.

Added value of this study

The present prospective study used both conventional and

genetic approaches within the same population. The genetic

analyses minimised artefacts of confounding and reverse

causation and assessed potential causal relevance across a wide

range of alcohol intakes, from negligible to moderate and

heavy intakes. Among Chinese men, conventional

observational analyses showed characteristic J-shaped

associations of self-reported alcohol intake categories with

overall and cause-specific mortality, with highest risks among

ex-drinkers, non-drinkers, and heavy drinkers, and lowest risks

among moderate drinkers, consistent with findings from

similar studies in populations of high-income countries.

Genetic analyses, using two genetic variants that predicted a

60-fold difference in mean intake (from 4 g/week in the lowest

category to 255 g/week in the highest category) showed that

higher alcohol intake was associated with a uniform dose–

response increase in risks of death overall and from some

cancers, CVD, and liver diseases. There were no genetic

associations with respiratory or non-medical (mainly accidents

and injuries) causes of death. There was no genetic evidence

that moderate alcohol intake (ie, 10–20 g/day) had substantial

protective effects for cause-specific or overall mortality,

including for ischaemic heart disease deaths. In separate genetic

analyses using a polygenic score to predict alcohol intake, there

were similar associations with mortality overall, and from

alcohol-related cancers and cardiovascular disease (CVD), across

different alcohol intakes.

Alcohol intake was extremely low among women in the study,

and the genetic variants had little effect on mortality overall or

from specific causes, suggesting that the higher risks in men

were chiefly mediated by alcohol, rather than by any pleiotropic

effects of the genotypes studied.

Implications of all the available evidence

Genetic studies, particularly in East Asian ancestry populations,

have helped to reliably clarify the causal relevance of alcohol

intake with mortality by accounting for biases in conventional

observational approaches. The genetic evidence provides

strong support for causal harmful effects of alcohol use, with

risks of deaths from CVD, cancer, liver disease, and all-causes.

There is no genetic evidence of any net beneficial effects of

moderate drinking compared with not drinking for any causes

of death, including CVD. These genetic studies that assess

causal relevance have improved our understanding of the

adverse effects of alcohol use on mortality, particularly at lower

intakes. This knowledge can improve estimation of the

regional and global burden of alcohol use, and inform public

health policies to address the risks of moderate and heavy

drinking

Mendelian randomisation uses genetic variants as

instrumental variables to assess the causal relevance of

alcohol intake while minimising the biases inherent to

conventional observational studies

Figure 1: Conventional and genetic associations of alcohol intake with major

cause-specific and all-cause mortality in men

Conventional epidemiological analyses relate self-reported drinking patterns at

baseline to mortality from major causes (all major causes are shown except for

infectious diseases, for which the number of deaths was lower) and all-causes.

Current drinkers with the lowest mean alcohol intake are the reference group.

The black squares represent findings from the main model adjusted for age-at

risk, area, education, household income, smoking, physical activity, and fresh

fruit intake, with exclusion of participants with previous chronic disease. The

HRs for current drinkers are plotted against usual alcohol intake, and a weighted

linear regression through the plotted estimates gives the HR (95% CI) per

100 g/week (~1–2 drinks per day, assuming 1 drink contains 10 g alcohol).

The grey squares represent findings from sensitivity analyses that further

exclude the first 5 years of follow-up. Genetic epidemiological analyses relate

mean alcohol intake in six categories of genotype-predicted intake to mortality

from major causes. The lowest mean intake group is the reference, and analyses

are adjusted for age-at-risk, area, and genomic national principal components.

HRs are plotted against the mean alcohol intake in each category. The HR

(95% CI) per 100 g/week is the inverse-variance-weighted mean of a weighted

linear regression through the plotted estimates within each study area, adjusted

for age-at-risk and genomic regional principal components. The HR (95% CI)

across six genetic categories in women applied the mean male intakes for each

category, and the heterogeneity of effects was compared between men and

women, to assess pleiotropy. The HR is plotted on a log scale. Each box

represents HR with the area inversely proportional to the variance of the group

specific log hazard within each subplot. The vertical lines indicate group-specific

95% CIs. HR=hazard ratio.

Figure 2: Conventional and genetic associations of alcohol intake with

mortality from aggregated cancers and cardiovascular disease types in men

Conventional epidemiological analyses relate self-reported drinking patterns at

baseline to mortality from aggregated cancers and cardiovascular disease types.

Alcohol-related cancers include lip, oral cavity, pharynx, larynx, oesophagus,

liver, and colon-rectum, defined as related to alcohol by the International

Agency for Research on Cancer. Current drinkers with the lowest mean alcohol

intake are the reference group. The black squares represent findings from the

main model adjusted for age-at-risk, area, education, household income,

smoking, physical activity, and fresh fruit intake, with exclusion of participants

with previous chronic disease. The HRs for current drinkers are plotted against

usual alcohol intake, and a weighted linear regression through the plotted

estimates gives the HR (95% CI) per 100 g/week (~1–2 drinks per day, assuming

1 drink contains 10 g alcohol). The grey squares represent findings from

sensitivity analyses that further exclude the first 5 years of follow-up. Genetic

epidemiological analyses relate mean alcohol intake in six categories of

genotype-predicted intake to mortality. The lowest mean intake group is the

reference, and analyses are adjusted for age-at-risk, area, and genomic national

principal components. HRs are plotted against the mean alcohol intake in each

category. The HR (95% CI) per 100 g/week is the inverse-variance-weighted

mean of a weighted linear regression through the plotted estimates within each

study area, adjusted for age-at-risk and genomic regional principal components.

The HR (95% CI) across six genetic categories in women applied the mean male

intakes for each category, and the heterogeneity of effects was compared

between men and women, to assess pleiotropy. The HR is plotted on a log scale.

Each box represents HR with the area inversely proportional to the variance of

the group-specific log hazard within each subplot. The vertical lines indicate

group-specific 95% CIs. HR=hazard ratio.