Alcohol consumption and risks of more than 200 diseases in Chinese men

Alcohol consumption accounts for ~3 million annual deaths worldwide, but uncertainty persists about its relationships with many diseases. We investigated the associations of alcohol consumption with 207 diseases in the 12-year China Kadoorie Biobank of >512,000 adults (41% men), including 168,050 genotyped for ALDH2-rs671 and ADH1B-rs1229984, with >1.1 million ICD-10 coded hospitalized events. At baseline, 33% of men drank alcohol regularly. Among men, alcohol intake was positively associated with 61 diseases, including 33 not defined by the World Health Organization as alcohol-related, such as cataract (n = 2,028; hazard ratio 1.21; 95% confidence interval 1.09–1.33, per 280 g per week) and gout (n = 402; 1.57, 1.33–1.86). Genotype-predicted mean alcohol intake was positively associated with established (n = 28,564; 1.14, 1.09–1.20) and new alcohol-associated (n = 16,138; 1.06, 1.01–1.12) diseases, and with specific diseases such as liver cirrhosis (n = 499; 2.30, 1.58–3.35), stroke (n = 12,176; 1.38, 1.27–1.49) and gout (n = 338; 2.33, 1.49–3.62), but not ischemic heart disease (n = 8,408; 1.04, 0.94–1.14). Among women, 2% drank alcohol resulting in low power to assess associations of self-reported alcohol intake with disease risks, but genetic findings in women suggested the excess male risks were not due to pleiotropic genotypic effects. Among Chinese men, alcohol consumption increased multiple disease risks, highlighting the need to strengthen preventive measures to reduce alcohol intake.

Excluded -defined as population outliers for the study area based on genomic data analysis (n=6097) a The genetic instrument (six genetic categories C1-C6) was derived from the randomly selected subset of genotyped men excluding exdrinkers (n=61,046 randomly selected, of whom n=5229 were ex-drinkers). The six genetic categories was then assigned to all genotyped men and women according to their genotype and area, regardless of individual drinking patterns. b To increase study power, additional recorded disease events (mainly but not limited to CVD/COPD) recorded during follow-up were selected from the subset of CVD/COPD cases and added to the genetic analyses of the corresponding disease outcome. CVD, cardiovascular disease; COPD, chronic obstructive pulmonary disease; PC, principal components. Calculations assigned an intake of 5 g/week to those who drink sometimes but less than weekly at the time of survey (regardless of their past drinking patterns). a Regression dilution ratio is estimated among baseline current drinkers using baseline and usual alcohol intakes. It was calculated using the assumption-free, non-parametric McMahon-Peto method i.e. the ratio of the range of the usual alcohol intake levels to the range of the baseline alcohol intake levels of baseline-defined current drinker groups. Adjusted means and 95% CIs were estimated using multiple linear regression, adjusted for age (in ten-year intervals) and ten study areas, in male and female current drinkers respectively. P values for trend were obtained from multiple linear regression models assessing the dose-response per g/week increase in alcohol intake. All P values are two-sided. CI, confidence interval. Means and percentages are adjusted for the age and study area structure of the CKB female current drinker population, using direct standardisation. a Chronic diseases included self-reported history of coronary heart disease, stroke, transient ischaemic attack, diabetes, tuberculosis, emphysema/chronic bronchitis, liver cirrhosis/chronic hepatitis, peptic ulcer, gallstone/gallbladder disease, kidney disease, rheumatoid arthritis, and cancer.  ICD-10, International Classification of Diseases, 10th Revision; WHO, World Health Organization; FDR, false discovery rate. a Included disease associations from Cox regression analyses either significant (p<0.05 for diseases classified as alcohol-related by the WHO, FDR-adjusted p<0.05 for other diseases; two-sided) from the comparison of ever-regular vs. occasional drinking or dose-response association analyses within current drinkers. b The 2 diseases were intracerebral haemorrhage (ICD-10 code: I61) (HR 1.20, 95% CI 1.03-1.40) and essential primary hypertension (I10) (HR 1.08, 95% CI 1.01-1.16), both of which were classified as alcohol-related by the WHO and showed nominally significant associations with ever-regular drinking (i.e. P value <0.05). c The 2 diseases were other acute lower respiratory infections (J20-J22) (HR 0.88, 95% CI 0.78-0.99) which was classified as alcohol-related by the WHO and showed nominal association with ever-regular drinking, and spondylosis (M47) (HR 0.87, 95% CI 0.80-0.95) which was significantly associated with ever-regular drinking after multiple testing correction (i.e. FDR-adjusted P value < 0.05).   (excluding I10, I11, I20, I21, I24, I25, I27, I42, I46, I48-I50, I60-I63, I65-I67, I69, I70,  I83) 33 CKB new alcohol-associated diseases b

Supplementary
Any coded or uncoded disease events All morbidity All morbidity minus CKB WHO alcohol-related diseases, CKB new alcohol-associated diseases, and diseases negatively associated with alcohol in men or women Non alcohol-related diseases CKB, China Kadoorie Biobank; ICD-10, International Classification of Diseases, 10 th Revision; WHO, World Health Organisation; FDR, false discovery rate. a Diseases which were considered to be alcohol-related by the WHO and showed significant adverse associations at two-sided p<0.05 with alcohol drinking in the CKB were combined as "CKB WHO alcohol-related" diseases. b Diseases showing significant adverse associations with alcohol drinking in men in CKB after multiple testing correction (FDR-adjusted p<0.05), but had not been classified as alcohol-related by WHO, were combined as "CKB new alcohol-associated diseases".

Supplementary Figure 2. Adjusted HRs for ICD-10 Chapters I to VIII associated with usual alcohol intake, in male current drinkers
Cox models were stratified by age-at-risk and study area, and adjusted for education and smoking. Each solid square or diamond represents HR per 280 g/week higher usual alcohol intake among male current drinkers, with the area inversely proportional to the variance of the log HR. The horizontal lines indicate 95% CIs. All P values are two-sided. * indicate p<0.05; ** indicate FDR-adjusted p<0.05. CI, confidence interval; HR hazard ratio; ICD-10, International Classification of Diseases, 10th Revision.

Supplementary Figure 3. Adjusted HRs for ICD-10 Chapters IX to XI associated with usual alcohol intake, in male current drinkers
Cox models were stratified by age-at-risk and study area, and adjusted for education and smoking. Each solid square or diamond represents HR per 280 g/week higher usual alcohol intake among male current drinkers, with the area inversely proportional to the variance of the log HR. The horizontal lines indicate 95% CIs. All P values are two-sided. * indicate p<0.05; ** indicate FDR-adjusted p<0.05. CI, confidence interval; HR hazard ratio; ICD-10, International Classification of Diseases, 10th Revision.

Supplementary Figure 4. Adjusted HRs for ICD-10 Chapters XII to XX associated with usual alcohol intake, in male current drinkers
Cox models were stratified by age-at-risk and study area, and adjusted for education and smoking. Each solid square or diamond represents HR per 280 g/week higher usual alcohol intake among male current drinkers, with the area inversely proportional to the variance of the log HR. The horizontal lines indicate 95% CIs. All P values are two-sided. * indicate p<0.05; ** indicate FDR-adjusted p<0.05. CI, confidence interval; HR hazard ratio; ICD-10, International Classification of Diseases, 10th Revision. Figure 5. Adjusted HRs per 280 g/week higher usual alcohol intake for specific alcohol-associated diseases in male current drinkers, with further adjustments or exclusion of participants with baseline medical conditions Cox models were stratified by age-at-risk and study area, and were adjusted for education and smoking in (A). (B) had the same model specification as in (A) plus further adjustments for income, physical activity, fruit intake, and body mass index. (C) had the same model specification as in (A) and excluded participants with poor self-rated health or prior chronic disease (i.e. self-reported history of coronary heart disease, stroke, transient ischaemic attack, diabetes, tuberculosis, emphysema/chronic bronchitis, liver cirrhosis/chronic hepatitis, peptic ulcer, gallstone/gallbladder disease, kidney disease, rheumatoid arthritis, and cancer) at baseline. Each solid square represents HR per 280 g/week higher usual alcohol intake among male current drinkers, with the area inversely proportional to the variance of the log HR. The horizontal lines indicate 95% CIs. † Included less common ICD-10 codes within the corresponding ICD-10 chapter which were not individually investigated in the present study. CI, confidence interval; HR hazard ratio; ICD-10, International Classification of Diseases, 10th Revision.

Supplementary Figure 6. Adjusted HRs for CKB WHO alcohol-related diseases, CKB new alcohol-associated diseases and all morbidity per 280 g/week higher usual alcohol intake, by population subgroups in male current drinkers
Cox models were stratified by age-at-risk and study area, and adjusted for education and smoking. Each solid square represents HR per 280 g/week higher usual alcohol intake among male current drinkers, with the area inversely proportional to the variance of the log HR. The horizontal lines indicate 95% CIs. All P values are two-sided. CI, confidence interval; HR hazard ratio; ICD-10, International Classification of Diseases, 10th Revision.

Supplementary Figure 7. Associations of aggregate disease categories with self-reported alcohol intake and with genotype-predicted mean male alcohol intake categories, in women
Conventional epidemiological analyses relate self-reported drinking patterns to risks of diseases (reference group = occasional drinkers), using Cox models stratified by age-at-risk and study area and adjusted for education and smoking. Within current drinkers, HRs were plotted against usual alcohol intake and were calculated per 100 g/week higher usual alcohol intake. Genetic epidemiological analyses relate genetic categories to risks of diseases (reference group = genotype group with lowest mean genotype-predicted mean male alcohol intake), using Cox models by age-at-risk and study area and adjusted for 11 genomic principal components. The HR per 280 g/week higher genotype-predicted mean male alcohol intake is calculated from the inverse-variance-weighted mean of the slopes of the fitted lines in each study area. Each box represents HR with the area inversely proportional to the variance of the group-specific log hazard within subplot. The vertical lines indicate group-specific 95% CIs. All P values are two-sided. CI, confidence interval; HR hazard ratio; CKB, China Kadoorie Biobank; ICD-10, International Classification of Diseases, 10th Revision; WHO, World Health Organisation. MET-h/d, metabolic equivalent of task per hour per day. Prevalences and means are adjusted for study area and (where appropriate) age (in 10-year intervals) structure of the CKB randomly selected genotyped subset, using direct standardisation separately by sex. The P value for trend is from an inverse-variance-weighted meta-analysis across ten areas, with within-area per G-allele effect adjusted (where appropriate) for age and genomic principal components using multiple linear regression models. All P values are two-sided. a Calculations assign an intake of 5 g/week to occasional drinkers, and exclude ex-drinkers. MET-h/d, metabolic equivalent of task per hour per day. Prevalences or means (except for those for alcohol drinking) are adjusted for area, genomic principal components and (where appropriate) age, using multiple linear regression models. The P value for trend is from an inverse-variance-weighted meta-analysis across ten areas, with within-area slopes adjusted (where appropriate) for age and genomic principal components using multiple linear regression models. All P values are two-sided. a Prevalences or means of alcohol consumption are unadjusted. b Calculations assign an intake of 5 g/week to occasional drinkers, and exclude ex-drinkers. Figure 8. Total expected hospitalisations overall and by major disease categories in ever-and never-regular drinkers from age-at-risk of 35 years among men

Supplementary
The solid and dashed lines indicate the estimated total hospitalisation of specific causes in a cohort of 100 participants in male ever-regular drinkers and male occasional drinkers, respectively. Grey bands show 95% confidence intervals. CKB, China Kadoorie Biobank; ICD-10, International Classification of Diseases, 10 th Revision. "All CKB alcohol-related diseases" include disease events from "CKB WHO alcohol-related diseases" or "CKB new alcohol-associated diseases".