Abstract

Aims

Alcohol drinking is associated with central obesity, hypertension, and hyperlipidemia, which further causes metabolic syndrome (MetS). However, prior epidemiological studies on such associations lack experimental evidence for a causal relationship. This study aims to explore the causal relationship between drinking behavior and MetS in Taiwan population by using Mendelian randomization (MR) analysis.

Methods

A cross-sectional study was conducted using the Taiwan Biobank database, which comprised 50 640 Han Chinese who were 30–70 years old without cancer from 2008 to 2020. In MR analysis, we constructed weighted and unweighted genetic risk scores by calculating SNP alleles significantly associated with alcohol drinking. We calculated odds ratios and 95% confidence interval (CI) by using a two-stage regression model.

Results

A total of 50 640 participants were included with a mean age of 49.5 years (SD: 1.67 years), 36.6% were men. The adjusted odds ratio (aOR) of MetS per 5% increase in the likelihood of genetic predisposition to drink based on weighted genetic risk score with adjustment was 1.11 (95% CI: 1.10, 1.12, P < .001). Analysis was also conducted by grouping the likelihood of genetic predisposition to drink based on quartiles with multivariate adjustment. Using Q1 as the reference group, the aORs of MetS for Q2, Q3, and Q4 were 1.19 (1.12, 1.27, p < .001), 1.31 (1.23, 1.40, p < .001), and 1.87 (1.75, 2.00, p < .001), respectively, for the weighted genetic risk score.

Conclusions

This study shows a modest relationship between drinking behavior and MetS by using MR analysis.

Introduction

Metabolic syndrome (MetS) is a collective term for abnormal blood pressure, blood lipids, blood sugar, and waist circumference and is considered a multifactorial disorder that causes inflammatory responses, endocrine disorders, and cardiovascular diseases (CVD; Esser et al. 2014; Ghowsi et al. 2021). In recent years, the global prevalence and incidence of MetS have rapidly increased, and it has become an important public health issue that cannot be ignored. The prevalence rates of MetS are 13.6%–25.5% in the Taiwanese population, 11.9%–37.1% in the Asian population, 11.6%–26.3% in the European population, and 20%–25% in the global population (Alberti et al. 2006; Ranasinghe et al. 2017). MetS is significantly associated with lifestyle factors, including alcohol intake and diet. The prevalence of MetS in heavy drinkers is 13.9% (Hernández-Rubio et al. 2022). People with excessive drinking habits are prone to developing MetS (Slanovic-Kuzmanovic et al. 2013).

Alcohol drinking is an important risk factor for hypertension (Biddinger et al. 2022), central obesity (Park et al. 2017), and hyperlipidemia, which is determined by elevated levels of serum triglycerides (TG), total cholesterol (TC), and low-density lipoprotein (LDL) cholesterol or a reduced high-density lipoprotein (HDL) cholesterol (Ye et al. 2023). Alcohol may interact with other environmental factors (such as smoking) (Delgado-Lobete et al. 2020). Excessive alcohol intake leads to components of MetS and leads to the occurrence of this syndrome (Paulson et al. 2010; Slanovic-Kuzmanovic et al. 2013; Chang et al. 2016). Many traditional observational studies have explored the association between alcohol drinking and MetS and its components, but the association between alcohol drinking and the prevalence of MetS remains controversial (Yamaoka and Tango 2012; Barbaresko et al. 2018). A large population-based study in the United States reported that alcohol drinking exceeding the U.S. dietary guideline recommendations and binge drinking were associated with a high prevalence of MetS (Fan et al. 2008). By contrast, no association was observed between alcohol drinking and MetS in a random sample of Portuguese (Santos et al. 2012) and Korean (Lee et al. 2005).

At present, experimental evidence regarding the causal relationship between alcohol drinking and the onset of MetS is lacking (Kothari et al. 2002; Nelson et al. 1996; Pickering 2007). Therefore, providing experimental evidence is necessary for MetS prevention. The widespread application of Mendelian randomization (MR) analysis in exploring the causal association between lifestyle factors and disease outcomes provides experimental evidence criteria for causal inference (Singh et al. 2000). MR analysis utilizes genetic variations at conception as an instrument to make causal inferences (Lawlor et al. 2008) and assess the causality of an observed association between a modifiable risk factor and clinically relevant outcomes to rule out confounding or reverse causality in observational studies (Saccone et al. 2007). A comprehensive search of the literature on MR was conducted to assess the associations between drinking behavior and metabolic factors, such as diabetes, hypertension, and CVD. Nine MR studies investigated the association between alcohol drinking and coronary heart disease (Lawlor et al. 2013), CVD (Holmes et al. 2014; Cho et al. 2015; Biddinger et al. 2022), fasting blood glucose (Jee et al. 2016), heart failure (Thorgeirsson et al. 2008), stroke (Freathy et al. 2009; Christensen et al. 2018), diabetes (Amos et al. 2008; Peng et al. 2019), vascular disease etiology (Millwood et al. 2019), and heart failure (van Oort et al. 2020). Among the studies exploring the causal relationship between alcohol drinking and cardiovascular risk factors, prior MR studies assessed the effects of variant-related alcohol drinking on blood pressure (Lawlor et al. 2013; Cho et al. 2015; Millwood et al. 2019), waist circumference (Cho et al. 2015), fasting plasma glucose (FPG) (Lawlor et al. 2013; Cho et al. 2015; Jee et al. 2016), TG (Lawlor et al. 2013; Cho et al. 2015), and HDL cholesterol (Lawlor et al. 2013; Cho et al. 2015). The causal relationship between alcohol drinking and MetS by using MR analysis has not been investigated in Chinese or Asians. The present study aimed to assess the relationship between alcohol drinking and MetS by using MR in subjects aged 30–70 years from the Taiwan Biobank Dataset.

Materials and methods

Study subjects and data source

The study subjects were obtained from the Taiwan Biobank database based on Taiwan’s community population, comprising Han Chinese who were 30–70 years old without cancer history and enrolled from 2008 to 2020. A total of 116 066 individuals with 9 809 486 variants were identified from the databases of genome-wide TWB1.0 and TWB 2.0 (Fig. 1). Study subjects were excluded if GWAS data did not pass the quality control (QC) criteria, leaving 90 381 study subjects with 2 581 477 variants. Additional 25 685 persons with 7 228 009 SNPs were excluded because of extreme heterozygosity rate with a mean heterozygosity rate > 2 standard deviation (N = 865), duplicated chromosomes 1–22 with identity-by-decent cut point of .1875 (N = 24 820), SNPs with a high missing genotype rate or low frequency of <1%, or deviation from Hardy–Weinberg equilibrium (P < .05; variants = 7 228 009). A total of 121 alcohol drinking-related variants identified in the literature were extracted from the dataset. Furthermore, 39 489 persons with 21 variants were excluded because of missing allele (variants = 7), minor allele frequency <5% (variants = 14), and missing data on genotype information (n = 39 489), resulting in 50 892 persons with 100 variants (Supplementary Table 1). Finally, 252 persons were excluded because of missing data on alcohol drinking variables (N = 31), MetS components (N = 26), covariates (N = 143), basic sociodemographic factors (N = 2), and non-Chinese ancestry (N = 50), resulting in 50 640 eligible subjects. The present study obtained approval from the Ethics and Governance Council of Taiwan Biobank (approval number: TWBR10811–06) and the Ethical Review Board of China Medical University Hospital (CMUH109-REC3–187). The study was conducted in accordance with the relevant regulations and guidelines.

Research flowchart for study subject selection in the present study
Figure 1

Research flowchart for study subject selection in the present study

Measurements

Measurements consisting of data collected by standardized structured questionnaire and clinical laboratory were classified into sociodemographic and lifestyle factors, disease history, biomarkers, and genetic data. Sociodemographic factors included gender (men/women), age groups (30–39, 40–49, 50–59, and ≥60 years), educational attainment, marriage status (unmarried, married, divorce, separation, or widowed), and living alone status (yes/no). The study subjects were classified into those with and without alcohol use based on alcohol drinking behavior status. Those with alcohol use had been drinking at least 150 cc of alcohol per week for six consecutive months. Those who have abstained from alcohol for more than 6 months and who have been drinking for ≥2 years before quitting drinking were also defined as people with alcohol use. People with no alcohol use were defined as those who did not drink or drink less than 150 cc of alcohol per week for six consecutive months. Study subjects were classified into smokers and non-smoker based on smoking behavior status. Smokers were defined as having smoked more than one cigarette in his/her lifetime. Those who never smoked were defined as non-smokers. Study subjects were allocated into classes of betel nut chewing and non-betel nut chewing based on individuals’ self-report about betel nut chewing behavior. Betel nut chewers were those who had eaten at least three times in his/her lifetime. Non-betel nut chewers were those who had never chewed betel nut or had eaten once or twice in his/her lifetime. A list of disease history was obtained through questionnaires, including hypertension (yes/no), hyperlipidemia (yes/no), diabetes (yes/no), arthritis (yes/no), gout (yes/no), stroke (yes/no), liver or gall stones (yes/no), kidney stones (yes/no), valvular heart disease (yes/no), coronary artery disease (yes/no), arrhythmia(yes/no), cardiomyopathy (yes/no), congenital heart disease (yes/no), and other heart disease (yes/no).

Biomarkers were assessed by laboratory tests for TC, HDL cholesterol, LDL cholesterol, TG, FPG, and serum creatinine, which were performed in the clinical laboratory of Chang Gung Memorial Hospital, Linkou. SNPs were genotyped through DNA samples of Taiwan Biobank using TWB arrays and run on the Axiom Whole Genome Array Plate System (Affymetrix, Santa Clara, CA, USA). The number of SNPs of genome-wide TWB1.0 and TWB 2.0 was 653 000 and 750 000, respectively. Missing genotypes were imputed by using IMPUTE2 with reference to the 1000 Genomes Project. The selection of genetic variants was based on literature studies using candidate gene and GWAS approaches. We selected ADH1B, ADH1C, and ALDH2 genes from literature by using a candidate approach (Lawlor et al. 2013; Holmes et al. 2014; Cho et al. 2015; Christensen et al. 2018; Millwood et al. 2019; Peng et al. 2019) and SNPs of other genes from prior studies using a GWAS approach (Zhang et al. 2007; Larsson et al. 2020; van Oort et al. 2020). The number of SNPs found in each gene included ADH1B (15 SNPs), ADH1C (12 SNPs), ALDH2 (21 SNPs), WASF3 (1 SNP), OTX2 (1 SNP), AGBL1 (1 SNP), LRRC28 (1 SNP), near GPR139 (1 SNP), SEZ6L2 (1 SNP), SRR (1 SNP), TNFSF13 and TNFSF12-TNFSF13 (1 SNP), RPTOR (1 SNP), TCF4 (1 SNP), near ONECUT2 (1 SNP), ACSS1 (1 SNP), near ADH1C (1 SNP), C4orf17 (2 SNPs), INPP4B (1 SNP), near LINC02273 (1 SNP), near NAMPTP2 (1 SNP), SGCD (1 SNP), TENM2 (1 SNP), AUTS2 (1 SNP), near MLXIPL (1 SNP), ARPC1B (1 SNP), ORC5 (1 SNP), DPP6 (1 SNP), near LINC02153 (1 SNP), near LOC105375746 (1 SNP), near LINC01505 (1 SNP), LINC02661 (1 SNP), near LINC02641 (1 SNP), TRIM66 (1 SNP), BDNF-AS and LINC00678 (1 SNP), SPI1 (1 SNP), near LOC390251 (1 SNP), near LOC107984393 (1 SNP), near SORL1 (1 SNP), SLC4A8 (1 SNP), HNRNPA1 and CBX5 (1 SNP), ACSS3 (1 SNP), near POLR3H (1 SNP), PDE4B (1 SNP), PHC2 (1 SNP), PTGER3 (1 SNP), ZBTB37 (1 SNP), NUCKS1 (1 SNP), GCKR (1 SNP), GPN1 (1 SNP), WDPCP (1 SNP), CADM2 (1 SNP), near NSUN3 (1 SNP), RASA2 (1 SNP), RSRC1 (1 SNP), KLB (1 SNP), BEND4 (1 SNP), near BTF3P13 (1 SNP), ADH5 (1 SNP), CPNE4 and LOC 105374113 (1 SNP), ZBTB38 (1 SNP), near KLB (1 SNP), ARID4A (1 SNP), KIF26A (1 SNP), LINC01572 (1 SNP), near LINC017985452 (1 SNP), LINC01833 (1 SNP), ARHGAP15 (1 SNP), near LINC01612 (1 SNP), near RN7SKP135 (1 SNP), near LOC107984390 (1 SNP), LRRC28 (3 SNP), near HAS3 (1 SNP), and near MAMSTR (1 SNP) for alcohol drinking. The details of SNP ID numbers for each gene are listed in Table 2.

Outcome ascertainment

This study used a definition of MetS for Asians that was proposed by a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity (Alberti et al. 2009). MetS was defined as the presence of three or more of the following factors: waistline ≥90 cm (men) or ≥80 cm (women), blood pressure (systolic/diastolic) >130/85 mmHg or hypertension, FPG ≥100 mg/dL or diabetes, TG ≥ 150 mg/dL or hyperlipidemia, and HDL cholesterol <40 mg/dL (men) or <50 mg/dL (women). The status of hypertension, diabetes, and hyperlipidemia was based on the checklist of the disease history of the subjects.

Statistical analysis

First, QC was performed by removing individuals with high missing genotype rate; removing individuals with high heterozygosity rate; removing duplicated or related individuals; and removing SNPs with high missing genotype rate, low frequency, and deviation from Hardy–Weinberg equilibrium.

Two-sample t-tests and Chi-square tests were used to analyze the distribution of sociodemographic factors, lifestyle behaviors, biomarkers, and comorbidities among persons with and without alcohol drinking behavior and MetS. The association among sociodemographic factors, lifestyle behavior, biomarkers, and comorbidities with SNPs in genotypic forms was assessed using ANOVA and Chi-square tests.

We then used unconditional logistic regression analyses to determine the relationship between drinking behavior and MetS in traditional epidemiologic study. The analyses were adjusted for age, gender, and other confounding variables. The assumption of MR analysis was assessed by the following steps. First, we analyzed the relationship between alcohol drinking and alcohol drinking-related SNPs, that is, the SNP-level MR assumption 1. As for the SNP-level MR assumption 3, we analyzed the relationship between MetS and alcohol drinking-related genotypes to verify whether the selected SNPs can be used as instrumental variables for MR analysis. Only SNPs satisfying MR assumptions 1 and 3 were retained for deriving weighted and unweighted genetic risk scores. The association between selected SNPs and alcohol drinking or MetS was quantified by logistic regression models, with each SNP coded as 0, 1, or 2 based on the number of minor alleles, that is, in an additive model form.

Before estimating the genetic risk score, we analyzed the LD between SNPs that satisfy the MR assumptions 1 and 3. If two SNPs have r2 > .8, then only one was selected on the basis of which was the most common alcohol drinking-related gene in literature. Only SNPs with low LD were retained for deriving genetic risk score. The unweighted genetic risk score was derived by summing the number of minor alleles for retained SNPs. The weighted genetic risk score was constructed by multiplying the estimated coefficients of the logistic regression model of each genotype by the number of minor alleles for each retained SNP and the products of multiplication were then summed across all retained SNPs. Weighted and unweighted genetic risk scores were further divided into quartiles for data analyses in case the assumption of linear in logit was not satisfied. In addition, weighted and unweighted genetic risk scores were analyzed as continuous variables for testing the existence of dose–response relationship.

In verifying the genetic risk score-level MR assumption 1, logistic regression models were used to explore the association between the genetic risk score and alcohol drinking. In addition, multinomial logistic regression models were used to explore the association between the genetic risk score of alcohol drinking and the covariates to verify the genetic risk score-level MR assumption 2 and whether the selected covariates can be used as confounders for MR analysis. As for the genetic risk score-level MR assumption 3, logistic regression models were used to explore the association between genetic risk scores and MetS.

For MR analysis, the causal effect of alcohol drinking on MetS was quantified by instrumental variable analysis by using two-stage regression with multivariate adjustment. The first stage included logistic regression with alcohol drinking as the dependent variable and weighted and unweighted genetic risk scores as the independent variables to see whether genetic risk scores can predict the likelihood of alcohol drinking. The predicted alcohol drinking likelihoods were derived from the logistic regression, that is, the genetic variation-alcohol drinking. The second stage included the predicted likelihood of alcohol drinking estimated in the first stage as the independent variable, and the MetS as the dependent variable in logistic regression. We used unweighted and weighted genetic risk scores as instrumental variables to examine the robustness of these associations. In addition, the analyses were performed by adjusting for covariates. The covariates considered in this stage include residuals estimated from the first stage, covariates of demographic factors and lifestyle behaviors that did not satisfy MR assumption 2, along with top 10 principal components from PCA of all SNPs in TWB1.0 and TWB 2.0 (Supplementary Fig. 1). When unconditional or multinomial logistic regression models were used, the OR and 95% CIs were reported.

The main statistical analyses were performed using SAS Statistical Software Version 9.4 (SAS Institute, Cary, NC, USA), and a two-sided P-value less than .05 was defined as statistical significance. Pairwise LD between SNPs was quantified by using the correlation coefficient r2 in Haploview (v4.2). PLINK (v1.9) was used to assess each SNP in the Hardy–Weinberg equilibrium.

Results

This study included 50 640 participants, with a mean age of 49.5 years (standard deviation: 1.67 years), of whom 36.6% were men. Table 1 shows the comparison of sociodemographic factors, lifestyle behavior, clinical and biochemical markers, and comorbidities based on status of alcohol drinking and MetS. Table 2 shows the odds ratio (OR) of MetS for alcohol drinking status estimated in the observational study. The proportion of MetS was statistically higher in the alcohol drinking group than those in the non-alcohol drinking group (p < .05), and the crude OR was 1.69 (95% confidence interval (CI): 1.57, 1.81). After multivariate adjustment, MetS was statistically associated with alcohol drinking (1.08 [1.00, 1.17]).

Table 1

The comparisons of sociodemographic factors, lifestyle behaviors, clinical and biochemical markers, and comorbidities according to status of alcohol drinking and metabolic syndrome

Alcohol drinkingMetabolic syndrome
AllNoYesP valueNoYesP value
Variables(N = 50 640)(N = 46 468)(N = 4172)(N = 38 923)(N = 11 717)
Sociodemographic factors, n (%)
Gender<.001<.001
 Men18 535 (36.60)15 231 (32.81)3304 (79.21)13 284 (34.11)5251 (44.81)
 Women32 105 (63.40)31 237 (67.21)868 (2.81)25 639 (65.91)6466 (55.21)
Age, years, mean ± SD49.50 ± 1.6749.41 ± 1.7250.43 ± 1.12<.00148.29 ± 1.6153.50 ± 9.89<.001
Age group<.001<.001
 30–3911 277 (22.27)10 573 (22.75)704 (16.87)9951 (25.57)1326 (11.32)
 40–4913 090 (25.85)11 898 (25.60)1192 (28.57)10 658 (27.38)2432 (20.76)
 50–5915 608 (3.82)14 229 (30.62)1379 (33.05)11 439 (29.39)4169 (35.58)
 ≥6010 665 (21.06)9768 (21.02)897 (21.50)6875 (17.66)3790 (32.35)
Education attainment<.001<.001
 No formal education, illiterate80 (.16)76 (.21)4 (.11)55 (.11)25 (.21)
 Self-study, literacy27 (.05)25 (.11)2 (.01)19 (.01)8 (.11)
 Elementary school2515 (4.97)2327 (5.01)188 (4.51)1486 (3.81)1029 (8.81)
 National (junior) middle school3793 (7.49)3354 (7.21)439 (10.51)2550 (6.61)1243 (10.61)
 High school (vocational)14 900 (29.42)13 419 (28.91)1481 (35.51)11 079 (28.51)3821 (32.61)
 University (professional)23 957 (47.31)22 247 (47.91)1710 (41.01)19 293 (49.61)4664 (39.81)
 Institute and above5368 (10.60)5020 (1.81)348 (8.31)4441 (11.41)927 (7.91)
Marriage status<.001<.001
 Unmarried6598 (13.03)6212 (13.41)386 (9.31)5550 (14.31)1048 (8.91)
 Married37 605 (74.26)34 415 (74.11)3190 (76.51)28 724 (73.81)8881 (75.81)
 Divorce or separation4271 (8.43)3777 (8.11)494 (11.81)3209 (8.21)1062 (9.11)
 Widowed2166 (4.28)2064 (4.41)102 (2.41)1440 (3.71)726 (6.21)
Living alone0.7450.550
 Yes4121 (8.14)3776 (8.13)345 (8.31)3152 (8.11)969 (8.31)
Residential place<.001<.001
 Taipei10 260 (20.26)9534 (20.51)726 (17.41)8114 (2.81)2146 (18.31)
 North district8504 (16.79)7752 (16.71)752 (18.01)6435 (16.51)2069 (17.71)
 Central district9273 (18.31)8422 (18.11)851 (20.41)7123 (18.31)2150 (18.31)
 Southern district9439 (18.64)8627 (18.61)812 (19.51)7227 (18.61)2212 (18.91)
 Kaohsiung Pingtung10 588 (20.91)9806 (21.11)782 (18.71)8133 (20.91)2455 (21.01)
 East district2576 (5.09)2327 (5.01)249 (6.01)1891 (4.91)685 (5.81)
Lifestyle behaviors, n (%)
Duration of alcohol drinking (month), mean ± SD17.08 ± 7.430.02 ± .51207.11 ± 144.43<.00113.57 ± 61.2928.76 ± 93.72<.001
Smoking experience<.001<.001
 Yes13 898 (27.44)10 808 (23.31)3090 (74.11)9883 (25.41)4015 (34.31)
Betel nut chewing experience<.001<.001
 Never eaten, or only eaten once or twice47 464 (93.73)44 631 (96.01)2833 (67.91)37 031 (95.11)10 433 (89.01)
 Eaten many times3176 (6.27)1837 (4.01)1339 (32.11)1892 (4.91)1284 (11.01)
Leisure-time physical activity<.0010.006
 Yes20 184 (39.86)18 382 (39.61)1802 (43.21)15 385 (39.51)4799 (41.01)
Clinical and biochemical markers, mean ± SD
Waistline, cm83.16 ± 1.1382.77 ± 1.0787.54 ± 9.85<.00180.60 ± 8.9991.67 ± 9.00<.001
SBP, mmHg119.21 ± 17.88118.65 ± 17.78125.43 ± 17.80<.001115.48 ± 16.29131.61 ± 17.33<.001
DBP, mmHg73.36 ± 11.0372.92 ± 1.9078.25 ± 11.39<.00171.33 ± 1.2380.09 ± 1.95<.001
FPG, mg/dL95.70 ± 2.2895.29 ± 19.69100.23 ± 25.49<.00191.83 ± 12.85108.57 ± 31.84<.001
HbA1c, %5.75 ± .795.75 ± .785.83 ± .93<.0015.60 ± .546.27 ± 1.18<.001
TG, mg/dL115.68 ± 95.87112.69 ± 87.33148.99 ± 159.39<.00193.02 ± 55.88190.95 ± 148.25<.001
TC, mg/dL195.61 ± 35.77195.73 ± 35.63194.37 ± 37.280.024194.51 ± 34.62199.27 ± 39.14<.001
HDL-C, mg/dL54.42 ± 13.3854.65 ± 13.3251.86 ± 13.75<.00157.42 ± 12.8744.47 ± 9.74<.001
LDL, mg/dL121.13 ± 31.73121.38 ± 31.66118.32 ± 32.40<.001120.26 ± 31.01124.02 ± 33.86<.001
Albumin, g/dL4.52 ± .234.52 ± .234.55 ± .24<.0014.52 ± .234.54 ± .23<.001
SCr, mg/dL0.72 ± .310.71 ± .290.85 ± .42<.0010.71 ± .260.77 ± .42<.001
eGFR, mL/min/1.73 m2102.59 ± 14.79102.95 ± 14.6398.57 ± 15.87<.001104.14 ± 14.0697.43 ± 15.91<.001
Diseases of Metabolic components, n (%)
Hypertension5977 (11.80)5150 (11.11)827 (19.81)2259 (5.81)3718 (31.71)<.001
Hyperlipidemia3593 (7.10)3132 (6.71)461 (11.01)1248 (3.21)2345 (20.01)<.001
Diabetes2444 (4.83)2131 (4.61)313 (7.51)687 (1.81)1757 (15.01)<.001
Comorbidities, n (%)
Arthritis2367 (4.67)2194 (4.71)173 (4.11)1537 (3.91)830 (7.11)<.001
Gout1940 (3.83)1524 (3.31)416 (10.01)992 (2.51)948 (8.11)<.001
Stroke300 (.59)238 (.51)62 (1.51)144 (.41)156 (1.31)<.001
Liver or gall stones2253 (4.45)2073 (4.51)180 (4.31)1519 (3.91)734 (6.31)<.001
Kidney stones3188 (6.30)2767 (6.01)421 (10.11)2070 (5.31)1118 (9.51)<.001
Valvular heart disease2106 (4.16)1983 (4.31)123 (2.91)1718 (4.41)388 (3.31)<.001
Coronary artery disease574 (1.13)467 (1.01)107 (2.61)271 (.71)303 (2.61)<.001
Arrhythmia2276 (4.49)2057 (4.41)219 (5.21)1564 (4.01)712 (6.11)<.001
Cardiomyopathy367 (.72)305 (.71)62 (1.51)165 (.41)202 (1.71)<.001
Congenital heart disease109 (.22)98 (.21)11 (.31)83 (.21)26 (.21)0.821
Other heart disease74 (.15)72 (.21)2 (.01)49 (.11)25 (.21)0.038
Alcohol drinkingMetabolic syndrome
AllNoYesP valueNoYesP value
Variables(N = 50 640)(N = 46 468)(N = 4172)(N = 38 923)(N = 11 717)
Sociodemographic factors, n (%)
Gender<.001<.001
 Men18 535 (36.60)15 231 (32.81)3304 (79.21)13 284 (34.11)5251 (44.81)
 Women32 105 (63.40)31 237 (67.21)868 (2.81)25 639 (65.91)6466 (55.21)
Age, years, mean ± SD49.50 ± 1.6749.41 ± 1.7250.43 ± 1.12<.00148.29 ± 1.6153.50 ± 9.89<.001
Age group<.001<.001
 30–3911 277 (22.27)10 573 (22.75)704 (16.87)9951 (25.57)1326 (11.32)
 40–4913 090 (25.85)11 898 (25.60)1192 (28.57)10 658 (27.38)2432 (20.76)
 50–5915 608 (3.82)14 229 (30.62)1379 (33.05)11 439 (29.39)4169 (35.58)
 ≥6010 665 (21.06)9768 (21.02)897 (21.50)6875 (17.66)3790 (32.35)
Education attainment<.001<.001
 No formal education, illiterate80 (.16)76 (.21)4 (.11)55 (.11)25 (.21)
 Self-study, literacy27 (.05)25 (.11)2 (.01)19 (.01)8 (.11)
 Elementary school2515 (4.97)2327 (5.01)188 (4.51)1486 (3.81)1029 (8.81)
 National (junior) middle school3793 (7.49)3354 (7.21)439 (10.51)2550 (6.61)1243 (10.61)
 High school (vocational)14 900 (29.42)13 419 (28.91)1481 (35.51)11 079 (28.51)3821 (32.61)
 University (professional)23 957 (47.31)22 247 (47.91)1710 (41.01)19 293 (49.61)4664 (39.81)
 Institute and above5368 (10.60)5020 (1.81)348 (8.31)4441 (11.41)927 (7.91)
Marriage status<.001<.001
 Unmarried6598 (13.03)6212 (13.41)386 (9.31)5550 (14.31)1048 (8.91)
 Married37 605 (74.26)34 415 (74.11)3190 (76.51)28 724 (73.81)8881 (75.81)
 Divorce or separation4271 (8.43)3777 (8.11)494 (11.81)3209 (8.21)1062 (9.11)
 Widowed2166 (4.28)2064 (4.41)102 (2.41)1440 (3.71)726 (6.21)
Living alone0.7450.550
 Yes4121 (8.14)3776 (8.13)345 (8.31)3152 (8.11)969 (8.31)
Residential place<.001<.001
 Taipei10 260 (20.26)9534 (20.51)726 (17.41)8114 (2.81)2146 (18.31)
 North district8504 (16.79)7752 (16.71)752 (18.01)6435 (16.51)2069 (17.71)
 Central district9273 (18.31)8422 (18.11)851 (20.41)7123 (18.31)2150 (18.31)
 Southern district9439 (18.64)8627 (18.61)812 (19.51)7227 (18.61)2212 (18.91)
 Kaohsiung Pingtung10 588 (20.91)9806 (21.11)782 (18.71)8133 (20.91)2455 (21.01)
 East district2576 (5.09)2327 (5.01)249 (6.01)1891 (4.91)685 (5.81)
Lifestyle behaviors, n (%)
Duration of alcohol drinking (month), mean ± SD17.08 ± 7.430.02 ± .51207.11 ± 144.43<.00113.57 ± 61.2928.76 ± 93.72<.001
Smoking experience<.001<.001
 Yes13 898 (27.44)10 808 (23.31)3090 (74.11)9883 (25.41)4015 (34.31)
Betel nut chewing experience<.001<.001
 Never eaten, or only eaten once or twice47 464 (93.73)44 631 (96.01)2833 (67.91)37 031 (95.11)10 433 (89.01)
 Eaten many times3176 (6.27)1837 (4.01)1339 (32.11)1892 (4.91)1284 (11.01)
Leisure-time physical activity<.0010.006
 Yes20 184 (39.86)18 382 (39.61)1802 (43.21)15 385 (39.51)4799 (41.01)
Clinical and biochemical markers, mean ± SD
Waistline, cm83.16 ± 1.1382.77 ± 1.0787.54 ± 9.85<.00180.60 ± 8.9991.67 ± 9.00<.001
SBP, mmHg119.21 ± 17.88118.65 ± 17.78125.43 ± 17.80<.001115.48 ± 16.29131.61 ± 17.33<.001
DBP, mmHg73.36 ± 11.0372.92 ± 1.9078.25 ± 11.39<.00171.33 ± 1.2380.09 ± 1.95<.001
FPG, mg/dL95.70 ± 2.2895.29 ± 19.69100.23 ± 25.49<.00191.83 ± 12.85108.57 ± 31.84<.001
HbA1c, %5.75 ± .795.75 ± .785.83 ± .93<.0015.60 ± .546.27 ± 1.18<.001
TG, mg/dL115.68 ± 95.87112.69 ± 87.33148.99 ± 159.39<.00193.02 ± 55.88190.95 ± 148.25<.001
TC, mg/dL195.61 ± 35.77195.73 ± 35.63194.37 ± 37.280.024194.51 ± 34.62199.27 ± 39.14<.001
HDL-C, mg/dL54.42 ± 13.3854.65 ± 13.3251.86 ± 13.75<.00157.42 ± 12.8744.47 ± 9.74<.001
LDL, mg/dL121.13 ± 31.73121.38 ± 31.66118.32 ± 32.40<.001120.26 ± 31.01124.02 ± 33.86<.001
Albumin, g/dL4.52 ± .234.52 ± .234.55 ± .24<.0014.52 ± .234.54 ± .23<.001
SCr, mg/dL0.72 ± .310.71 ± .290.85 ± .42<.0010.71 ± .260.77 ± .42<.001
eGFR, mL/min/1.73 m2102.59 ± 14.79102.95 ± 14.6398.57 ± 15.87<.001104.14 ± 14.0697.43 ± 15.91<.001
Diseases of Metabolic components, n (%)
Hypertension5977 (11.80)5150 (11.11)827 (19.81)2259 (5.81)3718 (31.71)<.001
Hyperlipidemia3593 (7.10)3132 (6.71)461 (11.01)1248 (3.21)2345 (20.01)<.001
Diabetes2444 (4.83)2131 (4.61)313 (7.51)687 (1.81)1757 (15.01)<.001
Comorbidities, n (%)
Arthritis2367 (4.67)2194 (4.71)173 (4.11)1537 (3.91)830 (7.11)<.001
Gout1940 (3.83)1524 (3.31)416 (10.01)992 (2.51)948 (8.11)<.001
Stroke300 (.59)238 (.51)62 (1.51)144 (.41)156 (1.31)<.001
Liver or gall stones2253 (4.45)2073 (4.51)180 (4.31)1519 (3.91)734 (6.31)<.001
Kidney stones3188 (6.30)2767 (6.01)421 (10.11)2070 (5.31)1118 (9.51)<.001
Valvular heart disease2106 (4.16)1983 (4.31)123 (2.91)1718 (4.41)388 (3.31)<.001
Coronary artery disease574 (1.13)467 (1.01)107 (2.61)271 (.71)303 (2.61)<.001
Arrhythmia2276 (4.49)2057 (4.41)219 (5.21)1564 (4.01)712 (6.11)<.001
Cardiomyopathy367 (.72)305 (.71)62 (1.51)165 (.41)202 (1.71)<.001
Congenital heart disease109 (.22)98 (.21)11 (.31)83 (.21)26 (.21)0.821
Other heart disease74 (.15)72 (.21)2 (.01)49 (.11)25 (.21)0.038

Differences in continue variables were tested using the Student’s t test. Differences in categorical variables were tested using the chi-square test.

SBP = Systolic blood pressure; DBP = Diastolic blood pressure; FPG = Fasting plasma glucose; HbA1c = Hemoglobin A1c; TG = Triglyceride; TC = Total cholesterol; HDL-C = High-density lipoprotein-cholesterol; LDL = Low-density lipoprotein; SCr = Serum creatinine; eGFR = Estimated glomerular filtration rate.

Table 1

The comparisons of sociodemographic factors, lifestyle behaviors, clinical and biochemical markers, and comorbidities according to status of alcohol drinking and metabolic syndrome

Alcohol drinkingMetabolic syndrome
AllNoYesP valueNoYesP value
Variables(N = 50 640)(N = 46 468)(N = 4172)(N = 38 923)(N = 11 717)
Sociodemographic factors, n (%)
Gender<.001<.001
 Men18 535 (36.60)15 231 (32.81)3304 (79.21)13 284 (34.11)5251 (44.81)
 Women32 105 (63.40)31 237 (67.21)868 (2.81)25 639 (65.91)6466 (55.21)
Age, years, mean ± SD49.50 ± 1.6749.41 ± 1.7250.43 ± 1.12<.00148.29 ± 1.6153.50 ± 9.89<.001
Age group<.001<.001
 30–3911 277 (22.27)10 573 (22.75)704 (16.87)9951 (25.57)1326 (11.32)
 40–4913 090 (25.85)11 898 (25.60)1192 (28.57)10 658 (27.38)2432 (20.76)
 50–5915 608 (3.82)14 229 (30.62)1379 (33.05)11 439 (29.39)4169 (35.58)
 ≥6010 665 (21.06)9768 (21.02)897 (21.50)6875 (17.66)3790 (32.35)
Education attainment<.001<.001
 No formal education, illiterate80 (.16)76 (.21)4 (.11)55 (.11)25 (.21)
 Self-study, literacy27 (.05)25 (.11)2 (.01)19 (.01)8 (.11)
 Elementary school2515 (4.97)2327 (5.01)188 (4.51)1486 (3.81)1029 (8.81)
 National (junior) middle school3793 (7.49)3354 (7.21)439 (10.51)2550 (6.61)1243 (10.61)
 High school (vocational)14 900 (29.42)13 419 (28.91)1481 (35.51)11 079 (28.51)3821 (32.61)
 University (professional)23 957 (47.31)22 247 (47.91)1710 (41.01)19 293 (49.61)4664 (39.81)
 Institute and above5368 (10.60)5020 (1.81)348 (8.31)4441 (11.41)927 (7.91)
Marriage status<.001<.001
 Unmarried6598 (13.03)6212 (13.41)386 (9.31)5550 (14.31)1048 (8.91)
 Married37 605 (74.26)34 415 (74.11)3190 (76.51)28 724 (73.81)8881 (75.81)
 Divorce or separation4271 (8.43)3777 (8.11)494 (11.81)3209 (8.21)1062 (9.11)
 Widowed2166 (4.28)2064 (4.41)102 (2.41)1440 (3.71)726 (6.21)
Living alone0.7450.550
 Yes4121 (8.14)3776 (8.13)345 (8.31)3152 (8.11)969 (8.31)
Residential place<.001<.001
 Taipei10 260 (20.26)9534 (20.51)726 (17.41)8114 (2.81)2146 (18.31)
 North district8504 (16.79)7752 (16.71)752 (18.01)6435 (16.51)2069 (17.71)
 Central district9273 (18.31)8422 (18.11)851 (20.41)7123 (18.31)2150 (18.31)
 Southern district9439 (18.64)8627 (18.61)812 (19.51)7227 (18.61)2212 (18.91)
 Kaohsiung Pingtung10 588 (20.91)9806 (21.11)782 (18.71)8133 (20.91)2455 (21.01)
 East district2576 (5.09)2327 (5.01)249 (6.01)1891 (4.91)685 (5.81)
Lifestyle behaviors, n (%)
Duration of alcohol drinking (month), mean ± SD17.08 ± 7.430.02 ± .51207.11 ± 144.43<.00113.57 ± 61.2928.76 ± 93.72<.001
Smoking experience<.001<.001
 Yes13 898 (27.44)10 808 (23.31)3090 (74.11)9883 (25.41)4015 (34.31)
Betel nut chewing experience<.001<.001
 Never eaten, or only eaten once or twice47 464 (93.73)44 631 (96.01)2833 (67.91)37 031 (95.11)10 433 (89.01)
 Eaten many times3176 (6.27)1837 (4.01)1339 (32.11)1892 (4.91)1284 (11.01)
Leisure-time physical activity<.0010.006
 Yes20 184 (39.86)18 382 (39.61)1802 (43.21)15 385 (39.51)4799 (41.01)
Clinical and biochemical markers, mean ± SD
Waistline, cm83.16 ± 1.1382.77 ± 1.0787.54 ± 9.85<.00180.60 ± 8.9991.67 ± 9.00<.001
SBP, mmHg119.21 ± 17.88118.65 ± 17.78125.43 ± 17.80<.001115.48 ± 16.29131.61 ± 17.33<.001
DBP, mmHg73.36 ± 11.0372.92 ± 1.9078.25 ± 11.39<.00171.33 ± 1.2380.09 ± 1.95<.001
FPG, mg/dL95.70 ± 2.2895.29 ± 19.69100.23 ± 25.49<.00191.83 ± 12.85108.57 ± 31.84<.001
HbA1c, %5.75 ± .795.75 ± .785.83 ± .93<.0015.60 ± .546.27 ± 1.18<.001
TG, mg/dL115.68 ± 95.87112.69 ± 87.33148.99 ± 159.39<.00193.02 ± 55.88190.95 ± 148.25<.001
TC, mg/dL195.61 ± 35.77195.73 ± 35.63194.37 ± 37.280.024194.51 ± 34.62199.27 ± 39.14<.001
HDL-C, mg/dL54.42 ± 13.3854.65 ± 13.3251.86 ± 13.75<.00157.42 ± 12.8744.47 ± 9.74<.001
LDL, mg/dL121.13 ± 31.73121.38 ± 31.66118.32 ± 32.40<.001120.26 ± 31.01124.02 ± 33.86<.001
Albumin, g/dL4.52 ± .234.52 ± .234.55 ± .24<.0014.52 ± .234.54 ± .23<.001
SCr, mg/dL0.72 ± .310.71 ± .290.85 ± .42<.0010.71 ± .260.77 ± .42<.001
eGFR, mL/min/1.73 m2102.59 ± 14.79102.95 ± 14.6398.57 ± 15.87<.001104.14 ± 14.0697.43 ± 15.91<.001
Diseases of Metabolic components, n (%)
Hypertension5977 (11.80)5150 (11.11)827 (19.81)2259 (5.81)3718 (31.71)<.001
Hyperlipidemia3593 (7.10)3132 (6.71)461 (11.01)1248 (3.21)2345 (20.01)<.001
Diabetes2444 (4.83)2131 (4.61)313 (7.51)687 (1.81)1757 (15.01)<.001
Comorbidities, n (%)
Arthritis2367 (4.67)2194 (4.71)173 (4.11)1537 (3.91)830 (7.11)<.001
Gout1940 (3.83)1524 (3.31)416 (10.01)992 (2.51)948 (8.11)<.001
Stroke300 (.59)238 (.51)62 (1.51)144 (.41)156 (1.31)<.001
Liver or gall stones2253 (4.45)2073 (4.51)180 (4.31)1519 (3.91)734 (6.31)<.001
Kidney stones3188 (6.30)2767 (6.01)421 (10.11)2070 (5.31)1118 (9.51)<.001
Valvular heart disease2106 (4.16)1983 (4.31)123 (2.91)1718 (4.41)388 (3.31)<.001
Coronary artery disease574 (1.13)467 (1.01)107 (2.61)271 (.71)303 (2.61)<.001
Arrhythmia2276 (4.49)2057 (4.41)219 (5.21)1564 (4.01)712 (6.11)<.001
Cardiomyopathy367 (.72)305 (.71)62 (1.51)165 (.41)202 (1.71)<.001
Congenital heart disease109 (.22)98 (.21)11 (.31)83 (.21)26 (.21)0.821
Other heart disease74 (.15)72 (.21)2 (.01)49 (.11)25 (.21)0.038
Alcohol drinkingMetabolic syndrome
AllNoYesP valueNoYesP value
Variables(N = 50 640)(N = 46 468)(N = 4172)(N = 38 923)(N = 11 717)
Sociodemographic factors, n (%)
Gender<.001<.001
 Men18 535 (36.60)15 231 (32.81)3304 (79.21)13 284 (34.11)5251 (44.81)
 Women32 105 (63.40)31 237 (67.21)868 (2.81)25 639 (65.91)6466 (55.21)
Age, years, mean ± SD49.50 ± 1.6749.41 ± 1.7250.43 ± 1.12<.00148.29 ± 1.6153.50 ± 9.89<.001
Age group<.001<.001
 30–3911 277 (22.27)10 573 (22.75)704 (16.87)9951 (25.57)1326 (11.32)
 40–4913 090 (25.85)11 898 (25.60)1192 (28.57)10 658 (27.38)2432 (20.76)
 50–5915 608 (3.82)14 229 (30.62)1379 (33.05)11 439 (29.39)4169 (35.58)
 ≥6010 665 (21.06)9768 (21.02)897 (21.50)6875 (17.66)3790 (32.35)
Education attainment<.001<.001
 No formal education, illiterate80 (.16)76 (.21)4 (.11)55 (.11)25 (.21)
 Self-study, literacy27 (.05)25 (.11)2 (.01)19 (.01)8 (.11)
 Elementary school2515 (4.97)2327 (5.01)188 (4.51)1486 (3.81)1029 (8.81)
 National (junior) middle school3793 (7.49)3354 (7.21)439 (10.51)2550 (6.61)1243 (10.61)
 High school (vocational)14 900 (29.42)13 419 (28.91)1481 (35.51)11 079 (28.51)3821 (32.61)
 University (professional)23 957 (47.31)22 247 (47.91)1710 (41.01)19 293 (49.61)4664 (39.81)
 Institute and above5368 (10.60)5020 (1.81)348 (8.31)4441 (11.41)927 (7.91)
Marriage status<.001<.001
 Unmarried6598 (13.03)6212 (13.41)386 (9.31)5550 (14.31)1048 (8.91)
 Married37 605 (74.26)34 415 (74.11)3190 (76.51)28 724 (73.81)8881 (75.81)
 Divorce or separation4271 (8.43)3777 (8.11)494 (11.81)3209 (8.21)1062 (9.11)
 Widowed2166 (4.28)2064 (4.41)102 (2.41)1440 (3.71)726 (6.21)
Living alone0.7450.550
 Yes4121 (8.14)3776 (8.13)345 (8.31)3152 (8.11)969 (8.31)
Residential place<.001<.001
 Taipei10 260 (20.26)9534 (20.51)726 (17.41)8114 (2.81)2146 (18.31)
 North district8504 (16.79)7752 (16.71)752 (18.01)6435 (16.51)2069 (17.71)
 Central district9273 (18.31)8422 (18.11)851 (20.41)7123 (18.31)2150 (18.31)
 Southern district9439 (18.64)8627 (18.61)812 (19.51)7227 (18.61)2212 (18.91)
 Kaohsiung Pingtung10 588 (20.91)9806 (21.11)782 (18.71)8133 (20.91)2455 (21.01)
 East district2576 (5.09)2327 (5.01)249 (6.01)1891 (4.91)685 (5.81)
Lifestyle behaviors, n (%)
Duration of alcohol drinking (month), mean ± SD17.08 ± 7.430.02 ± .51207.11 ± 144.43<.00113.57 ± 61.2928.76 ± 93.72<.001
Smoking experience<.001<.001
 Yes13 898 (27.44)10 808 (23.31)3090 (74.11)9883 (25.41)4015 (34.31)
Betel nut chewing experience<.001<.001
 Never eaten, or only eaten once or twice47 464 (93.73)44 631 (96.01)2833 (67.91)37 031 (95.11)10 433 (89.01)
 Eaten many times3176 (6.27)1837 (4.01)1339 (32.11)1892 (4.91)1284 (11.01)
Leisure-time physical activity<.0010.006
 Yes20 184 (39.86)18 382 (39.61)1802 (43.21)15 385 (39.51)4799 (41.01)
Clinical and biochemical markers, mean ± SD
Waistline, cm83.16 ± 1.1382.77 ± 1.0787.54 ± 9.85<.00180.60 ± 8.9991.67 ± 9.00<.001
SBP, mmHg119.21 ± 17.88118.65 ± 17.78125.43 ± 17.80<.001115.48 ± 16.29131.61 ± 17.33<.001
DBP, mmHg73.36 ± 11.0372.92 ± 1.9078.25 ± 11.39<.00171.33 ± 1.2380.09 ± 1.95<.001
FPG, mg/dL95.70 ± 2.2895.29 ± 19.69100.23 ± 25.49<.00191.83 ± 12.85108.57 ± 31.84<.001
HbA1c, %5.75 ± .795.75 ± .785.83 ± .93<.0015.60 ± .546.27 ± 1.18<.001
TG, mg/dL115.68 ± 95.87112.69 ± 87.33148.99 ± 159.39<.00193.02 ± 55.88190.95 ± 148.25<.001
TC, mg/dL195.61 ± 35.77195.73 ± 35.63194.37 ± 37.280.024194.51 ± 34.62199.27 ± 39.14<.001
HDL-C, mg/dL54.42 ± 13.3854.65 ± 13.3251.86 ± 13.75<.00157.42 ± 12.8744.47 ± 9.74<.001
LDL, mg/dL121.13 ± 31.73121.38 ± 31.66118.32 ± 32.40<.001120.26 ± 31.01124.02 ± 33.86<.001
Albumin, g/dL4.52 ± .234.52 ± .234.55 ± .24<.0014.52 ± .234.54 ± .23<.001
SCr, mg/dL0.72 ± .310.71 ± .290.85 ± .42<.0010.71 ± .260.77 ± .42<.001
eGFR, mL/min/1.73 m2102.59 ± 14.79102.95 ± 14.6398.57 ± 15.87<.001104.14 ± 14.0697.43 ± 15.91<.001
Diseases of Metabolic components, n (%)
Hypertension5977 (11.80)5150 (11.11)827 (19.81)2259 (5.81)3718 (31.71)<.001
Hyperlipidemia3593 (7.10)3132 (6.71)461 (11.01)1248 (3.21)2345 (20.01)<.001
Diabetes2444 (4.83)2131 (4.61)313 (7.51)687 (1.81)1757 (15.01)<.001
Comorbidities, n (%)
Arthritis2367 (4.67)2194 (4.71)173 (4.11)1537 (3.91)830 (7.11)<.001
Gout1940 (3.83)1524 (3.31)416 (10.01)992 (2.51)948 (8.11)<.001
Stroke300 (.59)238 (.51)62 (1.51)144 (.41)156 (1.31)<.001
Liver or gall stones2253 (4.45)2073 (4.51)180 (4.31)1519 (3.91)734 (6.31)<.001
Kidney stones3188 (6.30)2767 (6.01)421 (10.11)2070 (5.31)1118 (9.51)<.001
Valvular heart disease2106 (4.16)1983 (4.31)123 (2.91)1718 (4.41)388 (3.31)<.001
Coronary artery disease574 (1.13)467 (1.01)107 (2.61)271 (.71)303 (2.61)<.001
Arrhythmia2276 (4.49)2057 (4.41)219 (5.21)1564 (4.01)712 (6.11)<.001
Cardiomyopathy367 (.72)305 (.71)62 (1.51)165 (.41)202 (1.71)<.001
Congenital heart disease109 (.22)98 (.21)11 (.31)83 (.21)26 (.21)0.821
Other heart disease74 (.15)72 (.21)2 (.01)49 (.11)25 (.21)0.038

Differences in continue variables were tested using the Student’s t test. Differences in categorical variables were tested using the chi-square test.

SBP = Systolic blood pressure; DBP = Diastolic blood pressure; FPG = Fasting plasma glucose; HbA1c = Hemoglobin A1c; TG = Triglyceride; TC = Total cholesterol; HDL-C = High-density lipoprotein-cholesterol; LDL = Low-density lipoprotein; SCr = Serum creatinine; eGFR = Estimated glomerular filtration rate.

Table 2

The odds ratios of metabolic syndrome and its components for alcohol drinking status estimated in the observational study

Alcohol drinkingCrudeAdjusted
VariablesNoYesχ2OR (95% CI)OR (95% CI)
Metabolic syndrome228.229
 Yes10 357 (22.29)1360 (32.60)1.69 (1.57, 1.81)***1.08 (1.00, 1.17)*
Metabolic syndrome’s components
Waistline ≥90 cm (men) or ≥ 80 cm (women)
 Yes21 202 (45.63)1963 (47.05)1.06 (.99, 1.13)1.08 (1.01, 1.15)*
Blood pressure (systolic/diastolic) >130/85 mmHg or hypertension
 Yes14 170 (30.49)2039 (48.87)2.18 (2.04, 2.32)***1.40 (1.30, 1.51)***
FPG ≥100 mg/dL or diabetes
 Yes9391 (20.21)1379 (33.05)1.95 (1.82, 2.09)***1.27 (1.17, 1.37)***
TG ≥150 mg/dL or hyperlipidemia
 Yes11 072 (23.83)1594 (38.21)1.98 (1.85, 2.11)***1.15 (1.07, 1.24)***
HDL-C <40 mg/dL (men) or <50 mg/dL (women)
 Yes12 149 (26.14)867 (20.78)0.74 (0.69, .80)***0.67 (0.62, .73)***
Alcohol drinkingCrudeAdjusted
VariablesNoYesχ2OR (95% CI)OR (95% CI)
Metabolic syndrome228.229
 Yes10 357 (22.29)1360 (32.60)1.69 (1.57, 1.81)***1.08 (1.00, 1.17)*
Metabolic syndrome’s components
Waistline ≥90 cm (men) or ≥ 80 cm (women)
 Yes21 202 (45.63)1963 (47.05)1.06 (.99, 1.13)1.08 (1.01, 1.15)*
Blood pressure (systolic/diastolic) >130/85 mmHg or hypertension
 Yes14 170 (30.49)2039 (48.87)2.18 (2.04, 2.32)***1.40 (1.30, 1.51)***
FPG ≥100 mg/dL or diabetes
 Yes9391 (20.21)1379 (33.05)1.95 (1.82, 2.09)***1.27 (1.17, 1.37)***
TG ≥150 mg/dL or hyperlipidemia
 Yes11 072 (23.83)1594 (38.21)1.98 (1.85, 2.11)***1.15 (1.07, 1.24)***
HDL-C <40 mg/dL (men) or <50 mg/dL (women)
 Yes12 149 (26.14)867 (20.78)0.74 (0.69, .80)***0.67 (0.62, .73)***

Adjusted odds ratio: Adjusting for sociodemographic factors of gender, age group, education attainment, marriage status, living alone, place of residence, life-style behaviors of smoking experience, betel nut experience, and leisure-time physical activity, and comorbidities of having arthritis, gout, stroke, liver or gall stones, kidney stones, valvular heart disease, coronary artery disease, arrhythmia, cardiomyopathy, congenital heart disease and other heart disease. Metabolic syndrome is defined as fitting three conditions under five components.

*P < .05; **P < .01; ***P < .001; OR = odds ratio; CI = confidence interval.

Table 2

The odds ratios of metabolic syndrome and its components for alcohol drinking status estimated in the observational study

Alcohol drinkingCrudeAdjusted
VariablesNoYesχ2OR (95% CI)OR (95% CI)
Metabolic syndrome228.229
 Yes10 357 (22.29)1360 (32.60)1.69 (1.57, 1.81)***1.08 (1.00, 1.17)*
Metabolic syndrome’s components
Waistline ≥90 cm (men) or ≥ 80 cm (women)
 Yes21 202 (45.63)1963 (47.05)1.06 (.99, 1.13)1.08 (1.01, 1.15)*
Blood pressure (systolic/diastolic) >130/85 mmHg or hypertension
 Yes14 170 (30.49)2039 (48.87)2.18 (2.04, 2.32)***1.40 (1.30, 1.51)***
FPG ≥100 mg/dL or diabetes
 Yes9391 (20.21)1379 (33.05)1.95 (1.82, 2.09)***1.27 (1.17, 1.37)***
TG ≥150 mg/dL or hyperlipidemia
 Yes11 072 (23.83)1594 (38.21)1.98 (1.85, 2.11)***1.15 (1.07, 1.24)***
HDL-C <40 mg/dL (men) or <50 mg/dL (women)
 Yes12 149 (26.14)867 (20.78)0.74 (0.69, .80)***0.67 (0.62, .73)***
Alcohol drinkingCrudeAdjusted
VariablesNoYesχ2OR (95% CI)OR (95% CI)
Metabolic syndrome228.229
 Yes10 357 (22.29)1360 (32.60)1.69 (1.57, 1.81)***1.08 (1.00, 1.17)*
Metabolic syndrome’s components
Waistline ≥90 cm (men) or ≥ 80 cm (women)
 Yes21 202 (45.63)1963 (47.05)1.06 (.99, 1.13)1.08 (1.01, 1.15)*
Blood pressure (systolic/diastolic) >130/85 mmHg or hypertension
 Yes14 170 (30.49)2039 (48.87)2.18 (2.04, 2.32)***1.40 (1.30, 1.51)***
FPG ≥100 mg/dL or diabetes
 Yes9391 (20.21)1379 (33.05)1.95 (1.82, 2.09)***1.27 (1.17, 1.37)***
TG ≥150 mg/dL or hyperlipidemia
 Yes11 072 (23.83)1594 (38.21)1.98 (1.85, 2.11)***1.15 (1.07, 1.24)***
HDL-C <40 mg/dL (men) or <50 mg/dL (women)
 Yes12 149 (26.14)867 (20.78)0.74 (0.69, .80)***0.67 (0.62, .73)***

Adjusted odds ratio: Adjusting for sociodemographic factors of gender, age group, education attainment, marriage status, living alone, place of residence, life-style behaviors of smoking experience, betel nut experience, and leisure-time physical activity, and comorbidities of having arthritis, gout, stroke, liver or gall stones, kidney stones, valvular heart disease, coronary artery disease, arrhythmia, cardiomyopathy, congenital heart disease and other heart disease. Metabolic syndrome is defined as fitting three conditions under five components.

*P < .05; **P < .01; ***P < .001; OR = odds ratio; CI = confidence interval.

An additive model was used to determine whether single-nucleotide polymorphisms (SNPs) satisfy the SNP-level MR assumption 1. Figure 2 presents the forest plot of ORs for significant SNPs associated with alcohol drinking, satisfying MR assumption 1 (all p < .05). We reversed the codes for SNPs with negative associations, that is, OR value less than 1, as 2, 1, or 0 based on the number of minor alleles to ensure that the direction of the OR value would be consistently positive. Supplementary Fig. 2 presents the forest plot of ORs for non-significant SNPs associated with alcohol drinking.

Forest plot of odds ratios for significant SNPs associated with alcohol drinking
Figure 2

Forest plot of odds ratios for significant SNPs associated with alcohol drinking

In testing whether SNPs satisfy the SNP-level MR assumption 3, the association between individual SNPs and MetS was assessed using an additive model. Alcohol drinking-associated SNPs that were not significantly associated with MetS are presented in Fig. 3, whereas Supplementary Fig. 3 presents the forest plot of ORs for significant SNPs associated with MetS, that is, SNPs not satisfying MR assumption 3. Considering that SNPs satisfy MR assumptions 1 and 3, 40 SNPs were found. The linkage disequilibrium (LD) of these SNPs were examined, and 13 SNPs were retained (Supplementary Fig. 4). The weighted and unweighted genetic risk scores were derived using the 13 alcohol drinking-associated SNPs. We also examined the genetic risk score-level MR assumption 1, that is, the associations between weighted and unweighted genetic risk scores and alcohol drinking (Supplementary Table 2). The weighted and unweighted genetic risk scores were significantly positively associated with alcohol drinking either in continuous or categorical forms with and without adjustment, that is, satisfying assumption 1. As the weighted or unweighted genetic risk scores increased, the likelihood of alcohol drinking increased.

Forest plot of odds ratios for non-significant SNPs associated with metabolic syndrome
Figure 3

Forest plot of odds ratios for non-significant SNPs associated with metabolic syndrome

We then examined the MR genetic risk score-level MR assumption 2, that is, the association between weighted and unweighted genetic risk scores and covariates. The ORs were presented for associations of alcohol drinking-related weighted and unweighted genetic risk scores with covariates, including sociodemographic factors, lifestyle behaviors, clinical and biochemical markers, and comorbidities by using multinomial logistic regression analysis (Supplementary Table 3). Weighted genetic risk score was significantly associated with smoking experience, leisure-time physical activity, LDL, and gout. Unweighted genetic risk score was significantly associated with education attainment, marriage status, smoking experience, betel nut chewing experience, and congenital heart disease. These significant covariates did not satisfy assumption 2, which would not be considered for adjustment in the first stage of model for deriving the likelihood of alcohol drinking by using genetic risk scores.

We also examined the genetic risk score-level MR assumption 3, that is, the association between weighted and unweighted genetic risk scores and MetS (Supplementary Table 4). The weighted genetic risk score for alcohol drinking was not significantly associated with MetS either in continuous or categorical form with and without adjustment, that is, satisfying assumption 3. The unweighted genetic risk score for alcohol drinking in categorical form was significantly associated with MetS with or without adjustment, that is, not satisfying assumption 3. The unweighted genetic risk score would not be considered for subsequent analysis, i.e. deriving the likelihood of alcohol drinking.

Table 3 shows the ORs of MetS for genetic-related alcohol drinking derived from the weighted genetic risk score with and without adjustment. Genetic-related alcohol drinking was the phat, indicating the likelihood of genetic predisposition to drink, which was derived from regressing alcohol drinking on the weighted genetic risk score. The crude OR of MetS per 5% increase in the phat of alcohol drinking derived from weighted genetic risk score without adjustment was 1.00 (.97, 1.03). The phat of alcohol drinking was not associated with MetS when it was categorized according to quartiles. After multivariate adjustment, the OR of MetS per 5% increase in the phat of alcohol drinking derived from weighted genetic risk score was 1.11 (1.10, 1.12). After grouping the phat of alcohol drinking derived from weighted genetic risk score with adjustment according to the quartiles, the highest MetS prevalence rate was observed in Q2 (24.10%), and the lowest was in Q1 (22.41%). Using Q1 as reference group, the adjusted ORs of MetS for Q2, Q3, and Q4 of the phat of alcohol drinking derived from weighted genetic risk score were 1.19 (1.12, 1.27), 1.31 (1.23, 1.40), and 1.87 (1.75, 2.00), respectively. After multivariate adjustment, the weighted genetic risk score showed a linear trend between genetic-related alcohol drinking and MetS. This finding indicates that a high likelihood of genetic-related alcohol drinking is related to a high risk of MetS.

Table 3

The odds ratios of metabolic syndrome for genetic-related alcohol drinking derived from weighted genetic risk score with and without adjustment

CrudeAdjusted
VariablesnYes, n (%)OR (95% CI)VariablesnYes, n (%)OR (95% CI)
Phat of alcohol drinking derived from weighted genetic risk score without adjustmentPhat of alcohol drinking derived from weighted genetic risk score with adjustment
Per 5% increase50 64011 7171.00 (.97,1.03)Per 5% increase50 64011 7171.11 (1.10, 1.12)***
Q1: <.05212 6142881 (22.84)1.00Q1: <.06893492095 (22.41)1.00
Q2: .052–0.07412 6172884 (22.86)1.00 (.94, 1.06)Q2: .068–0.07611 9652883 (24.10)1.19 (1.12, 1.27)***
Q3: .075–0.10512 7483001 (23.54)1.04 (.98, 1.10)Q3: .077–0.09319 6944516 (22.93)1.31 (1.23, 1.40)***
Q4: ≥.10612 6612951 (23.31)1.03 (.97, 1.09)Q4: ≥.09496322223 (23.08)1.87 (1.75, 2.00)***
CrudeAdjusted
VariablesnYes, n (%)OR (95% CI)VariablesnYes, n (%)OR (95% CI)
Phat of alcohol drinking derived from weighted genetic risk score without adjustmentPhat of alcohol drinking derived from weighted genetic risk score with adjustment
Per 5% increase50 64011 7171.00 (.97,1.03)Per 5% increase50 64011 7171.11 (1.10, 1.12)***
Q1: <.05212 6142881 (22.84)1.00Q1: <.06893492095 (22.41)1.00
Q2: .052–0.07412 6172884 (22.86)1.00 (.94, 1.06)Q2: .068–0.07611 9652883 (24.10)1.19 (1.12, 1.27)***
Q3: .075–0.10512 7483001 (23.54)1.04 (.98, 1.10)Q3: .077–0.09319 6944516 (22.93)1.31 (1.23, 1.40)***
Q4: ≥.10612 6612951 (23.31)1.03 (.97, 1.09)Q4: ≥.09496322223 (23.08)1.87 (1.75, 2.00)***

Adjusted odds ratio: adjusting for residual, PCA and covariates that satisfying the assumption 2, respectively.

*P < .05; **P < .01; ***P < .001; OR = odds ratio; CI = Confidence interval; Q1 = the first quartile; Q2 = the second quartile; Q3 = the third quartile; Q4 = the fourth quartile.

Table 3

The odds ratios of metabolic syndrome for genetic-related alcohol drinking derived from weighted genetic risk score with and without adjustment

CrudeAdjusted
VariablesnYes, n (%)OR (95% CI)VariablesnYes, n (%)OR (95% CI)
Phat of alcohol drinking derived from weighted genetic risk score without adjustmentPhat of alcohol drinking derived from weighted genetic risk score with adjustment
Per 5% increase50 64011 7171.00 (.97,1.03)Per 5% increase50 64011 7171.11 (1.10, 1.12)***
Q1: <.05212 6142881 (22.84)1.00Q1: <.06893492095 (22.41)1.00
Q2: .052–0.07412 6172884 (22.86)1.00 (.94, 1.06)Q2: .068–0.07611 9652883 (24.10)1.19 (1.12, 1.27)***
Q3: .075–0.10512 7483001 (23.54)1.04 (.98, 1.10)Q3: .077–0.09319 6944516 (22.93)1.31 (1.23, 1.40)***
Q4: ≥.10612 6612951 (23.31)1.03 (.97, 1.09)Q4: ≥.09496322223 (23.08)1.87 (1.75, 2.00)***
CrudeAdjusted
VariablesnYes, n (%)OR (95% CI)VariablesnYes, n (%)OR (95% CI)
Phat of alcohol drinking derived from weighted genetic risk score without adjustmentPhat of alcohol drinking derived from weighted genetic risk score with adjustment
Per 5% increase50 64011 7171.00 (.97,1.03)Per 5% increase50 64011 7171.11 (1.10, 1.12)***
Q1: <.05212 6142881 (22.84)1.00Q1: <.06893492095 (22.41)1.00
Q2: .052–0.07412 6172884 (22.86)1.00 (.94, 1.06)Q2: .068–0.07611 9652883 (24.10)1.19 (1.12, 1.27)***
Q3: .075–0.10512 7483001 (23.54)1.04 (.98, 1.10)Q3: .077–0.09319 6944516 (22.93)1.31 (1.23, 1.40)***
Q4: ≥.10612 6612951 (23.31)1.03 (.97, 1.09)Q4: ≥.09496322223 (23.08)1.87 (1.75, 2.00)***

Adjusted odds ratio: adjusting for residual, PCA and covariates that satisfying the assumption 2, respectively.

*P < .05; **P < .01; ***P < .001; OR = odds ratio; CI = Confidence interval; Q1 = the first quartile; Q2 = the second quartile; Q3 = the third quartile; Q4 = the fourth quartile.

Discussion

This study evaluated the causal association between alcohol drinking and MetS by using one-sample MR analysis with SNPs in alcohol dehydrogenase 1B (ADH1B), alcohol dehydrogenase 1C (ADH1C), aldehyde dehydrogenase 2 family member (ALDH2), and other genes identified from prior GWAS studies as instrument variables in adults enrolled in the Taiwan Biobank database. After multivariate adjustment for covariates and 10 principal components from principal component analysis (PCA), the chance of genetic-related alcohol drinking was positively associated with MetS. A linear trend was found between genetic-related alcohol drinking and MetS, that is, individuals with a genetic predisposition to consume alcohol had higher odds of MetS. A 5% increase in likelihood of genetic-related alcohol drinking derived from weighted genetic risk score was associated with a 10% increase in the risk of having MetS, indicating the modest effect of genetic-related alcohol drinking. This study is the first to provide experimental evidence supporting the association between alcohol drinking and MetS.

Prior epidemiologic studies showed conflicting results on the association between alcohol consumption and MetS (Djousse et al. 2004; Baik and Shin 2008; Sun et al. 2014; Vidot et al. 2016; Lin et al. 2021). Some studies demonstrated positive associations (Baik and Shin 2008; Sun et al. 2014; Lin et al. 2021), whereas some showed inverse associations (Djousse et al. 2004; Sun et al. 2014; Vidot et al. 2016). Consistent with our findings of traditional observational study, previous works found that alcohol drinking was associated with a 60%–84% increase in likelihood of having MetS (Baik and Shin 2008; Sun et al. 2014; Lin et al. 2021). These studies show that heavy or frequent alcohol drinking were associated with an increased risk or prevalence of MetS (Baik and Shin 2008; Sun et al. 2014; Lin et al. 2021), but such association was not observed in light or moderate alcohol drinking. By contrast, some observational studies show that light/low or moderate alcohol consumption is associated with a lower prevalence of MetS (Djousse et al. 2004; Sun et al. 2014; Vidot et al. 2016). In a study that explored different types of alcohol drinking, the protective effect is consistent across those who drink wine only, beer only, spirits only, and any combination (Djousse et al. 2004). Given these inconsistent findings, considerable research is needed to provide evidence in this line of research questions. The present results were supported by an observational study that used MR analysis to determine the relationship between alcohol drinking and MetS (Baik and Shin 2008; Sun et al. 2014; Lin et al. 2021). The association between alcohol drinking and MetS in MR analysis was compared with that estimated in the observational study. The direction and statistical significance of the association were similar, indicating a significant association between alcohol drinking and MetS after ruling out the possibility of reverse causality with experimental evidence.

The advantages of this study included a large sample size and standardized data collection procedures and laboratory techniques for blood and genetic analysis to ensure a high degree of reliability and consideration of many potential confounding factors. In addition, all participants are of the same race.

This work also has some limitations. First, the present study considered a binary variable of alcohol drinking status because of the high missing rates of frequency, volume, and type of alcohol drinking. Dichotomizing alcohol use might lose information of different drinking amounts. In addition, this study included people who have drunk alcohol for two or more years before quitting alcohol. Some people may have quit alcohol drinking for a long time. As such, the effect of alcohol drinking on MetS may be reduced by including those who quit alcohol drinking, resulting in an underestimated risk of MetS. Second, smoking status was defined as having smoked more than one cigarette in their lifetime, which may not be appropriate. This method of measurement could result in a significant number of individuals being categorized as smokers, potentially leading to a wide range of cardiovascular risk. However, these concepts have been the basis for establish questions about smoking in the data collection process. We utilized existing data from the Taiwan Biobank dataset, and we do not have control over the manner in which smoking status was measured. While acknowledging that the method of measuring smoking status may not be ideal, it serves as a covariate rather than a key independent variable in our analysis. Additionally, we adjusted for comorbidities, such as CVDs, which may help mitigate the effect of misclassification errors related to smoking status. Third, this study adopted a cross-sectional design, which cannot provide a time sequence analysis similar to a cohort study. However, reverse causality can be preliminarily ruled out because genes are innately determined. Fourth, the findings were obtained from the Han Chinese population, which cannot be extrapolated to other populations because of differences in genes, race, and alcohol intake behavior. Based on previous reports, alcohol drinking consumption in Taiwan is not as high as that in other race population. The per capita annual alcohol consumption (in liters) in the country of origin is low in Taiwan (4.45) compared with those in Japan (7.38) and South Korea (7.71) (Cook et al. 2012). Fifth, although we included all variants that were thoroughly searched in literature, our study did not include some alcohol-related loci (rs698, rs1799971, rs1800497, rs1800759, and rs6265) (Lawlor et al. 2013; Katsarou et al. 2017; Christensen et al. 2018). If these loci satisfy MR assumptions and are considered in MR analysis, then the magnitude of associations assessed would be stronger, indicating that the effect of ignoring these loci on the study findings should be trivial. In the future, more alcohol-related loci should be found, and future studies should consider such newly identified loci. Finally, the sample of Taiwan Biobank may not be representative of Taiwan general population, indicating potential selection bias. In evaluating this possibility, we assessed the demographic characteristics of the sample of Taiwan Biobank and Taiwan population from the data of the Ministry of the Interior. The data collection period of Taiwan Biobank was from 2008 to 2020; thus, population of year 2013 from the data of the Ministry of the Interior was used. The age distribution of the population aged 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, and 65–70 years were 14.96%, 14.29%, 13.49%, 13.79%, 13.84%, 12.45%, 1.21%, and 6.97%, respectively (Registration, 2011). In addition, the corresponding percentages in Taiwan Biobank were 1.18%, 12.09%, 12.87%, 12.98%, 15.36%, 15.47%, 12.62%, and 8.44%, respectively. The percentages of men and women aged 30–70 years old in the data of Ministry of the Interior were 49.33% and 5.67%, respectively, and the corresponding percentage of Taiwan Biobank was 36.60% and 63.40%, respectively. Most of the differences in the distribution of age between sample and population is less than 3%; only one age group (30–34 years) is higher (4.78%). These results indicate that this kind of selection error caused by age distribution might be random; thus, the biased results in the effect may be toward the null, a lesser threat to validity. However, women were oversampling in the present study (63.40% in the sample vs. 5.67% in the population). The prevalence of alcohol drinking in the present study was lower than that estimated by National Health Interview Survey in 2013 (8.24% vs. 41.66%). The differences in alcohol drinking prevalence may be due to the definition of alcohol drinking. In the present study, alcohol drinking is defined as those who have been drinking at least 150 cc of alcohol per week for six consecutive months, whereas those who have alcohol use were those who drink any types of alcohol in the past 1 year based on the National Health Interview Survey without specifying the amount of alcohol and period of continuous drinking. Considering that the primary aim of the study is analytic, the most important consideration is whether the sample has enough study subjects with the major predictor of alcohol drinking and a sufficient number of persons with MetS to assess the potential relationship between alcohol drinking and MetS. The higher proportion of women and the lower prevalence of alcohol drinking would reduce the power of the study. By contrast, the definition of alcohol drinking increases the effect size, that is, the magnitude of the strength of association. Under the abovementioned conditions, the effect of the unrepresentativeness of gender and alcohol drinking on the validity of the study findings is trivial because differential selection bias is unlikely (i.e. women who drank alcohol with MetS is more likely to be selected as study subjects or vice versa). In addition, MR controlled potential confounders that satisfied the MR assumption 2, and this study has further adjusted for as many potential confounding variables as possible in multivariate analyses. Therefore, the confounding effects of gender should be excluded.

Conclusion

This study shows the modest relationship between drinking behavior and MetS and provides the first experimental evidence of research in the field of alcohol drinking and MetS in Han Chinese by using MR approach. Knowledge obtained can be integrated into health education intervention.

Acknowledgements

This research has been conducted using the Taiwan Biobank resource. We thank all the participants and investigators of the Taiwan Biobank.

Author contributions

Chuan-Wei Yang (Conceptualization [equal], Formal analysis [equal], Methodology [equal], Writing—original draft [equal]), Yu-Syuan Wei (Conceptualization [equal], Formal analysis [equal], Methodology [equal], Writing—original draft [equal]), Chia-Ing Li (Formal analysis [supporting], Writing—review & editing [equal]), Chiu-Shong Liu (Data curation [equal], Writing—review & editing [equal]), Chih-Hsueh Lin (Data curation [equal], Writing—review & editing [equal]), Cheng-Chieh Lin (Conceptualization [equal], Data curation [equal], Investigation [equal], Methodology [equal], Writing—review & editing [equal]) and Tsai-Chung Li (Conceptualization, Data curation [lead], Investigation [equal], Methodology [lead], Formal analysis [lead], Writing—review & editing [equal])

Funding

This study was supported primarily by the Ministry of Science and Technology of Taiwan (MOST 108–2314-B-039-035-MY3) and China Medical University (CMU110-MF-53).

Conflict of interest: The authors declare that they have no competing interests.

Data availability

The datasets generated during and analyzed during the current study are not publicly available due to the policy declared by Taiwan Biobank but are available from the corresponding author on reasonable request.

References

Alberti
KG
,
Zimmet
P
,
Shaw
J
.
Metabolic syndrome—a new world‐wide definition. A consensus statement from the international diabetes federation
.
Diabet Med
2006
;
23
:
469
80
. https://doi-org.libproxy.ucl.ac.uk/10.1111/j.1464-5491.2006.01858.x.

Alberti
KG
,
Eckel
RH
,
Grundy
SM
. et al. 
Harmonizing the metabolic syndrome: a joint interim statement of the international diabetes federation task force on epidemiology and prevention; National Heart, Lung, and Blood Institute; American Heart Association; world heart federation; international atherosclerosis society; and International Association for the Study of obesity
.
Circulation
2009
;
120
:
1640
5
. https://doi-org.libproxy.ucl.ac.uk/10.1161/CIRCULATIONAHA.109.192644.

Amos
CI
,
Wu
X
,
Broderick
P
. et al. 
Genome-wide association scan of tag SNPs identifies a susceptibility locus for lung cancer at 15q25.1
.
Nat Genet
2008
;
40
:
616
22
. https://doi-org.libproxy.ucl.ac.uk/10.1038/ng.109.

Baik
I
,
Shin
C
.
Prospective study of alcohol consumption and metabolic syndrome
.
Am J Clin Nutr
2008
;
87
:
1455
63
. https://doi-org.libproxy.ucl.ac.uk/10.1093/ajcn/87.5.1455.

Barbaresko
J
,
Rienks
J
,
Nothlings
U
.
Lifestyle indices and cardiovascular disease risk: a meta-analysis
.
Am J Prev Med
2018
;
55
:
555
64
. https://doi-org.libproxy.ucl.ac.uk/10.1016/j.amepre.2018.04.046.

Biddinger
KJ
,
Emdin
CA
,
Haas
ME
. et al. 
Association of Habitual Alcohol Intake with risk of cardiovascular disease
.
JAMA Netw Open
2022
;
5
:e223849. https://doi-org.libproxy.ucl.ac.uk/10.1001/jamanetworkopen.2022.3849.

Chang
SH
,
Chen
MC
,
Chien
NH
. et al. 
CE: original research: examining the links between lifestyle factors and metabolic syndrome
.
Am J Nurs
2016
;
116
:
26
36
. https://doi-org.libproxy.ucl.ac.uk/10.1097/01.NAJ.0000508662.88220.7a.

Cho
Y
,
Shin
SY
,
Won
S
. et al. 
Alcohol intake and cardiovascular risk factors: a Mendelian randomisation study
.
Sci Rep
2015
;
5
:
18422
. https://doi-org.libproxy.ucl.ac.uk/10.1038/srep18422.

Christensen
AI
,
Nordestgaard
BG
,
Tolstrup
JS
.
Alcohol intake and risk of ischemic and haemorrhagic stroke: results from a Mendelian randomisation study
.
J Stroke
2018
;
20
:
218
27
. https://doi-org.libproxy.ucl.ac.uk/10.5853/jos.2017.01466.

Cook
WK
,
Mulia
N
,
Karriker-Jaffe
K
.
Ethnic drinking cultures and alcohol use among Asian American adults: findings from a national survey
.
Alcohol Alcohol
2012
;
47
:
340
8
. https://doi-org.libproxy.ucl.ac.uk/10.1093/alcalc/ags017.

Delgado-Lobete
L
,
Montes-Montes
R
,
Vila-Paz
A
. et al. 
Individual and environmental factors associated with tobacco smoking, alcohol abuse and illegal drug consumption in university students: a mediating analysis
.
Int J Environ Res Public Health
2020
;
17
:3019. https://doi-org.libproxy.ucl.ac.uk/10.3390/ijerph17093019.

Djousse
L
,
Arnett
DK
,
Eckfeldt
JH
. et al. 
Alcohol consumption and metabolic syndrome: does the type of beverage matter?
Obes Res
2004
;
12
:
1375
85
. https://doi-org.libproxy.ucl.ac.uk/10.1038/oby.2004.174.

Esser
N
,
Legrand-Poels
S
,
Piette
J
. et al. 
Inflammation as a link between obesity, metabolic syndrome and type 2 diabetes
.
Diabetes Res Clin Pract
2014
;
105
:
141
50
. https://doi-org.libproxy.ucl.ac.uk/10.1016/j.diabres.2014.04.006.

Fan
AZ
,
Russell
M
,
Naimi
T
. et al. 
Patterns of alcohol consumption and the metabolic syndrome
.
J Clin Endocrinol Metab
2008
;
93
:
3833
8
. https://doi-org.libproxy.ucl.ac.uk/10.1210/jc.2007-2788.

Freathy
RM
,
Ring
SM
,
Shields
B
. et al. 
A common genetic variant in the 15q24 nicotinic acetylcholine receptor gene cluster (CHRNA5-CHRNA3-CHRNB4) is associated with a reduced ability of women to quit smoking in pregnancy
.
Hum Mol Genet
2009
;
18
:
2922
7
. https://doi-org.libproxy.ucl.ac.uk/10.1093/hmg/ddp216.

Ghowsi
M
,
Qalekhani
F
,
Farzaei
MH
. et al. 
Inflammation, oxidative stress, insulin resistance, and hypertension as mediators for adverse effects of obesity on the brain: a review
.
Biomedicine
2021
;
11
:
13
22
. https://doi-org.libproxy.ucl.ac.uk/10.37796/2211-8039.1174.

Hernández-Rubio
A
,
Sanvisens
A
,
Bolao
F
. et al. 
Prevalence and associations of metabolic syndrome in patients with alcohol use disorder
.
Sci Rep
2022
;
12
:
2625
. https://doi-org.libproxy.ucl.ac.uk/10.1038/s41598-022-06010-3.

Holmes
MV
,
Dale
CE
,
Zuccolo
L
. et al. 
Association between alcohol and cardiovascular disease: Mendelian randomisation analysis based on individual participant data
.
BMJ
2014
;
349
:g4164. https://doi-org.libproxy.ucl.ac.uk/10.1136/bmj.g4164.

Jee
YH
,
Lee
SJ
,
Jung
KJ
. et al. 
Alcohol intake and serum glucose levels from the perspective of a Mendelian randomization design: the KCPS-II biobank
.
PloS One
2016
;
11
:e0162930. https://doi-org.libproxy.ucl.ac.uk/10.1371/journal.pone.0162930.

Katsarou
MS
,
Karakonstantis
K
,
Demertzis
N
. et al. 
Effect of single-nucleotide polymorphisms in ADH1B, ADH4, ADH1C, OPRM1, DRD2, BDNF, and ALDH2 genes on alcohol dependence in a Caucasian population
.
Pharmacol Res Perspect
2017
;
5
:e00326. https://doi-org.libproxy.ucl.ac.uk/10.1002/prp2.326.

Kothari
V
,
Stevens
RJ
,
Adler
AI
. et al. 
UKPDS 60: risk of stroke in type 2 diabetes estimated by the UK prospective diabetes study risk engine
.
Stroke
2002
;
33
:
1776
81
. https://doi-org.libproxy.ucl.ac.uk/10.1161/01.STR.0000020091.07144.C7.

Larsson
SC
,
Burgess
S
,
Mason
AM
. et al. 
Alcohol consumption and cardiovascular disease: a Mendelian randomization study
.
Circ Genom Precis Med
2020
;
13
:
e002814
. https://doi-org.libproxy.ucl.ac.uk/10.1161/CIRCGEN.119.002814.

Lawlor
DA
,
Harbord
RM
,
Sterne
JA
. et al. 
Mendelian randomization: using genes as instruments for making causal inferences in epidemiology
.
Stat Med
2008
;
27
:
1133
63
. https://doi-org.libproxy.ucl.ac.uk/10.1002/sim.3034.

Lawlor
DA
,
Nordestgaard
BG
,
Benn
M
. et al. 
Exploring causal associations between alcohol and coronary heart disease risk factors: findings from a Mendelian randomization study in the Copenhagen general population study
.
Eur Heart J
2013
;
34
:
2519
28
. https://doi-org.libproxy.ucl.ac.uk/10.1093/eurheartj/eht081.

Lee
WY
,
Jung
CH
,
Park
JS
. et al. 
Effects of smoking, alcohol, exercise, education, and family history on the metabolic syndrome as defined by the ATP III
.
Diabetes Res Clin Pract
2005
;
67
:
70
7
. https://doi-org.libproxy.ucl.ac.uk/10.1016/j.diabres.2004.05.006.

Lin
Y
,
Ying
YY
,
Li
SX
. et al. 
Association between alcohol consumption and metabolic syndrome among Chinese adults
.
Public Health Nutr
2021
;
24
:
4582
90
. https://doi-org.libproxy.ucl.ac.uk/10.1017/S1368980020004449.

Millwood
IY
,
Walters
RG
,
Mei
XW
. et al. 
Conventional and genetic evidence on alcohol and vascular disease aetiology: a prospective study of 500 000 men and women in China
.
Lancet
2019
;
393
:
1831
42
. https://doi-org.libproxy.ucl.ac.uk/10.1016/S0140-6736(18)31772-0.

Nelson
RG
,
Bennett
PH
,
Beck
GJ
. et al. 
Development and progression of renal disease in pima Indians with non-insulin-dependent diabetes mellitus. Diabetic renal disease study group
.
N Engl J Med
1996
;
335
:
1636
42
. https://doi-org.libproxy.ucl.ac.uk/10.1056/NEJM199611283352203.

van
Oort
S
,
Beulens
JWJ
,
van
Ballegooijen
AJ
. et al. 
Modifiable lifestyle factors and heart failure: a Mendelian randomization study
.
Am Heart J
2020
;
227
:
64
73
. https://doi-org.libproxy.ucl.ac.uk/10.1016/j.ahj.2020.06.007.

Park
KY
,
Park
HK
,
Hwang
HS
.
Relationship between abdominal obesity and alcohol drinking pattern in normal-weight, middle-aged adults: the Korea National Health and nutrition examination survey 2008-2013
.
Public Health Nutr
2017
;
20
:
2192
200
. https://doi-org.libproxy.ucl.ac.uk/10.1017/S1368980017001045.

Paulson
QX
,
Hong
J
,
Holcomb
VB
. et al. 
Effects of body weight and alcohol consumption on insulin sensitivity
.
Nutr J
2010
;
9
:
14
. https://doi-org.libproxy.ucl.ac.uk/10.1186/1475-2891-9-14.

Peng
M
,
Zhang
J
,
Zeng
T
. et al. 
Alcohol consumption and diabetes risk in a Chinese population: a Mendelian randomization analysis
.
Addiction
2019
;
114
:
436
49
. https://doi-org.libproxy.ucl.ac.uk/10.1111/add.14475.

Pickering
TG
.
The natural history of hypertension: prehypertension or masked hypertension?
J Clin Hypertens (Greenwich)
2007
;
9
:
807
10
. https://doi-org.libproxy.ucl.ac.uk/10.1111/j.1751-7176.2007.tb00011.x.

Ranasinghe
P
,
Mathangasinghe
Y
,
Jayawardena
R
. et al. 
Prevalence and trends of metabolic syndrome among adults in the asia-pacific region: a systematic review
.
BMC Public Health
2017
;
17
:
101
. https://doi-org.libproxy.ucl.ac.uk/10.1186/s12889-017-4041-1.

Saccone
SF
,
Hinrichs
AL
,
Saccone
NL
. et al. 
Cholinergic nicotinic receptor genes implicated in a nicotine dependence association study targeting 348 candidate genes with 3713 SNPs
.
Hum Mol Genet
2007
;
16
:
36
49
. https://doi-org.libproxy.ucl.ac.uk/10.1093/hmg/ddl438.

Santos
KM
,
Tsutsui
ML
,
Galvao
PP
. et al. 
Degree of physical activity and metabolic syndrome: a cross-sectional study among the Khisedje group in the Xingu Indigenous Park, Brazil
.
Cad Saude Publica
2012
;
28
:
2327
38
. https://doi-org.libproxy.ucl.ac.uk/10.1590/S0102-311X2012001400011.

Singh
RB
,
Suh
IL
,
Singh
VP
. et al. 
Hypertension and stroke in Asia: prevalence, control and strategies in developing countries for prevention
.
J Hum Hypertens
2000
;
14
:
749
63
. https://doi-org.libproxy.ucl.ac.uk/10.1038/sj.jhh.1001057.

Slanovic-Kuzmanovic
Z
,
Kos
I
,
Domijan
AM
.
Endocrine, lifestyle, and genetic factors in the development of metabolic syndrome
.
Arh Hig Rada Toksikol
2013
;
64
:
581
91
. https://doi-org.libproxy.ucl.ac.uk/10.2478/10004-1254-64-2013-2327.

Sun
K
,
Ren
M
,
Liu
D
. et al. 
Alcohol consumption and risk of metabolic syndrome: a meta-analysis of prospective studies
.
Clin Nutr
2014
;
33
:
596
602
. https://doi-org.libproxy.ucl.ac.uk/10.1016/j.clnu.2013.10.003.

Thorgeirsson
TE
,
Geller
F
,
Sulem
P
. et al. 
A variant associated with nicotine dependence, lung cancer and peripheral arterial disease
.
Nature
2008
;
452
:
638
42
. https://doi-org.libproxy.ucl.ac.uk/10.1038/nature06846.

Vidot
DC
,
Stoutenberg
M
,
Gellman
M
. et al. 
Alcohol consumption and metabolic syndrome among Hispanics/Latinos: the Hispanic community health study/study of Latinos
.
Metab Syndr Relat Disord
2016
;
14
:
354
62
. https://doi-org.libproxy.ucl.ac.uk/10.1089/met.2015.0171.

Yamaoka
K
,
Tango
T
.
Effects of lifestyle modification on metabolic syndrome: a systematic review and meta-analysis
.
BMC Med
2012
;
10
:
138
. https://doi-org.libproxy.ucl.ac.uk/10.1186/1741-7015-10-138.

Ye
XF
,
Miao
CY
,
Zhang
W
. et al. 
Alcohol intake and dyslipidemia in male patients with hypertension and diabetes enrolled in a China multicenter registry
.
J Clin Hypertens (Greenwich)
2023
;
25
:
183
90
. https://doi-org.libproxy.ucl.ac.uk/10.1111/jch.14638.

Zhang
H
,
Zhang
H
,
Li
Z
. et al. 
Statistical methods for haplotype-based matched case-control association studies
.
Genet Epidemiol
2007
;
31
:
316
26
. https://doi-org.libproxy.ucl.ac.uk/10.1002/gepi.20212.

Author notes

Chuan-Wei Yang and Yu-Syuan Wei contributed equally as first author.

This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic-oup-com.libproxy.ucl.ac.uk/pages/standard-publication-reuse-rights)

Supplementary data