Abstract
We investigated whether higher amounts and frequencies of alcohol use were associated with increased risk of Coronavirus disease 2019 (COVID-19). A total of 15,980,069 individuals who underwent routine health examinations between 2009 and 2010 were followed from November 13, 2020 until July 6, 2021 for COVID-19 incidence through national database linkage. Alcohol use was assessed via a structured questionnaire. Cox proportional hazards model was used to estimate adjusted HRs. During follow-up, 47,308 individuals were diagnosed with COVID-19. Compared with current nondrinkers, individuals consuming <2, 2 to 3.9, 4 to 6.9, and ≥7 drinks per day had hazard ratios (HRs) of 1.04 (95% confidence interval: 1.01–1.06), 1.13 (1.09–1.18), 1.17 (1.12–1.23), and 1.22 (1.14–1.31), respectively (P for trend <.001). For alcohol use frequency, those drinking 1 to 2, 3 to 4, and 5 to 7 days per week had HRs of 1.08 (1.05–1.10), 1.19 (1.14–1.23), and 1.11 (1.05–1.17), respectively, compared with those drinking less than once per week (P for trend <.001). Associations were stronger in women; even <1 drink per day was associated with elevated risk (HR = 1.06; 95% confidence interval: 1.02–1.10). Age-specific associations differed by sex. Higher amounts and more frequent alcohol intake were associated with higher risk of COVID-19. Associations varied by sex and age. Further studies are needed to clarify whether sex- and age-specific public health guidance on alcohol consumption can reduce COVID-19 transmission.
Keywords: alcohol consumption, cohort studies, COVID-19, prospective studies, SARS-CoV-2
1. Introduction
Coronavirus disease 2019 (COVID-19), caused by infection with the novel coronavirus SARS-CoV-2, emerged in late 2019 and rapidly escalated into a global pandemic, resulting in substantial morbidity and mortality worldwide.[1] Although COVID-19 has since transitioned into an endemic phase – similar to seasonal influenza – it continues to circulate globally with ongoing public health consequences.[2] The virus remains highly transmissible and continues to cause illness and death, particularly among vulnerable populations. Moreover, the potential emergence of new variants with immune-evasive properties underscores the need for continued vigilance.[3] Several studies have sought to support COVID-19 prevention strategies by examining how demographic characteristics and occupational contexts shape patterns of transmission.[4,5] In this context, identifying modifiable risk factors that increase susceptibility to infection or transmission remains essential for pandemic preparedness.
Alcohol consumption has well-documented immunosuppressive effects and social consequences, both of which can increase vulnerability to both viral and bacterial infections.[6,7] Therefore, alcohol use has the potential to serve as a risk factor for COVID-19. However, the association between alcohol consumption and COVID-19 infection risk has received relatively limited attention.[8] Only a few prospective studies have examined this association in general or low-risk populations, and findings from population-based studies have been inconsistent.[9–11] For example, during the early phase of the pandemic, some analyses of UK Biobank data reported that current alcohol drinkers had a lower risk of COVID-19–related hospitalization compared to nondrinkers.[9] Overall, whether alcohol consumption increases the risk of COVID-19 in the general population remains unclear.
Using a large, population-based cohort in Korea, we aimed to examine the association between alcohol consumption and the risk of COVID-19. Specifically, we evaluated whether higher amounts and higher frequency of alcohol use were associated with increased infection risk, and whether these associations differed by sex and age.
2. Methods
2.1. Study population and follow-up
The National Health Insurance Service (NHIS) provides mandatory health coverage for approximately 97% of the Korean population and offers biennial health screening examinations.[12] Among 17,733,108 individuals aged 18 to 99 years who underwent routine health examinations between 2009 and 2010, we excluded those with missing information on cardiometabolic risk factors, alcohol use variables, or examination dates (n = 732,760). We further excluded individuals who died before January 1, 2020 (n = 902,223), those aged ≥100 years as of January 1, 2020 (n = 881), and individuals who died or were diagnosed with COVID-19 before the index date of November 13, 2020 (n = 117,175). The final analytic cohort comprised 15,980,069 individuals who were followed until July 6, 2021 (Fig. 1). COVID-19 incidence was identified through linkage to the Korea Disease Control and Prevention Agency (KDCA) database.[13] We assessed COVID-19 cases during Period 3 (November 13, 2020–July 6, 2021), which differed from periods 1 and 2. The earlier periods were characterized by outbreaks frequently associated with religious facilities and contributed only a small proportion of the total case count.[14,15] Although vaccination began in South Korea on February 6, 2021, its population-level impact was limited during Period 3.[16] Nationwide contact tracing and case confirmation using reverse transcription polymerase chain reaction (RT-PCR) testing were implemented throughout the follow-up period in accordance with government guidelines.[17]
Figure 1.
Study flowchart.
This study was approved by the Institutional Review Board of Catholic Kwandong University. The requirement for informed consent was waived due to the anonymized nature of data provided by the NHIS.
2.2. Data collection
Fasting serum glucose, total cholesterol, and triglyceride levels were measured using enzymatic methods. Systolic blood pressure was measured after at least 5 minutes of seated rest. Body mass index (BMI) was calculated as measured weight in kilograms divided by the square of measured height in meters (kg/m2).[18] Alcohol consumption, smoking history, and physical activity were assessed via a structured self-administered questionnaire. All measurements and data collection procedures followed standardized protocols established by the Korean government, and routine external quality assessments were conducted for clinical chemistry laboratories.
Alcohol consumption was assessed using 2 questionnaire items: “On how many days per week do you usually drink alcohol?” (response range: 0–7 days), and “On a typical drinking day, how many drinks do you consume?” (regardless of beverage type). Average daily alcohol intake was calculated by multiplying the number of drinking days per week by the number of drinks consumed per drinking day, then dividing by 7. Participants who reported drinking less than once per week but indicated a nonzero amount were assigned a weekly frequency of 0.23, reflecting estimated monthly consumption. In a previous study, we found that these alcohol measures were strongly associated with the incidence of alcoholic liver disease and considered them reasonably valid for epidemiologic use.[19]
We defined individuals as having known diabetes (ICD-10 codes E10–E14), hypertension (I10–I15), cardiovascular disease (heart disease [I20–I52] or stroke [I60–I69]), or cancer (C00–C97) at baseline if they had at least 1 medical visit for the relevant condition within the 12 months preceding their baseline health examination.
2.3. Statistical analysis
Alcohol consumption amount was categorized into 5 groups based on average number of drinks per day: 0 (reference), 0.1 to 1.9, 2 to 3.9, 4 to 6.9, and ≥7. Alcohol use frequency was categorized into 4 groups based on drinking days per week: <1 (reference), 1 to 2, 3 to 4, and 5 to 7. Hazard ratios for incident COVID-19 were estimated using Cox proportional hazards models.
The multivariable model was adjusted for the following covariates: age (continuous), sex, income status (quartiles: first [lowest], second, third, and fourth), smoking status (current, former, never, or missing), physical activity (none, 1–2, or ≥3 times per week of moderate activity for ≥30 minutes or vigorous activity for ≥20 minutes), hypertension status (SBP <120 mm Hg, 120–139 mm Hg, or ≥140 mm Hg or known hypertension), diabetes status (fasting glucose <100 mg/dL, 100–125 mg/dL, or ≥126 mg/dL or known diabetes), and hyperlipidemia (total cholesterol <200 mg/dL, 200–239 mg/dL, or ≥240 mg/dL or lipid-lowering medication use). Alcohol amount and frequency were modeled separately and were not included simultaneously in the same model. Sensitivity analyses were also conducted using different covariate adjustments and categorizations of alcohol consumption amount.
For trend analysis, alcohol consumption categories were treated as ordinal variables. Sex- and age-stratified subgroup analyses and trend tests were also conducted as sensitivity analyses. All p-values were 2-sided. Statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary).
2.4. Data availability
The data used in this study cannot be shared by the authors but are directly available from the NHIS (https://nhiss.nhis.or.kr/en/z/a/001/lpza001m01en.do).
3. Results
3.1. Baseline characteristics
During the 235 days of follow-up, a total of 47,308 individuals were diagnosed with COVID-19. As of January 1, 2020, the mean age of participants was 57.2 ± 13.6 years, and 49.1% were women (Table 1). At baseline health examination, the mean BMI was 23.7 ± 3.2 kg/m2. Current nonalcohol users were the oldest group, and among current alcohol users, those with higher consumption tended to be older. Alcohol users were more likely to be men and current smokers. Except for nonalcohol users, increasing alcohol use was generally associated with higher BMI, fasting glucose, total cholesterol, and systolic blood pressure levels.
Table 1.
Baseline characteristics of participants by categories of alcohol consumption amount.
| Alcohol consumption amount (drinks per day) | |||||||
|---|---|---|---|---|---|---|---|
| Noncurrent drinker | 0.1–1.9 | 2–3.9 | 4–6.9 | 7 or more | |||
| n = 15,980,069 | n = 7,986,620 | n = 5,041,691 | n = 1,747,381 | n = 875,400 | n = 328,977 | ||
| Age* | years | 57.2 ± 13.6 | 61.2 ± 13.5 | 52.7 ± 12.6 | 53.5 ± 12.0 | 53.6 ± 11.9 | 57.2 ± 12.8 |
| Body mass index | kg/m2 | 23.7 ± 3.2 | 23.7 ± 3.3 | 23.5 ± 3.2 | 24.2 ± 3.1 | 24.5 ± 3.1 | 24.6 ± 3.2 |
| Total cholesterol† | mg/dL | 195.5 ± 36.9 | 196.9 ± 37.7 | 192.7 ± 35.5 | 196.0 ± 36.3 | 197.0 ± 36.9 | 197.2 ± 38.4 |
| Fasting glucose‡ | mg/dL | 97.1 ± 22.8 | 96.9 ± 22.6 | 95.5 ± 21.0 | 99.1 ± 24.3 | 100.9 ± 26.1 | 104.3 ± 30.3 |
| Systolic blood pressure | mm Hg | 122.1 ± 15.0 | 121.6 ± 15.5 | 120.8 ± 14.3 | 124.8 ± 14.3 | 126.2 ± 14.4 | 128.0 ± 15.1 |
| Sex | Women | 7,841,137 (49.1) | 5663,624 (70.9) | 1,874,820 (37.2) | 218,232 (12.5) | 64,984 (7.4) | 19,477 (5.9) |
| Smoking status | Never smoker | 9,892,226 (61.9) | 6610,312 (82.8) | 2,623,019 (52.0) | 436,272 (25.0) | 162,553 (18.6) | 60,070 (18.3) |
| Past smoker | 2,223,425 (13.9) | 595,211 (7.5) | 916,866 (18.2) | 415,616 (23.8) | 215,969 (24.7) | 79,763 (24.2) | |
| Current smoker | 3,827,781 (24.0) | 765,668 (9.6) | 1,487,194 (29.5) | 891,494 (51.0) | 494,947 (56.5) | 188,478 (57.3) | |
| Missing | 36,637 (0.2) | 15,429 (0.2) | 14,612 (0.3) | 3999 (0.2) | 1931 (0.2) | 666 (0.2) | |
| Alcohol use, | <1 d/wk | 8,458,153 (52.9) | 7986,620 (100.0) | 471,497 (9.4) | 36 (0.0) | 0 (0.0) | 0 (0.0) |
| Frequency | 1–2 d/wk | 5,468,857 (34.2) | 0 (0.0) | 4,216,247 (83.6) | 1035,474 (59.3) | 209,265 (23.9) | 7871 (2.4) |
| 3–4 d/wk | 1,477,687 (9.2) | 0 (0.0) | 312,796 (6.2) | 585,246 (33.5) | 481,210 (55.0) | 98,435 (29.9) | |
| 5–7 d/wk | 575,372 (3.6) | 0 (0.0) | 41,151 (0.8) | 126,625 (7.2) | 184,925 (21.1) | 222,671 (67.7) | |
| Physical activity | No | 7,791,824 (48.8) | 4512,138 (56.5) | 2,046,845 (40.6) | 697,608 (39.9) | 371,970 (42.5) | 163,263 (49.6) |
| 1d/wk | 3,189,583 (20.0) | 1236,507 (15.5) | 1,283,701 (25.5) | 416,557 (23.8) | 195,140 (22.3) | 57,678 (17.5) | |
| ≥2 d/wk | 4,998,662 (31.3) | 2237,975 (28.0) | 1,711,145 (33.9) | 633,216 (36.2) | 308,290 (35.2) | 108,036 (32.8) | |
| Age group, years | 28–49 | 5,045,926 (31.6) | 1659,185 (20.8) | 2,233,563 (44.3) | 710,104 (40.6) | 347,250 (39.7) | 95,824 (29.1) |
| 50–64 | 6,336,329 (39.7) | 3156,677 (39.5) | 1,932,184 (38.3) | 729,095 (41.7) | 376,241 (43.0) | 142,132 (43.2) | |
| 65–99 | 4,597,814 (28.8) | 3170,758 (39.7) | 875,944 (17.4) | 308,182 (17.6) | 151,909 (17.4) | 91,021 (27.7) | |
| Income status, | Q1 (low-income) | 3,257,050 (20.4) | 1760,651 (22.0) | 979,774 (19.4) | 304,307 (17.4) | 149,527 (17.1) | 62,791 (19.1) |
| Quartile | Q2 | 3,394,731 (21.2) | 1603,964 (20.1) | 1,128,311 (22.4) | 385,952 (22.1) | 198,434 (22.7) | 78,070 (23.7) |
| Q3 | 4,265,763 (26.7) | 2023,228 (25.3) | 1,391,562 (27.6) | 504,715 (28.9) | 252,877 (28.9) | 93,381 (28.4) | |
| Q4 | 5,062,525 (31.7) | 2598,777 (32.5) | 1,542,044 (30.6) | 552,407 (31.6) | 274,562 (31.4) | 94,735 (28.8) | |
| Body mass index | <18.5 | 588,521 (3.7) | 313,448 (3.9) | 212,755 (4.2) | 39,878 (2.3) | 16,055 (1.8) | 6385 (1.9) |
| kg/m2 | 18.5–24.9 | 10,241,683 (64.1) | 5197,245 (65.1) | 3,325,120 (66.0) | 1044,032 (59.7) | 492,834 (56.3) | 182,452 (55.5) |
| 25–29.9 | 4,580,704 (28.7) | 2189,129 (27.4) | 1,347,688 (26.7) | 595,663 (34.1) | 325,064 (37.1) | 123,160 (37.4) | |
| ≥ 30 | 569,161 (3.6) | 286,798 (3.6) | 156,128 (3.1) | 67,808 (3.9) | 41,447 (4.7) | 16,980 (5.2) | |
| Hypertension§ | Yes | 3,580,850 (22.4) | 2017,325 (25.3) | 846,012 (16.8) | 397,617 (22.8) | 218,657 (25.0) | 101,239 (30.8) |
| Diabetes∥ | Yes | 1,366,286 (8.5) | 763,936 (9.6) | 318,940 (6.3) | 151,536 (8.7) | 87,880 (10.0) | 43,994 (13.4) |
| Hyperlipidemia¶ | Yes | 8,496,256 (53.2) | 3,996,307 (50.0) | 2,927,738 (58.1) | 939,186 (53.7) | 461,317 (52.7) | 171,708 (52.2) |
| Cardiovascular disease# | Yes | 637,933 (4.0) | 433,718 (5.4) | 131,310 (2.6) | 42,329 (2.4) | 20,788 (2.4) | 9788 (3.0) |
| Cancer# | Yes | 752,831 (4.7) | 526,163 (6.6) | 161,991 (3.2) | 39,056 (2.2) | 18,177 (2.1) | 7444 (2.3) |
Data were presented as mean ± SD or n (%).
The P-values were calculated by the chi-square test and one-way ANOVA across the alcohol amount groups; all P-values were <.001.
ANOVA = analysis of variance, ICD-10 = International Statistical Classification of Diseases and Related Health Problems, 10th Revision, n = number of participants, SD = standard deviation.
Age was calculated as of January 1, 2020.
To convert cholesterol from mg/dL to mmol/L, multiply by 0.02586.
To convert glucose from mg/dL to mmol/L, multiply by 0.0555.
Hypertension was defined as systolic blood pressure ≥140 mmHg or ≥1 clinical visit for hypertension (ICD-10: I10–I15) within 1 year before the health examination date.
Diabetes was defined as fasting glucose ≥126 mg/dL or ≥1 clinical visit for diabetes (E10–E14) within 1 year before the health examination date.
Hyperlipidemia was defined as total cholesterol ≥240 mg/dL or ≥1 prescription of statins, fibrates, ezetimibe, gemfibrozil, or omega-3-acid ethyl esters within 6 months before the health examination date.
Cardiovascular disease and cancer were defined as ≥1 clinical visit within 1 year before the health examination date for heart disease (I20–I52), stroke (I60–I69), or cancer (C00–C97).
3.2. Alcohol use and COVID-19 risk
Higher alcohol consumption was associated with an increased risk of COVID-19 (Fig. 2). Compared with current nondrinkers, individuals consuming <2, 2–<4, 4–<7, and ≥7 drinks per day had 4%, 13%, 17%, and 22% (HR = 1.22; 95% confidence interval [CI]: 1.14–1.31) higher COVID-19 risk, respectively (P for trend <.001). In the analysis of alcohol use frequency, compared with those drinking <1 day per week (including current nondrinkers), individuals consuming alcohol 1 to 2, 3 to 4, and 5 to 7 days per week had 8%, 19%, and 11% higher COVID-19 risk, respectively (P for trend <.001). In sensitivity analyses, adjustment for demographic, socioeconomic, and behavioral factors only (Model 1), or additional adjustment for high-density lipoprotein cholesterol, estimated glomerular filtration rate, and baseline cardiovascular disease and cancer (Model 2), did not substantially alter the associations (Table 2).
Figure 2.
Alcohol use and the risk of COVID-19. Higher alcohol-use amount and frequency were associated with an increased risk of COVID-19. Hazard ratios and 95% confidence intervals were calculated using Cox proportional hazards models adjusted for sex, age, household income, smoking, physical activity, hypertension status, diabetes status, and hyperlipidemia. CI = confidence interval, COVID-19 = coronavirus disease 2019, HR = hazard ratio.
Table 2.
Alcohol use and COVID-19 risk according to statistical model.
| Items |
No. of COVID-19 | Model 1 | Model 2 | ||||
|---|---|---|---|---|---|---|---|
| P-value | HR (95% CI) | P-value | HR (95% CI) | ||||
| Amount of alcohol use (drinks/day) | |||||||
| Noncurrent drinker | 19,233 | – | 1.00 | (Reference) | – | 1.00 | (Reference) |
| 0.1–1.9 | 12,263 | .012 | 1.03 | (1.01–1.06) | .003 | 1.04 | (1.01–1.06) |
| 2–3.9 | 4391 | <.001 | 1.14 | (1.10–1.18) | <.001 | 1.14 | (1.10–1.18) |
| 4–6.9 | 2238 | <.001 | 1.18 | (1.12–1.24) | <.001 | 1.17 | (1.12–1.23) |
| ≥7 | 859 | <.001 | 1.23 | (1.15–1.32) | <.001 | 1.23 | (1.15–1.32) |
| P for trend | – | – | <.001 | – | <.001 | ||
| Frequency of alcohol use (d/wk) | |||||||
| <1 | 20,189 | 1.00 | (Reference) | 1.00 | (Reference) | ||
| 1–2 | 13,568 | <.001 | 1.08 | (1.05–1.10) | <.001 | 1.08 | (1.05–1.11) |
| 3–4 | 3868 | <.001 | 1.19 | (1.14–1.23) | <.001 | 1.19 | (1.15–1.24) |
| 5–7 | 1359 | <.001 | 1.11 | (1.05–1.17) | <.001 | 1.12 | (1.06–1.19) |
| P for trend | – | – | <.001 | – | <.001 | ||
Associations were generally consistent across models, with no substantial differences observed according to the level of covariate adjustment.
HRs and 95% CIs were calculated using Cox proportional hazards models.
Model 1 was adjusted for sex, age, household income, smoking status, and physical activity.
Model 2 was further adjusted for hypertension status, diabetes status, hyperlipidemia, HDL-C, chronic kidney disease status (defined by estimated glomerular filtration rate), prevalent cardiovascular disease, and prevalent cancer.
CI = confidence interval, COVID-19 = coronavirus disease 2019, HDL-C = high-density lipoprotein cholesterol, HR = hazard ratio.
3.3. Sex- and age-stratified analysis
In each sex and age group, higher amounts of alcohol intake were consistently associated with an increased risk of COVID-19 (Fig. 3). The impact of alcohol use appeared stronger in women than in men. For example, men consuming ≥7 drinks per day had a 17% higher COVID-19 risk than nondrinking men (HR = 1.17; 95% CI: 1.09–1.26), whereas the corresponding increase in women was 36%. The age-specific effects of alcohol use differed between men and women (Fig. 3). Among men, the youngest age group (28–49 years) showed the strongest association, while the oldest group (≥65 years) had the weakest. For instance, men aged 28 to 49 consuming ≥7 drinks per day had 41% higher risk (HR = 1.41), while the increase was 6% (HR = 1.06) in men aged 65 to 99 years. In contrast, among women, the strength of the association did not decrease with age. Notably, even women consuming <1 drink per day had a higher risk of COVID-19 than nondrinkers (Fig. 4; HR = 1.06; 95% CI: 1.02–1.10). In analyses of alcohol use frequency, women again showed stronger associations than men (Fig. 5). Among men, the association between alcohol use frequency and COVID-19 risk was stronger in younger adults, whereas among women, the youngest age group exhibited the weakest association.
Figure 3.
Amount of alcohol use and the risk of COVID-19 by sex and age. The association between alcohol-use amount and COVID-19 risk appeared stronger in women than in men, with age-specific effects showing different patterns by sex. Hazard ratios and 95% confidence intervals were calculated using Cox proportional hazards models adjusted for age, household income, smoking, physical activity, hypertension status, diabetes status, and hyperlipidemia. CI = confidence interval, COVID-19 = coronavirus disease 2019, HR = hazard ratio.
Figure 4.
Amount of alcohol use and the risk of COVID-19 in women using an alternative categorization. Even women consuming <1 drink per day appeared to have a modestly higher risk of COVID-19 compared with nondrinkers. Hazard ratios and 95% confidence intervals were estimated using Cox proportional hazards models, with covariate adjustments identical to those in Figure 3. CI = confidence interval, COVID-19 = coronavirus disease 2019, HR = hazard ratio.
Figure 5.
Frequency of alcohol use and the risk of COVID-19 by sex and age. Analyses based on alcohol use frequency showed patterns similar to those observed for alcohol-use amount, with stronger associations in women than in men. Hazard ratios and 95% CIs were calculated using Cox proportional hazards models adjusted for age, household income, smoking, physical activity, hypertension status, diabetes status, and hyperlipidemia. CI = confidence interval, COVID-19 = coronavirus disease 2019, HR = hazard ratio.
4. Discussion
In our prospective cohort study, higher amounts of alcohol consumption were associated with an increased risk of COVID-19. This association was stronger in women than in men. In women, even consuming less than 1 drink per day was associated with elevated risk. The age-specific impact of alcohol use also differed by sex: among men, younger adults showed a stronger association than older adults, whereas in women, older adults exhibited a similar or even stronger association compared to younger adults. A higher frequency of alcohol consumption was likewise associated with increased COVID-19 risk, although the highest frequency was not associated with a greater risk than moderate frequency.
Higher amounts of alcohol consumption were associated with an increased risk of COVID-19 in our study. Although a relatively large number of studies have examined the association between alcohol use and COVID-19 severity among infected individuals,[20] only a few have investigated the relationship between alcohol consumption and the incidence of COVID-19 or susceptibility to SARS-CoV-2 infection. In a UK cohort study involving 16,559 COVID-19 events,[10] individuals who consumed <2, 2 to 3.9, and ≥4 units of alcohol per day had 4% (HR = 1.04; 95% CI: 0.97–1.12), 9% (HR = 1.09; 95% CI: 1.00–1.20), and 12% (HR = 1.12; 95% CI: 1.00–1.25) higher risk of COVID-19, respectively, compared to nondrinkers (Dai, 2022). In a U.S. study of college students,[11] consuming >14 drinks per week for men and >7 drinks per week for women was associated with a significantly higher risk of COVID-19 (HR = 2.32; 95% CI: 1.26–4.15), based on 42 events. The associations observed in our Korean participants appeared stronger than those reported in the UK cohort. This discrepancy may reflect differences in beverage preference: in Korea, alcohol consumption is dominated by soju and beer, whereas wine consumption is minimal. In contrast, a UK study found that wine drinkers had a lower risk of COVID-19, while beer and spirit drinkers exhibited a dose-dependent increase in risk.[10] For example, among Korean adults aged 20 years or older in 2009, the proportions consuming soju, beer, and wine (including Korean fruit wines) were 64.1%, 22.3%, and 1.5%, respectively.[21] A standard serving of soju (50 mL, with an alcohol content of approximately 20% during 2009–2010) or beer (200 mL) contains about 8 grams of pure alcohol, which is roughly equivalent to 1 UK alcohol unit. Overall, prospective studies suggest higher levels of alcohol consumption are associated with a moderately higher risk of COVID-19.
Women showed a stronger association between alcohol consumption and COVID-19 risk than men in our study. Those who consumed less than 1 drink per day had a higher risk of COVID-19 compared to nondrinkers, whereas men who consumed less than 2 drinks per day generally did not exhibit an elevated risk. Notably, in men, the strength of this association diminished with age, whereas in women, it remained consistent or even strengthened in older age groups. This sex difference may reflect selection effects among older Korean women: alcohol consumption declines sharply with age in women, resulting in a more selective group of older female drinkers. According to our cohort data, the prevalence of current alcohol use among women was 48.9% in those under 50 years, 28.1% in those aged 50 to 64 years, and only 10.6% in those aged 65 to 99 years. In contrast, the corresponding proportions in men were 79.6%, 73.0%, and 57.0%, respectively. Thus, older women who continue to drink may represent a distinct subgroup, potentially with different behavioral or biological profiles.[22] Prior literature supports greater biological sensitivity to alcohol in women. They experience stronger adverse effects at equivalent levels,[23,24] and harmful health and behavioral consequences occur sooner and at lower levels of consumption.[25] These consequences may in turn influence daily behaviors relevant to infection risk, such as mask-wearing and physical distancing.[11]
Higher frequency of alcohol consumption was associated with an increased risk of COVID-19. Our study found a consistent positive association between alcohol consumption frequency and COVID-19 risk (P for trend <.001), with all frequency categories (1–2, 3–4, and 5–7 days per week) associated with higher risk compared to nondrinkers. However, the highest frequency category (5–7 days per week) did not confer a higher risk compared to the moderate frequency category (3–4 days per week) in our study. Similarly, a UK cohort study reported that individuals consuming alcohol ≥3 times per week had a 10% lower risk of COVID-19 compared to nondrinkers (HR = 0.90; 95% CI: 0.83–0.98), while higher amounts of alcohol use were associated with higher risk.[10] The less pronounced dose–response relationship observed with alcohol consumption frequency, compared with amount, may reflect the greater importance of overall intake and drinking pattern rather than frequency alone. Chronic heavy alcohol consumption has been linked to impaired immune function and increased susceptibility to infections, whereas frequent but moderate drinking may not have the same effects as heavy episodic drinking.[26] Overall, our findings suggest that total alcohol consumption may be more relevant for COVID-19 risk than drinking frequency alone.
This study has several notable strengths that enhance the robustness and interpretability of the findings. Its large-scale population-based prospective cohort design provided high statistical power and enabled detailed subgroup analyses by sex and age, while minimizing biases inherent to retrospective studies. Nearly complete follow-up was achieved through a nationwide COVID-19 contact tracing program using RT-PCR–confirmed cases linked to a national database. In addition, dose–response relationships were examined for both the amount and frequency of alcohol consumption. Together, these features strengthen the reliability of the observed associations at the population level. Despite these strengths, several limitations should be acknowledged. First, the observational nature of the study limits causal inference. Second, alcohol consumption was assessed only once, more than 10 years before the index date, and changes in drinking behavior over time were not captured. This may have led to exposure misclassification and potential underestimation of the true association, particularly among heavy drinkers.[27] Although certain occupations have been associated with higher COVID-19 risk,[5,28] occupational status was not available in this study. Residual confounding by socioeconomic and behavioral factors also cannot be excluded, particularly given the 10-year time lag. The study population consisted almost entirely of Koreans, which may limit the generalizability of the findings. However, the consistent associations observed across sex and age groups may support the broader applicability of the results.
In conclusion, greater alcohol consumption – both in amount and frequency – was associated with higher risk of COVID-19, although the effect size was modest. The association appeared stronger in women than in men, and age-specific associations differed between the sexes. Further studies are needed to examine whether sex- and age-specific public health guidance on alcohol consumption can reduce COVID-19 transmission.
Acknowledgments
This research used data provided by the Korea Disease Control and Prevention Agency and National Health Insurance Service of Korea (KDCA-NHIS-2025-04-1-012).
Author contributions
Conceptualization: Daeho Kwon, Sang-Wook Yi.
Data curation: Sang-Wook Yi.
Formal analysis: Daeho Kwon, Sang-Wook Yi.
Methodology: Sang-Wook Yi.
Project administration: Daeho Kwon.
Visualization: Sang-Wook Yi.
Writing – original draft: Daeho Kwon, Sang-Wook Yi.
Writing – review & editing: Daeho Kwon, Sang-Wook Yi.
Abbreviations:
- BMI
- body mass index
- COVID-19
- coronavirus disease 2019
- ICD-10
- International Statistical Classification of Diseases and Related Health Problems, 10th Revision
- KDCA
- Korea Disease Control and Prevention Agency
- NHIS
- National Health Insurance Service
- RT-PCR
- reverse transcription polymerase chain reaction
- SBP
- systolic blood pressure
The study protocol was approved by the Institutional Review Board of Catholic Kwandong University (CKU-24-01-1308). Informed consent was waived because the study used de-identified data from the Korean National Health Insurance Service (NHIS). All analyses were performed within the secure NHIS analysis environment, and only aggregate statistical results approved through NHIS review could be exported. No individual participants were identifiable.
The authors have no funding and conflicts of interest to disclose.
The data that support the findings of this study are available from a third party, but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are available from the authors upon reasonable request and with permission of the third party.
How to cite this article: Kwon D, Yi S-W. Alcohol consumption and the risk of coronavirus disease 2019: A prospective cohort study. Medicine 2026;105:8(e47839).
References
- [1].Chung YS, Lam CY, Tan PH, Tsang HF, Wong SC. Comprehensive review of COVID-19: epidemiology, pathogenesis, advancement in diagnostic and detection techniques, and post-pandemic treatment strategies. Int J Mol Sci. 2024;25:8155. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Bartha I, Maher C, Lavrenko V, et al. Morbidity of SARS-CoV-2 in the evolution to endemicity and in comparison with influenza. Commun Med (Lond). 2024;4:244. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].Saha A, Roy SG, Dwivedi R, et al. Beyond the pandemic era: recent advances and efficacy of SARS-CoV-2 vaccines against emerging variants of concern. Vaccines (Basel). 2025;13:424. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Hancean MG, Ghita MC, Perc M, et al. Disaggregated data on age and sex for the first 250 days of the COVID-19 pandemic in Bucharest, Romania. Sci Data. 2022;9:253. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Hancean MG, Lerner J, Perc M, et al. Occupations and their impact on the spreading of COVID-19 in urban communities. Sci Rep. 2022;12:14115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Molina PE, Happel KI, Zhang P, Kolls JK, Nelson S. Focus on: alcohol and the immune system. Alcohol Res Health. 2010;33:97–108. [PMC free article] [PubMed] [Google Scholar]
- [7].Morojele NK, Shenoi SV, Shuper PA, Braithwaite RS, Rehm J. Alcohol use and the risk of communicable diseases. Nutrients. 2021;13:3317. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Friske MM, Spanagel R. Chronic alcohol consumption and COVID-19 infection risk: a narrative review. Alcohol Clin Exp Res (Hoboken). 2023;47:629–39. [DOI] [PubMed] [Google Scholar]
- [9].Hamer M, Kivimaki M, Gale CR, Batty GD. Lifestyle risk factors, inflammatory mechanisms, and COVID-19 hospitalization: a community-based cohort study of 387,109 adults in UK. Brain Behav Immun. 2020;87:184–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [10].Dai XJ, Tan L, Ren L, Shao Y, Tao W, Wang Y. COVID-19 risk appears to vary across different alcoholic beverages. Front Nutr. 2022;8:772700. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Kianersi S, Ludema C, Macy JT, Chen C, Rosenberg M. Relationship between high-risk alcohol consumption and severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) seroconversion: a prospective sero-epidemiological cohort study among American college students. Addiction. 2022;117:1908–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Seong SC, Kim YY, Khang YH, et al. Data resource profile: the national health information database of the national health insurance service in South Korea. Int J Epidemiol. 2017;46:799–800. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Lee H, Choi S, Eom E, Song KH, Kim SS. Introduction to the research results of coronavirus disease 2019 Big Data (K-COV-N) and the corresponding utilization plan. Public Health Wkly Rep. 2024;17:2147–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Ahn S, Jang J, Park S, et al. Outbreak report of COVID-19 during designation of class 1 infectious disease in the Republic of Korea (January 20, 2020 and April 24, 2022). Public Health Wkly Rep. 2022;15:1759–72. [Google Scholar]
- [15].Kim JY, Lee YM, Lee H, Kim JW, Kim SW. Epidemiological characteristics of a COVID-19 outbreak caused by religious activities in Daegu, Korea. Epidemiol Health. 2021;43:e2021024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Kwon SL, Oh J. COVID-19 vaccination program in South Korea: a long journey toward a new normal. Health Policy Technol. 2022;11:100601. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [17].Park Y, Huh IS, Lee J, et al. ; Seoul Metropolitan Government COVID-19 Rapid Response (SCoRR) Team. Application of testing-tracing-treatment strategy in response to the COVID-19 outbreak in Seoul, Korea. J Korean Med Sci. 2020;35:e396. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Yi SW, Ohrr H, Shin SA, Yi JJ. Sex-age-specific association of body mass index with all-cause mortality among 12.8 million Korean adults: a prospective cohort study. Int J Epidemiol. 2015;44:1696–705. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Yi SW, Kim SH, Han KJ, Yi JJ, Ohrr H. Higher cholesterol levels, not statin use, are associated with a lower risk of hepatocellular carcinoma. Br J Cancer. 2020;122:630–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].Wei B, Liu Y, Li H, Peng Y, Luo Z. Impact of alcohol consumption on coronavirus disease 2019 severity: a systematic review and meta-analysis. J Med Virol. 2023;95:e28547. [DOI] [PubMed] [Google Scholar]
- [21].Korea Health Promotion Institute. Alcohol Statistics 2022 [in Korean]. Korea Health Promotion Institute; 2022. [Google Scholar]
- [22].Erol A, Karpyak VM. Sex and gender-related differences in alcohol use and its consequences: contemporary knowledge and future research considerations. Drug Alcohol Depend. 2015;156:1–13. [DOI] [PubMed] [Google Scholar]
- [23].Mumenthaler MS, Taylor JL, O’Hara R, Yesavage JA. Gender differences in moderate drinking effects. Alcohol Res Health. 1999;23:55–64. [PMC free article] [PubMed] [Google Scholar]
- [24].Miller MA, Weafer J, Fillmore MT. Gender differences in alcohol impairment of simulated driving performance and driving-related skills. Alcohol Alcohol. 2009;44:586–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].McCaul ME, Roach D, Hasin DS, Weisner C, Chang G, Sinha R. Alcohol and women: a brief overview. Alcohol Clin Exp Res. 2019;43:774–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [26].Szabo G, Mandrekar P. A recent perspective on alcohol, immunity, and host defense. Alcohol Clin Exp Res. 2009;33:220–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].Jayasekara H, English DR, Room R, MacInnis RJ. Alcohol consumption over time and risk of death: a systematic review and meta-analysis. Am J Epidemiol. 2014;179:1049–59. [DOI] [PubMed] [Google Scholar]
- [28].Gaffney A, Himmelstein DU, McCormick D, Woolhandler S. COVID-19 risk by Workers’ occupation and industry in the United States, 2020–2021. Am J Public Health. 2023;113:647–56. [DOI] [PMC free article] [PubMed] [Google Scholar]