Abstract
Background
Alcohol use and cardiovascular disease (CVD) are major public health concerns, but their co-occurrence over time remains unclear.
Objectives
The objective of the study was to examine trends in the co-occurrence of alcohol use and CVD among U.S. adults.
Methods
We analyzed 43,465 U.S. adults from 11 cycles of the National Health and Nutrition Examination Survey (1999-2023). Alcohol use was categorized as no drinking (0 drinks/week), low-to-moderate drinking (≤7 drinks/week for women; ≤14 for men), and heavy drinking (>7 for women; >14 for men). Joinpoint regression estimated trends in co-occurrence of alcohol use and overall and type-specific CVD, stratified by age, sex, race/ethnicity, poverty-income ratio and educational attainment. Trends were quantified using annual percentage change (APC).
Results
Among 43,465 adults, 5,180 had CVD. The prevalence of heavy drinking among those with CVD increased from 17.1% in 1999 to 2000 to 33.4% in 2021 to 2023, whereas the prevalence of nondrinkers and low-to-moderate drinkers among participants with CVD declined. In overall population, the prevalence of co-occurrence of heavy drinking and CVD increased from 1.4% to 3.0% (APC: 3.4%; 95% CI: 2.0%-5.1%). Increases appeared more pronounced in adults aged 40 to 59 years (post-2007), women, non-Hispanic Whites, and those with lower poverty-income ratio or educational attainment. The co-occurrence of heavy drinking with heart failure (APC: 3.8%), angina (1.7%), and stroke (5.2%) increased. In contrast, nondrinking with CVD declined (−1.4%), and low-to-moderate drinking with CVD remained stable.
Conclusions
From 1999 to 2023, the co-occurrence of heavy alcohol use and CVD increased substantially, especially among certain demographic groups.
Key words: alcohol use, cardiovascular disease, public health, trends
Central Illustration
Alcohol consumption and cardiovascular disease (CVD) remain 2 of the most pressing public health challenges in the United States.1, 2, 3, 4, 5 Heavy alcohol use accounted for an estimated 13% of adult deaths between 2015 and 2019,1,2 whereas CVD has consistently ranked as the leading cause of death nationwide over the past 2 decades.3, 4, 5 Despite these potential substantial health burdens, alcohol use, including among individuals with diagnosed CVD, remains common.6 Historically, this has been partially influenced by observational evidence suggesting that low-to-moderate alcohol consumption might confer cardioprotective effects.7 However, even modest drinking can easily progress to heavier use, significantly increasing the risk of adverse health outcomes.8
Recent guidelines, including those from the World Health Organization, state that no level of alcohol intake is safe.9 Emerging evidence increasingly challenges the notion of cardioprotective effects from low-to-moderate drinking, indicating that even modest intake may exacerbate cardiovascular conditions and elevate the risk of other diseases, including cancer.10,11 For example, individuals with atrial fibrillation who are continuously monitored demonstrate an increased incidence of arrhythmia episodes shortly after alcohol intake.10 These findings raise serious concerns for individuals with existing CVD, who may face heightened risks from continued alcohol consumption.
Despite decades of national surveillance tracking alcohol use and CVD separately,5,12 little is known about long-term trends in the co-occurrence of CVD and alcohol use, especially heavy drinking. This knowledge gap has limited the ability of clinicians, health care systems, and policymakers to identify and respond to a potentially rapidly growing high-risk population.
To address this gap, we analyzed long-term trends in the co-occurrence of alcohol consumption and CVD among U.S. adults from 1999 to 2023, using data from 11 cycles of the National Health and Nutrition Examination Survey (NHANES). We further examined these trends across specific CVD types, including coronary heart disease (CHD), angina, heart attack, heart failure, and stroke, as well as across key sociodemographic subgroups, such as age, sex, race/ethnicity, poverty income ratio (PIR), and educational level. These findings may provide crucial evidence to guide clinical practice and inform public health interventions aimed at addressing the dual burden of alcohol-related harm and CVD.
Methods
Study design and participants
The NHANES is a continuous, nationally representative, serial cross-sectional survey of the civilian, noninstitutionalized U.S. population.13 It combines in-person household interviews with standardized physical examinations conducted in mobile examination centers (MECs). Ethics approval for NHANES was granted by the National Center for Health Statistics Research Ethics Review Board, and all adult participants provided written informed consent.13
Analyzed population
For this current study, we used data from 11 consecutive NHANES cycles: 1999 to 2000, 2001 to 2002, 2003 to 2004, 2005 to 2006, 2007 to 2008, 2009 to 2010, 2011 to 2012, 2013 to 2014, 2015 to 2016, 2017 to 2020, and 2021 to 2023.14 Among 119,555 participants across these cycles, we excluded individuals who were aged <20 years, and those with missing data on CVD status or alcohol use. We also excluded pregnant women and participants categorized as “other race/ethnicity,” including non-Hispanic Asians and multiracial individuals, due to small sample sizes and inconsistent racial/ethnic categorization across survey cycles. Finally, the final analytic sample included 43,465 participants with information on alcohol use, CVD status, age, sex, race/ethnicity, PIR, educational level, representing the U.S. civilian, noninstitutionalized adult population (Supplemental Figure 1).
Alcohol consumption
As part of the MEC protocol, we collected data on the current alcohol use (defined as alcohol consumption over the past 12 months) from respondents aged ≥20 years during 1999 to 2010 and ≥ 18 years during 2011 to 2023 using the Alcohol Use Questionnaire.14 According to the National Institute on Alcohol Abuse and Alcoholism guidelines, alcohol use was categorized as follows: nondrinkers (0 drinks/week for both men and women), low-to-moderate drinkers (≤14 drinks/week for men and ≤7 drinks/week for women), and heavy drinkers (>14 drinks/week for men and >7 drinks/week for women).15
Covariates
Information on age, sex, race/ethnicity, and educational level was obtained through self-reported questionnaires during the household interview. Family income was also self-reported during the household interview. The PIR was calculated by NHANES as the ratio of family income to the federal poverty level, adjusted for household size and survey year.16
Cardiovascular disease
CVD cases were identified based on participants’ self-reports of doctor-diagnosed conditions, including CHD, heart attack, angina, heart failure, and stroke.17 Participants answered whether a health professional had ever told them they had any of these conditions. We classified those who answered “yes” as having CVD.17
Statistical analysis
We applied sample weights from the National Center for Health Statistics to account for NHANES’ complex survey design and oversampling, ensuring nationally representative estimates. First, we examined secular trends in alcohol use categories (nondrinkers, low-to-moderate drinkers, and heavy drinkers) among participants with CVD. Then, we assessed trends in the co-occurrence of alcohol use and CVD across the full analytic sample.
We used Joinpoint regression analysis to assess temporal trends and identify any inflection points (ie, Join points) where statistically significant changes in trend occurred. This method, originally developed by Kim et al.,18 is widely used to evaluate temporal trends in epidemiological data such as disease prevalence. Using prevalence estimates as input, Joinpoint regression detects the years when significant trend changes start and end, and calculates the annual percentage change (APC) between each pair of joinpoints. It also calculates the average annual percent change (AAPC) over the entire study period as a summary measure. We estimated APC by applying least squares regression to the natural logarithm of weighted crude prevalence estimates.19 The model allowed up to 2 joinpoints, requiring at least 4 observations between joinpoints. The optimal number of joinpoints was determined using Monte Carlo permutation tests.20 Analyses were conducted using Joinpoint Trend Analysis Software (version 5.4.0). When no joinpoints were identified, trends were summarized using a single APC. When one or more joinpoints were detected, both segment-specific APCs and the AAPC were reported to describe overall trends across multiple segments.18 The APC was estimated overall and stratified by age, sex, race/ethnicity, PIR, and educational level to explore subgroup differences. We conducted Joinpoint regression with parallelism tests to evaluate whether the slopes (APCs) differed significantly between subgroups.15 We also examined trends in the prevalence of the co-occurrence of alcohol use with specific CVD outcomes, including CHD, heart attack, angina, heart failure, and stroke.
Sensitivity and exploratory analysis
We conducted sensitivity analysis to identify robustness of these results. First, to evaluate whether the trends in the co-occurrence of alcohol use and CVD were influenced by age, sex, and race/ethnicity, we fitted survey-weighted logistic regression models including these variables as covariates to estimate adjusted predicted prevalence for each NHANES cycle. The resulting adjusted prevalence estimates were then entered into Joinpoint regression models to examine temporal trends. Second, we repeated the analyses by excluding the 2017 to 2020 and 2021 to 2023 cycles to recalculate APCs. Finally, in exploratory analyses, we examined secular trends in alcohol use among participants without CVD, as well as trends in the co-occurrence of alcohol use and non-CVD conditions, stratified by age, sex, race/ethnicity, PIR, and educational level.
All analyses were conducted using R (version 4.4.1). and Joinpoint Trend Analysis Software (version 5.4.0). Statistical significance was set at P < 0.05 for all 2-sided tests.
Results
Baseline characteristics
Table 1 summarizes the weighted sociodemographic characteristics of participants across NHANES cycles from 1999 to 2023. A total of 43,465 adults were included (22,928 men, mean age 51.0 years; and 20,537 women, mean age 50.5 years), of whom 5,180 reported having CVD. Distributions of age, sex, race/ethnicity, PIR, and educational level varied across the 11 survey cycles. For instance, based on weighted estimates, the proportion of adults aged 20 to 39 years declined from 41.0% in 1999 to 2000 to 32.7% in 2021 to 2023; 40 to 59 years decreased from 37.2% to 34.2%; and ≥60 years increased from 21.8% to 33.1%. The proportion of Hispanics ranged from 10.8% to 17.3% and non-Hispanic Blacks from 9.7% to 12.1% across cycles. Supplemental Table 1 presents unweighted sociodemographic characteristics and Supplemental Table 2 shows detailed subgroup distributions of age, sex, PIR, and educational level across race/ethnicity.
Table 1.
Weighted Participant Sociodemographic Characteristics by NHANES Cycles
| Weighted, n (%) by NHANES Cycle |
|||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1999–2000 (n = 11,553,595) | 2001–2002 (n = 12,953,081) | 2003–2004 (n = 12,589,034) | 2005–2006 (n = 13,411,687) | 2007–2008 (n = 13,593,566) | 2009–2010 (N = 13,692,903) | 2011–2012 (N = 13,834,218) | 2013–2014 (N=14,394,356) | 2015–2016 (N = 14,235,428) | 2017–2020 (N = 25,204,162) | 2021–2023 (N = 11,523,025) | |
| Age group (years) | |||||||||||
| 20-39 | 4,740,734 (41.0) | 4,654,694 (35.9) | 4,673,612 (37.1) | 4,748,038 (35.4) | 5,071,920 (37.3) | 4,691,887 (34.3) | 4,792,633 (34.6) | 4,877,165 (33.9) | 4,732,286 (33.2) | 8,870,691 (35.2) | 3,773,050 (32.7) |
| 40-59 | 4,293,025 (37.2) | 5,721,262 (44.2) | 5,205,730 (41.4) | 5,659,881 (42.2) | 5,512,253 (40.6) | 5,583,607 (40.8) | 5,404,133 (39.1) | 5,450,877 (37.9) | 5,338,827 (37.5) | 8,829,134 (35.0) | 3,937,626 (34.2) |
| ≥60 | 2,519,837 (21.8) | 2,577,124 (19.9) | 2,709,693 (21.5) | 3,003,767 (22.4) | 3,009,393 (22.1) | 3,417,409 (25.0) | 3,637,452 (26.3) | 4,066,315 (28.2) | 4,164,316 (29.3) | 7,504,337 (29.8) | 3,812,349 (33.1) |
| Sex | |||||||||||
| Men | 6,110,101 (52.9) | 6,855,766 (52.9) | 6,519,242 (51.8) | 7,018,137 (52.3) | 6,934,968 (51.0) | 7,136,488 (52.1) | 7,180,906 (51.9) | 7,369,040 (51.2) | 7,220,060 (50.7) | 12,275,455 (48.7) | 5,783,274 (50.2) |
| Women | 5,443,494 (47.1) | 6,097,315 (47.1) | 6,069,792 (48.2) | 6,393,550 (47.7) | 6,658,598 (49.0) | 6,556,415 (47.9) | 6,653,312 (48.1) | 7,025,316 (48.8) | 7,015,368 (49.3) | 12,928,707 (51.3) | 5,739,750 (49.8) |
| Race/ethnicity | |||||||||||
| Hispanics | 1,763,574 (15.3) | 1,627,283 (12.6) | 1,424,718 (11.3) | 1,451,965 (10.8) | 1,809,570 (13.3) | 1,838,063 (13.4) | 1,884,588 (13.6) | 2,143,243 (14.9) | 2,078,598 (14.6) | 4,368,561 (17.3) | 1,944,124 (16.9) |
| Non-Hispanic Whites | 8,615,880 (74.6) | 10,068,521 (77.7) | 9,848,833 (78.2) | 10,566,537 (78.8) | 10,240,411 (75.3) | 10,362,867 (75.7) | 10,390,544 (75.1) | 10,666,216 (74.1) | 10,515,895 (73.9) | 17,780,695 (70.5) | 8,335,618 (72.3) |
| Non-Hispanic Blacks | 1,174,141 (10.2) | 1,257,277 (9.7) | 1,315,483 (10.4) | 1,393,185 (10.4) | 1,543,585 (11.4) | 1,491,973 (10.9) | 1,559,086 (11.3) | 1,584,897 (11.0) | 1,640,935 (11.5) | 3,054,906 (12.1) | 1,243,283 (10.8) |
| PIR | |||||||||||
| <3 | 5,809,405 (50.3) | 5,879,282 (45.4) | 6,090,359 (48.4) | 5,856,958 (43.7) | 6,617,746 (48.7) | 6,527,431 (47.7) | 7,013,141 (50.7) | 7,051,564 (49.0) | 6,898,292 (48.5) | 11,568,031 (45.9) | 5,093,380 (44.2) |
| ≥3 | 5,744,190 (49.7) | 7,073,799 (54.6) | 6,498,675 (51.6) | 7,554,729 (56.3) | 6,975,820 (51.3) | 7,165,472 (52.3) | 6,821,077 (49.3) | 7,342,792 (51.0) | 7,337,136 (51.5) | 13,636,131 (54.1) | 6,429,645 (55.8) |
| Educational level | |||||||||||
| ≤ High school | 5,617,684 (48.6) | 5,611,006 (43.3) | 5,458,625 (43.4) | 5,622,710 (41.9) | 5,965,561 (43.9) | 5,636,822 (41.2) | 4,735,116 (34.2) | 5,159,543 (35.8) | 4,823,485 (33.9) | 9,433,802 (37.4) | 3,570,499 (31.0) |
| > High school | 5,935,911 (51.4) | 7,342,075 (56.7) | 7,130,409 (56.6) | 7,788,977 (58.1) | 7,628,006 (56.1) | 8,056,081 (58.8) | 9,099,102 (65.8) | 9,234,813 (64.2) | 9,411,943 (66.1) | 15,770,360 (62.6) | 7,952,526 (69.0) |
| CVD status | |||||||||||
| Living with CVD | 928,050 (8.0) | 1,052,415 (8.1) | 1,145,860 (9.1) | 1,169,640 (8.7) | 1,069,739 (7.9) | 1,132,389 (8.3) | 1,161,124 (8.4) | 1,316,638 (9.1) | 1,255,444 (8.8) | 2,534,480 (10.1) | 1,041,024 (9.0) |
| Living without CVD | 10,625,546 (92.0) | 11,900,665 (91.9) | 11,443,175 (90.9) | 12,242,047 (91.3) | 12,523,827 (92.1) | 12,560,514 (91.7) | 12,673,094 (91.6) | 13,077,719 (90.9) | 12,979,984 (91.2) | 22,669,682 (89.9) | 10,482,001 (91.0) |
NHANES survey weights were adjusted to generate nationally representative percentages.
Data were the weighted number of participants and percentages.
CVD = cardiovascular disease; PIR = poverty income ratio.
Trends in alcohol use among adults with CVD (1999-2023)
Among 5,180 participants with CVD, the prevalence of heavy drinking increased significantly over time, rising from 17.1% (95% CI: 10.2%, 23.9%) in 1999 to 2000 to 33.4% (95%CI: 26.9%, 39.9%) in 2021 to 2023. In contrast, the prevalence of low-to-moderate drinking in participants with CVD decreased from 46.7% to 38.7%, and nondrinking from 36.2% to 27.9% over the same period (Supplemental Table 3).
APC of heavy drinking among adults with CVD
Joinpoint regression analysis showed a significant upward trend in heavy drinking among adults with CVD, with an APC of 2.8% (95% CI: 1.8%-4.0%) from 1999 to 2023 (Figure 1, Supplemental Table 4). Subgroup analyses showed variation in APCs of heavy drinking among participants with CVD across age, sex, race/ethnicity, PIR, and educational levels. Specifically, AAPC was 5.0% (95% CI: 1.0%-10.0%) in younger adults aged 20 to 39 years, APCs were 5.0% (95% CI: 1.6%-8.6%) in those aged ≥60 years, and 2.4% (95% CI: 1.0%-4.1%) in individuals aged 40 to 59 years. Relatively higher APCs were also observed in non-Hispanic White adults (3.1%; 95% CI: 1.3%-5.5%), women (3.8%; 95% CI: 1.8%-6.9%), participants with PIR <3 (4.1%; 95% CI: 2.5%-6.2%), and those with lower educational levels (3.4%; 95% CI: 2.1%-5.1%). However, formal comparisons between subgroups using Joinpoint regression parallelism tests did not reach statistical significance (Supplemental Table 4).
Figure 1.
Alcohol Use and Annual Percentage Change Among Adults With/Without Cardiovascular Disease
Weighted prevalence of alcohol use and corresponding APC from 1999 to 2023 among U.S. adults with and without CVD. (A) APC for the prevalence of heavy drinking among CVD and non-CVD participants was 2.8% (1.8%, 4.0%) and 0.4% (0.2%, 0.7%), respectively. (B) APC for the prevalence of low-to-moderate drinking among CVD and non-CVD participants was 0.0% (−1.0%, 1.1%) and 0.2% (−0.1%, 0.5%), respectively. (C) APC for the prevalence of nondrinking among CVD and non-CVD participants was −1.8% (−2.9%, −0.4%) and −1.4% (−2.3%, −0.4%), respectively. CVD = cardiovascular disease.
APC of low-to-moderate drinking and nondrinking among adults with CVD
As shown in Figure 1 and Supplemental Table 5, no significant annual change of low-to-moderate alcohol use was observed in participants with CVD from 1999 to 2023. However, subgroup analysis revealed a significant annual increase of low-to-moderate among non-Hispanic Black adults with CVD (APC: 2.0%; 95% CI: 0.6%-3.8%) (Supplemental Table 5). In contrast, a significant annual decrease of nondrinking was observed over the study period in CVD populations (APC: −1.8%; 95% CI: −2.9% to −0.4%) (Figure 1, Supplemental Table 6). Supplemental Table 6 shows largely consistent downward patterns across age, sex, race/ethnicity, PIR, and educational-level subgroups. Notably, 1 subgroup aged 20 to 39 years experienced a sharp annual decline of −9.6% (95% CI: −16.8% to −4.4%).
Joint trends of combined alcohol consumption and CVD (1999-2023)
Joint trends of combined heavy alcohol consumption and CVD
The overall weighted prevalence of adults with both diagnosed CVD and heavy alcohol use increased from 1.4% (95% CI: 0.8%-1.9%) in 1999 to 2000 to 3.0% (95% CI: 2.2%-3.8%) in 2021 to 2023, representing an approximate 114% relative increase (Figure 2).
Figure 2.
Co-occurrence of Alcohol Consumption and Cardiovascular Disease
Weighted prevalence of co-occurrence of alcohol consumption and CVD among U.S. adults from 1999 to 2023. Abbreviation as in Figure 1.
Joinpoint regression analysis confirmed a significant upward trend, with the prevalence of the co-occurrence of heavy drinking and CVD rising by 3.4% per year (APC: 3.4%; 95% CI: 2.0%-5.1%) between 1999 and 2023 (Table 2). Subgroup analyses suggested distinct patterns. Among adults aged 40 to 59 years, a significant annual increase of 4.9% (APC: 4.9%; 95% CI: 2.9%-9.6%) after 2007 followed a preceding period of nonsignificant decline. APCs were 4.2% (95% CI: 1.6%-7.8%) in women and 2.8% (95% CI: 0.1%-5.3%) in men; 3.8% (95% CI: 1.0%-7.0%) in non-Hispanic White adults, 4.0% (95% CI: −1.5% to 10.7%) in Hispanic adults, and 1.9% (95% CI: −0.8% to 5.1%) in non-Hispanic Black adults; and 4.5% (95% CI: 3.0%-6.1%) in participants with lower PIR and 1.3% (95% CI: −3.0% to 5.8%) in those with higher PIR (Table 2). In addition, annual increases were observed in both ≤ high school and >high school education, with a relatively higher APC in the ≤high school group (APC: 4.0%; 95% CI: 2.4%-5.9%) (Table 2).
Table 2.
Weighted Prevalence of the Co-occurrence of CVD and Heavy Drinking and APC Overall and by Age Group, Sex, Race/Ethnicity, PIR, and Educational-Level Group: NHANES 1999–2023
| Weighted Prevalence, % | Joinpoint Segment |
APC, % | P Value | ||
|---|---|---|---|---|---|
| Year Start | Year End | ||||
| Overall | 2.0 (1.8-2.2) | 1999 | 2023 | 3.4 (2.0-5.1) | <0.01 |
| Age group (years) | |||||
| 20-39 | 0.6 (0.5-0.7) | 1999 | 2023 | 2.7 (−6.9 to 16.5) | 0.47 |
| 40-59a | 2.3 (2.0-2.6) | 1999 | 2007 | −3.6 (−12.4 to 0.5) | 0.09 |
| 2007 | 2023 | 4.9 (2.9-9.6) | <0.01 | ||
| ≥60 | 3.5 (3.0-4.0) | 1999 | 2023 | 3.6 (−0.1, 7.5) | 0.05 |
| Sex | |||||
| Men | 2.0 (1.8-2.3) | 1999 | 2023 | 2.8 (0.1-5.3) | 0.05 |
| Women | 2.0 (1.8-2.3) | 1999 | 2023 | 4.2 (1.6-7.8) | <0.01 |
| Race/ethnicity | |||||
| Hispanics | 1.7 (1.4-1.9) | 1999 | 2023 | 4.0 (−1.5 to 10.7) | 0.13 |
| NHW | 2.0 (1.8-2.2) | 1999 | 2023 | 3.8 (1.0-7.0) | 0.01 |
| NHB | 2.6 (2.3-2.9) | 1999 | 2023 | 1.9 (−0.8 to 5.1) | 0.14 |
| PIR | |||||
| <3 | 2.7 (2.4-3.0) | 1999 | 2023 | 4.5 (3.0-6.1) | <0.01 |
| ≥3 | 1.4 (1.2-1.6) | 1999 | 2023 | 1.3 (−3.0 to 5.8) | 0.53 |
| Educational level | |||||
| ≤ High school | 2.8 (2.4-3.1) | 1999 | 2023 | 4.0 (2.4-5.9) | <0.01 |
| > High school | 1.5 (1.4-1.7) | 1999 | 2023 | 3.7 (2.0-5.7) | <0.01 |
Parallelism tests were used to evaluate whether the slopes (APCs) differed significantly between subgroups, and we did not observe significant differences by age, sex, race/ethnicity, PIR, and educational level.
APC = annual percentage change; NHANES = National Health and Nutrition Examination Survey; NHB = non-Hispanic Black; NHW = non-Hispanic White; other abbreviations as in Table 1.
Average annual percentage change (AAPC) for combined heavy drinking and CVD aged 40 to 59 years from 1999 to 2023: 1.7% (95% CI: 0.4%-3.2%).
Joint trends of combined low-to-moderate and no alcohol consumption and CVD
The overall weighted prevalence of adults with both diagnosed CVD and low-to-moderate alcohol use decreased slightly from 3.8% in 1999 to 2000 to 3.5% in 2021 to 2023 (Figure 2), but with no statistically significant annual change (APC: 0.6%; 95% CI: −1.1% to 2.5%) (Supplemental Table 7). However, subgroup analyses revealed significant upward trends by sex and race/ethnicity. Among men, the prevalence of the combined patterns of low-to-moderate and CVD increased significantly (APC: 1.2%; 95% CI: 0.3%-2.2%), as it did among non-Hispanic Black adults (APC: 2.6%; 95% CI: 1.1%-4.8%) (Supplemental Table 7). In contrast, the prevalence of adults with diagnosed CVD and no alcohol use declined from 2.9% in 1999 to 2000 to 2.5% in 2021 to 2023, with a significant annual decrease (APC: −1.4%; 95% CI: −2.3% to −0.2%). This downward trend appeared more pronounced among adults aged 20 to 39 years, who experienced an annual decline of −11.4% (95% CI: −16.2% to −7.0%), and more modest in adults aged 40 to 59 years and those aged ≥60 years (Supplemental Table 7).
Joint trends of combined alcohol consumption with CHD, heart failure, angina, heart attack, and stroke
Figure 3 and Supplemental Tables 8 and 9 present trends in the prevalence of the co-occurrence of alcohol consumption and specific CVD, including CHD, heart failure, angina, heart attack, and stroke. Significant annual increases were observed in the prevalence of adults with both heavy drinking and heart failure (APC: 3.8%; 95% CI: 0.1%-7.8%), angina (APC: 1.7%; 95% CI: 0.3%-3.4%), and stroke (APC: 5.2%; 95% CI: 2.8%-7.9%). The prevalence of adults with both low-to-moderate drinking and CHD increased (APC: 1.3%; 95%CI: 0.3%-2.3%) and stroke increased (APC: 1.7%; 95% CI: 0.2%-3.6%). In contrast, significant decrease was observed in the prevalence of the co-occurring low-to-moderate and heart attack (AAPC: −2.1%; 95% CI: −3.0% to −1.0%). Similarly, downward trends were seen for the co-occurring nondrinking and angina (APC: −3.8%; 95% CI: −6.1% to −1.3%), and heart attack (APC: −2.6%; 95% CI: −4.5% to −0.6%).
Figure 3.
APC of Co-occurrence of Alcohol Consumption and Specific Cardiovascular Disease
(A) APC for the co-occurrence of heavy drinking and overall and specific CVD. (B) APC for the co-occurrence of low-to-moderate drinking and overall and specific CVD. (C) APC for the co-occurrence of nondrinking and overall and specific CVD. APC = annual percentage change; CHD = coronary heart disease; other abbreviations as in Figure 1.
Sensitivity and exploratory analysis results
Adjusting for age, sex, and race/ethnicity did not substantially change the direction or magnitude of the trends of co-occurrence of alcohol use and CVD among overall participants or alcohol use in participants with CVD (Supplemental Table 10). For instance, the adjusted APC and 95% CI of co-occurrence of heavy, low-to-moderate, and nondrinking patterns and CVD were 2.6% (1.1%, 4.3%), −0.3% (−1.9%, 1.2%), and −2.3% (−3.2%, −1.1%), respectively. Results remained consistent when excluding the 2017 to 2020 and 2021 to 2023 cycles. Specifically, from 1999 to 2016, the prevalence of the combined patterns of heavy drinking and CVD increased by 2.2% annually (APC: 2.2%; 95% CI: 0.6%-4.1%), whereas low-to-moderate drinking and nondrinking with CVD remained stable during the same period (APC: 0.8%; 95% CI: −1.1% to 3.1%; APC: −1.0%; 95% CI: −3.0% to 1.9%).
Among participants without CVD (Supplemental Tables 3-6), the prevalence of heavy drinking slightly increased from 43.8% in 1999 to 2000 to 47.8% in 2021 to 2023 (APC: 0.4%; 95% CI: 0.2%-0.7%), substantially lower than that in individuals with CVD. Low-to-moderate drinking among participants without CVD remained relatively stable. Nondrinking in non-CVD group decreased significantly (APC: −1.4%; 95% CI: −2.3% to −0.4%).
Discussion
Main findings
In this nationally representative sample of U.S. adults, among participants with CVD, the prevalence of heavy drinking increased from 17.1% to 33.4% between 1999 and 2023, whereas the prevalence of both low-to-moderate and no alcohol consumption declined over the same period. Across the overall population, the proportion of adults with both CVD and heavy drinking notably rose from 1.4% to 3.0% (3.4% annual increase). Descriptive trends suggested higher increases among adults aged 40 to 59 years (post-2007), women, non-Hispanic White individuals, and populations with lower PIR and educational level, although parallelism tests were not statistically significant. Co-occurrence of heavy drinking with heart failure, angina, and stroke increased, whereas low-to-moderate use remained stable and nondrinking declined by 1.4% annually (Central Illustration).
Central Illustration.
Trends in the Co-occurrence of Alcohol Use and Cardiovascular Disease
APC = annual percentage change; CHD = coronary heart disease; M = men; NHANES = National Health and Nutrition Examination Survey; PIR = poverty income ratio; W = women.
Compared with previous studies
Prior population-based studies have reported trends in CVD prevalence and alcohol use among U.S. adults, separately.21, 22, 23 However, investigations specifically examining temporal trends of alcohol use among individuals with CVD, as well as the trends in the co-occurrence of alcohol consumption and CVD in the U.S. population, remain scarce.
To our knowledge, our study is the first to show that the prevalence of heavy alcohol use among individuals with CVD has increased markedly over time. Notably, heavy alcohol use increased approximately 7 times faster among adults with CVD than without, possibly due to increased stress or misperceptions about alcohol’s cardioprotective effects.
Importantly, this is also the first study to examine the trends in the prevalence of the co-occurrence of alcohol use and CVD and specify in demographic subgroups over time. We observed that in the overall population, the prevalence of co-occurring heavy alcohol consumption and CVD significantly increased between 1999 and 2023. Notably, adults aged 40 to 59 years showed a trend reversal post-2007, with increases following a prior nonsignificant decline. Several factors may underline this trend reversal. One possibility is the confounding impact of tobacco cessation efforts post-2007, as former smokers, now abstinent, may have altered their alcohol consumption patterns.24,25 Shifting preferences in alcohol type may also play a role, as spirit consumption is associated with higher CVD risk than wine.26,27 In addition, the increase coincided with the U.S. economic recession of 2008 to 2009, which aligns with Catalano’s “provocation hypothesis” that economic downturns elevate heavy drinking due to financial stress and insecurity.28,29 Consistent with this, our study observed an increase in combined patterns of heavy drinking and CVD among individuals with lower socioeconomic status (PIR <3), in line with prior research linking economic adversity to heavier alcohol consumption.29,30 Notably, the increase was slightly more pronounced among those with ≤high school education, which may suggest potential socioeconomic disparities in alcohol use and cardiovascular risk.
Furthermore, trends in combined patterns of heavy drinking and CVD appeared to vary by sex and race. Although formal tests for subgroup differences were not statistically significant, the magnitude of increase appeared higher in women than in men, possibly reflecting a complex interplay of biological, psychological, and sociocultural factors.21 For instance, women in high-stress jobs are more likely to engage in heavy drinking.31,32 The increasing marketing of alcoholic beverages has specifically targeted women.33 These patterns may suggest ongoing gender disparities in alcohol use and CVD burden. Moreover, the increase appeared greater among non-Hispanic White adults than among Hispanic and non-Hispanic Black adults. This may reflect cultural permissiveness toward heavy drinking, greater disposable income, and differential exposure to alcohol marketing within these populations.34,35
In addition, in line with existing evidence that has also indicated that total alcohol consumption increased overall, driven primarily by a rise in heavy drinkers accompanied by a decline in light drinkers,30 our analysis also reported decreasing trend of low-to-moderate and no alcohol use among individuals with CVD.
Public health and clinical implications
Our findings indicate that heavy alcohol consumption among individuals with CVD, and the co-occurrence of heavy drinking and CVD, has increased over time, posing growing challenges for clinical care and public health. Clinically, brief 5- to 10-minute counseling sessions and routine alcohol screening in CVD care may reduce hazardous drinking. At the population level, broader strategies, such as alcohol taxation, marketing restrictions, and warning labels, may be particularly relevant for subgroups with relatively higher observed increases, including middle-aged adults, women, non-Hispanic White individuals, and those with lower income or educational attainment. Together, these interventions may help reduce the growing dual burden of alcohol use and CVD.
Strengths and limitations
A major strength of this study is the use of approximately 25 years of nationally representative NHANES data collected using standardized protocols, ensuring robust population-level estimates and generalizability to U.S. adults. The extended time frame allowed for the assessment of long-term secular trends across key demographic subgroups. In addition, the use of APC provided a clear quantification of these trends over time.
This study has several limitations. First, alcohol intake was self-reported and subject to recall bias and potential misclassification, which may have underestimated actual consumption. In addition, beverage-specific alcohol consumption data (eg, wine, beer, spirits) were not consistently available from 1999 to 2023, preventing analysis by alcohol type. Second, in our study, CVD was defined based on self-reported physician diagnosis obtained during the household interview rather than examination-based or biomarker-confirmed measures from MEC. Although self-reported data may be subject to misclassification, the consistent definition across survey cycles enhances comparability for long-term trend analyses. Third, NHANES uses a repeated cross-sectional design. Different individuals are sampled in each survey cycle, so we could not assess within-person changes in alcohol use or CVD status over time. As such, the observational nature of the study precludes causal inference regarding the relationship between alcohol consumption and CVD trends. Fourth, the 2017 to 2020 and 2021 to 2023 survey cycles deviated from the standard 2-year format due to COVID-19 disruptions. Data from 2019 to March 2020 were merged with 2017 to 2018 to form a prepandemic composite sample. Similarly, the 2021 to 2023 cycle differed in duration, which may affect comparability despite adjustments. However, excluding these cycles did not alter results. Fifth, although Joinpoint regression with parallelism tests showed no statistically significant differences across subgroups (eg, age, sex, race/ethnicity, PIR, or education), observed variations in trends should be interpreted as descriptive. Such differences can still have public health relevance by highlighting subgroups that may warrant closer attention.36 Further research is needed to better understand these trends and assess their potential implications for targeted public health interventions. Finally, participants with missing data on CVD or alcohol consumption, as well as those classified as “other race/ethnicity”, were excluded due to small sample sizes and unstable estimates. These exclusions may introduce selection bias and limit the generalizability of our findings. In addition, NHANES includes only the civilian, noninstitutionalized U.S. population, so our findings may not directly generalize to institutionalized populations within the United States (eg, residents of long-term care facilities, nursing homes) and populations outside the United States.
Conclusions
In this nationally representative study, the prevalence of heavy alcohol use among U.S. adults with CVD nearly doubled from 1999 to 2023. The co-occurrence of heavy drinking and CVD also increased substantially over the past 2 decades, with relatively greater increases observed among adults aged 40 to 59 years, women, non-Hispanic White individuals, and those with lower socioeconomic status. These trends underscore the urgent need for targeted public health strategies and clinical interventions to mitigate alcohol-related risks in individuals with CVD.
Perspectives.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: Among U.S. adults with CVD, heavy alcohol use nearly doubled from 17.1% to 33.4% between 1999 and 2023, and the co-occurrence of heavy drinking and CVD rose from 1.4% to 3.0%, particularly among adults aged 40 to 59 years, women, non-Hispanic Whites, and lower socioeconomic groups. These trends identified a rapidly growing high-risk population.
TRANSLATIONAL OUTLOOK: Targeted interventions are needed to reduce alcohol-related risks in CVD patients. Routine alcohol screening and brief counseling, combined with population-level policies such as taxation, marketing restrictions, and warning labels, may curb the rising dual burden of heavy drinking and CVD.
Funding support and author disclosures
This study was supported by the National Natural Science Foundation of China to ZC (No. 82304148), grant from the Scientific Research Program of FuRong Laboratory to ZC (No. 2023SK2107-1), and American Diabetes Association’s Pathway to Stop Diabetes Award to LW (1-24-INI-04). The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
Footnotes
The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the Author Center.
Appendix
For supplemental tables and figures, please see the online version of this paper.
Supplementary material
References
- 1.World Health Organization Alcohol: key facts 2024. https://www.who.int/news-room/fact-sheets/detail/alcohol Accessed August 6, 2025.
- 2.Esser M.B., Leung G., Sherk A., et al. Estimated deaths attributable to excessive alcohol use among US adults aged 20 to 64 years, 2015 to 2019. JAMA Netw Open. 2022;5(11) doi: 10.1001/jamanetworkopen.2022.39485. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Lichtenstein A.H., Appel L.J., Vadiveloo M., et al. 2021 dietary guidance to improve cardiovascular health: a scientific statement from the American Heart Association. Circulation. 2021;144(23):e472–e487. doi: 10.1161/CIR.0000000000001031. [DOI] [PubMed] [Google Scholar]
- 4.He J., Zhu Z., Bundy J.D., Dorans K.S., Chen J., Hamm L.L. Trends in cardiovascular risk factors in US adults by race and ethnicity and socioeconomic status, 1999-2018. JAMA. 2021;326(13):1286–1298. doi: 10.1001/jama.2021.15187. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Centers for Disease Control and Prevention Heart Disease Facts. 2024. https://www.cdc.gov/heart-disease/data-research/facts-stats/index.html
- 6.Beulens J.W.J., Algra A., Soedamah-Muthu S.S., Visseren F.L.J., Grobbee D.E., van der Graaf Y. Alcohol consumption and risk of recurrent cardiovascular events and mortality in patients with clinically manifest vascular disease and diabetes mellitus: the Second Manifestations of ARTerial (SMART) disease study. Atherosclerosis. 2010;212(1):281–286. doi: 10.1016/j.atherosclerosis.2010.04.034. [DOI] [PubMed] [Google Scholar]
- 7.Ronksley P.E., Brien S.E., Turner B.J., Mukamal K.J., Ghali W.A. Association of alcohol consumption with selected cardiovascular disease outcomes: a systematic review and meta-analysis. BMJ (Clinical research ed) 2011;342 doi: 10.1136/bmj.d671. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Piano M.R., Marcus G.M., Aycock D.M., et al. Alcohol use and cardiovascular disease: a scientific statement from the American Heart Association. Circulation. 2025 doi: 10.1161/CIR.0000000000001341. [DOI] [PubMed] [Google Scholar]
- 9.World Health Organization No level of alcohol consumption is safe for our health 2023. https://www.who.int/europe/news/item/04-01-2023-no-level-of-alcohol-consumption-is-safe-for-our-health
- 10.Marcus G.M., Vittinghoff E., Whitman I.R., et al. Acute consumption of alcohol and discrete atrial fibrillation events. Ann Intern Med. 2021;174(11):1503–1509. doi: 10.7326/M21-0228. [DOI] [PubMed] [Google Scholar]
- 11.The US Surgeon General’s Advisory Alcohol and cancer risk. 2025. https://www.hhs.gov/surgeongeneral/reports-and-publications/alcohol-cancer/index.html
- 12.Centers for Disease Control and Prevention Alcohol use and your health. https://www.niaaa.nih.gov/alcohols-effects-health/alcohol-drinking-patterns
- 13.Zipf G., Chiappa M., Porter K.S., Ostchega Y., Lewis B.G., Dostal J. National health and nutrition examination survey: plan and operations, 1999-2010. Vital Health Stat. 2013;1(56):1–37. [PubMed] [Google Scholar]
- 14.Centers for Disease Control and Prevention About NHANES. https://www.cdc.gov/nchs/nhanes/about_nhanes.htm Accessed August 6, 2025.
- 15.Raza S.A., Sokale I.O., Thrift A.P. Burden of high-risk phenotype of heavy alcohol consumption among obese U.S. population: results from National Health and Nutrition Examination Survey, 1999-2020. Lancet Reg Health Am. 2023;23 doi: 10.1016/j.lana.2023.100525. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.National Center for Health Statistics . Poverty – National Health and Nutrition Examination Survey. CDC; 2024. https://www.cdc.gov/nchs/hus/sources-definitions/poverty.htm [Google Scholar]
- 17.Chobufo M.D., Singla A., Rahman E.U., Michos E.D., Whelton P.K., Balla S. Temporal trends in atherosclerotic cardiovascular disease risk among U.S. adults. Analysis of the National Health and Nutrition Examination Survey, 1999-2018. Eur J Prev Cardiol. 2022;29(18):2289–2300. doi: 10.1093/eurjpc/zwac161. [DOI] [PubMed] [Google Scholar]
- 18.Kim H.J., Fay M.P., Feuer E.J., Midthune D.N. Permutation tests for joinpoint regression with applications to cancer rates. Stat Med. 2000;19(3):335–351. doi: 10.1002/(sici)1097-0258(20000215)19:3<335::aid-sim336>3.0.co;2-z. [DOI] [PubMed] [Google Scholar]
- 19.Yu B., Barrett M.J., Kim H.-J., Feuer E.J. Estimating joinpoints in continuous time scale for multiple change-point models. Comput Stat Data Anal. 2007;51(5):2420–2427. [Google Scholar]
- 20.Clegg L.X., Hankey B.F., Tiwari R., Feuer E.J., Edwards B.K. Estimating average annual per cent change in trend analysis. Stat Med. 2009;28(29):3670–3682. doi: 10.1002/sim.3733. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Abdalla S.M., Yu S., Galea S. Trends in cardiovascular disease prevalence by income level in the United States. JAMA Netw Open. 2020;3(9) doi: 10.1001/jamanetworkopen.2020.18150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Biddinger K.J., Emdin C.A., Haas M.E., et al. Association of habitual alcohol intake with risk of cardiovascular disease. JAMA Netw Open. 2022;5(3):e223849. doi: 10.1001/jamanetworkopen.2022.3849. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.van de Luitgaarden I.A., van Oort S., Bouman E.J., et al. Alcohol consumption in relation to cardiovascular diseases and mortality: a systematic review of Mendelian randomization studies. Eur J Epidemiol. 2022;37(7):655–669. doi: 10.1007/s10654-021-00799-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Aubin H.-J., Farley A., Lycett D., Lahmek P., Aveyard P. Weight gain in smokers after quitting cigarettes: meta-analysis. BMJ (Clinical research ed) 2012;345 doi: 10.1136/bmj.e4439. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Inoue-Choi M., Freedman N.D. When to adjust for potentially confounding variables—reply. JAMA Intern Med. 2017;177(6):892. doi: 10.1001/jamainternmed.2017.1086. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Wannamethee S.G., Field A.E., Colditz G.A., Rimm E.B. Alcohol intake and 8-year weight gain in women: a prospective study. Obes Res. 2004;12(9):1386–1396. doi: 10.1038/oby.2004.175. [DOI] [PubMed] [Google Scholar]
- 27.Wannamethee S.G., Shaper A.G. Alcohol, body weight, and weight gain in middle-aged men. Am J Clin Nutr. 2003;77(5):1312–1317. doi: 10.1093/ajcn/77.5.1312. [DOI] [PubMed] [Google Scholar]
- 28.Catalano R. An emerging theory of the effect of economic contraction on alcohol abuse in the United States. Soc Justice Res. 1997;10(2):191–201. [Google Scholar]
- 29.Mulia N., Zemore S.E., Murphy R., Liu H., Catalano R. Economic loss and alcohol consumption and problems during the 2008 to 2009 U.S. recession. Alcohol Clin Exp Res. 2014;38(4):1026–1034. doi: 10.1111/acer.12301. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Bor J., Basu S., Coutts A., McKee M., Stuckler D. Alcohol use during the great recession of 2008-2009. Alcohol Alcohol. 2013;48(3):343–348. doi: 10.1093/alcalc/agt002. [DOI] [PubMed] [Google Scholar]
- 31.Kawakami N., Araki S., Haratani T., Hemmi T. Neurobehavioral methods and effects in occupational and environmental health. Elsevier; 1994. Relations of work stress to alcohol use and drinking problems in male and female employees of a computer factory in Japan; pp. 639–649. [DOI] [PubMed] [Google Scholar]
- 32.Esper L.H., Furtado E. Gender differences and association between psychological stress and alcohol consumption: a systematic review. J Alcohol Drug Depend. 2013;1(3):116–120. [Google Scholar]
- 33.Atkinson A., Meadows B., Emslie C., Lyons A., Sumnall H. ‘Pretty in Pink’and ‘Girl power’: an analysis of the targeting and representation of women in alcohol brand marketing on Facebook and Instagram. Int J Drug Policy. 2022;101 doi: 10.1016/j.drugpo.2021.103547. [DOI] [PubMed] [Google Scholar]
- 34.Azagba S., Shan L., Latham K., Manzione L. Trends in binge and heavy drinking among adults in the United States, 2011–2017. Subst Use Misuse. 2020;55(6):990–997. doi: 10.1080/10826084.2020.1717538. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Sudhinaraset M., Wigglesworth C., Takeuchi D.T. Social and cultural contexts of alcohol use: influences in a social-ecological framework. Alcohol Res Curr Rev. 2016;38(1):35–45. doi: 10.35946/arcr.v38.1.05. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Vandenbroucke J.P., Von Elm E., Altman D.G., et al. Strengthening the reporting of observational studies in epidemiology (STROBE): explanation and elaboration. Int J Surg. 2014;12(12):1500–1524. doi: 10.1016/j.ijsu.2014.07.014. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.