Do Beliefs about Alcohol and Cancer Risk Vary by Alcoholic Beverage Type and Heart Disease Risk Beliefs?

Alcohol consumption is a leading modifiable risk factor for cancer in the United States (1), contributing to an average of more than 75,000 cancer cases and almost 19,000 cancer deaths per year between 2013 and 2016 (2). All beverage types containing ethanol (e.g., wine, beer, liquor) increase cancer risk, and alcohol consumption has been linked to seven cancer types, including cancers of the breast, mouth, and colon (3–5). There is a dose–response relationship between alcohol consumption and cancer risk, such that the more one drinks the greater the risk of developing cancer (4). Even light drinking (≤12.5 g ethanol/day) increases risk of some cancers (e.g., esophageal and breast cancers; ref. 6). Multiple carcinogenic pathways linking alcohol and cancer have been identified, which vary by cancer site. For instance, alcohol consumption is associated with increased blood levels of estrogen, which has been implicated in heightened breast cancer risk (7). Furthermore, ethanol is metabolized to acetaldehyde, which damages DNA, increasing risk of head and neck, esophagus, and liver cancers (7).

Despite alcohol's significant contribution to cancer-related morbidity, mortality, and health care costs (1, 8), awareness of the alcohol–cancer link remains low in the United States (9–11). Data from the 2017 Health Information National Trends Survey found that only 38% of U.S. adults were aware of alcohol's carcinogenic risk. Similarly, analysis of the National Survey of Family Growth (2011–2015) revealed that only 25% of women were aware that alcohol increased the risk of breast cancer (10). In contrast, the same survey found 88% of women were aware that a family history of breast cancer was a risk factor (10). Furthermore, a 2017 survey by the American Institute for Cancer Research found that 93%, 80%, and 39% of U.S. adults were aware that tobacco, asbestos, and alcohol, respectively, increased cancer risk (11). Low awareness of the alcohol-cancer link is not unique to the United States, as recognizing that alcohol is a cancer risk factor has been found to be low in Australia and many European countries (e.g., Greece, Portugal, United Kingdom) as well (12). Moreover, there is some evidence that awareness may be higher for some types of cancer (e.g., liver cancer) and lower for others (e.g., breast cancer; ref. 13).

One factor that may be contributing to low awareness of the alcohol–cancer link is perceptions of potential health benefits of alcohol consumption. Meta-analyses of epidemiologic studies have reported that light to moderate drinkers have a reduced risk of heart disease (14, 15) and ischemic stroke (16), relative to nondrinkers. Despite the fact that these purported protective associations could be due to selection bias and residual confounding (17–19), cardioprotective benefits are not found in studies that control for lifestyle heart disease risk factors (20), and research using mendelian randomization has suggested that alcohol use of all amounts is associated with increased cardiovascular disease risk (21), the “heart health” message is reaching the public. Whitman and colleagues analyzed data from the Health eHeart Study and among participants whom indicated that alcohol was “heart healthy,” 80% reported lay press as a source of this information (22). Furthermore, the same study also reported that two thirds of participants believed that the type of alcohol mattered for heart health, and of these, 92% reported that red wine was exclusively cardioprotective (22).

It is currently not known whether perceptions of potential cardiovascular benefits of alcohol are associated with awareness of the alcohol–cancer link. Moreover, given that wine may be perceived as healthier than other alcoholic beverage types (e.g., beer, liquor; ref. 22), perceptions about the alcohol–cancer link may vary by beverage type. To help advance understanding of U.S. adults’ awareness of the alcohol–cancer link and inform future educational campaigns and messaging, we examined awareness of the alcohol–cancer link by beverage type, as well as its association with beliefs about alcohol and heart disease risk using nationally representative survey data.

This study used data from the National Cancer Institute's Health Information National Trends Survey (HINTS) 5 Cycle 4, a nationally representative mailed survey. HINTS used a two-stage sampling design. In stage one, a stratified random sample of addresses was selected, with an oversample of addresses from a stratum containing high minority census tracts. In stage two, a single adult was selected from each chosen household. A detailed description of the HINTS design has been published elsewhere (23, 24). The HINTS 5 Cycle 4 survey was administered February 27 – June 15, 2020 and a total of 3,865 adults participated (response rate = 36.7% using the American Association for Public Opinion Research Response Rate 4 formula; ref. 24).

Awareness of the link between alcohol and cancer was measured for three alcoholic beverage types: wine, beer, and liquor. The items asked, “In your opinion, how much does drinking the following types of alcohol affect the risk of getting cancer?” Response options were: decreases risk a lot, decreases risk a little, no effect, increases risk a little, increases risk a lot, and don't know. Responses were coded as increases risk (a little/a lot), no effect, decreases risk (a little/a lot), and don't know.

Awareness of how wine, beer, and liquor affect heart disease risk was measured by asking “In your opinion, how much does drinking the following types of alcohol affect the risk of getting heart disease?” Response options were the same as the cancer items and similarly coded as increases risk (a little, a lot), no effect, decreases risk (a little, a lot), and don't know.

Other variables included as potential confounders were sociodemographic measures including sex, age, race/ethnicity, education, income perceptions (e.g., “living comfortably” or “finding it difficult” on present income), region, cancer survivorship status, and drinking status. Current drinking status was measured by asking participants three items: (i) “During the past 30 days, how many days per week did you have at least one drink of any alcoholic beverage?”; (ii) “During the past 30 days, on the days when you drank, about how many drinks did you drink on average?”; (iii) “During the past 30 days, how many times did you have [5 or more for males, 4 or more for females] alcoholic drinks on one occasion?” (binge drinking). Nondrinkers were defined as those indicating that they drank on 0 days per week (in past 30 days) and no binge drinking. Participants indicating recent drinking were classified as drinkers or heavier drinkers based on the 2020–2025 Dietary Guidelines for Americans (25). The Guidelines recommend drinking no more than one drink per day (for men and women). Therefore, participants indicating they drank on >0 days per week, drank one drink on average, and did not binge drink were coded as “drinkers,” while those drinking >1 drink on average or binge drinking were coded as “heavier drinkers.”

In addition, previous research found that cancer information seeking and cancer fatalism beliefs were associated with awareness of the alcohol–cancer link (9). Therefore, we included a measure of cancer information seeking (“Have you ever looked for information about cancer from any source?”; yes/no) and cancer fatalism (“There's not much you can do to lower your chances of getting cancer”; agree/disagree; ref. 9).

All analyses were performed using Stata v16. Sample weights and 50 jackknife replicate weights were applied to produce weighted estimates and design-adjusted standard errors. In addition, design-corrected Pearson χ² tests with the second-order correction of Rao and Scott were used to assess bivariate relationships of awareness with predictor variables and covariates. Design-corrected Pearson χ² tests with the second-order correction of Rao and Scott were also used to compare the weighted prevalence of awareness by beverage type. For all design-corrected Pearson χ² tests, we dichotomized the responses to the three items measuring awareness of the link between alcohol and cancer (wine, beer, liquor) to “increases risk” versus all other responses.

We estimated three weighted multivariable logistic regression models predicting awareness that wine, beer, and liquor increase cancer risk (vs. all other responses). We then calculated predicted probabilities (and marginal effects) of awareness that each alcoholic beverage increases cancer risk by beliefs about how each alcoholic beverage affects heart disease risk. For each of the three models, we included the single item about heart disease risk that matched the beverage type in the dependent variable. For instance, in the model predicting awareness that beer increases cancer risk, we included the item measuring beliefs about how beer affects heart disease risk (and not the items about wine or liquor affecting heart disease risk). All models controlled for the covariates previously mentioned and whether the survey was completed before or during the COVID-19 pandemic (before or after March 11, 2020). Missingness among all variables included in models ranged from 0% to 9.2%, and was highest for Hispanic ethnicity (9.2%), drinking status (8.4%), and race (7.5%). All models used list-wise deletion.

HINTS 5 received an expedited approval from the Westat IRB on March 28, 2016, and was subsequently reviewed by the NIH Office of Human Subjects Research and given a non-human subjects determination via exemption #13204 on April 25, 2016.

The data analyzed from this study are publicly available and can be accessed from the HINTS website: https://hints.cancer.gov/. Stata code used for our analyses can be accessed here: https://osf.io/qshu4/files/osfstorage/635fdbadb233120e8d4313d9.

Awareness that alcohol consumption increased cancer risk was low overall and varied by beverage type. Awareness was lowest for wine (20.3%) and highest for liquor (31.2%); about a quarter of U.S. adults were aware that beer (24.9%) increased cancer risk (all differences significant at P < 0.001). A greater number of U.S. adults believed that wine (10.3%) decreased cancer risk, compared with beer (2.2%) and liquor (1.7%). The percentages of U.S. adults believing that wine, beer, and liquor had no effect on cancer risk were 17.6%, 18.8%, and 15.5%, respectively. Across all three beverage types, most U.S. adults (>50%) indicated they did not know how each beverage affected cancer risk (Fig. 1A).

Beliefs about how alcohol consumption affects heart disease risk followed the same pattern as cancer risk and varied by beverage type (Fig. 1B). Although 38.9% and 36.4% of U.S. adults believed that liquor and beer, respectively, increased heart disease risk, fewer U.S. adults (25.1%) believed that wine increased heart disease risk. In contrast, 17.3% of U.S. adults believed that wine consumption reduced heart disease risk, whereas 3.5% and 2.7% believed that beer and liquor reduced heart disease risk, respectively. A little less than half of U.S. adults did not know how wine (44.3%), beer (48.5%), and liquor (47.6%) affected heart disease risk (Fig. 1B).

Table 1 reports the weighted percentage of U.S. adults aware of the alcohol–cancer link and results from the design-corrected Pearson χ² tests examining bivariate relationships between awareness of the alcohol cancer-link (by beverage type) with sociodemographic characteristics and beliefs about how each beverage affects heart disease risk. Table 2 reports the weighted adjusted odds of awareness of the alcohol–cancer link by beverage type. Across all three beverage types, older adults (ages 40–59 and ≥60) had lower odds of awareness compared to younger adults (ages 18–39). In addition, U.S. adults reporting cancer information seeking from any source had greater odds of awareness of the alcohol–cancer link for beer [aOR = 1.80; 95% confidence interval (CI) = 1.21–2.68] and liquor (aOR = 1.69; 95% CI = 1.12–2.56), but not wine (aOR = 1.43; 95% CI = 0.95, 2.14), compared with U.S. adults reporting no cancer information seeking. Beliefs about how alcohol affects heart disease risk were also associated with awareness of the alcohol–cancer link (all three beverage types). U.S. adults believing each beverage type had no effect on heart disease risk or decreased heart disease risk, or do not know its effect on heart disease risk, had lower odds of awareness compared with U.S. adults that believed each beverage type increased heart disease risk. Drinking status was not associated with odds of awareness of the alcohol–cancer link (P > 0.05) for all beverage types. With one exception, race and ethnicity were not associated with awareness across all three beverage types. While black adults had lower odds (OR = 0.50; 95% CI = 0.25–0.99) of awareness that wine increases cancer risk (relative to white adults), all other relationships between race and ethnicity and awareness across all other products were nonsignificant (P > 0.05). In addition, for all three beverage types, awareness of the alcohol link was not associated with (P > 0.05) sex, education, census region, cancer fatalism beliefs, cancer survivorship status, or survey return date (before or after the COVID-19 pandemic was declared).

Beliefs about how each beverage type affected heart disease risk were associated with awareness of the alcohol and cancer link, which followed a consistent pattern across all three alcoholic beverages. Awareness was greatest among those believing that each beverage increased heart disease risk (unadjusted range: 53.6%–61.1%). Awareness of the alcohol–cancer link was substantially lower among U.S. adults believing alcohol decreased heart disease risk (unadjusted range 16.4%–22.2%), had no effect on heart disease risk (unadjusted range: 9.8%–16.0%), or did not know how alcohol affected heart disease risk (unadjusted range: 5.0%–10.7%).

The adjusted predicted probabilities of being aware of the alcohol–cancer link by beliefs about how consuming each beverage type affects heart disease risk is reported in Table 3. Awareness of the alcohol–cancer link was highest among those who believe alcoholic beverages increased heart disease risk (wine: 58.6%; beer: 52.4%; liquor: 59.4%), and were lower among U.S. adults who believed alcoholic beverages decreased heart disease risk (wine: 16.2%; beer: 21.6%; liquor: 23.8%), or had no effect on heart disease risk (wine: 10.2%; beer: 12.0%; liquor: 16.9%); or did not know how alcohol affects heart disease risk (wine: 6.0%; beer: 8.6%; liquor: 13.2%).

This is the first study to examine awareness of the alcohol–cancer link by beverage type, as well as the relationship of beliefs about alcohol and heart disease with awareness in a national U.S. sample. We found that awareness of the alcohol–cancer link in the United States was low and varied by beverage type. Awareness was highest for liquor (31.2%) and lowest for wine (20.3%). In addition, we found that beliefs about how alcohol affects heart disease risk were associated with awareness of the alcohol–cancer link, such that U.S. adults believing that alcoholic beverages increased heart disease risk reported higher cancer awareness than U.S. adults who believed alcoholic beverages reduced/had no effect or were unsure of how alcohol affects heart disease risk.

The finding that awareness of the alcohol–cancer link was low is consistent with previous studies. For instance, two previous versions of HINTS found that <40% of U.S. adults were aware that alcohol consumption increased cancer risk (9, 26). In addition, for all three alcoholic beverage types measured in this study, >50% reported that they “don't know” how consuming these beverages affected cancer risk. Interventions are clearly needed to educate the public about the cancer risks associated with alcohol consumption.

Although all beverage types containing ethanol increase cancer risk, this study found that fewer U.S. adults were aware that wine increases cancer risk (20.3%) compared with beer (24.9%) and liquor (31.2%). In addition, this study found that 10.3% of U.S. adults believed that wine consumption actually reduced cancer risk, compared with only to 2.2% and 1.7% for beer and liquor, respectively. These findings suggest that future educational interventions may need to underscore that all alcoholic beverage types increase cancer risk given that all have ethanol as a common ingredient. Research is also needed to determine whether different messaging strategies are needed to increase awareness among individuals who “don't know” how alcohol affects cancer risk versus those who believe alcohol has no effect or reduces cancer risk.

Despite public media attention to the purported heart health benefits of wine consumption (22), between 44.3% and 48.5% of U.S. adults were not sure how the three beverage types (including wine) affect heart disease risk. Fewer U.S. adults believed that drinking wine (17.3%), beer (3.5%), and liquor (2.7%) reduced heart disease risk. Nonetheless, this study found that beliefs about alcohol and heart disease were associated with awareness that alcohol consumption increased cancer risk. In particular, awareness was highest among those believing alcohol increased heart disease risk relative to those believing alcohol had no effect or decreased heart disease risk, or do not know how alcohol affects heart disease risk. Although the impact of alcohol consumption on heart disease risk is complex and controversial, U.S. adults believing that alcohol reduces/has no effect or are unsure of how alcohol affects heart disease reported very low awareness of the alcohol–cancer link and might be targeted with educational interventions.

In adjusted models, age was associated with awareness of the alcohol–cancer link, with young adults (ages 18–39) having greater odds of awareness than older adults. This finding held across all three alcoholic beverage types. However, awareness among young adults was still low (<40%). Young adults have higher rates of binge drinking than older adults, which could affect cancer risk (27). While educational interventions are needed for all age groups, targeted messaging for different age demographics may be needed given age-related differences in drinking behaviors, norms, and social pressure.

This study found no association between drinking status and awareness of the alcohol-cancer link (for all three beverage types), which is consistent with previous studies. For instance, Buykx and colleagues measured alcohol use using the Alcohol Use Disorders Test short form (AUDIT-C) and found no relationship between AUDIT-C score (<5 vs. 5+) and awareness of the alcohol–cancer link among adults in England (28). Thomsen and colleagues surveyed Danish adults and reported that odds of being aware that alcohol increased cancer risk were similar across different levels of alcohol consumption (29). Moreover, an earlier version of HINTS (2019) found no association between self-reported alchohol consumption and awareness of the alcohol–cancer link (26). Thus, it appears that current drinking level might not be a good indicator of awareness of alcohol's role in cancer risk. Given that nondrinking behaviors may be temporary, former drinkers (i.e., current nondrinkers) may be at risk for an alcohol-related cancer, and drinking behaviors may be influenced by the attitudes and knowledge of close friends and family, educating nondrinkers (in addition to drinkers) about the alcohol–cancer link may help reduce the burden of alcohol-related cancers.

Mass media campaigns have been used outside the United States to increase awareness of the alcohol–cancer link. For instance, Martin and colleagues evaluated a television campaign in England and found that awareness increased postcampaign (45%) versus precampaign (33%; ref. 30). A separate study found that a television and print media campaign increased odds of awareness among Western Australian women (31). Similar campaigns could be initiated in the United States to help increase awareness of the carcinogenic effects of alcohol. In addition, although clinicians are viewed as the most trustworthy source of health information, a recent analysis of national survey data found that less than half of U.S. adults report a medical provider discussing the harms of alcohol use with them in the past year (32, 33). Clinician discussions could be an effective way to increase awareness, as the same study found clinician discussions about the harms of alcohol were associated with greater odds of being aware of the alcohol–cancer link (33).

In the United States, alcoholic beverages are required to feature warnings about the dangers for pregnant women consuming alcohol and how alcohol consumption “impairs your ability to drive a car or operate machinery, and may cause health problems.” At the present time, no specific “health problems” are stated. Adding cancer warning labels to alcoholic beverages could also be used to increase awareness of the alcohol–cancer link, as recommended by several public health groups in a recent petition (34). Moreover, a separate analysis of the same 2020 HINTS data found that 65% of U.S. adults support adding health warnings to alcohol containers (35). In addition to increasing awareness, adding cancer warning labels may also reduce alcohol use. For instance, using an interrupted time-series design, researchers evaluated the impact of adding cancer risk warnings and drinking guidelines to alcoholic beverages sold in a sample of stores in Yukon, Canada, and found that the labels reduced alcohol sales (36).

Increasing awareness of the link between alcohol and cancer may also increase public support for alcohol control policies. Studies from the United States, Australia, England, Canada, and Denmark have all found awareness of the alcohol–cancer link to be associated with increased support for a variety of alcohol policies, including adding warning labels, restricting advertising, and pricing policies (35, 37–40). Therefore, educating the public about the cancer risks of alcohol may also help expedite alcohol policy adoption and implementation.

Strengths of the current study include the use of a large, diverse, national probability sample, and measurement of awareness of the alcohol–cancer link by beverage type. Measurement of awareness also included a “don't know” option, which yielded a more nuanced picture of lack of awareness. One limitation was the use of unconditional measures regarding awareness of the alcohol–cancer link and beliefs about how alcohol affects heart disease risk (i.e., quantity of alcohol consumed and length of drinking were not specified). In addition, these measures did not ask about how alcohol consumption affected risk of specific cancer types, which could have created confusion given that alcohol consumption is linked to some but not all cancer types. This study found an association between beliefs about how alcohol affects heart disease risk with awareness of the alcohol–cancer link. However, the cross-sectional design prevents assessment of temporality of this association. Another limitation is that some data collection occurred during the COVID-19 pandemic, which may have affected study participation and survey responses. Notably, adjusted analyses did control for whether the survey was returned before or during the pandemic.

Alcohol use is a leading modifiable risk factor for cancer in the United States, yet awareness that alcohol consumption increases cancer risk remains low (1, 9, 26). This study's findings underscore the need to develop interventions for educating the public about the cancer risks of alcohol use, particularly in the prevailing context of national dialog about the purported heart health benefits of wine. Educating the public about how alcohol increases cancer risk will not only empower consumers to make more informed decisions, but may also prevent and reduce excessive alcohol use, as well as cancer morbidity and mortality.

No disclosures were reported.

A.B. Seidenberg: Conceptualization, formal analysis, writing–original draft. K.P. Wiseman: Writing–review and editing. W.M. Klein: Writing–review and editing.

K.P. Wiseman is an iTHRIV Scholar. The iTHRIV Scholars Program is supported in part by the National Center for Advancing Translational Sciences of the National Institutes of Health under Award Numbers UL1TR003015 and KL2TR003016. Opinions expressed by the authors are their own and this material should not be interpreted as representing the official viewpoint of the U.S. Department of Health and Human Services, the National Institutes of Health, or the National Cancer Institute.

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.