Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2025 Mar 31;49(5):1117–1125. doi: 10.1111/acer.70046

Risk of stroke accompanying alcohol consumption with or without single‐occasion drinking

Midori Takada 1,2,3, Kazumasa Yamagishi 2,4,5,✉, Isao Muraki 2,4,6, Yuji Shimizu 1, Mari Tanaka 7, Tomomi Kihara 2, Mitsumasa Umesawa 2,8, Hironori Imano 7, Tomoko Sankai 9, Takeo Okada 10, Akihiko Kitamura 11, Masahiko Kiyama 10, Hiroyasu Iso 12
PMCID: PMC12098811  PMID: 40165519

Abstract

Background

Previous research on the relationship between alcohol and stroke has highlighted several issues. Notably, the conventional categorization based on average consumption, which categorizes both those who consume 20 g/day of alcohol daily and those who engage in risky single‐occasion drinking (RSOD) only on weekends into low drinkers, makes it difficult to account for individuals who fall into distinct characteristics. This study examined the association between alcohol and stroke, accounting for both average drinking levels and RSOD occurrences.

Methods

In a community‐based prospective cohort study in Japan, 8026 men and 12,461 women were followed from 1989 to 2018. The outcome was the first‐ever stroke event during the follow‐up period. Alcohol consumption was divided into seven categories: never drinkers; former drinkers; low drinkers (<20 g/day on average for men and <10 g/day for women) without RSOD; moderate drinkers (20–59 g/day on average for men and 10–39 g/day for women) without RSOD; low drinkers with RSOD; moderate drinkers with RSOD; and heavy drinkers (≥60 g/day on average for men and ≥40 g/day for women). RSOD was defined as consuming ≥60 g for men and ≥40 g for women on a single occasion. We calculated sex‐specific hazard ratios (HRs) and 95% confidence intervals (CIs) for stroke across these drinking categories compared with never drinkers.

Results

Low‐to‐moderate drinkers with RSOD experienced a significantly higher hazard of stroke; the multivariable HR (95% CI) of 1.47 (1.01–2.13) among men and 3.41 (1.50–7.79) among women. Overall, although some were not significant, low‐to‐moderate drinkers with and without RSOD tended to be associated with a higher hazard of stroke, except for low drinkers with RSOD in women.

Conclusions

RSOD potentially increases the risk of stroke among men and women, even if their usual amount of alcohol consumption is low to moderate.

Keywords: alcohol consumption, cohort studies, epidemiology, risky single‐occasion drinking, stroke

Previous studies on alcohol and stroke have highlighted several concerns. Notably, the conventional categorization based on average consumption, which categorizes both those who consume 20 g/day of alcohol daily and those who engage in risky single‐occasion drinking (RSOD) only on weekends into low drinkers, makes it difficult to account for differently characterizing individuals. This study addresses these concerns and shows that RSOD increases stroke risk among men and women, even if their usual amount of alcohol consumption is low to moderate.

graphic file with name ACER-49-1117-g001.jpg

INTRODUCTION

A comprehensive review of the literature has explored the effects of alcohol consumption on stroke risk (Griswold et al., 2018; Larsson et al., 2016; Patra et al., 2010; Rehm et al., 2010, 2017; Reynolds et al., 2003; Roerecke, 2021; Ronksley et al., 2011; Zhang et al., 2014). Chronic heavy drinking is associated with an increased risk of total stroke (Larsson et al., 2016; Rehm et al., 2010, 2017; Reynolds et al., 2003; Roerecke, 2021; Ronksley et al., 2011; Zhang et al., 2014). Moreover, a positive linear relationship has been identified between alcohol intake and the risk of hemorrhagic stroke (Patra et al., 2010; Rehm et al., 2010, 2017; Reynolds et al., 2003; Roerecke, 2021). Conversely, the relationship between alcohol consumption and ischemic stroke appears J‐shaped, suggesting that light‐to‐moderate drinking may offer a protective effect against ischemic stroke (Griswold et al., 2018; Larsson et al., 2016; Patra et al., 2010; Rehm et al., 2010, 2017; Reynolds et al., 2003; Roerecke, 2021; Zhang et al., 2014). Concerning drinking patterns, risky single‐occasion drinking (RSOD) (Gmel et al., 2011), defined as excessive alcohol consumption on a single occasion, has been linked to increased risks of all stroke types (Rehm et al., 2010, 2017; Roerecke, 2021).

However, prior studies have identified several methodological issues. Primarily, traditional classifications of alcohol consumption fail to differentiate between average daily intake and RSOD. For instance, individuals who consume 20 g of alcohol daily and those who consume 60 g only on weekends are grouped under the same category, calculated as an average consumption of less than 20 g/day. The conventional categorization system based on average consumption did not account for individuals who fall into distinct alcohol use exposure levels. Thus, the classification system that accounts for both average drinking levels and RSOD is needed.

Secondly, drinking patterns have traditionally been categorized merely by the presence or absence of RSOD (Leppälä et al., 1999; O'Donnell et al., 2010, 2016; Palomaki & Kaste, 1993; Smyth et al., 2015; Sull et al., 2010; Sundell et al., 2008; Wannamethee & Shaper, 1996) or by a basic frequency of such episodes (Hansagi et al., 1995; Longstreth et al., 1992; Sull et al., 2009). The independent risk of RSOD on stroke, irrespective of regular drinking amounts, remains underexplored. Several studies have only noted the presence or absence of RSOD, grouping those who drink heavily daily with those who do so just once a week (Leppälä et al., 1999; O'Donnell et al., 2010, 2016; Palomaki & Kaste, 1993; Smyth et al., 2015; Sull et al., 2010; Sundell et al., 2008; Wannamethee & Shaper, 1996). Others have assessed RSOD but failed to account for average alcohol consumption, which can lead to misleading conclusions (18–20). For example, categories, such as “lifelong abstainer, never binge drinker (current drinker without RSOD), yearly/more rarely binge drinker, weekly/monthly binge drinker” have been used (Hansagi et al., 1995), where sometimes the average alcohol consumption of a never binge drinker surpasses that of a weekly/monthly binge drinker. An example of this is someone who drinks a bottle of wine daily without bingeing (never binge drinker), compared with someone who consumes two bottles in one session weekly but abstains on other days (weekly/monthly binge drinkers).

Third, there has often been a conflation of never drinkers and former drinkers as a single reference group labeled nondrinkers. This is problematic as former drinkers might have ceased alcohol consumption due to health issues, leading to potential biases in studies. This “sick quitter effect” has fueled ongoing debates regarding the J‐shaped association observed between alcohol consumption and the risk of ischemic stroke (Puddey et al., 1999). Of the meta‐analyses reporting this J‐shaped curve, only one has distinctly classified former drinkers from never drinkers (Patra et al., 2010) among those that have reviewed the relationship between alcohol intake and ischemic stroke risk (Larsson et al., 2016; Patra et al., 2010; Reynolds et al., 2003; Ronksley et al., 2011; Zhang et al., 2014). Another meta‐analysis concluded that the J‐shaped association was not statistically significant PEVuZE5vdGU (Griswold et al., 2018). Furthermore, no study has simultaneously addressed the classification of drinking amount and drinking patterns, and the differentiation of never drinkers in the reference group.

Furthermore, most previous studies examining the association between RSOD and stroke have focused on Western populations, which generally consume more alcohol than Asian populations (Carvalho et al., 2019; OECD, 2019). Exploring this association in non‐Western contexts could enhance our understanding, as the impact of inappropriate drinking (heavy drinking or RSOD) on stroke risk varies regionally: it is lowest in Western Europe, North America, and Australia; moderate in Southeast Asia and South America; and highest in Africa and South Asia (O'Donnell et al., 2016). In Asia, research has primarily targeted the Korean population (Sull et al., 2009, 2010), where the prevalence of RSOD is notably high, comparable to Western levels (OECD, 2020).

This study aimed to explore the relationship between alcohol consumption and stroke risk within the Japanese population, employing a classification system that accounts for both average drinking amounts and patterns while distinctly categorizing nondrinkers into never drinkers and former drinkers.

The hypothesis of this study is that even for low or low‐to‐moderate drinking, which has previously been suggested to have a potential protective effect against certain types of stroke, engaging in RSOD may increase the stroke risk.

MATERIALS AND METHODS

Study cohort

The study participants were residents aged 30–74 from four communities where annual health check‐ups for cardiovascular disease prevention have been conducted. This is part of the Circulatory Risk in Communities Study (CIRCS), an ongoing dynamic community cohort study initiated in 1963. The communities include: (1) Ikawa town, Akita Prefecture (a northwestern rural community in Japan); (2) Minami‐Takayasu district, Yao City, Osaka Prefecture (a mid‐western suburb); (3) Noichi town, Kochi Prefecture (a southwestern rural community); and (4) Kyowa town, Ibaraki Prefecture (a mid‐eastern rural community). The specifics of CIRCS are detailed elsewhere PEVuZE5vdGU (Yamagishi et al., 2019). Between 1989 and 2010, 8303 men and 12,645 women were recruited in Ikawa, Minami‐Takayasu, and Kyowa, and from 1989 to 2005 in Noichi.

The study protocol was approved by the institutional review boards of the Osaka Institute of Public Health (2305‐01‐2).

Baseline survey

During the baseline health check‐up, interviewers collected data on participants' alcohol consumption habits, including both usual and unusual drinking occasions per week. Alcohol drinking status was determined through the question, “Do you habitually (more than once a week) drink alcoholic beverages?” Participants could select from three responses: “I currently drink,” “I do not currently drink, but I drank in the past (more than 3 months ago),” or “I have never drunk alcohol (including in the past).” These responses were categorized as “current drinker,” “former drinker,” and “never drinker,” respectively. For usual and unusual drinking occasions, the frequency of drinking per week and amount of drinking per occasion were recorded among current and former drinkers (as past drinking conditions), respectively. Weekly drinking frequency and amount of drinking per occasion for current drinkers were used in the analysis. This distinction between usual and unusual drinking occasions was important, as it is common for Japanese individuals to consume different amounts of alcohol on weekends compared with weekdays. Specifically, the initial question on usual drinking occasion was, “How many times a week do you drink alcohol and how much do you drink per occasion?” This was followed by an inquiry into unusual drinking occasion: “We are asking about occasions when you might drink more or less than usual in a week. How many times a week do such occasions occur, and how much do you drink on these occasions?”

For both usual and unusual drinking occasions, the amount of alcohol consumed was measured in units of “go”. A “go” is a traditional Japanese volume unit, equating to 180 mL of sake (Japanese rice wine), which contains 23 g of ethanol. This is comparable to one bottle (633 mL) of beer, two single shots (75 mL) of whiskey, or two glasses (180 mL) of wine. The number of “go” was then converted into grams of ethanol intake per occasion by multiplying by 23. The average daily intake was calculated by summing the total ethanol from both usual and unusual drinking occasions per week and dividing by seven. RSOD per week was defined for men as consuming 60 g or more of ethanol per occasion and for women, 40 g or more based on either usual or unusual drinking occasion. These cut‐offs are based on the WHO's definition of heavy episodic drinking, which is consuming 60 g of ethanol on at least one occasion for men (World Health Organization, n.d.), and aligned with the cut‐off from previous studies for women (Iwamoto et al., 2010; Puka et al., 2023; Sundell et al., 2008).

In order to devise a classification that integrates both average alcohol intake and drinking patterns, we initially categorized drinking status into five groups: never drinkers, former drinkers, low drinkers (average <20 g/day for men, <10 g/day for women), moderate drinkers (average 20–59 g/day for men, 10–39 g/day for women), and heavy drinkers (average ≥60 g/day for men, ≥40 g/day for women). We further segmented low and moderate drinkers based on their drinking patterns, distinguishing those with or without RSOD. The Ministry of Health, Labour and Welfare (MHLW, n.d.) in Japan recognizes an average daily intake of less than 20 g/day as reasonable and 60 g/day or more as heavy drinking for Japanese men, establishing these as the cut‐off values for low and heavy drinkers, respectively. The thresholds for women were set at less than 10 g for low drinkers and 40 g or more for heavy drinkers, based on evidence suggesting that women develop organ damage from alcohol at roughly half the rate of men (Becker et al., 1996). Since the number of low drinkers with RSOD and moderate drinkers with RSOD was modest, we further created the category low‐to‐moderate drinker with RSOD, which combines low drinker with RSOD and moderate drinker with RSOD into one category.

Body mass index (BMI) was calculated by dividing the weight in light clothing (kg) by the square of the height (m2). Blood pressure was measured in the right arm using standard mercury sphygmomanometers and unified epidemiological methods by trained physicians or nurses. Blood pressure was classified into six categories based on the 2019 Japanese Society of Hypertension guidelines: normal (<120/<80 mmHg), high normal (120–129/<80 mmHg), elevated (130–139/80–89 mmHg), grade I hypertension (140–159/90–99 mmHg), grade II hypertension (160–179/100–109 mmHg), and grade III hypertension (≥180/≥110 mmHg). Blood samples were collected from seated participants into plastic serum separator gel tubes and centrifuged within 30 min of collection. Serum total cholesterol and triglyceride levels were measured using enzymatic methods. Glucose levels were assessed enzymatically with hexokinase from 1989 to 1992 and after 2001 and with the glucokinase method between 1993 and 2000. Diabetes mellitus was defined by one or more of the following criteria: (1) fasting serum glucose ≥7.0 mmol/L (≥126 mg/dL), (2) nonfasting serum glucose ≥11.1 mmol/L (≥200 mg/dL), or (3) the use of medication for diabetes. Participants were required to fast for at least 8 h before testing. Interviews also covered the number of cigarettes smoked per day and medication usage for hypertension and diabetes mellitus. Standard 12‐lead electrocardiograms (ECGs) were conducted while the patient was at rest in the supine position, and two or more experienced physicians coded each ECG using the Minnesota codes.

Follow‐up and confirmation of stroke

The follow‐up continued until the end of 2018 in Ikawa, 2018 in Minami‐Takayasu, 2010 in Noichi, and 2015 in Kyowa, with the aim of identifying the first incidence of stroke, relocation out of the community, or death (shown in Figure 1).

FIGURE 1.

FIGURE 1

Open in a new tab

Follow‐up period for each study region. The follow‐up continued until the end of 2018 in Ikawa, 2018 in Minami‐Takayasu, 2010 in Noichi, and 2015 in Kyowa, with the aim of identifying the first incidence of stroke, relocation out of the community, or death.

The CIRCS stroke registry collected data on both fatal and nonfatal strokes within the study area, including stroke subtype, date of stroke onset, and date of death. Detailed descriptions of the registry system are available in previous CIRCS publications (Kitamura et al., 2008; Shimamoto et al., 1989).

Information on potential stroke cases was systematically gathered from various sources provided primarily by local municipalities, including death certificates, national health insurance claims, annual household questionnaires, reports from local physicians, public health nurses, health volunteers, and annual health check‐ups. These sources were consolidated for each individual using the Basic Resident Registers and research registration numbers assigned to residents of the study areas. If any source mentioned a stroke, the individual was considered a survey candidate.

All living candidates were contacted by phone, visited, or invited to annual health check‐ups to gather clinical histories, including symptoms at the time of stroke onset, to facilitate stroke diagnosis according to the standardized criteria of the National Survey of Stroke (Walker et al., 1981). This diagnosis required focal neurological symptoms with rapid onset persisting for at least 24 h or until death. For patients meeting these criteria, medical charts from local clinics and hospitals, including computed tomography and magnetic resonance imaging, were reviewed. For deceased candidates, clinical histories were obtained from the patients' families and/or attending physicians, and their medical charts were reviewed similarly to those of living candidates.

The final diagnosis was established through consensus among three physician‐epidemiologists familiar with the epidemiological diagnosis and registration of stroke, based on the compiled information and according to the standardized National Survey of Stroke criteria. They were blinded to data from annual health check‐ups, including any history of stroke and other cardiovascular diseases, to ensure adherence to these standardized criteria.

Statistical analysis

Baseline characteristics are presented as mean values or prevalence. Person‐years were calculated by summing the individual follow‐up durations from the date of the initial health check‐up until the occurrence of stroke, death, emigration, or the end of the follow‐up period, whichever came first. Hazard ratios (HRs) and 95% confidence intervals (CIs) for stroke were computed using Cox proportional hazards models, with never drinkers serving as the reference group. In the initial model, HRs were adjusted for age and stratified by community as Model 1. The multivariable adjustment model further included potential confounding factors, such as smoking status (categorized as never, former, current ≤19 cigarettes/day, and current ≥20 cigarettes/day) and survey year as Model 2. Blood pressure, antihypertensive medication use, diabetes, total cholesterol, triglyceride, BMI, and atrial fibrillation were not included in Model 2 because they were considered intermediate factors in the association between alcohol consumption and stroke. Indicator variables were employed for missing data.

As a sensitivity analysis to assess the impact of potential misclassification due to changes in drinking status during follow‐up, we recalculated HRs using updated categories of drinking status and confounders (except for sex), employing health check‐up data from 5 ± 2 years and 10 ± 2 years from the baseline, with participation rates of 63% and 35% of baseline participants, respectively (Yamagishi et al., 2022). For participants who did not attend the health check‐ups at 5 ± 2 years or 10 ± 2 years, data from the baseline or the 5 ± 2 years from the baseline were used, respectively.

Also, sex‐specific RSOD‐stratified analysis and average ethanol intake‐stratified analysis were performed. Multiplicative interactions with the presence or absence of RSOD and average ethanol intake were tested using a cross‐product term.

We also examined the HR of stroke using a classification system combining RSOD frequency and average ethanol intake.

All statistical analyses were conducted using SAS software (version 9.4; SAS Institute Inc., 2013). All p‐values were two‐tailed, with statistical significance set at p < 0.05.

RESULTS

After excluding participants with a history of stroke or coronary heart disease (214 men and 104 women) and those with missing alcohol consumption data (63 men and 80 women), 8026 men and 12,461 women were eligible for analysis.

During a median follow‐up period of 16.3 years, 549 (6.8%) incident strokes were recorded among men. For women, during a follow‐up of 19.1 years, 512 (4.1%) strokes were documented (shown in Table 1).

TABLE 1.

Incident of total stroke according to alcohol consumption, Circulatory Risk in Communities Study (CIRCS).

Alcohol consumption
Never drinkers Former drinkers Low drinkers without risky single‐occasion drinking Moderate drinkers without risky single‐occasion drinking Low drinkers with risky single‐occasion drinking Moderate drinkers with risky single‐occasion drinking Heavy drinkers
Risky single‐occasion drinking (−) (−) (+) (+)
Average ethanol intake <20 g/day 20–59 g/day <20 g/day 20–59 g/day ≥60 g/day
Men
No. at risk 1770 505 1627 2719 64 530 811
Person‐years 28,684 7375 26,301 45,641 1040 9566 14,224
Total stroke (n) 93 45 95 220 6 34 56
Risky single‐occasion drinking (−) (−) (+) (+)
Average ethanol intake <10 g/day 10–39 g/day <10 g/day 10–39 g/day ≥40 g/day
Women
No. at risk 10,062 298 1342 560 23 97 79
Person‐years 191,097 4632 22,831 8799 386 1560 1222
Total stroke (n) 457 9 18 19 1 5 3
Open in a new tab

Table S1 displays the baseline characteristics according to drinking status. Drinkers with RSOD tended to be younger and had lower total cholesterol levels, regardless of their average alcohol consumption. Additionally, low drinkers with RSOD exhibited a higher prevalence of diabetes mellitus, elevated BMI, and increased triglyceride levels. With higher average alcohol consumption, there were increases in both systolic and diastolic blood pressure, as well as in the proportions of current smokers and cases of atrial fibrillation.

The proportionality assumption of the Cox proportional hazards model was tested using a Schoenfeld global test. The assumption was confirmed for all variables among women and for all variables except community among men, leading us to stratify the model by community.

Table 2 shows sex‐specific HRs of total stroke according to alcohol consumption. In Model 2, low‐to‐moderate drinkers with RSOD were associated with a significantly higher hazard of stroke among both men and women; the multivariable HRs (95% CI) were 1.47 (1.01–2.13) for men, p = 0.04, and 3.41 (1.50–7.79) for women, p = 0.004. Among men, low and moderate drinkers without RSOD and heavy drinkers were also associated with a significantly higher hazard of stroke. Among women, moderate drinkers without RSOD tended to be associated with a higher but nonsignificant stroke hazard. Conversely, low drinkers without RSOD among women tended to be associated with a nonsignificant reduced stroke hazard. These results were attenuated in the model adjusted for intermediate variables, but low‐to‐moderate drinkers with RSOD were still associated with a significantly higher hazard of stroke among women; the multivariable HR (95% CI) was 3.10 (1.35–7.08), p = 0.007 (data not shown in tables). Similar results were observed when the category of drinking status was updated, as shown in Table S2.

TABLE 2.

Sex‐specific age and community stratified, and multivariable hazard ratios (HRs) and 95% confidence intervals (95% CIs) of total stroke according to alcohol consumption, Circulatory Risk in Communities Study (CIRCS).

Alcohol consumption
Never drinkers Former drinkers Low drinkers without risky single‐occasion drinking Moderate drinkers without risky single‐occasion drinking Low drinkers with risky single‐occasion drinking Moderate drinkers with risky single‐occasion drinking Low‐to‐moderate drinkers with risky single‐occasion drinking Heavy drinkers
Risky single‐occasion drinking (−) (−) (+) (+) (+)
Average ethanol intake <20 g/day 20–59 g/day <20 g/day 20–59 g/day <60 g/day ≥60 g/day
Men
Age and community stratified HR (95% CI), Model 1 1.0 1.44 (1.01–2.06) 1.28 (0.97–1.71) 1.45 (1.13–1.84) 2.27 (0.99–5.21) 1.45 (0.97–2.15) 1.52 (1.05–2.22) 1.53 (1.09–2.14)
Multivariable and community stratified HR (95% CI), Model 2 a 1.0 1.49 (1.04–2.14) 1.36 (1.02–1.82) 1.44 (1.13–1.84) 2.27 (0.99–5.22) 1.38 (0.93–2.05) 1.47 (1.01–2.13) 1.49 (1.06–2.10)
Risky single‐occasion drinking (−) (−) (+) (+) (+)
Average ethanol intake <10 g/day 10–39 g/day <10 g/day 10–39 g/day <40 g/day ≥40 g/day
Women
Age and community stratified HR (95% CI), Model 1 1.0 1.20 (0.62–2.33) 0.63 (0.39–1.01) 1.46 (0.92–2.31) — 3.76 (1.54–9.14) 3.37 (1.49–7.59) 2.18 (0.70–6.80)
Multivariable and community stratified HR (95% CI), Model 2 a 1.0 1.24 (0.64–2.43) 0.64 (0.39–1.02) 1.50 (0.93–2.41) — 3.82 (1.55–9.43) 3.41 (1.50–7.79) 2.38 (0.74–7.65)
Open in a new tab

Abbreviations: CI, confidence interval; HR, hazard ratio.

a

Model 2: further adjusted for smoking status, and survey year as potential confounding factors.

As for potential effect modification by RSOD on the association between average ethanol intake and stroke, Table S3 showed the results of the analysis to estimate the HR of stroke in moderate‐ and heavy drinkers compared with low drinkers stratified by RSOD among current drinkers. Compared with low drinkers, there was no substantial difference in HR of stroke for moderate drinkers among men without RSOD, and a nonsignificant lower HR of stroke for moderate drinkers among men with RSOD. The interaction was not statistically significant (p for interaction = 0.82). As for women, the modest number of women with RSOD did not allow for interaction to be investigated.

Additionally, the association between RSOD and stroke stratified by average ethanol intake has been shown in Table S4. Compared to those without RSOD, there was a nonsignificant increased HR of stroke in those with RSOD in male low drinkers, with no difference in HR of stroke for those with RSOD in male moderate drinkers. The interaction was not significant (p for interaction = 0.22). As for women, there was a nonsignificant increased HR of stroke in those with RSOD in moderate drinkers. The modest number of low drinkers with RSOD among women did not allow for interaction to be investigated.

Compared with never drinkers, those with a higher frequency of RSOD even at the same level of average ethanol intake showed a higher trend of HRs of stroke in Table S5.

DISCUSSION

In this community‐based prospective cohort study, we observed that low‐to‐moderate drinkers with RSOD were significantly associated with a higher stroke hazard. Overall, although some were not significant, low‐to‐moderate drinkers with and without RSOD tended to be associated with a higher hazard of stroke, except for low drinkers with RSOD in women. The findings of an increased risk of stroke in low‐to‐moderate drinkers with RSOD were consistent with the hypothesis. These associations remained consistent after updating the drinking status categories.

The finding that RSOD is linked to a higher stroke risk aligns with the results of previous research focusing on the relationship between RSOD and stroke risk. Six prospective studies and four case–control studies have indicated that RSOD is associated with either a significant or nonsignificant increase in stroke risk (Hansagi et al., 1995; Leppälä et al., 1999; Longstreth et al., 1992; O'Donnell et al., 2010, 2016; Palomaki & Kaste, 1993; Smyth et al., 2015; Sull et al., 2009, 2010; Sundell et al., 2008; Wannamethee & Shaper, 1996). However, these studies often failed to address whether RSOD increased the risk of stroke independently of regular drinking amounts (Hansagi et al., 1995; Longstreth et al., 1992; O'Donnell et al., 2010, 2016; Smyth et al., 2015; Sull et al., 2009, 2010; Wannamethee & Shaper, 1996) and frequently merged former drinkers and never drinkers into the same reference group PEVuZE5vdGU (Longstreth et al., 1992; Palomaki & Kaste, 1993; Sull et al., 2009, 2010; Sundell et al., 2008). To our knowledge, this is the first study to address these specific concerns.

Various studies have suggested potential mechanisms behind the association between RSOD and stroke risk. Firstly, heavy drinking can lead to acute hypertension by raising blood pressure (Puddey et al., 1999). Secondly, irregular heavy drinking may disrupt the cardiac conducting system, increasing the risk of atrial fibrillation (Liang et al., 2012). Additionally, irregular drinking patterns are linked to a higher risk of thrombosis, evidenced by increased thromboxane B2 formation by platelets, a lowered threshold for adenosine diphosphate concentration to trigger irreversible platelet aggregation, and reduced bleeding time during the withdrawal period following RSOD (McKee & Britton, 1998). RSOD has also been associated with an increased risk of diabetes (Cullmann et al., 2012).

Meta‐analyses have demonstrated a J‐shaped relationship between alcohol consumption and stroke risk, with the lowest risk occurring at light‐to‐moderate consumption levels (Griswold et al., 2018; Larsson et al., 2016; Patra et al., 2010; Reynolds et al., 2003; Ronksley et al., 2011; Zhang et al., 2014). In our study, women exhibited a J‐shaped curve, whereas men did not, and the reasons for this discrepancy are unclear. Mendelian randomization analyses have not supported a protective effect of light‐to‐moderate consumption on stroke PEVuZE5vdGU (Holmes et al., 2014; Larsson et al., 2020; Millwood et al., 2019).

The strength of our study lies in its large general population cohort, which was followed up for the first stroke event using well‐established criteria and supported by the availability of CT/MRI examinations. However, this study has some limitations. First, alcohol consumption is likely to change over the follow‐up period, potentially leading to misclassification and reduced validity. For example, 63% of participants had their drinking status reassessed 5 years after baseline. The proportions of those who remained in the same drinking category were 92% for never drinkers, 56% for former drinkers, 53% for light drinkers without RSOD, 63% for moderate drinkers without RSOD, 6% for light drinkers with RSOD, 12% for moderate drinkers with RSOD, and 33% for heavy drinkers. A portion of heavy episodic drinkers may have altered their drinking patterns due to health issues. These shifts in drinking status suggest that relying solely on baseline drinking status may not be sufficient. Nonetheless, the sensitivity analysis using the updated categories of drinking status and confounding variables produced similar results to those using baseline categories, suggesting that the internal validity of the study was partially ensured. Second, the assessment of drinking status was based on a question about weekly drinking frequency, which did not capture those who drank less frequently than once a week. Third, the unusual drinking measured in this study could only detect consumption that was either higher or lower than usual. Notably, very few current drinkers reported drinking less than their usual amount (0.2% of men and 0% of women). It appears that most current drinkers who engaged in unusual drinking consumed more than their usual level, and it would be rare for them to consume both more and less than usual. Fourth, potential confounding factors, such as socioeconomic status and psychosocial factors were unfortunately not collected. The presence of residual confounding can undermine internal validity by limiting the accuracy of causal inferences. Fifth, the number of stroke cases in each drinking category was small, indicating that further studies are needed. Finally, this study was conducted on a general population in Japan, so the external validity is limited, and caution is needed when applying the results to other populations with different regional or cultural backgrounds.

In conclusion, RSOD may increase the risk of stroke among men and women, even if their usual amount of alcohol consumption is low to moderate.

AUTHOR CONTRIBUTIONS

Midori Takada: conceptualization (lead); formal analysis (lead); funding acquisition (lead); investigation (equal); investigation (equal); methodology (lead); project administration (lead); visualization (lead); writing—original draft (lead); writing—review and editing (lead). Kazumasa Yamagishi: funding acquisition (equal); investigation (equal); writing—review and editing (equal). Isao Muraki and Mitsumasa Umesawa: investigation (equal); writing—review and editing (equal). Yuji Shimizu, Tomomi Kihara, Tomoko Sankai, and Akihiko Kitamura: investigation (equal). Mari Tanaka: data curation (lead). Hironori Imano: investigation (equal); writing—review and editing (equal); data curation (equal). Takeo Okada: supervision (equal). Masahiko Kiyama: investigation (equal); supervision (equal). Hiroyasu Iso: funding acquisition (equal); investigation (lead); supervision (lead); writing—review and editing (equal).

FUNDING INFORMATION

This research was partly supported by a Grant‐in‐Aid from the Ministry of Health, Labour and Welfare, Health and Labour Sciences Research Grants, Japan (22FA1012; 23FA1006).

CONFLICT OF INTEREST STATEMENT

None to declare.

Supporting information

Tables S1–S5

ACER-49-1117-s001.docx (96.4KB, docx)

ACKNOWLEDGMENTS

The authors are grateful to all the participants and the Circulatory Risk in Communities Study staff. We would like to thank Editage (www.editage.com) for the English language editing.

Takada, M. , Yamagishi, K. , Muraki, I. , Shimizu, Y. , Tanaka, M. , Kihara, T. et al. (2025) Risk of stroke accompanying alcohol consumption with or without single‐occasion drinking. Alcohol: Clinical and Experimental Research, 49, 1117–1125. Available from: 10.1111/acer.70046

DATA AVAILABILITY STATEMENT

Research data are not shared.

REFERENCES

  1. Becker, U. , Deis, A. , Sørensen, T.I. , Grønbaek, M. , Borch‐Johnsen, K. , Müller, C.F. et al. (1996) Prediction of risk of liver disease by alcohol intake, sex, and age: a prospective population study. Hepatology, 23, 1025–1029. [DOI] [PubMed] [Google Scholar]
  2. Carvalho, A.F. , Heilig, M. , Perez, A. , Probst, C. & Rehm, J. (2019) Alcohol use disorders. The Lancet, 394, 781–792. [DOI] [PubMed] [Google Scholar]
  3. Cullmann, M. , Hilding, A. & Östenson, C.G. (2012) Alcohol consumption and risk of pre‐diabetes and type 2 diabetes development in a Swedish population. Diabetic Medicine, 29(4), 441–452. Available from: 10.1111/j.1464-5491.2011.03450.x [DOI] [PubMed] [Google Scholar]
  4. Gmel, G. , Kuntsche, E. & Rehm, J. (2011) Risky single‐occasion drinking: bingeing is not bingeing. Addiction, 106, 1037–1045. [DOI] [PubMed] [Google Scholar]
  5. Griswold, M.G. , Fullman, N. , Hawley, C. , Arian, N. , Zimsen, S.R.M. , Tymeson, H.D. et al. (2018) Alcohol use and burden for 195 countries and territories, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. The Lancet, 392(10152), 1015–1035. Available from: 10.1016/s0140-6736(18)31310-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Hansagi, H. , Romelsjö, A. , Gerhardsson de Verdier, M. , Andréasson, S. & Leifman, A. (1995) Alcohol consumption and stroke mortality: 20‐year follow‐up of 15,077 men and women. Stroke, 26, 1768–1773. [DOI] [PubMed] [Google Scholar]
  7. Holmes, M.V. , Dale, C.E. , Zuccolo, L. , Silverwood, R.J. , Guo, Y. , Ye, Z. et al. (2014) Association between alcohol and cardiovascular disease: Mendelian randomisation analysis based on individual participant data. BMJ, 349, g4164. Available from: 10.1136/bmj.g4164 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Iwamoto, D.K. , Corbin, W. & Fromme, K. (2010) Trajectory classes of heavy episodic drinking among Asian American college students. Addiction, 105, 1912–1920. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Kitamura, A. , Sato, S. , Kiyama, M. , Imano, H. , Iso, H. , Okada, T. et al. (2008) Trends in the incidence of coronary heart disease and stroke and their risk factors in Japan, 1964 to 2003: the Akita‐Osaka study. Journal of the American College of Cardiology, 52(1), 71–79. Available from: 10.1016/j.jacc.2008.02.075 [DOI] [PubMed] [Google Scholar]
  10. Larsson, S.C. , Burgess, S. , Mason, A.M. & Michaelsson, K. (2020) Alcohol consumption and cardiovascular disease: a Mendelian randomization study. Circulation: Genomic and Precision Medicine, 13(3), e002814. Available from: 10.1161/CIRCGEN.119.002814 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Larsson, S.C. , Wallin, A. , Wolk, A. & Markus, H.S. (2016) Differing association of alcohol consumption with different stroke types: a systematic review and meta‐analysis. BMC Medicine, 14, 178. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Leppälä, J.M. , Paunio, M. , Virtamo, J. , Fogelholm, R. , Albanes, D. , Taylor, P.R. et al. (1999) Alcohol consumption and stroke incidence in male smokers. Circulation, 100, 1209–1214. [DOI] [PubMed] [Google Scholar]
  13. Liang, Y. , Mente, A. , Yusuf, S. , Gao, P. , Sleight, P. , Zhu, J. et al. (2012) Alcohol consumption and the risk of incident atrial fibrillation among people with cardiovascular disease. CMAJ: Canadian Medical Association Journal, 184, E857–E866. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Longstreth, W.T., Jr. , Nelson, L.M. , Koepsell, T.D. & Van Belle, G. (1992) Cigarette smoking, alcohol use, and subarachnoid hemorrhage. Stroke, 23, 1242–1249. [DOI] [PubMed] [Google Scholar]
  15. McKee, M. & Britton, A. (1998) The positive relationship between alcohol and heart disease in eastern Europe: potential physiological mechanisms. Journal of the Royal Society of Medicine, 91, 402–407. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Millwood, I.Y. , Walters, R.G. , Mei, X.W. , Guo, Y. , Yang, L. , Bian, Z. et al. (2019) Conventional and genetic evidence on alcohol and vascular disease aetiology: a prospective study of 500 000 men and women in China. The Lancet, 393, 1831–1842. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Ministry of Health, Labour and Welfare in Japan . (n.d.) Health Japan 21 (alcohol) . Available from: https://www.mhlw.go.jp/www1/topics/kenko21_11/b5.html (in Japanese). [Accessed 28th October 2024].
  18. O'Donnell, M.J. , Chin, S.L. , Rangarajan, S. , Xavier, D. , Liu, L. , Zhang, H. et al. (2016) Global and regional effects of potentially modifiable risk factors associated with acute stroke in 32 countries (INTERSTROKE): a case‐control study. The Lancet, 388, 761–775. [DOI] [PubMed] [Google Scholar]
  19. O'Donnell, M.J. , Xavier, D. , Liu, L. , Zhang, H. , Chin, S.L. , Rao‐Melacini, P. et al. (2010) Risk factors for ischaemic and intracerebral haemorrhagic stroke in 22 countries (the INTERSTROKE study): a case‐control study. The Lancet, 376(9735), 112–123. Available from: 10.1016/S0140-6736(10)60834-3 [DOI] [PubMed] [Google Scholar]
  20. OECD . (2019) OECD health statistics 2019. Paris: OECD Publishing. Available from: http://stats.oecd.org/ [Accessed 28th October 2024]. [Google Scholar]
  21. OECD . (2020) Tackling harmful alcohol use. In: OECD reviews of public health: Korea: a healthier tomorrow. Paris: OECD Publishing. Available from: https://www.oecd‐ilibrary.org/sites/3add27c4‐en/index.html?itemId=/content/component/3add27c4‐en [Online]. [Accessed 28th October 2024]. [Google Scholar]
  22. Palomaki, H. & Kaste, M. (1993) Regular light‐to‐moderate intake of alcohol and the risk of ischemic stroke. Is there a beneficial effect? Stroke, 24, 1828–1832. [DOI] [PubMed] [Google Scholar]
  23. Patra, J. , Taylor, B. , Irving, H. , Roerecke, M. , Baliunas, D. , Mohapatra, S. et al. (2010) Alcohol consumption and the risk of morbidity and mortality for different stroke types – a systematic review and meta‐analysis. BMC Public Health, 10, 258. Available from: 10.1186/1471-2458-10-258 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Puddey, I.B. , Rakic, V. , Dimmitt, S.B. & Beilin, L.J. (1999) Influence of pattern of drinking on cardiovascular disease and cardiovascular risk factors – a review. Addiction, 94, 649–663. [DOI] [PubMed] [Google Scholar]
  25. Puka, K. , Buckley, C. , Mulia, N. , Purshouse, R.C. , Lasserre, A.M. , Kerr, W. et al. (2023) Behavioral stability of alcohol consumption and socio‐demographic correlates of change among a nationally representative cohort of US adults. Addiction, 118, 61–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Rehm, J. , Baliunas, D. , Borges, G.L. , Graham, K. , Irving, H. , Kehoe, T. et al. (2010) The relation between different dimensions of alcohol consumption and burden of disease: an overview. Addiction, 105, 817–843. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Rehm, J. , Gmel, G.E., Sr. , Gmel, G. , Hasan, O.S.M. , Imtiaz, S. , Popova, S. et al. (2017) The relationship between different dimensions of alcohol use and the burden of disease‐an update. Addiction, 112, 968–1001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Reynolds, K. , Lewis, B. , Nolen, J.D. , Kinney, G.L. , Sathya, B. & He, J. (2003) Alcohol consumption and risk of stroke: a meta‐analysis. JAMA, 289(5), 579–588. Available from: 10.1001/jama.289.5.579 [DOI] [PubMed] [Google Scholar]
  29. Roerecke, M. (2021) Alcohol's impact on the cardiovascular system. Nutrients, 13(10), 3419. Available from: 10.3390/nu13103419 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Ronksley, P.E. , Brien, S.E. , Turner, B.J. , Mukamal, K.J. & Ghali, W.A. (2011) Association of alcohol consumption with selected cardiovascular disease outcomes: a systematic review and meta‐analysis. BMJ, 342, d671. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. SAS Institute Inc . (2013) SAS/STAT®13.1 user's guide. Cary, NC: SAS Institute Inc. Available from: https://support.sas.com/documentation/onlinedoc/stat/131/phreg.pdf [Accessed 28th October 2024]. [Google Scholar]
  32. Shimamoto, T. , Komachi, Y. , Inada, H. , Doi, M. , Iso, H. , Sato, S. et al. (1989) Trends for coronary heart disease and stroke and their risk factors in Japan. Circulation, 79(3), 503–515. Available from: 10.1161/01.cir.79.3.503 [DOI] [PubMed] [Google Scholar]
  33. Smyth, A. , Teo, K.K. , Rangarajan, S. , O'Donnell, M. , Zhang, X. , Rana, P. et al. (2015) Alcohol consumption and cardiovascular disease, cancer, injury, admission to hospital, and mortality: a prospective cohort study. The Lancet, 386(10007), 1945–1954. Available from: 10.1016/s0140-6736(15)00235-4 [DOI] [PubMed] [Google Scholar]
  34. Sull, J.W. , Yi, S.W. , Nam, C.M. , Choi, K. & Ohrr, H. (2010) Binge drinking and hypertension on cardiovascular disease mortality in Korean men and women: a Kangwha cohort study. Stroke, 41, 2157–2162. [DOI] [PubMed] [Google Scholar]
  35. Sull, J.W. , Yi, S.W. , Nam, C.M. & Ohrr, H. (2009) Binge drinking and mortality from all causes and cerebrovascular diseases in Korean men and women: a Kangwha cohort study. Stroke, 40, 2953–2958. [DOI] [PubMed] [Google Scholar]
  36. Sundell, L. , Salomaa, V. , Vartiainen, E. , Poikolainen, K. & Laatikainen, T. (2008) Increased stroke risk is related to a binge‐drinking habit. Stroke, 39, 3179–3184. [DOI] [PubMed] [Google Scholar]
  37. The Japanese Society of Hypertension . (2019) The guideline for the management of hypertension .
  38. Walker, A.E. , Robins, M. & Weinfeld, F.D. (1981) The National Survey of Stroke. Clinical findings. Stroke, 12, I13–I44. [PubMed] [Google Scholar]
  39. Wannamethee, S.G. & Shaper, A.G. (1996) Patterns of alcohol intake and risk of stroke in middle‐aged British men. Stroke, 27, 1033–1039. [DOI] [PubMed] [Google Scholar]
  40. World Health Organization . (n.d.) Data, Gho, Indicator Metadata Registry list, alcohol, heavy episodic drinking (drinkers only) past 30 days . Available from: https://www.who.int/data/gho/indicator‐metadata‐registry/imr‐details/458 [Accessed 28th October 2024].
  41. Yamagishi, K. , Iso, H. , Sairenchi, T. , Irie, F. , Takizawa, N. , Koba, A. et al. (2022) Diabetes mellitus modifies the association of serum triglycerides with ischemic cardiovascular disease mortality: the Ibaraki prefectural health study (IPHS). Journal of Atherosclerosis and Thrombosis, 29, 1319–1327. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Yamagishi, K. , Muraki, I. , Kubota, Y. , Hayama‐Terada, M. , Imano, H. , Cui, R. et al. (2019) The Circulatory Risk in Communities Study (CIRCS): a long‐term epidemiological study for lifestyle‐related disease among Japanese men and women living in communities. Journal of Epidemiology, 29, 83–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Zhang, C. , Qin, Y.Y. , Chen, Q. , Jiang, H. , Chen, X.Z. , Xu, C.L. et al. (2014) Alcohol intake and risk of stroke: a dose‐response meta‐analysis of prospective studies. International Journal of Cardiology, 174, 669–677. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Tables S1–S5

ACER-49-1117-s001.docx (96.4KB, docx)

Data Availability Statement

Research data are not shared.

Articles from Alcohol, Clinical & Experimental Research are provided here courtesy of Wiley

RESOURCES