Alcohol consumption is part of many cultures around the world. Approximately 7% of the world’s population live with an alcohol use disorder and approximately 3.7% of the world’s population live with alcohol dependence (WHO 2024), and this is a growing international problem.
Alcohol is widely recognised as a risk factor for a variety of diseases and is associated with mortality and morbidity around the world. Disproportionately, this affects younger people (WHO 2024), however there is growing evidence for the role of alcohol use in the development of dementia, a disease with a considerable burden on healthcare systems, caring networks and quality of life in old age (Alzheimer’s Society 2024). Previous studies have pointed to the increased risk of dementia in people who have a higher consumption of alcohol, even if this is infrequent (Adams, 2018; Crabbe, 2018; Koch et al 2019). The role of alcohol is not only of particular importance because of its ubiquitous nature in many people’s lives, but also represents a modifiable risk factor. Therefore, understanding the role which it plays in the development of dementia can help develop strategies to address it.
The paper reviewed here by Jeon et al (2023) sought to stratify the risks in the context of how people’s drinking habits changed over the duration of the study and what can be inferred from the changes in dementia risk, relative to drinking habits.
We know that alcohol is the number one modifiable risk factor for dementia, but how do changes in alcohol consumption affect the incidence of dementia?
Methods
The paper was a large-scale retrospective cohort study using the Korean National Health Insurance Service (NHIS) database and followed adults aged 40 years and older from 2009 to 2018. Analysis of the results took place in December 2021. Participants underwent two health examinations, one in 2009 and one in 2011, which were offered as part of the free biennial cardiovascular health screening provided by the NHIS. The final 3,933,382 participants were selected from an original group of 4,961,817, with exclusion criteria of previous diagnosis of dementia, cancer, cardiovascular disease, or death within one year of their second examination. Medical records for participants that lacked key variables, including information on alcohol consumption, were also excluded. The primary outcome was newly diagnosed dementia: either Alzheimer’s disease, vascular dementia or other type. The cohort was assessed from one year after the second health examination, continuing until dementia diagnosis, death or the end of the study in 2018.
Information relating to alcohol consumption was gained via participants completing self-reported questionnaires, which documented frequency and quantity of alcohol consumed. Total weekly alcohol consumption was approximated by converting the alcohol content of drinks to grams of alcohol and multiplied by frequency. Participants were then put into one of four groups relating to their alcohol consumption – none (0g per day), mild (30g per day) – and further put into groups relating to the change in their alcohol consumption from 2009 to 2011 (sustained non-drinkers, quitters, reducers, sustainers and increasers).
Results
The paper presents the crude results alongside two models which factor in increasingly more co-morbidities and show more nuanced and clinically informative results. The second model provides the results that are referred to in the paper and shows a statistically significant pattern of mild alcohol consumption in 2011 being associated with a reduced incidence of all-cause dementia, relative to either being a non-drinker or having moderate alcohol consumption in 2009.
The degree of biographical information about participants allows for interesting patterns to be identified. For example, the authors note that compared with people who continued to drink, people who stopped drinking alcohol tended to be older, female, non-smokers, be more engaged in regular exercise, and have lower incomes.
Mild alcohol consumption was associated with a reduced incidence of dementia relative to being a non-drinker or having moderate alcohol consumption.
Conclusions
The authors note that a J-shaped (or U-shaped) association between alcohol consumption and risk of all-cause dementia was present in their results, which is a phenomenon that has been noted in previous research (Sabia et al 2018). Although the results are consistent with previous studies, the exact mechanism of how mild-alcohol consumption can be associated with a reduced incidence of dementia remains unclear. The study comments on speculation that there may be promotion of prosurvival pathways and a reduction in neuroinflammation with mild alcohol intake, however the authors also note there is continuing debate on “numerous other [negative] outcomes [associated with alcohol consumption]”. There is discussion of the “sick quitter” phenomenon, in which a person stops a hazardous activity due to health issues. In this case, a person might stop drinking alcohol due to alcohol-related comorbid illness, but the impact on the person’s risk of dementia remains, even after stopping. Given the myriad associations between alcohol use and long-term health issues this seems a reasonable consideration, and the authors acknowledge this as a potential bias.
The lack of a clear reason for the U-shaped incidence of dementia is particularly important as the direct neurotoxic effect of alcohol is well described in other research and acknowledged by the authors in their own conclusions. Whilst these results are thought provoking and add further weight to the information about the association between alcohol and dementia, it is not as simple as suggesting that mild alcohol intake is inherently protective against the risk of dementia or that abstinence increases the risk.
People may stop consuming alcohol due to alcohol-related comorbid illnesses; however, their risk of developing dementia may persist.
Strengths and limitations
This paper has two main strengths; firstly the size of the population used and secondly looking at the impact of changes to alcohol consumption, which has not been well described previously. Using the NHIS, the study authors were able to capture data from nearly four million people, which goes a long way to minimising the issues of spurious results being amplified (as they might be in smaller studies). It is worth noting that a large sample size does not necessarily modify an underlying bias, although it will lead to narrower confidence intervals and consequently a result closer to the population average. Prior studies generally looked at the impact of a consistent level of alcohol consumption relative to the risk to health. Having access to the NHIS and the repeated nature of the health screenings allowed the authors to look at the impact of changing alcohol intake.
The authors concede that participants were self-selecting as not everyone would attend a health screening, and those that did so were more likely to be taking better care of their overall health. They also acknowledge that the consumption of alcohol was self-reported, which leaves the results open to consumption being underestimated. Whilst imperfect, other studies indicate that quantity-frequency measurements are likely to provide reliable measures in population-based surveys (McKenna et al 2018). They also comment that type (and therefore presumably, alcohol content) of drinks were not recorded, although previous studies suggest that there are not convincing arguments for one type of alcohol to be less harmful than another (Mäkelä et al 2011). The authors also comment on the degree to which genetics may impact the applicability of these results to other ethnic groups, due to difference in alcohol metabolism and the unmeasured cofounder of genetic risk factors such as the APOE genes.
It is also important to note that this study was a retrospective cohort study. This was a reasonable choice as the outcome measure was relatively common and a large amount of information about the cohort members was available to aid with identifying cofounding factors. However, the observational nature of the study means that correlation only, rather than causality, can be inferred from the results.
Alcohol consumption was self-reported in this study; it is possible that people underreported the levels of alcohol they were consuming.
Implications for practice
Alcohol consumption has been recognised as a modifiable risk factor for a number of diseases for many years (Koch et al 2019) and given the levels of consumption around the world, remains a topic of interest for many countries. It is also a topic reported on in the mainstream press both as a benefit and a harm (Guardian 2022). This paper helps add more information around the topic of alcohol consumption as a risk factor for dementia. This information could be used in support of discussions with older adults regarding the risks of moderate alcohol use and the benefits of reduction, when complete cessation may be undesired or unachievable. It may, however, further confuse the understanding of the risks of alcohol consumption by the general public, depending on how it is reported in the mainstream press.
Careful interpretation of these results is needed to ensure that practice is best informed.
Statement of interests
Dr Kube-Clare has no undisclosed interests.
Links
Primary paper
Jeon KH, Han K, Jeong SM, Park J, Yoo JE, Yoo J, Lee J, Kim S, Shin DW. Changes in Alcohol Consumption and Risk of Dementia in a Nationwide Cohort in South Korea. JAMA Netw Open. 2023 Feb 1;6(2):e2254771. doi: 10.1001/jamanetworkopen.2022.54771. PMID: 36745453.
Other references
Adams, S. Moderate and heavy alcohol consumption: what impact on later life brain and cognition? Alcohol consumption: what impact on later life brain and cognition? The Mental Elf, April 2018.
Alzheimer’s Society, The economic impact of dementia. (Accessed November 2024)
Crabbe, M. Alcohol is the number one modifiable risk factor for dementia. The Mental Elf, April 2018.
Guardian 14th July 2022 “Alcohol is never good for people under 40, global study finds”
Koch M, Fitzpatrick AL, Rapp SR, et al. Alcohol Consumption and Risk of Dementia and Cognitive Decline Among Older Adults With or Without Mild Cognitive Impairment. JAMA Netw Open. 2019;2(9):e1910319. doi:10.1001/jamanetworkopen.2019.10319
McKenna, H., Treanor, C., O’Reilly, D. et al. Evaluation of the psychometric properties of self-reported measures of alcohol consumption: a COSMIN systematic review. Subst Abuse Treat Prev Policy 13, 6 (2018). https://doi.org/10.1186/s13011-018-0143-8
Mäkelä P, Hellman M, Kerr W, Room R. A bottle of beer, a glass of wine or a shot of whiskey? Can the rate of alcohol-induced harm be affected by altering the population’s beverage choices? Contemp Drug Probl. 2011 Winter;38(4):599-619. doi: 10.1177/009145091103800408. PMID: 24431477; PMCID: PMC3888958.
Sabia S, Fayosse A, Dumurgier J, Dugravot A, Akbaraly T, Britton A et al. Alcohol consumption and risk of dementia: 23 year follow-up of Whitehall II cohort study BMJ 2018; 362 :k2927 doi:10.1136/bmj.k2927
WHO 2024, Alcohol Fact Sheet, (accessed November 2024)
