Alcohol 'linked to lower dementia risk'

Friday March 4 2011

A glass of wine a day “cuts dementia risk”, according to The Daily Telegraph. The newspaper reports that a study found that people who drank two to three units a day were 29% less likely to develop dementia over three years.

The study followed 3,202 German people aged 75 years or older who were free of dementia. Researchers assessed their alcohol consumption and then followed them for three years to look for new diagnoses of dementia. In total, 217 developed dementia, with 111 of these cases due to Alzheimer’s disease. Overall, those drinking any amount of alcohol had a lower risk of dementia compared with teetotallers. However, when participants were subdivided by the amount they drank, only an intake of 20-29g a day was associated with a significantly reduced risk.

While this research is of some value, uncertainty behind some of the results, along with limitations in the study’s design, means that the association between alcohol and lower dementia risk should be interpreted with caution. UK alcohol recommendations do not change in light of this research. These specify a daily maximum of three to four units for men, and two to three for women, with one unit being equal to 10ml (8g) of alcohol, or half a pint of weak beer or lager.

Where did the story come from?

The study was carried out by researchers from the Central Institute of Mental Health in Mannheim and other institutions in Germany. It was funded by the German Federal Ministry of Education and Research. The study was published in the peer-reviewed medical journal Age and Ageing.

The Daily Telegraph has accurately reported the calculated findings of this paper. However, it did not discuss the wider limitations of this type of study. These limitations mean that the claim made by the authors that a glass of wine a day “cuts dementia risk” is not so clear cut.

What kind of research was this?

This was a cohort study that aimed to investigate whether current alcohol consumption among a cohort of dementia-free elderly people was associated with the development of overall dementia and Alzheimer’s disease over a period of follow-up.

As the authors say, it is known that long-term excessive use of alcohol is detrimental to brain function and could be associated with neurodegenerative diseases. However, previous studies have demonstrated that light-to-moderate alcohol intake could actually decrease the risk of cognitive decline. Previous studies have looked at people under the age of 75 years, but this study aimed to look at people over this age.

A cohort study can be used to examine whether a particular exposure is associated with a certain disease outcome during a follow-up period. However, this study assessed self-reported alcohol intake at one point, which may not be truly representative of life-long drinking patterns. Also, as the exposure was assessed when participants were already in old age, with dementia incidence assessed only a few years later, it is possible that those people who went on to develop dementia were already undergoing the processes of cognitive decline when the study started and drinking was assessed.

What did the research involve?

The cohort study was conducted across six primary care centres in Germany. GPs recruited people of 75 years whom they did not consider to have dementia (the authors say that the GPs used a “screening instrument with good psychometric properties” to identify cases).

A total of 3,202 participants were included. They were interviewed at home by trained assessors, and then interviewed again 1.5 and 3 years later. At their first interviews they were asked “At present, on how many days per week do you drink alcohol?”, with responses of “never”, “1–2 days”, “3–4 days”, “5–6 days”, “7 days” or “I do not know”. Those who drank were then asked about the quantity and type of alcohol. Based on this the researchers calculated their intake of pure alcohol, which was categorised as abstinent, 1–9g, 10–19g, 20–29g, 30–39g or 40g or more. The type of alcohol they consumed was classed as abstinent, wine only, beer only or mixed (wine, beer and other alcoholic beverages).

For the first follow-up session 85% of the cohort was available and 74% for the second. For those that could not be reassessed the researchers contacted GPs, relatives and caregivers. GPs were also contacted for every participant to obtain information on dementia diagnoses or mild cognitive impairment (diagnosed according to standard criteria). GPs were also asked to complete questionnaires on the person’s other medical illnesses, which they defined as no comorbidity, one to four other medical diagnoses or five or more other medical diagnoses.

The researchers calculated the associations between alcohol intake at the start of the study (the baseline) and the incidence of dementia (due to any cause) and specifically due to Alzheimer’s disease. In their calculations they took into account the potential confounders of sex, age, education, living situation, impairment in activities of daily living, comorbidity, depression, mild cognitive impairment, smoking, and apolipoprotein E (apoE) genetic status (some research has linked the gene to Alzheimer’s disease).

What were the basic results?

At baseline, 50% were abstinent, 24.8% consumed less than 1 drink per day (10g of alcohol), 12.8% drank 10–19g and 12.4% drank 20g or more. A small subgroup of 25 participants fulfilled the criteria of harmful drinking (more than 60g of alcohol per day for men, and more than 40g for women). Of those who drank any amount almost half (48.6%) drank only wine, 29% drank beer only and 22.4% drank mixed alcohol beverages.

Over three years of follow-up, there were 217 new cases of dementia (6.8% of subjects), and of these 111 people (3.5% of total subjects) were diagnoses of Alzheimer’s disease. Compared with abstinence, drinking any alcohol was associated with a 29% reduced risk of overall dementia (hazard ratio [HR] 0.71, 95% confidence interval [CI] 0.53 to 0.96) and a 42% reduced risk of Alzheimer’s (HR 0.58, 95% CI 0.38 to 0.89).

Assessment by subcategory of volume of alcohol consumed gave a variable association. For example, compared with abstinence, categories of 1–9g, 10–19g, 30–39g and 40g or more were associated with non-significant reduced risk of overall dementia, but 20-29g was associated with a significantly reduced risk (HR 0.40, 95% CI 0.17 to 0.94). Similarly, for Alzheimer’s disease, categories of 1–9g, 10–19g, 30–39g and 40g or more were associated with non-significant reduced risk, but 20-29g gave a significant reduced risk (HR 0.13, 95% CI 0.02 to 0.95).

How did the researchers interpret the results?

The researchers conclude that, in agreement with previous studies that included younger age groups, their study suggests that light-to-moderate alcohol consumption is related to a lower incidence of dementia in people aged 75 years and older.


This study is valuable in that it has included a large number of elderly people without dementia and used thorough follow-up methods over three years to identify all new cases of dementia (due to any cause), and Alzheimer’s disease in particular. The main findings were a reduction in the incidence of dementia and Alzheimer’s in people with any alcohol intake compared with those consuming none. However, understanding the true relationship between quantity of alcohol intake and associations with dementia is more difficult:

  • When subdividing participants into groups by level of alcohol intake, it should be remembered that the number of dementia diagnoses in each of these six categories becomes much smaller, reducing the chance of a precise risk estimate being calculated.
  • The pattern seen across these different drinking levels was also unclear, with a trend for reduced risk with any amount of alcohol consumption compared with abstinence, but a significant risk reduction only for those consuming 20-29g daily.
  • Even for this significant association with 20–29g daily (both for overall dementia and Alzheimer’s) the confidence intervals were very wide. This suggests that this risk association should be interpreted with some caution.

Overall, this makes it difficult to conclude with any certainty that light-to-moderate intake is the most beneficial quantity of alcohol to drink to reduce dementia risk.

There were also some limitations within the design of the study that must be considered:

  • People self-reported their alcohol intake at only one point in time. It is difficult to know how truly representative this is of life-long drinking patterns. As with assessment of other types of food and drink, reporting of quantities of alcohol and strength of alcohol consumed may vary between people, resulting in inaccurate categorisation.
  • Alcohol exposure was assessed when participants were already in old age, and dementia incidence assessed only three years later. It is therefore possible that people who went on to develop dementia were already undergoing cognitive impairment at the time that the study started and the exposure was assessed. It could be the case that they stopped drinking as a consequence, which might be one alternative explanation for the link demonstrated in the data. It is not possible to imply cause and effect from this sort of study.
  • Lastly, the study was conducted in Germany, and alcohol drinking patterns, dementia patterns, and incidence of other lifestyle and medical risk factors may vary between countries.

The possible association between alcohol intake and cognitive and memory effects is worthy of further study. However, UK alcohol recommendations do not change. These specify a daily maximum intake of three to four units for men, and two to three for women. One unit contains 10ml (8g) of pure alcohol, the amount typically found in half a pint of weak beer or lager.

The study does not provide a reason to drink. As well as the recognised health dangers of heavy drinking, high daily intake of alcohol has been linked to an increased risk of dementia.

Analysis by Bazian
Edited by NHS Choices