It is particularly impressive that just 1-3 drinks per day can reduce the risk of vascular dementia

Dr Richard Harvey, Alzheimer's Society

Out of 8,000 people who took part, 197 individuals developed dementia - of these, 146 had Alzheimer's disease, 29 developed vascular dementia and 22 got other types of dementia, it is reported in the Lancet medical journal.

The team suggests alcohol may have a direct effect on brain activity by stimulating the release of the chemical acetylcholine in the hippocampus area of the brain.

Acetylcholine is known to facilitate memory and learning processes, however high alcohol intake inhibits acetylcholine production.

Monique Breteler, who led the research, said: "In recent years, evidence has been accumulating that vascular factors may be involved in the cause of dementia, both vascular dementia and Alzheimer's disease.

"Our findings lend further support to the vascular hypothesis of dementia.

Limited intake

"We saw some indication for a stronger relation with alcohol in persons with a genetically determined susceptibility for Alzheimer's disease.

"Our findings can help focus research into the specific mechanisms that underlie the development of dementing illnesses."

Alzheimer's disease is the most common cause of dementia, accounting for 50% of all cases.

Vascular dementia accounts for about 20% of cases.

The Alzheimer's Society has welcomed the survey findings.

The society's research director Dr Richard Harvey said: "This interesting new study confirms the results of previous research which has suggested that light to moderate alcoholic consumption is actually good for our health.

"It is particularly impressive that just 1-3 drinks per day can reduce the risk of vascular dementia.

"Clearly, however, excessive alcohol consumption is not good for our long term health and increases the risk of serious diseases such as cirrhosis of the liver.

"It is very much the case of a little of what you fancy appears to do you good."

All those taking part in the research were aged 55+ and did not have dementia at the start of the study.