Drinking fruit and vegetable juice on a regular basis can dramatically reduce the chances of developing Alzheimer's, according to a study appearing today in the American Journal of Medicine.

Researchers in Japan and the US followed almost 2,000 dementia-free volunteers for up to 10 years while monitoring their consumption of fruit and vegetable juices. The risk of Alzheimer's was reduced by 76% for those who drank juices more than three times a week, compared with those who drank them less than once a week. Qi Dai, from Vanderbilt University in Nashville, Tennessee, wrote: "Fruit and vegetable juices may play an important role in delaying the onset of Alzheimer's disease."

Previous studies have suggested that polyphenols, strong antioxidants found in fruit juices, might disrupt biological processes in the brain that lead to Alzheimer's.

The findings come as three senior psychiatrists say today that memory clinics which assess whether people with Alzheimer's are suitable for controversial drug treatment are taking NHS resources and staff away from the care of those with more severe problems. In the British Medical Journal they criticise the rush to prescribe drugs that have little more than a placebo effect for people with mild to moderate Alzheimer's.

After a public battle, the National Institute for Clinical Excellence (Nice) recommended that cholinesterase inhibitors, the best-known of which is Aricept, should be given to people with mild to moderate disease.

But, say Anthony Pelosi, consultant psychiatrist at Hairmyres hospital in East Kilbride, and colleagues, the drugs offer little real benefit. Results of a number of trials "show quite consistently that these medicines have modest beneficial effects compared with placebo".

Carers of people with dementia report improvements when they are given the drugs, and with an identical but inert pill. One trial showed that fewer people were admitted to care homes within a year of taking the drugs - 9% compared with 14% who were not on medication - but within three years the numbers were identical. Dr Pelosi and his team say that in the drive to get these drugs to patients, memory clinics have been set up, sometimes with funding from the manufacturers. Medical staff which used to help those with dementia in the community were now running tests on those with early Alzheimer's to establish whether they are suitable for drugs.

It has been claimed that Nice's proposal to limit the drugs to those with mild to moderate disease would be devastating, he and his colleagues write. Clive Ballard, director of research at the Alzheimer's Society, said: "Anthony Pelosi et al make some sensible points but to make general claims about memory clinics is a mistake."