The experimental study at UCL examined whether the drug — used by doctors as a sedative or for pain relief — could “reprogramme” the craving for alcohol.
Ninety people involved in “harmful” drinking behaviour but not diagnosed as alcoholics were enrolled in the study.
All preferred beer and were drinking an average of 30 pints a week — five times the recommended limit.
Members of the group who were given ketamine after being reminded how much they liked alcohol showed “significant reductions” in their urge to drink over the next 10 days and halved their alcohol consumption over the next nine months.
The volunteers were divided into three groups. On the first day, all were given a beer and told they could drink it after being shown images of drinks and rating their “anticipated pleasure”.
On the second day they were given a beer which was taken away unexpectedly to “destabilise” the brain’s “reward memory” system.
A third of volunteers were given ketamine in a bid to reboot the way they thought about alcohol.
So-called “maladaptive reward memory” is thought to be central to drug and alcohol addiction.
Other participants were given a placebo or were given ketamine but without being reminded how much they liked drinking.
Dr Ravi Das, the study’s lead author, said: “Learning is at the heart of why people become addicted to drugs or alcohol.
"The drug hijacks the brain’s reward-learning system so that you end up associating environmental ‘triggers’ with the drug. These produce an exaggerated desire to take the drug.
“Unfortunately, once these reward memories are established it’s very difficult to re-learn more healthy associations but it’s vital in order to prevent relapse.”
The launch of the study was first revealed two years ago in the Standard.
The results were published this afternoon in the Nature Communications journal.
Professor Matt Field, of the University of Sheffield, said: “These findings are worthy of further investigation.
"However, the authors’ claim that they have found ‘lasting reductions in alcohol consumption after a single brief manipulation’ should be interpreted with caution.”