An attempt to make a vaccine to treat Alzheimer’s disease, abandoned after it caused dangerous brain inflammation in some patients, may have done the patients some good, researchers reported on Monday.

They found evidence that the vaccine, made by Dublin-based Elan Corp. and Wyeth Pharmaceuticals, helped clear some of the brain-destroying “plaques” that characterize the disease.

Some vaccinated patients scored better on memory and a few other tests used to diagnose the fatal, incurable condition.

“We now need to see whether we can create an immune response safely and in a way that slows the progression of Alzheimer’s disease and preserves cognition,” Dr. Sid Gilman of the University of Michigan Medical School, who led one of the studies, said in a statement.

Alzheimer’s affects 15 million people around the world and 4.5 million in the United States. It gradually destroys the brain, causing memory loss and confusion that becomes so severe patients cannot care for themselves.

Elan’s experimental vaccine, AN-1792, targeted the beta amyloid proteins that form the brain-clogging plaques.

It has been shown to generate antibodies against beta amyloid, and Swiss researchers found the patients whose bodies made antibodies in response to the vaccine did not get any worse over the year following vaccination.

The vaccine trials were stopped in 2002 after 17 of 300 patients being tested developed a life-threatening inflammation of the brain called meningoencephalitis.

But the researchers continued to monitor the patients for a year after their last vaccination and were also able to examine the brains of some patients who died after the trial.

Writing in the journal Neurology, the researchers said volunteers whose immune systems mounted a response against beta amyloid did significantly better on a series of memory tests than patients given a placebo injection.

Magnetic resonance imaging scans also showed that the brains of patients who had a strong immune response shrank - possibly because the brain-clogging material was removed.

And they had reduced levels of tau, another Alzheimer’s-associated protein, in their spinal fluid.

“Three participants died of causes unrelated to the vaccination, two of whom had developed encephalitis and one other did not develop encephalitis,” Gilman said.

“All had large patches of their brains where beta amyloid had apparently been cleared out - the tangles of tau protein were still visible.”

Now the researchers are trying a new approach that uses antibodies against beta amyloid instead of the protein itself.

Last month another team reported that a small trial using antibodies seemed safe and may have delayed or even halted progression of the fatal disease.