WASHINGTON -- Eli Lilly & Co. reported Wednesday that an experimental medication might slow mild Alzheimer's if people take it early enough, one of a handful of drugs in late-stage testing in the frustrating hunt for a better treatment.

The new findings don't prove that Lilly's solanezumab really works; a larger study is underway that won't end until late 2016. On Wednesday, researchers at the Alzheimer's Association International Conference updated ongoing research into Lilly's effort, and those of two competitors, that aim to fight Alzheimer's by targeting a sticky protein that clogs the brain.

The bottom line: It's going to take more time to figure out if this approach works, but scientists think it's still the right target despite high-profile failures in recent years.

"These are not definitive reports that are going to lead to medications being approved tomorrow. What they represent is an important foundation for us moving forward," said Dr. David Knopman of the Mayo Clinic, who has monitored some of Lilly's data.

A look at the candidate drugs:

SOLANEZUMAB

In previous studies, solanezumab failed to help Alzheimer's patients overall. But there was a signal that the drug might help people with very mild disease, apparently slowing their mental decline by about a third. Lilly continued to track that subset of 1,300 milder patients, giving everyone the drug for another two years, including patients who had initially received a placebo.

The group that took solanezumab from the beginning fared better than the group that started later, although the difference was small -- about two points on cognitive test scores, Lilly researchers reported. They couldn't say what that might mean for daily functioning.

"Our hope is that slowing will be increasingly noticeable over time," said lead researcher Dr. Paul Aisen of the University of Southern California.

GANTENERUMAB

Roche stopped one late-stage study of gantenerumab last December because it didn't meet its main goal of a cognitive benefit in people with early Alzheimer's symptoms. Wednesday, researchers said a closer analysis found some patients did show signs of less amyloid in their brains, and another Alzheimer's-related protein named tau improved, too.

Probably, the gantenerumab doses were too low, concluded lead researcher Dr. Philip Scheltens of the VU University Medical Center in Amsterdam.

Gantenerumab is being tested in two additional large studies, and doses used in them will be raised, said Roche's Dr. Robert Lasser.

ADUCANUMAB

Biogen released results of its small, first-stage study of aducanumab last spring, showing that it appeared to slow cognitive decline in some patients. On Wednesday, it provided additional details confirming a higher dose worked better. One side effect, a brain fluid build-up in some people, tended to clear in several weeks and was "both monitorable and manageable," said Biogen clinical director Jeff Sevigny.

A large late-stage study is beginning.