Even the merely forgetful, whose lives have yet to be derailed by the disease, have become candidates for clinical tests. Scanning the brain for telltale signs of Alzheimer’s earlier — and ratcheting up drug dosage — may help turn failures into breakthroughs, said Andrea Pfeifer, chief executive officer of AC Immune SA. The Swiss company’s partner on its experimental drug, pharmaceutical giant Roche Holding AG, said this week it will embark on a second advanced test, undeterred by a failed intermediate trial.

Pfeifer’s optimism is a case study of the industry’s unusually high risk tolerance on Alzheimer’s, where more than 100 experimental drugs have already failed — with two recent setbacks from Merck and Eli Lilly. Like Roche and its partner, Biogen is chipping away at the same target that has eluded rivals. Merck and Lilly have opted to keep testing their medicines in patients showing only hints of the degenerative disease. Each advanced trial can cost hundreds of millions of dollars, according to Lilly.

“It’s always a question — are you throwing good money after bad?” said Birgit Kulhoff, a fund manager for Rahn & Bodmer in Zurich. “If you were to be successful, it would be an extremely big market.”

Drugs that halt the progress of Alzheimer’s disease could be a market worth as much as $30 billion in the U.S. alone, Sanford C. Bernstein analysts estimate.

Cancer is the only other disease for which companies are this willing to look deep into a failed study for a positive trend to chase, Kulhoff said, and Alzheimer’s tests are probably more expensive because researchers follow patients for longer and monitor them closely.

The drugmakers wouldn’t disclose how much they have spent on clinical tests. A final-stage trial in Alzheimer’s disease can cost north of $200 million, according to Eric Schmidt, an analyst with Cowen. Such a study in a cancer like melanoma would tend to be much smaller and could cost about $50 million to $100 million, he said.

Just arranging and administering the brain scans needed to make sure people have signs of the disease cost about $15,000 to $20,000 per person in a small test, according to AC Immune.

Roche made its cost-benefit calculation two and a half years ago. Crenezumab, the medicine it’s working on with AC Immune, had failed to help patients think more clearly or retrieve memories in a mid-stage trial, a hurdle potential drugs usually must clear before they go into the last, largest and most expensive stage of clinical studies.

Instead of abandoning the compound, Roche looked at a trend toward benefit in the least sick patients and spent the next two years figuring out how high it could safely ratchet up the dose. Then last year, it started another 750-patient study, with results expected in 2021.

Now it’s embarking on a separate trial, also with 750 patients with a very mild form of the disease. It’s unclear how that test will differ from the first one. It’s also trying a higher dose for another Alzheimer’s compound, gantenerumab, after a late-stage trial failure.

Both Roche drugs are aimed at protein pieces called beta amyloid that clump together in the brain and are thought to play a role in triggering the disease. Trouble is, so did the Lilly medicine that failed a test in patients last year. Last month’s unsuccessful trial from Merck also centered on beta amyloid as a target, although the type of medicine was thought to prevent the substance from forming. Those failures have cast doubt on the whole approach.

“Clearly those are two strikes against that theory,” Brent Saunders, the chief executive of U.S. specialty-drug maker Allergan Plc, said in New York last week. “My hope is that we’ll keep going, but we’ll see.”

Allergan is thinking hard before starting patient trials on an experimental drug that belongs to the same family of drugs as Merck’s after seeing the two recent failures, according to Saunders. The company also sells medicines that temporarily ease symptoms of the disease — the sole form of treatment currently available.

Alzheimer’s disease is on the rise across wealthy countries as their populations age. Globally, dementia unseated AIDS as one of the top 10 killers in the World Health Organization’s latest rankings.

“We’re not going to have a reversible cure any time in the near future,” said Phyllis Ferrell, who leads Lilly’s Alzheimer’s program. Still, “I think in our lifetime we will have something that slows this disease.”

Darryle Schoepp, the head of neuroscience research at Merck, acknowledges that the industry may have been “a little too optimistic” about Alzheimer’s trials, but said all the years of painful outcomes have taught companies how best to test their theories. While Merck is also looking at what’s next beyond beta amyloid, Schoepp says the big unanswered question is whether stopping the buildup before dementia begins can make a dent in the disease — something it’s testing in another big final-stage trial.

“The disappointing news is that once you have mild to moderate dementia, stopping the production of plaque will not help you,” Schoepp said.

Similarly, for AC Immune CEO Pfeifer, all the failures are signposts on the path toward eventual success, as scientists get better at peering into the brain, identifying Alzheimer’s potential before patients become very sick and figuring out which antibodies attack amyloid beta at its most toxic stages. She compared the Alzheimer’s field to where the study of cancer stood when she was a young oncology researcher at the National Institutes of Health in the U.S. some 30 years ago.

“We are now finally getting to the molecular level where I think the science in Alzheimer’s has to go,” Pfeifer said. “We were just not scientifically exact enough. That’s where I have big hopes.”

Another drugmaker that’s faced questions (despite largely positive results) is still gung-ho about the beta amyloid route. Gilmore O’Neil, an executive in clinical development at Biogen, says many industry studies so far have looked at patients who were already so sick that they’d probably be beyond saving even if the hypothesis were true — or patients that may not have had Alzheimer’s disease at all.

Brain Swelling

After promising initial findings, Biogen pushed its top compound, aducanumab, into final-stage studies immediately without running an intermediate study to confirm the early findings. Subsequent analysis wasn’t all positive: lower doses didn’t show a cognitive benefit and the drug was linked to a swelling of the brain.

Lilly also hasn’t abandoned the quest. It has 10 projects being tested in humans, including a molecule, partnered with AstraZeneca Plc, in the same family as the Merck compound that failed most recently. Like other drugmakers, it’s seeking earlier patients — including in some cases people who merely have brain scans showing amyloid and aren’t showing signs of forgetfulness at all.

“It’s like cancer,” said Ferrell. “You want to catch it at stage one, not at stage four.”

(c) 2017, Bloomberg · Naomi Kresge, Doni Bloomfield