Earlier today, Eli Lilly and Co. conceded that topline results from its Phase 3 EXPEDITION3 trial indicate the anti-Aβ antibody solanezumab failed to slow cognitive decline in people with mild Alzheimer’s disease. According to a Lilly press release and a subsequent media call, results on primary and secondary outcomes trended in the direction of a treatment benefit, but the effects were small. A Lilly spokeswoman told Alzforum the company will not continue open-label extensions for EXPEDITION, EXPEDITION2, and EXPEDITION3 and will work with investigators to conclude those studies. 

Three additional trials are ongoing to test solanezumab at earlier disease stages. They are the DIAN trial in autosomal-dominant mutation carriers, the A4 trial in cognitively normal people with brain amyloid deposition, and EXPEDITION-PRO, a trial in 2,450 people with prodromal AD that Lilly just started this past summer. Researchers running the A4 and DIAN trials testing solanezumab in a prevention strategy told Alzforum that they expect those studies to continue. In a conference call, Lilly’s Eric Siemers said the company needs to sort through the data before making a decision on those other trials, which are funded jointly through public-private partnerships. (In 2014, when the gantenerumab prodromal AD trial SCarlet-RoAD was cut short, the DIAN trial was able to continue in part because joint funding of DIAN left the decision in the hands of that public-private partnership, as well (see Dec 2014 news). 

“The news is terribly disappointing for patients who have mild Alzheimer’s,” noted Reisa Sperling, Brigham and Women’s Hospital, Boston. Sperling is principal investigator of the A4 trial. “However, that the data all trended in the right direction in a population with dementia supports the idea that we need to treat earlier [in disease],” she told Alzforum. Though cognition declined more slowly in patients treated with solanezumab than in those taking placebo, the difference missed statistical significance (p=0.095). The company released no further data in today’s announcement

Lilly researchers set up EXPEDITION3 when subgroup analysis of EXPEDITION and EXPEDITION2 trials of people with mild to moderate AD suggested a benefit only for patients with mild disease. Lilly then designed what it hoped would be a “confirmatory” trial of that pooled mild AD subgroup, largely repeating the EXPEDITION trial design. EXPEDITION3 recruited 2,100 patients with a diagnosis of mild AD; the main difference was that participants had to have a positive florbetapir PET scan for brain amyloid. In the earlier EXPEDITION trials about 30 percent of the subset of participants who had amyloid PET scans turned out to have no brain amyloid deposition. EXPEDITION3 was powered to detect a 1.5 to 2 point difference on the ADAS Cog14 over 18 months.

“Clearly, if all the stars aligned then the trial would have been positive,” noted Lon Schneider, University of Southern California, Los Angeles. “But that was based on the assumption that the trial would have enrolled the same kind and distribution of AD patients that they obtained by pooling EXPEDITION1 and 2, and that the patients would have had the same responses” he said. “This goes to show just how variable AD and clinical trials can be,” he added.

After the EXPEDITION1/2 pooled data came out, there were calls for the FDA to approve solanezumab, and some criticism that its insistence on a confirmatory trial might deprive patients of an effective treatment. Today’s news validates the agency’s call, said Schneider. “This shows the FDA’s wisdom in not bowing to pressure,” he said.

Even if the trial had met its primary endpoint, Schneider questioned how effective solanezumab would be. “Would a 1.5 point difference on the ADAS-Cog over 18 months be meaningful?” he asked, noting that acetylcholinesterase inhibitors can show a 2 point difference or more within six months. “We should be aiming for drugs that have larger and more robust effects earlier in development,” he told Alzforum.

Lilly’s stock dropped about 11 percent since the announcement. Shares of Biogen, which develops the anti-Aβ antibody aducanumab, currently in Phase 3, dropped as well, though scientifically speaking, the closest relative to solanezumab is crenezumab (see May 2015 news). Crenezumab earlier this year entered a Phase 3 trial in prodromal AD called CREAD.

In the press call today, Lilly’s Dave Ricks said that the company had spent $3 billion on Alzheimer’s research and development in the past 27 years.

Alzheimer’s researchers today were grappling with the implications. “Though the news is disappointing, it is not completely unexpected to those who have supported prevention trials,” said Randall Bateman, Washington University, St. Louis, who leads the DIAN-TU prevention trial. “The evidence is that treating early is likely to be more effective. Work over the past decade supports that conclusion, and has also emphasized the need to consider higher doses of monoclonal antibodies than were used in some earlier trials,” he said. That’s not to say Bateman and colleagues were not disappointed. “It would have been great if there was a larger effect size,” he said. Bateman expects the DIAN solanezumab trial to continue.

What does this mean for the amyloid hypothesis? “This trial was clearly a shot on goal, so this result is a worry,” said John Hardy of University College London. To Hardy’s mind, today’s news raises questions about BACE inhibitors because, like solanezumab, they reduce soluble Aβ. The stock price of Merck, which develops the Phase 3 BACE inhibitor verubecestat, dropped as well today. Hardy believes that implications for other therapies are less clear. Aducanumab, for example, targets plaques, not soluble Aβ.

What next? “We need to learn from this,” Hardy said. “Lilly have taken the high road so far with their openness with clinical trial data. Let's hope that we can use the data on biomarkers and secondary analyses to understand what we need to do next.” Lilly’s Siemers said during the press conference: “We anticipate working with our academic colleagues to look at the data together, to learn as much as we can from the EXPEDITION3 trial.”

Lilly will present additional data at the upcoming CTAD meeting on December 8, in San Diego.—Tom Fagan and Gabrielle Strobel


  1. After consultation with Lilly, the DIAN-TU leadership has decided the EXPEDITION3 results do not immediately impact the DIAN-TU trial (DIAN-TU-001). Solanezumab will continue to be tested in the DIAN-TU trial for the following reasons:

    • The DIAN-TU trial is a prevention study treating individuals at much earlier stages of the disease. The main hypothesis of the DIAN-TU trial is that treating at earlier stages of disease will be more effective. We continue to hypothesize that intervention earlier in the disease, such as the preclinical stage of AD being tested in the DIAN-TU trial, may have a better chance of demonstrating slowing of cognitive decline with a larger effect. 
    • The DIAN-TU trial is studying a different population of mutation-causing early onset AD compared to sporadic late-onset AD in EXPEDITION3. These younger patients with a pure form of Alzheimer’s may have a larger benefit from drugs that target Aβ.
    • The DIAN-TU trial primary outcome includes novel cognitive measures and composites and also biomarker measures of Alzheimer’s disease that will be important to understand how much solanezumab may slow or prevent Alzheimer’s disease progression.
  2. Lilly researchers set up EXPEDITION3 when a subgroup analysis of EXPEDITION1 and EXPEDITION2 trials suggested a benefit only for patients with mild disease. Why did EXPEDITION3 not confirm apparent positive findings of the two prior trials in mildly affected patients?

    The main difference between EXPEDITION3 and previous studies was that each participant had to have a positive florbetapir-PET scan for brain amyloid to be enrolled. I guess this inclusion criteria was imposed to ensure that participants had proof of a typical AD pathological process ongoing. In addition, I guess there was hope that the solanezumab's cognitive and clinical effects could be greater in patients with confirmed Aβ deposition in the brain.

    Indeed, we know that in EXPEDITION1 and 2, about 30 percent of the subset of participants who had amyloid PET scans had no brain amyloid deposition. On the other hand, there are no indications from these prior studies that solanezumab’s positive effects were greater in amyloid-positive than amyloid-negative subjects. More importantly, I see a strident contradiction in the strategy of enrolling subjects at an earlier clinical stage (“mild AD patients”) and at the same time enrolling only amyloid-positive patients who we know are more advanced in the disease process.

    Thus, while the approach of enrolling AD patients with confirmed brain Aβ deposition load will ensure more homogenous inclusion criteria in terms of AD pathology, at the same time this criterion determines enrollment of patients at a more advanced stage. This apparent contradiction may affect other studies that are targeting early stages of the disease process, even at a presymptomatic phase, but at the same time are including only subjects with confirmed presence of Aβ deposition in the brain.

    To confirm this hypothesis that Aβ positive subjects were at a more advanced stage of the disease, it would be interesting to see if mean cognitive scores at baseline of the EXPEDITION3 study suggest that a more severely affected population was enrolled compared with the population of mild AD patients in the EXPEDITION1 and EXPEDITION2 trials.

  3. All anti-Aβ antibodies failed to slow cognitive decline in people with mild Alzheimer’s disease. It's time to begin another line of research. In the future we´ll have to try combined treatment and only in young patients with early Alzheimer's. Alzheimer´s disease in older patients needs to be re-explored.

  4. We are grateful that Lilly tried so hard, and will be interested in their conclusions about the utility of amyloid imaging and biomarkers.

    With the 2025 "deadline" looming, next steps might include a more aggressive empiric repurposing approach, with recent forward-looking strategy reviews, the NIA's RFP for a robust clinical trials organization, and the recent Alzheimer's Combination Therapy Opportunity RFP. Aggressive clinical leadership could be helpful (see Seattle Times op-ed).

  5. Unfortunately in the absence of means to image soluble Aβ in brain, there is no way to know whether the treatment actually diminished this target. In fact, antibodies directed against soluble Aβ might cause enhanced release of soluble Aβ from insoluble plaques and might also cause increased production from APP. Abandoning patients with well-developed AD may be tragically premature if negative results are misinterpreted because of the unproven assumption that soluble Aβ was significantly reduced. This was extensively discussed in Rosenblum, 2014.


    . Why Alzheimer trials fail: removing soluble oligomeric beta amyloid is essential, inconsistent, and difficult. Neurobiol Aging. 2014 May;35(5):969-74. Epub 2013 Oct 19 PubMed.

  6. It is an attractive idea to remove small soluble species of toxic Alzheimer’s Aβ from an Alzheimer’s brain. On paper it would have seemed quite efficacious—providing that the different species of Aβ are not being generated by an ongoing chronic infectious disease. An infectious origin is a possibility subscribed to by Itzaki et al., Mawanda and Wallace, and others. What is making the Aβ toxic? Mawanda and Wallace maintained that there is emerging evidence that supports an infectious pathogen. They said this:

    “In addition, amyloidopathy—a condition characterized by elevated levels of serum amyloid and by amyloid deposition and aggregation in tissues—is a frequent occurrence in several acute and chronic systemic inflammatory conditions, especially chronic infections like tuberculosis and leprosy.”

    Before a condition such as Alzheimer’s disease can be treated successfully, its pathophysiology needs to be better defined. In that area, there is much room for improvement.


    . Microbes and Alzheimer's Disease. J Alzheimers Dis. 2016;51(4):979-84. PubMed.

    Are the Infectious Roots of Alzheimer’s Buried Deep in the Past? . J of MPE Mol Pathol Epidem 2017, Feb Vol 3 No S2:2: 1-19

    . Can Infections Cause Alzheimer's Disease?. Epidemiol Rev. 2013 Jan 24; PubMed.

    . CWD Tuberculosis Found in Spongiform Disease Formerly Attributed to Prions: Its Implication towards Mad Cow Disease, Scrapie and Alzheimer’s. J of MPE Mol Pathol Epidem 2017, April Vol 3 No 3:3: 1-13.

Make a Comment

To make a comment you must login or register.


Therapeutics Citations

  1. Solanezumab
  2. Aduhelm
  3. Crenezumab
  4. Verubecestat

News Citations

  1. End of the RoAD for Gantenerumab? Roche Declares Prodromal Alzheimer’s Trial Futile
  2. Shape of a Hug: How the Embrace of a Therapeutic Aβ Antibody Really Matters

Conference Citations

  1. Clinical Trials on Alzheimer's Disease 2016

External Citations

  4. announcement

Further Reading

No Available Further Reading