A year after announcing an early end to the EPOCH trial of the BACE inhibitor verubecestat in mild to moderate Alzheimer’s disease, Merck pulled the plug on the APECS study, a trial of prodromal AD patients. In a February 13 press release, the company said the trial would shut down a year early based on the results of an interim safety analysis. An external data-monitoring committee judged it “unlikely that positive benefit/risk could be established,” according to the release. Researchers at Merck did not respond to requests for comment, and the future for verubecestat’s clinical development remains unclear.

Scientists in the field said the results were disappointing but not surprising, given that animal studies have now shown that BACE inhibitors do little to halt the growth of established plaques (Jan 2018 news). People with cognitive symptoms are already nearing the plateau of amyloid plaque deposition, noted Robert Vassar at Northwestern University in Chicago. They have had amyloid deposition in their brains for well over a decade. Stefan Lichtenthaler at the German Center for Neurodegenerative Diseases (DZNE) in Munich voiced the prevailing view. “Prodromal AD is simply not early enough,” he wrote to Alzforum. Inclusion criteria for APECS included evidence of brain amyloid, a history of subjective cognitive complaint, and impaired episodic memory.

Like other BACE inhibitors in development, verubecestat slashes production of Aβ42, lowering levels in cerebrospinal fluid and blood by up to 90 percent (Nov 2016 news). Nonetheless, for AD patients in the EPOCH trial, the compound provided no clinical benefit. Treatment nudged down plaque load by 4 percent, but some side effects cropped up, most notably skin rashes and behavioral changes such as anxiety, insomnia, and falls (Dec 2017 conference news). 

These EPOCH data fueled the debate over how early BACE inhibitors would have to be given to be effective. The APECS results now push this proposed treatment window to even earlier stages. Vassar said the findings support the idea that inhibitors will have their greatest effect near the beginning of amyloid deposition, and will most likely need to be taken at low doses for years.

For his part, Jochen Herms at DZNE noted that the age of participants could be a complicating factor. He believes participants should be at most 70, because older adults often have comorbidities such as α-synuclein, TDP-43, or vascular pathology that could cloud cognitive outcomes. “It is unlikely that a trial will ever be positive if we continue to enroll very old MCI/AD patients,” Herms predicted (see comment below). The Merck trial enrolled participants as old as 85.

Other ongoing trials are already testing BACE inhibition in younger, presymptomatic populations. The A5 EARLY trial administers Janssen’s JNJ-54861911 to amyloid-positive people age 60 or older, and the Alzheimer’s Prevention Initiative tests Novartis’ CNP520 in a similar population (Dec 2017 conference news). Meanwhile, DIAN researchers are planning to give BACE inhibitors to young adults with a familial AD mutation but no detectable amyloid deposition (Aug 2017 conference news). “Primary prevention is the ideal stage for an [Aβ] production inhibitor,” Randall Bateman at Washington University in St. Louis explained (see comment below).

Despite this latest setback, Paul Aisen at the University of Southern California ATRI Institute in La Jolla, California, remains optimistic about the future of this class of drugs. “This discontinued trial does not really dampen our enthusiasm for BACE inhibitors … it remains very likely that they will be a mainstay of AD therapy, particularly in primary prevention,” he wrote (see comment below).

The APECS findings also renewed calls by some to pair BACE inhibition with plaque removal at later disease stages. “I would like to see this outcome encourage consideration of combination strategies, which may limit risk and allow treatment at ever-so-slightly later stages of disease where success may be easier to measure, if admittedly harder to achieve,” Joanna Jankowsky at Baylor College of Medicine, Houston, wrote to Alzforum (see comment below).—Madolyn Bowman Rogers


  1. Too much was expected from this clinical trial!

    We know that amyloid pathology alters axons at plaques—a pathology that we cannot rescue either with BACE inhibition or Aβ immunization. We can only stay axonal pathology at plaques with BACE inhibition, we can’t reverse it (Peters et al., 2018). 

    If we start with BACE inhibition when plaque load is low (and BACE inhibition stops the formation of new plaques as well as the further growth of existing) then I am convinced that we hold back amyloid pathology-related secondary pathology (though this does not mean we will see a benefit in cognition in short term trials). So yes, primary prevention is surely the best approach (if at low doses in order to avoid synaptic side effects of BACE inhibition), for example in ApoE positives 10–15 years before cognitive symptoms.

    I do not see that this trial result casts any doubt on the amyloid hypothesis or other anti-amyloid approaches. I would rather say that the design of the trial was flawed. If they had treated selected MCI patients with a mean age of 70, as Biogen did with the aducanumab immunization study (Sevigny et al., 2016), the chances of seeing a beneficial effect would have been much higher. Treating patients with a mean age of up to 80–85 years means lots of patients within the cohort have other pathology as well—especially α-synuclein and TDP43 deposits and vascular damage. To me it is unlikely that a trial will ever be positive if we continue to enroll very old MCI/AD patients. From a perspective of a neuropathologist (and maybe more aware of the age-related co-pathology) it does not make sense. We should face that we may only be able to treat younger MCI/AD patients once—and should plan the trials accordingly.  


    . BACE1 inhibition more effectively suppresses initiation than progression of β-amyloid pathology. Acta Neuropathol. 2018 May;135(5):695-710. Epub 2018 Jan 11 PubMed.

    . The antibody aducanumab reduces Aβ plaques in Alzheimer's disease. Nature. 2016 Aug 31;537(7618):50-6. PubMed.

  2. The Merck announcement that APECS is terminating early is, of course, a big disappointment. Still, we will learn a great deal from eventual analysis of the study data.

    This discontinued trial does not really dampen our enthusiasm for BACE inhibitors. We know that drugs in this class can largely eliminate the generation of the Aβ42 peptide. These drugs are administered orally and are generally tolerable (though safety issues have occurred with some BACE inhibitors). It remains very likely that BACE inhibitors will be a mainstay of our AD therapy, particularly in primary prevention of sporadic and genetically determined disease. The ongoing EARLY trial is testing efficacy in the preclinical population, and the A3 trial, which should launch next year, will move into a pre-preclinical population (individuals with “intermediate” brain amyloid by PET SUVR). Overall, the prospects for effective anti-amyloid therapy for AD remain very bright.

  3. The Merck announcement is especially disappointing because the drug looked so promising when it was first described. Although clinical studies have made a lot of progress toward understanding the early phases of AD, I think the field is still struggling to identify who will progress from MCI to AD and therefore who would be best served by pharmacologic intervention. The Merck team smartly limited their study to participants with episodic memory impairment and biomarker evidence of Aβ accumulation, but even among this group, benefit may be difficult to assess given expected heterogeneity in rate of progression. I am not a trialist and have no clinical training (my work primarily relies on mouse models), but I will be curious to learn more about the details of the trial population, the treatment duration prior to ending the study, and the risk profile that led to this decision. I certainly hope this is not a death knell for anti-amyloid approaches at large; instead, I would like to see this outcome encourage consideration of combination strategies which may limit risk and allow treatment at ever-so-slightly later stages of disease where success may be easier to measure, if admittedly harder to achieve.

  4. The test of BACE inhibition at earlier stages of AD, such as this study in prodromal AD, and earlier-stage prevention efforts, is necessary for the field. Merck did an outstanding job of running top-notch trials to help address the question in prodromal to mild AD. The results of the trial will be essential for the field to understand and to determine on-target and off-target effects.

    Although the outcome is very disappointing, it had been predicted that a BACE inhibitor at presymptomatic (or preclinical) stages of Alzheimer’s disease would have a much greater pathologic, and potentially clinical, benefit than in prodromal AD. This is partly based on data indicating that by the time symptoms first appear in AD, there are already near-maximal amyloid plaques. The results of the symptomatic AD BACE trials, data from animal models, and observational clinical studies indicate that inhibitors of Aβ production mechanisms need to be administered earlier in the disease course to stop amyloid plaques from growing. A production inhibitor, such as a BACE inhibitor, is likely to have a large impact while plaques are still growing in the preclinical stage, such as during the 10–15 years before clinical onset (i.e., secondary prevention), and most likely to have the largest impact on amyloid plaques in a primary prevention trial, i.e., before plaques are formed (Aug 2017 conference news). These are currently being tested and planned in several trials, including the DIAN-TU, EARLY, and API prevention studies.

  5. Like others in the field, I was disappointed to read about the external data monitoring committee’s (eDMC’s) finding that the BACE inhibitor verubecestat was unlikely to demonstrate “a positive benefit/risk” if the APECS study was completed in persons with prodromal AD. In order to understand what the findings may mean, it will be helpful to find out the extent to which the decision to discontinue treatment was based on efficacy, safety, or a combination of safety and efficacy findings. In addition to reporting its findings in future meetings and peer-reviewed articles, it would be an invaluable contribution to the field if Merck would share the data and a subset of biological samples from its trials with the field within the next 12 months, following Collaboration for Alzheimer’s Prevention (CAP) principles.

    If the eDMC decision was based primarily on the finding that verubecestat would be unlikely to demonstrate a significant benefit in prodromal AD, the findings would be disappointing but not unexpected. Even if the amyloid hypothesis is correct and the treatment turned out to be sufficiently safe, it remains unclear how this treatment—which is expected to have a greater effect on the production of Aβ42 than the clearance of existing plaques at this disease stage—would have limited damage from pre-existing Aβ aggregates and any biological consequences of this pathology, which are already extensive in persons with prodromal AD.

    It remains possible that a sufficiently safe, tolerable, and dosed BACE1 inhibitor could have a significant benefit at an earlier stage of the disease (e.g., in cognitively unimpaired persons at genetic risk for AD before the development of appreciable plaques). If the amyloid hypothesis is correct, BACE1 inhibitors could help to clarify whether we could extinguish the kindling that leads to AD before the development of a self-perpetuating fire. If the amyloid hypothesis is correct, a BACE inhibitor could also be used in combination with an amyloid aggregate-clearing immunotherapy at a later disease stage.

    In the five-year Alzheimer’s Prevention Initiative (API) Generation Study 1, we have the privilege of evaluating Novartis’s relatively selective BACE1 inhibitor CNP520 (which is being co-developed with Amgen) in cognitively unimpaired 60- to 75-year-old APOE4 homozygotes, some of whom will not have an appreciable Aβ plaque burden prior to treatment. In the five-year Generation Study 2, we are also evaluating this BACE inhibitor in APOE4 homozygotes and amyloid-positive APOE4 heterozygotes.

    We and others look forward to finding ways in which to give this promising drug class the best chance to work. We hope that sponsors maintain the courage and conviction to see this through, put the amyloid hypothesis to the best possible test, discover a more diversified portfolio of promising treatments, and find effective ways to treat and prevent this devastating disease. Indeed, I believe there is still a fighting chance to find and support approval of an effective AD prevention therapy by 2025. 

  6. Early intervention is key, but we need the right targets, too. Too many downstream "targets" have been implicated in AD to believe that any one of these is the actual cause of sporadic disease. Perhaps it really is time for a paradigm shift?

  7. This is a clearly a very disappointing result and a further setback to our therapeutic progress. It is difficult to know how to fully evaluate these early results without understanding how the interim analysis was constructed and what the data actually show, so we await this with great interest. This result certainly adds to the growing number of failed monotherapy amyloid-based therapeutic trials. Importantly, this verubecestat BACE inhibitor program had demonstrated clear target engagement with significant CSF Aβ lowering (1-42, 1-40, and sAPPβ) and with clear dose-response relationships, a level of data support for proof of concept that has not been achieved by many other programs.

    The lack of clinical efficacy in prodromal AD continues to underscore that while amyloidopathy is necessary, it is unlikely to be sufficient to cause the disease or serve as a monotherapy treatment target on its own. It does not diminish the fundamental role of amyloidopathy, but does suggest that there are other interrelated pathogenic mechanisms that will need to be addressed for successful treatment of this disease. These results should also lead us to further reflect on the move to diagnose Alzheimer’s disease in asymptomatic individuals with amyloidopathy who are cognitively normal. The proposal of the International Working on Diagnostic Research Criteria to consider this state of being cognitively normal with amyloidopathy as "asymptomatic at risk" seems a better tack than a disease label until there is more evidence that its treatment has an impact.

  8. The discontinuation of this project should not come as a surprise. For those who attended the presentation on the EPOCH trial at the Boston 2017 CTAD meeting, and have ever worked with a responsible drug company, it was quite obvious that the risk/benefit profile of the compound was unfavorable. Side effects reported were rash, falls, and weight loss in mild to moderate AD patients—i.e., clinical problems that cannot be taken easily. To hope that these unwanted effects of verubecestat would disappear or be attenuated once the drug is given over years (sic!) to so-called high-risk, but asymptomatic, trial participants, is wishful thinking. Merck should be commended for making a business decision that most likely is in the interest of potential future study participants.

  9. The termination of this trial is sad news for the Alzheimer patients, but does not surprise me. The termination simply shows that BACE inhibitors need to be given early—as a secondary prevention—to be successful. Thus, clinical trials with BACE inhibitors need to start several years before symptom onset. Prodromal AD is simply not early enough. In parallel, it is important to consider even primary prevention—exactly as DIAN currently plans on doing it.

    The termination compromises neither BACE1 as an excellent drug target for AD, nor the amyloid hypothesis.

  10. The latest trial failure with verubecestat /therapeutics/verubecestat underscores why sometimes a linear simple hypothesis is compounded by complex non-linear biology in the human brain. To assess the differential biology of short and long forms of Aβ on glutamate and nicotinic neurotransmission, we used a mechanism-based and ADAS-Cog calibrated Quantitative Systems Pharmacology computer model (Geerts et al., 2018). The introduction of a beneficial effect of shorter Aβ isoforms on glutamate neurotransmission, originally reported in preclinical studies, is absolutely necessary to explain three clinical data sets in human AD patients. When such a model applied to reported pharmacodynamic effects of BACE inhibitors, solanezumab /therapeutics/solanezumab and GSI semagecestat, the model suggest that the clinical effect is very dependent upon baseline amyloid load, with subjects at low baseline experiencing a cognitive worsening and subjects at high baseline improving. If one does not enrich for a patient population with relatively high baseline amyloid, average treatment group outcomes with amyloid lowering might tend to not differ from placebo. The model also suggests that interventions that spare shorter Aβ forms and reduce the toxic Aβ42 have a modest clinical benefit.

    It would be great if we could test these predictions in the analysis of the failed trials, because if these assumptions hold true, indiscriminate amyloid reduction in cognitively normal subjects with low or zero amyloid will likely worsen outcome, therefore questioning the rationale to start treatment of healthy elderly subjects with low amyloid levels “before the pathology starts.”  


    . Impact of amyloid-beta changes on cognitive outcomes in Alzheimer's disease: analysis of clinical trials using a quantitative systems pharmacology model. Alzheimers Res Ther. 2018 Feb 2;10(1):14. PubMed.

  11. I agree with Dr. Geerts. Because the human nervous system is a complex system, complexity theory can and should be applied to the study of age-related neurodegenerative diseases. Complexity theory provides a framework that takes into account the interactions between self-organization, adaptive systems, networks, and non-linear responses of organisms.

Make a Comment

To make a comment you must login or register.


Therapeutics Citations

  1. Verubecestat
  2. Atabecestat
  3. Umibecestat

News Citations

  1. BACE Block Nips New Plaques in the Bud, Old Ones Keep Growing
  2. Paper Alert: Verubecestat Preclinical and Phase 1 Data Published
  3. Verubecestat Negative Trial Data: What Does it Mean for BACE Inhibition?
  4. 10th CTAD: Finally, Alzheimer’s Field Is Serious About Prevention Trials
  5. Planning the First Primary Prevention Trial for Alzheimer’s Disease

External Citations

  1. EPOCH trial 
  2. APECS
  3. press release

Further Reading