jpad journal

AND option

OR option

TRC-PAD: ACCELERATING RECRUITMENT OF AD CLINICAL TRIALS THROUGH INNOVATIVE INFORMATION TECHNOLOGY

G.A. Jimenez-Maggiora1, S. Bruschi1, R. Raman1, O. Langford1, M. Donohue1, M.S. Rafii1, R.A. Sperling2, J.L. Cummings3, P.S. Aisen1 and the TRC-PAD Investigators*

1. Alzheimer’s Therapeutic Research Institute, University of Southern California, San Diego, CA, USA; 2. Center for Alzheimer Research and Treatment, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA; 3. Department of Brain Health, School of Integrated Health Sciences, University of Las Vegas, Nevada; Cleveland Clinic Lou Ruvo Center for Brain Health, USA; * TRC-PAD investigators are listed at www.trcpad.org

Corresponding Author: GA Jimenez-Maggiora, Alzheimer’s Therapeutic Research Institute, University of Southern California, San Diego, CA, USA, gustavoj@usc.edu

J Prev Alz Dis 2020;4(7):226-233
Published online August 16, 2020, http://dx.doi.org/10.14283/jpad.2020.48

 


Abstract

BACKGROUND: The Trial-Ready Cohort for Preclinical/Prodromal Alzheimer’s Disease (TRC-PAD) Informatics Platform (TRC-PAD IP) was developed to facilitate the efficient selection, recruitment, and assessment of study participants in support of the TRC-PAD program.
Objectives: Describe the innovative architecture, workflows, and components of the TRC-PAD IP.
Design: The TRC-PAD IP was conceived as a secure, scalable, multi-tiered information management platform designed to facilitate high-throughput, cost-effective selection, recruitment, and assessment of TRC-PAD study participants and to develop a learning algorithm to select amyloid-bearing participants to participate in trials of early-stage Alzheimer’s disease.
Setting: TRC-PAD participants were evaluated using both web-based and in-person assessments to predict their risk of amyloid biomarker abnormalities and eligibility for preclinical and prodromal clinical trials. Participant data were integrated across multiple stages to inform the prediction of amyloid biomarker elevation.
Participants: TRC-PAD participants were age 50 and above, with an interest in participating in Alzheimer’s research.
Measurements: TRC-PAD participants’ cognitive performance and subjective memory concerns were remotely assessed on a longitudinal basis to predict participant risk of biomarker abnormalities. Those participants determined to be at the highest risk were invited to an in-clinic screening visit for a full battery of clinical and cognitive assessments and amyloid biomarker confirmation using positron emission tomography (PET) or lumbar puncture (LP).
Results: The TRC-PAD IP supported growth in recruitment, screening, and enrollment of TRC-PAD participants by leveraging a secure, scalable, cost-effective cloud-based information technology architecture.
Conclusions: The TRC-PAD program and its underlying information management infrastructure, TRC-PAD IP, have demonstrated feasibility concerning the program aims. The flexible and modular design of the TRC-PAD IP will accommodate the introduction of emerging diagnostic technologies.

Key words: Alzheimer’s, clinical trials, informatics, recruitment.


 

Background

Alzheimer’s Disease (AD) has emerged as one of the most significant public health issues of the 21st century. In 2020, it was estimated that 5.8 million people in the United States (U.S.) were living with Alzheimer’s Disease; this number was expected to rise to 13.8 million by 2050 (1). The development of effective disease-modifying interventions for Alzheimer’s disease (AD) remains an enormously important world health need. As therapeutic research has expanded to early-stage interventions, the recruitment of minimally affected optimally selected study participants has been challenging (2). The challenges include: identifying potential asymptomatic participants with elevated amyloid levels who meet study criteria, expanding recruitment and participation of groups that have been traditionally underrepresented in AD clinical trials, and reducing the time and cost of screening failures (3). Novel approaches to improve participant selection and recruitment using sophisticated informatics platforms have been developed (4, 5). The Trial-Ready Cohort for Preclinical/Prodromal Alzheimer’s Disease (TRC-PAD) program was initiated with the overarching goal of accelerating therapeutic development for AD through the establishment of an infrastructure to ensure timely recruitment of targeted individuals into optimally designed trials (6). More specifically, TRC-PAD aimed to establish a biomarker-confirmed trial-ready cohort to facilitate recruitment into preclinical and prodromal AD trials by using an efficient, low-cost multi-stage selection process driven by an adaptive risk algorithm which combined web-based and in-person longitudinal assessment of participants. The TRC-PAD Informatics Platform (TRC-PAD IP) was developed to facilitate the selection, recruitment, and assessment of study participants in support of the TRC-PAD program. The purpose of this article is to describe the innovative architecture, workflows, and components of the TRC-PAD IP, demonstrate its utility, and discuss broader implications for the field of AD clinical trials.

 

Methods

TRC-PAD Informatics Platform

The TRC-PAD Informatics Platform (TRC-PAD IP) (Figure 1) was conceived as a secure, scalable, multi-tiered information management platform designed to facilitate high-throughput, cost-effective selection, recruitment, and assessment of TRC-PAD study participants. The program aims called for the construction of a multi-stage process that encompassed 1) a public-facing web-based registry, the Alzheimer Prevention Trials (APT) Webstudy registry, 2) an analytics platform capable of supporting the development and implementation of risk-based referral and screening algorithms, 3) a web-based referral management system, the Site Referral System (SRS), and 4) a regulatory-compliant clinical data management system to collect data for the standing Trial Ready Cohort (TRC). Given the complex structure and pathways of this process, tracking participants as they entered the program and transitioned from one stage to the next was also a critical requirement. The component systems underpinning this process are described in the following sections.

Figure 1. TRC-PAD Informatics Platform

 

APT Webstudy

The APT Webstudy (www.aptwebstudy.org) is a public-facing online longitudinal study that invited members of the general population who were over 50 years of age and had sufficient interest in participating in AD research to register for a user account. Once registered, users were presented with a guided walk-through during which they provided their contact information and preferences, current AD-related diagnosis (if known), and interest in participating in AD-related research. Subsequently, users were offered additional study information, asked to consider the benefits and risks of study participation, and allowed to consent and enroll in the APT Webstudy using an online consent process. Participants were subsequently asked to provide additional information regarding their demographics, family and medical history, lifestyle, current AD biomarker status (if know), and complete assessments of their subjective memory perceptions using the Cognitive Function Instrument (CFI), and cognitive performance using the Cogstate Brief Battery (7-9). To minimize participant burden, the registration, consent, and assessment process was designed to take approximately 15 to 20 minutes to complete. Follow-up assessments, scheduled every 3 months, were designed to be completed in 10 to 15 minutes to promote participant retention.

Website Design

The APT Webstudy was designed with several features to support usability and engagement. Users had the option to create their accounts using a «1-click» social login instead of a traditional approach which requires a username, password, and email address. Participants were provided with a dashboard to view their previous assessment scores. They could also download a report with all their information and assessment scores. Participants received a quarterly newsletter via electronic mail developed by leading AD researchers which highlighted new developments in the field as well as news related to the TRC-PAD program. The website was designed to support multiple screen sizes (phone, tablet, desktop) using responsive web design principles (10).

Longitudinal Assessments

To improve the prediction of participant risk, the APT Webstudy was designed to incorporate unsupervised cognitive assessments as part of a longitudinal observational study. Thus, participants are invited to return to the website every 3 months to complete a new round of assessments and update their information as needed. These follow-up assessments are designed to take 10 to 15 minutes to complete. To encourage compliance, participants received quarterly electronic mail reminders to log-in to the APT Webstudy website to complete their web-based assessments.

Website Architecture

An important consideration in a public-facing website is the variability of website traffic patterns. Websites with highly volatile traffic patterns can exhibit occasional periods of poor performance or unavailability. This is especially true when an unexpected media event, such as a viral social media mention or a celebrity endorsement, draws a burst of traffic toward the website. To mitigate the effects of these unexpected traffic spikes, the APT Webstudy has an architecture that dynamically adjusts its capacity to respond to incoming requests as changes in traffic patterns emerge. This architecture combines a cloud-based fleet of webservers, a multi-region database cluster, and a dynamic networking configuration to allow the website to automatically activate and deactivate computational resources on-demand to maintain performance during high-traffic periods in a cost-effective manner.

Recruitment and Retention

Recruitment into the APT Webstudy relied on a combination of traditional and digital media strategies. These efforts included community-based events, local and national paid media campaigns, social media campaigns, and earned media. Additionally, a group of preexisting registries which included the Brain Health Registry (brainhealthregistry.org), the Alzheimer’s Prevention Registry (endalznow.org), Trial Match (trialmatch.alz.org), and Healthy Brains (healthybrains.org), partnered with the TRC-PAD program to refer participants from their cohorts. These registries, collectively known as the “feeder” registries, provided the APT Webstudy with access to a large, engaged group of potential participants. The performance of these recruitment efforts was tracked by using Urchin Tracking Module (UTM) codes which associated participants with the campaign that referred them to the APT Webstudy (11).

Participant Technical Support

APT Webstudy participants were provided with both telephone- and email-based support channels. Telephone-based support requests were automatically transcribed and forwarded to the email-based support management system for follow-up. This architecture allowed the APT Webstudy support team to manage all incoming support requests via a single interface. Support cycle times and participant satisfaction were among the Key Performance Indicators (KPIs) used to assess and manage the performance of the support team.

Site Referral System

APT Webstudy participant data were regularly evaluated using an adaptive algorithm that assessed each participant’s risk of AD biomarker positivity and ranked participants based on predicted risk (12). Participants determined to have the highest risk were referred to the nearest TRC-PAD performance site based on their self-reported 5-digit Zip Code. Participant referrals were provided to performance sites via the Site Referral System (SRS), a secure website that allowed authorized site personnel to manage the site’s referral queue and ensure the timely disposition of each referral. Using the SRS, site personnel contacted participants and invited them to schedule an initial in-person screening visit to determine their eligibility for enrollment into the TRC-PAD standing cohort.

Website Design

A key concept guiding the design of the SRS was the idea that site referral queues would be managed collaboratively among multiple site personnel. To support collaboration, the SRS website incorporated several features that optimized team-based management and disposition of participant referrals. For example, as new referrals became available in the referral queue, site personnel received notifications via electronic mail. The status of every referral in the queue was summarized in a dashboard view facilitating management and reporting. Each referral provided site personnel with access to a summary of participant-reported demographic, medical history, lifestyle, and contact information, as well as status changes. As a final step, referrals were assigned an outcome code and marked complete.

Trial Ready Cohort

The TRC was conceived as the final stage in the TRC-PAD process. The initial target for the TRC was to enroll a standing longitudinal cohort of 2,000 biomarker-confirmed participants (50% preclinical and 50% prodromal) (6). An Electronic Data Capture (TRC EDC) system was developed to manage TRC participant data, based on the Alzheimer’s Therapeutic Research Institute (ATRI) EDC system. The TRC EDC was used to operationalize a multi-stage screening process and collect a rich set of clinical, neuropsychological, neuroimaging, and biospecimen data collected in a multi-site setting. The TRC EDC was validated to comply with CRF Title 21 Part 11 (13). Doing so allowed the TRC data to be eligible for use as run-in data in downstream clinical trials.

TRC EDC Design

The TRC EDC was built to provide centralized management of all critical cohort study dataflows and workflows. This approach provided study teams with broad transparency and management capabilities over the study’s complex dataflows and processes. Implementing this approach, however, has proven challenging for traditional systems, which struggle to scale up as larger and more complex data types are introduced. Historically, a solution to this problem has been the implementation of interconnected purpose-specific systems. While feasible, this solution suffers from several shortcomings: 1) supporting evolving study requirements requires complex multi-system impact analysis, 2) training on multiple systems increases the burden on study teams, and 3) data integrations require ongoing maintenance as systems are updated. The TRC EDC was designed to avoid these issues by implementing a cloud-native architecture, which allowed it to harness the full range of capabilities of the underlying cloud platform on which it was hosted. In this architecture, the TRC EDC served as an orchestration engine that coordinated and delegated computational workloads, such as uploading and processing large binary objects (e.g. medical images, sensor data, multimedia data, genetic data, graph data), to the underlying cloud-based service best-suited for the required function, without impacting the study team’s interaction with the system.

Site Network

An initial step toward the establishment of the TRC was the creation of the TRC-PAD site network, a set of 35 academic clinical sites distributed across the large population centers in the contiguous United States. Sites were selected based on multiple criteria including study team research experience and expertise, amyloid positron emission tomography (PET) imaging and radiotracer availability, and track record in AD/AD and related disorders (ADRD) clinical trials. Site selection and activation were conducted in two stages, the “vanguard” or initial phase, which included 8 pilot sites, and the broad activation phase. The vanguard phase was used as a learning exercise to test and refine TRC-PAD data management tools and processes. These learnings were applied during the broad activation phase to ensure the full site network was optimized.

Multi-stage Selection and Screening Process

The TRC-PAD selection and recruitment processes were managed by a series of learning algorithms that assessed participant data at multiple points. Newly acquired data were used to update participant risk predictions and rankings (12). These results informed the decision-making process used to graduate participants from one stage of TRC-PAD to the next, culminating with the final determination on enrollment eligibility into the standing TRC.

Referral to Downstream Clinical Trials

TRC participants who meet eligibility criteria will be referred for screening and potential enrollment to downstream clinical trials, temporarily suspending additional longitudinal assessments and data collection activities in the TRC EDC. In these cases, a link between a participant’s TRC data record and their downstream clinical trial data record will be established via the use of the National Institute on Aging’s (NIA) Global Unique Identifier (GUID) (14, 15). This link will allow for participant data to be integrated across cohorts and used to further inform the TRC-PAD learning algorithms.

Analytical Platform

TRC-PAD IP supported a single Analytical Platform (AP) that aggregated data from multiple sources in a single semi-structured repository, also known as a «data lake» (16). This approach combined the use of serverless computing methods with fast, low-cost object storage to facilitate the development and operationalization of multiple analysis workloads such as machine learning algorithms, statistical analyses, and reports.

Information Architecture

The TRC-PAD IP was constructed using open source web development and scientific computing tools hosted on Amazon Web Services (AWS), a public cloud computing platform (Table 1). The TRC-PAD program partnered with AWS’s consulting group, AWS Professional Services (AWS ProServ), to construct a secure, scalable, cost-effective computational infrastructure. The component systems of the TRC-PAD IP were built using a phased approach. Unique system-generated identification numbers (IDs) were assigned to each participant as they moved from one stage of TRC-PAD to the next. These IDs were linked across component systems to maintain the integrity of each participant’s data record while protecting confidentiality.

Table 1. Technology Components of the TRC-PAD Informatics Platform

 

Scalability and Cost

The TRC-PAD IP was built on cloud-based computational infrastructure that was optimized to support each component system. The infrastructure supporting the APT Webstudy, which was subject to temporary bursts of traffic due to recruitment campaigns or unexpected media events, was designed to dynamically adapt to changing web traffic patterns by increasing or decreasing its fleet of webservers, within set cost parameters. Likewise, the TRC EDC, which managed a large, regulatory-compliant cohort database, used a cost-effective, highly durable storage strategy. In this architecture, computational resources were used on-demand, reducing idle capacity and providing flexibility as the TRC-PAD program’s requirements evolved.

Security and Compliance

Ensuring participant confidentiality and data security were the primary requirements for the TRC-PAD IP. Thus, the TRC-PAD Informatics team worked with AWS ProServ to build a Health Insurance Portability and Accountability Act (HIPAA)-eligible architecture and implement AWS’s best practices (17). This architecture was designed to take advantage of multiple strategies to ensure data security and durability such as a multi-account structure, multi-region, encrypted data storage, and automated policy management. To further ensure data security, the TRC-PAD IP underwent annual security audits by an independent security firm.

Regulatory Oversight

Regulatory oversight of the APT Webstudy was provided by the University of Southern California Institutional Review Board (USC IRB). Regulatory oversight for the TRC was provided by Advarra, Inc., under a single IRB (sIRB) model. The sIRB model was established as a National Institute of Health (NIH) requirement for multi-site studies starting in 2018 to streamline the review of research that involves human subjects (18).

 

Results

The APT Webstudy was launched on December 22, 2017, after a 6-month development period. The SRS and TRC EDC launched in May 2019. The TRC-PAD development team worked closely with the study team employing agile software development methods to design, build, test, and deploy these systems (19, 20).
As of July 6, 2020, 36,955 users had registered for an APT account. Of these registered users, 33,259 (90.0%) enrolled in the study via online consent and had completed more than 280,000 remote assessments (Figure 2). Recruitment into the APT Webstudy was driven by earned, owned, and shared media as well as feeder-based referrals. The APT cohort was geographically distributed across all 50 U.S. states, with participants concentrated in the coastal, midwestern, and southwestern states. Participant mobile device usage (43.0%) on the APT Webstudy was higher than initially expected. The demographic characteristics of the cohort are female (73.0%), non-Hispanic White (92.4%), with a mean age of 64.6 years (SD = 8.3). 87.7% agreed to have their contact information shared with the TRC sites. After one year of quarterly follow-up, 44.7% of participants were retained. The retention rate after two years of quarterly follow-up was 29.7%.

Figure 2. APT Webstudy Enrollment (December 22, 2017 to July 6, 2020)

 

TRC sites began in-person screening of participant referrals in August 2019. As of July 6, 2020, 27 of 35 TRC sites were activated and had received 1,675 risk-ranked participant referrals via the SRS. Of these, 246 (14.7%) participants were referred to the TRC for initial screening, 123 (50.0%) participants completed the initial screening visit, 99 (80.5%) participants were authorized for amyloid testing, 55 (55.6%) participants were biomarker-confirmed using amyloid PET or CSF assessment, 26 (47.3%) participants were found to be amyloid elevated, and 23 (88.5%) participants were enrolled into the TRC (Figure 3). The demographic characteristics of the standing TRC are female (51.2%), non-Hispanic White (94.2%), with a mean age of 72 years (SD = 7.8), and a mean SUVr of 1.14 (SD = 0.22). The median (interquartile range) cycle time from initial site referral via SRS to enrollment decision in the TRC was 28 (15 to 84) days. During this period, the TRC-PAD IP proved to be a scalable, cost-effective solution to support all stages of the TRC-PAD selection and recruitment.

Figure 3. TRC-PAD Participant Flows by Stage (as of July 6, 2020)

 

Discussion

Our early experiences with the TRC-PAD program and its underlying information management infrastructure, TRC-PAD IP, have demonstrated feasibility concerning the program aims. The TRC-PAD IP supported recruitment, screening, and enrollment of TRC-PAD participants by leveraging a secure, scalable, cost-effective cloud-based information technology (IT) architecture. The APT Webstudy has demonstrated effectiveness in terms of selecting and recruiting individuals from the general population who have an elevated risk of amyloid positivity. The APT Webstudy has proven effective in remotely assessing participants on a longitudinal basis. The SRS has demonstrated effectiveness in terms of allowing TRC sites to manage site referrals promptly. The TRC EDC has supported the coordination of the multi-stage risk-based screening and enrollment process into the standing TRC. Much work remains to demonstrate the effectiveness of the TRC-PAD program in accelerating recruitment into downstream AD clinical trials.
Several limitations should be considered. First, the TRC-PAD program has struggled to select and recruit a representative sample of participants across multiple sociodemographic dimensions (21). Efforts to address this challenge have included updating the APT Webstudy to support Spanish-speaking participants, however, more work is needed. Second, the architecture has proven to be resilient during a few web traffic spikes but has yet to be subjected to the type of surge (>10-100x daily web traffic) associated with an unexpected mention in a national media platform. Third, the integrations with third-party platforms, such as the Cogstate Brief Battery, have proven to be fragile, requiring frequent maintenance and technical support for participants. Addressing these technical issues may serve to improve both participant retention and risk algorithm predictive performance.
In May 2019, the TRC-PAD principal investigators selected the first set of downstream clinical trials slated to utilize the standing TRC to recruit participants. These clinical trials are scheduled to begin recruiting participants in North America (U.S. and Canada) in May 2020. The efficiencies that will accrue to these clinical trials are two-fold: 1) by drawing participants from the TRC, the expectation is that recruitment into these clinical trials will be accelerated by reducing the number of screen failures; and 2) the TRC clinical, neuropsychological, biofluid, and imaging data will be available to use as high-quality run-in data for downstream clinical trials (22). These combined efficiencies should yield significant savings both in terms of resources and time.
As the TRC-PAD program has been established over the past few years, promising new AD biomarkers have been developed. Plasma-based AD biomarkers, for example, have been shown to effectively predict amyloid positivity (23). When fully validated, the introduction of these diagnostic tools into the TRC screening process may be used to further increase effectiveness. The flexible and modular design of the TRC-PAD IP will accommodate the introduction of these new technologies.
Finally, as news of the TRC-PAD program’s progress has spread, a global network of programs modeled on its approach has begun to take shape. Investigators in several countries have been collaborating with the TRC-PAD program leadership to establish similar programs in their home regions. By using global cloud computing infrastructure and the TRC-PAD IP as a model architecture, these programs have been able to rapidly establish similar platforms in their local jurisdictions. Once fully operational, the global TRC-PAD network aims to provide AD clinical trials with a steady stream of well-characterized, biomarker-confirmed participants yielding savings in time, effort, and expense.

 

Acknowledgments: The authors would like to thank the TRC-PAD participants and their families, sponsors and partners, investigators, site and coordinating center personnel for their contributions in support of the program. Dr. Cummings is supported by Keep Memory Alive (KMA); NIGMS grant P20GM109025; NINDS grant U01NS093334; and NIA grant R01AG053798.

Funding: This work was funded by the U.S. National Institute on Aging (NIA) (grant number 1R010AG053798).

Ethical standard: Institutional Review Boards (IRBs) approved these studies, and all participants gave informed consent before participating.

Conflict of interest: The authors report grants from National Institute on Aging, during the conduct of the study. None of the authors have additional financial interests, relationships or affiliations relevant to the subject of this manuscript.

Open Access: This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, duplication, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

 

References

1. 2020 Alzheimer’s disease facts and figures. Alzheimers Dement. 2020.
2. Watson JL, Ryan L, Silverberg N, Cahan V, Bernard MA. Obstacles And Opportunities In Alzheimer’s Clinical Trial Recruitment. Health Affairs. 2014;33(4):574-579.
3. Grill JD, Galvin JE. Facilitating Alzheimer Disease Research Recruitment. Alzheimer Disease & Associated Disorders. 2014;28(1):1-8.
4. Ritchie CW, Molinuevo JL, Truyen L, et al. Development of interventions for the secondary prevention of Alzheimer’s dementia: the European Prevention of Alzheimer’s Dementia (EPAD) project. Lancet Psychiatry. 2016;3(2):179-186.
5. Lovestone S, Consortium E. The European medical information framework: A novel ecosystem for sharing healthcare data across Europe. Learn Health Syst. 2020;4(2):e10214.
6. Aisen PS, Sperling RA, Cummings J, et al. The Trial-Ready Cohort for Preclinical/prodromal Alzheimer’s Disease (TRC-PAD) Project: An Overview. J Prev Alz Dis 2020; DOI: 10.14283/jpad.2020.45
7. Walsh SP, Raman R, Jones KB, Aisen PS, Group AsDCS. ADCS Prevention Instrument Project: the Mail-In Cognitive Function Screening Instrument (MCFSI). Alzheimer Dis Assoc Disord. 2006;20(4 Suppl 3):S170-178.
8. Li C, Neugroschl J, Luo X, et al. The Utility of the Cognitive Function Instrument (CFI) to Detect Cognitive Decline in Non-Demented Older Adults. J Alzheimers Dis. 2017;60(2):427-437.
9. Lim YY, Ellis KA, Harrington K, et al. Use of the CogState Brief Battery in the assessment of Alzheimer’s disease related cognitive impairment in the Australian Imaging, Biomarkers and Lifestyle (AIBL) study. J Clin Exp Neuropsychol. 2012;34(4):345-358.
10. Lane J, Barker T, Lewis J, Moscovitz M. Responsive Design. In: Foundation Website Creation with HTML5, CSS3, and JavaScript. Berkeley, CA: Apress; 2012.
11. Walter S, Clanton T, Langford O, et al. Recruitment into the Alzheimer Prevention Trials (APT) Webstudy for a Trial-Ready Cohort for Preclinical and Prodromal Alzheimer’s Disease (TRC-PAD). J Prev Alz Dis 2020; DOI: 10.14283/jpad.2020.46
12. Langford O, Raman R, Sperling RA, et al. Predicting amyloid burden to accelerate recruitment of secondary prevention clinical trials. J Prev Alz Dis 2020; DOI: 10.14283/jpad.2020.44
13. Title 21 Part 11 — Electronic Records; Electronic Signatures. In: Electronic Code of Federal Regulations; 1997.
14. Babcock D. The Role of GUIDs in Data Sharing. The NIA Alzheimer’s Disease Centers Program – National Alzheimer’s Coordinating Center. https://www.alz.washington.edu/NONMEMBER/FALL19/Dir/Babcock.pdf. Published 2019. Accessed May 5, 2020.
15. Johnson SB, Whitney G, McAuliffe M, et al. Using global unique identifiers to link autism collections. J Am Med Inform Assoc. 2010;17(6):689-695.
16. O’Leary D. Embedding AI and Crowdsourcing in the Big Data Lake. IEEE Intelligent Systems. 2014;29(5):70-73.
17. Amazon Web Services: AWS Well-Architected Framework. https://d1.awsstatic.com/whitepapers/architecture/AWS_Well-Architected_Framework.pdf. Published 2019. Accessed April 27, 2020.
18. National Institutes of Health: NIH Policy on the Use of a Single Institutional Review Board for Multi-Site Research. https://osp.od.nih.gov/wp-content/uploads/2016/06/NIH_sIRB_Policy_Multi_site_Research_UPDATED2016.pdf. Accessed April 28, 2020.
19. Beck K, Beedle M, van Bennekum A, et al. Manifesto for Agile Software Development. http://agilemanifesto.org/. Accessed May 13, 2020.
20. Hohl P, Klünder J, van Bennekum A, et al. Back to the future: origins and directions of the “Agile Manifesto” – views of the originators. Journal of Software Engineering Research and Development. 2018;6(1):1-27.
21. Walter S, Langford O, Clanton T, et al. The Trial-Ready Cohort for Preclinical/Prodromal Alzheimer’s disease (TRC-PAD): Experience from the first 3 years. J Prev Alz Dis 2020; DOI: 10.14283/jpad.2020.47
22. Laursen DRT, Paludan-Müller AS, Hróbjartsson A. Randomized clinical trials with run-in periods: frequency, characteristics and reporting. Clin Epidemiol. 2019;11:169-184.
23. Ovod V, Ramsey KN, Mawuenyega KG, et al. Amyloid β concentrations and stable isotope labeling kinetics of human plasma specific to central nervous system amyloidosis. Alzheimers Dement. 2017;13(8):841-849.

What do you want to do ?

New mail

What do you want to do ?

New mail

What do you want to do ?

New mail

What do you want to do ?

New mail

What do you want to do ?

New mail

What do you want to do ?

New mail

What do you want to do ?

New mail

What do you want to do ?

New mail

What do you want to do ?

New mail

What do you want to do ?

New mail

What do you want to do ?

New mail

What do you want to do ?

New mail

What do you want to do ?

New mail

What do you want to do ?

New mail

What do you want to do ?

New mail

What do you want to do ?

New mail

What do you want to do ?

New mail

What do you want to do ?

New mail

What do you want to do ?

New mail

GEROSCIENCE AND THE ROLE OF AGING IN THE ETIOLOGY AND MANAGEMENT OF ALZHEIMER’S DISEASE

 

F. Sierra

 

Correspondance author: Division of Aging Biology, National Institute on Aging (NIA), National Institutes of Health (NIH) – 7201 Wisconsin Ave, Suite 3N300, Bethesda MD 20892, USA, sierraf@nia.nih.gov

J Prev Alz Dis 2020;(7) in press
Published online December 5, 2019, http://dx.doi.org/10.14283/jpad.2019.49

Key words: Geroscience, aging, Alzheimer’s, neurodegeneration.

 


 

With the relentless aging of the population worldwide, two major concerns need our immediate attention: the expected dramatic increase in disease and disability burden, and the decrease in the ratio of working individuals vs. retirees. A comprehensive approach involving experts in many disciplines will be required to tackle these issues. Here we will concentrate on the role of basic biological science in averting the increase in disease and disability, using Alzheimer’s Disease (AD) as a model.
The spectrum of AD represents a serious threat and a psychological burden on people at all ages. The US Alzheimer’s Association estimates that 5.8 million Americans are currently living with Alzheimer’s, and 1 out of 3 seniors dies with the disease or other dementias. Because of its insidious effect on both the individual and his/her surroundings, as well as the associated healthcare cost, AD has been singled out for special efforts by funding agencies and scientists alike, and while some progress has been made, it is clearly not sufficient. Indeed, in the past few decades, scientists have been able to identify the molecular composition of the telltale plaques and tangles and have created a large number of mouse models that, to some extent, recapitulate the pathological characteristics of the disease (1). Furthermore, early efforts identified some major drivers of the familial (rare) form of the disease, and more recently, a large number of genes suspected to play a role on the late-onset, non-familial form of the disease have also been identified (2). The role of these new genes is just beginning to be unraveled. Yet a cure has been elusive, and even prevention strategies have been less than hoped for.
One aspect of the etiology of the disease that has been largely neglected until recent years is the role of aging. It is not by mere chance that so many chronic diseases appear simultaneously, in many species, as individuals reach approximately 2/3 of the lifespan for their species (around 60 years for humans). Most of these chronic diseases differ dramatically from the diseases we were able to conquer in the 20th century, in that they are not caused by external agents such as pathogens and poor environmental quality but rather, they are the result of failures within our own organism. For that reason, these diseases have proven to be less tractable, and fighting them is more complex. But the age-dependency in the appearance of symptoms from multiple chronic diseases belies the fact that aging is by far the major risk factor for most of these chronic conditions (3), including Alzheimer’s disease (4). Importantly, the fact that such diseases occur in multiple species at different chronological times (days in flies, months in mice, years in humans), but always at the same physiological time (about 2/3 of the expected lifespan) indicates that it is the process of aging, not the passage of time, that is central. The passage of time indeed allows the accumulation of damage that can lead to disease and disability. However, this accumulation is often rather slow while the organism is young and resilient. It is only after the process of aging starts weakening that resilience that serious accumulation of damage – and thus disease – occurs. So, it is not simply that as we age, damage has accumulated to an extent that causes disease; rather, it is that as we age, we had lost part of our defenses, thus allowing the damage to accumulate. Taking AD as an example, we know that even individuals with the worst genetic predisposition to the disease won’t develop symptoms when they are toddlers or teenagers, they will develop them late in life (earlier than other, non-genetically afflicted populations, but usually not earlier than their 40s or 50s) (5). Yet, because of their genetic burden, they are producing enormous amounts of deleterious aggregation-prone proteins from before birth! Minimal accumulation and no disease occurs because, while young, their resilience capacity allows them to counteract this burden, and resolve much of the damage through proteostasis mechanisms.
This is not unique to AD, and a similar argument can be brought to bear in many other chronic diseases of the elderly, including cancer, cardiovascular, chronic kidney disease, etc. This is the central tenet of the new field of geroscience: since aging is at the core, and the most important risk factor for so many chronic diseases and conditions, it follows that addressing aging will produce a better outcome than addressing each disease individually (6). Indeed, it is expected that, by slowing down the pace of aging we can delay all such chronic ailments, all at the same time. This is nothing new, since we have always known about the fragility and illnesses that often accompany old age. What is new is the amazing advancements we have had in the last couple of decades in our efforts to understand the biological underpinnings of the aging process. Indeed, scientists have now identified a handful of molecular and cellular pathways that drive the process of aging (7, 8). Moreover, those discoveries have led to the identification of pharmacological and dietary means to slow down aging processes, and some of these interventions are already being tried in the clinic, including rapamycin (9) and senolytics (10).
The AD field has been slow to recognize these developments, but changes are being implemented, among others, through several new initiatives promoted by the US National Institute of Aging, aimed at promoting research into the geroscience underpinnings of AD. In fact, pre-clinical data in various mouse models of AD suggest that interventions aimed at slowing down the aging process might be effective in delaying or slowing down disease progression. As in other diseases that affect preferentially the elderly population, these pre-clinical interventions have focused primarily on rapamycin (11, 12) and senolytics (13). In fact, a strong argument has been made to test rapamycin in clinical trials of the disease (14), and two small phase I trials of senolytics are being planned for the near future (J. Kirkland, pers. comm.). In the accompanying paper by Guerville et al., a vigorous argument is made for the inclusion of geroscience principles in our fight to conquer Alzheimer’s disease. Importantly, the paper also outlines specific areas where attention to the pillars of aging might be fruitful in our efforts against Alzheimer’s.

 

Disclosures: Dr. Sierra has no conflicts to disclose. The ideas discussed represent Dr. Sierra’s views and do not represent the views of the U.S. government.

 

References

1. Hardy J. A hundred years of Alzheimer’s disease research. Neuron. 2006;52:3–13.
2. Naj A.C., Schellenberg G.D. and the Alzheimer’s Disease Genetics Consortium (ADGC). Genomic variants, genes, and pathways of Alzheimer’s disease: An overview. Am J Med Genet B Neuropsychiatr Genet. 2017;174:5-26.
3. Sierra F. and Kohanski R. Geroscience and the trans-NIH Geroscience Interest Group, GSIG. Geroscience 39:1-5.
4. Qiu C., Kivipelto M. and Strauss E. Epidemiology of Alzheimer’s disease: Occurrence, determinants and strategies toward intervention. Dialogues Clin. Neurosci. 2009;11:111-128.
5. Ryman D.C., Acosta-Baena N., Aisen P.S. et al. Symptom onset in autosomal dominant Alzheimer’s disease: a systematic review and meta-analysis. Neurology 2014;83:253-260.
6. Austad SN. The geroscience hypothesis: Is it possible to change the rate of aging? in Advances in Geroscience, Sierra & Kohanski eds., Springer, 2016;pp1-36.
7. López-Otín C., Blasco M.A., Partridge L. et al. The hallmarks of aging. Cell 2013;153:1194-217.
8. Kennedy B.K., Berger S, Brunet A. et al. Geroscience: linking aging to chronic disease. Cell 2014;159:709-713.
9. Mannick JB, Del Giudice G, Lattanzi M et al. mTOR inhibition improves immune function in the elderly. Sci Transl Med. 2014;6:268ra179
10. Justice JN, Nambiar AM, Tchkonia T et al. Senolytics in idiopathic pulmonary fibrosis: Results from a first-in-human, open-label, pilot study. EBioMedicine. 2019;40:554-563.
11. Caccamo A., Majumder S., Richardson A. et al. Molecular interplay between mammalian target of rapamycin (mTOR), amyloid-ß and tau: Effects on cognitive impairments. J. Biol. Chem. 2010;285:13107-13120.
12. Lin A-L., Jahrling J.B., Zhang W. et al. Rapamycin rescues vascular, metabolic and learning deficits in apolipoprotein E4 transgenic mice with presymptomatic Alzheimer’s disease. J. Cereb. Blood Flow Metab. 2017;37:217-226.
13. Bussian T.J., Aziz A., Meyer C.F. et al. Clearance of senescent glial cells prevents tau-dependent pathology and cognitive decline. Nature 2018;562:578-582.
14. Kaeberlein M. and Galvan V. Rapamycin and Alzheimer’s disease: time for a clinical trial? Sci. Transl. Med. 2019;11, eaar4289

CLINICAL EFFECTS OF TRAMIPROSATE IN APOE4/4 HOMOZYGOUS PATIENTS WITH MILD ALZHEIMER’S DISEASE SUGGEST DISEASE MODIFICATION POTENTIAL

 

S. Abushakra1, A. Porsteinsson2, P. Scheltens3, C. Sadowsky4, B. Vellas5, J. Cummings6, S. Gauthier7, J.A. Hey1, A. Power1, P. Wang8, L. Shen8, M. Tolar1

 

1. Alzheon Inc., Framingham, Massachusetts, USA; 2. University of Rochester, Rochester, New York, USA; 3. Vrije University Alzheimer Center, Amsterdam, Netherlands; 4. Nova Southeastern University, Ft. Lauderdale, Florida, USA; 5. University of Toulouse, Toulouse, France; 6. Cleveland Clinic Lou Ruvo Center for Brain Health, Las Vegas, Nevada, USA; 7. McGill University and McGill Center for Studies in Aging, Montreal, Canada; 8. Pharmapace Inc., San Diego, California, USA

Corresponding Author: Susan Abushakra, MD, Alzheon, Inc., 111 Speen Street, Suite 306, Framingham, MA 01701, USA, Phone:  508.861.7709, Fax:  508.861.1500, susan.abushakra@alzheon.com

J Prev Alz Dis 2017;4(3):149-156
Published online June 22, 2017, http://dx.doi.org/10.14283/jpad.2017.26

 


Abstract

Background: Alzheimer’s Disease (AD) patients homozygous for the APOE4 allele (APOE4/4) have a distinct clinical and biological phenotype with high levels of beta amyloid (Aβ) pathology and toxic Aβ oligomers. Tramiprosate, an oral agent that inhibits Aβ monomer aggregation into toxic oligomers, was evaluated in two Phase 3 Mild to Moderate AD studies which did not show efficacy in the overall population. Re-analyses of these trials showed the most consistent clinical benefits in APOE4/4 patients. We analyzed efficacy in the APOE4/4 patients with Mild disease.
Objectives: To determine the optimal stage of AD for future trials in APOE4/4 homozygotes.
Design: Two randomized, double-blind, placebo-controlled parallel-arm multi-center studies of 78-weeks duration.
Setting: Academic Alzheimer’s disease centers, community-based memory clinics, and neuropsychiatric research sites.
Participants: Participants included 2,025 AD patients with MMSE 16-26.  Approximately 13-15% had APOE4/4 genotype (N= 147 and 110 per study), mean age 71.1 years, 56% females. Almost all were on stable symptomatic drugs.
Intervention: Randomized subjects received oral placebo, 100mg BID, or 150mg BID of tramiprosate.
Measurements: Co-primary outcomes were change from baseline in the ADAS-cog11 and CDR-SB. Disability assessment for dementia (DAD) was a secondary outcome.
Results: In APOE4/4 homozygotes receiving 150mg BID tramiprosate, efficacy in the traditional Mild AD patients (MMSE 20-26) was higher than the overall group (MMSE 16-26) and efficacy in the Mild patients (MMSE 22-26) was highest. Tramiprosate benefits compared to placebo on ADAS-cog, CDR-SB, and DAD were 125%, 81% and 71%, respectively (p<0.02). The Mild subgroup (MMSE 22-26) showed cognitive stabilization with no decline over 78 weeks, both ADAS-cog and DAD effects increased over time. Tramiprosate safety in APOE4/4 patients was favorable. Most common adverse events were nausea, vomiting, depression and decreased weight.
Conclusions: The Mild subgroup of APOE4/4 AD patients (MMSE 22-26) showed larger benefits on the high dose of tramiprosate than the overall Mild and Moderate group. Consistent with its preclinical effects on Aβ oligomers, tramiprosate seemed to stabilize cognitive performance, supporting its disease modification potential. Confirmatory studies using ALZ-801, an improved pro-drug formulation of tramiprosate, will target APOE4/4 patients with Mild AD.

Key words: Tramiprosate, Alzheimer’s, APOE4, amyloid oligomers.

Abbreviations: AE: Adverse Event(s); AD: Alzheimer’s Disease; Aβ: Beta Amyloid; ADAS-cog: Alzheimer’s Disease Assessment Scale-cognitive Subscale; APOE4 = Apolipoprotein E4; ε4 allele of the apolipoprotein E gene; ARIA-E: Amyloid Related Imaging Abnormalities-Edema; CBL: Change from Baseline; CDR-SB: Clinical Dementia Rating Scale-Sum of Boxes; DAD: Disability Assessment for Dementia; EU: European; GI: Gastrointestinal; ITT: Intent-to-Treat; LS: Least Squares; MedDRA: Medical Dictionary for Regulatory Activities; MMRM: Mixed Effects Model with Repeated Measures; MMSE: Mini-Mental State Examination; MRI: Magnetic Resonance Imaging; NA: North American; NPI: Neuropsychiatric Inventory; SAP: Statistical Analysis Plan; SAE: Serious Adverse Event(s); TEAE: Treatment-Emergent Adverse Event(s)


 

 

Introduction

The ε4 allele of the apolipoprotein E gene (APOE4) is an important risk factor for Alzheimer’s disease (AD), second only to age (1). The APOE4 genotype confers a 4 to 12-fold higher risk of AD and lowers the age of onset by approximately 10-15 years (2, 3). APOE carrier proteins, which play an important role in membrane maintenance, are produced by astrocytes in response to injury and regulate neuronal Beta amyloid (Aβ) metabolism (4, 5). The APOE4 isoform reduces clearance of Aβ monomers, promotes their aggregation into soluble toxic oligomers and insoluble fibrils (5, 6), and may directly stimulate Aβ synthesis (7). These effects on Aβ are thought to underly the high risk of AD in APOE4 carriers. .
APOE4 carriers have a distinct biological and clinical phenotype across imaging, biomarker and clinical studies. APOE4 carriers have higher rates of positive amyloid scans across all stages of disease (8, 9), show greater cortical amyloid burden, regional hypometabolism (10), and more marked hippocampal atrophy than non-carriers (11). Their clinical profile reflects hippocampal deficits with delayed memory affected earlier than executive function or language (12, 13). APOE4/4 homozygotes exhibit significantly greater amyloid burden than heterozygotes (14). APOE4/4 homozygotes thus represent a biologically homogenous population of AD patients that is enriched for amyloid pathology.
Tramiprosate is an oral amyloid targeted agent that inhibits Aβ oligomer formation and fibrillar (plaque) deposition in transgenic animal models (15-17). Tramiprosate was previously evaluated in two Phase 3 trials in Mild to Moderate AD. The North American trial results did not show efficacy in the overall study population (18), and the European trial was terminated before completion. Analyses of these studies based on the number of APOE4 alleles showed a gene-dose effect revealing the largest efficacy signals in APOE4/4 homozygotes, intermediate in APOE4 heterozygotes and lowest in APOE4 non-carriers (19). This parallels the rank order of amyloid positivity in APOE4 subgroups from recent AD trials, with highest rates (~98%) in APOE4/4 homozygotes, and lowest rates (~60%) in non-carriers (20). In APOE4/4 homozygotes, tramiprosate showed promising clinical benefits in the overall Mild to Moderate AD group (19). Since Aβ toxicity is thought to play a larger role at the early stages of AD, we explored the efficacy of tramiprosate in two Mild subgroups of APOE4/4 patients.

 

Methods

Study Design

The Phase 3 tramiprosate studies included a North American (NA) study (NCT00088673) and a European (EU) study (NCT00217763). The NA and EU study results were previously reported (18, 19). These studies were conducted in accordance with ICH guidance and all local regulations. Both the NA study (Study CL-758007) and EU study (CL-758010) were multi-center, randomized, placebo-controlled, double-blind, parallel-arm studies of 78 weeks duration. The EU study had similar design and outcome measures, but was terminated before completion when the NA study did not show significant efficacy in the overall study population.

Study Participants

Both studies enrolled AD patients with baseline Mini-Mental State Examination (MMSE) scores of 16-26, inclusive. The NA study included 67 centers in the US and Canada, and the EU study included centers from 11 Western European countries. The AD diagnosis was based on clinical criteria with brain MRI or CT imaging as supportive evidence, since amyloid imaging was not widely available at the time. Participants were enrolled and randomized into one of three arms (placebo, 100mg BID and 150mg BID of tramiprosate) and received study drug for 78 weeks, with study visits occurring every 13 weeks. The NA study allowed treatment with cholinesterase inhibitors alone, or with memantine, at stable doses for at least 12 weeks. The EU study allowed cholinesterase inhibitors only, and prohibited memantine use.

Outcome Measures

Co-primary outcomes in both studies were changes from baseline (CBL) to Week 78 on the 11-item Alzheimer’s Disease Assessment Scale-Cognitive Subscale (ADAS-cog) (21), and on the Clinical Dementia Rating – Sum of Boxes (CDR-SB) (22). The CDR-SB combines cognitive and functional measures. Secondary outcomes included the Disability Assessment for Dementia (DAD) (23) which evaluates function and disability; the MMSE (24) which is a cognitive staging instrument; and the 12-item Neuropsychiatric Inventory (NPI) (25) which evaluates the presence and severity of 12 neuropsychiatric symptoms that commonly occur in dementia patients.  Higher values on the ADAS-cog, CDR and NPI indicated worsening, while higher values on the DAD and MMSE indicate improvement.

Clinical Datasets: NA and EU Studies

The NA and EU studies included 973 and 870 participants, respectively, with known genotype, of whom 147 and 110 were APOE4/4 homozygotes. In these APOE4/4 homozygous subgroups, efficacy and safety datasets were analyzed separately and as a pooled dataset (discussed below).

Analyses by APOE4 Genotype

APOE4 status was one of the pre-specified covariates in the statistical analysis plan (SAP). The APOE4 genotype-treatment interaction was found to be significant for both primary outcomes (p< 0.05). In addition, there was an overall significant treatment effect by APOE4 by visit interaction for ADAS-cog (p< 0.05). These results suggested that treatment effects are dependent on APOE4 status. Therefore, an analysis in APOE4 carriers and non-carriers was performed, and showed significant efficacy in APOE4 carriers. We recently further analyzed and published the efficacy results of both studies based on the number of APOE4 alleles (19). A subset of patients from both studies had MRI at baseline and end of study, the MRI safety analyses for vasogenic edema (ARIA-E) were recently reported (19).

Efficacy analyses

Efficacy analyses in this report are for subjects in the intent-to-treat (ITT) population of each study, who had a baseline and at least one post-baseline assessment, received at least one dose of study drug, and carried the APOE4/4 genotype.
The mixed effect model with repeated measures (MMRM) was the analytical method specified in the prospectively developed SAP for the NA study. This SAP had undergone Food and Drug Administration review prior to database lock on March 30, 2007. This model included the following terms: treatment, visit, treatment by visit interaction, study site, and baseline value of the primary endpoint to be analyzed. The analyses were performed using this original MMRM. The SAP of the NA study included the following factors as covariates to be tested for interaction with drug effect: age, gender, use of symptomatic drugs, and APOE4 genotype. The SAP also required that efficacy results be summarized by APOE genotype.
An additional MMRM analysis termed ‘updated MMRM’ was performed that included baseline MMSE values in the model to be consistent with current practice (19). Both MMRM methods yielded similar results and conclusions. Results presented in this paper are from the updated MMRM analyses. Study sites recruiting few patients were pooled to avoid having a site with only one subject from a dose arm in an APOE4 subgroup.
The EU Phase 3 study was prematurely discontinued when the results of the NA study became available. Most patients had completed the Week 52 visit while approximately 30% had completed Week 78 evaluations. Therefore, the last two visits for subgroup analyses with each of the Mild subgroups had a small sample size and MMRM analysis was not suitable. The EU dataset was combined with the NA dataset and analyzed by MMRM, and is considered a supportive sensitivity analysis.

Safety analyses

The combined NA and EU (Pooled) safety population included all subjects who received at least one dose of study drug and who had the APOE4/4 genotype. Safety analyses included treatment-emergent adverse events (TEAE) and serious adverse event (SAE). Adverse events were coded using MedDRA version 9.0. An exploratory comparison of incidence of TEAE between placebo and the two active arms was performed using Fisher’s exact test.

Exploratory Analyses Based on Baseline Disease Severity

Additional exploratory analyses were performed based on baseline disease severity, and evaluated efficacy on ADAS-cog, CDR-SB, DAD, NPI and MMSE. Patients were classified by baseline MMSE category (16-26; 20-26; 22-26), and data analyzed using the same MMRM method as the primary analysis, with the exception that site was not included as a factor in the model due to the smaller number of patients per site in the Mild subgroups.
Since these APOE4 subgroup analyses are exploratory and are being used to inform the design of future studies, the p-values presented were not adjusted for multiple statistical comparisons. In this paper, the MMSE category of 20-26 is termed the ‘traditional Mild AD’ group, and the MMSE category of 22-26 is termed ‘Mild AD’.

Exploratory Analyses of Divergence of Slopes between Dose Arms

To explore the disease modification potential of tramiprosate, the difference in slopes between treatment arms was evaluated for statistical significance. An MMRM model was used for these analyses, with model terms including treatment, visit days as a numerical variable, baseline value of the response variable, baseline MMSE value, and treatment by visit-days interaction.

Specification of Model Variables (MMRM)

Response Variables

The primary outcome variables for all inferential statistical analyses were the changes from baseline (CBL) to each visit i for ADAS-cog and CDR-SB, as defined by:
CBL(ADAS-cog)Visit i=ADAS-cog)Visit i-ADAS-cog)Baseline
CBL(CDR-SB)Visit i=(CDR-SB)Visit i-(CDR-SB)Baseline

Independent Variables

The treatment group was regarded as a class variable with three levels: (1) tramiprosate 150 mg BID, (2) tramiprosate 100 mg BID, and (3) placebo. The visit variable was included as a class variable with values standing for the visit numbers (Visit 5 to Visit 10, corresponding to Weeks 13 to 78), and was used to index the within-subject observations over time.

Covariates

The baseline value of the dependent variable was entered as a continuous covariate since both absolute and percent changes over time in the outcome measure correlate with its value at the baseline visit. Baseline MMSE score was used as a measure of AD severity at baseline and was entered as a continuous covariate. The autoregressive order 1 covariance structure was used to model the within-subject correlation across the different visits.

 

Results

Demographics and Baseline Characteristics

The NA and EU studies enrolled a total of 1052 and 973 patients, respectively. The ITT population from the NA and EU studies with known APOE genotypes consisted of 973 and 870 patients, of whom 147 and 110 were APOE4/4 homozygotes, respectively. The demographics of APOE4 non-carriers, heterozygotes and homozygotes were balanced except for a younger mean age in APOE4/4 homozygotes (19).
Demographics and baseline scores of the APOE4/4 homozygous subgroups in each study are shown in Table 1. In the NA study, the mean age was 72 years, 57% were female, and 97% were caucasian. In the EU study, the mean age was 70 years, 56% were female, and 100% were Caucasian. In both studies ~99% of subjects were on cholinesterase inhibitors. The NA study had ~48% of patients taking memantine, while the EU study did not allow memantine use. APOE4/4 subgroups in each study showed similar patient demographics across all three dose arms.
Baseline clinical scores in the NA and EU studies were respectively: MMSE 20.9 and 21.5; ADAS-cog 22.3 and 21.6; CDR-SB 5.8 and 5.7; DAD 76.3 and 76.1; and NPI 8.8 and 9.5. In the NA study the low dose arm showed higher ADAS-cog but lower CDR-SB baseline scores than the other two arms. In the EU study the placebo showed higher ADAS-cog and CDR-SB scores than the active dose arms.

Table 1. Demographics and Baseline Characteristics in APOE4/4 Homozygotes in Each Study (All Enrolled)

Table 1. Demographics and Baseline Characteristics in APOE4/4 Homozygotes in Each Study (All Enrolled)

Higher values for ADAS-cog, CDR-SB and NPI indicate greater disease severity, and lower values in DAD and MMSE indicate greater severity. APOE4 = apolipoprotein E4; AChEIs = acetylcholinesterase inhibitors; MMSE = Mini Mental State Examination; ADAS-cog = Alzheimer’s Disease Assessment Scale-cognitive subscale; CDR-SB = Clinical Dementia Rating Sum of Boxes; DAD = Disability Assessment for Dementia; NPI = Neuropsychiatric Inventory.

 

Efficacy

Efficacy Outcomes in the APOE4/4 Subgroup by Baseline MMSE Range- NA study

In Table 2 the effects of both tramiprosate doses (100mg BID and 150mg BID) on the co-primary outcomes ADAS-cog and CDR-SB are shown for the last two visits (Weeks 65 and 78) for the NA Study. The overall group includes Mild and Moderate subjects (MMSE 16-26), while the two Mild groups include subjects with MMSE 20-26 and 22-26 (Table 2). At the high dose the effects on both ADAS-cog and CDR-SB increase progressively and become more significant as the MMSE range is limited to the traditional Mild group (20 and above), and then to the Mild group (22 and above).

Table 2. Tramiprosate Effects on Co-Primary Outcomes in APOE4/4 Homozygotes: Mild/Moderate and Mild Subgroups (NA Study)

Table 2. Tramiprosate Effects on Co-Primary Outcomes in APOE4/4 Homozygotes: Mild/Moderate and Mild Subgroups (NA Study)

Note: *MMSE 20-26 is the “traditional Mild group”. Change from Baseline (CBL), is the LS mean differences in ADAS, CDR-SB between active arm and placebo. Negative values for CBL of ADAS-cog and CDR-SB indicate benefit in favor of drug. Bolded values indicate drug effects with either nominal statistical significance (p< 0.05) or positive trends (p< 0.1).

 

The time course of tramiprosate effects on ADAS-cog, CDR-SB and DAD is shown at each visit in Figure 1. In the Mild group (MMSE 22 and above) at the high dose, cognitive scores improve and remain above baseline for 78 weeks; CDR-SB also remains above baseline for 65 weeks and shows a small decline compared to baseline at 78 weeks. In this Mild group the DAD also shows increasing benefit with time and reaches significance at 78 weeks. Due to this observation of divergent treatment effects between the 150 mg and placebo groups, an additional analysis was performed to examine the difference in the slopes of these two arms. The difference in slopes between placebo and the high dose arm was significant for ADAS-cog (p=0.015), non-significant for CDR-SB (p=0.114), and showed a positive trend for DAD (p=0.068).

Figure 1. Time Course of Effect in APOE4/4 Homozygotes in the Overall and Mild Subgroup (MMSE 22-26, NA Study)

Figure 1. Time Course of Effect in APOE4/4 Homozygotes in the Overall and Mild Subgroup (MMSE 22-26, NA Study)

CBL determined by MMRM analysis, error bars are SEM. For ADAS-cog and CDR-SB, negative CBL indicate improvement, for DAD positive CBL indicates improvement.  DAD was performed every 26 weeks only. # indicates p< 0.1; * indicates p< 0.05; ** indicates p< 0.01.

 

The percentage benefit compared to placebo in the Mild subgroup (MMSE 22-26) that showed the largest effects is shown in Table 3. For ADAS-cog and CDR-SB, and for the secondary outcomes of DAD, NPI and MMSE, results are shown from the last three visits at which they were assessed. In AD trials a benefit of at least 25% compared to placebo is considered the minimal clinically relevant benefit.  The percentage benefits on ADAS-cog, CDR-SB and DAD (125%, 81% and 71%, respectively) are well above that threshold. For NPI and MMSE the percentage benefit of 53% and 24%, respectively, was not statistically significant, but was directionally positive.
In the Mild group (MMSE 22-26) at the 78-Week endpoint, the effect size (Cohen’s d) and 95% confidence intervals (CI) were as follows: for ADAS-cog -0.54 (Lower CI: -0.82; Upper CI: -0.25); for CDR-SB -0.39 (Lower CI: -0.71; Upper CI: -0.06); and for DAD 0.52 (Lower CI: 0.15; Upper CI: 0.89).

Table 3. Effects of Tramiprosate 150mg BID in APOE4/4 Patients with Mild AD (MMSE 22-26, NA Study): Percent Drug Benefit Compared to Placebo

Table 3. Effects of Tramiprosate 150mg BID in APOE4/4 Patients with Mild AD (MMSE 22-26, NA Study): Percent Drug Benefit Compared to Placebo

* % Drug benefit is indicated as positive when drug is better than placebo. For ADAS-cog, CDR-SB, and NPI negative differences indicate drug benefit, for DAD and MMSE positive differences indicate drug benefit. Bolded values indicate either statistical significance (nominal p< 0.05) or positive trends (p< 0.1).  CBL: Change from Baseline.

 

Efficacy Outcomes in the APOE4/4 Subgroup by Baseline MMSE Range- Combined Studies

In the combined datasets, the time course of tramiprosate effects on ADAS-cog, CDR-SB and DAD is shown in Figure 2, for the overall (Mild and Moderate group) and Mild group (MMSE 22-26). For the high dose least squares (LS) mean differences in the overall group at 78 weeks on ADAS-cog, CDR-SB and DAD were -2.3, p=0.04; -0.4, p=0.26; and -1.8, p=0.49, respectively; and in the Mild group were -4.8, p=0.001; -0.9, p=0.05; 8.4, p=0.02, respectively. The effects of the high dose on all three outcomes at the study endpoint were again higher in the Mild group and achieved nominal significance.

Figure 2. Time Course of Effect in APOE4/4 Homozygotes in the Overall and Mild Subgroup (MMSE 22-26, Combined Studies)

Figure 2. Time Course of Effect in APOE4/4 Homozygotes in the Overall and Mild Subgroup (MMSE 22-26, Combined Studies)

CBL determined by MMRM analysis, error bars are SEM. For ADAS-cog and CDR-SB, negative CBL indicate improvement, for DAD positive CBL indicates improvement. DAD was performed every 26 weeks only. # indicates p< 0.1; * p< 0.05; ** p< 0.01.

 

Safety

Across the two studies there were 263 APOE4/4 patients in the safety population. The most common TEAE are shown in Table 4. The nature of adverse events in the APOE4/4 group is similar to the overall study population and there were no events of ARIA-E on active drug (19). The incidence of nausea, vomiting, depression and weight loss was higher in the high dose arm than placebo, but was not statistically significant (nominal p-value >0.05). The majority of nausea and vomiting events were mild or moderate and less than five percent led to discontinuation. The incidence of serious AE was lower in active arms (14% low dose, 11% high dose) than placebo (22%). The incidence of headache was lower in the high dose (1%) arm than placebo (11%, nominal p-value = 0.013).

Table 4. APOE4/4 Homozygotes: TEAE with incidence > 5% in overall group* (Combined NA and EU Safety Population)

Table 4. APOE4/4 Homozygotes: TEAE with incidence > 5% in overall group* (Combined NA and EU Safety Population)

TEAE with incidence > 5% in the 3 combined dose arms, and higher incidence in an active arm than placebo, are listed by descending frequency in the high dose arm. The incidence of nausea, vomiting, depression and decreased weight were not significantly higher in active dose arms than placebo (p >0.05). *Note: Headache is included since incidence in the high dose arm was significantly lower than placebo p= 0.013.

 

Discussion

In APOE4/4 homozygotes the sensitivity analysis of tramiprosate efficacy based on disease stage suggests that the Mild subgroup with MMSE 22 and above is more likely to have larger and more sustained benefits compared to the overall Mild and Moderate group. Based on both the NA and NA/EU datasets, the ADAS-cog drug-placebo differences increase with time, and cognitive scores remain stable for 78 weeks (Figures 1D, 2D). The CDR-SB effect is also sustained at more than one point clinical benefit over placebo at 78 weeks (Figure 1E). In both datasets, the differences in disability (DAD) also increase with time (Figures 1F and 2F). The low tramiprosate dose in the Mild subgroup shows smaller and less consistent numerical benefits. At the high dose, the ADAS-cog divergence of slopes was statistically significant, and the DAD showed a positive trend.
The magnitude of benefits over placebo on ADAS-cog, CDR-SB, and DAD (125%, 81% and 71%, respectively) was large and well above the 25% threshold that is considered the minimum clinically meaningful difference by clinical and regulatory experts. This is supported by the effect sizes at 78-Week endpoint being > 0.5 for ADAS-cog and DAD, and ~ 0.4 for CDR-SB.
AD patients who are APOE4/4 homozygotes represent a population with a significant burden of amyloid pathology including oligomers, and this may explain their reported preferential response to tramiprosate, which inhibits assembly of Aβ oligomers (6, 7). The improved clinical response in the Mild AD subgroups is also consistent with preclinical effects of tramiprosate on Aβ aggregation (17), since soluble Aβ oligomers are thought to play an important and early role in synaptic toxicity and neurodegeneration in AD (26, 27). This is supported by Amyloid imaging data from clinical trials showing ongoing plaque deposition in APOE4 carriers at the Mild but not the Moderate stage of disease (28).
The magnitude of the tramiprosate effect on CDR-SB in this analysis is consistent with the efficacy signal from a study in Early/Mild AD with the anti-amyloid antibody aducanumab (29). However, the efficacious doses of aducanumab in the latter study were associated with a high incidence of vasogenic edema/ARIA-E in APOE4 carriers; even with a dose titration regimen the incidence of ARIA-E was 35% (30).
The main limitation of the current efficacy analyses is that they are post-hoc subgroup analyses from the NA study that did not achieve its primary objectives. The sample size of the APOE4/4 dataset was also limited, especially in the Mild subgroups. Within the APOE4/4 homozygous patients, randomization was not stratified based on baseline severity and the three dose arms were not precisely comparable, although the MMRM analysis included baseline severity in the model. Therefore, these findings require confirmation in prospectively defined studies in the APOE4/4 homozygous patients at the Mild AD stage (MMSE 22 and above).
From the safety perspective, tramiprosate showed a favorable safety profile in APOE4/4 homozygotes, with a similar profile to the other groups. The most common gastrointestinal events of nausea and vomiting are most likely mediated by a direct, local gastrointestinal (GI) irritation effect of tramiprosate in some subjects. This GI tolerability is being addressed by development of ALZ-801, a new pro-drug formulation (31). ALZ-801 has shown substantially improved GI tolerability as well as more consistent plasma levels of tramiprosate in Phase 1b studies in over 160 healthy elderly subjects (31). For future studies, ALZ-801 will be used at a dose that provides bioequivalent exposures to tramiprosate 150 mg BID. Regarding the risk of vasogenic edema, MRI analyses from the two Phase 3 studies showed no evidence of vasogenic edema/ARIA-E with tramiprosate (19), even in the high risk group of APOE4 carriers (29).

 

Conclusions

Reanalysis of two prior tramiprosate studies showed promising efficacy in the APOE4/4 homozygous patients with Mild and Moderate AD. In this sensitivity analysis, the subgroup of Mild AD patients (MMSE 22 and above) shows larger and more sustained clinical effects on the high dose of tramiprosate. The cognitive scores improve and remain above baseline throughout the 78 weeks of treatment. The divergence of slopes (non-parallel slopes) between the two arms was significant. These effects on ADAS-cog were associated with significant and sustained benefits on function and disability over the study duration. This overall efficacy profile is supportive of a potential disease modifying effect and is consistent with the molecular mechanism of tramiprosate, inhibiting formation of toxic Aβ oligomers. The safety profile in the overall APOE4/4 homozygous group was favorable. Prospective confirmation of these clinical effects in APOE4/4 homozygotes at the Mild stage of AD is warranted. The focus on this genetically-defined population represents a precision medicine approach to AD that has been successful in other challenging diseases.

 

Acknowledgements: We thank the investigators, their staff, and the individuals with AD who participated in the trials and their caregivers. We thank Dr. Sandra Saouaf and Christine Rathbun for editorial assistance. We thank our scientific advisors for their valuable review and commentary.

Funding: Alzheon Inc. funded the above analyses and manuscript preparation. Bellus Health (previously Neurochem), the original sponsor, had funded the tramiprosate Phase 3 studies.

Conflict of interest: Drs. Abushakra, Hey, Power and Tolar are employees of Alzheon, Inc. and hold stock or stock options of Alzheon, Inc. Drs. Porsteinsson, Scheltens, Sadowsky, Cummings, Gauthier, Vellas were investigators in the tramiprosate Phase 3 program and serve as advisors to Alzheon, Inc., and receive advisory fees and/or stock options of Alzheon. Drs. Shen and Mr. Wang serve as statistical consultants to Alzheon.

Ethical standards: The study protocols were approved by local ethics committees, and were conducted in accordance with ICH standards and all local regulations.

 

References

1.     Roses AD. Apolipoprotein E alleles as risk factors in Alzheimer’s disease. Annu Rev Med. 1996; 47:387-400.
2.    Breitner JC, Wyse BW, Anthony JC, et al. APOE-epsilon4 count predicts age when prevalence of AD increases, then declines: the Cache County Study. Neurology. 1999; 53:321-331.
3.    Raber J, Huang Y, Ashford JW. ApoE genotype accounts for the vast majority of AD risk and AD pathology. Neurobiol Aging. 2004; 25:641-650.
4.    Mahley RW, Huang Y. Apolipoprotein E Sets the Stage: Response to Injury Triggers Neuropathology. Neuron. 2012; 76:871-885.
5.    Holtzman DM, Herz J, Bu G. Apolipoprotein E and Apolipoprotein E Receptors: Normal Biology and Roles in Alzheimer Disease. Cold Spring Harb Perspect Med. 2012; 2:a006312.
6.    Hashimoto T, Serrano-Pozo A, Hori Y, et al. Apolipoprotein E, especially apolipoprotein E4, increases the oligomerization of amyloid β peptide. J Neurosci. 2012; 32:15181-15192.
7.    Huang YA, Zhou B, Wernig M, Sudhof TC. ApoE2, ApoE3, and ApoE4 Differentially Stimulate APP Transcription and Aβ Secretion. Cell. 2017; 168:427-441 e421.
8.    Jansen WJ, Ossenkoppele R, Knol DL, et al. Prevalence of cerebral amyloid pathology in persons without dementia: a meta-analysis. JAMA. 2015; 313:1924-1938.
9.    Ossenkoppele R, Jansen WJ, Rabinovici GD, et al. Prevalence of amyloid PET positivity in dementia syndromes: a meta-analysis. JAMA. 2015; 313:1939-1949.
10.    Ossenkoppele R, van der Flier WM, Zwan MD, et al. Differential effect of APOE genotype on amyloid load and glucose metabolism in AD dementia. Neurology. 2013; 80:359-365.
11.    Liu Y, Yu JT, Wang HF, et al. APOE genotype and neuroimaging markers of Alzheimer’s disease: systematic review and meta-analysis. J Neurol Neurosurg Psychiatry. 2015; 86:127-134.
12.    Caselli RJ, Dueck AC, Osborne D, et al. Longitudinal modeling of age-related memory decline and the APOE epsilon4 effect. N Engl J Med. 2009; 361:255-263.
13.    Whitehair DC, Sherzai A, Emond J, et al. Influence of apolipoprotein E ε4 on rates of cognitive and functional decline in mild cognitive impairment. Alzheimers Dement. 2010; 6:412-419.
14.    Morris JC, Roe CM, Xiong C, et al. APOE predicts amyloid-beta but not tau Alzheimer pathology in cognitively normal aging. Ann Neurol. 2010; 67:122-131.
15.    Gervais F, Paquette J, Morissette C, et al. Targeting soluble Abeta peptide with Tramiprosate for the treatment of brain amyloidosis. Neurobiol Aging. 2007; 28:537-547.
16.    Martineau E, de Guzman JM, Rodionova L, Kong X, Mayer PM, Aman AM. Investigation of the noncovalent interactions between anti-amyloid agents and amyloid beta peptides by ESI-MS. J Am Soc Mass Spectrom. 2010; 21:1506-1514.
17.    Kocis P, Tolar M, Yu J, et al. Tramiprosate, an Anti-Amyloid Agent for the Treatment of Alzheimer’s Disease, Inhibits Aß42 Oligomerization via a Novel Enveloping Mechanism. CNS Drugs. 2017. doi:10.1007/s40263-017-0434-z.
18.    Aisen PS, Gauthier S, Ferris SH, et al. Tramiprosate in mild-to-moderate Alzheimer’s disease – a randomized, double-blind, placebo-controlled, multi-centre study (the Alphase Study). Arch Med Sci. 2011; 7:102-111.
19.    Abushakra S, Porsteinsson A, Vellas B, et al. Clinical benefits of tramiprosate in Alzheimer’s disease are associated with higher number of Apoe4 alleles: the “Apoe4 gene-dose effect”. J Prev Alz Dis. 2016; 3(4):219-228.
20.    Degenhardt EK, Witte MM, Case MG, et al. Florbetapir F18 PET Amyloid Neuroimaging and Characteristics in Patients With Mild and Moderate Alzheimer Dementia. Psychosomatics. 2016; 57:208-216.
21.    Mohs RC, Cohen L. Alzheimer’s Disease Assessment Scale (ADAS). Psychopharmacol Bull. 1988; 24:627-628.
22.    Morris JC. The Clinical Dementia Rating (CDR): current version and scoring rules. Neurology. 1993; 43:2412-2414.
23.    Gelinas I, Gauthier L, McIntyre M, Gauthier S. Development of a functional measure for persons with Alzheimer’s disease: the disability assessment for dementia. Am J Occup Ther. 1999; 53:471-481.
24.    Folstein MF, Folstein SE, McHugh PR. «Mini-mental state». A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975; 12:189-198.
25.    Cummings JL. The Neuropsychiatric Inventory: assessing psychopathology in dementia patients. Neurology. 1997; 48:S10-16.
26.    Selkoe DJ, Hardy J. The amyloid hypothesis of Alzheimer’s disease at 25 years. EMBO Mol Med. 2016; 8:595-608.
27.    Viola KL, Klein WL. Amyloid beta oligomers in Alzheimer’s disease pathogenesis, treatment, and diagnosis. Acta Neuropathol. 2015; 129:183-206.
28.    Liu E, Schmidt ME, Margolin R, et al. Amyloid-beta 11C-PiB-PET imaging results from 2 randomized bapineuzumab phase 3 AD trials. Neurology. 2015; 85:692-700.
29.    Sevigny J, Chiao P, Bussiere T, et al. The antibody aducanumab reduces Abeta plaques in Alzheimer’s disease. Nature. 2016; 537:50-56.
30.    Viglietta V, O’Gorman J, Williams L, et al. Aducanumab titration dosing regimen: 12-month interim analysis from prime, a randomized, double-blind, placebo-controlled phase 1b study in patients with prodromal or mild Alzheimer’s disease. J Prev Alz Dis. 2016; 3:278.
31.    Hey J, Abushakra S, Power A, Yu J, Versavel M, Tolar M. Phase 1 development of Alz-801, a novel beta amyloid anti-aggregation prodrug of tramiprosate with improved drug properties, supporting bridging to the phase 3 program. Alzheimer’s & Dementia: The Journal of the Alzheimer’s Association. 2016; 12:P613.

CLINICAL BENEFITS OF TRAMIPROSATE IN ALZHEIMER’S DISEASE ARE ASSOCIATED WITH HIGHER NUMBER OF APOE4 ALLELES: THE “APOE4 GENE-DOSE EFFECT”

 

S. Abushakra1, A. Porsteinsson2, B. Vellas3, J. Cummings4, S. Gauthier5, J.A. Hey1, A. Power1, S. Hendrix6, P. Wang7, L. Shen7, J. Sampalis8, M. Tolar1

 

1. Alzheon Inc., Framingham, MA, USA; 2. University of Rochester, Rochester, NY, USA; 3. University of Toulouse, Toulouse, France; 4. Cleveland Clinic Lou Ruvo Center for Brain Health, Las Vegas, NV, USA; 5. McGill University and McGill Center for Studies in Aging, Montreal, Canada; 6. Pentara Corporation, Salt Lake City, UT, USA; 7. Pharmapace Inc., San Diego, CA, USA; 8. McGill University and JSS Medical Research Inc., Montreal, Canada

Corresponding Author: Susan Abushakra, MD, Alzheon, Inc., 111 Speen Street, Suite 306, Framingham, MA 01701, USA, Phone:  508.508.7709, Fax:  508.861.1500, susan.abushakra@alzheon.com

J Prev Alz Dis 2016;3(4):219-228
Published online October 24, 2016, http://dx.doi.org/10.14283/jpad.2016.115

 


Abstract

Background: Tramiprosate is an oral amyloid anti-aggregation agent that reduces amyloid oligomer toxicity in preclinical studies and was evaluated in two 78-week trials in North America and Western Europe that enrolled 2,025 patients with Mild to Moderate Alzheimer’s Disease. The completed North American study did not achieve its efficacy objectives, but a pre-specified subgroup analysis suggested potential efficacy in apolipoprotein E4 (APOE4) carriers. To further explore this observation, we analyzed tramiprosate Phase 3 clinical data based on the number of APOE4 alleles.
Objectives: To analyze tramiprosate efficacy, safety, and occurrence of vasogenic edema in the three APOE4 subgroups: homozygous, heterozygous and non-carriers.
Design: Randomized, double-blind, placebo-controlled parallel-arm multi-center studies.
Setting: Academic Alzheimer’s disease & dementia centers, community-based dementia and memory clinics, and neuropsychiatric clinical research sites.
Participants: Subjects included 2,025 patients, 50 years of age or older, with approximately 60% having APOE4 carrier status (10-15% homozygotes and 45-50% heterozygotes), and mild to moderate disease.  All subjects were on stable symptomatic drugs.
Intervention: Randomized subjects received placebo, 100 mg BID, or 150 mg BID of tramiprosate.
Measurements: Co-primary outcomes in both studies were change from baseline in the ADAS-cog11 and CDR-SB assessment scales.
Results: Highest efficacy was observed in APOE4/4 homozygotes receiving 150 mg BID of tramiprosate, showing statistically significant effects on ADAS-cog and positive trends on CDR-SB (respectively, 40-66% and 25-45% benefit compared to placebo). APOE4 heterozygotes showed intermediate efficacy, and non-carriers showed no benefit. In 426 patients with MRI scans, no cases of treatment-emergent vasogenic edema were observed. In the three subgroups, the most common adverse events were nausea, vomiting, and decreased weight.
Conclusions: The “APOE4 Gene-Dose effect” is likely explained by the high prevalence of amyloid pathology in symptomatic APOE4 carriers. In APOE4/4 Alzheimer’s disease patients, the high dose of tramiprosate showed favorable safety and clinically meaningful efficacy in addition to standard of care.

Key words: Tramiprosate, Alzheimer’s, APOE4.

Abbreviations: AD: Alzheimer’s disease; Aβ: Beta amyloid; ADAS-cog: Alzheimer’s Disease Assessment Scale-cognitive subscale; APOE4: Apolipoprotein E4, ε4 allele of the apolipoprotein E gene; ARIA-E: Amyloid-Related Imaging Abnormalities-Edema; ARIA-H: Amyloid-Related Imaging Abnormalities-Haemosiderin; CBL: Change from Baseline; CDR-SB: Clinical Dementia Rating Scale-Sum of Boxes; DAD: Disability Assessment for Dementia; EU: European; FLAIR: Fluid-Attenuated Inversion Recovery; ITT: Intent to Treat; MedDRA:  Medical Dictionary for Regulatory Activities; MMRM: Mixed Effects Repeated Measures Model; MMSE: Mini-Mental State Examination; MRI: Magnetic Resonance Imaging; NA: North American; NPI: Neuropsychiatric Inventory; OC: Observed Cases; SAE: Serious Adverse Event; SAP: Statistical Analysis Plan; TEAE: Treatment-Emergent Adverse Events


 

Introduction 

The ε4 allele of the apolipoprotein E gene (APOE4) is the most important genetic risk factor for Alzheimer’s disease (AD), second only to age in determining the risk for developing AD dementia (1).  APOE alleles encode carrier proteins that are important in cholesterol and β-amyloid (Aβ) metabolism and clearance, with the APOE4 isoform showing reduced clearance of Aβ peptides and promoting their aggregation (2). APOE4 carriers have deficient Aβ clearance from brain and excessive amyloid deposition associated with markers of neurodegeneration (3). The APOE4 genotype confers a 4-fold to 12-fold higher risk of developing AD, and decreases the mean age of onset of AD by approximately 10-15 years (4, 5). APOE4 carriers also show more rapid progression from early AD to dementia (6). Neuroimaging as well as neuropathological and cerebrospinal fluid (CSF) samples from AD patients who are APOE4 carriers indicate high levels of Aβ pathology with amyloid deposition in both cortex and cerebral vasculature (3, 7, 8). The high burden of vascular amyloid in APOE4 carriers has been associated with increased risk of vasogenic edema and microhemorrhage or Amyloid Related Imaging Abnormalities (ARIA-E or ARIA-H) with some amyloid-targeted agents (9). Due to the susceptibility of APOE4 carriers, especially homozygotes, to occurrence of ARIA, clinical trials with most amyloid targeted agents now include APOE4 genotyping and brain MRI as a safety evaluation. Amyloid PET-imaging studies of patients with mild cognitive changes have shown highest rates of positive scans in homozygotes, intermediate rates in heterozygotes, and lowest rates in non-carriers (10). Similar findings were reported in Mild to Moderate AD patients (11). It is therefore plausible that an amyloid-targeted AD drug could have differential efficacy based on the number of APOE4 alleles in treated subjects. We evaluated clinical data from two Phase 3 trials with tramiprosate based on the number of APOE4 alleles.          

Tramiprosate is an oral amyloid anti-aggregation agent that reduces oligmeric and fibrillar (plaque) amyloid in transgenic animal models (12, 13). In a Phase 2 study in AD patients, tramiprosate was found to cross the blood-brain barrier and to dose-dependently reduce CSF Aβ42 levels, with maximum reductions at the highest tested dose of 150 mg BID (14). This allowed selection of 100 mg BID and 150 mg BID doses for the Phase 3 program. The Phase 3 program included two similarly designed, 78-week trials in patients with mild to moderate AD. The North American (NA) trial results that became available in 2007 did not show significant efficacy on the Alzheimer’s Disease Assessment Scale-cognitive subscale (ADAS-cog) and Clinical Dementia Rating Scale-Sum of Boxes (CDR-SB) co-primary outcomes (15). The European (EU) trial was therefore terminated prior to completion and the study results were not previously published. In both trials the diagnosis of AD was made based on clinical criteria, with brain MRI or CT imaging as supportive evidence. APOE4 status was one of the pre-specified covariates which was found to be significant for both primary outcomes (p < 0.05), and an initial analysis showed an efficacy trend in APOE4 carriers. We have further systematically analyzed the primary, secondary and safety outcomes of both studies based on the number of APOE4 alleles in APOE4 homozygous (2 alleles), heterozygous (1 allele) and non-carrier patients (0 alleles).

 

Methods

Study Design

The NA study was a randomized, placebo-controlled, double-blind, parallel-arm study (Study CL-758007) which enrolled subjects with baseline Mini-Mental State Examination (MMSE) of 16-26, inclusive, at 67 centers in the US and Canada. Subjects were randomized into one of three arms (placebo, 100 mg BID and 150 mg BID) and received study drug for 78 weeks, with visits occurring every 13 weeks (15). Subjects were allowed stable doses of cholinesterase inhibitors, alone or with memantine, for at least 12 weeks. The EU study (CL-758010) had a similar design, but prohibited memantine use, and enrolled subjects from 10 European countries. Co-primary outcomes of both studies were changes from baseline (CBL) to week 78 on the 11-item ADAS-cog (16) and on the CDR-SB (17). The CDR-SB combines cognitive and functional measures. Secondary outcomes included the Disability Assessment for Dementia (DAD) (18), the MMSE (19), and the 12-item Neuropsychiatric Inventory (NPI) (20) which evaluates neuropsychiatric symptoms that are common in dementia patients.

Clinical Data Analyses

APOE4 status was one of the pre-specified covariates which was found to be significant for both primary outcomes (p < 0.05). In addition, there was an overall significant treatment by APOE4 by visit interaction for ADAS-cog (p < 0.05).  Treatment by APOE4 interaction at individual visits was significant for both outcomes at weeks 52 and 65, and for CDR-SB at week 78 as well (p < 0.05).  These results suggest that placebo-corrected treatment effects are dependent on APOE4 status. Therefore an analysis in APOE4 carriers and non-carriers was performed, and showed an efficacy trend in APOE4 carriers.
Efficacy and safety data from each study were analyzed based on APOE4 genotype status in homozygous patients (APOE4/4 genotype), heterozygous (mostly APOE3/4) and APOE4 non-carriers (mostly APOE3/3). Safety analyses for treatment-emergent adverse events (TEAE) and serious adverse event (SAE) were analyzed in each study and in the combined dataset. Adverse events were coded according to MedDRA version 9.0.

Centralized Brain MRI Scan Analyses

Both studies included brain MRI imaging in a subset of patients for volumetric assessments (21), but centralized MRI assessments for vasogenic edema were not planned since the occurrence of ARIA in AD trials was not yet described. Blinded paired MRI scans were available for ARIA assessment. MRI scans were performed at screening and at week 78 or at the early termination visit, if the patient discontinued after 6 months in study. The scans included Fluid-Attenuated Inversion Recovery (FLAIR) sequences that were used for ARIA-E assessments. Subjects who had MRI scans after study treatment only were included in the analysis since spontaneous ARIA-E is uncommon. Neuroradiologists (Synarc/BioClinica®), who were blinded to patients’ treatment and genotype, evaluated all scans according to a standardized protocol.

Statistical Analyses

Efficacy analyses

Two independent biostatistical consulting groups were retained to independently conduct the efficacy analyses to ensure reproducibility (Pharmapace, Inc., San Diego CA and Pentara Corporation, Salt Lake City UT). Both groups started with the same raw datasets and agreed on conventions for visit windows and pooling of study sites with low subject numbers. The mixed effects repeated measures model (MMRM) was the analytical method specified in the statistical analysis plan (SAP) of the NA study. This SAP was developed by JSS Medical Research Inc., and had undergone Food and Drug Administration review prior to database lock. This MMRM included the following terms:  treatment, visit, treatment by visit interaction, study site, and baseline value of the primary endpoint to be analyzed. The analyses were performed using this “original” MMRM.  Since it is now known that baseline disease severity may influence drug effect, current AD studies include baseline disease severity in the MMRM model. Therefore, an additional analysis was performed that included baseline MMSE in the model (“updated MMRM”). In addition, due to smaller sample sizes in the APOE4 subgroups, study sites were pooled for sites recruiting few patients to avoid having only one subject at a site in a dose arm from an APOE4 subgroup. Both MMRM methods yielded similar results and conclusions.
The NA study SAP included the following factors as covariates to be tested for interaction with drug effect: age, gender, use of symptomatic drugs, and APOE4 genotype; and required that efficacy be summarized by APOE genotype.  
The EU study was prematurely discontinued when the NA study results became available and the program was terminated. Most patients in the EU study had completed week 52 when the NA study results became public, while approximately 30% had completed week 65 and week 78. Therefore, the last two visits had a small sample size, and the evaluations were possibly biased since they were also termination visits after public release of the data (temporal bias). The MMRM method that includes efficacy from all visits may not be appropriate for the EU study, due to the introduction of potentially biased data from the later visits. EU efficacy results are presented using summary descriptive statistics based on observed cases (OC analysis) without imputations for missing data.
Efficacy analyses are shown for subjects in the intent to treat (ITT) population who had a baseline assessment and at least one post-baseline assessment, received at least one dose of study drug, and had a known APOE genotype. The pooled safety population included all subjects who received at least one dose of study drug regardless of APOE genotype status.
In the NA study, sensitivity analyses were performed using the completer set (subjects who completed week 78) and summary statistics using OC analysis. Exploratory analyses were performed that evaluated efficacy on ADAS-cog, CDR-SB, and DAD for 3 MMSE categories (16-26; 20-26; 22-26), using the same MMRM method as the primary analysis. Since these APOE4 subgroup analyses are used for hypothesis generation to inform future studies, the p-values presented were not adjusted for multiple statistical comparisons.

Specification of Model Variables (MMRM)

Response Variables

CBL(ADAS-cog)Visit i = (ADAS-cog)Visit i – (ADAS-cog)Baseline
CBL(CDR-SB)Visit i = (CDR-SB)Visit i – (CDR-SB)Baseline

Independent Variables

The treatment group was regarded as a class variable with three levels:  (1) tramiprosate 150 mg BID, (2) tramiprosate 100 mg BID, and (3) placebo. The visit variable was included as a class variable with values standing for the visit numbers (visit 5 to visit 10, corresponding to weeks 13 to 78), and was used to index the within-subject observations over time.  Study site (SITE) was entered in the model as a categorical, non-ordinal variable.

Covariates

Baseline value (BASE) of the dependent variable was entered as a continuous covariate in order to account for the fact that both absolute and percent changes over time in the outcome measure correlate with its value at the baseline visit. Baseline MMSE score was entered as a continuous covariate. The autoregressive order 1 (AR1) covariance structure was used to model the within-subject correlation across the different visits.

 

Results

Demographics and Baseline Characteristics

The NA and EU studies enrolled 1,052 and 973 subjects, respectively. The ITT population for  these analyses included 973 and 870 subjects from the NA and EU studies with known APOE genotypes. Patient demographics and baseline values in the APOE4 subgroups (NA Study) are shown in Table 1.

 

Table 1. Patient Demographic and Baseline Clinical Characteristics in APOE4 Subgroups (NA Study, All Enrolled)

Higher values for ADAS-cog, CDR-SB and NPI indicate greater severity, while lower values in DAD and MMSE indicate greater severity. APOE4 = apolipoprotein E4; AChEIs =  acetylcholinesterase inhibitors; MMSE = Mini Mental State Examination; ADAS-Cog =  Alzheimer’s Disease Assessment Scale-cognitive subscale; CDR-SB = Clinical Dementia Rating Sum of Boxes; DAD = Disability Assessment for Dementia; NPI = Neuropsychiatric Inventory.

 

In the NA study, baseline demographics of the three APOE4 groups were balanced except for the APOE4/4 group having a younger mean age and highest proportion of patients under age 75 years. The younger age in APOE4/4 patients is consistent with established epidemiologic observations and attributed to earlier amyloid accumulation and onset of symptoms in this population (10, 22).
The APOE4/4 group also had a higher proportion of patients in the moderate AD category with 39% versus 34% in the non-carriers subgroup. In the APOE4/4 sub-group the mean baseline ADAS-cog score is slightly higher than in non-carriers, while the mean CDR-SB score is similar.
In the APOE4/4 subgroup, which showed an efficacy signal, baseline ADAS-cog and CDR-SB scores in the 3 treatment arms are shown (Table 2). Mean ADAS-cog is higher and CDR-SB is lower in the low dose arm than in the other two arms. The MMRM model adjusts for baseline severity by including the baseline score of the outcome measure as a covariate in the analysis.

 

Table 2. Baseline ADAS-Cog and CDR-SB Scores in the APOE4/4 Homozygous Subgroup (NA Study, ITT Population)

Higher values for ADAS-cog and CDR-SB indicate greater severity. ADAS-Cog = Alzheimer’s Disease Assessment Scale-cognitive subscale; CDR-SB = Clinical Dementia Rating Sum of Boxes

 

Efficacy

Co-Primary Outcomes in the APOE4 Subgroups

The primary analyses of the 11-item ADAS-cog and CDR-SB are shown for each APOE4 subgroup in the ITT population with the least squares mean differences between each dose and placebo shown at week 65 and week 78 (the primary endpoint). The results of the updated MMRM model are shown in Table 3. The p-values presented were not adjusted for multiple statistical comparisons and are thus considered nominal p-values.

Table 3. Effects of Tramiprosate on ADAS-Cog and CDR-SB in APOE4 Subgroups (NA Study, ITT Population, N=973)

Note: Change from Baseline (CBL), is the LS mean differences in ADAS, CDR-SB between active arm and placebo. Negative values for LS mean differences in CBL of ADAS and CDR-SB indicate clinical benefit in favor of drug. Bolded values indicate clinical benefit with either nominal statistical significance (p < 0.05) or positive trends (p values between 0.05 and 0.1).

 

In APOE4/4 homozygotes in the high dose arm, a significant drug-placebo difference was observed on ADAS-cog at the last visits. The CDR-SB shows numerical benefit at week 78 and a positive trend at week 65 (p = 0.063). In the homozygous group in the low dose arm, there is numerical benefit on ADAS-cog and non-significant effects on CDR-SB at weeks 65 and 78. The APOE4 heterozygous group shows non-significant drug effects on ADAS-cog at both doses, but shows significant benefit on CDR-SB at the low dose, and a positive trend at the high dose. The non-carrier group shows non-significant effects on both outcomes at the low dose. In the non-carrier group placebo is significantly better than 150 mg tramiprosate, the high dose. The APOE4/4 homozygous group on high dose tramiprosate, therefore, shows the most consistent benefit on co-primary outcomes. Figure 1A and 1B shows the efficacy results in the APOE4 subgroups on the high dose at the last  visits.

 

Figure 1. Effects of Tramiprosate 150 mg BID in APOE4 Subgroups on Co-Primary Outcomes, ADAS-cog (Panel A) and CDR-SB (Panel B). Change from Baseline (CBL): LS mean differences between active arm and placebo are shown at weeks 52, 65 and 78 (ITT, MMRM Analysis), error bars are standard error of mean. Negative CBL indicates clinical benefit for both outcomes

P values: *p < 0.05, **p < 0.01 (indicate nominal statistical significance) and #p < 0.1 (indicates positive trend).
 

Efficacy Analyses in APOE4/4 Homozygous Patients: Time Course of Effect in NA Study

The time course of treatment effects in the APOE4/4 subgroup is shown in Figure 2 (panels A-B). At the high dose, the ADAS-cog shows a progressive separation from placebo starting at week 26 and reaching the largest drug-placebo difference at week 65; this treatment effect remains significant at weeks 52-78. The CDR-SB shows a similar time course, with the drug-placebo difference showing a positive trend at week 26 and again at week 65. At the low dose, the ADAS-cog shows a biphasic effect with early significant benefit at week 26 and numerical benefit at weeks 65 and 78, that is smaller than the high dose effect. The CDR-SB at the low dose shows minimal separation from placebo.

Sensitivity Analyses of Co-Primary Outcomes in APOE4/4 Homozygous Patients (NA Study)

Efficacy analysis in the completers showed similar results to the ITT analysis (data not shown). Descriptive statistics in the OC analysis also showed similar results to the MMRM analysis, as shown in Figure 2 (panels C-D).

 

Figure 2. Time Course of Tramiprosate Effects on co-primary outcomes in APOE4/4 Population at 100 mg and 150 mg BID in NA (Panels A-D) and EU studies (Panels E, F). NA study panels A and B show MMRM analysis, C and D show summary statistics with observed cases (OC). EU study panels E and F show summary statistics with observed cases. (ITT population, NA study: N= 147; EU study: N= 110). For both ADAS-cog and CDR-SB, positive CBL (downward graph) indicates worsening.

P values: *p < 0.05 (indicate nominal statistical significance) and # p < 0.1 (indicates positive trend); The hashes in panels E and F, indicate that the study was terminated at the time when most patients had completed week 52 visit, but only 30% had completed week 65 and week 78 visits.

Primary and Secondary Efficacy Outcomes in APOE4 Homozygous Patients and Percent Benefit Compared to Placebo (NA Study)

To further understand the efficacy in the APOE4/4 homozygous subgroup, effects of the high dose on the co-primary and key secondary outcomes (DAD, NPI, MMSE) were analyzed by determining the percent benefit compared to placebo (Table 4). The secondary outcomes were assessed at weeks 25, 52 and 78, and are presented at the last 3 visits, together with ADAS and CDR. The percent benefit compared to placebo is calculated as:  [CBL drug – CBL placebo]/CBL placebo. Note that percent benefit is indicated as positive if the result favors drug (i.e. negative LS mean differences for ADAS, CDR, and NPI; and positive LS mean differences for DAD and MMSE).

 

Table 4. Effect of Tramiprosate 150 mg BID in APOE4/4 homozygous group on primary and secondary outcomes and % drug benefit compared to placebo (NA study, ITT)

* % Drug benefit is indicated as positive when drug is better than placebo. For ADAS, CDR, and NPI negative LS means differences favor active drug, for DAD and MMSE positive differences favor active drug. Bolded values indicate clinical benefit with either nominal statistical significance (p < 0.05) or positive trends: p values between 0.05 and 0.1

 

The DAD showed consistent numerical benefit at weeks 26, 52, and 78. The NPI showed numerical benefit at endpoint, and a positive trend at week 52 (p <0.05). The MMSE showed small effects that were not significant at these same visits.

Efficacy Analysis in APOE4/4 Homozygous Patients: Time Course of Effect in EU Study

The baseline demographics of the three APOE4 groups were balanced except for the APOE4/4 group having a younger mean age, and highest proportion of patients under age 75 years (58% under age 75 years in non-carriers, 64% in heterozygous, and 77% in APOE4/4 groups). The APOE4/4 group also had the highest proportion of patients in the mild AD category (68% mild AD in non-carriers, 64% in heterozygous, and 74% in APOE4/4 groups).
Summary statistics with observed cases are shown for both outcomes in Figure 2 (panels E-F). In the OC analysis the ADAS-cog effect at the low dose shows numerical benefit at all visits up to week 52. At the high dose results suggest benefit over placebo at week 52, but not at earlier time points that showed positive trends in the NA study. The effect at week 52 is similar in magnitude to the NA study effect (latter is 3.0 points at week 52). Patients in the EU study were only allowed cholinesterase inhibitors, while the NA study also allowed memantine use. This difference in background AD medications may have contributed to the differences in ADAS-cog results between the two studies.
For the CDR-SB OC analysis, the low dose shows no benefit for visits up to week 52, similar to the NA study. The high dose shows consistent numerical benefit over placebo, and reaches maximum benefit at week 52, where the effect of 1.1 points is of similar magnitude to the NA study (latter is 0.7-1.1 at weeks 52 and 65).
The EU efficacy data are limited by the early termination of the study and the potential bias introduced at the early termination visits at weeks 65 and 78, but show numerical trends on CDR-SB that are of similar magnitude to the NA study.

Sensitivity Analysis for APOE4/4 Subgroup by Baseline MMSE

Analysis of efficacy based on baseline MMSE group was performed to evaluate for differences in drug response between mild and moderate patients. MMRM analyses were performed on the following MMSE categories: overall mild and moderate population (MMSE 16-26, inclusive); mild population (20-26, inclusive); and the milder population (22-26, inclusive). The data are shown in Figure 3 for ADAS-cog, CDR-SB, and DAD.

 

Figure 3. Effects of Tramiprosate 150 mg BID in APOE4/4 subgroup: Sensitivity analysis of effects on ADAS-cog and CDR-SB (Panels A, B) at weeks 52, 65 and 78. Effects on DAD are shown at weeks 26, 52 and 78 (Panel C). Sensitivity based on MMSE categories with increasing proportion of Mild Subjects (16-26, 20-26, and 22-26, NA Study). For ADAS-cog and CDR-SB outcomes, negative values in LS mean differences between drug and placebo indicate drug benefit; for DAD, positive values indicate drug benefit. Error bars are standard error of mean. Bolded (nominal) p-values are at week 78 endpoint.

These analyses suggest that APOE4/4 patients with mild AD (MMSE ≥20) may show greater benefit on the high dose of tramiprosate than those with lower MMSE at baseline. Additionally, patients with MMSE ≥22 (milder patients) showed the highest efficacy with a sustained cognitive improvement and less disability compared to placebo over the 78 weeks of the study.

Safety Analyses

Safety in the Combined NA and EU Studies

The nature of adverse events was similar between the two studies and between the three APOE4 subgroups. Therefore, the safety datasets were analyzed for the combined NA and EU studies (N= 1,052 and 973, respectively), and included a total of 2,025 unique AD patients who received at least 1 dose of study drug. The combined safety population (N= 2,025) included drug exposures up to 78 weeks.
In this safety population, the overall incidence of treatment emergent adverse events (TEAE) was slightly higher in the active dose arms than placebo (86-88% versus 81-82% in the APOE4 non-carriers and heterozygotes, respectively). In the APOE4/4 homozygous group the incidence was similar across the three dose arms, approximately 90%.  The most common TEAEs in the overall safety population are shown in Table 5. The most common TEAEs with incidence more than double placebo rates were: nausea, vomiting, and decreased weight; and the majority were either mild or moderate in severity. The other TEAEs were similar in incidence between active arms and placebo.

 

Table 5. Treatment emergent adverse events (TEAE) with incidence > 5% in any active dose arm in the safety population (combined safety population, all genotypes)

The overall rates of discontinuation due to AE in the APOE4 subgroups in the NA study were: 9% in homozygotes, 12% in heterozygotes and 13% in APOE4 non-carriers. In the EU study, the rates were: 14% in homozygotes, 9% in heterozygotes and 15% in APOE4 non-carriers.
In the overall safety population, the incidence of SAEs was similar across the three dose arms. In the APOE4/4 subgroup the incidence of SAEs was lower in the active dose arms than placebo. The most common SAEs that were double the placebo rate were syncope and pneumonia. The other SAEs are commonly reported in AD studies in the elderly population. Across both studies, there were a total of 14 deaths that were divided equally among the treatments arms: 5 in placebo, 5 in 100 mg BID, and 4 in 150 mg BID. The causes of death were typical of the elderly population in AD trials.

Central MRI Safety Assessments for ARIA

This MRI analysis included a total of 409 patients with scans at screening and week 78, or early termination, and an additional 17 subjects with MRI after study treatment only. MRI evaluations for ARIA revealed no cases of ARIA-E on active drug, and only one placebo APOE4/4 subject had possible ARIA-E. The distribution of subjects in each treatment arm by APOE4 genotype is presented in Table 6.

 

Table 6. MRI Safety Assessment for Vasogenic Edema: Distribution of Subjects across Treatment Arms and APOE4 Subgroups. In brackets are subjects who had only post-treatment MRI. * One subject with suspected ARIA-E.

 


Discussion

In AD patients clinical trials with amyloid-targeting agents regularly include APOE4 genotyping since APOE4 status influences age of onset and rate of decline, as well as susceptibility to drug-induced vasogenic edema and microhemorrhage. To our knowledge this is the first published analysis of drug efficacy in a clinical trial suggesting differential efficacy based on the number of APOE4 alleles. The findings suggest a gene dose-related benefit of tramiprosate with the largest drug-placebo difference in APOE4/4 homozygotes, intermediate benefit in APOE4 heterozygotes, and no benefit in APOE4 non-carriers.
The “APOE4 gene-dose effect” follows the well-established prevalence of β-amyloid positivity among the APOE4 homozygotes, heterozygotes and non-carrier subgroups in similar clinical trial populations, where study inclusion was based on a clinical diagnosis of AD. In the two solanezumab EXPEDITION trials in mild to moderate AD, the rate of positive amyloid PET (florbetapir) scans was 98% in homozygous group, 88% in heterozygous group, and 64% in non-carriers (11). In the two bapineuzumab studies (23), the rate of positive PET-PIB scans in non-carriers was similarly low at 62%. Since β-amyloid imaging was not available at the time of initiation of tramiprosate studies, the APOE4 genotype serves as a proxy for amyloid burden as well as accuracy of clinical AD diagnosis, which is very high in APOE4/4 homozygotes (>90%). Since tramiprosate is an amyloid-targeting agent, evaluating a population that is naturally enriched for amyloid pathology is critical to demonstrate drug efficacy. Another possible explanation for the gene-dose effect is that tramiprosate may have a direct effect on the APOE4 protein, a mechanism that is being investigated further in nonclinical studies. The latter potential mechanism may explain the difference in efficacy between the APOE4 homozygous and heterozygous subgroups.  
The results of tramiprosate in the APOE4/4 homozygous subgroup suggest dose-dependent efficacy. Tramiprosate at the high 150 mg BID dose showed a beneficial cognitive effect that was nominally statistically significant and was supported by positive trends on functional outcomes. The low dose showed a lower numerical benefit on cognition but not on functional outcomes.
Drug effects in AD trials are considered clinically meaningful if they provide at least 25% benefit over placebo. The cognitive effect of the 150 mg BID dose in APOE4/4 AD patients corresponds to 40% benefit over placebo at week 78, and is thus clinically meaningful. This cognitive effect is also supported by positive trends on global function, which corresponds to 25% benefit on CDR-SB at the week 78 endpoint, as well as a numerically consistent effect on disability with 25% benefit on DAD. These effects were in patients who were already receiving current standard of care with either one or two AD symptomatic drugs, and thus have no other treatment options available at present. The additional tramiprosate efficacy could then provide a meaningful benefit in slowing the rate of decline and managing disease.
Based on the MMSE sensitivity analyses, tramiprosate efficacy appears to be greater and more sustained in mild AD patients (MMSE 20 and higher), as previously reported in several anti-amyloid antibody trials (24, 25). Clinical improvement appears to be largest in the mildest AD patients with baseline MMSE of 22 and higher, as reported in studies with other amyloid-targeting agents, scyllo-inositol and crenezumab (26, 27). In the mildest AD group ADAS-cog effects increase progressively from early visits to week 78, a pattern supportive of a potential disease modifying effect.
The above finding suggesting better efficacy on both cognition and function in mild AD patients is consistent with the mechanism of tramiprosate observed in preclinical studies. Tramiprosate inhibits aggregation of amyloid monomers into soluble oligomeric species that cause synaptic toxicity. Mild patients who have larger synaptic numbers and more synaptic integrity are therefore more likely to exhibit clinical benefit from an earlier protective effect of tramiprosate on synapses. The moderate AD subgroup appeared to show lesser effects on function but still showed potentially meaningful cognitive benefit.
The main limitation of the above efficacy analyses is that they are based on subgroup analyses from the NA study that did not achieve its primary objectives in the overall study population. The sample size of the APOE4/4 subgroup was also limited. Randomization was not stratified based on genotype and the genetically defined subgroups were not precisely comparable.  Amyloid imaging was not available to determine the amyloid status of patients in the genetically-defined subgroups. Therefore, these findings require confirmation in prospectively defined studies in the APOE4/4 homozygous population.
From the safety perspective, tramiprosate showed a favorable safety profile based on placebo-controlled data in 2,025 patients in the combined Phase 3 studies. There was no evidence of dose-limiting vasogenic edema in the MRI subgroup. This could be an important advantage of tramiprosate, especially in APOE4 carriers who are at the highest risk of vasogenic edema with some anti-amyloid candidate agents (9, 28).
The most common adverse events were gastrointestinal: nausea, vomiting and weight loss, which were mostly mild or moderate in severity. This gastrointestinal tolerability issue is being addressed with the development of ALZ-801, a new tramiprosate formulation which is a valine prodrug of the active agent tramiprosate. ALZ-801 has shown substantially improved gastrointestinal tolerability as well as more consistent plasma levels of tramiprosate in Phase 1b studies in over 170 healthy elderly subjects (29). ALZ-801 at a dose that provides bioequivalent exposures to tramiprosate 150 mg BID will be used in future confirmatory studies focusing on the symptomatic APOE4/4 AD population.

 

Conclusions

APOE4/4 homozygotes constitute approximately 10-15% of all AD patients in memory clinics. Development of disease modifying drugs is challenging for this population due to their high burden of cortical and vascular amyloid pathology as well as their susceptibility to development of vasogenic edema and hemorrhage. Oral tramiprosate shows promising efficacy in symptomatic homozygous APOE4/4 AD patients at a dose that has a favorable safety profile with no observed ARIA-E. Future studies will focus on the homozygous APOE4/4 population with symptomatic AD. The use of the new prodrug ALZ-801, which provides more consistent plasma exposures and improved gastrointestinal tolerability, may further enhance efficacy and tolerability in these patients. Prospective confirmation of these promising efficacy findings in the homozygous APOE4/4 population with symptomatic AD, together with a favorable safety profile and the convenience of an oral formulation, could provide a major therapeutic advance for this genetically-defined population with a large unmet medical need.

 

Acknowledgements: We thank the investigators, their staff, and the individuals with AD who participated in the trials and their caregivers. We thank Dr. Sandra Saouaf and Christine Rathbun for editorial assistance.
Funding: Alzheon Inc. funded the above analyses and manuscript preparation. BELLUS Health (previously Neurochem), the original sponsor, had funded the tramiprosate Phase 3 studies.

Ethical standards: The study protocols were reviewed by ethics committees and complied with ICH standards for clinical trials and with all local requirements.

References

1.    Roses AD. Apolipoprotein E alleles as risk factors in Alzheimer’s disease. Annu Rev Med. 1996;47:387-400.
2.    Holtzman DM, Herz J, Bu G. Apolipoprotein E and apolipoprotein E receptors: normal biology and roles in Alzheimer disease. Cold Spring Harb Perspect Med. 2012;2:a006312.
3.    Liu Y, Yu JT, Wang HF, et al. APOE genotype and neuroimaging markers of Alzheimer’s disease: systematic review and meta-analysis. J Neurol Neurosurg Psychiatry. 2015;86:127-134.
4.    Breitner JC, Wyse BW, Anthony JC, et al. APOE-epsilon4 count predicts age when prevalence of AD increases, then declines: the Cache County Study. Neurology. 1999;53:321-331.
5.    Raber J, Huang Y, Ashford JW. ApoE genotype accounts for the vast majority of AD risk and AD pathology. Neurobiol Aging. 2004;25:641-650.
6.    Petersen RC, Thomas RG, Grundman M, et al. Vitamin E and donepezil for the treatment of mild cognitive impairment. N Engl J Med. 2005;352:2379-2388.
7.    Lautner R, Palmqvist S, Mattsson N, et al. Apolipoprotein E genotype and the diagnostic accuracy of cerebrospinal fluid biomarkers for Alzheimer disease. JAMA Psychiatry. 2014;71:1183-1191.
8.    Chalmers K, Wilcock GK, Love S. APOE epsilon 4 influences the pathological phenotype of Alzheimer’s disease by favouring cerebrovascular over parenchymal accumulation of A beta protein. Neuropathol Appl Neurobiol. 2003;29:231-238.
9.    Sperling RA, Jack CR, Jr., Black SE, et al. Amyloid-related imaging abnormalities in amyloid-modifying therapeutic trials: recommendations from the Alzheimer’s Association Research Roundtable Workgroup. Alzheimers Dement. 2011;7:367-385.
10.    Jansen WJ, Ossenkoppele R, Knol DL, et al. Prevalence of cerebral amyloid pathology in persons without dementia: a meta-analysis. JAMA. 2015;313:1924-1938.
11.    Degenhardt EK, Witte MM, Case MG, et al. Florbetapir F18 PET Amyloid Neuroimaging and Characteristics in Patients With Mild and Moderate Alzheimer Dementia. Psychosomatics. 2016;57:208-216.
12.    Gervais F, Paquette J, Morissette C, et al. Targeting soluble Abeta peptide with Tramiprosate for the treatment of brain amyloidosis. Neurobiol Aging. 2007;28:537-547.
13.    Martineau E, de Guzman JM, Rodionova L, Kong X, Mayer PM, Aman AM. Investigation of the noncovalent interactions between anti-amyloid agents and amyloid beta peptides by ESI-MS. J Am Soc Mass Spectrom. 2010;21:1506-1514.
14.    Aisen PS, Saumier D, Briand R, et al. A Phase II study targeting amyloid-beta with 3APS in mild-to-moderate Alzheimer disease. Neurology. 2006;67:1757-1763.
15.    Aisen PS, Gauthier S, Ferris SH, et al. Tramiprosate in mild-to-moderate Alzheimer’s disease – a randomized, double-blind, placebo-controlled, multi-centre study (the Alphase Study). Arch Med Sci. 2011;7:102-111.
16.    Mohs RC, Cohen L. Alzheimer’s Disease Assessment Scale (ADAS). Psychopharmacol Bull. 1988;24:627-628.
17.    Morris JC. The Clinical Dementia Rating (CDR): current version and scoring rules. Neurology. 1993;43:2412-2414.
18.    Gelinas I, Gauthier L, McIntyre M, Gauthier S. Development of a functional measure for persons with Alzheimer’s disease: the disability assessment for dementia. Am J Occup Ther. 1999;53:471-481.
19.    Folstein MF, Folstein SE, McHugh PR. «Mini-mental state». A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12:189-198.
20.    Cummings JL. The Neuropsychiatric Inventory: assessing psychopathology in dementia patients. Neurology. 1997;48:S10-16.
21.    Gauthier S, Aisen PS, Ferris SH, et al. Effect of tramiprosate in patients with mild-to-moderate Alzheimer’s disease: exploratory analyses of the MRI sub-group of the Alphase study. J Nutr Health Aging. 2009;13:550-557.
22.    Blacker D, Haines JL, Rodes L, et al. ApoE-4 and age at onset of Alzheimer’s disease: the NIMH genetics initiative. Neurology. 1997;48:139-147.
23.    Salloway S, Sperling R, Fox NC, et al. Two phase 3 trials of bapineuzumab in mild-to-moderate Alzheimer’s disease. N Engl J Med. 2014;370:322-333.
24.    Doody RS, Farlow M, Aisen PS. Phase 3 trials of solanezumab and bapineuzumab for Alzheimer’s disease. N Engl J Med. 2014;370:1460.
25.    Lasser R, Ostrowitzki S, Scheltens P, et al. Efficacy and safety of gantenerumab in prodromal Alzheimer’s disease: Results from scarlet road—a global, multicenter trial. Alzheimer’s & Dementia: The Journal of the Alzheimer’s Association. 2015;11:P331-P332.
26.    Cummings J, Cho W, Ward M, et al. A randomized, double-blind, placebo-controlled phase 2 study to evaluate the efficacy and safety of crenezumab in patients with mild to moderate Alzheimer’s disease. Alzheimer’s & Dementia: The Journal of the Alzheimer’s Association. 2014;10:P275.
27.    Salloway S, Sperling R, Keren R, et al. A phase 2 randomized trial of ELND005, scyllo-inositol, in mild to moderate Alzheimer disease. Neurology. 2011;77:1253-1262.
28.    Sevigny J, Chiao P, Bussiere T, et al. The antibody aducanumab reduces Abeta plaques in Alzheimer’s disease. Nature. 2016;537:50-56.
29.    Hey JA, Abushakra S, Power A, Yu JY, Versavel  M, Tolar M. Phase 1 development of ALZ-801, a novel beta amyloid anti-aggregation prodrug of tramiprosate with improved drug properties, supporting bridging to the phase 3 program. Alzheimer’s & Dementia: The Journal of the Alzheimer’s Association. 2016;in press.

PROGRESS IN TREATMENT DEVELOPMENT FOR NEUROPSYCHIATRIC SYMPTOMS IN ALZHEIMER’S DISEASE: FOCUS ON AGITATION AND AGGRESSION. A REPORT FROM THE EU/US/CTAD TASK FORCE

M. Soto1, S. Abushakra2, J. Cummings3, J. Siffert4, P. Robert5, B. Vellas1, C.G. Lyketsos6 and Task Force Members*

1. Gerontopole, INSERM U1027, Alzheimer’s Disease Research and Clinical Center, Toulouse University Hospital, Toulouse, France; 2. Transition Therapeutics, San Matteo, California, USA; 3. Cleveland Clinic Lou Ruvo Center for Brain Health, Las Vegas, USA; 4. Avanir Pharmaceuticals, Inc. Aliso Viejo, USA; 5. EA CoBTeK/ICMRR University of Nice Sophia Antipolis – CHU, France; 6. Department of Psychiatry, The Johns Hopkins Bayview Medical Center, Baltimore, MD, USA. 

Corresponding Author: M.E. Soto-Martin, Gerontopole, INSERM U1027, Alzheimer’s Disease Research and Clinical Center, Toulouse University Hospital, Toulouse, France, soto-martin.me@chu-toulouse.fr

*Task Force Members:  Susan Abushakra (San Matteo), Sandrine Andrieu (Toulouse), Joanne Bell (Cambridge), Gene Bowman (Lausanne), Sasha Bozeat (Utrech), Robert Brashear (San Francisco), Marc Cantillon , Maria Carrillo (Chicago), Jesse Cedarbaum (Cambrdige), Er Chen (San Francisco), Isabelle Clavier (Chilly Mazarin), Caroline Cohen (Chilly Mazarin), Eskild Colding-Jorgensen (Valby), Csilla Csoboth (San Francisco), Jeffrey Cummings (Las Vegas), Rachelle Doody (Houston), Bruno Dubois (Paris), Jane Durga(Vevey), Michael Egan (North Wales), Laura Eggermont (Utrech), Laura Gault (Chicago), Serge Gauthier (Verdun), Bram Goorden (Vevey), Mark Gordon (Ingelheim), Michael Grundman (San Diego), Harald Hampel (Paris), Paul Hartung (Acton), Roza Hayduk (San Diego), Suzanne Hendrix (USA), Robert Hoerr (Karlsruhe), Michael Keeley (San Francisco), Ara Khachaturian (Potomac), Zaven S. Khachaturian (Potomac), Robert Lasser (Basel), John Lawson (Malvern), Valérie Legrand (Nanterre), Constantine Lyketsos (Baltimore), Richard Meibach (East Hanover), Annette Merdes (Munich), Mark Mintun (Philadelphia), Hans Moebius (Brunnen), Cristina Murat (Marly le Roi), Philip Nichols (Lausanne), Pierre Jean Ousset (Toulouse), Jana Podhorna (Ingelheim), Maria Pueyo (Suresnes), Christopher Randolph (Hamilton), David  Raunig (New Hope), Vanessa S. Reddy (Basel), Philippe Robert (Nice), Gary Romano (Titusville), Allen Roses (Chapel Hill), Juha Rouru (Turku), Ivana Rubino (Basel), Michael Ryan (East Hanover), Stephen Salloway (Providence), Philip Scheltens (Amsterdam), Rachel J.  Schindler (New York), Achim Schneeberger (Vienna), Lon Schneider (Los Angeles), Jeffrey Sevigny (Cambridge), Klaudius Siegfried (Langen), Eric Siemers (Indianapolis), João Siffert (Aliso Viejo), Chang-Heok Soh (Basel), Maria E. Soto (Toulouse), Johannes Streffer (Beerse), Joyce Suhy (Newark), Jacques Touchon (Montpellier), Gabriel Vargas (Thousand Oaks), Estelle Vester-Blokland (Basel), Michael Weiner (San Francisco), Glen Wunderlich (Ingelheim)

J Prev Alz Dis 2015;2(3):184-188
Published online June 24, 2015, http://dx.doi.org/10.14283/jpad.2015.77


Abstract

Background: The management of neuropsychiatric symptoms (NPS) such as agitation and aggression is a major priority in caring for people with Alzheimer’s disease (AD). Agitation and aggression (A/A) are among the most disruptive symptoms, and given their impact, they are increasingly an important target for development of effective treatments. Considerable progress has been made in the last years with a growing number of randomized controlled trials (RCTs) of drugs for NPS. The limited benefits reported in some RCTs may be accounted for by the absence of a biological link of the tested molecule to NPS and also by key methodological issues. In recent RCTs of A/A, a great heterogeneity design was found. Designing trials for dementia populations with NPS presents many challenges, including identification of appropriate participants for such trials, engagement and compliance of patients and caregivers in the trials and the choice of optimal outcome measures to demonstrate treatment effectiveness.  The EU/US -CTAD Task Force, an international collaboration of investigators from academia, industry, non-profit foundations, and regulatory agencies met in Philadelphia on November 19, 2014 to address some of these challenges. Despite potential heterogeneity in clinical manifestations and neurobiology, agitation and aggression seems to be accepted as an entity for drug development. The field appears to be reaching a consensus in using both agitation and aggression (or other NPS)-specific quantitative measures plus a global rating of change for agitation outcomes based on clinician judgment as the main outcomes.

Key words: Behavior, agitation, aggression, Alzheimer’s, measurement, therapeutics, clinical trial.  


 

Introduction 

The management of neuropsychiatric symptoms (NPS) such as agitation and aggression is a major priority in caring for people with Alzheimer’s disease (AD). NPS are frequent and associated with a number of adverse outcomes including accelerated transition from prodromal AD to AD dementia (1), and faster progression from early dementia to severe dementia or death (2). NPS have serious consequences for patients and caregivers such as greater disability, worse quality of life, earlier institutionalization, increased caregiver burden, and higher health care costs (3).    

Given their impact, NPS are increasingly an important target for development of effective treatments. However, the heterogeneity of NPS complicates treatment development; they are heterogeneous in both phenomenology and cause. NPS consist of distinct clinical syndromes, which are thought to have a common neurobiological basis. Therefore, pharmacological intervention must focus on specific syndromes. Agitation and aggression (A/A) are among the most disruptive symptoms, and are currently the focus of several drug development programs. There are limited non-pharmacological and pharmacological options available for the management of these symptoms. The most studied and widely used medication class has been atypical antipsychotics for the treatment of A/A and psychosis in AD. However, their efficacy is modest and use is associated with harmful adverse events and mortality (4-6). In North America there are no approved drugs for the treatment of NPS in AD. In the European Union, only risperidone is indicated for the short-term treatment of severe aggression in AD. As a result, most agents are used off-label due to the lack of other options (7). Thus, currently the management of clinically significant, persistent or recurrent dementia-related A/A unresponsive to non-pharmacologic intervention is a major challenge for clinicians and caregivers waiting for newer therapies.

Emerging evidence from neurobiological research about pathogenesis, such as links to monoaminergic system degeneration or specific neuronal circuit dysfunction, has led to the investigation of both repositioned and new therapeutics for NPS in AD, including A/A.  Considerable progress has been made in the last years with a growing number of randomized controlled trials (RCTs) of drugs for NPS. The majority of recent RCTs focused on A/A (8). The limited benefits reported in some RCTs may be accounted for by the absence of a biological link of the tested molecule to NPS and also by key methodological issues. In recent RCTs of A/A, primary endpoints were based on different behavior rating scales or subscales, with proxy-based scales being more common than direct clinical observation. A high response on placebo was observed in many trials. Moreover, variable definitions of “clinically significant A/A” were used, but most required at least moderate severity of A/A (8).

Designing trials for dementia populations with NPS presents many challenges, including identification of appropriate participants for such trials, engagement and compliance of patients and caregivers in the trials and the choice of optimal outcome measures to demonstrate treatment effectiveness.  The EU/US -CTAD Task Force, an international collaboration of investigators from academia, industry, non-profit foundations, and regulatory agencies met in Philadelphia on November 19, 2014 to address some of these challenges.

Definition of Agitation in Dementia

A range of NPS have been reported although they tend to aggregate into predictable groups such as A/A, depression, apathy, psychosis, and sleep disturbances (9). Recent treatment development has targeted presumptive or proposed syndromes in these areas (10, 11).

Until recently, there were no widely accepted diagnostic criteria for a syndrome of agitation associated with dementia, which had been a topic of discussion in past FDA meetings.  The lack of a consensus definition for A/A in trials has contributed to the lack of progress in the field. To address this gap, The International Psychogeriatric Association (IPA) formed an Agitation Definition Work Group (ADWG) to develop a consensus definition of agitation in patients with cognitive disorders that could be applied in different studies such as epidemiologic, pharmacologic, non-pharmacologic interventional, and neurobiological (12). This consensus proposed the following criteria for a provisional definition shown in table 1.

The key elements of this definition to be considered are its provisional intent awaiting validation studies; the subjective aspect (emotional distress) associated with observable behaviors; that agitation must be sustained for 2 weeks and represent a change from previous behavior; the presence of excess disability from agitation, which in the clinician’s opinion is beyond that due to the cognitive impairment; and exclusion of delirium or other aggravating medical or iatrogenic conditions.

However, this definition presents some areas of controversy that were discussed by members of the task force:

• Is it a syndrome or a complication of another syndrome, such as psychosis, or depression?

• Is the phenomenology of agitation specifically related to Alzheimer’s dementia? Or is agitation common to AD and other type of dementias?

• Are agitation and aggression the same?

• Is there more than one type of agitation?

It should be no surprise that NPS are universal in diseases that affect key brain areas regulating behavior, or that disrupt multiple brain areas over time. Therefore, A/A is most likely not specific to AD but occurs in other diseases with cognitive impairment.

Accordingly to recent IPA consensus, aggression is considered to be subsumed under the broad symptom cluster of agitation. Better understanding the phenotypes of agitation and identifying variables that may help differentiate sub phenotypes will be crucial for targeting specific treatments to sub phenotypes.  Consequently, there is a need for RCTs to target both the phenomenology of agitation as well as its underlying neurobiology. There are probably more than one type of agitation and more than one underlying neurobiological pathway. Currently, The Food and Drug Administration (FDA) is accepting agitation as a clinical target for treatment development, with several programs now entering Phase 3.

Table 1. Provisional IPA definition criteria of Agitation

Adapted from Cummings et al., 2015 (12); IPA: The International Psychogeriatric Association

Determining severity of NPS as inclusion criteria

Once agitation is defined: How is the severity of “clinically significant agitation/aggression” as inclusion criteria in RCTs best defined?  In recent and on-going trials two approaches have been used: 1) judgment of experienced clinicians that medication is deemed necessary and/or 2) severity rating above a cut-off indicative of moderate or more severe agitation on a scale (e.g. the Neuropsychiatric Inventory (NPI-A/A) (13) and global measures, such as clinical global impression of severity. The CitAD (citalopram for agitation in AD) trial combined both approaches (14). The recent AVP-923 phase II trial (study 12-AVR-131; NCT 01584440) required “agitation that interferes with daily routine and for which a prescription medication is deemed indicated, in the opinion of the investigator with a Clinical Global Impression of severity (CGI-S) score ≥ 4”. The HARMONY-AD trial (NCT01735630) and the brexpiprazole (NCT 01862640) trial, utilized the NPI-A/A ≥ 4 cutoff which indicates at least moderately severe A/A occurring at least weekly.  It will be interesting to compare the final patient characteristics once these trials will be completed. Despite that the eligibility criteria differ in these recent RCTs, a consensus definition for the minimum agitation severity for RCT inclusion seems to be emerging.

Outcomes Measures

Central to treatment development for NPS is clinical measurement. The choice of the efficacy outcome measure has varied across RCTs. In fact there has been no gold standard for assessing the response to treatment.  Direct assessments of behavior until now available (e.g., actigraphy, audio and video sensors) are mostly developed in research settings but are not fully compatible with multicenter RCT requirements. Recently, it seems that this situation is changing for the actigraphy. In most cases, measurement relies on reporting of observable behaviors and mental state by patients and caregivers/informants. This approach is affected by aspects of the cognitive disorder that may limit the patient’s ability to report their mental state, or may lead to forgetting prior experiences and behaviors. As a result, measurement depends on input from caregivers who themselves are “filters,” which may be biased by the caregiver’s emotional state. Further, because NPS are episodic, relapsing and remitting frequently, often “real-time” in response to environmental situations, the quantification of NPS frequency and severity over longer time frames may be difficult.

Two approaches have been used to assess treatment response in RCTs of NPS: 1) outcomes based on judgment of experienced clinicians such as the Clinical Global Impression of Change (CGIC); and/or 2) the use of outcomes measuring the severity of NPS over the treatment period such as the NPI. Validated scales for the measurement of NPS fall into two categories: narrow spectrum measures, for example of depression or agitation, and broad-spectrum measures (15), including the NPI that covers several NPS domains: 10 NPS domains (in its original version (13)) or 12 (in a later version (16)).

The NPI provides a comprehensive assessment of NPS in dementia, and has been widely used in epidemiological and treatment trials. It is familiar to most AD clinicians and simple to administer. The NPI is based on caregiver input obtained during a clinical interview and does not include clinician judgment as part of the assessment. Recently, the Neuropsychiatric Inventory Clinician (NPI-C) Rating was developed based on the NPI (17). The NPI-C further increased the number of assessed NPS (or domains) from 12 to 14, and included clinician assessment as the basis for scoring overall NPS severity. A trained clinician provides an overall rating based on both patient and caregiver interviews, direct observation, and additional chart or other information. The NPI-C adds additional details to the profile of some NPI domains behaviors, for example the agitation and aggression domains are assessed separately by 13 and 8 questions/items (total 21) versus 8 on the NPI-A/A domain. The NPI-C development has followed the LED standard (longitudinal, expert, all data), and has shown a better reliability and concurrent validity than the NPI in a validation study (17).  NPI-C is a versatile measure that can be used as a broad-spectrum or as a narrow spectrum measure of a particular domain, such as agitation or aggression. The main limitations of NPI-C are the longer administration time if the full Inventory is used, the requirement of expert clinician raters and the lack of data from longitudinal or interventional studies. Anticipated data from the 12-week HARMONY-AD trial population will allow describing NPS based on the NPI-C. Another widely used measure, specific to A/A, is the Cohen-Mansfield Agitation Inventory (CMAI), which was originally developed for use in institutionalized patients (18). It is also based on caregiver ratings and does not include a clinician assessment. The Neurobehavioral Rating Scale (NBRS) is a 28-item observer-clinician rater instrument derived from the Brief Psychiatric Rating Scale (19). The agitation sub-scale of the NBRS (NBRS-A) combines subscores of agitation, disinhibition/aggression and hostility/uncooperativeness. The NBRS-A was used as co-primary outcome in CitAD trial.

In order to complement NPS ratings based on caregiver report, clinical global ratings are used. Their strength is their being derived from experienced clinicians (20). Several versions allow global ratings in agitation specific domains (or other NPS) over time by study clinicians masked to treatment assignment. These scales have been used in recent and on-going trials, as key secondary or co-primary outcomes.

In summary, there are 3 types of outcome measures to assess NPS: 1) those based on structured caregiver interviews, 2) those based on structured clinician (global) ratings, and 3) those that utilize a combination of these approaches such as the NPI-C. Recent trials have used a combination of these measures as primary outcome measures. Recent positive data from two A/A completed trials (AVP-923 and CitAD) indicate that these scales are sensitive to drug effects.

Other Issues Discussed

Participation of caregivers 

A key question in this kind of RCTs is the role and participation of the caregiver. First, a standard definition of caregiver (family, formal/professional caregiver, how much time spent with the patient…) is needed. Second, caregivers are essential in efficacy assessment since they are the closest to the patient and observe frequency and severity of NPS closely, especially episodic and fluctuant NPS. However, as previously mentioned, a potential drawback is the inexperience of caregivers in performing efficacy assessments and their lack of objectivity—which could partly contribute to the high placebo effect observed in RCTs. Therefore, it is crucial to train caregivers (informal and professional) in identifying and rating NPS, since their information is precious.  Thus, innovative approaches that include training and support of caregivers (including psychosocial interventions) are needed to support trial engagement.

Study length 

Almost all RCTs for NPS were designed with a treatment duration of 9 to 12 weeks. However, for drugs that show efficacy, it will be important to evaluate the persistence of efficacy over a longer period. Time to relapse can be assessed in discontinuation phases of trials.

Allowed rescue medication to treat NPS

As a minimum, trials should allow for rescue medications for acute exacerbation of symptoms during the trial and to decrease patient dropout.  The requirement for rescue medications can serve as a proxy measure of the efficacy of the intervention.

Use of concomitant medications 

Some trials have allowed continuation of use of concomitant medications for agitation/aggression (AVP-923, 12-AVR-131 study) whereas others excluded patients taking other medications or required patients “wash out” their medications prior to study entry (CitAD). Like in other areas of CNS research (e.g. pain, depression), polymedication is not uncommon and therefore testing adjunctive therapy is important. Drug interactions (both pharmacokinetics and pharmacodynamics) can potentially confound assessment of efficacy and/or safety. Assessment of new treatments as monotherapy is important and can provide a clearer picture of specific drug effects, but monotherapy trials limit the pool of study participants in terms of symptom severity and use of medications and may limit generalizability of results with respect with prescription patterns in clinical practice.

Design Experience of Recent Completed Trials

A well designed and executed model trial evaluating the antidepressant citalopram, a selective serotonin reuptake inhibitor (SSRI), for agitation in patients with AD without major depression recently reported its findings (CitAD). The study evaluated the effect of up to 30 mg daily of citalopram on patient functioning, caregiver distress, and safety parameters, (14).  In CitAD a psychosocial intervention was used to ensure that patients and caregivers received appropriate « enhanced usual care » in both groups. This was expressly designed to be practical and easily standardized for a research setting (21). CitAD results suggested that citalopram led to a significant reduction on the agitation domain of the NBRS-A, with a meaningful clinically relevant response in 40% on citalopram (vs 26% on placebo) on the modified Alzheimer Disease Cooperative Study-Clinical Global Impression of Change scale (mADCS-CGIC).

More recently, a 10-week phase 2 clinical trial of AVP-923 for agitation in AD (study 12-AVR-131; NCT 01584440) reported efficacy for dextromethorphan combined with quinidine. Results were presented at the American Neurological Association Meeting on October 13, 2014 (22). The study enrolled 220 patients. The study employed a sequential parallel comparison design (SPCD) to reduce the impact of a potential placebo response on the ability to detect a treatment effect (23). The NPI A/A domain was the primary efficacy endpoint. AVP-923 was associated with a clinically meaningful and statistically significant improvement in agitation on the primary endpoint and also in key secondary endpoints (patient/caregiver and clinician global measures and NPI-4 domain agitation clusters). Analysis of clinical characteristics of study participants will provide insights to help plan future studies.

Conclusion

In summary, treatment development for NPS, including agitation and aggression, has accelerated in the last few years with the promise of more effective novel treatments on the immediate horizon. Despite potential heterogeneity in clinical manifestations and neurobiology, agitation and aggression seems to be accepted as an entity for drug development. The field appears to be reaching a consensus in using both agitation and aggression (or other NPS)-specific quantitative measures (such as relevant domains of NPI/NPI-C) plus a global rating of change for agitation outcomes based on clinician judgment as the main outcomes.

In parallel with the considerable efforts in crafting appropriate designs for RCTs of therapeutic agents for NPS, in particular A/A, EU-US Task Force members expressed the urgent need to gain more clarity regarding the underlying neurobiology and affected circuitry of NPS in AD. This better understanding of the neuropathogenesis may offer the opportunity to develop better targeted drug treatments, as well as biomarkers as intermediate outcomes or endpoints to measure treatment efficacy.

Conflict of interests: Dr. Soto has received grants from the National French Projet Hospitalier de Recherche Clinique (PHRC N° 13 7031 08), the European Commission (FP7-HEALTH-F3-2010-242153) and Ethypharm, and has served as a consultant/advisor to Ethypharm. Dr Lyketsos Grant has received support (research or CME) from the NIMH, NIA, Associated Jewish Federation of Baltimore, Weinberg Foundation, Forest, Glaxo-Smith-Kline, Eisai, Pfizer, Astra-Zeneca, Lilly, Ortho-McNeil, Bristol-Myers, Novartis, National Football League, Elan and Functional Neuromodulation. He has served as a consultant/advisor to Astra-Zeneca, Glaxo-Smith Kline, Eisai, Novartis, Forest, Supernus, Adlyfe, Takeda, Wyeth, Lundbeck, Merz, Lilly, Pfizer, Genentech, Elan, NFL Players Association, NFL Benefits Office, Avanir, Zinfandel, BMS, Abvie, Janssen, Orion, Otsuka, Servier, Astellas. He has received honorarium or travel support from Pfizer, Forest, Glaxo-Smith Kline and Health Monitor. Dr Abushakra is a full time employee and hold stocks and stock options of Transition Therapeutics, San Mateo, California, USA. Dr Joao Siffert is a full time employee of Avanir Pharmaceuticals, Inc., Aliso Viejo, CA, USA. The Gérontopôle (chair  Pr Vellas) has received grant support from the PHRC, ANR, European Comission as well as: Abbvie, Affiris, Avid, BMS,  Eisai,  Elan, Envivo, Exhonit, Genentech, GSK, Ipsen, Lilly, Lundbeck, Médivation, MSD, Nutricia, Otsuka, Pharnext, Pfizer, Pierre-Fabre, Régénéron, Roche, Sanofi, Servier, TauRx Therapeutics, Wyeth. Dr Vellas has served as consultant/advisor to Biogen, GSK,  Lilly, Lundbeck, Medivation, MSD, Nestlé, Nutricia,  Pfizer, Roche, Sanofi, Servier, TauRx Therapeutics, Novartis. Dr Robert has received grants from European commission (projet FP7 VERVe). He has received honorarium from Roche, Servier and Lilly. Dr. Cummings has provided consultation to Abbvie, Acadia, ADAMAS, Alzheon, Anavex, Avanir, Biogen-Idec, Biotie, Boehinger-Ingelheim, Chase, Eisai, Forum, Genentech, Grifols, Intracellular Therapies, Lilly, Lundbeck, Merck, Neurotrope, Novartis, Nutricia, Otsuka, QR Pharma, Resverlogix, Roche, Suven, Takeda, and Toyoma companies.  Dr. Cummings owns stock in ADAMAS, Prana, Sonexa, MedAvante, Neurotrax, and Neurokos.  Dr. Cummings owns the copyright of the Neuropsychiatric Inventory. The Task Force was partially funded by registration fees from industrial participants. These corporations placed no restrictions on this work.

References

1. Taragano FE, Allegri RF, Krupitzki H, Sarasola DR, Serrano CM, Loñ L, Lyketsos CG: Mild behavioral impairment and risk of dementia: a prospective cohort study of 358 patients. J Clin Psychiatry. 2009; 70:584–592.

2. Peters ME, Schwartz S, Han D, Rabins PV, Steinberg M, Tschanz JT, et al. Neuropsychiatric Symptoms as Predictors of Progression to Severe Alzheimer’s Dementia and Death: The Cache County Dementia Progression Study. Am J Psychiatry. 2015 ; 172:460-465.

3. Lyketsos CG, Carrillo MC, Ryan JM, Khachaturian AS, Trzepacz P, Amatniek J, Cedarbaum J, Brashear R, Miller DS: Neuropsychiatric symptoms in Alzheimer’s disease. Alzheimers Dement. 2011; 7:532–539.

4. Schneider, LS., Dagerman, K. and Insel, P. S. Efficacy and adverse effects of atypical antipsychotics for dementia: meta-analysis of randomized, placebo-controlled trials. Am J Geriatr Psychiatry. 2006; 14, 191–210.

5. Langballe EM, Engdahl B, Nordeng H, Ballard C, Aarsland D, Selbæk G. Short- and Long-Term Mortality Risk Associated with the Use of Antipsychotics Among 26,940 Dementia Outpatients: A Population-Based Study. Am J Geriatr Psychiatry. 2014 ; 4:321-323.

6. Maust DT, Kim HM, Seyfried LS, Chiang C, Kavanagh J, Schneider LS, Kales HC. Antipsychotics, Other Psychotropics, and the Risk of Death in Patients With Dementia: Number Needed to Harm. JAMA Psychiatry. 2015 ; 72:438-445

7. Maher AR, Maglione M, Bagley S, Suttorp M, Hu JH, Ewing B, et al. Efficacy and comparative effectiveness of atypical antipsychotic medications for off-label uses in adults: a systematic review and meta-analysis. JAMA. 2011; 306:1359-9. 

8. Soto M, Andrieu S, Nourhashemi F, Ousset PJ, Ballard C, Robert P, et al. Medication development for agitation and aggression in Alzheimer disease: review and discussion of recent randomized clinical trial design. Int Psychogeriatr. 2014; 16: 1-17.

9. Geda YE, Schneider LS, Gitlin LN, Miller DS, Smith GS, Bell J, et al.: Neuropsychiatric symptoms in Alzheimer’s disease: Past progress and anticipation of the future. Alzheimers Dement 2013; 9: 602-608.

10. Robert P, Onyike CU, Leentjens AFG, Dujardin K, Aalten P, Starkstein S, et al: Proposed diagnostic criteria for apathy in AlzheimerŁs disease and other neuropsychiatric disorders. European Psychiatry. 2009; 1–7.

11. Olin JT, Schneider LS, Katz IR, Meyers BS, Alexopoulos GS, Breitner JC, et al: Provisional diagnostic criteria for depression of Alzheimer disease. Am J Geriatr Psychiatry. 2002; 10:125–128.

12. Cummings J, Mintzer J, Brodaty H, Sano M, Banerjee S, Devanand DP, et al. Agitation in cognitive disorders: International Psychogeriatric Association provisional consensus clinical and research definition. Int Psychogeriatr. 2015; 27:7-17. 

13. Cummings JL, Mega M, Gray K, et al. The Neuropsychiatric Inventory: comprehensive assessment of psychopathology in dementia. Neurology. 1994; 44: 2308–2314.

14. Porsteinsson, A., Drye, LT, Pollock, BG., Devanand, DP, Frangakis, C, Ismail, Z, et al. Effect of citalopram on agitation in Alzheimer Disease; the CitAD randomized clinical trial, JAMA. 2014; 311:(7) 682-691.

15. Steinberg, M.; Lyketsos, CG. Neuropsychiatric measures for cognitive disorders. In: Rush, AJ.; First, MB.; Blacker, D., editors. Handbook of Psychiatric Measures. Arlington, VA: American Psychiatric Publishing; 2008. p. 864

16. Cummings JL.. The Neuropsychiatric Inventory: assessing psychopathology in dementia patients. Neurology. 1997; 48: S10–16.

17. de Medeiros, K, Gauthier, RP, Stella,F, Politis,A, Leoutsakos, J,  Taragano F, et al. The Neuropsychiatric Inventory-Clinician rating scale (NPI-C): reliability and validity of a revised assessment of neuropsychiatric symptoms in dementia, Int. Psychogeriatrics. 2010; 22: 984-994. 

18. Cohen-Mansfield J, Marx MS, Rosenthal AS. A description of agitation in a nursing home. J Gerontol. 1989;44(3):M77-84.

19. Levin, H.S., et al. The neurobehavioural rating scale: assessment of the behavioural sequelae of head injury by the clinician. Journal Neurology Neurosurgery Psychiatry 1987; 50,183–93.

20. Guy, W. ECDEU Assessment Manual for Psychopharmacology, revised. Bethesda, MD: US Department of Health, Education, and Welfare. 1976

21. Drye LT, Ismail, Z, Porsteinsson, AP, Rosenberg, PB, Weintraub, D, Marano, C, et al. Citalopram for agitation in Alzheimer’s disease: Design and Methods. Alzheimers Dement. 2012; 8: 121-130.

22. Cummings J, Lyketsos C, Tariot P, et al. Dextromethorphan/quinidine (AVP-923) Efficacy and Safety for Treatment of Agitation in Alzheimer’s disease: Results from a Phase 2 Study. In: [Poster] 2014 American Neurological Association (ANA) 139th Annual Meeting, October 10-12, 2014, Baltimore, MD. M1336WIP.

23. Baer L, Ivanova A. When should the sequential parallel comparison design be used in clinical trials? Clin Invest. 2013;3(9):823-33.