Lead Investigator: Wiesje van der Flier, Amsterdam UMC, VUmc
Title of Proposal Research: Optimizing Trial design to Achieve Personalized prevention of Alzheimer’s disease
Vivli Data Request: 6063
Funding Source: Funded by PPP Allowance made available by health-Holland, Top Sector Life Sciences & Health to stimulate public-private partnerships and Brain Research Center (grant no. LSHM19051).
Potential Conflicts of Interest: None
Summary of the Proposed Research:
With more than 40 million worldwide, Alzheimer disease (AD) is among the largest health care challenges of our century. However, curative therapy is not yet available. This may be due to a number of factors, that are slowly becoming clear as our understanding of the disease grows. First, AD develops gradually, in the course of decades. Studies using biomarkers (Amyloid or TAU; e.g. abnormal proteins linked to the development of AD and other diseases) and imaging (Magnetic resonance imaging (MRI) or positron emission tomography scan (PET)) have shown that brain changes associated with AD, are present until 20 years before clinical manifestation of the disease. The stage of dementia is too late to reverse the brain damage which has accumulated over the decades before. This novel knowledge implies that trials should focus on pre-dementia stage, and hence that future treatment strategies for AD will have the form of secondary prevention.
Second, AD is a complex, diverse disease. Most drugs tested, have focused on removing/reducing amyloid plaques. It could be that amyloid is simply the wrong target. While this notion cannot be excluded, literature strongly supports an important role for amyloid in onset and progression of the disease. Nonetheless, it is essential to select the right patients most likely to benefit from anti-amyloid therapy with the right mode of action. In addition, it is increasingly recognized that amyloid does not explain the disease in its entirety. Therefore, trials need to evaluate different mechanism-based approaches as well, e.g. anti-tau with active or passive immunization, anti-inflammatory drugs and neuroprotective compounds (targeted neurons designed to decrease their vulnerability to neurodegenerative mechanisms occurring in Alzheimer’s disease) and we should find out which patients benefit most from which strategy.
Finally, taking into account, one realizes that trial designs have been too crude; inclusion criteria do not reflect the mode of action of specific drugs and outcome measures lack sensitivity. To bring closer a future of personalized prevention of AD, we need to focus on early, pre-dementia disease stages, taking into account the diverse patient group.
Statistical Analysis Plan:
Regarding our statistical analysis approach, since we have multiple datasets, we think that the first approach – a two stage meta analyses by each data set – is most appropriate. Of course, we would also like the second approach (i.e. analysing the individual patient data including a covariate in the model to account for each study e.g. in a hierarchical regression model) but that is if we can merge all data. However, we do not know what the data looks like and if it is possible to merge. Also, we have other data sets to account for; therefore, we think approach one is for now the best option.
In two steps, we aim to identify a combination of patient characteristics (demographic, clinical, biomarker) associated with a positive response to treatment.
In the first step, we will define responders using different approaches:
1) early endpoints:
- a) biomarker improvement/stabilization;
- b) cognitive improvement/stabilization (primary cognitive outcome measure),
- c) cognitive improvement/stabilization and functional improvement/stabilization (primary cognitive & functional outcome measures),
2) late endpoints:
- a) change of cognitive status, i.e. clinical progression to Mild Cognitive Impairment (MCI);
- b) change of cognitive status, i.e. clinical progression to AD dementia
Different responder definitions (1abc and 2ab) will be used to dichotomize treatment response. In the second step, We merge data sets from different studies and analyze them jointly and we will model treatment effect, by applying a Causal Forest machine learning model. This predictive technique is a flexible and powerful predictor, in particular when higher order interactions are expected [Wager and Athey, 2018]. Using this model, we can model treatment-specific effects of medication. The following characteristics will be included in the model: age, sex, disease status, ethnicity, presence or absence of co-morbidities, baseline cognitive status, biomarkers of tau and amyloid. “Study cohort” will be included as an indicator covariate in the model since a random effects representation, as is perhaps common in classical meta analysis, is not possible. Based on these analyses, we will develop new knowledge on which combination of patient characteristics (demographic, clinical, biomarker) predisposes for a treatment effect to which type of drugs.
Requested Studies:
Modulation of Beta-amyloid Levels in CSF and Plasma by GSK933776 in Patients With Mild Alzheimer’s Disease or Mild Cognitive Impairment
Data Contributor: GlaxoSmithKline
Study ID: NCT01424436
Sponsor ID: BA1113043
A Randomised, Single-Blind, Placebo-Controlled Study to Investigate the Safety, Tolerability, Immunogenicity, Pharmacokinetics and Pharmacodynamics of Intravenous Infusion of GSK933776 in Patients With Alzheimer’s Disease.
Data Contributor: GlaxoSmithKline
Study ID: NCT00459550
Sponsor ID: BA1106006
Open-Label Extension for Alzheimer’s Disease Patients Who Complete One of Two Semagacestat Phase 3 Double-Blind Studies (H6L-MC-LFAN or H6L-MC-LFBC)
Data Contributor: Lilly
Study ID: NCT01035138
Sponsor ID: 5930
Effect of LY450139 a y-Secretase Inhibitor, on the Progression of Alzheimer’s Disease as Compared With Placebo
Data Contributor: Lilly
Study ID: NCT00762411
Sponsor ID: 11271
Effect of γ-Secretase Inhibition on the Progression of Alzheimer’s Disease: LY450139 Versus Placebo
Data Contributor: Lilly
Study ID: NCT00594568
Sponsor ID: 7666
A Multi-centre, Double-blind, Parallel-group, Randomised Controlled Study to Investigate Efficacy, Safety and Tolerability of Orally Administered BI 409306 During a 12-week Treatment Period Compared to Placebo in Patients With Cognitive Impairment Due to Alzheimer’s Disease
Data Contributor: Boehringer Ingelheim
Study ID: NCT02337907
Sponsor ID: 1289.7
A Phase 3, Multicenter, Randomized, Double-Blind, Placebo-Controlled, Parallel-Group, Efficacy and Safety Trial of Bapineuzumab (AAB-001, ELN115727) In Patients With Mild to Moderate Alzheimer’s Disease Who Are Apolipoprotein E4 Carriers.
Data Contributor: Johnson & Johnson
Study ID: NCT00575055
Sponsor ID: ELN115727-302
A Phase 3, Multicenter, Randomized, Double-Blind, Placebo-Controlled, Parallel-Group, Efficacy and Safety Trial of Bapineuzumab (AAB-001, ELN115727) In Patients With Mild to Moderate Alzheimer’s Disease Who Are Apolipoprotein E4 Non- Carriers.
Data Contributor: Johnson & Johnson
Study ID: NCT00574132
Sponsor ID: ELN115727-301
A Multi-centre, Double-blind, Parallel-group, Randomized Controlled Study to Investigate the Efficacy, Safety and Tolerability of Orally Administered BI 409306 During a 12-week Treatment Period Compared to Placebo in Patients With Alzheimer Disease
Data Contributor: Boehringer Ingelheim
Study ID: NCT02240693
Sponsor ID: 1289.5
Public Disclosures:
Dubbelman, M.A., Vromen, E.M., Tijms, B.M., Ottenhoff, L., Vijverberg, E.G., Prins, N.D., van der Flier, W.M. and Sikkes, S.A., 2023, July. P1-743 – Pooling trial data to identify heterogeneity and characteristics of patients most likely to respond to treatment: a causal forest approach. In Alzheimer’s Association International Conference. ALZ.