Most trials for potential Alzheimer disease (AD) treatments fail—and the reason may be poor protocol design.
After all, typical AD studies are double-blind placebo-controlled parallel group clinical trials with a dual outcome, including a cognitive measure and a global impression of aptitude for the activities of daily living—a trial design originally developed to study cholinesterase inhibitors and memantine. While this is an effective protocol for a drug designed to treat cognitive symptoms, the majority of compounds under investigation today are designed to be disease-modifying.
How can AD trials separate symptomatic effects of potential agents from disease-modifying effects? Harnessing a different set of trial protocols can be an effective first step—and there are several to choose from:
- A wash-in analysis compares the change in clinical outcome measures between the treatment group and the placebo group over the first few weeks or months of the study. If the treatment group shows greater improvement than placebo-treated subjects this could potentially indicate an early symptomatic effect, because a disease-modifying effect would not be seen so soon. This trial design is often combined with other design strategies.
- A washout analysis begins when treatment is withdrawn from both the active agent- and placebo-treated groups at the end of the study. The active agent is assumed to have disease-modifying properties if patients in the treatment group show slower disease progression throughout the double-blind treatment period than those treated with placebo and less severe deterioration when treatment is withdrawn.
- Delayed-start design randomizes one group of patients to receive the active agent from the beginning while the second group initially receive the placebo and later receive the active agent. In theory, if the active agent has a purely symptomatic effect, then when the second group receives it, the progression curves for both groups should meet. However, if the compound has a purely disease-modifying effect then the progression curve of the delayed start group will never catch up with those treated from the beginning.
- Futility studies compare the outcome of a single treated group against a pre-determined threshold value reflective of a clinically meaningful change. Such studies may use a placebo control arm, but often use historical controls to establish the threshold for clinical meaningfulness. Thus, fewer patients are observed for a shortened period of time in Phase 2, which facilitates decisions about which agents to prioritize for advancement to Phase 3.
- Well-designed long-term follow-up trials enable thesis modification to be inferred from any sustained divergence in outcome measures over time. This may well be the best current trial design to inform thesis modification; however, these studies are time-consuming and expensive and it’s not entirely clear how long they should last.
- Adaptive design is a very attractive protocol design which can minimize the overall sample size and duration of a study by stopping recruitment early in response to strong signals of success or failure, based on regular interim analysis. An important challenge for this design is the difficulty of assessing the outcome measure early enough so that any modification of the randomization can occur well before recruitment is complete. This challenge can be addressed by using a Bayesian design, which has numerous advantages over other statistical approaches—but requires extensive planning, a lengthy simulation process and a rapid flow of data from sites to the database to enable interim analysis in real-time.
While these six designs can help researchers identify AD-disease-modifying effects, there are also novel protocol designs for behavioral disturbances studies in AD—some of which overlap with disease-modification trial protocols.
- Withdrawal design is a novel tactic compared to the usual approach to behavioral disturbance studies.
- Delay to onset design also may be of value in AD trials focusing on symptom management
- Parallel sequential comparative design is a two-phase design that seeks to minimize the placebo group response.
Certainly, these protocol designs require further investigation and modification—but they each represent an opportunity to eliminate some issues with current AD trials, and to improve study.