Treating Alzheimer's Early

In February, the FDA offered a glimpse (http://1.usa.gov/14FSbAN) of its vision for bringing to market drugs targeting the earliest phases of Alzheimer's. Focusing on pre-dementia stages of Alzheimer's may offer the best chance of managing the disease as significant physical changes have yet to occur in the brain, offering the possibility of successfully treating, or even preventing, full-blown Alzheimer's. This move is echoed by a number of drug companies, notably Eli Lilly and Janssen Pharmaceuticals, looking to revitalize failed Alzheimer's drugs by testing them in early stage patients.

Such a goal, however, creates new problems in assessing drug efficacy. In particular, clinician reported outcomes (ClinROs), which are still central to gathering data in Alzheimer's trials, present some unique issues. Technology, particularly electronic clinical outcome assessments (eCOA), can help life science companies in meeting some of these challenges.

The draft guidance suggests using a ClinRO, specifically the Clinical Dementia Rating—Sum of Boxes (CDR-SB), as a scale for assessing Alzheimer's in patients nearing the overt dementia phase (i.e., the patients dementia has become apparent).

The proposal of CDR-SB for pre-dementia Alzheimer's trials marks a shift for the FDA. In trials of patients with overt dementia, the FDA looks for changes to co-primary outcomes, patient cognition, and function. Yet, as the FDA acknowledges in its draft guidance, the milder functional impairments of patients in the earliest phases of the disease make this very problematic. Currently, there is a lack of validated tools for testing functional impairment in these early phase patients.

In recognition of these challenges, the FDA is considering accepting the composite cognition and function scale of the CDR-SB. A CDR-SB score is generated through semi-structured interviews of Alzheimer's patients and their close associates, and assesses both cognition and function. While the FDA appears ready to accept marketing applications based on CDR-SB data, it remains committed to co-primary outcomes as the ideal. The FDA focus is still on improving how patients function and feel, and ClinROs will play a central role in assessing these changes.

The use of ClinROs gives life science companies a familiar process to work with, but also raises old problems. Research shows different clinicians tend to rate patients differently. In fact, there is even a lack of consistency in the ratings given by individual clinicians. These inter-/intra-rater reliability issues are a potential source of bias in Alzheimer's trials. The duration of Alzheimer's clinical trials and long gaps between patient visits also create logistical challenges for sites.

Improving Alzheimer's assessments with eCOA

Using eCOA, rather than paper, overcomes a lot of these issues in trials of post-dementia Alzheimer's, and the technology is equally applicable to early stages of the disease. An electronic system can automatically track what questionnaire a patient, caregiver, or clinician is meant to be completing at any given study visit. If paper-based systems are used, it becomes more difficult to keep track of what questionnaires need to be completed when. For sites, eCOA reduces the risk of failing to comply with the protocol because of a clerical error.

An electronic system also addresses problems around the reliability of assessments. Basic errors, such as adding up the scores of complicated ClinROs incorrectly are eliminated by the automation of the eCOA system. Clinicians using the technology are free to concentrate on their assessment of the patient for the duration of the interview, which for the CDR can last more than an hour.

eCOA can also guide a clinician through an assessment in a consistent and user-friendly manner. With branching questionnaires, eCOA systems can hide or display questions based on earlier responses, guaranteeing the form is completed correctly. The use of prompts and reminders to help clinicians complete the more complex questionnaires is also possible. Each of these features increases the uniformity of the administration process, cutting the likelihood of variation from clinician to clinician and improving data quality in an Alzheimer's trial.

Correlation of the answers given to questions assessing the same domain offers another way to see if a clinician is rating a patient accurately and consistently. Different Alzheimer's questionnaires often assess similar aspects of cognition and functioning. One would expect that for these related questions, clinicians should be giving a patient similar scores. If the eCOA system detects significant variation, it can flag the concern for study teams to investigate. Reminder training could then help improve the clinician's testing.

Alzheimer's clinical trials, particularly studies of pre-dementia patients, are looking for small improvements to patient cognition and function within noisy datasets. Life science companies can ill afford inconsistent ClinRO data jeopardizing their Alzheimer's clinical trials. eCOA can help by improving ClinRO reliability, resulting in cleaner data. Physician familiarity with eCOA is on the rise, increasing the acceptability of the systems while reducing the need for site training. The simplicity of the technology means first-time users quickly recognize the benefits too.

Alternatives to ClinROs

While ClinROs combining cognition and function are the FDA's preferred method for gathering data from patients with dementia, the FDA proposes alternatives for even earlier stages of the disease. A single primary efficacy measure showing cognition improvement could support approval. The FDA is also willing to consider "time to dementia" as a primary measure of efficacy. However, such a trial would last a long time and the FDA indicates a ClinRO jointly assessing cognition and function is more practical.

The guidance also addresses biomarkers, while raising doubts about their (current) practicality. As the FDA notes, despite lots of research into biomarkers in Alzheimer's, there is a lack of consistent evidence to show a measurement can predict clinical benefit. Until there is widespread, evidence-based consensus that a measurement predicts clinical benefits, the FDA will not consider approval based on biomarkers as a surrogate outcome measure. For now the role of biomarkers is limited to diagnostics and, if new evidence emerges, as a secondary outcome.

The full version of this article is available at www.crfhealth.com/alzheimer-article.

Paul O'Donohoe is Manager of Health Outcomes at CRF Health, e-mail: paul.odonohoe@crfhealth.com.