The Researcher’s Dilemma: Finding Alzheimer’s Patients without Alzheimer’s Symptoms

In 2016, Alzheimer’s Disease International reported that over 47 million people globally were living with dementia.¹ This number is projected to increase to more than 131 million by 2050, as the population ages. This has a huge economic impact – in 2017 worldwide costs were estimated at US$818 billion, and this was expected to exceed a trillion dollars by 2018.¹ Dementia is an emotive issue. In 2012 a Marist Institute Public Opinion survey of more than 1200 Americans was carried out to gain a perspective on the most pressing concerns around Alzheimer’s disease (AD), including diagnosis, and the difficulties faced by carers. The survey revealed that people fear developing AD more than any other life-threatening disease, with 44% of Americans citing AD versus 33% citing cancer, 11% stroke, 8% heart disease and 4% diabetes.² Many people with AD are cared for by their own families, and the burden of this care was also reflected in the survey, with 61% saying that they felt unprepared to care for a loved one with an AD diagnosis, and the fear of becoming unable to care for oneself and therefore burdening others being a concern for 68% of respondents.

In recent years we have seen steady inroads being made to reduce mortality due to cardiovascular diseases, cancer and many other highprofile conditions. Now, however, AD kills more individuals than breast and prostate cancer combined.³ Over 5 million Americans currently live with (AD), and this number is increasing steadily.³

The pipeline for promising drug candidates for AD is poor and has been dogged by a history of failures.⁴ In January this year, Pfizer announced that it is pulling out of early stage research into Alzheimer’s and Parkinson’s diseases, with spend on these areas being reallocated to other areas of the pipeline where there is stronger scientific expertise.⁵ Historically, Pfizer has invested heavily in Alzheimer’s research, but a succession of promising candidate drugs failed in trials, including bapineuzumab, a joint venture with Johnson and Johnson which failed to show an improvement over placebo at Phase III,⁶ and PF-05212377 which was halted at Phase II due to lack of efficacy.^7,8 More recently, Axovant’s intepirdine, a candidate molecule with the same target as PF-05212377, also failed to meet its primary endpoint and further development was stopped at Phase II⁹ and Lundbeck’s idalopirdine was removed from the company’s pipeline after failure at Phase III.¹⁰

In any drug development process, the percentage of candidate molecules that fail to make it to commercialization is high – for example 81% of oncology therapies are likely to fail.^4,11 However in AD this figure has been much higher with a 99.6% failure rate reported over the 10- year period 2002-2012.⁴ A key issue here is also that for CNS-targeted products, failure comes to light much later in the development process, when considerable resources have already been expended in bringing a candidate molecule to clinical trial. Consequently, many companies are rethinking their strategies around potential AD products.

Should a promising drug candidate enter into clinical studies, there are more challenges ahead. Recruiting and retaining patients for these studies is difficult. For studies in patients already exhibiting AD symptoms there is a significant challenge in managing and gaining informed consent from patients who are already suffering from the effects of the disease and may have diminished decision-making skills.¹² In 2014, Grill and Galvin examined the status of recruitment into clinical trials in AD and identified many reasons why this can be difficult - caregivers often have to be involved, and this can add an extra burden to their daily routine, so clinic visit times need to be kept to a minimum. Also, many patients are concerned about the nature, risk, and potential discomfort of the study procedures, which may include invasive testing such as PET scans (which require use of an intravenously-administered radioactive tracer medium) or cerebrospinal fluid testing (requiring a lumbar puncture). Some patients may also be reluctant to take part in studies because of concerns that they may be randomized to a placebo arm rather than receiving active drug, or indeed may be unaware of research opportunities.¹³

A number of strategies to improve recruitment have been adopted by investigators, including educating physicians, community outreach, use of mailing lists and recruitment through satellite clinics, although few have any empirical evidence to support them.¹³

Physician Referrals – How Engaged are Healthcare Providers?

Traditionally, referrals through healthcare professionals (HCPs) – physicians and nurses – have been a common way that patients learn about and become involved in clinical trials. However the evidence suggests that although HCPs are well informed about studies and comfortable to discuss opportunities with their patients, ultimately they refer only a small percentage of patients.¹⁴

A recent study of 2,000 physicians and nurses, primarily in the United States and Europe, found that nearly all physicians (91%), and the majority of nurses (72%) feel comfortable discussing the opportunity to participate in a clinical trial with patients, but physicians refer less than 0.2% of their patients into clinical trials, and nurses refer even fewer. Prof. Ken Getz, who led the study, suggests that the results challenge the long-held notion that health care providers are a barrier to recruitment. What was also apparent from the study was that follow-up for referring HCPs was poor. Nearly 30% of physicians and 45% of nurses reported never receiving initial or follow-up contact from investigative site staff following a referral, and a higher percentage - 39% and 56% respectively - reported never receiving their patients’ clinical trial results. Similarly, 42% and 53% respectively report not receiving any kind of contact from the study center of investigators at the end of a clinical trial.^14,15 So it is not education that is required, but perhaps an understanding that HCPs may be more likely to refer patients to clinical trials if they were more invested in the process through provision of information and updates, and there are likely to be opportunities for improving relationships between clinical research staff and HCPs.

A Shifting Focus

However good the referral rate from physicians, however, not all individuals of interest for clinical studies will have an existing diagnosis. As we learn more about the AD disease process, the focus of research is shifting from treatment of symptomatic patients to primary prevention (studies in asymptomatic patients with known risk factors – for example those who are genetically predisposed to develop AD) and secondary prevention (treatment of patients diagnosed with very Mild Cognitive Impairment (MCI) to slow disease progression). Patients with an existing diagnosis will already be engaged with the healthcare system, however less than a quarter will be eligible for clinical trials. However those asymptomatic individuals that are increasingly being sought for primary prevention studies, in the absence of a known family history of dementia or preexisting genetic testing, are likely to have had no contact with the health system regarding AD or dementia, and so are essentially invisible to the research community.

A gene known as APOE can influence a person’s risk for the more common late-onset type of AD. The gene has three variants – APOE ε2, ε3 and ε4. Everyone has two copies of the gene (one inherited from each parent) and the combination affects the risk of developing AD. APOE ε4, present in 13.7% of people, increases the risk for AD and lowers the age of onset. Having one copy of ε4 can increase risk by 2 to 3 times. However having two copies (known as genetically homozygous) can increase the risk by up to 20 times.^16-19 Approximately 2% of the overall population are homozygous for APOE ε4.¹⁷

Although this may seem like a reasonable number of individuals, the attrition rate in clinical trial recruitment is high, so very few patients who are put forward will ultimately take part in a study. By way of illustration, if 100 individuals came forward, 72% (72 patients) may be within the designated age range for study participation. Of this 72, some 42% would agree to the genetic testing required to establish if they fell into the high-risk category (so 30% of the original cohort remains). Following genetic testing, approximately 4% would be referred to the clinic (so just over 1% of the original cohort remains). Of these, approximately 50% would be eligible to take part in the study (so 0.6% remains) and of those who are eligible, 70% may ultimately enrol in the study. At the end of this process, just over 0.4% of the original 100 patients remain. The outcome of this is that if a study requires 1300 homozygotic patients, then the screening process must start with almost 29,000 individuals.

Are Registries Part of the Answer?

There are few sources of this size, and this has inspired a greater focus on the use of participant registries as a repository of patients for these types of study. A number of registries exist that include populations that may be suitable for clinical trial enrollment. However all have different models of engagement and different target populations. Some contain minimal information about registered participants, and relatively few include any of the genetic information required to identify the enriched patient population that would provide the greatest insight in clinical trials of new AD treatments.

But what makes a good registry? There are many criteria that will separate the good from the mediocre – and key to this are proactivity, flexibility and detail. Keeping participants engaged with the registry is essential if the information held is to be up to date, so that changes in medical history are logged for instance. Similarly there needs to be the flexibility to update entries with changes in medication, clinical status or outcomes of assessments. And additional biological data could be immensely beneficial – ideally DNA data, but there may also be the opportunity to add biomarker data. While this is not currently as predictive as genetic data, ongoing advances in knowledge mean that there may well be a place for this information in the future, and so it should not necessarily be ruled out.

The attrition rates that are seen with clinical trial registration apply as much to registry-derived populations as to the general population. In 2016, Zhong and Cummings reported figures from the healthybrains. org A4 study which showed that of a registered population of 7,142 participants, ultimately only 4 (0.056% of the original cohort) could be randomized. Not all participants were opted in for clinical trials, and of those who were, there were losses to the population due to location (only those close to the study site went forward) and two levels of screening.²⁰ However what registries do offer is the ability to pre-screen a population very simply to identify, for example, those who are at higher risk of developing dementia. A big difference to the ultimate patient yield for clinical trials in AD can be made by the addition of a simple cognitive test.

In 2013, Brandt and Sullivan reported outcomes from the Dementia Risk Assessment (DRA) Model – an internet-based assessment that collects information to determine if a respondent is at an increased risk of dementia.²¹ The model considers a history of known risk factors, a validated informant report and a brief test of associative memory. Of 3168 respondents screened, 309 were found to have some form of cognitive impairment. While some 50% or these would be excluded from any clinical trial as they would fail to meet some of the inclusion criteria, this type of screening could produce a patient yield of nearer 1%, which although still a small fraction of those taking part, is a significant improvement on yields from those populations who do not undergo cognitive screening. The entirely automated and highly accessible format of the DRA may, therefore, make it useful in large-scale screening programs, such as those that might be required to identify at-risk elderly for inclusion in AD prevention studies.

Table 1. Examples of currently available registries including normal, MCI and AD patients.

The accessibility of on-line resources makes easy provision of simple cognitive testing for individuals in their own homes a very real proposition. And similarly the inclusion of more useful cognitive and risk-related data within patient registries is becoming more of a reality. Although there are many registries available, few are big enough, or comprehensive enough to meet a burgeoning research need. In the coming years, an estimated 50,000 participants will be needed for clinical trials in AD in the US alone.²² However recruitment is painfully slow and current methods are not effectively identifying the right kind of patient, particularly for prevention studies where asymptomatic or very early stage patients are required. There are many factors to be addressed, including site referral efficiency, site infrastructure and processing of information. Patient registries are likely to be only part of the answer, but those which have some level of cognitive assessment or hold genetic information, a measure of flexibility suitable to keep pace with changing needs and of course the ability to service the large-scale international cohort studies required for AD trials may be a significant help in the search for a solution to dementia.

Acknowledgements

The author would like to thank Lesley Jacques of Watermeadow Medical, an Ashfield Company, part of UDG Healthcare plc for medical writing and editorial assistance (funded by Synexus).

References

1. Alzheimer’s Disease International (ADI). World Alzheimer Report 2016. Improving healthcare for people living with dementia. Coverage, quality and costs now and in the future. Available at: https://www.alz.co.uk/research/WorldAlzheimerReport2016.pdf (Accessed 16 February 2018).
2. Marist Institute for Public Opinion https://www.scribd.com/document/113284885/HomeInstead-Marist-Poll-Alzheimers-Tables (Accessed 16 February 2018).
3. Alzheimer’s Association. 2017 Alzheimer's disease facts and figures. Available at: https://www.alz.org/documents_custom/2017-facts2017_infographic.pdf (Accessed 20 February 2018).
4. Cummings JL, Morstorf T, Zhong K. Alzheimer's disease drug-development pipeline: few candidates, frequent failures. Alzheimers Res Ther. 2014 Jul 3;6(4):37.
5. Hawks, N. Pfizer abandons research into Alzheimer’s and Parkinson’s diseases. BMJ 2018;360:k122 http://www.bmj.com/content/360/bmj.k122 (Accessed 16 February 2018).
6. Salloway S, Sperling R, Fox NC, et al. Two phase 3 trials of bapineuzumab in mild-tomoderate Alzheimer's disease. N Engl J Med. 2014 Jan 23;370(4):322-33.
7. Fullerton T, Binneman B, David W, et al. A Phase 2 clinical trial of Pf-05212377 (Sam-760) in subjects with mild to moderate Alzheimer's disease with existing neuropsychiatric symptoms on a stable daily dose of donepezil. Alzheimer's & Dementia: The Journal of the Alzheimer's Association, Vol. 12, Issue 7, P419.
8. US National Library of Medicine ClinicalTrials.gov, ClinicalTrials.gov Identifier: NCT01712074. Available at https://www.clinicaltrials.gov/show/NCT01712074 (Accessed 23 February 2018)
9. Biomedical Research Forum, DLB: Intepirdine a no-go, nelotanserin shrouded in controversy. https://www.alzforum.org/news/research-news/dlb-intepirdine-no-gonelotanserin-shrouded-controversy (Accessed 16 February 2018).
10. Biomedical Research Forum, Therapeutics: Idalopirdine. https://www.alzforum.org/ therapeutics/idalopirdine (Accessed 16 February 2018).
11. DiMasi J. Clinical success rates for new cancer drugs double while more enter testing. Tufts Center Study Drug Dev Impact Rep 2013:15;1–4.
12. PhRMA. Medicines in development for Alzheimer’s disease: 2013 Report. Available at: http://phrma-docs.phrma.org/sites/default/files/Alzheimer%27s%202013.pdf (Accessed 16 February 2018).
13. Grill JD and Galvin JE. Facilitating Alzheimer’s disease research recruitment. Alzheimer Dis Assoc Disord. 2014;28(1):1–8
14. Tufts Center for the Study of Drug Development, Impact REPORT 2017; 19:(1). Poor physician and nurse engagement contributes to low patient recruitment rates. Available at http://csdd.tufts.edu/reports/description/ir_summaries (Accessed 18 February 2018)
15. Tufts Center for the Study of Drug Development. Poor physician and nurse engagement contributes to low patient recruitment rates. Press release available at http://csdd.tufts. edu/news/complete_story/pr_ir_jan_feb_2017 (Accessed 23 February 2018)
16. Michaelson, DM. APOE Epsilon4: The most prevalent yet understudied risk factor for Alzheimer's disease. Alzheimers Dement, 2014;10(6):861-8.
17. Farrer LA, Cupples LA, Haines JL, et al. Effects of age, sex, and ethnicity on the association between apolipoprotein E genotype and Alzheimer disease. A meta-analysis. APOE and Alzheimer Disease Meta Analysis Consortium. JAMA 1997: 278 (16);1349–56.
18. Hauser PS, Ryan RO. Impact of apolipoprotein E on Alzheimer's disease. Current Alzheimer Research. 2013; 10 (8): 809–17.
19. Liu CC, Liu CC, Kanekiyo T, Xu H, Bu G. Apolipoprotein E and Alzheimer disease: risk, mechanisms and therapy. Nat Rev Neurol. 2013 Feb;9(2):106-18.
20. Zhong K and Cummings J. Healthybrains.org: From registry to randomization. J Prev Alz Dis 2016;3(3):123-126.
21. Brandt J, Sullivan C, Burrell LE II, Rogerson M, Anderson A. Internet-based screening for dementia risk. PLoS ONE 2013;8(2).
22. Buckholtz NS, Recruiting Patients for Alzheimer’s Clinical Trials. Aging Well 2012;5(6):34.