Preclinical detection of Alzheimer’s disease poses some unique ethical concerns
Predictive testing raises "very hot ethical and social issues"
Executive Summary
Preclinical diagnosis of Alzheimer’s disease presents significant ethical concerns, such as the tests’ limited accuracy and lack of effective treatment for individuals identified at higher risk.
- Labeling persons as having preclinical Alzheimer’s disease who have a good chance of never developing the condition in their lifetimes is potentially harmful.
- Informed consent processes need to address the possibility of stigma and discrimination.
- It is difficult to establish the value of predictive testing in the absence of effective therapies.
Preclinical diagnosis of Alzheimer’s disease shares some of the same controversies surrounding preclinical states of other diseases such as cardiovascular disease, osteoporosis, and diabetes, says Jason Karlawish, MD, professor of medicine, medical ethics, and health policy at University of Pennsylvania’s Perelman School of Medicine. However, there are also some unique ethical considerations.
"We operationalize our ethic of autonomy through our brain," he explains. "So as we talk about labeling people’s brain at risk of decline before they are ill, we are playing with very hot ethical and social issues."
From a clinical point of view, the primary ethical consideration is whether diagnosing preclinical disease will make patients better or worse off, says Kenneth Covinsky, MD, MPH, professor at the School of Medicine at University of California, San Francisco. He says in most cases, imaging modalities cause more harm than benefit.
There is currently no treatment for individuals at higher risk for Alzheimer’s disease. "So it is not clear how the early identification of those at risk will lead to benefit," says Covinsky. "But the more important reason is that the tests are of limited accuracy."
The tests identify amyloid in the brain, which is thought to be part of the pathology that leads to Alzheimer’s, but many persons who have positive tests will never develop the condition. Labeling persons as having preclinical Alzheimer’s disease who have a good chance of never developing the condition in their lifetimes is more likely to cause harm than benefit, argues Covinsky.
"We know that many persons who die of other causes have amyloid in their brains, and some who develop dementia do not have amyloid," says Covinsky. "At this point, brain amyloid may be a risk factor, but not destiny."
Should results be disclosed?
Karlawish says that with longitudinal studies, which measure various biological and clinical features of individuals and follow them over time to look for change, the primary concern is preserving the validity of the study, while at the same time protecting subjects from the harms of the study.
"There is a strong argument that you don’t want subjects to learn preclinical biomarker results, for two reasons," says Karlawish. First, the study’s validity could be harmed. Once participants learn their result, it could affect how they view their cognition and how they perform on a test. It could also affect how the clinical researchers assessing them view their performance.
"So if biomarker information is given out to individuals who are cognitively normal on cohort studies, you run the risk that you are going to corrupt the validity of the study, which is ethically inappropriate," says Karlawish.
Another reason not to reveal the results is that the information could result in stigma or discrimination. Karlawish is co-author of a 2014 paper outlining the current lack of preparedness to address the health care policy and legal implications of preclinical detection of Alzheimer’s disease.1 "The general ethic to govern cohort studies is a policy that data are gathered, but not returned to individuals," says Karlawish.
Karlawish says that researchers would want to disclose results to participants in clinical trials looking at whether an intervention affects individuals with one or more features that might be diagnostic for a preclinical stage of the disease. "There is a good case to be made that a valid study requires this," he says.
In clinical practice, individuals would learn their results, and based on these results, would get the intervention. "So you’d want your study to test that paradigm — I tell you the result and based on the result, I give you a drug. Knowing the result and getting the drug, how do you do?" says Karlawish.
Researchers have to address the possibility of stigma and discrimination, he adds, by setting up steps in the enrollment and screening processes to identify individuals who understand the limits of the information, and are psychologically prepared to receive it.
This results in "a very informed consent-heavy process, to make sure people fully understand what they’re getting involved in," says Karlawish. "Researchers can draw analogies to the world of genetic testing for some guidance on how they ought to do that."
Two possible futures
Karlawish says there are two possible futures involving preclinical diagnosis of Alzheimer’s disease. One is that there will be rapid progress in personalized medicine for the brain, driven by therapeutics that successfully treat people.
In that scenario, says Karlawish, "the tight link between the thera-peutic and the biomarker will rapidly carve out a space for patients to be diagnosed, labeled, and treated."
Ethical questions would include whether fringe or borderline conditions should be treated, and the cost and duration of therapy. "But it’s generally a positive future, because it would show that at least some proportion of the population is benefiting from early intervention," says Karlawish.
The second possibility is that over time, longitudinal data accumulate in the absence of any effective therapies. "People will make the case that we should be using the data to diagnose people just because they want to know the information," says Karlawish. "But that will be a very difficult, steep, rocky road for us to climb, as a society and as a profession."
Risk prediction models are only so certain, he explains, and it is very difficult to translate those into clinical practice. "Large issues loom about what the value of the diagnostic is in the absence of a biologically validated therapy," says Karlawish. "That scenario is a clinical and policymaking nightmare."
One reason is the difficulty of validating a preclinical marker without a therapeutic. "The therapeutic is not just to help people — it’s also valuable to validate the construct of preclinical disease," he says. "I think a lot of people don’t recognize that second aspect of the role of therapeutics."
For this reason, Karlawish is generally pessimistic about a future of predictive testing that is driven simply by individuals’ desire to know the information.
"It strikes me as very difficult to establish value in that setting," he says. "With competing resources for health care, I don’t see how you can make the argument that that is a high priority."
- Arias JJ, Karlawish J. Confidentiality in preclinical Alzheimer disease studies: When research and medical records meet. Neurology 2014;82(8):725-729.
- Kenneth Covinsky, MD, MPH, Professor, School of Medicine, University of California, San Francisco. Phone: (415) 221-4810 ext. 4363. E-mail: [email protected].
- Jason Karlawish, MD, Professor of Medicine, Medical Ethics and Health Policy, University of Pennsylvania, Perelman School of Medicine. Phone: (215) 898-8997. E-mail: [email protected].
