The promise of the Alzheimer's vaccine -- that it could slow, or even reverse, Alzheimer's disease -- was powerful enough to persuade 300 patients to volunteer for a study of the experimental therapy in 2001.
Scientists thought the vaccine might stimulate the body's immune system to clear up the toxic plaques that destroy brain cells in Alzheimer's patients.
Four months after the study began, however, a safety board shut it down. Several patients had suffered a life-threatening side effect: encephalitis, a dangerous swelling of the brain.
The story is proof that research can be dangerous, even deadly.
Other studies of promising treatments for Alzheimer's disease have ended in less drama than the vaccine study, but with equal disappointment.
Since Alzheimer's disease clinical trials began about 20 years ago, thousands of patients have sacrificed their time and effort to attend frequent visits for studies lasting as long as two years.
For many studies, the result was the same: a promising drug failed to treat the disease.
How do researchers justify exposing people with Alzheimer's disease to these risks, hassles and disappointments?
A core justification is informed consent, a term that describes a competent subject who agrees to join a study after the investigator discloses all the relevant details about the research, such as its risks and benefit.
But informed consent -- that core justification -- is a challenge in research that enrolls persons with Alzheimer's disease.
By definition, Alzheimer's patients have trouble with memory, concentration and attention. They may not be competent to give an informed consent.
Without the patient's informed consent, how do we ethically enroll patients with Alzheimer's disease in research?
In the United States, the answer to this question remains the source of a loud and still unresolved debate. We have rules for researchers who study so-called vulnerable populations, children, prisoners and pregnant women. We do not have rules for research that involves people with Alzheimer's disease.
Jason Karlawish is associate professor of medicine at the University of Pennsylvania.
The debate centers around the use of proxy consent -- when one person provides informed consent on behalf of someone who cannot give his own consent. For example, the adult child of an Alzheimer's patient might offer proxy consent.
Some ethicists want strict limits on this practice. For example, the proxy could only enroll a patient in research if he or she had clear guidance from that person that they would want to be in research. Perhaps several years ago the patient wrote down what kinds of research they did or did not want to participate in. Their concerns include that a proxy will act according to their own interests, rather than according to the patient's best interests.
Others disagree with this position. They argue that to restrict research enrollment only to people who are either competent to give consent or who previously expressed a willingness to participate limits the number of potential research subjects. This will in turn slow the progress of discovering better treatments.
Each side in this debate has it merits. But neither side addresses adequately the critical questions: Who should be a patient's proxy? What limits, if any, should we place on the risks of research that enrolls patients with proxy informed consent?