fb tracking

Unreliable Alzheimer’s Diagnoses

If you’re not outraged,
you’re not paying attention!

Read what Public Citizen has to say about the biggest blunders and outrageous offenses in the world of public health, published monthly in Health Letter.


Unreliable Alzheimer’s Diagnoses

May 2011

Sidney M. Wolfe, M.D.

In a letter published in the May Journal of the American Medical Association (JAMA), Public Citizen criticized the authors of a study in a previous issue of the journal who tested the effectiveness of brain scans using an experimental dye to detect brain abnormalities found in Alzheimer’s disease patients. The authors withheld critically important data from the journal when documenting the extent to which interpretation of these scans varies from one physician to another, Public Citizen said.

The public health implications of these unreliable tests would be two-fold: First, substantial numbers of people could be incorrectly diagnosed as having such brain abnormalities when they really do not, and by the same token, others could be falsely reassured that they do not have such brain abnormalities when they actually do. By withholding important information, the authors misled the journal and its readers to believe that this test for identifying brain abnormalities was more reliable than it actually is.

Given what is known about Alzheimer’s disease and the lack of effective treatments, any proposed test for diagnosing this disease must be proven to be highly accurate in large, rigorously conducted studies before it is approved by the Food and Drug Administration for marketing.

In the JAMA study criticized by Public Citizen, researchers tested whether injection of an experimental radioactive dye, called florbetapir F 18, can be used to detect beta-amyloid plaques in the brains of living people using an imaging procedure called a positron emission tomography (PET) scan. The researchers proposed that the injected dye would bind to beta-amyloid plaques in the brain and be visible on a brain PET scan.

For the key part of the study, the researchers performed florbetapir-PET brain scans in 35 elderly, terminally ill subjects (only 17 of whom had been diagnosed with Alzheimer’s disease). After their deaths, the researchers compared those scan results to measurements of beta-amyloid plaques in their brains during autopsy. Importantly, three nuclear medicine physicians, who had undergone extensive training, independently read and scored each florbetapir-PET scan for the amount of beta-amyloid seen in the subject’s brain.

In their JAMA article, the researchers presented only the median score of the three physician readers for the amount of beta-amyloid seen on each florbetapir-PET scan. They withheld critical data on the scores for each individual physician reader – data that Avid Radiopharmaceuticals (Avid), a wholly owned subsidiary of Eli Lilly Inc. and the company that makes florbetapir, had submitted to the FDA on Sept. 17, 2010, four months before the JAMA article was published.

FDA analyses of the individual physician reader data showed the substantial variation among the three independent, highly trained physician readers. This led FDA reviewers to seriously doubt the validity and usefulness of florbetapir-PET scan for detecting beta-amyloid plaques in the brain.

Given the poor performance of the test at the individual physician reader level in a rigorously controlled setting, the variability and unreliability will only get worse when the test is widely distributed across the country. As a result, tens of thousands of patients will have either a false-positive test (told they have beta-amyloid deposits when they do not) or a false-negative test (told they don’t have beta-amyloid deposits when they do). This will lead to unnecessary anxiety and treatment in some patients, false reassurance in others and a waste of millions of health-care dollars.