Study taps cerebrospinal fluid to identify dementia

Measuring beta amyloid protein levels in cerebrospinal fluid in conjunction with brain amyloid imaging using PET might become the best diagnostic test for presymptomatic dementia, particularly Alzheimer's disease, according to a study in the March issue of Annals of Neurology.

Previous animal research showed an inverse relationship between amyloid-containing plaques in the brain and a specific protein fragment, beta amyloid 42 (Ab42), in CSF. The new study is the first to demonstrate that same inverse relationship in humans.

Amyloid plaques are associated with Alzheimer's disease. Consequently, individuals with brain amyloid tend to have lower levels of Ab42. But neither marker alone is very sensitive or specific for AD. Together, they could overcome their solo shortcomings.

In the study, Anne M. Fagan, Ph.D., and colleagues from Washington University School of Medicine and the University of Pittsburgh recruited 24 individuals from the Alzheimer Disease Research Center at WUSM. Eighteen were diagnosed as cognitively normal, three as very mildly demented, two as mildly demented, and one as moderately demented. Participants underwent PET imaging using the radiotracer Pittsburgh compound B (PIB), which binds to brain amyloid plaques. They also underwent lumbar puncture for CSF samples.

Individuals fell into two nonoverlapping groups, according to Fagan. Seven participants with positive PIB binding (high levels of amyloid plaque) had the lowest levels of CSF Ab42, and 17 subjects with negative PIB binding had the highest levels of CSF Ab42. Interestingly, three of the seven positive PIB scans were clinically diagnosed as cognitively normal. The hypothesis is that these subjects have preclinical AD, Fagan said. These subjects will be followed over time to see if they develop dementia.

"If they do develop AD, it would mean that we have a way to screen individuals for the presence of AD pathology before the appearance of clinical symptoms," said Fagan, a research associate professor of neurology at WUSM.

Another interesting finding involved two subjects whose impairment was diagnosed as due to AD. A year later, however, another clinician who was blinded to the PIB-PET results determined their impairment was due not to AD but rather to frontotemporal dementia in one subject and possible abuse of sleeping pills in the other. The PIB scans of both subjects were negative (amyloid plaques are not associated with frontotemporal dementia), and their CSF Ab42 values were in the normal to high range, Fagan said.

Clinicians increasingly have to decide whether memory impairment is due to normal aging or dementia, and if dementia is AD. This study suggests a sequential process to do that, according to Dr. William Klunk, an associate professor of psychiatry at Pittsburgh and the codeveloper of PIB. A suspicious clinical evaluation would lead to a spinal tap. Low CSF Ab42 levels would lead to a PIB-PET scan for confirmation of AD.

Performing the tap in the middle of the process could increase the yield and justify the use of amyloid imaging when it's available for routine clinical use or even for identifying a homogeneous research population, he said.

"What we're shooting for in Alzheimer's disease isn't a better way of diagnosing a moderately demented AD patient. We can do that just fine," he said. "But waiting till patients are moderately demented hampers our efforts to develop new therapies. We need to be running trials in people who are diagnosed at the very first sniff of the disease."