Scientists may have found protein code that indicates Alzheimer's
1:55 p.m. Wednesday, January 17, 2007
We all misplace keys and forget things. But when does simple forgetfulness indicate the start of Alzheimer's disease? Even clinicians have trouble answering that.
"The best available techniques we have right now, which are based on clinical judgement, depending upon the individual who does it, can be between 50 to about 85 percent accurate," said Dr. Norman Relkin of the Weill Cornell Medical College.
Relkin, a behavioral Neurologist, says the only 100 percent accurate way to diagnose Alzheimer's is an autopsy. But, he and Cornell University's Kelvin Lee may have found a way to diagnose Alzheimer's in living patients almost as accurately.
They compared proteins from spinal fluid of 34 people who died of autopsy proven Alzheimer's with 34 who did not have the disease. They identified a unique pattern of 23 proteins in the Alzheimer's group, that could serve as something like a barcode for Alzheimer's.
"By comparing barcodes from patients that have Alzheimer's disease from patients that don't have Alzheimer's disease, we're able to recognize certain patterns that tell us whether the subject has the disease or not," Lee said.
The barcode was also found in nine out of ten living patients suspected of having Alzheimer's.
"So the overall accuracy that we've observed is about 90 percent, a little bit higher than 90 percent, in assessing whether a patient has the disease or not," Lee said.
Relkin and Lee say their test still needs further validation before doctors can start using it. But they hope eventually it can be used to track the progress of the disease in patients, and even predict how well new Alzheimer 's drugs will work.
Relkin says the procedure used to collect the fluid -- a spinal tap -- is not as painful or risky as many people think.
And for patients, detecting and treating the disease early, even before obvious symptoms appear, would be a huge benefit.
Video courtesy of Sciencentral.








Post a comment
(Requires free registration.)