Alzheimer’s blood test breakthrough

Monday October 15 2007

A diagnostic test has been developed that can predict Alzheimer’s before the onset of any symptoms, newspapers reported. The newspapers said that the new test – a simple blood check – may be used to predict the onset of the disease six years in advance of symptoms.

The laboratory study behind these stories is one that has appeared to identify an “Alzheimer’s-specific signature”; a set of 18 proteins in the blood that appear to act as a marker for Alzheimer’s disease. Markers are indicators of the presence or risk of disease.

The researchers only tested the “Alzheimer’s signature” on samples of blood plasma that were available at the time of the study. In most of these, a certified diagnosis of the disease (only possible post mortem) had yet to be established because the studies participants were still alive.

Further research will give us a more useful idea of the application of such a test, but for now, this is an exciting laboratory finding. It could take several years before a test based on these findings is available.

Where did the story come from?

Sandip Ray and colleagues from several medical institutions across the United States conducted this research. The study was funded by the John Douglas French Alzheimer’s Foundation, the Alzheimer’s Association, the US National Institute on Ageing and Satoris, Inc. The authors declare that they have competing financial interests, and some are employed by Satoris (a manufacturer of pharmaceuticals and other goods) which part funded this study. The study was published in the peer-reviewed medical journal Nature Medicine.

What kind of scientific study was this?

This is a laboratory study conducted on samples of blood plasma from 259 individuals. 85 of these people were known to have Alzheimer’s disease while 79 served as “non-demented” controls. The remaining 95 people had other conditions, including other types of dementia, mild cognitive impairment, another neurological disease such as Parkinson’s or multiple sclerosis, or rheumatoid arthritis

The researchers wanted to investigate the difference between the (blood) plasma of people who have Alzheimer’s and those who were ‘non-demented’ controls. To do this, they compared the concentrations of 120 different plasma proteins (molecules that perform a variety of functions in the body) between 43 of the Alzheimer’s samples and 40 of the control samples.

Using various analyses, they identified a set of proteins that were particularly different between the two groups as an “Alzheimer’s-specific signature’

They then used this “signature” to predict the Alzheimer’s status of the remaining 42 members of the Alzheimer’s group and the 134 other samples from both the other groups. By doing this, they could see how accurate their signature set was at predicting whether a blood plasma “sample” had Alzheimer’s.

To assess further the accuracy of their test, they took samples of blood plasma from people in two previous studies. These people had mild cognitive impairment at the start of the studies and were then followed up for two to six years to see whether they converted to Alzheimer’s disease. The researchers were interested in whether applying their test to the plasma samples taken at the beginning of the study could predict those which had converted to Alzheimer’s.

What were the results of the study?

The researchers found that 18 of the 120 proteins that they analysed appeared in different concentrations in the plasma of Alzheimer’s patients compared to the non-demented controls.

They considered this to be the best  “signature” of Alzheimer’s. When they used it to predict the diagnoses of the remaining samples, they found that it predicted 90% of the Alzheimer diagnoses and 88% of the non-Alzheimer diagnoses.

When using their test on a cohort of people with mild cognitive impairment, the researchers found that the test was able to predict 20 out of 22 (91%) of those who developed Alzheimer’s two to five years later. However, in this same group of people, the test misclassified seven out of 17 of them as having Alzheimer’s when their diagnosis remained as mild cognitive impairment.

What interpretations did the researchers draw from these results?

The researchers state that they have identified an Alzheimer’s biomarker that could potentially be used for the diagnosis of early Alzheimer’s disease. They believe that similar “signatures” may exist for other diseases of the central nervous system and that these could hold clues for both treatment and diagnosis.

What does the NHS Knowledge Service make of this study?

This is a well-conducted laboratory study that has identified a set of proteins that appear to act as markers for the development of Alzheimer’s disease.

  • The most important point to highlight, is that it is not clear how many of the people who tested positive for having Alzheimer’s, actually had a definitive diagnosis of the disease. A definitive diagnosis of dementia can only be made with certainty by examining the brain post mortem. The researchers state that “many patients from our study are still alive and we cannot be 100% certain about the diagnosis for each of them”. They mention that in the first part of their test, the protein signature was able to identify “eight of nine post-mortem-confirmed subjects with Alzheimer’s disease”, implying that a definitive diagnosis was available in this very small number of samples. To test their accuracy, the results of new diagnostic tests must be compared with those from a “gold-standard” test (one that gives a definitive diagnosis). Admittedly, this is more difficult with Alzheimer’s disease, but further research when diagnoses are available after death will give more information that is useful.
  • When it is trialled in clinical practice, it will be important to assess the numbers of people who the test wrongly diagnoses as having Alzheimer’s and those who do have the disease, but which the test misses. These false positives and false negatives of a test can cause distress to patients and therefore need careful and accurate assessment.

Sir Muir Gray adds...

Even if the assay does turn out to be a reliable predictor, I don’t think I want to know I am going to develop Alzheimer’s six years hence unless an effective treatment has been developed by that time.

Analysis by Bazian
Edited by NHS Choices