27 Apr 2003

In this week's Journal of the American Medical Association, researchers report a correlation between Alzheimer's disease (AD) and the appearance in the cerebrospinal fluid (CSF) of two proteins thought to play major roles in the progression of the disease-amyloid β42 (Aβ42) and tau.

This is not the first time CSF levels of these proteins have been touted as possible predictors of the disease (see ARF related news story and ARF story), but to date, it is the largest study to test the relationship. Trey Sunderland and colleagues from the National Institute of Mental Health, Bethesda, Maryland, together with collaborators from Pfizer, and Vertex Pharmaceuticals, measured tau and Aβ42 in the CSF of 100 patients with probable AD, 31 patients who had since passed away and were confirmed by autopsy to have had the disease, and 72 control volunteers. Sunderland and colleagues demonstrate that Aβ1-42 levels are significantly lower in AD patients vs. controls, while the opposite is true for tau. However, as the authors point out, for each protein there is considerable variance within each group, making solid predictions difficult. For example, though the mean tau level in the control group was 224 pg/ml, several normal volunteers had over 600 pg/ml of tau, which is above the mean level in the AD group (587 pg/ml). This overlap between control and disease groups are a major roadblock in developing this test. Taking both sets of data, Sunderland and colleagues were able to determine cut-off points that predict AD with some degree of certainty, i.e., more than 200 pg/ml tau together with less than 444 pg/ml Aβ42, achieve a sensitivity of 92 and a specificity of 89 percent, respectively. This is comparable to values obtained from clinical diagnosis.

Even so, scientists still have a way to go before such measurements become truly useful for diagnosis. As the authors point out, the differences in protein levels may be due to the starkly different populations studied; moreover, using control volunteers with different types of dementia may be more informative, realistic, and challenging. The major benefit of measuring these markers to date, Sunderland suggests, may be in tracking the progression of the disease in individuals most at risk.—Tom Fagan