The brain changes that lead to Alzheimer's disease probably begin
years, and possibly even more than a decade, before symptoms such
as memory impairment appear. For this reason, Alzheimer's
researchers have long hoped to find biomarkers — early biological
signs of disease pathology — that could help identify people at
risk before symptoms develop.
Heart disease provides one example of how biomarkers can be useful.
Elevated levels of unhealthy LDL cholesterol or total cholesterol,
detected with a simple blood test, are biomarkers that indicate
someone may be at increased risk of having a heart attack. Statins
(anti-cholesterol drugs) reduce both LDL levels and risk of heart
As yet, however, there is nothing close to an Alzheimer's
equivalent of a cholesterol blood test. The tests for two of the
most promising categories of Alzheimer's biomarkers — brain imaging
scans and spinal fluid sampling — are more cumbersome to perform
than blood tests. Although the research on these biomarkers is
interesting and may be promising, none are yet reliable enough to
use in the clinical setting.
Perhaps most significant for patients concerned about their own
risk, no treatments are available to stop Alzheimer's disease
progression. Although a number of "disease-modifying" compounds are
under investigation, so far all have failed in the late-stage
clinical trials that are necessary to prove efficacy. The most
recent flop occurred in August 2010, when a drug company halted
development of a compound known as semagacestat after patients with
Alzheimer's participating in a clinical trial fared worse after
taking the drug than those taking placebo.
Still, the research on biomarkers continues because some
investigators believe that the disease-modifying drugs must be
given earlier in the Alzheimer's disease process to be effective.
Whether or not that is true remains to be seen.
To continue reading this article, you must login
Subscribe to Harvard Health Online for immediate access to health news and information from Harvard Medical School.