A biomarker is "an abnormal signal from a body fluid or tissue that can provide distinguishable pathologic information."
1 The characteristics for an ideal AD biomarker are listed in
Table 1.
Because of the "probabilistic" nature current criteria, there is an unmet need for exclusionary (ie, high-sensitivity) diagnostic biomarkers whose presence signify categoric diagnosis AD. Especially important would be a biomarker profile characteristic of AD in incipient stages, eg, amnestic mild cognitive impairment (MCI), since this might enable appropriate intervention limit or to prevent neurologic damage (assuming such interventions are available; see "
Treatment" section).
An extensive research effort has been mounted to identify AD biomarkers in accessible clinical samples, such as cerebrospinal fluid (CSF), blood plasma, and urine. It should be noted, however, that CSF is not routinely collected in AD assessment and that lumbar puncture is not a common procedure in primary care, psychiatric, or geriatric practice sites. Nevertheless, since the composition of CSF closely resembles that of the extracellular brain space, it would be expected to contain biomarkers associated with AD pathophysiology.
3 Phosphotau is significantly increased in the CSF of AD patients compared with that of healthy controls, and this may even distinguish AD-related MCI from non- AD-related MCI.
4 Beta amyloid, on the other hand, is decreased in the CSF of AD patients. This, however, may be due to sequestration in the neuritic plaques that are characteristic of AD brains.
5 Potential AD biomarkers in blood plasma would necessarily be limited to small molecular species, lipophilic molecules, and molecules attached to specific transporters because of the "filtering" action of the blood-brain barrier.
6 Amyloid beta—an obvious AD biomarker candidate—is present in both CSF and blood plasma. However, although plasma levels of amyloid beta are elevated in familial AD patients,
7 such levels do not reliably distinguish between sporadic AD patients and controls.
8 Deposition of amyloid beta in the brain causes reactive inflammatory responses including up-regulation of proinflammatory cytokines. One of these, interleukin-6 (IL-6), has been extensively investigated as a potential AD biomarker in blood plasma. However, results so far have been inconclusive.
9 Urinary biomarkers, of course, would be ideal for the detection of AD. Only one molecule, neural thread protein (NTP; specifically, AD7c-NTP, the form associated with AD), has so far been investigated. NTP is a brain protein that is selectively elevated in AD, and overexpression of the gene for NTP is associated with apoptotic cell death similar to that found in AD brains. In a recent study,
10 elevated NTP was detected in the urine of 91% of subjects with probable AD, 38% of those with possible AD, and 49% of those with MCI. Among definite non-AD controls, 91% had normal NTP levels. This suggests that NTP may be a useful biomarker for well-developed AD but is less useful in early AD, because it is unknown how many MCI subjects will eventually convert to AD.
3 [Editor's Note: A simple urine test for NTP has been developed with reported sensitivity of >80% and specificity of 90% in distinguishing between probable AD and definite non-AD (patients categorized as definite non-AD were determined to have conditions that included agerelated memory decline, pseudodementia due to depression, and metabolic disorders). Although the proposed label states that "the results are intended for use in conjunction with, and not in lieu of, current standard diagnostic procedure," marketing of the test kit in the US has been blocked by the Food and Drug Administration, mostly because of equivocal results between possible AD and MCI subjects, giving rise to concern over the value of the test in clinical practice if not used in conjunction with established clinical approaches to diagnosis. However, the kit has been approved in Europe.11] 
