Cerebrospinal fluid biomarkers in Alzheimer’s disease, vascular dementia and ischemic stroke patients: a critical analysis

The CSF analysis offers an excellent opportunity to detect early signs of neuronal degeneration and this has been widely used for AD and other dementia disorders. However, for correct interpretation, information on potential confounding factors is extremely important. Thus, we studied CSF alterations in CSF marker profiles of commonly used AD biomarkers in patients with AD, VaD and stroke.

Increased values of tau and a decrease of Aß42 were detected in AD in a similar manner as in VaD and acute ischemic events. With special regard to total tau, which is known as an indicator of neuronal damage and discussed to be specifically altered in AD, an increase after stroke was not expected to this degree. In our literature research, we found only limited data dealing with CSF biomarkers after stroke, all of them well in line with results obtained here [2‐5]. Tau was described to be significantly increased at day 2–3 (179 %), showing a peak after 1 week (257 %) and after 3 weeks (425 %), normalizing not until 3–5 month (140 %). No significant alteration of p-tau was reported.

For the distinction between VaD and AD, CSF proteins were found to be more altered in AD, such as Aß42 decrease and tau increase, especially phosphorylated forms (Table 2) [6‐8]. According to the literature, to differentiate AD from VaD, a combination of all three is recommended [9]. With special regard to AD, a huge amount of data was published. One important aspect is MCI and its progression to AD, which can potentially be predicted with the help of CSF profile [5, 10, 11]. Apparently those MCI patients with a more AD-like CSF profile progress to AD in a shorter period of time, mostly showing low Aß42 and high tau levels [12, 13]. Because CSF biomarkers are already altered very early, there is the possibility to detect AD patients at risk or at very early stages. More precisely, Amyloid-peptide burden and changes in APP metabolism are altered at first up to 10 years before clinical symptoms. Tau proteins are supposed to be rather late markers [14].

The pathogenetic basis for our findings is not clear. CSF Aß42 values are supposed to predict disease progression in AD, but can be altered in VaD, too. Up to now studies have mostly focused either on AD or on stroke and directly comparative analysis is rarely given. Similar to AD, imbalances concerning the total cleavage of APP, with tendencies towards accumulation to amyloid plaques, was reported in VaD. Analogies between cognitive decline in patients after stroke with or without cerebrovascular pathologies prior to this can be seen in comparison to amyloid impact in AD [15].

Total tau might reflect the degree of neuronal damage after ischemia [3, 4]. CSF samples taken immediately after stroke, some days up to several weeks after stroke document an increase of t-tau with a peak after 1 week and a re-normalization 3–5 months later [4]. The level of tau alteration depends on stroke size as well as on time passed since the event [3].

Neuronal apoptosis caused by ischemia is supposed to lead into hyperphosphorylation of tau, and p-tau is not increased directly after stroke [16]. Moreover, as a biomarker it is more useful for AD diagnosis but can be altered in the context of a chronic process like VaD [17]. Whether it can be considered as a new therapeutic target that should be regulated after ischemia and reperfusion process, remains to be determined [18].

Some research has been done using serum in patients after stroke and recently also in AD. Blood samples are of special research interests, because they are much easier to obtain and the analyses can be done sequentially. Increased serum tau was detected in 48 % stroke patients. Those patients with detectable serum tau developed more severe neurological deficits [19‐21]. Some other biomarkers were supposed to reflect processes of oxidative stress and inflammation followed by blood–brain barrier dysfunction after cerebral hypoxia. Acute-phase proteins like CRP and NSE were analyzed, too. A larger review concentrated on processes after ischemic brain injury, inflammatory processes contributing to neurodegeneration or by the ischemic event itself [16]. Amyloid balance is supposed to be hampered because of the down-regulation of α-secretase resulting into the non-amyloidogenic pathway of total cleaving of APP, but accumulating soluble neurotoxic amyloid peptides via striking the second pathway using β- and γ-secretases. This explains the finding of AD-like pattern in rodents, starting some days after ischemia—an increase of 200 % of APP in the penumbra on the seventh day post-stroke is described up to 1 year after the event [16]. It also raises the question of the cause of progressive cognitive decline following stroke, inflammatory or degenerative processes.

A striking overlap is found in VaD and AD, such as hypertension, hyperlipidemia, diabetes mellitus and white matter changes subsequently lead into VaD on one hand; on the other hand patients with AD showed more often large vessel abnormalities, like carotid artery stenosis, carotid intimal-medial thinning and bilateral present carotid plaques. Prevalence of cardiovascular risk factors, atrial fibrillation, vessel stiffness and microinfarction increase with age and as result cognitive impairment is not distinctly due to AD or microvascular brain damage alone, rather than a mixture of both [22].

As a drawback of our study the deviation of our control group concerning age has to be mentioned (arithmetic mean 58.9 years). Being around one decade younger than AD or stroke patients with less developed age related changes of the vascular system or degenerative processes, their CSF analysis is expected to show less affected values of the proteins a priori. A control group with more similar age in comparison with the rest would have been more significant; however, since white matter lesions are common in the elderly, such a cohort will be difficult to obtain.