Abstract
We evaluated cerebrospinal fluid amyloid-β 1-40 (Aβ40), amyloid-β 1–42 (Aβ42), total and phosphorylated-tau (t-tau and p-tau) in patients with symptomatic isolated cortical supratentorial superficial siderosis (SS), by prospectively recruiting ten patients with SS in the absence of pre-existing cognitive dysfunction, and comparing biomarkers with lobar hematoma cerebral amyloid angiopathy patients (LH-CAA, n = 13), Alzheimer’s disease patients (AD, n = 42), and controls (n = 16). Compared to controls, SS patients showed statistically significant higher t-tau (p = 0.019) and lower Aβ42 (p = 0.0084). Compared to other groups, SS showed statistically significant lower t-tau, p-tau, and Aβ40 compared to AD (p = 0.0063, p = 0.0004, and p = 0022, respectively), and higher p-tau compared to LH-CAA (p = 0.012).
INTRODUCTION
Cerebrospinal fluid (CSF) Alzheimer’s disease (AD) biomarkers, amyloid-β 1–42 (Aβ42), total and phosphorylated-tau (t-tau and p-tau) are considered as in vivo markers of brain amyloid load and neurodegenerative processes [1]. Sharing a common pathophysiological process, cerebral amyloid angiopathy (CAA) was also determined by a CSF specific profile [2–6]. Superficial siderosis (SS) tends to occur in individuals with a relatively low number of microbleeds.
None of the earlier CSF biomarker studies focused on patients with isolated (i.e., without associated intracranial hemorrhage) supratentorial cortical SS (further simply called SS). So far, reports on amyloid imaging by 11C-Pittsburgh compound B (PiB) PET scan in demented and non-demented patients with SS were in favor of underlying CAA in these SS patients [7, 8].
Our aim was to characterize the CSF AD biomarkers profile in SS patients, in comparison with that in LH-CAA, AD, and patients without neurological disease primary related to amyloid deposition (further called simply “controls”).
METHODS
Subjects
We prospectively included 10 SS and 13 LH-CAA patients, and retrospectively selected 42 AD patients and 16 non-degenerative controls in our registered CSF certified biobank (#DC-2008-417). One out of the 10 SS and 7 out of the 13 LH-CAA patients in this study were already analyzed and described in an earlier report of our group [3]. All SS patients presented with transient positive or negative neurological (e.g., motor, sensory, or visual) signs. SS was defined as focal when restricted to 3 or fewer sulci or as disseminated when affecting at least 4 sulci. Possible/probable LH-CAA patients presented with symptomatic LH according to the Boston criteria. SS was considered acute/subacute if hyperintense and chronic if hypointense on FLAIR. The number of microbleeds on T2*-weighted imaging in the SS patients was also noted. Supplementary Figure 1 shows the MRI images of the 10 SS patients. Standard CSF data (i.e., red and white blood cell count, protein levels) and MRI characteristics (focal/disseminated SS, acute/chronic SS, and microbleeds) in the SS group are shown in Supplementary Table 1. In the LH-CAA group, mean CSF red cell blood count (RBC) was 125/mm3 (including 10 patients with RBC <25/mm3, and only one patient with RBC >500/mm3 [i.e., 980/mm3]).
In SS/LH-CAA patients, brain imaging excluded other causes of SS/LH. Pre-existing cognitive deficit (reported by the patients and their family) was absent, cognitive testing was performed after SS/LH, and repeated during follow-up. AD patients met the NIA core clinic and imaging diagnosis criteria blinded to CSF results. In controls, final diagnoses were depression (n = 7), normal pressure hydrocephalus (n = 4), peripheral nerve disorder (n = 2), meningeoma (n = 1), cervical myelopathy due to arthropathy (n = 1), and ischemic brain infarction (n = 1). None of the controls, except patients with normal pressure hydrocephalus, had cognitive dysfunction. All AD and control patients underwent MRI (1.5 Tesla or more, including DWI/ADC/FLAIR/TOF/T2*-weighted imaging; SWI sequences and gadolinium-injection were optional) showing absence of SS/LH. The presence of microbleeds only was not an exclusion criteria for AD and controls.
Patients with anticoagulation treatment, recent trauma, imaging suggesting arteriovenous malformation, cavernoma, vasculitis, or brain tumor were excluded.
We only used CSF data of AD patients and controls older than 60 years since the analyzed CSF markers may vary with age. In AD patients at time of CSF analysis, mean, standard deviation, median, and range values were respectively 19/30, 5.5, 21/30, and 5–27/30 for the Mini-Mental State Examination (MMSE) examination and 34.5, 22.5, 34.5, and 2–84 for disease duration (months).
All subjects (or legal representatives in case of severe cognitive deficit) provided written informed consent to participate. The study was approved by the local ethics committee.
Data on lumbar puncture/CSF analyses are given in the Supplementary Material [3].
Statistical analysis
Graphic results were presented as medians and interquartile ranges. We performed statistical pairwise comparisons with the non-parametric Kruskal–Wallis test using the Conover post-hoc method [9]. The H-score corrected for ties is given in the text after the p-values. Results were compared between the different patients groups, and also between LH-CAA patient with one and with more than one LH, and between SS patients with acute/subacute versus chronic SS.
RESULTS
Median age for SS, LH-CAA, AD, and control patients were 78.5, 73, 73, and 70, respectively. The 10 SS patients (5 men, 5 women) were significantly older than controls (p = 0.0098) while age did not differ between other groups.
During hospitalization for transient neurological deficit, 9 SS patients had MMSE above 27/30 and one presented a MMSE of 23/30 in the absence of subjective cognitive impairment noticed by the patient and his family. Mean clinical follow-up in SS patients was 10 months (range 2–17), in the absence of appearance of cognitive deficit during follow-up even for the patient with initial MMSE at 23.
In SS and LH-CAA patients, mean time between diagnostic MRI and CSF analysis was 24 (range 2 to 66) and 44 days (range 10 to 265) respectively (without statistically significant difference). Of the 13 LH-CAA patients, LH was unique in 8 patients while 5 patients showed, in addition to the acute LH, one or more chronic LH onMRI.
None of the SS or LH-CAA patients had histological analysis except one LH-CAA patient with histologically confirmed CAA despite his young age (50 years).
CSF results are provided in the Fig. 1. Median levels and ranges are shown in Table 1. For Aβ40 analysis, data for SS patients were based on 9 patients only since in the remaining patient the Aβ40 value was not available.
A higher CSF t-tau (p = 0.019) and lower Aβ42 (p = 0.0084) was observed in SS versus controls. When comparing SS to AD, lower t-tau (p = 0.0063), p-tau (p = 0.0004), and Aβ40 (p = 0.0022) were observed. Interestingly, p-tau was higher in SS than in LH-CAA (p = 0.012).
AD patients showed statistically significant higher t-tau and p-tau compared to each other group, statistically significant lower Aβ42 compared to controls, and statistically significant higher Aβ40 than SS and LH-CAA.
No statistically significant differences were found in biomarker levels neither between patients with one LH and patients with more than one LH nor between patients with and patients without FLAIR hyperintense SS.
DISCUSSION
To the best of our knowledge, this is the first study analyzing AD biomarkers in a large clinically-based cohort of patients with isolated SS compared to LH-CAA, AD, and non-degenerative controls. We found similar results as observed earlier with CAA patients (including essentially LH-CAA patients) [2, 3]. The more relevant CSF biomarkers to discriminate SS from controls were t-tau and Aβ42, with SS patients showing intermediate levels between controls and AD, i.e., t-tau higher than controls but lower than AD and Aβ42 lower than controls but higher than AD. T-tau, p-tau, and Aβ40 levels differentiated SS best from AD (i.e., all showing lower levels in SS).
P-tau is the most specific biomarker of AD process. It was, therefore, expected that its concentration was lower in SS than in AD. Regarding Aβ40, it is possible that amyloid precursor protein processing is modified in SS or that amyloid peptides are somehow reduced due to tissue accumulation which would be not related to aggregation as in AD (and therefore not affecting Aβ42 primarily).
A drawback of our study was that the so-called “controls group” consisted of patients with heterogeneous diseases (and thus not healthy controls). Although control patients did not have neurological disorders primary related to amyloid deposition, some of these patients (especially patients with normal pressure hydrocephalus) had disorders in which associated altered AD biomarker profiles have been reported [10–12]. There was no clear correlation between the biomarker profile and the different disorders within the control group.
Although the risk of modifying CSF biomarkers following directly blood contamination is probably remote (since blood concentration of amyloid and tau are hundreds of time lower than in the CSF), the kit provider recommended CSF red blood cell count <500/mm3. Only 2 of our SS patients had red blood cell count >500/mm3 and most of the differences in CSF biomarker values between the groups were significant despite the relative small patient numbers. In our opinion, therefore, blood contamination probably did not play a major role in the biomarker values observed in our study patients.
CSF biomarkers (especially Aβ42) may vary with age (or is associated with prodromal AD more frequently present in older age). Age-related variation of biomarker concentrations in healthy patients, however, is still a matter of debate. Our SS patients had no prior cognitive testing and were older than controls, leading to a possible underestimation of pre-existing prodromal stage of AD and CSF AD data in our SS patients. However, we believe that this could not explain differences observed in biomarker levels since SS patients had no signs of prodromal AD (which is the pathological condition mostly responsible for age related differences in CSF biomarkers) and since most of the differences in CSF biomarker values between the groups were significant despite the relative small patient numbers.
That fact that SS CSF biomarker profile was strongly remaining of CAA suggests that this condition seems to be a CAA-specific radiological feature as suggested by earlier reports [13].
DISCLOSURE STATEMENT
Authors’ disclosures available online (http://j-alz.com/manuscript-disclosures/16-0400r2).