TY - JOUR
T1 - Neuropathologic correlates of cognitive impairment in Alzheimer’s disease with discordant CSF biomarker profiles
T2 - co-pathologies in focus
AU - Alzheimer’s Disease Neuroimaging Initiative
AU - Ioannou, Konstantinos
AU - Perrin, Richard J.
AU - Abdullaieva, Khadidzha
AU - Bluma, Marina
AU - Leuzy, Antoine
AU - Poulakis, Konstantinos
AU - Religa, Dorota
AU - Rodriguez-Vieitez, Elena
AU - Chiotis, Konstantinos
N1 - Publisher Copyright:
© The Author(s) 2025.
PY - 2025/12
Y1 - 2025/12
N2 - CSF Aβ reflects Alzheimer’s disease neuropathologic change (ADNC), while CSF p-tau offers an indirect indication of tangle pathology. However, interpretation can be challenging when cognitive impairment is present alongside Aβ positivity (Α +) but p-tau negativity (T −). We examined neuropathologic differences between CSF A + T − and A + T + profiles, defined by CSF Aβ42 and p-tau181 levels, hypothesizing that cognitively impaired older adults with a CSF A + T − profile would exhibit a greater co-pathology burden, suggesting alternative contributing disease processes. We identified 77 ADNI participants with available CSF biomarkers and neuropathologic assessments (median age = 79.8 years; IQR = 74.7–84.5). Depending on the presence–alone or in combination–of ADNC intermediate/high and non-ADNC pathologies (e.g., Lewy bodies (LB), argyrophilic grain disease (AGD), limbic-predominant age-related TDP-43 encephalopathy-neuropathologic change (LATE-NC)), individuals were classified as ADNC dominant, mixed ADNC, or non-ADNC dominant. The two CSF A + profiles were similar in demographics, frequency of cognitive impairment, longitudinal cognitive performance, clinical comorbidities, CSF Aβ42 levels, CSF α-synuclein positivity rates, and Aβ PET burden. ADNC intermediate/high was significantly more frequent in the CSF A + T + profile than in the CSF A + T − profile (100% vs. 78%, p value = 0.008). The most common co-pathologies contributing to cognitive impairment in the CSF A + T − profile were LATE-NC (stages 2–3) (47%), LB limbic/neocortical (44%), and AGD (stages II–III) (33%), while in the CSF A + T + profile, LB limbic/neocortical (28%) and LATE-NC (stages 2–3) (22%) predominated. The CSF A + T − profile showed 17% ADNC dominant, 61% mixed ADNC, and 22% non-ADNC dominant pathology, whereas the CSF A + T + profile showed 51% ADNC dominant and 49% mixed ADNC pathology (p = 0.001). Within the mixed ADNC subgroup, individuals with a CSF A + T − profile more often exhibited two or more non-ADNC co-pathologies compared to those with a CSF A + T + profile (73% vs. 21%, p = 0.009). Despite clinical similarities among cognitively impaired individuals with CSF A + T − and A + T + profiles, the CSF A + T − profile may reflect a greater burden of non-ADNC pathology. Extending biomarker profiling beyond Aβ and tau may facilitate more personalized care.
AB - CSF Aβ reflects Alzheimer’s disease neuropathologic change (ADNC), while CSF p-tau offers an indirect indication of tangle pathology. However, interpretation can be challenging when cognitive impairment is present alongside Aβ positivity (Α +) but p-tau negativity (T −). We examined neuropathologic differences between CSF A + T − and A + T + profiles, defined by CSF Aβ42 and p-tau181 levels, hypothesizing that cognitively impaired older adults with a CSF A + T − profile would exhibit a greater co-pathology burden, suggesting alternative contributing disease processes. We identified 77 ADNI participants with available CSF biomarkers and neuropathologic assessments (median age = 79.8 years; IQR = 74.7–84.5). Depending on the presence–alone or in combination–of ADNC intermediate/high and non-ADNC pathologies (e.g., Lewy bodies (LB), argyrophilic grain disease (AGD), limbic-predominant age-related TDP-43 encephalopathy-neuropathologic change (LATE-NC)), individuals were classified as ADNC dominant, mixed ADNC, or non-ADNC dominant. The two CSF A + profiles were similar in demographics, frequency of cognitive impairment, longitudinal cognitive performance, clinical comorbidities, CSF Aβ42 levels, CSF α-synuclein positivity rates, and Aβ PET burden. ADNC intermediate/high was significantly more frequent in the CSF A + T + profile than in the CSF A + T − profile (100% vs. 78%, p value = 0.008). The most common co-pathologies contributing to cognitive impairment in the CSF A + T − profile were LATE-NC (stages 2–3) (47%), LB limbic/neocortical (44%), and AGD (stages II–III) (33%), while in the CSF A + T + profile, LB limbic/neocortical (28%) and LATE-NC (stages 2–3) (22%) predominated. The CSF A + T − profile showed 17% ADNC dominant, 61% mixed ADNC, and 22% non-ADNC dominant pathology, whereas the CSF A + T + profile showed 51% ADNC dominant and 49% mixed ADNC pathology (p = 0.001). Within the mixed ADNC subgroup, individuals with a CSF A + T − profile more often exhibited two or more non-ADNC co-pathologies compared to those with a CSF A + T + profile (73% vs. 21%, p = 0.009). Despite clinical similarities among cognitively impaired individuals with CSF A + T − and A + T + profiles, the CSF A + T − profile may reflect a greater burden of non-ADNC pathology. Extending biomarker profiling beyond Aβ and tau may facilitate more personalized care.
KW - ADNC
KW - AT biomarker system
KW - Alzheimer
KW - CSF
KW - Co-pathologies
KW - p-tau
KW - β-amyloid
UR - https://www.scopus.com/pages/publications/105023213719
U2 - 10.1007/s00401-025-02960-w
DO - 10.1007/s00401-025-02960-w
M3 - Article
C2 - 41298940
AN - SCOPUS:105023213719
SN - 0001-6322
VL - 150
JO - Acta Neuropathologica
JF - Acta Neuropathologica
IS - 1
M1 - 57
ER -