Outcome associations of CSF total tau in suspected non-Alzheimer pathophysiology.

INTRODUCTION: Cognitively impaired patients with negative biomarkers of amyloidosis but with neurochemical evidence of Tau pathophysiology or neurodegeneration or both are classified as "suspected non-Alzheimer pathophysiology" (SNAP), according to the 2018 research framework for a biological definition of Alzheimer's disease (AD). The SNAP concept remains incompletely characterized, and associations of amyloid and tau biomarkers with survival outcomes in the SNAP context remain unclear. METHODS: We conducted a retrospective study in patients with a SNAP biomarker constellation, recruited from two tertiary centers between 2019 and 2024. We extracted clinical demographic variables, biomarker results, and survival times to all-cause-mortality or last seen from electronic health records. SNAP patients were defined based on normal cerebrospinal fluid (CSF) levels of amyloid-β1-42 (Aβ42) and a normal Aβ42/Aβ40 ratio, and abnormal levels of phospho-Tau-181 (pTau) and/or total Tau (tTau), as measured with the Lumipulse G chemiluminescence assay. We computed the Youden index to dichotomize SNAP subgroups by CSF biomarkers and demographic variables in a data-driven approach and assessed survival using Kaplan-Meier curves and multivariable Cox hazard models. Patients were categorized as having either idiopathic or secondary neurodegenerative diseases, while those with prion disease were excluded from the final analyses. RESULTS: Compared across amyloid-tau-neurodegeneration-defined categories, patients with SNAP had a shorter median survival than those with AD (31 vs. 42 months), while the Alzheimer's pathological change with concurrent neurodegeneration (APC + N) category exhibited the shortest median survival at 16 months. Patients with SNAP (n = 99) had a median age of 72 (inter-quartile range, IQR: 64; 77) years and a median follow-up time of 12 (IQR 1;27, max. 60) months. Among selected variables, tTau performed best in predicting survival during the follow-up period (area under the curve, AUC = 0.83) in non-prion SNAP, followed by the pTau/tTau ratio (AUC = 0.82). The optimal prognostic tTau cutpoint of ≥ 600 pg/mL was higher than the age-adjusted diagnostic reference (300-500 pg/mL). Patients with SNAP exhibiting higher tTau levels or a reduced pTau-181/tTau ratio had significantly shorter median survival times. Multivariable Cox hazard modeling confirmed tTau to be independently associated with survival with a hazard ratio of 6.1-6.8 (p < 0.001). CONCLUSION: We have identified CSF tTau as a novel predictor of survival in patients with SNAP. Although this descriptive diagnosis encompasses primarily various idiopathic neurodegenerative causes, we also demonstrate tTau's prognostic value in the context of secondary neurodegenerative processes. This finding can enhance prognostication in clinical settings.