Diagnostic performance and clinical applicability of plasma pTau181 to predict dementia risk in a large real‐world cohort of a memory clinic

Abstract

AbstractBackgroundThe Alzheimer Association has recently remarked the diagnostic and prognostic revolution that plasma biomarkers will mean for AD diagnosis, and highlighting the need of further research to translate them as a realistic option in the clinical practice. However, many studies to date have assessed them in research cohorts. Here we evaluated the diagnostic performance and clinical applicability of plasma pTau181 as a predictive marker of dementia risk in a real‐world memory clinic cohort.MethodThe ACE cohort encompass 1649 patients with subjective cognitive decline (SCD), mild cognitive impairment (MCI), AD dementia and other dementias, with a complete battery of neurological tests, biochemistry data, and paired CSF and plasma samples of the Ace Alzheimer Center Barcelona memory clinic. CSF AB42, AB40, pTau181 and tTau, and plasma pTau181 were measured using the Lumipulse G1200 automatic platform (Fujirebio Inc.). Complete cohort were divided in the Testing cohort (n = 1000) and Validation cohort (n = 649).ResultPlasma pTau181 significantly correlated (P<0.0001) with CSF pTau181 in MCI, AD dementia and other dementias, but not in SCD. Sex appeared not significant in plasma pTau181 levels, whereas age significantly influenced in the biomarker’s performance. ROC curve of A+T+ patients vs rest showed an AUC = 0.8250 and P<0.0001. Specifically for MCI patients, ROC curve of MCI A+T+ vs MCI A‐T‐ showed and AUC = 0.8874 and P<0.0001. A cut off value of 1.300 pg/ml of plasma pTau181 exhibited a 93.6% sensibility, 72.4% specificity, 77.8% precision and only 8.3% of false negatives in MCI patients. Hazard ratio of survival of conversion analysis showed that MCI patients above this cut off value exhibited a 4.570 higher risk to convert to dementia than MCI under the cut off value. Validation cohort confirmed these findings.ConclusionPlasma pTau181 could be useful as a pre‐screening biomarker of MCI patients with high risk to progress to dementia. However, positive patients will also require a CSF/PET confirmatory test. Plasma pTau181 by itself does not present a suitable performance in SCD patients. Another biomarker or a combination of biomarkers would be needed to this population. Further studies are need to obtain more clinical data and improve the biomarker performance to finally substitute CSF/PET biomarkers.